Problems, problems, problems: But Alkermes will fight it out this week over the FDA’s objections to its big depression drug ALKS 5461
Alkermes tried several new things when their researchers tested ALKS 5461 for major depression in clinical trials. Now they have to see if they can convince the FDA and a group of outside experts that the drug is actually as effective — and as safe — as they insist it is.
Alkermes $ALKS is attempting to gain an approval for this drug — which is seen as a lynch pin to its future revenue — despite key trial failures and an arguably thin layer of controversial data backing the drug.
In the agency’s internal review of the drug, initially rejected by the FDA and then taken in for a review after a sudden about-face on its demand for new trials ahead of a marketing decision, regulators make it clear that Alkermes pushed ahead with a study design they clearly insisted was inadequate to provide the insight on efficacy they need to approve a drug for major depression.
Those objections helped drive down the stock price today, which slid 6.5% on the frosty — though not forbidding — reception.
Most importantly, the regulators highlight Alkermes’ move to use an abbreviated scoring system, cutting out essential points that have scuttled other depression drugs — including tracking suicidal thinking on the part of patients.
From the review:
The Montgomery Åsberg Depression Rating Scale is a 10-item diagnostic questionnaire used to measure the severity of depressive episodes in patients with mood disorders (MADRS-10). The Applicant used an abridged 6-item version of the MADRS-10 for the primary endpoint of one of the principal studies (Study 207). The Division had rendered advice explicitly against this plan, based on analyses of the MADRS-10 and MADRS-6 by both the Division and the Agency’s Clinical Outcomes Assessment (COA) Staff. The COA Staff had concluded that the MADRS-6 could not replace the MADRS-10 for use as a primary endpoint because the abridged questionnaire excludes concepts that are relevant and important in MDD, specifically “reduced sleep,” “reduced appetite,” “concentration difficulties,” and “suicidal thoughts.”
Not only that, but the agency takes exception to Alkermes’ view that the drug is a non-opioid.
(B)uprenorphine is an opioid. Although the Applicant has made several arguments that the other component of their drug, samidorphan, negates the μ-opioid properties of buprenorphine, this has not been conclusively proven. Although we agree with the Applicant that there are fewer opioid properties for this combination product than there likely would have been from the opioid alone, there remains some evidence of a mild opiate effect (including mild withdrawal effects) from the trials.
In several cases, the FDA spotlights messy data that they clearly have taken issue with. That includes dropping the classic approach of assessing patients at particular time points in a study and instead shooting for average scores over a period of time. That may not fly with the FDA, but the regulators do want to hear the experts’ opinions on that subject.
Alkermes also used a novel trial design to select patients for the study, and now regulators want the outside experts to discuss their opinions on how this worked. Explaining the move, regulators noted:
In Stage 1, patients who meet study entrance criteria are randomized to drug vs. placebo, with a skewed randomization ratio that places far more patients in the placebo group. Stage 2 includes only non-responders from the Stage 1 placebo group, and these patients are re-randomized to drug vs. placebo. Results from the two stages are merged to provide a single overall test of hypothesis. There are a number of unresolved statistical questions regarding the most appropriate method for analyzing the results of an SPCD study.
The expert panel will gather on Thursday to hash it all out and offer their opinion to the FDA overlords. In the meantime, Paul Matteis at Stifel says Alkermes is facing some severe headwinds going into the panel review Thursday.
Overall, the docs are definitely negative and suggest that FDA is leaning towards a CRL, but they also don’t seem to completely close the door on approval (for example they discuss, hypothetically, what a REMS would look like – only relevant if the drug were approved). We continue to believe that the probability of approval is well below 50%, and we don’t think the docs today are likely to majorly change investors’ views on the issue.