Thirty-nine RCTs were included (see Results of the Review for the number of women used in the analysis of each drug class).
The results focused on treatment effects at 2 years.
There were no apparent differences in changes in lumbar spine BMD between the various predetermined treatment groups. Almost all treatment regimens at least maintained or increased lumbar spine and hip BMD at 2 years. The results were similar for analyses adjusting for sample size and drop-outs.
Tibolone appeared to be as effective as any other estrogen regimen (ANOVA P=0.944).
Change in BMD at lumbar spine at 2 years. Analyses weighted by the ratio of sample size to drop-outs.
Conjugated equine estrogens (11 RCTs, more than 1,927 women): the mean change in BMD was 0.076 (95% CI: 0.062, 0.091).
Estradiol, estradiol valerate, esterified estrogens, estrone sulphate and estriol (11 RCTs, more than 2,004 women): the mean change in BMD was 0.072 (95% CI: 0.055, 0.089).
Transdermal estrogens (7 RCTs, 1,047 women): the mean change in BMD was 0.075 (95% CI: 0.063, 0.088).
Tibolone (5 RCTs, 1,150 women): the mean change in BMD was 0.064 (95% CI: 0.030, 0.099).
Ethinyl estradiol (1 RCT, 1,265 women): the mean change in BMD was 0.078 (95% CI: 0.047, 0.110).
There were insufficient data to compare the effects of these classes of pHT on femoral neck or hip BMD, or to compare the effects of adding progestins to a specific estrogen.
There appeared to be a trend towards reduced BMD change with increasing age, but this was not statistically significant (P=0.0962).