Determinants of left atrial local impedance: Relationships with contact force, atrial fibrosis, and rhythm

The relationships between baseline tissue local impedance (LI), contact force (CF), atrial fibrosis, and atrial rhythm are uninvestigated in a clinical setting. We compared the relationship of LI and CF between atrial fibrillation (AF) and sinus rhythm (SR) accounting for the effects of atrial fibrosis as assessed by bipolar voltage and LI.

Local impedance (LI) is a novel parameter that has shown clinical utility in radiofrequency ablation.Drops in LI (ΔLI) correlate with ablation lesion size in preclinical studies [1][2][3] and targets consistent with successful achievement of pulmonary vein isolation (PVI) have been developed in clinical trials. 4,5The greatest correlate of ΔLI before ablation is the starting tissue LI. 6,7 Consequently, there is merit in establishing the relationships of factors affecting pre-ablation tissue LI.
In vitro and computer modeling studies have shown that tissue LI is affected by catheter proximity, contact force (CF), and tissue fibrosis. 8,9However, this has not been established clinically.
Furthermore, a clinical study has shown underlying rhythm to affect contact with the atrial wall, and thus may also affect tissue LI. 10 The IntellaNav Stablepoint (SP) is an ablation catheter that combines LI and CF technology, allowing for the relationships of CF, tissue fibrosis, and atrial rhythm with tissue LI to be explored in a clinical setting.

| Patient recruitment
Patients listed for radiofrequency ablation for persistent AF at our center were prospectively recruited for the study.Ethical approval was granted by the UK Research and Ethics Committee (Reference: 18/SC/0077).

| Procedure
Procedures were performed under general anesthetic or local anesthetic with conscious sedation.A decapolar catheter was placed into the coronary sinus as a reference for the creation of 3D electroanatomical maps.Mapping was conducted using the INTELLAMAP Orion catheter using the RHYTHMIA HDx system (Boston Scientific).The Agilis steerable sheath (Abbott Laboratories) was used for all cases.Maps were created in AF and then SR (pacing at 600 ms from the proximal coronary sinus) following direct current cardioversion.
To minimize inter-patient variability, LI values were referenced to a patient's blood pool reading (LIr).Catheter tip presence in the blood pool was confirmed by position on the map, a 0 g CF reading and lack of electrograms on the SP catheter tip electrode.LI and CF were measured using previously described methods. 1 To examine the relationship between CF and LI, the SP catheter was moved to a point on the LA wall selected by the operator.For 30 s, the CF was then gently adjusted between 0 and 40 g (Figure 1).A manual tag placed at the beginning of this maneuver allowed for any macrodisplacement to be easily noticed.If macro-displacement did occur, that site was excluded from analysis, and a new site was found to repeat the maneuver.Up to six locations were examined per patient.This procedure was initially conducted in AF, and following DC cardioversion, these specific points were then returned to and re-examined in SR.Assistance for this was provided by the Rhythmia tag distance tool, which minimized the distance between evaluated points in SR and AF (Figure 1).Up to 10 corresponding CF and LI values were manually obtained per maneuver.Following export, the distance between xyz coordinates of corresponding SR and AF tags was calculated.Any points >2.5 mm (catheter tip width) were excluded from the analysis.
To assess the effect of atrial fibrosis on the CF-LI relationship, the commonly used surrogate bipolar voltage was used alongside the novel parameter LI.Bipolar voltages sampled from the Orion catheter were obtained by co-locating the xyz coordinates of the manual tag and the mapping data from the Rhythmia export.This data were processed using custom-written MATLAB scripts (Mathworks Inc.).Bipolar voltages were categorized as normal-(NVM) and intermediate/low voltage myocardium (I-LVM) using 0.5 mV as a cut-off.
Ideally, a 0.05 mV cut-off would have also been used; however, this was merged with <0.5 mV due to a paucity of data.As above, LI values were referenced to the blood pool (LIr) and taken at 10 g of CF.Locations were then classified as electrically inert (EI), partially active (PA), and electrically active (EA), using cut-offs of ≤0Ω and ≥11Ω based on concurrent work, 11 which assessed the relationship of baseline LIr to a bipolar pacing threshold assessed starting at 5 mA and using a 2 ms pulse width from the tip to the first ring electrode of the SP catheter at 10 g of CF.EI tissue was defined as no pace captures at 20 mA, EA tissue was defined as pace capture <5 mA, and PA tissue was between.

