Study design and participants
The study was performed at a single center - University Hospital in Hradec Králové, which serves as a regional hospital for a district of around 160 thousand inhabitants. We enrolled patients examined in the emergency department of our hospital for one year (from January 1 to December 31, 2016). The ethical committee of University Hospital approved the study. The trial was conducted following the principles of the Declaration of Helsinki. The inclusion criteria were chest pain as a dominant symptom, age above 70 years, twelve lead electrocardiogram (ECG) recorded, and at least one blood sample for high sensitive troponin T (hs-TnT) analyzed. We excluded patients with apparent ST-segment elevations on the ECG (these were recommended for acute coronary angiography) and those unable to follow up.
Data collection
Demographic data, chest pain characteristics, ECG findings, past medical history, and hs-TnT concentrations for every patient were gathered from the hospital information system retrospectively. They were used to calculate T-MACS, HEART, EDACS, TIMI, GRACE, and ADAPT scores.
Laboratory analysis
Upon arrival to the emergency department, the blood sample was taken from a peripheral vein and immediately sent to the local laboratory for analysis. We analyzed blood samples for the concentration of high sensitivity troponin T (hs-TnT) with the assay Elecsys from Roche Diagnostics (99th percentile upper reference limit of 14ng/L for men and women, coefficient of variation < 10% at 13 ng/L).
Decision aid rules
Each of these chosen aids uses different clinical, electrocardiographical, or biochemical variables. According to the predicted probability of acute myocardial infarction calculated by the test score, patients can be divided into different risk groups. Original decision aid rule authors proposed managing the patient at the emergency department following the calculated risk. For example, safe discharge from the emergency department for low-risk groups, admission to the ward and invasive approach for high-risk groups, and additional examination such as serial troponin testing or stress/ imaging tests for moderate-risk groups [3]. Table 1 shows the summary of the selected decision aids.
Troponin-only Manchester Acute Coronary Syndrom Score Decision Aid Rule (T-MACS) operates with six dichotomous variables and one single hs-TnT concentration (measured at the time of arrival to the emergency department) [4]. Patients could be divided into four groups according to their probability of acute coronary syndrome or 30-days major adverse cardiac events (MACE) - very low-risk, low, medium, or high risk. This protocol can essentially exclude MACE in the very low-risk patients (rule-out diagnostics) and predicts MACE in high-risk patients (rule-in diagnostics) with high probability [5].
HEART is the acronym for History, ECG, Age, Risk factors, and Troponin concentration. This model predicts the 6-week risk of MACE. Low-risk patients (score 0–3) should be discharged from emergency, moderate risk patients (score 4–6) are recommended for further examination during a hospital stay, and patients with high risk (score above 6) could undergo early invasive coronary angiography [6].
Emergency Department Assessment of Chest Pain Score (EDACS) identifies chest pain patients with a low risk for major cardiac events in 30 days. Risk stratification is based on age, sex, coronary artery disease risk factors, and symptoms and signs of myocardial ischemia. In the initial evaluation, troponin testing is not required. However, patients who do not meet the low-risk criteria should be ruled out for acute myocardial infarction with usual chest pain protocols with serial troponin testing [7, 8]. Only low-risk patients (score < 16) without ischemic changes on ECG and with negative troponins are safe for early discharge.
GRACE (Global Registry of Acute Coronary Events) is an extensive multicentre international database [9]. Formerly, the GRACE score was studied to estimate in-hospital mortality of patients with the confirmed acute coronary syndrome and is still recommended in current ESC guidelines for the management of acute coronary syndrome without ST-segment elevations [1]. In this study, we hypothesized about its predictive value in a general population of chest pain patients. We used an improved version of GRACE 2.0, which involves eight variables (age, pulse, systolic blood pressure, ECG ischemia, abnormal cardiac enzymes, creatinine value, cardiac arrest at admission, and Killip class) [10, 11].
TIMI (Thrombolysis in Myocardial Infarction) risk score for unstable angina/ non-ST elevation myocardial infarction is one of the first widely implemented chest pain decision rules. Seven variables - age 65 years or older, at least three risk factors for coronary artery disease, prior coronary stenosis of 50% or more, ST-segment deviation on the electrocardiogram at presentation, at least two anginal events in last 24 hours, use of aspirin, and elevated serum cardiac markers - were identified from huge international database using multivariate logistic regression. The score predicts a percentual risk of 14-days MACE (all-cause mortality, myocardial revascularization, or myocardial infarction) [12]. Patients with a score of 0 and 1 point are at low risk of adverse outcomes. High-risk score patients (6 and 7 points) require aggressive pharmacological treatment and an early invasive approach.
Accelerated Diagnostic Protocol to Assess Patients With Chest Pain Symptoms (ADAPT) evaluates electrocardiogram, 0-hour and 2-hour troponin value, and TIMI score. Initially, this protocol was created to identify patients at low risk for 30 days MACE, suitable for rapid discharge from the emergency department [13]. A combination of normal troponin, normal ECG and TIMI score of zero represents the low-risk population; abnormal troponin or abnormal ECG constitutes a high-risk group.
