It's hard not to rejoice when it's shown that a drug has a very real potential to help people with Alzheimer's ̵
. The drug, BAN2401, significantly slowed cognitive decline after 18 months of high-dose treatment, depending on how you measure it. Regardless of the qualification, this is an absolutely promising result and the data are better than I, a skeptic to Alzheimer's studies, in some ways expected. But there are many more tests before anyone can be sure of this result, and it will take a long time.
Taken together, this is not the home run that some investors have been looking for, and they have already begun adapting to this reality. After Biogen's shares were pushed to a three-year high ahead of the release of the results, Wednesday's shares fell by more than 10 percent after-hours.
There are big differences in the way how Alzheimer's affects people and how they progress, which is a reason for all previous efforts to treat them has failed spectacularly. Many disappointments in the past were based on the same theory of treatment of the disease, the BAN2401, known as the amyloid hypothesis.
Some of the numbers published in the latest study are very impressive, as was the finding that a high-dose drug slowed cognitive decline by 30 percent. But the measure in question is a novel composite designed by Eisai that some experts question. And the drug did not reach any other statistical significance. Also, according to Bloomberg Intelligence biotech analyst Asthika Goonewardene, the process had an imbalance in a particular category of patient that could have influenced the results.
Those who search for other topics to fool around in the data can find them. Thus, the best dose was given only to 161 patients, and other dosages of the drug were worse in number than a placebo.
The two companies would like to submit this data directly to the regulatory authorities and obtain a permit. But there are enough questions that a bigger and longer test in the final phase should be needed. And remember, the end-stage trial started with Biogen's most advanced Alzheimer's drug in 2015 and is unlikely to deliver conclusive data until 2019.
A larger study involves many risks as the middle-stage test was strange. It used an unusual adaptive design that made it harder to interpret. After a previous look at the experiment showed it had failed, it took a different statistical approach to achieve that success.
There is absolutely enough here to justify hope and a bigger test – the easier, the better.
This column does not necessarily reflect the opinion of the editors or Bloomberg LP and its owners.
Contact the editor responsible for this story:
Beth Williams at bewilliams @ bloomberg .net