Author + information
- Received June 25, 2017
- Revision received September 12, 2017
- Accepted October 4, 2017
- Published online November 29, 2017.
- Duy T. Nguyen, MD∗ (, )
- Matthew Zipse, MD,
- Ryan T. Borne, MD,
- Lijun Zheng, MS,
- Wendy S. Tzou, MD and
- William H. Sauer, MD
- Section of Cardiac Electrophysiology, Division of Cardiology, University of Colorado, Aurora, Colorado
- ↵∗Address for correspondence:
Dr. Duy T. Nguyen, Section of Cardiac Electrophysiology, Division of Cardiology, University of Colorado, B-132, Leprino Building, 12401 East 17th Avenue, Aurora, Colorado 80045.
Objectives Given a paucity of data, the aim of this study was to define predictors of steam pops (SPs) during open-irrigated radiofrequency ablation (RFA).
Background SPs during RFA can lead to dire consequences, including perforation and stroke.
Methods In an ex vivo bovine myocardium model, open-irrigated RFA was applied at 50 W for 60 seconds; intracardiac echocardiographic images for RFA with and without SPs was compared. Using an in vivo porcine model, open-irrigated RFA was applied at 50 W for 60 seconds, and RFA parameters of SPs were analyzed. A retrospective analysis was performed of recorded SPs during clinical ablation procedures over a 1-year period.
Results For RFA SPs, there was 32% greater intracardiac echocardiographic tissue echogenicity than for RFA without SPs (p < 0.001). In addition, RFA SPs had more rapid increases of tissue echogenicity, particularly in the last 5 seconds before SPs. Compared with RFA without SPs, RFA SPs had larger impedance reductions (33.0 ± 16.0 Ω vs. 23.0 ± 10.8 Ω, p = 0.032). SPs were also associated with more rapid initial impedance reduction (1.40 Ω/s vs. 0.38 Ω/s for RFA without SPs, p = 0.001). Clinical SPs during ablation procedures had a significantly faster impedance reduction during the first 5 seconds of ablation compared with matched control ablations (15.7 ± 6.7 Ω vs. 8.1 ± 4.7 Ω, p < 0.0001).
Conclusions Certain echocardiographic and biophysical parameters during open-irrigated RFA are associated with increased SP risks. These include greater tissue echogenicity, larger total impedance reduction, rapid rate of initial impedance reduction, and rapid increase in tissue echogenicity.
Radiofrequency ablation (RFA) can result in unpredictable steam-generated explosions, or steam pops (SPs), that can lead to cardiac perforation. Very high tissue temperatures during RFA can lead to steam formation within the myocardium. If this vaporization leads to sufficient pressure buildup, gas is released as an SP.
SPs can lead to cardiac perforation and potential for microembolism. Seiler et al. (1) reported that when they used open-irrigated ablation for ventricular tachycardia, SPs occurred in 1.5% of ablation lesions, resulting in 1 cardiac perforation. Thus, as with any radiofrequency (RF) technology used to ablate myocardial tissues, there is a balance between achieving efficacy, when tissue destruction occurs at >50°C, and the potential for increased risk for SPs.
Predicting SPs would be clinically useful but remains a challenge. Various studies have found a correlation between impedance decreases during RFA with increased tissue heating and effective lesion formation (2); SP risk increased with large impedance decreases (1,3,4). However, the data for correlating impedance reduction with SPs remain equivocal, with 1 study finding no statistically significant difference in impedance reduction for ablations with and without SPs (5). Furthermore, most of these studies were performed with closed irrigation or were applied in the epicardium, and thus their relevance for endocardial open-irrigation RFA may be limited.
In this study, we sought to further characterize the RFA biophysical parameters that are associated with increased risk for SPs. Furthermore, we sought to establish imaging characteristics—specifically using a common tool used during RFA, intracardiac echocardiography (ICE)—that may correlate with an impending SP. Finally, we evaluated several strategies that can be used to minimize the risk for SPs.
