Well, I will avoid listing my social science degrees here, since they will probably not add to my credibility. Just a couple of questions and a couple of observations.
First the question.
>>Despite the variability in insulin dosing, 46% of patients evaluable at 52 weeks had a statistically significant, 12-month reduction in HbA1c > 0.5% (p =0.0060) compared to 29% of placebo patients. <<
I remember learning that a figure like .5% meant 1/2 of 1%. Is that correct? If so, then it looks to me like this release has consistently mis-stated AMLN's trial results. I assume that by .5% they mean 50% since they are comparing it to 29%. 1/2 of 1% is not much of a reduction. If I am correct, then I wonder who is writing and proofreading AMLN's press releases.
Now, in regard to the trial results. I agree with Henry and David that they are mixed. The results for Type 1 are clearly significant and though it is true the market is small compared to Type 2, it is still large enough to make a small company like AMLN a lot of money if the drug is widely adopted. That said, I must agree with you that it is troubling that only 75% of the Type 1 patients agreed to continue using the drug, even though they will get it free of charge. Still, overall, Type 1 is a significant market and the drug is significant in reducing glucose levels and the insurance companies will pay for it.
I have another question regarding the "intent-to-treat group." How does that differ from the pramlintide treatment group?
>>In the intent-to-treat analysis for type 2 patients who entered the study with poor glucose control (entry HbA1c > 8.0%), the reductions in HbA1c at 12 months for patients treated with pramlintide compared to placebo were 0.35% (p = 0.0418, n = 115) in the 75-microgram dose group, and 0.42% (p = 0.0048, n = 105) in the 150-microgram dose group. The result from the 150-microgram dose group was statistically significant, and this entry criteria corresponds to that which is being used for the ongoing Phase III studies.<<
It seems there was statistical significance for the 150mg group compared to placebo. But not in the pramlintide treatment group.
Perhaps the reason for the lack of statistical significance was as AMLN suggested, the variablility in insulin dosage--the next trials should help control that factor. While it appears pramlintide does help Type 11, lack of statistical significance will not get this indication from the FDA.
Conclusions? It appears, at the very least, that pramlintide is not the blockbuster diabetes drug that was hoped for. There is considerable uncertainty over its efficacy for Type 11. Previously, in shorter trials, it had appeared efficacious. My guess is that the market will be disappointed in this regard. |
