We observed in a controlled 2 year longitudinal trial in 248 perimenopausal women that a daily calcium supplement of either 1000 or 2000 mg Ca2+ significantly reduced lumbar bone loss and bone turnover in the first year of calcium supplementation. In the second supplementation year the rate of lumbar bone loss in the treated subjects was not significantly different from that in the control group, although two of the three biochemical parameters of bone turnover remained decreased throughout the study. To quantify further the long-term effect of calcium supplementation, we extended the study for another year in 214 women. In the women of the control group who were menstruating until the last year of the trial, the mean change in lumbar bone mineral density after 3 years was -3.2% of the initial value versus 1.6% in the calcium-supplemented groups (p < 0.01). The decrease in lumbar bone loss in these supplemented premenopausal and early perimenopausal women remained statistically significant in the second and third years of supplementation. In the women who stopped menstruating before or during the study, the long-term reduction in lumbar bone loss was not significant (mean difference between control and treatment groups < 0.6% points after 3 years). The decrease in metacarpal cortical thickness (MCT) in the treated subjects during 3 years was on average -3.0% of the initial value in the control versus -2.0% in the supplemented subjects (P < 0.01). The effect of calcium supplementation on MCT was not significantly related to the menopausal status of the subjects.(ABSTRACT TRUNCATED AT 250 WORDS)