Thanks. Is this the article?
https://www.bloomberg.com/gadfly/articles/2016-11-25/eli-lilly-alzheimer-s-failure-the-right-lesson
[But the more broadly applicable lesson is one pharma seems incapable of learning: Digging through failed drug trials to find justification for moving forward with medicines is a recipe for failing again.
DEJA VUToo rarely these days is a drug's clinical-trial failure the end of the line. Drugmakers almost always find a "but" -- "But the drug appeared to work in a certain population of patients." "But the placebo effect was unexpectedly large." "But, but, but."
These are variations of post-hoc analysis, in which firms blame trial failures on everything but the drug itself in order to justify more trials, focusing on narrower sub-groups of data. It's the
"this time is different" of the pharma industry, and it's endemic.
Sola had failed in two previous Phase 3 trials. But in poring through those results, researchers
came to believe the trials had failed the drug, not the other way around, and the latest trial selected patients more carefully. Sola was given every chance to succeed -- Lilly even moved the goalposts of success -- but it failed once again.
Sola joins a large graveyard of drugs that have repeatedly failed Phase 3 trials, including
Takeda's motesanib. The list of drugs that have been pushed into Phase 3 trials despite iffy Phase 2 data is even longer. That group includes Lilly's own schizophrenia drug candidate pomaglumetad.]
It is easy to pan LLY in hindsight, but they have disproved something, at immense cost to themselves and patients. Between Sola and semagacestat, some 5000 patients have volunteered for their trials. $3B in development funds gone. They made it clear from the start it was a gamble and I think the justification, was biomarkers, the immense potential earnings and the dire need. Saying that, one would think that the penny may have dropped several years ago when Tanzi and Moir demonstrated amyloid to be part of the innate immune system of the brain. A direct attack on amyloid is probably not a good thing in hindsight, but there were warnings.
Using data mined sub groups for trial justification is another thing. Remember the stat sig effect seen in early stage HD patients was a pre specified analysis, which is very different from data mining, and it was replication of the same effect seen in the Euro AD trial.
(20min delay)