Objective: To determine whether apolipoprotein E (Apo E) phenotype is associated with cognitive decline in community-dwelling nondemented elderly women.
Design: Prospective cohort study.
Setting: A university-affiliated clinic near Pittsburgh, Pa.
Patients: A total of 1750 nondemented community-dwelling women, aged 65 years and older, who were enrolled in the Study of Osteoporotic Fractures.
Main outcome measures: The women completed a baseline interview and performed 3 cognitive tests: the modified Mini-Mental State Examination, Trials B, and Digit Symbol. Serum samples were obtained for Apo E typing. Baseline cognitive scores and repeated scores approximately 6 years after study enrollment were compared in women with and without Apo E epsilon 4. Cognitive decline, defined as the worst 10th percentile change scores, was assessed for each test and by phenotype group.
Results: After adjustment for age, education, presence of severe tremor, and depression, baseline scores did not differ by Apo E epsilon 4 status except for lower scores on Trails B in the homozygous epsilon 4 group (mean score, 159.7 compared with 127.7 for the heterozygous epsilon 4 group and 125.4 for the no epsilon 4 group; P = .01). However, repeated test performance on follow-up examination was worse on all tests in those women with 1 or more epsilon 4. Reduction on the modified Mini-Mental State Examination was 0% for no epsilon 4 allele, 1.9% for 1 epsilon 4 allele, and 3.7% for 2 epsilon 4 alleles (P < .001); reduction on Digit Symbol was 6.2% for no epsilon 4 allele, 9.0% for 1 epsilon 4 allele, and 17.5% for 2 epsilon 4 alleles (P = .04); and reduction on Trials B was 5.9% for no epsilon 4 allele, 25.0% for 1 epsilon 4 allele, and 10.9% for 2 epsilon 4 alleles (P = .002). Women with at least 1 epsilon 4 had an odds ratio of 1.6 (95% confidence interval, 1.1-2.3) of having cognitive decline during the study period.
Conclusion: Apolipoprotein E epsilon 4 is associated with cognitive decline in community-dwelling nondemented women.