The objective of this study was to assess the importance of time since prior fracture as a risk factor for future osteoporotic fractures and how it affects 10-year fracture rates. We identified 39,991 women 45 years of age or older undergoing baseline bone mineral density (BMD) testing (1990-2007) from a regional database that contains dual-energy X-ray absorptiometry (DXA) results for Manitoba, Canada. Health service records were used to identify nontrauma ICD-9-CM fracture codes preceding DXA, grouped as "major" fractures (n = 5178; hip, spine, forearm, and humerus) or "minor" fractures (n = 3479; ribs, sternum, pelvis, trunk, clavicle, scapula, patella, tibia/fibula, and ankle). Time since prior fracture was coded in years as less than 1, 1 to 5, 5 to 10, and more than 10. Incident fractures (ie, hip, spine, forearm, and humerus) after BMD testing were identified (mean follow-up 4.2 years, maximum 10 years) and studied in Cox proportional-hazards models adjusted for age, BMD T-score, and other covariates. After BMD testing, n = 1749 (4.4%) women experienced an incident fracture. Prior major fracture was a strong risk factor for incident fracture, greatest risk in the first year [hazard ratio (HR) 1.90, 95% confidence interval (CI) 1.60-2.25], declining by more than 10 years (HR 1.62, 95% CI 1.25-2.10). Prior minor fracture was a weaker risk factor, greatest in the first year (HR 1.45, 95% CI 1.13-1.87) and no longer significant by 1 to 5 years. Major and minor fractures both showed a time-dependent decline in importance as risk factors. In conclusion, time since prior fracture modifies future fracture risk, but prior fractures of the hip, spine, forearm, and humerus remain strong risk factors even 10 years later. Fracture risk assessment should emphasize the importance of prior fractures at these sites.