Catheter Ablation -> Ventricular Arrhythmias -> Quality Measures & Complications D-PO06 - Poster Session VI (ID 26) Poster

D-PO06-115 - Poor Inter- And Intra-rater Agreement In The Praetorian Score For Subcutaneous Defibrillator Implantation (ID 674)


Background: Subcutaneous defibrillator (S-ICD) implantation generally requires VF conversion testing (CT). Proximity of the coil and generator to the thoracic cage and posterior position of the generator are essential for high shock efficacy. The PRAETORIAN score (PS) consists of 3 parameters measured on post-implant chest X-ray (CXR). Some experts recommend forgoing CT when coil/generator are optimally placed with the PS < 90, the value known to correlate with high CT success, and the ongoing PRAETORIAN-DFT trial is assessing this hypothesis. However, inter- and intra-rater agreement of the PS calculation have not been previously evaluated.
Objective: Assess the inter- and intra-rater agreement in the PRAETORIAN score
Methods: Thirteen cardiac electrophysiologists [EPs] and 13 nurses [RNs] individually calculated PS on 20 sets of posterior-anterior and lateral post-implant CXRs for 18 patients. Two sets of CXRs were included twice to assess intra-rater agreement. The median PS rated by 13 EPs for each CXR was used as “gold standard.” Inter-class R (ICR) statistic values > 0.75 are considered excellent agreement whereas values < 0.4 indicate poor agreement. PS agreement was assessed by fitting a mixed model to the PS, treating each CXR as a subject and rater as a random effect. Inter-rater agreement analysis was performed for all 18 S-ICDs and those with optimal lead/generator placement (PS < 90), respectively.
Results: Median PS rated by EPs for each patient ranged 30 to 200 (median: 60). Two out of 18 S-ICD (11%) placement was suboptimal based on the median PS > 90. Inter-rater agreement of the PS for 18 S-ICD placement was poor among all raters (ICR=0.008), EPs (ICR=0.0095), and RNs (ICR=0.199). Inter-rater agreement did not improve when only the 16 optimally placed S-ICDs (PS < 90) were included: ICRs for all raters, EPs, and RNs were 0.0015, 0.0033, and 0.0017, respectively. Only 6 EPs (46%) and 4 RNs (31%) rated all 3 components of the PS for duplicate CXRs identically, consistent with poor intra-rater agreement.
Conclusion: Both inter- and intra-rater agreement of the PS are poor, and significant variability was observed both among EPs and RNs. PS-based algorithm to forgo CT appears unreliable primarily due to inconsistencies in the PS calculation on post-implant CXRs.