Osteoporos Int (2012) 23 (Suppl 5):S521–S611 DOI 10.1007/s00198-012-2005-y

OSTEOPOROSIS AND BONE CONFERENCE 2012

Abstracts of the Osteoporosis and Bone Conference 2012 Oral Presentations

Invited Plenary Lecture Abstracts IS1 FRACTURE HEALING D Marsh; University College London, United Kingdom Abstract unavailable at the time of going to print. IS2 GENETIC DETERMINANTS OF PAGET’S DISEASE OF BONE S. Ralston; Rheumatic Diseases Unit, Molecular Medicine Centre, University of Edinburgh, United Kingdom Paget’s disease is a common skeletal disorder with a strong genetic component, which is characterized by focal increases in disorganized bone remodeling, predominantly affecting the axial skeleton. Current evidence suggests that classical Paget’s disease of bone (PDB) is caused by a combination of rare alleles of large effect size that cause autosomal dominant inheritance of the disease and more common alleles of smaller effect size. Mutations of SQSTM1 are the most common cause of classical PDB, occurring in about 10 % of patients. The causalmutations cluster in the ubiquitinassociated domain and impair its ability to bind ubiquitin. Other loci that predispose to PDB have recently been identified by genome-wide association studies, which have identified variants at seven loci that predispose to the disease. These increase the risk of PDB individually by 1.3- to 1.7-fold, but have combined effects that account for about 86 % of the population-attributable risk of PDB in SQSTM1 negative patients. IS3 OSTEOPOROSIS - A PRIMARY CARE PERSPECTIVE G. Davenport; Keele University, United Kingdom Osteoporosis is a global problem due to an ageing population, which results in increasing numbers of falls and consequently fragility fractures. General practitioners have an estimated 100

osteoporotic patients each to look after but have had no incentive, financial or otherwise, to look after them. Also, GPs’ undergraduate education in musculoskeletal medicine has been minimal. The result has been very little pro-active care in identifying and treating patients at risk of osteoporosis despite a limited directed enhanced service. However, the inclusion of osteoporosis into the QOF this year will at last reward GPs for identifying all patients over 50 years who have sustained a fragility fracture after the 1st April 2012, to investigate them appropriately with DXA if indicated, and to treat the identified at risk patients. Another part of QOF, the Quality and Productivity organizational indicators, will reward GPs for reducing emergency admissions into secondary care. Falls and fractures account for more emergency bed-days than heart attacks, heart failure and stroke put together so the development of an FLS with integral falls assessment clinics should reduce acute admissions. What do GPs require? The education of GPs and practice nurses is a priority and a new web-based resource has been developed by the NOS and RCGP and will be available from March 2012. GPs need clear guidelines on the use of FRAX, Read coding, the cardiovascular risk of calcium and vitamin D, the risk of atypical fractures and osteonecrosis with long-term bisphosphonate therapy, which have all caused considerable confusion in primary care. GPs need full access to resources e.g. DXA facilities, serum vitamin D measurement and the prescribing of new drugs such as denosumab. What do GPs need to do? We need to acknowledge that these at risk patients have general frailty and often considerable co-morbidity and to treat them holistically. We need to improve the monitoring of at risk patients, especially addressing the problems of poor compliance and adherence with bisphosphonates. Hopefully, the QOF will provide GPs with the impetus to integrate with secondary care and move towards universal access to a seamless FLS.

S522

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

IS4 EPIGENETICS IN THE PREDICTION, PREVENTION AND TREATMENT OF BONE HEALTH

IS6 CALCIUM AND VITAMIN D: KILL OR CURE? B. Abrahamsen; University of Southern Denmark, Denmark

C. Relton; University of Newcastle, United Kingdom

Calcium absorption declines with age and is strongly dependent on vitamin D sufficiency. Combined supplementation with calcium and vitamin D (CaD) modestly decreases the risk of fractures at key osteoporotic sites but also modestly increases the risk of renal stones. Thus, in an ITT analysis of the WHI CaD trial (Jackson, NEJM 2006) the absolute reduction in fractures (from 11.9 % to 11.6 %) was offset by the increase in renal stones (from 2.1 % to 2.5 %). Most studies of osteoporosis medications have addressed a scenario where the intervention was given together with calcium and vitamin D, a logical precaution as the net flux of calcium to the skeleton is positive during osteoporosis treatment, where it is negative in the untreated elderly. Vitamin D and analogues have the capacity to enhance vascular calcification in vitro and in animal models. Initial concerns that CaD supplementation could lead to adverse cardiovascular outcomes were based on observations of accelerated vascular calcification in patients with renal failure. Due to the size, aims and adjudication procedures in the controlled studies, it is much less clear if CaD supplementation in subjects with normal renal function is associated with cardiovascular harm. The majority of RCTs with CaD supplements were not set up to address cardiovascular endpoints and outcomes were in some cases selfreported or based on death certificates. Meta-analyses have suggested an excess risk of myocardial infarction with calcium supplements given with or without vitamin D (Bolland BMJ 2010, 2011). The strongest outcome information comes from the WHI, but an excess risk could only be inferred by excluding subjects who also used own supplements, a process that meant that balanced randomisation may no longer be present. Moreover, there was a survival advantage when CaD was added to personal supplements. It is difficult to conclude whether CaD supplements are associated with cardiovascular harm or not in subjects with normal or moderately reduced kidney function. Until this issue has been clarified, clinicians should probably replace the one-sizefits-all policy for CaD with an individualized approach where the dietary intake of calcium is taken into account.

Epigenetics has recently taken the world of medical research by storm, offering the promise of prediction, prevention and treatment of a wide spectrum of common complex diseases. Epigenetic mechanisms can be considered as mediators of environmental and lifestyle influences on disease risk but major challenges are posed in defining the importance of epigenetic mechanisms. Epidemiological approaches can be applied to improve our understanding of the role of epigenetics in development and disease. Population-based studies provide a framework for epigenetic studies with well-established methods for circumventing fundamental issues such as confounding and reverse causation as well as more recent developments of approaches to strengthen causal inference. The current status of role of epigenetic variation in bone health will be reviewed and the application of epigenetic epidemiology in this emerging field this will be presented. IS5 SARCOPENIA: CAUSES AND CONSEQUENCES Avan Aihie Sayer; Academic Geriatric Medicine, MRC Lifecourse Epidemiology Unit, University of Southampton, UK Sarcopenia is the loss of skeletal muscle mass and function with age. There is growing consensus about how it should be defined and it has become clear that it is common in both older women and men. It has serious health consequences in terms of disability, morbidity and mortality as well as significant healthcare costs. Adult influences on sarcopenia include levels of physical activity, nutritional intake, setting of the immune-endocrine axis and genetic variation. However there remains considerable unexplained variation in muscle mass and strength between older individuals of the same age which might be partly explained by the observation that muscle mass and strength in later life reflect not only the rate of loss but also the peak attained in earlier life. To date most observational and interventional epidemiological studies of sarcopenia have focused on factors modifying decline in later life but a lifecourse approach to understanding sarcopenia additionally focuses attention on the determinants of peak muscle mass and strength attained in early adulthood such as early growth and development and their long term consequences. This would be consistent with current understanding of the developmental origins of other age-related disorders such as osteoporosis, type 2 diabetes and cardiovascular disease. A lifecourse approach to sarcopenia is important because it suggests that the opportunity for instituting beneficial interventions may not just lie in later life. Birth cohorts such as the Hertfordshire Cohort Study are an invaluable resource for research in this area.

IS7 BONE: A TARGET FOR SYSTEMIC DISEASES J. Compston; Cambridge University Hospitals NHS Foundation Trust, Cambridge United Kingdom A growing list of medical conditions has been associated with adverse effects on bone health and increased risk of fracture. Diseases affecting the heart, lungs, nervous system, kidneys, liver and gastrointestinal tract have all been implicated, with pathogenetic mechanisms that are overlapping but not identical. These include immune- and inflammatory-mediated factors, changes in nutritional status, hypogonadism, reduced mobility, increased risk

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

of falling, effects on the vitamin D/parathyroid hormone axis, altered glucose/insulin metabolism and bone-active medications. Although it is often assumed that the increase in fracture risk associated with secondary osteoporosis is mediated by reduced bone mineral density (BMD), there is increasing evidence that many diseases have effects on bone fragility that are at least partially independent of BMD. Prediction of fracture risk by algorithms such as FRAX® may therefore be improved by addition of co-morbidities to other clinical risk factors. Until recently, obesity was widely believed to be protective against fractures but recent studies have challenged this perception. Thus in postmenopausal women, overall fracture incidence is similar in non-obese and obese women and the risk of certain fractures, for example ankle and upper leg, is higher in obese women whilst that of wrist and hip fractures is lower. The pathophysiology of fractures associated with obesity is incompletely understood but increased risk and different patterns of falling, higher rates of co-morbidities, and production by visceral fat of pro-resorptive cytokines and adipokines may all contribute. Increased visceral fat is also associated with insulin resistance and reduced production of insulin-like growth factor-1. Although BMD is often normal in obese women with fracture, it may nevertheless be inappropriately low for body weight. The association of bone fragility and diseases affecting multiple organ systems has important implications. With increasing longevity of the population, the contribution of co-morbidities that increase fracture risk will rise, emphasising the need for education across specialities. Increasing rates of obesity highlight the importance of developing fracture prevention strategies in the obese population. Finally, interactions between systemic diseases and bone have the potential to provide new insights into pathogenetic mechanisms of bone disease.

Oral Communication Abstracts O1 HABITUAL LEVELS OF HIGH, BUT NOT MODERATE OR LOW, IMPACT ACTIVITY ARE POSITIVELY RELATED TO HIP BONE MINERAL DENSITY AND STRENGTH: RESULTS FROM A POPULATIONBASED STUDY OF ADOLESCENTS Kevin Deere (1) presenting Adrian Sayers (1) Joern Rittweger (2) Jon Tobias (1) University of Bristol, Bristol, UK (1) Institute of Aerospace Medicine, Cologne, Germany (2) Introduction: Whether a certain level of impact needs to be exceeded for physical activity (PA) to benefit bone accrual is currently unclear. Materials and methods: To examine this question we performed a cross-sectional analysis between PA and hip BMD and structure (as measured by DXA) in 724 adolescents (292 boys, mean 17.7 yrs) from the Avon Longitudinal Study of

S523

Parents and Children, partitioning outputs from a Newtest accelerometer (Newtest Oy, Oulu, Finland) into six separate impact bands. Results: Counts within 2.1-3.1 g, 3.1-4.2 g, 4.2-5.1 g and >5.1 g were positively related to femoral neck (FN) BMD in our minimally adjusted model including age, height and sex (2.13.1 g, β00.042, p00.08; 3.1-4.2 g: β00.058, p00.009; 4.25.1 g: β00.070, p00.001; >5.1 g: β00.080, p<0.001) (β0SD change per doubling in activity). Similar positive relationships were observed between impact bands above 2.1 g and BMD at other hip sites (eg ward’s triangle, greater trochanter), indices of hip structure derived from DXA scans (eg FN width, crosssectional area, cortical thickness), and predicted strength (crosssectional moment of inertia). In subsequent adjusted analyses, impacts were re-categorised into low (0.5-2.1 g), moderate (2.14.2 g) or high (>4.2 g) bands, and adjusted for impacts in other bands. Whereas high impacts were positively related to FN BMD, if anything, moderate and low impacts were inversely related to hip BMD (low: β0−0.052, p00.2; medium: β0−0.058, p00.2; high: β00.137, p<0.001).The positive association between high impact PA and FN BMD was still observed after further adjustment for fat mass, lean mass, and socio-economic position (β0 0.096, p00.016). Results from a separate sample of 22 adolescents wearing Newtest accelerometers while performing supervised activities indicated that 4 g, the putative impact threshold which needs to be exceeded to affect hip BMD, is typically experienced when running faster than 10 km/hour, or jumping. Conclusion: We conclude that PA may only improve hip BMD and strength in adolescents when associated with high impacts like running and jumping, whereas lower impacts, such as those accrued during jogging, may be of little benefit. O2 BRIEF HIGH IMPACT EXERCISE IMPROVED FEMORAL NECK BONE MINERAL DENSITY AND HIP STRUCTURAL PARAMETERS IN OLDER MEN: A RANDOMISED UNILATERAL INTERVENTION Sarah J Allison (1), Jonathan P Folland (1), Winston J Rennie (2), Gregory D Summers (3), Katherine BrookeWavell (1); Loughborough University, Leicestershire, UK (1) University Hospitals of Leicester, Leicester, UK (2) Royal Derby Hospital, Derby, UK (3) Introduction:There is mixed evidence as to whether exercise can increase BMD in older men with little investigation of high impact exercise. Individual differences and lifestyle changes may confound exercise trials but can be minimised using a within-subject unilateral design (exercise vs. control leg). This study examined the influence of a 12-month high impact unilateral exercise intervention on femoral neck BMD in older men. Materials and Methods: Fifty, community-dwelling older men with no history of musculoskeletal disease or injury were

S524

recruited. They were prescribed a 12-month high impact unilateral exercise intervention which progressed to 50 multidirectional hops, performed daily on one randomly allocated leg. Dual energy-X-ray absorptiometry scans of both femurs were performed at baseline and after 12-months by an observer blinded to the leg allocation. Repeated measures ANOVA with post hoc tests was used to detect significant differences over time (pre vs. post); between legs (exercise leg [EL] vs. control leg [CL]) and detect any leg x time interactions. Results: Thirty-five men (mean±SD age 69.9±4.0 yrs) exercised for 12-months, attending 90.0±9.1 % of prescribed sessions. Fourteen men withdrew due to: health problems or injuries unrelated to the intervention (n09), time commitments (n02), or discomfort during exercise (n03), whilst BMD data were missing for one man. Femoral neck BMD increased in the EL (+0.7 %) compared to the CL (−0.9 %); interaction effect P00.003. Femoral neck BMC and cross-sectional area also increased significantly in the EL compared to the CL (P00.022 and P00.012 respectively). Although the interaction term was not significant (P>0.05), there were significant main effects of time for section modulus (P00.044) and minimum neck width (P00.006). Section modulus increased significantly in the EL (P00.016) but not the CL (P00.465); mean change +2.3 % and +0.7 % respectively, whereas minimum neck width increased significantly in the CL (P00.004) but not in the EL (P00.166); mean changes being +0.7 % and +0.3 % respectively. Conclusion: Brief, high impact exercise improved femoral neck BMD and hip structural parameters in older men. O3 THE RISK OF FRACTURE IN INCIDENT MULTIPLE SCLEROSIS PATIENTS: THE DANISH NATIONAL HEALTH REGISTERS Marloes Bazelier (1) presenting Joan Bentzen (2) Peter Vestergaard (3) Egon Stenager (2) Frank de Vries (1,4) Utrecht University, Utrecht, The Netherlands (1) University of Southern Denmark, Copenhagen, Denmark (2) Aarhus University Hospital, Aarhus, Denmark (3) University of Southampton, Southampton, UK (4) Background: Patients with Multiple Sclerosis (MS) may be at increased risk of fractures owing to osteoporosis and falling. Objective: To evaluate the risk of fracture in incident MS patients from a dedicated registry, and population-based controls. Methods: We conducted a population-based cohort study (1996–2007) utilising the Danish National Health Registers that were linked to the Danish MS Registry, and the Danish MS Treatment Registry. Incident MS patients (2,963 cases) were 1:6 matched by year of birth, gender, calendar time and region to persons without MS (controls). Cox proportional hazards models and logistic regression were used to estimate risk of fracture in MS. Time-dependent adjustments were made for age, history of diseases and drug use.

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

Results: Compared with controls, patients with MS had no overall increased risk of fracture: adjusted Hazard Ratio (adj. HR) 1.0 [95 % CI 0.9-1.2]. However, the risk of femur / hip fracture (adj. HR 1.9 [1.1-3.4]) was significantly increased versus controls. As compared with unexposed patients, MS patients who had been exposed to a short-course of methylprednisolone in the prior year had no significantly increased risk of osteoporotic fracture: adj. HR 1.2 [0.5-2.9]. Disabled MS patients (with expanded disability status scale [EDSS] from 6 to 10) had a 2.6-fold increased risk of osteoporotic fracture (adj. Odds Ratio 2.6 [1.0-6.6]) versus patients with low EDSS (0–3). Conclusion: Patients with MS had a higher risk of femur / hip fracture than controls. Disability status is probably more important than glucocorticoid use in the aetiology of MS and osteoporotic fracture. O4 CAN FUNCTIONAL MUSCLE TESTING IMPROVE FRACTURE RISK ASSESSMENT IN AN AGEING FEMALE POPULATION? Nicola Crabtree (1) presenting Natalie Bebbington (1) Katie Stant (2) Helen Duffy (2) Jim Parle (2) Neil Gittoes (1) Queen Elizabeth Hospital, Birmingham, UK (1) Primary Care Clinical Sciences, University of Birmingham, Birmingham, UK (2) Introduction: Osteoporotic fracture risk increases with age and decreasing bone density. Additionally, ageing causes muscle mass to decline and the risk of falling to increase. However, falls risk is currently not included in the frequently used fracture risk calculator(FRAX™). The aim of this study was to investigate the relationship between bone strength, muscle function and previous fracture in a sub group of the SCOOP study cohort. Material and Methods: The SCOOP study is a RCT which aims to demonstrate whether community-based screening for osteoporosis reduces the incidence of fractures and is cost effective in older women. It is a 7-year prospective study which has recruited 12,495 women between 70 and 85 years. As part of the study DXA scanning was offered to women considered to be at high risk of fracture estimated from clinical risk factors. At our centre, 1692 subjects were recruited of whom 390 attended for DXA scanning. Of these, 290 had muscle function assessed using the chair rising test(CRT) on the Leonardo™ ground force reaction plate. Hip strength index(HSI) was estimated using the advanced hip analysis tools on the GE Lunar iDXA™. Discussion: A significant decrease with age was observed for femoral neck BMD, yet HSI remained constant (1.42[0.37]). A significant decrease was noted in muscle power (p<0.0001) and this correlated with a small but significant increase in chair rise time (p00.015). Women with HSI≤1were significantly slower (3.0 vs. 2.4 s per chair rise, p00.005) and had less muscle power (4.4 vs.5.8 Watts/kg, p<0.0001) than those with HSI >1. Previous fracture incidence was greatest

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

(73.9 %) in those women where FSI≤1 and CRT>2.5secs and lowest (34.6 %) in those where FSI>1 and CRT≤2.5secs, suggesting that those with the lowest bone strength and highest risk of falling had the greatest risk of fracture. Conclusion: Identifying those persons who have both reduced bone strength and reduced muscle strength may better predict future fracture in an ageing female cohort and timed chair rises may prove to be a simple but useful addition to FRAX™ when considering future fracture risk. O5 ACCRUAL OF LEAN MASS IN EARLY CHILDHOOD IS ASSOCIATED WITH HIP STRENGTH AND GEOMETRY AT SIX YEARS OLD: THE SOUTHAMPTON WOMEN'S SURVEY Elizabeth Curtis (1) presenting Zoe Cole (1) Sarah Crozier (1) Georgia Ntani (1) Sian Robinson (1) Keith Godfrey (1,2) Avan Sayer (1) Hazel Inskip (1) Cyrus Cooper (1) Nicholas Harvey (1) MRC Lifecourse Epidemiology Unit, University of Southampton, Southampton, UK (1) Southampton Nutrition Biomedical Research Unit, Southampton, UK (2) Background: We evaluated longitudinal and cross-sectional relationships between body composition and bone size, density and indices of hip strength using DXA in a population cohort of children (Southampton Women’s Survey), characterised at birth, four and six years. Materials and methods: At 6 years, bone size and density at the total hip site, and body composition, were assessed by DXA (Hologic Discovery, Bradford, USA). Femoral neck geometry and strength were estimated using hip structural analysis. Height, weight, diet (questionnaire) and physical activity (accelerometer) were assessed. Discussion: 499 children had complete body composition and hip structure data at six years, with DXA body composition available on a subset at birth (n0202) and 4 years (n0377). After adjustment for child’s age, height and sex, statistically significant positive associations were observed between total body lean at age six years and total hip bone area (β00.96 cm2/sd, p<0.001), bone mineral content (β01.30 g/sd, p<0.001), areal bone mineral density (β00.04 g/cm2/sd, p<0.001), bone mineral content adjusted for body size (β00.14 g/sd, p00.002), and femoral neck cross sectional area (β00.15 cm2/sd, p<0.001), cross-sectional moment of inertia [CSMI] (β00.11 cm4/sd, p<0.001) and cortical thickness (β00.009 cm/sd, p<0.001). These associations persisted after adjustment for total fat mass, milk intake physical activity and maternal social class. Changes in fat and lean mass between birth and 4 years were statistically significantly positively associated with indices of hip structure and strength at six years (CSMI increased by 0.59 cm4 for each 1 sd increase lean mass and by 0.60 cm4 for each 1 sd increase fat mass (both p<0.001)). Conclusions: We have demonstrated that childhood total lean mass at six years is positively related to proximal

S525

femoral size, shape, density and strength, independently of fat mass. Furthermore change in body composition through early childhood was also associated with hip geometry at age six. These findings suggest that hip geometry, an independent risk factor for osteoporotic fracture in later life, might be partly determined by body composition in childhood, and thus potentially amenable to modification through lifestyle intervention. O6 THE EFFECT OF OBESITY ON BONE MINERAL DENSITY MEASURED AT FOUR FRACTURE SITES Amy Evans (1) presenting Jennifer Walsh (2) Richard Eastell (1) University of Sheffield, Sheffield, UK (1) Sheffield Teaching Hospitals Foundation NHS Trust, Sheffield, UK (2) Introduction: Obesity is associated with reduced risk of hip and spine fracture, increased risk of ankle fracture and no change in wrist fracture. We aimed to 1) determine BMD at these four sites, 2) identify the microarchitectural characteristics of any differences using high resolution peripheral quantitative computed tomography (HR-pQCT), 3) compare the effect of obesity on BMD in men and women. Methods: We studied individually matched pairs of lean (BMI 18.5-24.9) and obese (BMI>30) individuals, aged 55–75 years (13 male and 17 female pairs). DXA was used to determine hip and lumbar spine areal BMD. HR-pQCT was used to determine bone density and microarchitectural parameters at the radius and tibia. Discussion: Obesity was associated with higher BMD at each sites in women (all p<0.01) and men (all p<0.05 except tibia; p00.1). At the radius and tibia, higher BMD in obesity is due to greater cortical thickness and density and greater trabecular number and density (women all p<0.05, whilst in men, only trabecular number and spacing were significantly different at the tibia (p<0.01)). Bone size (total area and cortical perimeter) did not differ between lean and obese subjects. ANOVA did not identify any significant interaction of gender with the effect of obesity on BMD. Conclusion: BMD and microarchitecture do not explain the increased risk of ankle fracture in obesity. Contributors to the increased risk might include increased force through the ankle or changes in bone quality. Percentage mean difference in BMD in obese versus lean individuals Site Lumbar spine Total hip Distal radius Distal tibia *p<0.05, **p<0.001

Males 12.1* 11.6* 23.5** 10.32

Females 26.8** 27.7** 25.5** 25.3**

S526

O7 WHAT ARE THE ASSOCIATIONS BETWEEN WAIST HIP RATIO, BODY MASS INDEX AND HIP FRACTURES? COHORT OF NORWAY Anne Johanne Søgaard (1), Tone K Omsland (1&2), Haakon E Meyer (1&3) Division of Epidemiology, Norwegian Institute of Public Health, Oslo, Norway (1) Department of Public Health and Primary Health Care, University of Bergen, Bergen, Norway (2) Institute of General Practice and Community Medicine, University of Oslo, Oslo, Norway (3) NOREPOS Core Research Group, Oslo, Norway (4) Introduction:The purpose of this study was to examine the association between Waist Hip Ratio (WHR) and subsequent hip fractures controlling for Body Mass Index (BMI) and other potential confounders. Material and Methods: In Cohort of Norway – comprising participants in ten population-based health surveys in Norway 1994–2003, 180,546 individuals aged 20 years and above participated (58 % of the invited). Collection of anthropometric data allowed the calculation of BMI and WHR. Hip fractures (cervical, trochanteric or subtrochanteric) treated in Norwegian hospitals from 1 January 1994 to 31 December 2008 were retrieved from computerised patient registers, and death and emigration dates came from the National Population Register. The analyses in this paper are limited to participants 60–79 years with valid baseline data on BMI and WHR and subsequent hip fractures - i.e. 23,368 men and 20,230 women, followed for 8.1 years (median). Cox’s proportional hazards models were used to estimate hazard ratios (HR) of fractures with 95 % confidence interval (CI). BMI (kg/m2) was divided in tertiles with cut-off 25.1 and 27.9 in men and 24.9 and 28.8 in women. Discussion: During follow up, 900 men and 1,531 women suffered a hip fracture. Adjusting for age, the risk of fractures decreased with increasing BMI (p<0.001), whereas no association was found between WHR and hip fractures. However, after additional adjustment for BMI, WHR was positively related to the risk of fracture. In stratified analyses, after adjustment for age and BMI, the HR for hip fracture per 1 SD increase in WHR in men were 1.35 (CI 1.20,1.52), 1.26 (CI 1.09,1.45) and 1.17 (CI 1.01,1.36), respectively in the low, middle and high tertile of BMI. The corresponding HR in women were 1.22 (CI 1.11,1.33), 1.16 (CI 1.05,1.29) and 1.07 (CI 0.96,1.19). Additional adjustments for height, smoking, physical activity and use of hormone substitution (women), did only slightly attenuate the associations. Conclusion: WHR was a risk factor for hip fractures when BMI was adjusted for, but the predictive value of WHR seemed to decrease with increasing BMI.

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

O8 RISK OF FRACTURE IN PATIENTS WITH BARIATRIC SURGERY AND MATCHED CONTROLS: A POPULATION-BASED COHORT STUDY IN THE UNITED KINGDOM Arief Lalmohamed (1) presenting Cyrus Cooper (2) Marloes Bazelier (1) Alun Cooper (2) Corinne Klop (1) Frank de Vries (1) Nick Harvey (2) Utrecht University, Utrecht, The Netherlands (1) Southampton General Hospital, Southampton, UK (2) Introduction: Bariatric surgery can be considered among patients with morbid obesity. The surgery has been associated with reduced bone mineral density, probably as a result from malabsorption, and from secondary hyperparathyroidism. However, risk of fracture has not been determined. Objective of this study was to estimate fracture risk in bariatric surgery patients versus matched controls. Methods: A retrospective cohort study was conducted within the UK GPRD (1987–2010). Patients had a record of bariatric surgery (n02,346), and were matched to up to six controls without bariatric surgery by age, sex, practice, and body mass index. They were followed from the date of bariatric surgery for the occurrence of any fracture. Disease and medication adjusted (adj.) hazard ratios (HRs) were calculated using time dependent Cox regression. Results: No increased risk of fracture was seen in patients who underwent bariatric surgery, compared to matched controls (adj. HR 0.92; 95 % CI 0.73-1.15). Fracture risk remained steady over time since bariatric surgery. We found a significantly increased risk among patients aged≥60 years (adj. HR 1.52; 95 % CI 1.02-2.27), and those with a history of fracture (adj. HR 1.84; 95 % CI 1.29-2.63), and cerebrovascular diseases (adj. HR 6.31; 95 % CI 3.53-11.28). In contrast, a decreased risk of fracture was found in subjects using antidiabetic drugs (adj. HR 0.41; 95 % CI 0.18-0.91). Conclusion: This is probably the first and largest study that estimates risk of fracture after bariatric surgery as compared to controls. Our data did not find an overall association between bariatric surgery and risk of fracture. However, fracture risk assessment may be considered among patients who have undergone bariatric surgery, and who have a history of fracture, cerebrovascular disease or who are 60 years or older. O9 THE MEPE-ASARM AXIS REGULATES CHONDROCYTE MATRIX MINERALISATION Katherine Staines (1) presenting Neil Mackenzie (1) Matt Prideaux (4) Claire Clarkin (2) Lesya Zelenchuk (3) Lynda Bonewald (4) Peter Rowe (3) Vicky MacRae (1) Colin Farquharson (1) The University of Southampton, SoThe Roslin Institute and R(D)SVS, The University of Edinburgh, Edinburgh, UK (1) The University of Southampton, Southampton, UK (2) KUMC, Kansas City, USA (3) University of Missouri-Kansas City, Kansas City, USA (4)

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

Matrix Extracellular Phosphoglycoprotein (MEPE) belongs to the SIBLING protein family which play key roles in biomineralization. Here we document for the first time the function of MEPE and its ASARM peptide in chondrocyte matrix mineralization. In situ hybridization and immunohistochemistry localized MEPE and the ASARM peptide preferentially to the hypertrophic chondrocytes of the growth plate and this was confirmed by RT-qPCR analysis of microdissected epiphysis where there was a 9-fold increase in Mepe mRNA in the hypertrophic chondrocytes (compared to proliferative chondrocytes, P<0.05). Cathepsin B, which cleaves the ASARM peptide from MEPE, was immunolocalized to the chondrosseous junction. Treatment of ATDC5 chondrocytic cells with phosphorylated (p)ASARM peptide caused an inhibition of mineralization (P < 0.01) which was confirmed in MEPEoverexpressing ATDC5 cells (P<0.01). Interestingly, nonphosphorylated (np)ASARM promoted ATDC5 matrix mineralization (P<0.01). The enlargement of the diaphyseal mineralization zone of E17 metatarsals, a more physiological model, was inhibited by 20 μM pASARM (P<0.001). This was further examined in E15 metatarsals, a developmental time point in which the central mineralized core is not formed but develops in culture. Treatment with 20 μM pASARM prevented mineralization as confirmed by μCT where BV/TV was significantly lower in the pASARM treated bone (P<0.001). The linear growth of E15 and E17 metatarsals and the proliferation of chondrocytes within were normal in pASARM treated bones indicating that the decreased mineralization observed was not secondary to cell toxicity. mRNA expression of differentiation (Coll10a1, Coll2a1, Atf3, PthIh, Ihh) and mineralization (Alpl, Ank, Enpp1, Phospho1) markers were unchanged in pASARM treated bones as was ALP activity. The pASARM peptide, however, reduced endothelial cell markers (CD31, CD34, VEGFR2/Flk1) and VEGF120/164 and Mmp13 mRNA expression in E15 metatarsals (P< 0.05). Conditioned media from sclerostin expressing IDGSW3 osteocyte-like cells inhibited diaphyseal mineralization of E15 metatarsals implementing sclerostin as a negative regulator of chondrocyte matrix mineralization, potentially through a MEPE-dependent mechanism. These results indicate MEPE to be an important regulator of growth plate chondrocyte matrix mineralization through its cleavage, and subsequent phosphorylation to the ASARM peptide. O10 INTRA-ARTICULAR AMPA/KAINATE GLUTAMATE RECEPTOR ANTAGONISTS ALLEVIATE INFLAMMATION, PAIN AND PATHOLOGY IN RAT ANTIGEN INDUCED ARTHRITIS Cleo Bonnet (1) presenting Anwen Williams (1) Ann Harvey (1) Sophie Gilbert (1) Abdul Moideen (1) Mari Nowell (1) Deborah Mason (1) Cardiff University, Cardiff, UK (1)

S527

Introduction: Concentrations of the neurotransmitter glutamate are greatly increased in synovial fluids of RA and OA patients, where they correlate with cytokine concentrations. Previously, we demonstrated that human synoviocytes express functional glutamate receptors (GluRs) and that activation of NMDA GluRs decreases proMMP2 expression, whilst activation of kainate GluRs increases IL-6 release. High glutamate concentrations in the joint cause arthritic pain which is alleviated by intra-articular injection of GluR antagonists. However, the effects of such inhibitors on arthritis progression via GluRs on other joint tissues have not been considered. We are investigating the hypothesis that specific GluR subunits in the arthritic synovium mediate proinflammatory and degradative responses, and may be therapeutically targeted to reduce disease progression and pain. Methods: Using the mono-articular antigen induced arthritis (AIA) rat model, we used intra-articular (i.a.) injection of NBQX to inhibit AMPA/kainate (KA) receptors at the time of arthritis induction, prior to peak IL-6 levels. Over a 21 day period, we measured knee swelling and gait patterns of AIA, AIA+NBQX and normal rats. On day 21, joint tissues were taken for QRTPCR, x-ray, magnetic resonance imaging (MRI), histology and GluR/transporter immunohistochemistry to examine GluR expression and joint destruction. A modified Mankin score was used to quantify cartilage destruction in coronally sectioned, toluidine blue and safranin-O stained rat knees. Results: Less knee swelling (P<0.0006) was found in NBQX treated rats compared to AIA rats. Using paint footprints, NBQX treated rats displayed less pain related behaviour during the initial flare of arthritis compared to AIA rats. Ionotropic and metabotropic GluR mRNA was differentially expressed in cartilage, synovium, meniscus, fat pad, patella, femoral head and shaft of rat knees. X-ray and MRI analysis revealed a less severe joint erosion and synovial inflammation phenotype in NBQX treated rats compared to AIA rats. Cartilage degradation in NBQX rat knees appeared less severe compared to AIA knees (P00.07). The Glu transporters GLAST and EAAC1 were localised to joint tissues in all rats, however, greater levels of staining were observed in AIA rats in areas of bone remodeling which was less apparent in NBQX treated rats. Discussion: We have shown that i.a. NBQX treatment alleviates inflammation, pain and pathology in arthritis in vivo and that GluRs are differentially expressed in rat knees, supporting our hypothesis that KA GluRs may be specifically targeted to ease pain, inflammation and pathology in arthritis. O11 NON-ENZYMATIC OMEGA-3-FATTY ACID BREAKDOWN PRODUC TS A RE DIFFER ENT IA LLY EXPRESSED IN RHEUMATOID ARTHRITIS AND SUPPRESS NFΚB ACTIVITY TO INHIBIT OSTEOCLAST FORMATION Seint Lwin (1) Joshua Brooks (1,2) Lynett Danks (1) Karen Lyndberg (1) Ginger Milne (1,2) Jason Morrow (1,2) James

S528

Edwards (1) presenting University of Oxford, Oxford, UK (1) VUMC, Nashville, USA (2) Excessive osteoclastic resorption is a defining feature of rheumatoid arthritis (RA), resulting in tissue damage and pain. Omega-3 fatty acids (Ω3-FA) protect against RA and reduce bone pain, however the mechanism through which this occurs remains unknown. Cyclopentenone isoprostanes (IsoPs) form in situ as oxidation products of Ω3-FA. We hypothesize that a beneficial effect of Ω3-FA in RA occurs following the formation of IsoPs, to suppress osteoclast formation. To test this, we used GC-MS to analyze the IsoP profile in serum from age-, sex-matched RA patients or controls (n07). RA patient samples were confirmed as CEP1 and CCP positive. In addition we employed osteoclast formation assays using RAW 264.7 cells treated with 15-A3-IsoP (+ RANKL, M-CSF). NFκB signalling was quantified using an NFκB-reporter macrophage cell line and supported by immunofluorescent staining for nuclear/ cytoplasmic p65 and western blot analysis for phosphorylated IκBα. To confirm the role of NFκB in IsoP mediated osteoclast inhibition, site-directed mutagenesis was employed to alter a putative IsoP binding region of IKK (C179). Mutated RAW 264.7 cells were assessed for osteoclast formation capacity and NFκB activation status, compared to unmodified cells. Our studies show that the IsoP levels within RA serum are significantly altered, compared to normal control patients. In addition, osteoclast precursor cells treated with 15-A3-IsoP (15uM) formed fewer TRAP+multinucleated cells (26.3±5.3) compared to vehicle treated controls (46.3±4.3, p<0.05). Moreover, osteoclast precursor cells treated with 15-A3-IsoP demonstrated up to 60 % decrease in NFkB reporter activity (25uM, p<0.01), phosphorylated IκBα protein levels and nuclear localization of the p65 protein. Mutation of C179 on IKK resulted in a loss of inhibition of NFκB activity and osteoclast formation, following 15-A3-IsoP treatment compared to WT cells (p<0.01). These findings show that the 15-A3-IsoP molecule inhibits osteoclast formation through interactions with C179 on the IKK molecule, to reduce IκBα phosphorylation and suppress NFκB activation. Moreover, we identified a significant change in IsoP levels in RA patients compared to controls, suggesting that reduced IsoP levels may contribute in part, to the increased osteoclastic resorption of RA. O12 DEATH RECEPTOR 3 AND TL1A: A REGULATORY PATHWAY OF BONE TURNOVER AND TARGET FOR OSTEOPOROSIS THERAPY? Fraser Collins (1) presenting Michael Stone (2) Edward Wang (1) Anwen Williams (1) Institute of Infection and Immunity, School of Medicine, Cardiff University, Cardiff, UK (1) Bone Research Unit, Cardiff University Academic Centre, University Hospital Llandough, Penarth, UK (2)

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

Introduction: Patients with inflammatory disease such as rheumatoid arthritis are at an increased risk of developing osteoporosis, as inflammation exerts significant influence on bone turnover. Pro-inflammatory cytokines such as TNFα and IL-8 have been proposed as pathogenetic factors in the development of osteoporosis. Death Receptor 3 (DR3) and its only known ligand TNF-like protein 1A (TL1A) have been shown to play a role in inflammatory disease and our own studies, in a murine model of inflammatory arthritis, have highlighted a key function in regulating osteoclastogenesis, implicating DR3 as a significant regulator in bone (Bull MJ, Williams AS et.al. J.Exp.Med. 2008). We investigated the potential mechanisms by which the DR3/ TL1A pathway orchestrates osteoclastogenesis in an in vitro human model. Materials and Methods: CD14+ monocytes were obtained from healthy, pre-menopausal, females. Cells were cultured on ivory discs in the presence of M-CSF for 7 days and for a further 7 or 14 days in M-CSF and RANKL± TL1A at 10 ng/ml or 100 ng/ml. At termination, %osteoclastogenesis and %disc resorbed were quantified. Supernatants were collected and levels of TNFα, TL1A, IL-8 and MMP-9, a collagenase involved in bone resorption, were measured by ELISA. Results: TL1A was not produced endogenously by CD14 + monocyte cultures. A significant increase in %osteoclastogenesis was observed at days 7 and 14 after addition of exogenous TL1A (p0<0.01). These OC were functionally active with the % disc resorbed at day 14 showing a dose response, significantly rising with increasing TL1A concentrations (p 0<0.05). Addition of TL1A increased levels of total MMP-9, but not IL8 (p0>0.01). TNFα was not detected in any culture conditions. Conclusion: We have shown that TL1A has a direct effect on osteoclastogenesis and increased the level of bone resorption in a dose-dependent manner in an in vitro human system. The increase in resorption correlated with an increase in total MMP-9 and these effects were independent of TNFα and IL-8. Our findings highlight the DR3/TL1A pathway as a regulator of bone turnover and imply that DR3 could contribute to the pathogenesis of osteoporosis. O13 VITAMIN D, BUT NOT PTH, MODULATES VEGF AND IL-6 SECRETION BY OSTEOCYTES IN VITRO Madeleine Adams (1) presenting Meriel Jenney (2) John Gregory (1) Bronwen Evans (1) Institute of Molecular and Experimental Medicine, Cardiff University, Cardiff, UK (1) Paediatric Oncology Department, Children's Hospital for Wales, Cardiff, UK (2)

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

Introduction: The role of VEGF as a coupling factor between angiogenesis and osteogenesis is well documented, although most in vitro studies have used osteoblasts. It is well established that osteocytes modulate osteoblast activity, thus the role of VEGF in osteocyte biology is key to further understanding this crosstalk. Studies have shown that i) vitamin D upregulates VEGF production by smooth muscle cells and ii) parathyroid hormone (PTH) stimulates osteoblastic VEGF expression. We hypothesised that VEGF may influence osteocyte number and differentiation, and that vitamin D and/or PTH may affect osteocyte VEGF secretion. IL-6 secretion was also measured as its involvement in angiogenesis has previously been reported. Materials and Methods: MLO-Y4 osteocytes were incubated (72 hours) with VEGF164 (10-100 ng/ml). Cell numbers were measured (MTS) and osteocyte (DMP1, E11, and CX43) gene expression assessed (qRTPCR). MLO-Y4 osteocytes were also treated with 1α-25-dihydroxy-vitamin D (10-11-10-8 M) or PTH (1–34) (10-9-10-6 M). Media were collected (7 and 24 hours) and VEGF164 and IL-6 secretion measured (ELISA). Results were expressed as pg/ml/μg protein and analysed using SPSS using one way ANOVA. Discussion: Initial studies showed that 100 ng/ml VEGF164 significantly reduced osteocyte cell number (14 % reduction, 95%CI 6-22 %, p00.001), however no effect was shown on osteocyte gene expression. Vitamin D (10-8 M) significantly increased VEGF164 production by MLO-Y4 osteocytes at 7 (23 % increase, 95 % CI 11-35 %, p00.01) and 24 hours (41 % increase, 95 % CI 18-65 % p00.007) when compared to control. Although not statistically significant at lower concentrations, results appear to show a dosedependent effect. 10-8 M vitamin D treatment reduced IL-6 production by MLO-Y4 osteocytes at both time points although the results only reached significance after 24 hours (24 % reduction, CI 13-35 %, p00.04). PTH did not alter either VEGF or IL-6 secretion. Conclusion: This study reveals the novel findings that VEGF164 reduced osteocyte number and that Vitamin D up-regulated VEGF and down-regulated IL-6 production by MLO-Y4 osteocytes in vitro. We have previously shown that osteocyte VEGF and IL-6 secretion is reduced by dexamethasone suggesting a potential therapeutic role for vitamin D in modulating corticosteroid-induced bone disease. O14 HOST-DERIVED MMP-7 DECREASES MYELOMA PROGRESSION IN VIVO: AN UNEXPECTED ROLE FOR MMP-7 IN MYELOMA BONE DISEASE Seint Lwin (1,2) Conor Lynch (2) Jessica Fowler (2) James Edwards (1,2) Claire Edwards (1,2) presenting University of Oxford, Oxford, UK (1)Vanderbilt University, Nashville, TN, USA (2)

S529

Introduction: Defining interactions within the host bone marrow microenvironment that promote multiple myeloma (MM) is essential to identify new therapeutic approaches. Matrix metalloproteinases (MMPs) are key regulators of tumour: host interactions due to their ability to alter the activity of multiple substrates. MMP7 has recently been implicated in breast and prostate cancer-mediated osteolysis and we hypothesized that host-derived MMP7 may play a role in MM bone disease in vivo. Methods: Rag-2−/−/MMP7−/− (MMP7−/−) mice or wildtype controls (WT) were inoculated with 5TGM1-GFP MM cells. Tumour burden was assessed by ELISA and flow cytometry, bone disease was assessed by bone histomorphometry and microCT. MMP-7 activity and TIMP-1 levels were measured in serum of patients with MM by caesin zympography and ELISA respectively. Discussion: Inoculation of 5TGM1-GFP MM cells into MMP7 −/−mice unexpectedly resulted in a significant increase in the rate of tumour growth and tumour burden within bone, when compared to MM-bearing WT mice (p<0.01). MicroCT analysis demonstrated a 47 % decrease in BV/TV and 67 % increase in osteolytic lesions in MM-bearing MMP7−/−mice as compared with MM-bearing WT mice (p<0.01). No significant difference in osteoclast or osteoblast number was found. In support of an effect of host-derived MMP7 on tumor growth and/or survival, a 51 % decrease in MM cell apoptosis was observed in MM-bearing MMP7−/− mice (p<0.05). Furthermore, in vitro studies found no MMP7 expression in MM cells, and no direct effect of MMP7 on MM cell proliferation or apoptosis. Using a candidate list of MMP7 substrates, we identified that MMP7−/−mice had increased serum osteopontin (OPN), compared to WT (233.0 ±5.7 ng/ml vs. 167.0± 12.3 ng/ml, p<0.001). In vitro studies demonstrate that MMP7-cleaved OPN increases myeloma cell apoptosis. OPN is a MM cell survival factor and we propose that processing of OPN by MMP7 may abrogate the survival effect of full-length OPN. To determine the clinical relevance of our observations, we find that MMP-7 activity is decreased in patients with myeloma (p<0.003), and that the MMP inhibitor, TIMP1- is increased (p<0.0001), as compared to age-matched controls. Conclusion: Our studies suggest that host-derived MMP7 inhibits MM in vivo, demonstrating that MMPs can play differential roles in specific disease contexts and supporting the rationale for the design of selective MMP inhibitors for the treatment of MM bone disease. O15 BMP2-ENOS AXIS REGULATES BONE VOLUME AND STATIN-INDUCED BONE FORMATION Megan Moore-Weivoda (1,2) Seint Lwin (1) Ross Garrett (1,2) Jonathon Lowrey (1,2) Gregory Mundy (1,2) Gloria Gutierrez (1,2) James Edwards (1) presenting University of Oxford, Oxford, UK (1) VUMC, Nashville, USA (2)

S530

BMP2 is essential for normal skeletal development and bone formation. BMP2 also mediates the bone anabolic properties of statins although the mechanism by which this occurs is unclear. As statins induce beneficial vascular effects via endothelial nitric oxide synthase (eNOS), and eNOS KO mice have a skeletal phenotype strongly resembling that of BMP2 KO mice, we hypothesize that eNOS plays a central role in mediating the anabolic effect of statins by controlling BMP2. To investigate this, we employed in vitro studies, ex vivo calvariae assays, and genetically modified mice to assess the effect of statins and/or NO donors or NOS inhibitors on gene expression and bone formation in tissues from normal, DN-BMPRII, and eNOS KO animals. In addition, we characterized the skeletal phenotype of DN-BMPRII:eNOS KO mice by μCT scanning and histomorphometry. Simvastatin (10 μM) increased eNOS and BMP2 mRNA and led to a 50 % increase in NO in osteoblastic cells (p<.05). Similarly, lovastatin (5 μM) increased eNOS and BMP2 mRNA 2-fold (p<.001) and 5-fold (p<.02), respectively.Simvastatin significantly increased bone formation over vehicle control (6.7±0.6 vs 2.5±0.6 sq mm x10-3) in calvarial culture. Inhibition of NOS with L-NAME (10uM) returned bone growth to levels comparable with vehicle alone (2.8± 0.5 sq mm x10-3). Simvastatin treatment of eNOS KO murine calvariae demonstrated no significant increase in bone growth. Osteoblastic cells treated with NO donors increased BMP2 protein by 2-fold, while BMP2 treatment did not alter eNOS. Also, osteoblasts isolated from eNOS KO mice exhibited decreased BMP2 expression. These data support a linear pathway in which statin treatment increases BMP2 via eNOS. To assess the skeletal effects of this pathway in vivo, we examined DNBMPRII and eNOS KO mouse models. DN-BMPRII mice displayed a decrease in BV/TV by μCT analysis compared to WT (17.83±.16 % vs 5.6±1.12 %, p<0.01). They also exhibited decreased TbN, TbTh, and increased TbS. While the eNOS KO mice also displayed a decreased BV/TV, there was no difference between the WT and DN receptor mice. Together these data support a signalling system in which statins trigger eNOS activity to stimulate BMP2 expression. These findings demonstrate a clear mechanism through which the bone anabolic effects of statins may be mediated. O16 THE ROLE OF DUAL SPECIFICITY PHOSPHATASE1 IN LIMITING INFLAMMATORY OSTEOLYSIS AND ITS ROLE IN COLLAGEN-INDUCED ARTHRITIS Youridies Vattakuzhi, Andy Clark, Nikki Horwood; Kennedy Institute of Rheumatology, London, UK Activation of the p38 mitogen-activated protein kinase (MAPK) pathway is crucial in driving inflammatory osteolysis by inducing pro-inflammatory gene expression and regulating differentiation and activation of osteoclasts (OCLs). Dual specificity phosphatase 1 (DUSP1) dephosphorylates and

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

inactivates p38 MAPK. It is up-regulated by different proinflammatory stimuli and acts as a negative regulator to limit the duration and strength of inflammatory responses. DUSP1 expression is also up-regulated by glucocorticoids (GCs) and contributes to the anti-inflammatory effects of GCs. The aim of the present study was to investigate whether DUSP1 negatively regulates inflammatory osteolysis in collagen-induced arthritis, a model of rheumatoid arthritis, and to determine whether DUSP1 mediates anti-inflammatory effects of GCs. Dusp1+/+ and Dusp1−/− mice were immunized with type II chicken collagen to induce chronic arthritis. After onset of disease mice were treated with dexamethasone, injected intra peritoneally every second day for ten days. Development of disease and bone resorption of the affected paws was assessed by histologic examination, TRAP staining and micro-CT analysis. Furthermore, anti-collagen IgG antibodies were measured in the serum and T cell responses of immunized mice were analysed ex vivo. Compared to Dusp1+/+ mice, Dusp1−/− mice showed higher incidence, earlier development and more severe disease, characterised by increased numbers of OCLs and bone loss in affected joints. Dexamethasone treatment reduced clinical and histological scores in Dusp1+/+ mice but was less effective in Dusp1−/− mice. Surprisingly, serum anticollagen IgG2a levels were significantly lower in Dusp1−/− compared to Dusp1+/+ mice. These observations show that DUSP1 negatively regulates inflammatory osteolysis and mediates therapeutic effects of GCs in an experimental model of RA. DUSP1 does not appear to exert its protective effect by limiting the humoral immune response to collagen immunisation. O17 POWER COMPARISON OF DIFFERENT SCREENING TOOLS (FRAX WITHOUT BMD, OST, ORAI, ORISIS, SCORE AND AGE ALONE) TO IDENTIFY WOMEN WITH INCREASED RISK OF FRACTURE. ARE COMPLEX TOOLS BETTER Katrine Hass Rubin (1,2), Bo Abrahamsen (1,3), Teresa Friis-Holmberg (4), Jacob V.B. Hjelmborg (5), Mickael Bech (6), Anne Pernille Hermann (2), and Kim Brixen (1,2); Institute of Clinical Research, University of Southern Denmark (1) Department of Medical Endocrinology, Odense University Hospital (2) Department of Medicine F, Copenhagen University Hospital Gentofte (3) National Institute of Public Health, University of Southern Denmark, Copenhagen, Denmark (4) Department of Biostatistics, Institute of Public Health, University of Southern Denmark, Odense, Denmark (5) 6Institute of Public Health, Health Economics, University of Southern Denmark (6) Introduction:Fracture prediction tools help to identify individuals, who may benefit from treatment. One of the most recently released tools is the Fracture Risk Assessment Tool

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

(FRAX®) developed by the World Health Organisation (WHO), but several simpler tools are available to assess fracture risk, e.g. OST, ORAI, ORISIS and SCORE. Purpose: To compare the power of FRAX (without BMD) and simpler screening tools (OST, ORAI, OSIRIS, SCORE and age alone) in predicting fractures. Material and Methods: This study was a prospective, population-based study performed in the Region of Southern Denmark including 3,860 women aged 40–90 (mean±SD; 63 ±14) years, who have returned a questionnaire concerning items on risk factors for osteoporosis in March 2009. Fracture risk was calculated using the different screening tools (FRAX, OST, ORAI, ORISIS and SCORE) for each woman. The women were followed from the Danish National Patient Register registering new major osteoporotic fractures (vertebral, humerus, forearm and hip fractures) during 32 months, counting only the first fracture per person. ROC areas under the curve (AUC) statistics were calculated. Results: Complete data to calculate fracture risk were available for 3,614 women. A total of 4 % of experienced a new major osteoporotic fracture during the follow up period. There were no differences in the AUC values between FRAX without BMD and the more simple tools; AUC values ranged for all tools from 0.71 to 0.73 (p00.86). Restricting the analysis to women aged 50+ years did not change the results. Conclusion: During short-term follow-up (32 months) the more complex FRAX tool did not perform better in the fracture risk prediction compared with simpler tools (OST, ORAI, OSIRIS and SCORE or age alone) in a screening scenario where BMD was not measured. This conclusion supports the findings from respectively Ensrud et al. (2009) and Hillier et al. (2011) even though the present study included women ranging from 40–90 years. O18 ANXIETY, FRACTURE RISK AND ADHERENCE TO TREATMENT: THE MRC SCOOP TRIAL Nicholas Harvey (1), Janet Cushnaghan (1), Wendy Lawrence (1), Georgia Ntani (1), Karen Collins (1), Liz Lenaghan (1), Lee Shepstone (2), Eugene McCloskey (3), John Kanis (4), Cyrus Cooper (1); MRC Lifecourse Epidemiology Unit, Southampton, UK (1); University of East Anglia, Norwich, UK (2); University of Sheffield, Sheffield, UK (3); WHO Collaborating Centre for Metabolic Bone Diseases, Sheffield, UK (4) Background: We examined the relationships between anxiety, fracture risk and adherence to treatment in the MRC Scoop Trial, a multi-centre RCT of a screening strategy using FRAX in 12,486 women aged 70–86 years. Materials and methods: In those randomised to screening, a ten-year fracture risk was computed from baseline information

S531

followed by BMD by DXA in selected participants at higher risk. This yielded three groups within the screening arm: low risk without DXA (LR), low risk after DXA (LRD) and high risk after DXA (HRD). Anxiety [State-trait Anxiety Inventory (STAI), range 20 (low anxiety) to 80 (high anxiety)] was assessed by postal questionnaire at baseline (before randomisation) and at 6 and 12 months post randomisation. Adherence was assessed by questionnaire, asking whether participants had stopped taking osteoporosis medication if they had been prescribed it. Discussion: The LR group contained 3170 participants (52.9 % of the treatment group); the LRD and HRD groups contained 1926 (32.1 %) and 898 (15 %) participants respectively. The prevalence of risk factors for fracture increased from LR to LRD to HRD. The median anxiety score assessed at all three time points was lower in those participants categorised as low risk without DXA than in those categorised as low risk after DXA or high risk after DXA. [Median(IQR) STAI score at baseline: LR:30(20 to 40); LRD:33.3(23.3 to 43.3); HRD:33.3(23.3 to 43.3),p < 0.001]. Amongst all participants who had been prescribed osteoporosis medication, STAI assessed at baseline was lower in participants who were adherent rather than nonadherent at 12 months (baseline median(IQR) STAI033.3 (23.3 to 43.3) in adherent vs 38.3(26.7 to 43.3) in nonadherent participants,p00.08), with similar findings for STAI assessed at 6 months (p00.058). Conclusions: Anxiety score at all times was lowest in lowest fracture risk group. Baseline and 6 month anxiety score was greater in non-adherent than adherent patients at 12 months follow up. This suggests that greater levels of anxiety may be associated with higher fracture risk and poorer adherence to treatment, and that psychological factors might be profitably addressed in the assessment and treatment of osteoporosis. O19 VERTEBRAL FRACTURE ASSESSMENT (VFA) IS MORE USEFUL IN POPULATION-BASED SETTINGS THAN AS PART OF FRACTURE LIAISON SERVICES (FLS) Emma Clark (1,2) presenting Virginia Gould (1) Louise Carter (2) Leigh Morrison (1) Jon Tobias (1) University of Bristol, Bristol, UK (1) North Bristol NHS Trust, Bristol, UK (2) Introduction:VFA is potentially useful in screening patients for vertebral fractures (VFs), but the context in which this is best used is unclear. In this study, we applied our recent findings of a screening programme for VFs in older women from Primary Care, by evaluating the use of this screening tool to select patients for referral for VFA. In particular we aimed to determine whether VFA screening is more efficient when applied to such a novel screening strategy in Primary

S532

Care, as compared with use of VFA in patients referred for DXA scans after low trauma fracture as part of FLS. Methods: Two cohorts of women were recruited in parallel: Group 1 (251 women) identified from our primary carebased screening study as being at high risk of VFs; Group 2 (377 women) identified from the NHS fracture clinic (FLS) with a recent non-VF. The screening tool for Group 1 was the published COSHIBA tool. The intervention was DXA with VFA. Outcome was the impact of VFA results by ABQ on putative treatment decisions. This was assessed by comparing the proportion of patients identified with VF with osteopaenia who would otherwise not be treated in the 2 groups. Ethics approval was obtained. Results: Women in Group 1 were older and had lower BMD than Group 2 (FLS). In Group 1, VFA identified 15 women with VF, of whom only 6 had osteoporosis on DXA, so the results of VFA would change management in 9/251 (3.6 %). In Group 2, VFA identified 10 women with VF, of whom 5 had osteoporosis on DXA. Results of VFA performed in FLS would change management in 5/377 (1.3 %). Conclusions: There may be a role in using VFA in a population-based setting after screening for risk of VF. There is little justification for using VFA as a routine part of FLS as it would only change management in 1 % of cases. VFA is only likely to be cost-effective if it is used in a targeted way in high-risk groups. O20 VERTEBRAL FRACTURE ASSESSMENT BY THE GE LUNAR iDXATM VERSUS RADIOGRAPHIC ASSESSMENT IN CHILDREN Nicola Crabtree, Nick Shaw, Wolfgang Hogler, Natalie Bebbington, Dee Chapman, Steve Chapman; Birmingham Children's Hospital, Birmingham, UK Introduction:Vertebral fracture assessment (VFA) by DXA is an accepted tool for use in adults. However, its use in children has not been validated. The iDXA is a high resolution bone densitometer capable of producing good high images with minimal radiation exposure. The aim of this study was to validate VFA using iDXA against spinal radiographic assessment(RA) for the identification of vertebral fractures in children. Material and Methods: Spine radiographs and VFA(L5-T2) by iDXA were acquired on the same day in 25 children. An additional 24 children had VFA alone performed. Agreement between RA and VFA was assessed by an expert paediatric radiologist and two metabolic bone specialists; agreement between raters for VFA alone was additionally assessed by an expert, an experienced and a naive DXA operator. Vertebrae were ranked as normal, mild, moderate or severe if they had <10 %, 11-25 %, 26-50 % and >50 % deformity, respectively. Levels of agreement were calculated using the kappa statistic and consistency by the intra-class correlation coefficient (ICC).

