Research Letter 

January 27, 2025

Cardiac Biomarker Testing in US Emergency Departments

Cian P. McCarthy, Nathan W. Watson, Jason H. Wasfy, et al

JAMA Intern Med. Published online January 27, 2025. doi:10.1001/jamainternmed.2024.7554

Patient characteristics and presenting symptoms are used in risk stratification for individuals evaluated in the emergency department (ED) as to their likelihood of having acute coronary syndrome (ACS). Identifying those at risk of ACS can inform selective cardiac biomarker testing. While undertesting could result in missed myocardial infarction,¹ overtesting may lead to incorrect diagnoses, prolonged hospital stays, and unnecessary cascades of testing.¹ Little is known regarding recent temporal trends in cardiac biomarker testing in US EDs.

Methods

This serial cross-sectional analysis used adult ED visit data (eFigure in Supplement 1) from the National Hospital Ambulatory Medical Care Survey (NHAMCS) to examine cardiac biomarker testing trends from 2014 to 2022.² Using reason for visit, individuals were classified into 3 groups: chest pain, non–chest pain anginal equivalent, and no anginal symptoms. ED diagnosis (first or second) and hospital primary discharge diagnosis were used to identify ACS (eTable in Supplement 1). The primary outcome was the proportion of ED visits with cardiac biomarker testing (troponin or creatine kinase MB).

Weights, strata, and primary sampling unit design variables were used for all analyses. Survey-weighted logistic regression was performed to investigate trends in cardiac biomarker testing and ACS diagnosis with year of visit included as a continuous independent variable and generate 2-sided P values. Variables for cardiac biomarker testing were identified using survey-weighted logistic regression adjusting for patient characteristics, hospital characteristics, year, and clinician type. A P value of <.05 was considered statistically significant. This study was deemed exempt from review by the Beth Israel Deaconess Medical Center Institutional Review Board as NHAMCS contains publicly available deidentified data; informed consent was waived. The STROBE reporting guideline was followed. All analyses were performed using R version 4.1.0 (R Foundation for Statistical Computing).

Results

In total, 97 085 ED visits were sampled, representing a weighted total of 731 977 412 visits (Table); 52.4% (95% CI, 50.8-54.0) included laboratory testing. Cardiac biomarker testing occurred in 7.0% (95% CI, 6.1%-8.1%) of all ED visits, without significant change over time (P = .22 for trend) (Figure, A). An electrocardiogram was performed in 27.1% (95% CI, 25.9%-28.4%) of all visits; 22.2% (95% CI, 19.3%-25.4%) of these visits also included cardiac biomarker testing. Among individuals undergoing cardiac biomarker testing, 34.7% had chest pain, 39.3% had non–chest pain anginal equivalent symptoms, and 26.0% had no anginal symptoms. Among those visiting the ED with chest pain, 25.6% (95% CI, 22.3%-29.2%) underwent cardiac biomarker testing (Figure, B). Testing was more frequent among older adults and men (Figure, C and D).

Abnormal vital signs such as hypotension (odds ratio [OR], 4.13; 95% CI, 1.52-11.20), Asian compared with White race (OR, 2.61; 95% CI, 1.57-4.36), and cerebrovascular disease (OR, 2.03; 95% CI, 1.68-2.44) were among the largest ORs in cardiac biomarker testing among individuals without anginal symptoms. Among individuals evaluated with cardiac biomarker testing, 2.8% were diagnosed with ACS (95% CI, 2.3%-3.4%), with a similar trend over time (P = .70 for trend). ACS was more common among those with chest pain symptoms (5.6%; 95% CI, 4.5%-7.0%) than among those with non–chest pain anginal equivalent (1.1%; 95% CI, 0.7%-1.8%) and nonanginal symptoms (1.5%; 95% CI, 0.8%-2.9%).

Discussion

In this study, 7% of all patients presenting to the ED underwent cardiac biomarker testing. A transient increase in testing occurred in 2020 early in the COVID-19 pandemic. Two-thirds of patients undergoing testing did not have chest pain. Among this subgroup, ACS incidence was lower than 2%, a threshold below which the harms of evaluating for myocardial infarction may outweigh the benefits.³ Conversely, 75% of patients with chest pain did not undergo cardiac biomarker testing. Study limitations include a lack of granular cardiac biomarker data and an inability to adjudicate the appropriateness of testing and ACS diagnosis. The study data suggest refinements in cardiac biomarker testing strategies are needed to optimize ACS diagnosis.