Obesity virus — a new risk factor?

With more than 301,000 articles on Google reporting on an obesity virus, it clearly has created a media sensation. Surprisingly few people have noticed how correlations have been built to make us believe that the science for a fat virus is far more significant than it really is.

The best way to tell this story is probably to go back to the beginning.

Animal observations

The term infectobesity was invented by Nikhil Dhurandhar, Ph.D.. His father was a doctor who headed a weight loss clinic in India. While a young student at the University of Bombay, India, in 1988, Dhurandhar learned from a veterinary friend at Bombay Veterinary College about a plague that was killing thousands of chickens across India. The chicken virus behind the plague was SMAM-1. But instead of wasting away, the chickens seemed to have more fat on their abdomen. To confirm this finding, Dhurandhar infected chickens with the virus and they got fatter, with low levels of cholesterol and triglycerides. He published his findings in the Journal of Bombay Veterinary College [1990; 2:131-132] and in the June 1992 issue of Veterinary Microbiology.

In 1992, he moved his family to the United States in search of a lab to support his research and to create a new field of viral obesity. In 1994, he got a job collaborating with Dr. Richard Atkinson, M.D., then at the University of Wisconsin in Madison. In 1997 in Obesity Research, they published an observational report of an association between positive SMAM-1 antibodies in ten people in Bombay who weighed about 35 pounds more than a group who tested negative for the antibodies. No other information about these subjects was revealed, making any interpretation of this purported correlation impossible.

Dhurandhar and Atkinson were unable to get the SMAM-1 virus to continue his research here in the United States, they explained, because it only existed in India and the U.S. Department of Agriculture wouldn’t allow them to import the virus into this country. So, they turned to a laboratory-supply company catalog and randomly chose one of the more than fifty human adenoviruses in the American Type Culture Collection virus bank. They ordered Ad-36.

Human adenoviruses are a large group of 51 common viruses that cause a wide range of infections in people, including colds and upper respiratory, gastrointestinal and eye infections. The Ad-36 virus had been first isolated in 1978 in Germany from the feces of a diabetic girl suffering from enteritis. Dhurandhar and Atkinson began injecting Ad-36 into animals, in three experiments in chickens and one in mice. The very first adenovirus they tried, Ad-36, worked like SMAM-1. They were the first researchers, they said, to report that this human virus injected in animals was associated with greater adiposity. And the only researchers to date. They published these findings in the August 2000 issue of the International Journal of Obesity, the journal of the International Association for the Study of Obesity (IASO).

Few seemed to take notice of a small note in the discussion section of their animal study that said a lot of other human adenoviruses might also have the same affect on animals. Ad-36 may not be unique. But Ad-36 was the only one anyone at that time had ever tried to inject into animals, and focus was quickly turning to it as the obesity virus. Media largely didn't question if the fat buildup in animals exposed to human adenoviruses might be more a veterinary concern and might not translate to anything meaningful for humans.

Sure enough, in an abstract presented at the 2002 Nutrition Week Scientific and Clinical Forum and Exposition, these researchers noted that when they tested a different adenovirus, Ad-37, it produced similar effects in chickens as Ad-36, raising visceral fat in the chickens by 3.5 grams on average. These findings were finally published in the online September 2005 and in the January 2006 issue of the American Journal of Physiology — Regulatory, Integrative and Comparative Physiology. By then, another human adenovirus, Ad-5, had also been found by researchers at Imperial College London in 2005, to induce adiposity when injected into mice. This made three human adenoviruses found to affect animals, complicating the Ad-36 “obesity virus” hypothesis. In their report, Dr. Atkinson admitted that the “effects of the other 49 human adenoviruses on adiposity and serum lipids in vivo and on adipogenic potential in vitro have not been assessed.”

An editorial in that issue by Frank Greenway, M.D., with the Obesity and Functional Foods divisions at Pennington Biomedical Research Center, acknowledged that “the development of obesity in chickens is not an effective screening tool to identify human adenoviruses capable of causing obesity in humans.” Defining all of the adenovirus serotypes that might be linked to obesity by screening large populations is a daunting task, making an in vitro assay screening test more efficient. If there was one. But when testing an in vitro assay test [3T3-L1 cells], Dr. Atkinson and colleagues reported disappointing findings that “suggests that the in vitro assay will not be a good screening test to predict response to human adenoviruses in animals.” So, at the present time, they wrote, testing for human antibodies appears to be the only method available to screen for adenoviruses.

