Osteoporos Int (2014) 25:923–932 DOI 10.1007/s00198-013-2565-5

ORIGINAL ARTICLE

Prospective evaluation of renal function, serum vitamin D level, and risk of fall and fracture in community-dwelling elderly subjects D. Rothenbacher & J. Klenk & M. D. Denkinger & F. Herbolsheimer & T. Nikolaus & R. Peter & B. O. Boehm & K. Rapp & D. Dallmeier & W. Koenig & for the ActiFE Study Group

Received: 1 June 2013 / Accepted: 21 October 2013 / Published online: 13 November 2013 # International Osteoporosis Foundation and National Osteoporosis Foundation 2013

Abstract Summary This prospective study in elderly showed that kidney function plays a minor role in explaining the high prevalence of vitamin D deficiency seen in noninstitutionalized elderly subjects. However, 25-hydroxyvitamin D levels were clearly inversely associated with risk for first fall, which was especially seen in subjects with calcium levels above median.

Thorsten Nikolaus, the Director of the Agaplesion Bethesda Clinic Ulm, passed away last 26 September 2013. He will be warmly remembered and dearly missed by all. This work is dedicated to him. D. Rothenbacher (*) : J. Klenk : F. Herbolsheimer : R. Peter : D. Dallmeier Institute of Epidemiology and Medical Biometry, Ulm University, Helmholtzstr. 22, 89081 Ulm, Germany e-mail: [email protected] M. D. Denkinger : T. Nikolaus Agaplesion Bethesda Clinic Ulm, Ulm, Germany R. Peter Institute of the History, Philosophy and Ethics of Medicine, Ulm University, Ulm, Germany B. O. Boehm Division of Endocrinology and Diabetes, University of Ulm Medical Center, Ulm, Germany K. Rapp Clinic for Geriatric Rehabilitation, Robert-Bosch Hospital, Stuttgart, Germany D. Dallmeier : W. Koenig Department of Internal Medicine II-Cardiology, University of Ulm Medical Centre, Ulm, Germany

Introduction Few prospective studies in elderly exist that have investigated the association of renal dysfunction and vitamin D status on risk of falls. The aim of this study is to evaluate the association of renal function with 25hydroxyvitamin D (25-OH-D) levels and, secondly, to assess the role of both factors on the risk of falls and subsequent bone fractures. Methods This is a prospective population-based cohort study among noninstitutionalized elderly subjects during a 1-year follow-up. 25-OH-D levels and renal function were estimated, the latter by cystatin C-based equations. Information on falls was assessed prospectively. Results Overall, 1,385 subjects aged 65 and older were included in the study (mean age 75.6 years), of whom 9.2 % had a 25-OH-D serum level above 75 nmol/L (US units 30 ng/ mL); 41.4 %, between 50 and 75 nmol/L (US units 20 to 29 ng/mL, insufficiency); and 49.4 %, <50 nmol/L (US units <20 ng/mL, deficiency). We found no association of chronic kidney disease with risk of first fall. In contrast, 25-OH-D serum categories were clearly associated with risk of first fall and we found evidence of effect modification with calcium levels. In the group with a calcium level above the median (≥9.6 mg/dL), subjects with 25-OH-D serum level between 50 and 75 nmol/L and with concentrations <50 nmol/L had a hazard rate ratio (HRR) of 1.75 (1.03–2.87) and 1.93 (1.10– 3.37) for risk of first fall. 25-OH-D serum levels were also associated with several markers of inflammation and hemodynamic stress. Conclusions We demonstrated an association of 25-OH-D serum levels and risk of first fall, which was especially evident in subjects with serum calcium in upper normal, independent of renal function.

924

Keywords Elderly . Falls . Prospective cohort study . Renal function . Vitamin D

Introduction Falls in elderly subjects are common and pose a special threat as they may lead to muscular injury or bone fracture. They are among the second leading cause of unintentional injury death worldwide. Adults 65 years or older are at particular high risk of falls, and even on a global scale, falls are a major public health problem [1]. In addition, chronic kidney disease (CKD) emerges as a global public health problem and affects a large proportion of the adult population worldwide [2, 3]. CKD is an independent risk factor for cardiovascular diseases (CVD) as well as for allcause mortality [2]. Outcomes of CKD include not only progression to end-stage renal disease (ESRD) but also complications such as hypertension, malnutrition, anemia, and reduction of quality of life [4, 5]. Effects on bone health representing complications of CKD and renal osteodystrophy may develop as long-term sequela [6]. Vitamin D status and levels of other hormones such as parathyroid hormone (PTH), which are also associated with bone health, are also influenced by kidney function [7]. Not only both factors seem to be influenced by renal function but high PTH levels and low 25-hydroxyvitamin D (25-OH-D) were independently associated with an adverse profile of circulating markers of mineral and bone disorders in patients with CKD [7]. In addition, beside many known actions on auto- and paracrine functions on various tissues and the immune function, vitamin D status may have a direct effect on muscle strength [8]. So far, few prospective studies in elderly subjects have investigated the association of CKD and vitamin D status on risk of falls and subsequent bone fractures, information that may help to better design intervention strategies in this vulnerable group. The objective of the current study therefore is to, firstly, investigate the association of renal function with 25hydroxyvitamin D levels and, secondly, to prospectively assess the role of both factors on the risk of falls and bone fractures in a large population-based group of noninstitutionalized elderly subjects during a 1-year follow-up.

