The Diagnostic Value of ST Segment Analysis in Ambulatory ECG Monitoring Evaluated Using Bayes' Theorem in Patients Screened for Coronary Artery Disease

Abstract

Background: The diagnostic value of ambulatory ECG monitoring in screening for coronary artery disease has been studied in diverse and usually small groups of patients. There are no studies evaluating the diagnostic value of Holter recorded ST depression using the Bayes' theorem of probability, which accounts for the prevalence of the disease in prespecified populations.Purpose: Applying the Bayes' theorem, this study aimed to examine the diagnostic value of Holterrecorded 1‐mm ST depression in patients screened for coronary artery disease (CAD) and to identify groups of patients who may benefit from diagnostic ST segment monitoring in ambulatory ECG recordings.Methods: The ST segment analysis was performed in 24‐hour ambulatory ECG monitoring of 460 subjects (375 males; aged 35–65, mean 48.6 years), who were screened for CAD and had coronary angiography. The Bayes' formulae were used to calculate the predictive value of ST segment monitoring (posttest likelihood of CAD) in comparison to pretest likelihood of the disease based on age, gender, and symptoms.Results: The 1‐mm ST depression was identified in the ambulatory ECG monitoring in 203 (44%) patients. CAD was angiographically confirmed in 279 (61%) patients. The 1‐mm ST depression had 54% sensitivity, 71% specificity, 74% positive predictive value, and 50% negative predictive value for CAD. The Bayes' theorem analysis with adjustment for pretest likelihood of the disease in relation to age, gender, and symptoms showed that 1‐mm ST segment depression is significant diagnostically in patients with pretest likelihood of the disease exceeding 75%, i.e., in males aged 35–45 years and females aged 56–65 years, both with typical angina. In other groups of patients regardless of the symptoms, age, and gender, detection of ST segment depression does not improve the diagnostic process. A negative result (absence of ST segment depression) can be helpful, confirming the absence of CAD in patients with 15%–25% pretest likelihood of the disease, i.e., in females aged 35–45 years with atypical angina pains and in males aged 46–55 years with nonanginal chest pains.Conclusions: Based on our observations, the 1‐mm ST segment depression detected on 24‐hour ambulatory ECG monitoring indicates a high likelihood of CAD in patients with < 75% pretest likelihood of the disease. ST segment analysis in other age and gender relative groups of patients, regardless of the nature of their symptoms, does not significantly improve diagnosis process.