From STEMI to occlusion MI: paradigm shift and ED quality improvement

A generation ago the ST-elevation myocardial infarction (STEMI) paradigm led to quality improvement (QI) in the emergency department (ED). Now, insights from angiography and advances in electrocardiogram (ECG) interpretation have led to the new paradigm of occlusion myocardial infarction (OMI), creating the possibility of further QI. This article reviews the current STEMI paradigm, the emergence of the OMI paradigm, and the use of QI to continuously improve care for acute myocardial infarction (AMI) patients in the ED. STEMI paradigm and QI


STEMI paradigm and QI
Thrombolytic therapy in the 1990s led to a paradigm shift in the treatment of AMI through emergent reperfusion. This changed the use of the ECG, from retrospectively classifying AMI into Q-wave/non-Q wave to prospectively identifying those with ST elevation, as a marker of AMIs with persistent occlusion without collateral circulation, which need emergent reperfusion. ED providers responded with QI initiatives to reduce reperfusion delays for AMIs with ST elevation, or STEMI, from emergency nurse-initiated ECG acquisition to emergency physician-initiated cath lab activation.
However, from the beginning of the STEMI paradigm there were questions about ECG interpretation at the heart of the diagnostic process. A 1994 report on ED delays published in Annals of Emergency Medicine summarized, "ECG abnormalities may be subtle or open to different interpretation, such as early repolarization or pericarditis. Only borderline or minimal ST-segment elevation may be present, and the emergency physician may be uncertain of its significance. The presence of left bundle branch block or left ventricular hypertrophy may complicate ECG diagnosis. The emergency physician may suspect that the ST elevation is old, but a previous ECG may be unavailable for comparison. The computer interpretation of the ECG on which some physicians rely may be incorrect. The emergency physician may not be sufficiently trained to recognize certain ECG patterns as signs of AMI" [1].
At the time little could be done to improve on these quality issues. Those that did not meet STEMI criteria were labeled "non-STEMI" (NSTEMI) and did not receive emergent reperfusion. But in the nearly 30 years since this paradigm emerged, insights from angiography and advances in ECG interpretation have identified the limits of this paradigm and given rise to a new one.

From STEMI to OMI
Whereas the original thrombolytic trials were limited by rudimentary ECG analysis and AMI diagnosed by CK-MB (not angiography, and not even troponin), studies using angiography and formal STEMI criteria have put the paradigm to the test. For patients with STEMI, as adjudicated retrospectively by cardiologists, a recent prospective validation of STEMI criteria found that automated interpretation of the first ED ECG was only 35% sensitive for STEMI and 21% sensitive for any occlusion [2]. In a meta-analysis of 40,777 NSTEMIs in highly-monitored randomized-controlled trials, Khan et al. found a quarter of patients had a completely occluded coronary artery at the time of delayed angiography and had a nearly double mortality rate compared to NSTEMI patients with an open artery [3].
In response to these limitations, advances in ECG interpretation have identified signs of acute coronary occlusion that do not meet STEMI criteria. Emergency physicians such as Dr. Stephen Smith have played a leading role in these advances, which are summarized in his article in the Canadian Journal of Cardiology by Miranda et al. [4], and most recently in an article that provides step by step instructions in the diagnosis of OMI, and exclusion of mimics [5]. Examples include reciprocal ST depression in aVL, which can identify subtle inferior OMI and exclude pericarditis; a decision rule can differentiate between subtle left anterior descending (LAD) coronary artery occlusion and normal variant ST elevation in leads V2-V4; the modified Sgarbossa criteria can identify acute coronary occlusion in the presence of left bundle branch block and ventricular paced rhythms; the T/QRS ratio can differentiate LV aneurysm morphology from acute infarct; and primary ST depression maximal in V1-4 can identify posterior OMI. Table 1 demonstrates examples of these OMI ECG findings, and the full range can be found in these references by Miranda et al. [4] and Aslanger et al. [5].
These advances have given rise to a new paradigm, shifting the focus from the surrogate marker of ST segment millimeter criteria to the underlying pathology: occlusion MI [6]. Recent studies have now directly compared these paradigms. In the DIFOCCULT study, Aslanger et al. found that advanced ECG interpretation by cardiologists could reclassify 28% of NSTEMI as OMI, and this subgroup had a higher mortality rate than NSTEMIs whose ECGs had no evidence of OMI [7]. Meyers et al. showed that STEMI (+) OMI and STEMI (−) OMI have the same infarct size, mortality, number of wall motion abnormalities, and coronary interventions, which are significantly different than NSTEMI and especially NSTEMI that are non-OMI [8]. Furthermore, they found that emergency physicians expertly trained in ECG interpretation could identify OMI with twice the sensitivity as STEMI criteria, and significantly earlier [9].
These developments have answered the questions raised by Annals in 1994: computer interpretation and the STEMI paradigm on which it is based have limited accuracy for identifying acute coronary occlusion, evidence-based advances in ECG interpretation can differentiate between different causes of ST elevation and identify OMIs that do not meet STEMI criteria, and emergency physicians can be trained in this new paradigm. This new knowledge needs to be translated to the ED through QI approaches.

