Galantamine (also called galanthamine, marketed as Reminyl (Janssen)) can be isolated from several plants, including daffodil bulbs, and now synthesized. Galantamine is a specific, competitive, and reversible acetylcholinesterase inhibitor. It is also an allosteric modulator at nicotinic cholinergic receptor sites potentiating cholinergic nicotinic neurotransmission. A small number of early studies showed mild cognitive and global benefits for patients with Alzheimer's disease, and recently several multicentre clinical trials have been published with positive findings. Galantamine has received regulatory approval in Sweden, is available in Austria, and awaits marketing approval in the United States, Europe, and other countries. The objective of this review is to assess the clinical effects of galantamine in patients with probable Alzheimer's disease, and to investigate potential moderators of an effect. The Cochrane Dementia Group specialized register of clinical trials was searched using the terms 'galantamine,' and 'galanthamine' (15 February 2000) as was the Cochrane Controlled Trials Register (2000, Issue 2). These terms were also used to search the following databases: EMBASE, MEDLINE, PsychLit; Combined Health Information Database, NRR (National Research Register), ADEAR (Alzheimer's Disease Education and Referral Centre clinical database, BIOMED (Biomedicine and Health), Glaxo-Wellcome Clinical Trials Register, National Institutes of Health Clinical Trials Databases, Current Controlled Trials, Dissertation Abstracts (mainly North American dissertations) 1961-1994, Index to UK Theses (British dissertations) 1970-1994. Published reviews were inspected for further sources. Additional information was collected from an unpublished investigational brochure for galantamine. Trials selected were randomized, double-blind, parallel-group, and unconfounded comparisons of galantamine with placebo for a treatment duration of greater than 4 weeks in people with Alzheimer's disease. Data were extracted independently by the reviewers and pooled where appropriate and possible. The pooled odds ratios (95%CI) or the average differences (95%CI) were estimated. Intention-to-treat and observed cases data were both reported, if the data were available to be reported. -Outcomes of interest include the Alzheimer's Disease Assessment Scale-cognitive subscale (ADAS-cog), clinical global impression of change (CIBIC-plus or CGIC), Alzheimer's Disease Cooperative Study/Activities of Daily Living (ADCS-ADL), Disability Assessment for Dementia scale (DAD) and Neuropsychiatric Inventory (NPI). - Potential moderating variables of a treatment effect included trial duration and dose. Seven trials were identified that met criteria for entry, with 6 being Phase II or III industry-sponsored multicentre trials. One was of 12 weeks duration; one of 5 months; one of 29 weeks; and the rest of 6 months duration. Trials of 5 months or more were aggregated in the analyses as '6 months'. Overall, galantamine showed significant treatment effects at daily doses of 16-32 mg/d for trials of 3- to 6-months duration. For global ratings, trials of 3 months duration with doses of 24-32mg/d (Odds Ratio (OR) 2.2; 95%CI 1.4 to 3.7) and 36mg/d (OR 3.3; 95%CI 1.2 to 9.3) were statistically significant in favour of treatment. For trials of 6 months duration (5-months to 29 weeks), only doses of 8mg/d failed to be statistically significant (24mg: OR 2.0; 95%CI 1.5 to2.5; 32mg: OR 1.9; 95%CI 1.4 to 2.5). For cognitive function over 6 months duration: at a 24mg/d, improvements measured -3.5 points (k=3; 95%CI -4.3 to -2.8) on weighted mean difference on the ADAS-Cog scale, and -4.0 points at 32mg/d (k=2; 95%CI -5.0 to -3.0). Both observed cases (WMD 3.8; 95%CI 0.3 to 7.3) and intention to treat analyses using the Disability Assessment of Dementia gave statistically significant results in favour of treatment for daily doses of 32mg for 6 months duration. The small number of trials available for analysis, however, limited the power of analyses to detect differences. Galantamine consistently failed to show statistically significant treatment effects at doses of 8mg/day. Galantamine's adverse effects appear similar to those of other cholinesterase inhibitors, in that it tends to produce gastrointestinal effects acutely and with dosage increases. Overall, people treated with galantamine at doses of 24-32 mg/d were more likely to discontinue participation in most trials than were people treated with lower doses or placebo, but in the one trial with a slower rate of titration the discontinuation rate was not significantly greater than placebo for the 16 mg/day dose. (ABSTRACT TRUNCATED)