Background: Dementia is an emerging public health issue. Growing evidence emerged on the association between social integration and the risk of dementia. However, the magnitude of the association between different aspects of social integration and the risk of dementia is unclear.

Methods: Five databases were systematically searched. Newcastle-Ottawa scale for assessing the quality of the reporting was used for quality appraisal. Longitudinal cohort studies examining the association between social integration and the risk of dementia were analyzed using random effects model. A series of sensitivity analyses was conducted to enhance the robustness of the findings.

Results: Forty publications generated from 32 studies/databases were included. The meta-analysis showed that strong social engagement (overall RR = 0.81, 95% CI = 0.74-0.89, p < 0.001) and frequent social contact (overall RR = 0.86, 95% CI = 0.76-0.97, p = 0.018) were positively associated with decreased risk of dementia. The influence of social support (overall RR = 0.92, 95% CI = 0.80-1.06, p = 0.238) and close social contact (overall RR = 0.74, 95% CI = 0.48-1.13, p = 0.167) was not significant. Loneliness was significantly associated with an increased risk of dementia (overall RR = 1.42, 95% CI = 1.26-1.60, p < 0.001), whereas the influence of social isolation (overall RR = 1.58, 95% CI = 0.80-3.12, p = 0.192) was not significant. A larger social network size (RR = 0.75, 95% CI = 0.59-0.97, p = 0.028) was a promising influencing factor even though the number of studies was insufficient for a meta-analysis. However, the heterogeneity among studies was generally high even though sensitivity analysis was conducted.

Conclusions: Our findings reveal that high social engagement and frequent social contact are significantly associated with a lower risk of dementia, whereas loneliness is associated with a higher risk. The promising impact of large social network size is also identified. Substantial heterogeneity appeared in most of the analysis, making the inference tentative. Nevertheless, the sensitivity analysis provided valuable implications that enhancing social engagement and reducing loneliness may prevent or delay the onset of dementia among middle-aged and older adults.