While Sanofi's new diet drug Acomplia (rimonabant) may help with some cardiometabolic risk factors, lowering triglycerides and slightly increasing good cholesterol, a comprehensive review in the journal Circulation suggests Acomplia is no more effective than older diet pills in reducing bad cholesterol.

"Although some features of metabolic syndrome have been shown to improve modestly, no reduction in LDL-C occurs, although this appears to be the case with all centrally acting antiobesity drugs," reports Dr. Kishore M. Gadde, a Duke University obesity researcher.

"It is rather disappointing that weight loss achieved primarily with the help of centrally acting antiobesity drugs does not lead to reduction in LDL-C (bad cholesterol) in most studies, and rimonabant is yet another example of this," said Gadde. "In contrast, plant-based and high-fiber diets have been shown to
reduce LDL-C significantly and as much as the statins in some studies. . ."

Gadde's also raises detailed questions about Acomplia's psychiatric side effects -- which he recently surfaced in the Journal of the American Medical Association -- while concluding Acomplia is no better for weight loss than "the modest effects observed with currently approved antiobesity drugs."

Writing in the Aug. 28 issue of the prestigious journal, Gadde takes a look not just at all published trial data for Acomplia, but at the most recent information available on the three diet drugs currently approved for U.S. sale -- Meridia (sibutramine) and Xenical (orlistat), both approved for long-term obesity treatment, and phentermine.

Acomplia, while approved for sale in Europe, has still not been approved for sale in the United States by the Food and Drug Administration for reasons Sanofi has chosen not to divulge.

In his review, Gadde acknowledged that in studies with 6,600 overweight and obese patients, rimonabant demonstrated "consistent success" in aiding weight reduction, and "offers a novel mechanism of action, which may make it well suited as an alternative for people who do not respond well to other agents and for combination treatment with other antiobesity agents."

But he further elaborated on a point he made in his letter to JAMA: "Although rimonabant appeared to be reasonably well tolerated in general, psychiatric symptoms were the most common adverse effects that led to early withdrawal of patients in . . .trials."

In the Circulation review, Gadde noted:

• In the RIO-Europe trial, "almost half (42 of 87) of the adverse event–related dropouts in the rimonabant 20 mg group were attributed to psychiatric events, including mood alterations."

Mood was assessed with the Hospital Anxiety and Depression Scale (HADS)58 at baseline and every 3 months. HADS contains 7 items to assess depressive symptoms and 7 items to assess anxiety symptoms, with scores ranging from 0 to 3 for each item. Depressive and anxiety symptoms are separately totaled. Scores of 0 to 7 are considered normal, 8 to 10 represent borderline symptoms, and (equal to or over) 11 are considered significant enough to warrant further assessment of the patient. Questions probing suicidal thoughts are absent in HADS. Thus, HADS is generally not used as the primary outcome measure in clinical trials of depression but is considered an acceptable tool to screen for depression and anxiety in primarily nonpsychiatric patients. In RIO trials, when HADS scores reached 11 during the treatment period, patients were required to be seen by a psychiatrist for further assessment. However, none of the published papers reported the number of subjects who were removed from RIO trials after psychiatric consultations. Whereas the RIO-Europe report stated that there were no significant changes in HADS scores for depression or anxiety, given the above-described methodology of removing patients with mild depressive symptoms, one would wonder if the LOCF (last observation carried forward) imputation method might have underestimated mood changes over time. One would have liked to have known the proportions of patients by treatment with scores reaching (equal to or over) 11 or those with HADS scores that increased by (equal to or over) 25% relative to baseline.

• In the RIO-Lipids trial, "psychiatric adverse events in RIO-Lipids accounted for half (26 of 52) of the adverse event–related dropouts in the rimonabant 20-mg group compared with one third (8 of 24) of the withdrawals in the placebo group."

The authors of this article presented mood changes separately as “depression,” “major depression,” and “depressed mood,” when all these could have been combined as “depression”; if so combined, 14 patients in the rimonabant 20-mg group had treatment-emergent depression versus 2 in the placebo group. According to the lead author (personal communication, January 2006), patients with HADS scores (equal to or over) 11 were not excluded from participation in the study, but when the HADS score reached 11 or higher during the study, the patient was referred to a psychiatrist, and the study drug was discontinued for patients needing treatment with antidepressants. It is unclear how patients with HADS scores (equal to or over)11 were allowed to participate in the study when this score was considered high enough to require the patient to be seen by a psychiatrist during the course of the study.

• In the RIO-North America trial, "as seen in other RIO trials. . . more patients dropped out owing to psychiatric adverse events with rimonabant 20-mg treatment (6.2% versus 2.3%).

"Although rimonabant appeared to be reasonably well tolerated in general, psychiatric symptoms were the most common adverse effects that led to early
withdrawal of patients in RIO trials," Gadde concluded.

"Given the exclusion of patients with a past history of significant depression and current presence of even mild depressive symptoms (as reflected by normal mean HADS scores at baseline) and the removal of patients from the RIO trials after a slight increase in depressive symptoms, it is unclear whether this drug is suitable for obese patients with coexisting depression," he added.