Lancet Questions Antioxidants' Cancer-Preventive Abilities

NIS, Serbia & Montenegro--National and international media trumpeted results of a meta-analysis published in the Oct. 1 The Lancet (www.thelancet.com), in which researchers from University of Nis here concluded antioxidant supplements are not effective in protecting against gastrointestinal (GI) cancer (364:1219-28, 2004). In addition, editorial comment from two English researchers was included on the cover of the magazine: "The prospect that vitamin pills may not only do no good but also kill their consumers is a scary speculation given the vast quantities that are used in certain communities." News headlines ranged from the relatively moderate ("Vitamin Pills Don't Cut Stomach Cancer Risk" on Reuters) to extreme ("Vitamins Only Take You Closer to Death" on Asian News International).

For the meta-analysis, led by Goran Bjelakovic, M.D., researchers included 14 randomized trials, representing 170,525 individuals, from 681 references originally identified. The analysis incorporated data on beta-carotene, vitamin A, vitamin E, vitamin C and selenium, and compared it to outcomes including separate or combined incidence of esophageal, gastric, colorectal, pancreatic or liver cancers. Neither the fixed-effect nor random-effects meta-analyses showed significant effect of antioxidants alone or in combination on GI cancer outcomes; however, using the fixed-effect model on seven "high quality" trials, the researchers found antioxidants significantly increased mortality.

The researchers noted the detrimental effects of antioxidants on mortality were "unexpected," but added the results suggest, "For every million people exposed to toxic combinations or amount of antioxidant supplements, about 9,000 premature deaths could have occurred." Immediately after this suggestion, the researchers list the array of potential limitations with the analysis, including inconsistencies in reports between trials, high dropout rates, inability to corroborate information with original researchers, and use of populations at high risk of cancer (i.e., smokers, patients with previous history of skin cancer, hepatitis B surface antigen carriers).

These drawbacks were noted in the editorial comment, authored by David Forman, University of Leeds, and Douglas Altman, Centre for Statistics in Medicine, Oxford (364:1193, 2004). The commentary noted the results of the GI cancer study--particularly the mortality issue--must be regarded as preliminary and "[do] not offer convincing proof of hazard." Forman also wrote, "Heterogeneity arose almost wholly from the outlying result from one trial in an anomalous population of smokers, ex-smokers and occupationally exposed asbestos workers. The other high-quality trials show little heterogeneity, and do not suggest increased mortality." However, it was the "killing consumers" comment that was used on the Lancet cover and picked up by most media outlets.

The Council for Responsible Nutrition (CRN) immediately issued a response to the Lancet articles, taking the journal to task for highlighting the commentary quote and not emphasizing that the study results are preliminary. "By using a single incendiary and scientifically inaccurate quote from the commentary in not only their press release, but also as the sole focus of their cover, The Lancet has effectively moved from prestigious scientific journal to tabloid-style journalism in an effort to create visibility for their magazine," the association wrote.

Annette Dickinson, Ph.D., CRN president, noted it is not a surprise that science doesn't yet know what compounds might prevent cancer. "At the same time, it's unfortunate that of the 14 studies reviewed in the meta-analysis, only three focused on healthy people," she said. "Antioxidant supplements in and of themselves should not be expected to reverse the negative effects created by a lifetime of smoking or poor dietary habits."

During SupplySide West in Las Vegas, featured speaker Jeffrey Blumberg, Ph.D., from Tufts University, also addressed the Lancet article, echoing CRN's concerns about the analytical methods. "It's important to find out how people are weighing the evidence," he said. Blumberg broke down the types of studied included, particularly given the use of GI cancer as a secondary or tertiary consideration. He also noted Lancet's decision to use the editorial commentary quote on the cover of the magazine should give readers pause as to the journal's possible bias.