Op-Ed

More thoughts about Mammograms

November 2010 -- Mammography screening to detect breast cancer early enough to reduce death rates significantly is trashed in a September 23 issue of The New England Journal of Medicine.  The conclusion of this study of breast cancer deaths in the nineteen counties of Norway is that the clear reduction in likelihood of deaths from breast cancer that has occurred in the last thirty years is due mainly to better treatment; according to this study, mammogram screening played only a small, and perhaps virtually no, role in the reduction in deaths.

This is a geographic study, comparing counties in which mammogram screening was offered with those in which no such screening was offered.  Ordinarily, epidemiologists urge that the results of geographic studies not be over-interpreted, since such studies are rarely definitive.  But, the conclusions of this study have been almost uncritically applauded; one consequence is that women are being told they really do not have to get annual mammograms.

Although the conclusions have been greeted with enthusiasm with the assertion that there are, at worst, only minor flaws, there are, in fact, major flaws and questions.  There is absolutely no reason for women to abandon regular mammograms on the basis of this study.

Here are some of the issues.

1.  The senior investigator published a study one year ago, based on the same women, in which she found virtually the same reduction in deaths and concluded that roughly 60 percent of the substantial benefit was due to mammogram screening.  Same population, same cases, same deaths, different analysis, and strikingly different conclusions.

2.  The analysis is based on those offered mammograms.  It does not compare women who actually had regular mammograms with women who did not.  The authors say 77 percent of women offered mammograms complied and had the test, but we are given no details.  We do not know what percentage of women got a mammogram, but did not follow the recommended every two-year mammogram schedule.  And, we do not know what percentage of women not offered screening actually had mammograms.  All that will seriously bias the results against screening.  If a women was offered screening, never had a mammogram, developed breast cancer and died, she was considered a screening failure – even if she never had a mammogram.

3.  The interval between mammograms was two years.  In the United States, the recommendation is a yearly mammogram.  The two-year interval would be expected to reduce the mammogram benefit by increasing the percentage of women found on screening to have advanced stages of breast cancer that will kill them.

4.  The average followup was only 2.2 years; the longest any woman was followed was 8.9 years.  Since averages are pulled up by those with the longest followup, that 2.2 year average guarantees that the majority of screened women were followed for less than two years.  With breast cancer, that is a grossly inadequate followup and undercuts the whole study and its conclusions.

5.  Last, and perhaps most important, when the results were analyzed by stage of the breast cancer when diagnosed among those offered screening and those not offered screening, the findings undercut the negative inclusions.  No benefit was found for screening among those with very early cancer (Stage I), that is, no differences in death rates.  That is not surprising.  With very early breast cancer, no differences could possibly be found with such a short followup.  Similarly, those with advanced cancer (Stage III) did not benefit from screening; of course, they would not benefit; screening is done to detect early, not advanced, cancer.  That leaves Stage II, still early stage.  Those found to have Stage II tumors on screening showed a highly significant reduction in deaths.  That result strongly suggests screening was effective in detecting breast cancer early and reducing deaths.

The bottom line is that this study is full of flaws and limitations.  It must not be used to tell women they really do not need regular mammograms.  The evidence still indicates women should have annual mammograms starting at age 40.

Annual mammograms do save lives – and that is the real bottom line.

Op-Ed Archives

October 2010: Malnutrition can make invading microbes more venomous

September 2010: Are we losing faith in the future?

August 2010: Extraordinary Longevity: New Technologies

July 2010: Drug Abuse Policy

July 2010: Omega-3 Fatty Acids

June 2010: More on mammograms

May 2010: The Dumb New Recommendations for Breast Cancert Screening

March 2010: The Pill to End All Pills - True, Fals or In Between

December 2009: Is Pre-paying Medicare the way to go?

Nov. 30, 2009: Teach Young People to Rethink