Harvard Health Letter

War on cancer won't be won in the produce aisle

A large European study finds fruits and vegetables are only modestly protective against cancer, but it's a different story for heart disease.

We've gotten used to flip-flops in health advice by now, but the recommendation to eat plenty of fruits and vegetables looked to be incontrovertible. Science seemed to support the folk wisdom about what an apple per day would keep away. Actually, make that five apples (or oranges or peaches or servings of arugula), because in 1990 the World Health Organization (WHO) recommended that people eat five daily servings of fruits and vegetables (about 14 ounces worth) to prevent cancer and other chronic diseases. A year later, the National Cancer Institute came up with its 5-a-Day program to promote fruit and vegetable consumption. For years, cancer prevention had revolved around telling people what not to do — smoke cigarettes. Now the experts and public health officials had an opportunity to tell us what we should do — load up on the fruits and veggies.

But since those salad days, study findings have been chipping away at the notion that fruit and vegetable consumption protects against cancer. It didn't get much attention at the time, but Harvard researchers reported results in The Journal of the American Medical Association in 2001 that suggested fruit and vegetable consumption in adulthood really didn't have any effect on the risk of getting breast cancer. Two years later, the International Agency for Research on Cancer, a branch of the WHO, said the evidence for overall cancer protection was limited.

Experts knew that a shift was under way, but the public didn't, so study results reported in The Journal of the National Cancer Institute in April 2010 took many people by surprise. The headlines accurately reflected the study's main conclusion: fruit and vegetable intake has a very small effect on cancer risk. In an accompanying editorial, Dr. Walter C. Willett, chair of the nutrition department at the Harvard School of Public Health and a member of the Health Letter's editorial board, said that "any association of intake of fruits and vegetables with the risk of cancer is weak at best."

It wasn't a complete reversal; after all, "weak at best" is better than nothing. (Thanks goodness we didn't find ourselves trying to figure out how fruits and vegetables might cause cancer!) But there are some lessons to be learned from how what seemed a sure thing now seems less so. And for reasons we'll explain, giving up on fruits and vegetables is not one of them.

Not-so-epic results

The results reported in April 2010 were from the European Prospective Investigation into Cancer and Nutrition study, also known as the EPIC study. Ordinarily, the results from a single study wouldn't be given much weight, but in health research, there's credibility in numbers, and the size of the EPIC study matches its acronym. By cobbling together the cohorts of various studies, the EPIC researchers have assembled a juggernaut of a study that includes data on almost half a million people from 10 different countries.

The fruit and vegetable investigation was fairly straightforward. The researchers pulled together data on fruit and vegetable intake from answers to dietary questionnaires. Cancer registries showed that about 30,000 people in the EPIC cohorts had gotten cancer over the follow-up period of about nine years. And some sophisticated, but standard, statistical techniques were used to calculate how fruit and vegetable consumption might be related to the likelihood of being among those who got cancer. The EPIC researchers had conducted similar studies for particular types of cancers, but this time they lumped the cancers together, which increased the statistical certainty.

The main finding was that each 7-ounce increment in fruit and vegetable consumption — or about two servings' worth — lowered cancer risk by a modest 4%. That may seem okay, especially if you were to eat five servings a day. But it pales in comparison to earlier estimates of a 50% risk reduction from fruits and vegetables.

Some further analysis by the EPIC researchers showed that vegetables may have more of an effect on cancer risk than fruits, and that fruits and vegetables together may have more of an effect among people who drink lots of alcohol. The possible benefit for heavy drinkers fits with evidence from other research that alcohol creates some vulnerability to cancerous processes that may be countered by folate, a B vitamin found in many fruits and vegetables.

Four morals to the story

There are any number of morals to the story of how cancer protection from fruit and vegetables came down to earth. Here are four of them:

1. Case-control studies may mislead. As Dr. Willett explained in his editorial, much of the evidence for a large protective effect from fruits and vegetables came from case-control studies, and case-control studies tend to amplify the effect of risk factors for a couple of reasons. For one thing, they rely on memories of past behaviors and events, and people are prone to recall bias: selective memory for whatever they believe caused their condition (if they're cases) or prevented it (if they're the healthy controls).

Another problem is that the healthy controls in such studies tend to be people who are exceptionally health-conscious. That means a habit like high fruit and vegetable intake may seem more responsible for good health than it actually is.

2. Likable ideas win popularity contests. In health, as in other matters, we're more likely to accept ideas and follow advice that's appealing and leaves our other notions intact. Fruit and vegetables have a wholesome image. Studies showed — and experts believed — that antioxidants and fiber had important anticancer properties. So why wouldn't the foods that contain them also? It was only later that estimations of fiber and antioxidants as cancer fighters fell.

3. Something that comes "out of the blue" usually hasn't. The EPIC study results seemed to come out of nowhere to upend views about diet and cancer prevention. But, as mentioned earlier, it was just the latest in a number of findings that have called into question the size of a protective effect from fruits and vegetables.

4. Every study leaves stones unturned. One way to limit the frustration with the back-and-forth in health news is to recognize that health research is almost always a work in progress. No study answers all the questions.

Consider the EPIC study results and all the questions it raises. Is nine years a long enough follow-up period for a cancer study? How applicable are results for Europeans to the rest of humanity? Could "exposure" to fruit and vegetables in adulthood be too late? Perhaps consumption in childhood and the teen years matters more.

Don't give up on them

Even if the EPIC study results hold up, and fruits and vegetables prove to be minor players in cancer prevention, we shouldn't give up on them.

Cancer isn't the only foe, and there's good evidence from reliable studies (not just case-control comparisons) that eating fruits and vegetables lowers the risk of having a heart attack, stroke, and some other forms of cardiovascular disease. A large Harvard study found a 30% difference in cardiovascular risk between high and low fruit and vegetable intake.

Nor is it time to give up cancer prevention. We can choose not to smoke — and quit if we do. We can work to stay slim: being overweight is associated with a higher risk for many cancers. In fact, by some measures, extra heaviness now accounts for about the same number of cancers as active smoking. And we can exercise. Physical activity is associated with lower risk of several important cancers.