Ex vivo model of SP generation during irrigated ablation
Experimental protocols have been approved by the Institutional Animal Care and Use Committees of the University of Colorado and conform to the National Institutes of Health’s Guide for the Care and Use of Laboratory Animals. A load cell, used to standardize application of energy and force, was submersed in a circulating bath of saline; the bath contained a section of viable bovine ventricular myocardium excised within 1 h of experimentation. An open-irrigated ablation catheter (Biosense Webster, Diamond Bar, California) was positioned with 10 g of force perpendicular to the myocardium using a deflectable sheath, and tissue contact was visually confirmed. This model has been validated and described in further detail elsewhere (6–8). RF energy was applied to tissue at 50 W for 60 s or until an SP occurred.
ICE was performed and recorded throughout the ablation using a phased-array ultrasound catheter (AcuNav, Siemens, Mountain View, California). Images were obtained at 6.0 MHz and at a depth of 3.8 cm. Intracardiac echocardiographic images for RFA resulting in SPs were compared with images for RFA without SPs. Change in echogenicity was recorded in time intervals leading to, but not including, the SP. Tissue echogenicity was calibrated to baseline images and quantified by measuring pixels using Photoshop CS software (Adobe, Mountain View, California). Each image during an ablation was normalized to its baseline image by taking the ratio of the image’s tissue density to the baseline image’s tissue density; multiplying this ratio by 100 yielded a unit of standardized brightness intensity (SBI). Rate of echogenicity change was calculated as SBI per second.
In vivo porcine ablation and SP generation
Yorkshire pigs were anesthetized, and porcine thighs were prepared bilaterally, modified from previously described canine thigh preparation (9,10). Briefly, skin and connective tissues were dissected to expose the underlying muscle. The skin was raised to form a cradle, and heparinized, warmed porcine blood was circulated at 350 ml/min.
An open-irrigated ablation catheter was placed perpendicular to the muscle surface, and tissue contact was visually confirmed. Ablations were delivered at 50 W for 60 s with the same amount of force, as measured by a force-sensing, open-irrigated-tip RF catheter (SmartTouch ThermoCool, Biosense-Webster); ablation lesions were tagged by the electroanatomic mapping system and averaged between 10 and 20 g of force. After animals were sacrificed, thigh preparations were resected, and ablation lesion sizes were measured for non-SP lesions.
In vivo porcine ablation modulation to minimize SPs
Two separate experiments were performed to assess ablation strategies that may decrease SP risks. First, an open-irrigated ablation catheter was placed perpendicular to thigh muscle surface. Ablations were delivered starting at 30 W and increasing in 10-W increments every 10 s, reaching 50 W maximum, for 60 s with the same amount of force, as measured by a force-sensing, open-irrigated-tip RF catheter. The incidence of SPs, ablation parameters, and lesion sizes were compared with open-irrigated RFA with initial power at 50 W, which was not titrated.
For the second set of experiments, an open-irrigated ablation catheter was placed perpendicular to the thigh muscle surface. Ablations were delivered with initial power at 50 W, with the same amount of force, as measured by a force-sensing, open-irrigated-tip RF catheter. However, if the decrease in impedance was >4 Ω/s in the first 5 s, the power was immediately decreased to 30 W. The incidence of SPs, ablation parameters, and lesion sizes were compared with open-irrigated RFA with continuous power at 50 W for 60 s.
This study was approved by the Institutional Review Board at the University of Colorado. A retrospective analysis was undertaken for all ablation procedures performed from January 2016 to May 2017. Ablation procedures were included in the analysis if SPs were noted and recorded. Ablation settings, including RF power and duration, were at the discretion of the operator. All cases except 1 (4-mm nonirrigated catheter) used ThermoCool or ThermoCool SF catheters. Ablation parameters for each SP were analyzed. An SP was defined as an audible pop associated with a sudden impedance spike or visualization of gas on ICE during RFA. For each ablation with an SP, 2 matched control lesions were selected from the same procedure and heart chamber. The closest ablation lesion (in time and location) preceding the ablation with an SP was chosen.
Ablation lesion volume measurements
Lesion volumes were acquired by analyzing tissue sections by an investigator blinded to the experimental group. Because of altered tissue architecture and cavitations, lesions resulting in SPs were not measured and not included in lesion analyses. Single-lesion volumes were calculated using the equation for a truncated oblate ellipsoid. For each lesion, maximum depth (A), maximum diameter (B), depth at maximum diameter (C), and lesion surface diameter (D) were measured. Equation 1 calculates the volume of a truncated oblate ellipsoid:(Equation 1)where A is maximum depth, B is maximum diameter, C is depth at maximum diameter, and D is lesion surface diameter, in millimeters.
