Is there a new cutoff value for high sensitivity troponin to better risk stratify patients with non-ST elevated myocardial Infarction?

Background: The diagnosis of Non-ST Elevated Myocardial Infarction (NSTEMI) is dependent on elevation of high sensitivity troponin (Hs-troponin). The current cutoff point for Hs-troponin is highly sensitive but not specic for obstructive coronary artery disease (CAD). This study aims to determine the best cutoff point for diagnosing CAD in patients presented with NSTEMI. Methods:Our study included all patients admitted as NSTEMI that underwent coronary angiography (CAG). They were grouped into two groups; A and B. Group A has obstructive CAD of 70% or more stenosis and group B with non-obstructive CAD. Patients were assessed for their demographics, clinical history, laboratory and imaging results. Using SPSS version 22, the pooled cohort of patients were analyzed at signicant level <0.05 and the data were tested for signicant correlations between two predetermined groups. Results: Group A comprised 87.6% of the patients and both groups had a median age of 53 years. In Group A, 91% were males, 54% diabetics, 54% hypertensives, and median Hs-troponin was 145 ng/L. While in group B, 88% were males, 39% diabetics, 60% hypertensives, and median Hs-troponin was 54 ng/L. There was signicant correlation between the two groups in the percentage of diabetes and median troponin level (p < 0.05). A ROC curve has identied a level of 127 ng/dL as the best cutoff of Hs-troponin in detecting obstructive CAD (p=0.03). Interestingly, 60% of patients in group B had alternative diagnoses. Conclusion:Hs-troponin is sensitive but less specic for obstructive CAD. However; increasing its cutoff value will improve its specicity.


Background
Acute coronary syndrome with normal or near normal coronary arteries, commonly known as Myocardial Infarction/Injury with Non-obstructive Coronary Arteries (MINOCA), can be caused by many etiologies like coronary disorders, myocardial disorders, or non-cardiac disorders [1][2][3]. These etiologies include; coronary spasm, microvascular dysfunction, myocarditis, hypertensive heart disease, stroke, sepsis, and pulmonary embolism [1,4].All these conditions result in rise of cardiac enzymes; thus, the challenge arises in making a decision whether the patient is suffering myocardial injury, ischemic infarction requiring further cardiac intervention [1].These challenging situations have increased currently with the introduction of high sensitivity cardiac troponin assay in view of its low speci city [1]. In a review article by Pasupathy et al. the prevalence of MINOCA was reported as 6%, ranging from 1-15%. Moreover, 14% of patients with acute myocardial infarction have been found to have non-obstructive coronary artery disease de ned as coronary stenosis of less than 50%. Research has shown that two third of patient with MINOCA present as NSTEMI [1,3].
Risk strati cation in patients with NSTEMI is of extreme importance as VIRGO study found the prevalence of MINOCA is ve times more in women than men and higher in patients with lower traditional cardiac risk factors. Another study showed that the prevalence of MINOCA is higher in younger patientsas well [5][6][7][8]. This has resulted in multiple risk scores being developed such as: Thrombolysis in Myocardial Infarction score, HEART score and Global Registry of Acute Coronary Events. These scores incorporate Hs-troponin as inherent factor included in the assessment of cardiovascular risks in addition to other clinical risk factors of coronary artery disease [4].All these risk scores use the current cutoff value of Hstroponin of 14 ng/dL which is sensitive but lack speci city for myocardial infarction.The American College of Cardiology recently published an article suggesting that an absolute value of Hs-Troponin of 52 ng/dL is an indictive to rule in acute myocardial injury [9]. Furthermore, the European Society of Cardiology recent 2020 guidelines have incorporated four confounders that result in rise of Hs-troponin; renal dysfunction, time to chest pain, age and gender [10].
The aim of this research is to better risk stratify patients by determining statically if a new cutoff value of Hs-troponin is of more signi cance than the current value of 14 ng/dL in identifying obstructive coronary artery disease in patients with NSTEMI.

Methods
This is a cohort of patients admitted between 1st of April 2018 through 31st of March 2019 in Rashid hospital. It included patients who were above 18 years old, having documented diagnosis of NSTEMI, and underwent coronary angiography regardless of the result and management strategy. The sample size was calculated at 80% power with marginal error of 5%. According to the literature there is a 15% possibility of normal CAG in patients undergoing coronary angiography this will equate to a sample size of 15 patients per arm for paired T test comparing the two group means [1][2][3]. Data was collected using a structured questionnaire present in an excel spreadsheet that included: demographics (age, gender and ethnicity), previous risk factors for cardiac disease, laboratory ndings (Low Density Lipoprotein value, Hemoglobin-A1C, and highest Hs-troponin value), and CAG nding. Patients were divided into two groups according to the CAG ndings. Group A with obstructive CAD de ned as coronary artery stenosis of 70% or more underwent coronary intervention, or advised for surgical or medical management. And Group B with non-obstructiveCAD de ned as normal or non-obstructive coronary artery disease of less than 50%.
Data was analyzed using SPSS software version 22. Statistical tests were carried out at 95 % con dence interval and results considered signi cant at a p-value of 0.05 or less. For variables where their numerical data is non-skewed the mean is reported (Hemoglobin and LDL) while for variables with skewed data the median is reported (all other variables with numerical data). On the other hand, for categorical variables the percentage is reported. Correlation was investigated using T-test for non-skewed numerical data, Mann-Whitney test for skewed numerical data and chi-square test for categorical data.

