Objective: To determine whether factors associated with moderate chronic kidney disease can be used to independently predict fracture risk in postmenopausal white women by analyzing the effect of adding renal function measurements to traditional risk factors (eg, age, body weight, bone mineral density) for fracture risk assessment.
Participants and methods: In a prospective, population-based cohort study, postmenopausal women residing in Rochester, MN, with baseline measurements of bone mineral density and renal function were followed up for as long as 25 years for occurrence of fractures. Participants were enrolled in 1980-1984 or 1991-1993, and outcomes were analyzed in 2007. Standardized incidence ratios were used to compare the number of observed fractures with the number of predicted fractures, and potential risk factors were evaluated with Andersen-Gill time-to-fracture regression models.
Results: During 5948 person-years of follow-up of 427 women, 254 women (59.5%) experienced a total of 563 fractures, 394 (70.0%) of which resulted from moderate trauma. Excluding incidentally diagnosed fractures, the 186 clinically diagnosed fractures were statistically undifferentiated from the 195 predicted fractures (standardized incidence ratio, 0.95; 95% confidence interval, 0.82-1.10). No significant trends were observed toward increasing fracture risk with inclusion of quintiles of declining renal function (P>.10). In univariate analyses, serum creatinine concentration, creatinine clearance rate, and estimated glomerular filtration rate were associated with greater risk of some fractures. In multivariate analyses, however, decreasing renal function was not found to be a significant risk factor, after adjusting for age, body weight, and bone mineral density.
Conclusion: The addition of serum creatinine concentration, creatinine clearance rate, or estimated glomerular filtration rate does not improve fracture risk prediction in postmenopausal white women who have moderate chronic kidney disease. This result can be partly explained by the fact that important risk factors for decreased renal function (eg, advanced age, lower body weight) are already accounted for in most fracture prediction models.