JFS Exclusive: What’s the evidence cancers are our own fault?

We haven’t been able to turn on the television or radio or open a newspaper without being hit with news of the Second Expert Report just issued by the World Cancer Research Fund and American Institute for Cancer Research. The Report is called “Food, Nutrition, Physical Activity and the Prevention of Cancer: a Global Perspective,” and is said to have found the most convincing evidence yet that being fat causes six cancers. Cancers, we’re told, “are mostly preventable” by following their diet and lifestyle recommendations.

Its press release, “Landmark report: Excess body fat causes cancer,” was reported around the world as if a script. The evidence “is clearer than ever. Even small amounts of excess body fat, especially around the middle, can make cancer more likely,” Dr. Emily Senay told viewers on the CBS Early Morning Show. “Never before has the evidence been more compelling and the advice more specific,” said Charles Gibson during ABC World News Tonight.

The Report issued the most stringent cancer-preventive recommendations of yet any organization, including limit red meat, high-caloric foods, alcohol, refined carbohydrates and salt; avoid sugary drinks, fast foods and processed meats; eat mostly a plant-based diet with at least 5 servings of fruits and vegetables a day; be active and stay as thin as possible within the ‘normal’ range of weight.

This Report may have impressed or frightened some because of its size, the numbers of experts involved, and the bold assertive claims that it was “the most comprehensive review ever published on the evidence linking cancer risk to diet, physical activity and weight.” But wade through the 517 pages and you will discover that not one of the key recommendations were supported in clinical studies. In fact, the Report not only failed to provide what most scientists would consider convincing evidence, its conclusions were often the opposite of the evidence.

Rather than take the word of experts or the media, let’s see for ourselves.

Meta-study

While CBS News said the authors had “analyzed more than 7,000 clinical trials” and USA Today called it “a landmark report,” it was nothing more than a giant meta-analysis.

“This is a technique for trying to get a convincing result by combining the results of a lot of unconvincing studies,” said Dr. John Brignell, Ph.D., a British scientist and engineer, publisher of Number Watch, and author of The Epidemiologists — Have they got scares for you!

This may be the world’s biggest example of the foibles of meta-analyses: pooling weak studies — all of which use different techniques, populations, time periods, measures and statistical models; few of which account for confounding factors such as socioeconomic status or smoking — hoping to create a significant effect. Or, trying to build a convincing case by the sheer number of studies. The authors chose which studies to include in their review and these were not all randomized clinical trials as many consumers may believe, but almost entirely epidemiological ones. As we know, these studies dredge through characteristics among populations looking for correlations, in this case to cancers. Of course, epidemiological studies can pull out thousands of meaningless associations that often contradict each other and this Report is filled with confusing contradictions.

Close examination also reveals that this Report repeatedly excluded from its review stronger studies which contradicted the panel’s recommendations. It also included only a handful of randomized controlled clinical trials, even when well-designed trials were available. There had actually been nearly half a million studies in their original literature search, but they pared it down to the ones they determined to be relevant. Only 2,471 cancer-related dietary studies were referenced in this paper (with another 1,363 related to process, opinion summations, or body composition/obesity theories), and many of those studies were repeats in several chapters. There were nowhere close to 7,000 studies referenced.

Methodology — defining the evidence

The authors with the World Cancer Research Fund Global Network acknowledged that they were the first cancer group to “create awareness of the relationship between diet and cancer” as highlighted in its first Report in 1997, and they worked from that predetermined belief and “what is already known.”

As they wrote in the Introduction:

The Panel is aware of the general consensus shared by scientists, health professionals, and policy-makers on the relationships between food, nutrition, physical activity, body composition, and the risk of cancer... It is abundantly clear that the incidence of all the common cancers in humans is determined by various potentially controllable external factors.... it means that cancer is, in large part, a preventable disease.”

They admitted that the methods used in this Report were new and others had been developed in their first Report. One of their tasks in this Report, they said, was to develop a method that “enabled evidence of relationships between diet, physical activity, body fatness, and associated factors, and cancer to be judged as causal, with varying degrees of confidence.”

