Mammogram data unlikely to change screening policy: overview analysis concludes that mammograms don't increase life expectancy

OB/GYN News, Dec 1, 2001 by Mitchel L. Zoler

Mammography screening policies in the United States are unlikely to change despite a new overview analysis that concluded that mammography has not proved to cut overall mortality risk.

Even if the new analysis were correct and mammography did not increase the overall life expectancy of women, other study results indicate that mammography is able to cut the risk of death from breast cancer, several experts told this newspaper.

"The data are solid that screening mammography saves lives," said Robert Smith, Ph.D., director of cancer screening for the American Cancer Society in Atlanta.

"If you combine the results of all studies, you probably need to screen about 800 women aged 50 or older every year over 10 years to cut the rate of breast cancer deaths by 1," said Dr. Barnett S. Kramer, a senior scientist at the National Cancer Institute in Bethesda, Md.

The new analysis was a revision of a controversial report published last year by a pair of researchers at the Nordic Cochrane Centre in Copenhagen. Both last year's report and the new revision carefully reviewed seven published, randomized trials that assessed the efficacy of screening mammography. The more recent version provided a fuller analysis of the data.

Both analyses found flaws in several of these trials. The overviews focused on the proven ability of mammography to reduce total mortality rate because this is the only "reliable" measure of benefit, wrote Dr. Ole Olsen and Dr. Peter C. Gotzche (Lancet 358[9290]:1340-42, 2001). Studies that focused on breast cancer deaths missed tallying fatal cases that were misclassified or that were triggered by cancer treatment such as radiotherapy, they wrote in their latest article, which was published in late October.

An accompanying editorial written by Dr. Richard Horton, editor of the Lancet, praised the "rigorous and well-developed method" for conducting systematic literature reviews that is used by researchers in the Cochrane Collaboration. Based on the new review, "there is no reliable evidence from large, randomized trials to support screening mammography," Dr. Horton concluded (Lancet 358[9290]1284-85, 2001).

But several U.S. experts disagreed. Dr. Smith cited a recent analysis that he did of data collected over a period of 29 years in Sweden by the Two-County study This analysis showed that mammography cut the risk of women dying of breast cancer by 50%.

"In the United States, there is an overwhelming feeling among experts that mammography is effective," said Dr. Paul S. Frame, a family physician at the University of Rochester (N.Y.), and a member of the U.S. Preventive Services Task Force. Recommendations for annual mammography screening in women aged 50 and older are in place from at least four U.S. groups: the U.S. Preventive Services Task Force, the American Cancer Society the National Cancer Institute, and the American College of Radiology Some of these groups also call for regular mammography among women aged 40 and older. None of these groups revised their recommendations following the first overview report from Dr. Olsen and Dr. Gotzsche last year, and no one from these groups contacted by this newspaper anticipated revisions in response to the October report.

Critics of the Cochrane analysis took issue with its focus on overall mortality rather than on deaths due to breast cancer. "It's true that the ultimate goal of mammography is to prolong the life expectancy of women. But none of the trials were designed to detect this because breast cancer accounts for a small proportion of total deaths among women," Dr. Kramer said. "What they proposed is not practical."

By using total mortality differences as its benchmark, the Cochrane analysis discounted the results of five trials, including the Swedish Two-County study and focused primarily on the results of two trials, including one done in Canada. U.S. experts also thought that this was misguided.

"The Canadian study was the worst, in terms of the quality of the mammography used. The Two-County study was the best," said Dr. Stephen A. Feig, director of breast imaging at Mount Sinai Medical Center in New York.