A variety of electrocardiographic (ECG) changes have been thought to have diagnostic value in patients with suspected pulmonary embolism (PE), but most investigators have studied only patients with confirmed PE. The diagnostic value of the ECG can only be determined by applying it to patients with suspected PE and then determining if the test is predictive of an outcome of confirmed PE.

Rodger and associates evaluated all patients with suspected PE who were seen at a large tertiary-care hospital and referred for ventilation-perfusion scanning or pulmonary angiography. The referring physician first assigned an index of clinical suspicion for PE based on all available clinical data, including the history, physical examination, blood gases, ECG, d-dimers and chest radiograph. All patients then underwent ventilation-perfusion scanning. After the scan, patients with a posttest probability of PE of less than 5 percent and a low index of pretest clinical suspicion were considered not to have PE. Patients with a posttest probability of greater than 88 percent and an intermediate or high index of pretest clinical suspicion were considered to have PE. All other patients were recommended to proceed to pulmonary angiography, which some did, depending on the patient's treating physician. Patients with indeterminate scores who did not have angiography were excluded from the analysis.

Standard 12-lead ECGs were obtained on 189 of the 212 classified patients (PE or no PE) and analyzed for 28 features thought to be more common with PE. Only tachycardia and incomplete right bundle branch block were significantly more frequent in patients with PE than those without PE. S₁Q₃T₃ was equally likely to be found in patients without PE who were initially suspected to have PE. Previous investigators found other ECG changes to be significantly more common in patients with PE, looked only at specific patient populations or used inadequate diagnostic criteria to identify patients with PE.

The authors conclude that the ECG is of limited diagnostic value in patients with suspected PE. Many of the classically described ECG changes in patients with suspected PE are equally common in patients suspected of having PE but in whom the diagnosis is ultimately excluded. Even the two statistically significant ECG changes noted in this study, tachycardia and incomplete right bundle branch block, are rarely observed and are only slightly more frequent in patients with PE.