| Statistical analysis
IBM SPSS Statistics (Version 27, IBM Corp) was used for statistical analysis.A p-value of < 0.05 was considered statistically significant.
Variables were assessed as parametric or nonparametric by visual inspection of histograms and a Shapiro-Wilks test.Continuous data were expressed as mean ± SD.Count data were expressed as number (%).As multiple CF and LIr measurements were taken from each location and patient, repeated measures correlation (RMCorr) was used when data was not independent.The CF-LIr relationship was mathematically modeled.The effects of rhythm, bipolar voltage categories and LIr categories upon the CF-LIr relationship were examined using linear mixed modeling.

| Patient and data point characteristics
Patient and data point characteristics are described in Tables 1-3.
From 20 patients, 109 locations were sampled in both SR and AF, obtaining a total of 1903 data points.

| Effect of rhythm on the CF and LI relationship
At an individually sampled location, CF correlates strongly with LIr (RMCorr = 064).Linear mixed modeling was used to examine the effects of rhythm and CF on LIr accounting for repeated data at individual locations.Both rhythm and CF had a significant main effect on LIr (both p < .0005)and interaction effect (p = .022).Higher CF values meant higher LI values, with AF demonstrating a higher LIr than SR for the same CF values (mean difference: 7.5Ω, [95% CI 6.5-8.5Ω]).

| Effect of atrial fibrosis on the CF and LI relationship
The degree of tissue fibrosis assessed using a cut-off of 0.5 mV had no significant main effect on LIr in SR or AF, (p = .11/.50) or interaction effect with CF, (p = .61/.98, Figure 3).| 1063 In contrast, when using LIr to assess tissue fibrosis, a significant main effect (p < .0005both) and interaction effect with CF is seen for both SR and AF (p < .0005/.01, Figure 4), with higher LIr per CF in healthier myocardium.Assessment of atrial fibrosis using bipolar voltage was not shown to affect this relationship; however, an effect was seen if fibrosis was assessed by LIr itself.

| CF and referenced LI
The biophysical reasoning behind the CF-LIr correlation reflects the improved catheter-tissue coupling offered by greater CF.LI readings are formed from a combination of conductivities at the electrodetissue and electrode-blood pool interfaces, similar to two resistors in parallel.As CF increases, more of the catheter tip is embedded within the myocardium, decreasing the effect of the higher blood pool conductivity and raising LI readings.Following on from this, one may anticipate a plateau in this relationship as whilst CF can increase, tissue coupling, and LI would be expected to reach a maximum level.
This was not clearly found in our study, perhaps as we limited CF to 40 g to minimize the risk of perforation and to best reflect clinical practice.It may be that at higher levels of CF, this relationship is seen, or the tissue properties may change dynamically to alter the nature of the relationship.
The strong correlation between CF and LIr reproduces the results of laboratory models in a clinical setting.Using the same catheter, Unger et al. demonstrated how both approaching tissue and increasing CF lead to a rise in LI using an in silico model. 12Their work produced similar results to Sulkin et al., who noted in vitro a nonlinear monotonic increase in LI when the IntellaNav MiFi catheter approached tissue, before a linear relationship following tissue contact. 8Although these studies use catheter displacement in mm rather than measuring CF in grams, and thus a direct comparison is not possible, the relationship between CF and LIr closely resembles our findings.
Additionally, two studies using ex vivo animal models and the SP catheter have demonstrated increasing LI with CF in clinical ranges.
However, the absolute increase in LIr with CF is much greater than seen in our study.For example, using a porcine model in a saline