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

Discussion: Depending on rater, 84.0-90.5 % of the visible vertebrae were analysable by RA. In contrast, 97.2 % of the visible vertebrae were analyzable by VFA. Moderate agreement was noted between raters for RA[kappa 0.523-0.588] and between RA and VFA[kappa 0.511-0.584], substantial agreements were observed between raters for VFA assessment[kappa0.637-0.704]. Agreement improved if the vertebral deformities were dichotomised as normal or mild versus moderate or severe, with substantial agreement observed between raters for RA and between RA and VFA and almost perfect agreement observed between raters for VFA alone[ICC ranged from 0.802-0.895, p<0.0001]. Similar patterns were observed between all 6 raters for VFA with no discernable difference detected for rater experience. Poorer agreement was seen between quantitative and visual assessment[kappa 0.302-0.408], however this did improve when the classifications were dichotomised. Conclusion: VFA is accurate and consistent when identifying vertebral fractures, particularly moderate and severe fractures in chronically sick children. Moreover, its accuracy appeared to be independent of reader experience. Consequently, VFA should prove to be a useful tool in the assessment of bone health in children. O21 PLACENTAL SIZE AT 19 WEEKS PREDICTS NEONATAL BONE MASS: RESULTS FROM THE SOUTHAMPTON WOMEN’S SURVEY Christopher Holroyd (1) presenting Sarah Crozier (1) Pamela Mahon (1) Nicola Winder (1) Keith Godfrey (1,2) Hazel Inskip (1) Nicholas Harvey (1) Cyrus Cooper (1,3) MRC Lifecourse Epidemiology Unit, Southampton, Hampshire, UK (1) Southampton NIHR Biomedical Research Unit in Nutrition, Diet and Lifestyle, Southampton, UK (2) NIHR Musculoskeletal Biomedical Research Unit, Institute of Musculoskeletal Sciences, Oxford, UK (3) Introduction: Poor early growth is associated with reduced adult bone mineral content and an increased risk of hip fracture. In mother-offspring studies, maternal smoking, nutrition and physical activity during pregnancy appear to influence offspring bone mineral accrual. It is likely that these relationships are mediated by the placenta, but it is unclear whether the important determinant is placental size or function. In this study we investigate the relationship between placental size and offspring bone mass using the Southampton Women’s Survey (SWS) Material and Methods: The SWS is a unique prospective cohort of 12,583 non-pregnant women aged 20–34 years, of which 3,156 subsequent singleton pregnancies were followed. Detailed placental measurements were obtained by fetal ultrasound scanning at 19 weeks. In a subset of births, dual x-ray absorptiometry (DXA) was performed on the neonate within the first two weeks. Pearson correlation and linear regression were

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

used to relate placental measurements to neonatal body composition and bone mass. Discussion: We identified 757 mother-neonate pairs with complete ultrasound and DXA data. The statistically significant, positive predictors of offspring bone area (BA), bone mineral content (BMC) and bone mineral density (BMD) (P <0.05) included placental circumference, length of attachment to the uterine wall, cross-sectional area and volume. Placental volume at 19 weeks was positively associated with neonatal total lean mass, and percent and total fat mass (p<0.0001), however was negatively associated with percent lean (p<0.0001). This indicates that as placental volume increased, total neonatal size increased but with an increase in percentage fat and a reduction in percentage lean within the overall size envelope. All associations remained after adjusting for maternal factors known to affect offspring bone mass. Conclusion: Placental size is positively associated with intrauterine bone mineral accrual. These associations appear independent of maternal factors known to influence neonatal bone mass, suggesting that these factors might act through modulation of placental function rather than placental size. Low placental volume early in pregnancy may be a marker of a smaller postnatal skeletal envelope and potentially increased risk of fracture in older age. O22 LUMBAR SPINE AND HIP BONE MINERAL DENSITY ARE IMPORTANT RISK FACTORS FOR STRESS FRACTURE IN ROYAL MARINE RECRUITS Trish Davey (1,2) presenting Susan A. Lanham-New (2) Adrian. J. Allsopp (1) Pat Taylor (4) Cyrus Cooper (3) Joanne L. Fallowfield (1) Institute of Naval Medicine, Gosport, UK (1) University of Surrey, Guildford, UK (2) University of Southampton, Southampton, UK (3) Southampton General Hospital, Southampton, UK (4) Introduction: Royal Marine (RM) recruits (aged 16–32 y) undergo an arduous 32-week military training programme. The prevalence of stress fracture (SF) amongst RM recruits is of the order of 5 %. The aim of this study was to investigate whether bone mineral density (BMD) at the lumbar spine (LS), femoral neck (FN), femoral trochanter (FT) and Ward’s triangle (WT) sites were important risk factors for SF in RM recruits. Materials and Methods: A total of 1090 recruits were monitored throughout RM training. Recruits who suffered a SF during training were identified (n 65) and subsequently control-matched for age, height, body mass, aerobic fitness and smoking history with non-injured recruits (n 65). All recruits underwent Dual X-Ray Absorptiometry (DXA) scanning of the LS, FN, FT and WT sites. All recruits had undergone Broadband Ultrasound Attenuation (BUA) assessment of the dominant foot calcaneus at the start of

S533

training. A Mann–Whitney U test was undertaken to investigate differences in BMD and BUA between SF and control recruits. Results and Discussion: SF recruits had lower BMD at the LS, FN, FT and WT than controls (unadjusted and adjusted for body mass) (P<0.05). Of the SF group, 49 % (n 32) had a T-score of the LS below −1, compared to 20 % of controls (n 13) (P<0.05). The odds ratio of a recruit suffering a SF if the T score was below zero was 4.6 for the LS and 1.8 for the FN. SF recruits did not differ from controls in smoking status, lean body mass or body fat. Despite a correlation between BUA and BMD (r00.3, p<0.01), there was no difference in BUA between SF and matched control recruits. Conclusion: Low LS and hip BMD are important risk factors for SF in RM recruits. Further analysis is currently underway to explore differences in volumetric BMD using pQCT, bone resorption and nutritional status. Nevertheless, the DXA BMD data suggest that potential recruits should be advised on strategies to promote bone health and minimise stress fracture risk. These include smoking cessation, appropriate physical training and a balanced diet providing at least 1000 mg.d-1 of calcium. O23 RARE MUTATIONS ASSOCIATED WITH OSTEOCLAST-POOR OSTEOPETROSIS PROVIDE MOLECULAR INSIGHTS INTO RECEPTOR ACTIVATOR OF NFKB SIGNALLING Subhajit Das, Ilnaz Sepahi, David Mellis, Angela Duthie, Miep Helfrich, Julie Crockett; University of Aberdeen, Aberdeen, UK Activation of the signalling pathway downstream of receptor activator of NFkB (RANK) by RANK ligand (RANKL) is critical for osteoclast formation, function and survival as highlighted by the osteopetrotic phenotypes of Rank and Rankl-knockout mouse models that lack osteoclasts. Mutations in TNFRSF11A (encoding the RANK receptor) have been associated with osteoclast-poor autosomal recessive osteopetrosis in patients. This condition is often fatal and characterised by susceptibility to fracture and high bone mineral density, as a consequence of an absence of osteoclasts. To date, 12 different mutations within TNFRSF11a have been identified. Two of these mutations cause premature termination of translation of the RANK protein resulting in loss of critical domains that are required for signal transduction downstream of RANK. We have therefore been investigating the effects of these mutations on receptor processing and downstream activation of NFkB to inform on the relative contribution of each domain to proper activation of the RANK signalling pathway. Two truncated RANK proteins resulting from substitution mutations (W434X and G280X) within the C-terminal, intracellular region of the protein result in loss of the

S534

intracellular ligand-independent oligomerisation motif (546 I-IV-V-Y-V 551). In addition, RANK containing the G280X mutation also lacks the TRAF6 binding domain. When these proteins were overexpressed in Hela cells, immunostaining and confocal microscopy revealed that, like wildtype-RANK, they could be detected at the plasma membrane, but by contrast to wildtype-RANK, they did not accumulate predominantly in the golgi apparatus and exhibited a more diffuse pattern of staining throughout the rest of the cell. When cells were transfected with myc and FLAG tagged RANK constructs, immunoprecipitation revealed that the mutant proteins failed to oligomerise. Finally, when we analysed signalling downstream of RANK, immunoprecipitation showed that wildtype-RANK and W434X-RANK interact with TRAF6, whereas G280XRANK does not. Furthermore, p65 translocation experiments revealed that RANKL activates NFkB in cells transfected with wildtype-RANK or W434X-RANK, but not G280X-RANK. Taken together, these results strongly suggest that, whilst the TRAF6 domain is critical for activation of the RANK signalling pathway, oligomerisation via the I-I-V-V-Y-V motif is not essential. O24 A 55KDA ISOFORM OF SEQUESTOSOME-1 LACKING THE N-TERMINAL PB1 DOMAIN INTERACTS WITH 62KDA SEQUESTOSOME-1 (P62) AND EXHIBITS DIFFERENTIAL EXPRESSION Eman Azzam, Miep Helfrich, Lynne Hocking; Institute of Medical Sciences · University of Aberdeen, Aberdeen, UK Introduction:Paget’s disease of bone (PDB) is characterized by focal lesions of increased bone turnover, driven by overactive osteoclasts. PDB is caused by mutations in the gene encoding Sequestosome-1 (SQSTM1) affecting the Cterminal ubiquitin-associated domain, but the mechanism by which these cause PDB is not yet clear. SQSTM1, also known as p62, is a ubiquitously-expressed scaffold protein of 62 kDa that functions in multiple signalling pathways important for cell survival and osteoclast activity. In this abstract we report for the first time the presence of a short isoform of SQSTM1 lacking the N-terminal PB1 domain. Methods: Western blotting, immunofluorescent staining and immunoprecipitation were used to investigate SQSTM1 expression in bone biopsies from Pagetic patients, HEK293 cells, HeLa cells and human osteoclast-like cells (hOCL). Cells were tested under basal and bafilomycin-treated (inhibits autophagic protein degradation) conditions. Antibodies against different domains of SQSTM1 were used: αSQSTM1-N detects 62 kDa-SQSTM1 only; α-SQSTM1-C detects both isoforms. SQSTM1 transcripts were detected by PCR and sequencing of cDNA. Results: Western blotting using α-SQSTM1-C specifically detects two bands in HEK293, HeLa and hOCL, one at

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

62 kDa (p62) and another at ~55 kDa. Both isoforms accumulate upon bafilomycin treatment, indicating both are degraded by autophagy. In all cell types, four SQSTM1 transcripts were detected, which differ in their 5’-untranslated region; one transcript encodes p62, while the other three encode 55 kDa-SQSTM1 and have identical coding sequences. The 55 kDa isoform of SQSTM1 is predicted to lack the N-terminal PB1 homo-dimerisation domain present in 62 kDa-SQSTM1. Immunofluorescent staining demonstrated some colocalisation between the two SQSTM1 isoforms, and functional interaction between endogenous isoforms was confirmed by immunoprecipitation. Differential cellular localisation of SQSTM1 isoforms was observed in Pagetic bone biopies. Cell-specific differences in expression were observed by Western blotting: in hOCL, 55 kDaSQSTM1 was ~45x more abundant than the 62 kDa isoforms. Conclusion: A second, shorter isoform of SQSTM1 is expressed by osteoclasts and other cells. Both isoforms contain the C-terminal UBA domain mutated in PDB. The isoforms exhibits differential cellular expression, with the short isoform most abundant in osteoclasts. The role of mutations in 55 kDa-SQSTM1 has yet to be characterised. O25 A GENOME-WIDE ASSOCIATION META-ANALYSIS AND A MOUSE GENE DELETION MODEL IDENTIFY WNT16 AS A POTENTIAL REGULATOR OF CORTICAL BONE THICKNESS Jonathan Tobias (1) presenting Claes Ohlsson (2) Lavinia Paternoster (1) Terho Lehtimaki (3) Mika Kahonen (3) Olli Raitakari (4) Marika Laaksonen (5) Jorma Viikari (4) Dan Mellstrom (2) Sofia Moverare-Skitric (2) Magnus Karlsson (6) Osten Ljunggren (1) John P Kemp (1) Maria Nethander (2) Liesbeth Vandenput (2) Robert Brommage (8) Jeff Liu (8) David M Evans (1) Mattias Lorentzon (2) University of Bristol, Bristol, UK (1) University of Gothenburg, Gothenburg, Sweden (2) University of Tampere, Tampere, Finland (3) University of Turku, Turku, Finland (4) University of Helsinki, Helsinki, Finland (5) Lund University, Malmo, Sweden (6) Uppsala University, Uppsala, Sweden (7) Lexicon Farmaceuticals, Texas, USA (8) Background: Several loci associated with areal bone mineral density (aBMD) (as measured by dual x-ray absorptiometry (DXA)) have previously been identified in large bonerelated genome-wide association (GWA) studies. Cortical bone has been reported to be the main determinant of bone strength, but little is known of its genetic determinants, and as of yet, no GWA studies have been performed searching for genetic signals for cortical thickness. Methods: In order to determine genetic markers (Single Nucleotide Polymorphisms-SNPs) for cortical thickness,

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

we performed a GWA study of this trait, as measured by peripheral quantitative computed tomography (pQCT) in the tibial diaphysis, among 5,878 Caucasian subjects from three cohorts. We used an inverse variance weighted fixed-effect model meta-analysis of study-specific results. Results: An SNP (rs9525638) near the RANKL locus (Chr13) was identified as independently associated with cortical thickness (β0−0.11(0.02) (SD per allele and (SE)), p03.8x10-10). We also discovered a SNP (rs2707466) located in the WNT16 gene (Chr7), associated with cortical thickness (effect size of 0.11 SD per allele and p06.2x10-9). This SNP was also associated with cortical thickness (0.11 SD per allele and p00.008) of the tibia in a replication cohort of 75-year-old men (n01032). Wnt signalling is known to regulate bone turnover, but knowledge of the individual roles of the nineteen wnts in bone metabolism is very limited. In order to investigate the role of Wnt16 on skeletal traits, we developed Wnt16 inactivated mice (Wnt16−/−). Female Wnt16−/− mice had 27 % (p<0.001) reduced cortical thickness at the femur midshaft compared with their wild type littermates. Conclusions: We identified an SNP at the WNT16 locus consistently associated with cortical thickness in young, middle aged and elderly subjects. A Wnt16 gene deletion in mice resulted in reduced cortical thickness strongly supporting an important role of Wnt16 in regulating cortical bone thickness. O26 ECTO-NUCLEOTIDE PYROPHOSPHATASE/ PHOSPHODIESTERASE-1 IS EXPRESSED BY OSTEOCLASTS BUT DOES NOT AFFECT OSTEOCLAST FUNCTION Mark Hajjawi, Isabel Orriss, Timothy Arnett; University College London, London, UK Extracellular nucleotides, including adenosine triphosphate (ATP), adenosine diphosphate (ADP), uridine triphosphate (UTP) and uridine diphosphate (UDP) act through the P2 receptors which can be found on both the osteoblast and the osteoclast. In the extracellular compartment adenine nucleotides are dephosphorylated in stages to adenosine by ectonucleotidases. Ecto-nucleotidases are enzymes expressed on the surface of cells, there are four major families: ectonucleoside triphosphate diphosphohydrolase (E-NTPdase), alkaline phosphatase (ALP), ecto-nucleotide pyrophosphatase/ phosphodiesterase (ENPP) and ecto-5’nucleotidase. The ENPP family consists of seven isoenzymes. ENPP1-3 hydrolyzes 5’triphosphates to 5’-monophosphates (ATP to AMP) with the release of pyrophosphate (PPi). PPi prevents the formation of hydroxyapatite crystals and therefore the mineralisation of osteoid deposited by osteoblasts. The aim of this study was to determine if mouse osteoclasts express ENPPs and if the knockout of the ENPP1 gene has an effect on the development

S535

and resorptive activity of osteoclasts. Osteoclasts, formed from the bone marrow of 8 week old knockout or wild type mice were cultured on ivory slices with 200 ng/ml M-CSF and 3 ng/ml RANKL. RT-PCR was used to study the expression of ENPP enzymes at 3, 6, 8 and 10 days of culture. Thymidine 5’-monophosphate p-nitropheyl was used to determine the total ENPP activity of osteoclasts. ATP release was measured by a luminescent method. RT-PCR demonstrated that osteoclasts express ENPP1 from day 6 onwards, with increasing expression with maturity. RT-PCR did not show the expression of ENPP2; mouse brain was used as a positive control. Osteoclasts were found to possess ENPP activity, indicating the presence of the enzyme. Knockout of the ENPP1 gene did not have an effect on the number of osteoclasts at day 10 of culture, or on their resorptive activity. The release of ATP per osteoclast was not affected by knocking out the ENPP1 gene. These results show that mouse osteoclasts express the PPi producing enzyme ENPP1. With the loss of ENPP1 it could be expected that more ATP would be hydrolysed by NTPdase, hence there would be a greater transient concentration of ADP, however, the removal of ENPP1 does not effect the formation or activity of mouse osteoclasts in vitro. O27 DIET MANIPULATION AND BONE HOMEOSTASIS; LOW CALORIES PRESERVE BONE WITH AGE AND HIGH FAT DIETS DECREASE BONE MASS Seint Lwin, Barbara Rowland, Claire Edwards, James Edwards; University of Oxford, Oxford, UK The quantity and quality of the food we consume has major implications for health and well-being throughout life. Obesity has a severe impact on health with 1 in 4 UK adults considered obese. Therefore, it is crucial to understand the effect of such dietary regimes on bone, and the implications for increased risk of musculoskeletal disease. To investigate this, we compared normal mice fed regular or high fat chow (42%Kcal fat) from 6 to 15 weeks. Also we compared mice fed ad libitum (AL) to those on a low caloric diet (60 % of normal)(n010). Mice were analysed for body composition, and bone was assessed by uCT scanning and histomorphometric analysis. Gene expression was assessed by qPCR. Mice on the low calorie diet were considerably leaner at 25 g±1.1 compared to controls (33 g± 1.5, p<0.001) with higher BMD (195.7±10.5) compared to 148.4±8.0 (p<0.01). This corresponded with increased BV/ TV in the trabecular bone of low calorie-fed mice (10.8 % ±0.4 Vs 6.7 %±0.5, p<0.0001), but no significant change in cortical bone was observed. In the obesity-mimic high fat group, a considerable increase in body fat composition was observed (7.7 %±0.7) compared to controls (2.4 %±0.4, p< 0.001). Also, the high-fat diet led to a decrease in BV/ TV (2.9 %±0.3) compared to mice fed regular diets (6.1 %±

S536

0.6 p<0.01), and significant changes in trabeculae and BMD. A high-fat diet led to decreased adiponectin (ADPN) expression. Interestingly, ADPN is elevated following caloric restriction, suggesting that nutritional regulation of ADPN may impact bone. These studies describe striking effects of diet on bone, and suggest that a low calorie diet may protect against osteoporosis, but a high fat diet will impair bone homeostasis and reduce bone mass, possibly through the regulation of adipokines such as ADPN. O28 THE ANTI-DIABETIC DRUG METFORMIN HAS NO DELETERIOUS EFFECT ON BONE MASS IN VIVO BUT DOES NOT INDUCE OSTEOGENESIS NOR FRACTURE HEALING Jeshmi Jeyabalan (1) presenting, Benoit Viollet (2), Peter Smitham (3), Yasin Undre (1), Stephanie Ann Ellis (1), Allen Goodship (3), Chantal Chenu (1); Royal Veterinary College, London, UK (1), INSERM U567, Paris, France (2), Institute of Orthopaedics & Musculoskeletal Science, Stanmore, UK (3) There is strong clinical evidence that patients with type 2 diabetes mellitus (T2DM) have increased risk of fractures and augmented incidence of serious fracture complications, including delayed healing and non unions. To worsen the situation, thiazolidinediones (TZDs), a class of medications commonly used in the treatment of T2DM, increase the risk of secondary osteoporosis and prevalence of fractures in those patients. In view of these adverse skeletal effects of TZDs, more studies are necessary to understand the actions on bone of other T2DM therapies, such as metformin, the only biguanide in clinical use. While the osteogenic potency of metformin has been well documented in vitro, few studies have investigated the effects of metformin on bone mass and fracture healing in vivo. We first subjected 12 week-old female C57B/6-129 Sv mice to ovariectomy (OVX). 4 weeks after OVX, mice were randomly divided into two groups, one receiving saline and the other metformin administered by gavage (100 mg/kg/daily). After 4 weeks treatment, tibia were harvested from these mice and bone micro-architecture determined by micro-computed tomography (micro-CT). In another experiment, 24 female Wistar rats aged 3 months were randomly divided in two groups, one group receiving saline, the other group receiving metformin administered orally for 8 weeks via the drinking water at a concentration of 2 mg/ml. After 4 weeks metformin treatment, a mid-diaphyseal, transverse osteotomy was performed in the left femur, stabilized via a precision miniature external fixator. Rats were sacrified 4 weeks after osteotomy. Right contralateral and left osteotomised femora were excised, bone architecture analysed by micro-CT in the right femur while fracture healing and callus volume were determined in the left femur by x-ray analysis and micro-CT,

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

respectively. Our results show no significant differences in trabecular bone volume, number, separation and thickness in metformin-treated mice compared to saline. Metformin had also no significant effect on cortical parameters in mice. Similarly, our results in rat showed no significant differences for all cortical and trabecular parameters between the metformin and saline-treated groups. Mean X-ray scores assessed on four cortices of both control and metformin groups showed no significant differences of healing between the groups. Fracture callus volume and mineral content after 4 weeks were similar in metformin and saline groups. Our results indicate that while metformin has no adverse effects on bone, it does not neither promotes bone mass, as suggested by in vitro studies. This confirms clinical data which have not shown direct links between metformin and decreased fracture risk. O29 HIGH CIRCULATING SEROTONIN IN CARCINOID SYNDROME IS NOT ASSOCIATED WITH ALTERATIONS IN BONE TURNOVER OR BONE DENSITY Jennifer Walsh (1) presenting John Newell-Price (2) Miguel Debono (2) Richard Eastell (2) Sheffield Teaching Hospitals NHS Foundation Trust, Sheffield, UK (1) University of Sheffield, Sheffield, UK (2) Introduction: Gut-derived serotonin has been proposed as a regulator of bone formation, and inhibition of gut serotonin synthesis increases bone formation in rodents. Carcinoid neuroendocrine tumours can produce very high levels of circulating serotonin and so offer a model of serotonin excess in humans. There is very little information on the effects of carcinoid syndrome on bone. Methods: We studied 26 patients with carcinoid syndrome and 26 healthy controls, individually matched to carcinoid patients by gender, age, height and BMI. We measured BMD of the lumbar spine and hip with DXA, bone geometry density and microarchitecture of the distal radius and tibia with HR-pQCT. We measured 24 h urine 5HIAA (a serotonin metabolite) to assess circulating serotonin levels, and measured bone turnover with serum CTX, osteocalcin and PINP. We used paired samples t-tests to compare cases and controls. Results: Mean time since diagnosis of carcinoid disease was 5 years and 14 patients were currently treated with somatostatin analogues. Median 24 h urine 5HIAA was 105 μmol in carcinoid patients (range 27–1411) and 23 μmol in controls (range 11–49) (p<0.001). 24 carcinoid patients and 2 controls had 5HIAA above the upper limit of the reference range (37 μmol). Lumbar spine and total hip BMD did not differ between cases and controls (0.982 vs 1.028 g/cm2, 95 % CI of the difference −0.126 to 0.035, and 0.928 vs 0.981 g/cm2, 95 % CI of the difference −0.119 to 0.014). There were no differences between

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

cases and controls in any of the measures of bone density, geometry or microarchitecture at the radius or tibia. There were no differences between cases and controls in any of the three bone turnover markers (CTX 0.45 vs 0.41 ng/ml, 95 % CI of the difference −0.08 to 0.11; OC 25.4 vs 22.4 ng/ml, 95 % CI of the difference −2.3 to 8.5; PINP 48.8 vs 51.1 ng/ml, 95 % CI of the difference −13.5 to 9.1). Conclusion: We conclude that high circulating serotonin in carcinoid syndrome is not associated with reduced bone formation or alterations in bone density or structure. O30 STRONTIUM DIRECTLY INHIBITS MINERAL DEPOSITION IN BONE-FORMING PRIMARY OSTEOBLAST CULTURES Daniel Wornham, Mark Hajjawi, Timothy Arnett, University College London, London, UK Strontium ranelate (SrR) is now widely used for the prevention of osteoporotic (OP) fractures. The mechanisms by which this occurs, however, remain unclear. We investigated the actions of Sr2+ salts in long-term bone-forming cultures of primary rat osteoblasts. Primary osteoblasts were obtained from the calvariae of two day old Sprague–Dawley rats by sequential digestion with trypsin and collagenase. The cells were allowed to expand to confluence for 4–5 days in DMEM with 10 % FCS, then cultured in 24-well plates for a further 14 days in the same medium, supplemented with 10nM dexamethasone, 2 mM β-glycerophosphate and 50 μg/ml ascorbate, at an initial density of 25,000 cells / well. Cells were treated continuously with either SrR or SrCl2 in the range 100nM – 1 mM. Abundant, discretely mineralised ‘trabecular’ bone structures were evident in alizarin red-stained control cultures after 14 days. Surprisingly, SrR at 10, 100 and 1000 μM inhibited mineralisation, assessed morphometrically, to 75 %, 16 % (p<0.01) and 1 % (p<0.001) of control values, respectively; SrCl2 at the same concentrations caused 76 %, 47 % (p < 0.05) and 1 % (p < 0.001) inhibitions, respectively. Collagen deposition (assessed by Masson’s trichrome staining) and soluble collagen (measured using a sirius red assay) were unaffected by SrR or SrCl2 at any concentration. Osteoblast cell number (assessed by lactate dehydrogenase assay) or alkaline phosphatase activity were also unaltered. The selective inhibitory action of strontium salts on mineralisation was confirmed by inspection of unstained osteoblast cultures, revealing numerous unmineralised collagenous trabeculae. These data suggest that Sr2 + salts at high concentrations may exert a physicochemical inhibitory action on mineral deposition in humans treated with SrR. Maximum circulating Sr2+ concentrations reach about 20 μM in patients but local concentrations within bone could be higher. Whether our finding has any relevance to the in vivo situation remains to be determined.

S537

O31 ACCELERATION OF FRACTURE REPAIR WITH MUSCLE James Chan, Lorraine Harry, Graeme Glass, Nikki Horwood, Jagdeep Nanchahal; The Kennedy Institute of Rheumatology, London, UK Introduction: A significant proportion of fractures, particularly open injuries, are complicated by delayed or non-union. Using a murine fracture model, we compared the effects of muscle and fasciocutaneous tissue coverage in fracture repair. We aimed to identify the factors that lead to any differences between these two options in order to develop a translational strategy to accelerate fracture repair. Methods: Our model comprises periosteally-stripped, open tibial osteotomy with intramedullary fixation in C57Bl/6 mice. To compare the efficacy of fasciocutaneous versus muscle coverage of open fracture, inert PTFE was inserted to achieve exclusive coverage of 1/3 rd of cortical circumference at the fracture site by either fasciocutaneous tissue or muscle. Bones were tested mechanically by 4-point bending. The osteogenic potential of human stromal cells derived from different tissues, including muscle, bone marrow, fat and skin were compared using alkaline phosphatise (ALP) quantification and bone nodule staining. Results: The muscle group demonstrated accelerated repair and had a significantly higher load to failure (p<0.01). Muscle and bone marrow-derived stromal cells led to significantly greater ALP production and bone nodule formation than fat and skin-derived stromal cells (P<0.001). Discussion: Muscle is superior to fasciocutaneous tissue in both the rate and quality of fracture repair in open injuries. The muscle-derived stromal cells demonstrate a high osteogenic potential, making them an important source of osteoprogenitor cells in fracture repair and likely explains the accelerated healing in the muscle group. They may represent a therapeutic target cell population in bone regeneration. The greatest clinical need is in osteoporotic patients and future work is underway to investigate whether this paradigm pertains to fragility fractures. O32 EFFECTS OF AGE ON GENETIC INFLUENCE ON BONE DENSITY AND BONE LOSS OVER 17YEARS IN WOMEN: A LONGITUDINAL TWIN STUDY Alireza Moayyeri, Deborah Hart, Chris Hammond, Tim Spector; Department of Twin Research and Genetic Epidemiology, King's College London, London, UK Objectives: Rate of bone loss varies considerably between early post-menopausal and older women and different mechanisms may be involved in each of these periods. In

S538

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

this study we aimed to estimate the genetic component of bone loss (heritability) in hip and spine across different ages. Material and Methods: Starting in 1992 and during 17 years of follow-up in TwinsUK and Healthy Ageing Twin Study (HATS), 15,491 hip and lumbar spine DXA scans were performed in 7,056 twin volunteers. Out of these subjects, 2,810 female twins aged >35 years with at least two DXA scans with an interval period of >4 years were included in this analysis. We utilised a mixed-effects random-coefficients regression model to predict hip and spine BMD values for the included twins in exact ages of 40, 45, 50, …, to 80 years based on the available data in the close age ranges and with further adjustment for baseline age, weight, height and years of hormone replacement therapy (HRT) as fixed-effects covariates. We then estimated heritability of the changes in BMD measures between these age ranges (40 to 45, 45 to 50, …) using univariate variance component analysis. Results: Consistent with previous studies, heritability estimates for BMD were high and ranged from 63 % to 88 % at the ages of 80 and 40 years, respectively. Heritability of change of BMD was lower, generally ranging from 3050 %; between 40 and 45 years of age genetic factors explained 47.0 % (95 % CI: 29-60 %) of variance of BMD loss for total hip, 44.4 % (25-57 %) for femoral neck, and 68.1 % (56-76 %) for lumbar spine BMD loss. These estimates decreased with increasing age, and there appeared no heritability of changes after the age of 65 for total hip/ femoral neck BMD and after age of 70 for lumbar BMD. Conclusions: While genetic factors appear to have an important role in bone loss in early post-menopausal women, their effects weaken with age and completely disappear with advanced ageing. This is distinct from the behaviour of genetic component of bone mass in the elderly women and suggests involvement of different genetic mechanisms in the process of bone loss with ageing.

based on measurement of exposure to defined levels of impact using a Newtest device (Newtest Oy, Finland). Methods: Participants attending the ALSPAC research clinic underwent total body DXA and pQCT of the mid-tibia, and were subsequently invited to wear a Newtest accelerometer for seven days. Accelerometer results were partitioned into low (0.5-2.1 g), medium (2.1-4.1 g) and high (>4.1 g) impact activity. Results: 675 participants (272 boys) had valid accelerometer recordings and information on pQCT data and other covariates (mean age017.7 years). In our fully adjusted model (for age, height, fat mass and lean mass), moderate impact activity was positively related to periosteal circumference (PC) in boys but negatively related in girls [0.039 (95%CI −0.013,0.090.p 00.14) and −0.023 (−0.023,0.017.p00.26) respectively. p00.03 for gender interaction, coefficient0SD change per doubling in activity]. High impact activity showed a stronger positive association with PC in boys but no association was seen in girls [0.054 (0.007,0.100.p00.024) and 0.007 (−0.028,0.041. p00.707) respectively]. In further analyses additionally stratified by fat mass, an independent interaction was observed, such that the positive relationship between high impact activity and PC was greatest in those with highest fat mass [high impact versus PC in boys: 0.01 (−0.064, 0.085. p00.783), 0.045 (−0.040, 0.131.p00.298), 0.098 (0.012, 0.185.p00.027); high impact versus PC in girls: -0.041 (−0.101, 0.020,p00.187), -0.028 (−0.077,0.022.p 00.271), 0.082 (0.015,0.148. p 00.017) (p00.01 for fat mass interaction); lower, middle and upper fat-tertiles respectively]. Conclusions: Female gender and low body fat are associated with reduced periosteal expansion in response to high impact activity, via independent pathways. Our results suggest that skeletal response to high impacts is particularly impaired in girls with low body weight, which may partly explain the excess of stress fractures observed in this group after undergoing strenuous exercise.

O33 THE RESPONSE OF CORTICAL BONE TO HIGH IMPACT ACTIVITY IS ATTENUATED IN GIRLS: FINDINGS FROM A CROSS-SECTIONAL PQCT STUDY IN ADOLESCENTS Kevin Deere (1) presenting Adrian Sayers (1) Joern Rittweger (2) Jon Tobias (1) University of Bristol, Bristol, UK (1) Institute of Aerospace Medicine, Cologne, Germany (2)

O34 MUSCLE SIZE, STRENGTH AND PHYSICAL PERFORMANCE AS PREDICTORS OF BONE STRUCTURE IN THE HERTFORDSHIRE COHORT STUDY Mark Edwards, Karen Jameson, Celia Gregson, Nicholas Harvey, Avan Aihie Sayer, Elaine Dennison, Cyrus Cooper; MRC Lifecourse Epidemiology Unit, Southampton, Hampshire, UK

Background: The factors which govern skeletal responses to physical activity remain poorly understood, conceivably because accelerometers measuring physical activity are calibrated against energy expenditure rather than mechanical strain. We investigated whether a more precise understanding of these factors, including gender and fat mass, can be gained

Introduction: Sarcopenia is common in later life and has recently been clinically characterised in terms of both low muscle mass and reduced function (strength or performance) in guidance from the European Working Group on Sarcopenia in Older People. However, few studies have examined the

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

relative associations of these different components with bone mineral content (BMC), density, size and strength. In order to clarify these relationships, we studied participants from a prospective population cohort, the Hertfordshire Cohort Study. Material and Methods: 313 men and 318 women underwent baseline assessment of health (questionnaire) and detailed anthropometric measurements. Grip strength was measured using a Jamar hand-held dynamometer and gait speed was assessed by 3 metre walk test. Participants subsequently returned for peripheral quantitative computed tomography (pQCT) examination of the calf and forearm, performed using a Stratec 2000XL pQCT, to assess muscle crosssectional area (mCSA) (66 % level) and bone structure (radius and tibia, 4 % and 66 % levels). Discussion: The mean age (SD) at baseline was 65.0 (2.7) years. After adjustment for age, height, weight-adjusted-forheight, social class, smoker status, alcohol consumption, activity score, and dietary calcium, larger forearm mCSA was positively associated with distal radius TBA (β 0 0.330[0.206, 0.454]; β0unit change per SD increase in mCSA [95%CI]), radial BMC (β 00.421[0.313, 0.529]) and bone strength (strength strain index(β00.460[0.343, 0.576])) amongst women (all p<0.001). Similar strong associations were seen in men and between the calf mCSA and analogous tibia measures. No evidence of an association between mCSA and total or cortical bone density was detected. Higher grip strength was associated with a larger distal radial TBA and a tendency towards higher bone strength in both sexes. Gait speed did not show consistent associations with bone structure. Conclusion: Muscle size (muscle area) is strongly positively associated with BMC, bone size and strength, although not density. Muscle strength (grip strength) was also associated with bone structure however associations were weaker. Physical performance (gait speed) did not reliably predict bone structure. O35 HIGH BONE MASS IS ASSOCIATED WITH A GREATER PREVALENCE OF HIP BUT NOT KNEE REPLACEMENT Sarah A Hardcastle (1) presenting Celia L Gregson (1,2) Kevin Deere (1) George Davey-Smith (3) Jon H Tobias (1) MRC CAiTE Centre, University Musculoskeletal Research Unit, University of Bristol, Bristol, UK (1) MRC Lifecourse Epidemiology Unit, University of Southampton, Southampton, UK (2) MRC CAiTE Centre, University of Bristol, Bristol, UK (3) Introduction: Epidemiological studies have shown an inverse relationship between osteoarthritis (OA) and osteoporosis, implying that high bone mass may confer increased OA risk. We aimed to study clinical variables associated with OA in our

S539

High Bone Mass (HBM) population, recruited by systematically screening 335,115 DXA scans across England and Wales to identify index cases, and through them their family members. Material and Methods: HBM status in index cases was defined by DXA scan interpretation as (excluding known artefactual causes of elevated BMD measurement), an a) L1 Z-score of≥+3.2 and total hip Z-score≥+1.2 or b) total hip Z-score≥+3.2 and L1 Z-score≥+1.2. In relatives of known HBM index cases the definition of HBM was L1 Z-score plus total hip Z-score≥+3.2. Controls comprised unaffected relatives and spouses. Clinical indicators of OA (joint pain, joint replacement and NSAID use) were determined at structured interview; passive knee crepitus was assessed by doctor examination. Analyses used logistic regression adjusting for age, gender, smoking, alcohol use and BMI in Stata v11. Results: 339 HBM cases (mean age 61.7 years, 77 % female) and 188 controls (mean age 54.8 years, 46 % female) were included. Crude odds ratios suggested an increase in overall joint pain in HBM cases versus controls, which did not persist after adjustment. Adjusted NSAID use was more prevalent in HBM cases versus controls (OR 2.20 [1.09, 4.44], p00.03, 95 % CIs shown). The prevalence of hip replacement was 6.5 % (cases) and 1.1 % (controls) with adjusted OR 5.04 [1.11-22.84], p00.04. The prevalence of knee replacement was 6.8 % and 3.7 % in cases and controls respectively with adjusted OR 1.15 [0.44-2.99], p00.77. In participants who were examined (n0404), the prevalence of moderate/severe knee crepitus was no different between cases and controls (OR 1.16 [0.70-1.90], p00.57). Conclusion: HBM within this population was associated with increased prevalence of hip replacement surgery and NSAID use compared with unaffected family and spouse controls, suggesting a preferential effect of HBM on OA risk at the hip compared with the knee which warrants further exploration. O36 META-ANALYSIS OF GENOME-WIDE SCANS FOR TOTAL BODY BMD REVEALS AN INTERACTION WITH WEIGHT BEARING AT THE WNT16 LOCUS John P Kemp (1,2), Carolina Medina-Gomez (3,4,5,6), Karol Estrada (3,5,6), Joel Eriksson (7) , Jeff Liu (8), Sjur Reppe (9), David M Evans (1,2), Denise Heppe (4,5,10), Liesbeth VandenPut (7), Lizbeth Herrera (3), Susan M Ring (2), Claudia Kruithof (4,5), Nicholas J Timpson (1,2), M. Carola Zillikens (3,6), Ole K Olstad (9), Hou-Feng Zheng (11,12), Brent Richards (11,12,13), Beate St Pourcain (1), Albert Hofman (4,5,6), Vincent W Jaddoe (4,5,10), George Davey Smith (1,2), Mattias Lorentzon (7), Kaare M Gautvik (9,14), André G. Uitterlinden (3,4,5,6), Robert Brommage (8), Claes Ohlsson (7), Jonathan H Tobias (15),

S540

Fernando Rivadeneira (3,4,5,6); MRC CAiTE centre, School of Social and Community Medicine, University of Bristol, Bristol, UK (1); ALSPAC, School of Social & Community Medicine, University of Bristol, Oakfield House, Oakfield Grove, Bristol BS8 2BN, UK (2); Department of Internal Medicine, Erasmus University Medical Center, Rotterdam, The Netherlands (3); The Generation R Study Group, Erasmus University Medical Center, Rotterdam, The Netherlands (4); Department of Epidemiology, Erasmus University Medical Center, Rotterdam, The Netherlands (5); Netherlands Genomics Initiative (NGI)-sponsored Netherlands Consortium for Healthy Aging (NCHA), The Netherlands (6); Center for Bone and Arthritis Research, Institute of Medicine, Sahlgrenska Academy, University of Gothenburg, Sweden (7); Lexicon Pharmaceuticals, The Woodlands, Texas, USA (8); Department of Medical Biochemistry, Oslo University Hospital, Ullevaal, Oslo, Norway (9); Department of Pediatrics, Erasmus University Medical Center, Rotterdam, The Netherlands (10); Department of Medicine, Human Genetics, McGill University, Montreal, Quebec, Canada (11); Department of Epidemiology and Biostatistics, Lady Davis Institute for Medical Research, Jewish General Hospital, Montreal, Quebec, Canada (12); Twin Research and Genetic Epidemiology, King’s College London, London, United Kingdom (13); Department of Medical Biochemistry, Oslo Deacon Hospital, Oslo, Norway (14); School of Clinical Sciences, University of Bristol, Bristol, United Kingdom (15). Introduction: Bone Mineral Density (BMD) is a highly heritable trait used to assess skeletal health. It is well established that acquired peak bone mass influences the risk of osteoporosis. Material and Methods: To identify genetic loci influencing bone accrual we performed a genome-wide association scan for total-body bone mineral density (TB-BMD) variation in 2,660 children of different ethnicities. Discussion: We discovered variants in the region of the WNT16 gene associated with BMD measurements, with the lowest P04.1x10-11 observed for rs917727, which had a minor allele frequency of 0.37. We sought replication for all SNPs located +/−500 kb from rs917727 in 11,052 additional individuals from five independent studies including children and adults; together with de novo genotyping of rs3801387 (in perfect linkage disequilibrium (LD) with rs917727) in 1,014 mothers of children from the discovery cohort. The top signal above was replicated across studies with a meta analysis P02.4x10-31 and an effect size explaining between 0.61.8 % of TB-BMD variance. Conditional analyses on this signal revealed a secondary signal for total body BMD (P01.42x10-10) for rs4609139 within the same region, but upstream of the WNT16 gene in an open

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

reading frame (C7orf58). We also examined this genomic region for association with skull BMD to test if the associations were independent of skeletal loading. We identified two signals influencing skull BMD variation including rs917727 (P01.9x10-16) within WNT16, and a stronger signal, rs7801723 (P08.9x10-28), mapping to the upstream open reading frame C7orf58, in partial linkage with the secondary signal for total body BMD above (r200.50 with rs4609139). Conclusion: We detected two independent signals within the region of the WNT16 gene influencing total body and skull BMD variation in children and adults. The primary signal for total body BMD was within WNT16 while the secondary signal was within an upstream open reading frame, whereas the converse held for skull BMD, pointing to distinct genetic control at this locus of bone accrual at weight bearing and non-weight bearing sites. O37 CIRCULATING SCLEROSTIN AND DICKKOPF-1 IN PRE-DIALYSIS CHRONIC KIDNEY DISEASE : RELATIONSHIP WITH BONE DENSITY AND ARTERIAL STIFFNESS Rakshita Roplekar, Subashini Thambiah, Padmini Manghat, Ignac Fogelman, William Fraser, David Goldsmith, Geeta Hampson; Renal UDepartment of clinical chemistry, Guy's and St Thomas' Hospital, london, UK (1) Osteoporosis Clinic, Guy's Hospital, london, UK (2) Renal Unit, Guy's Hospital, london, UK (3) Nuclear Medicine, Guy's Hospital, Kings College london, london, UK (4) Norwich Medical School, University of East Anglia, Norwich, UK (5) Department of Pathology, university Putra, Selangor, Malaysia (6) Introduction: Abnormalities of bone metabolism and increased vascular calcification are common in chronic kidney disease (CKD) and are important causes of morbidity and mortality. Recent evidence over the last decade suggests a molecular link between the dysregulation in bone metabolism and the process of VC, although the mechanistic pathways involved are still incompletely understood. The Wnt signalling pathway may play a role in the bone and vascular disturbances seen in CKD, termed collectively CKD-MBD. The aim of the study was to investigate the possible association between circulating concentrations of the secreted Wnt signalling inhibitors; DKK1 and sclerostin with BMD and arterial stiffness in pre-dialysis CKD. Material and method: Seventy-seven patients (48 M, 29 F) aged (mean [SD]) 57 [14] years with CKD stages 3B (n032) and 4 (n045) were studied. Sclerostin, DKK1, PTH, 1, 25 (OH)2 vitamin D were analysed. Bone mineral density (BMD) was measured at the lumbar spine (LS), femoral

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

neck (FN), total hip (TH) and fore-arm (FARM). Arterial stiffness was determined by contour analysis of digital volume pulse (SI DVP). Discussion: There was a positive correlation between sclerostin and age ( r00.47, p <0.000). Sclerostin was higher in men than women (p 00.013). Following correction for age and gender, there was a negative association between GFR and sclerostin (p00.002).We observed a positive association between sclerostin and BMD at the LS (p00.0001), FN (p00.004), TH (p00.002). In contrast, DKK1 was negatively associated with BMD at the FN (p00.038). A negative association was seen between DKK1 and SI DVP (p00.027). Conclusion: Our findings suggest that the Wnt signalling pathway may be involved in the pathogenesis of CKDMBD. Further larger, prospective studies are needed to establish the clinical relevance of sclerostin and DKK1 as serological markers in CKD-MBD. O38 MATERNAL VITAMIN D LEVELS IN PREGNANCY AND OFFSPRING BONE MASS AT AGE 9: FINDINGS FROM A UK PROSPECTIVE BIRTH COHORT STUDY Andrew Wills (1) presenting Adrian Sayers (1) William Fraser (2) Jon Tobias (1) Debbie Lawlor (1) University of Bristol, Bristol, UK (1) University of East Anglia, Norwich, UK (2) There is a suggestion that higher levels of maternal vitamin D (as measured by dietary intake or circulating 25(OH)D) during pregnancy are related to higher offspring bone mineral density (BMD) in later childhood, and that the 3 rd trimester may be the important period of exposure. To address this question we examined whether, in the Avon Longitudinal Study of Parents and Children (ALSPAC), maternal 25(OH)D is associated with skeletal phenotypes measured by total body less head (TBLH) DXA at age 9.9 years. Methods: Data are from 2840 mother-offspring pairs from the ALSPAC study. 25(OH)D2 and 25(OH)D3 were measured by HPLC on serum obtained from blood samples routinely collected during clinic attendances throughout pregnancy. Total 25(OH)D was adjusted to the respective trimester midpoint (6, 20 & 34 weeks). Whole body DXA scans were carried out in the offspring (mean age: 9.9y). The association between total 25(OH)D and TBLH bone mineral density, content, area and bone area adjusted bone mineral content was assessed using multivariable linear regression. Results: There was no evidence for an association between maternal 25(OH)D and any of the offspring childhood DXA outcomes in minimally adjusted models (offspring age at DXA, sex, and maternal age at pregnancy) or a series of adjusted models (p>0.3 for all tests). There was no evidence of non-linearity (p>0.05) in these associations, or any evidence that deficient

S541

compared to replete 25(OH)D mothers had off spring with different bone outcomes. In addition we could find no evidence for an association after restricting analyses to those mothers with measures in the 3 rd trimester (p for trimester 25(OH)D interactions>0.05), or after accounting for seasonal variation in 25 (OH)D in an attempt to measure habitual vitamin D levels. Conclusion: Our study does not support the hypothesis that later childhood bone mass might be directly affected by prenatal exposure to varying levels of vitamin D. O39 VITAMIN D LEVELS IN HIP FRACTURES: RATIONALE AND GUIDELINES FOR RAPID SUBSTITUTION THERAPY Andy de Jong, Matthew Porteous; West Suffolk Hospital, Bury St. Edmunds, UK Introduction: Assessment for and treatment of osteoporosis is recommended following fracture, however all forms of osteoporosis treatment assume that those who receive it have an adequate calcium intake and are vitamin D replete. This study initially assessed vitamin D levels in patients with femoral neck fractures and from that we have developed guidelines on how to safely and effectively manage low vitamin D levels with high dose oral vitamin D3 (cholecalciferol). Materials and methods: Circulating 25-hydroxyvitamin D levels were measured in consecutive patients admitted with a hip fracture over an 18 month period. Substitution therapy with high dose oral vitamin D3 was started in a selected cohort of consecutive patients. Exclusion criteria for substitution were age<75 years, prior vitamin D substitution, corrected calcium>2.50 mg/dl. Discussion: 381 patients with 387 hip fractures were included. Only 27 patients had normal (>75 nmol/l) circulating vitamin D levels (mean 91.2 nmol/l, SD 20.0 nmol/l, range 75.6 – 171 nmol/l), and of these 22 were taking vitamin D supplements. The remainder, 353 patients, had abnormally low vitamin D levels, with a mean value of 26.4 nmol/l (SD017.9 nmol/l, range <10 - 74.4 nmol/ l). Substitution with 50.000 IU cholecalciferol daily for 3 days in 12 patients resulted in a rapid increase in circulating vitamin D levels from 27.6 nmol/l to 81.2 nmol/l (SD018.4 nmol/l, range 47.4 – 108.0 nmol/l, p<0.0001), at a mean of 13 days. Two-thirds of the patients achieved levels above the desired threshold of 75 nmol/l. No clinical or biochemical side effects were reported. Conclusion: This study shows that virtually all patients who are not taking vitamin D supplements sustaining a hip fracture have abnormally low circulating vitamin D levels and that they all require substitution before osteoporosis treatment can be initiated. This finding seems unaffected by the age of the patient. The routine measurement of vitamin D levels in patients with a corrected calcium <2.5 mg/dl would seem unnecessary. Substitution with 50.000 IU oral cholecalciferol daily for 3 days increases vitamin D levels fast, safe

S542

and consistently in these patients. We are now optimising our treatment algorithm for patients with corrected calcium above 2.50 mg/dl. O40 ACTIVE VITAMIN D (1,25-DIHYDROXYVITAMIN D) AND BONE HEALTH IN MIDDLE AGED AND OLDER MEN Stephen Pye (1) presenting Dirk Vanderschueren (2) Terence O'Neill (1) David Lee (1) Ivo Jans (2) Jaak Billen (2) Evelien Gielen (2) Frank Claessens (2) Judith Adams (3) Kate Ward (4) Gyorgy Bartfai (5) Felipe Casanueva (6) Joseph Finn (1) Gianni Forti (7) Aleksander Giwercman (8) Ilpo Huhtaniemi (9) Krzysztof Kula (10) Margus Punab (11) Frederick Wu (1) Steven Boonen (2) The University of Manchester, Manchester, UK (1) Katholieke Universiteit Leuven, Leuven, Belgium (2) Radiology and Manchester Academic Health Science Centre in The Royal Infirmary, Manchester, UK (3) MRC Human Nutrition Research, Cambridge, UK (4) Albert Szent-Gyorgy Medical University, Szeged, Hungary (5) Santiago de Compostela University, Santiago de Compostela, Spain (6) University of Florence, Florence, Italy (7) University of Lund, Malmo, Sweden (8) Imperial College London, London, UK (9) Medical University of Lodz, Lodz, Poland (10) United Laboratories of Tartu University Clinics, Tartu, Estonia (11) Introduction:1,25-dihydroxyvitamin D (1,25(OH)2D) is the active metabolite of vitamin D, however, little is known about its association with bone health. The aim of this study was to determine the influence of 1,25(OH)2D on bone health in middle-aged and older European men. Material and Methods: Men aged 40–79 years were recruited from population registers in eight European centres for participation in a prospective study of ageing: the European Male Ageing Study (EMAS). Subjects completed lifestyle and health questionnaires and were invited to attend for quantitative ultrasound (QUS) of the heel, assessment of height and weight, and a fasting blood sample. Dual energy x-ray absorptiometry (DXA) of the hip and lumbar spine was performed in two centres. 1,25(OH)2D was measured using liquid chromatography tandem mass spectrometry and 25(OH)D using radioimmunoassay. Parathyroid hormone (PTH) and the bone markers, serum N-terminal propeptide of type 1 procollagen (P1NP) and cross-links (β-cTX) were also measured. Associations between the bone and vitamin D parameters (all expressed as per standard deviation) were assessed using linear regression, with the bone parameters as the dependent variable and results expressed as β coefficients and 95 % confidence intervals (95%CI). Discussion: 2783 men, mean age 60.0 years (SD011.0) were included in the analysis. Mean 1,25(OH)2D was 59.3 pg/ml and mean 25(OH)D was 24.4 ng/ml. After adjustment for age

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

and centre, 1,25(OH)2D was positively associated with 25 (OH)D (β per SD 00.402; 95%CI00.368, 0.437) though not with PTH (β per SD0−0.034; 95%CI0−0.070, 0.003). After adjustment for age, centre, height, weight, lifestyle factors and season of measurement, 1,25(OH)2D was negatively associated with QUS (speed of sound) & DXA BMDa at the lumbar spine (β0−0.077; 95%CI0−0.116, -0.038 and β0−0.111; 95%CI0−0.209, -0.013 respectively) and positively associated with β-cTX (β00.162; 95%CI00.124, 0.201). 1,25(OH) 2D was not associated with P1NP (β00.017; 95%CI0−0.025, 0.058). 25(OH)D was positively associated with QUS & DXA parameters, but not related to either bone turnover marker. Conclusion: In this population sample of European men, serum 1,25(OH)2D was associated with increased bone turnover and lower bone mass. O41 ESTABLISHING REFERENCE INTERVALS FOR BONE TURNOVER MARKERS IN HEALTHY POSTMENOPAUSAL WOMEN Fatma Gossiel (1) presenting Judith Finigan (1) Richard Jacques (1) David Reid (2) Dieter Felsenberg (3) Christian Roux (4) Claus Glueer (5) Richard Eastell (1) University of Sheffield, Sheffield, UK (1) University of Aberdeen, Aberdeen, UK (2) Charite Universitatsmedizin, Berlin, Germany (3) Rene Descartes Universite, Paris, France (4) Universitaetsklinikum Schleswig-Holstein, Kiel, Germany (5) Introduction: There are several reference intervals for bone turnover markers (BTMs) in healthy younger women. There are fewer for older women, because they often suffer from diseases or take medications that affect bone turnover. We aimed to a) establish a healthy postmenopausal reference interval b) compare this to a healthy premenopausal reference interval and c) determine if the exclusion criteria are valid. Material and Methods: We studied postmenopausal women ages 55 to 80 years (n02422), from the OPUS study, a 5-European centre population-based study. We excluded women who: were osteoporotic (spine or hip T-score>−2.5), were vitamin D deficient (25 (OH) D<50 nmol/L), had low eGFR (<30 mL/minute/1.73 m2), had high serum calcium (>2.60 mmol/L) or suffered from any disease known to affect bone turnover. Women aged 30 to 40 years (n0205) were also recruited, and the same exclusions applied. We measured CTX, intact PINP and bone ALP using the IDS-iSYS automated immunoassays on non-fasting serum samples. Discussion: The reference intervals (table) were significantly higher in postmenopausal compared to premenopausal women (p < 0.001). Multiple linear regression analyses in all postmenopausal women showed that the 3 BTMs were all significantly higher if subjects were vitamin D deficient (n01208) or osteoporotic at spine or hip (n0127 and 339, respectively) (p≤0.01). CTX was higher if

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

S543

eGFR was low (n022). BTMs were not higher if calcium levels were high (n0181). Table: Geometric mean and 95 % reference intervals for bone turnover markers for healthy postmenopausal and premenopausal women. Bone turnover markers

Postmenopausal Women (n0343)

Premenopausal Women (n0 08)

Intact PINP ng/ml

40.5(16.0–102.0)

30.1(8.5–106.6)*

Bone ALP ug/L

14.1(7.3–27.2)

9.8(5.2–18.6)*

CTX ng/ml

0.31(0.10–0.98)

0.19(0.05–0.63)*

SA compared to E (−5.2 [−9.0, -1.5]) and B (−22.1, [−27.2, -17.1]). Cortical BMD was higher in SA compared to E (1.5 [0.3, 2.6]) but lower than B (−1.7 [−3.2, -0.1]). SA had thinner cortices than B (−21.1 [−26.3, -15.9]) and E groups (E −7.21 [−11.0, -3.4]). Strength was lower in SA compared to B (−19.2 [−25.4, -12.6]) but not significantly different to E. Compared to B, E had smaller bones (−6.1 [−10.2, -2.0]), thinner cortices (−13.7 [−18.0, -9.5]), lower BMC (−16.8, [−21.0, -12.7]), BMD (−3.1 [−4.4, -1.8]) and strength (−20.9 [−26.4, -15.5]). Conclusion: There are ethnic differences in male bone phenotype. Future work includes understanding how environment relates to these differences and how different phenotypes relate to fracture risk.

