"Our study ... exposes the failure of peer review," Dr. Veronica Yank of Stanford University and colleagues write in BMJ Online First for November 16. "Editors and peer reviewers, as well as policymakers, meta-analysts, and readers should closely scrutinize the conclusions of meta-analyses to ensure that they are supported by the data," they conclude.

The researchers investigated financial ties and concordance between results and conclusions in 124 meta-analyses involving anti-hypertensive drugs published through December 2004. Forty-nine (40%) of these meta-analyses had financial ties to one drug company.

Dr. Yank and colleagues found that meta-analyses with favorable conclusions, but not favorable results, were more likely to have financial ties to one drug company than other ties, even when controlling for other characteristics of meta-analyses.

Specifically, they report that 27 of 49 (55%) meta-analyses funded by a single drug company had favorable results, but 45 of 49 (92%) had favorable conclusions.

In contrast, meta-analyses with financial ties to non-profit groups had "excellent concordance between results and conclusions."

These findings, Dr. Yank and colleagues say, suggest discordance for drug company-sponsored studies between the data that underlie the results and the interpretation or "spin" of these data that constitute the conclusions.

In an editorial, Richard A. Epstein of the University of Chicago Law School says this study offers "further proof of the potential influence that the drug industry has on the outcomes of the studies they fund."

While original data are apt to be sound, "conclusions should be interpreted with caution," he notes.

However, the suggestion that drug companies should have a more restricted role in financing clinical trials, brings up a dilemma, Mr. Epstein writes. He asks: "Do we want fewer studies of presumably better quality, or do we want more studies whose quality may be more biased? I would opt for the last option."

"Nothing in the work of Yank and colleagues suggests that the raw data from the drug sponsored studies were defective," he explains. "The criticisms are directed to the optimistic inferences dhttp://cmstools.medscape.com/mta/viewindexfromsearch?ArticleID=566069 Edit Meta Datarawn from the data."

Therefore, Mr. Epstein suggests that a "sensible approach" might be to encourage further dialogue by asking for editorial comment. These editorials might be published in different journals than the original work. "In all likelihood, these critiques will subtly induce original authors to soften their basic claims."

BMJ Online First 2007.