Background Social isolation has been associated with increased mortality, and a much-cited previous review stated that the risk is comparable to smoking 15 cigarettes per day. However, the available evidence is inconsistent. We examined social isolation in relation to all-cause mortality in two UK prospective cohorts, and assessed whether the excess risk associated with social isolation was comparable to that for smoking 15 cigarettes per day.

Methods After excluding people with vascular disease, cancer or low self-rated health, to minimise reverse causation bias, 326,169 Million Women Study (MWS) participants (mean age=68 years) and 296,913 UK Biobank (UKB) participants (mean age=56 years), were followed for death. Social isolation was measured using an index of self-reported frequency of contact with family or friends, social group contact, and living alone. Adjusted RRs for all-cause mortality were calculated using Cox regression, comparing most isolated participants to the least isolated. Analyses adjusted for smoking and 12 other possible confounding factors.

Results Within each cohort, about 12% of participants were classified as most isolated and 44% as least isolated. Over 5.9 years of follow-up, 9667 MWS participants died; the most isolated had about a 30% excess risk of all-cause mortality compared to the least isolated (RR =1.28, 1.19–1.38). Over 6.8 years follow-up, 4694 UKB participants died; the most isolated had about a 40% excess risk of mortality compared to the least isolated (RR =1.38, 1.27–1.51). Of the constituent measures contributing to isolation, living alone was most consistently associated with an excess mortality. Previous analyses in MWS found smoking 15 cigarettes per day was associated with about a 180% excess risk of mortality compared to never smokers (RR =2.77, 2.72–2.82).

Conclusion While social isolation was associated with about a 30–40% excess risk of all-cause mortality, the excess risk associated with smoking 15 cigarettes per day was 4–6 fold greater, at around 180%.