Osteoporos Int DOI 10.1007/s00198-016-3836-8

ORIGINAL ARTICLE

Impact of body mass index on risk of acute kidney injury and mortality in elderly patients undergoing hip fracture surgery A. B. Pedersen 1 & H. Gammelager 1 & J. Kahlert 1 & H. T. Sørensen 1 & C. F. Christiansen 1

Received: 10 June 2016 / Accepted: 7 November 2016 # International Osteoporosis Foundation and National Osteoporosis Foundation 2016

Abstract Summary The literature is limited regarding risk factors for acute kidney injury (AKI) and mortality in hip fracture patients, although AKI is common in these patients. While obese patients were at increased risk of AKI, underweight patients with and without AKI had elevated mortality for up to 1 year after hip fracture surgery, compared with normal-weight patients. Introduction This study aimed to examine risk of postoperative AKI and subsequent mortality, by body mass index (BMI) level, in hip fracture surgery patients aged 65 and over. Methods A regional cohort study using medical databases was used. We included all patients who underwent surgery to repair a hip fracture during the years 2005–2011 (n = 13,529) at hospitals in Northern Denmark. We calculated cumulative risk of AKI by BMI level during 5 days postsurgery and subsequent short-term (6–30 days postsurgery) and long-term (31–365 days post-surgery) mortality. We calculated crude and adjusted hazard ratios (aHRs) for AKI and death with 95% confidence intervals (CIs), comparing underweight, overweight, and obese patients with normal-weight patients. Results Risks of AKI within five postoperative days were 11.9, 10.1, 12.5, and 17.9% for normal-weight, underweight, overweight, and obese patients, respectively. Among those

who developed AKI, short-term mortality was 14.1% for normal-weight patients compared to 23.1% for underweight (aHR 1.7 (95% CI 1.2–2.4)), 10.7% for overweight (aHR 0.9 (95% CI 0.6–1.1)), and 15.2% for obese (aHR 0.9 (95% CI 0.6–1.4)) patients. Long-term mortality was 24.5% for normal-weight, 43.8% for underweight (aHR 1.6 (95% CI 1.0– 2.6)), 20.5% for overweight (aHR 0.8 (95% CI 0.6–1.2)), and 21.4% for obese (aHR 1.1 (95% CI 0.7–1.8) AKI patients. Similar associations between BMI and mortality were observed among patients without postoperative AKI, although the absolute mortality risk estimates by BMI were considerably lower in patients without than in those with AKI. Conclusions Obese patients were at increased risk of AKI compared with normal-weight patients. Among patients with and without postoperative AKI, overweight and obesity were not associated with mortality. Compared to normal-weight patients, underweight patients had elevated mortality for up to 1 year after hip fracture surgery irrespective of the presence of AKI. The absolute mortality risks were higher in all BMI groups with the presence of AKI. Keywords Acute kidney injury . Body mass index . Cohort study . Hip fracture . Mortality

Introduction Electronic supplementary material The online version of this article (doi:10.1007/s00198-016-3836-8) contains supplementary material, which is available to authorized users. * A. B. Pedersen [email protected]

1

Department of Clinical Epidemiology, Aarhus University Hospital, Olof Palmes Alle 43-45, 8200 Aarhus N, Denmark

Hip fractures are a frequent cause of hospital admission among the elderly [1]. Thirty-day mortality among patients aged ≥65 years hospitalized with hip fracture is approximately 10% [2]. Presence of comorbid diseases at time of fracture and postoperative complications, including acute kidney injury (AKI), may explain the high mortality [3, 4]. We previously reported that 12.7% of hip fracture patients sustained postoperative AKI and that mortality was 2.8-fold

Osteoporos Int

higher in patients with AKI compared with patients without postoperative AKI [5]. Obesity is a frequent comorbidity, with a prevalence of more than 30% among US adults over age 65 [6]. Obesity has been associated with increased risk of postoperative AKI in patients undergoing cardiac surgery [7], as well as general orthopedic, colon, and thoracic surgery [8]. The proportion of hip fracture patients treated with total hip replacement who were obese increased three-fold from 1991 to 2008 [9]. Available data on the association between body mass index (BMI) and postoperative AKI among elderly hip fracture patients are limited [10]. In the general population, overweight and obesity have been associated with increased risk of a variety of diseases [11] and higher mortality [12, 13]. However, several studies have suggested a link between overweight and reduced risk of complications or in-hospital mortality among patients with chronic heart failure, peripheral arterial disease, thromboembolism, stroke, end-stage kidney disease, osteoporosis, rheumatoid arthritis, and critical illness (the Bobesity paradox^) [14–20]. No previous studies have examined the association between BMI and mortality among a large population of older hip fracture patients with postoperative AKI. A better understanding of the effects of both obesity and underweight on postoperative outcomes, including AKI and mortality, may permit timely intervention, improve quality of treatment, and reduce the healthcare burden. We therefore conducted a population-based cohort study to examine AKI risk in relation to BMI level among elderly patients following hip fracture surgery. In addition, we examined the association between BMI level and mortality among hip fracture patients with and without postoperative AKI.

surgery to repair a hip fracture. The DHFD is a nationwide population-based clinical quality database containing information on all patients with femoral neck, pertrochanteric, or subtrochanteric fractures surgically treated in Denmark since 2003. It is mandatory for all orthopedic departments in Denmark to provide data to the DHFD [2]. Preoperative and perioperative data are recorded on a standardized form by the operating surgeon and other members of the treatment team. Detailed definitions of data elements have been developed to ensure uniform registration of data across departments. Standard operative treatment involves either open reduction and internal fixation of the hip fracture or insertion of a primary total or partial hip replacement (please see Appendix for diagnostic and surgical procedure codes). We used the DHFD to identify hip fracture surgery patients undergoing surgery in Northern Denmark between January 1, 2005 and December 31, 2011. The study period was chosen based on the availability of laboratory data (described below) in Northern Denmark and to allow collection of information on drugs prescribed during the 100 days before hip fracture. Body mass index Information on body weight and height were obtained from the DHFD to calculate BMI (weight in kilograms (kg) divided by the square of height in meters (m)). Patients were categorized as underweight (BMI was <18.5 kg/m2), normal weight (BMI was ≤18.5–24.9 kg/m2), overweight (BMI was 25– 29.9 kg/m2), and obese (BMI was ≥30 kg/m2) [23]. For the approximately 15.6% of patients lacking information on BMI status, this information was imputed to allow their inclusion in the regression model [24]. Postoperative AKI

