Abstract
OBJECTIVES: To examine the interactive effect of cognition and body weight on hip fracture. DESIGN: A 7-year (1993-2000) prospective cohort study. SETTING: Five southwestern states (Texas, New Mexico, Arizona, Colorado, and California). PARTICIPANTS: Noninstitutionalized Mexican Americans (N=2,653) aged 65 and older and free of hip fracture at baseline interview. MEASUREMENTS: Incidence of hip fracture at 2-, 5-, and 7-year follow-up interviews. Body weight and cognition were measured using body mass index (BMI) and Mini-Mental State Examination score, respectively. Covariates included sociodemographics, self-reported medical conditions, visual acuity, and Short Physical Performance Battery. RESULTS: A significant interaction between BMI and hip fracture was found in persons with cognitive impairment (hazard ratio =0.91, 95% confidence interval=0.85-0.98; P=.02), after adjusting for covariates. In the lowest BMI category, the hip fracture rate in cognitively impaired subjects was more than four times the hip fracture rate for subjects who were not cognitively impaired with the same BMI (34.6% vs 8.7%). Hip fracture rates in the highest BMI category were similar in persons with and without cognitive impairment (9.3% vs 6.1%). CONCLUSION: Low cognitive function increased the conditional association between BMI and hip fracture in older Mexican Americans. The relationship between BMI and cognition is potentially important in identifying persons at risk for hip fracture and supports the need to include cognitive and anthropometric measures in the assessment of hip fracture risk into osteoporosis screening programs.