Methodological issues in assessing the association between periodontitis and caries among adolescents

Abstract

We aimed to investigate the association between dental caries and periodontitis among adolescents participating in a case-control study of periodontitis. In addition, we compared 2 analytical approaches to estimate the association: conventional regression and structural equation modelling (SEM). Along with periodontal recordings, data were obtained on caries, just as subgingival plaque samples were collected. Sociodemographic information was collected with a questionnaire. We used factor analyses to express the many correlated clinical periodontal dimensions in a smaller number of factors. The association between caries (counts of enamel and dentin lesions, or dentin lesions only) and periodontitis was tested using negative binomial regression with robust variance (conventional regression) and weighted least squares (SEM) estimation. Factor analysis revealed 2 different latent periodontal variables: "extent" and "severity" of periodontitis. Using conventional regression, the "extent" of periodontitis was positively associated with higher counts of dentin caries lesions, even after adjustments for maternal education and subgingival microbial composition (rate ratio 1.34; 95% CI 1.07-1.68). The "severity" of periodontitis was associated with lower counts of enamel and dentin caries lesions (rate ratio 0.85 95% CI 0.77-0.92). The SEM revealed a positive association between periodontitis "extent" and number of dentin caries lesions (coefficient 0.29; P < .0001). The "severity" of periodontitis was negatively associated with enamel and dentin caries (coefficient -0.44; P < .0001). We found an association between caries and periodontitis among adolescents. The "severity" of periodontitis was negatively associated with enamel/dentin caries, whereas the "extent" of periodontitis was positively associated with dentin caries irrespective of the analytical approach employed.