Solo Troponin T (TnT) strategy
we hypothesized about the accuracy of strategy using only one variable - single high-sensitivity Troponin T (hs-TnT) concentration with the cut-off value at the 99th percentile upper reference limit for used assay (14ng/L for men and women) for determination. Patients with hs-TnT value below 15ng/L were identified as low-risk, others as high-risk.
Table 1 Summary of chosen decision aids - evaluated variables, clinical outcomes in original trials, risk stratification, and value of age for prediction of ACS/ MACE
Protocol
|
T-MACS
|
HEART
|
EDACS
|
TIMI
|
GRACE
|
ADAPT
|
Variables
|
ECG ischemia
|
History
|
Age
|
Age ≥ 65
|
Age
|
Abnormal troponin at 0 or 2 hours
|
Crescendo angina
|
ECG
|
Sex
|
≥ 3 risk factors for CAD
|
Heart Rate
|
Ischemic changes on ECG
|
Pain radiating to shoulder
|
Age
|
Known CAD or ≥ 3 risk factors for CAD
|
Know CAD
|
Systolic Blood Pressure
|
Age ≥ 65
|
Pain associated with vomiting
|
Risk factors
|
Diaphoresis
|
ASA use
|
Creatinine
|
≥ 3 risk factors for CAD
|
Sweating observed
|
Initial Troponin
|
Pain radiating to arm, shoulder, neck, or jaw
|
Severe angina
|
Cardiac Arrest at Admission
|
Know CAD
|
Hypotension
|
Pain worsened with inspiration
|
ECG ischemic changes
|
ST-segment deviation
|
ASA use
|
Initial Troponin
|
Pain reproduced by palpation
|
Positive cardiac marker
|
Abnormal cardiac enzymes
|
Severe angina
|
Killip class
|
Outcomes
|
30 days ACS or MACEa
|
6 weeks MACEa
|
30 days MACEb
|
14 days MACEa
|
6 months death
|
30 days MACEb
|
Risk stratification
|
Very low risk (0-1%)
|
Low risk (score 0-3)
|
Low risk (score <16)
|
Low risk (score 0-1)
|
Low risk ( < 109 pts)
|
Low risk (normal TnT, normal ECG and TIMI 0)
|
Low risk (2-4%)
|
Moderate risk (score 4-6)
|
Not-low risk (score ≥ 16)
|
Moderate risk (score 2-5)
|
Medium risk (109-140 pts)
|
Intermediate (normal TnT, normal ECG, and TIMI 1)
|
Moderate risk (5-94%)
|
High risk (score >6)
|
High risk (score 6-7)
|
High risk (> 140 pts)
|
High (abnormal TnT or abnormal ECG, any TIMI)
|
High risk (>94%)
|
Value of age on total score
|
none *
|
age ≥ 65 = 2 pts
|
70 years = 12 pts
|
age ≥ 65 counts = 1 pt
|
70-79 years= 73pts
|
age ≥ 65 counts for TIMI score ≥1
|
71-75 years = 14 pts
|
76-80 years = 16 pts
|
81-85 years = 18 pts
|
>79years = 91pts
|
≥86years = 20pts
|
Maximum score
|
100%
|
10 pts
|
38 pts
|
7 pts
|
340 pts
|
NA
|
ACS - acute coronary syndrome; AIM - acute myocardial infarction; ASA - acetylsalicylic acid; CAD - coronary artery disease; ECG - electrocardiogram; MACE - major adverse cardiac event; MACEa - a composite of acute myocardial infarction, myocardial revascularization, and all-cause death (applicable for T-MACS, HEART, and TIMI); MACEb - a composite of acute myocardial infarction, myocardial revascularization, cardiac death, cardiogenic shock, cardiac arrest, ventricular arrhythmia and high-grade atrioventricular block (applicable for EDACS and ADAPT); NA - not applicable; * age is not involved in the calculation.
Outcomes
As in most of the original trials of decision aid rules mentioned above, we followed patients for one month. The primary endpoint was an occurrence of acute myocardial infarction (AMI) and major adverse cardiac events (MACE). MACE includes composite of AMI, myocardial revascularization, and all-cause mortality. The diagnosis of AMI was made out according to the current guidelines [1, 14].
Statistical analysis
Categorical data were presented as proportions and percentages, continuous data as median and interquartile range. Selected patient characteristics between the groups of patients with and without AMI were compared, the continuous data were tested with the Mann-Whitney U test and categorical data with the Pearson's Chi-square test for independence. The diagnostic accuracy of decision aid rules was evaluated by calculating sensitivity, specificity, odds ratios, negative and positive predictive values with respected 95% confidence intervals and constructing the receiver operator characteristics (ROC) curves. Sensitivities and specificities of decision aids were compared using McNemar`s test. Statistical analysis was performed using M.S. Office Excel (Microsoft, Redmont, Washington, USA) and open-source software SOFA Statistics v1.4.6 (Paton-Simpson & Associates Ltd).