Skip to main content
Browse Subject Areas

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Global Metabolomic Analysis of Human Saliva and Plasma from Healthy and Diabetic Subjects, with and without Periodontal Disease

  • Virginia M. Barnes,

    Affiliation Colgate Palmolive Technology Center, Piscataway, NJ, United States of America

  • Adam D. Kennedy,

    Affiliation Department of Oral Biology, School of Dental Medicine, University at Buffalo, State University of New York, Buffalo, NY, United States of America

  • Fotinos Panagakos,

    Affiliation Colgate Palmolive Technology Center, Piscataway, NJ, United States of America

  • William Devizio,

    Affiliation Colgate Palmolive Technology Center, Piscataway, NJ, United States of America

  • Harsh M. Trivedi,

    Affiliation Colgate Palmolive Technology Center, Piscataway, NJ, United States of America

  • Thomas Jönsson,

    Affiliation Department of Oral Biology, School of Dental Medicine, University at Buffalo, State University of New York, Buffalo, NY, United States of America

  • Lining Guo,

    Affiliation Department of Oral Biology, School of Dental Medicine, University at Buffalo, State University of New York, Buffalo, NY, United States of America

  • Shannon Cervi,

    Affiliation Metabolon, Durham, NC, United States of America

  • Frank A. Scannapieco

    Affiliation Metabolon, Durham, NC, United States of America


20 Nov 2014: The PLOS ONE Staff (2014) Correction: Global Metabolomic Analysis of Human Saliva and Plasma from Healthy and Diabetic Subjects, with and without Periodontal Disease. PLOS ONE 9(11): e114091. View correction


Recent studies suggest that periodontal disease and type 2 diabetes mellitus are bi-directionally associated. Identification of a molecular signature for periodontitis using unbiased metabolic profiling could allow identification of biomarkers to assist in the diagnosis and monitoring of both diabetes and periodontal disease. This cross-sectional study identified plasma and salivary metabolic products associated with periodontitis and/or diabetes in order to discover biomarkers that may differentiate or demonstrate an interaction of these diseases. Saliva and plasma samples were analyzed from 161 diabetic and non-diabetic human subjects with a healthy periodontium, gingivitis and periodontitis. Metabolite profiling was performed using Metabolon's platform technology. A total of 772 metabolites were found in plasma and 475 in saliva. Diabetics had significantly higher levels of glucose and α-hydroxybutyrate, the established markers of diabetes, for all periodontal groups of subjects. Comparison of healthy, gingivitis and periodontitis saliva samples within the non-diabetic group confirmed findings from previous studies that included increased levels of markers of cellular energetic stress, increased purine degradation and glutathione metabolism through increased levels of oxidized glutathione and cysteine-glutathione disulfide, markers of oxidative stress, including increased purine degradation metabolites (e.g. guanosine and inosine), increased amino acid levels suggesting protein degradation, and increased ω-3 (docosapentaenoate) and ω-6 fatty acid (linoleate and arachidonate) signatures. Differences in saliva between diabetic and non-diabetic cohorts showed altered signatures of carbohydrate, lipid and oxidative stress exist in the diabetic samples. Global untargeted metabolic profiling of human saliva in diabetics replicated the metabolite signature of periodontal disease progression in non-diabetic patients and revealed unique metabolic signatures associated with periodontal disease in diabetics. The metabolites identified in this study that discriminated the periodontal groups may be useful for developing diagnostics and therapeutics tailored to the diabetic population.


Periodontal disease is a chronic bacterial infection causing persistent gingival inflammation, and in some cases connective tissue destruction and bone resorption around the teeth. It is also characterized by pocket formation and recession. Although advances in dental care has resulted in improved periodontal status in certain populations, disparities persist as severe periodontitis is often found to be exaggerated in certain segments of the population, for example those from a low socio-economic background [1]. It is well known that bacteria colonize the teeth to form a biofilm, called dental plaque, which initiates gingivitis, and sometimes progresses to periodontitis. Release of bacterial products from the biofilm induces local inflammation. Without treatment, periodontal tissue destruction, bone resorption and tooth loss may ensue [1]. Periodontal disease also has been associated with several systemic diseases, including cardiovascular disease, diabetes mellitus, respiratory disease, rheumatoid arthritis, chronic kidney disease, and adverse pregnancy outcomes [2][6].

The gingival epithelium forms a crevice around each tooth that provides a protected space for bacterial colonization and proliferation [7][10]. Analysis of gingival crevicular fluid (GCF) [11] and saliva [12][14] from periodontal patients has identified a variety of inflammatory mediators and tissue-destructive molecules, including metalloproteinases [15][23] and metabolic signatures associated with host-bacterial interactions to be elevated when compared to periodontally health patients.

Diabetic patients have a high prevalence of gingivitis, periodontitis, oral candidiasis, and xerostomia, and the severity of these diseases are correlated with the duration of diabetes and degree of glycemic control [24], [25]. Poor glycemic control has been shown to be associated with poor periodontal health, of which the molecular signatures may be monitored using modern “omics”-based methods [26][30].

