Swedes doubt mammography trial Disparities found in landmark study By John Crewdson Tribune senior correspondent March 15, 2002 Swedish mammography researchers for the first time are questioning the validity of that country's historic "Two County" mammography trial, the study largely responsible for the popularity of mammography in Europe and the United States. The questions are raised in an updated overview of five other long-running Swedish studies that, more than 20 years later, collectively represent the best source of information on whether mammography saves lives. The overview, which does not include the most successful part of the Two County Trial because of its director's refusal to participate, concludes on the basis of the remaining studies that regular mammograms can reduce a woman's relative risk of dying from breast cancer by 20 percent. The report by epidemiologist Lennarth Nystrom of Umea University and other Swedish researchers is to be published Saturday in The Lancet, a British medical journal. That the Swedish overview revealed the same relative benefit as an analysis by the U.S. government in February should settle the question of what the mammography studies show. The 20 percent figure is a third lower than the 30 percent cited by mammography advocates for more than a decade. The newer number, described as "modest" in an accompanying editorial to be published in The Lancet, is the equivalent of preventing seven deaths per year in a female population of 250,000. In addition, the Lancet editorial cautions that "the latest analysis does not tell us whether the massive effort to develop screening programs in at least 22 countries and to encourage participation is worthwhile." "That issue," it says, "must be honestly confronted by those organizations and individuals with an interest in maintaining programs for mammography screening." The question facing health-care policymakers is not whether mammography prevents some breast cancer deaths--it does--but whether it saves enough lives to justify the economic and psychological costs, including false-positive results and the treatment of women with harmless tumors. The Lancet editorial acknowledges that the overview "cannot address these issues." No clear answer from study Despite having monitored nearly 250,000 Swedish women for more than 15 years, the studies that were reanalyzed for the overview still provide no clear answer to the central question of whether regular mammograms improve an individual woman's chances of living longer even if she avoids death from breast cancer. Among women invited to have mammograms, there were 2 percent fewer deaths from any cause than among the women who were not invited and who served as a control group. But that number is not statistically significant, meaning that it could be the result of chance rather than mammography. In addition, not all of the women invited to have mammograms responded to the invitations. The overview, which includes studies begun in the late 1970s and early 1980s in the Swedish cities of Malmo, Stockholm, Goteborg and the county of Ostergotland, also found that the benefits of mammography do not appear to begin until age 55, rather than 50 or even 40, as many mammography advocates have believed. Only last month, Health and Human Services Secretary Tommy Thompson responded to the controversy over the effectiveness of mammography by encouraging women over 40 to seek mammograms every one or two years. Thompson also cited a 20 percent relative benefit in favor of mammography. The overview's findings represent a glaring departure from the Two County Trial's dramatic announcement in 1985 that only two or three mammograms per woman over six years had cut breast cancer deaths by up to 40 percent for women over 50 in the Swedish counties of Kopparberg and Ostergotland. Using methods that have never been made completely clear, the female residents of the two counties were divided into two groups, one of which received mailed invitations every two or three years to have a free mammogram. It was on the basis of the landmark Two County findings that Sweden and several other European countries with national health-care programs began offering regular mammograms to all women over 40 at a nominal cost. Although the U.S. does not have a national mammography program, Medicare and private health insurance plans will pay an estimated $3 billion this year to provide some 33 million American women with mammograms. Although inconsistencies in the Two County Trial that could account for the inflated expectations it raised have been the subject of private speculation among researchers for years, they have not found their way into the public debate until recently. In an unusually pointed comment for a scholarly article, the overview states that "the Kopparberg part of the Two County Trial was not available for this overview . . . due to a decision not to continue with the collaboration with the Swedish collaborative group by the Kopparberg trialists . . . " The overview adds that "We regret this decision." Doctor tried to stop paper Dr. Laszlo Tabar, the director of the Kopparberg study and lead author on nearly a dozen publications reporting the Two County Trial's findings, did not respond to an e-mailed request for comment. One of the overview's co-authors, Dr. Ingvar Andersson of Malmo University, said in a telephone interview Thursday that Tabar requested some weeks ago "that we withdraw the paper" before publication, presumably because of the criticisms it contains. "I immediately said there is no way that we negotiate in any way like this," Andersson said, "or that this suggestion is even going to be discussed." Some of the discrepancies in Tabar's publications were first reported two years ago by a pair of Danish