Back in 2002, the Atkinson-Dhurandhar group had also proceeded to inject the human Ad-36 into three male marmoset moneys. They reported in the October issue of the Journal of Nutrition that the infected monkeys were 12% heavier and had 4.7% greater total fat at 28 weeks than three noninfected controls. But their animal research was finding problems with antibody tests, too. Few seemed to notice that they’d found that tests for human adenovirus antibodies were proving imprecise in identifying monkeys who’d been infected. One in three of the infected marmoset monkeys, they reported, failed to ever develop detectible levels of antibodies. An antibody screening test with one-third false negatives would not be a useful or valid test in the real world.

Ad-36 proved to be such a common and widely spread virus that its practicality proved to be challenging in another way in animal studies. In their monkey report, the researchers noted that they didn’t use rhesus monkeys because the entire rhesus monkey colony at WRPRC had Ad-36 antibodies, “suggesting a covert epidemic of Ad-36 infection.” They noted similar “naturally occurring antibodies to Ad-36 in chickens, as well as rats (unpublished data).” Ad-36 has the ability to infect and replicate in widely disparate vertebrate species, they said. A test for a virus that most subjects have, and large numbers of people could test positive for, isn’t very useful for a screening test.

So, animal research up to this point has only shown that human adenoviruses (how many isn’t known) are associated with slightly higher fat deposits in infected animals. But, people aren’t animals, so what is the link to humans?

Human link to a “fat virus?”

The only published study to date linking the “fat virus,” Ad-36, to people was first described and presented as an abstract to IASO members in 1998, entitled “Evidence for an association of an obesity virus with human obesity at three sites in the United States.” [International J Obesity 1998; 22: S57 (unavailable to nonmembers)]

This study remains the only study of Ad-36 in humans and is the study being used to support the entire promotion of the human “obesity virus” and the antibody tests for it. This study wasn’t published until March, 2005 — and a lot happened in the interim. But before we get to that, let’s briefly look at this study.

Published in the International Journal of Obesity, it reported that “Ad-36 is associated with increased body weight and lower serum lipids in humans.” In brief, they selected 360 “massively obese” people who were dieting and enrolled in weight loss programs at clinics in New York, Florida and Wisconsin; and 142 lean people recruited from the University and area communities from 1995 to 1999. Among these groups, 11% of the lean people tested positive for the virus and 30% of the obese dieters tested positive.

Testing for Ad-36 antibodies, along with fasting blood lipids, they reported:

● a correlation between positive antibodies for the virus and higher BMIs: Those testing negative had average BMIs of 35.8 ±12.3 (standard deviation) and those testing positive had average BMIs of 44.9 ±16.3 (SD).

● a correlation between positive antibodies for this virus and lower total cholesterol levels, and triglycerides: Those testing negative had average total cholesterol levels of 5.5.1 ±0.07 (standard error) and those testing positive had average total cholesterol levels of 4.64 ±0.11 (SE).

Did you catch the fallacies and why this wasn’t a “fair test” to conclude the adenovirus was responsible for weight differences? Regular readers will recognize the most glaring ones, such as nonrandomized samples and ‘correlations aren’t causation’, but even these correlations were questionable.

The authors compared two very different groups of people, yet made no attempt to control for any confounding factor other than age. None. The weak correlations were made meaningless. They only looked at one correlation and could have chosen hundreds of other things to compare a group of younger university students and area adults, and a group of older, extremely ‘obese’ people undergoing weight loss interventions at medical centers. It’s easy to come up with countless clinically contradicting or meaningless correlations.

They made no effort to rule out nosocomial infections at those medical centers as a potential confounder. Did they look at the prevalence of virus antibodies among lean employees working at those weight loss centers, for example, for comparison?

They clearly didn’t consider dieting as a potential confounder. Nearly all extremely fat people in the United States have undertaken repeated diets. Did they consider that people dieting, their bodies stressed and being undernourished by insufficient calories, might be more susceptible to infections? Reverse causation would, in fact, be a much more logical explanation because undernutrition has been soundly shown to reduce immunities.

The lower cholesterol levels among the fattest people could not credibly be attributed to Ad-36. They are most readily explained by the temporary changes in cholesterol levels seen during caloric restrictions and dieting — but the authors didn’t control for caloric intake or weight loss. Another possible explanation could be viral infections, themselves, rather than some unique ability of the Ad-36. The authors, however, discounted the fact that virtually all infections have long been associated with lower total cholesterol levels. [Of course, few medical professionals would advocate giving people infections as a sound medical treatment to lower a surrogate health index! Would they?]

