Pre-existing fractures and bone mass predict vertebral fracture incidence in women

Abstract

Objective: To determine the independent contributions of bone mass and existing fractures as predictors of the risk for new vertebral fractures.

Subjects: Postmenopausal Japanese-American women.

Measurements: Baseline measurements of the distal radius, the proximal radius, and the calcaneus were obtained in 1981 using single-photon absorptiometry. Measurements of the lumbar spine were obtained in 1984 using dual-photon absorptiometry. Prevalent vertebral fractures were identified using dimensions measured on lateral radiographs; vertebral height values more than 3 SD below vertebra-specific means were considered to indicate fracture. Statistical models were used to evaluate the utility of bone mass and existing (prevalent) fractures to predict the risk for new fractures during an average follow-up of 4.7 years.

Main results: Differences of 2 SD in bone mass were associated with fourfold to sixfold increases in the risk for new vertebral fractures. A single fracture at the baseline examination increased the risk for new vertebral fractures fivefold. Presence of two or more fractures at baseline increased the risk 12-fold. A combination of low bone mass (below the 33d percentile) and the presence of two or more prevalent fractures increased the risk 75-fold, relative to women with the highest bone mass (above the 67th percentile) and no prevalent fractures. Stature, body mass index, arm span, and spinal conditions such as scoliosis, osteoarthritis, and sacroiliitis did not predict fracture incidence (P greater than 0.05). Weight was marginally predictive (P = 0.04) of fracture incidence but became nonpredictive after adjusting for bone mass (P greater than or equal to 0.05).

Conclusions: Both bone mass and prevalent vertebral fractures are powerful predictors of the risk for new vertebral fractures. Combining information about bone mass and prevalent fracture appears to be better for predicting new fractures than either variable alone. Physicians can use these risk factors to identify patients at greatest risk for new fractures.