Identifying Undiagnosed Diabetes and Prediabetes in the Dental Setting in an Asian Population-A Clinical Risk Model.

Abstract

To assess the glycaemic status of Asian patients in a tertiary care dental setting and develop a risk model for undiagnosed diabetes mellitus (DM). A total of 1074 participants completed a diabetes risk test questionnaire, full-mouth periodontal examination and a point-of-care HbA1c finger-prick blood test. Univariable logistic regression was performed to assess the effect of potential factors to predict DM, with confirmed diabetes as the outcome. Subsequently, multivariable logistic regression analysis with stepwise variable selection was employed to develop the final models for predicting DM. Sixty-five (6.1%) and 83 (7.7%) of the 1074 participants were medically confirmed with T2DM and prediabetes, respectively. The 'best' predictive risk model for DM included body mass index (BMI), family history of diabetes, smoking and a diagnosis of Stage III/IV or severe periodontitis with an area under the curve (AUC) of 0.717 (95% confidence interval, CI [0.689-0.744]) and 0.721 (95% CI [0.693-0.748]), respectively. Including the oral health measure marginally increased the AUC. Dental patients clinically diagnosed with advanced periodontitis in combination with high BMI, positive family history of DM and smoking are potentially at high risk for DM and should be screened for DM and referred for medical confirmation and management.