Methods Setting We conducted this cohort study in Northern Denmark (Central and Northern Denmark Regions), using prospectively collected data from regional population-based medical databases [21]. During the study period (2005–2011), the two regions had a population of 1.9 million persons, encompassing approximately one third of the Danish population. The Danish National Health Service provides tax-supported healthcare for all Danish residents, guaranteeing free medical care for emergency and general hospital admissions and for visits to hospital outpatient clinics and general practitioners [22]. Study population We used the Danish Hip Fracture Database (DHFD) [2] to identify patients aged 65 years and older who underwent

AKI was ascertained using data from the Laboratory Information Systems (LABKA) database of the Central and Northern Denmark Regions [25]. This database includes information on virtually all blood samples collected during hospitalizations, outpatient visits, and visits to general practitioners in the two regions, including patients’ personal identification number; name of the laboratory test performed; and nomenclature, properties and units in laboratory medicine (NPU) codes and/or local Danish laboratory codes [26], test results, the measurement unit, and the sample date. We used the LABKA database to obtain and compare each hip fracture patient’s baseline plasma creatinine level with his/her postoperative creatinine level (see Appendix 1 for NPU codes and local Danish laboratory codes.) The baseline creatinine level was defined as the median value of all available creatinine measurements taken during visits to an outpatient clinic or general practitioner in the period from 1 year to 7 days before the hospitalization for hip fracture [27]. Patients who

Osteoporos Int

developed AKI up to 7 days before hospitalization for hip fracture were excluded from the study. This ensured that postoperative AKI most likely occurred due to the hip fracture or ensuing surgery rather than preoperative causes such as dehydration, infection, or other underlying reasons for the fall resulting in hip fracture. For patients lacking a baseline creatinine measurement (31% of the study population of 13,529 patients), we estimated it based on the four-variable version of the Modification of Diet in Renal Disease (MDRD) Eq. [28], as suggested by Bellomo et al. [28]. AKI stage was defined according to the creatinine criteria in the Kidney Disease Improving Global Outcome (KDIGO) classification: AKI stage 1 (1.5 to 1.9 times the baseline creatinine level or an increase of ≥0.3 mg per deciliter (mg/dl) (≥ 26.5 μmol/l (μmol/l)) within 48 h), AKI stage 2 (2.0 to 2.9 times the baseline creatinine level), and AKI stage 3 (≥ 3.0 times the baseline creatinine level or an increase in serum creatinine to ≥4.0 mg/dl (≥ 354 μmol/l) or initiation of renal replacement therapy) [29]. Patients who did not fulfill these criteria were classified as not having AKI. We obtained data on postoperative creatinine measurements within the first five postoperative days, as AKI occurring during this time window is more likely related to hip fracture and surgery. If a patient had several measurements within the 5 days, we used the highest measurement. Mortality The Danish Civil Registration System provided information on deaths. This national registry has maintained records on vital status, migration, and residence for the entire Danish population since 1968, with daily updates [30]. We studied short-term mortality (6–30 days postsurgery) and subsequent long-term mortality (31–365 days postsurgery) in patients with and without AKI following surgery. Covariates We obtained information on gender and age (in two categories: 65–74 and ≥75 years) from the DHFD. The Danish National Patient Registry (DNPR) provided data on all hospital admissions since 1977 and on all hospital outpatient and emergency visits since 1995, including admission and discharge dates, outpatient visit dates, and up to 20 discharge diagnoses recorded according to the International Classification of Diseases (eighth revision until the end of 1993 and tenth revision thereafter). We used the DNPR to obtain a complete hospital history for all patients in the study cohort from 1977 until the hip fracture surgery date. As a measure of comorbidity, we computed a Charlson Comorbidity Index (CCI) score [31] for each patient at time of surgery. We defined three comorbidity levels: a score of 0 (low), given to patients with no previous record of conditions included in the CCI; a score of 1–2 (medium); and a

score of 3 or more (high) [32]. Diagnoses of kidney diseases were excluded from the CCI index score. Instead, we used the baseline creatinine value obtained from LABKA, as discussed above, to assess chronic kidney disease (CKD). CKD was defined as an estimated glomerular filtration rate (GFR) below 60 ml/min per 1.73 m2, using the four-variable MDRD Eq. [33]. Information on preoperative hemoglobin and albumin measurements were also collected from LABKA during the preceding year of the hip fracture surgery. For categorization of these measurements, see Appendix 2. We obtained information from the Danish National Health Service Prescription Database (DNHSPD) on medications dispensed during the 100 days preceding hip fracture surgery. The DNHSPD [34] has maintained information on all prescriptions for reimbursed drugs dispensed by community pharmacies in Denmark since 2004, recorded according to the Anatomical Therapeutic Chemical classification system (ATC codes). Additional variables in the DNHSPD include drug name, package identifier (permitting identification of brand, quantity, and drug formulation), date of refill, code identifying the prescribing physician, and code identifying the dispensing pharmacy. We considered prescriptions for the following drugs as possible confounders of the association between BMI and AKI and/or mortality: non-steroidal antiinflammatory drugs (NSAIDs), antihypertensive drugs, antibiotics, systemic glucocorticoids, antidepressants, statins, anticoagulants, and opioids (ATC codes are provided in Appendix 1.) We chose a 100-day period because these drugs are most commonly prescribed for 3 months’ use. Exclusions Of the 16,111 eligible hip fracture patients, we excluded patients lacking information on creatinine level during the first five postoperative days (n = 1749), patients receiving dialysis treatment due to chronic kidney disease or with a previous kidney transplant (n = 44), as well as 789 patients diagnosed with AKI up to 7 days before hospitalization for hip fracture. Thus, 13,529 hip fracture patients were included in our analyses. Statistical analysis We tabulated characteristics of the study population in relation to BMI level. Age and length of hospital stay were presented as median values in years or days with an interquartile range (IQR). We computed the cumulative incidence of AKI within 5 days of surgery with 95% confidence intervals (CIs) for hip fracture patients in each BMI group, using the cumulative incidence method and considering death as a competing risk [35]. Cumulative incidence was based on patients with available information on BMI. We used Cox proportional hazards regression to compute 5-day hazard ratios (HRs) for AKI and

