Thirty-three studies met the inclusion criteria: 27 cross-sectional (3,227 females and 384 males), 2 longitudinal (n=396) and 4 intervention studies (n=193).
In the cross-sectional studies in females, the correlations were significantly heterogeneous (p<0.05) for both the correlation coefficent (R) and partial correlation coefficient (PR), as tested by means of a chi-squared test based on Fisher's Z-transformation. The results of the cross-sectional studies in pre-menopausal females showed a significant R (0.13, 95% confidence interval, CI: 0.09, 0.16) and PR (0.08, 95% CI: 0.05, 0.12), both weighted by sample size.
The results for the 3 cross-sectional studies in men showed a non significant R (0.01, 95% CI: -0.11, 0.13) and PR (0.02, 95% CI: -0.12, 0.16), weighted by sample size.
The weighted mean in the 2 longitudinal studies in pre-menopausal women was 0.04 (95% CI: -0.06, 0.14), i.e. non significant.
The 4 intervention studies in pre-menopausal women showed that the percentage of bone loss per year in the calcium-supplemented group was less than that of the control group: the weighted value was 1.32% (95% CI: 1.21, 1.42).