We conducted a single-center, prospective, observational study in fifteen patients (13% females) with atrial arrythmias undergoing catheter ablation. All patients were included from December 2021 to June 2022 and the inclusion criteria were: (1) Age of comprised between 18 and 85 years old, (2) Understanding of the informed consent, (3) That they did not present any contraindication and pass the exploration and tests prior to bioimpedance measurements. The exclusion criteria were: (1) Age outside the range described in the inclusion criteria, (2) Subjects who presented any type of complication during the procedure, (3) Pregnancy. The protocol was approved by the ethics committee and the study was conducted in accordance with the principles of the Declaration of Helsinki. Written informed consent was obtained from all patients prior to study participation.

An offline analysis of the bipolar electrograms (EGM) and of the impedance signals was done using custom-made Matlab scripts (Mathworks, USA) and the following parameters extracted: Bipolar amplitude (in mV) of the EGMs, contact force data, the magnitude and phase angle of the local multifrequency impedance at all current frequencies (in ohms), the amplitude of the cyclic changes of the impedance magnitude at 5 kHz (in ohms), and the impedance magnitude and phase angle drops at all current frequencies (in ohms and percentage). We also analyzed the effects of catheter orientation on the studied variables. For this purpose, we considered that the catheter was orthogonal to the tissue when the angle (ϴ) between the catheter and the contact force was ϴ<|15°|, and that the catheter was parallel to the tissue when the angle between the catheter and the contact force was ϴ>|70°| and ϴ<|110°|.

Quantitative data were expressed as the mean ± standard deviation (SD). The ordinary two-way ANOVA test with Sidak's multiple comparison correction was used to assess the statistical significance of changes in bipolar voltages, impedance magnitude and phase angle at different current frequencies (factor 1: blood, pre- or post-ablation; factor 2: current frequencies/catheter used/parallel or orthogonal condition). A p-value <0.05 was considered significant. All analyses were performed using SPSS v.22.0 software (IBM-SPSS, USA).

The LMI magnitude of pre-ablated myocardium was significantly higher than of post-ablated myocardium and of blood pool, from 1 kHz to 209 kHz, in both catheters (ANOVA p < 0.05). LMI magnitude of blood pool and post-ablated did not show differences. The LMI phase angle showed similar results (Figure 2). All LMI magnitude values acquired with the QDOT catheter were lower than the measurements performed with the SmartTouch catheter (ANOVA p < 0.001). However, the LMI drop was similar among both catheters. The current frequency range that better discriminated pre-ablated from post-ablated regions was 1–371 kHz for the LMI magnitude and 173–1,000 kHz for the LMI phase angle. Detailed impedance spectra for each patient are shown in Supplementary Figure S1.

Figure 3 illustrates the biphasic pattern of the LMI magnitude during the cardiac cycle and its relationship with the contact force, the bipolar electrograms and the surface ECG. The pre-ablated myocardium had an LMI amplitude higher than the post-ablated myocardium and the blood pool from 1 to 95 kHz, in both catheters (LMI magnitude amplitude using QDOT @ 41 kHz: Z_PRE = 12 ± 8 Ω vs. Z_POST = 7 ± 4 Ω; ANOVA p < 0.001). There were no statistical differences between post-ablated tissue and blood, despite the catheter being used. As expected, the mean amplitude of the bipolar electrograms was higher in the pre-ablated sites than in the post-ablated sites (M1-M2 amplitude using STOUCH: V_PRE = 2 ± 2 mV vs. V_POST = 0.9 ± 0.7 mV; ANOVA p < 0.05). The mean contact force recorded was comparable between the pre- and post-ablated sites and was higher than in blood pool for both catheters (Contact force measured using STOUCH: F_PRE = 8 ± 6 g and F_POST = 9 ± 7 g vs. F_BLOOD = 1 ± 1 g; in both, ANOVA p < 0.05).

In a subset of the explored sites, we analyzed the influence of catheter orientation on LMI, the contact force and bipolar electrograms. In Figure 4 we can observe that LMI Magnitude changed only in pre-ablated myocardium when using the SmartTouch catheter (LMI measured using STOUCH @ 5 kHz Orthogonal vs. Parallel: 114 ± 21 Ω vs. 131 ± 10 Ω; ANOVA p < 0.001). The contact force was higher in the pre-ablated sites only when using the QDOT catheter if it was orthogonal to the tissue. Local bipolar electrograms showed the same behavior despite catheter orientation.

This study shows for the first time that local multifrequency impedance can be used to distinguish pre-ablated from post-ablated atrial myocardial tissue, in patients submitted to radiofrequency ablations in PVI procedures.