Osteoporos Int

corresponding 95% CIs, both crude (not shown) and adjusted for gender, age, CCI level, CKD, and medication use before fracture. All patients were followed for 5 days in order to ascertain AKI or death. We calculated HRs for AKI overall and for AKI stages 1, 2, and 3 separately, comparing the four BMI groups. For hip fracture patients who developed AKI within five postoperative days and those who did not, we used the Kaplan-Meier method to compute the cumulative short-term mortality risk (time to death from 6 to 30 days after surgery, given that the patient was alive on day 5) and long-term mortality risk (time to death from 31 to 365 days after surgery, given that the patient was alive during the short-term period). We also used Cox proportional hazards regression to compute HRs for death and corresponding 95% CIs, both crude and adjusted, for gender, age, CCI level, CKD, and medication use before fracture, comparing the four BMI groups among patients who sustained AKI. In Cox proportional hazards regression analyses, we applied the multiple imputation method to handle missing BMI data, using all information available from the patients to generate five imputed datasets. All HRs (with both AKI and death as outcomes) were calculated as the geometric mean of the HRs of the five datasets, with corresponding confidence intervals corrected for between- and within-imputation variation [36]. All statistical analyses were performed using SAS version 9.2 (SAS Institute Inc., Cary, North Carolina).

Results Risk of postoperative AKI by BMI level Among the 13,529 hip fracture patients included in the analyses of postoperative AKI, 9.4% were underweight, 48.7% were normal-weight, 20.5% were overweight, and 5.8% were obese. Table 1 presents the baseline characteristics of the 13,529 patients according to BMI level, including the 2118 patients with missing BMI information. Compared with obese patients, underweight patients were slightly older, included fewer males, were less likely to have chronic kidney disease before surgery, included more patients with a moderate comorbidity level (scores 1–2) and fewer with a high comorbidity level (score ≥ 3), and had less medication use before surgery. Risks of AKI within 5 days of surgery were 10.1% (95% CI 8.5–11.8%) for underweight patients, 11.9% (95% CI 11.1– 12.7%) for normal-weight patients, 12.5% (95% CI 11.3– 13.7%) for overweight patients, and 17.9% (95% CI 15.3– 20.3%) for obese patients. The adjusted HRs for AKI (based on the multiple imputation method) were 0.9 (95% CI 0.8– 1.1) for underweight patients, 1.1 (95% CI 0.9–1.2) for overweight patients, and 1.6 (95% CI 1.3–2.0) for obese patients,

compared with normal-weight patients. We observed no clear differences in the risk of postoperative AKI stages 1, 2, and 3 between underweight and overweight patients compared to normal-weight patients. However, an increased risk of all three AKI stages was evident in obese compared with normal-weight patients (Table 2). BMI and mortality in patients with and without postoperative AKI Baseline characteristics of the 1717 hip fracture patients who developed AKI are presented according to BMI level in Table 3. Underweight AKI patients were slightly older, included fewer males, were less likely to have chronic kidney disease before surgery, and had less comorbidity and lower use of medication before surgery, compared with obese patients. Among the 1717 hip fracture patients who developed AKI, 205 (11.9%) died in the period between AKI development and 5 days postsurgery. These included 24 (18.7%) underweight patients with AKI, 92 (11.8%) normal-weight patients with AKI, 18 (5.2%) overweight patients with AKI, and 8 (5.7%) obese patients with AKI. A total of 64 (19.9%) of patients with missing BMI information died during this period. Cumulative short-term and long-term mortality in AKI patients by BMI level are presented in Fig. 1. Short-term mortality among AKI patients who survived five postoperative days was 23.1% for underweight patients, 14.1% for normal-weight patients, 10.7% for overweight patients, and 15.2% for obese patients. This corresponds to adjusted HRs (based on the multiple imputation method) of 1.6 (95% CI 1.0–2.6) for underweight AKI patients, 0.8 (95% CI 0.6– 1.2) for overweight AKI patients, and 1.1 (95% CI 0.7–1.8) for obese AKI patients prior to surgery, compared with normal-weight patients. Cumulative long-term mortality among hip fracture patients with AKI who survived 30 days postsurgery was 43.8% for underweight patients, 24.5% for normal-weight patients, 20.5% for overweight patients, and 21.4% for obese patients. This corresponds to adjusted HRs (based on the multiple imputation method) of 1.7 (95% CI 1.2–2.4) for underweight AKI patients, 0.9 (95% CI 0.6–1.1) for overweight AKI patients, and 0.9 (95% CI 0.6–1.4) for obese AKI patients, compared with normal-weight patients (Table 4). The same patterns of HR estimates were observed for associations between BMI groups and both short- and long-term mortality, calculated separately for patients with AKI stages 1, 2, and 3 (data not shown). HRs estimated for the risk of AKI and mortality based on the complete case method did not differ from HRs estimated based on the multiple imputation method, presented above. Baseline characteristics of the 11,812 hip fracture patients without AKI are presented according to BMI level in

2486 (37.7%) 1377 (20.9%)

1123 (40.6%) 536 (19.4%)

363 (46.4%) 175 (22.4%)

595 (76.1%) 179 (22.9%) 66 (8.4%) 182 (23.3%) 202 (25.8%)

b

a

Patients not treated with open reduction and internal fixation received a hip replacement

Estimated glomerular filtration rate (eGFR) <60 ml/min/1.73 m2

Underweight: BMI < 18.5 kg/m2 ; normal weight: BMI 18.5–24.9 kg/m2 ; overweight: BMI 25–29.9 kg/m2 ; and obese: BMI > =30 kg/m2

IQR interquartile range

440 (34.6%) 414 (32.5%)

Anticoagulants Opioids

1886 (68.1%) 668 (24.1%) 278 (10.0%) 617 (22.3%) 605 (21.8%)

146 (18.7%)

3995 (60.6%) 1628 (24.7%) 615 (9.3%) 1545 (23.5%) 1016 (15.4%)

422 (15.2%)

