Background: Metrifonate is a long-acting irreversible cholinesterase inhibitor, originally used to treat schistosomiasis. Its potential to enhance central nervous system cholinergic neurotransmission led to clinical trials for the treatment of people with Alzheimer's disease (AD). Although low incidence of serious side effects occurred during short-term use as an antihelmintic, in studies of the treatment of AD extending over 6 months, 20 patients experienced respiratory paralysis and problems with neuromuscular transmission. These findings have led to a halt to trials of metrifonate for AD and Bayer, the pharmaceutical company, has withdrawn its FDA application.

Objectives: 1) To establish the efficacy of metrifonate for patients with Alzheimer's disease, in terms of cognition, global impression, functional activity, non cognitive symptoms, rate of institutionalization and mortality.2) Assess the safety and tolerability of metrifonate.

Search strategy: The Cochrane Dementia and Cognitive Improvement Group's Specialized Register was searched on 5 December 2005 using the term metrifonat*. This Register is regularly updated with records from all major health care databases (MEDLINE, EMBASE, CINAHL, PsycINFO) and many trials databases. One of the authors (LS), as member of the Metrifonate Study Group has had the opportunity to contact other metrifonate trialists to obtain data from potentially non published data of metrifonate clinical trials.

Selection criteria: All unconfounded, randomized double-blind clinical controlled trials comparing metrifonate to placebo in people with AD.

Data collection and analysis: Data were extracted by the two reviewers, cross-checked, and pooled when appropriate and possible.

Main results: Most studies assessed changes in cognitive function, global function, activities of daily living, behavioural problems, severity of disease and adverse events. Occasionally the results were not reported in sufficient detail to allow extraction of data for the meta-analyses. The treatment regimens were varied: loading doses were used in some trials. The range of maintenance doses and studies were not pooled unless the treatment regimens were considered comparable. The lengths of treatment varied from 6 to 26 weeks and studies were not pooled unless the treatment duration was similar. The results are derived from the ITT populations. Metrifonate at various doses, fixed and loading doses, was associated with significant cognitive improvement compared to placebo, except for weekly doses where there was no difference from placebo: MMSE (metrifonate 60-80 mg/day with initial loading at 26 weeks; metrifonate 50 mg/day fixed dose with no initial loading at 26 weeks MD 1.85, 95% CI 1.06 to 2.64, p<0.00001); ADAS-Cog (metrifonate 60-80 mg/day with initial loading at 26 weeks MD -3.24, 95% CI -4.40 to -2.08, p<0.00001)In most trials, there was improvement in clinical global impression: CIBIC-Plus (metrifonate 30-55 mg/day, approximately 0.65 mg/kg body weight, with initial loading at 26 weeks MD -0.25, 95% CI -0.41 to -0.09 p=0.002; metrifonate 50 mg/day fixed dose with no initial loading at 26 weeks MD -0.20, 95% CI -0.39 to -0.01, p=0.04). There were generally-significant drug-placebo differences in activities of daily living but this often depended on sample size and the characteristics of the instrument used: DAD (metrifonate 30-55 mg/day, 0.65 mg/kg body weight, with initial loading at 26 weeks MD 2.72, 95% CI 0.66 to 4.77, p=0.01; metrifonate 50 mg/day fixed dose with no initial loading at 26 weeks MD 4.07, 95% CI 0.29 to 7.85, p=0.03)Also there were differences associated with metrifonate compared with placebo for different doses of metrifonate in scores on a behavioural symptom scale, caregiver burden scale, and severity of disease scale. Adverse events occurring more often with metrifonate included abdominal pain, bloating, bradycardia, diarrhoea, leg cramps, nausea and rhinitis and were described as mostly mild and transient, but occasionally moderately severe, and infrequently severe and serious. Analysis of the number of patients suffering at least one mild, moderate, severe or serious adverse event before the end of treatment showed that there was usually no difference between placebo and metrifonate.

Authors' conclusions: Metrifonate given once per day appears to be related to clinical response in cognition, global improvement, and activities of daily living in patients with mild to moderate Alzheimer's disease. Tolerability is good with adverse events as expected from a cholinesterase inhibitor, but with a low incidence of neuromuscular dysfunction and respiratory failure, too low to be detected in this review. It has been withdrawn from further development.