Thursday, March 25, 2010

Saphris®/Asenapine: Did Merck's Researchers "Cherry-Pick" A 3-Week Data Point?

Salmon has provided (in comments, below) an excellent new piece of the asenapine (now branded as Saphris®, by legacy Schering-Plough/New Merck) puzzle. On March 1, 2010, in Munich, Germany, researchers hired by Merck presented some new study results, from an asenapine study designed to measure efficacy at a therapy duration of 12 weeks.

Oddly, the study's PI (or "principal investigator") reported out the data from week 3 (even though the study actually ran for 52 weeks), apparently. Do go read it all (at the link), but here is an interesting snippet:

. . . .Joseph Calabrese, MD, PhD, Mood Disorders Program, Case Western University School of Medicine, Cleveland, Ohio, and colleagues determined the efficacy of asenapine co-treatment with lithium or valproate in a 12-week primary study and a blinded extension study that followed for 40 weeks.

The primary-study endpoint was to evaluate the efficacy, safety, and tolerability of adjunctive asenapine for acute manic or mixed episodes in patients with bipolar disorder who incompletely responded to lithium or valproate monotherapy. The extension study had the same endpoints in addition to determining the maintenance of efficacy of adjunctive asenapine treatment.

The primary study was a randomised, double-blind, placebo-controlled trial wherein patients who had incompletely responded to treatment with a mood stabiliser received sublingual flexible-dose adjunctive asenapine 5 or 10 mg BID (n = 155) or placebo (n = 163). The 40-week blinded extension included 38 patients from the asenapine group and 33 from the placebo group. . . .

The incidence of treatment-emergent adverse events with asenapine and placebo was 73% and 69% in the core study and 78% and 69% in the extension. . . .

Is Salmon right, here (as he so often is) -- is this the use of the classic early effect magnification version of "cherry-picking" a study's results?

Quoth Salmon:
. . . .Plus, at 52 weeks PLACEBO BEAT ASENAPINE.

Now of course the author could have it wrong. Initial 3 week evaluations are more common with acute mania with extensions to 12 and 52 weeks. But if the message is supposed to be that asenapine is useful for extension beyond three weeks then I don't see it. Besides I've discussed other reports of extension studies here before and they also had major problems.

All in all so far everything I've seen so far does argues against any extended use beyond three weeks. . . .

I'll do some more digging -- but I doubt the study is registered at We'll see. Great find, Salmon!


Anonymous said...

Thanks for the reminder I forgot to check, but here's a summary of the info available:

Clintrials abbreviated title:

12 Week Study of the Safety/Efficacy of Asenapine When Added to Lithium/Valproate in the Treatment of Bipolar Disorder (A7501008)(COMPLETED)(P05844)

NCT number: NCT00145470

Official Title:

A Phase 3, Randomized, Placebo-Controlled, Double-Blinded Trial Evaluating the Safety and Efficacy of Asenapine in Subjects Continuing Lithium or Valproic Acid/Divalproex Sodium for the Treatment of an Acute Manic or Mixed Episode

Primary Outcome Measures:

• Improvement in bipolar I disorder manic or mixed symptoms as measured by the Young Mania Rating Scale (YMRS). [ Time Frame: after three weeks of treatment ] [ Designated as safety issue: No ]

Secondary Outcome Measures:

• Safety and tolerability (including extrapyramidal symptoms) of asenapine when used as an adjunct to Lithium or Valproate [ Time Frame: 12 weeks (end of trial) ] [ Designated as safety issue: No ]

Study Completion Date: April 2007
Primary Completion Date: February 2007 (Final data collection date for primary outcome measure)

Extention Study

Clintrials abbreviated title:

40 Week Trial to Study the Safety of Asenapine When Added to Lithium or Valproate in the Treatment of Bipolar Disorder (A7501009)(COMPLETED)(P05786)

Official Title:

A Phase 3, Placebo-Controlled, Double-Blinded Continuation Trial Evaluating the Safety and Efficacy of Asenapine in Subjects Completing Trial A7501008 and Continuing Lithium or Valproic Acid/Divalproex Sodium for the Treatment of an Acute Manic or Mixed Episode

NCT number: NCT00145509

Primary Outcome Measures:

• Characterize long-term safety and tolerability of asenapine in bipolar I disorder subjects [ Time Frame: 40 weeks ] [ Designated as safety issue: Yes ]

Secondary Outcome Measures:

• Longterm maintenance of effect on the Young -Mania rating scale (Y-MRS) and Montgomery-Asberg Depression Rating Scale (MADRS) scores (in the A7501008 lead-in trial) with adjunctive therapy. [ Time Frame: After the initial 12 weeks of treatment ] [ Designated as safety issue: No ]

Study Completion Date: December 2007
Primary Completion Date: December 2007 (Final data collection date for primary outcome measure)

In summary there was an initial study that looked at efficacy at 3weeks in patients who were taking Lithium or valproate (possibly relapsed). The responders were then continued to 12 weeks and then 52 weeks (12 + 40). There was efficacy shown at 3 weeks, as asenapine is approved for. Although the FDA reviews questions whether it really works in about half the population it's approved for who likely don't have psychotic symptoms. However beyond 3 weeks likely no efficacy. In spite of the primary endpoint being at 3 weeks Merck titles the abstract "[Presentation title: Asenapine as Adjunctive Treatment for Bipolar Mania: A Placebo-Controlled 12-Week Study and 40-Week Extension. Abstract PW01-28]".

So maybe they didn't "Cherry Pick" but rather used a misleading title to suggest to people quickly perusing the poster that it was effective for extended use. Why do I say that? Well at these scientific conventions people often quickly walk by these posters look at the titles and then quickly glance at a table or graph of the results as they're walking by or if they stop they might quickly look at the conclusions. If they see a title that says 12 and 40 weeks and then see a graph or conclusion showing efficacy they're likely to think it's effective longer than the 3 weeks the study actually assesses.


Anonymous said...

P.S. These studies were completed in April and December 2007.

Why is Merck only reporting results 3 years later?

Why are no results posted to


Condor said...

These are GREAT questions, Salmon! -- will make them (and the info) a new post -- right after a dinner meeting.


Anonymous said...

Thanks. Although we need to check if the results haven't been reported previously. These studies might be ones we discussed here previously.