Results
Two hundred and two patients ful lled the inclusion criteria out of all 263 patients admitted as NSTEMI (   Table. 1 ) in the study period. Group A included 177 patients (87.6%) of which 91% were males, meanwhile Group B had 25 patients (12.4%) of which 88% were males. The median age in both groups was 53 years old. There is signi cant difference between group A and B in percentage of diabetics (P = 0.014) and Median Hs-troponin (P = 0.03), but no signi cant differences reported in relation to hypertension, smoking, HbA1C, or LDL (Table.2). The confounding variables were presented in Table.3 that showed no statistical signi cance between both groups.
Troponin as a diagnostic biomarker for myocardial infarction A receiver operating characteristic curve (ROC) was obtained to statistically measure a diagnostic cutoff value for troponin as a cardiac biomarker for NSTEMI ( Figure.1) considering CAG as the gold standard test. The ROC analysis showed an Area Under the Curve (AUC) of 0.634 (0.51-0.751) with a p value of 0.03 at con dence interval of 95%. Studying the sensitivities and speci cities of the ROC curve; it was deduced that the best cutoff value for the Hs-troponin is 127 ng/L.

Discussion
Troponin has been acknowledged as a cardiac biomarker of choice for diagnosis and management of NSTEMI by the American Heart Association (AHA), American College of Cardiology (ACC) and the European Society of Cardiology [11,12]. However, the ACC and AHA emphasize that elevation of troponin is not always due to atherosclerotic coronary artery disease [11]. Currently, 14 ng/L is considered as the cutoff value for a positive Hs-troponin result when evaluating patients with NSTEMI. This value is highly sensitive but lack speci city. When conducting a ROC curve on our data a Hs-troponin value of 14 ng/L had a sensitivity of 99.9% with an only 4% speci city. Further studying the ROC curve, a level of 127 ng/L was the best cutoff for Hs-troponin not compromising speci city, which showed a sensitivity of 58% and speci city of 68%. The test speci city will increase with increasing Hs-troponin level, while sensitivity decreases and vice versa. This curve has an AUC of 0.634 indicating a low diagnostic accuracy (<0.7) despite being statistically signi cant (P = 0.03) [13]. This directs that Hs-troponin is statistically signi cant as a biomarker for acute myocardial infarction but not without using additional clinical risk strati cation.
The presence of higher percentage of diabetes, which is a known risk factor for atherosclerotic cardiovascular events, as well as a higher median Hs-Troponin value in group A patients representing a higher chance of acute coronary syndrome and the likelihood of an underlying obstructive coronary artery disease as the myocardial damage is greater. The absence of signi cant difference between the groups in the confounding variables such as smoking status, history of hypertension, LDL value, creatinine, hemoglobin and CPK; gives more weight to use Hs-troponin as a determinant factor for predicting obstructive coronary artery disease.
An alternative diagnoses were found in those with a non-obstructive CAD (group B). Cardiac diagnoses were found in 32% of them which included valvular diseases, arrhythmogenic cardiomyopathy, myopericarditis, myocarditis, chest trauma and arrythmias; bigeminy, ventricular tachycardia, and supraventricular tachycardia. This indicates that the elevated troponin in these patients was due to a myocardial injury rather than a myocardial ischemic infarction. Furthermore, 20% of group B had noncardiac diagnoses which included bladder cancer, sarcoidosis, H.Pylori infection, gastric ulcer, and anemia. Several published articles support that these alternative cardiac and non-cardiac diagnoses are associated with elevated hs-troponin without any signi cant obstructive coronary artery disease [14][15][16][17][18].
Although our sample size is calculated to confer 80% power, our research is uni-centered and conducted in one-year time interval; a limitation that can be easily overcome by further research. This will confer more power and clarify the accuracy of Hs-troponin in detection of obstructive coronary artery disease in patients admitted with NSTEMI.

Conclusion
Our study concludes that at the current cutoff point of 14 ng/L is less speci c in detecting obstructive coronary artery disease. A new level of 127 ng/L is proposed as it holds a good speci city for detecting obstructive coronary artery disease. Further study is recommended to evaluate this result in a large-scale population. The additional use of clinical cardiac risk strati cation can improve the accuracy of Hstroponin in predicting obstructive coronary artery disease.

Declarations
Ethical Approval and consent to participate: An ethics approval was obtained from Dubai Scienti c Research Ethics Committee, DHA holding the reference number of DSREC/RRP/2019/13.

Availability of Data and material
Data is available with the authors in an excel sheet.

Consent for publication
Not applicable

Competing interests
Authors have no relationships relevant to the contents of this paper to disclose.

Funding
No funding source was for this manuscript.