Can Exercise, Calcium Supplementation, And Changes In Body Composition Explain Bone Responders And Nonresponders?

Abstract

PURPOSE: To determine whether 4-year bone mineral density (BMD) changes were different between responders (R) and nonresponders (NR) after adjusting for age, 4-year exercise frequency (EX Freq), 4-year calcium supplementation (CA) and 4-year changes in body composition (BC). METHODS: Sedentary postmenopausal women (40-65 years) with HT (1-3.9 years) or without HT (>1 year) were recruited to participate in a study designed to determine the effects of strength and weight-bearing exercise training on BMD in calcium- supplemented postmenopausal women. After year 1, women randomized to control were invited to self-select exercise or control. One hundred and sixty-seven women completed 4 years of the study. Regional BMD and BC were measured using dual-energy x-ray absorptiometry at baseline and yearly thereafter. EX Freq, CA and HT use were monitored throughout the study. Hierarchical multiple regression analyses were used to predict 4-year BMD changes from EX Freq, CA, age, and 4-year changes in BC. Analyses were conducted separately for HT users (n = 115, 55.3 ±4.3 yrs) and non-users (n = 52, 57.5 ± 4.7 yrs). Factor analyses were used to identify groups of women who had similar patterns of 4-year BMD changes from: those who did not train (Ctrl, n = 23) and those who did train (EX, n = 52) for 4years of the study. Those with increasing BMD patterns were denned as Rs and those with decreasing BMD patterns were denned as NRs. Adjusted 4-year BMD changes were calculated from the above regressions. Independent t-tests were used to determine whether R and NR differed in their 4-year BMD changes at the spine, trochanter and femoral neck. RESULTS: For the Ctrls, the mean differences between R and NR for the spine, trochanter and neck BMD changes were 0.09, 0.03, and 0.08 g/cm2 (p = 0.00) prior to adjustment and were 0.02, 0.01, and 0.02 g/cm2 (p > 0.05) after adjustment. For the EXs, the mean differences between R and NR for the spine, trochanter and neck BMD changes were 0.13, 0.08, and 0.08 g/cm2 (p = 0.00) prior to adjustment and were 0.02 (p = 0.05), 0.02 (p = 0.003), and 0.01 g/cm2 (p = 0.16) after adjustment. CONCLUSIONS: EX NRs and Ctrl group Rs existed. Age, EX Freq, CA and changes in BC accounted for the differences between R and NR in all cases except for the trochanter and spine BMD changes in EXs where mean differences were reduced but not eliminated.