Predictions of arrhythmic, heart failure and mortality outcomes in pericarditis using automatic electrocardiogram analysis

BackgroundPericarditis is a relatively rare disease with a global burden. Despite its strong association with adverse cardiovascular outcomes, identification of patients at risk of future heart failure or arrhythmic events is difficult. In the following study, automated electrocardiogram (ECG) were used to predict new onset ventricular tachycardia/fibrillation (VT/VF), atrial fibrillation (AF) and heart failure with reduced ejection fraction (HF) in an Asian cohort of pericarditis patients. MethodsConsecutive patients admitted to a single tertiary center in Hong Kong, China, for a diagnosis of pericarditis between 1st January 2005 and 31st December 2019 with baseline ECG measurements were included. Patients with existing AF or HF were excluded. The follow-up period was until the 31st December 2020, or death. Cox regression was applied to identify significant predictors of the primary outcomes (incident VT/VF, AF or HF). ResultsA total of 874 patients were included. The cohort was 57% male and had a median age of 59 (IQR: 50-70) years old. During follow-up, 57 patients (6.5%), 156 (17.8%) and 168 (19.2%) suffered from VT/VF, AF and HF, respectively. Cox regression identified baseline VT/VF, terminal angle of the QRS vector in the transverse plane, mean QRS duration and mean QTc intervals as significant predictors of incident VT/VF events, with only the former most maintaining significance in multivariate analysis. In contrast, baseline age, prior diagnoses of hypertension, initial angle and magnitude of the QRS vector in the transverse plane, P-wave and QRS axis in the frontal plane, ST segment axis in the frontal and horizontal planes, mean PT interval, mean PR segment duration and QTc intervals were all univariate predictors of incident AF, albeit only baseline age and initial angel of the QRS vector in the transverse plane retained significance after multivariate adjustment. As it pertains to new-onset HF, several clinical and electrocardiographic parameters demonstrated an association with HF in univariate analysis, with prior diagnosis of HT or DM, initial QRS angle in transverse plane, I 40 in horizontal axis, ST-segment axis in the horizontal plane, T-wave frontal axis and atrial rate, of which, except for prior diagnosis of DM, I40 in horizontal axis and T-wave frontal axis, all variables showcased significant relationships in multivariate analysis InterpretationAF and HF are relatively common complications VT/VF occurs less frequently in the context of pericarditis. Different clinical and ECG predictors of these outcomes were identified. Future studies are still needed to evaluate their use for risk stratification in the clinical setting.