But, correlations are never proof of causation. This needs to be repeated over and over again, said Dr. Brignell, because it’s one of the greatest contributions to confusion, scares and deliberate deceptions. Randomized, controlled clinical trials have long been recognized in the scientific process as the gold standard for determining causation. Countless times, interventional trials have proven beliefs based on epidemiological studies to be false.

The authors didn't use established epidemiological techniques but created their own method for proving causation from correlations, which they described:

The Panel endorses the view of the panel responsible for the previous report, that causal relationships between food and nutrition, and physical activity can be confidently inferred when epidemiological evidence, and experimental and other biological findings, are consistent, unbiased, strong, graded, coherent, repeated, and plausible. Individually, none of these factors is likely to be sufficient to infer a causal relationship with confidence.

How did they define associations as being causal? They created three grades of evidence:

· Convincing: Associations deemed as strong enough evidence to call ‘convincing’ of a causal relationship included: “at least two independent cohort studies...and a plausible biological gradient (‘dose response’) in the association. Such a gradient need not be linear or even in the same direction across the different levels of exposure, so long as this can be explained plausibly.”

· Probable: They defined associations as being strong enough to label as ‘probable’ of a causal relationship if it included: “at least two independent cohort studies, or at least five case control studies” and they could find a biological plausibility.

· Limited: The label of “limited, but suggestive” included associations “too limited to permit a probable or convincing causal judgement, but where there is evidence suggestive of a direction of effect. The evidence may have methodological flaws, or be limited in amount, but shows a generally consistent direction of effect. This almost always does not justify recommendations.” It included “at least two independent cohort studies or at least five case control studies.” And evidence labeled “limited, no conclusion” was that so limited no conclusion could be made.

But, as we’ll see, none of the relative risks reported in the epidemiological studies they used were strong enough to be valid, as defined by the National Cancer Institute (October 26, 1994), and credibly due to anything more than chance, statistical bias or the effects of confounding factors. Yet, they were sufficient evidence of causation in this Report.

The first line of evidence

According to their press statement, “the evidence that excess body fat increases the risk of developing cancer is much stronger now than ever before.” Expert panel member Phillip James, whose promotion of obesity issues is well known, as chairman of the International Obesity Task Force, said: “The most striking finding in the report is that excess body fat increases risk for numerous cancers. That is why body weight is the focus of our first recommendation.” So, their first recommendation was not just to be within the “normal” weight range of BMI 18.5-24.9, but to “be as lean as possible” within that range.

“The first lines of evidence suggesting that cancer is a largely preventable disease” are international variations and trends in cancers, these experts said. In Chapter One, twelve countries were profiled, including descriptions of the rates of cancer deaths, life expectancies and body mass indexes. But look at this data — do you see any support for their claims that BMIs near 18.5 means less cancer or longer life?

On the contrary, as BMIs slightly rise in each country, so do the average ages and life expectancies among the populations, but there is no relationship between cancer deaths and BMIs. The average BMIs in Egypt (with the lowest cancer deaths), for instance, are about the same as in Australia (with one of the highest cancer deaths) — but people live about 13 years longer in Australia. Average BMIs in India are the lowest they exampled, but men and women in the U.S. live 14 to 17 years longer. The Report supplied no support that the entire world needs to be on a diet to prevent cancer.

We can’t review each of the thousands of studies used in this Report — most of you have lives. So, we’ll look at some of the deadliest cancers and foods highlighted in this Report.

Breast cancer

These cancers are the fifth leading cause of death in women and the leading cancer-related death in women. The Report notes that this diagnosis is made more in high-income countries, and age is the most significant risk factor, with individual risk factors varying by age of menarche, childbearing and menopause.

Concerning body fatness, they used 43 cohort studies and more than 100 case-controlled studies but found the data “inconsistent.” Look closely at the relative risks of BMI for breast cancers reported in these studies — they all hug a relative risk of 1.0 (a null finding) on either side, and none approaches 3, essential to be considered tenable in epidemiological studies.