Saliva is a complex secretory fluid that contains trace metals, metabolites, biochemicals, proteins, glycoproteins, lipds, etc., that serve a spectrum of physiological needs. Saliva is a critical source of tissue lubricants, tooth mineralizing factors, acid buffers, toxin neutralizers, and antimicrobial components [31][35]. The ability to use saliva to evaluate physiological conditions, follow disease progression, and monitor post-treatment therapeutic results through noninvasive methods is an important objective for healthcare in general and periodontology in particular. We have previously performed a series of untargeted global metabolomic profiling tests of GCF and saliva samples from subjects with healthy gingiva, gingivitis, and periodontitis that have suggested a rigorous set of potential biomarkers for monitoring periodontal disease status and examining the effectiveness of oral care treatment that resolves the metabolic signature of inflammation [11], [12], [36], [37]. Many metabolites associated with inflammation, oxidative stress, tissue degradation, and bacterial metabolism were found to be significantly elevated in periodontal disease and reduced by treatment [37]. Validation of such biomarkers will provide an objective phenotype to allow practitioners to diagnose disease, monitor patient disease activity and determine the effectiveness of treatment.

Metabolomic profiling is a rapidly evolving technology that has been increasingly used to discover early markers of disease [11], [12], [37][40]. Whole saliva can be easily collected through noninvasive means and has considerable potential to monitor general health and disease status. With the advancement of technological means such as metabolomics, saliva can be leveraged as a clinical and diagnostic tool because of its potential to mirror oral and systemic health conditions [41][45]. To this end, metabolomic signatures were determined in saliva and compared to those in plasma from orally healthy subjects and periodontal subjects, with and without type 2 diabetes, in an effort to identify unique metabolomic profiles and biomarkers of periodontal disease in both systemic health and diabetes. Based on our previous results from biomarker identification of periodontal disease progression [11], [12], [37], we performed a discovery study to identify biomarkers of decreased periodontal health as monitored by biochemical profiling of saliva and plasma collected from orally healthy subjects and those with gingivitis and periodontitis using an unbiased metabolomic profiling approach based on liquid and gas chromatography/mass spectrometry (LC/MS and GC/MS).

Materials, Participants, and Methods

Experimental Design and Participants

This cross-sectional clinical study enrolled adult male and female subjects from the Buffalo, New York area and was conducted in accordance with the Helsinki Declaration. The Institutional Review Board at the University at Buffalo approved the protocol (ORB0650111E), including the clinical examination, collection of saliva and plasma samples. All subjects read and signed a written informed consent before their enrollment into the study. The inclusion criteria were: age between 18 and 65 years, in good general health (n = 81), or subjects diagnosed with diabetes (n = 80); and a minimum of 20 natural teeth (excluding third molars) present. Exclusion criteria included: inability or unwillingness to sign the informed consent form, diagnosis of a medical condition which required pre-medication prior to dental visits/procedures, 5 or more decayed untreated teeth at screening (cavities), diagnosis of other diseases of the hard or soft oral tissues, impaired salivary function, use of antibiotics or antimicrobial drugs within 30 days prior to the first study visit, a history of uncharacterized systemic disease, pregnant or nursing women, participation in any other clinical study within 1 week prior to enrollment into this study, present use of tobacco products, subjects who required dental treatment during the duration of the study, immune compromised individuals (e.g. those diagnosed with HIV, AIDS, or taking immunosuppressive drug therapy), or subjects who received periodontal treatment within the previous 30 days.

Subjects were evaluated by a single dental examiner (SC) for inclusion/exclusion criteria. At the first visit, participants received a tube of standard, non-anti-microbial fluoridated toothpaste with instructions for use, in order to minimize variations in results that might be driven by the use of different toothpastes. A fasting blood sample was then collected from subjects self-identifying as diabetic to confirm diabetic status. During the second visit, subjects gave an update on their medical and medication history, contributed saliva and blood samples, had an assessment of their oral tissues to assess periodontal status, to monitor adverse events since the first visit and answer questions regarding their compliance as to the use of the standard toothpaste. A minimum of 0.5 ml of unstimulated saliva and 10 ml of blood was collected. Subjects were asked to refrain from eating or drinking (excluding water) from 11:00 PM the previous night; to brush their teeth and entire mouth the evening before the saliva donation, but not the morning of collection. Blood samples were obtained after collection of the saliva. Saliva was collected into sterile polypropylene tubes and immediately frozen in a dry ice bath and stored at −80 degrees C until shipped to Metabolon. No preservatives were added to the saliva samples

Definition of oral status groups

Within the diabetic or non-diabetic cohorts, three sub-groups were established: periodontal health, gingivitis and periodontitis. Subjects were stratified according to the gingivitis index: Healthy: These subjects had an average full mouth Modified Gingival Index (MGI) [46] score below 1.0, with fewer than 3 bleeding on probing (BOP) sites and minimal dental plaque present as scored by the method of Sillness and Lőe [47]. Gingivitis: These subjects had an average full mouth MGI score between 1 and 2 with multiple BOP sites and periodontal pockets <4 mm. Periodontitis: These subjects had an average full mouth MGI of 2.0 or greater with multiple BOP sites, and 2 or more periodontal pockets with probing depths of 5 mm or more in at least two quadrants. Plaque at any level was present. Measurement of oral indices was performed as per Lobene et al. [46] as follows: 0 = normal, 1 = mild inflammation of the gingival unit (slight change in color, little change in texture of any portion of the gingival unit), 2 = mild inflammation of the entire gingival unit, 3 = moderate inflammation (moderate change in texture, redness, edema and hypertrophy of the gingival unit, 4 = severe inflammation (marked redness and edema/hypertrophy, spontaneous bleeding or ulceration) of the gingival unit. Plaque indices were assessed for 6 surfaces on each tooth and the plaque index score was the average of the scores obtained from all teeth. The amount of plaque observed on each tooth was scored 0 to 3 as follows: 0 = no plaque notes; 1 = plaque seen only on the tip of an explorer passed over the tooth surface; 2 = plaque obvious with the naked eye; 3 = gross deposits of plaque present over the entire tooth.