Osteoporos Int (2014) 25:923–932

the local registry office, a random sample of 8,613 noninstitutionalized inhabitants was contacted by mail and asked for participation. Exclusion criteria were severe deficits in cognition, vision, or hearing that precluded the accomplishment of most assessments or serious German language difficulties; institutionalized persons were also excluded. Between March 2009 and April 2010, 1,506 eligible individuals agreed to participate and underwent the baseline assessments. Further details are described elsewhere [9, 10]. All participants provided written informed consent. The study was approved by the ethical committee of Ulm University. Data collection

Methods

Participants satisfying the inclusion criteria were contacted by a field worker to make an individual appointment for an interview at home. Participants who did not want to be visited in their home were given the alternative to meet the interviewer in a designated room located at the Bethesda Geriatric Clinic, Ulm. In total, there were three visits incorporated in the ActiFE Ulm study, all to be completed within 7 days. The first and last visits were conducted by a study nurse, and the second visit was managed by a physician who carried out a physical exam. During the first visit, the interviewer obtained informed consent from the participants, provided information about the study procedure, and conducted the first half of the baseline interview. In the last visit, the interviewer accomplished the second half of the baseline interview. Briefly, the core questionnaires included questions on sociodemographic characteristics, diagnosis and related respiratory symptoms of asthma and COPD, physical functioning and activity, comorbidity (“has a doctor ever told you…”), exposures and potential risk factors related to asthma, COPD and physical activity, clinical management (treatment and self management issues), and accessibility and use of health services. During the third visit, handgrip strength, as a measure of the upper extremity function, was assessed using a JAMAR hydraulic hand dynamometer (Sammons Preston, Bolingbrook, IL). Participants performed two measurements for the left and right hands, respectively. The maximum of the mean value of each side was used for analyses. If the participants had separately agreed, blood was collected during the second visit. Participants who were successfully instructed on visit 1 also provided midstream urine specimen (morning void).

Study population

Laboratory methods

The Activity and Function in the Elderly in Ulm (ActiFE Ulm) study is a population-based cohort study in subjects aged 65 years and older who were randomly selected in Ulm and adjacent regions in Southern Germany. Based on the data from

Blood at baseline was drawn under standardized conditions. The 25-OH-D serum level was measured by an electrochemilumineszenz immunoassay (ECLIA) on a Roche E 2010 (interassay coefficient of variation (CV)

Osteoporos Int (2014) 25:923–932

4.96–5.43 %). This method was standardized against liquid chromatography–tandem mass spectrometry (LC-MS/MS), and the results obtained by LC-MS/MS again were compared to the National Institute of Standards and Technology (NIST) standard [11]. Serum creatinine (Cr) was measured by the kinetic Jaffe method (interassay CV 1.2–3.0 %) on an IMDS-traceable reference standard. Serum cystatin C (CysC) was measured by immunonephelometry on a Behring Nephelometer II (interassay CV 2.9–3.2 %). Creactive protein (CRP) was determined by a high-sensitivity assay on the same device (interassay CV 5.2–6.4 %). Serum calcium was measured photometrically on a cobas 6000 of Roche. Blood lipid measurements and other measurements were done by routine methods. All markers were measured in a blinded fashion. Assessment of chronic kidney disease Kidney function was assessed by means of estimated glomerular filtration rate (eGFR) according to a cystatin C-based eGFR according to the CKD-EPI collaboration [12]: CysCeGFR(CKD-EPI)=127.7×(CysC)−1.17 ×age−0.13 ×(0.91 if female)×(1.06 if black). CKD was defined as eGFR of less than 60 mL/min/1.73 m2 [13]. In equations, CysC is in milligrams per liter, age in years, weight in kilograms, and eGFR in milliliters per minute/ 1.73 m2. Stages of CKD followed the classification of the K/ DOQI guidelines: stage 1 was kidney damage with eGFR of ≥90 mL/min/1.73 m2, stage 2 was kidney damage with mildly decreased eGFR of 60–89 mL/min/1.73 m2, stage 3 was moderately decreased eGFR of 30–59 mL/min/1.73 m2, stages 4 and 5 (4/5) were eGFR of ≤29 mL/min/1.73 m2. Considering the few cases of severe stages of kidney disease (eGFR <15 mL/min/1.73 m2) in the present study, we combined stages 4 and 5. Renal damage was defined as an albumin–creatinine ratio (ACR) in spot urine sample (microalbuminuria: ACR 30 to <300 mg/g, macroalbuminuria: ACR ≥300 mg/g). Assessment of falls and fractures Information on falls was assessed prospectively by a falls calendar over a time period of 1 year which was handed over at baseline. The falls calendar was filled out by the participants on a weekly basis and sent in to the study centre every 3 months. In order to maximize the participants’ compliance, reminders were sent out every 6 weeks by the study centre. If the calendars were not returned within 2 weeks, the participants were contacted by phone. Bone fractures were also assessed by means of the same falls calendar. With each documented fall, it was also evaluated whether an injury occurred (contusion, bleeding, bone fracture).

