Sex Differences in the Associations between Education, Neuropathology, and Cognition
IMPORTANCE Autopsy evidence on sex-specific associations of education with neuropathology and cognition is limited. OBJECTIVE To test sex differences in education associations with neuropathologic burden and cognition. DESIGN Cohort study of data collected from September 2005 through August 2024. SETTING Multicenter autopsy cohort from the National Alzheimer's Coordinating Center. PARTICIPANTS Adults aged 55 years or older; cross-sectional eligibility required a final CDR-SB assessment within 2 years before death, and longitudinal eligibility additionally required at least 3 assessments. EXPOSURES Years of formal education. MAIN OUTCOMES AND MEASURES Neuropathologic measures and the Clinical Dementia Rating Scale Sum of Boxes (CDR-SB). Linear regression and mixed-effects models tested associations and interactions of education with sex, neuropathology, and linear and quadratic time. Benjamini-Hochberg correction addressed multiple comparisons. RESULTS Among 2592 participants (mean [SD] age at death, 81.6 [10.8] years; 1188 [45.8%] female), 1969 contributed to longitudinal analyses. In female participants, higher education was associated with lower Thal amyloid phase, diffuse plaques, neuritic plaques, and Braak stage ({beta} range, -0.097 to -0.071; qFDR<.026); none was significant among male participants. Education-by-sex interactions supported differences in Thal amyloid phase, and diffuse and neuritic plaques ({beta} range, -0.119 to -0.097; qFDR<.045). Education was unrelated to final CDR-SB among female ({beta}, -0.007; 95% CI, -0.052 to 0.038; P=.767) and male participants ({beta}, -0.026; 95% CI, -0.071 to 0.019; P=.254), with no education-by-sex interaction (P=.660). The education-by-Braak-stage-by-sex interaction indicated that higher education was associated with a weaker Braak stage-CDR-SB association among females than males ({beta}, -0.104; 95% CI, -0.171 to -0.038; qFDR=.022). Longitudinally, education was unrelated to linear CDR-SB change in either sex. Education-by-quadratic-time associations were observed among females ({beta}, 0.002; 95% CI, 0.001-0.003; P<.001) and male participants ({beta}, 0.001; 95% CI, 0-0.002; P=.037). The corresponding interaction with sex was also significant (P=.043). CONCLUSIONS AND RELEVANCE Among female participants, higher education was associated with lower selected Alzheimer's disease neuropathologic measures and weaker associations of tau pathology with cognition. Longitudinal trajectories were nonlinear, with slower earlier decline followed by later acceleration at higher education levels.