Systemic Health Selection Criteria/Assessment for Diabetes Status

For this study, diabetes status was established by self-reported history. Subjects who reported a history of diabetes were asked to confirm diabetic status by providing blood samples to measure HbA1c values. An individual was defined with Type 2 diabetes if hemoglobin A1C≥6.5%. Due to financial limitations, HbA1c levels were not assessed for those not reporting a history of diabetes. As a recent study suggested, the prevalence of diabetes (as measured by HbA1c>6.4) is likely low (extrapolated to be 5.0% with diabetes and 9.5% with pre-diabetes) in dental patients unaware of their diabetes status [48], we assumed that the prevalence of diabetes in the non-diabetes subject in this study would be similar and therefore not compromise the analysis.

Saliva and Plasma Sample Collection

Saliva was collected in 50 mL polypropylene tubes and plasma was collected in Vacutainer tubes and anti-coagulated with EDTA. Samples were immediately frozen in a dry ice bath and stored at −80°C.

Metabolomic Profiling Technology

Metabolomic profiling was performed as described previously [12], [49]. In summary, after the extraction of metabolites from each sample, the extracts were analyzed by GC/MS and LC/MS. We carried out chromatographic separation, followed by full-scan mass spectroscopy, to record and quantify all detectable ions presented in the samples [50]. We identified metabolites with known chemical structure by matching the ions' chromatographic retention index and mass spectral fragmentation signatures with reference library entries created from authentic standard metabolites under the identical analytical procedure as the experimental samples [51]. For ions that were not covered by the standards, additional library entries were added based on their unique ion signatures (chromatographic and mass spectral). After this, these ions could be routinely detected and quantified.

Statistical Analysis

After normalization and imputation, the data were log-transformed. We then performed ANOVA and t tests to compare data obtained from the healthy, gingivitis, and periodontitis samples. Multiple comparisons were accounted for with the false discovery (FDR) rate method, and each FDR was estimated by q-values.


A total of 174 subjects were enrolled in this study (Table 1). The mean age of the non-diabetic subjects tended to be younger than the diabetic subjects (Table 2). Diabetic subjects had a higher BMI for all groups (Table 2). HbA1C testing confirmed diabetic status of the subjects self-reporting as diabetic.

An untargeted metabolomic profiling approach was used to assess the chemical milieu of saliva and plasma. In total, 161 individual saliva samples were collected, 80 from subjects with diabetes and 81 from subjects without diabetes. We detected 475 and 772 metabolites in saliva and plasma, respectively, of which 370 and 430 could be mapped to known chemical structures (Tables S1 and S2).

Characterization of Diabetic and Non-Diabetic Samples

Welch's two sample t tests and ANOVA identified statistically significant differences in these saliva samples between non-diabetic subjects and those with diabetes. Approximately 14% of the detected metabolites (69 of 475) in saliva showed altered levels (p≤0.05) among the healthy samples when compared to the diabetic cohort, suggesting that these may be markers of diabetic status in humans (Table 3). Alternatively, comparison of the plasma from the same subjects resulted in 174 statistically significant changes. ANOVA analysis with either matrix did not produce a list of metabolites largely different from the t tests (data not shown).

Table 3. Overall statistical analysis of saliva and plasma samples from the healthy periodontal cohorts.

Glycemic control can be monitored by changes in 1,5-anhydroglucitol (1,5-AG). 1,5-AG is reabsorbed in the renal tubules, a process competitively inhibited by glucosuria, which leads to a reduction in its level in plasma. Thus, the greater the degree of glucosuria, the lower the amount of 1,5-AG present in the plasma [52][54]. The correlation between this reduction and the amount of glucose present in urine is so close that 1,5-AG represents a sensitive, real-time marker of glycemic control [55]. Thus, elevations of 1,5-AG in plasma are highly indicative of reduced glucosuria as a consequence of improved glycemic control. Not surprisingly, subjects with diabetes had significantly lower 1,5-AG and significantly higher levels of glucose in their plasma (Table 4). Additionally, diabetic subjects had significantly lower 1,5-AG in their saliva for all periodontal disease groups and higher levels of glucose in their saliva (Table 4). These observations lend confidence to the appropriateness of the control group, which was based on self-report of diabetes. None of the non-diabetic subjects had metformin in their plasma or saliva (data not shown).

Table 4. Relative levels of biochemicals linked to diabetic status in saliva and plasma*.

α-hydroxybutyrate (AHB) is an early marker for insulin resistance in humans [52] as it arises from states of lactic and ketoacidosis or disrupted energy metabolism [56][61]. AHB levels were increased significantly in the diabetic cohort in plasma (Table 4) and were increased significantly in saliva in the healthy and gingivitis cohorts (data not shown). This may be a reflection of the increased energetic stress induced by the infection during the development of periodontitis or a translation of insulin resistance markers to saliva.