The Student t test was used to compare continuous variables (lesion characteristics, measured impedances, and temperatures) between different groups with equal variances within the groups not assumed. When comparisons were made using pooled results from the same animal or slab, hierarchical analysis with adjustment for possible bias due to clustering was performed (Table 1). The p values in Tables 1 to 3⇓⇓ are from paired comparisons within each group adjusted for clustering. Two-way analysis of variance was performed for Figures 1 and 2. We used chi-square and Fisher exact tests for comparison of dichotomous variables with a dichotomous outcome and analysis of variance for comparisons of continuous variables with a dichotomous outcome. A multivariate logistic regression including baseline impedance and the impedance change during the first 5 s of RF application was performed to assess the relationship of these variables on odds of subsequent SP. Statistical analyses were performed using SPSS version 24.0 (IBM, Armonk, New York), and statistical significance was defined as a 2-sided p value <0.05. All results are presented as mean ± SD.
Receiver-operating characteristic curves were constructed to determine optimal cutoff values for impedance reductions and echogenicity on ICE to predict SPs during ablation. The curve point with the highest sum of sensitivity and specificity was labeled as the optimal cutoff point and was used in reporting the positive and negative predictive value.
Echocardiographic changes in cardiac tissue density can predict SP risk
For RFA lesions with SPs evaluated by ICE, there was 32% greater tissue echogenicity (Figure 3) compared with the maximum echogenicity (at the end of ablation) in lesions without SPs (p < 0.001). In addition, SP RFA lesions had a more rapid rate of change in tissue echogenicity (1.48 ± 1.38 SBI/s vs. 0.25 ± 0.45 SBI/s for lesions without SPs, p < 0.001). Changes in tissue echogenicity were analyzed for every 5-second interval before an SP or to the end of ablation (Figure 1). The most significant and rapid increase in echogenicity occurred within the last 5 s before SP generation.
Measured tissue impedance changes can predict risk for SPs during in vivo ablation
Compared with RFA without SPs, RFA resulting in SPs had a higher starting impedance (130.0 ± 21.2 Ω vs. 111.0 ± 10.7 Ω, p = 0.002) and a larger total impedance reduction (33.0 ± 16.0 Ω vs. 23.0 ± 10.8 Ω, p = 0.032) (Table 1). Not only were SPs associated with a larger absolute impedance change, but they also were associated with a more rapid rate of impedance reduction after the initial delivery of RFA: there was an average impedance reduction of 33 Ω over 23.0 ± 16.3 s, or 1.4 Ω/s for RFA with SPs, compared with an impedance decrease of 0.38 Ω/s for RFA without SPs (p = 0.001). After adjusting for the initial impedance observed, the rate of change in the first 5 s of ablation remained statistically significant between SP and non-SP lesions (p = 0.02).
In addition, changes in impedance were analyzed for every 5-s interval from initiation of ablation to an SP or to the end of ablation (Figure 2). The most significant and rapid reduction in impedance occurred within the first 5 s for RFAs resulting in SPs (19% of initial impedance vs. 10% for RFAs without SPs, p < 0.001), even though SPs did not occur until later during RFA delivery.
There were no significant differences in peak temperatures, average force, and maximum force for those RFAs with or without SPs.
A strategy of titration of RF power can decrease the risk for SPs
Incremental power up-titration significantly decreased SP incidence compared with RFA with initial power of 50 W without power titration (5 of 24 lesions vs. 19 of 36 lesions, p = 0.01). The overall impedance reduction for up-titrated RFA lesions was similar to RFA at 50 W that did not result in SPs (19 ± 13 Ω vs. 19 ± 9 Ω, p = 1.00). Up-titrated RFA lesions did not have a rapid reduction in impedance (<4 Ω/s) in the first 5 s of ablation or throughout the ablation, compared with RFA at 50 W with SPs. Predictably, because they had less time at 50 W, ablation lesion volumes were overall smaller for up-titrated RFA lesions (495.4 ± 114.7 mm3 vs. 607.5 ± 119.4 mm3 for RFA lesions at continuous 50 W that did not result in SPs, p = 0.01) (Table 2). However, for continuous 50-W ablations, 16 of 36 SPs occurred before 40 s (3 of 19 SPs occurred after 40 s) compared with only 5 of 24 SPs (p = 0.09) occurring in the 40 s of 50 W for up-titrated ablations.