The second part of the study [for which Dhurandhar received a $1.88 million NIH grant] went on to test 178 selected sets of twins and found only 28 that differed in Ad-36 antibodies. Again not controlling for any confounders, they reported that the average BMIs among these 28 twins differed by a mere 1.6 units and total cholesterols by 0.02 mmol, but the standard deviations were so large (9.5-9.8 and 1.37-1.26, respectively) that there were actually no tenable associations at all. Further discounting the role of Ad-36 in the twin part of their paper, they admitted: “Most pairs were not obese and the length of time from exposure to the virus to our measurements could not be calculated.”

Despite the fact that neither of the groups included in their paper were randomized or representative of people in the general population, they went on to conclude that their data provide evidence of an association between Ad-36 and human adiposity.

Without conducting well-designed prospective clinical studies, of similar groups of people, this poorly controlled observational study failed to support the astounding media claims surrounding a human obesity virus that have been building for more than a decade. — Each flurry of media attention following a press release, of course.

This week’s media blitz — science by press release

It was impossible to miss the latest obesity virus media blitz. Big Liberty made the best observation that someone had sent out a press release. In an entertaining post she titled, “This is what a press release looks like,” she made screen shots of pages of headlining news articles that appeared last week about the obesity virus (with her own poignant annotations in red). Whenever every news outlet in the world all report the same story in lockstep, we are witnessing marketing, not science.

This week’s media flurry was about a small animal study, again by Dhurandhar’s group, that was e-published ahead of the print issue of the journal Obesity, the official journal of the Obesity Society. Professor Dhurandhar is on its Editorial Board. It was also timed with a BBC-2 special that aired on Monday night — “Why are thin people not fat?” — about catching obesity from a virus and the need to speed development of an anti-obesity vaccine. It featured professor Dhurandhar.

[Obesity researcher Dr. Rudy Leibel from Columbia University in New York, countered by explaining the genetic diversity among people and the “natural body weight” everyone is programmed to have. “Individuals have a biology which determines how tall or short they will be and how skinny or fat they will be, and wishing it one way or the other really cannot change it that much,” he explained.]

In this laboratory study, Dhurandhar’s group used cultured mouse preadipocytes and showed that infecting them with human Ad-36 resulted in the fat cells accumulating, and killing the virus stopped the process. This was merely a fancier way to show what is already known: that this human virus can promote fat cells in mice. It didn’t provide additional information for human obesity. But it did help capture media attention.

Stepping back and looking for the red flags

In nearly two decades, the science hasn’t moved forward much beyond showing a human adenovirus gives animals fat deposits. The media coverage far exceeds the merits of the science in providing health information we can use. Anytime we see that, it’s our red flag that something else may be up.

It’s impossible not to notice that all of the articles about infectious obesity and an Ad-36 “obesity virus” have originated from Dhurandhar and Atkinson. No independent scientists have replicated the findings. Even Tam Fry, of the UK was skeptical, telling the BBC last week that “this theory has been around for ten years and no one has come up with a comparable study to back this up.”

Not even Dhurandhar and Atkinson have been able to give a biological plausible explanation for how a common virus could make people fat, let alone animals. “The mechanisms by which Ad-36 increases adiposity in animals are not yet known,” they concluded in their paper in humans.

Many, if not most, scientists remain skeptical of an obesity virus having any meaningful role in human obesity or an “obesity epidemic,” not just because adenoviruses are so common, but because there has not even been an association shown between adenoviruses and obesity among any population.

In the Western world, these viruses are seasonal, self-limiting and short duration, according to doctors at the University of Pittsburgh who were examining the difficulties in developing effective antivirals. There’s no growing epidemic of adenoviral infections or correlations to population-wide changes in obesity rates. Quite the contrary. Adenoviral infections are endemic in Japan, China and Korea, they wrote. These countries are not hotbeds of obesity, either, with China’s rate at 3% and Japan’s at 3.4%.

At first, it appeared the intense media marketing and journal articles might be a case of “publication planning.”* But, despite news reports going back years saying that a vaccine for the obesity virus was just years away, a vaccine currently appears speculative and a long way off. There continues to be no clinical trials registered at clinicaltrials.gov involving Obetech, Atkinson or Dhurandhar, Ad-36 or an obesity assay screening test.

The media marketing might have a simpler explanation.

Even before the human study was published, Dr. Atkinson had moved to Richmond, Virginia and founded Obetech, LLC in August, 2004. Professor Dhurandhar moved to Baton Rouge, Louisiana, and became head of the country’s first department of viruses and obesity at Pennington Biomedical Research Center at Louisiana State University.

Obetech. Obetech almost immediately began marketing the first lab test for Ad-36 antibodies in people. Many medical professionals will probably be surprised to learn that the public has been sold, via the internet, screening tests for the “obesity virus” since 2004, without a single clinical trial. Obetech’s website, obesityvirus.com, tells consumers that obesity is due to a virus and that people infected with Ad-36 get fatter.

