Electrocardiographic diagnosis of myocardial infarction in patients with left bundle branch block.

STUDY
OBJECTIVE: To validate ECG criteria previously proposed by Sgarbossa et al for the detection of myocardial infarction (MI) in patients with left bundle branch block (LBBB) and suspected ischemia.
METHODS: A retrospective cohort study was performed at an urban teaching hospital. All patients admitted with suspected ischemia and LBBB were eligible. MI was defined as an elevated creatine kinase (CK) isoenzyme MB (>14 IU/L) that was at least 5% of total CK level. ECGs were interpreted by 2 physicians blinded to patient outcome. Interpreters were asked to rate ECGs for the presence of each of the 3 criteria proposed by Sgarbossa et al: (1) ST-segment elevation greater than or equal to 1 mm concordant with the QRS complex; (2) ST-segment elevation greater than or equal to 5 mm discordant with the QRS complex; and (3) ST-segment depression in leads V(1) through V(3). Interobserver agreement was assessed.
RESULTS: Of 190 eligible patients, 25 (13%) had MI. Sensitivities of the 3 criteria varied from 0 to 16%, with specificities of 93% to 100%. Only the first criterion demonstrated a clinically useful likelihood ratio (positive likelihood ratio=16 [95% confidence interval 4 to >100]). Patients with new LBBB were more likely to have MI (relative risk=5. 1 [95% confidence interval 2.6 to 10]). Interobserver agreement among ECG interpreters ranged from 93% to 98%.
CONCLUSION: The criteria of Sgarbossa et al cannot be used to exclude MI in patients with LBBB because of low sensitivities and poor negative likelihood ratios. ST-segment elevation concordant with the QRS complex had a high positive likelihood ratio for identification of MI. Patients with new LBBB and suspected ischemia are 5 times more likely to have MI than patients with LBBB of chronic or unknown duration.