In discussing possible biological explanations, they suggested hormones might be a predominant mechanism, and described the theories by which fatness could prevent premenopausal breast cancer due to the role of progesterone and estrogen. Their selectivity was evident, however, when they rebuffed the preventive role of fatness, saying “there is no single well established mechanism through which body fatness could prevent premenopausal breast cancer.”

The panel concluded: “There is a substantial amount of consistent epidemiological evidence with a dose response relationship, but the mechanistic evidence is speculative. Greater body fatness probably protects against premenopausal breast cancer [but] the evidence that greater body fatness is a cause of postmenopausal breast cancer is convincing.” Interestingly, they say the same things about height (the other part of the BMI equation) as being a cause for breast and ovarian cancers. Fatness isn’t even mentioned as a risk factor for ovarian cancer, the seventh most common cause of cancer death in women.

Look again at their chart. Not one relative risk of BMI to breast cancers deviates sufficiently from 1.0 to be beyond what could happen by chance or statistical error. Convincing as a cause?

Lung cancer

Lung cancer is the single most common cancer and is the deadliest form of cancer in the world. The Report attributed smoking as being related to 85% of lung cancers and occupational exposures to carcinogens next. In lung cancers, they found the evidence linking dietary vegetables, red meat, processed meats, total fats and butter, like physical activity, to be limited and inconsistent. Evidence for dietary supplements was sparse and inconsistent. Body fat was found to be associated with a lower risk, but they discounted this relationship as possibly confounded by cigarette smoking.

Instead, they reported that pooled meta-analyses of cohort data suggested fruits might lower risk and concluded: “The evidence is ample and consistent. A dose-response relationship is apparent from both cohort and case control studies and there is evidence for plausible mechanisms operating in humans. The evidence that fruits protect against lung cancer is convincing.”

However, among the studies not included anywhere in this Report, for instance, was the massive study led by Dr. Walter C. Willett, M.D., at Harvard School of Public Health in the Journal of the National Cancer Institute reviewed here. Examining the data on 71,910 women and 37,725 men followed for 15 years, Willett and colleagues found no relationship between fruits and vegetables and cancer.

Look at their chart. Not one relative risk of fruit to lung cancers deviates sufficiently from 1.0 to be beyond what could happen by chance or random fluke. Convincing of cause?

Pancreatic cancer

While this is the ninth most common cause of cancer deaths, age-adjusted incidences have been stable since the 1970s, they reported. The Report found the evidence for all of the foods they considered, including red meat and fruits, or body fatness to be inconsistent and any correlations not beyond chance. Ample evidence examining coffee, for instance, “fails to show an association. It is unlikely that coffee has a substantial effect on the risk of pancreatic cancer,” they concluded.

Look at their graph. Do you see any appreciable differences between their admittedly untenable associations with coffee, and those between fruits and lung cancer above, which they claimed were ‘convincing’ evidence of preventing cancer? Of course not.

Stomach cancer

Stomach cancers are the second most common cause of death from cancer. Age-adjusted rates for these cancers are dropping and in 2002 were half of what they were just 30 years ago. The bacterium H.pylori and various infectious agents are associated with stomach cancers and the increased availability of refrigeration has likely aided the decline of these cancers, they reported.

In fact, in contrast to the Report’s conclusions that all cancers are largely preventable by diet, exercise and weight control, it actually stated that “approximately 1 in 4 cancers in low-income countries are estimated to be attributable to infection. In 2002, this represented some 1.9 million cancers or close to 1 in 5 of all cancers worldwide.” The role of diet comes in — not from eating too much, but — in ensuring people get enough to eat to prevent nutritional deficiencies which “can lead to immunodeficiencies and increased susceptibility to infections.”

In reviewing epidemiological studies surrounding fruits and vegetables, they reported nonsignificant relationship to stomach cancers, with none of the studies finding relative risks reaching relevance. The evidence for legumes, soy products, chili and other foods were limited and inconsistent. That included processed meats, smoked meats, and grilled or barbecued meats.

