Right bundle branch block and ventricular septal fibrosis in patients with hypertrophic cardiomyopathy

Machine-learning computer-assisted ECG analysis to predict myocardial fibrosis in patients with hypertrophic cardiomyopathy

AIMS

The application of computer assisted techniques to the electrocardiogram (ECG) analysis is showing promising results. Our main aim was to apply a machine learning approach to the ECG analysis in patients with hypertrophic cardiomyopathy (HCM), to identify predictors of macroscopic fibrosis, a marker of ventricular arrhythmias and sudden cardiac death.

METHODS

136 patients diagnosed with HCM were included. The main clinical and echocardiographic variables were collected. All patients underwent cardiac magnetic resonance (CMR) and the presence of macroscopic fibrosis was assessed on late gadolinium enhancement (LGE) sequences. From the 12‑lead digitized ECGs of each patient 468 morphological variables were quantified with a dedicated software.

RESULTS

The mean age of the population was 62.6 ± 14.1 years, and in 82 patients (60.3 %) LGE was observed. After preselecting significant ECG variables from the univariate analysis, a multivariate regression was performed, obtaining a predictive model composed of five parameters: the duration of the QRS in I, the duration of the QT interval in V3, the duration of the T wave in aVF, the peak-to peak amplitude of the QRS in V1, and the amplitude of the S wave in V4. A random forest algorithm confirmed that the duration of the QRS was the strongest predictor of fibrosis.

CONCLUSION

In patients with HCM the addition of a computer-assisted ECG analysis can help to identify predictors of LGE, being the duration of the QRS the strongest one. Our findings can be especially useful when access to CMR is scarce, to select patients at higher risk.

Left ventricular ejection fraction and mortality in patients with ST-elevation myocardial infarction and bundle branch block

BACKGROUND

The aim of our study is to assess the effect of bundle branch block (BBB) on mortality and left ventricular ejection fraction (LVEF) in ST-elevation myocardial infarction (STEMI) patients treated in the current era of percutaneous reperfusion therapy.

PATIENTS AND METHODS

In this retrospective cohort study, a total of 1123 STEMI patients treated in the University Medical Center Groningen from January 2011 until May 2013 were included. The follow-up duration was 2-4 years. Transthoracic echocardiography was performed within 2 weeks after STEMI.

RESULTS

In total, 23 (2.0%) patients presented with left BBB and 49 (4.4%) patients presented with right BBB. Two-year mortality after STEMI was 25.0% (n=18) in patients with BBB and 9.2% (n=97, P<0.001) in patients without BBB. Patients with BBB had more frequently a severely reduced LVEF (<30%) [20.0% (n=6) compared with 4.2% (n=21), P=0.002] and less frequently a normal LVEF [16.7% (n=5) compared with 35.7% (n=179), P=0.046]. After multivariable analysis, BBB did not remain an independent predictor of mortality, but was an independent predictor of reduced LVEF.

CONCLUSION

The presence of a BBB was an independent predictor of a reduced LVEF. However, we found no effect of BBB on 2-year mortality in the current era of percutaneous reperfusion therapy.