Abbreviations
AI – Ablation Index
AF – atrial fibrillation
CTI – cavotricuspid isthmus
ID – impedance drop
Ω - Ohms
PVI – pulmonary vein isolation
LVEF – left ventricular ejection fraction
Wgs – Watts/grams/second
Introduction
The advent of contact force-sensing catheters has delivered important data on ablation lesion size, safety and efficacy, and their application is well-described during pulmonary vein isolation (PVI) for atrial fibrillation (AF) . Using a weighted formula, catheter contact force (g) can be combined with the duration (s) and power (W) of a radiofrequency application to calculate Ablation Index (AI). Expressed as a continuous value in Wgs, AI has been shown to predict lesion diameter and depth during AF ablation , and left atrial procedures guided by site-specific AI targets have demonstrated more enduring PVI and a comparable safety profile versus conventional ablation techniques . AI infers the energy delivered by the ablation catheter; this differs from impedance drop (ID) which describes local impedance changes at the blood-tissue interface and infers tissue receipt of injury. Whilst AI is known to correlate with ID, the strength of this relationship attenuates with procedural variables such as catheter angle of incidence and irrigation techniques; as such, there may be important limitations when relying on AI alone to guide lesion delivery .
Radiofrequency ablation of cavo-tricuspid isthmus (CTI) dependent (‘typical’) atrial flutter achieves acute success (i.e. bidirectional block) in over 90% of cases, and carries a class IA recommendation as a treatment strategy in the ESC’s 2019 Supraventricular Tachycardia (SVT) guidelines . Despite this, a significant proportion of CTI ablations can prove technically challenging, and hence novel predictors of acute and long-term efficacy remain desirable. Autopsy studies have demonstrated marked heterogeneity in CTI architecture; Klimek-Piotrowska et al. (2016) dissected 140 human hearts and found that, when compared to the anterior or posterior margins, the middle CTI frequently harbours distinct morphological variations such as trabeculae (62.1%) or recesses (25%) . Peri-ablation imaging data has also shown that structural anomalies – such as the presence of pouches, angular crypts or tricuspid regurgitation – result in prolonged procedure times and poorer outcomes . Accordingly, data from contact-force sensing catheters suggests that site-specific inconsistencies in isthmus tissue contact may be responsible for procedural failure . The prevalence of complex CTI anatomy has led some authors to suggest that pre-procedural imaging, such as cardiac MRI or right atrial angiography, would promote more patient-specific ablation strategies and hence improve outcomes .
Hypothesis
We hypothesised that AI may provide important insights into lesion delivery across the CTI during ablation of typical atrial flutter. Using established 3D electro-anatomical mapping systems and ablation catheters, examining local variations in ablation lesion characteristics may elucidate the mechanisms which impede enduring bidirectional block or contribute to complications. An appreciation of these relationships may promote a more prescriptive approach to CTI ablation without the use of additional resources such as peri- or intraprocedural imaging techniques.
Methods
Ethics
This project was registered with the local clinical effectiveness unit. Consenting patients underwent procedures which were clinically indicated, without randomisation or allocation, and which made use well-established mapping and ablation techniques. As such, the work was consistent with Clinical Service Development in line with the UK’s Health Research Authority recommendations, and no specific additional ethical approval was required.
