I realize this study is SO last week, which is about two months in dog years and a decade in internet years, but seeing as I’m about to lend my dulcet tones and my delicate opinions to Skeptically Speaking on this topic, I feel I must needs blog this paper.
That, and it’s about coffee. How could I NOT blog this paper!?
Freedman et al. “Association of coffee drinking with total and cause-specific mortality” NJEM, 2012.
From some of the coverage I’ve read of this study, you’d think that all you had to do to live happily til 100 is drink MOAR COFFEE. Far be it from me to prevent the intake of one of my favorite beverages, but really…it’s just not that simple. And I was very pleased by a few pieces of coverage (like this one) that went into the caveats of the study. Because if it seems as simple as “do this one thing and live longer!”…well it’s probably not true.
So what was this study? It’s a study taken from a large set of data: the AARP Diet and Health study. This is older adults, 50-71. The authors excluded people who were already obese, had diabetes, heart disease, or cancer. They looked at the data in the study concerning how much coffee people drank, and whether it was caffeinated or decaf.
During the time of the data collection (1995-2008), of the total 402,260 people, 52,515 of them died. At first blush, the risk of death (comparing the people who died to the people who didn’t and their demographics) was higher among the coffee drinkers.
But when you break it down, a large number of the heavy coffee drinkers (more than 2 cups/day) were also smokers, which is a very high risk of death in and of itself. When you controlled for the smokers, the authors got the OPPOSITE effect, this time coffee drinking (more than 2 cups per day), decreased the risk of mortality by 10{9f43b4361d9a125bc126dd2a2d1949be02545ec69880430bc4fed2272fd72da3} in males and 15{9f43b4361d9a125bc126dd2a2d1949be02545ec69880430bc4fed2272fd72da3} in females.
Keep in mind there that “risk” is risk in the epidemiological sense. So the study found that you are less likely to die (by a very little bit, 10{9f43b4361d9a125bc126dd2a2d1949be02545ec69880430bc4fed2272fd72da3} is not a lot) if you are an older adult if you drink coffee (caffeinated vs decaf did not make a difference).
When they broke it down by causes of death, they found that mortality risk was decreased in coffee drinkers for most causes: accidental death or injury (more alert and less car accidents?), heart disease, stroke, diabetes, and infections. The only exception was cancer.
So does this mean that we should all be grabbing coffee to extend our lives? Honestly, probably not. Coffee drinking was also associated with a number of other unhealthy habits, not just smoking. Coffee drinkers eat fewer vegetables, exercise less, consume more calories, drank more alcohol, took fewer vitamins, and ate more red meat. Downing a few more cups of coffee isn’t going to make up for other bad habits if you have them. And of course, this is a study in humans, and humans are full of variables. There are lots of other things.
Things this study did NOT take into account:
1) How long people had been drinking coffee for vs when they died
2) They took measures of diet but did not control for it, they didn’t compare with similar caloric intakes.
3) Meausures of other health outcomes OTHER THAN diabetes, obesity, cancer, etc. 4) They never asked about blood pressure or cholesterol
5) Adjusting for levels of exercise
6) They compared caffeinated vs decaf but only if you had one or the other “more than 50{9f43b4361d9a125bc126dd2a2d1949be02545ec69880430bc4fed2272fd72da3} of the time”
While they took measures of all these things, they only corrected for the smoking. And adjustments for variables, though helpful, will never give you the real picture. And of course, there are two big issues aside from all the little ones.
A: correlation is not causation. People with chronic disease, for example, are often told to avoid caffeine and stop drinking coffee. And in general, it’s impossible to prove a cause here, what we have instead is an association. That’s a good start, but it doesn’t prove…
B: What the mechanism is. What’s going on? Coffee is an amalgamation of a hundreds of different chemicals, far more than caffeine alone. I don’t think you could figure this out from a human study, you’d have to break down the chemicals and test the unknown ones against various combinations.
So what CAN we tell from this? Well, so far it appears that coffee doesn’t increase the risk of death. So that’s a start. And should you drink coffee because you like it? Sure! But should you screw up your nose and start chugging coffee and hope to live to 100? Probably not yet.
Freedman, N., Park, Y., Abnet, C., Hollenbeck, A., & Sinha, R. (2012). Association of Coffee Drinking with Total and Cause-Specific Mortality New England Journal of Medicine, 366 (20), 1891-1904 DOI: 10.1056/NEJMoa1112010