researchers who had been asked to evaluate the Swedish trials by a Danish government agency that was considering whether to recommend a national mammography program in that country. Two months ago, an article in the Journal of the National Cancer Institute suggested that some women who died of breast cancer in the Two County study had been incorrectly recorded as dying of something else, based on the observation that Two County women who received mammograms had a slightly higher chance of dying from any cause. The disparities noted by the Danish researchers, led by Dr. Peter Gotzsche, included the fact that Tabar had reported different totals over the years for the number of women enrolled in the Ostergotland part of the Two County Trial. The explanation, according to the overview, is that Tabar's publications before 1989, including his landmark 1985 article, include more than 1,000 Ostergotland women who developed breast cancer before the study began and should never have been included. Because of Tabar's refusal to cooperate, the current overview does not comment on the disappearance of an additional 750 women from published reports of the Kopparberg portion of the Two County Trial after 1989. Official numbers differ Perhaps the most damaging observation is the discrepancy between the mammography benefit recently reported by Tabar for Ostergotland County and the benefit found by the overview committee using breast cancer and death records compiled by the Swedish government. Those records are reputed to be the most precise in the world, by virtue of the fact that every Swede is assigned a "person number" at birth that allows public health agencies to track her or his lifetime progress through the medical system. The overview notes that, in a published report two years ago, Tabar declared the average breast cancer death benefit for Ostergotland women aged 50-70 was 30 percent after 20 years of mammography. Using government statistics, the overview calculated an average benefit of only 17 percent for the same women during the same period. That discrepancy, the overview's authors suggest, may have been due to "the fact that the Two County team was not blinded for the screening status of the deceased women." "It's a strong statement," said one of the overview's co-authors, Dr. Nils Bjurstam of Goteborg University. A central tenet of any clinical trial is that, to avoid unconscious bias, those who determine the outcome of a study should be "blind," or not aware of whether a particular patient was among those treated or one of the untreated "controls"--in this case, whether a woman's breast cancer had been detected by mammography or not. Although Tabar was directly involved only with the Kopparberg portion of the Two County Trial, he has long claimed that the trial should be regarded as a single entity. As lead author on all of the trial's publications, he has assumed the role of the study's spokesman and has become a world-renowned advocate for mammography. Researchers `were not blind' Tabar's counterpart in Ostergotland, Dr. Gunnar Fagerberg, retired in 1993. But in an interview with the Tribune last month at his home in Mellerud, Sweden, Fagerberg confirmed that he and his colleagues knew before deciding their cause of death which women had received mammograms. "We were not blind," he said. "We had all the papers." Dr. Ulf Ljungquist, a retired surgeon who served as part of the Two County Trial committee in Kopparberg that was headed by Tabar, said that he and the others charged with deciding whether a Kopparberg woman had died of breast cancer also knew whether she had been screened. Another member of the Kopparberg committee, Dr. Lars Holmberg, agreed with Ljungquist, adding that the committee members did not check to see whether cases presented by Tabar for a cause-of-death determination included women who should have been excluded because of prior breast cancer or for some other reason. The 1985 report of the Two County Trial, which also appeared in The Lancet, does not reveal either that women with pre-existing breast cancer were included in the study totals or that those making cause-of-death determinations were not "blind" to whether a woman had received mammograms. Andersson said Thursday that the overview group had not intended to suggest that Tabar and his colleagues intentionally had altered the outcome of the Two County Trial, only that their knowledge of which women had received mammograms could have unconsciously biased them in favor of mammography. "Bias doesn't tell you that it's intentional," said Andersson, who presided over a similar mammography study in the Swedish city of Malmo. The Malmo trial, in which more of the women who received mammograms died of breast cancer during the first seven years than women who did not, used a "blinded" committee to determine cause of death. According to Lennarth Nystrom of Umea University, the Swedish biostatistician who headed the overview, his group did not use government statistics to calculate the actual benefit for the Kopparberg portion of the Two County Trial, although it could have done so. "I have always said in the overview group that we should look at that," said Dr. Nils Bjurstam, chief investigator for the Goteborg mammography trial, which also had a blinded death committee. The possibility of making such a calculation for Kopparberg County "was discussed," says Ingvar Andersson. "But because Dr. Tabar didn't want to collaborate with us, we dropped the issue." Copyright (c) 2002, Chicago Tribune Remember we are NOT Doctors and have NO medical training. This site is like an Encylopedia - there are many pages, many links on many topics. Support our work with any size DONATION - see left side of any page - for how to donate. You can help raise awareness of CAM.