Free Shipping on All Orders $75 Or More!

Your Trusted Brand for Over 35 Years

Life Extension Magazine

<< Back to June 2006

Media Bias, Conflicts of Interest Distort Study Findings on Supplements

June 2006

By Lyle MacWilliam, MSc, FP

Sloppy reporting, distorted editorial sensationalism, and conflicts of interest by researchers are unnecessarily alarming the public and threatening to destroy our trust in complementary health care. These injustices must be addressed before irreversible harm is done to an industry committed to natural approaches to wellness and to a public increasingly confused about where to turn for sound advice on preventing disease and achieving optimal health and well-being.

On February 23, the venerable New York Times—long regarded as an icon of journalistic integrity—trumpeted the headline, “2 Top-Selling Arthritis Drugs Are Found to Be Ineffective.” The “drugs” mentioned in the article are not drugs at all, but glucosamine and chondroitin, two popular, natural nutritional supplements successfully used by millions of people worldwide. Despite the fact that the study authors came to a distinctly different conclusion, this one article quickly raised questions in the public’s mind. The Associated Press and its Canadian counterpart, Canadian Press, also picked up the story and parroted the headline that glucosamine and chondroitin sulfate were “… no better than dummy pills” in relieving knee pain associated with osteoarthritis. The findings of the $12.5 million study, published February 23 in the New England Journal of Medicine, provide a very different view.

Preceding this was another egregious headline in New York Times that likely sold a lot of newspapers. On February 16, the Times declared, “Big Study Finds No Clear Benefit of Calcium Pills.” Reporting on an $18 million study conducted by the Women’s Health Initiative on the protective effect of calcium and vitamin D, the Times article dismissed the finding that those women who actually adhered to their supplementation regimen experienced a 29% reduction in hip fractures. Such a reduction is rarely achieved with the strongest pharmaceuticals, yet here was a natural supplement producing enviable results and the media failed to report it accurately.

A week prior, on February 9, an Associated Press newswire declared, “Palmetto No Help for Prostate.” Reporting on a study conducted by researchers at the San Francisco VA Medical Center, news media throughout the US and Canada regurgitated the story that the popular herbal remedy saw palmetto is of no help in relieving symptoms of an enlarged prostate. According to the New York Times, the study found “no benefit from saw palmetto by a variety of different measures.” Canada’s national daily, the Globe and Mail, condemned the extract as “no more effective than dummy capsules in easing symptoms.” The Los Angeles Times advised that men “might be better off taking FDA-approved medications.” Nowhere in these superficial and slanted accounts was there any more than a passing mention of the study’s substantial limitations.

On February 8, the New York Times reported on two related Women’s Health Initiative studies claiming that low-fat diets do not cut health risks. The studies, part of a multifaceted $415 million federal research project involving nearly 49,000 postmenopausal women, investigated the effects of a low-fat diet in reducing the risks of cardiovascular disease and colon cancer.

According to the Times article, low-fat diets have virtually no effect on invasive colon cancer, heart attacks, or strokes. No attention was paid to the studies’ numerous design weaknesses, the most obvious of which are their common failure to discriminate between different types of fat and their handicapped ability to detect change—known as the power of a test—due to actual reductions in fat intake that were far less than anticipated. These considerations, along with the fact that the studies were based on generally overweight postmenopausal women, render the findings inapplicable to the general population and belie the implications of articles describing the study results.

Not to be outdone, the February 28 edition of Canada’s Globe and Mail followed up on the findings of the Women’s Health Initiative fat-loss studies. Declaring “everything you know about your health is wrong (again),” journalist Margaret Wente espoused her bias that “despite everything you’ve been told for umpteen years by countless experts, the experts were wrong.” According to Wente, low-fat diets are of no benefit at all and the only people who benefit from calcium and vitamin D are those who work in the supplements industry. “Salads and supplements are useless in warding off the deadly diseases we all dread,” noted Wente. “Give up health advice,” she further advised. “You’ll feel 100% better in no time.”


