Preventive Medicine Column

Dr. David L. Katz

You might expect that a new study out of Sweden, in roughly a million women, suggesting a 26% mortality reduction with routine mammography in women in their 40s, would resolve long-standing controversy about the practice.  But even with these new data, routine mammograms before 50 are, in essence, a 50/50 proposition; a toss-up.  Let’s take a look at both sides of this flip-of-the-coin, and see which way it faces for you.

First, some context for the new study.  Approximately a year ago, the United States Preventive Services Task Force examined all of the evidence then available, and reached the controversial conclusion that routine breast cancer screening should begin at 50, not 40.

This recommendation- from a group that directly influences national policy- did not refute that mammography before age 50 can save lives- everyone knows it can, and has.  It simply compared the potential individual benefit to the costs borne by the whole population.

What costs?  You might think dollars- but that issue was not even considered by the Task Force, which should relieve anyone who thinks their recommendation carried the taint of “rationing.”  Not so; economic analysis is conducted separately from the Task Force evidence reviews.

The relevant costs were tallied as harms, and missed benefits.  In terms of missed benefits, mammograms are harder to interpret in pre-menopausal women- because of breast density- and breast cancer, when it does occur before menopause, often progresses faster than it does later in life.  The result is that even with screening, a significant proportion of pre-menopausal breast cancers will elude detection until they are somewhat advanced.

As for harms, we have long known that for every breast cancer found in women under 50, well over ten times as many women will have false positive results.  Quite a few of these women will have biopsies.  There are, of course, potential complications of the biopsies.  In addition, the scar tissue left behind makes future mammograms harder still to interpret, increasing the risks of missing a cancer when there actually is one.  And though small, the dose of radiation from routine mammograms can actually contribute, slightly, to breast cancer risk over time.

The new study, which ostensibly revisits the controversy, certainly seems compelling at first, as much for its size as its findings.  Researchers took advantage of a natural experiment in Sweden: some counties offered mammograms routinely over a 20-year span, others did not.  Populations were compared among the counties, and mortality was 26% less overall in those counties that screened.

But the study limitations become pretty salient on closer scrutiny.  First, you should immediately think to ask: 26% lower than what?  There were roughly 1200 breast cancer deaths in a half-million women followed for a decade or more without screening; and just over 800 deaths in a comparable group that was screened.   Well over 1,000 women under 40 needed to be screened for ten years before screening saved one life.

Second, the Swedish study reported the mortality benefit of screening, but did not look at its harms.  We must presume the usual high volume of false positive tests and biopsies.

Third, we don’t know from the Swedish study what else may have differed between counties that did and counties that didn’t screen.  Maybe those that did screen also provided better care- and the apparent survival benefit was more about treatment than screening.  Maybe the populations differed in other important ways.  And lastly, even with screening, 74% of the breast cancer deaths that would have occurred without it, occurred anyway.

The official recommendation in the US before this study was neither for, nor against, routine mammography in women under 40; rather, it was for individualized decision making because the risk/benefit trade-off is still like calling a coin toss.

The ultimate solution to the dilemma of a toss-up is, of course, a third choice.  In this case, that would be better screening that does not involve radiation; that is based on a personal risk profile; that reliably finds cancer at its earliest stages when present; that reliably rules out cancer when it is absent; and that distinguishes between cancer that requires treatment and cancer that might go away on its own.  By combining genomic advances with advances in imaging, and housing them together in comprehensive breast care centers, we will get there.  But we’re not there yet.

While waiting, there is no “right” answer for all women in their 40s.  But all you really need is an answer that’s right for you, and that’s what good clinical care is all about.  Confer with your doctor; consider pros and cons; your risk factors and preferences.  And together, make either heads, or tails, of this particular toss-up.




Dr. David L. Katz;