Editor's Choice

Abstract

The trial of moxibustion, the stimulation of an acupuncture point with heat from a herbal stick, (pp. 743–747) caused much editorial debate. Not only was the researchers' planned sample size of 260 probably too small to reliably tell us whether it helps turn breech babies, but the trial was stopped halfway, without a proper stopping rule, when no effect was seen. Some editors therefore recommended rejection. However, a previous widely-publicised small trial had apparently shown that moxibustion was effective, a finding which this Western sceptic, for one, found implausible. Other editors therefore argued that even this small trial` should go into the public domain. Two studies emphasise the difficulties of diagnosing and treating this important problem. On pages 737–742 some new tests for pre-term delivery turn out to have only modest predictive ability. In a logistic regression model derived from the original dataset they add a little information, but whether they would do so in fresh datasets remains uncertain. On pages 725–730 a group from Queen Charlotte's and St Thomas' Hospitals report a randomised trial of rofecoxib, a cox-2 inhibitor, for preventing pre-term labour. There were good reasons to believe that it might have been effective, but it's not. Rofecoxib does not postpone delivery; the trend was in the opposite direction. Three years ago Valerie Beral and her colleagues summarised the evidence from the Women's Health Initiative (WHI) trials of hormone replacement therapy (HRT) in an important Lancet paper.1 In this issue of BJOG they extend their analysis to include the recent estrogen-only WHI results as well as some other smaller trials. The message is timely since no other major trials are ongoing. Both types of HRT reduce fractures, and increase stroke and pulmonary embolus. Combined therapy seems to increase breast cancer and probably coronary heart disease as well, although it also protects against colorectal cancer. On pages 689–691 Mary Lumsden discusses other possible interpretations of the data, such as the suggestion that the cardiovascular harm may be less pronounced at younger ages, puts the risks in context, and gives the absolute risk estimates necessary for sensible decision making. She concludes that HRT should not usually be used for prophylaxis, and that for treatment of symptoms it should be used at minimum dose and for the minimum time. In the remainder of this large issue-the next six each have forty extra pages as we deal with a backlog of accepted papers-is a trial of amnio-infusion (pp. 759–763), a paper putting risk estimation for pre-eclampsia on the same scientific footing as Down's syndrome screening (pp. 703–709), and a debate about the ethics of the various ongoing vitamin E trials in pregnancy (pp. 684–688). However, I am writing this on April 25th and standing for parliament next week, so perhaps readers will forgive a short Editors Choice, and read the details for themselves.