734 (57.7%) 370 (29.1%) 166 (13.1%) 312 (24.5%) 143 (11.2%)

Antihypertensive drugs Antibiotics Glucocorticoids Antidepressants Statins

553 (70.7%)

1880 (67.9%)

275 (35.2%) 314 (40.2%) 193 (24.7%)

272 (34.8%) 10 (7–14)

206 (26.3%)

923 (33.3%) 10 (6–12)

80 (74–86) 201 (25.7%)

480 (17.3%)

429 (33.7%) 568 (44.7%) 275 (21.6%)

Charlson Comorbidity Index (CCI) score Low (score 0) Medium (score 1–2) High (score ≥ 3)

1913 (29.0%) 9 (6–13)

82 (76–87) 938 (33.9%)

Obese patients (n = 782) (5.8%)

Diabetes 70 (5.5%) 651 (9.9%) Type of surgery 990 (77.8%) 4646 (70.5%) Open reduction and internal fixationb Redeemed prescriptions for selected medications within 100 days of hip fracture Non-steroidal anti-inflammatory drugs 154 (12.1%) 866 (13.1%)

308 (24.2%) 9 (5–13)

Chronic kidney diseasea Hospital stay, median (IQR), days

84 (78–89) 1817 (27.6%)

Overweight patients (n = 2769) (20.5%)

1014 (36.6%) 1180 (42.6%) 575 (20.8%)

84 (78–89) 168 (13.2%)

Age at time of surgery, median (IQR) Gender, male

Normal-weight patients (n = 6588) (48.7%)

2469 (37.5%) 2817 (42.8%) 1302 (19.8%)

Underweight patients (n = 1272) (9.4%)

797 (37.6%) 469 (22.1%)

1305 (61.6%) 557 (26.3%) 220 (10.4%) 555 (26.2%) 308 (14.5%)

298 (14.1%)

1579 (74.6%)

282 (13.3%)

728 (34.4%) 930 (43.9%) 460 (21.7%)

597 (28.2%) 7 (4–11)

84 (78–89) 627 (29.6%)

Missing data on BMI (n = 2118), (15.6%)

Baseline characteristics of 13,529 hip fracture patients aged 65 years and older according to body mass index (BMI) level, Denmark, 2005–2011

Patient characteristics

Table 1

5209 (38.5%) 2971 (22.0%)

8515 (62.9%) 3402 (25.1%) 1345 (9.9%) 3211 (23.7%) 2274 (16.8%)

1886 (13.9%)

9648 (71.3%)

1689 (12.5%)

4915 (36.3%) 5809 (42.9%) 2805 (20.7%)

4013 (29.7%) 9 (6–13)

83 (77–88) 3751 (27.7%)

Total (n = 13,529) (100%)

Osteoporos Int

Underweight: BMI < 18.5 kg/m2 ; normal weight: BMI 18.5–24.9 kg/m2 ; overweight: BMI 25–29.9 kg/m2 ; and obese: BMI > =30 kg/m2

HR denotes hazard ratio as a measure of risk ratio adjusted for age, gender, Charlson Comborbidity Index score, redeemed prescriptions for selected medications, and presence of chronic kidney disease before hip fracture, estimated using multiple imputation methods

b

a

1.3 (0.8–2.2) 2.8 (1.5–4.9) 1.0% (0.7–1.4) 1.9% (1.1–3.1) 1.1 (0.8–1.5) 1.9 (1.3–3.0) 2.5% (1.9–3.1) 4.2% (3.0–6.0) 2769 782 Overweight patients Obese patients

9.0% (7.9–10.1) 11.8% (9.6–14.1)

1.0 (0.9–1.2) 1.4 (1.1–1.8)

1.0 (reference) 1.0 (0.5–1.8) 0.8% (0.6–1.1) 0.8% (0.4–1.4) 1.0 (reference) 0.9 (0.5–1.4) 2.4% (2.0–2.8) 1.7% (1.1–2.6) 6588 1272 Normal-weight patients Underweight patients

8.7% (8.0–9.4) 7.5% (6.2–9.1)

1.0 (reference) 0.9 (0.8–1.2)

Adjusted HR (95% CI) Cumulative risk (95% CI) Adjusted HR (95% CI) Cumulative risk (95% CI) Cumulative risk (95% CI)

Adjusted HRb (95% CI)

AKI stage 3 AKI stage 2 AKI stage 1 Patients at risk, n BMIa

Table 2 Cumulative risk and adjusted hazard ratio (based on multiple imputation method) for acute kidney injury (AKI) stages 1, 2, and 3 within 5 days of surgery among 13,529 hip fracture patients, according to body mass index (BMI)

Osteoporos Int

Appendix 2. Cumulative short-term and long-term mortality in hip fracture patients who did not develop AKI by BMI level are presented in Fig. 2. The associations between BMI groups and mortality were similar in patients without and with postoperative AKI, but the absolute mortality risk estimates were considerably lower in non-AKI patients.

Discussion In this regional population-based cohort study of 13,529 hip fracture patients aged 65 years and older, obese patients were at increased risk of postoperative AKI compared with normalweight patients. While overweight and obesity in AKI and non-AKI patients were not associated with increased shortterm or long-term mortality, underweight patients had elevated mortality up to 1 year following hip fracture surgery compared to patients with normal weight irrespective of the presence of AKI. However, the absolute mortality risks were higher in all BMI groups with the presence of AKI. Comparison with previous literature This is the first population-based cohort study to examine the association between BMI and risk of postoperative AKI and subsequent mortality among elderly hip fracture patients. One previous study of 165 hip fracture patients over age 65 years who underwent surgery at a single department in Turkey [10] found no association between BMI and postoperative AKI, unlike our study. However, the Turkish study did not clarify how BMI was measured and categorized. In addition, the effect of BMI on AKI was examined only in regression analyses with no adjustment for possible confounding, hampering comparison with our study. Our findings are in line with results of several studies on other types of surgical patients. In a case-control study of 514 patients with AKI and 694 matched controls without AKI who underwent general orthopedic, colon, and thoracic surgery, BMI over 35 kg/m2 was found to be associated with increased risk of AKI (adjusted odds ratio (aOR) = 1.65, 95% CI 1.16– 2.32)) within 30 days of admission [8]. The study did not provide a separate estimate for orthopedic patients. Jafari et al. reported that among 17,938 hip and knee patients receiving total joint arthroplasties at a single institution in USA, BMI over 35 kg/m2 was associated with increased risk of AKI (aOR = 1.08, 95% CI 1.03–1.12) [37]. In addition, Billings et al. reported that among 445 patients undergoing cardiac surgery [7], BMI was associated with increased risk of AKI within 30 days of surgery (aOR = 1.26 (95% CI 1.04– 1.54) per 5 kg/m2). Thus, our findings support the increasing evidence of an association between obesity and postoperative AKI. This may be explained by several mechanisms. Increased systemic

