The use of intracardiac echocardiography catheters in endocardial ablation of cardiac arrhythmia: Meta‐analysis of efficiency, effectiveness, and safety outcomes

Abstract Aims The optimal use of intracardiac echocardiography (ICE) may reduce fluoroscopy time and procedural complications during endocardial ablation of cardiac arrhythmias. Due to limited evidence in this area, we conducted the first systematic literature review and meta‐analysis to evaluate outcomes associated with the use of ICE. Methods and Results Studies reporting the use of ICE during ablation procedures vs without ICE were searched using PubMed/MEDLINE. A meta‐analysis was performed on the 19 studies (2186 patients) meeting inclusion criteria, collectively representing a broad range of arrhythmia mechanisms. Use of ICE was associated with significant reductions in fluoroscopy time (Hedges' g −1.06; 95% confidence interval [CI] −1.81 to −0.32; P < .01), fluoroscopy dose (Hedges' g −1.27; 95% CI −1.91 to −0.62; P < .01), and procedure time (Hedges' g −0.35; 95% CI −0.64 to −0.05; P = .02) vs ablation without ICE. A 6.95 minute reduction in fluoroscopy time and a 15.2 minute reduction in procedure time was observed between the ICE vs non‐ICE groups. These efficiency gains were not associated with any decreased effectiveness or safety. Sensitivity analyses limiting studies to an atrial fibrillation (AF) only population yielded similar results to the main analysis. Conclusion The use of ICE in the ablation of cardiac arrhythmias is associated with significantly lower fluoroscopy time, fluoroscopy dose, and shorter procedure time vs ablation without ICE. These efficiency improvements did not compromise the clinical effectiveness or safety of the procedure.

procedures may also require prolonged fluoroscopic guidance, which exposes the patient, operator, and laboratory staff to significant levels of radiation. 3 Estimates suggest that an average fluoroscopy time of 1 hour during ablation increases a patient's lifetime risk of fatal cancer by up to 0.1%. 4 It has been reported that cardiac electrophysiologists have an annual ionizing radiation exposure two to three times higher than that of diagnostic radiologists, which translates to a cumulative risk of one additional cancer diagnosis for every 100 exposed individuals following a full professional career. 5,6 Preprocedural computed tomography (CT) or magnetic resonance imaging (MRI) can be used to visualize the location of the left atrium, pulmonary veins, and surrounding structures before ablation. Preprocedural images can also be merged into three-dimensional electroanatomical mapping (EAM) systems to visualize the esophagus and its location along the posterior wall of the left atrium, and reduce radiation exposure during the ablation procedure. 7 However, because of changes in patient position, esophagus location, and/or cardiac rhythm, use of CT or MRI images obtained pre-procedurally to guide catheter manipulation may not accurately reflect the actual cardiac anatomy. [8][9][10] Despite advances in EAM and catheter technology, most complex procedures still utilize prolonged fluoroscopic guidance which exposes the patient, operator, and laboratory staff to radiation. 3 Intracardiac echocardiography (ICE) has multiple real-time applications during catheter ablation, including the ability to guide transseptal puncture, visualize the location of the esophagus, provide guidance of cardiac anatomy, and detect microbubbles as a result of overheating, thereby preventing complications. 11,12 Optimal use of ICE may reduce fluoroscopy time and procedural complications; however, there is limited evidence comparing the use of ICE with procedures dependent on fluoroscopy or other imaging modalities to procedures that do not use ICE. Published evidence comparing ablation with ICE vs ablation without ICE have also not been analyzed in a meta-analysis comparing clinical endpoints to date. Therefore, a systematic literature review and metaanalysis were conducted to evaluate the use of ICE for real-time imaging during endocardial ablation of various cardiac arrhythmias.

| METHODS
We conducted a systematic literature review following typical best practices, including the use of a prospective protocol specifying search terms and study eligibility. 13 A statistical analysis plan was also prepared a priori.

