ABC | Volume 113, Nº4, October 2019

Original Article Abolhasani et al. Serum levels of markers in CAD prediction Arq Bras Cardiol. 2019; 113(4):667-674 of isolated acute phase reactants increases the SA levels. SA is associated with atherosclerosis independently of other cardiovascular risk factors. 15 Previous studies have reported total serum levels of SA that were high in patients with acute coronary syndrome when compared to healthy controls. 34 Specifically, Govindarajan et al. 14 showed that the total plasma SA level was significantly higher in patients with myocardial infarction than in those with unstable and stable angina. In a recent 17-year follow-up study, elevated serum levels of SA were found to be predictive of cardiovascular events in apparently healthy individuals. 35 In addition, several studies suggested a positive relation between hs-CRP and CAD among healthy individuals. 36,37 Mahajan et al. 38 found a relation between inflammatory markers and coronary artery involvement on diabetic patients suffering from early onset CAD. 38 In addition, many evidences have indicated that hs‑CRP is a cautiously sensitive systemic marker for diagnosis of inflammation and a useful and potent predictive marker of cardiovascular events. 11,39 This study has shown that serum hs‑CRP and SA levels may be used as predictive or diagnosis biomarkers in patients with CAD. The findings in this study showed significant elevated serum levels of OX-LDL, MDA, PAI1-, VN, hs‑CRP and SA in patients with CAD, when compared to healthy individuals. hs‑CRP, SA and PAI-1 had the best sensitivity and specificity, suggesting the value of these biomarkers in patients with CAD diagnosis. ROC curve analysis showed satisfactory diagnostic power of all these six indexes, from highest to lowest: SA, hs-CRP, PAI-1, OX-LDL, MDA, and VN. This study also considered the diagnostic value of combined assay for all indexes for the best confirmative diagnosis value, including hs-CRP, SA, PAI-1, OX-LDL, MDA, and VN, resulting in elevated sensitivity and specificity values, without significant decrease of negative predictive value. These results supported the complementary role of combined assay in diagnosis of CAD. Table 3 – Area under curve of ROC Test Variable Area SD a p value b Asymptotic 95% Confidence Interval Lower bound Upper bound OX-LDL 0.870 0.032 .000 0.806 0.934 MDA 0.804 0.034 .000 0.738 0.869 PAI-1 0.951 0.010 .000 0.921 0.992 VT 0.799 0.036 .000 0.729 0.869 SA 0.962 0.021 .000 0.921 0.998 hs-CRP 0.971 0.032 .000 0.953 0.995 Note: a . Under the nonparametric assumption, b . Null hypothesis: true area = 0. OX-LDL: oxidation of low-density lipoprotein; MDA: Malondialdehyde; PAI‑1: Plasminogen Activator Inhibitor; VT: Vitronectin; SA: sialic acid; hs‑CRP: high-sensitivity C-reactive protein. Table 4 – Diagnostic efficacy of parameters index Positive criteria Sensitivity/% Specificity/% False negative /% False positive/% SA ≥ 89.7 (mg/dL) 94 96 6 4 hs-CRP ≥ 3.4 (mg/dL) 94 93 6 7 PAI-1 ≥ 67 (ng/mL) 92 90 8 10 VT ≥ 254 (ng/mL) 70 83 30 17 OX-LDL ≥ 2.67 (ug/mL) 70 75 30 25 MDA ≥ 5.49 (mmol/mL) 74 77 26 23 OX-LDL: oxidation of low-density lipoprotein; MDA: Malondialdehyde; PAI-1: Plasminogen Activator Inhibitor; VT: Vitronectin; SA: sialic acid; hs‑CRP: high-sensitivity C-reactive protein. Table 5 – Combined assay of PAI-1, VT, OX-LDL, MDA, SA and hs‑CRP Index Sensitivity/% Specificity/% OX-LDL, MDA, PAI-1, SA, hs‑CRP 97 95 OX-LDL, MDA, SA, hs‑CRP, VN 98 97 OX-LDL, VT, PAI-1, SA, hs‑CRP 98 97 VT, MDA, PAI-1, SA, hs‑CRP 97 97 OX-LDL, MDA, PAI-1, VT, SA, hs‑CRP 99 99 OX-LDL: oxidation of low-density lipoprotein; MDA: Malondialdehyde; PAI-1: Plasminogen Activator Inhibitor; VT: Vitronectin; SA: sialic acid; hs‑CRP: high-sensitivity C-reactive protein. 671