25 (19.5%) 37 (28.9%) 9 (5–13) 51 (39.8%)

Gender, male Chronic kidney diseasea Hospital stay, median (IQR), days

Charlson Comorbidity Index (CCI) score Low (score 0)

334 (42.7%) 136 (17.4%)

145 (42.0%) 60 (17.4%)

261 (75.7%) 89 (25.8%) 39 (11.3%) 75 (21.7%) 73 (21.2%)

62 (44.3%) 34 (24.3%)

115 (82.1%) 28 (20.0%) 10 (7.1%) 36 (25.7%) 39 (27.9%)

19 (13.6%)

87 (62.1%)

55 (39.3%) 41 (29.3%) 44 (31.4%)

44 (31.4%)

35 (25.0%) 67 (47.9%) 11 (7–16)

83 (77–87)

Obese patients with AKI (n = 140)

b

a

Patients not treated with open reduction and internal fixation received a hip replacement

Glomerular filtration rate (GFR) <60 ml/min/1.73m2

IQR interquartile range

Underweight: BMI < 18.5 kg/m2 ; normal weight: BMI 18.5–24.9 kg/m2 ; overweight: BMI 25–29.9 kg/m2 ; and obese: BMI > =30 kg/m2

46 (35.9%) 29 (22.7%)

Anticoagulants Opioids

571 (73.0%) 189 (24.2%) 78 (10.0%) 185 (23.7%) 115 (14.7%)

44 (12.8%)

88 (68.8%) 40 (31.3%) 14 (10.9%) 31 (24.2%) 13 (10.2%)

Antihypertensive drugs Antibiotics Glucocorticoids Antidepressants Statins

217 (62.9%)

139 (40.3%) 76 (22.0%) 64 (18.6%)

130 (40.3%)

125 (36.2%) 154 (44.6%) 10 (6–14)

85 (80–90)

Overweight patients with AKI (n = 345)

501 (64.1%)

92 (71.9%)

338 (43.2%) 186 (23.8%) 91 (11.6%)

258 (33.0%)

243 (31.1%) 287 (36.7%) 10 (5–14)

87 (82–91)

Normal weight patients with AKI (n = 782)

Redeemed prescriptions for selected medications within 100 days of hip fracture Non-steroidal anti-inflammatory drugs 11 (8.6%) 98 (12.5%)

Type of surgery Open reduction and internal fixationb

49 (38.3%) 28 (21.9%) 9 (7.0%)

88 (83–92)

Age at time of surgery, median (IQR)

Medium (score 1–2) High (score ≥ 3) Diabetes

Underweight patients with AKI (n = 128)

139 (43.2%) 67 (20.8%)

234 (72.7%) 87 (27.0%) 39 (12.1%) 81 (25.2%) 40 (12.4%)

40 (12.4%)

220 (68.3%)

140 (43.5%) 84 (26.1%) 55 (17.1%)

98 (30.4%)

110 (34.2%) 106 (32.9%) 7 (3–12)

87 (82–91)

Patients with AKI and missing data on BMI (n = 322)

Characteristics of 1717 hip fracture patients aged 65 years and older who developed acute kidney injury (AKI) within 5 days of surgery, according to body mass index (BMI)

Patient characteristics

Table 3

726 (42.3%) 326 (19.0%)

1269 (73.9%) 433 (25.2%) 180 (10.5%) 408 (23.8%) 280 (16.3%)

212 (12.3%)

1117 (65.1%)

721 (42.0%) 415 (24.2%) 263 (15.3%)

581 (33.9%)

538 (31.3%) 651 (37.9%) 9 (5–14)

84 (78–88)

Total (N = 1717)

Osteoporos Int

Osteoporos Int Fig. 1 Cumulative 1-year mortality in patients with acute kidney injury by body mass index (BMI). (underweight: BMI <18.5 kg/m2; normal weight: BMI ≤18.5–24.9 kg/m2; overweight: BMI 25–29.9 kg/m2; and obese: BMI > =30 kg/m2)

inflammation and oxidative stress are more characteristic of overweight and obese than non-obese patients [7]. It has been suggested that increased levels of a marker of oxidative stress (plasma concentration of F2-isoprostanes) predict clinical AKI-causing renal hypoperfusion, decreased creatinine clearance, and renal damage [19, 38]. Hypotension during surgery may cause pre-renal azotemia characterized by reduced renal perfusion, leading to a decrease in the GFR and an increase in serum creatinine [39]. Volume resuscitation and maintenance of renal perfusion [40] might be more challenging to achieve in obese than in non-obese patients, since fluid status is more difficult to monitor in obese patients [41]. Because of the redundant fatty tissue in obese patients, signs of fluid overload Table 4 Crude and adjusted hazard ratios (HRs) for short-term (6–30 days following surgery) and long-term (31–365 days following surgery) mortality in 1717 hip fracture patients who developed acute kidney injury within 5 days of surgery, according to body mass index (BMI)

such as peripheral edema may be hard to recognize. Further, we observed a strong association between obesity and diabetes, which could also explain why obese patients are at increased risk of AKI since it is well-known that diabetes is a risk factor for AKI [42]. Finally, obese patients are at higher risk of developing surgical complications during hospitalization, including bleeding, infection, anemia, coagulopathy, need for mechanical ventilation, and thromboembolic complications [43], which can trigger AKI [8, 44]. Previous studies among healthy adults from the general population [12, 13] reported that obesity increases mortality risk. In contrast, studies among patients with heart failure, rheumatoid arthritis, and critical illness reported that obesity

Patients at risk, n

Short-term mortalitya Normal-weight patients Underweight patients Overweight patients Obese patients Long-term mortalityb Normal-weight patients Underweight patients Overweight patients Obese patients

