Elan spokesman Jonathan Birt would only acknowledge that the primary cognitive endpoint of the phase III studies will be either the NTB or the ADAS-cog test. The companies will be measuring both and will choose one as the primary endpoint sometime during the course of the trials, he says. "This is all being done in close cooperation with the FDA and other regulators who are fine with the approach," Birt adds. This approach, however, is far from typical. In nearly all clinical studies, primary endpoints are chosen well before the first patient is ever enrolled. Birt did confirm that Elan and Wyeth have met with FDA officials to discuss use of the NTB as the primary endpoint of the bapineuzumab studies. Birt would not comment, however, when asked if Elan and Wyeth had reached agreement with the FDA on a Special Protocol Assessment, or SPA, for the bapineuzumab phase III studies. An SPA is a written agreement between a drug sponsor and the FDA that says the agency has signed off on an acceptable study design, including the clinical endpoints that could lead to a drug's approval. Typically, drug companies will negotiate an SPA with the FDA for drug trials involving unusual or little-used clinical endpoints, or in situations where the requirements for approval are not entirely clear. An SPA doesn't guarantee a drug's eventual approval, but it does provide some assurance that the drug company and the FDA are on the same page when it comes to a clinical trial design. In the case of bapineuzumab, an SPA between Elan, Wyeth and the FDA would specify that the NTB could be used as the primary measure of cognition in the phase III studies, and that data collected using the NTB, if positive, could form the basis of the drug's approval. Yet Elan and Wyeth declined to comment on whether they sought an SPA from the FDA. The FDA's own rules require SPAs to be in place before clinical trials begin enrolling patients.
If the ADAS-cog is a perfectly acceptable way to assess the effect of an Alzheimer's drug, why are Elan and Wyeth going to great lengths to devise their own test? The answer, says Elan's Birt, is that the companies believe the NTB may be more sensitive than the ADAS-cog in measuring cognitive function in patients with milder forms of Alzheimer's. But it should be noted that none of the drugs in Elan's highly touted Alzheimer's drug development program have demonstrated an improvement in memory or mental status when measured using traditional tests, including the ADAS-cog. That's likely one of the issues prompting Elan to come up with new ways of showing the benefit of its Alzheimer's drugs, including coming up with its own cognitive test. Elan researchers created the NTB by combining six older tests of memory and cognitive function to record nine measures of patient performance. While the individual components of the NTB are in wide use (mainly as standard memory tests for all types of patients), they have never before been used to assess the mental status or memory of patients in a phase III Alzheimer's disease clinical trial meant to lead to a drug's approval. Elan admits as much in a paper published in the Archives of Neurology in September. This paper describes the use of the NTB to measure the mental status and memory of patients in a study of AN-1792, an older Alzheimer's drug and predecessor to bapineuzumab that Elan was forced to shelve in 2002 due to toxicity.