Abstract

Study objective: To estimate the incidence of lower limb fractures in the United Kingdom and assess the relative importance of various risk factors for lower limb fractures.

Design: Cohort analysis and matched case-control study.

Setting: General practices contributing information to the General Practice Research Database.

Subjects: Individuals registered with these general practices who were at risk for a first time lower limb fracture from 1 January1990 to 31 December 2001.

Main outcome measures: Age, sex, and fracture site specific incidence rates; relative risks and population attributable risks for various medical risk factors.

Results: Overall, the risk of lower limb fracture was 17% higher in women then in men. Within age groups, men and women had generally similar proportions of fractures at specific sites in the lower limb. Among the risk factors evaluated, road collisions were associated with the highest relative risk for lower limb fracture, but only accounted for 3.1% or less of the population attributable risk for specific fracture types in any age group. The relative risk for lower limb fracture associated with a diagnosis of dementia was 2.3 (95% confidence interval 2.0 to 2.6), while relative risk estimates for other medical diagnoses were less than 2. Fracture risk was increased among current users of corticosteroids, antipsychotics, antidepressants, and hypnotic/sedatives, but the population attributable risks for each of these drug classes within fracture and age specific strata were only 3.0% or less.

Conclusions: Many risk factors for lower limb fracture have been identified, but population attributable risk estimates for various risk factors are small. These findings suggest that multifactorial prevention programs are needed to decrease the incidence of lower limb fractures in the general population.