Risk of atypical femoral fracture higher with longer duration of bisphosphonate use and in other groups of women

Reference

Black DM, Geiger EJ, Eastell R, et al. Atypical femur fracture risk versus fragility fracture prevention with bisphosphonates. N Engl J Med 2020;383(8):743-753.

Synopsis

Balancing benefits and harms when prescribing a therapy is, of course, central to shared decision-making. Inherent biases in cognition can lead to the avoidance of effective therapy due to the overvaluation of rare adverse events, while a failure to fully appreciate harms can work in the other direction and lead to overtreatment. In this retrospective cohort study, researchers used the records of the Kaiser Permanente health care system to identify 196,129 women who were 50 years or older and taking a bisphosphonate. Approximately 60% were 65 years or older, 53% were White, 24% were Hispanic, 13.5% were Asian, and 6% were Black. The authors identified 277 women who had experienced an atypical (subtrochanteric) femoral shaft fracture, excluding anyone with evidence of high-energy trauma. They also gathered data on potential confounders such as race, age, corticosteroid use, tobacco use, height, and weight. The risk of atypical fracture was somewhat higher for women aged 65 to 84 years than for older women (hazard ratio [HR] = 2.50 - 2.76), but the most important risk factors were Asian race (HR = 4.84; 95% CI 3.57 - 6.56) and increasing duration of bisphosphonate use (HR = 2.5 for 3 months to less than 3 years of use, HR = 8.9 for 3 years to less than 5 years of use, HR = 20 for 5 years to less than 8 years of use, and HR = 44 for 8 or more years of use). Atypical fractures were rare in Black women. The authors used modeling to estimate the number of hip fractures and total clinical osteoporotic fractures prevented in each racial or ethnic group. Since their assumptions were based on a study of women 65 years or older with a hip T score of less than -2.5, and since many of the women in the cohort who were prescribed bisphosphonates were probably at lower risk, the estimate of fractures prevented is probably too high for a real world setting. For White and Hispanic women, the authors report a very favorable ratio of osteoporotic fractures to atypical fractures; for example, at 3 years a ratio of 541 osteoporotic and 149 hip fractures prevented to 2 atypical fractures caused. However, for Asian women, these ratios were 330 and 91 to 8 (less benefit and more harm) at 3 years, with ratios of 831 and 360 to 236 at 8 years of use. Limitations of the study include a failure to adjust for activity level.