Longitudinal association between multimorbidity, participating activity and cognitive function: cross-lagged mediation analysis

Background
Previous studies have reported an association between multimorbidity and cognitive function, however, the specific direction and underlying mechanism remain unclear. The study aimed to explore the direction of this association and to examine the role of physical activity and leisure activity among older adults.

Methods
Data from 5,546 dementia-free Americans aged 60 or above of 2008 (T1) and 2016 (T2) of the Health and Retirement Study were used. Multimorbidity was measured by the multimorbidity weight index. Cognitive functioning was measured by the Telephone Interview of Cognitive Status. We used cross-lagged panel models to determine the associations between multimorbidity and cognitive function and examine the mediation effect of physical and leisure activity.

Results
There was a bidirectional association between multimorbidity and cognitive function. More severe multimorbidity predicted worse cognitive function (β = -0.064, SE = 0.016) and vice versa (β = -0.024, SE = 0.009). Paths from multimorbidity to cognitive function were stronger than those from cognitive function to multimorbidity. Physical and leisure activity mediated the association between multimorbidity (T1) and cognitive function (T2), and the association between cognitive function (T1) and multimorbidity (T2). The bidirectional association between multimorbidity and cognitive function was only observed in APOE ε4 noncarriers.

Conclusions
A negative bidirectional association was observed between multimorbidity and cognitive function. Additionally, the association is mediated by physical and leisure activity.