Periodontitis and Inflammatory Bowel Disease

A Meta-analysis

Yang-yang She; Xiang-bo Kong; Ya-ping Ge; Zhi-yong Liu; Jie-yu Chen; Jing-wei Jiang; Hong-bo Jiang; Si-lian Fang


BMC Oral Health. 2020;20(67) 

In This Article


Search Strategy

In this meta-analysis, we followed the method described in the Meta-Analysis of Observational Studies in Epidemiology guidelines.[21] The following major databases were screened for the following terms: ((((inflammatory bowel disease) OR Crohn's disease) OR ulcerative colitis)) AND ((((((((periodontal diseases) OR gingival diseases) OR periodontitis) OR gingivitis) OR edentulous) OR edentulism) OR tooth loss) OR teeth loss). English-language publications were extracted from Web of Science, PubMed, Cochrane, and Embase; while Chinese-language articles were retrieved from China National Knowledge Infrastructure (CNKI), Wangfang and CQVIP. The search was limited to literature from before August 1, 2019, including primary researches.

Study Selection

Eligible studies were examined by authors (Yang-yang She and Xiang-bo Kong) independently. Final selection was verified by a third author (Hong-bo Jiang) and disagreements were resolved by discussions. The inclusion criteria of an eligible study were as follow: (1) related to periodontal conditions in patients with IBD; (2) provided at least one of the clinical parameters: bleeding on probing (BOP), clinical attachment loss (CAL), oral plaque index (PI), gingival index (GI), gingival recession (GR), probing pocket depth (PPD), papilla bleeding index (PBI); (3) reported original data; (4) presented cross-sectional studies, cohort studies or case-control studies; (5) Full text in English or Chinese. Case reports, case series, in vitro studies, reviews, abstracts, editorials, and letters were excluded from the selection. In case where multiple publications were based on the same population, the more recent or complete report were considered.

Data Extraction

Data extraction conducted by Yang-yang She and Xiang-bo Kong was based on a standardized, pre-piloted data extraction form. The extracted information included: (1) last name of the first author, publication year, study location, study design and matched variables, (2) study participant demographics including male to female ratio, mean age, total number of IBD, CD, UC cases and controls, (3) periodontal measures including prevalence of periodontitis and risk estimate, BOP, CAL, GI, GR, PI, PPD, and PBI, (4) pharmacological treatments, (5) smoking status of the IBD patients and controls, (6) adjusted variables.

Quality Assessment

Newcastle-Ottawa Scale (NOS) was employed to evaluate the methodological quality of the included studies.[22] Studies with at least five scores were defined as moderate or high methodological quality.

Statistical Methods

The estimates (or adjusted estimates if applicable) and the corresponding 95% confidence interval (CI) between IBD and periodontitis were used to calculate the pooled estimates. If no estimates were available in the studies, the numbers of IBD cases (with periodontitis or not) and controls (with periodontitis or not) were used to calculate the pooled estimates. Forest plots were performed to assess the individual and pooled estimates with the corresponding 95% CI. The random effect model would be applied if the Cochrane Q statistic with a significant level of P < 0.10 or I2 > 50%.[23] The statistical analysis was accomplished with Stata Version 11.0 (Stata Corp, College Station Texas). The P values from the two-tailed tests are statistically significant when P < 0.05, except where otherwise specified.

Begg's test and Egger's test were applied to test potential publication bias.[24] Funnel plot was generated to assess publication bias by the visual inspection of asymmetry. The P values for Begg's test and Egger's test less than 0.05 revealed that publication bias existed.

To examine the impact of a single estimate on the pooled estimates, we performed sensitivity analysis by systematically removing one study at a time and recalculating the pooled results.