925

Statistical analysis Descriptive statistics were calculated to describe the main characteristics of the study population. In addition, prevalence of CKD was calculated according to the CysC-eGFR estimating equation. Least-square means and 95 % confidence intervals adjusted for age and gender were calculated for the various CKD stages. Correlations between 25-OH-D serum levels, creatinine, cystatin C, CysC-eGFR, and various other biochemical markers were calculated overall and in categories with and without CKD by Spearman’s partial correlation coefficient after adjustment for age and gender. The relation of 25-OH-D serum levels, CKD, and first fall during follow-up was assessed by the Kaplan–Meier and life table methods and quantified by means of the log-rank test. The proportional hazards assumption was assessed graphically. The relationship between 25-OH-D, CKD, and risk of first fall during follow-up was quantified using the Cox proportional hazards model. Mainly factors which were described as potential risk factors in the literature were considered. Beside adjustment for age and sex, the following factors were included in the final model: BMI, seasonality, living alone, walking aid use, mini mental state examination <25, hypertension, history of myocardial infarction, heart failure, neurological disease, diabetes, and intake of vitamin D supplements. Hazard rate ratios (HRRs) and 95 % confidence intervals (CIs) were estimated in crude and adjusted analyses. As calcium levels are suggested to impact the risk of falls [14], and an interaction between vitamin D and calcium had been described [15], a potential interaction between kidney function, serum calcium level, and 25-OHD serum level was assessed by adding an interaction term of both variables with 25-OH-D into the model. All analyses were performed using SAS 9.2.

Results Overall, 1,385 subjects aged 65 and older with complete data on vitamin D (25-OH-D) serum levels and renal function measures were included in the final analysis (mean age 75.6 years, SD 6.53) (Table 1). More males were included than females (57.2 vs. 42.8 %). Most of the subjects were married (66.0 %) and had a school education of 9 years or less (58.1 %). The mean body mass index was 27.5 kg/m2, and 24.0 % were obese. More than half of the participants were hypertensive, 8.6 % had a previous myocardial infarction, 14.7 % reported heart failure, and 13.1 % had physiciandiagnosed diabetes. Only 2.8 % reported CKD. Overall, only 55 subjects (4.0 %) of the 1,385 participants were identified with supplemental vitamin D intake. Median CysC-eGFR was 82.4 mL/min/1.75 m2 (interquartile range (Q1–Q3) 67.9–95.8). The median 25-OH-D serum

926 Table 1 Characteristics of study population (N =1,385)

Osteoporos Int (2014) 25:923–932

Age

Gender, n (%) Family status, n (%)

Duration of school education, n (%)

Alcohol consumption, n (%) Smoking status, n (%)

Body mass index History of Co-morbidity, n (%)

Medication, n (%)

Mini mental state examination, n (%) Creatininea Cystatin Ca C-reactive proteina Renal function (eGFR) Parathyroid hormonea

a

Median (interquartile range, Q1–Q3); categories were as per Holick (details are in ref. [14])

b

25-OH-D serum categories in SI units are equal to the following categories in US units: ≥30 ng/mL (normal), 20 to 29 ng/mL (insufficiency), and <20 ng/mL (deficiency) c

Only the first fall of a participant during a 1-year observation period

a

Calcium 25-hydroxyvitamin D serum concentration*

Number of fallsc, n (%) Number of bone fractures, n (%) Median follow-up timea

level was 20.0 ng/mL (Q1–Q3 15.5–25.3). The 25-OH-D serum level was above ≥75 nmol/L in 9.2 % of subjects, between 50 and 75 nmol/L (insufficiency) in 41.4 %, and <50 nmol/L (deficiency) in 49.4 %. In general, women had lower 25-OH-D levels compared to men (normal 6.2 vs.

Years, mean (SD) 65–69 years, n (%) 70–79 years, n (%) ≥80 years, n (%) Male Female Married Widowed Other ≤9 years 10–11 years ≥12 years

75.6 (6.53) 329 (23.8) 642 (46.4) 414 (29.9) 792 (57.2) 593 (42.8) 911 (66.0) 334 (24.2) 135 (9.8) 792 (58.1) 316 (23.2) 256 (18.8)

Daily Smoker Former smoker Never smoker kg/m2, mean (SD) ≥30 kg/m2, n (%) Hypertension Myocardial infarct Heart failure Chronic kidney disease Diabetes Psycholeptics Psychoanaleptics Glucocorticoids Supplemental vitamin D intake <25 mg/dL mg/L

424 (31.2) 123 (8.9) 573 (41.5) 684 (49.6) 27.5 (4.12) 332 (24.0) 730 (52.7) 119 (8.6) 203 (14.7) 39 (2.8) 182 (13.1) 50 (3.6) 117 (8.5) 39 (2.8) 55 (4.0) 60 (4.7) 0.98 (0.87–1.14) 0.88 (0.76–1.02)

mg/L mL/min/1.75 m2* pmol/L 6.89 pmol/L (US >65 pg/mL), n (%) 1.59 to 6.89 pmol/L (15 to 65 pg/mL), n (%) <1.59 (<15 pg/mL), n (%) mg/dL nmol/L ≥75 nmol/Lb, n (%) 50 to 75 nmol/L (insufficiency), n (%) <50 nmol/L (deficiency), n (%)

1.70 (0.86–3.69) 82.4 (67.9–95.8) 3.70 (2.95–4.71) 98 (7.4) 1210 (91.5)

Days

15 (1.1) 9.6 (9.2–10.0) 49.9 (38.7–63.1) 127 (9.2) 574 (41.4) 684 (49.4) 446 (32.2) 44 (3.2) 370 (162–370)

11.5 %, insufficiency 38.1 vs. 43.9 %, deficiency 55.8 vs. 44.6 %; data not shown). During the 1-year follow-up (median 370 days), 446 subjects reported a fall. Only the first fall was counted during the observation period. Bone fractures occurred in 44 cases.

