D and remodeled plaques (96 and 77 , respectively), which also surpassed the positive predictive value of each biomarker separately.Prediction of Clinical Outcomes by Plaque Composition and Biochemical MarkersDuring a mean follow-up duration of 3.260.9 years 19 MACE were recorded, including 6 deaths, 2 myocardial infarctions and 11 revascularization procedures in 147 patients. Five patients were lost during follow-up (3 ). Logistic regression analysis demonstrated that non-calcified plaque burden, hs-TnT and HMBG1 were significant predictors of clinical outcome (Table 4).Table 3. Calcium scoring, non-calcified plaque volume and plaque composition by hsTnT and HMBG1 tertiles.Calcium scoring (Agatston Units)HsTnT tertiles HsTnT lower tertile (3.0?.9 pg/ml) N = 50 HsTnT mid tertile (8.0?1.7 pg/ml) N = 51 HsTnT upper tertile (11.8?4.4 pg/ml) N = 51 1346185 1076186 2066203* HMBG1 tertiles HMBG1 lower tertile (1.2?.7 ng/ml) N = 51 HMBG1 mid tertile (4.8?.1 ng/ml) N = 50 HMBG1 
upper tertile (6.2?7.9 ng/ml) N = 51 1146194 1016150 2286209*Non-clacified plaque burden (mm3)Presence of non-calcified plaquePresence of remodeled plaque10.0617.9 6.5613.4 31.5627.6*37 33 7217 10 4211.5621.4 6.5613.7 29.3626.1*12 38 922 2 61Lower and upper tertiles for HMBG1 HsTnT HMBG1 HsTnT lower tertiles N = 20 
HMBG1 HsTnT upper tertiles N = 26 *p,0.05 versus mid and lower tertiles; { p,0.05 versus lower tertiles. doi:10.1371/journal.pone.0052081.t003 1716242 2716211{ 11.0620.6 43.7625.1{ 5 96 0 77HMGB1 and Atherosclerotic Plaque CompositionTable 4. Non-calcified plaque burden and biochemical markers for the prediction of clinical outcomes (combined endpoint for death, myocardial infarction and coronary revascularization).Variables Age(yrs.) Number of risk factors Non-calcified plaque burden Hs-CRP Hs-TnT Hmbg1 doi:10.1371/journal.pone.0052081.tCoefficient 0.05 0.31 0.02 0.06 0.08 0.Odds Ratio 1.05 1.37 1.02 1.06 1.08 1.95 Confidence Interval (CI) 0.99 to 1.10 0.90 to 2.09 1.00 to 1.04 0.98 to 1.14 1.00 to 1.16 1.07 to 1.p-value 0.06 NS ,0.05 NS ,0.05 ,0.Observer VariabilitiesInter- and intra-observer variability was 11 and 8 for the assessment of total plaque burden, and 1.8 and 1.3 for coronary calcium scoring, respectively. For differentiation by plaque composition inter- and intra-observer agreement rates were 90 (k = 0.76) and 95 (k = 0.88), respectively.DiscussionOur study demonstrates that besides hs-TnT, a well established marker of cardiovascular risk, HMGB1 is independently associated with non-calcified plaque burden in patients with clinically stable CAD. HMBG1 exhibits complementary value to hsTnT for the prediction of non-calcified plaque, which possibly implicates the involvement of different underlying pathophysiologic pathways in the release of the 2 biomarkers into the serum of such patients. Non-calcified plaque, hs-TnT and HMBG1 are all related to clinical outcome. HsCRP, on the other hand, a marker of 12926553 systemic low grade PHCCC inflammation is not independently associated with noncalcified plaque burden and with plaque composition. Several lines of evidence ZK 36374 site suggest that inflammation has a key role in the pathogenesis of atherosclerosis, plaque remodeling and transduction of the effects governed by conventional CAD risk factors [3,4]. In this regard, markers of vascular inflammation and metabolic risk factors have been previously identified as predictors of adverse clinical outcomes in patients with CAD [4]. However, since