A fifteen to twenty ohms drop in local impedance during RF ablation is associated with successful lesion creation, as shown in Table 2. This is in accordance with our mean impedance magnitude drop at the current frequency used by commercial RF generators (50 kHz).

Recent studies have shown that the optimal LI drop for effective lesion formation depends on the atrial region where the lesions are performed. Higher cut-off values are needed for the anterior wall which has higher wall thickness compared to the lower cut-off values needed on the thinner posterior wall (10, 16). In addition the baseline LI and the LI drop differ according to atrial tissue characteristics: normal-voltage areas show higher baseline LI and greater LI drops, suggesting that tissue substrate should be considered for tailored ablation strategies (9, 17, 19). In this sense, the measurement of impedance at different frequencies allows a better tissue characterization than LI since it permits the detailed characterization of the intra- and extracellular compartments (13). In our study we have described for the first time that human blood pool and post-ablated tissue have a lower LMI magnitude and phase angle than pre-ablated tissue, at all current frequencies studied. The lack of significant differences between post-ablated scarred tissue and blood pool suggests that LMI drops reflect cellular disruption and extracellular matrix disorganization post-ablation, consistent with irreversible lesion formation. This in accordance with previous large-animal studies where the impedance magnitude and phase angle showed the same behavior in the atrial and ventricular scarred tissue (9, 11). The frequency-specific responses, with lower frequencies reflecting extracellular changes and higher frequencies correlating with membrane integrity, further highlight LMI's potential for evaluating the tissue substrate characteristics to perform tailored applications and at the same time evaluate in real-time the lesion formation.

Finally, this study also shows that, despite the ablation power strategy used, 90W for 4 s (QDOT) or 30W for 30 s (Smarttouch), there is a similar LMI drop. This is in accordance with a recent experimental study where different high power short duration strategies and catheters were compared and showed similar impedance drops (20).

The biphasic pattern of LMI magnitude during the cardiac cycle highlights its dynamic interaction with mechanical and electrical activity (21, 22). Pre-ablated myocardium exhibited 171% greater cyclic impedance amplitude than post-ablated tissue, reflecting preserved tissue elasticity and contact force variability. The time-correlation between LMI fluctuations and contact force emphasizes the importance of stable catheter contact during ablation. Post-ablated tissue's attenuated cyclic changes may indicate loss of structural integrity, serving as a functional marker of lesion maturity. These findings suggest that integrating both static multifrequency impedance values and dynamic cyclic behavior could enhance lesion assessment.

Catheter orientation significantly impacted LMI measurements in pre-ablated tissue when using the SmartTouch catheter, but it did not have an effect when using the QDOT catheter. This discrepancy may stem from differences in electrode configuration or contact force sensing mechanisms between catheters. Bipolar electrogram voltages remained unaffected by orientation. Both results are in accordance with a recent in vitro study by Calzolari et al. where they found that changes in catheter orientation did not affect the lesion depth (23).

This study has several limitations. First, the sample size was modest, and the single-center design may limit generalizability. Second, differences in LMI values between catheters (QDOT vs. SmartTouch) complicate direct comparisons, requiring catheter-specific validation. Third, the fact that the impedance signature of a successful lesion converges with that of blood, potentially limits the specificity of the impedance measurement for distinguishing an effective ablation lesion from the pool of blood near the catheter tip. However, the integration of the magnitude and phase angle together with the cyclic changes of LMI and the contact-force data could overcome this issue. Finally, while LMI's cyclic behavior was characterized, its clinical utility for predicting lesion durability remains speculative without further validation using imaging methods.

This study shows that measuring local impedance at different current frequencies is feasible in the clinical scenario. The integration of LMI with electroanatomic mapping systems may enable detailed local tissue characterization and at the same time automated lesion tagging, optimizing ablation line continuity. Furthermore, cyclic impedance patterns could serve as intraprocedural biomarkers of contact stability and lesion transmurality. Recent studies have shown that the deepest penetrations can be achieved using an ablation strategy consisting of applying 50W for 10–15 s (24). Future studies should explore LMI-guided ablation strategies using this power setting and others to determine their impact on long-term arrhythmia-free survival.

This study shows for the first time that LMI can differentiate pre- from post-ablated tissue in a cohort of patients submitted to RF ablations. By combining frequency-specific impedance measurements with dynamic cyclic analysis, this new tool could be of potential clinical applicability for the characterization of the atrial substrate and at the same time monitor lesion quality during the ablations.