Skip navigation

Bone mass and exercise in women


Bandolier 62 examined a systematic review [1] of how exercise affects bone mass in postmenopausal women. It was disappointing, with only six studies and only two of those randomised. Another, better, review has now been published, with many more studies, much more data, and with comprehensible and useful results [2].

Searching


The new review had a comprehensive search strategy to identify randomised and non-randomised controlled trials looking at the effects of training programmes on bone mass. Journals were also hand searched.

Included studies had to meet several criteria:


Outcome


A study treatment effect was calculated, which was the difference between the percentage change in bone mass in one year in the training group minus the percentage change in bone mass in one year in the control group. A positive figure indicates a protective effect of exercise.

Results


There were 34 randomised comparisons and 19 non-randomised, beautifully laid out in tables in the paper, in training programmes ranging from six to 24 months.

The effects of training on bone mass at the lumbar spine in 552 postmenopausal women are shown in Figure 1. The overall treatment effect was a one-year percentage difference due to training of 0.79% for endurance and strength training programmes. For 204 premenopausal women there was a 0.91% benefit.

Figure 1: Lumbar spine - postmenopausal

Legend: Difference between the percentage change in bone mass in one year in training group minus the percentage change in bone mass in one year in control group. Open circles endurance training, filled circles strength training.
The effects of training on bone mass at the femoral neck in 409 postmenopausal women are shown in Figure 2. The overall treatment effect was a one-year percentage difference due to training of 0.89% for endurance and strength training programmes. For 174 premenopausal women there was a 0.90% benefit.

Figure 2: Femoral neck - postmenopausal

Legend: Difference between the percentage change in bone mass in one year in training group minus the percentage change in bone mass in one year in control group. Open circles endurance training, filled circles strength training.
In both cases, non-randomised studies produced an estimate that was roughly twice as large as for randomised studies. Data from non-random studies are not included here.

Comment


The bottom line from the new review is that exercise training programmes prevented or reversed bone loss of almost 1% per year compared with the controls. The effects were consistent for the lumbar spine and the femoral neck. The variability in the results was seen mostly in the smaller trials, and the larger trials showed results consistent with the overall effect (Figures 1 and 2).

The notable feature is that two systematic reviews published within about a year of each other had such a different clutch of randomised trials. In part this reflects the later cut-off point for searching, with several studies published in the mid-1990s that were likely to have been missed by the earlier review. In part it reflects a much more comprehensive search strategy.

The result is important - yet another benefit of exercise for womens' health. Most of the exercises were somewhat more vigorous than a brisk walk and included treadmill walking and running, and some resistance and back strengthening exercise or aerobics, for instance. Taken together, these are also a useful teaching aid for critical appraisal.

There is also a cautionary tale here for reading systematic reviews. If the amount of information found is small, perhaps with a few trials and tiny numbers, then the chances of the review being correct is diminished.

References:

  1. GA Kelley. Aerobic exercise and bone density at the hip in postmenopausal women: a meta-analysis. Preventive Medicine 1998 27: 798-807.
  2. I Wolff et al. The effect of exercise training programs on bone mass: a meta-analysis of published controlled trials in pre- and postmenopausal women. Osteoporosis International 1999 9: 1-12.
previous or next story in this issue