Domain-Specific Cognitive Concordance in Older Married Couples

Importance: Spousal similarity in late-life cognition has been reported, but whether concordance differs across cognitive domains remains unclear. Objective: To examine whether cognitive concordance among older married couples is domain specific by comparing real couples with demographically and genetically matched control pairs. Design, Setting, and Participants: This nationwide, multicenter, community-based cross-sectional study was conducted from January 2019 to December 2020 and included married couples from the Korean Longitudinal Study of Cognitive Aging and Dementia. Each spouse was individually matched to an unrelated control based on sex, cognitive diagnosis, apolipoprotein E ε4 status, age, and education. Data were analyzed from March to April 2026. Exposure: Marital pairing within a 4-group matched design (real couples, spouse-control pairs, and matched null couples). Main Outcomes and Measures: For 9 subtests from the Korean version of the Consortium to Establish a Registry for Alzheimer Disease Assessment Packet, intraclass correlation coefficients (ICCs) were calculated within each group, and the spousal concordance (ICC difference) was defined as the difference between the ICC of real couples and that of matched null couples. Significance was assessed with 2-sided permutation tests and Benjamini-Hochberg false discovery rate correction. Results: The sample included 783 married couples (1566 individuals; 783 [50.0%] female and 783 [50.0%] male; mean [SD] age, 74.0 [5.5] years) and 783 matched control couples. Among the 783 married couples, 4 of 9 subtests showed significant ICC differences after false discovery rate correction in unadjusted analyses: Verbal Fluency (ICC difference, 0.232; 95% CI, 0.138-0.324), Constructional Praxis (ICC difference, 0.145; 95% CI, 0.039-0.252), Boston Naming Test (ICC difference, 0.134; 95% CI, 0.038-0.231), and Constructional Recall (ICC difference, 0.110; 95% CI, 0.017-0.200); all remained significant after covariate adjustment. In contrast, the Mini-Mental State Examination and verbal memory recall and recognition showed little concordance beyond demographically matched null pairs. Conclusions and Relevance: In this cross-sectional study of 783 older married couples, spousal cognitive concordance was domain specific, with the greatest effect size estimates in language and visuospatial function rather than global cognition. These findings suggest that spousal cognitive similarity may be missed when analyses rely only on global screening measures, highlighting the need for longitudinal study of within-couple concordance.