Validation of Standard and New Criteria for the Differential Diagnosis of Narrow QRS Tachycardia in Children and Adolescents: Erratum

[This corrects the article DOI: 10.1097/MD.0000000000002310.].


INTRODUCTION
T he primary diagnoses in the differential of regular narrow QRS complex supraventricular tachycardia (SVT) in children and adolescents are atrioventricular nodal reentry tachycardia (AVNRT) and orthodromic atrioventricular reentry tachycardia (OAVRT). Analysis of a standard 12-lead electrocardiogram (ECG) is the first step toward determining treatment strategy, which may include choice of antiarrhythmic drugs, invasive procedure (radiofrequency catheter ablation [CA] vs cryoablation), selection of ablation center, and selection of an interventional cardiologist. [1][2][3][4][5][6][7][8] In 2009 and 2011, new electrocardiographic criteria for the differential diagnosis of SVT were reported to be highly accurate in adult populations. These criteria, which include the presence of a positive r 0 deflection in lead aVR during SVT which is absent during the sinus rhythm (SR) and notch of the QRS complex in lead aVL during SVT but not during the SR have not yet been tested and validated in a pediatric population. [2][3] The purpose of this multicenter study was to determine the value of the standard and new ECG criteria for the differential diagnosis of SVT in children and adolescents under 18 years of age who already underwent CA.

MATERIALS AND METHODS
The study included 219 consecutive pediatric patients (including those with AVNRT, OAVRT or atrial tachycardia [AT]) who underwent an electrophysiological study (EPS) and CA between January 2010 and December 2013. Patients provided written informed consent, and a local institutional review board approved the study, which complied with the Declaration of Helsinki. Patient ECGs were recorded at 3 different medical centers and retrospectively evaluated. In addition to the AVNRT and OAVRT groups, 7 ECGs that showed AT with 1 to 1 conduction were included in the evaluation but not in the statistical data. All patients on antiarrhythmic drugs discontinued them at least 5 half-lives prior to their planned EPS and ablation procedure. During EPS, arrhythmia was induced by pacing or medications if it had not occurred spontaneously. Patients with pre-excitation on the ECG during SR as well as those with atrial flutter, atrial fibrillation, and structural heart disease were excluded from the study. The mechanism of SVT was precisely established during EPS and then confirmed by the performance of a successful ablation.
The 12-lead ECGs were recorded during EPS using an electrophysiological recording system (EP-Tracer, CardioTek, Maastricht, The Netherlands and BARD Lab System, Lowell, MA). All ECGs were recorded at a paper speed of 25 or 50 mm/s with a gain of 10 mm/mV. Standard filter settings of 0.5 and 1000 Hz were used. A non-preexcitadat SR or SVT was recorded for at least 10 s or 10 beats at the beginning of every ECG.
Each ECG was analyzed for the presence of the standard and new electrocardiographic criteria. The standard criteria were: the presence of retrograde P waves wave in the ST segment in !1 lead (Fig. 1); T wave inversion defined as a change in polarization of a T wave for !1 mm in !1 lead (Fig. 1); QRS alternans defined as a beat-to-beat oscillation in QRS amplitude for !1 mm in !1 lead during SVT; horizontal or upsloping ST-segment depression !2 mm for !80 ms at the J point in leads II, III, aVF, or V1-V6 (Fig. 1); horizontal or upsloping ST-segment elevation !1 mm at the J point or downsloping ST-segment elevation !1.5 mm at the J point for !80 ms in lead aVR (Fig. 1); pseudo r 0 wave defined as an r 0 wave in V1 present during SVT but absent during the SR (Fig. 2); and the presence of pseudo s 0 wave defined as an s wave present during SVT but absent during the SR in leads II, III, or aVF (Fig. 2). The new criteria were the presence of a pseudo r 0 wave defined as an r 0 wave in lead aVR present during SVT but absent during the SR (Fig. 2) and an aVL notch defined as a positive deflection at the end of a QRS complex in lead aVL that is present during SVT but absent during SR (Fig. 2).   ). All researchers were blinded to the mechanism of SVT and the ablation results. The researchers were asked to specify the most likely mechanism of SVT for each ECG based on their clinical judgment. Data from 2 adult cardiologists (SMS and PK) were included into the final evaluation of the diagnostic performance of ECG criteria. Although age and sex were not ECG criteria those data were included in the evaluation as basic patient information.

