In 2003, the FDA approved Namenda (memantine) as a treatment for moderate to severe Alzheimer’s Dementia (AD). Backed by the legendary marketing expertise of Forest Labs, the medication has become remarkably popular, gradually eating into the market share of the cholinesterase inhibitors (CIs). The question, as always, is whether this popularity is justified by the data.

Recently, a meta-analysis was published by the Cochrane Collaborative reviewing pretty much everything that we know about Namenda to date (http://www.mrw.interscience.wiley. com/cochrane/clsysrev/articles/CD003154/frame.html, accessed 6/30/05), including nine published and three unpublished studies. For those who don’t know, the Cochrane reviews are considered the international “gold standard” in evidence-based health information.

Let’s look first at how robust the data is for Namenda’s only official FDA indication, moderate to severe Alzheimer’s dementia. The standard research definition of “moderate to severe dementia” is an MMSE score below 15; in the two Namenda studies, the average MMSE of participants ranged from 8 to 10 (NEJM, 2003; 348:1333-1341, and JAMA, 2004; 291:317-324). This is a level of impairment that most of us are unlikely to see in our offices, unless we specialize in dementia treatment. According to the study authors, these patients typically required assistance with basic ADLs, such as bathing and dressing, and many were incontinent.

In this population, Namenda 20 mg QD was statistically superior to placebo in measures of cognition, ADLs, and agitation, but the effect sizes, in all cases, were “small or very small,” according to the reviewers. The primary cognitive measure in these studies, the “SIB,” is a scale ranging in score from 0-100; the average improvement over placebo in the two trials was 4 SIB points, unlikely to yield a clinically significant improvement, particularly in such an impaired population.

There also exists a mysterious third Namenda study, named “MD-01” by Forest, but it hasn’t yet been published, even though the study was completed in June of 2003. Such a long lag in study publication is generally diagnostic of unfavorable results, leading the Cochrane reviewers to conclude that the published results may overestimate the effectiveness of Namenda (this phenomenon is known as “publication bias”). TCR urges you to nudge your Forest reps about the MD-01 results next time you see them!

What about Namenda for mild to moderate dementia? Just as we were about to go to press with this issue, Forest announced that the FDA rejected Namenda’s application for approval for mild dementia, which was no surprise to those who had looked at the data. No studies of Namenda for mild to moderate AD have been published. One study, presented as a poster, showed a small benefit over placebo, but the two other studies yielded no significant effects, according to the Cochrane review.

Data or not, Forest has found a creative way to produce a mild-dementia “halo effect” around Namenda. They did this by funding a study in which 404 patients with moderate to severe AD, all of whom had been taking Aricept for about two years, were randomized to receive add-on Namenda vs. add-on placebo. The Namenda group did a little bit better than the placebo group after six months (JAMA, 2004; 291:317-324). While Forest can officially use this study only to promote Namenda for moderate to severe AD, the fact that it was added to a drug only approved for mild to moderate AD (Aricept) has the effect of planting a seed in our minds that Namenda might be effective for both indications. Talk about subliminal advertising!

TCR VERDICT: Namenda: A small effect for the most impaired.