The WHI ET and EPT trials are of the largest and longest randomized controlled trials to examine the affects of postmenopausal HT, providing an opportunity to evaluate the affects of HT on lean body mass and the lean body mass relationship to falls and fractures post-administration of HT. Based on a prior evaluation of the WHI EPT trial (year 3)5
the ongoing preservation of lean body mass via EPT was considered likely and the affect of E alone on lean body mass was hypothesized to match or exceed that of EPT over 6 years.6
We were able to confirm that lean body mass loss was less with either ET or EPT treatment at 3 years compared to placebo, however, despite plausible biological linkages and prior short-term evidence, the early preservation of lean body mass was lost by 6 years and there were no differences between placebo and HT treatment groups. Thus, there was a delay of lean body mass loss with HT, but the benefit did not persist long-term.
Others have also found short term enhancement of lean body mass with HT. In the Bone Estrogen and Strength Training trial, postmenopausal women using and not using HT who were randomly assigned to the no exercise group were followed during the intervention year for changes in body composition. Women in the no exercise, no HT group lost significant lean body mass, whereas women in the no exercise, HT group preserved lean body mass and demonstrated a trend towards gain in lean body mass (p=0.08) over one year. A similar study by Sipila et al,19
also found that lean body mass was enhanced by one year of HT compared to control without exercise. Sorensen et al conducted a short-term crossover study of HT among postmenopausal women which implicated a role for HT in increasing lean body mass.4
These results from smaller trials support the early preservation of lean body mass noted in the WHI report of EPT at year 35
and in ET at year 3 from this analysis.
In contrast, several other 2–3 year trials with various measures of lean body mass following hormone therapy in postmenopausal women have found no effect of HT. Sites et al found no effect of EPT on lean body mass compared to placebo over 2 years.20
Kenny et al found no change in lean body mass over 3 years with ultra-low dose ET.21
The Women’s Health, Osteoporosis, Progestin, Estrogen (HOPE) Trial, a randomized, double-blind, placebo-controlled trial of women (n=749) randomized to various HT regimens (4 doses of ET, 3 doses of EPT, plus 1 placebo group), found no differences between groups for DXA derived changes in lean body mass over 2 years.22
All groups experienced similar, small losses in lean body mass. Finally, Tanko et al found no change in appendicular lean body mass with hormone replacement therapy over 3 years and noted that the trend towards a decrease in appendicular lean body mass in HT groups compared to placebo between years 2 and 3 may suggest a catabolic rather than anabolic effect of HT over time.23
Although not a focus of this manuscript, upon evaluation of the WHI HT intervention cohorts, appendicular lean body mass followed the same pattern as overall lean body mass, i.e. preservation at year 3, but no differences between treatment and control arms at 6 years, and thus is not in agreement with the findings of Tanko et al and does not support a catabolic affect of HT on lean body mass. In agreement with our findings, a 5 year randomized controlled trial of HT, the Danish Osteoporosis Prevention Study (DOPS), found no significant difference between treatment and control for change in lean body mass over 5 years.24
Importantly, the rapid loss of lean body mass seen by others during early menopausal years1
was not evident in these cohorts, regardless of treatment arm. All groups within the WHI BMD sites, regardless of assignment to HT or placebo, lost ≤0.5kg over six years, as opposed to 2–3kg per year in the 3 year randomized placebo controlled trial by Aloia et al.1
Our 3 year results contrast with that of Aloia et al, in that we found that HT was associated with preservation of lean body mass, but by year 6 of our analysis there were no differences between the HT and placebo arms for change in lean body mass. Although the DOPS study in younger women (45–58 years) also demonstrated minimal lean body mass loss over 5 years,24
we may speculate that the relatively long time since menopause (13–22 years) among the women of the WHI HT trials included in the WHI BMD sites and the small degree of change in lean body mass overall during this window of time may have affected our ability to detect a long term effect of HT on lean body mass. Data from women participating in The Third National Health and Nutrition Examination Survey (NHANES III) evaluation of lean body mass by bioelectrical impedance supports a general decline in the pace of lean body mass loss as menopause progresses if we compare by age groups spanning typical menopausal years. The prevalence of normal lean body mass decreased by 20% when comparing age groups 40–49 years to 50–59 years; the prevalence of normal lean body mass declined by another 11% by ages 60–69 years and then remained level until 80+ years.25
The average age of our cohorts at baseline was 63 years. Thus, it is also possible that if investigated earlier in the menopausal years, the affect of HT on lean body mass over 6 years may be more in line with our hypothesis of longer term preservation with HT administration.