Statistical Methods
Continuous variables were presented as means and standard deviations. Categorical variables were presented as counts and percentages. The equality of 2 independent proportions was tested using exact Chi-square test. The performance of binary decision rules was described using diagnostic accuracy, sensitivity, specificity, positive predictive value, negative predictive value, likelihood ratio of a positive test, and likelihood ratio of a negative test (with 95% confidence intervals). The ''multivariable decision rule'' was developed using logistic regression. Statistical analysis was performed using a ''BDTcomparator'' program and R 3.0 9,10. P-values <0.05 were considered to indicate statistically significant results.
In-depth statistical analysis was performer only on the ECGs of patients with AVNRT and OAVRT as they are the most common mechanisms of SVT in the pediatric population. 9,10

RESULTS
Of the 212 ECGs showing SVT, 97 (46%) were due to AVNRT and 115 (54%) were due to OAVRT. Forty-eight percent of the patients were female, and the mean age was 13.2 AE 3.5 years (range 1-18 years). In the AVNRT group, 54% of the patients was female (mean age, 13.6 AE 3.3) and 44% of the patients were female in the OAVRT group (mean age, 12.6 AE 3.8). In both groups gender were not statistically significantly different (P ¼ 0.188). In the AT group, 43% of the patients were female (mean age, 14.9 AE 1.5). Mean heart rate (HR) in the OAVRT group was 191 AE 25.9 beats per minute (BPM), in the AVNRT group 189 AE 33.8 BPM and in the AT group 187 AE 49.9 BPM.
Of the 7 standard criteria evaluated for the differential diagnosis of SVT, 1 were found significantly more often in the AVNRT group than in the OAVRT group (pseudo r 0 wave in V1) and 2 were found significantly more often in the OAVRT group than in the AVNRT group (retrograde P waves, and STsegment depression) ( Table 2). Retrograde P waves were found in 63% of the patients in the OAVRT group and 11% in the AVNRT group, ST-segment depression was found in 42% of the patients in the OAVRT group and 27% in the AVNRT group, and pseudo r 0 wave in V1 were found in 10% of the patients in the OAVRT group and 39% of the AVNRT group. The other standard criteria were not valuable for differentiating between AVNRT and OAVRT in children as they are in adult (Table 1). [4][5][6][7][8] The number of new criteria found in each group was similar and show no statistical differences. The occurrence of a pseudo s 0 wave was similar in the SR and SVT, found in 69% of AVNRT cases and 70% of OAVRT cases. Conversely, an r 0 wave in the aVR lead was present in 83% of AVNRT patients and 77% of OAVRT patients in the SR. Table 2 shows the diagnostic accuracy, sensitivity, specificity, positive predictive value, negative predictive value, likelihood ratio of a positive test and likelihood ratio of a negative test for all ECG criteria as well as age (age <11 vs age ¼>11), gender (male vs female), HR (HR between 160 and 230 vs HR below160 or above 230 BPM), and the multivariable decision rule.
Among the reviewers, the highest percentage of correct diagnoses of the SVT mechanism based on ECG criteria was approximately 60% in the OAVRT group obtained by SMS, AM and MG, and 73% in the AVNRT group obtained by PK. The rest of the reviewers made correct diagnoses approximately

