The association of multimorbidity and neuropsychological test scores in Canadian cohorts recruited in 1991 and 2015

Abstract

AbstractBackgroundMultimorbidity is associated with cognitive decline. Little information exists regarding whether the relationship between multimorbidity and cognitive decline has changed over the last few decades.ObjectiveTo compare estimates of the association between multimorbidity (MM) and performance on cognitive tests in two prospective cohorts of Canadian community‐dwelling older adults recruited 24 years apart.MethodWe analyzed the two datasets separately, including participants ≥65 years from both baseline and first follow‐up of the Canadian Study of Health and Aging (CSHA, 1991 – 2001; first follow up in 1996) and the Canadian Longitudinal Study on Aging (CLSA, 2015 – ongoing; first follow up in 2018). We excluded participants with baseline dementia. In both cohorts, we defined MM as two or more conditions from a list of 14. The neuropsychological tests were Animal Naming Test (ANT) and the Rey Auditory Verbal Learning Test (RAVL) for both cohorts. Tests of frontal function were the Digit Symbol Substitution Test (DSST) in the CSHA and the Mental Alternation Task (MAT) in the CLSA. We performed multilevel linear modelling. We controlled for confounders, which were detected using a Directed Acyclic Graph.ResultWe included 497 participants from the CSHA and 9308 from the CLSA. The mean age was 78.0 in the CSHA and 72.0 in the CLSA, and male accounted for 36.4% in the CSHA and 50.2 in the CLSA. Mean MM was 2.1 in both cohorts, with a higher prevalence of MM in women (CSHA 63.9%; CLSA 64.2%) than in men (CSHA 55.8%; CLSA 59.5%). When comparing both cohorts, we found that baseline MM was not associated with changes in the Z‐scores for the ANT (CSHA: 0.16, CI95: ‐0.014 to 0.34; CLSA: ‐0.002, CI95: ‐0.044 to 0.04), RAVL delayed recall (CSHA ‐0.033, CI95: ‐0.23 to 0.16; CLSA: 0.031, IC95: ‐0.013 to 0.074), or DSST in CSHA (0.1, CI95: ‐0.084 to 0.29) / MAT in CLSA (0.008, CI95: ‐0.035 to 0.052).ConclusionWe did not find differences in Z‐scores changes over 5 years follow up in the CSHA and 3 years follow up in the CLSA, for three neuropsychological tests among Canadians ≥65 years recruited over two decades apart.