Osteoporos Int (2014) 25:923–932

927

Table 2 shows the association of renal function with 25OH-D serum levels. CKD was present in 13.3 % (stage 3) and 0.9 % (stage 4/5) of subjects. With increasing stage of CKD, the mean serum levels of 25-OH-D decreased and the differences were statistically significant among the various CKD stages. Among all categories, between 25.0 % (stage 4/5) and 47.2 % (stage 1) of subjects had a 25-OH-D serum level considered to be insufficient and between 42.0 % (stage 1) and 75.0 % (stage 4/5) of subjects had a level considered to be deficient (p =0.0012 for difference among groups). Table 3 shows the association of 25-OH-D serum levels with different renal function parameters and various biochemical markers and handgrip strength after adjustment for age and gender in the overall population and in subjects with and without CKD. We found a consistent and statistically significant inverse correlation with cystatin C, albumin–creatinine ratio, CRP, PTH, and glucose. Notably, we found no consistent correlation with lipids, a correlation with white blood cell count (WBCC) mainly in patients without CKD. By contrast, a positive correlation of 25-OH-D serum concentration was seen with CysC-eGFR and calcium and with handgrip strength mainly in the participants with CKD. The association with N-terminal-pro-B-natriuretic peptide (NTproBNP), highly sensitive troponin T (hsTnT) (both mainly seen in subjects without CKD), and fasting glucose was inverse and reached statistical significance in the total population. Figure 1 shows the Kaplan–Meier estimates of CKD (panel a) and 25-OH-D serum levels (panel b) with risk of first fall during follow-up. There was no statistically significant association seen with CKD (p =0.438). In contrast, we found a clear inverse association with 25-OH-D (p =0.009) levels. The multivariable association of renal function, 25-OH-D serum levels, and risk of first fall is presented in Table 4. As we found clear evidence of effect modification between calcium and 25-OH-D (p =0.03 in the adjusted model), results were stratified according to calcium levels. We found no association of CKD with risk of first fall, both in partly and fully adjusted models. Results for 25-OH-D levels were

similar in the stratum with calcium levels below the normal median value (≥9.6 mg/dL, which were 33.1 % of the study population; unfortunately, we could not achieve more balanced strata because of the very narrow range of calcium). By contrast, 25-OH-D serum categories were clearly associated with risk of first fall in the stratum with calcium levels above median (which were 66.9 % in the study population). Subjects with 25-OH-D serum level between 50 and 75 nmol/ L and with concentrations <50 nmol/L had a HRR of 1.75 (95 % CI 1.03–2.87) and 1.93 (95 % CI 1.10–3.37), respectively, when compared to the reference group in the fully adjusted model. We found no evidence of effect modification between CKD and 25-OH-D serum level, and the associations stayed almost identical when CKD and 25-OH-D serum level were considered simultaneously in the models. The fracture rate in the respective categories was 8 (95 % CI 0–45), 36 (95 % CI 22–56), and 35 (95 % CI 23–53) per 1,000 person-years in the participants with normal, insufficient, and deficient 25-OH-D serum levels, respectively. However, due to the small number of events, the 95 % CI of the rates is wide and the difference is not statistically significant (Table 5).

Discussion In this large prospective study involving 1,393 communitydwelling subjects aged 65 years and over who were randomly selected in the area of Ulm (Germany), more than 90 % of subjects had combined prevalence of 25-OH-D insufficiency and deficiency. The study showed an independent and dose– response association of 25-OH-D serum levels and risk of first fall, independent of renal function, which was especially evident in the large proportion of subjects with calcium levels in the upper normal. Although renal dysfunction was related to lower vitamin D serum levels in a statistically significant manner, it does not explain the large vitamin D deficit in this relatively healthy population and showed no risk for fall itself. Potential pathways of the vitamin D-associated risk of falls may

Table 2 Association between chronic kidney disease (CKD) and vitamin D serum levels CKD stage

1 2 3 4/5

n (%)

481 (34.7) 708 (51.1) 184 (13.3) 12 (0.9)

Adjusted meana (95 % CI)

21.1 (20.4–21.8) 20.1 (19.6–20.7) 19.6 (18.5–20.7) 15.5 (11.4–19.7)

Vitamin D serum levels Normal (≥75 nmol/L), n (%)

Insufficiency (50 to 75 nmol/L), n (%)

Deficiency (<50 nmol/L), n (%)

52 (10.8) 65 (9.2) 10 (5.4) 0

227 (47.2) 276 (39.0) 68 (37.0) 3 (25.0)

202 (42.0) 367 (51.8) 106 (57.6) 9 (75.0)

a

Least-square means and 95 % confidence intervals adjusted for age and gender

b

Fisher’s exact test

p =0.0012b

928

Osteoporos Int (2014) 25:923–932

Table 3 Association of 25-hydroxyvitamin D with different renal function parameters and various biochemical markers after adjustment for age and gender and after stratification according to chronic kidney disease (Spearman’s partial correlation coefficient (p value)) N =1,350

Creatinine (mg/dL)

Cystatine C (mg/L)

CysC-eGFR equation (mL/min/ 1.75 m2)

Albumin– creatinine ratio (urine)

CRP (mg/L)

Cholesterol (mmol/L)

HDL (mmol/L)

LDL (mmol/L)

25-OH-D (in all) 25-OH-D (in patients with CKD) 25-OH-D (in patients without CKD) eGFR (CysC)