Decrease in power after rapid impedance reduction minimizes SP risk
The strategy of immediate down-titration of power to 30 W was also tested for its ability to prevent SPs when RFA with initial power at 50 W led to rapid impedance reduction (>4 Ω/s) in the first 5 s. This strategy resulted in a significantly decreased SP incidence compared with RFA at 50 W without down-titration (0 of 13 vs. 19 of 36, p < 0.01). The overall impedance reduction for down-titrated RFA lesions was greater than for RFA at 50 W that did not result in SPs (41 ± 13 Ω vs. 19 ± 9 Ω, p < 0.01) and similar to RFA at 50 W that did result in SPs (41 ± 13 Ω vs. 33 ± 16 Ω, p = 0.12). Ablation lesion volumes were significantly smaller compared with those created with RFA at continuous 50 W (186.6 ± 51.3 mm3 vs. 607.5 ± 119.4 mm3, p < 0.001) (Table 2).
Tissue characterization BEFORE SPs during clinical ablation procedures
From January 2016 to May 2017, 20 SPs were noted and recorded in 16 procedures. These procedures included 10 ablations for ventricular tachycardia: 10 SPs occurred in the left ventricle and 3 occurred along the right ventricular septum (Figure 4). SPs occurred in 4 ablations for cavotricuspid isthmus flutter, 1 pulmonary vein isolation, and 1 atrioventricular node ablation. The SmartTouch ThermoCool with force sensing was used in 9 cases, the SurroundFlow ThermoCool with force sensing was used in 6 cases, and a standard nonirrigated 4-mm catheter was used in 1 case. There were no perforations or pericardial effusions resulting from the SPs. No other complications were observed in these specific cases.
Compared with matched subjects (n = 40), ablations resulting in SPs (n = 20) had larger impedance reductions after 5 seconds of ablation (15.7 ± 6.7 Ω vs. 8.1 ± 4.7 Ω, p < 0.0001). Figure 5A shows representative impedance curves for an SP ablation compared with a control ablation. In this example, the control ablation had a steady reduction in impedance in the initial 5 s, and power was appropriately up-titrated throughout ablation duration. For SP ablation, a significantly rapid reduction in impedance occurred in the first 5 s; power continued to be up-titrated during the ablation, and an SP resulted. An SP may have been avoided if the initial rapid impedance reduction was considered a warning to not up-titrate power further during that ablation.
As a percentage of the starting impedance, the change in impedance after 5 seconds was also significant for ablations with SPs (12.0 ± 5.0% vs. 6.8 ± 3.7%, p < 0.001). Figure 5B is a scatterplot of the percentage changes (from starting impedances) at 5 seconds for individual ablations with SPs versus controls. There was a significantly larger overall impedance reduction for the entire ablation duration for those ablations with SPs (28.0 ± 11.0 Ω vs. 19.1 ± 10.0 Ω for controls, p = 0.01). There were no differences in maximum or average power, maximum or average force, force-time integral, ablation duration, or maximum temperatures (Table 3).
In this study, we performed ex vivo and in vivo experiments designed to assess echocardiographic and biophysical predictors of SPs; findings were validated with a retrospective analysis of clinical SPs during ablation procedures. We found that although absolute changes were important, the rate of these changes was even more important. RFAs resulting in SPs had an absolute increase in tissue echogenicity of 32%. Furthermore, increase in echogenicity for SP-generating RFAs occurred at >5 times the rate of echogenicity changes for non-SP lesions. Most of this increase occurred in the last 5 s before SP.
The absolute reduction in impedance was 43% greater for RFAs resulting in SPs compared with RFA without SPs. In addition, the rate of impedance decrease was almost 4 times faster for SP-causing RF lesions compared with those without SPs. Most of this rapid impedance decrease occurred within the first 5 s of ablation. The impedance reduction, as a percentage of the starting impedance, was almost twice as large as control impedance reductions (19% vs. 10%, respectively, for experimental in vivo studies and 9.3% vs. 5.3%, respectively, for clinical ablations).
