Abstract
CONTEXT: Poor social relationships have been associated with increased mortality risk. Configurations of social relationships vary across the life course, affecting the association with mortality. We investigated the relationship between social integration and mortality in four age groups and to examine three blocks of potential intermediate variables. METHODS: Tromsø Study wave 4 and 6 data and mortality events in the Norwegian Population Registry were linked. A three-level (low, medium, high) social integration index (SII) was constructed from scores to three related items: social participation, living with spouse and frequency of contact with friends. Adjusted cox-regression models and including intermediate variables in blocks (1:behavioral, 2:psychosocial resources, 3:biological markers) were conducted in the study population (N = 9879) and in age-stratified subsamples: [25-52]y. (N = 6785); [53-59]y. (N = 1679); [60-65]y. (N = 846); [66-74]y. (N = 569). Additional sex-stratified and causes-of-death models (cancer, cardiovascular diseases, neuropsychiatric conditions and other deaths) were constructed. RESULTS: A graded SII and mortality risk association was observed in confounder-adjusted models in the total study population: individuals with medium and low SII had incrementally higher risk of mortality compared to those with high SII (HR(mediumSII)[95%CI] = 1.15 [1.02; 1.31], HR(lowSII)[95%CI] = 1.42[1.14; 1.76]). This association was mainly observed in the [53-59]y. group and among men. Health behaviors and the psychosocial resources block partly attenuated the association, but biological markers did not. Low SII was particularly associated with deaths from "other causes", including respiratory conditions, and to a lesser extent with deaths from cardiovascular disease and cancer. CONCLUSION: Our study consolidates evidence on the relationship between social integration and mortality risk, especially in mid-adulthood, and suggests some potential underlying pathways. To reduce mortality risk inequalities, we encourage structurally addressing social marginalization.