Abstract
IMPORTANCE: Fragility fractures are often defined as those that occur after a certain age (eg, 40-50 years). Whether fractures occurring in early adulthood are equally associated with future fractures is unclear. OBJECTIVE: To examine whether the age at which a prior fracture occurred is associated with future fracture risk. DESIGN, SETTING, AND PARTICIPANTS: This observational, population-based cohort study included individuals from the Manitoba Bone Mineral Density Registry with a first bone mineral density (BMD) measurement between January 1, 1996, and March 31, 2018, with and without prior fracture in adulthood. Data analysis was completed between April 1, and May 31, 2023. EXPOSURE: Individuals with fractures before their first dual-energy x-ray absorptiometry were stratified by the age at first fracture (10-year intervals from 20-29 to ≥80 years of age). MAIN OUTCOMES AND MEASURES: Incident fractures occurring after dual-energy x-ray absorptiometry (index date) and before March 31, 2021, were identified using linked provincial administrative health data. RESULTS: The cohort included 88 696 individuals (80 066 [90.3%] female; mean [SD] age, 64.6 [11.0] years) with a mean (SD) femoral neck T score of -1.4 (1.0). A total of 21 105 individuals (23.8%) had sustained a prior fracture at a mean (SD) age of 57.7 (13.6) years (range, 20.0-102.4 years) at the time of first prior fracture. During a mean (SD) of 9.0 (5.5) years of follow-up, incident fractures occurred in 13 239 individuals (14.6%), including 12 425 osteoporotic fractures (14.0%), 9440 major osteoporotic fractures (MOFs) (10.6%), and 3068 hip fractures (3.5%). The sex- and age-adjusted hazard ratios for all incident fractures, osteoporotic fractures, and MOFs, according to age at first fracture, were all significantly elevated, with point estimates ranging from 1.55 (95% CI, 1.28-1.88) to 4.07 (95% CI, 2.99-5.52). After adjusting for the additional covariates, the effect estimates were similar and remained significantly elevated, with point estimates ranging from fully adjusted hazard ratios of 1.51 (95% CI, 1.42-1.60) to 2.12 (95% CI, 1.67-2.71) across age categories. Sensitivity analyses examining age at last prior fracture and in those with multiple prior fractures showed similar results. CONCLUSIONS AND RELEVANCE: In this cohort study, fractures in adulthood were associated with future fractures regardless of the age at which they occurred. Thus, fractures in early adulthood should not be excluded when assessing an individual's ongoing fracture risk.