OBJECTIVE To assess the possible association between tacrine (Cognex, manufactured by Parke-Davis, Morris Plains, NJ) dose and likelihood of nursing home placement (NHP) or death in patients with AD. METHODS A 30-week, randomized, double-blind, placebo-controlled, parallel-group multicenter clinical trial involving 663 patients, after which patients were treated openly and followed up a minimum of 2 years later. METHODS At baseline, outpatients were at least 50 years of age, met criteria for probable AD, with baseline Mini-Mental State Examination scores between 10 and 26 (inclusive), were otherwise healthy, and had a caregiver who could provide assessments and ensure medication compliance. METHODS mandomized assignment to placebo or one of three ascending dosage regimens of tacrine over 30 weeks, followed by open label treatment for all patients who began the double-blind trial. METHODS NHP and death were examined using logistic regression. RESULTS PATIENTS who remained on tacrine and were receiving doses > 80 mg/d or > 120 mg/d were less likely to have entered a nursing home than patients on lower doses (odds ratios > 2.7,2.8, respectively.) There was a trend for lower mortality for patients receiving > 120 mg/d (p = 0.063). CONCLUSIONS Treatment with tacrine at doses > 80 mg/d was associated with a reduced likelihood of NHP. These data demonstrate that tacrine's 30-week effects on cognitive function and clinicians' global ratings may generalize to effects on a major milestone of AD. Future studies should attempt to replicate these findings prospectively.