the unde.D and remodeled plaques (96 and 77 , respectively), which also surpassed the positive predictive value of each biomarker separately.Prediction of Clinical Outcomes by Plaque Composition and Biochemical MarkersDuring a mean follow-up duration of 3.260.9 years 19 MACE were recorded, including 6 deaths, 2 myocardial infarctions and 11 revascularization procedures in 147 patients. Five patients were lost during follow-up (3 ). Logistic regression analysis demonstrated that non-calcified plaque burden, hs-TnT and HMBG1 were significant predictors of clinical outcome (Table 4).Table 3. Calcium scoring, non-calcified plaque volume and plaque composition by hsTnT and HMBG1 tertiles.Calcium scoring (Agatston Units)HsTnT tertiles HsTnT lower tertile (3.0?.9 pg/ml) N = 50 HsTnT mid tertile (8.0?1.7 pg/ml) N = 51 HsTnT upper tertile (11.8?4.4 pg/ml) N = 51 1346185 1076186 2066203* HMBG1 tertiles HMBG1 lower tertile (1.2?.7 ng/ml) N = 51 HMBG1 mid tertile (4.8?.1 ng/ml) N = 50 HMBG1 upper tertile (6.2?7.9 ng/ml) N = 51 1146194 1016150 2286209*Non-clacified plaque burden (mm3)Presence of non-calcified plaquePresence of remodeled plaque10.0617.9 6.5613.4 31.5627.6*37 33 7217 10 4211.5621.4 6.5613.7 29.3626.1*12 38 922 2 61Lower and upper tertiles for HMBG1 HsTnT HMBG1 HsTnT lower tertiles N = 20 HMBG1 HsTnT upper tertiles N = 26 *p,0.05 versus mid and lower tertiles; { p,0.05 versus lower tertiles. doi:10.1371/journal.pone.0052081.t003 1716242 2716211{ 11.0620.6 43.7625.1{ 5 96 0 77HMGB1 and Atherosclerotic Plaque CompositionTable 4. Non-calcified plaque burden and biochemical markers for the prediction of clinical outcomes (combined endpoint for death, myocardial infarction and coronary revascularization).Variables Age(yrs.) Number of risk factors Non-calcified plaque burden Hs-CRP Hs-TnT Hmbg1 doi:10.1371/journal.pone.0052081.tCoefficient 0.05 0.31 0.02 0.06 0.08 0.Odds Ratio 1.05 1.37 1.02 1.06 1.08 1.95 Confidence Interval (CI) 0.99 to 1.10 0.90 to 2.09 1.00 to 1.04 0.98 to 1.14 1.00 to 1.16 1.07 to 1.p-value 0.06 NS ,0.05 NS ,0.05 ,0.Observer VariabilitiesInter- and intra-observer variability was 11 and 8 for the assessment of total plaque burden, and 1.8 and 1.3 for coronary calcium scoring, respectively. For differentiation by plaque composition inter- and intra-observer agreement rates were 90 (k = 0.76) and 95 (k = 0.88), respectively.DiscussionOur study demonstrates that besides hs-TnT, a well established marker of cardiovascular risk, HMGB1 is independently associated with non-calcified plaque burden in patients with clinically stable CAD. HMBG1 exhibits complementary value to hsTnT for the prediction of non-calcified plaque, which possibly implicates the involvement of different underlying pathophysiologic pathways in the release of the 2 biomarkers into the serum of such patients. Non-calcified plaque, hs-TnT and HMBG1 are all related to clinical outcome. HsCRP, on the other hand, a marker of 12926553 systemic low grade inflammation is not independently associated with noncalcified plaque burden and with plaque composition. Several lines of evidence suggest that inflammation has a key role in the pathogenesis of atherosclerosis, plaque remodeling and transduction of the effects governed by conventional CAD risk factors [3,4]. In this regard, markers of vascular inflammation and metabolic risk factors have been previously identified as predictors of adverse clinical outcomes in patients with CAD [4]. However, since the unde.