Diabetes‐Associated Periodontitis Molecular Features in Infrared Spectra of Gingival Crevicular Fluid. Issue 12 (1st December 2013)

• Record Type: Journal Article
• Authors: Xiang, Xiaoming; Duarte, Poliana Mendes; Lima, Jadson Almeida; Santos, Vanessa Renata; Gonçalves, Tiago Dias; Miranda, Tamires Szeremeske; Liu, Kan‐Zhi
• Is Part Of: Journal of periodontology. Volume 84:Issue 12(2013)
• Page Start: 1792
• Page End: 1800
• Publication Date: 2013-12-01
• Subjects: Diabetes mellitus -- diagnosis -- gingival crevicular fluid -- periodontitis; Periodontics -- Periodicals; 617.632
• DOI: 10.1902/jop.2013.120665
• Languages: English
• ISSNs: 0022-3492

Abstract

Abstract : Background: It has been established previously that infrared spectroscopy (IRS) can be used to identify periodontitis‐specific molecular signatures in gingival crevicular fluid (GCF) and to confirm clinical diagnoses. This follow‐up study is designed to assess whether this novel technique is also able to differentiate diseased from healthy sites in patients with diabetes mellitus (DM) by analyzing the molecular fingerprints embedded in the GCF. Methods: A total of 65 patients with DM with moderate‐to‐severe chronic periodontitis (CP) was recruited, and 15 individuals without DM (65 sites) without periodontal diseases were used as control. Clinical examination and GCF samples were taken from a total of 351 sites, including periodontitis (109), gingivitis (115), and healthy (127) sites. Corresponding absorption spectra of GCF samples were acquired and processed, and the relative contributions of key functional groups in the infrared spectra were identified and analyzed. The qualitative assessment of clinical relevance of these GCF spectra was interpreted with multivariate statistical analysis: linear discriminant analysis (LDA). Results: Spectral analysis revealed several molecular signatures representing vibrations in protein (amide I and II), lipid ester, and sugar moieties in the GCF of patients with DM with CP and non‐DM controls. The diagnostic accuracy for distinction between healthy and CP sites in patients with DM determined by LDA of GCF spectra was 95.3% for the training set of samples and 87.5% for the validation set. Additional LDA of GCF spectra from healthy sites of non‐DM controls and patients with DM revealed 100% diagnostic accuracy for the training set and 86.7% for the validation set. The regions robotically selected by LDA for the two analyses were slightly different in that first LDA identified major regions clustered with the side chain vibrations originating from protein and DNA contents, whereas the second was predominantly the glycation and protein components. Conclusion: IRS is a feasible method to differentiate disease‐specific molecular signatures in GCF in the presence of DM and to generate a complex biochemical profile of GCF to identify DM‐specific spectral features.