Risk factors for fracture in adult kidney transplant recipients

AIM: To determine the general and transplant-specific risk factors for fractures in kidney transplant recipients.METHODS: We conducted a cohort study of all adults who received a kidney-only transplant(n = 2723) in Ontario, Canada between 2002 and 2009. We used multivariable Cox proportional hazards regression to determine general and transplant-specific risk factors for major fractures(proximal humerus, forearm, hip, and clinical vertebral). The final model was established using the backward elimination strategy, selecting risk factors with a P-value ≤ 0.2 and forcing recipient age and sex into the model. We also assessed risk factors for other fracture locations(excluding major fractures, and fractures involving the skull, hands or feet). RESULTS: There were 132 major fractures in the follow-up(8.1 fractures per 1000 person-years). General risk factors associated with a greater risk of major fracture were older recipient age [adjusted hazard ratio(a HR) per 5-year increase 1.11, 95%CI: 1.03-1.19] and female sex(a HR = 1.81, 95%CI: 1.28-2.57). Transplant-specific risk factors associated with a greater risk of fracture included older donor age(5-year increase)(a HR = 1.09, 95%CI: 1.02-1.17) and end-stage renal disease(ESRD) caused by diabetes(a HR = 1.72, 95%CI: 1.09-2.72) or cystic kidney disease(a HR = 1.73, 95%CI: 1.08-2.78)(compared to glomerulonephritis as the reference cause). Risk factors across the two fracture locations were not consistent(major fracture locations vs other). Specifically, general risk factors associated with an increased risk of other fractures were diabetes and a fall with hospitalization prior to transplantation, while length of time on dialysis, and renal vascular disease and other causes of ESRD were the transplant-specific risk factors associated with a greater risk of other fractures.CONCLUSION: Both general and transplant-specific risk factors were associated with a higher risk of fractures in kidney transplant recipients. Results can be used for clinical prognostication.