*P<0.001 Independent sample T-test

Conclusion: Older women have higher BTMs than younger women. Many factors can affect BTMs, particularly in the presence of osteoporosis, vitamin D deficiency and chronic kidney disease. O42 ETHNICITY AND BONE PHENOTYPE IN UK MEN Kate Ward (1,2) presenting Judith Adams (2,3) Stephen Pye (2) Joseph Finn (2) Frederick Wu (2) Terence O'Neill (2) MRC Human Nutrition Research, Cambridge, UK (1) University of Manchester, Manchester, UK (2) The Royal Infirmary, Manchester, UK (3) Introduction: Understanding how bone phenotype differs between ethnic groups provides potentially important insights into how differences in fracture risk may occur. This cross sectional study aimed to characterise volumetric BMD, bone geometry and strength in UK men from different ethnic groups. Materials and methods: 40 to 80 year-old, community dwelling men (339) of European origin (E) were recruited from south Manchester for participation in a study of male ageing (European Male Aging Study). A further sample of black (B, 44) and south Asian (SA, 64) men were recruited from Manchester and completed the same study protocol. All subjects completed an interviewer assisted questionnaire and had assessment of height and weight. Peripheral quantitative computed tomography (Stratec) was performed at the 4 % and 50 % site of the non-dominant arm. Group differences in skeletal parameters were tested using linear regression adjusting for age, height and weight. Results are presented as beta co-efficient and 95 % confidence intervals (%) [95 % CI]. Discussion: Mean (SD) age of the three groups were E- 60.4 (11.0), SA- 59.1 (11.0) and B- 59.7 (11.3) years. At the 4 % radius there were no group differences in total or trabecular BMD. SA had smaller bones than the other two groups (E −6.5 [−10.5, -2.6]; B −12.8 [−18.2, -7.3]); E smaller than B (−6.0 [−10.5, -1.5]). At the 50 % radius there were no differences in size between E and SA (2.46 [−1.2, 6.2]). BMC was lower in

O43 DIFFERENTIAL EFFECTS OF TERIPARATIDE ON TRABECULAR AND CORTICAL BONE MEASURED BY 18 F-FLUORIDE POSITRON EMISSION TOMOGRAPHY Michelle Frost (1) presenting Musib Siddique (1) Glen Blake (1) Amelia Moore (1) Paul Schleyer (1) Paul Marsden (1) Richard Eastell (2) Ignac Fogelman (1) King's College London, London, UK (1) University of Sheffield, Sheffield, UK (2) Introduction:The functional imaging technique of 18 F-fluoride positron emission tomography (18 F-PET) allows the non-invasive assessment of regional bone metabolism at clinically important sites of the skeleton, which is not possible using conventional techniques. The aim of this 6-month prospective study was to examine the effects of teriparatide on bone metabolism at both trabecular and cortical sites including the spine, pelvis and hips. Methods: 12 treatment naive postmenopausal women with osteoporosis (mean age 64 years) had a dynamic 18 F-PET/CT scan of the lumbar spine and a static scan of the pelvis and hips at baseline and after 6-months treatment with teriparatide (20 μg/ day). The plasma clearance of 18 F- to bone tissue, termed Ki, reflecting the rate of bone metabolism was estimated using the Patlak method at the lumbar spine and a novel a two-point Patlak method was used to estimate Ki at the pelvis and 3 regions within the hip. Ki was measured in an elliptical ROI of purely trabecular bone within each vertebral body, a 6 cm section of purely cortical bone in the femoral shaft, areas equivalent to the DXA total hip and femoral neck ROIs, and the pelvis. Data for the right and left hips were averaged. Subjects also had DXA measurements at baseline and 6-months and biochemical markers were assessed at multiple time points during the study. Discussion: After 6 months treatment with teriparatide Ki increased by 23 % (p00.004) at the trabecular ROI within the lumbar spine. Ki at the femoral shaft increased by 83 % (p00.001), at the total hip by 38 % (p00.026), at the femoral neck by 33 % (p00.033) and at the pelvis by 37 % (p00.029). Changes in bone turnover markers and BMD were consistent with previous trials of teriparatide.

S544

Conclusion: This study demonstrates that changes in bone metabolism in response to teriparatide differ at different skeletal sites, with larger increases observed in cortical bone than in trabecular bone. 18 F-PET may provide a useful biomarker tool for assessing the differential response of the skeleton to novel treatments being developed for osteoporosis. O44 HIP FRACTURE PREDICTION USING ACTIVE SHAPE (ASM) AND APPEARANCE (AAM) MODELLING S Goodyear (1), R Barr (1), R Aspden (1), D Reid (1), I Reid (2), J Gregory (1) University of Aberdeen, Aberdeen, UK (1), University of Auckland, Auckland, New Zealand (2) Introduction:Hip fracture is a major cause of mortality and a significant cost to the economies of western nations. Hip fracture prediction is necessary to best match expensive treatments to the neediest individuals. BMD measurement from DXA imaging is often used to predict who is at greatest risk of a hip fracture but only extracts a single variable from each image. We set out to determine if the additional information ASM and AAM can extract from DXA images could be useful in predicting hip fracture. Material and Methods: Hip fracture cases and random controls (3 per fx) were identified from 3 retrospective studies: Auckland calcium (21 fractures), OPUS (21) and NOSOS (59) studies, over 5, 10 and 10 years follow-up respectively. A template of 83 points was used to record the coordinates of points outlining the proximal femur, acetabulum and parts of the pelvis. Baseline DXA images for each individual were analysed. The variation in point position and grey-level in the images was recorded using ASM and AAM, and shape and appearance mode scores calculated for each image. The mode scores and age, height, weight and BMD were analysed by logistic regression to determine which variables were useful in predicting hip fracture alone and in combination with other variables. Results: Shape modes 6 (0.79 (OR), 0.63-1.00 (CI), 0.047 (P)) and 9 (0.77, 0.61-0.97, 0.025) and appearance mode 10 (1.53, 1.22-1.93, <0.001) were significant predictors of fracture. Shape mode 6 remained a significant predictor of hip fracture after adjustment for BMD (0.70, 0.55-0.90, 0.005) and age and BMD (0.71, 0.56-0.92, 0.009). Additionally, shape mode 1 (1.37, 1.07-1.75, 0.013 and 1.30, 1.01-1.67, 0.044) and appearance mode 9 (0.74, 0.58-0.96, 0.023 and 0.76, 0.59-0.98, 0.036) became significant predictors after similar adjustment. Conclusion: ASM and AAM can generate modes that are good predictors of hip fracture, both alone and after adjustment for other risk factors.

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

O45 EPIDEMIOLOGY OF ATYPICAL FEMORAL FRACTURES IN THE BRISTOL POPULATION Cecilia Mercieca (1,2) presenting Virginia Gould (1) Jon Tobias (1,2) Ashley Blom (1,2) Emma Clark (1,2) Academic Rheumatology; Musculoskeletal Research Unit, University of Bristol, Bristol, UK (1) North Bristol NHS Trust, Bristol, UK (2) Introduction: Reports suggest a link between long term bisphosphonate (BP) use and atypical femoral fractures (AFF). A causal effect has not been proven and cases in non-BP users have been reported. Population-based studies are sparse and lack specific clinical and radiographic data, and consequently the incidence of AFF in the general population is unknown. Recently the ASBMR has developed a case definition for AFF, and we wished to apply this to a large population of hip fractures from the Bristol area of the UK to accurately identify the prevalence of AFF. Materials and Methods: A population based radiographic database was systematically searched for all hip and femur fractures occurring at a single hospital between 2007–2009. Fractures were classified as neck, trochanteric, subtrochanteric or shaft, and periprosthetic. Subtrochanteric and shaft fractures were classified as ‘atypical’ using the ASBMR definition radiographic features (non-comminuted, transverse or short oblique, partial or complete, located distal to lesser trochanter and proximal to supracondylar flare) or ‘typical’. Discussion: 1376 fractures in 1320 individuals were identified. Under 50s were excluded from further analysis as high trauma was reported in 89 % of their subtrochanteric/shaft fractures, compared to 6 % in the over-50’s. 1296 fractures occurred in 1246 patients (female 929: male 317) aged over 50 years. Of these, 57.0 % were neck, 32.2 % trochanteric, 3.2 % subtrochanteric and 3.1 % shaft, while 4.5 % were periprosthetic. The mean age at fracture was 82.3 (SD 9.8) years. Of the 83 shaft/ subtrochanteric fractures 7 (8.4 %) were classified as definite AFF. Ambiguity with the definition of shortoblique meant a further 4 (4.8 %) were probable AFF. Two of the seven definite AFF’s occurred in people with Paget's disease of the femur. Conclusion: This is the first study applying the ASBMR definition of AFF to a population-based cohort of hip fractures. The incidence of AFF that we found (13.2 % of subtrochanteric/shaft fractures) was comparable with other studies that used slightly different definitions of AFF. Further clinical analysis will provide better understanding about any association between BP and AFF.

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

O46 OUTCOME FOLLOWING HIP FRACTURE: LONG TERM MORTALITY AND POST-DISCHARGE RESIDENCE Antony Johansen, Maizura Mansor, Sue Beck, Heather Mahoney, Suzanne Thomas; University Hospital of Wales, Cardiff, UK Introduction: We set out to determine the pattern of mortality in a five year follow-up study of a cohort of over 1,000 patients admitted with hip fracture, and to examine how this was affected by loss of independence sufficient to require placement in institutional care following discharge. Methods: This is a prospective study of 1,050 patients (sequential admissions with hip fracture to the Cardiff Trauma Unit; between 1st January 2003 and 31st December 2004), using the Cardiff Hip Fracture Database to collect baseline demographic details of all admissions, along with admission clinical assessment data, information regarding progress care, and details of their discharge destination. In addition, in January 2009 we determines which patients were still alive, and defined their residential status at a mean of 5 years after their original admission. Exclusion of patients from outside Cardiff and the Vale of Glamorgan left 916 subjects. Trust IT system, local GP mortality audit data, and a postal questionnaire to patients, left only 48 patients (5.2 %) in whom we were unable to ascertain survival or mortality at follow-up. Results: We present data on these 869 patients (94.8 %). By the end of the study period 577 patients (66.4 %) had died and 292 (33.6 %) were alive. Survival probability was analysed using the Kaplan-Meier method - the median survival time is 2.83 years (95 % CI: 2.47, 3.2). A total of 139 patients (16 %) died in hospital, many being frailer older individuals who had originally been living in institutional care. 599 (69 %) of patients were admitted from their own home and only 448 (74.8 %) of these returned there on discharge. Patients who returned home had a median survival of 5.25 years (63 months), while those who needed new placement in institutional care had a median survival of only 1.33 years (16 months). Discussion: We have showed that mean survival was higher for those needing residential home placement (30.4 months) than for those needing nursing home care (19.6 months). For 2007–2008 the estimated weekly cost of care in a UK was £502 for a private residential home, and £712 for a nursing home. Combining these figures we are able to estimate the average lifetime cost of care home placement. This proved to be similar in both settings (£61,323 for residential care and £66,377 for nursing care). Thus, 70 % of hip fracture patients are admitted from their own home, and 10 % of these will need care home placement on discharge. Each such placement carried an estimated life-time cost of around £63,940. Such estimates might prove useful when trying to convince

S545

commissioners to invest in the multidisciplinary orthogeriatric care that has been shown to avoid such outcomes. O47 INVESTIGATING NOVEL REGULATORS AND INHIBITORS OF AORTIC VALVE CALCIFICATION Neil Mackenzie (1) presenting Dongxing Zhu (1) Daniel Lerman (1,2) Mark Dweck (2) David Newby (2) Vicky Macrae (1) The Roslin Institute, Edinburgh, UK (1) Centre for Cardiovascular Science, Edinburgh, UK (2) Introduction: Calcific aortic valve stenosis (CAVS) is the leading indication for aortic valve replacement surgery. Calcification of the valve occurs following trans-differentiation of the valve interstitial cells (VICs) through a myofibroblast stage into osteoblast cells. Here we investigate the effect of Denosumab, a monoclonal antibody with high affinity to Receptor Activator of Nuclear factor K-B Ligand (RANKL), on the calcification of VICs. Methods: Porcine VICs were extracted from the aortic valve leaflets by serial collagenase digestion and used as a model to characterise gene expression changes during calcification. Various treatments were used to induce calcification of the cells. Calcium deposition and alkaline phosphatase (ALP) activity was determined by Alizarin red staining and kinetic assay respectively. Results: Initial studies demonstrated that porcine VICs calcify spontaneously with a significant increase in calcium deposition (377 %; p<0.001) and ALP activity (3.7-fold; p<0.001) by day 14. mRNA expression of the osteoblast markers Runx2 (1.3 fold; P<0.05) and TGF-β (3.2 fold; P<0.001) were also increased at day 14 with similar increases seen in a number of myofibroblast markers including α-actin (1.7 fold; P<0.05) and RhoA (4.6 fold; P<0.001). Treatment of porcine VICs with Na3PO4 (3 mM) led to a marked increase in calcium deposition (535 %; P<0.05) and ALP activity (228 %; P< 0.05) by day 7. Denosumab (50 ng/mL) dramatically inhibited this Na3PO4 -induced calcification to baseline levels (P< 0.01) but had no significant effect on ALP activity when compared to Na3PO4 treated cells. Significant increases in calcium deposition were observed by day 14 following induction of VICs with both Na3PO4 (3 mM) and TGFβ (10 ng/mL) (700 %; p<0.05), which was also reduced to baseline following treatment with Denosumab (P<0.01). No difference in VIC growth rate was observed following treatment with Denosumab over a 6-day culture period, indicating no significant toxicity. Conclusion: These studies demonstrate the induction of VIC calcification in vitro by several treatments, and show for the first time that Denosumab inhibits this pathological process. A fuller understanding of the actions of Denosumab may identify a novel therapeutic strategy for clinical intervention against CAVS.

S546

O48 RAB27 KNOCKOUT MICE HAVE LOW BONE DENSITY DUE TO DEFECTIVE OSTEOBLAST FUNCTION Fraser Coxon (1) presenting Angela Douglass (1) Alun Hughes (1) Miguel Seabra (2) Tanya Tolmachova (2) University of Aberdeen, Aberdeen, UK (1) Imperial College, London, UK (2) The Rab family of small GTPases regulate intracellular vesicular transport pathways. Within this family, Rab27a and Rab27b play an important role in the trafficking and secretion of lysosome-related organelles in specialised cells, including melanocytes and cytotoxic T cells. Since secretory lysosomes, also considered a type of lysosome-related organelle, are important for both osteoclast and osteoblast function, we hypothesised that Rab27 may play an important role in bone physiology. To address this question, we have analysed Rab27 expression in osteoclasts and osteoblasts, and analysed the bone phenotype as well as bone cells derived from mice lacking both Rab27a and Rab27b (Rab27 KO). Rab27a was detected in osteoclast precursors, but expression markedly decreased during osteoclast differentiation in vitro. By contrast, Rab27a was detected in mouse calvarial osteoblasts, and increased during differentiation with ascorbic acid, and βglycerophosphate. Rab27b could not be detected in either cell type. Micro CT analysis of the tibiae of 8week old mice demonstrated that Rab27 KO mice had 25 % lower trabecular bone volume than WT mice. Given the expression of Rab27a in osteoblasts, we reasoned that this was due to impaired bone formation, rather than increased osteoclast activity. In support of the latter, there were no significant differences in the formation of TRAP-positive osteoclasts from bone marrow cells of WT or KO mice, or in the ability of these cells to resorb dentine in vitro. Although the alkaline phosphatase activity of calvarial osteoblasts isolated from KO mice was similar to WT mice, the ability of differentiated osteoblasts to mineralise, assessed by alizarin red staining, was slightly reduced. There were no differences in the ability of osteoblasts isolated from the WT or KO mice to stimulate osteoclast formation or resorption in vitro in co-cultures with WT mouse bone marrow cells, indicating that RANKL trafficking is unaffected in the KO mice. This contrasts with a recent study in mice lacking Munc13-4, an effector of Rab27, in which osteoclast formation was reduced due to impaired RANKL transport to the plasma membrane. Therefore, it appears that Rab27a is important for bone formation by osteoblasts, but not for stimulation of osteoclastogenesis.

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

O49 TNF-α: A THERAPEUTIC TARGET IN BONE REPAIR James Chan (1) presenting Graeme Glass (1) Nikki Horwood (1) Jagdeep Nanchahal (1) The Kennedy Institute of Rheumatology, London, UK (1) Introduction: Tissue injury, including fractures, leads to an acute inflammatory response. TNF-α, a major inflammatory cytokine, has been found to be important in the early phase of fracture repair but there are conflicting reports regarding its effects on osteogenesis. We sought to define the role of TNF-α in bone fracture repair. Method: The fracture cytokine environment was simulated in vitro using supernatants of fractured human tibial fragments. Muscle-derived stromal cells (MDSCs) were obtained from human subjects. 1. The role of TNF-α in osteogenesis in vitro was determined using MDSCs cultured in (i) fracture supernatant+neutralising antibody to TNF-α (AbTNF-α), and (ii) human serum containing medium (HSM)+rhTNF-α; osteogenesis was assessed by ALP production and bone nodule formation. 2. The role of TNF-α in chemotaxis: MDSC migration through an 8 μm pore transwell membrane in response to (i) fracture supernatant+AbTNF-α, and (ii) HSM+rhTNF-α, was measured. 3. Effect of TNF-α in vivo: 20μL of rhTNF-α at 50 ng/ml was injected at the fracture site on days 0 and 1. Fracture repair was assessed using histology and microCT. Results 1. AbTNF-α inhibited the osteogenic activity of fracture supernatant in a dose-dependent manner. rhTNF-α promoted osteogenic differentiation of MDSC (optimal dose 1 ng/ml). 2. Human fracture supernatants promoted MDSC migration; this effect was inhibited by AbTNF-α in a dose dependent manner. rhTNF-α induced MDSC migration (optimal dose at 1 pg/ml). 3. Local rhTNF-α injection accelerated fracture repair as indicated by resolution of the cartilaginous intermediary and remodelling of the mineralized callus on histology, and significantly greater percentage callus mineralization (p0 0.002) on microCT. Discussion: TNF-α enhances osteogenesis and migration of MDSCs. Manipulation of the local inflammatory environment is an effective way to enhance fracture repair and bone regeneration. The greatest unmet clinical need is in osteoporotic patients therefore we are currently studying the effects of exogenous TNF-α in fragility fractures. O50 IGF-2 PROMOTES BMP-9-INDUCED VASCULAR CALCIFICATION Dongxing Zhu (1) presenting, Neil Mackenzie (1), Jose Luis Millan (2), Colin Farquharson (1), Vicky Macrae (1), The

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

Roslin Institute, Edinburgh, UK (1) The Sanford-Burnham Research Institute, California, USA (2) The process of vascular calcification shares many similarities with that of skeletal mineralisation. However, the cellular mechanisms responsible have yet to be fully elucidated. IGF-2 has been shown to exert direct effects on both bone development and vascular function. In the present study we have investigated the role of IGF2 in vascular smooth muscle cell (VSMC) calcification. Murine VSMCs were cultured in calcifying medium containing Na2HPO4/NaH2PO4 for 14d. Calcium deposition was confirmed by alizarin red staining. Calcified VSMCs showed increased Runx2, Akp2 and PiT-1 mRNA expression, which are recognised markers of vascular calcification. mRNA expression of IGF-2 was significantly up-regulated by 7d (1.5 fold; P<0.05), with a concomitant increase in protein expression. Increased IGF-2 protein expression was also observed in the calcified media of Enpp1−/− mouse aortic tissue. Cell signalling studies in VSMCs revealed no effect of IGF-2 (50 ng/ml) on Smad signalling, whereas IGF-2 stimulated the phosphorylation of Akt and P44/42 MAP kinase. However, treatment with IGF-2 (1-100 ng/ml) or chromeceptin (0-3 μm) (an inhibitor of IGF-2 signalling) had no effect on VSMC calcification. Further studies were undertaken to examine whether IGF-2 was able to promote BMP-9 induced VSMC calcification, as this mechanism has recently been demonstrated in osteoblasts (bone forming cells). BMP-9 mRNA expression was significantly up-regulated by 7d (1.4 fold; P<0.05) in calcified VSMCs, and BMP-9 (50 ng/ml) treatment caused a small, significant increase in VSMC calcium content (1.4 fold; P<0.05). However, a marked increase in VSMC calcium content (9.4 fold; P<0.05) was induced by the combined treatment of IGF-2 (50 ng/ ml) and BMP-9 (50 ng/ml). This data suggests that IGF-2 acts through a novel mechanism in VSMCs to promote BMP-9 induced calcification. IGF-2 signalling can be negatively regulated by IGF binding proteins (IGFBPs). Calcified VSMCs showed significantly increased expression of IGFBP-1 (1.6 fold; P<0.05), IGFBP-3 (2.5 fold; P<0.05) and IGFBP-4 (2.1 fold; P<0.05) by d14. Future studies will determine whether these IGFBPs can antagonise the novel effect of IGF-2 on BMP-9induced VSMC calcification that we have reported. This may identify new potential therapeutic strategies for clinical intervention. O51 EXPRESSION OF 11β-HYDROXYSTEROID DEHYDROGENASE ENZYMES IN HUMAN OSTEOSARCOMA: POTENTIAL ROLE IN PATHOGENESIS AND AS TARGETS FOR TREATMENTS Pushpa Patel (1) presenting Rowan Hardy (1) Sumathi Vaiyapuri (2) Paul Stewart (1) Elizabeth Rabbitt (1) Neil Gittoes (1) Mark Cooper (1) University of Birmingham, Birmingham, UK (1) Royal Orthopaedic Hospital, Birmingham, UK (2)

S547

Introduction: Osteosarcoma (OS) is a malignant tumour of bone occurring predominantly in children & young adults. Despite chemotherapy relapse is common & mortality remains ~50 %. Non-transformed osteoblasts are highly sensitive to glucocorticoids (GCs) which reduce proliferation & induce apoptosis. Previously we observed that OS cells, but not normal osteoblasts, express 11β-hydroxysteroid dehydrogenase type 2 (11β-HSD2). This enzyme powerfully inactivates cortisol (active) to cortisone (inactive) & expression renders OS cells resistant to GCs. Some synthetic GCs (e.g. dehydrodexamethasone (DHD), inactive counterpart of dexamethasone (DEX)) have however been reported to be activated by 11β-HSD2. We therefore investigated the expression & activity of 11β-HSD enzymes in human OS tissue, determined if 11β-HSD expression has prognostic value in the response to therapy, & evaluated the potential utility of synthetic GCs to selectively target OS cells. Methods: We examined enzyme expression in OS tissue before (N09) & after (N021) chemotherapy using immunohistochemistry & real-time PCR. Controls included benign osteoblastic bone lesions. Enzyme activity studies were carried out on tissue and cells using radiolabelled GCs & thin layer chromatography. Results: OS samples expressed both 11β-HSD1 & 11β-HSD2. Expression of the 11β-HSD2 expression was much higher in OS tissue than in benign bone lesions (e.g. osteoblastoma). Expression & activity of 11β-HSD1 was unrelated to the degree of tumour necrosis following chemotherapy. However, high 11β-HSD2 expression correlated with a poor response to therapy (r0−0.48 for correlation of enzyme activity and % necrosis; p<0.05). OS cells that expressed 11β-HSD2 powerfully inactivated endogenous GCs & could not reactivate their inactive equivalents. However, these cells generated significant amounts of active DEX from inactive DHD (activity 15+ 2pmol/mg/hour). Conclusions: These results suggest that OS treatment response is related to 11β-HSD2 enzyme expression. Furthermore, OS cells expressing this enzyme could be targeted by treatment with synthetic GCs that are selectively reactivated by the enzyme. O52 AGE-RELATED STRUCTURAL CHANGES IN THE INTERTROCHANTER AND TROCHANTER OF THE HIP Benjamin Gabbott (1) presenting Ken Poole (2) Richard Eastell (1) Lang Yang (1) Academic Unit of Bone Metabolism, Sheffield, UK (1) Dept. of Medicine and Dept of Radiology, Cambridge, UK (2) Introduction:Resistance to hip fracture may depend on anatomic distribution of cortical / trabecular bone. Age-related bone loss at the femoral neck has been studied extensively. There is paucity on data at the intertrochanter and

S548

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

trochanter, and this study aims to quantify the age-related structural changes within these regions. Method: We studied quantitative computed tomography (QCT) scans from 100 healthy female volunteers aged 20– 90 from the Cambridge MRC study. The participants had QCT scans for unrelated non skeletal diseases. We quantified cortical, trabecular and integral vBMD and cortical thickness in quadrants of cross sections along the proximal intertrochanter (pIT), distal intertrochanter (dIT), midtrochanter (mTR) and lesser trochanter (lTR). Linear regression with age was performed. Discussion: There was a negative linear relationship between cortical thickness and age across all regions (p<0.05). The

steepest gradient of −0.13 mm thickness per decade (95 % CI −0.17, -0.08) was observed in the lTR cross-section (fig1 far right, ‘medioposterior quadrant’). Cortical and trabecular vBMDs in all lateroanterior and medioposterior quadrants showed strong age effects. The steepest gradient for cortical vBMD occurred at the medioposterior quandrant of the dIT at −28 mg/cm3 (95 % CI 35, 20) and for trabecular vBMD was within the medioposterior quardrant of the mTR, 21 mg/cm3 (95 % CI 25–18) per decade. Conclusion: Changes in bone structure of the trochanteric region may provide a basis to a better understanding of the increased risk of intertrochanteric fractures with ageing.

Poster Presentations

were created, adjusting for gestation, and conditional regression modelling yielded mutually independent growth variables for each time period. Proximal femoral size, density and geometry at 6 years were assessed using DXA (Hologic Discovery, Hologic Inc., Bedford, MA, USA) and hip structure analysis software. Discussion: After adjustment for child’s age, sex and milk intake, linear growth at all time intervals was positively related to total hip bone area and bone mineral content, and to femoral neck cross-sectional area and cross-sectional moment of inertia (CSMI). The strongest associations were with growth in late gestation (19–34 week linear growth and CSMI at 6 years: beta00.26 cm4/sd, p< 0.0001 versus 11–19 week linear growth and CSMI: beta00.11 cm4/sd, p00.06), and in the first two years of postnatal life (0–1 year linear growth and CSMI: beta 00.28 cm4/sd, p < 0.0001; 1–2 years: beta 0 0.34 cm4/sd, p<0.0001), with progressively weaker relationships over years 3 (beta00.23 cm4/sd, p00.0001) and 4 (beta00.11 cm4/sd, p00.14). Associations were similar when boys and girls were analysed separately. Conclusions: We have demonstrated that early growth predicts proximal femoral size, mineralisation, geometry and strength at six years old. The relationships are particularly strong in late pregnancy and in the first two years of postnatal life suggesting that these might be critical periods in which there is capacity for long term influence on the later skeletal growth trajectory.

P1 GROWTH IN UTERO AND EARLY CHILDHOOD PREDICTS HIP GEOMETRY AT SIX YEARS OLD Nicholas Harvey (1), Zoe Cole (1), Sarah Crozier (1), Georgia Ntani (1), Pam Mahon (1), Sian Robinson (1), Hazel Inskip (1), Keith Godfrey (1,2), Elaine Dennison (1), Cyrus Cooper (1); MRC Lifecourse Epidemiology Unit, Southampton, UK (1); NIHR Biomedical Research Unit in Diet, Lifestyle and Nutrition, Southampton, UK (2) Background: We have demonstrated that poor early growth predicts risk of hip fracture in older adulthood, and that this might be mediated through altered proximal femoral geometry. It is not clear whether changes in hip structure might already be present in childhood. We investigated the relationships between early growth and proximal femoral geometry at six years old in a prospective population-based cohort, the Southampton Women's Survey (SWS). Materials and methods: 493 mother-offspring pairs were recruited from the SWS. Fetal linear size at 11 weeks (crown-rump length) and at 19 and 34 weeks (femur length) was assessed using high resolution ultrasound; postnatal linear growth (length or height) was measured at birth, 6, 12, 24, 36 and 48 months. Size Z-scores

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

P2 COMPARISON OF QUANTITATIVE ULTRASOUND OF THE OS CALCIS AND STANDARD RADIOLOGICAL METHODS FOR MONITORING BONE MINERAL DENSITY CHANGES 18MONTHS POSTPARTUM William WK To (1) presenting Margaret WN Wong (2) United Christian Hospital, Hong Kong, Hong Kong (1) Prince of Wales Hospital, Hong Kong, Hong Kong (2) Objective: To evaluate the bone mineral density (BMD) changes within 18 months after delivery using standard dual energy X-ray absorptiometry (DXA), peripheral quantitative computerized tomography (pQCT) and quantitative ultrasound (QUS) of the os calcis Materials and Methods: Consecutive patients with uncomplicated singleton pregnancies were recruited from a general obstetric clinic over a period of 6 months. Standard BMD measurements of the lumbar spine and hips were performed using DXA and volumetric BMD of the distal radius and distal tibia were also performed using pQCT techniques within 3 weeks after the index delivery. The same measurements were then repeated between 15 to 18 months after the index delivery. Those that were born preterm before 36 weeks or who had significant medical problems antenatally or postnatally were excluded. Discussion: A total of 99 patients with complete data were analyzed. Uniformly significant gains in BMD were observed between the two assessments in the lumbar spine (L2-L4), neck of femur, as well as core and total BMD of the distal radius and tibia. The mean percentage increase in BMD at the different sites ranged from 1.15 to 7.34 % of the early postpartum value. The gain in general appeared to be higher in areal BMD values as measured by DXA compared to volumetric BMD values as measured by pQCT. QUS also showed a significant gain in the derived BMD value of 5.3 % in the os calcis. There was fair correlation between the measured BMD gain from the two techniques (correlation coefficient between QUS and spine 0.28, p 00.07) Conclusion: Significant gains in BMD at various bone sites were observable within 18 months after delivery, using both radiological methods (DXA and pQCT) for areal and volumetric BMD measurement as well as QUS of the os calcis with good agreement between the methods. The uniform gain in BMD at all sites demonstrated clearly the recovery of BMD loss from pregnancy. Quantitative ultrasound can be a valid and inexpensive means of monitoring BMD changes after delivery.

S549

P3 COMPARISON OF PATIENT POPULATIONS IDENTIFIED USING DIFFERENT FRAX THRESHOLDS: APPLICATION OF THE 20 % MAJOR AND 5 % HIP THRESHOLD IMPROVES TARGETING OF TREATMENT TO THOSE AT HIGH RISK Louise Carter (1) presenting Jon Tobias (2,1) North Bristol NHS Trust, Bristol, UK (1)Academic Rheumatology, University of Bristol, Bristol, UK (2) Background and aims: How to incorporate FRAX into DXA reporting strategies remains unclear. We compared the characteristics of patients selected for treatment by FRAX using three different fracture risk thresholds, as part of a strategy where treatment decisions were based on DXA in patients with normal or osteoporotic BMD, while FRAX was used in those with osteopenia. Methods: We examined DXA reports of 1884 consecutive patients (average age 63.0 years, 79 % female) attending our NHS DXA unit over 12 months. In those with osteopenia, treatment was recommended where 10 year fracture risk as calculated by FRAX exceeded 20/5 % major/hip fracture risk. Results were simulated when thresholds were applied based on NOGG (age-dependent thresholds) and NOF (20/ 3 % major/hip fracture risk) recommendations. Age, gender and FRAX scores of those identified for treatment using different cut-offs were analysed. Results: 24.2 % (n0456) had normal DXA scan results according to WHO criteria (T-score>−1) (no treatment recommended); 27.2 % (n0512) were osteoporotic (T-score≤−2.5) (treatment recommended); 48.6 % (n0916) were osteopenic (T-score −2.4 to −1 inclusive) (FRAX used to determine treatment). Local criteria identified substantially less patients requiring treatment [local criteria0156 (17.0 %;95 % female); N O G G 02 9 9 ( 3 2 . 6 % ; 8 9 % f e m a l e ) ; N O F 03 0 0 (32.8 %;88 % female)]. Patients identified as requiring treatment were substantially older following application of local criteria (mean 70.9, 61.6, 69.1 years with local criteria, NOGG and NOF respectively), had a higher 10 year major fracture risk (mean (%)024.1, 18.2, 19.4 respectively), and higher 10 year hip fracture risk (7.6, 4.7, 5.8 respectively). In contrast, femoral neck BMD of patients selected for treatment was similar (mean 0.61, 0.63 and 0.62 g/cm2 for local criteria, NOGG and NOF respectively). Conclusions: Using a threshold of osteoporosis by DXA, combined with a 20/5 % threshold following FRAX in those with osteopenia, resulted in 35.5 % of patients attending our DXA unit being recommended for treatment. Using the NOGG threshold increased this figure to 43.0 % due to the identification of substantially younger patients, of considerably lower fracture risk.

S550

P4 NHS RADIOGRAPH REPORTS OF VERTEBRAL FRACTURE (VF) COMPARE FAVOURABLY WITH STANDARDISED METHODS, BUT SOME NHS REPORTS ARE UNCLEAR Emma Clark (1) presenting Virginia Gould (1) Leigh Morrison (1) Jon Tobias (1) University of Bristol, Bristol, UK (1)

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

Department of Paediatrics, Faculty of Health Sciences, University of the Witwatersrand, Johannesburg, South Africa (3) MRC Keneba, Keneba, Gambia (4)

Introduction: We have recently reported the results of a large RCT of a screening programme aiming to identify older women with osteoporotic VFs, and results were encouraging as allocation to screening approximately doubled prescribing of medication for osteoporosis. However, despite this increase, less than half of women with a VF had received a new prescription for anti-osteoporosis medications within 12 months, with the main reason for this being the non-standardised nature of NHS radiograph reporting. We therefore undertook an in-depth analysis of the radiology reports produced as part of our RCT with an aim of identifying areas for improvement. Methods: This study consisted of 310 spinal radiographs performed on older women, all of whom had been identified as being at high risk for VF as part of a previous study (COSHIBA). The radiographs were interpreted using three methods: standard NHS reporting; the algorithm-based qualitative (ABQ) approach; and quantitative morphometry (QM) by SpineAnalyzer software, with presence of VF being defined as ≥25o height reduction. Ethics approval was obtained. Results: NHS reports were categorised into 230 (74.2 %) without VF and 28 (9.0 %) with definite VF. There were a further 52 (16.8 %) where the reports were unclear (using words such as 'collapse' instead of fracture, or use of 'probably' ). Analysis by ABQ and QM showed the NHS reports had good agreement for no VF (93.0-98.3 % agreement) and definite VF (71.4-82.1 % agreement). But the 'possible VF' category contained approximately 70-80 % without actual osteoporotic VF. Conclusions: Standard NHS spinal radiograph reports of VF compare favourably with ABQ and QM methods. However, approximately a fifth of reports were unclear. Clarifying this group and re-categorising into definite or no VF is needed to improve communication to GPs, and allow appropriate treatment with anti-osteoporosis medications for this group of women at very high risk for future fractures.

Introduction: HIV-infection and antiretroviral medication (ARV) are associated with low BMD and increased rates of bone loss. Proposed mechanisms include traditional risk factors e.g. low BMI but also poor vitamin D status and exposure to a pro-inflammatory milieu. Many ARVs, especially Tenofovir, are associated with low BMD because of an increased prevalence of proximal renal tubular dysfunction and urinary phosphate loss. The purpose of this study was to consider bone status and body composition (BC) in HIV-positive women living in urban South Africa. Materials and methods: 247 premenopausal (mean age 32.1 (±7.24) years), black African women were recruited in Soweto, South Africa and underwent baseline DXA measurement for bone mass and BC. Group 1: HIV-negative control (n098), Group 2: HIV-positive, preserved CD4 count (i.e. not requiring ARV, mean CD4: 412) (n074), Group 3: HIV-positive, low CD4 count (i.e. prior to ARV initiation, mean CD4: 161) (n075). Discussion: Group means were compared using ANOVA. Mean values (±SD) in groups 1, 2 & 3 respectively for: Height (m) 1.58 (±0.06), 1.59 (±0.06), 1.59(±0.05) (NS); Weight (Kg) 69.7 (±17.0), 72.0 (±17.4), 62.3 (±15.2) (p0 0.02 (Groups 1–3), p00.002 (Groups 2–3)); BMD (g/cm2) Total Hip 1.013 (±0.131), 0.985 (±0.124), 0.988 (±0.125) (NS); Femoral Neck 0.930 (±0.114), 0.916 (±0.125), 0.923 (±0.131) (NS); Lumbar Spine 1.018 (±0.118), 1.021 (±0.109), 1.006 (±0.128) (NS) & Whole body less head 0.958 (±0.079), 0.943 (±0.071), 0.947 (±0.080) (NS); Fat: Lean proportion 0.66 (±0.22), 0.60 (±0.19), 0.59 (±0.23) (NS). After full adjustment for bone area, weight, height and age there were no significant BMD differences at any site. Conclusion: HIV-positive women with a low CD4 count, requiring ARV initiation, are significantly lighter than HIVpositive women with preserved CD4 counts and HIVnegative women but have no significant difference in BC or BMD measures at any site. These baseline data provide a unique insight into bone health and BC in black, African women living with HIV and will allow for longitudinal assessment of changes. The subjects are currently being followed up to assess effects of ARV exposure on bone mineralisation and BC.

P5 BONE AND BODY COMPOSITION IN HIV-POSITIVE AND HIV-NEGATIVE BLACK, SOUTH AFRICAN WOMEN Matthew Hamill (1,2) presenting Kate Ward (1) John Pettifor (3) Shane Norris (2) Ann Prentice (1,4) MRC Human Nutrition Research, Cambridge, UK (1) Developmental Pathways for Health Research Unit, Johannesburg, South Africa (2)

P6 THE RELATIONSHIP BETWEEN AORTIC STIFFNESS AND ABDOMINAL AORTIC CALCIFICATION MEASURED ON SPINE DXA VFA SCANS IN POSTMENOPAUSAL WOMEN Sylvia Edwards (1) presenting Ignac Fogelman (1) Amelia Moore (1) Michelle Frost (1) King's College London School of Medicine, London, UK (1)

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

Introduction: It has been demonstrated that low BMD is associated with increased vascular calcification and aortic stiffness. Pulse Wave Velocity (PWV) is a validated noninvasive method for measuring aortic stiffness, which is dependent in part on the extent of calcification within the aorta. Pulse wave velocity has been shown to be an independent predictor of both cardiovascular events and cardiovascularrelated mortality. It is possible to assess abdominal aortic calcification (AAC), an important risk factor for subclinical cardiovascular disease, using DXA as calcification within the aorta can be visualised on lateral vertebral fracture assessment (VFA) scans in the lumbar spine region. The aim of this study was to assess the correlation between AAC measured on spine VFA DXA scans with aortic PWV. Methods: A total of 149 postmenopausal women with a mean age of 61.5 years had both aortic PWV and DXA VFA scans of the lumbar spine. Aortic PWV was measured on the SphygmoCor system using ECG and ultrasound tonometry. VFA DXA scans were performed at the lumbar spine and the extent of calcium was estimated in the anterior and posterior walls of the aorta using a previously validated 24-point and a simplified 8-point scoring method. Of the 149 a total of 114 women were included in the analysis. Results: The extent of calcium could not be assessed on 35 of the DXA VFA scans due to poor image quality caused by bowel gas and other image artefacts. Of the remaining 114 scans calcium deposits in the abdominal aorta were observed in 63 % of subjects. There was a significant correlation between aortic PWV and AAC using both the 24-point (r00.23, p00.015) and 8-point (r00.19, p00.043) scoring methods. Conclusion: This is the first study reporting a significant correlation between PWV and abdominal aortic calcification assessed on DXA VFA scans estimated using both a validated 24-point and a simplified 8-point scoring method. P7 QUANTITATIVE HEEL ULTRASOUND PARAMETERS AND MORTALITY IN EUROPEAN MEN Stephen Pye (1) presenting Dirk Vanderschueren (2) Steven Boonen (2) Evelien Gielen (2) Judith Adams (3) Kate Ward (4) David Lee (1) Gyorgy Bartfai (5) Felipe Casanueva (6) Joseph Finn (1) Gianni Forti (7) Aleksander Giwercman (8) Ilpo Huhtaniemi (9) Krzysztof Kula (10) Margus Punab (11) Frederick Wu (1) Terence O'Neill (1) The University of Manchester, Manchester, UK (1) Katholieke Universiteit Leuven, Leuven, Belgium (2) Radiology and Manchester Academic Health Science Centre in The Royal Infirmary, Manchester, UK (3) MRC Human Nutrition Research, Cambridge, UK (4) Albert Szent-Gyorgy Medical University, Szeged, Hungary (5) Santiago de Compostela University, Santiago de Compostela, Spain (6) University of Florence, Florence, Italy (7) University of Lund, Malmo, Sweden (8) Imperial College London, London,

S551

UK (9) Medical University of Lodz, Lodz, Poland (10) United Laboratories of Tartu University Clinics, Tartu, Estonia (11) Introduction: Low bone mineral density (BMD), as measured by dual energy x-ray absorptiometry, is associated with increased mortality in prospective cohort studies. Less is known about the relationship between other skeletal phenotypes and mortality particularly in men. The aim of this study was to determine the relationship between quantitative heel ultrasound parameters and mortality in European men. Material and Methods: Men aged between 40 and 79 years were recruited from population registers in 8 European centres for participation in a prospective study of male ageing: the European Male Ageing Study (EMAS). At baseline, subjects attended for quantitative ultrasound (QUS) of the heel (Hologic SAHARA) and completed questionnaires which included information on physical activity, smoking status and number of comorbidities. Height and weight were also measured. All subjects were invited to attend a follow-up assessment and reasons for non-participation, including death, were recorded. The relationship between baseline QUS parameters (broadband ultrasound attenuation [BUA] and speed of sound [SOS]) and subsequent mortality was assessed using Cox proportional hazards model with the results expressed as hazard ratios (HR) and 95 % confidence intervals (95%CI). Discussion: 2825 men (mean age 59.8, standard deviation [SD] 10.8 years) were included in the analysis, of whom 185 (6.6 %) died during a median of 4.3 years of follow-up. Mean QUS levels were lower in those who died compared to survivors: BUA (74.4 vs 80.2 db/MHz; p<0.001) and SOS (1538.8 vs 1550.8 m/s; p<0.001). After adjusting for age, centre, body mass index, physical activity, current smoking and number of comorbidites, each SD increase in BUA was associated with a 16 % reduction in risk of mortality (HR per SD 00.84; 95%CI00.71, 0.98). Similarly, each SD increase in SOS was linked with a 15 % reduction in mortality though this was of borderline significance only (HR per SD00.85; 95%CI00.71, 1.00). Compared to those in higher quintiles (2nd to 5th), those in the lowest quintile of BUA and SOS had an increased mortality risk (BUA: HR01.63; 95%CI01.16, 2.31 and SOS: HR01.66; 95%CI01.19, 2.31). Conclusion: Lower heel ultrasound parameters were associated with a significantly increased risk of all cause mortality in European men. P8 COMPARISON OF THREE FRACTURE PREDICTION TOOLS IN AN OSTEOPOROSIS CLINIC POPULATION Jennifer Thain (1) presenting Darren Aw (1) Lindsey Marshall (1) Opinder Sahota (1) Wei Mei Chua (1) Namal Weerasuriya (1) Thanda Aung (1) Fiona Kearney (1) Aamer

S552

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

Ali (1) Tahir Masud (1) Nottingham University Hospitals NHS Trust, Nottingham, Nottinghamshire, UK (1)

Exeter, UK (2) College of Life and Environmental Sciences, University of Exeter, Exeter, UK (3)

Introduction:Fracture prediction is used to assess fracture risk as an aid for discussion with patients and to help guide management decisions. The most commonly used tool is FRAX, although the Garvan and QFracture tools have also recently been advocated. This study compared fracture prediction between these tools in women attending osteoporosis clinics. Materials and Methods: Women consecutively attending osteoporosis clinics were studied. Clinicians completed a structured questionnaire designed to ascertain information required to populate the three tools. A researcher used the web-based tools to find the 10 year fracture predictions which were then compared. The FRAX and Garvan scores included the femoral-neck bone density T-scores, whereas the QFracture does not allow for this investigation. The relationship between fracture prediction and a history of at least one fall in the previous year was also explored for Garvan and QFracture (FRAX does not allow for a falls history). Discussion: 106 women (mean age069.1 years, SD012.4) were studied. 37 women (34.9 %) had fallen in the previous year. The mean 10 year ”major” fracture risks in percentage (95%CIs) for FRAX, Garvan and QFracture were 18.3 % (10.9-25.7), 46.4 % (35.6-57.3) and 12.1 % (5.5-18.7) respectively. The mean 10 year hip fracture risks were 7.1 % (2.2-12.0), 29.1 % (19.1-39.1) and 6.0 % (1.2-10.8) respectively. There was a significant difference in the mean 10 year “major” fracture risk between non-fallers versus fallers using the Garvan tool [39.4 % vs 63.7 %; p<0.001] whereas QFracture was unable to distinguish between the groups [12.0 vs 12.4 %; p00.8]. Conclusion: Compared to FRAX, the Garvan tool grossly overestimated both the 10 year risk of major fractures (more than doubled) and hip fractures (by four-fold). The QFracture tool tended to underestimate the risk of major fractures (by a third) compared to FRAX, although hip fracture prediction was similar. Only the Garvan tool could predict a higher fracture risk in those with falls in the previous year. Clinicians should be aware of the differences and variations in fracture prediction between the web-based tools. The comparative accuracy of the different tools in predicting fractures requires a prospective study.

Introduction:Disuse osteopenia is a known consequence of reduced weight-bearing on lower limbs. Osteoarthritis (OA) commonly necessitates joint replacement with limited mobility for a variable period post surgery. Although OA is generally associated with higher levels of BMD, a significant increase in hip fracture incidence in the year following total knee replacement (TKR) has been demonstrated; however the extent of disuse-related bone loss at the hip following TKR and its potential contribution to post-surgical fracture risk has not been reported. Material and Methods: This study investigated long-term changes in DXA BMD at the neck of femur (NOF) and total hip (TH), (GE Lunar Prodigy), and leg lean tissue mass (LLTM) in a population of 11 postmenopausal females following TKR compared to 45 postmenopausal controls. Relative ipsilateral/ contralateral weight-bearing, lower limb function, 3 day pedometer readings and falls were also recorded. Measurements were obtained at pre-surgery baseline, and six months post surgery. Discussion: No significant differences were demonstrated between groups at baseline bilaterally in LLTM or BMD at the NOF and TH. Significant losses in ipsilateral hip BMD and bilateral LLTM are demonstrated in the TKR group 6 months post surgery.

P9 A STUDY INVESTIGATING THE LONG-TERM EFFECTS ON FUNCTION, BONE MINERAL DENSITY AND LEAN TISSUE MASS POST TOTAL KNEE REPLACEMENT IN A FEMALE POSTMENOPAUSAL POPULATION Susan Hopkins (1) presenting Chris Smith (1) Andrew Toms (2) Mary Brown (2) Joanne Welsman (3) Karen Knapp (1) College of Engineering Mathematics and Physical Sciences, University of Exeter, Exeter, UK (1) Princess Elizabeth Orthopaedic Centre, Royal Devon and Exeter Hospital,

Controls Mean Baseline 6 months

TKR Mean Baseline 6 months

n Age at baseline (y) BMI at baseline (kg/m2) BMD

45 64.5 25.7

45 -

11 63.6 34.11

11 -

Ipsilateral NOF (g/cm2) Contralateral NOF (g/cm2) Ipsilateral Total Hip (g/cm2) Contralateral Total Hip (g/cm2) Leg Lean tissue (g) Ipsilateral Contralateral Other measurements % Ipsilateral weightbearing LEFS[1] (maximum score 80) Pedometer reading (steps per day) Mean Number of falls

0.899

0.900

0.941

0.917*

0.905

0.902

0.899

0.902

0.954

0.954

0.968

0.952*

0.959

0.954

0.964

0.967

6307 6281

6301 6257

6629 6642

6091* 5966*

50.39

50.52

41.09

46.44

73.8

73.1

24.7

47.0*

9973

8236

390

3502*

0.25

0.22

0.73

0.45

* p-<0.05 when compared to baseline for the same group [1] Lower Extremity Functional Scale, Binkley et al

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

Conclusion: Significant impairment in function and weight-bearing persisted in the TKR group 6 months post operatively alongside significant deficits in bilateral muscle mass and ipsilateral NOF and TH BMD. Falls incidence was not significantly higher in the TKR group suggesting that disuse-related reduction in hip BMD and LLTM may contribute to increased hip fracture incidence following TKR. P10 FACTORS UNDERLYING WOMEN'S PERCEPTION OF RISK FOR OSTEOPOROTIC FRACTURES - PERCEIVED RISK VERSUS FRAX Mette Juel Rothmann (1,2) presenting, Jette Ammentorp (2) Mickael Bech (3) Kim Brixen (4) Anne Pernille Hermann (1,2) Department of Endocrinology, Odense University Hospital, Odense, Denmark (1) Institute of Regional Health Services Research, University of Southern, Odense, Denmark (2) The Research unit of Health Economics, University of Southern Denmark, Odense, Denmark (3) Institute of Clinical Research, University of Southern Denmark, Odense, Denmark (4) Introduction:In most countries general practitioners refer persons with one or more risk factors to Dual energy Xray absorptiometry (DXA). Little is known about referred women´s perceived risk of fracture. The aim of this study was to investigate self-perception of fracture risk, potential factors associated with this, and to compare perceived risk and absolute risk estimated by FRAX Material and Methods: A self-administrated pilot tested questionnaire was sent to 1000 consecutive women DXA at our outpatient clinic, response rate 80 %. The questionnaire comprised 25 items on osteoporosis, risk factors and perceived risk. Women aged 40–90 years were included (n0592) in this analysis. Median age was 65 [41–88]. Sixty-three percent of the women had a high 10-year probability of major osteoporosis-related fractures (FRAX>15 %), and 28 % of the women perceived themselves at high risk. Univariate analyses, probit and logistic regression models were performed. Association between perceived risk and FRAX were analysed by odds model and kappa statistic. Data is presented as median [range], and odds ratio (CI 95 %) Discussion: We found that women with a family history of osteoporosis and a personal history of fracture had a 2-fold increased probability to perceive themselves at high risk. Also self-rated health was highly associated to perceived risk; the overall effect of self-rated health was highly significant, and the analysis showed that the predicted probability of a high perceived risk was 0.44 for the lowest self-rated health and 0.02 for the highest self-rated health. Furthermore housing was associated; women living in apartments had a 1.9-fold increased probability of perceived high risk. Finally, falls were

S553

slightly significant 1.8(1.0-3.5). Overall, the kappa statistic (k0 0.061) showed poor agreement between FRAX and selfperceived risk. The women slightly significantly underestimated fracture risk compared to FRAX, odds ratio1.5 (1.0-2.2) Conclusion: Perceived risk is highly significantly associated with family history of osteoporosis, personal history of fracture and self-rated health, and to a lesser degree with housing. Women referred for osteoporosis evaluation seem to underestimate fracture risk compared to absolute risk (FRAX) P11 PREVENTING FALLS AND FRACTURES IN NURSING HOMES: A PILOT PROJECT Alison Doyle (1) presenting Jane Beach (1) Neil Gittoes (1) Mark Cooper (1) Bola Ogunremi (1) Birmingham University, West Midlands, UK (1) Introduction:Falls and fractures are a major problem in South Birmingham. The region has a population of 371,000 of which 52,916 are over 65. It is estimated that 30 % of the over 65 population will have a fall and 1600 will have a major injury e.g. fracture. Nursing home residents are at the highest risk of falls and associated morbidity and mortality. In an attempt to reduce the burden of falls and fractures we piloted a multidisciplinary intervention in 13 of the 23 nursing homes within the region and assessed the impact that this had on falls and fractures. Methods: Initially each home was visited by members of the project team to establish specific issues. Potential solutions were identified in partnership with the homes and best practice and training from a multidisciplinary team was delivered. This involved the development of a ‘Falls and Fracture Prevention Toolkit’, establishing specific networks for sharing good practice and the development of a risk assessment tool and care plans. Systems for the reliable monitoring of falls and fractures within each home were established. Discussion: At the end of the pilot, data on rates of falls before and after the pilot were available for 12 of the 13 homes. The number of falls within all nursing homes combined decreased by an average of 22 % overall (decreasing from 123 to 96 falls in a population of 543 residents; p<0.05). The intervention reduced falls based on a paired analysis using data from each home (p< 0.05). Although not powered to detect an effect on fractures in matched time periods before and after the intervention the rate of hip fracture decreased from 4 to 0 (the first time on record that no hip fractures had occurred in this group of homes). Conclusion: This study suggests that targeted multidisciplinary intervention can be effective in reducing falls. Benefits are likely to extend to reduction in serious fracture. The overall direct cost of the intervention, ~£7000 suggests the intervention is cost-effective. The intervention was well received by nursing home management and staff.

