Pharma’s Cutting Edge

As you’re likely to have read by now, FDA is holding an advisory committee hearing tomorrow, June 13^th, to discuss rimonabant (formerly Acomplia; now Zimulti?), the cannabinoid CB1 receptor inverse agonist developed by Sanofi-Synthelabo, now Sanofi-Aventis. The rimonabant NDA was granted an approvable letter for the weight loss-related indications in February 2006. The advisory committee is being held to review in a public forum all information on the drug, including the additional information S-A supplied to address the issues raised in the approvable letter, which were heretofore undisclosed publicly.

I see that no shortage of pundits is commenting on the briefing documents FDA released yesterday on its web site and the markets have already reacted. That’s okay. Not everything deserving of a post by me is necessarily uninteresting to the mainstream press. So, let’s have a quick look and await the outcome together.

 The efficacy of rimonabant has not really been in doubt since the sponsor began releasing data from its RIO pivotal trials in 2005.  The drug causes people who are obese to lose weight, probably through diminished appetite and food intake, although I don’t believe the specific effector mechanism(s) of weight loss, nor the exact pharmacological actions accounting for it, has been demonstrated in clinical studies. The weight-loss effect is dose-related and is clinically meaningful at 20 mg daily. It’s also a modest effect with a 1 year weight drop of about 11 pounds overall relative to placebo (the heaviest subjects, those with a BMI above 40, lost the most weight), which was similar in the 4 pivotal studies. In year two a good bit of the weight lost was regained. In those with lipid disorders, hypertension or diabetes, the 20-mg dose of rimonabant either improved or did not worsen surrogate markers of benefit. The antihyperglycemic effect in the RIO Diabetes study was particularly good, with placebo-subtracted HbA1c reduction of 0.7%, despite a higher proportion of rimonabant users using lower doses of concomitant metformin or sulfonylurea.

Not surprisingly, the safety profile of rimonabant has been slower to emerge.  Earlier disclosures did not include an integrated look at all treatment groups side by side. When viewed this way, the distinction between the 20-mg dose and placebo re psychiatric events psychiatric events is clear. A total of 25 preferred AE terms related to psychiatric symptoms were reported by at least 5 subjects in the 20-mg treatment group. A preferred term represents a coded adverse event, essentially a coded translation of a verbatim AE report, although not one that is as rich in information content. Of the 25 preferred terms, the frequency of reported AEs was numerically higher in the 20-mg rimonabant group as compared with placebo for 24 of them. Only 1 category (decreased libido) was associated with a lower frequency of AE reports by subjects assigned to rimonabant. The relative risk of a psychiatric AE was roughly two for the 20-mg group compared with placebo when all events were combined. Also, rimonabant was associated with nearly four-fold higher frequency of associated psychiatric symptoms not captured under the preferred terms per se. Te above is convincing evidence of a generalized CNS folks; I don’t recall a drug not intended to treat psychiatric disease with such a one-sided AE profile. That said, a list of adverse event frequencies hardly gets at risk; it speaks volumes about likely tolerability and propensity to prescribe, but the committee is going to be focused on the benefits of rimonabant in relation to its risks (the draft questions of the committee have also been published).

Apparently, in its 2006 approvable letter, FDA asked Sanofi-Aventis to investigate suicidality further and asked for data to support a meta analysis by the Columbia University group of Posner.   The committee had done this recently with SSRIs in pediatric patients.  S-A did a concurrent analysis using similar methods.  Both studies found essentially the same thing: rimonabant doubles the incidence of suicidal ideation (i.e. thinking about suicide).  S-A undertook additional analyses to demonstrate that nearly half (40%) of the cases of suicidal ideation were reported among patients who self-reported depression at baseline, regardless of assigned therapy.  Presumably, S-A is looking to see if they can get by with restrictions against prescribing rimonabant in depressed patients if the alternative is non-approval.

A number of other related safety issues will be discussed tomorrow, but the bottom line for most of the panelists will undoubtedly be determined by their assessments of the risk of suicide versus the benefit of weight loss and its attendant benefits (on diabetes primarily).  They will be reminded, probably numerous times, that S-A is in the midst of conducting a large cardiovascular outcomes study with rimonabant.  Should they wait before seeing its results before giving the thumbs up?  Or should they give their blessings now and argue for strenuous warnings against its use in depressed patients?  What type of risk management strategy can guard against excessive population risk of suicide?  Would it have to look just like the Accutane strategy?  If so, would S-A even launch under such restrictions?  This is a  fascinating story waiting to play out.

June 12th, 2007 | Category: Drug Safety, FDA, For Investors | Subscribe to comments | Leave a comment | Trackback URL