High-Level Characterization of Diabetic and Non-Diabetic Periodontal Cohorts

The analysis of each periodontal cohort (diabetic and non-diabetic) separately resulted in the identification of several metabolites that were associated with the progression of gingival status (Table 5). The most striking find in the study was the direction of change in the periodontitis samples in the diabetic cohort. In the non-diabetic cohort, metabolites increased in abundance through disease progression, whereas many biochemicals increased in abundance when comparing the healthy to gingivitis groups, but then decreased in abundance in the diabetic subjects with periodontal disease suggesting an interaction between the diabetes signature and the periodontal signature.

Table 5. Overall statistical analysis of saliva samples comparing periodontal cohorts from diabetics and non-diabetics.

Both periodontal status and diabetes status of the subjects were major driving factors in the biochemical changes observed. For the majority of metabolites with altered concentrations in the non-diabetic cohort, the levels at gingivitis samples were found to be intermediate between those levels found in healthy and periodontitis samples (Table S1). This suggests that the metabolic changes induced by gingivitis are a continuum of those of periodontitis.

Analysis of the Non-Diabetic Saliva Samples

Barnes et al. found differences in the biochemical signatures in saliva between non-diabetic subjects with healthy gingiva or periodontitis [12]. This report showed increased levels of metabolites in the saliva of the periodontal population reflecting enhanced degradation of macromolecules, including proteins, lipids, nucleotides and polysaccharides. Comparison of the non-diabetic cohort from the current study with these previous studies showed similar results (Table 6). This metabolic signature, which includes elevated levels of ω-6 fatty acids, is indicative of a hyper-inflammatory environment resulting from chronic bacterial infection and further support for increased macromolecular degradation of the periodontal tissues resulting from inflammation. Several metabolic signatures supported the notion that a more diverse and active microbial flora accompanied the inflammatory response resulting in gingivitis and periodontitis. Aromatic amino acid metabolites likely derived from bacteria were increased in periodontitis samples (Table 6). Moreover, carnitine, which is often the sole source of carbon and nitrogen for bacteria, was also elevated. Periodontal saliva samples also contained elevated levels of 3-dehydrocarnitine, a bacterial degradation product of carnitine, further supporting the utilization of carnitine in oral bacterial metabolism.

Table 6. Comparison of biochemical changes to previously published results.

Comparison of Biochemical Signatures in Saliva Samples from Diabetics vs Non-Diabetic subjects

Oxidative Stress and Anti-Oxidant Capacity.

Purine degradation is regulated, in part, by oxidative stress [62], and increased purine degradation serves as an indication of an increased inflammatory response. Specifically, the conversion of hypoxanthine to xanthine and the subsequent conversion of xanthine to uric acid require O2 to activate xanthine oxidase [63], [64] and these reactions are coupled with a reduction of oxygen to generate superoxides in the form of O2- and H2O2. Urate is known to function in redox balance as a radical scavenger or antioxidant under conditions of oxidative stress and in the process converted to other products (e.g. allantoin). In the context of gingival disease, increased oxidative stress would result in higher levels of purine degradation metabolites, reflected in increased levels of hypoxanthine and xanthine. The results of this study are consistent with the results from previous studies [11], [37]. Metabolites in the purine degradation pathway were statistically elevated in the diabetic cohort compared to the non-diabetic cohort (Table 7).

Table 7. Relative levels of biochemicals linked to purine degradation and anti-oxidant status in saliva samples from diabetics.

The levels of cysteine-glutathione disulfide and oxidized glutathione increased with gingival disease in non-diabetics and diabetics (Table 7). Glutathione plays a central role in the cellular defense against free radicals and xenobiotics. The increased levels of these oxidized glutathione species in saliva indicated increased oxidative stress.

Lipid and Sphingolipid Metabolism.

Several lipid inflammatory mediators increased significantly in saliva from both the non-diabetic and diabetic groups during periodontal disease progression (Table 8). In this state of oxidative stress, lipoxygenase enzymes oxidize fatty acids to generate lipid signaling molecules such as 12-HETE. In saliva, arachidonate and 12-HETE increased in the non-diabetic cohort with periodontal disease and also increased in the diabetic gingivitis cohort compared to the healthy group, but was reduced in periodontitis (Table 8). Sphingomyelin, ceramides and glycosphingolipids are molecules that reside in the plasma membrane and participate in intracellular signaling cascades. Palmitoyl sphingomyelin was elevated in saliva of diabetic and non-diabetic subjects periodontitis (Table 8).

Table 8. Fatty acids and sphingomyelins inreased in saliva from subjects with diabetes and periodontal disease.


We used metabolomic profiling technology to analyze saliva from subjects with and without diabetes who had a healthy gingiva, gingivitis, or periodontitis. In addition to verifying previously published observations in saliva [12] and GCF [11], [37], we also identified in the present study salivary biochemical changes associated with diabetes, as well as changes associated with periodontal disease. The results from this analysis show that human diabetic subjects with periodontal disease have an increased purine degradation signature, a decreased redox balance capacity (Table 7) and altered ω-3/ω-6 fatty acid profiles (Table 8) in saliva.