S554

P12 DO HEAVY METALS IN MUNICIPAL DRINKING WATER INCREASE THE RISK OF HIP FRACTURES? THE NORWEGIAN EPIDEMIOLOGIC OSTEOPOROSIS STUDY (NOREPOS) Cecilie Dahl (1,2) presenting Anne Johanne Søgaard (1) Grethe Tell (2) Geir Aamodt (1) Division of Epidemiology, Norwegian Institute of Public Health, Oslo, Norway (1) Department of Public Health and Primary Health Care, University of Bergen, Bergen, Norway (2) Introduction: Hip fracture incidence has recently stabilized, and even declined in many western countries, including Norway. At the same time there has been a reduction in pollutants, such as cadmium and lead in the environment. Geographic variations indicate that a decreased exposure to harmful environmental factors in some areas could be one cause of this decline. Our aim was to investigate a possible effect of cadmium and lead in municipal drinking water on the incidence of hip fracture in the Norwegian population. Particularly we were interested in possible interactions with calcium. Materials and Methods: Geographic Information Systems was used to link a survey on heavy metals conducted in 556 waterworks to hip fracture incidence data from Norwegian hospitals (1994–2000). A total of 7,003 fractures in men and 19,727 in women, ages 50–111 were included. Poisson regression (incidence rate ratios) was used to investigate the effect of cadmium and lead on the incidence of hip fractures, stratifying on age-group, time-period, water-source, region, water acidity, calcium and other bone related factors. Discussion: Incidences of 33 and 75 per 10,000 were seen in men and women, respectively. Men exposed to cadmium above the measureable limit (>0.1 ug/l) in drinking water had a higher age adjusted incidence of hip fracture, particularly in the age-group 50–85 years (IRR 01.08, 95 % CI: 1.0, 1.17). When stratifying on high and low calcium content, a 25 % increased risk of hip fracture was seen in men exposed to water containing high cadmium and >15 mg/l calcium (IRR 01.25, 95 % CI: 1.04, 1.49) compared to low cadmium and high calcium, but no association was seen in the low-calcium group. No increased risk was seen in men by exposure to lead, and there was no increased risk in women by exposure to neither lead nor cadmium. Conclusion: Our findings indicate an increased risk of hip fracture when exposed to cadmium in drinking water; however the risk is not uniform across gender and age-groups. It is also dependent on the concentration of components such as calcium in the water.

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

P13 THE MANCHESTER CHILDREN'S BONE DENSITOMETRY (HOLOGIC DXA) REFERENCE DATABASE: AN UPDATE Judith Adams (1,2) presenting Zulf Mughal (1,2)Steve Roberts (4) Kate Ward (3,2) Central Manchester University Hospitals NHS Foundation Trust, Manchester, UK (1) Manchester Academic Health Science Centre, University of Manchester, Manchester, UK (2) MRC Human Nutrition Research, Cambridge, UK (3) Health Sciences Methodology, University of Manchester, Manchester, UK (4) Introduction: For interpretation of bone mineral content (BMC) and mineral density (BMD) in children robust reference databases are essential. DXA measured areal BMD (g/cm2) is size-dependent and various correction methods are suggested however no consensus of the optimum method is agreed(1). The International Society for Clinical Densitometry produced guidelines to clarify which measurements to use but require expert interpretation(1). We published the Manchester children’s BMD reference database (Hologic QDR Discovery DXA [n0442; 6–16 years] for the assessment of lumbar spine (LS), femoral neck (FN) and whole body (WB) DXA data, these data have been widely applied in the UK(2). We have updated these to increase subject numbers, extend age range, provide more sizecorrected data and use whole body less head (WBLH) data advocated in children(1). Subjects & methods: Children/young adults were recruited from the Greater Manchester area. DXA scans of LS, hip (total, FN) and WB were made and analysed (fast array mode, software version 12.1, adult analysis for WB). Bone mineral apparent density (BMAD) was calculated for LS and FN(3). Data are presented as centile charts calculated by LMS method and enable calculation of gender specific Z-scores(4). Results: 679 (additional 237 subjects) children/young adults (M0408, F0271 aged 5-22y and 5-21y respectively) were recruited. LMS curves and tables were generated for LS and FN BMAD, total hip BMD. WBLH was analysed by Molgaard method(5) which compares ‘height for age’ (short bones), bone area for height (narrow bones) & BMC for bone area (are bones under-mineralised?) and the Crabtree method assessing whether primary disease is muscle or bone(6). Conclusions: We apply these enhanced reference database LMS curves to children with various clinical disorders which affect the skeleton (e.g. osteogenesis inperfecta, Duchenne muscular dystrophy, inflammatory bowel disease, etc.). These data provide UK specific resource for bone density measurements in children/ young adults. 1 J Clin Densitom 11, 43 (2008); 2Arch Dis Child 92, 53 (2007); 3 J Bone Miner Res 7, 137 (1992); J Clin Endocrinol Metab 81, 1586 (1996); 4 Statist Med 11, 1305 (1992); 5Arch dis child 76, 9 (1997); 6Bone 35, 965 (2004).

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

S555

P14 IN-VIVO CROSS-CALIBRATION OF GE LUNAR iDXATM AND PRODIGY TM BONE DENSITOMETERS IN ADULTS AND CHILDREN Natalie Bebbington (1,2) presenting Dee Chapman (2,1) Nick Shaw (2) Wolfgang Högler (2) Chris Boivin (1) Nicola Crabtree (2,1) The Queen Elizabeth Hospital Birmingham, Birmingham, UK (1) Birmingham Children's Hospital, Birmingham, UK (2)

P15 MUSCLE SIZE, STRENGTH AND PHYSICAL PERFORMANCE AS PREDICTORS OF FALLS AND FRACTURES IN THE HERTFORDSHIRE COHORT STUDY Mark Edwards (1) presenting Karen Jameson (1) Celia Gregson (1) Nicholas Harvey (1) Avan Aihie Sayer (1) Elaine Dennison (1) Cyrus Cooper (1) MRC Lifecourse Epidemiology Unit, Southampton, Hampshire, UK (1)

Introduction: Cross-calibration between old and new technologies for bone density and body composition should be considered when upgrading densitometers, particularly when assessing longitudinal trends. Good agreement exists in adults between the Prodigy and iDXA, yet there is little published data examining whether this relationship extends to paediatrics. The aim of this study was to cross-calibrate these densitometers for both an adult and paediatric population. Material and Methods: 62 adults and 70 children (mean age 58.4 and 12.4 years) referred for clinical DXA assessment were scanned using Prodigy and iDXA densitometers. In accordance with standard practice, bone mineral density (BMD), bone mineral content (BMC) and bone area (BA) was evaluated in adults at the lumbar spine and proximal femur and in children at the lumbar spine and total body less head (TBLH). Prodigy and iDXA differences were assessed using paired t-tests and cross-calibration was achieved using a general linear regression model and Bland-Altman analysis. Discussion: Lumbar spine and TBLH BMD was significantly higher on the Prodigy than iDXA (Prodigy-iDXA mean difference: L2L4 00.008 g/cm2 [0.02 g/cm2]; TBLH 0 0.092 g/cm2[0.025 g/cm2];p≤0.05). There was no significant difference between Prodigy and iDXA means for total hip BMD. Marked differences were detected for body composition parameters. The iDXA consistently reported higher fat and consequently lower lean tissue values than the Prodigy (mean difference: fat 01205 g[801 g]; lean 0 433 g[863 g]).Of the evaluated sites, Bland-Altman analysis demonstrated most marked significant machine bias for TBLH parameters. At lower levels of mean bone density the iDXA consistently resulted in greater detected bone area, BMC and a significantly lower proportion of lean mass, and consequently higher fat levels. r200.445, 0.497, 0.073, 0.257 for mean difference versus mean BMD for BA, BMC, lean and fat mass respectively. These findings are consistent with differences in automated bone and soft tissue edgedetection algorithms between densitometers, resulting from improved resolution on the iDXA. Conclusion: Significant differences exist between Prodigy and iDXA densitometers, the most marked of which being associated with TBLH data. This has relevance for both clinical practice and research indicating the importance of maintaining consistency of densitometer.

Introduction: Sarcopenia is common in later life and associated with subsequent disability. However, the definition of sarcopenia is often problematic; hence recently the European Working Group on Sarcopenia in Older People proposed a practical clinical definition for agerelated sarcopenia requiring both low muscle mass and function (strength or performance). However, the extent to which these different components correlate with falls risk, and ultimately fracture, remains uncertain. We address this issue in a prospective population cohort, the Hertfordshire Cohort Study. Material and Methods: 1579 men and 1418 women underwent baseline assessment of health (questionnaire) and detailed anthropometric measurements. Grip strength was measured using a Jamar hand-held dynamometer and gait speed was assessed by 3 metre walk test. A subset of 313 men and 318 women, underwent peripheral quantitative computed tomography (pQCT) examination of the calf and forearm using a Stratec 2000XL pQCT to assess muscle cross-sectional area (mCSA) (66 % slice). Subsequently, 2299 participants completed a postal questionnaire, detailing fall and fracture history, a mean of 5.5 years after baseline (range 2.9–8.8 yrs). Discussion: The mean age (SD) at baseline was 66.2 (2.8) years. At follow-up weaker grip strength and, in men reduced gait speed, predicted a higher risk of falls reported since the age of 45 years and in the previous year. A similar but much weaker trend with gait speed was seen in men. Incident fractures were more common in those with lower grip strength (men OR 1.075[95%CI 1.028, 1.125] p 00.002; women OR 1.039[95%CI 1.002, 1.077] p00.04, for every 1 kg reduction in grip strength) after adjustment for age, height, weight-adjusted-forheight, social class, smoking status, alcohol consumption, activity score, dietary calcium (and oestrogen replacement use and years since menopause in women). However, no associations were detected between mCSA and incident falls or fractures. Conclusion: Both grip strength and gait speed but not mCSA were associated with falls risk; additionally grip strength also predicted incident fractures. Assessments of muscle function therefore may be better predictors of these clinical outcomes than muscle size.

S556

P16 SARCOPENIA AND BONE MINERAL DENSITY IN EUROPEAN MEN Sabine Verschueren (1) Evelien Gielen (1) presenting Terence O'Neill (2) Stephen Pye (2) Judith Adams (3) Kate Ward (4) FrederickWu (2) Pawel Szulc (5) Michaël Laurent (1) Frank Claessens (1) Dirk Vanderschueren (1) Steven Boonen (1) KU Leuven, Leuven, Belgium (1) University of Manchester, Manchester, UK (2) Radiology and Manchester Academic Health Science Centre in Manchester Royal Infirmary, Manchester, UK (3) MRC Human Nutrition Research, Cambridge, UK (4) University of Lyon, Lyon, France (5) Introduction: There is now considerable evidence that muscle and bone interact to impact on bone strength. The aim of this analysis was to determine the association between reduced muscle mass (sarcopenia) and areal bone mineral density (BMDa) in middle-aged and elderly communitydwelling European men. Materials and Methods: Men aged 40–79 years from the Manchester (UK) and Leuven (Belgium) cohorts of the European Male Ageing Study (EMAS) were invited to attend for physical performance tests (time to walk 50 feet, [s]) and dual energy X-ray absorptiometry, from which appendicular lean mass (aLM, [kg]), total fat mass (FM, [kg]) and BMDa at the total hip and lumbar spine were determined. Relative appendicular skeletal muscle mass (RASM) was calculated as aLM/height² and sarcopenia defined as RASM <7.26 kg/m2. Linear regression was used to determine the associations between LM, FM, physical performance and BMDa (standardized into Z-scores). Logistic regression was used to determine the association between sarcopenia and osteoporosis, defined as a T-score<−2.5 at the hip or lumbar spine. Discussion: 679 men, mean age 60.0 (SD011.0) were included in this analysis. Mean total hip BMDa was 1.015 (SD0 0.142) g/cm² and mean lumbar spine BMDa 1.049 (SD0 0.173) g/cm². 11.9 % were sarcopenic and 8.8 % osteoporotic. After adjustment for age and centre, increasing aLM, RASM and FM were associated with increasing BMDa at the total hip (β00.139, β00.433, β00.049 respectively; p<0.05) and lumbar spine (β00.102, β00.294, β00.034; p<0.05). Sarcopenic men had significantly lower BMDa at both sites compared to those with RASM ≥7.26 kg/m2. In a stepwise linear regression model which included age, height, aLM, FM and physical performance, aLM was most consistently associated with BMDa (β00.119 at total hip and β00.101 at lumbar spine; p<0.05). Fat mass was an independent predictor at the hip (β00.017; p<0.05) though not spine. After adjustment for age and centre, sarcopenic men were more likely to have osteoporosis compared to those with normal RASM (Odds Ratio03.0; 95%CI01.6, 5.8).

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

Conclusion: Sarcopenia is associated with low BMDa and osteoporosis in men. Further studies are necessary to assess whether the maintenance of muscle mass may contribute to prevent bone loss in older men. P17 USE OF THIAZOLIDINEDIONES AND RISK OF OSTEOPOROTIC FRACTURE: DISEASE OR DRUGS? Marloes Bazelier (1) presenting Arlene Gallagher (1,2) Peter Vestergaard (3) Cyrus Cooper (4) Frank de Vries (1,4) Utrecht University, Utrecht, The Netherlands (1) General Practice Research Database, London, UK (2) Aarhus University Hospital, Aarhus, Denmark (3) University of Southampton, Southampton, UK (4) Background: The use of thiazolidinediones (TZDs) has been associated with an increased fracture risk. In addition, the severity of type 2 diabetes mellitus (T2DM) is also a risk factor for osteoporotic fracture. Objective: To evaluate to which extent the association between TZD use and fracture risk is related to the drug, or to the severity of the underlying disease. Methods: We conducted a population-based cohort study using the Danish National Databases (1996–2007), which link pharmacy dispensing data to the national hospital registry. Oral antidiabetic users (n0180,049) were matched 1:4 by year of birth and sex to non-users. Cox proportional hazards models were used to estimate hazard ratios (HRs) of osteoporotic fracture. Time-dependent adjustments were made for age, comorbidity, and drug use. We created a proxy indicator for the severity of disease. The first stage was defined as current use of either a biguanide or a sulfonylureum, the second stage as current use of a biguanide and a sulfonylureum at the same time, the third stage was assigned to patients using TZDs and the fourth stage to patients using insulin. Results: The risk of osteoporotic fracture was increased 1.3fold for stages 3 and 4, compared with controls. Risk with current TZD use (stage 3, HR 1.27 [1.06-1.52]) and risk with current use of insulin (stage 4, HR 1.25 [1.20-1.31]) were similar. In the first (HR 1.15 [1.13-1.18]) and second stage (HR 1.00 [0.96-1.04]) risks were lower. Conclusion: Users of TZDs were at an increased risk of osteoporotic fracture. The association is probably partially confounded by the severity of the underlying T2DM. P18 THE INFLUENCE OF PAST DIET ON BONE QUALITY AND STRESS FRACTURE RISK IN ROYAL MARINE RECRUITS Trish Davey (1,2) presenting Anneliese Dziubak (1) Rosalyn Cobley (1) Susan A. Lanham-New (2) Adrian J. Allsopp (1) Joanne L. Fallowfield (1) Institute of Naval Medicine, Gosport, UK (1) University of Surrey, Guildford, UK (2)

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

Introduction: A high quality diet during childhood and adolescence is essential to ensure optimal development of bone. The aim of this prospective cohort study was to examine the influence of past dietary intake on bone quality and stress fracture risk in young males (aged 16–32 y) undergoing arduous Royal Marine (RM) recruit military training. Materials and Methods: RM recruits (n 870) underwent Broadband Ultrasound Attenuation (BUA) measurement of the dominant foot calcaneus at the start of training, as an assessment of bone quality. Past dietary intake was selfreported using a validated Food Frequency Questionnaire (FFQ). Spearman rank correlations and partial correlations were undertaken to explore associations between diet and BUA. Relative Risk (RR) was calculated to compare risk of fracture between levels of food and nutrient intakes. Underand over-reporters were excluded. Results and Discussion: Milk intake during childhood (0–12 y), and fruit and vegetable intakes during adolescence (12–18 y) were positively associated with BUA at the start of RM training (P<0.05). The correlations between fruit and vegetable intakes and BUA remained significant after adjustment for body mass, age and fitness (r=0.08 and r=0.03 respectively, P<0.05). When milk intake was divided into tertiles, low or moderate intake of milk during childhood was associated with an increased risk of stress fracture compared to a high intake (RR 3.2, Confidence Interval (CI) 1.5 – 6.7, and RR 2.4, CI 1.2 – 5.2, respectively). A high intake of fizzy soft beverages during adolescence was linked to an increased risk of stress fracture during RM training compared to a low intake (RR 1.5; CI 1.2 – 2.4). Recruits whose habitual calcium intake was less than 1000 mg.d-1 prior to RM training were at increased risk of suffering a stress fracture compared to recruits consuming more than 1000 mg.d-1 (RR 1.6; CI 1.2 – 2.1). Conclusion: Past diet (childhood, adolescence and prior to training) influenced bone quality and stress fracture risk in young males undergoing arduous military training. Further work is underway to explore the influence of current dietary intake and nutritional status on indices of bone health (DXA, pQCT, bone resorption) in RM recruits. P19 NEW BONE FOR OLD CARTILAGE; GROWTH PATTERNS OF THE BIZARRE 'SIXTH TOES' OF ELEPHANTS Alan Boyde (1) presenting Andrew Pitsillides (2) John R. Hutchinson (3) Barts and The London School of Medicine and Dentistry, QMUL, London E1 4NS, UK (1) Royal Veterinary College, London NW1 0 T, UK (2) Royal Veterinary College, Hatfield, Herts, AL9 7TA, UK (3) Hutchinson et al (Science 2011 334 1699) have recently considered the evolution, structure, function and development of an extra bony organ within the feet of elephants, which is

S557

usually lost during post mortem disintegration of the skeleton and has therefore been ignored and overlooked. We have studied the developmental microscopic anatomy of these organs in 57 Elephas and 6 Loxodonta, ages from neonatal to 55 years. We located areas of mineralisation in computed xray tomography scans, and, following dissection, by plain xray and/or by serial transverse sectioning into ~5 mm thick slabs. After photographic documentation, these blocks were decalcified in EDTA for conventional histology and polarised light study; or digested to leave mineralised components for 3D backscattered electron (BSE) SEM after carbon coating; or embedded in PMMA for correlative autofluorescence mode confocal laser scanning optical microscopy for histology and quantitative BSE SEM to characterise mineral content distribution. The latter blocks were repolished to remove carbon coating and stained with high atomic number elements to image additionally matrix and cell components in the block surface. Non-demineralised blocks were also imaged by high resolution x-ray microtomography. The elephant false 'sixth digit' appears in adult life in a bar of cartilage which is not part of the primary cartilaginous skeleton as this is usually understood. Our analyses of multiple samples from a range of ages suggests that the cartilaginous rods are slowly converted to bone by an unusual, seemingly haphazard, endochondral ossification mechanism. The initial hyaline cartilage shows no preferred orientation of chondrocytes or growth plate like organisation. Calcified cartilage regions are resorbed and replaced by bone which models to a porous spongy structure which exhibits no apparent preferential trabecular orientation. Additional calcified layers are added peripherally by mineralisation within ligamentous tissue. Otherwise, there is little distinction between erstwhile cortex and bulk cancellous bone. To our knowledge, no such tissue conversion mechanism has been previously described in any animal. Future work with better preserved tissue will attempt to identify growth factor involvement and deposition rates. P20 RISK OF VENOUS THROMBOEMBOLISM IN PATIENTS WITH TOTAL HIP / KNEE REPLACEMENTS AND MATCHED CONTROLS: A POPULATION-BASED COHORT STUDY IN DENMARK Arief Lalmohamed (1) presenting Peter Vestergaard (2) Corinne Klop (1) Marloes Bazelier (1)Anthonius de Boer (1) Frank de Vries (1) Utrecht University, Utrecht, The Netherlands (1) Aarhus University Hospital, Aarhus, Denmark (2) Introduction:Venous thromboembolism (VTE) is the most common cause for emergency hospital readmission following total hip and knee replacements (THR/TKR). Guidelines recommend thromboprophylaxis for 10–35 days. However VTE risk beyond this period against matched controls has

S558

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

not been studied extensively, and influence of outpatient anticoagulant use remains unknown. Consequently, the objective was to evaluate long-term VTE risk following THR/ TKR against matched controls. Methods: A Danish nationwide cohort study was conducted (1998–2007). Patients with a primary THR/TKR (n 0 95,255) were included, and three controls without any evidence for THR/TKR were matched by age, sex, and region. Time-dependent adjusted (adj.) hazard ratios (HR) for VTE were calculated. Outpatient use of anticoagulants was assessed in the previous six months. Results: Within six weeks following THR, a 13-fold increased risk of VTE (adj. HR 12.88; 95 % CI 11.25-14.76) was observed, as compared to matched controls. Risk remained substantially increased for at least six months following THR (adj. HR 2.23; 95 % CI 1.83-2.70 within 3–6 months postoperatively). In subjects who were prescribed outpatient warfarin, a substantially lower six-weeks risk was found (adj. HR 7.56; 95 % CI 4.51-12.68), compared to aspirin (10-fold increase), no use (14-fold increase), or other / mixed use (19-fold increase). All findings were similar for TKR patients. Conclusion: This study shows that VTE risk remains substantially elevated for at least six months following THR/TKR, which is well beyond the recommended duration of anticoagulant use. Furthermore, outpatient use of warfarin appeared to lower VTE risk. It may be worth investigating extended anticoagulant use (for up to six months) in clinical trials.

Results: A 2.3-fold increased risk of haemorrhagic stroke was found within two weeks following THR (adj. HR 2.25; 95 % CI 1.38-3.67), compared against matched controls. The risk remained slightly elevated within 6–12 weeks post-THR (adj. HR 1.52; 95 % CI 1.05-2.20), and dropped to baseline after this period. The six-week risk of haemorrhagic stroke was only increased in patients who underwent THR as a result of hip fracture adj. HR 2.81; 95 % CI 1.774.60), and not osteoarthritis (adj. HR 0.43; 95 % CI 0.161.14). For TKR patients, we were not able to detect an increased risk, nor could we observe a time trend for haemorrhagic stroke. Conclusion: This study shows, for the first time, that THR patients have a 2.3-fold increased risk of haemorrhagic stroke in the first two weeks post-surgery. Risk assessment for haemorrhagic stroke may be indicated, in particular in patients with a THR as a result of a hip fracture.

P21 RISK OF HAEMORRHAGIC STROKE IN PATIENTS WITH TOTAL HIP / KNEE REPLACEMENTS AND MATCHED CONTROLS: A POPULATION-BASED COHORT STUDY IN DENMARK Arief Lalmohamed (1) presenting Peter Vestergaard (2) Corinne Klop (1) Marloes Bazelier (1) Anthonius de Boer (1) Frank de Vries (1) Utrecht University, Utrecht, The Netherlands (1) Aarhus University Hospital, Aarhus, Denmark (2)

Background: Fracture Liaison Services are designed to identify patients, at the time of a fragility fracture who then may be at risk of further osteoporotic fractures and facilitate prompt fracture prevention management accordingly. One key outcome that might indicate success of an FLS is a reduced hip fracture incidence. Hip fractures are the most costly fractures to HealthCare systems given the high frequency of subsequent morbidity and the financial burden of hip fracture associated health and social care. However, hip fractures are preventable. Given that about 50 % of hip fractures occur in people who have had a previous (‘signal’) fragility fracture, such patients are potentially identifiable at the time of their ‘signal’ fracture by FLSs and offered preventive treatment to reduce the risk of hip fracture. Methods and Patients: We aimed to evaluate the role of our Fracture Liaison Service (FLS), in the context of associated local Healthcare initiatives, on hip fracture incidence over a 5y period (2005–10) through comparing the local observed hip fracture incidence with expected National UK fracture incidence. A FLS was implemented and developed at The Ipswich Hospital in 2003–4 to screen all patients with fragility fracture age ›50y. Patients with high risk for secondary fractures have since continuously been identified and managed accordingly with fracture prevention measures including osteoporosis therapy. Hip fracture data was

Introduction:Haemorrhagic stroke is a potentially fatal complication of total hip and knee replacements (THR/TKR), mainly associated with extensive thromboprophylaxis. However, timing of haemorrhagic stroke in THR/TKR patients, against matched controls, remains unknown. Objective of this study was to determine risk of haemorrhagic stroke in patients with THR/TKR against matched control subjects. Methods: A nationwide cohort study was conducted within the Danish registers (1998–2007). Patients included those with a primary THR/TKR in the study period (n095,359), and were matched by age, sex, and region to three referent subjects without THR/TKR. Time-dependent cox models were used to derive hazard ratios (HR), and were adjusted (adj.) for disease history and drug use.

P22 EVALUATING CHANGE IN HIP FRACTURE INCIDENCE IN A HOSPITAL WHERE A FRACTURE LIAISON SERVICE HAS BEEN OPERATIONAL FOR OVER 5 YEARS Gavin Clunie (1), Sonya Stephenson (1), Stephen Wilson (1), Johnathan Belsey (2); Ipswich Hospital NHS Trust, Ipswich, Suffolk, UK (1) JB Medical, Gt Cornard, Suffolk, UK (2)

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

collated by ward and FLS staff continuously since 2005. Local osteoporosis drug prescribing statistics (2006–10) were obtained from The UK NHS Business Service Authority. Expected fracture incidence was estimated for 2005–10 by applying age-sex specific hip fracture rates derived from UK Hospital Episode Statistics, to the demographic profile of our local population. Rate ratios were calculated for age groups: 55–64, 65–74, 75–84 and 85+, based on observed fracture incidence. SDs and CIs for the rate ratios were estimated using the Poisson function. Results: For all ages combined, the hip fracture rate ratio (Ipswich vs UK [England]) was unity or below (range from 0.85-1.00) though with no significant trends over time. This was similar across age groups though for some individual year/age group combinations the hip fracture rate was significantly below the expected values, notably for age ranges 55-64y. Considering all 24 data points (hip fracture rates for 4 age ranges over 6y), Ipswich hip fracture rates were less than expected on 17/24 occasions and significantly less (95 % CIs) on 5 occasions, 3 of which were recent years in the 55-64y age group. Conclusions: Hip fracture rates in Ipswich have been consistently lower than expected for England. A greater reduction in hip fractures is seen in the 55-64y old age group compared to older ages. The reduced hip fracture rates may reflect the success of continual FLS activity but results need to be viewed in the context of increased general osteoporosis drug prescribing. There is little reduction, if any, in hip fracture rates in the very elderly - an age group in which admissions to hospitals for falls is continually increasing in The UK. P23 NO EVIDENCE FOR UPREGULATION OF AUTOPHAGY DURING TRANSITION OF OSTEOBLASTS TO OSTEOCYTES Ruairidh Watson, Eman Azzam, John Greenhorn, Zhi Liu, Jenny Greenhorn; Miep Helfrich Musculoskeletal Programme, School of Medicine and Dentistry, University of Aberdeen, Aberdeen, UK (1) Introduction: Osteocytes are important for regulation of mineralisation, bone formation, phosphate metabolism, and overall energy metabolism. Paradoxically, while such activities indicate osteocytes are metabolically active cells, they are generally described as “small cells, with little cytoplasm and few organelles” and “to contain 70 % less cytoplasm than osteoblasts”. In general involution of organs is associated with upregulation of recycling mechanisms and we asked whether this might also be the case for the reduction in size during osteocyte “birth”. We specifically asked whether there was any evidence for upregulation of autophagy during the transition of osteoblast to osteocyte to recycle organelles no longer required.

S559

Materials and Methods: We used high pressure freezing (HPF) and conventional chemical fixation to prepare calvarial bones from newborn mice for ultrastructural examination and studied expression of the autophagy-related proteins LC3 and p62 by immunohistochemistry (IHC) in demineralised wax-embedded sections of newborn mouse and rabbit bone. Results: Ultrastructural analysis of more than 50 examples of osteocytes and osteoblasts showed that osteocytes embedded in calvarial bone appeared very similar to osteoblasts on the bone surface in terms of cellular content with abundant rough endoplasmic reticulum (RER), mitochondria and cytoplasm. Cellular features were then measured in randomly selected osteocytes and osteoblasts (n020 for each) and revealed no statistically significant differences between cell types in cytoplasmic area, area occupied by RER and numbers of mitochondria. Osteocytes did contain more autophagosomes than osteoblasts, but the variation between individual osteocytes was large (2–11 autophagosomes per cross section). IHC showed strong staining for LC3 and p62 in osteoclast and hypertrophic chondrocytes. Staining was also present in osteoblasts, not detected in transition cells and only in individual osteocytes, in keeping with the large variation seen in the autophagosome count in the ultrastructural analysis. Conclusion: We found no evidence for systematic induction of autophagy during the transition from osteoblast to osteocyte in calvarial bone, or long bones in young animals. We also conclude that in young calvarial bone, osteocyte organelle content is similar to that in osteoblasts, entirely consistent with their ability to synthesise a range of proteins in their role as regulators of bone remodelling. P24 ACTIVE SHAPE AND APPEARANCE MODELLING USED TO IDENTIFY STRUCTURAL CHANGES IN KNEE OSTEOARTHRITIS Susanna Delmonte (1) presenting Jenny Gregory (1) Richard Aspden (1) Kanako Yoshida (1) Rebecca Barr (1) David Reid (1) University of Aberdeen, Aberdeen, UK (1) Introduction: Development of knee osteoarthritis (OA) pharmaceuticals is restricted by the lack of methodology to measure disease progression. Current methods utilise joint space narrowing (JSN); however Active Shape and Appearance Modelling (ASM and AAM) offer a method to quantify the whole morphology of the knee joint, including features other than JSN. This study aimed to determine if ASM and AAM are sensitive to structural changes in the knee associated with OA severity, and to identify features other than JSN that could be used to measure progression of the disease. Material and Methods: Eighty-eight subjects with varying degrees of OA were selected. Knee DXA scans (iDXA GE

S560

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

Lunar) were obtained from each subject and assigned a Kellgren-Lawrence (KL) score from 0 (no OA) to 4 (severe OA): KL0 021; KL1 023; KL2 021; KL3&4 023. Each DXA scan was modelled by ASM and AAM using an 84 point template. Six scores of shape and appearance variance, ‘mode scores’, were assigned for each knee. Each mode represents an independent shape and set of characteristics, with mode scores representing how many standard deviations each knee lies from the average value (0) for each mode. Associations between modes and OA severity were tested using ANOVA with linear contrasts. Discussion: Shape modes 1, 2 and 6 (P<0.01) as well as appearance modes 1 (P00.026), 2, 4 and 5 (P<0.01) were linked to OA severity. In appearance mode 2, low mode scores were associated with severe OA and high mode scores associated with less diseased knees. Mode 2 represented several features associated with OA including medial JSN, misalignment of the tibio-femoral joint and medial osteophytes. Conclusion: This study indicates that ASM and AMM can identify associations between joint shape and OA severity, and are sensitive to features other than JSN which are affected by OA disease progression.

osteoblast outgrowths in this medium appeared later than with FCS. No osteoblasts were observed with K/O serum replacement. When osteoblasts from explants grown in FCS were maintained in test media, TheraPEAK increased (133.8 %±1.72 (SEM); p<0.001) whereas all other media decreased cell numbers to a similar extent (e.g. 85.6 %± 1.21; p<0.001 in human serum). All media increased ALP although the increase with TheraPEAK (997.8 %±44.26; p<0.001) was higher than with the other test media (e.g. human serum 232.7 %±10.2; p<0.001). Type 1 collagen expression was also significantly reduced in the presence of TheraPEAK. Other than with TheraPEAK (no mineralisation probably due to a reduced matrix), all test media resulted in good mineralisation. Similar results were obtained with human MSCs. Conclusions: We propose the following protocol to establish and maintain human osteoblasts in an ‘animal-free’ environment in vitro: i) explants in α-MEM plus human serum (1– 2 weeks), ii) seed and maintain osteoblasts in TheraPEAK (2 weeks) and iii) maintain osteoblasts in α-MEM plus human serum (1 week) prior to setting up assays. We plan to use this protocol for the study of primary human osteoblasts in 3D collagen gels.

P25 ESTABLISHMENT AND MAINTENANCE OF HUMAN OSTEOBLASTS IN VITRO WITHOUT THE USE OF ANIMAL-DERIVED MATERIAL Carole Elford (1) Deborah Mason (2) Jim Ralphs (2) John Gregory (1) Alastair Sloan (3) Bronwen Evans (1) presenting School of Medicine, Cardiff University, Cardiff, UK (1) School of Biosciences, cardiff University, Cardiff, UK (2) School of Dentistry, Cardiff University, Cardiff, UK (3)

P26 IN VITRO 3D OSTEOBLAST-OSTEOCYTE CO-CULTURE MECHANICAL LOADING MODEL Marisol Vazquez (1) presenting Bronwen AJ Evans (2) Sam Evans (3) Jim R Ralphs (4) Daniela Riccardi (4) Deborah J Mason (4) Arthritis Research UK Biomechanics and Bioengineering Centre, School of Biosciences, Cardiff University, Cardiff, UK (1) Department of Child Health, School of Medicine, Cardiff University, Cardiff, UK (2) School of Engineering, Cardiff University, Cardiff, UK (3) Division of Pathophysiology and Repair, School of Biosciences, Cardiff University, Cardiff, UK (4)

Introduction: Musculoskeletal replacement therapies require the use of media not containing animal-derived materials. It is also important to advance and implement the 3Rs (replacing, reducing, refining the use of animals in research and testing). We have developed an ‘animal-free’ protocol to establish/maintain human osteoblasts in vitro. Methods: Human tibial explants were set up in non-FCS containing media - 1) and 2) α-MEM plus human serum (Lonza), 3) α-MEM plus K/O serum replacement (Invitrogen), 4) TheraPEAK (defined medium; Lonza), 5) Stempro (defined medium; Invitrogen). α-MEM plus FCS was used as control. Osteoblasts established in this control medium were also maintained in test media and cell numbers (MTS), differentiation (qRTPCR, ALP activity) and mineralisation (alizarin red) measured. Furthermore, we investigated the effects of these media on human mesenchymal stem cells (MSC; Promocell). Disscusion: The best ‘animal-free’ supplement/medium for establishing osteoblasts was human serum, although

Introduction: Normal mechanical loading potently induces bone formation via effects on osteocytes. Current investigations of mechanical loading of bone do not reflect the interactions of the cells within it, with most focusing on mechanical loading of osteoblasts in monolayers. Of the 3D models that do exist, none elucidate the osteoblast-osteocyte interactions that regulate mechanically-induced bone formation. We developed a novel in vitro 3D co-culture model of bone1 and a loading device to investigate osteoblastosteocyte interactions. Methods: MLO-Y4 cells (1.5x106 cells/ml) were incorporated into acid-soluble rat tail tendon type I collagen (2 mg/ml in alpha MEM, pH7.4) gels and MC3T3-E1 (1.0x105 cells/well) layered on top and cultured at 37°C in DMEM (5 % FBS) for 1 week. Co-cultures were fixed with 1 % paraformaldehyde, infiltrated with OCT,

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

S561

cryosectioned and labelled with: 1) phalloidin and DAPI to assess cell morphology, 2) ethidium homodimer and DAPI to assess cell viability, 3) immunostained using anti-connexin 43 antibody to assess cell connectivity, or 4) immunostained with antibodies against osteoblast and osteocyte markers to assess phenotype. Osteoblast and osteocyte phenotypes were also determined by RT-PCR of RNA extracted (Trizol) separately from surface osteoblasts and encased osteocytes. Results: Data show co-cultures survive, for at least one week, with osteocyte cell death within gels averaging 16.86±3.56 % at day 1 and 14.11±2.69 % at day 7 comparable to monolayer cultures2. MLO-Y4 and MC3T3-E1 cells maintain their morphology; form a network through connexin 43, and express osteocyte and osteoblast phenotypic markers at mRNA and protein level. All data was obtained from 3 independent experiments of n03 or 4. A silicone based loading device was developed for the loading of these co-cultures and mineralisation conditions investigated. Discussion & Conclusions: We have established a mouse osteoblast-osteocyte 3D co-culture system and are developing a loading device to investigate mechanically-induced bone formation.

rugged surfaces. Subsequent histological examination revealed that adipocytes exactly fitted the contours of the bone surfaces. A role for adipocytes as bone lining cells has been suggested previously (Boyde, 2006), but there is no literature proposing a role for adipocytes in templating or moulding the surface of bone. Immunohistochemistry shows excrescences contain collagen VI, an ECM protein not normally present in mature bone matrix but which is known to be synthesized by adipocytes. Conclusions: These results suggest the deposition of the bone matrix within these mineralised excrescences is at least templated and modulated by adipocytes. However, the absence of osteoblasts from the forming surfaces raises the question of whether the matrix which is subsequently mineralised is partially synthesised by the adipocytes. Furthermore, it provides unequivocal evidence that adipocytes rather than inactive osteoblasts are the de facto lining cells on some bone surfaces. It is well recognised that adipocytes and osteoblasts can be derived from the same mesenchymal stem cells under the influence of different transcription factors. Furthermore there is evidence of transdifferentiation between the cell types. However, this is the first report of an intimate role for adipocytes in osteogenesis. Boyde A (2006). J Oral Biosci (Japan) 48 Suppl 57–59.

P27 IDENTIFICATION OF ADIPOCYTE MEDIATED OSTEOGENESIS Adam Taylor (1) presenting Lakshminarayan Ranganath (2) Jonathan Jarvis (2) James Gallagher (2) Alan Boyde (3) Lancaster University, LANCASTER, UK (1) University of Liverpool, LIVERPOOL, UK (2) Barts & the London School of Medicine & Dentistry, Queen Mary University of London, LONDON, UK (3)

P28 GENERATION OF PHYSIOLOGICALLY RELEVANT CELL LINES TO STUDY TNFRSF11A MUTATIONS ASSOCIATED WITH EARLY ONSET PAGETIC-LIKE SYNDROMES David Mellis (1) presenting Angela Duthie (1) Patrick Meraldi (1) Miep Helfrich (1) Julie Crockett (1) University of Aberdeen, Aberdeen, UK (1)

Introduction: It is the dominant concept that bone remodelling is a process in which resorption is coupled with subsequent formation by osteoblasts. We recently identified novel bone structures, termed excrescences, in human arthropathy samples. These appear in numerous forms; including new packets and projections laid upon resting trabecular surfaces. These structures are hypermineralised and often poorly integrated with pre-existing matrix. Others appear as partially resorbed trabeculae which have been smoothed over by new bone deposition. These structures appear to be formed by the non-coupled deposition of mineralised bone. Further study was instigated by the outline and shape of some of these structures not being congruent with typical osteoblastic bone formation. Materials & Methods: Using 3D SEM, qBSE-SEM & histology, we undertook further study of our catalogue of routinely prepared arthropathy samples. Discussion: SEM Analysis revealed close association of adipocytes with some excrescences, which showed indented,

Early onset Paget’s disease of bone (ePDB), familial expansile osteolysis (FEO) and expansile skeletal hyperphosphatasia (ESH) are related syndromes characterised by focal areas of increased bone turnover driven by bone-resorbing osteoclasts. They are caused by heterozygous tandem insertion duplication mutations within the signal peptide region of TNFRSF11a (encoding receptor activator of NFkB; RANK). Our studies have shown that when overexpressed alone (homozygous expression) the mutant proteins are sequestered within an extended endoplasmic reticulum and cannot translocate to the plasma membrane thus preventing RANKL-dependent signalling; in agreement with the osteopetrotic phenotype recently reported in the homozygous ePDB knockin mouse model. Given that patients are always heterozygous for these mutations we aimed to generate a physiologically relevant model to investigate the molecular consequences of these mutations in vitro. Bidirectional expression constructs containing cDNAs for both wild-typeRANK(wtRANK) and each mutantRANK cDNA

S562

were generated by Infusion PCR (Clontech) into the pBI-CMV vector (Clontech). The entire expression cassette was then excised and ligated into the pcDNA-FRT vector (Invitrogen). Given that each RANK cDNA was tagged with either myc or FLAG epitopes, we checked the subcellular localisation of each RANK protein when Hela cells were transfected with pcDNA-FRT-wtRANKFLAG-mutantRANKmyc constructs compared to “homozygous constructs” (pcDNA-FRT-mutantRANKFLAG-mutantRANKmyc or pcDNA-FRTwtRANKFLAG-wtRANKmyc). Heterozygous overexpression resulted in: detection of mutant protein at the plasma membrane; co-immunoprecipitation of wtRANK and mutantRANK, suggesting hetero-oligomer formation; and prolonged activation of signalling downstream of RANK compared to expression of wtRANK alone. We then used the Flp-In system (Invitrogen) to generate Hela cells lines expressing single copies of the bidirectional expression cassettes. Specific regions of genomic RANK DNA from each cell line have been PCR amplified and sequence verified to confirm the presence of each RANK construct. qPCR has confirmed that the RANK transcript is significantly more abundant in the RANK-transfected cell lines compared to the parental cell line. Finally, whereas the parental cell line does not respond to RANKL, cells expressing RANK constructs show increased nuclear translocation of p65. Taken together, these cell lines are an ideal model to further investigate the consequences of carriage of these heterozygous RANK mutations on RANKL-dependent signalling. P29 GENOME-WIDE ASSOCIATION STUDY OF PRIMARY TOOTH ERUPTION Ghazaleh Fatemifar (1) presenting Clive Hoggart (2) Inga Prokopenko (3) Momoko Horikoshi (4) Jon Tobias (5) Marjo-Riitta Jarvelin (2) David Evans (6) Department of Social and Community Medicine, University of Bristol, Bristol, UK (1) Department of Epidemiology and Biostatistics, Imperial College London, London, UK (2) Oxford Centre for Diabetes, Endocrinology and Metabolism, University of Oxford, Oxford, UK (3) Wellcome Trust Centre for Human Genetics, University of Oxford, Oxford, UK (4) Academic Rheumatology, University of Bristol, Bristol, UK (5) MRC Centre for Causal Analyses in Translational Epidemiology, School of Social and Community Medicine, University of Bristol, Bristol, UK (6) Introduction:Abnormalities of tooth development are one of the most common developmental malformations in humans affecting over 10 % of the population. Twin and family studies reveal that the timing of primary tooth eruption is highly heritable, with estimates of ~90 %. Materials and Methods: To identify variants involved in primary tooth eruption we performed a genome-wide-

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

association study of ‘time of first tooth eruption’ and ‘number of teeth at one year’ using ~6600 individuals from the Avon Longitudinal Study of Parents and Children and 5403 individuals from the 1966 Northern Finland Birth Cohort. We tested 2,589,439 imputed SNPs (HapMap PhaseII build36 release 22) common to both studies. All analyses were adjusted for gestational age, sex and age. Results from the two studies were combined using fixed effects inverse variance meta-analysis. Discussion: We identified eight loci reaching genome-wide significance (P<5X10-8) for ‘time to first tooth eruption’ and a further nine loci for ‘number of teeth at one year’. These loci included genes that play a role in tooth and other developmental pathways (KCNJ2, BMP4, EDA, MSRB3, IGF2BP2, HMGA2, CACNB2). The strongest association with ‘time of first tooth’ was SNP rs8080944 (P01.98E-33). This SNP is located down stream of the gene KCNJ2, which has been implicated in Andersen-Tawil syndrome. Individuals with this syndrome show abnormal tooth development. The strongest association for ‘number of teeth’ was rs8079702 (P01.22E-19), a SNP also located downstream of KCNJ2. Furthermore we report findings of a novel SNP in the protein-coding region of BMP4 (rs17563, P06.13E-16). This gene plays an important role in dental organ generation and morphogenesis as well as in the onset of bone formation, and reduction in its expression has been implicated in numerous bone diseases. Conclusion: Our results show that primary tooth eruption is a trait particularly amenable to mapping via genome-wide association, in that the common variants identified so far involve loci of unusually large effect (0.26 %-1.15 % of phenotypic variance). These findings suggest that a genomewide approach is a powerful strategy for detecting variants involved not only in dentition, but also skeletal growth and development. P30 DIFFERENTIATION OF OSTEOBLASTS TO OSTEOCYTES IN THREE DIMENSIONAL COLLAGEN GELS Nicole Scully (1) presenting Deborah Mason (2) Bronwen Evans (1) Institute of Molecular and Experimental Medicine, School of Medicine, Cardiff University, UK (1) Division of Pathophysiology and Repair, School of Biosciences, cardiff University, UK (2) Introduction:Bone remodelling, orchestrated by osteoclasts, osteoblasts and osteocytes, is a continuous process that maintains skeletal integrity. Osteocytes, derived from osteoblasts, are embedded in mineralised matrix and are difficult to isolate. To date there has been a dependence on the mouse MLO-Y4 cell line to study osteocyte biology. There is thus a

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

need to develop new methods to study these cells and at the same time mimic their 3D environment in vivo. Recent studies indicate that osteoblasts in 3D collagen gels may differentiate to osteocytes. Methods: We maintained mouse MC-3 T3 osteoblasts in type I collagen (rat tail tendon). Cells were set up in gels (250 μl; 48-well plates) and different seeding densities (3 x 105–2.3 x 106 cells/ml gel) assessed during 14 days. Cells were analysed for viability (trypan blue) and phenotype (sections or whole gels) by e.g. IHC (DMP-1, E11), qRTPCR (osteoblast and osteocyte markers), confocal microscopy (phalloidin and calcein labelling). Mineralisation was induced in some gels using a standard protocol. Furthermore, the effect of FGF-2 on cell differentiation was investigated. Discussion: MC-3 T3 cells set up in gels at the optimal seeding density (7 x 105 cells/ml gel) showed >85 % viability during 14 days. They appeared more dendritic when compared to monolayer cultures and formed connecting cellular networks. DMP-1 (osteoblast to osteocyte differentiation marker) was not expressed (IHC) at day 3 but then gradually increased in expression (days 7 – 14). E11 (IHC and qRTPCR; osteocyte marker localised to dendrites), however, was low at day 3, peaked at day 10, and returned to lower levels by day 14. Mineralisation was detected from day 7. FGF-2 significantly changed the cell differentiation patterns observed in the gels Conclusions: Our 3D gel methodology enables the in vitro maintenance of viable cells, and it is possible to mineralise these cultures. Osteoblasts maintained in gels differentiate along the osteocytic pathway. Furthermore, the addition of compounds such as FGF-2 will further enhance osteocytic differentiation. Differentiation of osteoblasts to osteocytes in 3D gels provides a potential new model to study osteocyte biology.

S563

phosphatase (ALP) expression was assessed at 7 days and bone nodule formation at 21 days on MSC cultured in the normal and osteogenic media, in the presence or absence of human peripheral blood mononuclear cells. Cytokine levels were measured by ELISA and Luminex, microarray analysis used Illumina whole human genome chips. Monocytes/Mφs potently induced MSC differentiation to OBs evidenced by increased ALP and mineralisation. The monocyte-induced osteogenic effect was mediated by cell contact as well as monocyte-derived soluble factors. This was not due to BMPs or TGFβ production or cytokines such as IL-6, TNF and IL-10 that were found to be up-regulated following monocyte:MSC contact. Monocyte:MSC cocultures also resulted in the up-regulation of PGE2 whilst inhibition of PGE2 production by COX2 inhibitor NS-398 resulted in the abrogation of monocyte-induced osteogenesis. However, PGE2 failed to directly induce osteogenic differentiation in MSC cultures indicating that it acts indirectly to induce the production of other factors by monocytes. Gene profiling microarray of monocytes identified Oncostatin M (OSM) as a mediator of monocyte-induced osteogenesis. The activation of STAT3 signalling by OSM in MSCs enhanced osteogenic differentiation; similar results were obtained using STAT3 constitutively active adenoviral overexpression. This study establishes a role for monocyte/Mφs as critical regulators of osteogenic differentiation via OSM production and the induction of STAT3 signalling in MSC. OSM, and other factors, induced by the contact between monocytes and MSC may be valuable new anabolic factors for use in osteoporosis and in localised bone remodelling during fracture and arthritis.

P31 MONOCYTES PROMOTE OSTEOGENIC DIFFERENTIATION OF MESENCHYMAL STEM CELLS VIA PGE2 AND ONCOSTATIN M Vicky Nicolaidou, Mei Mei Wong, Dilair Baban, Andy Cope, Nikki Horwood, The Kennedy Institute of Rheumatology, London, UK (1) Wellcome Trust Centre for Human Genetics, Oxford, UK (2) Kings College, London, UK (3)

P32 RISK OF GASTROINTESTINAL BLEEDINGS IN PATIENTS WITH TOTAL HIP / KNEE REPLACEMENTS AND MATCHED CONTROLS: A POPULATION-BASED COHORT STUDY IN DENMARK Arief Lalmohamed (1) presenting Peter Vestergaard (2) Corinne Klop (1) Marloes Bazelier (1)Anthonius de Boer (1) Frank de Vries (1) Utrecht University, Utrecht, The Netherlands (1) Aarhus University Hospital, Aarhus, Denmark (2)

Bone loss is a characteristic of chronic inflammatory and degenerative diseases such as rheumatoid arthritis and osteoporosis. A major challenge is how to replace bone once it is lost leading to the need for novel bone anabolic agents. The immune system is known to regulate bone as demonstrated by the ability of immune cells and cytokines to control the differentiation and activity of osteoclasts. However, less is known about the regulation of osteoblasts (OB). Mesenchymal stem cells (MSC) are multipotent progenitors that can be induced in culture to form OBs. Alkaline

Introduction: Gastrointestinal (GI) bleedings may impose a serious threat in patients undergoing total hip and knee replacements (THR/TKR). However, timing and risk factors of GI bleedings in THR/TKR patients versus matched controls have not been studied. Objectives were to derive time-dependent hazard ratios (HR) and determinants for GI bleedings following THR/TKR, against matched controls. Methods: In a nationwide Danish cohort study, we selected all patients with a primary THR/TKR between 1998 and

S564

2007 (n095,213). Three control subjects without THR/TKR were matched by age, sex, and region. We used timedependent cox models to calculate disease and medication adjusted (adj.) HRs for GI bleedings with THR/TKR versus controls. Results: We identified a 4.4-fold increased risk of GI bleedings during the first six weeks following THR (adj. HR 4.39; 95 % CI 3.79-5.09), versus matched controls. The risk remained elevated in the 6–12 weeks post-THR (adj. HR 1.61; 95 % CI 1.31-1.98), and dropped to baseline afterwards. The six-week risk was highest among NSAID users (10-fold increase). Among non-selective NSAID users, we found a lower risk in those concomitantly using proton pump inhibitors (PPI, 2.3-fold increase), as compared to those without a PPI (11-fold increase). In TKR patients, risk of GI bleedings was only increased in the first six weeks (2.3-fold increase). Conclusion: This study demonstrates an increased risk of GI bleedings during the first six weeks following THR (4.4fold), and TKR (2.3-fold). Risk assessment of GI bleedings may be considered during the first six weeks after THR/ TKR. The use of PPIs substantially reduces risk of GI bleedings when non-selective NSAIDs are prescribed to patients with a total joint replacement. P33 SERVICE EVALUATION: IMPROVING VERTEBRAL FRACTURE DETECTION RATE Rita Abdulkader (1) presenting Carolyn MacNicol (1) Karen Brixey (1) Sonya Stephenson (1) Gavin Clunie (1) Ipswich hospital NHS trust, Ipswich, UK (1) Introduction: Patients with vertebral fractures (VFs) are under-represented in Fracture Liaison Services (FLSs). Symptoms are often minimal with late presentation, even when radiographs are requested not all patients are identified owing to reporting variation. Between January and June 2010 only 16 patients with VFs were referred to our FLS. We initiated a pilot service to improve the referral rate. Material and Methods: Over a 6-month period, an arrangement was agreed with the Radiology Department to copy reports of VFs to FLS. Also ‘Osteoporosis risk assessment’ would be advised where ‘osteopenia’ alone was noted. In addition, a radiographer (FLS team) reviewed all reports on Thoracic and Lumbar spine radiographs weekly. When VFs were reported, patients <80y old were invited to FLS providing an osteoporosis assessment had not been done. . For patients ≥80y with VFs and all patients with osteopenia alone, management was delivered through a community nurse (CN). Discussion: 281 patients were identified. 157/281 (56 %) were <80y and 91/157 (58 %) had one or more VFs. FLS

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

was notified directly by Radiology for only 26/91 (29 %) patients. The other 65/91 (71 %) were identified by our radiographer. We invited 75/91 (82 %) patients to FLS. Sixty-three (84 %) attended, and 25/63 (40 %) were started on osteoporosis treatment (25/91 [27 %] of the original VF group). Existing treatment was changed in 3 (continued in 21). Of the 124 patients ≥80y, 80/124 (65 %) had VFs; FLS was only notified of 6. There were 57/80 (71 %) already on osteoporosis treatment. Treatment was started in 7/23 (30 %); the rest declining or further assessment was unsuitable. Of the 110 patients with osteopenia, 27 (25 %) were already on treatment. After assessment of the remaining 83, treatment was started in 14. Conclusion: In 6 months, 49 patients were identified for initial, or change of, osteoporosis treatment. The detection of patients with VFs requiring treatment was over 3 times the detection rate the year before. A proactive approach to reviewing radiograph reports is better in detecting VF cases, than relying on notification by The Radiology Department. P34 THE ROLE OF VERTEBRAL FRACTURE ASSESSMENT (VFA) IN ROUTINE CLINICAL CARE Louise Carter (1) presenting Virginia Gould (2) Jon Tobias (2,1) Emma Clark (2,1) North Bristol NHS Trust, Bristol, UK (1) Academic Rheumatology, Musculoskeletal Research Unit, Avon Orthopaedic Centre, Southmead Hospital, Bristol, UK (2) Introduction:Lateral DXA scanning for VFA is potentially useful in screening patients for VFs, but the context in which this should be used is currently unclear. The aim of this study was to audit how much the introduction of VFA into a routine DXA service changed management. Methods: Data was collected over an 8-month period. 1600 consecutive adults referred for a DXA were given a VFA if they were within the Sheffield protocol: (1) long-term steroids (2) hip fracture (3) >65 if female (4) >70 if male. VFAs were reported as being suspicious for VF, graded using quantitative morphometry (Genant method) and a recommendation was made that the referring clinician arrange spinal X-rays for confirmation. Results: Of the 1600 patients, 443 (27.7 %) had a VFA. 43 (9.7 %) were suspicious for VF with 69 % suspicious for one, 12 % for two and 19 % for multiple VFs, with the most common being Grade 2 fractures (44 %). Of the 443 patients who had VFA, 64 (14.4 %) had a normal BMD result (not recommended treatment) and 119 (28.9 %) had osteoporosis (recommended treatment), with 32 of the osteoporotics having a T-score≤−3.5. 260 (58.9 %) were osteopaenic. FRAX scores on all those with osteopaenia resulted in 79 recommendations for treatment. Of the 181 osteopaenics not recommended

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

treatment on the basis of FRAX alone, 14 (7.7 %) had VFAs that were suspicious for VF and where X-rays were requested 71.4 % were confirmed. Of the 32 with severe osteoporosis 7 had suspicious VFAs. Out of the 443 VFs performed, 15 (3.4 %) patients had changed management: 10/181 osteopaenic patients having a change in management from no treatment to treatment of their high fracture risk, and 5/32 patients with severe osteoporosis having alternative treatments considered such as anabolic therapies. Conclusion: VFA needs to be targeted more closely: those with osteopaenia who would otherwise not be treated; and those with severe osteoporosis where alternative treatments such as anabolic therapies may be indicated. Using this approach we would have performed 292 VFAs and changed management in 5.1 %. P35 BENEFICIAL EFFECTS OF PTH ON SPINE BONE MINERAL DENSITY (BMD) AND MICROARCHITECTURE (TBS) PARAMETERS IN POSTMENOPAUSAL WOMEN WITH OSTEOPOROSIS. A 2-YEAR STUDY Beatrice Günther1 (1) presenting Albrecht Popp1 (1) Delphine Stoll (2) Berengère Aubry Rosier (2) Romain Perrelet (1) Petr Kasalicky (3) Didier Hans (2) Kurt Lippuner1 (1) Osteoporosis Policlinic, University Hospital and University of Bern, Bern, Switzerland (1) Center of Bone disease, Bone and Joints Department, Lausanne University Hospitals, Lausanne, Switzerland (2) Mediscan Center, Prague, Czech Republic (3) Objective(s): Trabecular Bone Score (TBS, Med-Imaps, France) is an index of bone microarchitecture independent of BMD calculated from antero-posterior spine DXA scans. TBS was associated with fracture in prior case–control and prospective studies. In addition, an earlier study showed a positive maintenance of TBS in patients treated with alendronate while treatment-naïve controls were significantly losing bone microarchitecture. The aim of this study was to assess the effects of teriparatide, recombinant 1–34 human parathyroid hormone (PTH) on spine BMD and spine microarchitecture assessed by TBS in postmenopausal women with osteoporosis. Material and Methods: In this open label multi-center study (Bern, Lausanne and Prague, Centers of Bone Diseases), 111 women were treated with teriparatide (Forsteo®, Eli Lilly, USA) during 2 years. At the lumbar spine, BMD was assessed by DXA (Hologic Discovery) and TBS parameters were assessed by TBS iNsight (v1.9) at baseline and after 24 months of treatment. ISCD-like rules for individual vertebrae exclusion were applied independently for BMD and TBS. The analysis was by ITT. Results: Baseline characteristics (mean±SD) were similar between groups in term of age, 69.8±10.3 years; BMI, 24.1±

S565

4.4 kg/m2; T-score L1-L4 BMD, - 2.9±1.4 and TBS 1.190± 0.11. The correlation between BMD and TBS at the lumbar spine was very low (r200.13). Over 2 years, L1-L4 BMD increased significantly by +8.44 % (p<0.001) and Spine TBS increased by +4.33 % (p<0.001). At 2 years, there was no correlation between the changes in BMD and TBS from baseline. Conclusion(s): In postmenopausal women with osteoporosis, a 2-year treatment with teriparatide lead to an independent increase in BMD and TBS at the lumbar spine, suggesting that teriparatide has independent positive effects on spine bone mass and microarchitecture. P36 TREATMENT OF SEVERE OSTEOPOROSIS WITH TERIPARATIDE IS ASSOCIATED WITH IMPROVED CLINICAL OUTCOME AS COMPARED WITH STANDARD CARE Ailsa Oswald (1) presenting Jackie Berg (2) Gina de'Lara (2) Stuart Ralston (1,2) University of Edinburgh, Edinburgh, UK (1)NHS Lothian, Edinburgh, UK (2) Background: Teriparatide (TPTD) is an anabolic agent which in the UK is usually reserved for the treatment of patients with severe osteoporosis who fail to respond to other treatments. Our referral centre has routinely been able to offer TPTD as a first line therapy for the treatment of severe osteoporosis as well as for rescue therapy for patients who are intolerant of or fail to respond to other drugs. Here we report upon our clinical experience with this agent, as compared with those who received treatment with antiresorptive therapy. Patients and Methods: The study group comprised 376 patients with severe osteoporosis (T-score<−4.0); 258 (70.2 %) were treated with TPTD followed by antiresorptive therapy whereas 118 (29.8 %) received standard care (SC) with various antiresorptive drugs including Alendronate (66.1 % of cases), Zoledronic acid (17.8 %) and Risedronate (11 %). The reasons that patients received SC as opposed to TPTD were that they declined or were unable to self inject (57.6 %), or that they had already been prescribed antiresorptive therapy by the GP (33.1 %). Results: The treatment groups were matched for severity of osteoporosis; mean±sem lumbar spine T-score was −4.3±0.7 the TPTD group vs. -4.4±0.6 in the SC group (p00.42). 88 % of the TPTD and 89 % of the SC group had suffered fragility fractures (p00.90); most had postmenopausal osteoporosis (82.2 % TPTD vs. 78.8 % SC) and most of the TPTD group (79 %) were treatment naïve. The duration of follow up in both groups was about 36 months. The increase in spine BMD (mean % change±sem) was slightly greater in the TPTD group as compared with SC (7.1±0.47 vs. 5.8±0.59) but this was not significant (p00.09). There was no difference in response of femoral neck BMD (data not shown). Significantly fewer TPTD treated patients suffered fragility fractures during follow up compared with SC treated patients (5 % vs. 15.3 %, p<0.001).

