Role of Diet in Cancer Hard to Study, Expert Says

Although it is clear that diet plays a role in the etiology of manycancers, making dietary recommendations to reduce risk is highly complicated,if, indeed, it is possible at all, according to an international cancerprevention expert. With 30% to 50% of all cancers traceable to dietaryfactors, food is as important as smoking as a cause of human cancer, MichaelJ. Hill, phd, told attendees of the second Strang International CancerPrevention Conference, co-sponsored by Cornell University Medical College."Unfortunately, we can give up smoking, but we cannot give up eating,"said Dr. Hill, who is chairman of the European Cancer Prevention Organization.

Cancers tied to dietary factors can be divided into two classes, Dr.Hill said: those of overnutrition, such as breast, endometrial, ovarian,prostate, and colorectal cancers; and those of undernutrition, principallystomach, liver, esophageal, and oropharyngeal cancers. Worldwide, thosetied to overnutrition are more common in the West, whereas those relatedto undernutrition are more common in the East.

Pitfalls of Making Dietary Recommendations

The problem with making dietary recommendations to reduce cancer riskis two-fold, according to Dr. Hill. First, altering diet to avoid one classof cancer may put one at risk for the other class. He gave as an examplethe so-called Mediterranean diet, which may reduce the risk of colorectalcancer while increasing the risk for gastric cancer. "Given this inverserelationship, it is hard to say what to recommend dietarily, since there'sno use preventing one cancer if you're giving yourself another one. Obviously,we have to find a middle course," he said.

The second problem has to do with the quality of data available on therole of nutrition in the development of cancer. In view of the long latencyperiod of most cancers, the dietary relationship is not well reflectedin prospective studies that span only 5 to 10 years, Dr. Hill pointed out."We're always looking at late-stage events unless there is a verylong follow-up, which is what has been the problem with beta carotene studies,"he maintained.

Furthermore, population studies necessarily involve other environmentalfactors besides diet that may skew the data. Longer range prospective studiesare impractical, Dr. Hill said, and retrospective studies are extremelyvulnerable to recall bias on the part of the study subjects. Animal studiesare of limited relevance, he said, and in vitro studies are even more limited,since "mutagens are not necessarily carcinogens." Substancesthat may, in isolation, cause DNA damage to cells in vitro may not do soin the company of other substances, as they are found in food, and in vivo,they may not cause cancers at all, since the human organism has many effectiverepair mechanisms, Dr. Hill explained.

Noncontroversial Recommendations

Nonetheless, there are some observations that are so strong and consistentthat they can be believed, Dr. Hill said. It makes sense to avoid becomingoverweight, since there are many cancers related to obesity, and none forwhich it is protective. "Avoiding overweight is safe. It won't giveyou anything else," he said.

Second, it seems clear that fruit and vegetable intake is a good thing."From the standpoint of epidemiology, it looks like it's protective,and no causal effects have been observed in a wide range of cancers, includinggastric, colorectal, esophageal, oropharyngeal, pancreatic, breast, prostate,bladder, and liver cancers," he said. It is apparent, he asserted,that cereals are protective against colorectal and probably breast cancers,and it is fairly certain that dietary fiber, not crude fiber, is responsiblefor the benefit.

The role of fat and meat, however, is less clear, according to Dr. Hill.The recommendation that fat should represent less than 30% of the totaldaily caloric intake caused a furor, he pointed out, because European countrieswith the highest fat intake also have the lowest rate of fat-related cancers,and a study in Japan showed meat to be protective against gastric cancer."What is not clear is whether fat and/or meat intake is a genuineeffect or a surrogate marker of a good diet," he said.

Dr. Hill concluded by saying that dietary intervention is probably thebest tool for studying the issue, using a high-risk population with shortfollow-up. Nonetheless, such studies are really asking the question ofwhether dietary factors can save someone in whom the process of carcinogenesisis already well underway, and clearly the answer is that they cannot.