Diagnostic accuracy and smoking-driven modulation of salivary MMP-8, MMP-9, and HIF-1α in stage III grade B/C periodontitis: a cross-sectional study.

OBJECTIVES: Heavy cigarette smoking is a risk factor for Grade C periodontitis according to the 2018 World Workshop classification. This study evaluated the diagnostic accuracy and smoking-driven modulation of salivary HIF-1α, MMP-8, and MMP-9 across four phenotypically defined cohorts: periodontally healthy non-smokers, Stage III Grade B non-smokers, periodontally healthy heavy smokers, and Stage III Grade C heavy smokers. MATERIALS AND METHODS: Eighty systemically healthy adults (n = 20/group) underwent full-mouth periodontal examination, including probing depth (PD), bleeding on probing (BOP%), clinical attachment level (CAL), plaque index (PI), and radiographic bone loss (RBL). Unstimulated whole saliva was analyzed for HIF-1α, MMP-8, and MMP-9 by ELISA. Statistical analyses included Kruskal-Wallis with Dunn-Bonferroni post-hoc, log-transformed GLM adjusted for age, sex, and BMI, and ROC analysis with 2000-iteration bootstrap 95% CIs. A secondary ROC analysis addressed the smoker subgroup (Grade C(S) vs. Healthy(S), n = 40). RESULTS: MMP-8 achieved an AUC of 0.855 (95% CI: 0.756-0.934; sensitivity 82.5%, specificity 85.0%) for distinguishing periodontitis from healthy controls and retained robust accuracy in smokers (AUC = 0.831; sensitivity 75.0%, specificity 90.0%), significantly outperforming MMP-9 (AUC = 0.702; ΔAUC = 0.154, p = 0.005). HIF-1α showed no diagnostic value (AUC = 0.453) but was independently suppressed by smoking (β=-0.394, p = 0.007), consistent with chronic PHD2/PHD3-mediated hypoxia adaptation. Notably, BOP% was significantly higher in Grade B (NS) (44.6 ± 25.3%) than in Grade C (S) (30.9 ± 16.5%), consistent with tobacco-induced vascular suppression of gingival bleeding. The BOP% difference between Grade B (NS) and Grade C (S) did not reach statistical significance after Bonferroni correction (p = 0.593), consistent with the known vasoconstriction-mediated suppression of gingival bleeding in heavy smokers. CONCLUSIONS: Salivary MMP-8 demonstrated clinically meaningful diagnostic accuracy (AUC = 0.855) maintained in heavy smokers despite suppressed bleeding (AUC = 0.831). Paradoxical HIF-1α suppression by smoking, independent of periodontal status, positions this marker as a mechanistic indicator of tobacco-driven hypoxia adaptation rather than a periodontitis diagnostic tool. CLINICAL RELEVANCE: In Grade C periodontitis patients where tobacco suppresses gingival bleeding, salivary MMP-8 provides adjunctive diagnostic value unaffected by this masking effect (smoker subgroup AUC = 0.831), supporting its use as a point-of-care adjunct where conventional bleeding-based indices underestimate inflammatory burden. The HIF-1α suppression phenotype in healthy smokers represents a novel tobacco-host interaction target warranting further mechanistic investigation.