Reader Comments

Post a new comment on this article

Incorrect Interpretation of Studies in Table 1

Posted by michael_gold on 11 Mar 2011 at 01:27 GMT

There are several substantive errors in the interpretation of Table 1. and the inference that the amyloid hypothesis has been adequately tested.

The studies listed in Table 1 are described as being "anti-amyloid drugs" and if having either failed of having either failed or been abandoned due to toxicity.

1) Rosiglitazone is not an anti-amyloid drug and the study referenced was not designed to test the amyloid hypothesis
2) The Bapineuzumab study referenced was not designed or powered to detect efficacy, therefore the fact that no efficacy was detected cannot be claimed as a failure
3) The AZD-103 study was likewise, not designed or powered to detect a treatment effect.
4) There are data to support the hypothesis that the doses of tarenflurbil used in clinical trials was too low to have any effect in amyloid processing in the brain.

Before dismissing a hypothesis, it is important that the data used to test the hypothesis be drawn from appropriate studies. One cannot deny that in some cases (i.e. tramiprosate and semagestat), the data used to justify progressing into large clinical trials was not robust.

Competing interests declared: I am a former GSK employee and was involved in the Rosiglitazone-XR clinical trials.

RE: Incorrect Interpretation of Studies in Table 1

Kimmelman replied to michael_gold on 15 Mar 2011 at 19:38 GMT

We respectfully take issue with Dr. Gold’s characterization of our table 1.

Rosiglitazone was developed for type 2 diabetes, and advanced into Alzheimer’s disease trials at least in part on evidence of its effects on the processing of amyloid precursor protein. For instance, the first pilot study testing rosiglitazone in Alzheimer’s patients states “we hypothesized that rosiglitazone would modulate plasma [beta]-amyloid levels and rescue patients from the progressive cognitive decline that characterizes AD…;”[1] readers are also directed to reference 29 in our original article.

Bapineuzumab and AZD-103 were accurately reflected in our table as phase 2 studies. The fact that neither study met its primary endpoint is consistent with our thesis that causal premises underwriting these strategies do not yet evidence a high level of maturity. That doses of tarenflurbil used in trials may have been insufficient, once again, seems to testify to immature knowledge of causal systems.

Our article did not set out to adjudicate the amyloid hypothesis. Nor do we make any claims the “hypothesis has been adequately tested;” indeed we clearly state that such approaches may eventually succeed. But we think Dr. Gold’s criticisms reinforce our point about evidential conservatism. Tempting though it may be to view each trial result in isolation from the others, the totality of evidence accumulated thus far does not inspire confidence that proponents of the amyloid hypothesis have yet developed sufficient knowledge of causal pathways to intervene in clinically useful ways. We think this ought to influence ethical decision-making about risk and informed consent.

[1] Watson GS, Cholerton BA, Reger MA et al. Preserved Cognition in Patients With Early Alzheimer Disease and Amnestic Mild Cognitive Impairment During Treatment With Rosiglitazone: A Preliminary Study. Am J Geriatr Psychiatry 2005 Nov;13(11):950-8

No competing interests declared.