The impact of COVID‐19 on clinical outcomes among acute myocardial infarction patients undergoing early invasive treatment strategy

Abstract Background The implications of coronavirus disease 2019 (COVID‐19) infection on outcomes after invasive therapeutic strategies among patients presenting with acute myocardial infarction (AMI) are not well studied. Hypothesis To assess the outcomes of COVID‐19 patients presenting with AMI undergoing an early invasive treatment strategy. Methods This study was a cross‐sectional, retrospective analysis of the National COVID Cohort Collaborative database including all patients presenting with a recorded diagnosis of AMI (ST‐elevation myocardial infarction (MI) and non‐ST elevation MI). COVID‐19 positive patients with AMI were stratified into one of four groups: (1a) patients who had a coronary angiogram with percutaneous coronary intervention (PCI) within 3 days of their AMI; (1b) PCI within 3 days of AMI with coronary artery bypass graft (CABG) within 30 days; (2a) coronary angiogram without PCI and without CABG within 30 days; and (2b) coronary angiogram with CABG within 30 days. The main outcomes were respiratory failure, cardiogenic shock, prolonged length of stay, rehospitalization, and death. Results There were 10 506 COVID‐19 positive patients with a diagnosis of AMI. COVID‐19 positive patients with PCI had 8.2 times higher odds of respiratory failure than COVID‐19 negative patients (p = .001). The odds of prolonged length of stay were 1.7 times higher in COVID‐19 patients who underwent PCI (p = .024) and 1.9 times higher in patients who underwent coronary angiogram followed by CABG (p = .001). Conclusion These data demonstrate that COVID‐19 positive patients with AMI undergoing early invasive coronary angiography had worse outcomes than COVID‐19 negative patients.


| INTRODUCTION
There is a growing body of evidence regarding the cardiovascular manifestations of coronavirus disease 2019 (COVID-19) as well as the increased risk of cardiovascular morbidity and mortality. 1,2 These consequences may be related to the direct effects of the virus on the heart in addition to the pandemic's on healthcare systems leading to disruption in acute care pathways or alterations in health-seeking behaviors. 3 Hospitalized patients with COVID-19 have a 10%-30% risk of myocardial injury, with higher rates in severe cases. [4][5][6] This may be akin to the effects of other cardiotropic viruses that may induce direct myocardial injury or myocardial infarction. 7 Multiple mechanisms have been proposed including activation of inflammatory cells within atherosclerotic plaques, generation of a prothrombotic state leading to coronary thrombosis, and increase in metabolic demands from an activated immune system. 6 An early invasive strategy (24-72 h) for acute myocardial infarction (AMI) is associated with improved outcomes among high-risk patients. 8 Accordingly, consensus guidelines recommend an early invasive coronary angiogram with a revascularization strategy for patients presenting with high-risk AMI. 9 The most common pathways following an early invasive diagnostic coronary undergoing coronary angiography within 72 h of presentation to the hospital. We tested the hypothesis that there would be differential outcomes in COVID-19 patients who underwent coronary angiograms for AMI as compared to COVID-19 negative patients.

| Study design and participants
This study was a cross-sectional, retrospective analysis of the Given that the treatment guidelines for acute coronary syndromes recommend consideration for early risk stratification of AMI with coronary angiogram and or revascularization within 72 h, our cohort was constructed to capture patients who were considered for this treatment strategy. To evaluate whether our analysis should include control patients from the pre-COVID-19 era, we evaluated treatment groups' pre and post-COVID-19 era. We detected a variation in treatment groups between these two time periods (p < .001). As a result, we only included COVID-19− patients (comparison group) who had their initial AMI event during the COVID-19 era.
Groups of relevant medical codes were identified using the Observation Medical Outcomes Partnership (OMOP) common data model. These groups of medical codes, or concept sets, were created to identify COVID-19 diagnoses, AMI, as well as all criteria (comorbidities and outcome variables) used for matching. These concept sets were created using an open-source application called Atlas. Since sites use different medical coding, we only used codes standardized by the OMOP common data model to define our patient cohort and our variables because the N3C database also uses the OMOP common data model. Since Atlas allows us to collect standardized codes, we used this tool toto match the same data model as N3C. While many concept sets were custom built for this analysis, there were four concept sets that were already created that were reused to construct the data set for this analysis. Concept sets to identify cancer diagnoses and chronic obstructive pulmonary disease (COPD) were used in the N3C Consortium publication 13 on COVID-19 early severity prediction. The full list of concept sets is provided in supplemental materials, along with links to the code workbooks for reproducibility of variable derivation and statistical analysis (eMethods in the Supporting Information).

