ABC | Volume 110, Nº1, January 2018

Original Article Soeiro et al Troponin vs. coronary lesion Arq Bras Cardiol. 2018; 110(1):68-73 acute myocardial infarction were included. Presence of ST-segment elevation was the only exclusion criterion. The coronary lesion was considered significant when ≥ 70% on coronary angiography. Chronic renal failure was defined as a creatinine level > 1.5 mg/dL. The patients were divided into two groups: with (N = 681) and without (N= 310) significant coronary lesion. For Receiver Operating Characteristic (ROC) curve analysis, the patients were divided into two other groups: with (N = 184) and without (N = 807) chronic renal failure. The commercial ADVIA Centaur ® TnI-Ultra assay (Siemens Healthcare Diagnostics, Tarrytown, NY, USA) was used for current sensitive troponin with a 99 th percentile value of 0.04 ng/mL. The flowchart of the management of all patients with chest pain met the criteria established by the last American Heart Association guideline . 7-9 Non-ST-elevation acute coronary syndrome was defined as presence of chest pain associated with electrocardiographic changes or troponin elevation/drop on admission or, in the lack thereof, clinical findings and risk factors compatible with unstable angina (chest pain at rest or on minimal exertion, of severe intensity or occurring in a crescendo pattern). The highest troponin level during hospitalization before coronary angiography was considered for analysis, following the every 6-hour marker collection protocol of the institution. The following data were obtained: age, sex, presence of diabetes mellitus, systemic arterial hypertension, smoking habit, dyslipidemia, family history of early coronary artery disease, chronic coronary artery disease, previous acute myocardial infarction, creatinine, ST-segment depression or T-wave inversion on the electrocardiogram. This study was submitted to the Ethics Committee in Research and approved by it. All patients provided written informed consent. Statistical analysis The ROC curve analysis was performed to identify the sensitivity and specificity of the best cutoff point of troponin as a discriminator of the probability of significant coronary lesion, and 95% confidence interval (CI) was used. That analysis was performed for the general population and separately for patients with and without chronic renal failure. Descriptive analysis of the categorical variables was performed by use of percentages. Continuous variables with non-normal distribution were expressed as medians and interquartile intervals, and those with normal distribution, as means and standard deviations. The comparison between groups was performed by use of the chi-square test for categorical variables. The continuous variables, when the Kolmogorov-Smirnov test showed normal distribution, were assessed by using the unpaired T test, and when the distribution was not normal, the Mann-Whitney U test was used. Both troponin cutoff points analyzed (the 99 th  percentile of the method and the best cutoff point found in this study) were entered into the univariate analysis. Comparison between patients with versus without significant coronary lesion was performed. Multivariate analysis was performed with logistic regression, p < 0.05 being the significance level adopted. All baseline characteristics listed in Table 1 that reached statistical significance on univariate analysis were considered as variables in the analysis. Multivariate analysis was performed separately for each troponin cutoff point assessed (the 99 th percentile of the method and the best cutoff point found in this study). The calculations were performed with the SPSS software, version 10.0. Results The median age was 63 years, and 52% of the patients were of the male sex. The area under the ROC curve between the troponin levels and significant coronary lesions was 0.685 (95% CI: 0.65 – 0.72). In patients with or without renal failure, the areas under the ROC curve were 0.703 (95%CI:0.66–0.74)and0.608(95%CI:0.52–0.70),respectively. The best cutoff points to discriminate the presence of significant coronary lesion were: in the general population, 0.605 ng/dL (sensitivity, 63.4%; specificity, 67%; positive predictive value, 65.9%; negative predictive value, 64.7%; accuracy, 65.3%; and likelihood ratio, 1.9); inpatientswithout renal failure, 0.605ng/dL (sensitivity, 62.7%; specificity, 71%; accuracy, 66.9%; and likelihood ratio, 2.2); and in patients with chronic renal failure, 0.515 ng/dL (sensitivity, 80.6%; specificity, 42%; accuracy, 61.3%; and likelihood ratio, 1.4) (Figure 1). In the general population, the level of 0.05 ng/dL (immediately above the 99 th percentile) showed sensitivity of 93.7% and specificity of 23%. For patients with chronic renal failure to reach a specificity of 67% (as in the general population), an elevation in the troponin level to 1.58 ng/dL was necessary. Troponin was negative in 143 patients, and, in 40.6% of them, significant lesions were observed on coronary angiography. In addition, 10.5% of those patients with negative troponin showed ST-segment depression/T-wave inversion on electrocardiogram. Using the gold-standard procedure of cardiac catheterization, the acute coronary syndrome diagnosis was confirmed in 68.7% of the patients admitted due to chest pain. In 9.1% of those without significant coronary lesion on coronary angiography and with positive troponin, the acute coronary syndrome diagnosis was confirmed by cardiac magnetic resonance. The baseline characteristics of the population studied and the univariate analysis between the groups are shown in Table 1. In multivariate analysis, considering the 99 th percentile of the method, there were significant differences between the groups with and without coronary lesion regarding smoking habit (OR = 1.58, p = 0.002), ST-segment depression/T-wave inversion (OR = 2.05, p < 0.0001) and troponin positivity (OR = 3.39, p < 0.0001), respectively. However, when considering the best troponin cutoff point found in this study, there were significant differences between the groups with and without coronary lesion regarding the male sex (OR = 1.35, p= 0.039), smoking habit (OR= 1.64, p= 0.001), ST-segment depression/T-wave inversion (OR = 2.22, p < 0.0001) and troponin positivity (OR = 3.39, p < 0.0001), respectively. The multivariate analysis results are shown in Table 2. 69