−0.003 (0.919) −0.013 (0.856)

−0.094 (<0.001) −0.103 (0.166)

0.098 (<0.001) 0.102 (0.167)

−0.068 (0.012) −0.097 (0.191)

−0.142 (<0.001) −0.126 (0.088)

−0.022 (0.427) 0.032 (0.662)

0.060 (0.028) 0.051 (0.492)

0.047 (0.082) 0.075 (0.314)

0.041 (0.162)

−0.079 (0.007)

0.083 (0.005)

−0.058 (0.047)

−0.139 (<0.001)

−0.037 (0.202)

0.048 (0.103)

0.035 (0.232)

0.226 (<0.001)

0.055 (0.043)

−0.244 (<0.001)

N =1,350

WBC (giga/L)

25-OH-D (in all)

−0.076 (0.005) 0.045 (0.541) −0.089 (0.002) −0.098 (<0.001)

25-OH-D (in patients with CKD) 25-OH-D (in patients without CKD) eGFR (CysC)

Calcium (mg/dL)

0.098 (<0.001)

Glucosea (mg/dL)

PTH (pmol/L)

Handgrip strength (kg)

NTproBNP (pg/mL)

hs TNT (pg/mL)

0.093 (<0.001)

−0.300 (<0.001)

0.059 (0.030)

−0.060 (0.029)

−0.083 (0.002)

−0.112 (<0.001)

0.106 (0.153)

−0.421 (<0.001)

0.188 (0.011)

−0.024 (0.747)

−0.080 (0.278)

−0.233 (0.010)

0.093 (0.001)

−0.270 (<0.001)

0.031 (0.298)

−0.050 (0.089)

−0.071 (0.016)

−0.095 (0.008)

−0.046 (0.089)

−0.177 (<0.001)

0.075 (0.006)

−0.190 (<0.001)

−0.296 (<0.001)

−0.050 (0.132)

MDRD Modification of Diet in Renal Disease Study, CysC cystatin C, CRP C-reactive protein, HDL high-density lipoprotein, LDL low-density lipoprotein, WBC white blood cell count, PTH parathyroid hormone, NTproBNP N-terminal-pro-B-natriuretic peptide, hsTNT high-sensitivity troponin T a

Based on n =907 observations with fasting blood glucose measurements

include vitamin D-associated muscle weakness and hyperparathyroidism, resulting in impaired bone health. As a consequence of our study, vitamin D supplementation might pose a possible prevention target in order to decrease the fallassociated morbidity and mortality in elderly. Yet randomized controlled clinical trials would be needed to support this notion. According to a review by Holick [8] and the recommendations by the Endocrine Society [16] summarizing the results of several studies from the USA and Europe, between 40 and 100 % of elderly community-dwelling subjects are considered to have vitamin D deficiency, which is supported by our data, of whom 90 % had insufficient vitamin D levels based on reference values of the Endocrine Society. The main source of vitamin D in humans is exposure to sunlight and its production in the skin. Therefore, especially subjects who spend little time with outdoor activities might be at risk. In addition, old age is also associated with a decreased concentration of the precursor of vitamin D3 in the skin, resulting in a particular high risk of fall in the elderly [17]. Nutritional intake from foods or supplements represents another important source. Yet, in Germany, vitamin Dfortified foods are rather rare. Vitamin D is biologically inert and requires first hydroxylation in the liver into 25-OH-D (which is ideally measured in studies) and subsequently hydroxylated in the kidney into the active form of 1,25-(OH)2-D (calcitriol). CKD seems to be associated with a high incidence of nutritional vitamin D insufficiency or deficiency as manifested by decreased levels of 25-hydroxyvitamin D. This fact

and impaired 1-α-hydroxylation in the kidneys contribute to the inability to maintain the levels of 1,25-dihydroxyvitamin D. As the renal production of calcitriol is regulated by PTH, an increase in secretion of this hormone from the parathyroid gland results [17], leading to elevated PTH, abnormal calcium, and phosphorus levels. Especially subjects with CKD stage 4/5 and subjects requiring dialysis need regular vitamin D supplementation to maintain normal calcium levels and avoid secondary hyperparathyroidism [6]. The latter could result in an increased risk of renal bone disease. Although we observed an association between renal function and 25OH-D serum level in correlation analysis, the prevalence of advanced CKD in our study population was relatively low. Only 1.4 % had CKD stage 4/5 and it seems that factors other than CKD may play a more important role in this elderly population since almost all had low levels of 25-OH-D, even though the vast majority of subjects had no signs of kidney dysfunction. Falls are very common in the elderly, and it is estimated that 30 % of elderlies fall once within a given year, a number in line with our observations [18]. It is also known that several drug-related side effects are related to a higher risk of falls. [19]. We considered a relevant medication in our analysis. Psycholeptics (also include anxiolytics and sedatives) and psychoanaleptics (also include antidepressants) were not associated with the risks of falls in adjusted models in our study. Several recent meta-analyses show a clear association between low 25-OH-D serum levels and mortality [20] and also

Osteoporos Int (2014) 25:923–932

A 1.00 CKD non CKD

0.95

Proportion of event-free participants

Fig. 1 Kaplan–Meier estimates of risk of first fall associated with chronic kidney disease (CKD) (a) and serum levels of vitamin D (b)

929

0.90 0.85 0.80 0.75 0.70 0.65 Log-Rank p-value = 0.438

0.60

0

50

100

150

200

250

300

350

400

Follow-up [days]

