Analysis of Mammography Trials Renews Debate on Mortality Reduction

A recent Cochrane review and a related report in the Lancet have reignited the controversy about screening mammography. Based on a major review of clinical trials addressing the issue, the conclusions of Lancet authors Ole Olsen and Peter C. Gøtzsche, M.D., were strikingly unequivocal: “In 2000, we reported that there is no available evidence that screening for breast cancer reduces mortality. As we discuss here, a Cochrane review has now confirmed and strengthened our previous findings.” The accompanying editorial by Richard Horton, M.D., the journal’s editor, was as explicit and more succinct: “At present, there is no reliable evidence from large randomised trials to support screening mammography programmes.” The Cochrane review, which was released last month, was also authored by Olsen and Gøtzsche, of the Nordic Cochrane Center, Copenhagen. The two reports had similar conclusions. Olsen and Gøtzsche rated seven major randomized controlled trials of screening mammography—and found them all wanting. The Malmö (1976) and Canada (1980) studies were rated as medium quality; three other Swedish studies—the Two-County (1977), Stockholm (1981), and Göteborg (1982) studies—as poor quality; and the New York (1963) and Edinburgh (1978) studies as flawed. Gøtzsche and Olsen’s condemnation of screening mammography is largely based on their conclusion that the two best studies provide no statistically significant evidence for a decrease in breast cancer mortality, although the studies they rate as weaker showed decreased breast cancer mortality but not a decreased overall mortality. They suggest that overall mortality is a stronger measure of the impact of mammography because “breast cancer mortality is an unreliable outcome that is biased in favor of screening.” The authors also contend that screening leads to more aggressive treatment; they suggest that it increases the number of mastectomies by about 20%. Kathleen Pritchard, M.D., head of clinical trials and epidemiology at Toronto-Sunnybrook Regional Cancer Center, Toronto, and a member of the Cochrane Breast Cancer Group that reviewed the report, said she agrees with some of the Gøtzsche and Olsen’s points but not with their overall conclusion. “Personally, I feel that the authors had an agenda,” she said. “The trials they describe as being the best all have problems.” She noted that all of the major trials on mammography have flaws, but it is not likely that there will be more trials that address the issue. She added that she thinks that screening (particularly mammography), early diagnosis, and better systemic adjuvant therapy all play roles in decreasing mortality from breast cancer. Experts unconnected with the Cochrane Collaboration also question Gøtzsche and Olsen’s analysis. For example, the study authors tracked all-cause mortality, which, while appropriate in a therapeutic trial, has little or no place in a screening trial, said Robert Smith, Ph.D., director of cancer screening for the American Cancer Society. Bringing in overall mortality as a factor, he said, “is a bit of a red herring.” In addition, he said, the authors were prompted toward such an analysis by a 1999 Swedish study that reviewed breast cancer deaths since 1990, but Smith called that study “rife with methodological blunders.” For example, he said, it mixed screened and unscreened cohorts and presumably counted deaths that would not have been prevented by mammography. William Gradishar, M.D., chair of the American Society of Clinical Oncology’s Cancer Communications Committee and director of the breast cancer program at Northwestern University, Chicago, said there’s “a great deal of heterogeneity in how the trials were conducted,” making it problematic to group them for analysis. Anthony Miller, M.D., of the Division of Clinical Epidemiology at the German Cancer Research Center, Heidelberg, disagreed with Gøtzsche and Olsen’s “damning of the HIP [New York] trial,” but agreed that “a lot needs answering” in the Two-County study, where they expressed concerns about randomization of subjects. “Gøtzsche and Olsen have done a useful service by raising these issues,” especially in reference to the Swedish trials, he said. So where, if anywhere, is future research on the efficacy of screening mammography going? Gradishar noted that researchers are caught in a dilemma. The variations in the existing studies, such as how patients were followed and the equipment used, plus the fact that the raw data are often unavailable, make it difficult to mine data from the earlier research. The corollary is that, to readdress this issue, tens of thousands to hundreds of thousands of women would have to be recruited, placed under the same protocol, and screened using the same equipment, said Gradishar. “This would be a trial of enormous magnitude.” Others suggest that the development of new breast-imaging techniques might someday render debate on current mammography screening moot. The same question still lingers: Will these results change clinical practice? “We acknowledge that there are probably truths on both sides,” Gradishar said, yet “screening mammography remains our best tool for identifying abnormalities in the breast at an earlier time.” Dr. Kathleen Pritchard Dr. William Gradishar Dr. Anthony Miller

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.