| Study eligibility
Studies were screened for eligibility using a two-step process and prespecified criteria. During "level I" review, studies were screened for eligibility by a single reviewer, based on their title and abstract.
Potentially eligible studies with an eligible patient population and some indication that ICE was evaluated were reviewed in full text.
During "level II" screening, full-text articles were reviewed for use of the technology of interest and for eligible comparative data. At level II screening, two reviewers independently assessed each study for fit with the selection criteria.
Included studies were required to be in English-language, comparative (randomized or nonrandomized, including retrospective comparisons), and with at least 10 patients undergoing endocardial ablation with either sensor-based (SOUNDSTAR® Catheter, Biosense Webster, Inc) or non-sensor-based (eg, AcuNav TM Catheter, Biosense Webster, Inc; Ultra ICE TM Catheter, Boston Scientific; or ViewFlex TM Catheter, St. Jude Medical) catheters compared to each other, or to ablation procedures without the use of ultrasound. We deliberately included reports of ablation for any form of cardiac arrhythmia to broadly assess the use of ICE according to real-world clinical practice.
Studies were excluded if they did not include a procedure of interest, did not use real-time guidance during ablation (eg, preprocedural evaluation of left atrial appendage thrombus only), or did not report any outcomes of interest (ie, extractable data for ICE and comparator for at least one efficiency, effectiveness, or safety outcome).

| Data extraction and quality assessment
For the preparation of the meta-analysis data set, all data elements were extracted by one reviewer and confirmed by a second reviewer.
Key study, patient, and treatment characteristics were captured from each study using a standard template, and studies were assessed for quality using the Oxford Level of Evidence Centre for Evidence-Based Medicine (CEBM) level of evidence. 14 Randomized clinical trials (RCTs) were considered level 2, except where the ICE comparison was confounded by inclusion of some use of ICE in the control arm or by other imaging techniques which varied between groups. These "downgraded" RCTs were considered level 3 evidence, along with prospective non-randomized comparative studies. Retrospective non-randomized comparative studies were considered level 4 evidence. Efficiency, effectiveness, and safety outcomes were extracted from all eligible studies, as available. Safety events were divided into venous access vs all other peri-procedural complications, to isolate the setup-related outcomes from those associated with the ablation itself.

| Data synthesis and analysis
The primary endpoint for the meta-analysis was fluoroscopy time.
Secondary outcomes were fluoroscopy dose, procedure time, acute procedure success, peri-procedural complications, and freedom from arrhythmia at 6 months follow-up or longer.

| Outcome study measures
Intention-to-treat results were extracted for all binary outcomes, with between-group effect sizes compared using a risk ratio (RR).
Continuous and time outcomes were compared using standardized mean difference (with Hedges' g adjustment). The standardized mean difference allows the analysis of studies assessing the same outcome, but with different outcome definitions, data reporting formats, or measurement scales. Absolute Hedges' g values <0.2 indicate a small effect, 0.5 a medium effect, and >0.8 a large effect. 15 For this study, negative Hedges' g values are desirable. Mean difference (MD; in minutes) analysis was also performed on fluoroscopy and procedure time outcomes. In studies with three arms, we chose the two arms with the cleanest ICE vs non-ICE comparison (eg, a group including ICE for imaging during the ablation procedure vs a group without ICE, but with otherwise similar mapping and ablation techniques).
Since the choice was not always clear-cut, we also performed a sensitivity analysis using the alternate comparator arm.

| Analyses
Because heterogeneity was expected among studies, the main analyses were performed with random-effects inverse variance weighting models, as recommended by Fleiss et al 16 [18][19][20] The DL method is currently considered the standard for random-effects models. However, the HKSJ method has been found to provide more consistently adequate error rates, especially when the number of studies is small and there is moderate or substantial heterogeneity, and thus is increasingly accepted as a more appropriate method. 21 Since the acute procedure success and peri-procedural complications outcomes showed sparse data for events or non-events, they were estimated using a randomeffects Mantel-Haenszel (MH) risk ratio.
Sensitivity analyses were performed to understand the robustness of the initial effect size estimates and to assess potential sources of heterogeneity among studies. When data permitted, the following sensitivity analyses were conducted: effect size calculation using the HKSJ method, removal of outlying and statistically influential studies, imputation of non-reported study means and/or standard deviations (SDs), and use of an alternate comparator arm. For fluoroscopy time, sensitivity associated with the use of a sensor-based ICE catheter (SOUNDSTAR® Catheter) was also performed. In particular, extensive post-hoc sensitivity analyses were conducted for studies restricted to an AF population across all endpoints of interest, after our systematic review revealed that a majority of reports focused on this arrhythmia type. Doing so eliminated all studies with a pediatric cohort, as well as those with substrates distinctive from AF.
Data manipulation and statistical analyses were performed using SAS Software, Version 9.4 and the R meta-package, Version 4.9-4.
All P-values were two-sided, and values ≤.05 were considered statistically significant.

