Greenslade JH, Carlton EW, Hise CV, et al. Diagnostic accuracy of a new high-sensitivity troponin I assay and five accelerated diagnostic pathways for ruling out acute myocardial infarction and acute coronary syndrome. Ann Emerg Med. 2018;71(4):439-451.
The purpose of this study was to describe the accuracy of 5 accelerated chest pain pathways in conjunction with a high-sensitivity troponin I (hs-TnI) assay in assessing risk and diagnosing acute myocardial infarction (MI) and acute coronary syndrome (ACS) in patients with chest pain presenting to the emergency department (ED). The chest pain pathways studied are as follows: (1) the modified accelerated diagnostic protocol to assess patients with chest pain symptoms using troponin as the only biomarker (m-ADAPT); (2) the Emergency Department Assessment of Chest Pain Score (EDACS); (3) the History, Electrocardiogram (ECG), Age, Risk Factors, and Troponin (HEART) score; (4) the No Objective Testing Rule; and (5) the new Vancouver Chest Pain rule, with endpoints of 30-day diagnosis of MI or ACS.
The m-ADAPT and EDACS pathways are validated for identification of acute MI; and the HEART, new Vancouver Chest Pain Rule, and No Objective Testing Rule are validated for identification of ACS. One of the goals of this study was to determine how adding an hs-TnI assay (measured at 0 and 2 hours after presentation) to each of the pathways would affect the assessment and disposition of patients with chest pain who present to the ED. In this study and analysis of data collected in 2 previous studies, one-third of patients were identified that would have been safe for discharge with no additional testing with the new Vancouver Chest Pain Rule or the No Objective Testing Rule and negative hs-TnI at 0 and 2 hours. The EDACS, m-ADAPT, and HEART pathways and hs-TnI rapidly identified 50% of patients who could be referred for additional testing.
Approximately 13% of patients presenting to the ED with chest pain are diagnosed with ACS. Assessment includes the history, ECG, and measurement of troponin levels to rule out acute MI. In patients in whom acute MI is ruled out, the next step is to rule out other types of ACS such as unstable angina. These patients are sometimes admitted for further testing or procedures, including stress tests and cardiac catheterization. As these workups can be time- and resource-intensive, more efficient ways of assessing risk for and ruling out ACS safely in the ED, such as adding troponin levels to established diagnostic pathways, are needed.
This study was undertaken as a re-analysis of data from 2 previous studies performed at a tertiary hospital in Australia; the data were collected as part of the original studies with the goal of performing this re-analysis. The studies (undertaken to assess chest pain protocols in the ED) included a prospective observational study of 986 patients from the Australian cohort of the 2-Hour Accelerated Diagnostic Protocol to Assess Patients With Chest Pain Symptoms Using Contemporary Troponins as the Only Biomarker (ADAPT) study and an intervention trial cohort of 1366 patients from the Improved Assessment of Chest Pain Trial (IMPACT). These 2 trials were chosen because they had similar data collection processes and patient inclusion/exclusion criteria. The final number of patients included in the analysis of this current study was 1811; eligible patients were aged > 18 years, had ≥ 5 minutes of chest pain and were being assessed for possible ACS. Patients were excluded if the symptoms were determined to be caused by a condition other than ACS, consent could not be obtained, the patient could not be contacted after discharge. The current study also excluded patients who presented with an ST-segment elevation myocardial infarction (STEMI), as they required immediate intervention.
This study retrospectively classified patient risk utilizing the different pathways, blinded to final patient diagnosis. The HEART pathway data were calculated retrospectively, as they were not measured in the original studies upon which this study was based. A patient was determined to have a positive hs-TnI if the value was greater than the 99th percentile (18 ng/L). Investigators also included sensitivity analyses in this study to determine whether time from onset of symptoms to presentation or sex of the patient influenced the diagnostic accuracy. The 30-day follow-up data (such as the occurrence of cardiac events, follow-up testing, or consultation with a healthcare clinician) were obtained by trained research nurses by contacting the patients or their healthcare providers, or by reviewing hospital database information. There was no loss to follow-up for this study.
The primary outcomes measured were acute MI (determined by death from a cardiac cause during hospital stay), diagnosis of NSTEMI or STEMI, or the need for emergent revascularization within 30 days of presentation to the ED. The secondary outcomes were ACS, as determined by death from a cardiac cause during hospital stay, diagnosis of NSTEMI or STEMI, the need for emergent or unplanned revascularization, or symptoms of unstable angina within 30 days of presentation to the ED. The ability to identify MI and ACS were assessed for all pathways; however, the primary endpoints differed for each pathway based on the designed purpose of the protocol. Cardiac endpoints were assigned and agreed upon by 2 cardiologists, with any disagreement resolved by consensus between the cardiologists and an emergency physician.