Cumulative mortality, n (%)

Multiple imputation method Crude HR (95% CI)

Adjusted HR (95% CI)c

690 104 327 132

97 (14.1%) 24 (23.1%) 35 (10.7%) 20 (15.2%)

1.0 (reference) 1.7 (1.0–2.7) 0.8 (0.6–1.1) 0.9 (0.6–1.4)

1.0 (reference) 1.6 (1.1–2.2) 0.8 (0.6–1.2) 0.9 (0.6–1.5)

593 80 292 112

145 (24.5%) 35 (43.8%) 60 (20.5%) 24 (21.4%)

1.0 (reference) 1.9 (1.3–2.6) 0.8 (0.6–1.1) 0.8 (0.5–1.2)

1.0 (reference) 1.7 (1.3–2.3) 0.8 (0.6–1.0) 0.8 (0.6–1.3)

Underweight: BMI < 18.5 kg/m2 ; normal weight: BMI 18.5–24.9 kg/m2 ; overweight: BMI 25–29.9 kg/m2 ; and obese: BMI > =30 kg/m2 a

Only patients who survived 5 days following surgery were included (n = 1511)

b

Only patients who survived 30 days following surgery were included (n = 1271)

c

HR denotes hazard ratio as a measure of risk ratio adjusted for age, gender, Charlson Comborbidity Index score, redeemed prescriptions for selected medications, and presence of chronic kidney disease before hip fracture, estimated using multiple imputation methods

Osteoporos Int Fig. 2 Cumulative 1-year mortality in patients without acute kidney injury by body mass index (BMI) (underweight: BMI <18.5 kg/m2; normal weight: BMI ≤18.5–24.9 kg/m2; overweight: BMI 25–29.9 kg/m2; and obese: BMI > =30 kg/m2)

reduces mortality [18–20], suggesting the existence of an Bobesity paradox^ in these patient groups. The mechanism underlying this paradox is not clear, but selection bias has been suggested as a possible explanation. In contrast with findings of studies performed in other settings, we found that obesity in hip fracture patients with AKI or without AKI neither increased nor decreased their short-term or long-term mortality. Obesity has been recognized as a risk factor for a variety of diseases [11, 45]. Due to the potential risk of fluid overload, obese patients are more susceptible than are non-obese patients to cardiovascular and pulmonary complications, which are the main causes of death among hip fracture patients [46, 47]. One reason why obese hip fracture patients with or without AKI have a similar or even better outcome than normalweight hip fracture patients with or without AKI could be that high BMI reduces mortality through preservation of fat-free mass [48]. It is possible, however, that our findings reflect a Bhealthy survivor^ effect, i.e., persons who are most susceptible to obesity-related diseases and complications die before sustaining hip fracture and AKI. Thus, we may have included only Bhealthy^ obese patients. In underweight patients, loss of both peripheral and respiratory muscles followed by elevated risk of respiratory disease-specific mortality [49], as well as elevated risk of infections due to reduced immune response [50], has been observed. This may explain increased mortality in underweight compared to normal-weight hip fracture patients with AKI. Underweight also may be a marker of underlying chronic diseases [51], suggesting that unmeasured confounding could play a role in explaining higher mortality in underweight patients.

Limitations Although registration in the DHFD is mandatory for all hospital departments treating hip fracture, we may not have completely captured all hip fracture patients. However, extensive efforts are made to ensure completeness of DHFD data [52], including hospital reimbursement only after registration of a surgical procedure. Any failure to register hip fracture patients in the DHFD is probably independent of BMI since prospective registration of data would result in underestimation of the relationship between BMI and AKI/death. Another concern is our exclusion of patients lacking a postoperative creatinine measurement. This might have introduced selection bias if those with missing information were more likely to be underweight patients who developed postoperative AKI or obese patients who died postoperatively. We found that the comorbidity level of patients excluded due to a missing postoperative creatinine measurement was higher than that for included patients (data not shown). This was clearly a source of selection bias. Patients who developed AKI up to 7 days before hospitalization for hip fracture were excluded from the study (n = 789). Accordingly, we would expect the percentage of patients with chronic kidney disease to be slightly lower than expected. We cannot entirely exclude selection bias, meaning that some patients with chronic kidney disease did not live long enough to sustain hip fracture. We lacked information on the severity of some medical conditions included in the CCI score (diabetes and chronic lung disease), which may have introduced residual confounding into our analyses. Lack of data on lifestyle factors such as smoking, physical activity, and alcohol consumption, and other unmeasured conditions, which are closely linked to BMI status, could also cause unmeasured confounding leading to bias of the

Osteoporos Int

association between underweight and AKI or between obesity and mortality [53]. Although we have adjusted for a number of potential confounders, we cannot exclude residual or unmeasured confounding. The clinical impact of BMI can only be examined in non-randomized studies, thereby limiting our ability to identify the underlying cause of the associations. Significance and implications of the study results The association of obesity and underweight with worse prognosis has been shown in a number of clinical settings [8, 12, 13] but has received little attention in elderly patients. In addition, the prognosis of underweight frail patients with comorbidity and hip fracture, comprising approximately one out of four patients in Denmark [24], has not been studied in detail previously [49, 50]. Using unique data from the Laboratory Information Systems (LABKA) of the Central and Northern Denmark Regions, we were able to ascertain the incidence of AKI in 13,529 hip fracture patients. As AKI is a common complication after hip fracture surgery, any effort to identify relevant risk factors could affect the quality of treatment, improve our understanding of clinical course, and reduce mortality among hip fracture patients. To summarize

References 1.

2.

3.

4.

5.

6.

7.

8.

9. 10.

In our large regional population-based cohort study of 13,529 hip fracture patients, obese patients were at increased risk of postsurgical AKI compared with normal-weight patients. Among hip fracture patients who developed AKI within 5 days of surgery, as well as among those who did not, overweight and obesity were not associated with increased mortality, while underweight patients had elevated mortality up to 1 year following surgery, compared with normal-weight patients. The absolute mortality risks in all BMI groups are higher in AKI than in non-AKI patients. The underlying causal mechanisms of the associations remain to be clarified. Acknowledgements The authors wish to thank the orthopedic surgeons and other healthcare professionals at all hospitals in Denmark for their cooperation in submitting data to Danish national registries. The study was supported by a grant from Aarhus University Research Foundation and by the Program for Clinical Research Infrastructure (PROCRIN), established by the Lundbeck Foundation and the Novo Nordisk Foundation.