B 1.00

Vitamin D ≥30 ng/mL Vitamin D 20-29 ng/mL Vitamin D <20 ng/mL

Proportion of event-free participants

0.95 0.90 0.85 0.80 0.75 0.70 0.65 Log-Rank p-value = 0.009

0.60

0

50

100

150

200

250

300

350

400

Follow-up [days]

falls [21]. Besides decreasing bone density, vitamin D deficiency also adversely affects muscle health resulting in muscle weakness, which is also one of the first clinical symptoms [22]. In addition, vitamin D receptors expressed in human muscle tissue promote de novo protein synthesis influencing muscle strength. There may be a role for vitamin D in cardiac muscle, as in a mouse model; vitamin D receptor knockout results in cardiac hypertrophy [23]. Notably, we found a statistically significant inverse association of 25-OH-D serum levels with various markers of inflammation like CRP or white blood cell count and hemodynamic stress like NTproBNP, or cardiomyocyte necrosis like hsTnT, which also seem to carry prognostic value in patients with stable CHD

[24]. Such data may point to other potential health consequences of low 25-OH-D serum levels and suggest the involvement of further pathogenetic pathways. These findings certainly should be explored in more detail. Heart failure had been independently associated with a 30 % risk increase in major fractures in a cohort study of adults aged 50 years and older from Canada [25]. As shown in an updated metaanalysis, vitamin D and calcium supplementation subsequently decreased the risk of falls [26], probably by increasing muscle strength and balance, in line with our observed positive association of 25-OH-D serum levels with handgrip strength, a direct marker of muscle strength. The fact that increased levels of inflammatory biomarkers return to normal

930

Osteoporos Int (2014) 25:923–932

Table 4 Association of chronic kidney disease and vitamin D serum levels with risk of fall according to strata of calcium level Model 1 HRR (95 % CI)a

n falls/N Calcium <9.6 mg/dL (33.1 %) CKD 114/389 (29.3 %) 21/69 (30.4 %) Vitamin Dc 9/35 (25.7 %) 40/167 (24.0 %) 86/256 (33.6 %) Calcium ≥9.6 mg/dL (66.9 %) CKD 267/800 (33.4 %) 44/127 (34.7 %) Vitamin D 20/92 (21.7 %) 133/407 (32.7 %) 158/428 (36.9 %) a

≥60 mL/min per 1.75 m2 <60 mL/min per 1.75 m2 Vitamin D ≥75 nmol/L Vitamin D 50 to 75 nmol/L Vitamin D <50 nmol/L

1.00 1.09 (0.67–1.77) 1.00 0.93 (0.45–1.92) 1.32 (0.66–2.64)

≥60 mL/min per 1.75 m2 <60 mL/min per 1.75 m2 Vitamin D ≥75 nmol/L Vitamin D 50 to 75 nmol/L Vitamin D <50 nmol/L

1.00 0.97 (0.69–1.37) 1.00 1.55 (0.97–2.49) 1.71 (1.07–2.74)

p trend

Model 2 HRR (95 % CI)b

p trend

0.098

1.00 1.20 (0.72–2.01) 1.00 0.72 (0.34–1.53) 0.90 (0.43–1.90)

0.662

0.041

1.00 1.03 (0.71–1.49) 1.00 1.75 (1.03–2.97) 1.93 (1.10–3.37)

0.05

Hazard rate ratio (HRR) and 95 % confidence interval adjusted for age and sex

b

Hazard rate ratio (HRR) and 95 % confidence interval adjusted for age, sex, BMI, seasonality, living alone, walking aid use, mini mental state examination <25, hypertension, myocardial infarction, heart failure, neurological disease, diabetes, and prescribed vitamin D supplementation

values after daily supplementation with vitamin D has been shown in a high-risk group of patients with diabetes [27]. Improved bone health with reduced risk of fracture might represent an additional benefit. As potential negative health consequences of low vitamin D status are well documented and the toxicity from supplementation seems small, a positive benefit–risk ratio is quite obvious. A pooled analysis including 31,022 participants aged 65 and older found that high-dose vitamin D supplementation of more than 800 IU daily was favorable in the prevention of nonvertebral and hip fractures [28], a dose which is also in line with The Institute of Medicine’s recent recommendation. Nevertheless, especially in elderly subjects, future intervention studies should explore the effect of controlled vitamin D intake on serum levels and other relevant surrogate biomarkers. Our study has several strengths and limitations. We were able to use a prospective design and therefore had a reliable record of falls and fractures over a relatively long time period. Still, the number of fall-associated fractures was too low to show a statistically significant difference between groups with

normal vitamin D levels, vitamin D insufficiency, and vitamin D deficiency. In addition, as the main source of vitamin production is exposure to sunlight, outdoor activity represents an important factor to be considered. Although we adjusted our results for seasonality and walking aid use, with the latter being a marker of disability, we could not adjust for time spent with outdoor activity and thus residual confounding may still exist. However, with more time spent outdoor, the risk of falls may also increase. Furthermore, we had no information on diet (except intake of vitamin D supplements which is relatively uncommon in Germany), and therefore, we were not able to consider other sources of vitamin D. We have no explanation why the risk was especially seen in subjects with serum calcium levels above the median. As the interquartile range of calcium was very narrow and two-thirds of the subjects were above median, it might only reflect a statistical and a precision issue. This is supported by the fact that the 95 % CI of the risk estimate of the below-median group (which was only one-third of the population) was wide and also overlapped with the risk estimate of the above-median group.