Out of the 1811 patients evaluated in this study, there was a low prevalence of acute MI and ACS, with 96 (5.3%) patients diagnosed with acute MI within 30 days, and 139 patients (7.7%) diagnosed with ACS. Sensitivity outcomes of this study are presented in Table 1.
The pathways combined with the 2 troponin levels were also evaluated to determine the number of patients they found to be at low risk for MI or ACS. The m-ADAPT identified 64.3% of patients as low risk; EDACS, 62.5%; HEART score, 49.8%; new Vancouver Chest Pain Rule, 28.2%; and No Objective Testing Rule, 34.5%.
When compared with the full patient population of the original 2 studies from which these data were collected, the patients in this study had a lower prevalence of acute MI and ACS, and were younger. The sensitivity analyses of time of presentation and sex of the patient did not find a significant effect on the sensitivity and specificity calculations for the pathways. The scores of early and late presenters were similar.
The authors concluded that the new Vancouver Chest Pain Rule and the No Objective Testing Rule combined with the hs-TnI assay could rule out 30-day ACS in 25% to 30% of patients safely (ie, these patients had a near-zero risk for acute MI or ACS). Based on the stated aims of these pathways, they would work well for systems in which patients are discharged from the ED with no follow-up, or to help reduce outpatient testing. The hs-TnI assay combined with the EDACS, m-ADAPT, and HEART pathways did not achieve a high enough sensitivity to rule out acute MI or ACS safely, as the acceptable sensitivity was 99%. However, the authors suggest that these tools can be used to rule out MI and then refer the patient rapidly to outpatient cardiology, or for objective inpatient or outpatient testing. Although the HEART pathway was designed to determine which patients could be discharged with no further testing, the results of this study of including the 2 hs-TnI levels led the investigators to instead suggest that it may be a more useful tool for rapidly identifying patients without acute MI who could be discharged for further outpatient assessment. This is because the HEART pathway achieved lower sensitivity for ACS as compared to the new Vancouver Chest Pain Rule and the No Objective Testing Rule. It also failed to achieve the 99% sensitivity for acute MI. The m-ADAPT and EDACS pathways combined with the hs-TnI classified two-thirds of patients as low risk and had moderately high sensitivity for acute MI and ACS. This is consistent with previous research; however, these pathways have not performed as well on United States cohorts. This study also showed that measurement of troponin levels alone was not sensitive enough for acute MI; using hs-TnI levels with the pathways led to higher sensitivity.
Overall, these results found that the negative hs-TnI levels at 2 hours combined with the new Vancouver Chest Pain Rule or the No Objective Testing Rule allowed for approximately one-third of patients to be discharged safely with no further testing. The use of the troponin studies with the EDACS, m-ADAPT, and HEART pathways identified more than half of ED patients who could be referred rapidly for inpatient or outpatient objective testing.
Overall, this study presents important information regarding the application of chest pain decision tools and troponin levels in the workup of patients with chest pain presenting to the ED. One concern regarding this study is the exclusion of patients who would be difficult to reach following discharge, such as homeless patients. This raises concern about generalizability of the study to all patient populations, as patients who are hard to reach following discharge may have different risk factors for MI or ACS at 30 days.
This study was performed on a cohort of patients from a single hospital, also leading to concern regarding the generalizability of the results of this study. The authors mention within the paper that the EDACS and m-ADAPT pathways have not stood up as well to validation in cohorts in the United States. Additionally, there may be some selection bias, as patients were selected only between the hours of 8 AM and 5 PM, and overall the subjects in this study were younger and had a lower incidence of ACS or acute MI than the full cohorts in the original studies in which each of these pathways were validated.
The clinical decision rules were calculated retrospectively from the data from the separate original studies, which can be problematic. This is especially an issue for the HEART score, since the researchers had to piece together the history component of this score. Since these pathways all have strong subjective elements, it would be important to ensure that the subjective data were obtained appropriately from the patient directly and recorded in the notes. As mentioned by the authors, this study would be improved if it were a prospective study in which these scores were calculated during the actual patient encounters.
The EDACS, m-ADAPT, and HEART score pathways were determined to be able to identify more than half of ED patients who could be referred for later testing. However, this testing could be either inpatient or outpatient. Since one of the primary functions of the ED is to determine whether the patient should be discharged or admitted, this paper does not provide clear guidance for practicing emergency clinicians.
Overall, this study opens up discussion of the application of chest pain decision tools and troponin levels to aid in diagnosis and risk assessment in the workup of patients with chest pain presenting to the ED. Although this study would be improved by having a wider population of patients as well as by being a prospective study, it does present valuable information in determining potential next steps for chest pain patients.