Metabolic Risk Signals in Periodontal Clinics: Cross-Sectional Associations of PISA with hs-CRP and HOMA-IR

Background/Objectives: Periodontitis can elevate systemic inflammation and negatively impact glucose regulation. This cross-sectional study investigated relationships between periodontal inflammatory burden, measured by periodontal inflamed surface area (PISA), and insulin resistance and systemic inflammation in never-smokers attending a periodontal clinic with unknown diabetes status. Methods: A total of 154 adults underwent full-mouth periodontal assessments to determine PISA. Fasting blood samples were taken for glucose, insulin, and hs-CRP, and HOMA-IR was calculated. An inflammation composite was created from z-scored log(hs-CRP), IL-6, and TNF-α. Primary analysis involved regressing log(HOMA-IR) and log(hs-CRP) on PISA, scaled by interquartile range (IQR) increases, with sequential adjustments. Mediation analysis, adjusted for age, sex, and waist circumference, used bootstrap testing to evaluate inflammation as a mediator. Results: HOMA-IR and hs-CRP increased across PISA tertiles (HOMA-IR: 2.01 to 3.17; hs-CRP: 0.73 to 2.48 mg/L; both p < 0.001). In the prespecified primary adjusted model (age, sex, waist circumference, education, physical activity, family history of diabetes, and medication use), each IQR increase in log(PISA) was associated with 85.2% higher hs-CRP (95% CI 53.1% to 124.1%; p < 0.001) and 15.7% higher HOMA-IR (95% CI -0.8% to 34.9%; p = 0.064). Estimates were similar with additional adjustment for BMI and SBP, whereas adding plaque (% sites) attenuated associations (hs-CRP: +15.2%; p = 0.234; HOMA-IR: +1.0%; p = 0.925). Mediation analysis (adjusted for age/sex/waist circumference) was consistent with an indirect pathway via the inflammation composite (a × b = 0.212; 95% CI 0.094 to 0.337), while the direct effect was not supported (c' = -0.047; 95% CI -0.207 to 0.125). Conclusions: Higher periodontal inflammatory burden was strongly associated with systemic inflammation and showed weaker, model-dependent associations with insulin resistance; the findings were consistent with an inflammation-linked pathway in exploratory, partially adjusted mediation analyses; given the cross-sectional design, causal inference is not possible.