The presence of complete left bundle branch block (LBBB) is thought to limit the utility of the electrocardiogram (ECG) in the evaluation of suspected myocardial infarction. In a 1996 study, Sgarbossa and colleagues reported on the use of an algorithm based on ST-segment changes that had a sensitivity of 78 percent and a specificity of 90 percent for the diagnosis of myocardial infarction in patients with LBBB. Shlipak and associates investigated the clinical utility of the algorithm by retrospectively studying a cohort of 83 patients with LBBB and symptoms of myocardial infarction.

The study included patients with complete LBBB and acute chest pain, acute pulmonary edema or cardiac arrest. (Patients in cardiac arrest were included if the ECG revealed LBBB after conversion to an atrial rhythm.) Excluded from the study were patients in whom creatine kinase or troponin I levels were not determined.

According to the algorithm developed by Sgarbossa, an ECG is considered positive for myocardial infarction if its score is at least 3 points on the basis of three criteria: ST-segment elevation of at least 1 mm in the lead with concordant QRS complex—a score of 5 points; ST-segment depression of at least 1 mm in lead V₁, V₂ or V₃—a score of 3 points; and ST-segment elevation of at least 5 mm in the lead with discordant QRS complex—a score of 2 points.

A total of 103 events occurred in the 83 patients with LBBB. Of these presentations, 30 percent were classified as myocardial infarctions. On the basis of the scoring system of the proposed algorithm, only three of the ECGs met the criteria for myocardial infarction in these patients with LBBB. All of the patients did, in fact, have a myocardial infarction. A positive result therefore had a very high positive predictive value. However, a negative result did not decrease the likelihood of a myocardial infarction. Of the nine ECG findings assessed, none was found to effectively identify the 30 percent of patients with myocardial infarction.

The authors also performed a decision analysis to determine the optimal course of treatment for patients with LBBB and acute chest pain. Possible treatment strategies were (1) to treat all patients with thrombolysis, (2) to treat no patients with thrombolysis or (3) to apply the proposed algorithm before deciding on thrombolysis. The analysis revealed that if 1,000 patients with LBBB who presented with chest pain received thrombolysis, 929 would survive without a major stroke, whereas if the ECG algorithm were used, 918 would survive without stroke.

The authors conclude that, although the proposed algorithm had a high sensitivity, it missed myocardial infarction in many patients who would benefit from aggressive treatment. The findings show that the ECG cannot reliably be used to rule out myocardial infarction in patients with LBBB. This study supports the recommendations of the American College of Cardiology and the American Heart Association: that all patients with LBBB and symptoms of acute myocardial infarction should receive reperfusion therapy if there are no contraindications.