DISCUSSION
To our knowledge, this is the largest study, in terms of the number of patients and reviewers, to identify the criteria associated with the differential diagnosis of SVT in the pediatric population. Moreover, a clinically useful algorithm was developed to provide practicioners with simple and accurate method for the differential diagnosis of AVNRT and OAVRT.
This study demonstrated that 3 of the 7 standard criteria (retrograde P waves, ST-segment depression, and r 0 wave in V1) are valuable for distinguishing AVNRT-associated SVT from OAVRT-associated SVT in a pediatric population. Differences between children and adults such as smaller chest size, right ventricular domination (especially in the youngest patients), fast retrograde activation via the fast pathway, and a smaller distance between chest structures may be responsible for the low value of ECG criteria in children compared to adults. 11 Results of more than 1 point indicate OAVRT and results of 1 fewer points indicate AVNRT. Retrograde P waves, the presence of retrograde P waves in !1 lead; ST-segment depression, horizontal or upsloping ST-segment depression !2 mm for !80 ms at the J point in leads II, III, aVF, or V1-V6; r 0 wave in V1, pseudo r 0 wave defined as an r 0 wave in V1 present during the SVT but absent during the sinus rhythm. AVNRT ¼ atrioventricular nodal reentry tachycardia, BPM ¼ beats per minute, HR ¼ heart rate, OAVRT ¼ orthodromic atrioventricular reentry tachycardia.   Using the ECG criteria in the setting of the clinical examination and patient history might increase the odds of correctly diagnosing the mechanism of SVT in children using ECG criteria. In addition, the use of noninvasive methods of ECG monitoring over a prolonged period, such as a Holter monitor or tele-ECG, might increase the odds of making a proper diagnosis. A previous study of ECG criteria in a pediatric population found that retrograde P waves and ST-segment elevation in aVR were found more frequently in OAVRT, and a pseudo r 0 wave in V1 was found mostly in AVNRT. This study also showed the high value of ST-segment depression for the differential diagnosis of SVT, which was not noted in previous studies. These different results may have been due to the larger test group and lower average mean age 1 ; however, none of these criteria had the same sensitivity and specificity reported in the adults. [2][3][4][5][6][7][8] The low prevalence of the pseudo s 0 wave criteria in this study coincided with a high prevalence of s waves in both the SR and SVT (Figs. 1 and 2). This was not noted in previous studies of this kind in a pediatric population. 1 Similar findings were seen with the r 0 wave in aVR criteria (Fig. 1); however, this is the first study evaluating these criteria in a pediatric population.
Of all the possible standard and new criteria for the differential diagnosis of SVT, only 3 criteria were useful when only the ECG during tachycardia was used to diagnose the mechanism of SVT. These criteria are retrograde P waves, STsegment elevation in aVR, and ST-segment depression. Most of them were accurate for the identification of OAVRT, and they might be useful in situations when SVT onset is acute in a nonmedical setting.
In both the OAVRT and AVNRT groups, the reviewers in our study missed more diagnoses than in previous adult and pediatric studies. 1,2,6 Compared to the results of previous studies in pediatric populations, our results might be more authentic due to our study having a larger study population. The missed diagnosis of AT might be because it is less common than AVNRT and OAVRT, and there are no specific 12-lead ECG criteria to identify AT when a 1:1 atrial to ventricular ratio is present.
Recent study in the differential diagnosis of wide QRS tachycardia shows that using a scoring system increases the chance that a reviever will correctly diagnose arrhythmia compared to algorithms systems, suggesting that diagnosis based on our JUNIOR SVT SCORE will be more reliable than previously reported algorithms. 1,12 Differentiation between AVNRT and OAVRT based on ECG before performing CA might allow electrophisologist to prepare for different approaches, and reduce cost of the additional catheters and resources. Experience with simplified 2-catheter procedure with a non-fluoroscopic approach have reduces the need for lead-apron for medical staff and fluoroscopic exposure to less than 5% of cases with AVNRT. 13 Transseptal puncture in left-sided OAVRT may require more expertise in retrograde access or transseptal puncture, and some centers will prefer cryoablation for AVNRT.
Our study has several limitations. First, although we investigated the standard and new criteria for the differential diagnosis of SVT, we only investigated 2 of the possible mechanisms of SVT (AVNRT and OAVRT), excluding other mechanisms such as AT due to the low number of cases. Only short episodes of SVT (usually within the first minute of tachycardia induction) recorded in the EP lab was evaluated.

CONCLUSIONS
Visible P waves in the ST segment, ST-segment depression in II, III, aVF, and V1-V6, as well as pseudo-r 0 -waves in lead V1 are valuable for the differential diagnosis of SVT by ECG in a pediatric population; however, the new ECG criteria showed only limited diagnostic ability in the pediatric population. Unfortunately, the overall diagnostic ability of all ECG criteria for determining the mechanism of SVT in children and adolescents was very low despite the use of an adult cardiologist, pediatric cardiologist, and medical students as reviewers. The need for a simple, accurate tool to determine the differential diagnosis remains. In addition, an algorithm for the diagnosis of SVT in children and adolescents is yet to be developed. Based on the results of this study, a multivariable decision rule (JUNIOR SVT SCORE) for the differential diagnosis of SVT in a pediatric population will be validated.