AACR 2009: Diet, Nutrition, and Cancer — Don't Trust Any Single Study

Zosia Chustecka

April 22, 2009

April 22, 2009 (Denver, Colorado) — Numerous studies on diet and cancer were presented here at the American Association for Cancer Research (AACR) 100th Annual Meeting, but several of the findings that were highlighted in AACR press releases — and thus are likely to be picked up by the lay media — run counter to the accumulated body of evidence, and some of the comments based on these studies are untrue or premature. So said Walter Willet, MD, DrPH, from the department of nutrition at Harvard School of Public Health, in Boston, Massachusetts, in an exclusive interview with Medscape Oncology.

"No conclusions should be made on the basis of a single study," he said.

Dr. Willett presented an overview entitled "Diet, Nutrition, and Cancer: The Search for Truth," in which he reviewed many of the associations that have been suggested by epidemiologic studies. These include consumption of red meat, meat cooked at a high temperature, a high-fat diet, and alcohol all increasing the risk, and fruit and vegetables decreasing the risk. However, much of the evidence for these links is rather weak, he said; the most robust evidence supports a link between obesity and an increased risk for cancer.

"The estimate that diet contributes to around 30% to 35% of cancers is still reasonable," Dr. Willet said, "but much of this is related to being overweight and inactive."

"At this point in time, being overweight is second only to smoking as a clear and avoidable cause of cancer," he said. "People should stay as lean as they can, recognizing that it is more difficult for some than for others."

What we are looking at are little slices of life.

Beyond this clear message about obesity, there are only hints from the rest of the data. One of the main limitations of all of the studies so far is that they have looked at a specific time of life — for example, women after menopause — and they have had fairly short follow-ups, often less than 10 years. "So what we are looking at are little slices of life," Dr. Willet said, whereas the effect of diet is lifelong, and might be particularly important in the years before adulthood (e.g., during adolescence).

Barbequing and Other High-Temperature Cooking

One suspect that has been extensively studied as potentially increasing the risk for cancer is the high-temperature cooking of meat, such as barbequing, grilling, frying, and roasting, during which the meat is charred and can form carcinogens.

If there was a strong association, we would have seen it by now.

"But after more than 30 years of study, this link has not been refuted or confirmed in any clear way," Dr. Willet commented. "If there was a strong association we would have seen it by now, but we cannot exclude a mild or moderate effect."

One of the studies highlighted in an ACCR press release suggests that charred meat increases the risk for pancreatic cancer. The finding comes from a prospective analysis of 62,581 participants of the Prostate, Lung, Colorectal, and Ovarian multicenter screening trial, and was presented by Kristin Anderson, PhD, associate professor at the University of Minnesota School of Public Health, in Minneapolis. Her team looked at 208 cases of pancreatic cancer, and found that individuals who preferred very well done steak were almost 60% more likely to develop pancreatic cancer than those who ate their steak less well done or who did not eat steak at all. When the researchers considered overall consumption and doneness preferences, this rose to a 70% higher risk for pancreatic cancer.

"We cannot say with absolute certainty that the risk is increased due to carcinogens formed in burned meat," Dr. Andersen said in the AACR press release. "However, those who enjoy either fried or barbequed meat should consider turning down the heat or cutting off the burned portions when it's finished."

Dr. Anderson also advised "cooking meat sufficiently to kill bacteria without charring," and microwaving meat for a few minutes and pouring off the juices before cooking it on the grill to reduce the precursors of cancer-causing compounds.

But Dr. Willet said that these are very specific recommendations, and "I just don't think that this is appropriate on the basis of a single study."

Alcohol — Even 1 Glass Might Increase Risk

For alcohol, there have now been dozens of studies showing an increase in the risk for breast cancer, even with very low levels of consumption, "so this is now an established relationship," Dr. Willet said in his talk.

It's been known for a long time that alcohol increases the risk for cancers of the upper aerodigestive organs, but this is at high levels of consumption (around 3 or 4 glasses a day), he told Medscape Oncology. "What's unique about breast cancer is that the risk is increased at very modest levels of consumption," he said.

"There is strong evidence that even 1 glass a day can cause a small but significant increase in the risk of breast cancer," he said. A recent study from the United Kingdom suggests that the risk for many different cancers is increased with even 1 drink a day, and that the risk increases in a dose-dependent fashion, as reported by Medscape Oncology.

So the finding from another study highlighted by the AACR, that "drinking wine may increase survival among non-Hodgkin's lymphoma patients," is somewhat surprising. "This conclusion is controversial," admits first author Xuesong Han, a doctoral candidate at the Yale School of Public Health, in New Haven, Connecticut. "However, we are continually seeing a link between wine and positive outcomes in many cancers," she noted in the AACR press release.

"This is not true," said Dr. Willet. There have been benefits shown consistently for cardiovascular disease, but not for cancer, he told Medscape Oncology.

The study conducted by Han and colleagues involved 546 women with non-Hodgkin's lymphoma. Those who drank wine had a 5-year survival rate of 76%, and those who did not had a 5-year survival rate of 65%. In a subanalysis, the researchers found that the strongest link was seen in patients with diffuse large B-cell lymphoma. These patients had a 40% to 50% reduced risk for death, relapse, or secondary cancer.

The researchers also asked the patients about their wine-drinking habits in the 25 years before their diagnosis. In the overall group, patients who had been drinking wine at least this long had a 25% to 35% reduced risk for death, relapse, or cancer, whereas in the subgroup of patients with large B-cell lymphoma, this reduction was 60%.

"We cannot look at this 1 study is isolation," said Dr. Willet.

Specific Foods and Anticancer Effects

So far, there have been no specific foodstuffs that have been identified as having proven anticancer effects, Dr. Willet told Medscape Oncology.

Even the case for eating more fruit and vegetables, a message widely promulgated by many authorities, including the World Cancer Research Fund, is fairly weak when it comes to cancer. There have been studies showing a decrease in the risk for colon and breast cancer, but other studies have shown insignificant or no appreciable effects, Dr. Willet told the meeting. A 2004 meta-analysis by Hsin-Chia Hung and colleagues concluded that eating more fruits and vegetables decreases the risk for cardiovascular disease, but not the risk for cancer (J Natl Cancer Inst. 2004;96:1577-1584).

"So the message to eat fruits and vegetables is still a good message, but there appears to be more benefit for cardiovascular disease than for cancer," he said.

In that context, the claim made in another AACR press release, that "walnuts may prevent breast cancer" is premature, especially because it is based on an animal study, Dr. Willet said. The study was conducted in a mouse model of breast cancer, and mice fed a diet estimated to contain the human equivalent of 2 ounces of walnuts per day showed a significant decrease in the incidence of tumors, a significant decrease in tumor size, and a delay in the development of these tumors by about 3 weeks..

Lead researcher Elaine Hardman, PhD, associate professor of medicine at Marshall University School of Medicine, in Huntington, West Virginia, said: "It is clear that walnuts contribute to a healthy diet that can reduce breast cancer."

"That's a premature leap, that's for sure," said Dr. Willet.

Maybe the only foodstuff that does have some evidence suggesting a preventive anticancer effect is soy products. Its antiestrogen properties might protect against prostate cancer in men and against breast cancer in women, especially in premenstrual women, Dr. Willet noted. "But this is not yet in the category of convincing — it's possible," he said.

Dr. Willet has disclosed no relevant financial relationships.

American Association for Cancer Research (AACR) 100th Annual Meeting: Abstracts LB-224, LB-243, and LB-247. Presented April 21, 2009.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.