Association of mainly vegetarian and vegan diets with loneliness, social isolation and social withdrawal in a German population survey

by André Hajek, Razak M. Gyasi, Supa Pengpid, Karl Peltzer, Karel Kostev, Pinar Soysal, Ashwin Kotwal, Lee Smith, Dong Keon Yon, Hans-Helmut König

Aim There is a very limited number of studies examining how diet is associated with social factors. The aim of this study was therefore to examine the association between type of diet and loneliness, social isolation and social withdrawal.

Methods Cross-sectional data from the general adult population in Germany were used. The sample included 5,000 individuals aged 18, 74 years. To quantify loneliness, the De Jong Gierveld tool was used. Social isolation was measured using the Lubben Social Network Scale and Bude/Lantermann tool. Social withdrawal was operationalized using the 25-item Hikikomori Questionnaire. Three types of diet were considered (mainly vegetarian; mainly vegan; neither vegetarian nor vegan, reflecting an omnivorous diet). The analysis was adjusted for sociodemographic, lifestyle-related, health-related and mental health-related covariates in multiple linear regressions.

Results Overall, 15.4% of the individuals reported following a mainly vegetarian diet and 3.5% of the individuals surveyed followed a mainly vegan diet, with the remaining participants (81.1%) following an omnivorous diet. When adjusting for sociodemographic, lifestyle-related and health-related factors, individuals following a mainly vegetarian diet had significantly higher levels of perceived social isolation and social withdrawal compared to individuals following an omnivorous diet. However, these significant associations disappeared when adjusting for mental health-related covariates. Mainly vegan eaters feel more isolated and withdraw more from society compared to individuals following an omnivorous diet, even in the fully adjusted model.

Conclusion Following a mainly vegan diet was associated with perceived social isolation and social withdrawal, even when adjusting for sociodemographic, lifestyle-related, health-related and mental health-related covariates. However, the variance explained by diet type was minimal (R2 close to zero and partial eta2 < 0.01), indicating limited practical relevance.