RESUMO
OBJECTIVES: The study investigates whether periodontal parameters can identify subjects with undiagnosed diabetes mellitus (DM) or pre-DM in patients seeking dental treatment at a university dental hospital. MATERIALS AND METHODS: Adults older than 35 years, not being diagnosed with DM before and have at least one of the risk factors of DM were included in the study. All subjects received a complete periodontal examination, filled a medical history survey, and a fasting blood glucose measurement was obtained. A multiple logistic regression test using a backward elimination method to assess factors that predict if the participant is healthy, prediabetic or diabetic was done. A p-value of <0.05 was considered significant. RESULTS: A total of 61 subjects were enrolled with an average age of 42.9 ± 9.4 years. Having a family member diagnosed with DM was reported by 64.5% of the subjects; 59% were diagnosed with advanced periodontal disease. The final logistic regression model included smoking, hypertension, family history of DM, and percentage of clinical attachment loss >3 mm was statistically significant (P < 0.001). The model explained 47.7% of the diabetes condition and correctly classified 69.4% of cases. Participants with a family history of diabetes are 4.98 times more likely to exhibit prediabetic or diabetic status. Each unit increase in the percentage of clinical attachment loss increases the likelihood of participant to be prediabetic or diabetic by 1.104 times. CONCLUSION: Dental patients presenting with severe clinical attachment loss and family history of DM have increased likelihood of undiagnosed DM or pre-DM and would benefit from screening at the dental office.