OMI paradigm and QI
Among QI interventions, standardization and automation are higher on the hierarchy of effectiveness [10]. But we are currently operating with a paradigm based on a suboptimal standard, reinforced by inaccurate automation. All research, guidelines, and QI initiatives are designed only to improve care for patients with OMI that meet STEMI criteria on their ECG, ignoring those who don't. Ultimately, we need to complete the paradigm shift, with OMI as the new standard, aided by artificial intelligence ECG interpretation of the totality of the ECG, not only the ST segments. Until that time, other QI interventions assume greater importance.
EDs should assess ECG and OMI quality benchmarks. A Door-to-ECG time of less than 10 min has been a key quality benchmark that has helped emergency nurses improve the speed of triage ECG acquisition through multiple QI interventions [11]. But there is a surprising lack of complementary quality benchmark for emergency physicians, perhaps because of simplified STEMI criteria. ECG-to-Activation time reflects the diagnostic time of emergency physicians, is independent of cath lab capabilities, and can be compared across different settings; this metric can help identify preventable reperfusion delays and promote new advances in ECG interpretation [12]. In our QI project, including a grand rounds presentation based on the article by Miranda et al., followed by weekly ECG audit and feedback to all physicians on signs of OMI, ECG-to-Activation time was reduced by 20 min [13].
EDs should review the ECG-to-Activation time (whether this activates their own cath lab or activates transfer to another centre's cath lab) for all their patients with OMI. This includes the 25% or more of NSTEMI patients with occluded arteries on angiogram and the third of true STEMI patients that have an open artery by the time of angiogram. In order to identify all patients with an occluded artery at ED presentation, the definition of OMI includes the following: (1) confirmed OMI (angiographic culprit lesion with TIMI 0-2 flow), and (2) presumed OMI with significant cardiac outcome, defined as: (a) angiographic acute but non-occlusive culprit lesion with highly elevated troponin (as defined in several studies, between 70 and 300 times the 99th percentile upper reference limit, depending on the assay), (b) highly elevated troponin and new regional wall motion abnormality on echocardiography, in those without angiography, or (c) STEMI(+) ECG with death before angiogram [7][8][9]. EDs can design QI interventions based on this outcome, and target the different components of the reperfusion decision (Table 2).  [14]. It should be updated to include advances in OMI, which have been led by emergency physicians and empower emergency providers to better interpret ECGs at the bedside. As with point-of-care ultrasound (POCUS) skills [15], advanced ECG interpretation requires workshops and training to incorporate interpretation into clinical decision-making, in addition to ED administrative support and quality assurance. Moreover, the OMI paradigm shift is not just about the ECG. The entire outcome is changing from a single element of the ECG (i.e., certain ST elevation voltage) to a patient-oriented one (i.e. occlusion or not), and QI needs to reflect that. While POCUS is not needed for obvious STEMI(+)OMI and can unnecessarily prolong ECG-to-Activation time, advanced POCUS training can help identify regional wall motion abnormalities that complement subtle STEMI(−)OMI ECGs. Patient alerts for refractory ischemia could help identify OMI patients who require cath lab activation even in the absence of ECG changes (as current guidelines recommend). Protocols and audits of STAT cardiology consultations can help with joint decision-making for challenging cases that incorporate clinical, ECG and POCUS findings. Collaboration between emergency and cardiology departments on tracking OMI quality metrics, and implementing and assessing OMI quality improvement projects, can help emergency physicians and cardiologists advance towards the paradigm shift together.

Conflict of interest All authors report no conflict of interest.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.