Limited Use of Outpatient Stress Testing in Young Patients With Atypical Chest Pain

Discussion

In this retrospective, observational study of 1,036 young patients with atypical chest pain who had stress testing, there was relatively strong agreement between baseline ECG and index stress test results. Individuals also were 8 times more likely to have positive baseline ECGs and negative stress testing than having the opposite finding. Additional cardiac testing similarly demonstrated congruency with index stress testing and showed the propensity for false-positive stress tests. Further testing with CCTA demonstrated minimal nonobstructive CAD in < 1% of the study cohort and 2 LHC were negative. Despite the low prevalence of CAD and apparent low diagnostic use of stress testing in our young cohort, symptomatic service members still require stress testing to determine deployment suitability.

The low yield of outpatient stress testing in our young population is rooted in Bayes’ theorem, which highlights the importance of pretest likelihood in the diagnosis of CAD.^7,23 Because our cohort had a low prevalence and low pretest likelihood of CAD, positive index stress tests were often false-positive results and consequently did not increase the posttest likelihood of CAD, resulting in low positive predictive value. Additional cardiac testing had limited clinical value in our cohort. The 3 cases of nonobstructive CAD were unlikely to be pathologic given the minimal degree of observed stenosis and the 2 LHC did not require revascularization. These results are similar to those shown by Christman and colleagues and Mudrick and colleagues, which highlighted the low yield of additional cardiac studies and low rate of revascularization among symptomatic patients without known cardiac disease, respectively.^18,19

This is the first study, to our knowledge, to quantitatively demonstrate the low use of outpatient stress testing for young adults with atypical chest pain. Previous studies that assessed stress testing for young patients with chest pain in acute settings such as emergency departments and chest pain observation units, similarly demonstrated minimal yield of routine diagnostic testing.^23,24 This further highlights the premise that outpatient and even emergent-setting stress testing in low cardiac risk individuals may be of limited value and not always necessary.

Limitations

There were several study limitations. As a single-center, cross-sectional review, we may not be able to extrapolate our findings to the general population. However, given the low prevalence of CAD in young adults, stress testing would likely have limited value regardless of the sample distribution; so it may be possible to extend our findings beyond our cohort. Also, neither baseline ECG nor index stress test (irrespective of modality) could be given a diagnostic value in predicting ischemia alone; doing so would require comparison with the gold standard—heart catheterization. Although referral bias has been associated with diagnostic performance of stress testing, we did not adjust for this phenomenon.²⁵ Given the higher average metabolic equivalents achieved in our cohort, this potential bias likely did not affect diagnostic performance.

Conclusion

There was low diagnostic use of outpatient stress testing and additional cardiac testing for CAD among young patients with atypical chest pain. The limited value of cardiac stress testing is likely a function of the low CAD prevalence within this population, suggesting that younger patients may not necessarily require stress testing for chest pain evaluations as long as pretest likelihood is low. Despite our results, we maintain that the decision to perform stress testing should still be guided by clinical judgment, but perhaps our findings may alleviate physicians’ concerns over the urgency of when to refer low-risk patients for testing. Although we are cautious in inferring our findings to the general population, the similarity it shares with those from other published reports may suggest its applicability beyond our study cohort.