Is the new diet drug Acomplia (rimonabant) -- already approved in Europe and awaiting FDA action in the United States -- safe for the many obese individuals who suffer from varying degrees of depression?

A Duke University psychiatrist and obesity researcher, writing in the August issue of the Journal of the American Medical Association, contends that the clinical trials conducted to date do not provide a satisfactory answer.

Dr. Kishore M. Gadde noted that "psychiatric adverse events have accounted for about half of all early terminations attributed to adverse events with the 20 mg dose of rimonabant in all (Acomplia) trials published to date."

The director of Duke's Obesity Clinical Trials Program also noted that participants in the Acomplia trials had a mean subscore for depression of approximately 3 on the Hospital Anxiety and Depression scale, "suggesting that efforts were made to enroll individuals with minimal or no depressive symptoms."

Scores of 0 to 7 on this scale are considered within the normal range, a score of 8 to 10 suggests borderline symptoms, and 11 or higher suggests probably significant symptoms of depression.

"The prevalence of depression is high among obese persons seeking treatment," Gadde said.

"The question then is whether the psychiatric safety data of rimonabant presented in the reports of the (rimonabant) trials can be generalized to real-life clinical practices in which most obese patients are likely to have a Hospital Anxiety and Depression scale subscore for depression far higher than 3," Gadde said.

^"There is not sufficient evidence from the (rimonabant) trials to determine whether rimonabant is safe for use in obese individuals with even mild depression," he concluded.

In a reply to Gadde, Dr. F. Xavier Pi-Sunyer, lead researcher in the North American Acomplia trial, said patients were not excluded from his study of rimonabant based on their baseline Hospital Anxiety and Depression score.

"Most patients (who started the trial) with a normal subscore for depression of 7 or less at baseline had a normal subscore at follow-up," Pi-Sunyer said. But he conceded that "a small proportion, equally distributed in the placebo group and in the 20 mg of rimonabant daily group, shifted" from the normal range to the high range in the course of the trial.

He also said that among participants who started the trial with a depression subscore in the borderline range, "similar proportions in the 20 mg of rimonabant and placebo groups" moved to the high range during the course of the trial.

He said, however, that two thirds of the patients who started the trial with a high depression subscore moved into the borderline group, compared to half of the trial participants taking a placebo.

"Thus, no overall impairment in the Hospital Anxiety and Depression scale depression subscore was observed in patients in the rimonabant group compared with the placebo group," Pi-Sunyer concluded.

He conceded that the number of patients taking rimonabant who "withdrew for depressed mood" was almost double the number taking a placebo, but said "we did not record if this was following psychiatrist evaluation or investigator's judgment."