Fluke or Trend: Coffee vs Mortality

As you might have guessed, I read USAToday quite a bit, so of course, I stumbled on their report about a possible link between coffee consumption and mortality in which the reporter pitted a Harvard professor of public health against the chief of cardiology at the Cleveland Clinic.  While neither were involved in the study, the former found the findings intriguing while the latter picked apart the study and thus concluded the findings were nonsensical.  The lead author agreed that the study design did not prove that coffee affects mortality and thus could not be used as evidence to go out and start drinking coffee.  So will coffee really affect our longevity?

It turns out that a prospective cohort study published in the Annals of Internal Medicine addressed this same issue back in June 2008 when the authors followed 41,736 men in the Health Professional Follow-Up Study for 18 years and 86,214 women in the Nurses' Health Study for 24 years, all w/o any history of heart disease and/or cancer at baseline.  To their credit, the authors assessed coffee consumption every 2 to 4 years.  While some participants rarely drank any coffee, there were some who drank 6 or more cups/day!  The good news back then was the coffee consumption was not associated w/all-cause mortality in either men or women.  In fact, there was a hint of decrease risk in women.

Fast forward 4 years to the New England Journal of Medicine which published today the results of the prospective cohort analysis of 229,119 male and 173, 141 female participants in the National Institutes of Health - AARP Diet & Health Study who were 50-71yo (avg 61yo) at baseline and free of cancer, heart disease and stroke.  After 14 years of follow up, the authors actually found that greater coffee consumption was associated w/lower all-cause mortality, after taking into account the usual confounders.

But as expected, association does not prove causation.  What's worse, the authors only checked for coffee consumption just once at the beginning of the study.  So as cardiologist Dr. Steve Nissen pointed out, habits vary over time.  It makes no sense to base a long-term association upon just one assessment.  However, I have to agree w/Dr. Frank Hu that it makes sense to take into account the usual suspects, eg subtract out as many confounding factors as possible, including tobacco use.  After all, we do this for all other studies.

So the debate rages on.  The good news for those who love their morning cuppa joe is that we have two large, long-term studies, neither of which demonstrate any harm w/greater consumption.  Since this looks more like a trend than a fluke, go out and enjoy your morning brew.

Follow @alvinblin
Tweet