Quotes of the Day

Starting a low-fat diet in mid-life does not, by itself, decrease a woman's risk of heart disease or stroke or dramatically reduce her risk of breast or colorectal cancer, according to three studies published today of almost 50,000 healthy, post-menopausal women. But don’t break out the ice cream just yet. Once you get into the nitty-gritty of the study, the investigators stress, the findings still support the general idea that you need to pay attention to how much fat—and particularly what kinds of fat—you eat.

Still, the results of these studies, which were sponsored by the National Heart, Lung and Blood Institute and published in the Journal of the American Medical Association, cannot be ignored, or easily dismissed. This is by far the largest, most scientifically rigorous attempt ever to test a low-fat diet—defined as 20% or less of total calories coming from fat. Researchers randomly divided the participants into two groups, worked hard to get one group to cut its fat intake and then compared the results for each of them for the next several years.

Through focus groups and sessions with nutritionists, the so-called intervention group was given a lot of help and information on how to cut the amount of fat in their diet. As a group they weren’t able to get all the way down to 20% fat intake, but they did manage to reduce it to 24% in the first year, though it slid back up to 29% by the sixth year. The other, so-called control group received printed material on dietary guidelines but no other help on healthy eating. The fat content of their diets was 35% after one year and crept up to 37% after six years—and they gained a little weight to boot. (The intervention group lost weight or stayed the same.)

All told, the women in the two groups showed no overall difference in the number of heart attacks or strokes they suffered over the eight years of the study. Part of the reason may be that the level of LDL or "bad" cholesterol dropped in the intervention group by a tiny 2.4%—too small a reduction for the study to be able to detect a difference in outcome.

It’s also possible that if you want to prevent cardiovascular disease, it’s not enough to focus just on the total amount of fat in the diet; you also have to look at the types of fat—particularly saturated and trans fats—as well as exercise and physical activity levels. Why didn’t the study focus more intently on these variables? Researchers were not as aware of how important they were back in the late 1980s, when they were designing the trial.

Women in the low-fat group suffered 9% fewer breast cancers than those in the control group. Although that difference was not statistically significant, it is very suggestive. Given how long it takes for most tumors to grow, it may simply be that the study has not lasted long enough to show a significant effect. In addition, there was a clear benefit for one sub-group of women—those who began the study with the highest total fat consumption and who were able to make and maintain the greatest reduction in the number of fat calories in their diet. This group benefited from a 15% to 20% drop in the incidence of breast cancer.

Women in the low-fat group did not see a significant drop in their rates of colorectal cancer, but here again, it may mean that the study simply hasn’t lasted long enough. The low-fat group did have 9% fewer polyps and other precancerous lesions in their lower gastrointestinal tract—and that is a statistically significant finding. Given how long it takes colorectal cancer to develop, it’s reasonable to be heartened by those results.

Why is a 9% difference significant in the colorectal cancer group but not in the breast cancer group? Welcome to the wild and complex world of statistics. After looking at the variation in the data with various mathematical tools, research statisticians determined that they couldn't rule out the possibility that the 9% difference in breast cancer results was due to chance alone. (It could be chance; it could also be due to the low-fat diet.) On the other hand, after applying those same mathematical tools to the data in the colorectal cancer study, they determined that the 9% difference in the number of polyps was unlikely to be due to chance alone.

So, in many ways these studies raised more questions than they answered. (They’re on solid ground there, considering they were conducted as part of the Women’s Health Initiative, which famously concluded in 2002 that long-term hormone replacement therapy does not protect women’s hearts after menopause. There are plenty of doctors, and quite a few patients, who are still trying to figure what that actually means.) Strictly speaking, the results do not apply to men, who tend to get heart disease earlier than women do and for whom a low-fat diet might be particularly beneficial. But this study has set the gold standard for diet studies in the future. Fortunately, there are enough research funds to continue following the WHI participants for another five years. Here’s hoping that is long enough to get more answers that are statistically significant—and potentially life-saving.

• Christine Gorman
• Three major studies of 50,000 women show how hard it is to prove that eating healthfully makes a difference