The long-term protective effect of stronger back muscles on the spine was determined in 50 healthy white postmenopausal women, aged 58-75 years, 8 years after they had completed a 2 year randomized, controlled trial. Twenty-seven subjects had performed progressive, resistive back-strengthening exercises for 2 years and 23 had served as controls. Bone mineral density, spine radiographs, back extensor strength, biochemical marker values, and level of physical activity were obtained for all subjects at baseline, 2 years, and 10 years. Mean back extensor strength (BES) in the back-exercise (BE) group was 39.4 kg at baseline, 66.8 kg at 2 years (after 2 years of prescribed exercises), and 32.9 kg at 10 years (8 years after cessation of the prescribed exercises). Mean BES in the control (C) group was 36.9 kg at baseline, 49.0 kg at 2 years, and 26.9 kg at 10 years. The difference between the two groups was still statistically significant at 10 year follow-up (p = 0.001). The difference in bone mineral density, which was not significant between the two groups at baseline and 2 year follow-up, was significant at 10 year follow-up (p = 0.0004). The incidence of vertebral compression fracture was 14 fractures in 322 vertebral bodies examined (4.3%) in the C group and 6 fractures in 378 vertebral bodies examined (1.6%) in the BE group (chi-square test, p = 0.0290). The relative risk for compression fracture was 2.7 times greater in the C group than in the BE group. To our knowledge, this is the first study reported in the literature demonstrating the long-term effect of strong back muscles on the reduction of vertebral fractures in estrogen-deficient women.