Drug therapy for Alzheimer's disease

We would like to highlight several shortcomings in the meta-analysis of cholinesterase inhibitors in Alzheimer's disease presented by Krista Lanctôt and associates1 and to clarify information on the risk–benefit ratio for galantamine.

First, some of the data used in the meta-analysis were obtained with subtherapeutic or higher-than-recommended doses, and treatment duration varied widely (12 weeks to 1 year). Consequently, the heterogeneity reported by Lanctôt and associates1 is unsurprising. Focusing on data obtained with recommended maintenance doses for common treatment durations would yield more reliable and clinically relevant information.

Second, the use of global clinical measures for examining patient change is problematic because of the processes used to obtain these results (usually semistructured interviews): with this method, the domains covered and specific questions used may vary from one subject to another. Furthermore, the relative contribution of changes in different domains to the overall clinical judgement about change from baseline values varies between clinicians. Inter-rater reliability and differences between the 2 assessment scales used in the study by Lanctôt and associates are also of concern.2 Given the degenerative nature of Alzheimer's disease, limiting the definition of treatment success to improvement may be too restrictive. Rather, an overall positive response encompasses not only an increase in the number of patients showing improvement, but also a reduction in the number showing (greater) deterioration, especially over the longer term. For example, longer-term clinical trials of galantamine (5 to 6 months) have shown statistically significant benefit of 16-mg and 24-mg doses over placebo.3^,4^,5 More generally, given the importance of cognition to Alzheimer's disease and the relatively smaller inter-rater variation in this measure, we suggest that more weight be given to findings for the cognitive portion of the Alzheimer's Disease Assessment Scale (ADAS-Cog).

The data presented on the safety of galantamine do not reflect product labelling and recommended use in clinical settings. The clinical development of this drug has involved an evolution away from high doses (32 mg) and fast (weekly) titration3^,4^,6^,7 toward lower dosing (16 and 24 mg) with a slower (monthly) titration schedule.5 By combining data from early and later trials, Lanctôt and associates1 determined discontinuation rates that incorrectly appear higher than those for donepezil or rivastigmine. Safety data from a trial in which dosing was consistent with product labelling5 showed discontinuation rates due to adverse events similar to placebo (odds ratio [OR] for 16-mg dose 0.97, 95% confidence interval [CI] 0.51 to 1.86; OR for 24-mg dose 1.46, 95% CI 0.80 to 2.64).8 The similarity in safety profiles of galantamine and donepezil was confirmed in a recent comparative trial.9

Shane Kavanagh Johnson & Johnson Pharmaceutical Services Beerse, Belgium Patricia Kabathova Johnson & Johnson Pharmaceutical Services Raritan, NJ