Last Updated: 2009-10-26 17:53:54 -0400 (Reuters Health)

NEW YORK (Reuters Health) – Many randomized controlled trials reported in top medical journals do not report harms, or else do so in a fragmented or suboptimal way, according to a systematic review published in the October 26 issue of the Archives of Internal Medicine.

According to editorialist Dr. John P. A. Ioannidis, that’s despite the view that "differences in the profile of harms should play a key role in treatment choice."

For their review, senior author Dr. Philippe Ravaud and colleagues at Groupe Hospitalier Bichat-Claude Bernard in Paris searched MEDLINE via PubMed to identify all reports of randomized controlled trials published in 2006 in six general medical journals with a high impact factor (New England Journal of Medicine, The Lancet, JAMA, BMJ, Annals of Internal Medicine, and PLoS Medicine). Their analysis was restricted to trials with 2 parallel arms.

Based on their assessment of 133 articles, the researchers warn that "reporting of harm remains inadequate."

"Descriptions of adverse events with numerical data in each trial arm were missing in 18% of the reports, and information related to the severity of adverse events and the withdrawal of patients owing to adverse events was lacking in 27.1% and 47.4% of reports, respectively," they said.

Furthermore, roughly a third of the reports described only the most common events or the most severe or statistically significant results. Dr. Ravaud and associates note, however, that randomized controlled trials are "well known to have insufficient statistical power to assess safety outcomes."

Many articles also failed to clearly report the population considered for safety analyses.

In his editorial, Dr. Ioannidis, from the University of Ioannina School of Medicine, Greece, points out that many articles combined information on different harms per organ in composite outcomes, in "a stretched effort to gain power" at the expense of validity.

He continues, "A composite of neurological toxic effects may include harms as disparate as occasional numbness in the fingertips and deep coma."

Many authors blame printed space limitations for their inability to provide comprehensive lists of adverse events. In answer, Dr. Ioannidis maintains, "Web supplements can cater to any amount of information."

Arch Intern Med 2009;169:1737-1739,1756-1761.