11.

12.

13.

14.

15.

16.

17.

Compliance with ethical standards Ethics statement This study was approved by the Danish Data Protection Agency (record number 2013-41-1924). As the study did not involve any contact with patients or any intervention, it was not necessary to obtain permission from the Danish Scientific Ethical Committee. Conflicts of interest None.

18.

19.

Kanis JA, Oden A, McCloskey EV, Johansson H, Wahl DA, Cooper C (2012) A systematic review of hip fracture incidence and probability of fracture worldwide. Osteoporos Int 23:2239–2256 Nielsen KA, Jensen NC, Jensen CM, Thomsen M, Pedersen L, Johnsen SP, Ingeman A, Bartels PD, Thomsen RW (2009) Quality of care and 30 day mortality among patients with hip fractures: a nationwide cohort study. BMC Health Serv Res 9:186 Lawrence VA, Hilsenbeck SG, Noveck H, Poses RM, Carson JL (2002) Medical complications and outcomes after hip fracture repair. Arch Intern Med 162:2053–2057 Roche JJ, Wenn RT, Sahota O, Moran CG (2005) Effect of comorbidities and postoperative complications on mortality after hip fracture in elderly people: prospective observational cohort study. BMJ 331:1374 Pedersen AB, Christiansen CF, Gammelager H, Kahlert J, Sorensen HT (2016) Risk of acute renal failure and mortality after surgery for a fracture of the hip: a population-based cohort study. Bone Joint J 98-B:1112–1118 Flegal KM, Carroll MD, Kit BK, Ogden CL (2012) Prevalence of obesity and trends in the distribution of body mass index among US adults, 1999-2010. JAMA 307:491–497 Billings FT, Pretorius M, Schildcrout JS, Mercaldo ND, Byrne JG, Ikizler TA, Brown NJ (2012) Obesity and oxidative stress predict AKI after cardiac surgery. J Am Soc Nephrol 23:1221–1228 Kelz RR, Reinke CE, Zubizarreta JR, Wang M, Saynisch P, EvenShoshan O, Reese PP, Fleisher LA, Silber JH (2013) Acute kidney injury, renal function, and the elderly obese surgical patient: a matched case-control study. Ann Surg 258:359–363 Wolf BR, Lu X, Li Y, Callaghan JJ, Cram P (2012) Adverse outcomes in hip arthroplasty: long-term trends. J Bone Joint Surg Am 94:e103 Ulucay C, Eren Z, Kaspar EC, Ozler T, Yuksel K, Kantarci G, Altintas F (2012) Risk factors for acute kidney injury after hip fracture surgery in the elderly individuals. Geriatr Orthop Surg Rehabil 3:150–156 Hsu CY, McCulloch CE, Iribarren C, Darbinian J, Go AS (2006) Body mass index and risk for end-stage renal disease. Ann Intern Med 144:21–28 Calle EE, Thun MJ, Petrelli JM, Rodriguez C, Heath CW Jr (1999) Body-mass index and mortality in a prospective cohort of U.S. adults. N Engl J Med 341:1097–1105 Manson JE, Willett WC, Stampfer MJ, Colditz GA, Hunter DJ, Hankinson SE, Hennekens CH, Speizer FE (1995) Body weight and mortality among women. N Engl J Med 333:677–685 Barba R, Zapatero A, Losa JE, Valdes V, Todoli JA, Di MP, Monreal M (2008) Body mass index and mortality in patients with acute venous thromboembolism: findings from the RIETE registry. J Thromb Haemost 6:595–600 Park J, Ahmadi SF, Streja E, Molnar MZ, Flegal KM, Gillen D, Kovesdy CP, Kalantar-Zadeh K (2014) Obesity paradox in endstage kidney disease patients. Prog Cardiovasc Dis 56:415–425 Silva HG, Mendonca LM, Conceicao FL, Zahar SE, Farias ML (2007) Influence of obesity on bone density in postmenopausal women. Arq Bras Endocrinol Metabol 51:943–949 Vemmos K, Ntaios G, Spengos K, Savvari P, Vemmou A, Pappa T, Manios E, Georgiopoulos G, Alevizaki M (2011) Association between obesity and mortality after acute first-ever stroke: the obesitystroke paradox. Stroke 42:30–36 Curtis JP, Selter JG, Wang Y, Rathore SS, Jovin IS, Jadbabaie F, Kosiborod M, Portnay EL, Sokol SI, Bader F, Krumholz HM (2005) The obesity paradox: body mass index and outcomes in patients with heart failure. Arch Intern Med 165:55–61 Escalante A, Haas RW, del Rincon I (2005) Paradoxical effect of body mass index on survival in rheumatoid arthritis: role of comorbidity and systemic inflammation. Arch Intern Med 165:1624–1629

Osteoporos Int 20.

21.

22. 23.

24. 25.

26. 27.

28.

29.

30.

31.

32.

33.

34.

35.

36. 37.