Table 5 Association of vitamin D with risk of fracture 25-OH-D serum concentration

Number Person-years Fracture ratea

≥75 nmol/L 1 50 to 75 nmol/L (insufficiency) 20 <50 nmol/L (deficiency) 23 a

122.6 548.2 648.2

8 (0–45) 36 (22–56) 35 (23–53)

Fracture rate per 1,000 person-years and 95 % confidence limits

Conclusion Our prospective study in community-dwelling elderlies with a relatively low prevalence of severe CKD showed that kidney function plays a minor role in explaining the high prevalence of vitamin D deficiency seen in noninstitutionalized elderly subjects. However, 25-OH-D serum levels were clearly

Osteoporos Int (2014) 25:923–932

inversely associated with risk for first fall; this was especially seen in the large proportion of subjects with calcium levels above median. Interventional measures seem necessary to increase vitamin D status and subsequently decrease the risk of fall, associated fracture, as well as major injury. Acknowledgments The study was funded by a grant from the Ministry of Science, Research and Arts, state of Baden-Wuerttemberg, Germany, as part of the Geriatric Competence Center, Ulm University and by funds from the Department of Internal Medicine II-Cardiology, Ulm University. The authors would like to thank Mrs. Gerlinde Trischler for the expert technical assistance. The funders had no role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript. Conflicts of interest None.

References 1. Moyer VA, U.S. Preventive Services Task Force (2012) Prevention of falls in community-dwelling adults: U.S. Preventive Services Task Force recommendation statement. Ann Intern Med 157:197–204 2. Chronik Kidney Disease Prognosis Consortium, Matsushita K, van der Velde M, Astor BC, Woodward M, Levey AS, de Jong PE, Coresh J, Gansevoort RT (2010) Association of estimated glomerular filtration rate and albuminuria with all-cause and cardiovascular mortality in general population cohorts: a collaborative metaanalysis. Lancet 375:2073–2081 3. Zhang Q, Rothenbacher D (2008) Prevalence of chronic kidney disease in population-based studies: systematic review. BMC Public Health 8:117 4. Astor BC, Matsushita K, Gansevoort RT, van der Velde M, Woodward M, Levey AS, Jong PE, Coresh J, Consortium Chronic Kidney Disease Prognosis, Astor BC, Matsushita K, Gansevoort RT, van der Velde M, Woodward M, Levey AS, de Jong PE, Coresh J, ElNahas M, Eckardt KU, Kasiske BL, Wright J, Appel L, Greene T, Levin A, Djurdjev O, Wheeler DC, Landray MJ, Townend JN, Emberson J, Clark LE, Macleod A, Marks A, Ali T, Fluck N, Prescott G, Smith DH, Weinstein JR, Johnson ES, Thorp ML, Wetzels JF, Blankestijn PJ, van Zuilen AD, Menon V, Sarnak M, Beck G, Kronenberg F, Kollerits B, Froissart M, Stengel B, Metzger M, Remuzzi G, Ruggenenti P, Perna A, Heerspink HJ, Brenner B, de Zeeuw D, Rossing P, Parving HH, Auguste P, Veldhuis K, Wang Y, Camarata L, Thomas B, Manley T (2011) Lower estimated glomerular filtration rate and higher albuminuria are associated with mortality and end-stage renal disease. A collaborative meta-analysis of kidney disease population cohorts. Kidney Int 79:1331–1340 5. Nitsch D, Mann AG, Bulpitt C, Roderick PJ, Fletcher A (2011) Impairment of kidney function and reduced quality-of-life in older people: a cross-sectional study. Age Ageing 40:381–387 6. Moe S, Drüeke T, Cunningham J, Goodman W, Martin K, Olgaard K, Ott S, Sprague S, Lameire N, Eknoyan G, Kidney Disease: Improving Global Outcomes (KDIGO) (2006) Definition, evaluation, and classification of renal osteodystrophy: a position statement from Kidney Disease: Improving Global Outcomes (KDIGO). Kidney Int 69:1945–1953 7. Ureña-Torres P, Metzger M, Haymann JP, Karras A, Boffa JJ, Flamant M, Vrtovsnik F, Gauci C, Froissart M, Houillier P, Stengel B, NephroTest Study Group (2011) Association of kidney function,