S566

This was mainly driven by a reduction in clinical vertebral fractures (1.2 % vs. 5.9 %, p00.008), but was also significant for non-vertebral fractures (5.9 % vs. 9.3 %, p00.03). Conclusions: Although this was an observational study, our findings suggest that TPTD therapy may improve clinical outcome in patients with severe osteoporosis. The lack of a difference in BMD response between the TPTD and SC groups suggests that this may be due in part, to an effect on bone quality. P37 DO HEALTH CARE PROFESSIONALS APPROPRIATELY ADVISE MOTHERS AND CHILDREN UNDER 5 ABOUT THE IMPORTANCE OF VITAMIN D SUPPLEMENTATION ACCORDING TO NATIONAL GUIDELINES? AN INVESTIGATION Sara Patience (1) presenting Surrey University, Surrey, UK (1) Introduction: It has been suggested that Health Care Professionals are failing to give appropriate advice to mothers and babies about vitamin supplementation, specifically vitamin D. The principle aim of this project was to review the service given by primary health practitioners by examining supplementation habits in mothers and babies and children under 5 yrs who attend well baby clinics. This was achieved through face to face questionnaires, and investigated the source of advice received, assessed each participant for the need for supplementation using national guidelines, examined whether specific ethnicities were more likely to take a supplement containing vitamin D or have been advised by a Health Care Professional, and tested parents knowledge about the source of vitamin D and consequences of deficiency. An additional question was asked about maternal experience of bone pain. Methods: A total of 65 parent and baby/child pairs completed a questionnaire whilst attending Health Visitor well baby clinics at various locations in a London borough. Information was obtained about ethnicity, supplement use for mothers and babies/ children, baby/child feeding methods, parental knowledge of vitamin D and maternal experience of bone pain. SPSS v.18.1 was used to analyse the data using descriptive statistics such as frequency and crosstabs and non-parametric test, Chi-squared. Results and Discussion: A total of 70 % of the mothers took an appropriate antenatal supplement containing vitamin D in pregnancy. 30.7 % continued to take an appropriate postnatal supplement, of which 11.1 % were advised by a healthcare professional. 63.1 % of babies and children under 5 were not appropriately supplemented, (either through maternal supplementation or directly supplemented). 24 % of babies received advice from a healthcare professional. Being over 6 months of age was significantly associated (P<0.03) with receiving advice from a healthcare professional. Although the ethnicity in the study reflected the borough averages the numbers were too small to find statistical significance. It was noted however that whilst Asian

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

mothers were comparatively more likely to take an antenatal supplement containing vitamin D, (83.3 %) compared to white (75 %) or black (61.5 %) mothers, Asian mother were least like to take a postnatal (PN) supplement (16.7 %) compared to black (30.8 %) or white (37.5 %) mothers, only one mother covered for religious/cultural reasons took a PN supplement. For the babies, 62.5 % of white babies, 61.5 % of black babies and 66.7 % of Asian babies did not receive an appropriate supplement, it would appear that no consideration is given to ethnic risk. Conclusion: Healthcare Professionals are not appropriately advising mothers and babies/children to take vitamin D supplementation according to national guidelines. P38 THE IMPACT OF A NEW MODEL OF HIP FRACTURE CARE AT A TEACHING HOSPITAL Rahul Bhattacharyya (1) presenting Yuvraj Agrawal (1) Heather Elphick (1) Chris Blundell (1) Northern General Hospital, Sheffield Teaching Hospitals NHS Trust, Sheffield, UK (1) Introduction: In August 2010, a new model of shared care for hip fracture patients was implemented at the Northern General Hospital, Sheffield. In the new model, 48 hours post surgery, patients are allocated to an orthogeriatric consultant led team who review patients daily to manage medical comorbidities and coordinate multidisciplinary rehabilitation, with orthopaedic input if necessary. This differed from the previous model where these patients remained under the care of the orthopaedic teams until discharge. Aim: To compare the two models of care as perceived by different members of staff and compare clinical outcomes. Material and Methods: Prospective data was collected using questionnaires from medical, nursing and allied health professionals. Their opinions were rated using the Likert scaling system and analysed with the Mann Whitney U-test. In the second part of the study, clinical outcomes were obtained from the local hip fracture database. Discussion: 59 responses; 21 doctors (Group A) and 38 nurses and allied health professionals (Group B). 91 % of group A and 79 % of group B responses agreed that the quality of patient care has improved compared to the previous model. (p0.13, r0.19). 91 % of group A and 84 % of group B preferred to work in the new model (p0.31, r0.13). The mean length of stay in the previous model (pre-August 2010, n0 274) was 30 days compared to 25 days in the new model (post August 2010, n0249). (p00.2). 55.83 % of patients returned to their source of admission in the previous model compared to 72.69 % in the new model (p00.00007). Conclusions: The majority of staff perceived the new model to have improved quality of patient care and preferred to work in the new model. The new model had a shorter mean length of stay and improved in-patient rehabilitation as

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

indicated by significantly higher rates of return to the source of admission. This model can pioneer a change in the system of management of hip fracture patients nationally. P39 RURAL TO URBAN MIGRATION IN INDIAN WOMEN IS ASSOCIATED WITH SUBSTANTIAL GAINS IN HIP AND SPINE BONE MINERAL DENSITY Heli Viljakainen, Yoav Ben-Sholomo, Sanjay Kinra, Hannah Kuper, Jon Tobias Musculoskeletal Research Unit, University of Bristol, Bristol, UK (1) Children's Hospital, University of Helsinki, Helsinki, Finland (2) School of Social and Community Medicine, University of Bristol, Bristol, UK (3) Department of Epidemiology and Population Health, London School of Hygiene & Tropical Medicine, London, UK (4) Introduction: We used the sib-pair design of the Indian Migration Study as a novel method to investigate the contribution of obesity to BMD, by examining whether those who had migrated from a rural to an urban environment had a higher BMD than participants who had remained in a rural environment, and whether any difference is explained by fat mass. Material and Methods: Indian rural–urban migrants (RUM) were matched with rural non-migrated siblings (RNM) recruited as controls. The present analysis focused on female participants undergoing a clinical examination at, Hyderabad Southern India, along with blood sampling, anthropometric and DXA measures. Data on other lifestyle factors and standard living index (SLI) were collected. Lumbar spine (LS) and femoral neck (FN) BMD was compared between RUM and RNM with or without adjustment for confounders by linear regression. Discussion: 235 RUM and 114 RNM female participants (mean 47 years of age) were identified. After age-and height-adjustment LS BMD was 0.25 (95 % CI: 0.050.46)SD (p00.015) and FN BMD 0.36 (0.16-0.57)SD (p <0.001) higher in RUM compared RNM. Of the baseline factors BMI, fat mass, fasting insulin, parity, and proportion of subjects in highest SLI differed between groups: 27.3 and 25.1 kg/m2, 24.5 and 20.8 kg, 9.6 and 7.1 mU/ml, 2.8 and 3.1, and 99.5 and 78 %, in RUM and RNM, respectively. In the fully adjusted model (age, height, fat mass, parity, insulin, PA, SLI) the differences between groups were largely attenuated: 0.07 (−0.17-0.49)SD (p00.569) and 0.09 (−0.14-0.32) SD (p00.431) in LS and FN BMD, respectively. The most important appeared to be fat mass; after adjusting for this alone the difference decreased to 0.13 (−0.080.33)SD (p00.233) and 0.18 (−0.02-0.38)SD (p00.075) in LS and FN BMD, respectively, suggesting this explained a substantial proportion of the difference in

S567

BMD between groups. The second best predictors were SLI and parity; adjustment for these factors led to a reduction in differences at the LS and FN of 15 % and 19 %, respectively. Conclusions: Rural to urban migration is associated with a marked increase in LS and FN BMD, which reflects a range of factors, of which higher fat mass consequent on urban migration makes the greatest contribution. P40 IDENTIFICATION OF JOINT OCHRONOSIS IN ALKAPTONURIC MICE: A NEW EXPERIMENTAL MODEL OF OSTEOARTHROPATHY Craig Keenan (1) presenting Andrew Preston (1) Hazel Sutherland (1) Adam Taylor (2) Brenda Wlodarski (1) Peter Wilson (1) Lakshminarayan Ranganath (1) Dominic Williams (1) James Gallagher (1) Jonathan Jarvis (1) University of Liverpool, Liverpool, Merseyside, UK (1) Lancaster University, Lancaster, Lancashire, UK (2) Introduction: Alkaptonuria (AKU) is a rare autosomal recessive disorder with loss of function of homogentisate dioxygenase (Hgd) characterised by the accumulation of homogentisic acid (HGA) in plasma and urine. Although HGA is readily excreted, its concentration in plasma rises, and is gradually deposited as polymerised pigment in collagenous tissues, principally in the cartilage of loaded joints. This process (ochronosis) is associated with early onset severe osteoarthritis. Previous studies suggested AKU mice do not develop ochronosis. This study was designed to make a comprehensive survey of AKU mice, over the life-span and to use new histological methods to identify pigmentation. We also investigated whether nitisinone, a potential therapy for AKU, prevents the initiation and progression of ochronosis in mice. Materials and Methods: Hgd−/− mice on BALB/c and C57BL/6 backgrounds were culled between 4 and 18 months for whole life studies. For the nitisinone study, 4 mg/L nitisinone was added to the drinking water of 15 out of a group of 30 mice. Histological evaluation of the medial and lateral tibio-f em oral joints was performed with haematoxylin-eosin (H&E) and Schmorl’s staining. Discussion: We report the detection of early stage ochronosis in AKU mice, on BALB/c and C57BL/6 backgrounds. Ochronosis in the hyaline articular cartilage and the calcified articular cartilage was synonymous with early ochronosis in patients with AKU. Pigmentation of individual chondrocytes and their associated territorial matrix could be seen using Schmorl’s reagent, a modified stain for melanin-like pigment. Histological analysis showed that nitisinone prevented pigmentation of chondrocytes and

S568

associated territorial matrix, both of which are signs of ochronosis. The untreated mice all had varying degrees of pigmentation. Conclusion: The observation of early stage ochronosis in the AKU mice is significant as it provides a model to investigate the pathological change associated with AKU that progresses to joint degeneration. Furthermore AKU mice provide a new model to investigate initiating events in osteoarthritis. Our results also show that nitisinone, when given for the life-span, prevents ochronosis in AKU mice. P41 WHEN SIZE AND SPECIES MATTER: POSSIBLE BIPHASIC EFFECT OF SI NANOPARTICLES AND NON-MONOMERIC SPECIES ON THE OSTEOGENESIS OF HUMAN BONE MARROW STROMAL CELLS Priya Kalia (1,4) presenting Ravin Jugdaohsingh (2) Andrew Brown (3) Roger Brooks (4) Neil Rushton (4) Biomaterials, Biomimetics and Biophotonics, The Dental Institute, King's College London, London, UK (1) MRCHuman Nutrition Research, Cambridge, UK (2) Institute for Materials Research, Faculty of Engineering, University of Leeds, Leeds, UK (3) Orthopaedic Research Unit, Dept. of Surgery, University of Cambridge, Cambridge, UK (4)

Figure 1.

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

Introduction: Silicon (Si) has stimulatory effects on bone development and repair, yet its mechanism of action is poorly understood. Silicate solutions >2 mM contain nonmonomeric silicate ions and polymers (colloidal particulates). Our hypothesis was that different silicate species, similar to those released by Si-containing bioceramics, may have different effects on osteogenesis. Results: A reduction in monomeric Si was observed in Sicontaining cell culture media containing ≥2 mM (Figure 1), showing the presence of non-monomeric Si species (Fig. 1a). Transmission electron microscopy and electron energyloss spectroscopy confirmed the presence of differently sized Si nanoparticles in the 6 and 42 mM silicate solutions. When the silicate solutions were diluted to 1 mM Si and cultured with bone marrow stromal cells to observe speciesspecific effects, cell metabolism and cell proliferation increased significantly with increasing Si in the silicate solution (Fig. 1b), whereas, interestingly, ALP activity decreased (Fig. 1c). Runx-2 gene expression, an early osteogenic marker, was significantly increased at 6 mM Si, while Runx-2 and osteoprotegerin (OPG) decreased with increasing Si (12–42 mM) in the silicate solution (Fig. 1d). Conclusion: These results suggest that different silicate species may have different actions on osteogenesis and bone formation.

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

P42 OUTCOME OF OSTEOPOROTIC PELVIC FRACTURE IN 50 OLDER PATIENTS Sarah Marrinan (1) presenting Sheena Waters (1) Yousif Shanshal (1) Gateshead Health NHS Foundation Trust, Gateshead, UK (1) Introduction:The incidence of osteoporotic pelvic fractures in older people is increasing. With an ageing population this trend looks set to continue. Material and Methods: We prospectively collected data on 50 consecutive older patients with pubic rami fractures admitted to medical wards in a district general hospital. Data collection began in April 2010, ending one year later. Discussion: The mean age of patients was 84.3 years (median: 85, range: 64 to 98 years). The vast majority of patients were female (88 %). Half of patients had a previous fragility fracture but only 18 (36 %) were taking a bone-sparing agent at the time of their pelvic fracture. Acute medical problems (mainly infections) were common, affecting 70 % of patients. In-hospital mortality was 8 % (n04). Of the 46 patients who survived to discharge, a further 5 patients died in the three months following their fracture giving a total three month mortality rate of 18 %. The mean length of stay in hospital was 30 days (excluding those who died in hospital). For the 46 patients who survived to discharge, we compared mobility and accommodation type pre and post fracture. Levels of mobility were categorised into: ‘independent,’ ‘using walking aid’ or ‘dependent.’ We saw a major reduction in the number of patients mobilising independently upon discharge (8,7 % vs 41.3 % on admission) and accordingly more patients needed a walking aid on leaving hospital (78.3 % vs 56.5 % on admission). There was also a dramatic increase in people whose mobility was entirely dependent (13.0 % vs 2.2 % pre-fracture). Accommodation type was similarly categorised as independent or dependent. On admission 87.0 % of patients lived independently, on discharge this figure fell to 69.6 %. Conclusion: Osteoporotic pelvic fractures are typically managed conservatively but they should not be overlooked as a significant cause of mortality and morbidity in older patients. The increase in physical dependence and reduction in mobility seen with this fracture may have significant psychosocial and financial implications for health care provision at local and national levels. P43 THE EFFICACY AND SAFETY OF IBANDRONIC ACID FOR INTRAVENOUS ADMINISTRATION FOR TREATMENT OF SEVERE SYSTEMIC OSTEOPOROSIS IN PATIENTS WITH JUVENILE ARTHRITIS Ekaterina Alexeeva (1) presenting Rina Denisova (1) Tatyana Bzarova (1) Saniya Valiyeva (1) Tatyana Sleptsova

S569

(1) Elena Mitenko (1) Alexandra Chomahidze (1) Scientific Centre of Children’s Health, Moscow, Russia (1) Relevance: juvenile arthritis (UA) is one of the most common and disabling rheumatic disease in children. Severe manifestation of juvenile arthritis is a systemic osteoporosis. There is currently no developed approaches to the treatment of children with systemic osteoporosis. In adult patients with rheumatoid arthritis the firstline treatment drug for osteoporosis is bisphosphonates Objective: To evaluate the efficacy and safety of ibandronic acid for intravenous administration for treatment of severe systemic osteoporosis in patients with juvenile arthritis. Patients and Methods: The study included 25 patients with juvenile arthritis at the age of 7 to 17 years (10 girls, 15 boys), including 20 patients with systemic arthritis, 3 - with polyarthritis, 2 - with ankylosing spondylitis. The mean age was 12.5 (10.5, 14.0) years. Was allocated to 2 groups: 17 children treated with glucocorticoids (GC) and 8 - not receiving GC.Children of the first group had 12 vertebral fractures and patients of the 2-d group - 1. Ibandronic acid was administered intravenously at a dose of 3 mg every 12 weeks. Evaluation of treatment efficacy was conducted by a combined index of bone mineral density of tissue Z-score, assessing the patient / parent to the severity of pain on a visual analog scale, serum markers of bone resorption C-terminal telopeptide. Results: The treatment of ibandronic acid in children of both groups was marked increase in bone mineral density. In patients treated with GC, a statistically significant increase was registered in 76 weeks of therapy (p <0.01), while the second group of children - through 46 and 76 weeks (p <0.05). Within 6 months of treatment in both groups of ibandronic acid was a statistically significant reduction in pain index (p <0.05), after a year of treatment with this trend persisted (p <0.001). In both groups, reducing the concentration of C-terminal telopeptide in serum were detected through 52 weeks of reatment (p <0.05). After 76 weeks of treatment new vertebral fractures and fractures of the peripheral skeleton are not fixed. Tolerability of therapy ibandronic acid in patients was satisfactory. Conclusion: In the course of the study has identified high efficacy and an acceptable tolerability ibandronic acid for treatment of severe systemic osteoporosis in patients with juvenile arthritis. P44 W H AT S H O U L D B E T H E I N T E RV E N T I O N THRESHOLDS OF TRABECULAR BONE SCORE (TBS) WHEN USED AS A MAJOR CLINICAL RISK FACTOR (CRF) OF OSTEOPOROTIC FRACTURES (OPS)? A META-LIKE ANALYSIS Didier Hans (1) presenting Renaud Winzenrieth (1) Berengere Aubry-Rozier (1) Delphine Stoll (1) Olivier Lamy (1) MarcAntoine Krieg (1) Bone and Joint Department, Center of

S570

Bone Diseases, Lausanne University Hospital, Lausanne, Switzerland (1) More and more guidelines from medical societies give to CRF an important role to weight on the individual risk and thus on the medical decision. TBS predicts fracture independently of BMD as demonstrated from previous prospective studies. The aim of the study is to verify if TBS can be considered as a major CRF of osteoporotic fracture and to determine its intervention thresholds. Existing validated datasets of Caucasian women were reanalyzed in this meta-like analysis. These different datasets were weighted differently according to their design. This study involves more than 33000 women (≥50 years) with 2200 osteoporotic fractures from three prospective studies (OFELY, MANITOBA, SEMOF) and 12 cross-sectional studies including the OsteoLaus and Osteo-Mobile studies. Weighted relative risk (RR) for TBS was expressed for each decrease of one standard deviation and compared with those obtained for the major CRF included in FRAX®. TBS intervention thresholds were evaluated using a tertile approach. TBS thresholds obtained were 1.195 and 1.301 for the lowest and the highest tertiles respectively. Overall TBS RR after adjustment for age was 1.81 [95%CI-1.35–2.47]. For all women combined, RR for fracture for the highest compared with the middle TBS tertile was 1.8 and for the highest compared with the lowest TBS tertile was 2.8. Besides, the TBS lowest tertile was reached at 75 years as the BMD WHO cut-off point of −2.5 T-score. TBS is comparable to most of the major CRF (Fig 1) and thus could be used as one of them. Defined thresholds seem to be consistent amongst the different studies and could be a starting point for clinical use.

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

P45 LYCOPENE DOES NOT ALTER BONE TURNOVER IN HEALTHY MEN AND WOMEN - A RANDOMIZED CONTROLLED TRIAL Claire Macdonald (1) presenting Lindsey Masson (1) Garry Duthie (1) Amelia Rudd (1) Nicholas Vaughan (1) Frank Thies (1) Helen Macdonald (1) University of Aberdeen, Aberdeen, UK (1) Introduction: Lycopene, mainly originating from tomatoes, has been associated with lower incidence of osteoporosis, and high dietary lycopene intake has been reported to protect against fracture in cross sectional and longitudinal studies. A pilot intervention study showed that high serum lycopene was associated with decreased bone resorption in a population of postmenopausal women. The aim of this investigation was to determine the effects of lycopene intake on percentage change of bone turnover markers amino-terminal procollagen propeptides of type 1 collagen (P1NP) and C-terminal telopeptide of type 1 collagen (CTX). Materials and Methods: A randomized controlled trial (RCT) was conducted in a healthy population of men and women (N0239) aged 29-69y (mean age [±SD] 51.1 [±6.8] years). Participants were randomly assigned to 3 groups: 1. high lycopene/tomato diet (minimum 10 mg/day), 2. 10 mg lycopene supplement and 3. low lycopene/tomato diet. The serum bone formation marker P1NP and plasma bone resorption marker bone CTX were measured at baseline (week 4) and at 16 weeks for each of the participants. Investigators were blinded to the treatment groups. Discussion: Two hundred and twenty seven participants completed the final visit (95 %). Serum and plasma was available for 221 participants. One extreme outlier was removed leaving 76 in group 1, 66 in group 2, and 78 in group 3, for analysis. One-way ANOVA showed no significant difference in percent change in bone turnover between groups (mean [±SD]): plasma CTX: +7.7 % [±27.2 %], +6.8 % [±23.9 %] and +5.4 % [±26.4 %] and serum P1NP: -1.0 % [±24.3 %], +4.26 % [±22.0 %] and +2.8 % [±15.4 %] in groups 1, 2 and 3 respectively. Conclusion: These data suggest that a twelve-week lycopene supplementation or a diet high in lycopene does not alter bone turnover in a healthy population of men and women. Future research is required to determine whether lycopene effects bone turnover longer term or attenuates bone resorption in groups at high risk of osteoporosis. P46 THE INCIDENCE OF PREVIOUSLY UNDIAGNOSED CONDITIONS IN PATIENTS ATTENDING FRACTURE LIAISON SERVICE Rita Abdulkader (1) presenting Sonya Stephenson (1) Gavin Clunie (1) Ipswich hospital NHS trust, Ipswich, UK (1)

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

Introduction: The importance of a fracture liaison service (FLS) in secondary fracture prevention is well recognised. In a FLS, identification of medical conditions affecting bone metabolism, aims not only at improving bone health but also general wellbeing. We report the incidence of new medical conditions identified through routine investigations in FLS. Material and Methods: All patients who attended FLS clinic have an assessment including history, DXA scan, blood tests (U&Es, bone profile, PTH, thyroid function test, Vitamin D, CRP, and in patients >70y, serum protein electrophoresis. Male patients are tested for hypogonadism. We retrospectively reviewed patients attending FLS clinic identified with a previously-undisclosed medical condition between November 2008 and October 2011.We reviewed FLS and any subsequent clinic letters and the Hospital laboratory database. We did not include new cases of vitamin D insufficiency. Discussion: Over this 3-year period 74/2876 (2.5 %) patients who attended the FLS (age 50-80y) were identified to have a previously undisclosed medical condition relevant to poor bone health. 28 (1 %) were found to have primary hyperparathyroidism (PHPT) and 5 patients were found to have paraproteinemia (3 cases of monoclonal gammopathy of unknown significance and 2 still under investigation). Hypogonadism was identified in 19 men (15 primary hypogonadism, 4 secondary [2 with hyperprolactinemia]). Otherwise cases of: hyperthyroidism (4), celiac disease (1), hypothyroidism (11), inflammatory arthritis (2) and chronic kidney disease 3–5 (2) were identified. In 2 patients review was inconclusive owing to loss of follow-up; one may have had PHPT and a male patient with possible hypogonadism.

Measurement (reference range) 25(OH)D (50–125 nmol/L) Se Ca (2.2–2.6 mmol/L) PTH (14–72 ng/L) Alkaline phosphatase (30–130 IU) U Ca (2.5–7.5 mmol/24 hours)

S571

Conclusion: In our FLS, routine and simple screening disclosed new medical conditions, relevant to bone health in 2.5 % cases. Notably the incidence of PHPT in fracture patients >50y is at least 10x higher than it is in the general population (1 % vs 0.1 %); 1 % an underestimate as screening was not designed to capture any new PHPT in those >80y. P47 HIGH DOSE COLECALCIFEROL IN THE TREATMENT OF VITAMIN D DEFICIENCY Nicola Peel (1) presenting Mary Bull (1) Jennifer Walsh (1) Sheffield Teaching Hospitals NHS Foundation Trust, Sheffield, UK (1) High dose supplementation is frequently used in the treatment of vitamin D deficiency. Most regimes, however, are arbitrary and have not been systematically evaluated. In our clinical practice, patients with 25(OH)D <30 nmol/L are treated with oral colecalciferol 100,000 IU/month for 3 months (4 doses) in addition to calcium (1000 to 1200 mg) and colecalciferol (800 IU) daily, with a repletion target of 50 to 125 nmol/L. Our aims were to examine the safety and efficacy of this approach and examine whether serum PTH identifies those individuals who will benefit from treatment. We evaluated 42 patients (ages 19 to 93) without evidence of malabsorption, primary hyperparathyroidism or renal impairment (CKD 4 & 5). Forty-one patients achieved repletion by 3 months. The remaining patient, on long-term anticonvulsants, achieved a level of 48.6 nmol/L and received further high dose treatment.

Baseline mean (% out of range) 21.1 (100% low) 2.27 (26% low) 123.9 (64% high) 101.7 (17% high) 2.1 (73% low)

PTH decreased during therapy, even in the 15 (36 %) patients with normal baseline levels (17 % decrease,] P0 0.03). PTH remained high after D repletion in 13 patients with one showing biochemical evidence of primary hyperparathyroidism. Post-treatment 25(OH)D exceeded 125 nmol/L in 5 (12 %) patients (125.6 to 235 nmol/L) but none developed hypercalcaemia. Four of these patients had low BMI (16.6 to 18.5 kg/m2). We conclude that 1) this regime enables vitamin D repletion to be achieved in the majority of patients

3 months mean (% out of range) 97.8 (2% low, 12% high) 2.34 (1% low, 1% high) 69.6 (31% high) 86.3 (14% high) 3.0 (31% low)

P paired t-test <0.0001 <0.0001 <0.0001 0.006 0.003

and 2) serum PTH is not a useful surrogate to identify patients who will benefit from vitamin D supplementation. P48 EFFECT OF ORAL AND IV BISPHOSPHONATE TREATMENT AND DISCONTINUATION ON LEVELS OF URINARY NTX Diane Powell (1)Sally Evans (1) presentingMichael Davie (1)Clare Matthews (1)Mark Garton (1)Robert Jones &

S572

Agnes Hunt Orthopaedic Hospital NHS Foundation Trust, Oswestry, Shropshire, UK (1) Introduction: Bisphosphonate (BP) holidays have been suggested as a way to reduce the risk of side-effects. The mechanism for monitoring and when to reinitiate treatment has not been established: therefore we looked at the effect of BP treatment and discontinuation on bone turnover in an osteoporosis service setting. Material and Methods: Urinary N-telopeptide crosslink of type I collagen (uNTx) levels were measured in women on IV BP at 3 months (±14 days) post infusion and in women receiving oral BP. Data were also collected on 8 subjects on Zoledronate (Zol) at clinic review (−0.21±0.7 years post infusion (YPI)) and at follow-up (1.15±0.36 YPI), 25 subjects on Pamidronate (Pam) 0.42 YPI (range 0–0.9) and 1 YPR (range 0.8-2.2) and 16 subjects on Alendronate (ALN) prior to discontinuation and 1 year later (range 0.42-2). Discussion: The ages of the different BP groups were similar. uNTx levels were higher in patients on oral BP (n0818) compared to IV BP (n0167) (29.8 vs.18.6; p<0.001). Of the patients on IV BP uNTx levels were lower in those on Zol (n0134) compared to those on Pam (n033) (18.6 vs. 25.4; p<0.005). Of the patients on oral BP uNTx levels were lower in patients on Alendronate (ALN; n 0572) than those on Etidronate (n039) (28.3 vs. 35.4; p<0.05) or Risedronate (n0207) (28.3 vs. 32.9; p<0.001). When ALN was compared to the IV BP, uNTx levels were lower in the Zol group (28.3 vs. 18.6; p<0.001). In the patients on a BP holiday uNTx levels at discontinuation or clinic review were not significantly different to the treatment groups above. At follow-up, approx 1 year postdiscontinuation, uNTx levels increased in subjects on Zol (31.0 vs. 20.3; p<0.014) and ALN (45.4 vs. 28.7; p<0.005) but not Pam (31 vs. 29; ns). Conclusion: The levels of uNTx in patients on BP treatment is dependent on the BP given with Zol having the greatest effect and risedronate and etidronate the weakest. uNTx levels increased post discontinuation in subjects who had received Zol and ALN but not PAM: however, mean levels in all 3 groups were still in the premenopausal range after 1 year. P49 QUANTITATIVE ULTRASOUND AND BONE’S RESPONSE TO EXERCISE: A META ANALYSIS Opeyemi. O Babatunde (1) presenting Jacky. J Forsyth (1) Centre for Sports, Health & Exercise Research, Staffordshire University, Stoke-on-Trent, UK (1) Introduction: The restrictions to bone density testing and the enduring debate over repeat dual energy absorptiometry testing spells uncertainty over both clinical and non-

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

clinical monitoring of responses to exercise intervention for prevention and management of osteoporosis across the age spectrum. This study, via a systematic review and metaanalysis, aimed to paint a portrait of current evidence regarding quantitative ultrasound (QUS)’s application to monitoring and evaluation of bone’s adaptive response to exercise interventions. Material and Methods: A structured and comprehensive search of databases was undertaken along with handsearching of key journals and reference lists to locate relevant studies published up to September 2011. Twelve articles (2 cohorts and 10 controlled trials) met predetermined inclusion criteria. The effect of low-impact and or non-weight bearing exercise interventions for improving bone health as measured by QUS of the calcaneum were examined in the young, premenopausal and postmenopausal age groups. Trial quality was assessed using the Effective Public Health Practice Project quality assessment tool. Study outcomes for analysis, absolute change (dB/MHz) or relative change (%) in broadband ultrasound attenuation (BUA) and or os calcis stiffness index at the calcaneus were compared by calculating standardised mean difference (SMD) using fixed- and random-effects models. Discussion: Quality of included trials varied from medium to high on a scale of one to three. Four to 36 months of varying exercise interventions for improving bone health led to a significant improvement in calcaneum BUA (0.98 SMD, 95 % CI 0.80, 1.16, overall effect Z-value010.72, p00.001) across the age spectrum. The exact mechanism of assessment of bone quality and or strength by the QUS is still subject to evolving technological advancement. Conclusion: This meta-analysis, attest to the sensitivity of the QUS to exercise-induced changes in bone health across the age groups. The lack of exposure to ionizing radiation, timesaving capability, and low cost of the QUS could be seen as added advantage when considering its use for large-scale public health interventions that are primarily aimed at improving bone health for both preventive and or treatment purposes. P50 THE RELATIONSHIP BETWEEN SERUM VITAMIN D LEVELS AND BONE HISTOMORPHOMETRY IN ELDERLY PATIENTS WITH HIP FRACTURE Arvind Nune (1) presenting Abbas Ismail (1) A Freemont (1) N Makepeace (1) Stepping Hill Hospital, Manchester, UK (1) Introduction: Vitamin D deficiency is common in elderly people, studies suggesting up to 65 % of patients with hip fracture have severe vitamin D deficiency (1). However, routine bone biochemistry is frequently normal despite low or even undetectable serum Vitamin D levels (2). The clinical significance of low serum Vitamin D levels therefore remains uncertain.

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

Aims: To assess the relationship between serum vitamin D concentration and bone biochemistry with static bone histomorphometry parameters in elderly patients with hip fracture. A secondary aim was to determine whether osteomalcia a contributory factor in the pathogenesis of hip fracture. Materials and Methods: Hip fracture patients were identified prospectively from the Orthopaedic wards in Stockport, UK between July 2009 and July 2010. After obtaining consent, blood samples were taken for serum vitamin D and bone biochemistry prior to surgery. Parathyroid hormone levels were not measured. Serum 25-hydroxy vitamin D levels were measured by mass spectrometry. At the time of surgery for the hip fracture the operating orthopaedic surgeon retained either the femoral head in the case of subcapital fracture or took a small bone biopsy in the case of trochanteric hip fracture. The bone samples underwent histomorphometric analysis at the University of Manchester. Dynamic histomorphometry was not possible because of the clinical setting in which the bone samples were obtained. Discussion: 18 hip fracture patients were included in the analysis, 12 female and 6 male, mean age 84.2 years (range 70–92 years). There were 15 subcapital fractures and 3 trochanteric hip fractures. 15 patients had low serum Vit D levels of less than 20ug/L and 5 patients had very low levels of less than 12 ug/L (normal level>30ug/L). In all 18 patients, bone histomorphometry showed evidence of osteoporosis with the mean age and sex adjusted Z score for trabecular bone volume being - 2.8 and mean Z score compared to peak bone mass being - 5.6. However, there was no evidence of mineralization defect or osteomalacia in any of the bone samples. Serum calcium and alkaline phosphatase levels were all within normal limits. Conclusion: In this preliminary study of elderly hip fracture patients, low or very low serum vitamin D levels were not associated with either biochemical nor histological evidence of osteomalacia although osteoporosis was universal. In this study osteomalacia does not appear to be a contributory factor in the pathogenesis of hip fracture. P51 A POTENTIAL MAJOR ROLE FOR FRACTURE LIAISON SERVICES IN PRIMARY CARE Chris Dixon (1) presenting Karen Lewis (1) Karen Sampson (1) Nick Viner (1) Torbay Hospital, Torquay TQ2 7AA, UK (1) Introduction: Fracture liaison services (FLS) in secondary care can reduce the incidence of subsequent fractures in

S573

patients presenting with a fragility fracture. We have extended the remit of our FLS to include a role in education and improving medication adherence/compliance in primary care, particularly targeting care homes since they are the source of a disproportionate number of patients presenting with hip fractures locally. Methods: 33 of 117 local care homes were visited by FLS nurses to identify patients who might benefit from calcium and vitamin D (Ca/VitD) supplementation and to check adherence/compliance in patients already prescribed bone protection therapy. Discussion: 837 care home residents were assessed. 301 (36 %) were taking calcium and vitamin D supplements, but 41 (14 %) lacked adherence/compliance and were given advice about alternative preparations. Of the remaining 536, it was established that 415 (77 %) would benefit from Ca/VitD supplementation according to the Chapuy data (BMJ 1994;308:108) and were therefore advised to start treatment. Of 125 residents taking oral bisphosphonates, 17 (14 %) were not taking Ca/VitD supplements and 93 (74 %) were taking their weekly bisphosphonate incorrectly, often with other medication (including Ca/VitD supplements). Therefore, proper adherence/compliance was only observed in 12 % of patients prescribed long term oral bisphosphonates. 24 of the 33 homes visited were giving bisphosphonates incorrectly to all residents due to lack of awareness of the correct dosing procedure. Following the provision of education sessions and written information, 13 homes were revisited after 3 months. 12 of the 13 homes had changed their practice positively to reflect the education given. Conclusion: There is the potential for FLS to play a major role in education and improving medication adherence/compliance in primary care, particularly in nursing and residential homes. Liaison with care home staff is likely to be particularly valuable. Ensuring adherence/ compliance with oral bisphosphonates remains a major challenge and primary care teams should play a greater role in addressing this problem. P52 FUNCTIONAL LINK BETWEEN MATERIAL LEVEL STRUCTURAL ALTERATIONS IN STEROID INDUCED OSTEOPOROTIC BONE AND ITS INCREASED FRAGILITY Angelo Karunaratne (1) presenting Chris Esapa (3,4) Jennifer Hiller (2) Nick Terrill (2,5) Rajesh Thakker (3) Himadri Gupta (1) Queen Mary University of London, School of Engineering and Material Sciences, London, UK (1) Diamond Light Source Ltd., Diamond House, Harwell Science and Innovation Campus, Oxfordshire, UK (2) Academic Endocrine Unit, Nuffield Department of Clinical Medicine, Oxford Centre for Diabetes, Endocrinology and Metabolism (OCDEM), University of Oxford, Oxford, UK (3) MRC Mammalian Genetics Unit and Mary Lyon Centre, MRC Harwell, Harwell Science

S574

and Innovation Campus, Oxford, UK (4) Department of Chemistry, University of Sheffield, Sheffield, UK (5) Introduction: Glucocorticoid therapy is a widespread treatment mostly addressed to the elderly population who are suffering with overactive immune system disorders such as asthma, autoimmune diseases and arthritis. There are several short term (high blood glucose levels, insomnia and euphoria) and long term side effects (Cushing’s syndrome, glaucoma and cataracts) associate with this particular steroid therapy. One of the most serious long-term side effects of glucocorticoid treatment is secondary (drug-induced) osteoporosis, enhancing fracture risk in bone. The rapid increase in bone fracture risk is indicative both of loss of bone quantity and degradation of bone quality. Reduction in bone quantity can be assessed using bone mineral density (BMD) measurements by clinical tools like DXA and qCT. However, the alterations to the material level properties of bone due to glucocorticoid treatment are not well understood. Understanding the nanostructural origins of increased fracture fragility in drug induced osteoporosis is essential to correlate bone quality alterations in the fibrillar level to mechanical deteriorations. Material and Methods: Here we demonstrate alterations in the nanostructural mechanical response of the mineralized fibrillar collagen matrix (quality) and link to an increased fracture risk in glucocorticoid induced osteoporosis. Using bone from a murine model for glucocorticoid induced osteoporosis developed by ethylnitrosurea (ENU) mutagenesis, we measure the deformation of the mineralized fibrils occurring in-situ during external loading, by combining mechanical testing with synchrotron small angle X-ray scattering (SAXS), a form of functional imaging. These provide nano-mechanical parameters of bone quality, including fibril elastic modulus, maximum fibril strain and fibril to tissue strain ratio. Discussion: A significant reduction (67 %) of fibril modulus, enhancement (125 %) of maximum fibril strain occurs in osteoporotic mice. We also find a much larger fibril strain/tissue strain ratio in osteoporotic mice compared to the wild type mice. Conclusion: Our study demonstrates the ability of in-situ synchrotron X-ray nanomechanical imaging as a high resolution diagnostic technique to link the changes in steroidinduced osteoporotic bone to local mechanical competence and increases in bone fragility. P53 TIBIAL PLATEAU BONE SURFACE ROUGHNESS – A POTENTIAL NEW EARLY MARKER TO DETECT ARTHRITIS? Qi Guang Wang, Sindhu Mohanty, Ilaria Bellantuono, Alison Gartland; The University of Sheffield, Sheffield, UK (1) Introduction: Arthritis, whether inflammatory or degenerative, is a multifactorial disease which is difficult to diagnose

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

early. Often the only definitive diagnosis is given once the damage has already been done to the joint and the underlying bone. Therefore development of an imaging technique to detect bone damage at an earlier stage would be advantageous. In this study we present a novel quantitative approach using high-resolution micro-CT scanning to correlate tibial plateau bone surface roughness (BSR) with the progression of arthritis. Methods: We utilised the Balb/C Interleukin-1 receptor antagonist knock-out mice, which spontaneously develop an inflammatory arthritis-like disease in 100 % of mice by 20 weeks of age. Mice tibia at three different stages: 1) before-disease, 2) at disease-onset, or 3) late-disease were scanned using micro-CT (Skyscan). The scanned images were reconstructed into 3D models. The appearance of tibial plateau bone surface was visually inspected. BSR was calculated using the ratio of the raw and the smoothed bone surface area, which was achieved using 3D median filters. The BSR values were compared between each disease stage by student t tests. Discussion: From the 3D reconstructed models no bone erosion was observed at stage 1, whilst severe bone erosion was clearly visible at stage 3. Only minor bone surface erosion was evident from the 3D reconstructions at stage 2, which in isolation would be insufficient to classify disease.. Quantitative BSR analysis gave the mean BSR values of 1.059, 1.147 and 1.319 for the stages 1–3 respectively (p<0.05 and p<0.01 between each stage). Conclusion: We have developed a new approach to quantitate murine tibial plateau BSR using micro-CT. Based on our murine model, the quantitation of BSR correlated to the visual determination of disease progression. More importantly, significant statistical difference was detected between the before-disease and disease-onset stages. This finding indicates that this method could potentially be a new marker to detect early bone damage at an earlier stage of arthritis. P54 THE USE OF TERIPARATIDE IN THE TREATMENT OF POSTMENOPAUSAL OSTEOPOROSIS EXPERIENCE OF USING BONE MARKERS AND BONE MINERAL DENSITY TO MONITOR RESPONSE Rosemary J Hollick (1), David M Reid (2), Alison J Black (1) Aberdeen Royal Infirmary, Aberdeen, UK (1) University of Aberdeen, Aberdeen, UK (2) Aims: Teriparatide (1–34 N-terminal active fragment of parathyroid hormone) is recommended for women with severe osteoporosis and high risk of fracture who have failed/are intolerant of first line therapy (NICE, 2008). The aims of this audit were to assess (a) prescribing of teriparatide within our unit and explore deviations from NICE

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

guidelines (b) lumbar spine (LS) and total hip bone mineral density (BMD) response following 18 months of teriparatide and correlate with bone turnover marker (BTM) response at 6 months and (c) fracture outcome post teriparatide therapy. Methods: All patients commencing teriparatide from October 2004 until April 2009 were identified from the pharmacy register. DEXA scanning was performed on a GE Lunar Prodigy scanner. Procollagen type 1 Npropeptide (P1NP) and C-terminal telopeptides of type 1 collagen (CTX) were measured using the Elecsys 2010 auto-analyser (Roche,USA). Results: Case notes were available for 57/62 patients who had completed 18 months teriparatide. Female to male ratio 56>1. Mean age at start of teriparatide 73±7.5 yrs. Prevalent number of vertebral fractures 2.5±1.6. All patients fulfilled Part A of NICE guidelines in terms of intolerance to bisphosphonates/ unsatisfactory response. 23/57 patients did not fulfil BMD criteria for prescription of teriparatide and 2/57 patients were <55 years old. Median % change BMD LS and hip (n054) was 9.9 (32.1 to −6.5) and 0.4 (15.9 to −17.7) respectively. Median absolute change P1NP and CTX at 6 months (n042) was 92.6 (427.7 to −15.63) and 0.4 (1.5 to −0.3) respectively. No significant correlation was observed between BTM response and % change LS BMD. 53/54 patients have had no further fractures during/ since completion of therapy (median time to non-fracture 3 years). Conclusions: Deviations from NICE guidelines largely reflected artificial elevation of absolute T-scores by a number of factors e.g. vertebral fractures/degenerative change. Majority of patients had >3 % increase LS BMD with modest changes in total hip BMD. Magnitude of BTM response does not predict LS BMD response in this group, however, poor BTM response at 6 months can highlight issues with compliance. Teriparatide appears effective in reducing vertebral and non-vertebral fractures irrespective of BMD/BTM response. P55 FRACTURE LIAISON SERVICE AND COMPLIANCE WITH OSTEOPOROSIS TREATMENT Mashood Siddiqi (1) presenting Jacqueline O'Hare (1)University Hospital Aintree, Liverpool, UK (1) Introduction: The study was conducted to assess the compliance with the treatment, in patients seen in the Fracture Liaison Service clinic. Our initial audit had shown that the compliance at 12 months was a satisfactory 75 %1 and we decided to test whether the compliance is maintained over longer period and in larger cohort of patients. Material and Methods: Aintree Fracture Reduction Initiative (AFRI), service is similar to Glasgow model of FLS2,

S575

with slight difference. All patients of over 50 years, attending our fracture clinics are sent a risk-assessment questionnaire and patients meeting the criteria of low trauma fracture, are assessed by DEXA measurement, and those requiring treatment, are seen in the AFRI clinic. The primary care physicians are advised accordingly. There are no follow-up arrangements. The fracture patients, who are admitted to hospital, are assessed by ortho-geriatric team. Questionnaires were sent out to 1000 patients who had attended the AFRI clinic from November 2005 to October 2008. No reminders were sent. Discussion: 625(62.5 %) patients responded. Average duration of treatment was 40 months (range 24–60 months). 507 (81 %) of the respondents were on the same form of treatment as initially prescribed. Of these, 496(98 %) were on calcium and vitamin D supplements, 305(60 %) were on Bisphosphonates(oral/IV) and 11(2 %) were on other medications(Strontium,Raloxifene). Only 20(4 %) of patients required change in medication. 10 patients, were changed from Alendronate to Risedronate and one patient to Strontium, by their physicians. 9 patients were receiving intravenous Ibandronate after being changed from Alendronate (4), Risedronate(3), oral Ibandronate(1) and Strontium(1). Of the 118(19 %) non-compliant patients, only 21(18 %) had stopped taking medications on doctors advice. The rest 97(82 %) gave no specific reason for stopping treatment. Of the non-compliant patients, 54 (46 %) were initially prescribed oral bisphosphonates and 5 strontium. Conclusion: Significant number of patients, (81 %) are compliant with the treatment prescribed at the FLS clinic, even after over two years of treatment, indicating that this is a reliable way for the secondary prevention of osteoporotic fractures. References:1.Siddiqi M, Ipe A, O’Hare J (2010) Age & Ageing 39 suppl I, i14. 2.McLellan AR, et al (2003) Osteoporos Int 14:1028–1034. P56 AUDIT INTO FRAGILITY HIP FRACTURES IN LOCAL CARE HOME RESIDENTS Rachel Davies (1) presenting Radcliffe Lisk (1) Keefai Yeong (1) St Peter's Hospital, Chertsey, Surrey, UK (1) Introduction: Hip fractures exert a heavy toll on older adults, especially nursing home residents with annual incidence rates of hip fractures calculated at 3-6 % for this group of patients [1–2]. Several studies have investigated risk factors for hip fractures in residents of long term care facilities; they include increasing age, falls, cognitive impairment, urinary incontinence and impaired mobility [1,2]. Osteoporosis has been reported in approximately 80 % of nursing home residents [3].