Redox balance is vital for maintaining cellular homeostasis. Oxidative stress, caused by over production of reactive oxygen species and insufficient cellular anti-oxidant capabilities, can lead to cell damage. Glutathione and other anti-oxidant molecules serve to protect cellular proteins and structures from oxidative stress caused by reactive oxygen species. Glutathione is a thiol-containing tripeptide that plays a central role in cellular defense against free radicals and xenobiotics [65][68] and is the major intracellular redox mediator and it is pivotal for controlling the electrical or redox gradient across the mitochondrial inner-membrane [69], [70]. Glutathione in the mitochondrial intermembrane space plays a intricate role in the mechanism of redox balance in living cells through its linkage to the redox milieu in the cytosol [71]. Oral interventions that enhance redox balance have been shown to be effective at relieving oral symptoms. For example, Hansen and Grunnet showed that taurine plays a significant role in mitochondria in pH buffering that not only affects the status of glutathione, but also the NADH/NAD+ redox pair [72]. These molecules were all found to be elevated in saliva from the diabetic subjects.

Several lipid compounds are important for the initiation and propagation of inflammatory signaling in cells and tissues. Linoleic acid and arachidonic acid give rise to many of these metabolites. These lipid signaling inflammatory mediators are liberated from the plasma membrane phospholipids by phospholipase A2 enzymes ultimately resulting in the production of linoleic and arachidonic acid-derived eicosanoids which help to mediate the inflammatory response [73]. Arachidonic acid release can also be initiated by collagen, thrombin, bradykinin, serotonin, and adrenaline, which increase in tissues during tissue damage and the early stages in inflammation and oxidative stress. Immune cells such as macrophages and neutrophils produce reactive oxygen species and reactive nitrogen species during cell activation that have the potential to damage lipids, proteins, and other molecules. Interestingly, these compounds were found to be elevated in the saliva of diabetic subjects with gingivitis.

Mobilization of lipid signaling pathways results in the synthesis of ceramides or the liberation of ceramides from intracellular stores. The activity of ceramidases results in the formation of sphingosine and sphingomyelin molecules that function in an array of intracellular signaling or maintenance of cell membrane architecture including apoptosis [74], [75]. Depending on the specific cellular type and mechanism of activation, sphingomyelins and ceramides can serve activation or inhibitory purposes [76]. The formation of ceramides and sphingomyelins has been linked to cell death through nutrient starvation [74][76] as well as an important signaling molecule for cell differentiation and polarity [77]. Porphyromonas gingivalis is known to induce and alter sphingomyelin and ceramide signaling during infection [78].

Prediction of risk and accurate diagnosis of current disease activity may facilitate effective prevention and treatment of gingival diseases. GCF and saliva have been extensively examined in attempts to assess the oral disease status. The markers identified in this study may be leveraged to monitor the effectiveness of interventions as well as identify the mechanisms of action for these interventions. Several different approaches have confirmed that identifying a single marker characteristic of a disease is unlikely; rather, a combination of biomarkers would likely constitute an effective clinical test. For example, insulin resistance is a feature if the pre-diabetes stage in humans prior to the development of diabetes. Untargeted biochemical profiling to identify biomarkers of insulin resistance following the development of targeted assays and validation of the targeted assays in a clinical cohort resulted in the development of a diagnostic that monitors insulin sensitivity status in humans [52], [79]. By monitoring selected metabolites representing the diverse biochemical pathways, it is therefore possible to identify unique biomarkers and to monitor during the course of periodontal treatment in diabetics. In conclusion, this study demonstrated that biochemical profiling technology is a powerful tool for dental research and the results provide new mechanistic insights and treatment strategies that could be specific for oral health of diabetics. Additionally, the biochemicals identified in this study that segregated the individuals by periodontal status could be leveraged or developed into a rapid diagnostic to monitor periodontal disease activity.

Supporting Information

Table S1.

Saliva Metabolic Profiling and Pathway Analysis. Statistical analyses for each biochemical detected in saliva between each of the gingival and diabetic cohorts.


Table S2.

Plasma Metabolic Profiling and Pathway Analysis. Statistical analyses for each biochemical detected in plasma between each of the gingival and diabetic cohorts.


Author Contributions

Conceived and designed the experiments: VMB FP WD HMT TJ LG FAS. Performed the experiments: TJ LG ADK SC FAS. Analyzed the data: VMB LG ADK FAS. Contributed reagents/materials/analysis tools: LG ADK. Contributed to the writing of the manuscript: ADK FAS VMB FP LG.


