OR03-4 Atypical Femur Fracture Incidence Increases with Duration of Bisphosphonate Exposure

Abstract

BACKGROUND: Dell et al. reported that the incidence of atypical femur fracture (AFF) increased with duration of bisphosphonate (BP) exposure, ranging from 2 to 113 per 100,000 person-years with <2 to 8-10 years (yrs) treatment, respectively (JBMR 27:2544, 2012). These estimates were based on 128 AFF cases (over 95% women) identified in Kaiser Permanente Southern California. This study aims to validate Dell’s findings using female members of a different large integrated healthcare system and using a different method of capturing subjects (women were followed forward from the date of BP initiation and had health plan membership with no BP for 2 yrs prior). METHODS: Women age 45-89 of a large integrated healthcare delivery system who initiated oral BP in 2002-2014 were identified, excluding the following: receipt of intravenous BP, teriparatide or denosumab; skeletal disorder; advanced kidney disease; and metastatic cancer. The total period of subsequent BP exposure (TOTPERIOD) was determined by the time between the first and last prescriptions, inclusive of all treatment gaps. We also calculated the total days covered from the sum of all qualifying prescriptions (DAYSUPPLY). Women were followed for an AFF outcome during the period of BP exposure, censoring at membership termination, death, exclusion criteria met, age 90, end of 12 yr total exposure period, or end of study follow-up (2015). The incidence of AFF (with 95% confidence interval, CI) was calculated per person-years (p-y) exposure in 2-yr intervals, categorized as <2, 2 to <4, 4-<6, 6 to <8, 8 to <10 and 10 to <12 yrs exposure and adjusted for age using 2010 US Census data. RESULTS: Among 94,542 women, the median (interquartile range) TOTPERIOD and DAYSUPPLY of BP was 2.7 (IQR 0.7-5.6) and 1.6 (0.5-3.7) yrs. During follow-up for up to 12 yrs of BP exposure, a total of 85 women with AFF were identified. Mean age at AFF was 74 ± 7 yrs, and median TOTPERIOD was 8.3 yrs (interquartile range 6.5-9.6), with median DAYSUPPLY of 6.5 yrs (IQR 5.0-8.5). For each 2-yr increment, the age-adjusted incidence (95% CI) of AFF was: <2 yrs [n=94,542; 132,367 p-y; AFF incidence 2 (1-5) per 100,000 p-y]; 2 to <4 yrs [n=50,056; 81,823 p-y; AFF incidence 7 (4-12) per 100,000 p-y]; 4 to <6 yrs [n=32,219; 50,164 p-y; AFF incidence 24 (13-41) per 100,000 p-y]; 6 to <8 yrs [n=18,956; 28,858 p-y; AFF incidence 71 (29-158) per 100,000 p-y]; 8 to <10 yrs [n=10,505; 15,171 p-y; AFF incidence 101 (54-182) per 100,000 p-y]; 10 to <12 yrs [n=5132; 7627 p-y; AFF incidence 88 (42-177) per 100,000 p-y]. The incidence of AFF increased substantially with duration of BP exposure, increasing over 20-fold when comparing exposure <4 yrs to 8-12 yrs (4 to 93 per 100,000 p-y). CONCLUSION: The age-adjusted incidence of AFF increases progressively with duration of BP exposure. Overall, our data are strikingly similar to those reported by Dell et al. based on 128 AFF cases identified using a somewhat different approach.