It is important to note that on the basis of our findings, there are 2 separate time points at which SP risk can be predicted: at the initiation of ablation, when impedance changes are most dramatic, and right before SP generation, when echocardiographic imaging can detect impending steam generation.
We believe that our findings have direct applicability for clinicians. Using a discrete cutoff of a >9% reduction in baseline impedance within the first 5 s of RF application, we can predict subsequent SPs with good discrimination (positive predictive value, 75%; negative predictive value, 92%). The receiver-operating characteristic analysis using impedance change as a predictor of SP occurrence demonstrated a C statistic of 0.85, indicating good discrimination for SP risk. When used together with increasing echogenicity, the discrimination for SP prediction is enhanced, because the measurements are at different times and are independent of each other. On the basis of our results, an impedance change of >9% in the initial 5 s of RF application followed by a >26% in echogenicity at any time after this would yield a positive predictive value of 94.3% and a negative predictive value of 98.3%.
Mechanisms for altered tissue characteristics during lesion formation and subsequent SPs
When RF energy is applied to tissue interposed in an electric circuit running from the catheter tip to a surface patch electrode, it is heated because of its resistive properties (11,12). With heating, the tissue becomes necrotic and desiccated, resulting in altered electric properties. With the application of RF energy resulting in lesion formation, one can expect a reduction in impedance as the electric properties of the heated tissue change.
An audible SP occurs when there is a phase change from liquid to gas and gas is released. In the present study, we have shown that a steep reduction in impedance during initial RFA is associated with a subsequent SP later during RFA. This large impedance reduction likely reflects tissue desiccation. Our results indicate that immediately before an SP occurs, there is ultrasonic evidence of gas formation. The rapid impedance reduction seen earlier in lesion formation predicts the gas formation that can be seen by ICE immediately before explosion.
Prior studies on biophysical and ultrasonic parameters related to SP predication
There have been several studies evaluating SP creation in the clinical or laboratory setting. In 2004, Cooper et al. (5) demonstrated higher signal intensity using ICE immediately before SPs. Wright et al. (13) used a novel ablation catheter with the ability to visualize echogenic signals to identify increased signal intensity immediately before steam-generated cavitary lesion formation. In addition, Seiler et al. (1) noted a greater impedance reduction associated with subsequent SPs. However, the precise range of impedance decrease or the timing at which SP risk increases, relative to impedance reduction, is unknown. Our results are consistent with these prior studies, but we note additional features of lesions associated with impending steam-generated cavitation, including the importance of a rapid impedance change as well as initial rapid reduction in the first 5 s of ablation. In addition, Ikeda et al. (14) found increasing SP incidence with increasing contact force (CF) in a canine model. Given that CF is a known predictor of SP risk, CF was kept constant in our studies to assess other predictors. Of note, there were no statistical differences in mean CF, maximum CF, and force-time integral in the clinical cases between SP and non-SP ablations.
Building on experience described by others, we have quantified the rate of impedance reduction leading to an SP, which is more significant than absolute impedance change. In our studies, if an initial rapid impedance reduction is noted, and the ablation power is down-titrated, an SP can be avoided; thus, even though total absolute impedance change for ablation was large, comparatively as large as ablations with SPs, a SP did not occur.
In addition, although increased echogenic signals have been described using near field (>9-MHz) ultrasound, our study is the first to use phased-array ultrasound, which has become the clinical standard for use in ablation of atrial fibrillation and ventricular tachycardia.
Clinical strategies to reduce risk for SPs
In our study, we were able to measure impedance reductions and increased echogenic signal intensity when creating ablation lesions using different models of clinical lesion formation. On the basis of our results, we believe that careful attention to rapid rates of impedance reductions (>4 Ω/s) will lead to reduced SPs. Consistent with this hypothesis, our studies showed that immediate reduction of power, once this impedance trend was observed, decreased SP incidence. In addition, it is noteworthy that absolute impedance reduction for these lesions with power down-titration was large but did not cause SPs; in fact, the impedance change was similar to the reduction seen for ablation lesions at 50 W that resulted in SPs (41 ± 13 Ω vs. 33 ± 16 Ω, p = 0.12).
