Letters Section Editor: Robert M. Golub,
MD, Senior Editor.
In Reply: We appreciate the responses of Drs
Alekel, Sirtori, and Muanza and colleagues to our study. Regarding Alekel’s
methodological concerns, although the general assumption was that isoflavones
are as effective as hormone therapy, for BMD power calculations we used data
previously reported from an intervention that was quite similar to ours for
an even shorter duration that included smaller groups of women.1 When
we calculated that 200 participants were needed for this study, this calculation
took into account a dropout rate of 25% for all women, for an eventual number
of 75 participants in each intervention group, which is the number needed
for the secondary per protocol analysis. The sample-size calculations were
2-sided and in agreement with the above assumptions. Our statistical adjustment
for smoking history included current smoking and did not change the results.
Additional adjustment for the use of antihypertensive medication also did
not change our findings. As we noted in our discussion, we agree that the
timing of the intervention in relation to onset of menopause might be critically
important. Although the exclusion of 12 scans does decrease power, our modified
intention-to-treat analysis still reached the pre-calculated numbers of 75
scans per treatment group. Finally, compliance, defined as use of more than
80% of supplements, was 96% in the soy group and 94% in the placebo group.
van der Schouw YT, Grobbee DE, Kok L, Kreijkamp-Kaspers S. Isoflavones and Postmenopausal Women—Reply. JAMA. 2004;292(19):2338-2339. doi:10.1001/jama.292.19.2337-b