The myth of unhealthy belly fat

Before continuing with the obesity paradox series, one of the most important null studies of the year deserves mention… especially since the media universally ignored it. As the body mass index (BMI) is finally being recognized as an uncredible measure of health or predictive of premature death, other measures of body fat are being promoted because everyone “knows” that fat is unhealthy.

It’s inconceivable to contemplate that our condemnation of our body fat and of fat people might be little more than vanity, profit and prejudices. That increasingly seems to be the case, though, when we stop to think about why we remain so intent on finding a reason to condemn fat even when the null studies are far stronger than any others.

One of the most popularized new measures is waist circumference or waist-to-hip ratio, as an indicator of belly fat. It’s based on the belief that there is good and bad body fat, and that visceral fat — the fat that accumulates inside the abdomen — is the unhealthy, dangerous fat.

Null studies: No link, don’t blink

It’s not science itself that we can’t trust, it’s the stuff that often becomes pop science and cults masquerading as science.

Popular culture is topsy turvy when it comes to understanding science. Many people mistakenly believe that the strength of science comes in proving something. It’s actually the exact opposite. Science doesn’t prove anything. The ability of a carefully designed study to disprove a hypothesis is what sets science apart from pseudoscience. Those null findings that are unable to support a hypothesis or belief are the most important. Yet, we almost never about them and few people realize that those are the ones we should be paying attention to.

"As a scientist you are taught not to answer questions, but to question answers." — Anonymous

Null findings enable true scientists to know they’re looking in the wrong direction and that it’s time to go back to the drawing board and develop a different hypothesis. They also enable us to stop needlessly worrying about something that doesn’t matter. Null results are most vital to the progress of science and are the source of Albert Einstein’s famous saying: “No amount of experimentation can ever prove me right; a single experiment can prove me wrong.”

In fact, a big earmark of junk science and pseudoscience — or science being misused to sell us something or advance a special interest — is continuing to test and retest a belief long after it’s been disproven in well-designed studies. Building a body of research can lead unsuspecting consumers to believe that there’s a body of evidence in support of a belief. All studies are not created equal and the weight of the evidence does not come by tallying up the number of studies on one side or another.

Another common earmark of the misuse of science is trying to make the science appear conflicting and undecided, when it really isn’t, by burying us in an overwhelming number of conflicting studies. Unfortunately, these techniques can be quite effective when people don’t understand how to recognize a well-designed, strong study of merit versus a poorly-designed weak one.

The problem is accentuated when people aren’t armed with scientific literacy and fall back on fallacies of logic, especially the top five. When people don’t know the difference between studies that can be trusted from those that can’t, they get taken advantage of.

When scientists talk about a strong, sound study, they are referring to one that’s been well-designed to be a fair test of a hypothesis. That means one that has carefully controlled for biases. In science, “bias” doesn’t refer to the motives of a study’s author, but to specific aspects of study designs that can lead to faulty conclusions. Biases “stack the deck” and can be imposed in all sorts of ways, such as how the groups being compared are selected (randomized or cherry-picked, representative samples of the general population or from the Land of Ignognita); the quality of the data being measured (actually measured data or self-reported, recalled information or measured in real-time); and failing to control for confounding factors (contributing factors to mortality that could go along with the one you’re looking at, such as age or social-economic status).

Epidemiological (observational) studies — which use computer models to dredge through a chunk of information on a group of people to find correlations — are most rife with misinterpreted statistics, errors and biases, and are most easily manipulated to arrive at whatever conclusions researchers set out to find. That makes them especially helpful for marketing, too. Epidemiological studies make up the bulk of what’s funded, published and reported nowadays because they’re done in a computer and are cheap and easy to do. These are the studies most popularly reported as the health scare of the week and the health miracle of the week. They also cause epidemiological whiplash — coffee is bad for us one week and good for us the next.

They’re also the source of today’s never ending efforts to link our lifestyles, foods or the environment to some horrible disease.