S576

22 % of our patients with fractured neck of femurs (NOFs) are from care home settings, which is slightly higher than the national average [4]. We set out to investigate the baseline characteristics of these patients to determine whether a falls prevention package will help reduce hip fractures from these homes. Methods: We looked at 91 patients from local care homes admitted to St Peter’s Hospital over a 21-month period and analysed their comorbidities, falls risk, and medication. Results: Our findings showed that five care homes in particular accounted for 33 % of all care home admissions. Other important findings were that 75 % of patients had an unwitnessed fall, 35 % patients were known to be recurrent fallers and over half had a hospital admission in the preceding year. Only 14 % patients were on anti-resorptive therapy and only 21 % were taking calcium and vitamin D supplements. In addition, 30 % patients were on 6 or more medications and 70 % patients had a diagnosis of dementia. Conclusions: Our findings identified opportunities for intervention within care homes, particularly in those with the highest number of fractured NOFs. A falls prevention package, focusing on education, training and medication reviews, including secondary prevention of osteoporosis should help prevent falls and prevent harm from fractures and has been implemented in the 5 care homes with highest number of hip fracture admissions to our Trust. Ultimately, a falls prevention package targeting care homes nationally to ensure an enhanced access to community falls teams, geriatricians and education of care home staff should reduce hip fracture rates. P57 RISK OF ACUTE MYOCARDIAL INFARCTION IN PATIENTS WITH TOTAL HIP / KNEE REPLACEMENTS AND MATCHED CONTROLS: A POPULATION-BASED COHORT STUDY IN DENMARK Corinne Klop (1) presenting Arief Lalmohamed (1) Peter Vestergaard (2) Frank de Vries (1) Utrecht University, Utrecht, The Netherlands (1) Aarhus University Hospital, Aarhus, Denmark (2) Introduction: Although evidence is conflicting, risk of acute myocardial infarction (AMI) may be increased after total hip and knee replacement (THR/TKR). However, risk of AMI in these patients has not been compared against control subjects. The objective of this study was to estimate hazard ratios (HR) for AMI in THR/TKR patients versus matched controls. Methods: We carried out a nationwide cohort study in Denmark (1998–2007). Patients with a primary THR/TKR (n0 95,227) were matched to three controls by age, sex, and region. All subjects were followed for AMI, and time-

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

dependent cox models were used to derive disease history and drug use adjusted (adj.) HRs. Discussion: Among THR patients versus controls, risk of AMI was substantially increased in the first two weeks post surgery (adj. HR 9.20; 95 % CI 7.97-10.60). The risk remained elevated within 2–6 weeks postoperatively (adj. HR 2.30; 95 % CI 1.89-2.81), and then dropped to baseline levels. For TKR patients, AMI risk was only increased during the first two weeks (adj. HR 5.72; 95 % CI 4.217.79), and was similar to their matched controls two weeks after surgery. Among THR patients, the six-week AMI risk was seven-fold increased among patients aged≥80 years. The risk of AMI was 40-fold increased among patients with a previous AMI and 9-fold increased among patients with cerebrovascular disease. Conclusion: This is probably the first study that shows that THR (9-fold) and TKR patients (6-fold) are at increased risk of AMI during the first two weeks after surgery. These findings support evidence for risk assessment of AMI during the first six weeks after THR, and during the first two weeks after TKR. Conflict of Interest: The division of Pharmacoepidemiology & Clinical Pharmacology employing authors CK, AL, and FV has received unrestricted funding for pharmacoepidemiological research from GlaxoSmithKline, the privatepublic funded Top Institute Pharma (www.tipharma.nl, includes co-funding from universities, government, and industry), the Dutch Medicines Evaluation Board, and the Dutch Ministry of Health. P58 AN AUDIT OF ANDROGEN DEPRIVATION THERAPY AND BONE PROTECTION IN SECONDARY CARE May Yung Tiet (1) presenting Allyah Abbas (1) Andrew Whallett (1) Dudley group of hospitals, Dudley, West Midlands, UK (1) Background: Prostate cancer is the most common cancer in men, with an incidence of 37,000 in 2008 in the UK. Androgen deprivation therapy plays an important role in prostate cancer treatment for many patients. However, hormone treatment is known to reduce bone density, even with short term use. This may lead to morbidity from pathological fractures and has a negative correlation with survival. Bisphosphonates and denosumab can be used to ameliorate these changes in bone mineralisation and thus reduce the incidence of fractures during use of anti-androgens. Currently there are no national guidelines regarding bone protection during and following use of androgen deprivation therapy. Method: We audited the management of bone protection in prostate cancer patients starting androgen deprivation therapy over a nineteen month period at Russells Hall Hospital,

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

Dudley, West Midlands. Consideration for risk factors of osteoporosis by clinician use of risk stratification tools (FRAX®), dual-energy X ray absorptiometry (DEXA) scans and appropriate prescription of preventative treatment with calcium/vitamin D, bisphosphonates and denosumab were reviewed. Results: A total of 177 cases were analysed. During and following treatment FRAX® score had not been documented for any patient, even in cases of known osteoporosis, and only 2.8 % had undergone a DEXA scan. Only 2.3 % were prescribed calcium/vitamin D and 6.8 % received bisphosphonates. Denosumab had not been prescribed throughout this period. Recommendations: Although osteoporosis is a risk with androgen deprivation therapy, here patients at risk of fracture had not been identified by lack of use of risk stratification tools and baseline DEXA scans. Yet delays in preventative treatment may have detrimental effects on skeletal health. Agreed national guidelines and educating those managing prostate cancer care are required to make screening for at risk patients standard practice. Re-auditing following guidance is necessary. P59 WHOLE BODY 18 F-FLUORIDE PET MEASUREMENTS OF REGIONAL BONE METABOLISM Musib Siddique (1) presenting Michelle Frost (1) Amelia Moore (1) Ignac Fogelman (1) Glen Blake (1) King’s College London, London, UK (1) Introduction: 18 F-fluoride positron emission tomography (18 F-PET) studies to measure bone plasma clearance provide a unique way of investigating regional bone metabolism that gives insight into the changes that occur during treatment at specific regional sites of interest that it is not possible to obtain with biochemical markers of bone turnover. Quantitative 18 F-PET studies are often performed using a protocol that involves a 60-minute dynamic scan. However, with a single injection of 18 F-fluoride the region of interest (ROI) is restricted to the 15 cm axial field of view of the PET scanner so, for example, the lumbar spine or hip may be studied, but not both sites in the same examination. We present whole body 18 F-PET studies of the skeleton using a new method that enables estimation of regional 18 F-fluoride bone plasma clearance from a series of short (5-minute) scans acquired at multiple sites in the skeleton following a single injection of tracer. Material and Methods: The study included eleven untreated postmenopausal osteoporotic women without any disease affecting bone metabolism. For each subject a series of eight separate 5-minute 18 F-PET scans were acquired covering the body from the upper femora to the head. Blood samples

S577

were taken at 30, 40, 50 and 60 minutes after injection to derive the plasma input function. The scans were analysed to estimate 18 F-fluoride bone plasma clearance in ten different ROIs. Results: Mean values for 18 F-fluoride bone plasma clearance in each ROI were (units: mL min-1 mL-1): femoral shaft: 0.0087; total hip: 0.0107; femoral neck: 0.0094; pelvis: 0.0201; lumbar spine: 0.0245; thoracic spine: 0.0275; cervical spine: 0.0278; skull: 0.0160; humerus: 0.0056; forearm: 0.0039. Discussion and Conclusions: The new method provides a way of extending 18 F-PET studies so that measurements can be made at multiple sites in the skeleton with a single injection of 18 F-fluoride and a shorter total image acquisition time. P60 MINIMISING SUBSIDENCE IN LUMBAR TOTAL DISC REPLACEMENT THROUGH EFFECTIVE CEMENT PLACEMENT Adam Liddle (1) Vishal Borse (1) Daniel Skrzypiec (1) Jake Timothy (2) Nikil Kapur (1) Richard Hall (1) School of Mechanical Engineering, University of Leeds, Leeds, UK (1) Department of Neurosurgery, Leeds General Infirmary, Leeds, UK (2) Introduction: Osteoporosis currently excludes a large population of elderly patients from receiving total disc replacement (TDR) due to increased risk of post-operative implant subsidence. Cement augmentation has been combined with TDR clinically, using a modification of percutaneous vertebroplasty, a technique shown to increase vertebral body (VB) strength. However, there has been little research to optimise the volume of cement required and how it should be distributed within the VB. This is a biomechanical investigation exploring the affect of PMMA cement volume and distribution on TDR subsidence in augmented cadaveric vertebrae. Material and Methods: Thirty-two single vertebra specimens (L2-L5) were split into four groups 1) Control group with no cement augmentation, 2) Traditional bipedicular vertebroplasty approach with 10 % PMMA cement fill, 3) Bipedicular with 20 % fill, and 4) Anterior approach with 20 % distributed fill. Resistance to subsidence was ascertained via indentation tests, performed on the superior endplate using a generic TDR-shaped indenter that was free to rotate in all three planes. A strain rate of 1 mm/min was used until 25 % strain was achieved. Change in indenter angle was calculated in the sagittal and coronal planes using a digital height gauge. Discussion: Cement augmentation significantly affected failure stress (p0.002) with both 20 % fill groups showing increases of approximately 1.5 times than that observed in the control cohort. Maximum modulus was only marginally affected (p0.084) despite the same trends being shown. Indenter angle was significantly changed in the sagittal

S578

plane (p<.001). Both traditional approach cohorts showed increased tilting towards the posterior. Indenter angle in the coronal plane remained consistently low in all groups with no significant differences (p0.914). Conclusion: Only a 20 % cement fill was sufficient to significantly increase endplate strength in vitro. However, increased subsidence towards the posterior occurred when using a traditional bipedicular approach but remained level when distributing the cement evenly across the vertebra via an anterior approach. In summary, both cement distribution volume and distribution are important considerations in using vertebroplasty as an anti-subsidence strategy. P61 PATIENTS ACCEPT THE USE OF TOUCH SCREENS TO SELF REPORT OSTEOPOROSIS RISKS. RESULTS OF THE "CATCH BEFORE A FALL" STUDY Peter Smitham (1) presenting Felix Allen (1) Alun Cooper (1) Simon Grange (2) Graham Davenport (1) David Marsh (1) UCL, London, UK (1) University of Calgary, Calgary, Canada (2) Introduction:The prevention of osteoporotic fractures is a clinical and political priority. There have been some success stories in treating new fracture patients with bisphosphonates. However, finding and treating patients who have sustained a fracture in the past is considerably harder. This requires primary care trusts to trawl through medical records to determine if a previous fracture has occurred. This is an ineffective use of time and being retrospective relies on the data being accurate.The Lanarkshire Osteoporosis Questionnaire showed that patients were able to recollect a previous fracture and other risk factors, but again this required a form to be sent out and time expended to process the forms on their return. We aimed to see if patients could interact with a touch screen PC based in a GP's waiting room and self report their FRAX score and other questions surrounding osteoporosis. Method: A patient friendly iPad app was created and installed in two General Practice waiting rooms from around the country. A further ipad was used on ward based patients. Along with the results of the application patients were asked to complete a paper questionnaire assessing the applications acceptability and feasablity. Results: A total of 314 patients completed the application successfully on the ipad and a further 102 patients completed the paper questionnaire. Excluding under 55 year olds a total of 68 patients completed the paper questionnaire assessing the ipads acceptability and feasablity with a mean age of 68 (55-89 yrs). Although there was an increasing perception of difficulty in completing the questionnaire with age, the mean useability score was 2.6 (score 1– 10 with very easy being 1). 75 % of respondents said they would prefer using a touch screen device rather than paper or

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

phone questionnaires and 83 % stated they would use the touch screen if it was offered in GP surgeries. The questions patients had most difficulty answering surrounded medication and height loss of more than 5cms. Conclusion: This study showed that the above 55 year olds have readily accepted touch screen devices as a way of interacting and the majority of patients can self report their osteoporosis risk. This has the potential to collect data cheaply and empower patients through education of the risks of osteoporosis, allowing them to seek appropriate follow-up. P62 25(OH)D CONCENTRATION AND MUSCLE FUNCTION IN PREMENOPAUSAL SOUTH ASIAN AND CAUCASIAN WOMEN Ohood Hakim (1) presenting Andrea Darling (1) Kath Hart (1) Jacqueline Berry (2) Susan Lanham-New (1) University of Surrey, Guildford, UK (1) University of Manchester, Manchester, UK (2) Introduction: It has been shown that hypovitaminosis is associated with poor muscle strength and performance among elderly women1. Little is known in literature about middle age women and if there is an ethnic difference in muscle performance. Materials and Methods: As part of the D-FINES (Vitamin D, Food Intake, Nutrition and Exposure to Sunlight in Southern England) study, we aimed to determine if there was an association between serum 25(OH)D, and muscle function. Forty healthy premenopausal women (21 C and 19 SA), age ranges 18–55 yrs took part in the study. Fasted blood samples were collected for vitamin D. Muscle pain and physical activity were assessed via a questionnaire. Grip strength and the sit-to-walk test were measured to reflect muscle strength and performance. Results and Discussion: 25(OH)D status was significantly different between SA and C groups (p<0.001), with mean values of 31.53[16.32] and 80.91[20.08] nmol/l respectively. The Table below shows the mean value of the muscle function measurements according to ethnicity

Grip strength Right hand Grip strength Left hand Sit to walk test (seconds) Muscle pain score

C, n019 A, n016 C, n019 A, n016 C, n019 A, n015 C, n021 A, n015

Mean [SD] 24.91 [5.36]** 20.21 [5.03] 22.32 [5.03]* 19.00 [5.82] 6.17 [0.73] 6.63 [0.81] 16.00(10.84)** 35.46(27.15)

***p<0.001, **p<0.01, *p<0.05; Values were analysed by t-test

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

There was a significant difference between grip strength mean and muscle pain score between SA and C groups (P<0.01), with the South Asian women having much lower values. There was however, no significant correlation between vitamin D and muscle strength variables. Inability to observe association between vitamin D and muscle strength in the SA women may be due to widespread vitamin D deficiency in this group. Further investigation of the link between vitamin D and musculoskeletal health in younger women is underway. 1 Mastaglia et al (2011) J Nutr Health Aging 15(5):349–54 Conclusion: Vitamin D was inversely associated with bone size and had no correlation with bone density among premenopausal SA women. 1 Adami et al (2009) Bone 45(3):423–426. P63 STRUCTURE ANALYSIS OF THE HIP QCT SCANS AND HIP FRACTURE RISK IN OLDER WOMEN Will Udall, Richard Eastell, Eugene McCloskey, Margaret Paggiossi, Lang Yang presenting University of Sheffield, Sheffield, UK This cross-sectional case–control study aimed to map the 3Dimensional distribution of bone density in the proximal femur and examine its association with hip fracture with a view to improve current methods of fracture discrimination. Fifty postmenopausal women (55–89 yrs) who had sustained hip fractures due to low-energy trauma underwent QCT scans within 3 months of the fracture. For each fracture case a postmenopausal women, matched by age (±5 yrs), weight (±5 kg) and height (±5 cm) was recruited as a control. We quantified cortical, trabecular and integral vBMD and cortical thickness in four quadrants of cross sections along the length of the femoral neck (FN), intertrochanter (IT) and trochanter (TR). Compared with controls, cortical vBMD and thickness in most quadrants of the FN, IT and TR were significantly (p<0.05) lower in fracture cases, whereas lower trabecular vBMDs were found in 2 lateral quadrants of the FN. To examine the association of QCT measurements with hip fracture risk, quadrants in the medial and lateral aspects of the FN, IT and TR were merged due to the stress distribution during a fall and the pattern of fracture initiation in the superior aspect of the FN. At FN, IT and TR, QCT measurements were associated significantly (p<0.005) with hip fracture, the odds ratio (OR) for one standard deviation decrease ranged between 2.41 (95 % CI 1.32, 4.39) to 6.34 (95 % CI 2.22, 18.10). After adjustment for TH areal BMD, cortical vBMDs and thickness at the FN were still associated with hip fracture significantly (p<0.05), the ORs ranged from 2.79 (95 % CI 1.13, 6.86) to 3.70 (95 % CI 1.32, 10.36). In conclusion, QCT provides structural information that is additional to that provided by DXA.

S579

P64 NANOTOPOGRAPHY-INDUCED OSTEOGENIC DIFFERENTIATION OF HUMAN EMBRYONIC STEM CELLS Emmajayne Kingham (1) presenting Nikolaj Gadegaard (2) Matthew Dalby (2) Richard Oreffo (1,3) University of Southampton, Southampton, UK (1) University of Glasgow, Glasgow, UK (2) King Saud University, Riyadh, Saudi Arabia (3) Introduction: The efficient production of stem cell-derived cell types from adult and embryonic sources will be of significant medical and research benefit but differentiation protocols currently rely on chemical induction and genetic manipulation which may increase the risks associated with transplantation. Materials and Methods: Using discrete nanoscale surface topographies displaying nanopits arranged in a near-square geometry, we have directed the mesodermal differentiation of human Embryonic Stem Cells (hESCs) in the absence of chemical induction factors. Resulting cell types were characterised using immunofluorescent staining, qPCR and epigenetic techniques including pyrosequencing. Discussion: Near square nanotopography-differentiated cell types exhibited a down-regulation of hESC markers Nanog, OCT4, SOX2, TRA-1-60 and SSEA4, an up-regulation of adult skeletal stem cell-specific genes including STRO1 and CD44, and enhanced expression of osteogenic markers Type I Collagen, RUNX2 and Osteonectin. However, following induction of differentiation by formation of embryoid bodies, cells became less sensitive to nanotopographical cues. Epigenetic mechanisms for nanotopography-directed differentiation were also explored. While global methylation remained the same, changes in DNA methylation at specific promoter region sites (osteocalcin and OCT4) were detected, furthermore these methylation changes were akin to chemicallyinduced differentiation changes also detected. Conclusion: These studies using discrete nanotopgrpahical cues in the absence of chemical cues, provide a unique approach to overcome regenerative medical challenges and describe a reproducible platform for hESC directed differentiation with downstream applications in stem cell biology, small molecule screening and therapeutic use. P65 DOES PHOSPHO1 REGULATE INSULIN SIGNALLING IN OSTEOBLASTS? Karla Oldknow (1) presenting Carmen Huesa (1) Manisha Yadav (2) Vicky Macrae (1) Jose Luis Millan (2) Colin Farquharson (1) Roslin Institute, Edinburgh, UK (1) Sanford-Burnham Medical Research Institute, La Jolla, CA, USA (2) Genetic approaches to bone physiology utilising judicious gain and loss of function models have identified bone as a true endocrine organ, possessing the capabilities to regulate

S580

both energy metabolism and reproduction. These recent advances expand our understanding and identify a new and unconventional role of bone beyond its classical functions. Specifically, osteocalcin is regulated by the Esp gene which encodes osteotesticular protein tyrosine phosphatase (OST-PTP). A direct interaction between the insulin receptor (INSR) and OST-PTP has been demonstrated resulting in the dephosphorylation and subsequent inactivation of the INSR (Lee et al. 2007). Here, we investigate a possible role for PHOSPHO1 in insulin signalling in osteoblasts, potentially implicating a key role for PHOSPHO1 in energy homeostasis. A screen of phosphorylated peptides identified potential PHOSPHO1 substrates and a yeast-2-hybrid (Y2H) approach identified PHOSPHO1 binding partners. Primary osteoblasts (pOB) were isolated from wild-type and PHOSPHO1−/− mice and challenged in culture with insulin and IGF-I. pOB over-expressing PHOSPHO1 were also studied. Western blotting determined Akt phosphorylation and PHOSPHO1 protein expression. Additionally cultured pOB were examined by RT-qPCR to assess Phospho1, Esp, and BGLAP. PHOSPHO1 was shown to have the highest hydrolysing activity towards the INSR S (P) 1322, suggesting PHOSPHO1 can potentially dephosphorylate the INSR at serine 1322. Additionally Y2H studies revealed a potential PHOSPHO1 interaction with GRB10 interacting GYF protein 1 (GIGYF1) which forms a complex with GRB10 to regulate both insulin and IGF-1 receptor (IGF1R) signaling. We hypothesised that such associations will result in an inverse relationship between PHOSPHO1 expression and insulin/IGF-I induced Akt phosphorylation and IRS1 phosphorylation. However, similar pAKT levels in WT and PHOSPHO1 −/− and PHOSPHO1 over-expressing pOB were observed. This may suggest a compensatory mechanism occurring in the absence of PHOSPHO1. Indeed, gene expression studies showed increased expression of OST-PTP in mineralising PHOSPHO1−/− pOB suggesting OST-PTP and PHOSPHO1 may crosstalk in regulating IGF-I/insulin signalling and that OST-PTP may compensate for the loss of PHOSPHO1. Our findings indicate a novel potential role of PHOSPHO1 in the regulation of insulin/IGF-1 signalling in osteoblasts. However further investigations are required to examine the mechanism regulating the potential interaction between OST-PTP and PHOSPHO1. P66 INVESTIGATING THE REGULATORY ACTIONS OF NPP1 IN BONE Carmen Huesa (1) presenting Neil CW Mackenzie (1) DongXing Zhu (1) Geoff J Faulkner (1) Jose Luis Millan (1,2) Vicky E MacRae (1) The Roslin Institute / University of Edinburgh, Midlothian, UK (1) Sanford-Burnham Medical Research Institute, La Jolla, USA (2) Introduction:Nucleotide pyrophosphatase phosphodiesterase 1 (NPP1) is required to convert extracellular ATP into

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

inorganic pyrophosphate (PPi), a crucial negative regulator of hydroxyapatite (HA) crystal formation. Bone alkaline phosphatase, together with NPP1 are responsible for maintaining the balance between PPi and inorganic phosphate (Pi) and thereby regulating the rate of HA formation in the developing skeleton. NPP1 has also been shown to regulate glucose metabolism in muscle and adipose tissue. In the present study, microarray analysis has been undertaken in order to further elucidate the regulatory actions of NPP1 in bone. Methods: ENPP-1 null and wild-type osteoblasts were isolated from 3-day-old calvariae and cultured for 3 days. RNA was extracted and hybridised to Agilent Mouse Genome CGH Microarrays. Analysis of differential gene expression was undertaken using the Linear Models for Microarray Data command driven package using R, as part of the Bioconductor package. Pathway analysis was performed using Ingenuity IPA Software. Results: Significantly up-regulated genes (P<0.001) included Mtdh (118 fold change (FC)), Thrap3 (104 FC), Calm2 (95 FC), Hmgcr (86 FC), Emp1 (79 FC) and Fgfr1 (22 FC). Significantly down-regulated genes (P < 0.001) included Ank-1 (0.33 FC), Lgals2 (0.35 FC), Elk-1 (0.24 FC) Tab2 (0.26 FC), Nrxn2 (0.23 FC), Dnm3 (0.22 FC) and Mxi1 (0.18 FC). Alterations in insulin, eIF-2 and mTOR signalling were identified through pathway analysis. Conclusions: This study suggests that NPP1 is involved in the regulation of a number of novel genes and pathways. Further studies are required to fully understand the regulatory actions of NPP1 in bone. P67 TYPE I COLLAGEN DEGRADATION DURING TISSUE REPAIR: COMPARISON OF MECHANISMS FOLLOWING FRACTURE AND MYOCARDIAL INFARCTION Rachel Stansfield (1) presenting Fatma Gossiel (1) Allison Morton (1) Christopher Newman (1) Richard Eastell (1) University of Sheffield, Sheffield, UK (1) There is turnover of type I collagen during tissue repair. The degradation of type I collagen by matrix metalloproteinases (MMP) is reflected by serum CTX-MMP (or ICTP) and that by Cathepsin K by CTX-I. There is evidence for increases in CTX-MMP after myocardial infarction and in CTX-I during fracture repair. It isn’t known whether the MMP pathway might also be involved in fracture repair or cathepsins after myocardial infarction. We studied 62 men; 10 were admitted to hospital on the day of their myocardial infarction with ST elevation (mean age 56 yr, range 39 to 80) and 9 attended hospital on the day of their tibial shaft fracture (mean age 33 yr, range 21 to 79); we had 43 age-matched controls (mean age 54 yr, range 20 to 82). The subjects with myocardial infarction and tibial shaft fracture were followed for up to one year; the control subjects were used to establish a

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

reference interval. We measured serum CTX-MMP by ELISA (Orion; reference interval 1.1 to 17.6 ng/mL) and the CTX-I by chemoilluminscence (Roche Cobas; reference interval 0.094 to 0.991 ng/mL). After myocardial infarction, the mean CTX-MMP increased from 5.02 to 6.04 ng/mL within one day of admission (p>0.05); the mean CTX-I increased from 0.200 to 0.374 ng/mL (p<0.05). In one case, the CTX increased to above the reference interval. After tibial shaft fracture, the mean ICTP increased from 5.51 to maximum of 8.71 ng/mL within 28 days of admission (p<0.01) the mean CTX increased from 0.200 to 0.374 ng/mL (p<0.001). In four cases, the CTX increased to above the reference interval. We conclude that the MMP and cathepsin pathways are both used for tissue repair in the bone and heart. This may have clinical implications; drugs that block either pathway (TIMPs, cathepsin K inhibitors) may affect the repair of both tissues. P68 TOTAL HIP REPLACEMENT AFTER HIP FRACTURE IMPLICATIONS OF THE NEW NICE GUIDANCE Manju Krishnan (1) presenting Sue Beck (1) William Havelock (1) Eamonn Eeles (1) Ruth Hubbard (1) Antony Johansen (1) University Hospital of Wales, Cardiff, South Glamorgan, UK (1) Introduction: The Management of Hip Fracture in Adults (Clinical Guideline 124, NICE 2011) states: “Offer total hip replacements to patients with a displaced intra capsular (IC) fracture who: were able to walk independently out of doors with no more than the use of a stick, are not cognitively impaired and are medically fit for anesthesia and the procedure”. We set out to establish how many patients meet these criteria (to model appropriate staffing and equipment for our Orthopaedic unit) and to examine how data routinely reported to the National Hip Fracture Database (NHFD) might be used to monitor different units’ compliance with this element of the NICE Guidance. Methodology: We assessed sequential admissions using an extended version of the NHFD including Nottingham Hip Fracture Score, and Frailty Index (FI). FI was measured by an index of accumulated deficits giving a potential score of 0 (no deficits) to 1 (all 52 deficits). Results: 50/132 (37.0 %) had a displaced IC hip fracture. 50.0 % of all admissions were unable to walk outside using no more than a stick, 38.5 % had an AMT<7, and 14.8 % had an ASA grade >3. These figures total 164 %, but as physical, cognitive and anaesthetic exclusion criteria tended to coexist, only 86.8 % of admissions would be excluded. Thus, 13.2 % of hip fracture admissions appear to meet the criteria for THR. Such patients were significantly less frail than those excluded (mean Frailty Index 0.21 [sd 0.09] vs. 0.37 [sd 0.18]; p<0.01).

S581

Discussion: 13.2 % equates with 66 of our 500 admissions each year, but our clinical experience is that decisions over appropriate surgery require a more comprehensive approach. We are now prospectively assessing all patients using the Frailty Index to examine how this summary of patients' mental and physical impairments might perform as a predictor of hip fracture outcome - including the longer term mobility that is the main justification for THR in more robust individuals. P69 ASSOCIATIONS BETWEEN METABOLIC SYNDROME, MENOPAUSAL STATUS AND LOW BONE MINERAL DENSITY IN TAIWANESE WOMEN Yue-Yuan Lee (1,2) presenting Yen-Ching Chen (1,3) JengMin Chiou (4) Keh-Sung Tsai (5) Jen-Hau Chen (1,6) Institute of Epidemiology and Preventive Medicine, College of Public Health, National Taiwan University, Taipei, Taiwan (1) Department of Obstetrics and Gynecology, Buddhist Tzu Chi General Hospital, Taichung Branch, Taichung, Taiwan (2) Research Center for Genes, Environment and Human Health, College of Public Health, National Taiwan University, Taipei, Taiwan (3) Institute of Statistical Science, Academia Sinica, Taipei, Taiwan (4) Department of Internal Medicine, National Taiwan University Hospital, Taipei, Taiwan (5) Department of Geriatrics and Gerontology, National Taiwan University Hospital, Taipei, Taiwan (6) Introduction:Osteoporosis is a major public health issue because of it has been linked to increased fracture risk and the subsequent mortality. Metabolic syndrome (MetS) has been related to several health issues. However, the association between MetS and bone mineral density (BMD) has been inconsistent and no previous study has explored how menopausal status modified this association. Material and Methods: A total of 1,577 Taiwanese women aged 40 to 55 years old were recruited from the MJ Health Screening Center between 2009 and 2010. Bone mineral density (BMD, g/cm2) was measured at the lumbar spine by using dual-energy X-ray absorptiometry (DXA) and were tertiled (T1, T2, and T3). MetS is based on the definition from ATP III report with modification for the Asian criteria for waist circumference and fasting glucose. High BMD was defined as T2 plus T3, low BMD was defined as T1. This study was aimed to explore the association between MetS or individual component of MetS and BMD in Taiwanese women. We also explored how menopausal status modified this association. Results: The prevalence of MetS is 8.6 % in this female population. As compare to women without MetS, women with MetS were older (47.3 vs. 46.3 years old) and heavier (68.6 vs. 54.7 Kg), had higher body mass index (27.6 vs. 21.9 kg/m2), higher serum alkaline phosphatase level (69.8 vs. 58.5 IU), and more menopausal (33.6 vs. 20.4 %). Women with MetS had a decreased risk of low BMD

S582

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

[adjusted odds ratio (AOR)00.32, 95 % confidence interval (CI)00.19-0.56]. The risk of low BMD (T1 vs. T2 plus T3) decreased as the number of MetS components increased (≥ 2 components) (P for trend<0.0001).Women with abdominal obesity (AOR00.54, 95 % CI00.35-0.84), high blood pressure (BP, AOR00.63, 95 % CI00.43-0.94), or high fasting glucose (AOR00.66, 95 % CI00.48-0.90) had a decreased risk of low BMD. Menopause was associated with an increased risk of low BMD (T1 vs. T2 plus T3: AOR03.14, 95 % CI02.14-4.63). Menopausal status did not significantly modify the association of each individual components of MetS (P interaction >0.05), or MetS itself, with BMD. Conclusions: Women with MetS, abdominal obesity, high blood pressure, or high fasting glucose were significantly associated with a decreased risk of low BMD. Menopausal status did not significantly modify the associations between MetS or individual components of MetS and the risk of low BMD. Keywords: Metabolic syndrome, bone mineral density, osteoporosis, menopause, women. P70 RELATIONSHIP BETWEEN SPINE BONE MINERAL DENSITY AND TRABECULAR BONE SCORE IN POSTMENOPAUSAL POPULATIONS FOLLOWING TOTAL KNEE REPLACEMENT OR LEG FRACTURE Susan Hopkins (1) presenting Chris Smith (1) Andrew Toms (2) Mary Brown (2) Andrew Appelboam (3) Joanne Welsman (4) Karen Knapp (1) College of Engineering, Mathe-

matics and Physical Sciences, University of Exeter, Exeter, UK (1) Princess Elizabeth Orthopaedic Centre, Royal Devon and Exeter Hospital, Exeter, UK (2) Emergency Department, Royal Devon and Exeter Hospital, Exeter, UK (3) College of Life and Environmental Sciences, University of Exeter, Exeter, UK (4) Introduction:The trabecular bone score (TBS) is a novel grey-level texture measurement that is based on the use of experimental variograms of 2D projection images, and is able to differentiate between two 3-dimentional (3D) microarchitectures that exhibit the same bone density but different trabecular characteristics. This study investigated differences in bone quantity and quality assessed by spine bone mineral density (BMD) and TBS between four groups of white postmenopausal females, with no known history of vertebral fracture… Material and Methods: Participants included postmenopausal white females having sustained a leg fracture< 4 weeks previously (n09), leg fracture>1 year (3.0± 2.3 yrs) ago (n 023), total knee replacement (TKR) surgery at baseline (n028) and controls (n046). DXA (GE Lunar Prodigy, Madison, USA) BMD measurements (GE Lunar enCORE®, v9.40) and TBS scores (TBS iNsight®, v1.8. Med-Imaps, France) were obtained for L1-L4 vertebrae. The International Society for Clinical Densitometry (ISCD) recommendations for exclusion of abnormally different vertebrae were applied in the analysis.

Table 1. Results n0

Controls 46

TKR 28

Fractured < 4 weeks 9

Fractured > 1 year 23

Age (years) BMI Spine BMD (L1–L4) g/cm2 T score L1–L4 TBS L1–L4 Correlation BMD v TBS (r20)

64.3 25.7 1.146±0.19 -0.29 1.320±0.08 9%

67 33.2* 1.205±0.22 -0.2 1.291±0.15 9.30%

62.6 26.1 1.100±0.14 -0.66 1.299±0.10 NA

65 28.5* 1.034±0.16* -1.21* 1.249±0.15* 17.60%

*Difference from control group significant at 0.05 level

Conclusion: The correlation between Spine BMD and TBS is very low confirming that TBS is reflecting some bone quality properties not related to BMD. These results suggest that at baseline, leg fracture may not be due to bone fragility, but long-term reduction in weight-bearing activity following fracture may result in a significant reduction in spine density and quality. The TKR group exhibit higher BMD than the controls but poorer TBS scores. Although non-significant, and possibly related to arthritic changes rather than to bone fragility, these differences merit further investigation

P71 CHARCOT NEUROARTHROPATHY IS ASSOCIATED WITH LOW BONE MINERAL DENSITY Karen Knapp (1) presenting Lorraine Scaife (2) Richard Paisey (3) Roderick Warren (2) Mollie Donohoe (2) University of Exeter, Exeter, UK (1) Royal Devon and Exeter Hospital, Exeter, UK (2) Torbay Hospital, Torquay, UK (3) Introduction: Patients with diabetes and associated neuropathy are at risk of Charcot neuroarthropathy, a rare, but limb-

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

S583

threatening complication. It involves a destructive process most commonly of the foot and ankle joints and may be associated with reduced bone density (BMD). The aim of this pilot study was to investigate the BMD in patients with diabetes suffering with Charcot neuroarthropathy. Material and Methods: 71 men were recruited onto the study, a mixture of control participants without diabetes and patients with type 1 (T1) and type 2 (T2) diabetes,

Control

T1 Diabetics

with and without Charcot neuroarthropathy. All participants had Dual energy x-ray absorptiometry scans of their lumbar spine and bilateral proximal femora (GE Lunar Prodigy, Madison, USA). A two-tailed t-test was used to compare the differences between groups. Zscores were used in the analysis to account for age differences the mean hip Z-scores used to reduce the impact of any disuse osteopenia in the Charcot group.

T1 Charcot

T2 Diabetics

T2 Charcot

Mean (SD) 12

20

6

22

10

58.8 (9.12) 26.6 (4.5)

58.7 (12.0) 26.4 (3.8)

49 (11.3) 25.8 (3.0)

64.0 (12.3) 31.8 (3.3)

60.6 (7.7) 34.3 (4.6)

-0.3 (1.5)

0.0 (1.4)

0.1 (2.2)

1.1 (1.4)

0.0 (1.7)

NOF T. Hip

0.1 (1.1) -0.1 (1.0)

0.1 (0.8) 0.0 (0.8)

-0.7 (1.3) -0.7 (1.3)

0.9 (0.9) 0.8 (0.8)

0.3 (0.7)* 0.4 (0.7)

G.trochanter Ward’s

-0.2 (1.0) -0.2 (1.2)

0.0 (0.8) -0.1 (0.7)

-1.1 (1.5)* -0.8 (1.3)

0.9 (1.1) 0.5 (1.1)

0.5 (0.7) -0.5 (1.0)*

n Age (years) BMI (kg/m2) BMD Z-scores: Lumbar spine

p0<0.05 when compared to same type diabetes group.

Discussion: The patients with T1 and T2 diabetes have different BMD characteristics and require treating as two discrete groups of patients. The patients with Charcot neuroarthropathy have reduced hip Z-scores compared to patients with the same type diabetes on the study, with the same pattern seen at the lumbar spine for the T2, but not the T1 groups. Conclusion: These results suggest lower BMD may be associated with Charcot neuroarthropathy. Further investigation is required to evaluate the impact of disuse osteopenia during the treatment of this complex pathology. P72 PRIMARY CARE FRACTURE LIAISON SERVICE: CLOSING THE CARE-GAP Wendy Carr (1) presenting Wendy McMaster (1) TyneHealth Clinical Commissioning Group, Newcastle upon Tyne, UK (1) Introduction: The last decade has witnessed two paradigm shifts in the perception of osteoporosis placing their management firmly within Primary Care. Firstly the emphasis has moved from a management of reduced bone mineral density to the assessment and management of fracture risk. Secondly, fracture risk is now perceived as a chronic disease with a new fracture as the acute event. Material and Methods: A Clinical Commissioning Group in the North East of England elected to invest savings from Freed up Resources to investigate the potential for a purely Primary

Care based Fracture Liaison Service (PC FLS) closer to the patient to address the care-gap issues highlighted in previous national audits. Electronic primary care records are comprehensive, current and continually updated with information from all other healthcare providers, are an ideal means of identifying those at increased risk of fracture from current or previous events. This expanded on an on-going primary prevention programme of selective case finding, using clinical data and a risk assessment tool, with DEXA scanning and appropriate treatment recommendations. The pilot service ran for 10 months within the consortium of 14 practices encompassing 118,000 patients. The service was run by a GP with expertise in osteoporosis and afull time specialist nurse experienced in practice and research nursing and use of primary care IT systems. DEXA scanning, clinic facilities and laboratory investigations were hosted within a community PCT facility. We designed a pathway to collect all new information received by practices regarding fractures, new diagnoses, radiology results, medication changes and factors impacting on falls risk. A Read Coded results and recommendations proforma was sent to GPs. Discussion: 112,000 patients screened within the practices, 560 patients invited for assessment, 442 attended, all given lifestyle advice, 69 % recommended treatment with a 2.5 % complex case bone referral and 7.5 % falls services referral. Patient and clinician satisfaction was high on surveys Conclusion: This is a robust, cost-effective near patient service to reduce future fracture risk.

S584

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

P73 POSTMENOPAUSAL SOUTH ASIAN WOMEN SHOW ADAPTATIONS IN TIBIAL CORTICAL THICKNESS AND VOLUMETRIC BONE MINERAL DENSITY TO COMPENSATE FOR SMALL OVERALL BONE SIZE Andrea L Darling (1) presenting Ohood A Hakim (1) Kathryn H Hart (1) Jacqueline L Berry (2) Susan A Lanham-New (1) University of Surrey, Guildford, UK (1) University of Manchester, Manchester, UK (2) Introduction:It is unknown whether older South Asian women differ in bone geometry in comparison with same-age Caucasian women. This is the first study, to the authors’ knowledge, to use pQCT to measure radial and tibial bone geometry in postmenopausal South Asian women, and assess how this compares with same-age Caucasian women.

Radius SSIp 66%*mm3 ToA 4%mm2 ToA 66%* mm2 ToD 4%mg/cm3 ToD 66% mg/m3 Trab A 4% mm2

Materials and Methods: pQCT scans (Stratec X2000L) at the distal tibia and tibial shaft (4 %,14 % and 38 %) as well as at the distal and proximal radius (4 % and 66 %) were undertaken in n018 South Asian (mean age 63.5y+/−3.6) and n050 Caucasian women (mean age 65.9y+/−4.8). Discussion: The Asian women had a significantly smaller bone area at all parts of the radius and tibia than did Caucasians (p< 0.05). Also, the Asians had larger cortical thickness at the 38 % tibia, in relation to their bone size. At the distal tibia, and 14 % tibia there were increased total and cortical density in South Asians as compared to Caucasians (p<0.05) and increased total density at the distal radius (p<0.05). However, such adaptations are unlikely to be able to prevent highly reduced predicted strength strain indexes (polar and x axis) and reduced fracture load in Asians, caused by their smaller bone size. Indeed, predicted bending fracture load was significantly lower by 2040 % in the Asians than in the Caucasians (p<0.05).

C Mean 245 384 154 272 655 173

SD 79 63 38 54 99 28

A Mean 196 314 130 306 688 141

SD 57 45 26 35 82 20

% 79.9 81.8 84.6 112.8 105.1 81.8

p** 0.03 <0.01 0.04 0.01 0.21 <0.01

Tibia SSIp 38%*mm3 ToA 4%mm2 ToA 38%* mm2 ToD 4%mg/cm3 ToD 38% mg/m3 Trab A 4% mm2

Trab D 4% mg/ cm3 CoA 66% mm2 CoD 66% mg/cm3 CT 66% mm

168

44

183

33

108.6

0.21

68 1070 1.8

16 51 0.4

61 1088 1.8

11 40 0.4

89.7 101.7 97.8

0.04 0.18 0.73

Trab D 4% mg/ cm3 CoA 38% mm2 CoD 38% mg/cm3 CT 38% mm

FLx* 66% N

500

166

398

130

79.5

0.03

FLx* 38% N

C Mean 1462 1116 391 272 828 502

SD 283 238 49 49 91 107

A Mean 905 940 272 304 840 423

SD 522 121 116 31 80 54

% 61.9 84.2 69.7 111.9 101.6 84.2

p** <0.01 0.01 0.01 0.01 0.61 0.01

226

48

237

43

104.6

0.44

254 1135 4.6

36 35 0.7

183 1151 3.8

88 33 1.4

72.1 101.5 83.1

0.01 0.10 0.04

3370

661

2017

1147

59.8

<0.01

Radius site 4% (distal) 66% (shaft) Tibia site 4% (distal) 38% (shaft); SSIp0polar strength strain index. ToA0 total area. ToD0 total density. CoA0cortical area. CoD0cortical density. FLx0fracture load x axis. TrabA0trabecular area. TrabD0trabecular density. CT0Cortical Thickness*0log transformed for statistical analysis; **0Independent Samples T-test; A0Asian, C0Caucasion; %0value as a percent of C value

Conclusion: Overall, the smaller bone size in Asians is detrimental to bone strength, despite some adaptations in tibial cortical thickness and tibial and radial density to attempt compensation for this. P74 DYNAMIC EVALUATION OF VERTEBRAL CEMENT REINFORCEMENT FOLLOWING TOTAL DISC REPLACEMENT Adam Liddle (1) Vishal Borse (1) Daniel Skrzypiec (1) Ondrej Holub (1) Jake Timothy (2)Nikil Kapur (1)Richard Hall (1) presenting School of Mechanical Engineering, University of Leeds, Leeds, UK (1) Department of Neurosurgery, Leeds General Infirmary, Leeds, UK (2)

Introduction: Total disc replacement (TDR) is a popular alternative to interbody spinal fusion as a treatment for degenerative disc disease in younger patients. However, patients suffering from osteoporosis are currently denied the motion preserving benefits offered by the treatment due to elevated subsidence risk. This in vitro study aims to assess supplementary vertebral body (VB) cement augmentation following TDR in a dynamic environment, that more closely reflects the physiological environment. Materials and Methods: Twenty-four two-vertebra motion segments were harvested from cadaveric human spines (T11-L5). These were distributed into four groups; 1: control with no cement and a TDR with spiked fixation (control), 2: single level cement augmentation in the lower VB with spiked

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

TDR (single), 3: double level augmentation above and below a spiked TDR (double), 4: double level augmentation and a TDR with a keel fixation (keel). A 20 % cement fill was used in all cases, introduced via an anterior bilateral approach. All specimens underwent dynamic compression at 1 Hz for 24 hours. Loading regimes were tailored to each specimen according to fracture predictions. Subsidence depth and angle were calculated from micro-CT scans of post-test endplate impressions. Discussion: Average subsidence in the inferior VB was greatest in the control group, being significantly different to the single group (p0.041). Therefore, without the presence of cement, subsidence was greatest in the weaker superior endplate of the inferior VB. Analysis in the superior VB showed subsidence to be greatest in the single group, significantly higher than the double (p0.025) and keel cohorts (p0.037). Hence, when the inferior VB was augmented with cement, the subsidence was transferred to the superior VB. This, together with the overall reduced subsidence and increased stability displayed by the double augmentation specimens, provides evidence that cement placement is extremely important in reducing subsidence and maintaining TDR functionality. The use of keel fixation provided no benefit in terms subsidence reduction over spiked fixation but there was less variation in coronal tilting suggesting that the keel plays an anchoring role within the VB. Conclusion: Cement augmentation significantly reduced TDR device subsidence of which cement placement plays a crucial role. Variation in TDR fixation did not affect device subsidence. P75 BONE TURNOVER IN PATIENTS WITH CHRONIC SPINAL CORD INJURY Vladyslav Povoroznyuk (1) presenting Maryna Bystrytska (1) Institute of Gerontology AMS Ukraine, Kyiv, Ukraine (1) Objective(s): To determine the peculiarity of the bone turnover markers in patients with chronic spinal cord injury (SCJ). Material and Methods: 52 patients were divided into three groups. First group included 10 patients with early chronic SCJ (average age 26.0±1.44 years, duration of trauma period – under 1 year), the second group included 21 patients with late chronic SCJ (average age 32.4±1.6 years, duration of trauma period – over 1 year) and the third group consisted of 21 healthy individuals of appropriate age (average age of 33.8±1.5 years). Markers of bone turnover (osteocalcin, serum P1NP, serum β-CTx, and parathyroid hormone) were determined by the electrochemiluminescence method (Elecsys 2010, Roche). Results: The results of examination showed patients with early chronic SCJ had significantly higher bone formation

S585

and bone resorption markers than persons of second group: osteocalcin (41.5±10.1 vs 25.6±1.7 ng/ml, F– 4.5, p00.04), P1NP (235.4±68.7 ng/ml vs 57.1±4.2, F – 13.7 p<0.0001), serum β-CTx (1.77±0.17 vs 0.63±0.06 ng/ml, F – 14.6 p< 0.0001), parathyroid hormone (19.8±4.7 vs 42.4±5.3 ng/ml, F – 7.4 p00.01). The average level of bone turnover markers in patients of first group were significantly higher compared with healthy persons of control group: P1NP (235.4±68.7 vs 40.8±7.4 ng/ml, F – 13.1 p<0.0001), serum β-CTx (1.77± 0.17 vs 0.36±0.05 ng/ml, F – 15.3 p<0.0001), parathyroid hormone (19.8±4.7 vs 40.8±3.9 ng/ml, F – 4.9 p00.01). The level of osteocalcin in group 2 and 3 without difference (25.6 ±1.7 vs 27.7±4.4 ng/ml, F– 2.5, p00.09). It is important to note that the considerable difference between second and third group was only in serum β-CTx level (0.63±0.06 vs 0.36± 0.05 ng/ml, F – 10.1 p00.03). Conclusion(s): The bone formation and bone resorption markers in patients with early chronic SCJ were significantly higher than appropriate data of healthy individuals and persons with chronic SCJ with duration of trauma over 1 year. P76 SKELETAL STEM CELL ISOLATION USING PERIVASCULAR MARKER CD146, CD105 AND STRO-1: A COMPARATIVE INVESTIGATION BETWEEN SINGLE AND DUAL LABELLING David Gothard (1) presenting Richard Oreffo (1) University of Southampton, Southampton, UK (1) Introduction: Skeletal stem cells (SSCs) provide an ideal cell source for bone tissue engineering strategies, although their homogeneous isolation from human bone marrow (HBM) remains a significant unmet challenge. Stro-1 is a robust and reliable SSC surface marker. However, Stro-1+ populations exhibit variable colony forming unit-fibroblastic (CFU-F) capacity and osteogenic differentiation potential (bone stem cell pre-requisites). The present study has investigated the use of the pericyte marker CD146 and endothelial marker CD105 as alternative and additional markers to Stro-1 for homogeneous SSC enrichment. Material and Methods: Marker expression was quantified by flow cytometry within osteoarthritic HBM before magneticactivated cell sorting of single labelled populations and fluorescence-activated cell sorting of dual labelled populations. Isolated populations were characterised in vitro for CFU-F capacity, alkaline phosphatase (ALP) activity and osteogenic gene expression, and in vivo for osteogenic differentiation and function. Discussion: CD146+ and Stro-1+ populations exhibited both enriched CFU-F capacity determined by ALP+colony formation, and new bone matrix deposition in vivo following subcutaneous implantation of cell seeded diffusion chambers

S586

within immunodeficient mice, compared to control and CD105+ populations. Interestingly, gene expression analysis of select osteogenic genes ALP, CADM1, CLEC3B, DCN, LOXL4, OPN, POSTN and SATB2 showed Stro-1+ and CD146+ populations were not significantly different. Indeed, a small human bone marrow stromal cell fraction (2.53 %) exhibited positive immunolabelling for both Stro-1 and CD146. Conclusion: The data presents CD146 as an additional SSC enrichment marker to Stro-1, which targets a narrower cell population. However, molecular analysis demonstrated no significant difference between these enriched skeletal populations. The current studies highlight a further requirement for novel SSC specific markers to isolate populations for implementation in tissue engineering and regenerative medicine strategies. P77 OSTEOPOROSIS: IMPROVING THE COMMUNICATION BETWEEN SECONDARY AND PRIMARY CARE USING THE FRACTURE LIAISON SERVICE MODEL David Oni (1) presenting Malin Wijeratna (1) Dong Toon (1) Leslie Pilbrow (1) Juliet Compston (1) Addenbrooke's Hospital, Cambridge, UK (1) Introduction: Despite availability of effective drugs to reduce fracture, rates of treatment in high-risk patients are low. Fracture Liaison Services may improve treatment rates but there are no reported data on whether treatment recommended by Fracture Liaison Services in accordance with NICE guidance is implemented in primary care. Method and Materials: All women between the ages of 50 and 75 yrs presenting to the Fracture Liaison Service with a low trauma fracture in 2009 were included in the study (n0603). 473 women in total were excluded from the study: 123 women did not have a DXA scan and 350 women had a T-score≥−2.5. In the remaining women, a letter was sent to the GP from the Fracture Liaison Service recommending bone protective therapy. GP practices were contacted by telephone and the prescription rates of recommended therapy recorded. Results: A total of 130 women who were on the register of 57 general practices had a T-score≤−2.5 and were eligible for treatment. There was a 100 % response rate from GP practices. 123 (95 %) patients were found to be on appropriate bone protective medication for osteoporosis. 99 patients (74 %) were receiving calcium and vitamin D supplements. Conclusions: The majority of GP’s act on the recommendations of the Fracture Liaison Service and prescribe the recommended medication. The Fracture Liaison Service bridges the gap between secondary and primary care effectively.

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

P78 IS THERE AN ASSOCIATION BETWEEN ADOLESCENT FEMALES’ SKIN TONE AND BEHAVIOUR WITH RESPECT TO SUN EXPOSURE? Susan Fairbanks (1) Susan Fairbanks (1), Dr Katie E. Moss (1) presenting Catherine Collins (1) Philip Sedgwick (1) St George’s Hospital, University of London, London, UK (1) Introduction: Vitamin D deficiency is associated with metabolic bone disease. Most of the body’s vitamin D is produced in the skin following sun exposure. Levels of skin’s melanin content dictate the necessary exposure time for sufficient UVB absorption, with darker skin requiring longer periods. This research aimed to discover the behaviours behind the attitudes that influence teenage girls’ exposure to sunlight. Methods: In September 2011, using cluster sampling, a questionnaire was completed by 330 female students, from a cohort of 360 students, from two inner-city multi-ethnic schools in London.The age range was 16 to 18 years. Skin tone was self-selected. Discussion: Results, within groups analysis, of students not liking having a tan; 77 % of the ‘Dark’, 75 % of the ‘Medium’, 31 % of the ‘Olive’ and 31 % of the ‘Pale’. Reasons for covering the body; 'Religious'; 'Dark’ 14 (26 %), 'Medium’ 57 (44 %) and ‘Olive’ 20 (33 %) and 'Pale’ 4 (5 %). 'Cultural reasons'; 'Dark’ 20 (15 %), 'Medium’ 20 (15 %), 'Olive’ 5 (8 %) and 'Pale’ 5 (6 %). ] Time Spent Sunbathing Summer 2011

Never 1-3 Times week Most days Total

Skin Tone Pale Olive

Medium

Dark

Total

42 53.8% 32

40 67.8% 18

117 92.1% 10

49 90.7% 5

248 78.0% 65

41.0% 4 5.1% 78 100.0%

30.5% 1 1.7% 59 100.0%

7.9% 0 .0% 127 100.0%

9.3% 0 .0% 54 100.0%

20.4% 5 1.6% 318 100.0%

‘Dr Philip Sedgwick, 2011’ Count, % within Skin Tone Fisher’s Exact Test used to test the hypothesis. Test statistic049.58; P<0.0001. Conclusions: In our study population, the reasons for avoiding the sun varied with self-selected skin tone. There is a statistically significant difference in sun exposure behaviour revealing that groups requiring longer in the sun are not achieving adequate exposure.

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

S587

P79 MENOPAUSAL STATUS, LIPID PROFILE, AND BONE MINERAL DENSITY Yueh-Hsuan Sheng (1) Yen-Ching Chen (2) Keh-Sung Tsai (3) Jeng-Min Chiou (4) Yue-Yuan Lee (1) presenting ChienLin Mao (4) Shoei-Loong Lin (4) Jen-Hau Chen (4) Epidemiology & Institute of Preventive Medicine, College of Public Health, National Taiwan University, Taipei, Taiwan (1) Taipei Hospital, Department of Health,Taiwan, R.O.C,Taipei, Taiwan (2) Institute of Statistical Science Academia Sinica, Taipei, Taiwan (3) Department of Internal Medicine, National Taiwan University Hospital Bei-Hu Branch, Taipei, Taiwan (4) Department of Internal Medicine, National Taiwan University Hospital, Taipei, Taiwan (5) Introduction: Osteoporosis and low bone mineral density (BMD) have been related to elevated serum lipid levels, increased risk of atherosclerosis, vascular calcification, and stroke-related mortality. Studies explored the association between lipid profile and bone mineral density (BMD) have been inconsistent. In addition, no study has assessed this association longitudinally. Material and Methods: This is a retrospective longitudinal study. A total of 1,577 women aged 40 to 55 year-old were recruited from MJ health examination center in Taipei between 2009 and 2010, with an average follow up of 3.4 years. Spinal BMD was assessed by dual-energy x-ray absorptiometry. Demographic information on lifestyle and comorbidity were obtained from a self-reported questionnaire. Lipid profile was obtained from routine lab work. Logistic regression model was applied to assess the association of lipid profile with BMD for each follow-up. Stratification analyses were performed by menopausal status, which is an important risk factor of low BMD. Discussion: At baseline, BMD was significantly associated with body weights in premenopausal [adjusted odds ratio (AOR)00.96] and postmenopausal women (AOR00.87). But no significant was observed in women who were in the transition state (pre- to post-menopause) between baseline and follow-up. After follow-up, BMD was significantly associated with body weights (AOR 01.53) and LDL (AOR015.13) in both premenopausal and postmenopausal women. Waist circumference (AOR 01.02) and alkaline phosphatase (AOR01.04) were significantly associated with BMD in women who were in transition state. Serum alkaline phosphatase was significantly correlated with LDL

at baseline (r00.09, p00.006) and after follow up (r00.02, p<0.001). In premenopausal and postmenopausal status , BMD was associated with LDL while bone loss is in stable pace. In the menopause transition state, BMD was only associated with body weight and waist circumference. During this period women sex hormone was decreased rapidly, BMD may mainly be influenced by menopausal status and may not modify the relation between LDL and BMD. Conclusion: This study found that lipid profile was associated with BMD and menopausal status plays a role in modifying this association. P80 QUANTITATIVE ULTRASOUND MEASURES IN BANGLADESHI MOTHERS AND DAUGHTERS Diane Harper (1) presenting Katherine Brooke-Wavell (1) Barry Bogin (1) Loughborough University, Loughborough, UK (1) Introduction: One of the consequences of our ageing population is the high prevalence of osteoporosis. The aim of this study was to establish the bone status of a sample of Bangladeshi women, a minority ethnic group in the UK that suffers particularly poor health. Material and Methods: The data were taken from the study, MINA (Migration, Nutrition and Ageing across the Lifecourse in Bangladeshi families). 37 mother-daughter pairs resident in Cardiff, Wales were recruited. Mothers comprised both preand post- menopausal women. Quantitative ultrasound (QUS) measurements, Broadband Ultrasound Attenuation (BUA) and Speed of Sound (SOS), of the calcaneous were taken using an Osteometer DTU-one machine. Other variables recorded included age, height and weight. T tests were used to compare variables between mothers and daughters and to compare Bangladeshi women bone parameters with the DTU-one reference population (white North European). Pearson correlation coefficients were calculated for mother-daughter bone parameters. Discussion: Daughters had better bone parameter values than mothers but the significance disappeared when age-standardised z scores were used. There was more correlation between mother and daughter for SOS (r00.38, p00.03) than for BUA (r00.08, p0.67). The total sample of women (mothers and daughters) had significantly better SOS z-scores and worse BUA z-scores than the reference population, consistent with previous publications.

Table: Comparison of variables for mother and daughter (Mean±standard deviation) Mother (n037) Daughter (n037) T test results

Age (years) 55.7 ± 7.5 27.4 ± 5.5 p< .001

BUA (db MHz-1) 42.5 ± 7.4 46.5 ± 7.5 p0.01

BUA z-score -0.67 ± 0.78 -0.76 ± 1.05 P0.688

SOS (m s-1) 1549 ± 13.1 1559 ± 10.7 p0.001

SOS z-score 0.39 ± 0.78 0.29 ± 1.07 p0.710

S588

Conclusion: After taking age into account, mothers and daughters had similar QUS bone parameters with SOS correlating between mothers and daughters. QUS measurements from Bangladeshi women differed from the white reference population. P81 COX-2 SELECTIVE NSAIDS AND RISK OF HIP / KNEE ARTHROPLASTY: A POPULATION-BASED CASE–CONTROL STUDY Corinne Klop (1) presenting Paco Welsing (2) Arief Lalmohamed (1) Hubert Leufkens (1) Frank de Vries (1) Utrecht University, Utrecht, The Netherlands (1) University Medical Centre, Utrecht, The Netherlands (2) Introduction: Disease models of osteoarthritis (OA) have shown that COX-2 selective non-steroidal antiinflammatory drugs (NSAIDs; coxibs) may have beneficial effects on cartilage. Clinical or epidemiological evidence for this potential association is scarce. The objective of this study was to evaluate the risk of hip or knee replacement in users of coxibs as compared to non-selective NSAIDs. Methods: A population-based case–control study was conducted with the Dutch PHARMO Record Linkage System. Cases (n026,202) had a first replacement of the hip or knee after enrolment (2000–2009). Up to two controls (without hip or knee replacement) were matched by year of birth, gender and region. Using conditional logistic regression analysis, odds ratios (ORs) for hip or knee replacement were estimated by comparing long-term (≥ one year) nonselective NSAID use with long-term coxib use. Analyses were statistically adjusted for disease and drug history. Discussion: Long-term use of non-selective NSAIDs was not associated with a changed risk of hip replacement (adjusted OR 0.89; 95 % confidence interval (CI): 0.65 – 1.22) or knee replacement (adjusted OR 0.74; 95 % CI: 0.49 – 1.11) as compared to long-term coxib use. Results were not different after stratification to gender, age, cardiovascular or gastrointestinal disease. Conclusion: This study shows that long-term users of nonselective NSAIDs do not have a changed risk of hip or knee replacement as compared to long-term coxib users. Therefore, our results do not support that patients with osteoarthritis could benefit from using coxibs, in order to slow progression of this disease. Conflict of Interest: The division of Pharmacoepidemiology & Pharmacotherapy employing authors CK, AL, HL, FV has received unrestricted funding for pharmacoepidemiological research from GlaxoSmithKline, Novo Nordisk, the privatepublic funded Top Institute Pharma (www.tipharma.nl, includes co-funding from universities, government, and industry), the Dutch Medicines Evaluation Board, and the Dutch Ministry of Health.

