OBJECTIVE: Dementia with Lewy bodies (DLB) is the second commonest form of dementia. The response to acetylcholinesterase inhibition (AChEI) could be greater in DLB than in Alzheimer's disease (AD) because cholineacetyl-transferase levels are more reduced in the former. This preliminary trial seeks to compare performances in cognitive tasks before and after tacrine administration in DLB and AD subjects. METHODS: Six DLB and 6 AD patients were enrolled in an open, nonrandomized, intervention trial using 80 mg/day tacrine. Patients met ADRDA or DLB consortium criteria for probable diseases. Subjects were matched for Mini Mental State Examination (MMSE) score, age and sex. Mattis Dementia Rating Scale (DRS), Controlled Oral Word Association Test (FAS) and Boston Naming tests were administered at baseline and at 6 months into treatment. RESULTS: AD and DLB groups did not differ in initial mean total DRS scores. In the primary analysis, both groups declined during the course of treatment (-7.3 +/- 4.2 and -16.8 +/- 39.2 DRS points, respectively). Due to the large variability in DLB posttreatment scores, this group was divided post hoc into responders (DLBr) and nonresponders (DLBnr). The DLBr group outperformed the DLBnr group at baseline (p < 0.05) and, notably, in follow-up DRS test scores (p < 0.001). Two-way MANOVA comparing both DLB subgroups with either the entire AD cohort or similarly stratified AD subgroups showed a significant interaction (F = 7.6; p < 0.015), attributed mostly to declines in DLBnr group scores (p < 0.01). Surprisingly, on DRS memory subscale and FAS tests, there were significant improvements in DLBr scores (p < 0.02). A baseline MMSE (or DRS memory) score >/=15 predicted a positive response to tacrine in DLB. Acceleration of parkinsonism occurred in all DLB subjects. CONCLUSION: Results from a primary analysis of the therapeutic effect of 80 mg/day tacrine in DLB and AD were negative. However, post hoc analysis showed that mild to moderate DLB responds favorably to AChEI relative to AD through stabilization of global cognitive decline and improvements in specific cognitive areas. These results could be useful in the planning of a more definitive study.