The most common blunders when interpreting epidemiology are using correlations as evidence of causation and thinking that the findings are meaningful when they’re no greater than chance, computer error or random coincidence. In other words, failing to acknowledge a null study. No matter how popular, impressive or intuitively correct it might seem, a correlation to a disease is not evidence that it’s the cause. A correlation big enough to suggest a true link, and one that deserves our attention, is also much bigger than most people would guess. But it’s easy to trick people with statistics and lead them to believe that random chance findings mean something.

All studies are not created equal in another way, too. Each type of study is designed to answer specific questions and cannot credibly be used to conclude things it wasn’t designed to answer. When we hear about an epidemiological study finding something linked to higher risks for a disease, for example, it means the scientific process has barely begun to investigate a potential cause, let alone a treatment. Finding a correlation is just the first step in narrowing down potential factors in a disease. If a strong link is found, then an hypothesis about a potential cause is thought up and tested in a series of randomized, controlled clinical intervention studies. Only after an intervention has been tested in humans to see if it’s effective and if the benefits outweigh the harms does it suggest something we or our doctors might want to actually do.

In other words, when you hear about an association, don’t blink, sit back and wait for a study that actually tests something. Don’t YOU be the one who falls for the fallacy of correlations as evidence for causation. Even more importantly, look for those null studies. When a well-designed epidemiological study can’t even find a strong correlation, then that variable can’t possibly be the cause and you can relax and move on.

That’s why those null studies are so important for us.

The health risk factors said to be associated with obesity — high blood sugars and insulin resistance, high blood lipids (cholesterol and triglycerides) — and said to lead to diabetes, heart disease and premature death — are all blamed on visceral fat. These health indices have been lumped together and called the metabolic syndrome. The entire metabolic syndrome theory — which is being used to support endless preventive health screening tests and surveillance, "healthy eating" plans, exercise programs and prescription drugs (that are costly for us, but make gobs of money for those who want to manage our health) — is held up by beliefs about visceral fat.

This theory is evidence of the failure to understand risk factors and of how a belief can be built and take on a life of its own by ignoring null studies — in this case, layers of null studies.

Null link: BMI; waist, hip, and arm circumference; waist-hip ratio; waist-height ratio; skinfold thickness; and body fat — and all causes of death

Among the many studies showing no link, the most recent null studies were two independent analyses of the most precise measurements of body size, measurements and body composition available on a large representative sample of the U.S. population conducted by the Third National Health and Nutrition Examination Survey (NHANES III, 1988–1994) of the Centers for Disease Control and Prevention (CDC). As senior scientists at the National Center for Health Statistics at the CDC and the National Cancer Institute reported, the data shows that no higher measurement of body shape or size — BMI; waist, hip or arm circumference; waist-hip ratio; waist-height ratio; skinfold thickness or body fat composition measured by bioelectrical impedance — is predictive of higher risks of dying from all causes.

Nor was there a net benefit of using BMI versus another measurement. The data also found that NONE of the 21 diseases popularly attributed to obesity — those “obesity-related” diseases, including: cardiovascular disease, cancers (colon cancer, breast cancer, esophageal cancer, uterine cancer, ovarian cancer, kidney cancer, or pancreatic cancer) and diabetes or kidney disease — are actually associated with excess deaths at any BMI category, including obese.

Null link: waist circumference, waist-hip ratio, and visceral fat — and mortality, type 2 diabetes or cardiovascular disease

A second group of researchers examined the data using different statistical methods and reached the same results. They found no significant correlation between BMI, waist circumference or waist-hip ratio and risks for death from all causes among younger adults, while higher BMIs and waist measurements were associated with lower risks for dying among older adults. Tummy fat was not even linked to greater risks for premature death.

As they also noted in their review of the medical literature, while waist circumference has been more highly correlated with visceral fat than waist-hip ratio or waist-thigh ratio, studies to date have not consistently supported a correlation between abdominal fat or body fat distribution as predictive of mortality. Nor has waist circumference been consistently shown to be more strongly associated with type 2 diabetes, cardiovascular disease or mortality than waist-hip ratio.

