Everyone over 50 should take statins to lower their cholesterol, an editorial argued last week in The Lancet. The piece based its recommendation on a meta-analysis of 27 clinical trials published in the same issue that concluded statins significantly reduce the risk of heart attacks and other cardiovascular events in healthy people without posing substantial risks. Subsequent articles heralding the meta-analysis's findings were published in the Guardian, Forbes and the U.K. Telegraph. But based on the numbers, many experts still aren't convinced that the drugs' benefits outweigh their risks.

There's no question that statins save lives when they are prescribed to people with cardiovascular disease. But whether the drugs should also be given to healthy people who do not have high cholesterol or other cardiovascular risk factors has been a long-standing and controversial question. One large clinical trial known as JUPITER reported in 2008 that rosuvastatin (Crestor) lowers the risk of heart attacks and other events by 44 percent in healthy subjects but experts have since raised questions about its methodology in part because the trial was stopped early, which might have created the effect of overestimating the drug's benefits. The current meta-analysis was designed to help put the issue to rest. "Our aim was to bring together all the available evidence," explains co-author Colin Baigent, an epidemiologist at the University of Oxford in England.

After pooling the results of 27 trials involving 165,149 people, the meta-analysis reported that people are 21 percent less likely to suffer a serious vascular event such as a heart attack, stroke or bypass surgery after their cholesterol drops by the amount that might be expected after taking statins for a year than are similar people who do not take the pills. But such outcomes are rare in healthy individuals anyway, so the risk reduction actually translated to a small clinical benefit—reducing the overall risk from 4.04 percent to 3.27 percent per year, a difference of 0.77 percent.

In other words, approximately 130 people need to take statins for a year to prevent just one unwanted health outcome, and 500 people have to take them to prevent a single death. "Once you get down to very low levels of risk, the benefits are very small," Baigent admits.

Experts also raise questions about the subjects included in the meta-analysis. Although the review was supposedly designed to assess the effects of statins in people at low risk of vascular disease, 60 percent of its participants in fact already had vascular disease. "Why combine people who have heart disease with people who don't? It's really misleading," says Kausik Ray, a cardiologist at Saint George's University of London. In 2010 Ray and his colleagues published a meta-analysis of 11 statin clinical trials involving 65,229 subjects without cardiovascular disease and concluded that statins do not reduce the risk of death in healthy people. (By including people who had vascular disease, the Lancet meta-analysis overestimated statins' benefits: a subgroup analysis reveals that among people who did not have vascular disease, statins only reduced the absolute risk of a cardiovascular event by 0.4 percent per year.)

And what about side effects? It's long been known that statins increase the risk of hemorrhagic strokes, muscle pain and other severe (but rare) muscle and liver complications. In February 2012 the U.S. Food and Drug Administration warned consumers that the drugs might also increase the risk of diabetes and memory loss. Nevertheless, the Lancet analysis suggests that the probability of these side effects is quite low; for instance, only one person out of every 2,000 treated will suffer a hemorrhagic stroke. "The benefits are substantially bigger than the hazards, even at very low levels of risk," Baigent says.

Yet some experts worry the findings underestimate true risk. According to Rita Redberg, a cardiologist at the University of California, San Francisco, and chief editor of the Archives of Internal Medicine, prior to the start of one of the trials included in the analysis, potential subjects were given statins for several weeks to see how well they tolerated them. If any individuals experienced side effects, they weren't invited into the trial. This type of prescreening is "not clean science," says Vinay Prasad, an internist at Northwestern University Feinberg School of Medicine, because it makes drugs look safer than they really are.

There is also the issue of funding and bias. Almost all of the trials included in the meta-analysis were funded in part by pharmaceutical companies, and some of the meta-analysis's co-authors have received honoraria from drug companies, too. Although these facts do not mean that the results are invalid, a 2003 study published in the British Medical Journal suggests that trials funded by drug companies are more likely to report favorable results about their products than are trials funded by independent organizations. "Asking industry to conduct its own studies is like asking a painter to judge his own work as to whether or not it should win an award," Prasad says. "It's another reason to be skeptical."

And even if the findings of the meta-analysis are accurate and the side effects are rare, "one should question whether it is cost-effective" to put millions of people on the drugs for a small potential benefit, says James Liao, an expert in vascular medicine at Brigham and Women's Hospital in Boston. Generic low-cost versions of the drugs are available—Walmart sells a month's supply of low-dose lovastatin for $4—but many consumers choose costlier brand name options. In 2011, for instance, Americans spent $4.4 billion on AstraZeneca's Crestor and $7.7 billion on Pfizer's Lipitor.

So should more healthy people take statins? Baigent argues that doing so could save millions of lives, because many heart attacks occur in people who are considered low-risk. But at what cost? "There are a lot of people taking statins who are not getting any benefits from them, and they're subject to a lot of adverse events," Redberg says.