Five Criteria Predict Induction and Ablation of Supraventricular Tachycardia

Figures

• 

  Figure 1. Patient cohorts and overall results. (A) Derivation cohort. (B) Validation cohort. AF, atrial fibrillation; AFL, atrial flutter; AP, accessory pathway; DAVNNRT, dual AV nodal nonreentrant tachycardia; PAC, premature atrial contraction; PVC, premature ventricular contraction; SVT, supraventricular tachycardia; VT, ventricular tachycardia; VF, ventricular fibrillation.

• 

  Figure 2. Typical and non‐typical SVT EKG patterns. (A–D) Typical SVT EKGs. (E–J) non‐typical SVT EKGs. Sinus rhythm (SR) complexes from the same lead are shown for comparison. Arrowheads indicate P waves.

• 

  Figure 3. Preoperative factors predicting SVT induction and ablation. Patients who exhibited each feature were compared to all patients who did not, including those with no data, except as noted below. The RR values, 95% confidence limits, and p values were calculated using Poisson regression. Asympt, Asymptomatic; CI, Confidence Interval, Excl, Excluding; RR, Risk Ratio; Sxx, Symptoms; Sympt, Symptomatic; Term, Termination. (A) An EKG showing regular and uninterrupted SVT predicts SVT induction. (B) Termination with adenosine or vagal stimulation predicts SVT induction. ND, No Data. (C) Regular SVT episodes on monitoring predict SVT induction. Patients exhibiting each feature were compared to patients who had no episodes on monitoring. (D) Pre‐excitation predicts SVT induction after other variables are excluded. (+) factor indicates typical SVT EKG, termination with adenosine or vagal stimulation, or monitored event ≥ 30 s. Note change in scale. (E) Preoperative factors that do not independently predict SVT induction. (+) factor indicates typical SVT EKG, termination with adenosine or vagal stimulation, monitored event ≥ 30 s, or pre‐excitation. (F) Negative predictive factors are…

• 

  Figure 4. The duration of the longest episode on monitoring predicts SVT induction. (A) Patients with longer duration episodes are more likely to have SVT induced at EPS. Pastel lines show the results from different types of monitors. The solid black line shows the pooled results from all monitors. (B) Regular SVT events (black) tend to last longer than irregular events (gray). The duration of each patient's longest event is plotted. (C) The probability of SVT induction above each cutoff duration is plotted. The data points represent cutoff durations of 0, 5, 15, 30, 60, 120, 300, 600, 1200, and 3600 s. The open circle represents a cutoff of ≥ 30 s. When comparing the probability of SVT induction above and below the cutoff value, the p value was < 0.001 for all data points. (D) A cutoff duration of 30 s best distinguishes patients who will have SVT induced and ablated from those who will not. The vertical axis shows the area under the ROC curve for each cutoff duration. Black circles, SVT induction; red squares, SVT ablation. The open symbols represent a cutoff of ≥ 30 s.

• 

  Figure 5. The frequency of SVT episodes on monitoring does not predict SVT Induction. (A) The frequency of SVT events on monitoring does not differ between patients who have SVT induced (pink) and those who do not (blue). Note that the frequency of SVT events varies widely between different types of monitors (the vertical axis is a logarithmic scale). p values were calculated using Student's t‐test assuming unequal variances, excluding outliers that were > 3× the interquartile range from the median. (B) The probability of SVT induction does not increase at higher event rates. Pastel lines show the results from different types of monitors. The solid black line shows the pooled results from all monitors. The lower frequency range is dominated by Preventice monitors, the middle range by Ziopatch monitors, and the upper range by Holter monitors.

• 

  Figure 6. Scoring system for estimating the probability of SVT induction and ablation. (A) Each preoperative feature independently predicts SVT induction and ablation. The RR and p values for each feature were calculated after excluding patients who exhibited any of the other features. (B) Algorithm for estimating the probability of SVT induction or ablation. Points in the prediction score were calculated by dividing each regression coefficient (Supporting Information S1: Table S4) by the smallest coefficient, multiplying by 2 to be accurate within 0.5, and rounding to the nearest integer. (C) ROC curves showing the performance of the prediction score in predicting SVT induction (black circles) or ablation (red squares) in the derivation cohort. (D) Same as (C), validation cohort.