  1. 1. Leyden JJ, Thiboutot D, Shalita A (2004) Photographic review of results from a clinical study comparing benzoyl peroxide 5%/clindamycin 1% topical gel with vehicle in the treatment of rosacea. Cutis 73: 11–17.
  2. 2. Gutman DA, Cooper LA, Hwang SN, Holder CA, Gao J, et al. (2013) MR imaging predictors of molecular profile and survival: multi-institutional study of the TCGA glioblastoma data set. Radiology 267: 560–569.
  3. 3. Demmer RT, Desvarieux M (2006) Periodontal infections and cardiovascular disease: the heart of the matter. J Am Dent Assoc 137 Suppl: 14S–20S; quiz 38S
  4. 4. Tonetti MS, D'Aiuto F, Nibali L, Donald A, Storry C, et al. (2007) Treatment of periodontitis and endothelial function. N Engl J Med 356: 911–920.
  5. 5. Pussinen PJ, Tuomisto K, Jousilahti P, Havulinna AS, Sundvall J, et al. (2007) Endotoxemia, immune response to periodontal pathogens, and systemic inflammation associate with incident cardiovascular disease events. Arterioscler Thromb Vasc Biol 27: 1433–1439.
  6. 6. Pussinen PJ, Paju S, Mantyla P, Sorsa T (2007) Serum microbial- and host-derived markers of periodontal diseases: a review. Curr Med Chem 14: 2402–2412.
  7. 7. Nieman DC, Shanely RA, Gillitt ND, Pappan KL, Lila MA (2013) Serum metabolic signatures induced by a three-day intensified exercise period persist after 14 h of recovery in runners. J Proteome Res 12: 4577–4584.
  8. 8. Scannapieco FA (1994) Saliva-bacterium interactions in oral microbial ecology. Crit Rev Oral Biol Med 5: 203–248.
  9. 9. Scannapieco FA, Bergey EJ, Reddy MS, Levine MJ (1989) Characterization of salivary alpha-amylase binding to Streptococcus sanguis. Infect Immun 57: 2853–2863.
  10. 10. Scannapieco FA, Solomon L, Wadenya RO (1994) Emergence in human dental plaque and host distribution of amylase-binding streptococci. J Dent Res 73: 1627–1635.
  11. 11. Barnes VM, Teles R, Trivedi HM, Devizio W, Xu T, et al. (2009) Acceleration of purine degradation by periodontal diseases. J Dent Res 88: 851–855.
  12. 12. Barnes VM, Ciancio SG, Shibly O, Xu T, Devizio W, et al. (2011) Metabolomics reveals elevated macromolecular degradation in periodontal disease. J Dent Res 90: 1293–1297.
  13. 13. Christodoulides N, Floriano PN, Miller CS, Ebersole JL, Mohanty S, et al. (2007) Lab-on-a-chip methods for point-of-care measurements of salivary biomarkers of periodontitis. Ann N Y Acad Sci 1098: 411–428.
  14. 14. Miller CS, King CP Jr, Langub MC, Kryscio RJ, Thomas MV (2006) Salivary biomarkers of existing periodontal disease: a cross-sectional study. J Am Dent Assoc 137: 322–329.
  15. 15. Akalin FA, Baltacioglu E, Alver A, Karabulut E (2007) Lipid peroxidation levels and total oxidant status in serum, saliva and gingival crevicular fluid in patients with chronic periodontitis. J Clin Periodontol 34: 558–565.
  16. 16. Akalin FA, Isiksal E, Baltacioglu E, Renda N, Karabulut E (2008) Superoxide dismutase activity in gingiva in type-2 diabetes mellitus patients with chronic periodontitis. Arch Oral Biol 53: 44–52.
  17. 17. Pradeep AR, Kumar MS, Ramachandraprasad MV, Shikha C (2007) Gingival crevicular fluid levels of neopterin in healthy subjects and in patients with different periodontal diseases. J Periodontol 78: 1962–1967.
  18. 18. Pradeep AR, Manjunath SG, Swati PP, Shikha C, Sujatha PB (2007) Gingival crevicular fluid levels of leukotriene B4 in periodontal health and disease. J Periodontol 78: 2325–2330.
  19. 19. Waddington RJ, Embery G, Samuels RH (1994) Characterization of proteoglycan metabolites in human gingival crevicular fluid during orthodontic tooth movement. Arch Oral Biol 39: 361–368.
  20. 20. Embery G, Waddington R (1994) Gingival crevicular fluid: biomarkers of periodontal tissue activity. Adv Dent Res 8: 329–336.
  21. 21. Alpagot T, Bell C, Lundergan W, Chambers DW, Rudin R (2001) Longitudinal evaluation of GCF MMP-3 and TIMP-1 levels as prognostic factors for progression of periodontitis. J Clin Periodontol 28: 353–359.
  22. 22. Alpagot T, Silverman S, Lundergan W, Bell C, Chambers DW (2001) Crevicular fluid elastase levels in relation to periodontitis and metabolic control of diabetes. J Periodontal Res 36: 169–174.
  23. 23. Kardesler L, Buduneli N, Cetinkalp S, Lappin D, Kinane DF (2011) Gingival crevicular fluid IL-6, tPA, PAI-2, albumin levels following initial periodontal treatment in chronic periodontitis patients with or without type 2 diabetes. Inflamm Res 60: 143–151.
  24. 24. Acker-Hewitt TL, Shafer BM, Saunders MJ, Goh Q, Luden ND (2012) Independent and combined effects of carbohydrate and caffeine ingestion on aerobic cycling performance in the fed state. Appl Physiol Nutr Metab 37: 276–283.
  25. 25. Garcia-German D, Viloria F, Menendez P, Canillas F (2012) [Letter in reference to the article "surgery on the wrong side: implication for the patient and the professional. Experience in a major ambulatory surgery of the foot unit"]. Rev Esp Cir Ortop Traumatol 56 : 421; author reply 422.
