Livingston_2004_Int.J.Geriatr.Psychiatry_19_919

INTRODUCTION: Memantine is currently the only treatment approved for moderately severe to severe Alzheimer's disease (AD). There is still some discussion as to its place in clinical practice and many UK clinicians are discouraged for economic reasons from prescribing it. We adopt a 'number needed to treat' (NNT) approach to assess the benefits reported in memantine trials. METHOD: We searched Medline and the Cochrane Dementia and Cognitive Improvement Specialised Register for double-blind, randomised and controlled trials of memantine in AD. If efficacy was only reported in terms of mean change, rather than as number of individuals who responded or were harmed by an intervention, we contacted the drug companies (Merz and Lundbeck) to ask for more data. We also calculated effect size.
RESULTS: We found two trials of memantine in AD that met our criteria. We found that NNTs for global outcome were 3 and 6, for cognitive outcome 7 and for activities of daily living 4 and 8. The effect size for memantine varied between 0.32 and 0.62. For NNH memantine was no more harmful than placebo and significantly less so for the outcome of agitation. CONCLUSION: The small NNTs and the lack of harm shown by the NNHs strongly suggest that memantine, as with cholinesterase inhibitors, has a valuable place in the current clinical management of AD. The effect sizes are mainly in the 'medium' range for clinical effect, which also suggests that memantine has a clinical place in terms of cognition and dependency. There remains a need for more studies that examine carer burden, behavioural and psychological effects, and quality of life for both the person with dementia and the caregiver.