Background Social connections are recognized as a major determinant of survival and wellbeing. Around half of Canadians aged 80 and older live alone and more senior women live alone than senior men. The link between different social ties and central obesity in women and men is poorly understood, and the interplay of different social ties is seldom studied. This population-based study examined multiple structural social ties in relation to waist circumference (WC) among women and men in Canada.

Methods We used baseline data from the population-based Canadian Longitudinal Study on Aging (2012–2015) Comprehensive cohort of 28,238 adults (45–85 years), and gender-stratified multivariable linear regression models with interaction terms for each pair of social ties. Regression coefficients were used in post-estimation calculation of adjusted means and 95% confidence intervals.

Results The association between marital status and waist circumference was altered by living arrangement, social network size and social participation, with different combined effects for women and men. All non-partnered women who were co-living had higher mean WC than partnered women who were co-living. Mean WC was higher in lone-living than in co-living women in all marital status categories except for widowed. Lone-living widowed women had a lower WC (-0.97 cm [-3.85, 1.91]), compared to lone-living partnered, that was significantly different (p-interaction=0.005) from co-living widowed women (+3.57 [2.26, 4.88]) relative to co-living partnered women. We found that men who were partnered, single, or widowed had higher mean WC when they were also lone-living compared to counterparts who were co-living. Social network size was positively associated with mean WC in women for all marital status categories, except divorced, whereas social participation was inversely associated with mean WC in women for all marital status categories, especially divorced. Combined effects were less clear in men. Greater social participation (5 or more activities) also appeared to mitigate the health-harming influence of having a low social network size on WC in women but not men. Specifically, in the absence of social participation, mean WC reduced by 0.63 cm (-0.88, -0.38) in women and by 0.27 cm (-0.51, -0.03) in men as social network size increased. Notably, living arrangement did not appear to modify the link between either social participation or social network size and WC in either gender.

Conclusion The interplay of different types of social ties revealed an important source of heterogeneity with unique associations with visceral adiposity in women and men. Prevention efforts can be improved by understanding which modifiable social factors are most relevant for obesity in women and men.