Two research reports on politically sensitive issues this week -- one on abortion and breast cancer, the other on Persian Gulf war illnesses -- offer a cogent reminder that scientific studies differ widely in credibility and authority. The abortion study, which largely disproved the notion that abortions cause breast cancer, was the most authoritative research yet on the issue. It surveyed a very large number of women and used a research design that virtually eliminated the main cause of bias in previous studies.
By contrast, the gulf war illness study, which suggested that some gulf veterans may have suffered neurological impairments from exposure to multiple chemicals, was only preliminary and suggestive. It covered a very small group of veterans and its results may have been distorted by selective participation.
The abortion study, performed in Denmark, should go a long way toward resolving an issue that has been clouded in controversy for more than a decade. As long ago as 1980, some scientists theorized that women who abort their pregnancies may be left with vulnerable breast cells that, deprived of hormonal changes in late pregnancy, are prone to becoming cancerous. But dozens of studies of this possibility showed conflicting results.
Unfortunately, virtually all these studies were marred by likely reporting bias because they relied on women to tell the truth about whether they had had an abortion. The studies typically compared the abortion histories of a group of women who had breast cancer with a group of comparable women who did not. The breast cancer victims, as subsequent research showed, were far more apt to admit they had had an abortion, presumably because they wanted to give the doctor all relevant facts that might help in their treatment. The other women were less honest, presumably because they considered the abortion embarrassing.
The Danish study got around this inevitable bias by relying on official records rather than the testimony of the women. Abortion has been legal in Denmark since 1973 and mandatory registries are kept of births, cancer cases and abortions. The Danish researchers examined the records for 1.5 million women, of whom 280,000 had had abortions, some more than once. Over all, these women were no more likely to develop breast cancer than women who had never had an abortion.
The only uncertainty was a suggestion that women who had abortions in the second or third trimester did have an increased risk of breast cancer, but the number of women in this category was too small to warrant firm conclusions. The Danish work has also been challenged on the ground that even the registry does not fully eliminate bias because women with recent abortions have not had time to develop cancer and older women may have had abortions before the registry started. Further research is needed. But for now, this study shows that women need not shun a first-trimester abortion for fear of developing breast cancer.
The other study, involving 249 gulf war veterans, was more suggestive than authoritative. Researchers at the University of Texas Southwestern Medical Center surveyed veterans of a Navy seabees unit about their health, asked them about possible exposures to chemicals, and performed neurological tests on 23 ill veterans and 20 healthy controls. They found subtle differences in neurological function and concluded that some of the impairment may have been related to pesticides, chemical warfare agents or anti-chemical medications.
But the small size of this study and the possibility that only the sickest veterans participated undermines its authority. Research on multiple chemical exposures in the gulf war is in the early stages and will require more comprehensive follow-up, perhaps comparable to the Danish abortion study.
Copyright 1997 The New York Times Company