| Propensity matching
Data set was stratified into four groups based on the treatment received. For each strata, the propensity score for COVID-19 positivity was calculated via logistic regression with independent variables including demographics (age, sex, race, ethnicity), data partner ID, smoking status, and diagnosis of certain comorbidities on or before their first reported AMI. These comorbidities included diabetes type II, hyperlipidemia, uncontrolled hypertension, controlled hypertension, heart failure with reduced ejection fraction, heart failure preserved ejection fraction, peripheral arterial disease, obesity, atrial fibrillation/flutter, COPD, coronary artery disease (CAD), CABG procedures, cancer, and history of stroke. Height and weight measurements were also used to calculate body mass index (BMI) so that any patients with a BMI over 30 kg/m 2 on the measurement date most recent to their AMI were considered obese.
Exact matching was done based on race and data partner ID, while nearest neighbor matching was done based on other variables.
Demographic characteristics of patients pre-and postmatching according to treatment group are outlined in Supporting Information: Table S1. COVID-19 positive patients were matched to COVID-19 negative in a 1:3 ratio, using a calliper of 0.2 (Table 1). Only AMI (STEMI and NSTEMI) patients after March 15, 2020 were included in the study to ensure comparisons between the two groups (COVID-19 positive vs. negative) were made during the prevailing conditions of the pandemic.

| Outcome variables
The outcome variables included in this study were reported death, respiratory failure within 30 days of their AMI, cardiogenic shock within 2 days of their AMI, rehospitalization within 30 days of their AMI, and prolonged length of stay. Death was defined as a reported death in the patient record. Respiratory failure was defined using a concept set consisting of intubation of the respiratory tract, controlled mandatory ventilation, and veno-venous extracorporeal membrane oxygenation. If these procedures occurred within 30 days of the patient's index AMI event, then this was considered as having respiratory failure. Cardiogenic shock was defined as having a diagnosis of cardiogenic shock within 2 days of the index AMI event.

| RESULTS
The N3C registry included 1 222 296 adult patients of whom 10 520 had a diagnosis of AMI who met the prespecified inclusion criteria.
After excluding 14 patients with no identified age, there were 10 506 patients included in the final cohort ( Figure 1). The detailed descriptive results of demographics and comorbidities according to the invasive treatment strategy they received are displayed in   [18][19][20] In addition, some reports have described a persistent residual risk of cardiovascular complications even after the initial infection has resolved that may continue to predispose patients to other cardiovascular events such as myocarditis. 12,21 Our study highlights the adverse effects of COVID-19 on patients with CAD undergoing early invasive therapy with PCI. However, the subgroup T A B L E 2 Impact of COVID-19 status in AMI on outcomes for different treatment groups in the matched sample In the North American COVID-19 MI Registry, STEMI patients with COVID-19 who did not undergo coronary angiography had higher mortality than those who did. 15 Therefore, it is critically important to risk stratify patients with COVID-19 and AMI to ensure appropriate invasive treatment strategies and optimize clinical outcomes.

| Limitations
The limitations of this study include the real-world effects of medical record data, N3C data selection pipeline, and informatics limitations pertaining to computational processes on this large data set. There could also potentially be unidentifiable discrepancies in the data due to inconsistency in codes entered by physicians across multiple institutions. Studies have demonstrated a decrease in hospitalizations for AMI during the pandemic. 16 It would be reasonable to presume that the overall AMI admissions and interventions in our cohort were impacted by the pandemic. There is a wide range of false negative results based on the timing of the test. False negatives for PCR testing range from 20% to 100% while false negative for AG testing has been found to be <1%. 25 analysis to assess outcomes in STEMI and NSTEMI patients separately. However, these limitations are offset by the robust, quality-assured N3C data set which is the largest national database for COVID patients with centrally curate patient-level data. 10

| CONCLUSION
Our data reveals the differences in outcomes and deleterious effects of COVID-19 infection on AMI (STEMI and NSTEMI) patients undergoing early diagnostic coronary angiography. These data demonstrate that COVID-19 positive patients with AMI undergoing early invasive coronary angiography had worse outcomes than COVID-19 negative patients. The prolonged length of stay in COVID-19-positive patients has widespread implications on both patient outcomes and healthcare resource utilization. Timely diagnosis and intervention for AMI in this high-risk group of patients is essential to assure appropriate allocation of treatment strategies to improve outcomes.

ACKNOWLEDGMENTS
The analyses described in this (publication/report/presentation) were conducted with data or tools accessed through the NCATS N3C Data