The main findings of this study are as follows: 1. The middle back dispersive electrode position yielded the lowest BP-GI and Myo-GI, followed by the upper back and lower back positions. The GI variations among these positions were small yet statistically significant. Dual upper and middle back dispersive electrodes further lowered BP-GI by 9 Ω compared to a single middle back electrode. 2. BMI and Hct could predict BP-GI and Myo-GI with acceptable accuracy. The predictive accuracy for BP-GI was notably higher when using the back dispersive electrode positions (adjusted R2 = 0.70–0.80) compared with the hip position (adjusted R2 = 0.44). The same trend was observed in Myo-GI predictions, although the predictive accuracy was lower than that for BP-GI. 3. The contribution of Hct to GI was greater than that of BMI.

These findings suggest that dispersive electrode positions below or above the lower edge of the scapula (middle back and upper back positions) are optimal for attaining low GI, and the hip position should be avoided. Dual dispersive electrodes effectively decreased GI. The higher contribution of Hct compared to BMI to GI for each dispersive electrode position suggests that the impact of blood pool impedance is stronger than that of extracardiac impedance.

Effects of dispersive electrode position on GI

Dispersive electrodes with a large surface area and proximity to the ablation sites have been reported to reduce GI [2]. In the current study, among the four dispersive electrode positions, the middle back position (just below the lower edge of the scapula) showed the lowest BP-GI and Myo-GI values, followed by the upper back position (just above the lower edge of the scapula). Dual upper and middle back dispersive electrodes further lowered BP-GI by 9 Ω compared to a single middle back electrode. The three back dispersive electrode positions showed acceptably low GI values. The 3–7 Ω difference between these positions, though statistically significant, is likely of limited clinical impact. However, despite the small differences, the reason why the middle back dispersive electrode yielded significantly smaller GI values than the other back dispersive electrodes should be considered.

The difference between the lower back and middle back positions can be simply attributed to the distance from the heart. On the other hand, the difference between the middle back and upper back positions cannot be easily explained, as both electrodes were considered to be almost equally close to the heart. Although we could not definitively elucidate the cause of the difference, we suspect that air in the lung might have increased the extracardiac impedance when the upper back dispersive electrode was used. A disproportionately high GI was observed when the hip position was used compared with the other dispersive electrode positions, possibly due to a greater amount of fat tissue in the hip than in the back.

This study also suggested the usefulness of BP-GI as a GI indicator instead of Myo-GI during dispersive electrode modulation. As described in the introduction, Myo-GI is affected by myocardial tissue characteristics and catheter contact. This point is advantageous when predicting lesion formation or the risk of steam pop at the contacting myocardium; however, it can introduce noise into the assessment of the optimal dispersive electrode position. Evaluating both types of GI may help increase the overall efficacy of RFCA.

Predictability of GI by BMI and Hct

We also investigated the predictability of BP-GI and Myo-GI based on patient BMI and Hct because we considered that BMI and Hct could be used as indicators of extracardiac impedance and blood pool impedance in the GI measurement models (Figs. 1A and B), respectively [6]. The variability of extracardiac impedance is caused by fat-free mass, body fat, total body water, and other parameters [9]. As we could not preprocedurally collect these data, we adopted BMI as a surrogate for extracardiac impedance. Blood pool impedance is determined by Hct, electrolyte and fibrinogen concentrations, total protein, and temperature. We adopted Hct as a surrogate for blood pool impedance because Hct has been reported as a major determinant among them [10,11,12].

Using the middle back dispersive electrode position, identified as optimal based on the results above, the predictability of BP-GI by these two parameters was sufficiently high (adjusted R2 = 0.78). Notably, the relatively low predictability of Myo-GI (adjusted R2 = 0.55) with the same dispersive electrode position could be attributed to the heterogeneity of myocardial tissue characteristics and catheter contact. These results indicate that GIs with an optimal dispersive electrode configuration can be predicted before ablation.

The contributions of dispersive electrode configuration, extracardiac impedance, and blood pool impedance on RFCA

According to the current study, dispersive electrode configuration, BMI, and Hct were demonstrated to be major determinants of GI values. The variance of GI due to these factors may affect the effectiveness of RFCA. A high GI due to an inappropriate dispersive electrode position or a high BMI value (typically related to the amount of subcutaneous fat) leads to an increase in the impedance of the overall circuit and a decrease in RF current in the myocardium. Decreased RF current in the myocardium may impair RFCA effectiveness. These points contrast with a high GI caused by tight catheter contact, which is associated with an increased risk of steam pop. Conversely, decreasing the circuit impedance by modulating the dispersive electrode configuration may contribute to offsetting the excessive extracardiac impedance and thus lead to enhancing the ablation efficacy.

A high Hct is associated with high blood impedance [10,11,12], which can also contribute to an increase in the impedance of the overall circuit. However, in an ex vivo experiment, Takigawa et al. reported that higher local impedance of the saline pool increased lesion depth and the rate of steam pop [13]. Theoretically, an increase in blood pool impedance (or saline pool component in an ex vivo model) can change the current distribution between the myocardium and the blood pool. Therefore, an increased GI due to high blood pool impedance may have a different effect on ablation efficacy than that caused by excessive extracardiac impedance.

Clinical implications

According to the results of this study, the middle back dispersive electrode position, followed by the upper back position, may be optimal for lowering GI. The decrease in GI values by attaching an additional dispersive electrode was also demonstrated; additional back dispersive electrodes lowered BP-GI by 6 to 15 Ω compared to a single back dispersive electrode. This study also adds novel insight into the predictability of GI. With the dispersive electrode attached near the LA, BP-GI can be accurately predicted by BMI and Hct. A measured BP-GI greater than the predicted value possibly indicates some kind of excessive extracardiac impedance and suggests the need to modulate the dispersive electrode configuration.