Objective: Only scarce data are available on the long-term absolute risk (AR) of all clinical fractures, taking into account the time when they occurred. Therefore, we assessed during a 10-year follow-up the risk factors associated with the occurrence of any first or second clinical fracture.
Study design: This was a population-based study in 10 general practice centres. The sample comprised 2372 postmenopausal women, aged between 50 and 80 years at baseline, who completed a questionnaire about the incidence of radiographically confirmed fractures and fracture risks, analysed by multiple Cox regression.
Main outcome measure: AR for any clinical fracture.
Results: During the 10-year follow-up, 380 women (16%) had a fracture. A first fracture occurred in 267 women (11%). Osteoporosis at the lumbar spine (T-score <-2.5; hazard ratio (HR) 1.8, 95% confidence interval (CI) 1.4-2.3) and age over 60 years (HR1.4, 95% CI 1.1-1.8) were the only risk factors retained in the Cox analysis. The AR in the lowest-risk group was 10%, and it was 23% in the highest-risk group. A second fracture occurred in 113 women during follow-up (5%). The time when a fracture occurred was the only risk factor retained in the Cox analysis. The AR for a second fracture was 41% in the five years after any first fracture before baseline and 25% if the first fracture had occurred earlier (HR 1.8, 95% CI 1.3-2.7).
Conclusion: In postmenopausal women, over a 10-year follow-up, the AR of a second clinical fracture is highest in the five years after any first clinical fracture. The AR for a first clinical fracture is lower and depends on osteoporosis and age.