And what about salt, most popularly linked to stomach cancers? They reported that a “meta-analysis of case-control data showed an [untenable] 18 percent increased risk per gram of sodium per day; the meta-analyses for total salt indicated increased risk but were not statistically significant.” As they acknowledged, even in the higher-quality studies, confounding factors could explain the differences seen as “salt is inversely related to the availability of refrigeration in a population, and so to socioeconomic status, which is itself related to stomach cancer risk.” No food was related to this cancer, either.

Liver cancer

These tragic cancers are almost always fatal and are the third most common cause of death from cancer. Aflatoxins (which contaminate organic grains and legumes stored in moist, hot conditions) have long been recognized as a direct source of liver cancer and increase risks up to seven-fold. Hepatitis B and C infections are also a cause of liver cancer and about 80% of liver cancers develop in cirrhotic livers, they reported.

Not surprisingly, they found that any evidence for a role of fruits and vegetables, as well as body fatness, was sparse and inconsistent. Even no study of alcohol consumption they reviewed had been able to find statistically significant results. “The cancer disease path may be different in people with cirrhosis,” they noted, but studies were difficult to interpret. No food was related to this cancer.

Prostate cancer

This is the sixth most common cause of death among men. They reported that age-adjusted rates were increasing “largely because of the increased availability for screening for prostate specific antigen (PSA) in men...leading to the detection of many prostate cancers that are small and/or would otherwise remain unrecognized and which may or may not develop further into higher stage disease.” It is mainly a disease among high-income countries and risks increase with age, they said.

They found the evidence linking prostate cancer with legumes, soy products, processed meats, milk and dairy products, vitamin E or beta-carotene to be sparse or limited and inconsistent. Concerning foods containing lycopene, however, with no support they concluded: “There is a substantial amount of consistent evidence, in particular on tomato products, from both cohort and case-control studies. There is evidence for plausible mechanisms. Foods containing lycopene probably protect against prostate cancer.”

Not included in their review, however, were any of the studies examining lycopene and cancers discussed in a recent post, such as one conducted by the U.S. Food and Drug Administration (FDA) Center for Food Safety and Applied Nutrition which found no association between lycopene and any cancer evaluated. The Report also didn’t include the large study at the National Cancer Institute and Fred Hutchinson Cancer Research Center among 28,000 men enrolled in the Prostate, Lung, Colorectal, and Ovarian (PLCO) Cancer Screening Trial. This ongoing, randomized NCI trial found no association between serum lycopene and total prostate cancers or aggressive prostate cancers.

Colorectal cancer

Cancers of the colon and rectum are the fourth most common cause of death from cancer and incidences are highest in industrialized, high-income countries. The Panel determined “that food and nutrition have a highly important role in the prevention and causation of [colorectal] cancers.” While they said that as many as 10% are the consequence of recognized hereditary conditions and another 20% occur in those with family histories, inflammatory bowel diseases and certain medications have also been shown to increase risks. Nevertheless, these cancers are most preventable by appropriate diets, they said.

Let’s look.

Dietary fiber is the dietary change most commonly believed to protect against colorectal cancers. Based on 107 epidemiological studies, they said that most found lower risks — but none were significant, including pooled analyses. “Therefore no effect can be attributed specifically to fibre,” they wrote. So what did they conclude in their recommendations? They pretty much ignored their own findings: “A clear dose-response relationship is apparent from generally consistent cohort studies, supported by evidence for plausible mechanisms, but residual confounding could not be excluded. Foods containing dietary fibre probably protect against colorectal cancer.”

This Report failed to include in its review five randomized clinical trials of dietary fibre that have been conducted. Not one has found fibre to be effective at reducing the recurrence of polyps or the occurrence of colorectal cancer. JFS recently reviewed these studies and others on dietary fiber and whole grains.

The Report found no evidence for a relationship between colorectal cancers and vegetables, fruits, fish, cheese and sugars. The evidence was all inconsistent and limited, even by their own admissions. Evidence for animal fats was limited, with the 5 cohort studies they reviewed and a meta-analysis all showing nonsignificant associations, they reported.

