Blog | Monday, July 21, 2014

Medical 'reversals'


Twenty-five years ago, the New England Journal of Medicine issued a report on a stunning new medical discovery: Aspirin helps prevent heart attacks.

Yes, good ol’ aspirin. Known since the time of Hippocrates for its magical abilities to quell fever and pain, it took only 2000 years for us to understand the science of it well enough to design a ’sufficiently powered’ double-blind, placebo-controlled randomized trial on aspirin’s efficacy in preventing heart attacks. The Physicians’ Health Study, so named because the study subjects were randomly selected physicians from across the U.S. (whom it was correctly assumed would have higher adherence in swallowing daily pills), addressed the question of whether or not aspirin has true live-saving benefit.

It does. Citing aspirin’s “extreme beneficial effects on non-fatal and fatal myocardial infarction”—doctor speak for heart attacks–the study’s Data Monitoring Board recommended terminating the aspirin portion of the study early (the study also looked at the effects of beta-carotene). In other words, the benefit in preventing heart attacks was so clear at 5 years instead of the planned 12 years of study that it was deemed unethical to continue blinding participants or using placebo.

Confusingly, there is now strong evidence that what’s beneficial for hearts can be harmful in eyes: older people who routinely take aspirin are nearly three times as likely to develop macular degeneration than non-users. Macular degeneration is the leading cause of blindness in Americans 55 and older. It presents a vexing medical problem, in that once it’s discovered, it’s too late to do anything about it. Macular degeneration afflicts more than ten million people in the U.S., which is likely an underestimate of its true prevalence.

What we see in this situation is a classic story arc in modern American medicine: Wonder-drug saves lives. Study stopped. Practice adopted. Then, years later: wonder-drug causes harm. Edicts to stop prescribing it, or be much more selective about recommending it.

Adam Cifu, a doctor (and former colleague) at the University of Chicago, has written about the concept of “reversals” in medicine. In a fascinating paper, Cifu and colleagues catalogued several examples of new knowledge leading to “abandonment” of mainstream medical practices. A major example is hormone replacement therapy for post-menopausal women. For decades estrogen was given to American women as an elixir for many ills. All of the data in support of the practice was observational; outcomes conformed neatly to expectations about the drug. When good science (a randomized, controlled trial called the Women’s Health Initiative) finally challenged the practice, it was almost entirely abandoned.

I asked Cifu for a prediction about the near-blanket recommendation to people of a certain age to take aspirin, which is prescribed so widely, including to many who do not actually have heart disease or elevated cardiovascular risk. In an email, he wrote: The [Physicians’ Health Study] never showed mortality benefits but only benefit to lesser outcomes such as [heart attack]. Even in the earliest studies there was evidence that the benefit was only in limited populations, older ones, and was balanced by significant risk of bleeding (GI [stomach, etc.] and CNS [brain]). I think what has happened is that as the scope of the trials have expanded in terms of both patients and endpoints, we have gotten a more nuanced view. Certainly the idea that [not] everyone over 40 should be on aspirin is a reversal but we will probably always be giving prophylactic [aspirin] to a subset of patients.

Got that?

One thing seems to be sure in medicine: if we just wait long enough for excellent science to guide us ahead, things we trust as ironclad rules often change.

This post by John H. Schumann, MD, FACP, originally appeared at GlassHospital. Dr. Schumann is a general internist. His blog, GlassHospital, seeks to bring transparency to medical practice and to improve the patient experience.