931 vitamin D deficiency, and circulating markers of mineral and bone disorders in CKD. Am J Kidney Dis 58:544–553 8. Holick MF (2007) Vitamin D deficiency. N Engl J Med 357:266–281 9. Denkinger MD, Franke S, Rapp K, Weinmayr G, Duran-Tauleria E, Nikolaus T, Peter R, ActiFE Ulm Study Group (2010) Accelerometer-based physical activity in a large observational cohort-study protocol and design of the activity and function of the elderly in Ulm (ActiFE Ulm) study. BMC Geriatr 10:50 10. Rothenbacher D, Klenk J, Denkinger M, Karakas M, Nikolaus T, Peter R, Koenig W (2012) Prevalence and determinants of chronic kidney disease in community-dwelling elderly by various estimating equations. BMC Public Health 12:343 11. Phinney KW (2008) Development of a standard reference material for vitamin D in serum. Am J Clin Nutr 88:511–512 12. Inker LA, Eckfeldt J, Levey AS, Leiendecker-Foster C, Rynders G, Manzi J, Waheed S, Coresh J (2011) Expressing the CKD-EPI (Chronic Kidney Disease Epidemiology Collaboration) cystatin C equations for estimating GFR with standardized serum cystatin C values. Am J Kidney Dis 58:682–684 13. Levey AS, Eckardt KU, Tsukamoto Y, Levin A, Coresh J, Rossert J, De Zeeuw D, Hostetter TH, Lameire N, Eknoyan G (2005) Definition and classification of chronic kidney disease: a position statement from Kidney Disease: Improving Global Outcomes (KDIGO). Kidney Int 2005(67):2089–2100 14. Pfeifer M, Begerow B, Minne HW, Suppan K, Fahrleitner-Pammer A, Dobnig H (2009) Effects of a long-term vitamin D and calcium supplementation on falls and parameters of muscle function in community-dwelling older individuals. Osteoporos Int 20:315–322 15. Lips P (2012) Interaction between vitamin D and calcium. Scand J Clin Lab Invest 72(Suppl 243):60–64 16. Binkley NC, Bischoff-Ferrari HA, Gordon CM, Hanley DA, Heaney RP, Murad MH, Weaver CM, Endocrine Society (2011) Evaluation, treatment, and prevention of vitamin D deficiency: an Endocrine Society clinical practice guideline. J Clin Endocrinol Metab 96: 1911–1930 17. Holick MF, Chen TC (2008) Vitamin D deficiency: a worldwide problem with health consequences. Am J Clin Nutr 87:1080S–1086S 18. Deandrea S, Lucenteforte E, Bravi F, Foschi R, La Vecchia C, Negri E (2010) Risk factors for falls in community-dwelling older people: a systematic review and meta-analysis. Epidemiology 21:658–668 19. Huang AR, Mallet L, Rochefort CM, Eguale T, Buckeridge DL, Tamblyn R (2012) Medication-related falls in elderly. Drugs Aging 29:359–376 20. Schöttker B, Ball D, Gellert C, Brenner H (2012 )Serum 25hydroxyvitamin D levels and overall mortality. A systematic review and meta-analysis of prospective cohort studies. Ageing Res Rev 12: 708–718. PubMed PMID: 22343489 21. Murad MH, Elamin KB, Abu Elnour NO, Elamin MB, Alkatib AA, Fatourechi MM, Almandoz JP, Mullan RJ, Lane MA, Liu H, Erwin PJ, Hensrud DD, Montori VM (2011) Clinical review: the effect of vitamin D on falls: a systematic review and meta-analysis. J Clin Endocrinol Metab 96:2997–3006 22. Bischoff-Ferrari H, Stähelin HB, Walter P (2011) Vitamin D effects on bone and muscle. Int J Vitam Nutr Res 81:264–272 23. Chen S, Law CS, Grigsby CL, Olsen K, Hong TT, Zhang Y, Yeghiazarians Y, Gardner DG (2011) Cardiomyocyte-specific deletion of the vitamin D receptor gene results in cardiac hypertrophy. Circulation 124(17):1838–1847, PubMed PMID: 21947295 24. Koenig W, Breitling LP, Hahmann H, Wüsten B, Brenner H, Rothenbacher D (2012) Cardiac Troponin T measured by a highsensitivity assay predicts recurrent cardiovascular events in stable coronary heart disease patients with 8-year follow-up. Clin Chem 58: 1215–1224

932 25. Majumdar SR, Ezekowitz JA, Lix LM, Leslie WD (2012) Heart failure is a clinically and densitometrically independent risk factors for osteoporotic fractures: population-based cohort study of 45,509 subjects. J Clin Endocrinol Metab 97:1179–1186 26. Chung M, Lee J, Terasawa T, Lau J, Trikalinos TA (2011) Vitamin D with or without calcium supplementation for prevention of cancer and fractures: an updated meta-analysis for the U.S. Preventive Services Task Force. Ann Intern Med 155:827–838 27. Neyestani TR, Nikooyeh B, Alavi-Majd H, Shariatzadeh N, Kalayi A, Tayebinejad N, Heravifard S, Salekzamani S, Zahedirad M (2012) Improvement of vitamin D status via daily intake of fortified yogurt drink either with or without extra calcium ameliorates systemic inflammatory biomarkers, including adipokines, in the subjects with type 2 diabetes. J Clin Endocrinol Metab 97:2005–2011 28. Bischoff-Ferrari HA, Willett WC, Orav EJ, Lips P, Meunier PJ, Lyons RA, Flicker L, Wark J, Jackson RD, Cauley JA, Meyer HE, Pfeifer M, Sanders KM, Stähelin HB, Theiler R, Dawson-Hughes B

Osteoporos Int (2014) 25:923–932 (2012) A pooled analysis of vitamin D dose requirements for fracture prevention. N Engl J Med 367:40–49

The ActiFE Ulm study group consists further of K. ScharffetterKochanek and H. Geiger, Department of Dermatology and Allergology; A. Lukas, Agaplesion Bethesda Clinic, Ulm; M. Riepe, Division of Gerontopsychiatry, Department of Psychiatry and Psychotherapy II; L. Rudolph, Max-Planck Group for Stem Cell Research; Ch. Schumann, Department of Internal Medicine II-Pneumology; J.M. Steinacker, Department of Sports and Rehabilitation Medicine; A. Ludolph and C. von Arnim, Department of Neurology; and G. Nagel and G. Weinmayr, Institute of Epidemiology and Medical Biometry. All Institutes are located at Ulm University.