Cortexyme Alz GAIN Trial PR data analysis:
The Cortexyme GAIN trial was run with 643 subjects, with two primary endpoint tests ADAS-Cog and ADCS-ADL. MMSE testing was done as a secondary endpoint test, but the PR gave no MMSE data. There were three different arms, one a high dose of COR388(80mg), one a low dose of COR388(40mg), and a placebo. The thesis is that the bacteria P. gingivalis (aka PG) is the primary cause of Alzheimers, and that their drug COR388, will be efficacious in subjects that have PG.
As the PR says “In overall population, co-primary endpoints of ADAS-Cog11 and ADCS-ADL were not met”
That says a lot, but is not the whole story. Lets look at what exactly happened…
Keep in mind that the subgroup of patients WITH bacteria PG, are the primary intended subjects. Analogous to the A2-73 patients that WITH wild type genes. Comparison of these subgroups should prove interesting.
Regarding the cognitive test ADAS-COG11:
The subgroup with bacteria PG constituted 37.6% (242/643) of all the subjects. The high dose group with PG, showed a 57% slowing of cognitive decline for ADAS-COG11. This is a significant improvement, but did not meet their predicted improvement of 2.5 points. I think they actually attained ~1.71 points of improvement (57% of 2 to 4, being ~1.71 points). There was a dose improvement with the high dose over the 40mg dose, but the 40mg dose had a high p-value, emphasizing the need for the higher dose for better efficacy, but also indicating that the lower dose is likely near the bottom of the useable dose range.
In summary, if any random group of people with AD, took COR388, 37.6% of them would see a cognition decline of 43% versus placebo. This is compared to placebo and is much better than 100% decline. IIRC this high dose gives much better efficacy than the BIIB aducanumab!
Regarding the cognitive test ADCS-ADL:
There was also a primary endpoint test using Activities of Daily Living(ADL). The PR stated that this endpoint was also not met. Why? The company rightly raises the issue that many of the ADCS-ADL questions are affected by the pandemic lockdown. For example, there are questions about travel and shopping. Presumably, the company will raise this issue with the FDA. What will be the outcome?
How Does This Data Compare to A2-73
Amount of improvement:
The A2-73 super responders-----IMPROVED FROM BASELINE
COR388 super responders-------DECLINED 43% FROM BASELINE compared to placbo
This is not a numerical comparison, as A2-73 data is from the P2 trial which used ADCS-ADL. But one obviously did improve from baseline, while the other declined (though arguably much better than placebo).
How many improve?:
It is helpful to understand what percentage of a population could be expected to show improvement.
A2-73 super responders as percentage of AD population----- ~57%
COR388 super responders as percentage of AD population---~37.6%
SUMMARY
COR388 appears to be helpful with AD, but may be a victim of choosing the wrong predicted benefit value, needing a longer trial, and an endpoint affected by the pandemic! It will be interesting to see how it pans out.
Compared to A2-73, COR388 appears to help a smaller subset of AD patients, with less efficacy. However, it is probably preferable to the Biogen approved drug.
Anavex’s P3 ALZ trial also has primary endpoints using the ADAS-COG and ADCS-ADL tests. What if Anavex’s ADCS-ADL results are affected by the pandemic?
All imho.
