Introduction: Noninvasive postinfarction risk assessment for sudden cardiac death is limited. Standard analysis of the signal-averaged QRS complex can identify patients at risk for monomorphic ventricular tachycardia, but its value for discriminating patients at risk for sudden death is low.
Methods and results: The aim of this study was to prospectively investigate repeated late potential analysis of digital Holter ECG and compare it with standard analysis of the signal-averaged QRS complex within a short ECG period and with common clinical risk factors for sudden cardiac death in 756 consecutive patients after acute myocardial infarction. Digital Holter ECG were subdivided into 5-minute segments, and late potential analysis was performed on each 5-minute segment. During follow-up of 32 +/- 15 months, 35 patients died of sudden cardiac death and 50 patients died of nonsudden cardiac death. Sudden cardiac death was associated with ejection fraction < 40%, nonsustained ventricular tachycardia, creatine kinase > 1,000 IU/L, and late potentials in > 75% of analyzed Holter ECGs (abnormal LP75), but not with late potentials determined by only a short ECG period. According to multivariate analysis, the best independent significant predictor of sudden cardiac death was abnormal LP75 (P = 0.002, sensitivity 29%, specificity 96%, positive predictive value 40%, negative predictive value 97%). Nonsudden cardiac death was associated with ejection fraction < 40% (P = 0.001).
Conclusions: Late potential analysis of digital Holter ECG is a powerful tool that can be used to determine postinfarction patients at risk for sudden cardiac death and is optimized when combined with determination of ejection fraction.