Mass spectrometric analysis of gingival crevicular fluid biomarkers can predict periodontal disease progression

Abstract

Background and Objective: Gingival crevicular fluid has been suggested as a possible source of biomarkers for periodontal disease progression. This paper describes a technique for the analysis of gingival crevicular fluid from individual sites using mass spectrometry. It explores the novel use of mass spectrometry to examine the relationship between the relative amounts of proteins and peptides in gingival crevicular fluid and their relationship with clinical indices and periodontal attachment loss in periodontal maintenance patients. The aim of this paper was to assess whether the mass spectrometric analysis of gingival crevicular fluid may allow for the site-specific prediction of periodontal disease progression.

Material and Methods: Forty-one periodontal maintenance subjects were followed over 12 mo, with clinical measurements taken at baseline and every 3 mo thereafter. Gingival crevicular fluid was collected from subjects at each visit and was analysed using matrix-assisted laser desorption/ionization timeof- flight (MALDI-TOF) mass spectrometry. Samples were classified based upon pocket depth, modified gingival index (MGI), plaque index and attachment loss, and were analysed within these groups. A genetic algorithm was used to create a model based on pattern analysis to predict sites undergoing attachment loss.

Results: Three hundred and eighty-five gingival crevicular fluid samples were analysed. Twenty-five sites under observation in 14 patients exhibited attachment loss of > 2 mm over the 12-mo period. The clinical indices pocket depth, MGI, plaque levels and bleeding on probing served as poor discriminators of gingival crevicular fluid mass spectra. Models generated from the gingival crevicular fluid mass spectra could predict attachment loss at a site with a high specificity (97% recognition capability and 67% cross-validation).

Conclusions: Gingival crevicular fluid mass spectra could be used to predict sites with attachment loss. The use of algorithm-generated models based on gingival crevicular fluid mass spectra may provide utility in the diagnosis of periodontal disease.