Femur neck bone mineral density and fracture risk by age, sex, and race or Hispanic origin in older US adults from NHANES III.

UNLABELLED

Differences in the relationship between femur neck bone mineral density (FNBMD) and fracture risk were examined by age, sex, and race/ethnicity in the third National Health and Nutrition Examination Survey (NHANES III) cohort. FNBMD had similar, significant predictive utility for fracture in the different subgroups, but it did not completely account for subgroup differences in risk.

PURPOSE

Few previous studies of FNBMD and fracture risk examined the relationship by age, sex, and race within the same cohort. The present study examined the relationship between FNBMD and risk of incident major osteoporotic fracture (hip, spine, radius, and humerus) in older US adults from NHANES III (1988-1994).

METHODS

Incident fractures were identified using linked mortality and Medicare records obtained through 2007 for 2,743 men and women ages 65 years and older. FNBMD was measured by dual-energy X-ray absorptiometry. Cox proportional hazards models were used to estimate the hazards ratio (HR) of fracture for FNBMD and femur neck T score and risk of major osteoporotic fracture.

RESULTS

The sample included 380 incident major osteoporotic fractures. Fracture risk approximately doubled for each SD decrease in FNBMD. HR's for FNBMD were similar within age, sex, or race/Hispanic origin subgroups, and also for T scores calculated with either white female or sex- and race/ethnic-specific reference data. Adding FNBMD to Cox models slightly attenuated HR for age, sex, or race/Hispanic origin, but all three variables remained significant predictors of fracture risk.

CONCLUSIONS

FNBMD had similar, significant predictive utility within age, sex, and race/Hispanic origin subgroups. However, FNBMD did not appear to completely account for fracture risk differences in these subgroups. Similarity of HR's for T scores calculated with two different reference databases support use of a uniform reference database to calculate these scores.