Does mammography screening save lives?
September 13, 2012 — Over the past 40 years, many have come to view mammography screening as the “gold standard” for early breast cancer detection. But a number of recent studies have cast doubt on its benefits. Some suggest that decreased breast cancer rates may be more a function of improved treatment than increased screening. Others find that screening may lead to overdiagnosis and overtreatment—and question whether the mortality reduction that may result from screening outweighs the burden of overdiagnosis.
Lars Vatten, adjunct professor in the Department of Epidemiology at Harvard School of Public Health (HSPH) and a professor at the Norwegian University of Science and Technology in Trondheim, Norway, gave an overview of recent studies on mammography’s effectiveness at a September 5, 2012 talk titled “Mammography Screening: A Controversy That Never Ends,” as part of the Department of Epidemiology’s Lunchtime Seminar Series.
Several studies have found links between mammography screening and overdiagnosis or overtreatment of breast cancer, Vatten said. Mammography is a valuable tool for detecting early cancers, but it can’t distinguish which cancers will progress and which won’t. It’s therefore possible for mammography to lead to harm for women in the form of mental distress, biopsies, surgeries, or chemotherapy and hormone treatments for disease that would never have caused symptoms.
In a 2009 British Medical Journal (BMJ) study, Karsten Jørgensen and Peter Gøtzsche, of Denmark’s Nordic Cochrane Centre, Rigshospitalet, looked at breast cancer data from five different countries, both seven years before and seven years after publicly organized mammography screening was introduced. They concluded that 1 in 3 women with screen-detected cancer would not have been harmed by the disease.
The following year, the New England Journal of Medicine published a study by HSPH visiting scientist and Oslo University Hospital surgeon Mette Kalager and colleagues that examined Norwegian mammography data and concluded that for each life saved by mammography screening, between 5 and 15 women would be diagnosed and treated for a cancer that would not have harmed them.
Screening and mortality rates
Discussing countries where breast cancer mortality rates have declined, Vatten listed data from several studies that suggest the decline cannot clearly be attributed to the implementation of mammography screening. For example:
• A 2010 BMJ study by Jørgensen, Gøtzsche, and Per-Henrik Zahl of the Norwegian Institute of Public Health, comparing breast cancer mortality rates in both screened and unscreened areas in Denmark, found that declining rates were slightly stronger in the unscreened than in the screened areas. The authors concluded that the reduced mortality rates may have resulted not from screening but from changes in risk factors or improved treatment.
• Another 2011 BMJ study, led by Philippe Autier of the International Prevention Research Institute in France (Vatten was senior author), compared breast cancer mortality trends in pairs of countries similar with respect to socioeconomics, culture, health care, and access to treatment, but different in the timing of when they implemented mammography screening by at least 10 years. The hypothesis was that reduction in breast cancer mortality would start earlier in countries where screening began earlier. But the results did not support that assumption.
• A 2012 Journal of the National Cancer Institute study, also led by Autier, with Vatten as a co-author, examined declining breast cancer mortality rates in Sweden, looking for links between declines and the implementation of mammography screenings in individual counties. The researchers found no consistent patterns.
There are criticisms of these studies, Vatten noted. Some experts argue that gradual implementation of screening in various countries may complicate research. Others point out that some of the studies focus on whether women in particular counties or areas are offered screenings but don’t include information on which women actually had screenings. And some of the studies, critics say, did not follow women long enough to detect any meaningful effect of screening.
Benefits of treatment
Although mammography’s effects on breast cancer mortality reduction appear murky, Vatten said that the benefits of treatment are much clearer. Citing statistics from a 2010 British Medical Journal study by Autier, himself, and other colleagues, Vatten said that countries where there’s access to modern treatment had the greatest reduction in breast cancer rates between 1989 and 2006—Iceland, England/Wales, and Spain topped the list—and those countries without such access, including Romania, Estonia, and Lithuania, had either minimal reduction in rates or, more troubling, significant increases. Vatten also said that two large overviews published in the Lancet in 1998 showed significant survival benefits for women who receive either chemotheraphy or the drug tamoxifen after breast cancer surgery.
It’s possible that there is a combined downward effect on mortality rates from both treatment and screening, but it’s unclear if studies can accurately distinguish between both effects, Vatten said.
In response to questions about whether or not any of the studies took into account women who were still premenopausal over age 50, or differences in women’s ethnicity, Vatten said they did not. He acknowledged that “the design of many of these studies is not optimal” because they’re not based on individual follow-up, but on aggregated population data, taking into account only mortality rates, women’s ages, and screening dates. In the future, he said, it would be best to conduct cohort studies that follow individual women over time and that include factors such as women’s actual screening history and their menopause status.