Null link: visceral fat — and health outcomes or mortality or obesity or belly fat

Regardless of the fact that no association has been found between visceral fat and actual health outcomes or mortality, these null studies continue to be ignored. The visceral fat theory, which was first created by lumping together metabolic health risk factors, is now hanging its hat on a purported correlation with another surrogate measure for heart disease: carotid artery intima-media thickness.

We’re now really getting into risk factor (correlation) squared. Risk factors are being defined based on correlations with other risk factors based on other risk factors. All the while, ignoring the null studies.

With visceral fat found to have no association with obesity or belly fat, the theory now claims that even thin people with small waistlines are at risk from visceral fat based on a correlation with carotid IMT measurements. But this ignores the null studies that have found that carotid IMT measurements are not a measure of atherosclerosis, either. As Japanese researchers reported in the April 2001 issue of the journal Stroke found, “that increased intima-media thickness is a physiological effect of aging that corresponds to diffuse intimal thickening, especially in very elderly persons, and that IMT is distinct from pathological plaque formation.” Rotterdam Study researchers, using different methodology, found IMT added nothing to predicting coronary heart disease and cerebrovascular disease.

Null link: Metabolic syndrome — and cardiovascular disease or premature death

Even more significantly, while the metabo syndrome and all of its health indices are being used to promote a plethora of preventive health interventions and pills, none of the metabolic risk factors being attributed to visceral fat in the first place has been shown to have a meaningful correlation to cardiovascular disease or premature death.

Fallacy of looking for causal role where there’s not even a link

Abdominal fat and metabo. So, we have null studies that have found no tenable correlation between any “obesity-related” metabolic risk factor and waist circumference or any body measure or abdominal fat. But those null studies were ignored and the search for a causal role has continued to look where no link has even been found. That’s not how good science works.

So, what do you think researchers at Washington University School of Medicine in St. Louis and the Istituto Superiore di Sanità Rome in Rome, found when they performed a clinical study to test the hypothesis that abdominal fat was the cause of “obesity-related” metabolic risk factors and that reducing abdominal fat could lower those risk factors?

They recruited 15 obese women with abdominal obesity (average BMI 35.1) willing to have extensive liposuction to remove 28-44 percent of their abdominal fat (an average of an astounding 22 pounds of abdominal fat from each of them). Over the next four months, the women were clinically monitored. As the researchers reported in a June 2004 issue of the New England Journal of Medicine, even after physically removing a significant amount of abdominal fat they found no effect on any metabolic lab measure (insulin sensitivity of muscle, liver or adipose tissue; glucose; C-reactive protein; interleukin-6; tumor necrosis factor {alpha}; and adiponectin) or other risk factor for heart disease or diabetes (blood pressure, plasma glucose, insulin, and lipids).

This research had been led by Dr. Samuel Klein, M.D., who was actually medical director of the Weight Management Program at Washington University School of Medicine and at that time was president of the North American Association for the Study of Obesity (now the Obesity Society). With grant and research support from a range of pharmaceutical companies, he also served on the Consensus Conference Panel of the American Society of Bariatric Surgery (now called the American Society for Metabolic and Bariatric Surgery), in their 2004 Consensus Statement recommending bariatric surgery as “the most effective therapy available for morbid obesity and can result in improvement or complete resolution of obesity comorbidities.”

But as he admitted in a Clinical Cornerstone, the failure of liposuction to reduce metabolic risk factors and the seeming favorable effects of bariatrics on those metabolic risk factors has not been shown to be due to reducing fat, but by “inducing a negative energy balance.”

While bariatric surgery and diet and weight loss is said to resolve “comorbidities associated with obesity” and to cure type 2 diabetes, high cholesterol and high blood pressure, as JFS has examined at length, these health claims are based on temporary drops that occur in these health indices during any period of caloric restriction, starvation and weight loss — even before much weight or fat loss has occurred at all. It’s not about fat.

