New research on the effectiveness of folic acid in lowering the risk of heart disease highlights the danger of relying on observational studies.
When will we ever learn? Over and over, experts tell us, and the media reports, that people who engage in behavior X (let’s say it’s making paper dolls in their spare time) have a lower rate of disease Y (heart attacks, say) than people who do not make paper dolls. Ergo, conclude the experts and the press, making paper dolls prevents heart attacks.
Stated this way, it’s the height of absurdity. But that “logic” has fostered more useless and even harmful health advice than almost anything this side of homeopathy. Its latest victim: taking folic acid to decrease your blood levels of homocysteine, an amino acid in the blood, and thus of heart disease. A study in today’s Journal of the American Medical Association (JAMA) concludes that doing so has no effect on your risk of heart disease or stroke.
The idea that elevated levels of homocysteine raises the risk of heart attack became all the rage in the mid- to late 1990s. For good reason: observational studies going back to 1991 found that patients with heart disease—in particular, clogged arteries—have higher blood levels of homocysteine than healthy people, and that lower levels of homocysteine are associated with a lower risk of coronary heart disease and stroke. (Specifically, a 25 percent lower level of homocysteine was linked to an 11 percent lower risk of heart disease and a 19 percent lower risk of stroke.) For instance, a 1998 study in the journal Circulation reported that, of 800 healthy volunteers and 750 with vascular disease, the latter had, as a group, higher homocysteine levels as well as low B6 levels. Another 1998 study, in JAMA, found that among 80,000-plus women without heart disease who were followed for 14 years, those least likely to have a heart attack took folic acid and B6 supplements.
Although one editorial in Circulation was commendably cautious about whether elevated homocysteine causes cardiovascular disease or is just an innocent bystander (or whether cardiovascular disease causes elevated homocysteine, reversing the conventional take on what is cause and what is effect), another threw caution to the wind: everyone should take folic-acid supplements to reduce risk of cardiovascular disease, it said. The popular press also jumped on the homocysteine bandwagon. NEWSWEEK began beating the drums for homocysteine as an important cause of cardiovascular disease in a 1997 cover story. Noting the circumstantial evidence linking elevated homocysteine to heart disease, the story tut-tutted that “the government’s heart-health promoters say they’ll continue to ignore homocysteine…until the evidence is definitive. But that doesn’t mean you should.” The story then recommended folic acid, B12, and B6 to lower homocysteine. And in a 2005 “ask the doctor” feature, NEWSWEEK ran advice from a Harvard Medical School physician urging people who “worry about heart disease” and have elevated homocysteine levels to take folic acid, B12, and B6 to lower it.
I hope none of you took this advice, not because it’s dangerous, but because it’s useless. Today’s JAMA study, of 12,064 heart-attack survivors ages 18 to 80, finds that those who lowered their homocysteine levels with two milligrams of folic acid and one milligram of B12 supplements daily did not have a lower risk of another heart attack, of coronary death, or of stroke during the six to eight years that they were followed. (There was one bit of good news, though: taking the folic acid did not promote tumor growth, which has been theorized.) Lowering homocysteine levels, concluded the scientists, led by Jane Armitage of the University of Oxford, “produces no beneficial effects on cardiovascular disease.”
This is not the first word on whether homocysteine-lowering treatments reduce the risk of heart attacks and stroke, which makes its findings all the stronger. In a 2009 review of eight rigorous studies (randomized clinical trials, in which some of the 24,210 participants received homocysteine-lowering folic acid or other B vitamin, and others received a dummy pill), the Cochrane Collaboration reached the same negative conclusion. Lowering homocysteine, it found, “did not reduce the risk of non-fatal or fatal myocardial infarction, stroke, or death by any cause…[T]here is no evidence to support the use of [homocysteine-lowering interventions] to prevent cardiovascular events.”
Yet lowering homocysteine has been, and in some quarters remains, a pillar of cardiac health. (Just Google “lower homocysteine levels.”) How could so many be so wrong? Because of the paper-doll fallacy. Studies that simply observe two groups of people to see how they differ can’t distinguish correlations from causes: people who make paper dolls may be healthier, but it is not because they make paper dolls. (Maybe people who have time to make paper dolls are more relaxed and stress-free than the rest of us, and their lower levels of cortisol are the cause of their better health.)
The paper-doll fallacy reached its sorriest point in the decades-long crusade to get menopausal women to take hormone replacement therapy (HRT)—until the Women’s Health Initiative showed that HRT raises, rather than lowers, the risk of heart attacks and stroke, not to mention breast cancer. The advice rested in large part on observational studies finding that women who took HRT were healthier. But the reason they were healthier wasn’t because they took HRT, but because those who took HRT tended to be more educated and healthier to begin with, to have more contact with doctors, and similar factors.
In the case of homocysteine and heart health, it isn’t clear yet why the observational studies—high homocysteine equals higher risk of cardiovascular disease—were misleading. It might be that high levels of homocysteine are a marker for the real culprit, and fixing it leaves that culprit unscathed. But you have to wonder how many more times we—the press as well as supposed experts—will make the mistake of basing health advice on observational studies.