Procedural and Anatomical Determinants of Multielectrode Renal Denervation Efficacy

Abstract

Radiofrequency renal denervation is under investigation for treatment of hypertension with variable success. We developed preclinical models to examine the dependence of ablation biomarkers on renal denervation treatment parameters and anatomic variables. One hundred twenty-nine porcine renal arteries were denervated with an irrigated radiofrequency catheter with multiple helically arrayed electrodes. Nerve effects and ablation geometries at 7 days were characterized histomorphometrically and correlated with associated renal norepinephrine levels. Norepinephrine exhibited a threshold dependence on the percentage of affected nerves across the range of treatment durations (30-60 s) and power set points (6-20 W). For 15 W/30 s treatments, norepinephrine reduction and percentage of affected nerves tracked with number of electrode treatments, confirming additive effects of helically staggered ablations. Threshold effects were only attained when ≥4 electrodes were powered. Histomorphometry and computational modeling both illustrated that radiofrequency treatments directed at large neighboring veins resulted in subaverage ablation areas and, therefore, contributed suboptimally to efficacy. Account for measured nerve distribution patterns and the annular geometry of the artery revealed that, regardless of treatment variables, total ablation area and circumferential coverage were the prime determinants of renal denervation efficacy, with increased efficacy at smaller diameters.

Keywords: catheter ablation; hypertension; models, animal; norepinephrine; renal artery.

Figure 1. Correlation of nerve effects with NEPI response.

Average NEPI values (A) and corresponding reduction relative to historical baseline (B) for 125 kidneys of treated animals tracked %AN in sections with mural evidence of treatment. Diamonds – binned experimental values. Red lines – sigmoidal fit of the entire data set to (Eq. 1) with respect to AN (R²=0.331, AN_th= 46.3%).

Figure 2. Biomarker effects scale with the number of circumferentially staggered arterial treatments.

(A) Angiography illustrates electrode locations along the renal artery (B) % AN (diamonds) is consistent with each electrode affecting 12.94% of the nerves (line, R²=0.213, with p < 0.0001 for the slope). (C) Percent NEPI reduction (diamonds) increases sigmoidally (red line, Eq. 2) with the number of electrodes (#electrodes_th=2.95, R²=0.617) such that at least 4 electrode treatments were required to ensure >49% NEPI reductions. Asterisks denote statistically significant differences. All treatments used the same protocol (15W/30 sec).

Figure 3. Power dependence of single and dual treatments.

%AN (A) and NEPI (B) for single (empty diamonds) or dual (solid diamonds) 30sec/5-electrode treatments at 10–15W.

Figure 4. Neighboring veins can limit ablation area.

Histomorphometry of sections corresponding to the blue (A) and red (C) ablation zones depicted in Supplemental Fig S2. Tissue micrographs were contrasted with computationally predicted ablation zones for single electrode treatments that are either directed into the neighboring vein (B) or lateral to it (D). Artery locations are denoted by ‘Ar’.

Figure 5. Nerve distribution varies with angle and depth.

(A) Nerve (solid red circles) locations and relative sizes are depicted in a tissue cross section with a circular lumen (red dashes, r_lumen=2.0mm) and circular demarcations (purple dashes) at depths 4, 8, 12 mm from the lumen. (B) Cumulative radial nerve distribution: in vivo (open circles) vs a model (Eq. 1) of exponential radial distribution, (line) with L_min=0.5mm, L_nerve=4.0mm (R²=0.990). (C) Cumulative angular nerve distribution: in vivo (open circles) vs a model of uniform angular distribution (dashes).

Figure 6. Diminishing returns for larger ablation areas.

(A) Measured AN values for a total of 49 arteries (43 male, 6 female) are grouped in bins of 7 and plotted against the corresponding maximal ablation area (diamonds) and contrasted with a model best fit (red line) that accounts for non-uniform radial nerve distribution (Eq. 3 with L_min=0.5mm, L_nerve=4.0 mm, 2r_lumen=4.0mm and x₁=6.87 (R²=0.218), implying an average ablation arc of 52°. Model predicted AN is also presented for 4- (magenta) and 3- (orange) electrode treatments. (B) Model NEPI reductions (Eq. 2) corresponding to AN from panel (A). (C) Model (Eq. 3) predicted AN dependence on lumen diameter: 2r_lumen=2.0mm (blue), 4.0mm (red) or 6.0mm (green). Petal diagrams illustrate the inverse dependence of area coverage on lumen diameter (target tissue =pink, ablation zones= light blue, lumen = blue, red or green based on diameter value). (D) Model predicted NEPI reductions (Eq. 2) corresponding to AN from panel (C).