Low-risk young adult patients with chest pain may not benefit from routine cardiac stress testing: a Bayesian analysis.

INTRODUCTION

Low-risk emergency department (ED) chest pain patients with a nondiagnostic electrocardiogram (ECG) and negative cardiac biomarkers are commonly evaluated with cardiac stress testing to detect undiagnosed coronary artery disease. Provocative testing incurs certain costs and may require additional time investment either in the ED or in an observation setting. Recent research has questioned the utility of provocative testing in young adults with negative cardiac biomarkers and nondiagnostic ECG. We sought to evaluate the utility of cardiac stress testing in our population of young adult patients with chest pain.

METHODS

We performed a retrospective chart review of all chest pain patients aged 40 years and younger who were admitted to our ED observation unit over the 14-month period between April 2006 and May 2007. We included all patients who were admitted to the observation unit for serial biomarkers and provocative testing and had normal or nondiagnostic ECG, no history of coronary disease, and an initial negative troponin. We recorded baseline characteristics and stress test results of these patients and reviewed the patient charts for the 30-day period following discharge to identify repeat hospital visits and adverse events. We used Bayesian analysis to estimate the rate of true-positive stress testing in this population, using the only prior study of unit patients showing as high as 2 of 220 patients testing positive as a prior estimate.

RESULTS

A total of 36 patients met inclusion criteria; average age was 34.6 years old (range: 22-40 years) and 61% were male. Patient risk factors included hypertension (19%), diabetes (6%), family history (42%), and smoking (44%). All patients had negative serial cardiac biomarkers and a negative treadmill stress echocardiogram. Thirty-day follow-up demonstrated no adverse cardiac events. We performed Bayesian analysis through the addition of the 36 patients to the 220 patients represented by prior data. The posterior probability distribution changed slightly in location and scale gave a median estimated rate of positive stress testing in this population of 1.04% (95% credible interval, 0.24%-2.78%).

CONCLUSIONS

The population of chest pain patients younger than 40 years with no history of coronary disease, a nondiagnostic ECG, and negative serial biomarkers may not benefit from provocative testing. Our findings complement those reported previously on the limited utility of cardiac stress testing in this population.