The study protocol was approved by the Institutional Review Board of Hennepin County Medical Center. The requirement for informed consent was waived because the research involved a retrospective analysis of deidentified patient data. This study was conducted in accordance with the ethical standards outlined in the Declaration of Helsinki.

This retrospective case-control study was a secondary preplanned analysis of the DOMI ARIGATO (Diagnosis of Occlusion MI and Reperfusion by Interpretation of the ECG in Acute Thrombotic Occlusion) database (ClinicalTrials.gov identifier: NCT03863327), which is a two-site collaboration (Stony Brook University Hospital and Hennepin County Medical Center) designed to investigate OMI. Stony Brook University Hospital is a suburban, academic hospital which serves as a regional cardiac catheterization referral center. Hennepin County Medical Center is an urban academic emergency department (ED) with a cardiac catheterization lab on site. Both sites have more than 100,000 ED visits per year.

At both sites we retrospectively collected a cohort of patients who presented to the ED with symptoms suggestive of possible ACS. Each site searched through the cardiac catheterization laboratory activation database, which provided both cases and controls. At Hennepin County Medical Center, additional controls were identified by searching the UTROPIA (Use of Abbott High Sensitivity Troponin I Assay in Acute Coronary Syndromes) database (ClinicalTrials.gov identifier: NCT02060760) for patients without OMI but with repolarization abnormalities (ST elevation, ST depression, or T-wave inversion) in order to ensure that the database contains as many control patients with abnormal ECGs as possible; approximately one-third of these patients had a final diagnosis of AMI that did not meet our OMI definition, who were placed in the NOMI cohort. At Stony Brook University Hospital additional cases and controls were identified by a prospectively collected population of ED patients who were admitted to the cardiology service with suspected ACS. Patients were excluded if no ECG data were available in the electronic medical records.

Chart review was performed by four emergency medicine physicians after training with a standardized data coding manual. Authors (HPM and SWS) were available for on-demand questions, feedback, and retraining as necessary. Demographics, clinical and laboratory results, serial ECGs, and angiographic findings were collected using REDCap (Research and Electronic Data Capture; https://project-redcap.org/) hosted at Stony Brook University Hospital. We collected all available transfer, prehospital, and study site ECGs for each patient.

ECG interpretation was performed by authors (SWS and HPM), who were blinded to all patient information except age and sex, using a standardized data form including multiple specific ECG findings, objective measurements, and subjective interpretations as surrogate findings for the diagnosis of OMI. Neither interpreter was familiar with any data on any patient except for age and sex (required to determine whether the ECG met STEMI criteria which are age- and sex-based). STEMI criteria were defined according to the fourth universal definition of MI, and thus ST elevation was measured in millimeters, to the nearest 0.5 mm as measured at the J-point and relative to the QRS onset (PQ junction). Interobserver variation to the nearest 0.5 mm has been previously established within our author group [3–6]. If any of a patient’s ECGs prior to cardiac catheterization met STEMI criteria, and ST elevation was not considered to be due to a nonischemic etiology, the patient was considered to be STEMI(+). Otherwise, the patient was considered to be STEMI(–).

OMI was defined similarly to prior studies as either “confirmed OMI” on cardiac catheterization by the presence of an acute culprit lesion with Thrombolysis in Myocardial Infarction (TIMI) flow grades 0 to 2 or “presumed OMI with a significant cardiac outcome,” defined as any of the following: (1) an acute but nonocclusive culprit lesion with highly elevated cardiac troponin (cTn; contemporaneous cTnT ≥1.0 ng/mL [Roche Diagnostics Elecsys; reference range, ≤0.01 ng/mL]; or contemporaneous cTnI ≥10.0 ng/mL [Abbott Architect fourth generation; reference range, ≤0.030 ng/mL]); (2) if no angiography, then highly elevated cTn and a new or presumed new RWMA on echocardiography; or (3) ECG positive for STEMI with death before attempted emergent catheterization. Formal adjudication was based on all available data, including ECGs, cTns, and angiogram results. If TIMI flow was not reported, the cineangiogram was reviewed by a cardiologist (JAK at Stony Brook University Hospital, and GS at Hennepin County Medical Center) with experience interpreting angiograms. NOMI is defined by the presence (usually) or absence (less common) of culprit, but with TIMI grade 3 flow and peak troponin within the specified threshold. NOMI usually needs intervention, usually with stenting of the culprit; however, this need not be performed emergently because there is no ongoing infarction—intervention within 24 hours is acceptable. If the ECG was considered to reflect “OMI,” then the myocardial wall involved was specified (anterior, lateral, inferior, or posterior). AMI was considered to be “ruled out” (MI ruled out, MIRO) if the patient did not have a final diagnosis of AMI. We assessed the interobserver reliability between HPM and SWS for all cases interpreted by both. Furthermore, 108 consecutive OMI cases were reviewed for STEMI criteria by a cardiologist blinded to the outcome and the study goals.

