skip navigation nih record
Vol. LXIV, No. 18
August 31, 2012

next story

Willett Sharpens Scope of Diet/Cancer Link

On the front page...

Dr. Walter Willett at NIH on July 25

Dr. Walter Willett at NIH on July 25

Soft-spoken, data-driven and painstakingly careful about the accuracy of his recommendations, Dr. Walter Willett gives no outward appearance of opposing the American Way of Fun. It’s just that virtually every credible study of the causes of human cancer with which he is familiar tends to oppose the lifestyle of, say, Texas troubadour Jerry Jeff Walker, who once sang, “I like to smoke and drink and have fun, jump in my car and see how fast it’ll run.”

Willett visited NIH July 25 as a guest of NCI’s Cancer Prevention Fellowship Program. In a talk he titled “Diet & Cancer: The Fourth Paradigm,” he explained that every 10 years or so for the past 40 years, a new paradigm has emerged in the science of determining diet’s link to the incidence of cancer. The distinguished Harvard nutritionist and epidemiologist gave an overview of that era, arguing that the field’s focus would most profitably turn toward an examination of energy balance, or the problem of being overweight and under-active.


The idea that diet and cancer were linked emerged in the 1960s, he said, with barbecued meats considered a chief carcinogenic culprit.

“I don’t think we’ve totally resolved this question,” said Willett, conceding that lack of a good biomarker for carcinogenesis hampers the hypothesis. “However, we would have seen it by now if [barbecued meats] were a major cause of cancer.” Score one for Texas BBQ.

In the early 1980s, in a survey of causes of cancer death in the U.S., British scientists Richard Doll and Richard Peto found tobacco at the top of the list, but estimated that some 35 percent of cancers were dietary in origin. “They were quite unclear about the magnitude of that estimate and what the cause was,” noted Willett.

Comparisons of cancer incidence and mortality rates across countries showed large differences, as great as 10- or 20-fold. Scientists also found that when immigrants came to the U.S. from abroad, they tended to adopt the cancer rates common in their new environment, which was typically higher in rates of animal fat consumption than immigrants’ countries of origin.

“This was very critical evidence that genetic causes are not behind high rates of colon, breast, prostate and other types of cancer that we experience,” Willett observed. So what is it that’s causing variation in cancer rates from country to country? “These finding have been a huge stimulus to do more studies,” he said.

A number of large cohort studies and randomized trials such as the Women’s Health Initiative and the Nurses’ Health Study have failed to find much of an association between fat intake in midlife and later on the incidence of breast cancer. Although Willett said some of the studies were probably not measuring dietary fat effectively, the sheer weight of evidence over time has essentially “buried the low-fat paradigm,” he said.

Next up was the fruit-and-vegetable paradigm, which the National Cancer Institute embraced some years ago with its “5 servings a day” recommendation, which at the time was thought to offer as much as a 50 percent reduction in risk of major cancers.

Willett emphasized the long time frames required in getting reliable data about cancer, a disease that takes years to develop.

Willett emphasized the long time frames required in getting reliable data about cancer, a disease that takes years to develop.

Photos: Ernie Branson

“That was really a big promise that they made,” Willett said, “but a series of prospective studies have not supported the initial recommendations of that paradigm…the trends have not been significant.”

The fruit-and-vegetable dietary emphasis was helpful in preventing cardiovascular disease, he noted, but was not noticeably effective against cancer.

“That doesn’t mean there’s absolutely no benefit from fruits and vegetables,” he qualified—lycopene (found in tomato products) does seem protective against prostate cancer, he reported, but the effect is modest. “In the big picture, these benefits get lost…It’s not a totally closed door, but clearly it’s not going to be a major solution to cancer prevention.

”The fourth paradigm—energy balance—seems, Willett said, to have legs. “This one is, I think, here to stay.” Gaining weight during one’s adult years is risky; a stable weight is better than an increase, he said.

Body mass index measures in both men and women, especially waist circumference, are clearly predictive of cancer risk. “Increases in adiposity lead to a steady elevation of cancer risk of many kinds,” Willett said. “Very clearly, there is a relationship that’s been supported now in dozens and dozens of studies done across many countries.”

In fact, because overweight and obesity have become so common, they account for about the same number of cancers as does smoking. However, for an individual, smoking is still riskier than being obese, Willett said.

Throughout his talk, Willett emphasized the long time frames required in getting reliable data about a disease that takes years to develop. “Seven years is a short period of time,” he said, when studying the incidence of cancer in humans.

Back in 2003, a World Health Organization subcommittee on nutrition and cancer recommended a BMI of 18.5-25, increased physical activity and virtually no consumption of alcohol as prudent means of avoiding cancer. “A lot more is likely to be added to that list,” Willett said.

He concluded with five observations:

  • The estimate [made by Doll and Peto] is still reasonable that 30-35 percent of cancer is due to nutritional factors, but much of this is related to overweight and inactivity.
  • Alcohol consumption does increase the risks of breast and other cancers.
  • Low folate intake likely contributes to colon, and possibly other cancers.
  • Considerable evidence supports a role of high dairy consumption and low intakes of calcium, lycopene and vitamin D in human cancer.
  • We still have much to learn. Studies of diet during childhood and long-term follow-up will be important.

During a brief Q&A, Willett debunked the notion that the resveratrol found in red wine offers any health benefit.

“The idea that red wine is beneficial is French winemakers’ propaganda,” he said. “It’s the alcohol in wine that offers cardiovascular protection…You’d need a couple of gallons [of red wine] a day to reach the levels [of resveratrol] used in experiments showing benefit.”

Not that any number of American singer/songwriters haven’t tried.

back to top of page