Statistical power refers to the probability that one can detect an effect if, in fact, there is one. It is influenced by the size of a study (the number of subjects) as well as the variable measured (the endpoint of the study). Sometimes studies are discounted if they show elevated, but not statistically significant, effects. Therefore, one should be careful of dismissing a possible association on the basis of negative study results alone.2

Jumping to False Conclusions

A journalist’s obligation is to tease fact from hyperbole and truth from innuendo—particularly when it comes to issues of public health. In this regard, the articles just described are seriously wanting. That said, the media cannot be held solely accountable for the needless confusion and fear sown by these distorted accounts. Also to blame are those within the scientific community whose desire for 15 minutes of fame—not to mention future research funding from Big Pharma—displaces their scientific judgment.

Such is the case in a recent study showing that modest reductions in homocysteine did not reduce heart attack risk in those with significant pre-existing arterial disease. Last September, at the 2005 European Society of Cardiology Congress held in Stockholm, Dr. Kaare Bonaa’s pronouncement that “the homocysteine hypothesis is dead”1 certainly drew everyone’s attention. According to Dr. Bonaa, the study results “tell doctors that prescribing high doses of B vitamins will not prevent heart disease or stroke.”1

At the time of Dr. Bonaa’s pronouncement, the trial had been neither peer reviewed nor published. A detailed critique of this study is provided in this issue in the article, “Mainstream Doctors Still Confused About Homocysteine.” If we are to believe the conclusions of Dr. Bonaa, the homocysteine hypothesis is dead. The question is, should we believe Dr. Bonaa when scores of published studies suggest the exact opposite conclusion?

Factors Behind the Study Results

So what is going on here? Why the sudden blizzard of published studies that appear to refute the majority of the scientific evidence on the benefits of natural approaches to wellness?

For one thing, several long-term Women’s Health Initiative dietary intervention studies, developed in the early 1990s, are now coming to fruition. These studies failed to separate “good fats” (omega-3 fats and monounsaturated oils, such as fish oil and olive oil) from “bad fats” (trans fats), although much about their differences was known even at the time of the studies’ design. Furthermore, the studies make no attempt to reconcile the balance of omega-6 and omega-3 fatty acids, believed by many health experts to be a critical factor in inflammatory disease risk. According to these studies, fat is fat, and reducing fat means cutting down on all fat. If anything, the studies’ mixed findings simply serve to demonstrate the folly of such an indiscriminate approach.

Second, when testing an association or effect, statistics tell us that one time out of every twenty times the association or effect may seem to be real, but in fact is not. There is always bound to be a statistical fluke in the bunch.

Third, it is not unusual that clinical studies investigating a particular effect will not have the needed number of subjects to show a statistically significant result. This occurs because in most clinical trials, the probability of detecting a difference between variables, known as the “power” of a test, is set at 90%, usually with a minimum “power” of 80%. Consequently, there may be a 10% or 20% chance of missing your mark and failing to find a difference when one in fact exists. This is merely the gremlin of statistical probability at work.

Finally, some investigations are just bad science—improperly conducted, poorly reported, and inadequately reviewed. As has been the case lately, it is these studies that attract the undue attention of a news media hungry for sensational headlines.

Just what do the latest studies tell us? To find their real message, we need to look beyond the headlines, the skewed media spin, and the superficial analyses that pad these “news” stories.


A recent headline in the OB/Gyn News declared, “Vitamin E Shown Not to Reduce Cardiovascular Disease.” The article discussed the findings of a recent study of vitamin E, as reported in the Journal of the American Medical Association in July 2005.

The study’s objective was to test whether vitamin E supplementation decreased the risks of cardiovascular disease and cancer among healthy women. According to the article, the study concluded, “vitamin E showed neither benefit nor harm in all clinical parameters examined.”

However, if you read the study’s fine print, an entirely different picture emerges. The study also found that for cardiovascular death, there was a markedly significant 24% reduction in risk. This is a huge decrease in death!

So why was this finding not heralded as a significant discovery and headlined in all the major dailies? Because cardiovascular death was not one of the pre-specified clinical parameters set up by the study (although it was a component of a composite parameter). Instead, the authors concluded that vitamin E supplementation is not recommended for cardiovascular disease prevention, despite the fact that it reduced cardiovascular disease death by 24%.

Glucosamine/Chondroitin Findings Ignored

The Glucosamine/chondroitin Arthritis Intervention Trial (GAIT),3 hailed as the largest-ever clinical study of these supplements, was supposed to be the definitive word on the effectiveness of glucosamine and chondroitin in reducing the pain of osteoarthritis. Instead, the study results have only generated more controversy, due, in part, to poor experimental design and the media’s misrepresentation of the findings.

