BioEd Online

A prize-winning analysis that overturned accepted facts about contraception and health is under fire. The organization that owns the data used in the study is warning women not to believe that contraceptive pills can protect against heart disease and stroke.

The study in question, by Rahi Victory and colleagues from Wayne State University in Detroit, Michigan, was originally presented at October's annual meeting of the American Society for Reproductive Medicine (ASRM). It won the Society for Reproductive Endocrinology and Infertility prize and garnered a great deal of media attention (See " Oral contraceptive may cut risk of heart disease"). It was published in abstract form in Fertility & Sterility, the magazine of the ASRM¹.

The abstract uses data from the Women's Health Initiative (WHI), a long-running US study, which famously showed that women who take the hormones oestrogen and progesterone after menopause are more likely to get cardiovascular diseases. Using the same data set, Victory's study found that women who had, at some point, taken such hormones in the form of the contraceptive pill were actually less likely to get a cardiovascular disease.

The unexpected result was widely reported, but when officials at the WHI learned to what use their data had been put, they were sceptical. "The findings seemed implausible to those of us who have read the literature," says Ross Prentice, a biostatistician at the WHI.

Removing bias

"The WHI data is not well suited to their purpose because of a range of potential biases," says Prentice about Victory and his colleagues' work. For example, many of the older women in the study became menopausal before the pill was introduced, and never got a chance to take it. So the subset of women who never took the pill is generally a bit older, and therefore more likely to have developed cardiovascular trouble.

What's more, all the data Victory used was from when women first joined the WHI trial, and so is composed of their own memories about contraceptive use and disease, unverified by medical files. Even women who had cardiovascular disease before they began taking the pill were counted, according to the WHI.

The WHI is worried about being associated with the Victory paper. "These results have been used to impugn the results of the hormone trial," says Jacques Rossouw, a project officer at the WHI. "But it's chalk and cheese. The WHI hormone trial was a prospective controlled trial [whereas the other was not]."

The WHI will send a letter questioning the abstract to Fertility and Sterility in the next week or two. However, a spokesman for the ASRM says that there are no plans to retract the abstract or withdraw the prize.

A carefully worded statement was issued earlier this month by the WHI and Wayne State University, characterizing Victory's work as "exploratory" and promising further analysis. The school is standing by their researchers: John Oliver, the vice-president for research, says that the work shouldn't be dismissed just yet. "This was a preliminary study, and it needs to be looked at very carefully in considerably more detail."