[Can signal-averaged electrocardiograms be interpreted in cases of complete bundle branch block?].

The aim of this study was to evaluate the results of signal-averaged (SA) ECG in cases of complete right (RBBB) or left bundle branch block (LBBB). One hundred and seven patients had RBBB; 42 without cardiac disease (0), 56 with chronic myocardial infarction (MI) and 9 with primary cardiomyopathy (CMP). Seventy-four patients had LBBB: 20 without cardiac disease, 26 with chronic myocardial infarction and 28 with primary cardiomyopathy. A SA ECG (Cardionics, Fidelity) was performed with a 40 Hz band pass and compared with the recordings of 72 healthy controls without bundle branch block. The duration of the averaged QRS (QRS dur), the voltage of the last 40 milliseconds (RMS40) and duration of terminal activity < 40 microV (LAS) were measured. The analysis of results showed that QRS dur was significantly longer in subjects with ventricular tachycardia (VT) (p < 0.05) and in those with advanced cardiac disease (p < 0.05), whatever the type of bundle branch block, and that only the RMS40 distinguished patients with VT from those without VT, irrespective of the underlying cardiac disease and the type of bundle branch block. However, the study of the diagnostic value of each parameter showed very mediocre results: RMS 40 < 20 microV in myocardial infarction and < 17 microV in cardiomyopathy had sensitivities and specificities in RBBB of 73% and 50% respectively, incalculable in CMP, in LBBB 70% and 33%, 77% and 60% respectively; the LAS was unusable. The authors conclude that it is hazardous to interprete SA ECG in bundle branch block, especially in advanced cardiac disease where the specificity of the criteria becomes very low (< 50%).