Association between inflammation and cognition: triangulation of evidence using a population-based cohort and Mendelian randomization analyses

ABSTRACT Background There is evidence for an association of inflammation with cognitive functioning and dementia in older adults, but the association with cognitive functioning in youth and whether this is causal remains unclear. Methods In a population-based cohort (Avon Longitudinal Study of Parents and Children; ALSPAC), we investigated cross-sectional associations of inflammatory markers (C-reactive protein [CRP] and Glycoprotein acetyls [GlycA]) with measures of cold (working memory, response inhibition) and hot (emotion recognition) cognition at age 24 (N=3,305 in multiple imputation models). Furthermore, we conducted one-sample and two-sample bidirectional Mendelian randomization (MR) analyses to examine potential causal effects of genetically-proxied inflammatory markers (CRP, GlycA, Interleukin-6, soluble Interleukin-6 receptor) on cognitive measures (above) and general cognitive ability. Results In the ALSPAC cohort, there was limited evidence of an association between inflammatory markers and cognitive measures at age 24 after adjusting for potential confounders (N=3,305; beta range, -0.02 [95% confidence interval (CI) -0.06 to 0.02, p =.29] to 0.02 [95% CI -0.02 to 0.05, p =.38]). Similarly, primary MR analyses found limited evidence of potential effects of genetically-proxied inflammatory markers on working memory, emotion recognition or response inhibition in one-sample MR using ALSPAC data (beta range, -0.73 [95% CI -2.47 to 1.01, p =.41] to 0.21 [95% CI -1.42 to 1.84, p =.80]; or on general cognitive ability in two-sample MR using the latest Genome-Wide Association Study (GWAS) datasets (beta range, -0.02 [95% CI -0.05 to 0.01, p =.12] to 0.03 [95% CI -0.01 to 0.07, p =.19]. Conclusions Our findings do not provide strong evidence of a potential causal effect of inflammatory markers (CRP, Interleukin-6, GlycA) on the cognitive functions examined here. Given the large confidence intervals in the one-sample MR, larger GWAS of specific cognitive measures are needed to enable well-powered MR analyses to investigate whether inflammation causally influences hot/cold cognition.