Television viewing and risk of mortality: exploring the biological plausibility

Background and aims Television (TV) viewing is a major component of leisure sedentary time, and has been consistently associated with cardiovascular disease. We examined the extent to which metabolic biomarkers explain the association between TV viewing and mortality. Methods Participants (N = 8,451, aged 64.8 ± 9.9 yrs) were drawn from The English Longitudinal Study of Ageing (ELSA), a national prospective cohort study of community-dwelling men and women living in England. The individual participant data were linked with death records from the National Health Service registries from 2008 to 2012. Cox proportional hazards models were used to estimate the risk of death according to time spent watching TV, with biomarkers added in a stepwise fashion to estimate potential mediation. Results Over an average follow up of 4 years (33,832 person years), there were 370 deaths. In models adjusted for comorbidities, psychosocial factors, and health behaviours including physical activity, there was an association between TV viewing and mortality (≥6 h per day vs. < 2 h per day [Ref]; Hazard ratio = 1.98, 95% CI, 1.25, 3.15). Adjustment for inflammatory markers (C-reactive protein and fibrinogen) accounted for ∼15.7% of the association between TV viewing and mortality, but metabolic risk factors (HDL-cholesterol, triglycerides, glycated haemoglobin) did not contribute. Conclusions The association between TV viewing and mortality was partly mediated by inflammatory markers, although the relationship remains largely unexplained.