  26. 26. Keles GC, Cetinkaya BO, Eroglu C, Simsek SB, Kahraman H (2010) Vascular endothelial growth factor expression levels of gingiva in gingivitis and periodontitis patients with/without diabetes mellitus. Inflamm Res 59: 543–549.
  27. 27. Salvi GE, Franco LM, Braun TM, Lee A, Rutger Persson G, et al. (2010) Pro-inflammatory biomarkers during experimental gingivitis in patients with type 1 diabetes mellitus: a proof-of-concept study. J Clin Periodontol 37: 9–16.
  28. 28. Sima C, Rhourida K, Van Dyke TE, Gyurko R (2010) Type 1 diabetes predisposes to enhanced gingival leukocyte margination and macromolecule extravasation in vivo. J Periodontal Res 45: 748–756.
  29. 29. Xiang X, Duarte PM, Lima JA, Santos VR, Goncalves TD, et al. (2013) Diabetes-Associated Periodontitis Molecular Features in Infrared Spectra of Gingival Crevicular Fluid. J Periodontol.
  30. 30. Yoon AJ, Cheng B, Philipone E, Turner R, Lamster IB (2012) Inflammatory biomarkers in saliva: assessing the strength of association of diabetes mellitus and periodontal status with the oral inflammatory burden. J Clin Periodontol 39: 434–440.
  31. 31. Mandel ID (1989) The role of saliva in maintaining oral homeostasis. J Am Dent Assoc 119: 298–304.
  32. 32. Stuchell RN, Blitzer A, Mandel ID (1984) Medical management of non-neoplastic salivary gland disease. Otolaryngol Clin North Am 17: 697–703.
  33. 33. Tabak LA, Levine MJ, Mandel ID, Ellison SA (1982) Role of salivary mucins in the protection of the oral cavity. J Oral Pathol 11: 1–17.
  34. 34. Dunn W, Griffith JW, Morrison MT, Tanquary J, Sabata D, et al. (2013) Somatosensation assessment using the NIH Toolbox. Neurology 80: S41–44.
  35. 35. Hargreaves CE, Grasso M, Hampe CS, Stenkova A, Atkinson S, et al. (2013) Yersinia enterocolitica provides the link between thyroid-stimulating antibodies and their germline counterparts in Graves' disease. J Immunol 190: 5373–5381.
  36. 36. Barnes VM, Arvanitidou E, Szewczyk G, Richter R, DeVizio W, et al. (2011) Evaluation of the antiplaque efficacy of two cetylpyridinium chloride-containing mouthwashes. J Clin Dent 22: 200–203.
  37. 37. Barnes VM, Teles R, Trivedi HM, Devizio W, Xu T, et al. (2010) Assessment of the effects of dentifrice on periodontal disease biomarkers in gingival crevicular fluid. J Periodontal 81: 1273–1279.
  38. 38. Ganti S, Weiss RH (2011) Urine metabolomics for kidney cancer detection and biomarker discovery. Urol Oncol 29: 551–557.
  39. 39. Lobene RR, Kashket S, Soparkar PM, Shloss J, Sabine ZM (1979) The effect of cetylpridinium chloride on human plaque bacteria and gingivitis. Pharmacol Ther Dent 4: 33–47.
  40. 40. Fredericks MA, Lobene RR, Mundy P (1971) A model for teaching social concepts to dental auxiliaries. J Dent Educ 35: 232–235.
  41. 41. Bessonneau V, Bojko B, Pawliszyn J (2013) Analysis of human saliva metabolome by direct immersion solid-phase microextraction LC and benchtop orbitrap MS. Bioanalysis 5: 783–792.
  42. 42. Zhang A, Sun H, Wang X (2012) Saliva metabolomics opens door to biomarker discovery, disease diagnosis, and treatment. Appl Biochem Biotechnol 168: 1718–1727.
  43. 43. Al-Tarawneh SK, Border MB, Dibble CF, Bencharit S (2011) Defining salivary biomarkers using mass spectrometry-based proteomics: a systematic review. OMICS 15: 353–361.
  44. 44. Rathnayake N, Akerman S, Klinge B, Lundegren N, Jansson H, et al. (2013) Salivary biomarkers for detection of systemic diseases. PLoS ONE 8: e61356.
  45. 45. Rathnayake N, Akerman S, Klinge B, Lundegren N, Jansson H, et al. (2013) Salivary biomarkers of oral health: a cross-sectional study. J Clin Periodontol 40: 140–147.
  46. 46. Lobene RR, Weatherford T, Ross NM, Lamm RA, Menaker L (1986) A modified gingival index for use in clinical trials. Clin Prev Dent 8: 3–6.
  47. 47. Silness J, Loe H (1964) Periodontal Disease in Pregnancy. Ii. Correlation between Oral Hygiene and Periodontal Condtion. Acta Odontol Scand 22: 121–135.
  48. 48. Genco RJ, Schifferle RE, Dunford RG, Falkner KL, Hsu WC, et al. (2014) Screening for diabetes mellitus in dental practices: A field trial. J Am Dent Assoc 145: 57–64.
  49. 49. Judice PB, Matias CN, Santos DA, Magalhaes JP, Hamilton MT, et al. (2013) Caffeine intake, short bouts of physical activity, and energy expenditure: a double-blind randomized crossover trial. PLoS ONE 8: e68936.
  50. 50. Evans A, DeHaven CD, Barrett T, Mitchell M, Milgram E (2009) Integrated, nontargeted ultrahigh performance liquid chromatography/electrospray ionization tandem mass spectrometry platform for the identification and relative quantificationof the small-molecule complement of biological systems. Anal Chem 81: 6656–6667.
