The reciprocal associations between social deficits, social engagement, and inflammation: Longitudinal evidence comparing venous blood samples and dried blood spots and mapping the modifying role of phenotypic and genotypic depression

Social psychoneuroimmunology suggests an interplay between social deficits (loneliness and isolation) and chronic inflammation, but the direction of these relationships remains unclear. We estimated the reciprocal associations of social deficits and social engagement with levels of C-reactive protein (CRP), compared the consistency of the findings depending on the biological sampling method used, and examined the modifying role of phenotypic and genotypic depression. We used longitudinal nationally representative data from the US (Health and Retirement Study, 3 waves, 2006–16) and England (English Longitudinal Study of Ageing, 4 waves, 2004–18). Loneliness, social isolation, and social engagement were self-reported. CRP was measured using dried blood spots (US) and venous blood samples (England). Cross-lagged panel models were fitted and tested interactions with phenotypic depression (above-threshold depressive symptom scores) and genotypic depression (polygenic score for major depressive disorder). We included 15,066 participants (mean age = 66.1 years, SD = 9.8) in the US and 10,290 (66.9 years, SD = 10.5) in England. We found reciprocal associations between loneliness and CRP using dried blood spots and venous blood samples. Higher CRP predicted higher subsequent loneliness and higher loneliness predicted elevated CRP. Both phenotypic and genotypic depression modified this reciprocal association. There were also reciprocal associations for social engagement in venous blood samples: higher CRP predicted lower social engagement and greater social engagement predicted lower subsequent CRP. Associations between social isolation and CRP were inconsistent and unidirectional. Loneliness may increase chronic inflammation, whereas social engagement may reduce inflammation. As these relationships were reciprocal, there may be a loop between inflammation, loneliness, and social engagement. This loop was stronger in those with depression or at high genetic risk for major depressive disorder. This relationship for loneliness was present in both blood sampling methods despite contrasting methods of CRP measurement, indicating that the finding is not attributable to measurement bias in biomarkers.