“Obetech has the technology to test humans for Ad-36 antibodies,” it states. “A positive test means that a person has been exposed to this virus and the infection may have contributed to weight gain or difficulty maintaining a healthy weight.” Its test is also advocated for “normal weight” people, because a positive test result “is a risk factor for gaining weight in the future, just like cholesterol testing,” alerting people that they’re at risk for developing obesity.

The solution for both fat and thin people who test positive for the virus is…

We recommend that people who test positive for Ad-36 should contact their personal physician right away to start a treatment program to prevent further weight gain or to prevent becoming overweight. The standard treatments are eating a healthier diet with lots of vegetables and fruits, reducing calorie intake, increasing physical activity to obtain 30-90 min of vigorous activity (such as walking) at least 5 days per week, and in cases where it they are indicated, starting drugs that may lower body weight or prevent weight gain. For people who qualify, obesity surgery may cause dramatic weight losses.

This screening test may be the most clever way to draw people to weight loss interventions yet.

Obetech’s Business Plan Summary states its “current products are three types of lab tests to detect Ad-36 infection.”

The US markets for the tests, the size of the markets, and the expected penetration in 5 years are as follows:

1. Pharmaceutical clinical trials: 100,000 subjects/yr; 50% penetration (50,000 tests/yr)

2. Consumers-direct: 297 million US population; 0.1% penetration (297,000 tests/yr)

3. Companion animals: 147 million US pets; 0.1% penetration (147,000 tests/yr)

4. Company funded worker wellness programs: 150 million; 0.05% penetration (75,000 tests/yr)

5. Donated blood: 14.6 million units/yr; 100% penetration (14.6 million tests/yr)

Currently the price of the tests is $195 to direct consumers and $100 to companies, doctors, and other third parties with large volumes of tests…

Long term, Obetech will develop a safe, effective Ad-36 vaccine with a worldwide target market.

Obetech’s financial projections were $5 million in revenue by early 2008. Under its Research and Development strategy, Obetech’s business plan says its “first research priority is to develop a rapid, semi-automated ELISA test to reduce the price for the lab-based test to about $50, then to develop an OTC test kit to be priced at about $30 for the general public…” [ELISA stands for enzyme-linked immunoassay, a laboratory test to detect antibodies.] Under its capital needs, it states:

Obetech seeks to raise $2 million in the next two years to support development of a semi-automated ELISA test and OTC test kit ($350,000), sales and marketing of our Ad- 36 test and kits ($850,000), R&D for vaccine development ($400,000), intellectual property protection ($150,000), and corporate administration ($250,000).

Dr. Atkinson, president of Obetech, is also President and co-founder of the American Obesity Association, the main lobbying organization of obesity-related interests**, and past president of the North American Association for the Study of Obesity. As we know, these two groups merged on September 5, 2006 to become the Obesity Society. AOA was founded in 1995 with its “central goal,” according to its mission statement, to “recognize obesity as a disease” and work for weight loss treatments to be a medical deduction and covered by Medicare, increase NIH funding for obesity and actively lobby for coverage of bariatric surgeries, promote “healthy” foods, and work to set the clinical care guidelines healthcare providers must follow to treat obesity. Professor Dhurandhar is also a Councilor for the Obesity Society and chair of its Clinical committee.

Patents. Professor Dhurandhar and Obetech hold two U.S. patents (#6,127,113 and 6,664,050) for viral obesity lab tests. As Obetech’s business plan also states, “patent applications or provisional patent applications have been submitted to broaden protection and to protect additional types of lab tests, vaccines, and other intellectual property.” The World Intellectual Property Organization also records Dr. Atkinson and Obetech’s international application, filed on December 27, 2005, for adenovirus “as a biomarker for disease” of obesity and “obesity-related” diseases.

Obetech also holds U.S. patent (RE039544), filed on April 4, 2005, for an assay test for the virus Ad-36p. The application says “a source of viral induced obesity has been discovered” and Ad-36 causes obesity:

For purposes of this application, a person is obese if the person's body mass index equals or exceeds 27 kg/m²and the person has excess adipose tissues. [Note the new, even lower cut-off for obesity.] … Understanding of the etiologies underlying obesity and the corresponding recognition that obesity is a disease eventually will lead medical insurance companies, which now at least in the United States typically do not recognize the condition as a disease, to recognize it as such and reimburse persons for diagnosis and treatment of it in the same way that the companies now do so for conditions that have long been recognized as diseases.