| The effect of rhythm
In the presence of AF, higher LIr readings were seen for corresponding CF values in SR.This finding may be related to improved tissuecatheter coupling in AF, where the atrial wall does not have the rhythmic "kick" that would be seen in SR. impedance drop during ablation. 10This finding could have particular clinical significance.Pre-ablation LI values have been shown to strongly correlate with LI drop during ablation. 1,6,15Our study demonstrates that higher LI values can be achieved in AF over SR, suggesting AF may a preferable rhythm to ablate during to maximize ablation efficacy.Whether the higher LIr values seen in AF translate to greater LI drop with ablation is unclear and was outside the remit of this study.
In addition to this point, one must also consider that during AF, the atrium is effectively enlarged compared to SR and, therefore, would require more ablation lesions to complete PVI.This increases the chances of gaps appearing acutely or chronically and would counter, to an extent, the biophysical benefit to ablating in AF.

| The effect of fibrosis
In analyzing the effect of atrial fibrosis on the CF-LIr relationship, mixed results were obtained.Classifying fibrosis by bipolar voltage demonstrated no interaction on the CF-LIr relationship.In contrast, using LIr to classify atrial fibrosis did reveal a significant effect on the CF-LIr relationship.
In current clinical practice, bipolar voltage amplitude is used as a surrogate for atrial fibrosis.Using modern mapping systems and specialized catheters, thousands of electrograms can be sampled rapidly and automatically verified, creating exquisite ultrahigh-density maps familiar to electrophysiologists.However, the cut-off value to delineate fibrosis from healthy tissue (0.5 mV) has never been histologically verified.The lack of interaction effect of bipolar voltage categories on CF-LIr could therefore be explained by an inappropriate cut-off to determine fibrosis, or that bipolar voltage is a poor reflection of tissue tensile properties.
Using LI as a mapping parameter to assess tissue fibrosis on a point-by-point basis has been used in a small study, 9 and it is capable of finding gaps in PVI lines. 16The significant interaction effect of tissue fibrosis when classified by LIr (at 10 g) on CF-LIr is of clinical use.For example, if a tissue LIr at 10 g is classified as EI, one may be tempted to increase CF to improve LIr before ablation.However, our data suggests minimal changes in LIr will be seen in exchange for increased risk of perforation or steam pops.In contrast, classifying a point as EA, increasing CF beyond 10 g will continue to show increases in LIr which may improve LI drop.These findings suggest that LIr may represent a better reflection of tensile changes secondary to atrial fibrosis affecting catheter-tissue coupling.
Furthermore, for a potential LI mapping catheter, it should be considered that the amount of tissue contact made will affect readings considerably, especially in healthier tissue.

| Limitations
The measurements made used the IntellaNav Stablepoint catheter.

F
I G U R E 1 A simplified version of the Rhythmia interface demonstrating the response of local impedance (LI) to changes in contact force (CF).(A) CF reading; (B) real-time graphical representation of changes in CF; (C) LI reading; (D) Real-time graphical representation of changes in LI (white = raw LI reading, yellow = filtered LI).Note the peaks and troughs that correspond between CF and LI; (E) 3D electroanatomical map of the left atrium in a postero-anterior projection.The tip of the ablation catheter can be seen in the center of the posterior wall overlying a tag 'AF2'.A distance measurement can be seen above this tag (2 mm) and allows for the assessment of catheter drift.

T A B L E 1
Study population characteristics.

4 | DISCUSSION 4 . 1 |
Key findingsIn summary, our clinical study demonstrates a strong correlation between CF and LIr on the left atrial endocardium.The CF-LIr relationship is logarithmic in nature and affected by atrial rhythm.