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

P82 OSTEOPOROSIS TREATMENT ADHERENCE FOLLOWING HIP FRACTURE Thanda Aung (1) presenting Rowan Harwood (1) Tahir Masud (1) Nottingham University Hospitals NHS trust, Nottingham, Nottinghamshire, UK (1) Introduction:Amongst osteoporosis related morbidities, hip fracture has major physical, psychological and economical impacts on patients, families and societies. Effective pharmacological agents are available. Poor adherence is a major issue compromising the prevention of the second fracture. This study aimed to assess 3-month adherence (both “persistence” and “compliance”) with osteoporosis therapy after hip fracture, patient's self reported reasons and patient’s characteristics associated with non-adherence. Methods: A prospective observational study of hip fracture patients treated with oral bisphosphonates and calciumvitamin D was conducted. Patients with cognitive impairment, high impact and pathological fractures were excluded. Postal questionnaires identified patients' characteristics and telephone interviews 3 months following hospital discharge ascertained treatment adherence and patents' self reported reasons for non-adherence. Non-compliance was defined as taking ≤75 % of total prescribed medication during a 30-day period. Results: 125 participants (mean age 82 years, SD 6.69, 95 % females) were studied. 10 of 125 participants had stopped taking bisphosphonates by 3 months and 19 of 115 were non-compliant giving a persistence and compliance rate of 92%and 83 % respectively. For calcium-vitamin D, 2 discontinued and 6 were non-compliant (persistence 98 %, adherence 95 %). Combined adherence rate (bisphosphonate+calcium/vitamin D) was 72 %. Of the bisphosphonates non-persistent subjects, 40 % reported that medication ran out and 20 % had gastrointestinal side effects. Of the bisphosphonates non-compliant subjects, 84 % often forgot their tablets and 16 % reported gastrointestinal side-effect. All calcium-vitamin D non-persistent patients claimed they "disliked the taste”, and among non-compliant patients, two thirds reported "forgetting” and one third “inconvenience” as their reasons for poor compliance. Patients age ≥80 years, unmarried status, living alone, low tendency to go out, falls history, pain, further fractures, lack of osteoporosis knowledge, perception of good-excellent health were associated with bisphosphonate non-adherence and patients' lack of osteoporosis medication knowledge with calcium-vitamin D non-adherence. Conclusion: Three-month adherence with osteoporosis therapy was sub-optimal. Patients' lack of knowledge, forgetfulness, disliking tablets and gastrointestinal side effects were main reasons for treatment non-adherence. Improving patients’ education, regular review of patients and addressing

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

issues regarding medications, and prompts to remind people to take tablets may improve adherence. P83 USE OF A SIMPLE ELECTRONIC DATABASE FOR COHORT ANALYSIS OF OSTEOPOROSIS & FRACTURE RISK FACTORS INFORMS SERVICE DEVELOPMENT Nandkishor Athavale (1) presenting Bernd Franke (1) Kathryn Thompson (1) Lauren Bowden (1) Mary Holt (1) Rotherham NHS Foundation Trust, Rotherham, UK (1) Introduction: DXA axial bone density scans and the related osteoporosis risk factor questionnaires generate large volumes of data per patient, with multiple variables. We have designed and used a prospective, electronic database to collate and analyse this information. Analysis of one aspect of data stored in the first year (1.09.2009 – 31.08.2010) is presented and discussed. Materials and Methods: At the completion of the reporting process for the DXA scan, data for 54 variables are collected via a purpose designed, ‘tick-box’, A4 sheet, before being transferred to the electronic database, by optical character recognition. Discussion: For 1595 adult patients, paper reports were provided by our Consultant reporting team direct to the referring clinicians, of whom 63 % (1005) were GPs, and 35 % (560) were hospital clinicians. 1335 (83.6 %) patients were female and 255 (16 %) were male. 1336 patients had not been scanned previously by our service. Scan results showed that 32 % had normal bone density, 48 % osteopaenia and 19 % osteoporosis. For 389 (25 %) of patients, the primary reason for referral for DXA scanning was recent (within 12 months) low trauma fracture. For 563 patients (35 %) there was a history of previous low trauma fracture. The majority of the recent fracture patients 237 (61 %) were referred by GPs, with only 87 (22 %) coming from the Orthopaedics (including Fracture Clinic). For the 109 patients with vertebral fractures, the prevalence of osteoporosis was 37 % (40 patients) and thus twice that for the whole group. Forty six percent (50) had osteopaenia and only 19 (17 %) had normal bone density. These factors led us to introduce a system whereby case finding for our newly established ‘secondary fracture prevention service’ is facilitated by automatic notification by the Medical Imaging Department of all new fractures, including vertebral (but excluding skull and digital) occurring in adults age 50 years and above. Conclusion: Development and use of a simple electronic database has facilitated data collation and analysis of osteoporosis risk factors and bone density measurements, to the benefit of local service development.

S589

P84 BIRTH AND GROWTH OF THE MULTIDISCIPL INARY OSTEOGENESIS IMPERFECTA SERVICE Susan Stewart (1,2) presenting Umar Raja (1) Tarekegn Hiwot (1) John Ayuk (1) Mark Cooper (3) Trevor Cole (2,1) Neil Gittoes (1) Queen Elizabeth Hospital Birmingham, Birmingham, West Midlands, UK (1) West Midlands Regional Clinical Genetics Unit, Birmingham, West Midlands, UK (2) University of Birmingham, Birmingham, West Midlands, UK (3) Introduction: Osteogenesis imperfecta (OI) is a rare, complex, highly variable genetic disease characterised by bone fragility and tendency to fracture. Fractures occur most frequently in childhood and risk increases later in life due to age-related bone loss. Defects in the collagen genes also cause increased morbidity from joint hypermobility, dislocation, pain, dental problems and progressive deafness. Scoliosis and respiratory complications can feature. Patients require complex holistic care. Methods: In 2006 a newly combined clinical genetics and metabolic bone service prompted review of known OI patients within the region. 176 potential patients were identified from genetic databases. A small number of OI patients were known to the bone service, but only 1 patient was known to both services. Referrals to a specialist multidisciplinary OI clinic (metabolic bone consultants, clinical geneticist and specialist nurse) were invited from these adult OI patients. Detailed medical, fracture and family histories are taken from each patient. Thereafter clinical screening is individualised. Bone density scanning (DXA scan), localised hearing and dental assessment are often indicated. Discussion with patients includes the genetic inheritance of OI, prenatal testing, pregnancy, childbirth and risks for future offspring. Untreated relatives with OI are identified and offered further contact. Specific genetic mutation testing is not done routinely. Discussion: From 2007–2011, data were available for 67 patients seen, age 19–82 years. 46 DXA scans were performed (69 % of patients seen), 14 % had normal bone density, 20 % osteopenia and 64 % osteoporosis, indicating a need for ongoing assessment. 50 % patients required audiometry referral, and a smaller number specialist dental input. Conclusion: To help patients stay healthy and manage their OI, and positively impact on quality of life, this holistic clinic acts as an ongoing support and advice service. Follow-up is via an open door policy and open access to the specialist nurse. Brittle Bone Society contact details are provided. Where appropriate, pharmacological treatment is offered (as well as lifestyle advice) to minimise fracture risk. The service identified most affected adult members of the extended family were not under formal follow up.

S590

P85 QUALITY OF LIFE IN WOMEN WITHOUT A HISTORY OF FRACTURE REFERRED FOR BONE DENSITY ASSESSMENT Sally Wilson (1) presenting Christopher Sharp (1) Michael Davie (1) Robert Jones & Agnes Hunt Orthpaedic Hospital NHS Foundation Trust, Oswestry, UK (1) Introduction: The effect of fracture on Health Related Quality of Life (HRQoL) is of increasing interest, but the effect of osteoporosis by itself has received little attention. This cross-sectional study aimed to assess factors associated with HRQoL in women without a history of postmenopausal fracture referred for bone mineral density (BMD) assessment. Materials and Methods: SF36v2 was completed before BMD assessment by 154 postmenopausal women (64± 13y) without a history of fracture, cancer or any condition affecting bone metabolism. SF36v2 comprises 8 domains which form physical (PCS) and mental (MCS) scores. Correlations were performed between HRQoL and age, body mass index (BMI), Charlson co-morbidity index, 10y risk of osteoporotic fracture (FRAX®) and femoral neck BMD T-score (BMDFN). Age, BMI, co-morbidity index and BMDFN were used in multiple regression analysis to identify contributing factors towards HRQoL. Discussion: PCS scores in women with osteoporosis (n020) were not significantly different compared with those who were non-osteoporotic (n0134) (39±18 and 43±13 respectively). MCS scores were also not significantly different between the two groups (52±8 and 48±10 respectively). According to the SF36v2 interpretation guide (3 rd edition; 2011) the low PCS values for those with osteoporosis would be considered clinically relevant compared with normative data for age 60-64y (44 for PCS). Age (r 0−0.427, p00.000), BMI (r0−0.386, p00.002), FRAX® (r0−0.276, p00.026) and co-morbidity index (r00.265, p00.034) were negatively correlated with PCS scores. BMDFN was not significantly correlated with PCS scores (r0−0.019) and no variables correlated with MCS scores. In multiple regression analysis, age and BMI accounted for 27.6 % (p 00.006) of variance in PCS scores. BMDFN and FRAX® were not associated with HRQoL after controlling for age and BMI. Conclusion: Overall, patients referred for BMD assessment did not have reduced HRQoL, although in this small sample osteoporotic patients did have reduced PCS scores. Poor PCS scores were associated with increasing age and BMI and the inclusion of these variables in FRAX® may account for the relationship with HRQoL. These data suggest that in this group of patients, age and BMI have a greater effect on HRQoL than BMDFN. However, HRQoL in osteoporosis merits further investigation.

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

P86 AUDIT OF A FRACTURE LIAISON SERVICE DESIGNED TO BE PRIMARY CARE LED Bronwen Mackenzie-Green (1) presenting Richard Smith (1) Salisbury District Hospital Introduction: In this current economic climate there is increasing strain placed upon all services including fracture liaison. The introduction in April this year of a QOF tariff may lead to osteoporotic fractures being picked up within primary care. Working with our General Practitioner colleagues we had designed a fracture liaison service where GPs were requested to investigate and treat patients with fragility fractures for Osteoporosis. A letter was automatically generated to GPs from fracture clinic when the term “fragility” was used in the clinic letter. This novel service was subsequently audited after 12 months. We aimed to investigate whether all patients over 50 years who sustained a Colles fracture were being assessed for Osteoporosis at Salisbury District Hospital during 2010. Materials and Methods: 50 patients who had sustained a Colles fracture between the 1st January 2010 and 31st December 2010 were surveyed on whether they had been investigated for Osteoporosis or had received treatment. Clinical coding provided a list of 169 over 50 year olds who had sustained Colles fractures based on attendance to fracture clinic. The patients were telephoned and asked 5 questions. The audit was complete when the first 50 patients had been contacted. Discussion: The patients surveyed were from 29 of 58 GP practices. Their ages ranged from 51 to 100 years. The majority of patients were over 70 years and were female. Of the patients surveyed 56 % (28) had not been investigated for Osteoporosis and 8 % (4) were unsure. 26 % (13) of the patients surveyed had Osteoporosis. The majority of patients did not know whether they were affected as they had not been investigated. Of the 13 patients with Osteoporosis, 12 were on treatment. Two of the patients surveyed had sustained hip fractures following their Colles fracture. One of these had not been investigated for Osteoporosis and was not on any treatment. Conclusion: This audit clearly showed that pre-QOF, a fracture liaison service triggered by fracture clinic with the emphasis on action led by general practice resulted in insufficient osteoporosis investigation and treatment. P87 ASSESSMENT OF SOFTWARE TOOLS FOR MEASUREMENT OF BONE STRUCTURE Amy. J Riddell (1) presenting, Nigel. Loveridge (1), Mark. E Edwards (2), Elaine. M Dennison (2), Cyrus. C Cooper (2), Kate. A Ward (1) MRC Human Nutrition Research, Elsie Widdowson Laboratory, Cambridge, UK (1) MRC Lifecourse Epidemiology Unit, Southampton General Hospital, Southampton, UK (2)

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

Introduction: The analysis of images from bone density machines relies on manufacturer specific software. The software is often difficult to use and doesn’t easily allow interactive analysis for full interrogation of scans. Availability of generic image analysis software is limited. BoneJ, a free multi-platform and open access software plugin for Image-J, was developed as a more accessible and modifiable tool (Doube et al, Bone 2010). The aim of our study was to compare manufacturer software with BoneJ for analysis of high-resolution pQCT (HRpQCT) scans of whole bones. Material and Methods: Ten ex-vivo human femoral necks (FN) were analysed using HR-pQCT software (Scanco XTreme) and BoneJ. Initially the BoneJ data for trabecular bone volume/ tissue volume (BV/TV) was not independent of the area of bone within the image windows so that small bones had a lower BV/TV and cropping the image increased the BV/TV. Also mean trabecular thickness (Tb.Th) was over 1mm clearly inconsistent with the many reported values of between 100-400µm. Therefore, new macros were developed to address the technical issues and were compared to X-Treme results in two sets of images: 1) ex-vivo human FNs 2) in-vivo tibia scans (aged 72-80 years). All data were scanned at an isotropic resolution of 82µm with 100-110 slices per image. Correlations were used to test agreement between new macro and Scanco results for BV/ TV, Tb.Th and trabecular separation (Tb.Sp). Discussion: Results are presented as R2[slope, intercept]. The correlations in ex-vivo samples were; BV/TV = 0.75 [0.54, -2.95], Tb.Th = 0.63 [1.27, -0.24] and for Tb.Sp = 0.93 [0.74, 0.10]. The correlations for in-vivo tibia samples were; BV/TV = 0.87 [0.32, 5.23], Tb.Th = 0.67 [0.38, 0.034] and for Tb.Sp = 0.67 [0.35, 0.184]. Conclusion: The original comparative analysis in FN samples indicates the need for caution and a thorough assessment when applying new tools to image analysis. We have developed new macros to assess whole bone samples appropriate for assessment of BV/TV. The trabecular structure analysis requires further development. In conclusion BoneJ and the new macros are practical and easy to use and offer a manufacturer independent analysis of bone structure. P88 RELATIONSHIP BETWEEN SERUM IGF-I AND FRACTURE TYPES IN POSTMENOPAUSAL WOMEN Fatma Gossiel (1) presenting Judith Finigan (1) Richard Jacques (1) David Reid (2) Dieter Felsenberg (3) Christian Roux (4) Claus Glueer (5) University of Sheffield, Sheffield, UK (1) University of Aberdeen, Aberdeen, UK (2) Charite Universitatsmedizin, Berlin, Germany (3) Rene Descartes Universite, Paris, France (4) Universitaetsklinikum Schleswi-Holstein, Kiel, Germany (5)

S591

Introduction: Low IGF-I levels have been found to predict fractures in one previous study of women (Garnero et al, Lancet 2000;355:898–9). The aims of the present study were to 1) determine whether the association is with vertebral or non-vertebral fractures; and 2) to determine whether the effect is through low muscle strength. Material and Method: We studied a subset of the OPUS study who attended for a follow-up measurement of BMD and fracture assessment (including spine x-rays) after 6 years observation. OPUS is a 5-centre population-based study of women ages 55 to 80 years. Subjects who took antiresorptive treatments during the study period were excluded from this analysis. We had vertebral fracture data in 817 women; 23 had incident vertebral fractures; we had nonvertebral fracture data in 932 women; 103 had nonvertebral fractures. We measured IGF-I by immunoassay on non-fasting serum samples at baseline (IDS-iSYS) and BMD by DXA. The spinal radiographs were read centrally, the non-vertebral fractures were self-reported and verified. The IGF-I had a skewed distribution and so was log-transformed. Discussion: In a multivariable model, incident vertebral fracture was associated with IGF-I (RR00.467 per 2-fold increase, p00.027) and prevalent vertebral fracture (RR0 2.917, p00.040). Incident non-vertebral fracture was associated with femoral neck BMD (RR00.997 per mg/cm2, p0 0.004, prior self-reported fracture (RR01.893, p00.002) and weight (RR 01.022 per kg, p00.008), but not with IGF-I (p > 0.8). We examined the relationship between IGF-I and BMD and there was a significant correlation with total hip BMD (r00.13, p<0.001) and femoral neck BMD (r00.14, p<0.001) but not spine BMD, and percentage change in total hip BMD (r00.100, p00.003) and femoral neck BMD (r00.07, p00.030). We found a relationship between IGF-I and grip strength (r00.17, p<0.001) and history of falls (OR00.76 per 2-fold increase, p00.042). Conclusion: We conclude that low IGF-I is a risk factor for vertebral fracture but not for non-vertebral fracture. This effect may be mediated by both low muscle strength and low BMD. P89 EPIGENETIC MECHANISMS UNDERLIE NANOTOPOGRAPHICAL REGULATION OF SKELETAL STEM CELL FATE Emmajayne Kingham (1) presenting Nikolaj Gadegaard (2) Matthew Dalby (2) Richard Oreffo (1,3) University of Southampton, Southampton, UK (1)University of Glasgow, Glasgow, UK (2) King Saud University, Riyadh, Saudi Arabia (3) Introduction: The unmet need for autologous, patient-specific stem cells for regenerative therapies and improvements to

S592

osseointegration of implants could be achieved using unique nanotopographical surfaces. We have previously reported that a square arrangement of nanopits can maintain the skeletal stem cell state whilst displacement of this square arrangement by 50 nm in x and y axis can induce the osteogenic differentiation. However, an understanding of the mechanisms underlying nanotopographical regulation of skeletal stem cell fate will inform and provide new avenues for future developments in topography for regenerative medical applications. Epigenetic mechanisms are those which result in modifications to DNA and histones affecting gene expression without alteration of the DNA sequence. Materials and Methods: Human bone marrow derived STRO+skeletal stem cells were cultured on nanopatterned substrates displaying nanopits arranged in square and near square geometries in the absence of chemical osteogenic factors. Epigenetic mechanisms including microRNA expression, histone deacetylase (HDAC) expression and activity, global and site specific DNA methylation and chromatin structure were investigated using TaqMan assay, immunofluorescent staining and qPCR, methylation assay, pyrosequencing and chromatin accessibility assay. Results: Assessment of osteocalcin promoter region methylation revealed DNA methylation status to be developmental stage dependent with demethylation occurring between human fetal and human adult skeletal cell types. In addition chromatin accessibility within this region of DNA was enhanced following chemical or nanotopographical-induction of osteogenic differentiation and correlated with enhanced expression of osteocalcin. Conclusion: The use of nanotopographical surfaces negates the need for complex soluble chemical factors or cellular reprogramming in order to obtain clinically relevant cell types. P90 SPECIATION AND DISTRIBUTION OF COBALT AND CHROMIUM IN OSTEOBLASTS AND OSTEOCLASTS USING MICROFOCUS X-RAY FLUORESCENCE AND X-RAY ABSORPTION NEAR-EDGE STRUCTURE Karan Shah, Paul Quinn, Mark Wilkinson, Alison Gartland; The Mellanby Centre for Bone Research, University of Sheffield, Sheffield, UK (1) The NIHR Biomedical Research Unit for Musculoskeletal Disease, University of Sheffield, Sheffield, UK (2) Diamond Light Source Ltd, Harwell Science and Innovation Campus, UK (3) Introduction: We have recently shown that Co and Cr ions at clinically relevant concentrations have an adverse effect on human osteoblasts and primary osteoclasts function and survival in vitro. In this study we used Microfocus X-ray fluorescence (μ-XRF) and X-ray Absorption Near-Edge

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

Structure (XANES) to identify the intracellular species and distribution of these metals in human osteoblasts and osteoclasts. Methods: Human osteoblasts and osteoclasts were cultured on silica-nitride windows and treated with clinically relevant concentrations of metal ions. For each cell and treatment, two-dimensional elemental distribution map was generated by μ-XRF using a 9 keV beam and raster scanning samples with 4μmx2μm step-size and 1000 ms collection per step. XANES absorption spectra were generated at various points selected from μ-XRF maps to determine the chemical state. Results: The elemental distribution μ-XRF map of Co2+ treated osteoblasts demonstrated presence of Co throughout the cell body and the XANES spectra confirmed the species to be Co2+. The map of Cr3+ treated osteoblasts demonstrated presence of Cr; XANES spectra confirmed this to be Cr3+, predominantly Cr(III)phosphate. In Cr3+ treated mature osteoclasts, unlike osteoblasts, Cr localised to specific perinuclear sites. The speciation data indicated Cr presence mainly as Cr(III)phosphate. A similar distribution was observed with Cr6+ treated mature osteoclasts with the intracellular species being Cr(III)phosphate suggesting intracellular reduction. Neither metal was detected in negative controls of both cell types. Conclusion: This is the first study that investigates subcellular Co and Cr speciation in human osteoblasts and osteoclasts at concentrations found in patients post metalon-metal hip replacement. Our results suggest the oxidation state of Co2+ and Cr3+ remains stable upon entry into osteoblasts and osteoclasts, but Cr6+ undergoes intracellular reduction to Cr3+. Furthermore, Cr3+ thought to be relatively impermeable to the cell membrane, was detected within cells. The mechanism of cellular entry via various putative transporters or active processes through ion channels or endocytosis is under investigation. These studies help us better understand osteolysis due to prosthesis-derived metal ions and identify potential therapeutic opportunities. P91 THE EFFECTS OF GENDER AND OBESITY ON FRACTURE RISK AND CARDIOVASCULAR RISK Maslah Amin, Alice Manley, Amy Evans, Richard Eastell, Jennifer Walsh; NIHR Bone Biomedical Research Unit, Sheffield, UK (1) Introduction: Epidemiological evidence suggests that that fracture risk and cardiovascular risk are positively correlated. Fracture and cardiovascular disease have some common risk factors (e.g. age), but some divergent risk factors (gender and body weight). The aim of this study was to investigate how fracture risk relates to cardiovascular risk and carotid intima-medial thickness (CIMT) in older lean and obese men and women.

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

Methods: We recruited 92 healthy men and women ages 55– 75, with a body mass index 18.5 to 25 (lean) or greater than 30 (obese). Ten-year fracture risk was calculated with FRAX (including hip BMD), and ten-year cardiovascular risk was calculated using QRISK2 (including measurement of blood pressure and lipids). BMD was measured at the spine and hip by DXA. Ultrasound measurement of carotid intima-medial thickness was used as a surrogate of vascular risk. Results: Ten year fracture risk was higher in women than in men, and higher in lean people than obese people. Lumbar spine and hip BMD were higher in men than in women, and higher in obese people than in lean people. Ten year cardiovascular risk was higher in men than in women, and higher in obese people than in lean people. CIMT was higher in men than in women, and did not differ between lean and obese people. (All significant results two-way ANOVA p<0.05). Conclusions: Obese older adults have lower fracture risk and higher cardiovascular risk than lean older adults. Fracture risk was lower and cardiovascular risk was higher in men and in obese subjects. This was reflected by higher BMD in men and obese subjects and higher CIMT in men (although not in obese subjects). These results suggest that the positive correlation between fracture risk and cardiovascular risk may not apply in all groups within the population. P92 OSTEOPAENIA: SILENT BUT SERIOUS THREAT FOR BONE HEALTH - RETROSPECTIVE ANALYSIS OF FRAGILITY FRACTURES IN OSTEOPAENIC PATIENTS Myo Lynn (1) presenting Sarah Brannigan (2) Caje Moniz (3) Rheumatology Department, Queen Elizabeth Hospital, London, UK (1) Department of Rheumatology and Metabolic Bone Clinics, King's College Hospital, London, UK (2) Department of Rheumatology and Metabolic Bone Clinics, King's College Hopsital, London, UK (3) Introduction:Osteopaenia risk factors are the same as osteoporosis.Osteopaenia is normally asymptomatic, however a bone fracture could be the first presentation.Early diagnosis and prompt treatment for those with risk for fractures is ideal to prevent future morbidity and mortality.The aim of this study was to identify common fracture sites and re-fracture sites in osteopaenic patients. Material and Methods: Patients with bone fractures(total 134, male 41,female 93) presented in the fracture clinic with osteopaenia(established by Dual-emission X-ray absorptiometryDXA scan)between November 2007 and November 2010 at King’s College Hopsital, London were data-analysed retrospectively.Their demographics, ethnicity, risk factors for osteoporosis, trauma history and treament for bone health were

S593

analysed.Bone fracture sites and re-fracture sites were assessed.Re-fracture rate was calculated. Discussion: Mean age was 45.9 yrs (Age group: <60 yrs 095, 60–75 years025,>75 years013).Ethinicity showed White Caucasians 121,Black Caribbean 8, South Asian 5. 127 patients (94.8 %)gave falls history.7 (5.2 %)had no falls history. 60 (44.7 %)were on bone protection treatment at the time of fracture(34/134(25.3 %) on calcium +vitamin D, 26/134 (19.4 %) on bisphosphonate and calcium+vitamin D). The rest of 74 (55.3 %)were not on treatment. The study revealed that the most common fracture site was distal radius (27 %), followed by tibia, fibula, ankle (18 %), humeral(16 %), metatarsals(8 %), metacarpals(4 %), neck of femur (3 %), olecranon (3 %),calcaneum(3 %), scaphoid (3 %). A smaller group (15 %) had clavicle, lumber vertebra, glenoid, patella,phalangeal fractures.In patients with no falls,most common fracture site was metatarsals(4/7), followed by lumber vertebra(1/7), fibula(1/7), phalanges in foot(1/7).Total of 5 patients have re-fractured in either distal radius,calcaneum, clavicles, scaphoid or phalanges despite treatment.Re-fracture rate was 3.7 % (5/134). Conclusion: Bone fracture in osteopaenic patients is not an uncommon occurrence.Distal long bones are the most common fracture sites in this group.Non-trauma related fractures are seen in weight bearing metatarsals and thin bones.There is also a possibility of re-fracturing in those patients despite treatment.Fracture rate is the highest in white Caucasians and low in Black Carribeans.This study revealed that careful clinical evaluation and a FRAX score assessment are important in patients with fragility fractures even when they have a category of osteopenia by DXA scan. P93 PREVALENCE OF VITAMIN D DEFICIENCY IN PRE- AND POSTMENOPAUSAL SAUDI WOMEN- A CROSS SECTIONAL STUDY Hussein K.S (1,2), Al Kadi H.A (1,2), Lanham-New SA (4), Ardawi M.SM (2,3) Department of Physiology (1) & Centre of Excellence for Osteoporosis Research (2), Department of Clinical Biochemistry, Faculty of Medicine, King Abdul-Aziz University, Jeddah, Saudi Arabia (3); Nutrition and Metabolism Department, University of Surrey, Guildford. UK (4) Introduction and aims: Vitamin D deficiency is a global health problem, with a range of chronic conditions being associated with low circulating levels of 25(OH)D. The aim of this study was to determine vitamin D status in a sample of randomly selected pre- and postmenopausal Saudi women. Methods: This cross-sectional study was conducted in the Center of Excellence for Osteoporosis Research (CEOR). A total of 449 healthy women were randomly recruited from the city of Jeddah through Primary Health Care Centers. Data are presented on 226 premenopausal [20–39 years]; and 223 postmenopausal women [>51 years]. Menopausal

S594

grouping was confirmed by hormonal status. Fasting blood samples were collected for assessment of 25(OH)D status. Results & discussion: A total of 20 % of women had severe vitamin D deficiency with a serum level <12.5 nmol/L (29 % pre, 11 % post), 35 % of women had moderate deficiency with levels between 12.5- <25 nmol/L (42 % pre, 28 % post), 30.5 % of women had mild deficiency with levels between 25-<50 nmol/L (23 % pre, 38 % post), and 10.7 % had vitamin D insufficiency with levels between 50-<75 nmol/L (3.5 % pre, 18 % post). Only 3.8 % had sufficient vitamin D levels >75 nmol/L (2.5 % pre and 5 % post). Conclusion: Vitamin D deficiency is rather highly prevalent among both pre-and postmenopausal otherwise healthy Saudi women. Proper measures for the management and prevention of vitamin D deficiency are highly indicated to avoid the ill consequences of this deficiency. P94 EFFECT OF HYPOVITAMINOSIS D ON MUSCLE FUNCTION AND PHYSICAL PERFORMANCE IN SAUDI WOMEN Hussein K.S (1,2), Al Kadi H.A (1,2), Lanham-New SA (4), Ardawi M.SM (2,3) Department of Physiology (1) & Centre of Excellence for Osteoporosis Research (2), Department of Clinical Biochemistry, Faculty of Medicine, King Abdul-Aziz University, Jeddah, Saudi Arabia (3); Nutrition and Metabolism Department, University of Surrey, Guildford. UK (4) Introduction and aims: Vitamin D deficiency is associated with reduced muscle strength. Data on the effect of poor vitamin D status on physical performance in Saudi women are lacking. The aim of this study was to investigate the correlation between vitamin D level and measures of physical performance. Methods: This cross-sectional study was conducted in the Center of Excellence for Osteoporosis Research (CEOR). A total of 449 healthy women were randomly recruited from the city of Jeddah through Primary Health Care Centers. Data are presented on 226 premenopausal [20–39 years]; and 223 postmenopausal women [>51 years]. Fasting blood samples were collected for assessment of 25(OH)D status. Muscle function was assessed by the time taken to perform the following tests: get up and go (GUG); 8 feet walk (8FW); five-times sit to stand (5-STS). SPSS (version 16) was used for data analysis. Results & discussion: A total of 55 % of women were< 25 nmol/L (71 % pre, 39 % post), 86 %<50 nmol/L (94%pre, 77 % post), and 96 %<75 nmol/L (97%pre, 95 % post). Linear regression analysis was used to examine the association between vitamin D level and measures of physical performance adjusting for potential confounders. Vitamin D level was not among the predictors of physical performance measures in this study group. Conclusion: These data suggest that low vitamin D status is not associated with poor physical performance and may be a

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

reflection of muscle adaptation to prolonged, life-cycle of hypovitaminosis D. P95 IDENTIFYING RESPONDING TO SYSTEM FAILURES TO ACHIEVE BEST PRACTICE STANDARDS FOR HIP FRACTURE CARE Michael Harkness (1) presenting York Hospital, York, UK (1) The Department of Health offers a Best Practice Tariff (BPT) initiative for optimal management of hip fracture, which offers additional payment for cases of which the care meets agreed standards (surgery within 36 hours; joint care by surgeon and geriatrician, care protocol agreed by geriatrician, surgeon and anaesthetist; pre/perioperative assessment by geriatrician; geriatrician-led multi-disciplinary rehabilitation; secondary prevention including falls and bone health assessment) that are monitored by the National Hip Fracture Database. Over the past year, our Hip Fracture unit has consistently achieved a quarterly level of 50-60 % patients attaining BPT. A detailed quarterly review of patient records by a Consultant Orthogeriatrician was undertaken on the group not meeting BPT to identify the underlying causes. Failure to reach surgery within 36 hours was commonest (47 %), no orthogeriatric review within 72 hours (37 %) or both factors (16 %). Delays to surgery more frequently occur between Friday and Sunday. Patients on warfarin were often managed sub-optimally and delays were also common in those requiring further imaging. Limited weekend or Bank Holiday ward rounds and patients not being directly admitted to the Hip Fracture Unit resulted in the vast majority of delays for orthogeriatrican review. This process has been an effective use of time and has enabled our unit to identify key themes which require addressing to allow us to improve future care. There is an acknowledged need for a dedicated weekend trauma list but better planning of the weekday cases is also required. We aim to admit all newly diagnosed hip fracture patients from the Emergency Department to the Hip Fracture Unit to facilitate multi-disciplinary care at the earliest opportunity and to improve the routine senior medical review of patients outside of normal working hours. We have reviewed and highlighted protocols for reversal of warfarin prior to surgery and plan to streamline pathways for diagnostic MRI scanning. P96 HIP / KNEE ARTHROPLASTY AND RISK OF ACUTE MYOCARDIAL INFARCTION: A POPULATION-BASED COHORT STUDY Corinne Klop (1) presenting Paco Welsing (2) Arief Lalmohamed (1) Frank de Vries (1) Utrecht University, Utrecht, The Netherlands (1) University Medical Centre, Utrecht, The Netherlands (2)

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

Introduction: Joint arthroplasty surgery is associated with an increased risk of acute myocardial infarction (AMI), possibly due to bone marrow embolization. However, previous studies identified this risk only for short periods after surgery and without comparison to control subjects. Consequently, the objective of this study was to evaluate risk and determinants of AMI following hip or knee arthroplasty versus population-based controls. Methods: A cohort study was conducted using the Dutch PHARMO Record Linkage System (2001 – 2009). Patients with a first arthroplasty of the hip or knee (n022,021) were matched by age, gender and geographical location to three patients without arthroplasty. Time dependent Cox regression analyses were used to estimate risk of AMI in arthroplasty patients versus controls. Discussion: Risk of AMI was not increased within two weeks after arthroplasty of the hip (adjusted [adj.] HR 0.94; 95 % CI 0.23 – 3.79) or knee (adj. HR 0.76; 95 % CI 0.11 – 5.43), as compared to controls. This risk was not different for other time periods. Determinants of AMI among hip arthroplasty patients included male gender (adj. HR 1.52; 95 % CI 1.26 – 1.85) and age older than 60 years. Among knee arthroplasty patients, determinants included male gender (adj. HR 1.58; 95 % CI 1.20 – 2.7), prior ischemic heart disease (adj. HR 1.58; 95 % CI 1.06 – 2.33) and use of paracetamol (adj. HR 1.49; 95 % CI 1.08 – 2.04) in the six months before. Conclusion: Risk of AMI was not increased after hip or knee arthroplasty, as compared to controls. Determinants of AMI following hip or knee arthroplasty included male gender, age older than 60 years, recent use of paracetamol and prior ischemic heart disease. Conflict of Interest: The division of Pharmacoepidemiology & Pharmacotherapy employing authors CK, AL, HL, FV has received unrestricted funding for pharmacoepidemiological research from GlaxoSmithKline, Novo Nordisk, the privatepublic funded Top Institute Pharma (www.tipharma.nl, includes co-funding from universities, government, and industry), the Dutch Medicines Evaluation Board, and the Dutch Ministry of Health. P97 A REGIONAL ASSESSMENT OF DXA David Rawlings (1) Olwen Parry (2); Regional Medical Physics Department, Freeman Hospital, Newcastle-upon-Tyne, UK (1); Regional Medical Physics Department, University Hospital of North Durham, Durham, UK (2) Introduction: We wished to carry out a preliminary assessment of performance across 15 DXA machines in 13 NHS centres across the NE of England. We wished to compare bone density by machine and patient data by operator. Method: We compared operators through an observational audit of the

S595

most recently performed adult hip and spine scan. An experienced operator checked printouts against published indications. Machines were compared by scanning the European Spine Phantom (COMEC-ESP) and a pelvic skeletal phantom set in PMMA at proximal hip and lumbar spine. All scan programmes in routine clinical use were checked. Results: For the 51 operators contributing scans, 83 features inconsistent with published indications were found although some of these were most likely related to patient specific factors (such as scoliosis). Some vertebrae could not be identified with certainty. Long scan lengths sometimes led to unnecessary dose. There was no overall consistency in selection of the scan programme. Number of operators varied between 2 and 10 per machine. As expected, HOLOGIC and LUNAR bone mineral COMEC phantom densities did not agree, even after standard conversion. For the skeletal phantom there was some disagreement in T and Z score at all sites but particularly at the neck of femur. Conclusions: A further audit is proposed following circulation of a notionally “correct“ scan report in LUNAR and HOLOGIC format with key scan and analysis features annotated. Operators will be asked to score their own reports for a limited period and feed back results to complete the audit cycle. Guideline breakpoints for scan programme (based on patient weight) will be developed. Variation in T and Z score between machines may have implications for the increasing use of neck of femur DXA results as input to FRAX scoring. P98 BASELINE ANALYSIS OF USE OF DENOSUMAB IN OSTEOPOROTIC PATIENTS Myo Lynn (1) presenting, A L Dolan (2) Rheumatology Department, Princess Royal University Hospital, South London NHS Trust, UK (1), Queen Elizabeth hospital, London (2) Introduction: Denosumab is a human monoclonal antibody that inhibits osteoclast activity, reducing bone loss. Use of denosumab was approved by National Institute for Health and Clinical Excellence (NICE) from October 2010 (TA204). It is recommended in primary and secondary prevention of osteoporotic fragility fracture in post menopausal women. Denosumab was used in Rheumatology Department, Queen Elizabeth Hospital, London since October 2010. We audited indications for use of denosumab in these patients and compliance with NICE criteria. Material and Methods: 81 patients (female 77, male 4, mean age 72 years) who commenced between October 2010 and December 2011 were identified retrospectively. Demographics, risk factors, fracture history, bone mineral density (BMD), spine and hip T score and previous osteoporosis treatment were assessed. Renal function

S596

was monitored every six months. Side effects were documented. Discussion: Pre-treatment DXA was available on all patients (mean BMD spine and hip 0.686 and 0.705 , mean T score spine and hip −3.025 and −1.776). 87 % (75/81) failed treatment with both oral and intravenous bisphosphonates. 11 %(9/81) had breast cancer and were on aromatase inhibitors. 16 %(13/81) had mild renal impairment and 7 %(6/81) had moderate renal impairment. 35.8 %(29/81) of patients were found to have vitamin D deficiency at baseline and were treated. 9 %(7/81) were on treatment for primary prevention (T score range −2.3 to −4.1). All had failed oral medication previously. Four swapped from intravenous Ibandronate due to poor venous access. Use of denosumab was indicated clinically despite lack of specific risk factors suggested by NICE. All patients treated for secondary prevention (91 %,74/81) met NICE criteria by virtue of previous fracture and failure of other treatments. Four men were treated off-license. Conclusion: Denosumab seems well tolerated and no serious side effect has been observed. Patients’ renal function appears stable on treatment. Our patients satisfied NICE criteria for secondary prevention, not primary.

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

soluble triple helical collagen whereas dimers are likely acting on insoluble collagen fibers. The tetramer forms a central pore, which is sufficient to accommodate triple helical collagen. CD-analysis suggests that the tetrameric complex acts as a helicase by unwinding the collagen helix. Mutations within the pore, at sites of the chondroitin sulfate-protein and proteinprotein interactions of the complex significantly decrease the collagenase activity of the variant proteins without interfering with their non-collagenase activities. For the dimer, mutation at the protein-protein interaction sites specifically inhibits the degradation of insoluble collagen. AFM and degradation studies suggest that cathepsin K releases triple-helical collagen fragments from the fiber, which are subsequently degraded into smaller peptides. Conclusions: The collagenase activity of cathepsin K depends on specific dimer and tetramer forms of cathepsin K/GAG complexes. The inhibition of complex formation may represent a highly specific method to selectively inhibit the therapeutically relevant collagenase activity of this enzyme.

P99 MECHANISM OF COLLAGEN DEGRADATION BY CATHEPSIN K Vidhu Sharma (1) presenting Kevin Xin Du (1) Adleke Aguda (1) Preety Panwar (1) Gary D Brayer (1) Dieter Bromme (1) University of British Columbia, Vancouver, BC, Canada (1)

P100 EFFECT OF TERIPARATIDE ON SKELETAL REMODELLING IN TWO PATIENTS WITH CHRONIC KIDNEY DISEASE AND SURGICAL HYPOPARATHYROIDISM Ian Scott (2,3) presenting David Goldsmith (4) Ignac Fogelman (2) Geeta Hampson (1,2) Department of Clinical Chemistry, Guy's and St Thomas' Hospital. London, UK (1) Osteoporosis Clinic, Guy's hospital, london, UK (2) Rheumatology Unit, Guy's Hospital, london, UK (3) Renal Unit, Guy's Hospital, london, UK (4)

Introduction: Cathepsin K, a cysteine protease highly expressed in osteoclasts, is the only mammalian collagenase that effectively cleaves at multiple sites within triple-helical collagens. We have previously demonstrated that cathepsin K only exhibits its collagenase activity in the presence of glycosaminoglycans (GAGs). Based on X-ray structures of cathepsin K-chondroitin sulfate complexes, we suggest that the active collagenase forms either tetramers or dimers of cathepsin K molecules. These oligomers are responsible for the degradation of soluble and insoluble collagen. Methods: Various cathepsin K mutants were expressed in Pichia pastoris and purified. The enzymatic activity of the mutants as well as wild type cathepsin K were tested using soluble and insoluble type I collagen, gelatin and a fluorogenic peptide substrate. Atomic force microscopy (AFM) and circular dichroism were used to characterize the interactions between cathepsin K and collagen. Further, electron microscopy (EM) was used to characterize the degradation of collagen fibers. Discussion: The most effective degradation of soluble and insoluble collagen is achieved at molar ratios of 2:1 (cathepsin K : GAG). AFM studies indicate that a tetramer is active at

Introduction: Use of Teriparatide is contra-indicated in patients with chronic kidney disease (CKD) although it may be useful in these patients after parathyroidectomy. The objective of this case study was to investigate the efficacy of Teriparatide in 2 patients with CKD , low bone mineral density (BMD), fragility fractures and surgical hypoparathyroidism. Material and Methods: We describe 2 post-menopausal women with primary renal disease; patient A and B aged 67 and 61 years respectively. Patient A had rapidly progressive glomerulonephritis. Patient B was diagnosed with focal glomerulonephritis. Both patients developed end stage renal disease (ESRD) requiring haemodialysis and underwent total parathyroidectomy for severe secondary hyperparathyroidism. They had successful cadaveric renal transplant in 2003 and continued on immunosuppression, prednisolone and one-alfacalcidol. They had a BMD diagnosis of osteoporosis (Patient A ;lumbar spine : ‘T’ score −4.7, hip −2.6, Patient B; lumbar spine : -3.0, hip : -1.5). Patient A was treated with Alendronate but sustained a number of fractures despite treatment. Patient B stopped Alendronate after 1 year due to

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

side-effects. They were referred to the metabolic bone clinic where treatment with Teriparatide was initiated. Discussion: The bone turnover markers (serum P1NP and CTX) rose early and markedly in both patients with an interesting time course in Patient A (Patient A ;P1NP (ug/L) baseline : 62, 3 months : 692, 6 months : 843, 9 months : 547, 12 months : 171. CTX (ug/L) baseline 0.626, 3 months : 2.44, 6 months : 5.4, 9 months : 3.5, 12 months : 0.76, Patient B ; P1NP baseline :62, 3 months : 506, CTX baseline :.254, 3 months :.908). BMD increased by 10.4 % at the lumbar spine at 6 months in patient A. A bone scan revealed a metabolic ‘superscan’. Both patients developed hypercalcaemia which decreased following reduction of one-alfa calcidol. Conclusion: The skeletal response to Teriparatide is not impaired in patients with CKD and surgical hypoparathyroidism, despite previous treatment with bisphosphonates. The use of Teriparatide for fracture prevention should be considered in these patients. P101 HIP FRACTURE PREVENTION: WHY DO OLDER PEOPLE NOT ATTEND THE ‘FALLS CLINIC’ ? Atef Michael (1) presenting Joan Stellman (1) Russells Hall Hospital, Dudley, UK (1) Introduction: 90 % of hip fractures in the elderly result from a fall. Prevention in the form of bone protection and falls assessment is of proven value in reducing the incidence of hip fractures. The Falls Clinic is an ideal facility to provide clinical assessment and multidisciplinary intervention to prevent falls. It has been observed that some older people do not attend their falls clinic appointment. Aim: To study the “Do Not Attend (DNA)” rate at the Falls Clinic and explore the reasons why patients do not attend. Methods: The DNA rate and reasons for non-attendance, based on nursing staff and ward clerks’ telephone discussion with patients/family, were collected from a retrospective analysis of consecutive Falls Clinics in a UK teaching hospital in a 12-month period. Results: Over 12 months, there were 46 Falls Clinics. Data was incomplete for 3 clinics. 43 clinics were analysed. A total of 184 new patients were sent appointments, 112 attended (39%DNA). There were 192 follow-up appointments, 142 patients attended (26%DNA). Overall DNA rate was 33 %. Reasons of non-attendance were mainly classified into: * Poor communication and administrative error: Some patients did not receive the appointment letter. Some did not know why they were given the appointment. * Under-estimatation of the seriousness of falls or denial of the need for falls assessment: Some patients said they did not need to come as they had only had one fall. Some said that the fall was accidental "just tripped" and felt they did not need falls service.

S597

* Stigma: Some patients did not like the“stigma”of falls and denied falling. * Forgetfulness: Some patients had forgotten that they had a fall as it was a“while ago”and they are “Okay now”. * Health reasons: Some patients were unwell or hospitalized. Conclusions: One in three patients did not attend the Falls Clinic appointment. This high DNA rate may be because older people underestimate the significance or seriousness of falls or may deny that there is a problem. Some elderly do not like to be stigmatised as having falls. Lack of communication also plays a role. P102 EXTENT OF OBESITY IN SAUDI WOMEN AND THE ASSOCIATION BETWEEN VITAMIN D STATUS AND DIFFERENT MEASURES OF ADIPOSITY Hussein K.S (1,2), Al Kadi H.A (1,2), Lanham-New SA (4), Ardawi M.SM (2,3) Department of Physiology (1) & Centre of Excellence for Osteoporosis Research (2), Department of Clinical Biochemistry, Faculty of Medicine, King Abdul-Aziz University, Jeddah, Saudi Arabia (3); Nutrition and Metabolism Department, University of Surrey, Guildford, UK (4) Introduction and aims: Obesity is associated with vitamin D deficiency. The aims of this study were to: 1) determine the extent of obesity in Saudi pre- and postmenopausal women; 2) and to assess the association between different measures of adiposity and 25(OH)D in both groups. Methods: This cross-sectional study was conducted in the Center of Excellence for Osteoporosis Research (CEOR) during the year 2011. A total of 449 healthy women were recruited from Primary Health Care Centers. Data are presented on 226 premenopausal [20–39 years]; and 223 postmenopausal women [>51 years]. Menopausal grouping was confirmed by hormonal status. Fasting blood samples were collected for assessment of 25(OH)D status. Weight, height, waist circumference (WC), hip circumference (HC) and total body fat (TBF) by dual energy X-ray absorptiometry were measured. Waist-to-hip ratio (WHR) and body mass index (BMI) were calculated. Results & discussion: A total of 30.7 % of the women were overweight with BMI 25-<30 kg/m2 (28.3 % pre, 33.2 % post), and 38.5 % were obese with BMI≥30 kg/m2 (21.2 % pre, 56.1 % post). After adjusting for age, there was no significant correlation between 25(OH) D and any of the obesity measurements in the premenopausal women. A significant negative correlation between BMI (r0−0.203, P< 0.01), TBF (r0− 0.340, P<0.01) and WC (r00.140, P< 0.05) was found in the postmenopausal women. Conclusion: Obesity was more prevalent among the postmenopausal women. Obesity -associated vitamin D insufficiency is likely due to the decreased bioavailability of vitamin D3 because of its deposition in body fat compartments.

S598

Measures to reduce weight in this group may improve vitamin D status. P103 SIDE EFFECT PROFILE OF DENOSUMAB IN CLINICAL PRACTICE Paul Ryan (1) presenting Colletee Anderson (1)Medway Maritime Hospital, Gillingham, UK (1) Introduction:Clinical trials show a low side effect profile for Denosumab but there is limited data from clinical practice. This study assessed possible side effects of Denosumab following first injection in clinical practice. Methods: Patients were of average age 68 years, 6 male and 62 female. Treatment for Denosumab was approved and funded by the Kent PCT’s for osteoporotic patients who were in renal failure (CKD<30), had poor venous access or in whom bisphosphonates or strontium ranelate could not be used or were contraindicated. Patients were contacted by phone 1–2 weeks after the treatment and then seen in out patient clinic 2–3 months post therapy. Data was available on the first 37 patients. Side effects were reported in 11 patients. These were grouped as smelly urine: 3 pts which persisted for 2–3 months, urinary frequency: 1 pt, discomfort passing urine: 1 pt, urgency of micturition: 1 pt, UTI: 1 pt, chest infection: 2 pts (1 developed after 3 days and 1 after 2 weeks), intermittent diarrhoea for 3 months: 1 pt, unusual taste: 1 pt, itching 1 pt:, eczema patches for 2 months: 1 pt, worsening arthritis 1 pt, and profound hypocalcaemia requiring admission 1 pt. Discussion: Denosumab is well tolerated in most patients but there maybe a higher frequency of side effects than clinical trials would indicate, which can be prolonged. P104 EXPERIENCE OF STRONTIUM RANELATE IN THE TREATMENT OF OSTEOPOROSIS IN A TERTIARY REFERRAL CENTRE IN THE UK Author withdrawn P105 VITAMIN D DEFICIENCY AND BONE MINERAL DENSITY Vladyslav Povoroznyuk (1) presenting Nataliya Balatska (1) Fedir Klymovytsky (1) Omelyan Synenky (1) Volodymyr Vayda (1) Institute of Gerontology AMS Ukraine, Kyiv, Ukraine (1) Objective(s): to study determine the frequency of vitamin D deficiency and insufficiency and it influence to bone mineral density (BMD) in patients of different region of Ukraine. Material and Methods: It was examined 1575 people aged 20–95 yrs. old who lived in different regions of Ukraine. 25-

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

OH vitamin D and PTH level was evaluated by electrochemiluminescence method (Elecsys 2010, Roche). Vitamin D deficiency was diagnosed in level of 25-OH vitamin D below 49.5 nmol/l, vitamin D insufficiency – between 74.5 and 50.0 nmol/l. BMD was determined by ultrasound densitometry Sahara (Hologic) and DXA (Lunar). Results: Vitamin D deficiency was registered in 81.8 % persons, 13.6 % examined had vitamin D insufficiency. It was determined negative significant correlation between PTH and 25OHD (r 0−0.16, p < 0.0000001). Secondary hyperparathyroidism was diagnosed in 11.9 % patients. The mean level of 25OHD was significantly higher in southern resident of the country (p < 0.001) and during summer (p < 0.05). No significant correlation between 25OHD and BMD was found. But, only patients with vitamin D deficiency had significant negative correlations between PTH level and BMD at the level of femur neck (r 0−0.12, p < 0.004), dual femur (r 0−0.09, p < 0.004), upper and lower extremities (r 0−0.11, p < 0.01 and r 0−0.10, p < 0.01 accordingly), forearm 33 % (r 0−0.20, p < 0.001). Conclusion(s): In Ukrainian population the frequency of vitamin D deficiency is 81.8 %. Only patients with vitamin D deficiency have significant negative correlations between PTH level and BMD at the level of femur neck, dual femur, forearm 33 %, upper and lower extremities. P106 THE IMPACT OF DAILY ORTHOGERIATRIC INPUT ON TREATMENT FOR SECONDARY FRACTURE PREVENTION IN PATIENTS WITH HIP FRACTURE Jennifer Ruddlesdin, Amit Pramanik, Ray Hyatt; East Lancashire Hospitals NHS Trust, Blackburn, UK Introduction:Half of all hip fracture cases come from the 16 % of the post-menopausal female population with a history of previous fracture (1). A Royal College of Physicians audit however found that only 25 % of females over 75 years with prior fragility fracture had treatment for osteoporosis. Implementation of a systematic approach to secondary fracture prevention resulted in a significant reduction in expected hip fracture rate (2). Intervention is therefore both effective and cost efficient (3). Standards set out by the British Geriatrics Society and the British Orthopaedic Association include involvement of an orthogeriatrician in the care of patients with fragility fractures, and is one of the Best Practice Tariff (BPT) indicators. On this background, a second orthogeriatrician was appointed to East Lancashire Hospitals NHS Trust in August 2011, increasing input to the acute trauma ward from twice weekly to daily. Method: All patients with hip fracture are inputted onto the National Hip Fracture Database for the purposes of the BPT.

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

We analysed data in the 3 months previous to, and the 3 months subsequent to the intervention. Discussion: The number of patients commenced on bone protection doubled in the second half of the study (21 % vs 47 %), and those patients in whom it was felt inappropriate reduced in number (51 % vs 33 %). The patient population was essentially unchanged, implying that there were significant numbers previously untreated in whom treatment would have been appropriate, and that the orthogeriatrician was more confident in deciding on those who would benefit. Following intervention there were almost no patients (0.6 %) who had no assessment of their suitability for bone protection (14 % previously). Conclusion: Overall, bone protection assessment and prescription in patients with hip fracture has improved, thereby resulting in better patient care, improved compliance with the BPT and subsequent increased revenue for the Trust. We plan to further increase the focus on prevention through the development of fracture liaison services, and by greater integration with falls management within secondary care and between primary and secondary care. References: 1. British Orthopaedic Association. The care of patients with fragility fracture. 2007. Available for download from: http:// www.fractures.com/pdf/BOA-BGS-Blue-Book.pdf 2. The Clinical Effectiveness and Evaluation Unit of the Royal College of Physicians, London. National Audit of the Organisation of Services for Falls and Bone Health for Older People. 2009. Available for download from: http:// www.rcplondon.ac.uk/clinical-standards/ceeu 3. McLellan AR, Reid DM, Forbes K et al. NHS Quality Improvement Scotland. Effectiveness of Strategies for the Secondary Prevention of Osteoporotic Fractures in Scotland. 2004. Available for download from: http://www.nhshealth quality.org/nhsqis P107 FALLS AND FRACTURES IN PATIENTS WITH TYPE1 DIABETES Raashda A Sulaiman (1) presenting Parth Narendran (1) Mark Cooper (1) John Ayuk (1) Tarekegn Hiwot (1) Neil Gittoes (1) University Hospital Birmingham, Birmingham, UK (1) Introduction: Patients with type 1 diabetes have a higher risk of fracture and lower bone mineral density. We performed a postal survey to assess the incidence of fractures in a cohort of patients with type 1 diabetes in the West Midlands. Materials and Method: Questionnaires were sent to 90 patients with type 1 diabetes who were over the age of 50 years. All were actively attending a diabetes clinic at University Hospitals Birmingham. Anonymised data was collected for age, gender, ethnicity, duration of type 1 diabetes and its related complications, number of falls and fractures,

S599

exercise and other risk factors including smoking, alcohol intake, steroid use, family history and bone specific treatment. Discussion: Seventy one questionnaires were returned. Mean age of responders was 67.4 +/−8 years. Overall, 30 % reported fractures. Fractures were more common in females. Falls were reported significantly more in the group with history of fractures (50 %), as compared to those with no history of fracture (14 %). Of 21 patients with a history of fracture, only three were taking Bisphosphonates. Conclusion: There was a high prevalence of both falls and fractures in this cohort of patients with type 1 diabetes. We propose that specific and detailed assessments of falls and fracture risk should be integral to the management plan for patients with type 1 diabetes. Appropriate measures aimed at fracture prevention may be considered early in this ‘at risk’ and easily identifiable group. P108 APPROPRIATE USE OF IV ZOLEDRONATE IN OSTEOPOROSIS TREATMENT - A DISTRICT RHEUMATOLOGY UNIT EXPERIENCE Sreekanth Vasireddy (1) presenting Laura Bromilow (1) May Bishop (1) Bolton NHS Foundation Trust, Bolton, Greater Manchester, UK (1) Introduction: IV zoledronate is an effective treatment for osteoporosis. Infusion delivery costs make zoledronate more expensive than off-patent oral bisphosphonates. In Bolton, an IV infusion team delivers zoledronate at home or in a community health centre when prescribed from our department. We aimed to assess if appropriate patients were receiving zoledronate and if reasons were appropriately documented when choosing zoledronate. Methods: Details of patients receiving zoledronate on our database were reviewed excluding those with indications other than osteoporosis for zoledronate. Records of a random sample of 25 % were reviewed in greater detail. Statistical analysis was done in SPSS ver 17. Discussion: 121 patients (107 female) were currently receiving zoledronate. Mean age at 1st infusion was 72.1±10.8 yrs (range 46–98). 42.1 % had received 1 infusion, 24.8 % received 2, 24 % received 3, and 9.1 % received 4 (median02). Case records of 30 patients (23 female) were reviewed in detail. 25 had at least one fragility fracture. 24 had a DXA scan before 1st zoledronate infusion. Mean lumbar T Score was – 2.59, and mean total hip T Score was – 2.63. Despite high prevalence of vertebral fractures (50 %) in this sample, correlation remained high between lumbar and total hip BMD (Spearman correlation 0.60, P00.003). Before zoledronate, 13 (43.3 %) did not have any oral anti-resorptive treatment, 6 (20 %) had at least 1 anti-resorptive drug, 9(30 %) had 2 antiresorptive drugs, and 2(6.7 %) had 3 anti-resorptive drugs. Reason for choosing zoledronate was not documented in 2

S600

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

(6.7 %). Reasons documented were: GI intolerance to oral bisphosphonates (17, 56.7 %); contraindication for oral bisphosphonates (5, 16.7 %); other significant diagnoses (3, 10 %); and unable to comply with instructions (1, 3.3 %). Conclusion: Most patients had fragility fractures, but more than a third did not have documentation of a trial of oral anti-resorptive treatment before the first infusion of zoledronate. However, the reason for choosing zoledronate was documented for most patients, the majority being intolerance to or contraindication for oral bisphosphonates. It is therefore possible that documentation of previous oral antiresorptive therapy was incomplete. We would recommend that oral anti-resorptive therapy is considered and outcome appropriately documented before choosing zoledronate for osteoporosis treatment.

compounds. These compounds can be grouped into polyanionic and polyaromtic compounds. Of particular interest are the polyaromatic compounds, which contained flavonoid and steroid-like compounds. IC50 values of collagenase inhibitors were between 6–80 μM. Interestingly, some of the identified cathepsin K inhibitors are known antiresorptive drugs. Though the mechanism of those compounds has been primarily described as transcriptional modulators or compounds interfering various signal transduction pathways, our studies reveal a direct inhibition of cathepsin K. Conclusion: We identified a novel class of cathepsin K inhibitors, which specifically target the collagenase activity of the protease. These inhibitors may overcome intrinsic offtarget effects of active site –directed inhibitors presently in development.

