051 Organisational index is the best assessment of fractionation in persistent atrial fibrillation: analytic comparisons and ablation results. (16th May 2012)

• Record Type: Journal Article
• Authors: Finlay, M C; Lim, B; McCready, J; Ahsan, S; Gopalmurugan, A B; Xu, L; Segal, O; Lowe, M; Lambiase, P D
• Is Part Of: Heart. Volume 98(2012)Supplement 1
• Page Start: A31
• Page End: A32
• Publication Date: 2012-05-16
• Subjects: Atrial fibrillation -- catheter ablation -- fractionation; Heart -- Diseases -- Treatment -- Periodicals; Cardiology -- Periodicals; 616.12
• DOI: 10.1136/heartjnl-2012-301877b.51
• Languages: English
• ISSNs: 1355-6037

Abstract

Abstract : Introduction: Targeting fractionated electrograms (EGMs) is commonly used in catheter ablation (CA) of persistent atrial fibrillation (PersAF). Several automated algorithms exist to identify sites of fractionation. EGMs exhibiting continuous activity (ContA) and a high organisation index (OI) have also been proposed as features of significant sites. We investigated five indices of fractionation to assess the consistency of signal classification, and associated the effects of ablation lesions to these classification techniques. Methods: 16 patients undergoing first-time CA of persAF were studied. A NavX CFAE map was acquired after circumferential pulmonary vein isolation. EGM data were exported and analysed in custom software. CA was performed on sites showing high ContA. Offline analysis compared 5 algorithms: ContA, OI, Dominant Frequency (DF), CFAEmean (NavX equivalent) and shortest complex interval (SCI, Carto equivalent) after signal screening. The effect of ablation lesions on AF cycle length was correlated with underlying fractionation. Results: CA on ContA sites terminated AF in 50% cases. 2589 8-s EGMs and 471 ablation lesions targeting fractionation were analysed. Varying the refractory periods (RP) of algorithms produced large changes in signal classification for CFAEmean (Pearson R=0.81±0.05 for 10 ms variation in RP, R=0.08 for 90 ms variation in RP) and SCI (R=0.95±0.04 for 10 ms, R=−0.34 for 90 ms), whereas ContA was unaffected (R=0.99±0.01 for 10 ms, R=0.80 for 90 ms). There was a modest inverse correlation between DF and OI of EGMs (R=−0.58, p<0.001). There was no correlation between these indices & the same EGMs classified according to other measures (OI vs CFAEmean, R<0.01, OI vs SCI, R=−0.11, OI vs ContA, R=−0.06, all p>0.5, similar for DF). Only 2±1% of EGMs were in the top quintile of fractionation by all 3 indices of CFEmean, ContA and SCI. Only high OI (ROC AUC=0.64 p<0.01) at ablation site and lesion number (p<0.01) predicted an increase in AF cycle length. The image shows a typical left atrial map following analysis. Yellow areas represent 20% most fractionated by OI, blue represents CFAEmean, Red indicates areas of both high OI and CFAEmean. Lesions are shown, green if AFCL increased with lesion and red if AFCL was unchanged. Conclusions: The classification of fractionated EGMs is very sensitive to user-selected characteristics. Interpretation of results of "fractionation maps" must take into account analysis techniques, OI appears the most promising at guiding appropriate ablation site.