F I G U R E 2
Relationship of contact force versus local impedance referenced to blood pool for sinus rhythm (SR) and atrial fibrillation (AF).Contact force is given in 2 g categories.Data points are mean ± standard deviation for each CF category.R 2 values: SR = 0.13, AF = 0.21.p = .022for the interaction effect of rhythm upon the contact force to referenced local impedance relationship.F I G U R E 3 Box plot demonstrating contact force versus referenced local impedance divided by degree of atrial fibrosis (assessed by bipolar voltage categories, [cut-off 0.5 mV]).There was no significant interaction effect of bipolar voltage category with contact force on referenced local impedance (Sinus Rhythm: p = .61,Atrial Fibrillation: p = .98).solution, Matsuura et al. demonstrated a rise in LI of 60-80Ω at 10 g of CF from a baseline of 140Ω depending on catheter orientation. 2Undertaking a similar experiment, Tsutsui et al. recorded values of approximately 220Ω at 10 g of CF. 13 In comparison, using curves of best fit, our clinical study obtained values of 10.5Ω (SR) and 16.0Ω (AF).The discrepancy in these results may be explained by the natural variation in conductivity of human tissue, differences in tissue deformation between clinical and in vitro experiments and the thickness of tissue examined (atrial vs ventricular).In contrast, Gutbrod et al. demonstrated closer absolute results to our study using a left atrial in vivo canine model. 14This highlights the importance of translating concepts discovered in thoroughly controlled, well-designed preclinical studies into the more difficult-tocontrol, complex clinical setting.Interestingly, the correlation between CF and tissue LI has not been shown in other clinical studies.However, the objectives of Solimene et al. and Yasumoto et al. were to examine the relationship of CF to LI drop with ablation. 6,7Consequently, single readings of LI and CF were taken at multiple locations before ablation.In contrast, in our study, multiple LI and CF measurements were taken from each location and referenced to the patient blood pool, therefore reflecting the relationship between CF and LI at individual sites more clearly.
In a previous left atrial study, Ullah et al. noted greater CF variability (the difference between peak and trough CF values) using the Thermocool SmartTouch catheter (Biosense Webster Inc.) in SR compared to AF, and this correlated with a comparatively smaller generator F I G U R E 4 Box plot demonstrating contact force versus referenced local impedance divided by degree of atrial fibrosis (assessed as referenced local impedance categories).There was a significant interaction effect between these categories with contact force on referenced local impedance (Sinus Rhythm: p < .0005,Atrial Fibrillation: p = .01).EI = electrically inert (LIr < 0.5Ω), PA = partially active (0.5Ω < LIr < 10.5Ω), EA = electrically active (LIr > 10.5Ω).
Data point characteristics.Distribution of locations sampled by local impedance category.
T A B L E 2Note: Displayed as n (%) or mean ± standard deviation.Abbreviations: g, grams; lL-IVM, low and intermediate voltage myocardium; mV, millivolts; NVM, normal voltage myocardium.T A B L E 3Abbreviations: LIr, local impedance referenced to blood pool; WACA, wide area circumferential ablation.
2xtrapolation to any future CF and LI-equipped catheters should be performed cautiously.Although increasing CF results in greater LI values, it is unclear if this would result in a greater LI drop with ablation.Orientation was not considered in our analysis and may affect LI readings at certain CF levels.2Thickness of atrial tissue could not be considered in our analysis, and its effect on deformation may affect the CF-LI relationship.Unipolar voltage was not assessed as another surrogate of underlying tissue fibrosis, and therefore the relationship between this is unclear.5 | CONCLUSIONSCF correlates with LIr in the left atrial endocardium.Rhythm affects this relationship with measurements during AF showing higher LIr values than in SR at the same locations for corresponding CF.Optimal catheter-tissue coupling, as judged by tissue LI may be better achieved in AF than SR.Atrial fibrosis does not affect the CF-LIr relationship when assessed by bipolar voltage but does when assessed by LIr values at 10 g.