A population-based cohort study of fracture risk among 122 255 adults before and after cancer diagnosis compared with matched controls

Individuals with cancer have numerous risk factors for fracture. The objective of this study was to compare fracture risk trajectories in individuals with and without cancer before and after cancer diagnosis. We conducted a matched retrospective cohort study using administrative healthcare and cancer registry data for Manitoba, Canada. We identified residents diagnosed with cancer (excluding nonmelanoma skin cancer) at age >18 years between 1987-2014. Each cancer patient was uniquely matched with up to four non-cancer individuals based on age, sex, length of residency, and area of residence. Incident clinical fractures were identified using linked healthcare records and validated case definition, before and after cancer diagnosis (index date) up to March 31, 2021. Our cohort comprised 122,255 cancer cases and 460,029 matched controls (average 3.8 non-cancer matches per case). The incident rate of major fracture in the year prior to cancer diagnosis was 10.45 per 1000 person-years (95% CI 9.77-11.19), increasing to a peak of 20.30 (95% CI 19.63-21.00) at 1-5 years post-diagnosis. Incident rate ratios (IRRs) for major, humerus, and vertebral fractures were significantly elevated in cancer versus non-cancer individuals in the year before the index date and remained significantly elevated for up to 5-10 years. Fracture IRRs were increased across all cancer types except gynecological, melanoma, and thyroid cancers. In conclusion, fracture risk is elevated in the year preceding cancer diagnosis and persists at a higher level compared to the non-cancer population for up to 10 years. This study emphasizes the necessity of timely and ongoing fracture risk assessment and prompt anti-osteoporosis therapy initiation.