As examined here last year, when you stop eating, eat insufficient calories, or suffer from malabsorption, what happens to your blood sugars? They drop, of course. Does that mean the underlying disease pathology of diabetes has been cured or arrested? Of course not. Blood sugars are a symptom, a health index, not the disease of diabetes itself. Dr. Francesco Rubino, now at New York Presbyterian Weill Cornell Medical Center, and Dr. Jacques Marescaux, M.D. at the University Louis Pasteur in France, wrote in the 2004 issue of Annals of Surgery that claims “a direct antidiabetic effect of bariatric surgery... is not supported scientifically.” As they explained, “it is admittedly impracticable to rule out that the rapid normalization of plasma glucose and improved insulin resistance after these surgeries be simply the effect of decreased caloric intake.”

Despite the null studies, the belief that belly fat and metabo are bad have continued virtually unchecked.

Visceral fat and metabolic syndrome. The “obesity is deadly” belief still rests on visceral fat as being “particularly pernicious to health” because of its purported causal role in the metabolic syndrome. The metabolic syndrome theory mistakenly attributes obesity-related risk factors (including blood sugar, insulin resistance, blood pressure and blood lipids) as causing diabetes, heart disease and premature death. Visceral fat is largely genetic, as well as appears greater with aging, yo-yo dieting and emotional stress.

Yet, the strongest, well-controlled null epidemiological studies have found no correlation between visceral fat itself and “obesity-related” metabolic risk factors, adverse health outcome or premature death. The importance of null studies hasn’t been understood.

So, what do you think researchers found when they conducted the definitive study (a randomized clinical trial, the gold standard) to confirm once and for all if visceral fat has anything to do with metabolic health risk factors?

If visceral fat had a real link to metabolic risk factors, then surgically removing it would change those metabolic measures and finally settle the question of whether the visceral fat theory has any merit. As Dr. Klein had written: “The true test needed to determine whether visceral fat is deleterious is to surgically remove visceral fat and see if that results in a beneficial metabolic effect.” The largest portion of visceral fat in our bodies is found in the omentum, a large apron of fat inside the abdomen.

Incredibly, 70 adults (average age 37) agreed to participate in a randomized clinical trial at the University of Chile in Santiago, where half would have their visceral fat (the entire omentum) surgically removed and half would have surgery without removing the omentum. The study results were published in the April issue of Obesity Surgery.

The study participants were all obese, with an average BMI of 43. Were the participants told that their risks of premature mortality are no different than people with “normal” BMIs or did they agree to participate in such an extreme clinical trial, and a surgery that even the bariatric surgeons admitted was difficult, because they believed their obesity put them at greater risk for dying? All of the participants otherwise received that same treatment, gastric bypass, and all had their metabolic risk factors — weight, blood pressure, blood sugar, insulin, serum cholesterol and triglycerides — evaluated before the surgery and two years afterwards.

Two years later, there was no significant difference in percentage of weight loss between the groups, or in changes of blood pressure, blood sugar levels, insulin levels, and cholesterol or triglyceride levels. Visceral fat proved unrelated to metabolic risk factors. As the authors, led by Dr. Attila Csendes, M.D., concluded, there is no scientific basis for removing visceral fat and the theoretical benefits were not supported in this clinical trial.

How long do you think it will take for the importance of the null findings of even the most radical clinical trial every conducted on visceral fat to be recognized? How long will people continue to be frightened that their belly or having bad numbers or not eating "healthy" means they’re at risk for type 2 diabetes, heart disease and premature death? How long will these fears be used to compel heightened monitoring of metabolic lab indices in fat and older people, and interventions to get their numbers increasingly lower, under the belief they’ll reduce their risks for developing chronic diseases of aging or dying prematurely? There is no evidence to suggest it will happen anytime soon.

Until more people understand the importance of null studies and risk factors, how to recognize strong studies, and what makes science they can trust, entire belief systems can be built on things other than the soundest evidence.

© 2009 Sandy Szwarc. All rights reserved.