Utility of salivary biomarkers for demonstrating acute myocardial infarction

Abstract

The comparative utility of serum and saliva as diagnostic fluids for identifying biomarkers of acute myocardial infarction (AMI) was investigated. The goal was to determine if salivary biomarkers could facilitate a screening diagnosis of AMI, especially in cases of non-ST elevation MI (NSTEMI), since these cases are not readily identified by electrocardiogram (ECG). Serum and unstimulated whole saliva (UWS) collected from 92 AMI patients within 48 hours of chest pain onset and 105 asymptomatic healthy control individuals were assayed for 13 proteins relevant to cardiovascular disease, by Beadlyte technology (Luminex(®)) and enzyme immunoassays. Data were analyzed with concentration cut-points, ECG findings, logistic regression (LR) (adjusted for matching for age, gender, race, smoking, number of teeth, and oral health status), and classification and regression tree (CART) analysis. A sensitivity analysis was conducted by repetition of the CART analysis in 58 cases and 58 controls, each matched by age and gender. Serum biomarkers demonstrated AMI sensitivity and specificity superior to that of saliva, as determined by LR and CART. The predominant discriminators in serum by LR were troponin I (TnI), B-type natriuretic peptide (BNP), and creatine kinase-MB (CK-MB), and TnI and BNP by CART. In saliva, LR identified C-reactive protein (CRP) as the biomarker most predictive of AMI. A combination of smoking tobacco, UWS CRP, CK-MB, sCD40 ligand, gender, and number of teeth identified AMI in the CART decision trees. When ECG findings, salivary biomarkers, and confounders were included, AMI was predicted with 80.0% sensitivity and 100% specificity. These analyses support the potential utility of salivary biomarker measurements used with ECG for the identification of AMI. Thus, saliva-based tests may provide additional diagnostic screening information in the clinical course for patients suspected of having an AMI.

Keywords: biological markers; coronary disease; diagnosis; early diagnosis; saliva; serum.

Figure 1.

Box plots showing analyte levels expressed as log values in serum (A) from 90 acute myocardial infarction (AMI) patients and 102 non-acute myocardial infarction adult control individuals, and in unstimulated whole saliva (B) from 73 AMI and 85 control individuals. Concentrations are pg/mL for all analytes, except Adip, sICAM-1, CRP, TnI, CK-MB, and MYO, which are ng/mL. Samples were obtained within 48 hr of chest pain onset and analyzed in duplicate. Seventeen AMI patients were unable to provide sufficient saliva for analysis. The p values shown above the bars indicate significant differences between the 2 groups as determined by Wilcoxon Rank Sum statistics.

Figure 2.

CART decision trees. (A) Based on serum biomarkers in which 90 cases and 102 control individuals were included for the primary endpoint of AMI. Two cases and 3 control individuals had insufficient sample volume for analysis of all biomarkers. (B) The CART decision tree based on unstimulated whole saliva (UWS) biomarkers in which 73 cases and 85 control individuals were included for the primary endpoint of AMI; analysis done without ECG. Nineteen cases and 20 control individuals had insufficient sample volume for analysis of all biomarkers. (C) CART analysis with UWS biomarkers from 45 cases and 51 matched control individuals with the primary endpoint of AMI. (D) Forty-five cases and 51 matched control individuals, UWS biomarkers, confounders, and ECG to discriminate STEMI and NSTEMI.