Exercise training and bone mineral density in postmenopausal women: an updated systematic review and meta-analysis of intervention studies with emphasis on potential moderators

The aim of this systematic review and meta-analysis was (1) to determine exercise effects on bone mineral density (BMD) in postmenopausal women and (2) to address the corresponding implication of bone and menopausal status or supervision in postmenopausal women. A comprehensive search of eight electronic databases according to the PRISMA statement up to August 9, 2022, included controlled exercise trials ≥ 6 months. BMD changes (standardized mean differences: SMD) at the lumbar spine (LS), femoral neck (FN), and total hip (TH) were considered as outcomes. Study group comparisons were conducted for osteopenia/osteoporosis versus normal BMD, early versus late postmenopausal women, and predominantly supervised versus predominantly non-supervised study arms. We applied an inverse heterogeneity (IVhet) model. In summary, 80 studies involving 94 training and 80 control groups with a pooled number of 5581 participants were eligible. The IVhet model determined SMDs of 0.29 (95% CI: 0.16–0.42), 0.27 (95% CI: 0.16–0.39), and 0.41 (95% CI: 0.30–0.52) for LS, FN, and THBMD, respectively. Heterogeneity between the trial results varied from low (I(2) = 20%, TH BMD) to substantial (I(2) = 68%, LS-BMD). Evidence for publication bias/small study effects was negligibly low (FN-, TH-BMD) to high (LSBMD). We observed no significant differences (p > .09) for exercise effects on LS-, FN-, or TH-BMD-LS between studies/study arms with or without osteopenia/osteoporosis, early versus late postmenopausal women, or predominantly supervised versus non-supervised exercise programs. Using robust statistical methods, the present work provides further evidence for a positive effect of exercise on BMD in postmenopausal women. Differences in bone status (osteopenia/osteoporosis versus normal bone), menopausal status (early versus late postmenopausal), and supervision (yes versus no) did not significantly affect the exercise effects on BMD at LS or proximal femur.