SUMMARY OF THE INVENTION. We have discovered that some obesity in humans is caused by viruses….In another of its aspects, the invention is an anti-obesity vaccine… the invention is a method of preventing obesity caused by a virus in a human susceptible thereto which comprises administering to the human an amount of an anti-obesity vaccine of the invention… the invention entails a method of reducing the serum levels of triglyceride, (total) cholesterol and low-density-lipoprotein-associated cholesterol... the invention entails methods for diagnosing whether a non-obese person is at risk of acquiring viral obesity, a method for diagnosing whether an obese person suffers from viral obesity, and a method of screening body fluids or organs and tissues (especially donated blood or donated organs or tissues) for the presence of obesity-causing adenoviruses.

Online screening tests

Remember the ovarian cancer screening test story? The development of a lab test involves the same rigorous clinical trial testing for safety and effectiveness, appropriate sensitivity and specificity, and used for conditions with proven medical interventions. A positive screening test labeling someone as being at risk for a disease has far-reaching effects and potentials for harm, including being subjected to costly, risky or ineffective interventions (all of which describe weight loss interventions); adversely affecting people’s insurance or employment status; and leading to the emotional stress and persistent anxiety that accompany being labeled as at risk.

The ovarian screening test had run afoul of the U.S. Food and Drug Administration when commercial interests stepped ahead of the science. Not only had the science been poor, the biomarker to have poor predictive value, and the test to not be clinically validated, the FDA found; the company hadn’t gotten approval from the FDA’s Office of In Vitro Diagnostic Device Evaluation and Safety (OIVD) before commercially marketing the lab screening test. The OIVD helps ensure that diagnostic products are safe and effective.

Even though Obetech has been marketing its obesity virus screening tests to the public for years, a search of the FDA’s IVD Over-The-Counter Database finds no record of FDA-approval for Obetech’s test or any test for an obesity virus or Ad-36.

The FDA has an article, “Buying Diagnostic Tests From the Internet: Buyer Beware!,” warning consumers about buying tests on the internet. It cautions that most IVD tests sold over the internet have not been approved or cleared for sale directly to the public. “FDA has cleared or approved many tests for use in a doctor's office or for professional use only, but internet marketers are selling them OTC or for unapproved uses,” they caution. It advises people talk to their doctors before purchasing any medical test on the internet.

The FDA also offers consumer tips for telling if a test is legitimate, as well as resources for anyone with questions or complaints about a product or website. “Don’t be fooled by a professional-looking website,” they say. You cannot tell by the popularity, credentials or prestige of a source if the science is sound.

An obesity virus and infectious obesity has made a good media story. But how many mainstream media, medical journals, or healthcare professionals have spoken out and tried to give the public information to help balance the marketing, to help protect people from being possibly taken advantage of or hurt, and to help prevent limited healthcare resources being spent on tests that haven’t yet been clinically shown to have validity and to offer benefits that outweigh the potential harms?

When it comes to obesity, consumers are left to protect themselves.

© 2009 Sandy Szwarc

* What is Publication Planning?

As Dr. Adriane Fugh-Berman, M.D., associate professor at the Department of Physiology and Biophysics at Georgetown University Medical Center in Washington, DC, described in Open Medicine:

Within the pharmaceutical industry, the term describes the finely calibrated process by which clinical trials, commentaries and other articles supporting the efficacy of particular products are written and released into the biomedical literature… industry uses publication planning to sway medical and public opinion through the medium of medical journals…

The controlled production and release of pre-clinical studies, clinical trials, reviews and commentaries may begin years before a drug is launched… Pharmaceutical companies cannot legally promote a drug before it has been approved by a regulatory authority, nor can they legally promote a marketed drug for off-label use. However, the US Food and Drug Administration (FDA) does not consider articles in the medical literature as promotional. As one industry article states, “Peer-reviewed publications offer pharma companies shelter from often-stormy regulatory waters. FDA views published articles as protected commercial speech so doesn’t regulate their content.”

For this reason, the generation of subtly persuasive opinion pieces that can be distributed to prescribers in the pages of medical journals is an extremely important component of publication planning. Sponsored articles can be difficult for journal editors and readers to spot… For example, if a drug is the only treatment for a given condition, articles that review the prevalence, severity or complications of that condition will prepare the market by raising physician awareness of specific issues… Reviews and commentaries are the Trojan horses bearing these messages.

Publication planning, as it is currently practised by pharmaceutical companies, can undermine the medical literature. Industry control over the timing, content and authorship of studies and opinion pieces including reviews and commentaries distorts medical discourse… Sponsored writing reflects sponsored messages. Even the most vigilant editor could not uncover all of the marketing messages embedded in manuscripts by publication planning professionals.

**