Pickkers P, de KN, Dusseljee J, Weerheijm D, van der Hoeven JG, Peek N (2013) Body mass index is associated with hospital mortality in critically ill patients: an observational cohort study. Crit Care Med 41:1878–1883 Sorensen HT (1997) Regional administrative health registers as a resource in clinical epidemiology. A study of options, strenghts, limitations and data quality provided with examples of use. Int J Risk Safety Med 10:1–22 Frank L (2000) Epidemiology. When an entire country is a cohort. Science 287:2398–2399 WHO (1995) Physical status: the use and interpretation of anthropometry. Report of a WHO expert committee. WHO technical report series 854. World Health Organization. 1995, Geneva Danish Hip fracture database (2015) Annual report 2014 Grann AF, Erichsen R, Nielsen AG, Froslev T, Thomsen RW (2011) Existing data sources for clinical epidemiology: the clinical laboratory information system (LABKA) research database at Aarhus University, Denmark. Clin Epidemiol 3:133–138 International Union for Pure and Applied Chemistry (2013). http://www.iupac.org/ Go AS, Parikh CR, Ikizler TA, Coca S, Siew ED, Chinchilli VM, Hsu CY, Garg AX, Zappitelli M, Liu KD, Reeves WB, Ghahramani N, Devarajan P, Faulkner GB, Tan TC, Kimmel PL, Eggers P, Stokes JB (2010) The assessment, serial evaluation, and subsequent sequelae of acute kidney injury (ASSESS-AKI) study: design and methods. BMC Nephrol 11:22 Bellomo R, Ronco C, Kellum JA, Mehta RL, Palevsky P (2004) Acute renal failure—definition, outcome measures, animal models, fluid therapy and information technology needs: the second international consensus conference of the acute dialysis quality initiative (ADQI) group. Crit Care 8:R204–R212 Kidney Disease: Improving Global Outcomes (KDIGO) Acute Kidney Injury Work Group (2012) KDIGO clinical practice guideline for acute kidney injury. Kidney International Supplements 2:1–138 Schmidt M, Pedersen L, Sorensen HT (2014) The Danish civil registration system as a tool in epidemiology. Eur J Epidemiol 29: 541–549 Charlson ME, Pompei P, Ales KL, MacKenzie CR (1987) A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 40:373–383 de Groot V, Beckerman H, Lankhorst GJ, Bouter LM (2003) How to measure comorbidity. A critical review of available methods. J Clin Epidemiol 56:221–229 Levey AS, Eckardt KU, Tsukamoto Y, Levin A, Coresh J, Rossert J, De ZD, 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 67:2089–2100 Johannesdottir SA, Horvath-Puho E, Ehrenstein V, Schmidt M, Pedersen L, Sorensen HT (2012) Existing data sources for clinical epidemiology: the Danish National Database of reimbursed prescriptions. Clin Epidemiol 4:303–313 Ranstam J, Karrholm J, Pulkkinen P, Makela K, Espehaug B, Pedersen AB, Mehnert F, Furnes O (2011) Statistical analysis of arthroplasty data. II. Guidelines. Acta Orthop 82:258–267 Carpenter J, Kenward M (2013) Multiple imputation and its application Jafari SM, Huang R, Joshi A, Parvizi J, Hozack WJ (2010) Renal impairment following total joint arthroplasty: who is at risk? J Arthroplast 25:49–53, 53

38.

Takahashi K, Nammour TM, Fukunaga M, Ebert J, Morrow JD, Roberts LJ, Hoover RL, Badr KF (1992) Glomerular actions of a free radical-generated novel prostaglandin, 8-epi-prostaglandin F2 alpha, in the rat. Evidence for interaction with thromboxane A2 receptors. J Clin Invest 90:136–141 39. Brienza N, Giglio MT, Marucci M, Fiore T (2009) Does perioperative hemodynamic optimization protect renal function in surgical patients? A meta-analytic study. Crit Care Med 37:2079–2090 40. Sinclair S, James S, Singer M (1997) Intraoperative intravascular volume optimisation and length of hospital stay after repair of proximal femoral fracture: randomised controlled trial. BMJ 315:909–912 41. Piccoli A, Brunani A, Savia G, Pillon L, Favaro E, Berselli ME, Cavagnini F (1998) Discriminating between body fat and fluid changes in the obese adult using bioimpedance vector analysis. Int J Obes Relat Metab Disord 22:97–104 42. James MT, Grams ME, Woodward M, Elley CR, Green JA, Wheeler DC, de JP, Gansevoort RT, Levey AS, Warnock DG, Sarnak MJ (2015) A meta-analysis of the Association of Estimated GFR, albuminuria, diabetes mellitus, and hypertension with acute kidney injury. Am J Kidney Dis 66:602–612 43. Choban PS, Flancbaum L (1997) The impact of obesity on surgical outcomes: a review. J Am Coll Surg 185:593–603 44. Bihorac A, Yavas S, Subbiah S, Hobson CE, Schold JD, Gabrielli A, Layon AJ, Segal MS (2009) Long-term risk of mortality and acute kidney injury during hospitalization after major surgery. Ann Surg 249:851–858 45. Henegar JR, Bigler SA, Henegar LK, Tyagi SC, Hall JE (2001) Functional and structural changes in the kidney in the early stages of obesity. J Am Soc Nephrol 12:1211–1217 46. Brandstrup B, Tonnesen H, Beier-Holgersen R, Hjortso E, Ording H, Lindorff-Larsen K, Rasmussen MS, Lanng C, Wallin L, Iversen LH, Gramkow CS, Okholm M, Blemmer T, Svendsen PE, Rottensten HH, Thage B, Riis J, Jeppesen IS, Teilum D, Christensen AM, Graungaard B, Pott F (2003) Effects of intravenous fluid restriction on postoperative complications: comparison of two perioperative fluid regimens: a randomized assessor-blinded multicenter trial. Ann Surg 238: 641–648 47. Choban PS, Heckler R, Burge JC, Flancbaum L (1995) Increased incidence of nosocomial infections in obese surgical patients. Am Surg 61:1001–1005 48. Bouillanne O, Dupont-Belmont C, Hay P, Hamon-Vilcot B, Cynober L, Aussel C (2009) Fat mass protects hospitalized elderly persons against morbidity and mortality. Am J Clin Nutr 90:505–510 49. Kvamme JM, Holmen J, Wilsgaard T, Florholmen J, Midthjell K, Jacobsen BK (2012) Body mass index and mortality in elderly men and women: the Tromso and HUNT studies. J Epidemiol Community Health 66:611–617 50. Weiss A, Beloosesky Y, Boaz M, Yalov A, Kornowski R, Grossman E (2008) Body mass index is inversely related to mortality in elderly subjects. J Gen Intern Med 23:19–24 51. Mahan KL, Escott-Stump S, Raymond JL (2011) Krauses’s Food and the Nutrition Care Process, 13th edn 52. Mainz J, Krog BR, Bjornshave B, Bartels P (2004) Nationwide continuous quality improvement using clinical indicators: the Danish National Indicator Project. Int J Qual Health Care 16(Suppl 1):i45–i50 53. Mokdad AH, Marks JS, Stroup DF, Gerberding JL (2004) Actual causes of death in the United States, 2000. JAMA 291:1238–1245