RWMA was defined by formal echocardiographic report as performed by a certified echocardiographer, usually with bubble contrast, and interpreted by a cardiologist. We did not define RWMA by one or more of the 17 standard segments from the American Heart Association (AHA) consensus because they were rarely specified in the formal echocardiography report. Rather, we defined the involved wall by the terminology that was reported, using the terms “inferior,” “anterior,” “septal,” “apical,” “lateral,” or “posterior.” Posterior WMA has been classified by some cardiology societies as a lateral and/or inferior region simply because the AHA consensus naming of ventricular walls does not include the word posterior to describe the myocardial region that is directly opposite the anterior wall [7]. However, the most recent American College of Cardiology guidelines on STEMI equivalents continue to use the term “posterior STEMI” to refer to this area of myocardium [8]. We used the term “posterior” because it is in agreement with our current guidelines for the purpose of identifying acute coronary occlusion, identifies a region that is opposite the anterior leads that does not manifest as ST elevation on any of the standard 12-leads, and because cardiologists use that terminology in the formal echocardiography reports available in our data. Electrocardiographic findings of anterior, lateral, inferior, and posterior were considered to correlate positively to echocardiographic findings of anterior/septal/apical, lateral, inferior, and posterior/inferobasal, respectively (Fig. 1).

We collected and compared expert ECG findings of patients with OMI and WMA by individual LV wall in order to explore the relationship between ECG and echocardiographic manifestations according to affected anatomical territory (Fig. 2). For any given OMI, more than one territory is usually affected. We also collected information on time to cardiac catheterization, and time to echocardiography for those with STEMI(+) OMI, STEMI(–) OMI, and MIRO. Descriptive analytics were used for statistical analysis.

There are several limitations to this study. First, it is retrospective and was performed at two tertiary care hospitals. Also, the retrospective nature of this study could have led to both selection and timing biases. Furthermore, most patients were White male, and all patients were at high risk for ACS, which may limit generalizability to all ED patients with possible ACS symptoms. Furthermore, RWMA may resolve after restoration of previously impaired coronary blood flow. Since STEMI(+) OMI patients underwent percutaneous intervention more rapidly than did those with STEMI(–) OMI (a median of 41 minutes vs. 152 minutes after presentation, respectively), wall motion may have been affected. It would have been useful, but less practical, to record the echocardiogram prior to PCI. This difference in timing might have affected the incidence of RWMA in either cohort. Furthermore, we used a regional wall motion convention different from that of the 17 AHA consensus segments. While this correlation is not optimal because some “apical” WMA, and some “septal” WMA can be associated with either an inferior or anterior infarct location, ECG was only classified into four locations; since this limitation applies to both STEMI(–) OMI and STEMI(+) OMI, this should not have affected our results. More importantly, the echocardiographers did not use the 17-segment convention. Additionally, nearly every patient with OMI undergoes formal echocardiography during index hospitalization at the study sites. However, nine patients with OMI in our study did not undergo formal echocardiography. It is unclear why these patients did not undergo formal echocardiographic assessment, although it may have been due to clinical circumstances (e.g., patient or family preference, death prior to echocardiography). Finally, although the STEMI(–) OMI group had a higher prevalence of preexisting CAD, and since we did not assess for new WMA, versus any WMA, it remains possible that some of the WMA attributed to acute OMI were instead a result of past infarction. Furthermore, this higher prevalence was in subgroups of those without OMI (both NOMI and MIRO) who underwent echocardiography; such patients were likely higher risk for WMA than NOMI and MIRO patients who did not undergo echocardiography.

The prevalence of RWMA was similar between STEMI(+) OMI and STEMI(–) OMI in this cohort, and both groups had a higher prevalence than the no-occlusion group. Location of RWMA was well-correlated anatomically with ECG findings regardless of the presence or absence of STEMI criteria. These data support the notion that acute OMIs are similar whether or not they meet ECG ST elevation criteria, and that ECG is accurate not only for diagnosing OMI but for predicting the myocardial territory affected. Furthermore, prospective research is warranted to determine the utility of echocardiography in the detection of STEMI(–) OMI.