Differential diagnosis of wide QRS tachycardias: comparison of two electrocardiographic algorithms

• Published in: Europace (London, England) England), 2015-09, Vol.17 (9), p.1422-1427
• Main Authors: Kaiser, Elisabeth, Darrieux, Francisco C C, Barbosa, Silvio A, Grinberg, Rodrigo, Assis-Carmo, Andre, Sousa, Julio C, Hachul, Denise, Pisani, Cristiano F, Kosa, Eva, Pastore, Carlos A, Scanavacca, Mauricio I
• Format: Article
• Language: English
• Subjects: Algorithms; Diagnosis, Differential; Electrocardiography; Heart Ventricles - physiopathology; Humans; Prospective Studies; Sensitivity and Specificity; Tachycardia, Supraventricular - diagnosis; Tachycardia, Ventricular - diagnosis
• ISSN: 1099-5129; 1532-2092
• DOI: 10.1093/europace/euu354

This study's aim is to compare the ability of two ECG criteria to differentiate ventricular (VT) from supraventricular tachycardia (SVT): Brugada et al. [horizontal plane (HP) leads] and Vereckei et al. [frontal plane (FP), specifically aVR lead], having electrophysiological study (EPS) as gold standard. After comparing, suggestions for better diagnosis of wide QRS-complex tachycardia (WCT) in emergency situations were made. Fifty-one consecutive patients with 12-lead ECG registered during EPS-induced regular WCT were selected. Each ECG was split into two parts: HP (V1-V6) and FP (D1-D3, aVR, aVL, and aVF), randomly distributed to three observers, blinded for EPS diagnosis and complementary ECG plane, resulting in total 306 ECG analyses. Observers followed the four steps of both algorithms, counting time-to-diagnosis. Global sensitivity, specificity, percentage of incorrect diagnoses, and step-by-step positive/negative likelihood ratios (+LR and -LR) were calculated. Kaplan-Meier curve was plotted for final time-to-diagnosis. Inter-observer agreement was assessed with kappa-statistic. Global sensitivity was similarly high in FP and HP algorithms (89.2 vs. 90.1%), and incorrect classifications were 27.4 vs. 24.7%. Forty-eight correct analyses by Vereckei criteria took 9.13 s to diagnose VT in the first step, showing that first step was fast, with high +LR, generating nearly conclusive pre- (72.6%) to post-test (98.0%) changes for VT probability. Both algorithms as a whole are similar for diagnosis of WTC; however, the first step of Vereckei (initial R in aVR) is a simple, reproducible, accurate, and fast tool to use. The negativity of this step requires a 'holistic' approach to distinguish VT from SVT.