Memantine is a selective blocker of the N-methyl-D-aspartate type of glutamate receptor. Repeated activation of this receptor by the neurotransmitter glutamate can cause degeneration of those same neurons that are most vulnerable in AD. Senior author Barry Reisberg and colleagues at the New York University School of Medicine launched the original U.S. clinical trial to determine if this drug, which had been used for years in Germany, has any benefit for people suffering from the disease (see ARF related news story). Favorable results from that initial 6-month trial led to FDA approval for the drug in October 2003 (see ARF related news story).

Now, results from the open-label phase extend observations from the original trial. They also provide an opportunity to test the notion that memantine may actually tackle some of the underlying pathology of AD. This is possible because open-label extension is similar to the random-start design of a clinical trial that Paul Leber, then at the FDA, proposed to distinguish between drugs that have “disease-modifying” effects and ones that merely relieve symptoms (see Leber, 1997). In theory, a disease-modifying drug will not elicit as robust a response in patients who start therapy later, because by then the disease will have progressed further along. In other words, those patients may not “catch up” in terms of cognitive or neuropsychological ability to those who start treatment earlier. On the other hand, if the drug simply relieves symptoms, then patients who are started later may indeed catch up.

“The results suggest the mechanism of action may be symptomatic because the group switched from placebo ends up the same as the group who were on memantine from the start,” comments Martin Farlow, Indiana University School of Medicine, by e-mail. This is because at the end of the 52 weeks, there was no statistical difference in either the Alzheimer’s Disease Cooperative Study-Activities of Daily Living (ADCS-ADL) scale or the New York University version of the Clinician’s Interview-Based Impression of Change plus caregiver input (CIBIC-Plus), the two main yardsticks the authors used to measure the efficacy of memantine. In other words, the original placebo group has “caught up” with those who had been on memantine all along. This is disappointing, because if memantine has no effect on the underlying progression of disease, then the benefits of memantine may peter out as the disease advances, leaving clinicians and caregivers the unenviable task of deciding if, and when, the medication should be stopped. Longer trials may shed some more light on this issue.

But the symptomatic benefits seem real. In both the ADCS-ADL and the CIBIC-Plus, patients who had received placebo in the original phase of the trial and then were switched over to memantine had significantly slower rates of decline during the second phase of the trial. “The results are consistent with memantine having a continued beneficial effect over one year but in the absence of a placebo group don’t prove it,” suggested Farlow. This speaks to one of the major shortcomings of open-label extensions, that is, the lack of placebo. In fact, Jeffrey Cummings, University of California, Los Angeles, addressed this point in an accompanying Archives of Neurology editorial. Reisberg and colleagues addressed the lack of placebo by extrapolating the rate of decline seen in the placebo group before they switched to drug. This method is “straightforward,” but "fails to take into account the observation that the rate of change in patients with moderate to severe [Alzheimer] disease accelerates over time until a stage of severe disease is reached," Cummings writes. The upshot is that “the result would be a slight underestimate of treatment effects,” he adds.—Tom Fagan.

References:
Reisberg B, Doody R, Stoffler A, Schmitt F, Ferris S, Mobius HJ. A 24-week open-label extension study of memantine in moderate to severe Alzheimer disease. Arch. Neurol. January 2006;63:49-54. Abstract

Cummings J. What we can learn from open-label extensions of randomized clinical trials. Arch. Neurol. January 2006;63:18-19. Abstract