Association of prevalent vertebral fractures, bone density, and alendronate treatment with incident vertebral fractures: Effect of number and spinal location of fractures

Vertebral fractures are the most common osteoporotic fracture and are assoc iated with significant pain and disability. Prior vertebral fracture and lo w bone mineral density (BMD) are strong predictors of new vertebral fractur e. Using data from 6082 women, ages 55-80 years, in the Fracture Interventi on Trial (a randomized, placebo-controlled trial of the antiresorptive agen t, alendronate), we explored the association of the number of prior vertebr al fractures with the risk of new fractures and whether this association is influenced by the spinal location of fractures. The risk of future vertebr al fractures increased with the number of prevalent fractures, independentl y of age and BMD; in the placebo group, more than half of the women with fi ve or more fractures at baseline developed new vertebral fractures, compare d to only 3.8% of women without prior vertebral fractures. The magnitude of association with an increased risk of future vertebral fractures was equal for prevalent fractures located in either the "lower" (T12-L4) (relative r isk [RR] = 2.9; 95% CI = 1.9, 3.6) or "upper" (T4-10) spine (RR = 2.6; 95% CI = 1.9, 3.6). We found no evidence that the effectiveness of alendronate in reducing the risk of future vertebral fracture was attenuated in women w ith up to five or more prevalent fractures, or that it varied by the locati on of prevalent fractures. However, prevalent vertebral fractures in any lo cation were more strongly associated with risk of new fractures in the uppe r (RR = 5.2; 95% CI = 3.2, 8.3) than in the lower spine (2.3; 1.6, 3.3). In addition, each 1 SD decrease in spinal BMD was associated with a 2.1 (1.7, 2.6) times greater odds of new fracture in the upper spine, compared with 1.5 (1.3, 1.8) for the lower spine. These findings suggest that, in older w omen, osteoporosis may be a stronger risk factor for new fractures in the u pper (vs. lower) thoracolumbar spine, although we found no evidence that th e location of prior fractures should influence treatment decisions. Physici ans should recognize that prior vertebral fractures are a strong risk facto r for future fractures, and consider treating such patients to reduce their risk of subsequent fractures. (C) 1999 by Elsevier Science Inc. All rights reserved.