OBJECTIVE: The relationship between number of apolipoprotein E epsilon 4 (APOE epsilon 4) alleles and the rate of cognitive decline in patients with Alzheimer's disease was examined. METHOD: Rate of decline in score on the Mini-Mental State was measured during the active phase of the decline curve between Mini-Mental State scores of 23 and 0. To characterize onset, the authors also estimated for each subject the age at which the Mini-Mental State score fell below 23 and obtained a retrospective report of age at onset from the caregiver. The number of APOE epsilon 4 alleles carried by each subject was determined from genomic DNA samples. The study included 86 subjects with probable Alzheimer's disease who had had at least two cognitive evaluations (a mean of 5.6 evaluations per subject over an average period of 3.6 years). RESULTS: The results did not support an association between APOE epsilon 4 dosage and rate of cognitive decline. Age at onset and age at which the Mini-Mental State score fell below 23 were also not related to APOE epsilon 4 dosage. The APOE allele frequencies were similar to those in other studies of subjects with Alzheimer's disease, showing an enrichment of the epsilon 4 allele. CONCLUSIONS: Although the APOE epsilon 4 allele is a risk factor for Alzheimer's disease, there is no support of a strong association between APOE epsilon 4 dosage and rate of cognitive decline. The epsilon 4 allele did not predict age at onset. Methodological inconsistencies may account for discrepancies between these results and previous findings.