Another clinically relevant finding in the present study was the reduced number of SPs observed when power was titrated to 50 W compared with fixed maximum-power delivery. This strategy of power titration could allow the dispersion of heat deeper within tissue before achieving higher power. Of course, lesion size will be comparably smaller because of the lower average power delivered, but the up-titration strategy should result in fewer SPs when used clinically.
Finally, careful attention to lesion formation using ICE can possibly avoid impending SPs with identification of microscopic gas formation within the lesion. This is especially pertinent when a rapid increase in tissue echodensity is observed.
The limitations of ex vivo studies have been detailed previously and include variability in catheter contact or angulation, passive catheter cooling, intracardiac ultrasound angulation, and the presence of ischemic myocardium due to nonperfusion (6–8,15,16). These variables were nondifferential among control and test groups, and repeated measurements within groups were performed to reduce the impact of these variables. Our ex vivo and in vivo experimental models were performed on normal cardiac and muscle tissue and do not necessarily replicate pathophysiologic substrates during clinical ablation. The lack of precise real-time brightness information on ICE precludes the immediate application of our findings. However, an extension of this research would be that future versions of intracardiac echocardiographic equipment may allow this potential safety feature by automatically assessing the rate of increase in echogenicity and establishing warning parameters.
For our case series of SPs, the retrospective nature is a limitation. The small sample size is a limitation, but differences were significant. Procedural details were not standardized, as they were at the operator’s discretion. Furthermore, which catheter was used was dependent on availability and on the proceduralist’s approach. Furthermore, individual clinical cases varied depending on patient habitus, underlying pathological substrates, and CF. There were individual power variations and the amount of impedance change that was tolerated before ablation was halted or not further titrated.
Our study builds on limited published data for SP monitoring and prevention. Certain biophysical and echocardiographic parameters during open-irrigated RFA are associated with increased SP risk. Although absolute changes in impedance and echocardiographic intensity are significant, rates of change in these parameters are additional predictors of importance. Further clinical studies are needed to confirm these findings as predictors of impending SPs.
COMPETENCY IN MEDICAL KNOWLEDGE: SPs during RFA can lead to complications, but predicting SP risk can be difficult. In this study, we found that although absolute changes were important, the rate of these changes was even more predictive. RFAs resulting in SPs had rapid impedance reductions during the first 5 seconds of ablation. Furthermore, increase in echogenicity for SP-generating RFAs occurred at >5 times the rate of echogenicity changes for non-SP lesions. Most of this increase occurred in the last 5 s before SPs. When these findings are apparent, various strategies to mitigate SP risk should be used.
TRANSLATIONAL OUTLOOK: There are potential clinical applications of our findings for predicting and mitigating SP risk during RFA. These applications use tools, such as trending impedances over time and ICE, that are already commonly used during electrophysiological procedures.
Drs. Sauer and Nguyen have received significant research grants from Biosense Webster and CardioNXT; and educational grants from Biosense Webster, Boston Scientific, and Medtronic. Drs. Sauer and Nguyen have a provisional patent on partially insulated focused catheter ablation; and have nonpublic equity interests and stock options in CardioNXT. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
All authors attest they are in compliance with human studies committees and animal welfare regulations of the authors’ institutions and Food and Drug Administration guidelines, including patient consent where appropriate. For more information, visit the JACC: Clinical Electrophysiology author instructions page.
- Abbreviations and Acronyms
- contact force
- intracardiac echocardiography
- radiofrequency ablation
- standardized brightness intensity
- steam pop
- Received June 25, 2017.
- Revision received September 12, 2017.
- Accepted October 4, 2017.
- 2017 American College of Cardiology Foundation
- d’Avila A.,
- Houghtaling C.,
- Gutierrez P.,
- et al.
- Nakagawa H.,
- Wittkampf F.H.,
- Yamanashi W.S.,
- et al.
- Nguyen D.T.,
- Tzou W.S.,
- Zheng L.,
- et al.
- Ikeda A.,
- Nakagawa H.,
- Lambert H.,
- et al.
- Nguyen D.T.,
- Barham W.,
- Moss J.,
- et al.