  51. 51. DeHaven CD, Evans AM, Dai H, Lawton KA (2010) Organization of GC/MS and LC/MS metabolomics data into chemical libraries. J Cheminformatics 2: 9–21.
  52. 52. Gall WE, Beebe K, Lawton KA, Adam KP, Mitchell MW, et al. (2010) Alpha-hydroxybutyrate is an early biomarker of insulin resistance and glucose tolerance in a nondiabetic population. PLoS ONE 5: 1–10.
  53. 53. Nerby CL, FStickle DF (2009) 1,5-anhydroglucitol monitoring in diabetes: A mass balance perspective. Clin Biochem 42: 158–167.
  54. 54. Toth BI, Geczy T, Griger Z, Dozsa A, Seltmann H, et al. (2009) Transient receptor potential vanilloid-1 signaling as a regulator of human sebocyte biology. J Invest Dermatol 129: 329–339.
  55. 55. Buse JB, Freeman JL, Edelman SV, Jovanovic L, McGill JB (2003) Serum 1,5-anhydroglucitol (GlycoMark): a short-term glycemic marker. Diabetes Technol Ther 5: 355–363.
  56. 56. Brass EP (1986) Effect of alpha-ketobutyrate on palmitic acid and pyruvate metabolism in isolated rat hepatocytes. Biochim Biophys Acta 888: 18–24.
  57. 57. Landaas S (1975) The formation of 2-hydroxybutyric acid in experimental animals. Clin Chim Acta 58: 23–32.
  58. 58. Landaas S, Pettersen JE (1975) Clinical conditions associated with urinary excretion of 2-hydroxybutyric acid. Scand J Clin Lab Invest 35: 259–266.
  59. 59. Pettersen JE, Landaas S, Eldjarn L (1973) The occurrence of 2-hydroxybutyric acid in urine from patients with lactic acidosis. Clin Chim Acta 48: 213–219.
  60. 60. Silva AR, Ruschel C, Helegda C, Wyse AT, Wannmacher CM, et al. (2001) Inhibition of in vitro CO2 production and lipid synthesis by 2-hydroxybutyric acid in rat brain. Braz J Med Biol Res 34: 627–631.
  61. 61. Yang W, Roth KS (1985) Defect in alpha-ketobutyrate metabolism: a new inborn error. Clin Chim Acta 145: 173–182.
  62. 62. Quinlan GJ, Lamb NJ, Tilley R, Evans TW, Gutteridge JM (1997) Plasma hypoxanthine levels in ARDS: implications for oxidative stress, morbidity, and mortality. Am J Respir Crit Care Med 155: 479–484.
  63. 63. Quick KL, Hardt JI, Dugan LL (2000) Rapid microplate assay for superoxide scavenging efficiency. J Neurosci Methods 97: 139–144.
  64. 64. Becker BF (1993) Towards the physiological function of uric acid. Free Radic Biol Med 14: 615–631.
  65. 65. Forman HJ, Dickinson DA (2003) Oxidative signaling and glutathione synthesis. Biofactors 17: 1–12.
  66. 66. Pastore A, Federici G, Bertini E, Piemonte F (2003) Analysis of glutathione: implication in redox and detoxification. Clin Chim Acta 333: 19–39.
  67. 67. Wu G, Fang YZ, Yang S, Lupton JR, Turner ND (2004) Glutathione metabolism and its implications for health. J Nutr 134: 489–492.
  68. 68. Townsend DM, Tew KD, Tapiero H (2003) The importance of glutathione in human disease. Biomed Pharmacother 57: 145–155.
  69. 69. Garcia-Ruiz C, Colell A, Morales A, Kaplowitz N, Fernandez-Checa JC (1995) Role of oxidative stress generated from the mitochondrial electron transport chain and mitochondrial glutathione status in loss of mitochondrial function and activation of transcription factor nuclear factor-kappa B: studies with isolated mitochondria and rat hepatocytes. Mol Pharmacol 48: 825–834.
  70. 70. Garcia-Ruiz C, Morales A, Colell A, Ballesta A, Rodes J, et al. (1995) Feeding S-adenosyl-L-methionine attenuates both ethanol-induced depletion of mitochondrial glutathione and mitochondrial dysfunction in periportal and perivenous rat hepatocytes. Hepatology 21: 207–214.
  71. 71. Kojer K, Bien M, Gangel H, Morgan B, Dick TP, et al. (2012) Glutathione redox potential in the mitochondrial intermembrane space is linked to the cytosol and impacts the Mia40 redox state. Embo J 31: 3169–3182.
  72. 72. Hansen SH, Grunnet N (2013) Taurine, glutathione and bioenergetics. Adv Exp Med Biol 776: 3–12.
  73. 73. Calder PC (2006) n-3 polyunsaturated fatty acids, inflammation, and inflammatory diseases. Am J Clin Nutr 83: 1505S–1519S.
  74. 74. Guenther G, Peralta E, Rosales K, Wong S, Siskind L, et al. (2008) Ceramide starves cells to death by downregulating nutrient transporter proteins. Proc Natl Acad Sci U S A 105: 17402–17407.
  75. 75. Peralta E, Edinger A (2009) Ceramide-induced starvation triggers homeostatic autophagy. Autophagy 5: 407–409.
  76. 76. Mao C, Obeid LM (2008) Ceramidases: regulators of cellular responses mediated by ceramide, sphingosine, and sphingosine-1-phosphate. Biochim Biophys Acta 1781: 424–434.
  77. 77. Bieberich E (2011) Ceramide in stem cell differentiation and embryo development: novel functions of a topological cell-signaling lipid and the concept of ceramide compartments. J Lipids In Press.
  78. 78. Nichols FC, Rojanasomsith K (2006) Porphyromonas gingivalis lipids and diseased dental tissues. Oral Microbiol Immunol 21: 84–92.
  79. 79. Cobb J, Gall W, Adam KP, Nakhle P, Button E, et al. (2013) A novel fasting blood test for insulin resistance and prediabetes. J Diabetes Sci Technol 7: 100–110.