Targeted exercise against osteoporosis: A systematic review and meta-analysis for optimising bone strength throughout life

Abstract

Background: Exercise is widely recommended to reduce osteoporosis, falls and related fragility fractures, but its effect on whole bone strength has remained inconclusive. The primary purpose of this systematic review and meta-analysis was to evaluate the effects of long-term supervised exercise (> or =6 months) on estimates of lower-extremity bone strength from childhood to older age.

Methods: We searched four databases (PubMed, Sport Discus, Physical Education Index, and Embase) up to October 2009 and included 10 randomised controlled trials (RCTs) that assessed the effects of exercise training on whole bone strength. We analysed the results by age groups (childhood, adolescence, and young and older adulthood) and compared the changes to habitually active or sedentary controls. To calculate standardized mean differences (SMD; effect size), we used the follow-up values of bone strength measures adjusted for baseline bone values. An inverse variance-weighted random-effects model was used to pool the results across studies.

Results: Our quality analysis revealed that exercise regimens were heterogeneous; some trials were short in duration and small in sample size, and the weekly training doses varied considerably between trials. We found a small and significant exercise effect among pre- and early pubertal boys [SMD, effect size, 0.17 (95% CI, 0.02-0.32)], but not among pubertal girls [-0.01 (-0.18 to 0.17)], adolescent boys [0.10 (-0.75 to 0.95)], adolescent girls [0.21 (-0.53 to 0.97)], premenopausal women [0.00 (-0.43 to 0.44)] or postmenopausal women [0.00 (-0.15 to 0.15)]. Evidence based on per-protocol analyses of individual trials in children and adolescents indicated that programmes incorporating regular weight-bearing exercise can result in 1% to 8% improvements in bone strength at the loaded skeletal sites. In premenopausal women with high exercise compliance, improvements ranging from 0.5% to 2.5% have been reported.

Conclusions: The findings from our meta-analysis of RCTs indicate that exercise can significantly enhance bone strength at loaded sites in children but not in adults. Since few RCTs were conducted to investigate exercise effects on bone strength, there is still a need for further well-designed, long-term RCTs with adequate sample sizes to quantify the effects of exercise on whole bone strength and its structural determinants throughout life.

Figure 1

Flow diagram of the search process of exercise RCTs to improve or maintain bone strength.

Figure 2

Effects of exercise on indices of bone strength (standard mean difference, 95% CI) in young girls at the distal tibia, tibial shaft and femoral neck. The squares and diamonds represent the test values for individual studies and the overall effect, respectively.

Figure 3

Effects of exercise on bone strength (standard mean difference, 95% CI) in young boys at the distal tibia, tibial shaft and femoral neck. The squares and diamonds represent the test values for individual studies and the overall effect, respectively.

Figure 4

Effects of exercise on bone strength (standard mean difference, 95% CI) in adolescent boys and girls at the femoral neck. The squares represent the test values for the individual study.

Figure 5

Effects of exercise on bone strength (standard mean difference, 95% CI) in premenopausal women at the proximal tibia and femoral midshaft.

Figure 6

Effects of exercise on bone strength (standard mean difference, 95% CI) in postmenopausal women at the distal tibia, mid tibia, femoral neck, midfemur and proximal tibia. The squares and diamonds represent the test values for individual studies and the overall effect, respectively.

Figure 7

The changes in cortical bone structure in response to exercise at the mid- and distal humerus in female tennis players. The increase in cortical area in the prepubertal players was the result of greater periosteal (outer bone surface) than endocortical (inner bone surface) expansion at the midhumerus, but greater periosteal expansion alone at the distal humerus. During the peri- to postpubertal years, loading resulted in both periosteal expansion and endocortical contraction at both sites (adapted from Bass et al. [44]).

Figure 8

Athletes representing high-impact (H-I) and odd-impact (O-I) type of exercise loadings clearly have thicker cortices than their sedentary counterparts at the femoral neck. In a recent cross-sectional study [66], athletes in high-impact sports had 60% thicker inferior cortex; however, athletes representing odd-impact sports had 20% thicker cortex uniformly around the femoral neck, whereas athletes in high-magnitude (H-M), low-impact (L-I), and nonimpact (N-I) sports did not have thicker cortices than their nonathletic counterparts (top) (adapted from Nikander et al. [66]).