Myriad Genetics Doesn't Make the Grade

Myriad says, simply, that the drug takes longer than 12 months to show full effects. The longer a patient takes Flurizan, the greater the benefit.

But I have an alternative -- and more skeptical -- explanation: Notice that the size of the high-dose Flurizan patient group decreased by 52% from 48 patients at the start of the study to 23 patients after 12 months -- when the dramatic improvements began to be recorded. (Look at the yellow "N's" on the chart.)

This is a classic example of selection bias, where patients who are doing better remain in the study and artificially skew longer-term results. Patients doing badly at 12 months dropped out and ceased being tracked.

When Myriad announced the initial results from the Flurizan phase II study in May 2005, it also announced an amendment to the ongoing phase III studies to extend the study period to 18 months from 12 months, claiming the longer time period boosted chances for the study's success. (Yes, Myriad actually started its phase III studies before having data from the phase II study, yet another red flag ...)

As I mentioned above, the Flurizan phase II study failed at 12 months to meet its primary endpoint, which was to show a statistically significant treatment effect between Flurizan 800 mg taken twice a day compared with placebo on three different measures of Alzheimer's disease severity. (Those tests were the Alzheimer's Disease Cooperative Study -- Activities of Daily Living inventory (ADCS-ADL), Clinical Dementia Rating -- Sum of Boxes (CDR-sb), and the Alzheimer's Disease Assessment Scale -- Cognitive function subscale (ADAS-cog).)

Myriad describes Flurizan as a Selective Amyloid Beta-42 Lowering Agent. As a reminder, the "beta-amyloid" theory of Alzheimer's believes damaging beta-amyloid plaques form when certain enzymes cut amyloid precursor protein (APP) into fragments. Instead of being cleared from the body, some of these fragments clump together into plaques, which damage nerve cells in the brain.

Flurizan is supposed to work by attaching itself to gamma secretase, one of the enzymes responsible for cutting up APP. Once attached, Flurizan forces gamma secretase to modulate its behavior, cutting APP into smaller -- and less sticky -- fragments that can't form into brain-damaging plaques.