Amyloid beta focus – Three solanezumab questions that Lilly must answer

The glimmer of efficacy that Lilly’s solanezumab showed in Alzheimer’s disease last week came like a bolt from the blue, most investors having already written the drug off as another dud in the company’s in-house R&D pipeline.

Although the dust has started to settle, full data from solanezumab’s two phase III studies will not be released until the American Neurological Association meeting in Boston on October 7-9, and the share prices of Lilly and other companies developing amyloid-targeting Alzheimer’s therapies could remain volatile until then. As expectations around the approvability of solanezumab wax and wane, EP Vantage spells out three key issues that investors will need to have clarified.

The two solanezumab studies, Expedition and Expedition 2, failed their primary endpoints, which had initially targeted meaningful changes in cognitive and functional measures. After unblinding the Expedition study and seeing it fail, Lilly amended the primary endpoint in Expedition 2 to a change only in cognition and only in the mildly impaired population; despite this Expedition 2 also failed.

However – crucially for what is now the future of the molecule – a pre-specified analysis of pooled data across both trials showed a statistically significant slowing of cognitive, but not functional, decline in the mild and mild-to-moderate populations (Sola sees surprise glimmer of hope in mild Alzheimer’s, August 24, 2012). The answers to three questions could determine whether solanezumab has a future or not.

How significant was the pooled improvement?

Although Lilly says the pooled secondary analysis hit statistical significance it does not say by what margin, and it will be interesting to find out how far below 0.05 the p value was.

Neither has the company said whether the statistical significance it saw was corrected for multiplicity – to adjust for the fact that another analysis effectively amounts to an additional look at the data and should thus be penalised in statistical terms.

Moreover, we do not know how many pre-determined secondary analyses there were. If there had been several, of which only the one disclosed happened to show an effect, this might amount to data mining; a very strong p value would then be needed to show that the result was not just a fluke.

Of course, the real-world effect will be suggested when solanezumab’s absolute reduction of the decline in cognition versus placebo is revealed.

In a recent investor webinar organised by ISI Group, Dr Adam Fleischer, associate professor at the Department of Neurosciences at University of California, San Diego, said he would expect a placebo group to show a seven to 10-point worsening in the Adas-cog cognition scale over 18 months.

Against this background, he said he would regard a three to four-point improvement in the active groups as being clinically relevant. He also cautioned that cognitive function in Alzheimer’s patients does not decline linearly, and the FDA will likely also consider the difference in the curves for active and placebo and the point at which they diverge.

However, he expects the absolute improvement with solanezumab in the pooled analysis to be relatively small – perhaps of the order of two points on Adas-cog – although in milder patients the placebo change was likely also smaller.

By how much did Expedition 2 miss?

Given how nebulous this all is, any additional signs of activity will be seized upon as evidence that the positive result is not just a one-off. One such signal will come when Lilly discloses by how much Expedition 2 missed the cognition endpoint in mild patients.

If it missed by a whisker and if there was a positive effect without statistical significance this will count as a major bonus to sentiment around the drug. Moreover, the miss might have been caused simply by there being too few mild patients in this study.

As an aside, it might also be interesting to see whether Expedition 2 would have hit significance had the primary endpoint not been changed. Although this is highly unlikely, if it were the case this could call the pooled analysis into serious question, because it would go against the logical hypothesis that the milder patient population was driving the positive result.

It would also leave Lilly with egg on its face.

Will regulators accept cognition alone?

When all is said and done, perhaps the most pressing question will be whether, given the lack of a hit on any functional score, the US and/or European regulators will accept a filing based on an effect on cognition alone.

Previous guidance from the FDA has been that an effect in both cognition and function has to be shown, but in a disease with a large unmet need the goalposts might be moved. One question that Lilly has so far avoided answering is whether the FDA had blessed its change of the Expedition 2 primary endpoint to cognition alone; if it did, this strongly suggests that an effect on function might not be necessary, although Lilly’s obfuscation suggests that the FDA did not approve the change.

In any case, Dr Fleischer said that if a cognition effect is seen sooner or later this will translate into a benefit in function; it is likely that in the Expedition trials the functional benefit was just too small to show significance, he suggested.

As such, regulators will likely look for additional factors to strengthen their conviction that the clinical benefit is real. One such factor is solanezumab’s effect on biomarkers: a reduction in amyloid protein in the brain along with a reduction in Tau protein – a marker of neurodegeneration – would be a strong help.

Still, however positively Lilly is able to answer these questions it is still a very long shot for solanezumab to be approved on the pooled Expedition data alone. A new study will most likely be needed in mild patients, perhaps just based around Adas-cog, so it will be important to test the regulatory position to such an approach, as well as delving more deeply into the data to gauge the likelihood of achieving a positive result in the new trial.

On the debit side this might put solanezumab behind competitor projects, some of which are already being studied in a predominantly mild patient population (see table).

But the fact is that, against all odds, solanezumab is not dead. Moreover, it still has the chance to breathe new life into other amyloid-targeting projects.

To contact the writer of this story email Jacob Plieth in London at [email protected]