Psychobiological dimensions of risk and protection in cognitive aging: Social factors, inflammation, and cognitive test performance in middle-aged and older adults

Background: The prevalence of Alzheimer’s disease and age-related cognitive impairment is on the rise and no treatment is currently available; focus on potentially modifiable risk and protective factors is crucial. Psychosocial determinants including social support and stress have been associated with cognition across the lifespan, with stress-related inflammatory processes proposed as the most likely mechanism behind these associations. We examined relationships between life stress, social support, stress biomarkers, and cognition in late middle-aged adults enrolled in the Wisconsin Registry for Alzheimer’s Prevention (WRAP) study.

Methods: WRAP participants completed neuropsychological tests and psychosocial questionnaires and provided serum samples at two-year intervals. Outcomes included six cognitive factor scores: speed and flexibility (SF), working memory (WM), immediate memory (IM), verbal learning and memory (VLM), visual learning and memory (VIS), and story recall (SR). Predictor variables included index scores for social support and life stress. Potential mediators included two stress biomarkers, serum interleukin-6 (IL-6) and c-reactive protein (CRP). Using mixed models to account for within-subject correlation, we regressed cognitive factor scores on each psychosocial predictor variable, then assessed mediation by stress biomarkers using established Baron and Kenny methods. In secondary analyses, we examined possible interactions between the two psychosocial predictors, and between individual predictors and stress biomarkers.

Results: Sample mean (SD) age of participants (N=1,052) at baseline was 60.2 (6.7) years. After adjusting for demographic and health covariates, primary analyses revealed that social support was associated with SF (B=0.08; p=0.003) but not with memory scores. Life stress was not associated with cognitive scores at the p<0.05 level. Although high IL-6 independently predicted poorer SF (B=-0.13; p=0.008) and IM (B=-0.14; p=0.02), evidence did not support inflammatory mediation of the social support-SF relationship. In secondary analyses, a marginally significant interaction was seen between social support and IL-6 (B=0.08; p=0.05).

Conclusions: Relationships were observed between social support, stress-related inflammation, and cognition; associations varied by cognitive domain, with processing speed and flexibility most strongly linked with markers of stress and coping. Inflammation did not mediate associations between predictors and cognition as proposed, but directly impacted SF; social support, however, may buffer much of that detrimental impact.