Concerning red meat, they found most of the 16 cohort studies they considered showed higher risks, but they only cited relative risks below relevance, such as 15% and 43%. Similarly, they claimed an equally untenable 21% higher risk had been reported in a meta-analysis of cohort data on processed meats, but they included very few studies, most of which were more than a decade old.

Despite no epidemiological evidence showing a tenable connection, they concluded that both red meat and processed meats are a “convincing cause of colorectal cancer.” The risks of processed meats were so severe, that they recommended people eat “very little if any.”

Oh really?

Here again, larger, stronger and newer studies were not included in this Report, including the largest meat study to date. For example, researchers at Harvard examined 14 studies on 725,258 people in North America and Europe investigating meat and fat and associations with colorectal cancer risk. As they reported in the Proceedings of the American Association of Cancer Research in 2004: “Greater intake of either red meat (excluding processed meat) or processed meat was not related to colorectal cancer risk.” This study was found nowhere in the list of references in this Report.

Also not included was the May 2000 or the May 2001 analyses “Toxicology and Carcinogenesis Studies of Sodium Nitrite,” by the National Toxicology Program, a division of the U.S. Department of Health and Human Services and considered the gold standard for determining cancer risks, which found “no evidence of carcinogenic activity” due to sodium nitrite. As also reported in the Carcinogenic Potency Database, the research to date has found nitrites do not cause cancer. While it’s popular among consumers to think that nitrates and nitrites mostly come from processed meats, in actuality they are a very small source of our intakes (less than 5%). Nitrates occur naturally in vegetables and plants as a result of the nitrogen cycle where nitrogen is fixed by bacteria. Dietary studies around the world have found 70% (in UK) to over 97% (New Zealand) of our daily consumption of nitrates and nitrites comes from vegetables. Nitrites used in cured meats, by the way, prevent the growth of botulism-causing bacteria and prevent spoilage that could make people sick. Since it was added to cured meats in 1925, there’s been no cases of botulism from commercially cured meats.

And concerning fatness, in the Report’s review of epidemiological studies, the associations they found weren’t even close to tenable — relative risks from a meta-analysis of an insignificant 15%. Yet, they concluded: “There is abundant and consistent epidemiological evidence with a clear dose-response relationship, and evidence for plausible mechanisms that operate in humans. The evidence that greater body fatness is a cause of colorectal cancer is convincing.”

Convincing of cause?

None of the recently reviewed studies, which showed no association between BMI and colorectal cancers, were included in the Report. These included the National Institutes on Health’s “Annual Report to the Nation on the Status of Cancer, 1975-2003,” which found that the incidences of colon and rectal cancers (despite increased surveillance) have been dropping among both men and women, as weights rise. Nor did the Report include the study by the Kansas Cancer Registry which found “obesity is associated with lower incidence rates of colorectal cancer.” Nor did it include the study of colonoscopy results for more than 3,000 adults conducted at 13 Veterans Affairs medical centers across the country.

We could go on, but you get the idea. Overall, none of the 17 cancers reviewed in this Report found even statistically meaningful associations with foods (sugars or sugary drinks, fats, meats, fruits, vegetables, legumes, vitamins, alcohol or processed foods) or body fatness.

Nor did it find evidence that losing weight, avoiding practically every food and beverage imaginable (except moldy, diseased foods), and eating their idea of a healthy diet prevents cancers. Did we mention that the Report also didn’t even consider the largest, longest randomized clinical dietary intervention trial ever conducted on ‘healthy’ eating? We already know what that one found, too.

Alan Caruba at the National Anxiety Center had it right when he made us laugh at all of this. As he wrote this week in USA Daily:

Eating Food Will Kill You

It is now a proven fact that eating food—any kind of food—will kill you. No one who has eaten food in the past is alive today and everyone currently eating food will die. Therefore, those noble people who seek to save us from eating every kind of food that the earth provides should be hailed and saluted for their efforts to keep us alive.

I say this as the son of a woman who taught the art of haute cuisine for over three decades and authored several cookbooks. That poor woman died at age 98 and I am convinced it was all that fabulous food that killed her. Ridiculous? YES!

© Sandy Szwarc. All rights reserved.