A Mechanism-anchored Framework Linking Local Pharmacokinetics to Health Economics in Periodontitis

Background Locally delivered periodontal adjuncts are usually compared on isolated clinical endpoints, while local pharmacokinetics and economic uncertainty are assessed separately. We developed an auditable proof-of-concept pipeline linking published local-exposure summaries to structural clinical and economic scenarios. Methods Published gingival crevicular fluid (GCF) summaries for doxycycline gel (Atridox), chlorhexidine chip (PerioChip) and minocycline microspheres (Arestin) were represented in a local mini physiologically based pharmacokinetic (PBPK) model. Fits to the same summaries were treated as calibration, not validation. Published probing pocket depth (PPD) effects entered four alternative structural mappings to pocket closure. A 20-year illustrative Markov model and 5000-draw probabilistic sensitivity analysis (PSA) propagated uncertainty. The base case excluded any separate matrix metalloproteinase-8 (MMP-8) quality-adjusted life-year (QALY) gain. Results Calibration absolute average fold errors were 1.00 for Atridox, 1.59 for PerioChip and 1.36 for Arestin; independent validation was not performed. Multistart and residual-Jacobian diagnostics indicated that the Atridox release parameters were not identifiable and that PerioChip precision remained weak with sparse anchors. Across linear, logistic, probit and Emax closure mappings, the deterministic preferred adjunct was Atridox when the chlorhexidine PPD effect was 0.30 mm and PerioChip when it was 0.75 mm. At US$50,000/QALY, probabilities of being optimal were 0.543 for PerioChip, 0.350 for Atridox and 0.107 for Arestin. Illustrative expected value of perfect information was US$419 per treated site (Monte Carlo standard error US$9). Conclusions The framework is reproducible, but its clinical and economic layers remain proof-of-concept simulations. The product ranking is sensitive to comparative PPD evidence, and the current results do not establish clinical or reimbursement superiority. Independent patient-level pharmacokinetic (PK), closure, recurrence and health-system data are required before decision use.