We describe the follow-up of a cohort of 255 Alzheimer's disease (AD) patients (81 males, 174 females) treated by tacrine during 4 years. We performed the survey of hepatic, cholinergic and general tolérance. Drug efficacy was measured by MMS examination on weeks 0, 18, 30, 52, 104, 156 and 208. A total of 190 patients (74.5 percent) were dropped out of this study, 75 (29 percent) for adverse events. We found 85 hepatic (33 percent), 79 cholinergic (31 percent), 31 (12 percent) neuropsychiatric and 72 general (28 percent) side effects. In term of drug efficacy we observed a global decline of 2.5 MMS points during the first year and 2 MMS points between W52 and W156. Tacrine's symptomatic efficacy, defined as the number of patients improved or stabilized at W30, was present in 50 patients (46 percent) among the 109 patients reaching W30. The intensity of symptomatic efficacy was expressed by a 2.7 MMS points increase in 37 patients improved on W30. The long term effects of Tacrine, measured by the MMS score at one year, showed a positive impact as the MMS was 2.5 points above the expected score in non treated AD patients. This study raises the practical problem of optimal cholinesterase inhibitors use in AD and the theoretical question of long term action of cholinesterase inhibitors on cerebral lesions of AD.