Bidirectional association between periodontal disease and diabetes mellitus: a systematic review and meta-analysis of cohort studies

Periodontal disease has been reported to be associated with diabetes mellitus. However, the direction of the association and the influence of bias are not clear. Thus, the aim of this systematic review and meta-analysis was to summarize the existing evidence on the bidirectional prospective association between periodontal disease and diabetes mellitus by accounting for the risk of bias of the original studies. The literature search was conducted on the electronic data sources PubMed and Web of Science up to February 9th, 2021. We included observational studies, which investigated the prospective association between diabetes mellitus and periodontal disease or vice versa. The risk of bias of the primary studies was evaluated by applying the Quality in Prognosis Studies (QUIPS) tool. Random effects models were used to calculate summary relative risk (SRR) with 95% CI. Subgroup analyses were applied to investigate heterogeneity and the robustness of the finding. In total, 15 studies were included . The SRR for incident diabetes mellitus was 1.26 (95% CI 1.12, 1.41; I(2): 71%, n = 10; participants = 427,620; identified cases = 114,361), when comparing individuals with periodontitis to individuals without periodontitis. The SRR for incident periodontitis was 1.24 (95% CI 1.13, 1.37; I(2): 92%, n = 7; participants = 295,804; identified cases: > 22,500), comparing individuals with diabetes to individuals without diabetes. There were no significant differences between subgroups after stratification for risk of bias. The findings show a positive bidirectional association between periodontal disease and diabetes mellitus, and thus, underline the need for screening of patients with periodontitis regarding diabetes mellitus and vice versa. The main limitation of the study is the high unexplained heterogeneity between the studies including the different assessment methods of the disease diagnosis.