| RESULTS
Our initial search retrieved 1349 articles ( Figure 1). After exclusion by title and abstract ("Level I" screening), 101 potentially eligible studies were reviewed in full text. Eighty-two articles were excluded F I G U R E 1 Flow diagram of the stages of the systematic literature search. ICE, intracardiac echocardiography during the "Level II" screening. The most common reasons for exclusion were studies that did not use ICE during the procedure or did not include an ICE vs no ICE comparison.

| Procedure time
Data for procedure time were available from 14 studies, with significant heterogeneity among studies (I 2 = 86%) and no evidence of publication bias or small study effects. 10 Figure S2). These results were consistent in a sensitivity analysis restricting studies to an AF population (Hedges' g −0.43; 95% CI −0.74 to −0.13; P < .01). Other sensitivity analyses including use of the HKSJ method (P = .10), removal of one outlying and influential study (P = .09), 27 inclusion of two studies with imputed means and/or SD values (P = .09), 23,34 and use of an alternate group for 3-arm studies (P = .07), 10,11,29 were directionally similar but failed to reach statistical significance (Table S4).
3.3 | Effectiveness outcomes: acute success and freedom from arrhythmia

| Acute success
Thirteen studies were included in the analysis, 10

| Freedom from arrhythmia
Eleven studies reported freedom from arrhythmia outcomes, with no observable heterogeneity among studies (I 2 = 0%). 10,11,[24][25][26][27]30,31,33,37,38 All studies for this outcome were in the setting of AF. The use of ICE was not associated with a change in freedom from arrhythmia compared to ablation without the use of ICE (RR 1.04; 95% CI, 0.97-1.11, P = .24) ( Figure S3B). No outlying or influential studies were identified to include in a sensitivity analysis. Sensitivity analyses using an alternate group for the 3-arm study 10 (P = .33) and using the HKSJ difference method (P = .23) did not demonstrate any significant differences in freedom from arrhythmia between groups (Table S5).

| Safety outcome: peri-procedural complications (excluding venous access)
Six studies reported zero (0) peri-procedural complications in one or both arms 10 95% CI, 0.81-4.60; P = .14) ( Figure S3D). No studies with an alternate group for 3-arm studies were available for sensitivity analysis, and limiting reports to an AF population yielded nonsignificant results consistent with the main analysis (P = .21) (Table S6). 10 that radiation dose must be as low as reasonably achievable (ALARA). 44 Certain patient groups have been identified that are more vulnerable to radiation risks. For example, obese patients were shown to require nearly three times the amount of radiation exposure than what is required for nonobese patients. 45 Radiation risks are also higher for children and pregnant women. 46,47 Operators and staff who perform many ablation procedures over time accumulate significant exposure to radiation; therefore, these individuals are also highly susceptible to the risks associated with heavy fluoroscopy use, which may be even greater than that experienced by patients. 48 Brain tumors, breast cancer in female cardiologists, and cataracts have been reported in interventional cardiologists and staff potentially due to their increased radiation exposure. [49][50][51] In addition, a higher prevalence of orthopedic injuries have been reported in interventional cardiologists, who wear lead apparel to protect themselves from radiation, compared to noninterventional cardiologists. 52 Studies indicate that over one-third of GOYA ET AL. In this study improved efficiency (ie, reductions in fluoroscopy and procedure times) using ICE did not compromise effectiveness. However, the difference between groups was not statistically significant (RR 1.93, P = 0.14). In our meta-analysis, peri-procedural complications were rare, with many studies reporting no events in one or both comparator arms. Due to such rarity, it is difficult to definitively assess a statistical difference in the rate of complications among procedures that use ICE vs those that do not. In a 2013 systematic review examining complications of catheter ablation in AF, the incidence of peri-procedural complications was reported at 2.9%. 55 Overall, sensitivity analyses of outcomes limited to AF ablation were consistent with findings from the main analyses, which broadly However, it is difficult to generalize the determination of "costeffectiveness" from country to country. Each healthcare system is subject to diverse perspectives of value, influenced by current cultural and political factors, existing healthcare financing models, level of economic development, and local unmet medical needs.