In this analysis we also evaluated the contribution of the early preservation of lean body mass with HT (baseline to year 3) to the risk of annual falls and the incidence of fracture in later years. We hypothesized that the reduction in fractures known to accompany HT 10,11
may be partially explained by the preservation of lean body mass and that lean body mass preservation may also contribute to reduced falls. These hypotheses were based on earlier work by Baumgarter et al and Melton et al demonstrating associations between critically low lean body mass, termed sarcopenia, and falls and fractures.26,27
In the presence of HT, change in lean body mass has also been a predictor of BMD, which is related to fracture risk.24
These hypotheses were primarily rejected by results from this analysis. We did not find a relationship between the preservation of lean body mass up to year 3 and subsequent annual falls and fractures between years 3 and 6.
In agreement with the findings of a large observation study, The Study of Osteoporotic Fractures, we found that postmenopausal HT use did not significantly reduce the risk of falling.28
HT was associated with a non-significant risk reduction in both falls and fractures between years 3 and 6. However, there was a suggestion of decreased risk of total falls over 6 years with HT which may occur by an unknown mechanism and is worthy of further investigation.
Although not significant in this sub-sample, the fracture risk reduction was within range of the main WHI EPT 10
and ET trials.11
However, these findings were not explained by early preservation of lean body mass via hormone therapy for falls or for fractures. The WHI Observational Study points towards other soft tissue changes in risk of falling and fractures; they recently reported that the more obese may be at greater risk of falling and site specific fractures,29
which was not evaluated in this substudy of the hormone therapy intervention arms of WHI. In this analysis we found a non-significant 8% increase in risk of falling and 5% increase in total fractures with increased lean body mass, which may be an indication of increased overall body size and would support the recent WHI Observational Study results. When we evaluated change in weight and change in body fat in the respective statistical models, there was no independent effect of weight or body fat on falls or total fractures and the relationships between HT, lean body mass, and falls and fractures were not changed, however site specific fractures were not evaluated.
The present analysis was focused on change in lean body mass which was not predictive of falling and fractures and other long-term trials relating hormone therapy induced changes in lean body mass to risk of falling and fracture could not be located for comparison. However, there is evidence that exercise training can both improve muscle function and reduce risk of falling.30
A recent meta-analysis suggested that hormone therapy may also improve strength,31
which may lead to reduced risk of falling and subsequent fractures. Further research is needed to evaluate strength enhancement by HT and fall and fracture patterns over several years.
The randomized, placebo controlled trial design, relatively large sample size, and lengthy duration (6 years) of the analysis are major strengths of this substudy of the WHI HT trials. This substudy is an important addition to the previously published 3 year analysis of EPT affects on body composition5
which included baseline and year 3 measurements only. Herein we were able to demonstrate that ET alone conferred similar lean body mass benefits at 3 years. This 6 year analysis also allowed us to examine trends in HT affects on lean body mass, and although it may be informative to examine even longer periods of HT treatment, the analysis was limited to the time period in which we had the greatest number of blinded subjects in both the ET and EPT trials.
Importantly, the mean BMI and mean age in the subsample were representative of the entire WHI ET and EPT cohorts by intervention assignments at baseline.10,11
Although other baseline characteristics, such as ethnicity, varied slightly between the WHI cohorts at large and the WHI BMD sites, block randomization occurred by clinical center, such that we would expect the similar results if these analyses were repeated in the entire WHI cohort. Additionally, this study was limited by the small number of fractures for specific sites so the affects of lean body mass on fracture risk at hip, lumbar spine and forearms could not be assessed separately. Our relatively small number of events (falls and fractures) was a limitation of our study and future studies with larger sample sizes and longer follow-up may provide more insight into the relationship between lean mass, falls, and fractures because a greater number of events will be more likely. The use of DXA-derived lean body mass to proxy skeletal muscle mass is also a limitation of this study. Although a good relationship between DXA-derived lean body mass and skeletal muscle mass measured on MRI has already been reported,32
whether DXA is sensitive enough to detect changes in lean body mass is unknown.