The results are in, and Alkermes’ dream of becoming a speciality pharma company lies, if not in tatters, then at least badly torn. The depression project ALKS 5461 has failed in two of its three phase III studies, and the company’s shares are down 33% in early trading as a result.
Alkermes had designed the Forward-3 and Forward-4 trials to minimise the placebo effect, but to no avail. The group plans to tweak the design of the third co-pivotal trial, increasing enrolment and altering the statistical analysis, but success here seems a long shot. And it gets worse: Alkermes’ schizophrenia candidate ALKS 3831 incorporates one of the same molecules as ’5461, so there could be read-across.
Nothing really MADRS
The primary endpoint of both Forward-3 and 4 was change versus placebo from baseline on the Montgomery–Åsberg Depression Rating Scale (MADRS). Alkermes had made much of the sequential parallel comparison design employed by the trials as a means to control for high placebo response.
This is a two-step process in which the trial starts conventionally, but at a certain point the placebo group is evaluated to identify those who have not responded. Only these patients are allowed to continue into the second part of the trial – they are re-randomised to either drug or placebo – eliminating those who felt better without drug treatment.
Such is the theory. In actuality no treatment effect was seen in Forward-3, which tested a 2mg/2mg sublingual dose, while the results from Forward-4 were more complicated. This trial tested two doses of ’5461, 0.5mg/0.5mg and 2mg/2mg, both of which missed the primary endpoint.
Alkermes said that there was “a clear trend toward efficacy” with the 2mg/2mg dose, and added that post-hoc analyses achieved statistical significance for the entire 2mg/2mg dose group on the MADRS endpoint. This provides “supportive evidence” of efficacy, it says, adding that 2mg had succeeded in phase II. investors seem not to share the company’s optimism.
According to clinicaltrials.gov the third pivotal trial in the programme, Forward-5, is to conclude in July. This will now be pushed back as Alkermes alters its design and signs up more patients; the company says it will announce the new conclusion date on its fourth-quarter conference call in the next couple of months.
Regulators are often flexible when it comes to depression drugs, as a strong placebo effect is a perennial hazard – but there is a limit to flexibility. Forward-5 will have to be a huge success if ’5461 is to gain FDA approval, its fast-track status notwithstanding. Even before then the company will have to get the agency’s buy-in on the protocol changes.
So attention now turns to the second of Alkermes’ new assets: ALKS 3831 for schizophrenia. This is a combination of samidorphan, a mu opioid receptor antagonist, and the generic antipsychotic olanzapine. ALKS 5461 also contains samidorphan – along with buprenorphine – so investors might well fear that the failure of the antidepressant bodes poorly for ’3831.
Before the Forward failure ’5461 was expected to sell $335m in 2020 according to EvaluatePharma’s consensus forecasts, and ALKS 3831 $105m. Not blockbusters, but still set to become Alkermes’ third and fourth-biggest sellers in five years’ time.
Analysts from Leerink suggest that with around $780m in cash on the books Alkermes might want to make up for the failure of one and the potential loss of the other by bolstering its CNS pipeline via an acquisition. However, they add that potential targets, like Acadia Pharmaceuticals and Intra-Cellular Therapies, might still be too expensive.
Alternatively maybe the company will want to shelve its ambitions to develop innovative drugs and return to its traditional strength of reformulations.
|Forward-5||Under review; initially 350||NCT02218008|