New study questioning screening mammography’s value stirs up opposition

A new study questioning the efficacy of screening mammography has stirred up an international debate with researchers from around the world challenging its scientific validity.

The study, published in the British Medical Journal, compares breast cancer death rates in screened and unscreened regions of Denmark and finds no real difference.

Written by Karsten Jorgensen and Peter Gotzsche from the Nordic Cochrane Centre in Copenhagen, the study is based on data collected from two public screening programs (Copenhagen and Funen County) and regions across the rest of Denmark that have no screening program. This is not the first time the two have published a study criticizing mammography. Last year they said one in three breast cancers is overdiagnosed.

In the current study, among women who could benefit from screening (which they defined as those 55 to 74 years old), they found breast cancer mortality declined by 1% per year in the screened areas and by 2% per year in the nonscreened areas. In women they deemed too young to benefit from screening (35 to 54 years), breast cancer mortality declined by 5% per year in the screened areas and by 6% per year in the nonscreened areas during the same period.

For the older age groups (75 to 84 years), there was little change over time in either screened or nonscreened areas.

Their analysis covered 10 years after screening could have had an effect on breast cancer mortality. For comparison, they also looked at the 10-year period before screening was introduced.

“We were unable to find an effect of the Danish screening program on breast cancer mortality,” the authors said. “The reductions in breast cancer mortality we observed in screening regions were similar or less than those in nonscreened areas and in younger age groups, and are more likely explained by changes in risk factors and improved treatment than by screening mammography.”

The findings are similar to results found in other countries with nationally organized programs, the authors said.

“We believe it is time to question whether screening has delivered the promised effect on breast cancer mortality,” they said.

The article, published online, has already garnered a slew of responses, none in support of the study.

The authors’ claim that mammography screening has no impact on breast cancer mortality in Denmark is unsubstantiated because the authors used very crude data, and because the analysis was not geared to answer the question, said Elsebeth Lynge, Ph.D., a professor at the University of Copenhagen.

“First, breast cancer screening can only possibly have an effect on women not already diagnosed with breast cancer prior to screening,” Lynge said. “Furthermore, they merge data from three screening areas starting screening at different points in time, and used age groups instead of cohorts. Together this gave quite ‘polluted’ data.”

Dr. Daniel Kopans, a professor of radiology at Harvard Medical School and a staunch screening advocate, also challenged the study.

“The recent article in the British Medical Journal by Jorgensen et al is another example of the failure of peer review that permitted the publication of a methodologically unsupportable analysis,” he said.

Like Lynge, Kopans said the data were diluted by mixing women who were not offered screening with those who were. In addition, previous data has shown mortality reductions of as high as 44% when using screening mammography, he said.

Robert A. Smith, Ph.D., director of cancer screening for the American Cancer Society in Atlanta, said there are multiple failures to actually compare screened and unscreened groups.

“In the screening period, the authors do not distinguish breast cancer deaths attributable to diagnosis before screening was introduced and breast cancer deaths that were due to diagnoses after the initiation of screening,” he said. “Experience has shown that more than half of all breast cancer deaths occurring within a 10-year period of observation are associated with diagnoses that occurred before the period began.”

Responding to the critiques, Jorgensen and Gotzsche said, “Lynge et al spread a smokescreen, using irrelevant arguments. We believe they simply do not wish to see our results for what they clearly are.”

If there had been an effect of screening, they would have expected to see a widening gap in mortality rates over time between the screened and the nonscreened areas, but the two curves were parallel in the time period, they said.

Of the comments from Kopans, Jorgensen and Gotzsche said, “Daniel Kopans offers much criticism but little substance. He is cherry-picking when he refers to the effect of mammography screening as being ‘as high as 44%.’ The best trials showed little or no effect, and the estimate from independent researchers is a 15% effect.”

The comments in full can be read at BMJ’s website.