TCS Daily

Cancer Report

By John Luik - December 27, 2007 12:00 AM

Ever since the recent release of the massive report from the World Cancer Research Fund ("Food, Nutrition and Physical Acitivty and the Prevention of Cancer") the media have been busy pounding into our heads the three new truths about cancer, food and obesity- namely that being fat increases our risk for cancer, that eating certain foods gives us cancer and that cancer is in the words of the report "mostly preventable".

But before committing ourselves to a life of little red meat, alcohol, soft drinks, milk shakes, French fries, pastries, chips and other such "bad" foods, and to being as "thin as possible", it's worth spending a minute to go behind the scary headlines and ask whether science really supports these three "truths" about cancer, fat, and food.

The report's authors tell us that it is just obvious since they looked at over half a million studies, and then concentrated on 7000 that were most relevant. But that is not quite true, for they actually reference slightly less than 2500 studies on disease and diet, and more importantly, the conveniently omit many major studies that don't support their three truths theory.

Most crucially, they reference almost exclusively epidemiological studies- studies which can never establish that "being fat gives you cancer" or that eating "red meat gives you cancer", since that's not what these sorts of studies do. Indeed, the very nature of these sorts of studies means that the margin of error arising from the nature of the data exceeds the supposed relationships that the study has found. And while they admit this- sort of- in the report, they somehow forget to mention it at the press conference. Perhaps because it might well have spoiled an otherwise really good story.

So what about the big headline attention-getting claim that being fat gives one cancer. What the report actually claims is that being overweight or obese can increase your risk for six cancers- cancers of the oesophagus, pancreas, colon/rectum, breast, endometrium and kidney. But when you look at the report's support for this conclusion the evidence is extremely limited.

Take pancreatic cancer, for example. The report cites 20 case control studies, but only three show a statistically significant association between obesity and pancreatic cancer. Similarly, of 42 cohort studies on colorectal cancer- the third most common cancer in US men and women- only 13 show a link with obesity. Again, with breast cancer and obesity, of 16 studies, only three are statistically significant, while eight show a DECREASED risk between breast cancer and obesity. Even for oesophageal cancer the increased risk was generally found in the morbidly obese with BMI's of at least 40. And with endometrial and kidney cancers the relative risks were below two, risks which according to the National Cancer Institute are so small that they may be due to "chance, statistical bias or the effects of confounding factors."

A similar pattern of results is found in the just-published Million Women Study from the UK( Reeves et al "Cancer incidence and mortality in relation to body mass index in the Million Women Study: cohort study" BMJ 7/11/07) which examined the evidence for a link between 17 of the most common cancers and BMI. In this study the incidence of ten of the cancers does not show a statistically significant association with either higher levels of overweight or obesity. Of the remaining seven cancers, the association between overweight and the cancer is nonsignificant in four, and where the results are significant the RR's except for endometrial and oesophagal (adenocarcinoma) cancer are never stronger than 2 except in the obese.

But just as important is the fact that the obesity- cancer link is contradicted by a just-published study from the US National Cancer Institute and the Centres for Disease Control ( Flegal et al "Cause-Specific Excess Deaths Assocaited with Underweight, Overweight, and Obesity" JAMA 7/11/07) This study found that being overweight was not associated with increased mortality from cancers considered obesity-related, and further noted "little or no association of excess all-cancer mortality with any of the BMI categories. Indeed, the study suggests that overweight might in fact be protective against cancer. For instance, in individuals aged 25-59, obesity appeared to be protective against death from cancer. Even for those individuals aged 70 and over, BMI's in excess of 35 were not significantly linked with a higher risk of cancer mortality.

What then about the second truth, that eating certain foods increases our risk for cancer? Of the 17 cancers discussed in the report, virtually all have statistically nonsignificant associations with every type of food, which means that they provide no evidence of a link between a particular food and a particular cancer. For instance, of the 17 studies cited which looked a link between colon cancer and processed meat, 13 are not statistically significant. Despite the scary headlines about red meat, the report concludes that "there is limited evidence... suggesting that red meat is a cause of oesophageal cancer." Or again, "there is limited, inconsistent evidence... that grilled ... or barbecued animal foods are causes of stomach cancer." Or "there is limited evidence suggesting that processed meat is a cause of stomach cancer." Given the limited nature of this evidence, it is difficult to see just how the report justified its advice to avoid red and processed meat.

And just in case these findings are thought to be anomalous, there is the American Cancer Society's 2001 study ("A Prospective Study of Diet and Stomach Cancer Mortality in the United States Men and Women) of diet and stomach cancer which looked at 436,000 men and women and found no increased risk of stomach cancer associated with eating processed meats. What that study did find, however, was an increased risk of stomach cancer with women who consumed more vegetables!

And what then, finally, of the report's claim that cancer is "mostly preventable"? This is perhaps the most curious claim since there is massive evidence of the best kind that suggests that it is simply not true. One of the largest and most expensive randomized controlled studies of the effect of diet and weight on breast cancer, colon cancer, heart disease and strokes- is the Women's Health Initiative Dietary Modification Trial. Almost 49,000 US women were followed over an eight year period. The women in the intervention group ate diets that were low fat, and high fiber with six servings of grains and five of vegetable and fruits per day. The results? There were no statistically significant differences between the intervention and the control group in the incidence of breast cancer, colon cancer, strokes or heart attacks. Ironically, the women following the healthy diet designed to reduce cancer and heart disease didn't even weigh less than they did at the beginning or less than the group that continued to eat as they always had.

So, unlike the epidemiological studies cited in the report, a gold standard randomized and controlled intervention to test the claim that there is a connection between eating certain foods and being a certain weight and preventing cancer found nothing. Again, this study is not unique. A just published analysis of data from 14 studies involving 756,000 men and women followed from 6-20 years in the Journal of the National Cancer Institute (Savage et al) found that fruit and vegetable intake was not associated with a reduced colon cancer risk. Some prevention.

Far from showing some connection between cancer and being fat, or eating certain foods, or even more extravagantly preventing cancer, what this report and its studies show most conclusively is that, as one epidemiologist put it, people who eat die.


TCS Daily Archives