Procedure
Patients with no previous CTI ablation underwent radiofrequency ablation of CTI-dependent atrial flutter at two tertiary Cardiothoracic hospitals in the UK from 2019-2020. Procedures were performed on uninterrupted oral anticoagulation, either under general anaesthetic or with conscious intravenous sedation. Ultrasound-guided femoral venepuncture was performed following local anaesthetic administration, and a quadripolar or decapolar catheter was positioned in the coronary sinus. A contact-force sensing ablation catheter (Thermocool SmartTouchTM, Biosense Webster, Diamond Bar, California) was passed to the right atrium. Electro-anatomical mapping was performed using CARTO software (v3, Biosense Webster). For patients presenting in atrial flutter, entrainment discerned CTI dependence, whereas patients in sinus rhythm underwent empirical CTI ablation during pacing of the proximal coronary sinus at 600ms. Point-by-point ablation was delivered from the ventricular margin of the CTI progressing towards the inferior vena cava, maintaining a 6 o’clock alignment in the left anterior oblique projection as per standard clinical protocol. The use of additional sheaths to guide ablation was at the operators’ discretion. Ablation VisiTagTM settings were pre-specified to accept 5mm of catheter drift, and force-over-time (FOT) constraints of 5s, 25% in conjunction with 3g minimum force. Operators were advised to deliver ablation lesions with a peak AI of 600Wgs at 45-50W. These recommendations were based on a retrospective analysis performed at our institution which demonstrated safe and effective CTI ablation with clusters of lesions in this range of AI (18). In the event of visual macro-displacement, locations were discarded. Saline irrigation flow rate was 2ml/min during mapping and 17ml/min during ablation. Impedance was measured between the catheter tip and a ground patch on the patient’s right thigh. If bidirectional CTI block was achieved, this was reassessed after 15 minutes’ observation and consolidative ablation lesions delivered as required. In the absence of complications, patients were discharged the same day if the procedure was performed under sedation, or the following day if performed under general anaesthetic.
Data extraction
Post hoc, VisiTags were anatomically trisected according to their position on the CTI: IVC end (IVC), middle CTI (Mid), or ventricular end (V). Lesion characteristics were subsequently extracted and aligned with time stamps to allow assessment of temporal changes, including contact force (every 50ms), impedance, impedance drop and ablation index (every 10-20ms), catheter angle in the axial and lateral planes (every 10-20ms), and power (every 100ms). It has been demonstrated that precipitous rises in impedance are associated with thrombus and steam formation, and a variety of methods for assessing change in impedance during catheter ablation have been published previously, including total impedance drop, overall median impedance drop, or median impedance drop after 10 seconds of ablation (2) (9) (19). During our data cleaning, very rare transient spikes and troughs (lasting <50ms) in impedance – and consequently the running calculations of impedance drop – were noted. The examination of other contemporaneous lesion characteristics (e.g. contact force) suggested that these data points were real, and as their ramifications on ablation safety were potentially significant, they were included in further analyses. However, these outlying values rendered the measurement of total impedance drop less reliable, and so a mean ID was instead calculated from all the values recorded throughout the duration of each radiofrequency application. The mean ID was assigned as a surrogate marker of ablation efficacy. Total energy delivery was assessed in terms of peak ablation index; this was defined as the maximum recorded AI (Wgs) measured during each lesion; for all VisiTags, this corresponded to the final recorded value.
Follow-up
Patients underwent clinical review with ECG analysis at 3 months post procedure, with a further review at 12 months or sooner if necessitated by symptoms. Anticoagulation was continued according to CHA2DS2-VASc score, and anti-arrhythmic drugs were adjusted according to patient preference and physicians’ discretion.
Statistical analysis
Data were analysed in R (x64, v. 3.5.2.) and XLSTAT (Addinsoft, v. 2020.1). Categorical group parameters were compared using Chi square tests. The Shapiro-Wilk test identified whether or not data were normally distributed. Subsequently, continuous data were compared with two-tailed T tests or analysis of variance (ANOVA) with post-hoc Tukey HSD testing for normally distributed data, or with the Mann-Whitney U or Kruskal-Wallis test for non-normally distributed data. Correlation was assessed using Pearson’s coefficient for normally distributed data or the Spearman rank correlation if non-normally distributed. Data are presented as mean ± SD or median (interquartile range). The level of significance was set at p<0.05.
Results
38 individuals were included in this study; clinical characteristics and baseline ablation data are shown in table 1. 14 cases made use of additional long sheaths (11 Swartz SR0, Abbott/St Jude Medical; two AgilisTM, Abbott/St Jude Medical, one VizigoTM, Biosense Webster) to assist ablation. Acute success (bidirectional block) was achieved in all 38 cases (100%). There were no complications, and no steam pops were recorded. After a mean follow-up of 6.6 ± 3.3 months, 97.4% (n=37) of patients were in sinus rhythm.
Table 1: clinical characteristics and baseline ablation data