The trial was a randomized, double-blind, placebo- and celecoxib (Celebrex®)-controlled intervention trial with 1,583 patients with symptomatic osteoarthritis of the knee. The primary outcome was a 20% reduction in knee pain over 24 weeks. From a clinical perspective, the study appears well designed, with a projected 85% probability of detecting change and high adherence to the treatment protocol.

Unfortunately, an inordinately high placebo effect of 60.1%, which almost doubled the expected rate of 35%, virtually destroyed the trial’s validity. The fact that 6 of 10 patients in the placebo group found significant pain relief from a dummy pill is an enormous placebo effect!

Another issue is the form of glucosamine used in the study. While glucosamine sulfate is the standard form used in supplements, the type used in this study was glucosamine hydrochloride. This form of glucosamine does not contain the sulfur moiety, found in the sulfate part of the glucosamine sulfate molecule, which may amplify its analgesic properties.

Finally, little was mentioned about the potential confounding effects of the use of pain relievers such as aspirin and acetaminophen. Despite the well-known fact that acetaminophen enhances the efficacy of osteoarthritis treatment, researchers allowed patients to take up to 4000 mg of acetaminophen daily, a decision that likely contributed to the outsized placebo effect noted previously.


One of the flaws of the New England Journal of Medicine study may have been that the form of glucosamine used did not provide any sulfur.3

Animal studies have shown that joints affected by osteoarthritis have lower sulfur content,4 and that arthritic mice given the sulfur-containing nutrient MSM (methylsulfonylmethane) experience less joint degeneration.5 In a double-blind trial in people with osteoarthritis, study participants who received MSM alone experienced significant pain relief.6

In a study published in 2004, the combination of glucosamine and MSM was more effective in improving the signs and symptoms of osteoarthritis than either agent alone.7 After 12 weeks of treatment, the average pain score in the glucosamine-only group dropped from 1.74 to 0.65, a 63% reduction. In the MSM-only group, it fell from 1.53 to 0.74, a 52% reduction. However, in the group taking glucosamine and MSM, the average pain score dropped from 1.7 to 0.36—an astounding reduction of 79%! The researchers also found that the combination therapy had a faster effect on pain and inflammation than either glucosamine or MSM alone.

Despite these limitations, the study did find that for those individuals with moderate-to-severe knee pain, the combination of glucosamine and chondroitin sulfate provided a 25-26% improvement in pain relief—a response that exceeded the projected 20% design measure to prove efficacy. According to the study authors, “treatment with chondroitin sulfate was associated with a [statistically] significant decrease in the incidence of joint swelling, effusion, or both.”

In fact, for those participants with moderate-to-severe pain, the only treatment that bore significant benefit was the combination of glucosamine and chondroitin sulfate, which outpaced the anti-arthritis drug Celebrex® by a large margin.

Despite these findings, newswire coverage of the study chose to celebrate the positive effects of Celebrex®. In fact, Celebrex® did not relieve arthritis pain in those needing it the most—patients with moderate-to-severe pain—whereas the combination of glucosamine and chondroitin was effective in this group of patients. At the same time, these news articles claimed that “nutritional supplements show no overall benefit in treating arthritis.”

It appears that the news media took its cue from an editorial appearing in the same issue of the New England Journal of Medicine.8 Not once did the news coverage mention that the author of this editorial criticizing the use of glucosamine and chondroitin is someone who has received financial compensation from Pfizer, the maker of Celebrex®. Nor did the news media explore the disclosure that a number of the authors of the GAIT study received compensation from Pfizer and McNeil Pharmaceuticals (the maker of Tylenol®).


Here’s a plan: take $18 million of taxpayer money and 36,000 postmenopausal women in various stages of osteoporosis. Divide them randomly into two groups and give one group a sugar pill and the other a pill containing low amounts of vitamin D and a poorly absorbed form of calcium.

Complicate the study design by mixing together women who are already taking calcium and still others who are on hormone replacement therapy and other forms of supplements.

Now, tell the intervention group to take their pills every day—but don’t worry if they don’t take them every day, because you’ll include their results anyway.
What definitive study result do you get? None.

Are you surprised? You shouldn’t be.