Efficacy assessment of local doxycycline treatment in periodontal patients using multivariate chemometric approach.

The aim of our study was application of chemometric algorithms for multivariate data analysis in efficacy assessment of the local periodontal treatment with doxycycline (DOX). Treatment efficacy was evaluated by monitoring inflammatory biomarkers in gingival crevicular fluid (GCF) samples and clinical indices before and after the local treatment as well as by determination of DOX concentration in GCF after the local treatment. The experimental values from these determinations were submitted to several chemometric algorithms: principal component analysis (PCA), partial least square discriminant analysis (PLS-DA) and orthogonal projection to latent structures-discriminant analysis (OPLS-DA). The data structure and the mutual relations of the selected variables were thoroughly investigated by PCA. The PLS-DA model identified variables responsible for discrimination of classes of data, before and after DOX treatment. The OPLS-DA model compared the efficacy of the two commonly used medications in periodontal treatment, chlorhexidine (CHX) and DOX, at the same time providing insight in their mechanism of action. The obtained results indicate that application of multivariate chemometric algorithms can be used as a valuable approach for assessment of treatment efficacy.