P109 IDENTIFICATION OF EXOSITE INHIBITORS OF CATHEPSIN K AS SELECTIVE COLLAGENASE BLOCKERS Kevin Xin Du (1) Jadwiga Kaleta (1) Vidhu Sharma (1) Allison Nicholls (1) Tom Pfeifer (1)Dieter Bromme (1) presenting University of British Columbia, Vancouver, BC, Canada (1)

P110 ORAL BONE SPARING THERAPY IN A HIP FRACTURE PATIENT COHORT Clare Batten (1) presenting Mary Lawrence (1) Nuttan Tanna (1) Northwick Park Hospital, Harrow, Middlesex, UK (1)

Introduction: Cathepsin K is the major collagen degrading proteolytic enzyme in osteoclasts and therefore a major pharmaceutical target for the treatment of osteoporosis. Several cathepsin K inhibitors are presently in clinical trials for osteoporosis. However, some drug candidates failed in phase II trials due to off-site effects. Besides its function in bone degradation, cathepsin K has been implicated in extracellular matrix degradation in various non-osteoclastic cells as well as in the processing of regulatory proteins which may count for the observed side effects. Thus, the selective inhibition of the collagenolytic activity and the sparing of its non-collagenolytic activities would be a desirable feature of novel cathepsin K inhibitors. The collagenase activity of cathepsin K depends on the formation of oligomeric complexes with glycosaminoglycans. The inhibition of the formation of such complexes would lead to the selective inhibition of their collagenase activity. Using specific assays discriminating between the inhibition of complex formation and the inhibition of the active site of the protease allowed us to identify selective anti-collagenase inhibitors of cathepsin K. Methods: A fluorometric polarization assay was developed to screen for anti-complex formation inhibitors of cathepsin K. Fluorogenic peptide substrate assays were used to exclude inhibitors affecting the active site of the protease. A collagen-degradation assay was used to verify the anticollagenase activities of inhibitors. Discussion: We have screened a 3500-member drug library and have identified several selective anticollagenolytic

Introduction: This paper reports on an audit, undertaken within a secondary care setting, to assess bone-sparing discharge medication in hip fracture patients. Method: 102 consecutive patients discharged between May to October 2011 from the orthopaedic service in a district general hospital were identified using patient codes. Computerised discharge summaries were reviewed for criteria including age, sex, survival, and discharge medication. Specifically data on oral bone-sparing agents and calcium and vitamin D supplementation were collated. Results and Discussion: Of 102 patients, 4 were duplicate entries, one an overseas patient and one had no discharge summary, resulting in 96 for review. A further 9 juvenile cases were excluded. Ten patients had died within two months of their hip fracture and 5 were transferred to other units. Therefore a total of 72 records (49 female; 23 male) were available for final assessment. Given that the literature cites lack of consideration of osteoporosis management by orthopaedic units (Brit Ortho Assoc 2007; Bukata SV et al 2011; Moroni A et al 2011), it was encouraging to note that 78 % of our patient cohort had been discharged on specific bone-sparing therapy (43 % bisphosphonate; 35 % strontium ranelate). 59 patients had in addition been prescribed Adcal D3. Owing to hypercalcaemia, three had no calcium and vitamin D supplementation. NICE TA 161 (2011) recommends treatment with bisphosphonates, raloxifene and strontium ranelate, with bisphosphonates rated the most cost-effective therapy. None of these patients, age range 57–99, were prescribed raloxifene. A large proportion (22; 31 %) had dementia; 10 were prescribed alendronate and 7 strontium ranelate. One patient prescribed strontium ranelate

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

had a history of thrombotic episodes, whilst one developed pulmonary emboli after starting this medication. Clinical governance suggests a need to improve training for junior doctors prescribing bone-sparing therapy. Another strategy to ensure safer prescribing would be to have the discharge pharmacist flagging up compliance and adherence issues in dementia with bisphosphonates and strontium ranelate, and thrombotic risk factors with strontium ranelate. Conclusion: This audit highlights a need for training to ensure safer bone-sparing therapy prescribing for hip fracture patients. P111 HIGH DOSE ERGOCALCIFEROL IN VITAMIN D DEFICIENCY Paul Ryan (1) presenting Medway Maritime Hospital, Gillingham, UK (1) Introduction: The treatment of vitamin D deficiency with conventional calcium and vitamin D therapies is often insufficient to consistently raise vitamin D levels to the normal range. This study investigated the value of high dose ergocalciferol. Methods: 52 patients, 15 male and 37 female were treated with ergocalciferol 50,000 units weekly for 8 weeks. Patients were of mean age 61 years (range 24–91). Vitamin D measurements were made with an IDS Elisa assay which detected 100 % Vit D3 and 75 % Vit D2. Most patients had vitamin D deficiency or marked insufficiency with mean pre treatment 25OH Vit D 26 nmol/l (SD 7.6). Baseline measurements were distributed 0–25: 24 pts, 25–50: 27 pts, 50 – 75: 1 pt. Repeat vitamin D measurements were made 12 weeks post therapy with the mean value 78 nmol/l (SD 27.5) a rise of 52 nmol/l (SD 28) or 224 %. Post therapy values were distributed 0–25 nil pts, 25–50: 5pts, 50–75: 25 pts and>75: 22 pts. Discussion: A short course of high dose oral ergocalciferol substantially raised vitamin D measurements with no patients deficient and few significantly insufficient post treatment. Given the assay underestimated D2 the actual levels of Vitamin D2 and D3 combined would be even higher. The above regime was well tolerated and achieved greater responses than usually prescribed calcium and vitamin D. P112 ASSESSING THE RISK OF FRAGILITY FRACTURE IN A MEDICAL IN-PATIENT POPULATION AT A DISTRICT GENERAL HOSPITAL Ian Lyons (1) presenting Rashmi Mathur (1) Grantham and District Hospital, Grantham, Lincolnshire, UK (1) Introduction:At present the National Osteoporosis group guidelines recommend screening for osteoporosis be

S601

performed on an ‘opportunistic’ basis. We considered the frequency of risk factors for osteoporosis in a medical inpatient population at a district general hospital and the proportion of patients for whom further investigation and/or treatment would be recommended, to determine whether this should be a routine assessment in all patients. Materials and Methods: Information on clinical risk factors for osteoporosis was obtained from patient notes for all medical in-patients at Grantham and District Hospital on 29/11/11. FRAX scores for these patients were calculated as indicated by the guidelines and the investigation and treatment of patients audited against guideline recommendations Discussion: 101 patients, representing the entire general medical in patient population and Grantham and District hospital, were assessed for the presence of risk factors for Osteoporosis; 6 of these were excluded on the basis of age, or being pre-menopausal. Of the remainder 75 patients (76 %) were found to have 1 or more risk factor for osteoporosis – of these treatment for osteoporosis without assessment of fragility fracture risk was indicated in 10 patients. FRAX score calculation (prior to Bone mineral density (BMD) assessment) indicated that 35 % of patients required a Dual X-ray absorptiometry (DXA) scan and 5 % required treatment for osteoporosis without DXA scan. Of the 35 patients for whom FRAX calculation recommended BMD assessment by DXA scan, no patients were found to have previously had, or been referred for, this procedure. In addition, 8 patients were on pharmacological treatment for osteoporosis without a FRAX score or risk factors to support this intervention. Conclusions: Our results indicate that a large proportion of the in-patient displayed clinical risk factors of osteoporosis. We recommend that assessment of risk of fragility fracture a should be routinely performed in all medical admissions to improve the identification of osteoporosis at an earlier stage. P113 FUNCTIONAL TESTS AND QUALITY OF LIFE IN PATIENTS WITH VERTEBRAL FRACTURES Nataliia Grygorieva (1) presenting Vladislav Povoroznyuk (1) Helena Rybina (1) Institute of Gerontology AMS Ukraine, Kyiv, Ukraine (1) The purpose of the study was to examine the functional activity and quality of life in patients with vertebral fractures. We examined 153 women aged 60–89 years old in postmenopausal period (mean age 69.67±0.54 years). Patients were divided into two groups: the first (control group) – without osteoporotic fractures, and the second – with vertebral fractures. Methods of research - questionnaires (to assess life style, Euro-Qul-5D, Roland-Morris, ECOS-16), functional tests (dynamometry, static balancing, 15-meter test), Thomayer’s,

S602

Schober’s, Ott’s tests, orthopedic examination (range of movement assessment in the thoracic/lumbar spine, determination of the chest excursion and breath holding spell), dual- energy X-ray absorptiometry (DXA). Bone mineral density (BMD) of lumbar spine and femoral neck in patients with vertebral fractures was significantly lower than appropriate data in control group. Indexes of quality of life and daily activity in patients of the second group were considerably lower compared to the control group. It was found significant differences in Schober’s test (p00.04) and parameters of movement of the thoracic and lumbar spine (p00.04). Others functional tests were without significant difference. In patients without vertebral fractures it was found the significant positive correlation between BMD of the femoral neck and lumber spine and data of functional tests indexes such as dynamometry, Thomayer’s, Schober’s tests, maximum and average chest excursion. In contrast, patients with vertebral fractures didn’t have significant correlation between BMD data and indexes of functional tests and orthopedic examination data. In patients with vertebral fractures was determined significant correlation between ECOS-16 indexes and Schober’s tests (p 00,006) and breath holding spell (p 00,03) in contrast to patients without vertebral fractures. Our study found significant correlations between BMD and some functional tests in patients without vertebral fractures. Vertebral fractures leads to reducing of functional ability and decreasing of quality of life. P114 THROUGH THE CRACKS: INVESTIGATION AND MANAGEMENT OF OSTEOPOROTIC FRACTURES AT BARWON HEALTH Shannon McCarthy (1) presenting Barwon Health, Geelong, Australia (1) Introduction: An ortho-geriatrics registrar could improve the low rates of evaluation and management of osteoporotic fractures. Materials and Methods: Medical records of 227 consecutive patients with a discharge diagnosis of neck of femur fracture from Septemer 1 2008 to September 30 2009 were included. Traumatic fractures, fractures from greater than standing height, patients with incomplete records and incomplete coding were excluded, therefore 199 patients were included. Discussion: The average age was 83 (33–103). 74 % were female. 8 patients died in hospital. 76 [38 %] had a documented history of osteoporosis and/or low trauma fracture, yet only 35 of those [46 %] were on any osteoporosis treatment [calcium, vitamin D, bishosphonate, strontium,

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

raloxifene] and only 24 [32 %] on bone-modifying treatment [bisphosphonate, strontium, raloxifene]. Investigations slightly increased with physician review: serum calcium 89 % vs 45 %, vitamin D 66 % vs 31 %, TFT 52 % vs 53 %, PTH 17 % vs 13 %, SPEP 7 % vs 3 %.46 patients [44%of those tested] were diagnosed with vitamin D deficiency (<50 nmol/L); only 24 [52 %] of these were treated. 37 [19 %] patients had a history of a solid organ malignancy; operative histology was sent on 7 patients [4 %]. 122 [61 %] were reviewed by a general medical registrar and/or consultant, and were only slightly more likely to be investigated, but far more likely to have their osteoporosis treatment escalated [43 % vs 4 % patients]. 41 patients were discharged on triple therapy of calcium, vitamin D replacement if required, and a bone-modifying agent [21 %]; of these, 10 patients were receiving this treatment prior to fracture. 21 % of bone-modifying treatment-naïve patients had bonemodifying treatment introduced. 16 patients were discharged directly home [not including nursing home], and of these, only 1 was discharged on triple therapy, which was in place pre-fracture. 23 patients were discharged from inpatient rehabilitation on no medical therapy [22 % of patients undergoing rehabilitation]. 19 patients suffered further low-trauma fractures, in the subsequent 3 year period; 9 of these were still receiving no medical therapy for osteoporosis. Conclusion: rates of investigation and treatment of osteoporotic fractures remain low overall. Review by medical registrar slightly increases rates of investigation but greatly increases rates of treatment. P115 BONE LOSS DEPSITE IV ZOLEDRONATE PAUL RYAN (1) presenting Medway Maritime Hospital, Gillingham, UK (1) Introduction: Clinical trials suggests very few patients lose bone mass if properly treated with bisphosphonates and this has suggested monitoring of patients with serial BMD is unnecessary in clinical practice. IV Zoledronate therapy provides a good model tom investigate this notion as there is no concern regarding compliance. An impression of some patients losing significant bone mass despite therapy prompted the study. Methods: Patients who lost bone on serial BMD following first or subsequent IV Zoledronate therapy were carefully considered to assess reasons for treatment failure. Study entry required bone loss of more than 5 % lumbar spine or 8 % total hip between BMD measurements made at baseline and 1 year after therapy.16 patients, 1 male and 15 female of average age 73 years were identified over a 12 month period from a population of about 300 therapies. In 10 patients a

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

low BMI or marked loss of weight more than 5 kilos was felt a possible cause. 7 patients had a body weight<45 kilos at 1 year. 2 patients were on continuous corticosteroids and 1 was very immobile. Discussion: Although the vast majority of patients respond well to bisphosphonate treatment, a few patients do appear to keep losing bone and those with low body mass may be those most at risk. P116 A COMPARISON OF THE PRESCRIPTION OF SECONDARY PREVENTION OF OSTEOPOROSIS ON A GENERAL ELDERLY WARD COMPARED TO A WARD WITH SPECIALIST ORTHOGERIATRIC INPUT Toby Jennison (1) presenting Katrina Topp (1) Laura Hunt (1) Shivan Sivakumar (1) Leeds Teaching Hospitals, Leeds, UK (1) Aims: The secondary prevention of osteoporosis is a major target in geriatric medicine. Patients admitted to hospital with a fragility fracture should be assessed for the prescription of secondary osteoporosis prevention. Whether this occurs is generally down to junior doctors. In this audit we compared the prescription of secondary prevention of osteoporosis medications on a general geriatric ward, and with those on a specialised orthopaedic ward, with senior orthogeriatric input. Material and Methods: A prospective audit was completed over a 4 week period. Patients were included from 3 elderly wards, and one orthopaedic ward with orthogeriatic input. All patients greater than 75 years old, and admitted with a fragility fracture were included. Discussion: 16 medical and 13 orthogeriatric patients were found. The average age was 93 and 85 respectively. No patients in either group were known to have osteoporosis. All patients had routine bloods performed. 11 (69 %) patients on the medical ward had a calcium profile, and 9 (69 %) on the orthogeriatric ward. In the medical group 9 (56 %) patients were prescribed Calcium supplementation, compared to 9 (69 %) on the orthogeriatric ward, whilst only 3 (19 %) were prescribed bisphosphonates on the elderly ward, compared to 8 (62 %) on the orthogeriatric ward. Conclusions: The secondary prevention of osteoporosis is of the upmost importance for prevention of further fragility fractures. This study found, that the prescription of secondary prevention is considerably improved with the input of a senior orthogeriatrician when compared to a standard elderly ward. P117 VITAMIN D DEFICIENCY AND INSUFFICIENCy IN CHILDREN OF UKRAINE Vladyslav Povoroznyuk (1) presenting Nataliya Balatska (1) Oleksandra Tyazhka (1) Inga Kubey (1) Tetyana Budnyk (1) Institute of Gerontology AMS Ukraine, Kyiv, Ukraine (1)

S603

Objective(s): To determine the frequency of vitamin D deficiency and insufficiency in children of different region of Ukraine and to evaluate the influence of 25OHD and PTH level to bone mineral density (BMD). Material and Methods: It was examined 220 practically healthy children aged 10–18 yr. old who lived in different regions of Ukraine. 25-OH vitamin D and parathyroid hormone (PTH) were evaluated by electrochemiluminescence method (Elecsys 2010, Roche). Vitamin D deficiency was diagnosed in level of 25-OH vitamin D below 50.0 nmol/l, vitamin D insufficiency – between 74.5 and 50.0 nmol/l. BMD was determined by ultrasound densitometry Sahara (Hologic). Children were examined during October and November 2011 yr. Results: Vitamin D deficiency was registered in 92.2 % children, 45.9 % of schoolchildren had 25-OH vitamin D level below 25 nmol/l. 6.4 % examined patients had vitamin D insufficiency. 0.9 % of examined children had secondary hyperparathyroidism. Low mineral density was registered in 4.8 % children. No significant correlation between 25-OH vitamin D and BMD. Only in children 12–15 yrs. old with vitamin D deficiency it was significant correlation between PTH level and all data of ultrasound densitometry (stiffness index (r0−0.32, p<0.01), broadband ultrasound attenuation (r0−0.26, p<0.05), speed of sound (r0−0.31, p<0.02)). Conclusion(s): High level of vitamin D deficiency (92.2 %), secondary hyperparathyroidism (0.9 %), negative significant correlation between PTH level and data of ultrasound densitometry in pubertal children with vitamin deficiency make doctors to research the effective methods of treatment and prophylactics of revealed disorders. P118 MORPHOLOGICAL DISSECTION OF THE COLONY FORMING UNIT-FIBROBLASTIC POPULATION AND THE POTENTIAL FOR SKELETAL STEM CELL ENRICHMENT David Gothard (1) presenting Jonathan Dawson (1) Richard Oreffo (1) University of Southampton, Southampton, UK (1) Introduction: Skeletal stem cells (SSCs) constitute a selfrenewing and multipotential population within human bone marrow (HBM). SSCs hold great promise for bone tissue engineering however, conventional isolated populations lack the required homogeneity for clinical application. Current antibody-based enrichment generates heterogeneous populations with variable colony forming unit-fibroblastic (CFU-F) capacity (a stem cell pre-requisite). The present study investigated the potential of CFU-F dissection for SSC enrichment through categorisation of different colony morphologies. Material and Methods: Distinct cell populations were isolated from HBM using flow cytometry according to surface expression of Stro-1, CD146 (pericyte marker)

S604

and CD105 (endothelial marker) individually and in combination. CFU-F capacity within resultant populations was assessed and morphological differences within colonies categorised according to size, density and circularity. A two-step CFU-F assay was employed to assess SSC enrichment following isolation according to categorised colony morphologies. Discussion: Immunolabelled populations generated colonies exhibiting four morphology categories based on size and density including large and dense (LD), large and sparse (LS), small and dense (SD), and small and sparse (SS). Parent colonies isolated according to these parameters generated progeny with little difference in CFU-F capacity and thus were not predictive of cell growth potential. Further analysis depicted a positive correlation between parent colony circularity and progeny growth potential. However, parent circularity was not predictive of progeny CFU-F capacity. Conclusion: Characterisation of CFU-F colonies according to morphology provides an alternative non-invasive selection criteria to current antibody-based approaches for SSC enrichment. The present study highlights this potential but also the requirement for further refinement of selective morphological parameters. P119 AN AUDIT TO ASSESS THE VALUE OF DUAL HIP SCANNING Marion Barber (1) presenting Andrea Smith (1) Judith Cordingley (1) Lesley Hordon (1) Sharon Wilson (1) Mid Yorkshire NHS Trust, Dewsbury, West Yorkshire, UK (1) Background: Patients referred for DXA routinely have single hip and spine scanned. There are no national guidelines but the reasons given by various sites were time constraints and waiting list pressures. When the DXA Unit was set up in 2007, the scanner that was purchased by the Trust to support the DXA service had the facility to assist in positioning the patient for dual hip scanning. The decision was made to perform dual hip scanning with the intention to audit the results after a period of time and use the results to develop local guidelines. Aim: The aims of the study was to determine whether dual hip scanning at baseline is clinically beneficial and that should follow up scans concentrate on the hip with the lowest bone mineral density (BMD), to best utilise scan time and radiation dose. In the event of a fracture or hip replacement at one hip the remaining hip can be scanned to continue to provide a rate of change, from baseline, to aid clinical assessment. Objective: To maximise the clinical information provided while minimising scan time allowing a greater number of patients to be scanned during the working day. There are currently no Dose Reference Levels (DRL) for DXA but it is anticipated that these will be introduced in the

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

future. This work will be invaluable in ensuring that the DRL’s in the Trust are considered and consistent with best practice. Methodology: 100 scans were independently checked for positional comparability by two of the scan team. The neck of femur and total hip T-scores were recorded on an excel spreadsheet. Patients were identified on the DXA scanner and clinical history identified using RIS. Patients who had a single hip scanned or whose scans of dual hips demonstrated some potential positioning issues at baseline, were excluded from the audit. Results: There is a measurable difference between values at the hip and femoral neck in some patients and this can affect the diagnosis or treatment pathway for the individual patient. Conclusion: The proposal to the clinical team is that at baseline both hips are scanned. At follow up the hip with the lowest value is scanned unless surgery or injury would require the alternative hip to be used. P120 EFFECT OF STRONTIUM RANELATE ON VERTEBRAL PAIN SYNDROME AND FUNCTIONAL ABILITIES IN POSTMENOPAUSAL WOMEN WITH SYSTEMIC OSTEOPOROSIS Vladyslav Povoroznyuk (1) presenting Nataliia Dzerovych (1) L Bondarenco (1) V Verych (1) A Gnylorybov (1) H Hrytsenko (1) S Kosterin (1) O Kuhtei (1) D Recalov (1) O Synenkii (1) S Trubina (1) I Chizwikova (1) N Shpilevaya (1) E Jashina (1) Institute of Gerontology AMS Ukraine, Kyiv, Ukraine (1) Aim: To evaluate the effect of strontium ranelate in treatment of systemic osteoporosis in postmenopausal women. Materials and Methods: There were examined 894 postmenopausal women with systemic osteoporosis (average age 59,97±10,57 years, average height 161,82±7,09 cm, average weight 71,32±13,44 kg). Evaluation of pain syndrome and level of physical activity was carried out with visual analogue scale (VAS). Examination was performed before onset of treatment and after a four, eight and twelve month treatment course. Strontium ranelate (Bivalos, «Servier») was taken in a dose of one 2 g sachet as a suspension in water once a day and 1 tablet of Calcemin-advance (Calcium – 500 mg, Vit. D – 400 IU) 2 times a day during 12 months. Results: The patients had the risk factors of osteoporosis: 28 % of patients had osteoporotic fractures in their anamnesis; 17 % – hip fractures in mother or father of patients, 12 % – smoking, у 8 % – alcohol abuse, 27 % of patients have taken corticosteroid tablets for more than 3 month. We observed a reliable decrease of vertebral pain syndrome (after treatment – 2.97±0.77, after four months – 2.24±0.85, after eight months – 1.61±0.94; after twelve months – 1.24 ± 1.04; p < 0.00001) and increase of

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

functional abilities of patients (after treatment – 1.50± 0.67, after four months – 2.08±0.52, after eight months – 2.67 ± 0.53; after twelve months – 2.88 ± 0.63; p < 0.00001). Conclusion: It has been demonstrated that strontium ranelate treatment significantly decreases pronounced vertebral pain syndrome and improves functional abilities of patients in the postmenopausal women. P121 AUDIT OF BLOOD TESTS PERFORMED FOR LOW BONE DENSITY. THE EFFECT CLINICALLY AND FINANCIALLY OF CHANGING THE T-SCORE THRESHOLD FOR REQUESTING THEM Lesley Radford (1) presenting Haywood Hospital, Stokeon-Trent, UK (1) Introduction: Performing appropriate blood tests to exclude underlying metabolic bone disease or secondary causes in patients with low bone density is essential. Recommendations were used to develop a local protocol for requesting blood tests at the time of the bone density scan in patients with a T-score of<−2. The audit’s purpose was to identify if appropriate blood tests were being done, and the effect of lowering the threshold to a T-score of<−2.5. Patients suffering a confirmed low trauma fracture had blood tests done whatever their T-score. Material and Methods: 135 consecutive new routine scan patients were identified. Information collected included: low trauma fracture, referral source, requestor indicated blood tests already done, requested at the time of the scan, patient fulfilled the criteria at T-score<−2 / <−2.5, blood tests actually requested, and results within / outside the normal range. Discussion: 59 patients had a T-score of<−2. The dxa service requested 44 blood tests. The discrepancy was because 2 had been requested by the referrer, and 1 consultant referral (it is assumed that they have already done the blood tests unless they indicate otherwise on the referral). 10 referrers performed incomplete blood tests: 4 had not done TFT, 2 ESR, and 5 electrophoresis and immunoglobulins. 10 were not requested by the scan staff when they did fit the protocol: 4 referrers stated already done, 3 consultant referrals assumed done, 3 not done in error but recommended in all 3 scan reports. 46 had a T-score of<−2.5, leading to a potential 13 fewer requests, bringing a potential saving of £364.13 per month. £605.70 is spent on electrophoresis and immunoglobulins per month to detect 4 significant abnormal results. Conclusion: Requiring the lower T-Score to request blood tests has the potential to reduce a number of unnecessary blood tests without missing significant or unexpected abnormal results. There is a potential for cost savings. However, prior fragility fractures and other clinical factors must be taken in to account.

S605

P122 DEVELOPMENT OF A NEW TECHNIQUE FOR THE INVESTIGATION OF THE 3D SPATIAL DISTRIBUTION OF THE CALCIUM/PHOSPHORUS RATIO EFFECT IN OSTEOPOROSIS Andria Hadjipanteli (1) presenting Jie Huang (1) Paul Fromme (1) Nikolaos Kourkoumelis (2) Alessandro Olivo (1) Margaret Tzaphlidou (2) Robert Speller (1) University College London, London, UK (1) University of Ioannina, Ioannina, Greece (2) Introduction: In order to enhance the understanding, diagnosis and prevention of osteoporosis the chemical factors of bone which affect the load-bearing capability of bone have been studied. Bone quality is strongly related to the molecular groups and atomic elements that comprise the bone and to their spatial distribution within the bone mass. Previous studies have shown a relation between osteoporosis and a lowered and non-homogeneously distributed ratio of the two main constituents of bone, calcium (Ca) and phosphorus (P) especially in cortical regions and this project evaluates the significance of this distribution. Materials and Methods: The dual energy analysis technique (DEA), a form of subtraction x-ray imaging quantification that can be used for the non-invasive assessment of different material components within a substance, was developed for the quantification of the 3D spatial distribution of the Ca/P ratio in bone. The technique has been simulated to find the optimum conditions for the x-ray imaging of bone to achieve the most accurate results. Bone phantoms of known Ca/P ratio were synthesised for calibration purposes. Normal and inflammation mediated osteoporotic rabbit bones were imaged to investigate the relation of Ca/P ratio and its spatial distribution to osteoporosis. All studies were approved by the Ioannina University Institutional Animal Care and Use Committee. All images were acquired using a high resolution micro-computed tomography system (X-tek BenchTop). Discussion: Measurements of Ca/P ratio using synthesised bone have validated the newly developed technique and have allowed the Ca/P ratio distribution to be found in rabbit bone. Results on the measurements in rabbit bone show the relation of a lowered and non-homogeneously spread Ca/P ratio and osteoporosis. Conclusion: The assessment of 3D Ca/P ratio spatial distribution can be studied and could potentially be used clinically to enhance the diagnosis of osteoporosis. P123 MATERNAL DIET ALTERS FETAL BONE DEVELOPMENT S.A. Lanham (1) F.R. Lock (1) E.S. Lucas (1) A.J. Watkins (2) T.P. Fleming (1) R.O.C. Oreffo (1) S. Harding (1) Presenting University of Southampton, Southampton, UK (1) University of Nottingham, Nottingham, UK (2)

S606

Introduction: epidemiological studies suggest skeletal growth is programmed during intrauterine and early postnatal life. We used a mouse model to investigate different diet components and timing on the fetal bone development. Materials and Methods: dams received either a low (9 %) protein diet, high (30 %) protein diet, or high (22.5 %) fat diet for either the entire pregnancy or only for the first 3.5 days of gestation. Fetal samples were studied on day 17 of pregnancy by micro-computed tomography to assess bone development. Discussion: the low and high protein diets fed throughout pregnancy showed no alteration to bone development, however, feeding of a low or high protein diet for only the first 3.5 days of gestation produced accelerated bone development in the offspring in the limbs and vertebrae. In contrast, feeding a high fat diet for the first 3.5 days of gestation showed no alteration in fetal bone development, whereas feeding the high fat diet throughout pregnancy produced accelerated bone development in the offspring in the limbs and vertebrae. In addition, the proportions of low and high density bone within the whole fetal skeleton was altered by the diets. Conclusion: the data suggests a dietary checkpoint for bone development at around 3.5 days of gestation with regard to protein, and altering dietary protein levels around this time point accelerated bone development. In contrast, a high fat diet accelerated bone development only if maintained through pregnancy. P124 DOES ANTI-TNF THERAPY CAUSE A LESS SIGNIFICANT FALL IN BONE MINERAL DENSITY IN PATIENTS WITH RHEUMATOID ARTHRITIS? Myo Lynn (1) presenting Philip Sherrard (2) Lowel Ling (3) Katie Moss (4) Rheumatology Department, Princess Royal University Hospital, South London NHS Trust, UK (1) King's College Hospital, London, UK (2) St George's Hospital, London, UK (3) St George's Hospital, London, UK (4) Introduction: Since there is a better understanding of the immunological factors in pathogenesis of rheumatoid arthritis (RA), it has always been an interest to determine whether antitumour necrotic factor (anti-TNF) therapy causes a less significant fall in bone mineral density (BMD) in these patients. This study will investigate role of anti-TNF therapy in BMD compared to conventional disease modifying anti-rheumatic drugs (DMARDs) in RA patients. Material and Methods: 58 patients(9 male,48 female)with RA were identified retrospectively from database of St George’s Hospital, London, and divided into two groups: A and B. Group A received anti-TNF. Group B received DMARDs. Patients who received osteoporosis treatments including bisphosphonates were excluded. Changes in spine BMD and T-score between baseline and repeat

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

Dual Energy X-ray Absortiometry (DXA) scan were assessed. All data were presented as mean±S.D. Mean percentage changes of spine BMD and T score per year were calculated. Discussion: Mean age(yrs)were 56.57±11.15(A)vs.63.46± 11.33(B),P 00.023.Mean post-menopausal age(yrs)were 47.98 ± 2.49(A) vs.48.14 ± 1.37(B), P 00.942). Group A patients had higher disease activity (ESR 37±6.8 (A) vs. 16±3.2 (B), P00.007) and they needed more daily steroid dose (mg) (7.5±3.1 (A) vs. 6.1±2.8 (B), P00.013).Group A received repeat DXA scan earlier than B(months between scans 46.5 ± 17.00(A) vs. 77.76 ± 38.97(B), P 00.001). Group A have used more concomitant calcium and vitamin D (78 % vs.73 %).There was no significant difference of mean duration of disease (yrs)(11.32±5.56(A)vs.10.36± 5.56(B),P00.287). In group A, there was no significant reduction in spine BMD and T-score between two scans (BMD 0.90±0.15 vs.0.97±0.16, P00.353) (T-score −0.70± 1.50 vs.-0.81±1.60, P00.344). In group B, there was almost a significant reduction in spine BMD (1.03±0.20 vs. 1.00±0.17, P00.058) and significant reduction in T-score (−0.05±1.83 vs. -0.47±1.53,P00.013)between two scans. However, in terms of mean percentage change of spine BMD and Tscore per year, there were no significant differences between both groups (BMD −0.20±1.46 (A) vs. -0.40±1.37 (B), P00.593, T-score −4.20±17.59 (A) vs. -13.41±32.44 (B), P00.185). Conclusion: Effect of anti -TNF therapy on BMD might be beneficial if accompanied by bisphosphonates, however, anti-TNF alone may not deter falling in BMD in RA patients. Beneficial effect of anti-TNF may not be obvious because of small sample sizes. Age–matched groups, equal time interval between two scans and further studies with larger power are preferable. P125 SYSTEMATIC REVIEW OF TIP APEX DISTANCE (TAD) IN DYNAMIC HIP SCREW (DHS) FIXATION OF OSTEOPOROTIC HIP FRACTURES Imran Haruna Abdulkareem (1) presenting Leeds University Teaching Hospitals, Leeds, West Yorkshire, UK (1) Introduction: Femoral neck fractures are common problems seen in elderly osteoporotic patients, among which the extra-capsular ones, are treated with Dynamic Hip Screw (DHS). This (DHS) is based on tension band principle which allows the screw to slide within the barrel to enable compression of the fracture when the patient bears weight. This only works with intact medial wall and so cannot be used for reverse oblique fractures. It is important that the technique of screw placement is precise and should ideally be central in the femoral neck, on both AP and lateral radiographs, and

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

the Tip Apex Distance (TAD) is critical to the outcome of fixation and can accurately predict failure or survival of the screw. Methodology: A systematic review of articles published in PubMed, from 1991 to 2011, was carried out to analyse common practice with regards to DHS fixation of extra-capsular femoral fractures in relation to TAD. Search words include Tip Apex Distance, Dynamic Hip Screw, Sliding Hip Screw, Femoral neck fractures, Pertrochanteric fractures, Tension band principle, Fracture collapse, Screw cut-out, DHS failure, and Failure of fixation. Results: Forty eight articles were found relevant to the review. 2009 had the highest number of relevant publications (10), followed by 2010 (8), 2005 (7), and 2011 (5). Along the same line, Injury journal has the highest number of relevant published articles (9), followed by International Orthopaedics (6), JBJS Br (4), JBJS Am (3), and Orthopaedics (3). Conclusion and Recommendations: Although some new devices have theoretical and biomechanical advantages over the DHS, overall, they have not been found to be superior in terms of failure rate and functional outcomes, in the treatment of extra-capsular fractures, provided the principles of TAD and adequate fixation are adhered to. Therefore, DHS still remains the gold standard for the treatment of stable intertrochanteric fractures in suitable patients. However, intra-medullary devices are generally better for sub-trochanteric or very osteoporotic fractures. P126 PATIENTS ON STEROIDS -ARE THEIR BONES PROTECTED? DGH AUDIT Venkiah Kavuri (1) presenting Anthony Egboh (1) Thushani Wickramaratne (1) Epsom General Hospital, Epsom, Surrey, UK (1) Introduction: Steroids, widely prescribed are the commonest cause of drug-induced osteoporosis. Therefore attention to glucocorticoid-induced osteoporosis has substantially increased. Bone loss occurs rapidly in the first few months of glucocorticoid therapy and the risk of non-vertebral fractures increases by over 50 % in the first year of treatment irrespective of the dose. 1 % of the adult population in UK are on steroids, increasing to 2.4 %>70 years. Our prospective audit aimed to identify if bone protective therapy was given to patients new or on long term oral steroids. Material and Methods: All patients admitted to the medical and surgical wards above the age of 16+ from October to December 2010(three months) were analysed into gender, age, new steroid, maintained steroids, whether on bone protective therapy and the type of therapy. Exclusion criteria: short term steroids and parenteral single doses.

S607

Discussion: Total number patients on steroids: 29. Total on bone protective therapy: 52 12 (23 %) new to steroids. None on bone protective therapy.8 (67 %) males, 4 (33 %) females. 2 (16 %) <50 yrs, 10 (82 %) >50 yrs. Maintenance steroids (17): 6(35 %) males, 11 (65 %) females. 1 (6 %) <50 yrs 16(94 %) >50 yrs. 8 (47 %) not on bone protective therapy. 5 (29 %) only on calcium/Vit D 42 % on Bisphosphonates. None on strontium or Teripartide. 23(44 %) on bone protective therapy despite not on steroids. Conclusion: RCP guidelines advice that individuals at high risk e.g. those aged 65 years or over and those with a prior fragility fracture, should be advised to commence bone-protective therapy at the time of starting glucocorticoids. Measurement of bone density is not required before starting treatment. Our study concluded that bone protective therapy offered to patients new to steroids was none. 47 % on maintenance dose do not receive any bone protective therapy. However a significant proportion of patients were on bone protective therapy, despite not being on steroids. Education of the importance of bone protective therapy with the health care professionals and the public would be the way forward. We would strongly recommend a reaudit. P127 A NEW MODEL OF CONSULTANT LED FRACTURE LIAISON SERVICE ABHAYA GUPTA (1) presenting Srinivas Chenna (1) Alison Lorch (1)Glangwili Hospital, Carmarthen, UK (1) Introduction: Royal College of Physicians(RCP) have issued guidelines for osteoporosis management since 1999. RCP Audit and National Osteoporosis Society have recommended nationwide Fracture Liaison services(FLS)for secondary prevention of falls and fractures in UK. RCP 2007 Audit results locally showed only 12 % had osteoporosis assessment,8 % had DEXA scan, 44 % prescribed medications. We assessed the impact of new Fracture Liaison service started for assessment and management of osteoporosis patients at a district hospital in Carmarthen, Wales. Material and Methods: A new secondary care service was commenced from October 2010. Fragility fracture patients >50 years attending A&E are proactively screened and appropriate patients seen in Consultant led clinic supported by specialist nurse. Referrals are also received from GPs/clinicians/ward. Vertebral fractures are identified from hospital RADIS. Inpatient fragility fractures are assessed by Orthogeriatrician/specialist nurse. New treatment protocols and latest evidence based guidelines are being implemented.

S608

Discussion: From October 2010 upto December 2011, from 305 AE cards screened 123 were suitable for clinic. Including referrals, total0211. Seen in clinic0110. Awaiting(40) declined (47) Followed up(77). SpineRADIS reports screened (1469) advice letters to GPs sent for 156. Alendronate calcium and vitamin D prescribed as first line for osteoporosis. Other medications prescribed are Strontium, Risedronate, Aclasta(34) Denosumab(17) PTH (4). All had standardised assessment on new proforma, written leaflets provided, lifestyle and medication advice given. Conclusion: A new and complete FLS –previously nonexistent has been established within existing resources to close secondary fracture prevention management gap. Unlike other FLS, this is highest quality Consultant led, uses proactive case finding approach and fallers benefit from orthogeriatricians expertise on comorbidity, polypharmacy. A large inpatient and outpatient service is being provided locally by enthusiasm, innovation and strong leadership of Orthogeriatrician. P128 TRABECULAR BONE SCORE IN NORMAL UKRAINIAN WOMEN OF DIFFERENT AGE Vladyslav Povoroznyuk (1) presenting Nataliia Dzerovych (1) Alla Palamarchuk (1) Anna Musienko (1) Institute of Gerontology AMS Ukraine, Kyiv, Ukraine (1) The aim of this study is evaluating Trabecular Bone Score (TBS) in normal women of different age. Materials and Methods: We’ve examined 176 normal women aged 40–79 years (mean age – 53,4±0,6 yrs; mean height – 163,5±0,5 cm; mean weight – 80,4±1,1 kg). The patients were divided into the following age-dependent groups: 40–49 yrs (n053), 50–59 yrs (n089), 60–69 yrs (n017), 70–79 yrs (n017). TBS (L1-L4), total body, lumbar spine, femoral neck bone mineral density (BMD), lean and fat masses were measured by DXA using a densitometer Prodigy, GE. Results: We have determined the significant decrease of TBS (L1-L4) in women with age (40–49 yrs – 1,334± 0,016 mm-1; 50–59 yrs – 1,289±0,013 mm-1; 60–69 yrs – 1,194±0,034 mm-1; 70–79 yrs – 1,205±0,050 mm-1; F0 6,56; p00,0003). BMD of spine is significantly increase with age (BMD of spine: 40–49 yrs – 1,126±0,015 g/cm2; 50–59 yrs – 1,234 ± 0,013 g/cm2; 60–69 yrs – 1,343 ± 0,053 g/cm2; 70–79 yrs – 1,348±0,100 g/cm2; F04,04; p00,008). BMD of femoral neck didn’t show significant differences. The significant correlation was observed between TBS (L1-L4) and age, fat and lean masses: – – –

TBS01,64-0,007*Age; r0−0,34; t04,41; p00,00002. TBS01,47-0,000005*Total fat (g); r0−0,37; t04,86; p00,000003. TBS01,90-0,00001*Lean mass (g); r0−0,59; t08,98; p<0,000.

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

We did not find significant correlation between TBS and BMD of spine and femoral neck: – –

TBS01,36-0,05*BMD of spine; r0−0,05; t00,66; p00,5. TBS01,53-0,22*BMD of femoral neck; r0−0,16; t01,94; p00,05.

Conclusion: The significant correlation between TBS and lean mass indicates that bone quality can be associated with muscular system. TBS was significantly decreased with age. TBS is independent parameter which has potential diagnostic value without bone mineral density. P129 COMBINATION OF QUANTITATIVE ULTRASOUND AND FRAX® IN EVALUATION OF STRUCTURAL-FUNCTIONAL STATE OF BONE IN POSTMENOPAUSAL WOMEN Vladyslav Povoroznyuk (1) presenting Nataliia Grygorieva (1) Vasyl Povorozniuk (1) Institute of Gerontology AMS Ukraine, Kyiv, Ukraine (1) The aim of the study was to estimate the informative value of quantitative ultrasound and its combination with FRAX® in evaluation of structural-functional state of bone in Ukrainaian postmenopausal women. Material and Methods: 363 postmenopausal women aged 45–87 years were examined, average age 65,1 ± 0.5 years, duration of postmenopausal period 16,5 ± 0.5 years. Bone mineral density (BMD) was measured by Dual-energy X-ray absorptiometer (DXA) “Prodigy” and calcaneus quantitative ultrasound (QUS) “Sahara”. The ten years probability of hip fracture calculated with FRAX® tool. Results: There is difference in distribution of bone indexes in depending of used methods. Among women which had osteoporosis of femoral neck by DXA, 34 % had osteoporosis, 57 %−osteopenia, 9 %−norma data by QUS. Sensitivity of QUS indexes ranging was from low to moderate, but specificity was low (with femoral neck – 38 % and 39 %, total hip – 63 % and 34 %, lumbar spine – 45 % and 34 %, total body – 56 % and 34 % accordingly). Such sensitivity and specificity increased when combining QUS with the ten years probability of hip fracture without BMD (FRAX®) (with femoral neck – 81 % and 74 %, total hip – 78 % and 67 %, lumbar spine – 73 % and 63 %, total body – 81 % and 70 % accordingly). Conclusions: QUS of is informative method in evaluation of structural-functional state of bone in postmenopausal women. Sensitivity and specificity increased when combining QUS with FRAX® from 38 % and 34 % up to 81 % and 74 % accordingly.

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

P130 PSEUDO PSEUDO FRACTURE Sonia Panchal (1) presenting Yuthandar Aung (1) Peter Sheldon (1) University Hospitals of Leicester, Leicester, UK (1) Background: Pseudo fractures are radiological evidence of a thickened periosteum and new bone formation. Looser’s zone (LZ) are considered insufficiency fractures and are radiologic hallmark of osteomalacia, usually secondary to vitamin D deficiency. Case Report: A 70 year old Asian woman with a history of polymyalgia rheumatic presented with a seven week history of left hip pain and difficulty mobilising. She was on low dose prednisolone with Alendronic acid and Adcal D3 supplements for bone protection. Initial investigations included a left hip xray, which showed an active LZ in the superior margin of the femoral neck and an old pubic ramus fracture, a typical area for a LZ. Unfortunately the pain became severe within a month leading to a hospital admission. A repeat left hip xray showed progression of the LZ to a full intra-capsular neck of femur (NOF) fracture. It was assumed this was secondary to osteomalacia and she was treated with IM Ergocalciferol. A dynamic hip screw was inserted. However no vitamin D deficiency was detected (vitamin D 61, sample having been drawn prior to the injection). Bone profile was normal (Ca2+ 2.47, ALP 50 PO4 1.09, PTH 4.3). A DEXA scan confirmed osteoporosis (lumbar T score −2.5 Z score −1.0 femur T score −2.0, Z score −0.5). She had been maintained on bisphosphonates for five years. Discussion: True LZ occurs usually in the presence of severe osteomalacia with gross clinical and biochemical abnormalities. This case did not demonstrate biochemical features of osteomalacia. Many reports have highlighted the close resemblance of atypical insufficiency fractures with LZ. This confusion may be due to insufficiency fractures appearing similar to LZ or the delay in fracture healing may lead to widening of the fracture margins and replacement of the radiolucency with a sclerotic band. Therefore this may represent an insufficiency fracture secondary to prolonged bisphosphonate use although typically these are subtrochanteric femoral fractures. There have been several case series and case reports of these fractures. Approximately 400 cases have been registered in WHO records with the suspicion that bisphosphonate was the causative agent, usually Alendronate. P131 EFFECTIVENESS OF FRACTURE LIAISON SERVICES IN MIDDLESBROUGH Monica Clarke (1) presenting Henry Waters (1) Stephen Tuck (1) Middlesbrough Primary Care Trust, Middlesbrough, UK (1) James Cook University Hospital, Middlesbrough, UK (2)

S609

Introduction: The identification and treatment of patients sustaining fragility fractures is the key componenet of secondary prevention of osteoporosis. Previous audits in Middlesbrough have demonstrated that only a third of those over 65 years with hip fractures and 12 % of those with non-hip fractures were being given appropriate therapy. Despite the best efforts of the falls team and locally agreed pathways of care these figures did not improve. This provided the impetus for the the employment of orthogeriatricians to care for those over 65 years with hip fractures and the commission by all 3 local PCTs of a fracture liaison service (FLS). The aim of the FLS is to identify all those over 50 years sustaining fragility fractures except hip fractures over 65 years. Methods: audit of the activity of the first 9 months of the FLS. Discussion: 879 subjects were identified. In the same time period approximately 900 fractures would be expected suggesting successful identification. Some 40 % required scanning, 8 % did not attend their assessments and 20 % were from outside the area. Of those who attended 43 % required treatment. Conclusions: The FLS seems to have been successful in identifying those people over 50 years who have had fractures and improved the treatment rate. P132 INVESTIGATING LEVELS OF PRECISION IN GENE EXPRESSION MEASUREMENT BY DIGITAL PCR Rebecca Sanders (1,2), Carole Foy (1), Deborah Mason (2), Jim Huggett (1), LGC, Teddington, UK (1) Cardiff University, Cardiff, UK (2) Introduction: Gene expression studies profiling mRNA are central to biomolecular research offering considerable potential for research, diagnostics and prognostics. In musculoskeletal research, gene expression, measured by reverse transcriptase (RT) linked PCR, is commonly used to interpret the effects of specific signals both in vitro and in vivo, often on low copy transcripts extracted from tissues such as bone or cartilage. The RT step, necessary to convert mRNA to cDNA, is widely accepted as both inefficient and imprecise, with studies reporting variabilities up to 17-fold. This study tested RT measurement variability of synthetic RNA transcripts (ERCC developed targets; European RNA Control Consortium) using RT digital (d) PCR evaluated through one-step reactions; where RT-dPCR is performed in a one-step (one reaction vessel) process rather than the standard two-step approach (independent reactions in different tubes for the RT and then PCR stages). dPCR applies single molecule amplification achieved by sample partition for absolute quantification. While many published studies characterise dPCR capabilities for DNA quantification, little work has yet been done to determine similar parameter values for RNA measurement.

S610

Material and Methods: The measurement capability and reproducibility of RT to aid quantification of synthetic RNA targets spiked into human cell-line derived total RNA was assessed. Furthermore, to compare measurement accuracy of RT-dPCR with more established DNA dPCR, six well-characterised ERCC targets of known concentration were evaluated, enabling assay bias assessment. Discussion: Our results demonstrate different quantification values may be determined for each target despite evaluation of equal copy numbers. Variabilities between 19-56 % of the expected quantification value were measured, depending on RNA transcript, leading to investigation of the effect different factors, such as RT enzyme and target length, on the RT-dPCR result. Conclusion: RT reaction sensitivity is assay dependent, highlighting the need for further RT variability assessment in gene expression analysis. Furthermore, this study is one of the first to demonstrate application of RT-dPCR for absolute quantification of low copy RNA targets. This approach allows RT precision, sensitivity and linearity to be monitored alongside reaction efficiency, facilitating accurate interpretation of musculoskeletal gene expression data. P133 STUDY OF OSTEOPOROSIS IN GERIATRIC MALE POPULATION OF BIKANER CITY veer bahadur singh (1) presenting kusum singh (1) babu lal meena (1) vijay tundwal (1) s.p.medical college, Bikaner, India (1) Introduction: Osteoporosis is a metabolic condition that affects the bones, causing them to become thin and weak. About 300 million people in India have osteoporosis. Approximately three million people in the UK and there are over 230,000 fractures every year. Aims: To measure the Bone Mineral Density (BMD), determine the prevalence of osteoporosis in elderly men and correlate various risk factors with osteoporosis. Material & Methods: Subjects were randomly selected for this study after explaining the detailed objective of the study, informed consent was taken. A total 140 subjects age ≥60 years male were included in this study. Results: The prevalence of osteoporosis in study population was 14.3 %. Significant positive correlation of BMD with BMI (p<0.05) was seen. Significant negative correlation of BMD was seen with diabetes mellitus, smoking and androgen deprivation. However, there was insignificant negative correlation of BMD with age, alcohol intake and sedentary life style. Conclusion: There is high prevalence of osteoporosis in elderly male population. Risk factor for osteoporosis includes advancing age, hyperglycemia, smoking, low BMI, androgen deprivation, alcohol intake and lack of physical activity. However among these parameters age,

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

alcohol and physical activity had insignificant association with BMD. Key Words: Osteoporosis, Bone Mineral Density, Fractures P134 COMPOSITION AND STRUCTURAL HIERARCHIY IN BONE - DETERMINANTS OF MECHANICAL COMPETENCE Philipp Thurner (1) presenting Orestis Katsamenis (1) Orestis Andriotis (1) University of Southampton, Southampton, UK (1) The evaluation of fracture risk in osteoporotic patients, still mostly based on Bone Mineral Density (BMD) measurements, is a challenging task. During the past decades the research community has identified that not only bone mass, i.e. BMD, but also bone quality needs to be evaluated in order to achieve a reliable diagnosis of bone fracture risk for the individual. Bone quality includes among many other parameters the matrix material properties, which are dependent on the ultra- and microstructural arrangement of components that make up bone tissue. These components are non-stoichiometric carbonated apatite, collagen type I, water and nonollagenous proteins (NCPs). While the relative amount of NCPs is small they seem to accumulate in interfaces i.e. interlamellar areas and cement lines and densely populate fracture surfaces, which are mostly located within those areas. During past research the hypothesis had been formulated that besides their role in tissue organization NCPs could even “directly” influence bone matrix material properties, especially in the post-yield regime. A first validation of this proof was recently given through an investigation of osteopontin knockout mice, who display no changes in bone mass or geometry but a significant reduction in fracture toughness. To understand the origin of this reduction nanoscale investigations are necessary. However, given the limitations of size for performing mechanical experiments it currently seem highly challenging to investigate the post-yield and fracture behavior of bone at the smallest building block, i.e. the mineralized collagen fibril. For this reason we investigated bone fracture and post-yield behavior one hierarchical level above, i.e. at inter-lamellar areas (interfaces) and cement lines, which are enriched in NCPs. Atomic Force Microscopy combined with in situ micromechanical testing of bovine bone nanostructure provides the insight that cracks in cortical bone indeed are largely propagating through interfaces as cement lines and interlamellar areas. Micro-Raman microscopy further reveals a quantitative enrichment of these areas with NCPs and cantilever-based nanoindentation reveals a selective stiffening of these interfaces when loaded, which is not evident in lamellae. Overall, this points to previously unknown property of soft interfaces in bone and to the future potential for diagnostics and therapy currently not taken into account.

Osteoporos Int (2012) 23 (Suppl 5):S521–S611

P135 ORCHID–BONE: A NEW TOOL FOR BONE HEALTH CARE AND RESEARCH Tadeusz Jones (1), John Chelsom (2), Ian Gaywood (1), Ira Pande (1) Presenting, Rheumatology Department, Queen’s Medical Centre, Nottingham University Hospitals, Nottingham, UK (1), Centre For Health Informatics, City University, London, UK (2) Clinical records of patients with osteoporosis contain large amounts of information about bone health, risk factors, comorbidities and medications. Because data are not collected or stored in any standardised fashion they cannot easily be used for secondary purposes such as research, audit or provision of data to Hospital Episode Statistics (HES). HES data for 2009-10 show that only 3 % of 67.4 million outpatient episodes generated any diagnostic code. In 2010-11 there were only 4078 outpatient episodes recorded in England where the primary diagnosis was osteoporosis. More than 98 % of these episodes were recorded as “osteoporosis unspecified” (ICD 10-code M81.9). Better organisation of bone health data would create a more complete and accurate picture of bone diseases, its causes and social and healthcare consequences. ORCHID-Bone captures routine clinical data in a structured and standar-

S611

dised manner and organises it to maximise clinical and research potential. Data are organised in an indexed hierarchy linked to core data sets which capture key clinical characteristics to assist clinical decision making. Additional longitudinal data sets store information on sequential indices of bone health and treatment response such as bone turnover markers and BMD. Users do not need to understand the complexities of the data model as data collection, searching and analysis are carried out using simple, intuitive, customised web based interfaces. The osteoporosis module of ORCHID-Bone is in real time clinical use by the multidisciplinary osteoporosis team. Time-saving features include automatic downloading of demographic and laboratory data, pre-defined graphing facility for longitudinal data, automated production of clinic letters and relevant clinical alerts. Data analysis is enhanced by a bespoke search engine. ORCHID applications generate automatic cross mapping to existing coding systems (ICD and SNOMED). Patients offer positive feedback on its ability to provide explanation and education during consultations. ORCHID-Bone is implemented using open source software and is compliant with agreed standards for electronic health records. It will run on any platform and is scalable. It could therefore provide the infrastructure for a unified national bone health register.