What's the fuss about coffee?

Why did coffee study make headlines?

Why is it that so much scientific research ends up making headlines? Because they ask us to report on their research, that’s one reason. Medical journals and research institutions work hard to make sure that their scientific papers appear in the news.  

This is how it works: science and medical reporters are given special access to websites that give advance notice of upcoming research. There are dizzying lists of dozens of scientific papers from hundreds of journals, volumes of new research that is published every single day. There are armies of helpful public relations people who arrange interviews with the scientists, so we can all be ready for the moment when the embargo lifts.

Really? Drinking coffee helps you live longer? CBC's Kelly Crowe asks. (John Rieti/CBC)

The embargo is a strict deadline, enforced by threat of future exclusion, and reporters who mistakenly break the embargo can be punished for years by the offended journal, which will refuse all access to future papers. This fear of embargo-breaking keeps the international media in line. We all hold back on the story until the designated day and time, say 5 p.m. on Wednesday, when the embargo magically lifts, and the headlines fly out around the world, giving the impression that news has just broken wide open.

Take this past Wednesday, for example. At the magical hour of 5 p.m. when the embargo on this week’s New England Journal of Medicine was lifted, more than 500 headlines suddenly appeared, all announcing some variation on the theme of coffee and longer life.  "Coffee Drinkers May Live Longer" (New York Times), "Coffee Drinkers Have Lower Risk of Overall Death, Study Shows" (ABC News), "Coffee Drinkers Live Longer, Big Study Finds" (Vancouver Sun).

The world's science and medical reporters had been reading the study for a few days, in advance of the embargo deadline, trying to decide whether and how to report on it. For CBC TV News, we tried to book an interview with the lead author, but he seemed to be camera shy and refused to talk on television. But I did have some questions about the research.

First question: Really?  Drinking coffee helps you live longer? That seems like pretty big news. But that's not what the study said. The headlines should have read "study shows that coffee doesn't seem to hurt you" even if you smoke,  drink too much,  eat unhealthy food, or avoid exercise, all habits that seem to go along with drinking coffee.

Once the epidemiologists squeezed the data through their statistical models, and adjusted for all the smoking, red meat eating, vegetable avoiding, boozing couch potatoes who also drink coffee, it appears that coffee is "associated" with a modest decrease in risk of dying from everything but cancer. Not that it keeps you from dying, just that the rate of death in coffee drinkers is slightly lower.

The authors reported a 10 per cent lower risk of death in men who drank two or more cups of coffee a day, compared to men who didn’t drink coffee. For women, coffee drinking was associated with a 15 per cent lower risk of death.

In the paper the authors admit that the study doesn't prove anything. "It is not possible to conclude that the inverse relationship between coffee consumption and mortality reflects cause and effect" the paper states.

Although he wouldn't do an on camera interview, author Neal Freedman, from the National Cancer Institute did answer a few of my questions in an email and over the phone, and he confirmed that the observed associations could reflect another exposure, just like all observational studies of this type.

The research is based on a questionnaire from 1995 asking more than 500,000 American seniors, over age 50, how much coffee they consumed, along with a long list of other lifestyle and nutrition questions. The original database was established to compare dietary habits and cancer, but these coffee researchers used it for a more specific purpose, to study the relationship between coffee and death, and they sifted through the data to see whether drinking coffee was associated with both overall death, and death from specific causes, like heart disease.

Before they used the data, the researchers excluded anyone who suffered from cancer, heart disease, or stroke, and anyone with extremely high or low caloric intake, which prompts a naïve reporter question: "If they separated out all the people who were sick with cancer, heart disease and stroke from the study group, does that mean that only healthy people were considered? And if the research subjects made it to age 50 and beyond without cancer, heart disease or stroke, could it be that they were in good health anyway and that was why they had a lower risk of dying? Maybe coffee had nothing to do with it."

In his email to me, Freedman said he excluded people who were already sick, because that might have affected how much coffee they drank.

But how much coffee were these people drinking anyway? The researchers admit in the paper that they aren't sure. The study says "coffee consumption was assessed by self report at a single time point and may not reflect long-term patterns of consumption". 

So, another naïve reporter question:   "If the study subjects only filled out one questionnaire, once in their lives, how does anyone know that their coffee consumption didn't change over time?  Maybe they stopped drinking coffee, or maybe they switched to herbal tea?"  In his email to me,  Neal Freedman said it is likely that at least some of the people changed their coffee drinking habits, although other studies suggest people stick to the coffee habits for years. And he said if there were changes, that would make the coffee drinking and non-coffee drinking groups more similar, "which would likely have attenuated our findings."  In other words, changes in coffee drinking habits would have weakened the findings.

The authors also admit they couldn’t be confident about whether people were drinking caffeinated or decaffeinated coffee: "The distinction between persons who drank caffeinated coffee and those who drank decaffeinated coffee was subject to misclassification," the study said. The authors also weren't sure what kind of coffee they drank, even though they suggest it might matter: "We lacked data on how coffee was prepared (espresso, boiled, or filtered), and the constituents of coffee may differ according to the method of preparation of either beverage more than half the time."

So what can we learn from this paper? Apparently not much beyond the fact that coffee probably won't hurt you.

"Our results provide reassurance with respect to the concern that coffee drinking might adversely affect health," the authors write.

Haven't we heard this before? It turns out that a 2008 paper by a Harvard University researcher came to a similar conclusion, that "regular coffee consumption was not associated with an increased mortality rate in either men or women."  

But this is the biggest study to show that coffee isn't dangerous. And other researchers we contacted said that it was a well done study, for its type, and that it confirms the findings of previous studies. So although it's not clear that drinking coffee is good for you, this study confirms what others have shown, that it won't harm you.

When I asked Freedman, in the email, why he did the study in the first place, this is what he wrote:

"Coffee is one of the most widely consumed beverages, but the association between coffee consumption and health has been unclear. Several recent studies suggested that there might be a modest inverse association between coffee drinking and the risk of death, but these possible modest inverse associations mostly did not reach statistical significance. Our study is larger than previous studies and therefore is well suited to evaluating modest associations."

Note he said "modest association" between coffee and mortality, not "proof that drinking coffee extends life." A rather different headline than "Coffee drinkers live longer."