Nutrition retracts 2001 paper

Study by Chandra had been questioned since 2003

Mar 3, 2005
Doug Payne(dougpayne@islandtelecom.com)

Nutrition this month retracted a 2001 paper by immunologist Ranjit Chandra that claimed a specific combination of vitamins and minerals significantly improved seniors' ability to think and reason.

Editor Michael Meguid wrote that the journal was "forced to act" after a series of "serious questions" were raised. The retraction outlines eight specific reasons for the decision, "all of which [Chandra] either ignored or dealt with inadequately in his responses to his critics," according to Meguid, over whose signature the withdrawal appears.

The retraction cited significant statistical errors in Chandra's Nutrition study. It also cited errors in his 1992 study in The Lancet, which the 2001 study followed. Among other concerns were that while Chandra claimed that the participating subjects were normal, "the average MMSE score reported placed them below normal, in the demented category" and that while assignment of subjects to the placebo and supplement groups was said to be randomized, "there were statistically significant differences between the two groups on seven measures at the start of the experiment."

The paper—done when Chandra was working at Memorial University in Newfoundland—had been published by Nutritionafter being turned down by the BMJ, which said that its statistical reviewer had found the study showed "all the hallmarks of being entirely invented." The Nutrition study had been cited at least 16 times, while The Lancet study had been cited more than 350 times.

"A group of scientists and investigators found some of the claims made [were] implausible, not reproducible, that the basis on which the data was analyzed was not appropriate and could not yield the results claimed," Meguid told the Canadian Broadcasting Corporation. Contacted by The Scientist, Meguid declined further comment.

There was also concern that "Chandra failed to declare that he holds a patent on the tested supplement formula and has a financial stake in it because the supplement was licensed to Javaan Corporation, a company founded by his daughter, that sells the supplement," Meguid wrote in the retraction.

Although Javaan's toll-free order line was working this morning, the company's main listed number was not in service when The Scientist sought comment. Chandra, now living in India, did not respond to E-mails seeking comment.

Last year, the Canadian Institutes of Health Research (CIHR) also tried to look into the concerns over Chandra's study, but Chandra continued to refuse to provide his research data to Memorial University, "so the investigation is incomplete," said CIHR spokeswoman Janet Weichel-MacKenzie.

When BMJwas raising its concerns last year, then-editor Richard Smith said, "I have urged the university to investigate," something that the university said was not its role.

"We're pleased that [Nutrition] is taking some responsibility," Memorial University spokesperson Jack Strawbridge told The Scientist. "If there was a problem with the paper that was published, then the problem was with the vetting process of the journal. [Meguid] gives a very detailed set of reasons why they are doing this, and I think that the reasons are a fairly brave statement on the part of the editor."

"Our point has been, all along, that we have a responsibility to create conditions that allow research to happen, but we don't vet it directly; we don't say that any piece of research done by any particular researcher should or should not be published. That's the role of the peers and the journal editors," Strawbridge said. "If The Lancet did something in parallel, it would really put a stronger face on the responsibility of the journals in a case like this."