Reassuring Findings continued...
But he's not advising anyone to start drinking coffee on the basis of his results.
"Coffee is a complicated beverage," Freedman says, noting that it contains 1,000 different compounds, most of them little-studied.
Besides, because study participants weren't assigned to drink coffee or not drink it, the researchers can't be sure what caused the lower death rate -- the beverage itself or some unmeasured characteristic of the people who chose to drink it. This type of study can show the association, but it can't say that coffee caused people to live longer.
If you are thinking about starting to drink coffee, talk to your doctor about whether you might have any reason not to, Freedman says, adding that "there are some data that show coffee might cause a short-term increase in blood pressure."
Although their study can't prove that coffee itself lowers drinkers' risk of dying, Freedman and his colleagues speculate about how it might. Caffeine probably is not a factor, he says, because death rates linked to decaf, preferred by a third of the coffee drinkers, were similar to those associated with caffeinated. But other compounds in coffee, such as antioxidants, might be important, the researchers write.
The researchers note several limitations of their study. For one, they lacked information about how the coffee was prepared, which could affect its compounds. And they asked about coffee consumption only once at the beginning of the study, but participants' habits might have changed over time.
"People do quit drinking coffee as they get older, or they switch to decaf ... because they don't sleep as well or they get palpitations," says Arthur Klatsky, MD, senior consultant in cardiology for Kaiser Permanente of Northern California. Klatsky was not involved with Freedman's research but has conducted his own studies of coffee and health.
"It's a big study with a remarkable result," Klatsky says. "It's a little bit hard to believe that coffee drinking is protective against all those different causes of death."
His own research found no such connection overall. Perhaps Freedman's study would have come up with different findings if it had enrolled a younger and more racially diverse population, Klatsky says.