Author + information
- Received December 19, 2013
- Revision received March 1, 2014
- Accepted March 11, 2014
- Published online August 19, 2014.
- Kenichi Sakakura, MD∗,
- Elena Ladich, MD∗,
- Qi Cheng, MD∗,
- Fumiyuki Otsuka, MD∗,
- Kazuyuki Yahagi, MD∗,
- David R. Fowler, MD†,
- Frank D. Kolodgie, PhD∗,
- Renu Virmani, MD∗ and
- Michael Joner, MD∗∗ ()
- ∗CVPath Institute, Inc., Gaithersburg, Maryland
- †Office of the Chief Medical Examiner, Baltimore, Maryland
- ↵∗Reprint requests and correspondence:
Dr. Michael Joner, CVPath Institute, Inc., 19 Firstfield Road, Gaithersburg, Maryland 20878.
Background Although renal sympathetic denervation therapy has shown promising results in patients with resistant hypertension, the human anatomy of peri-arterial renal nerves is poorly understood.
Objectives The aim of our study was to investigate the anatomic distribution of peri-arterial sympathetic nerves around human renal arteries.
Methods Bilateral renal arteries were collected from human autopsy subjects, and peri-arterial renal nerve anatomy was examined by using morphometric software. The ratio of afferent to efferent nerve fibers was investigated by dual immunofluorescence staining using antibodies targeted for anti–tyrosine hydroxylase and anti–calcitonin gene–related peptide.
Results A total of 10,329 nerves were identified from 20 (12 hypertensive and 8 nonhypertensive) patients. The mean individual number of nerves in the proximal and middle segments was similar (39.6 ± 16.7 per section and 39.9 ± 1 3.9 per section), whereas the distal segment showed fewer nerves (33.6 ± 13.1 per section) (p = 0.01). Mean subject-specific nerve distance to arterial lumen was greatest in proximal segments (3.40 ± 0.78 mm), followed by middle segments (3.10 ± 0.69 mm), and least in distal segments (2.60 ± 0.77 mm) (p < 0.001). The mean number of nerves in the ventral region (11.0 ± 3.5 per section) was greater compared with the dorsal region (6.2 ± 3.0 per section) (p < 0.001). Efferent nerve fibers were predominant (tyrosine hydroxylase/calcitonin gene–related peptide ratio 25.1 ± 33.4; p < 0.0001). Nerve anatomy in hypertensive patients was not considerably different compared with nonhypertensive patients.
Conclusions The density of peri-arterial renal sympathetic nerve fibers is lower in distal segments and dorsal locations. There is a clear predominance of efferent nerve fibers, with decreasing prevalence of afferent nerves from proximal to distal peri-arterial and renal parenchyma. Understanding these anatomic patterns is important for refinement of renal denervation procedures.
Renal sympathetic denervation is a promising new therapy for patients with resistant hypertension, which is defined as failure to achieve control of blood pressure (BP) despite treatment with optimal doses of ≥3 antihypertensive medications (1). Catheter-based radiofrequency renal denervation has demonstrated both safety and efficacy for the treatment of resistant hypertension, with 93% of patients having a reduction in office-based systolic BP of ≥10 mm Hg at 3 years (2). Furthermore, other denervation technologies, such as catheter-based ultrasound, externally applied focused ultrasound, or catheter-based microinfusion of neurotoxic drugs, have been developed (3). Although all renal denervation technologies target renal sympathetic nerves around the renal artery, our understanding of human anatomy of peri-arterial renal nerves remains limited.
The aim of the present study was: 1) to examine the morphological characteristics of nerve fibers with respect to density, size, and distance from renal artery lumen; 2) to investigate the influence of hypertension on peri-arterial renal nerve distribution; and 3) to determine the proportion of efferent and afferent fibers along the peri-arterial neuronal network of renal arteries and within the kidney.
Bilateral renal arteries with attached abdominal aorta and kidneys were collected from 25 autopsy subjects. The first 20 cases were used for the investigation of peri-arterial nerve anatomy, and the other 5 cases were used to investigate the ratio of efferent and afferent nerve fibers within nerve fascicles. Identification of hypertension was on the basis of patient histories and histological examination of kidney sections (4). In the first 20 cases, renal arteries were perfusion-fixed ex vivo under physiological pressure (80 to 100 mm Hg) with 10% neutral-buffered formalin. The use of dyes demarcated the ventral, dorsal, superior, and inferior regions around the renal artery. Each artery with surrounding soft tissue was sectioned at 4- to 5-mm intervals and equally divided into proximal, middle, and distal segments. At least 2 sections distal to the arterial bifurcation also were submitted. Each segment was dehydrated, embedded in paraffin, cut at 5 μm thickness, and stained with hematoxylin and eosin (H&E) and Movat pentachrome. To minimize the effect of autolysis on immunohistochemistry (5), the last 5 cases were collected from subjects within 24 hours of death. After cutting, each section was fixed in paraformaldehyde (4%), followed by microwave fixation.
Digital images from H&E-stained histological sections were acquired at 1.25× magnification. The images were divided into 4 quadrants on the basis of the dye labeling and analyzed with image analysis software (IP Lab for Mac OS X, Scanalytics, Rockville, Maryland). Measurements of the distance from the luminal surface of the renal arteries to each nerve were performed in each quadrant around the renal artery. Details of the methods used for immunohistochemistry are described in the Online Appendix.
Results for continuous variables are expressed as mean ± SD. The Shapiro-Wilk test was used to statistically assess violations of the normal distribution assumption. Each individual distance from lumen to nerve was used in the tables for whole distribution or figures for cumulative percentile of nerves, whereas mean values per renal artery were used for statistical comparison. For statistical comparison of spatial nerve distribution, mean values of nerve counts were derived for proximal, middle, and distal regions, as well as for ventral, dorsal, superior, and inferior location, and matched comparisons were performed by using paired Student t tests or repeated measures analysis of variance for normally distributed parameters. For skewed data distribution, a matched comparison using the Wilcoxon signed rank or Friedman test was applied. Comparisons between hypertensive and nonhypertensive subjects were performed by using the independent Student t test or Wilcoxon rank sum test. Categorical data were analyzed by using the chi-square test or the Fisher exact test. The Spearman correlation coefficient was calculated to assess the correlation between nerve counts identified by using H&E staining and those identified by using neurofilament protein (NFP) staining.
All analyses were performed by using SPSS version 19 (IBM SPSS Statistics, IBM Corporation, Armonk, New York) and JMP 5 (SAS Institute, Inc., Cary, North Carolina). All reported p values were determined by 2-sided analysis, and values <0.05 were regarded as statistically significant.
A total of 40 renal arteries from the first 20 patients were allocated for histological evaluation to determine the renal artery anatomy and nerve distribution. Patient characteristics and renal arterial anatomy are shown in Table 1. Mean individual lumen diameter between the proximal, middle, and distal segments was 4.4 ± 0.9 mm, 4.0 ± 0.7 mm, and 4.1 ± 0.9 mm, respectively (p = 0.004).
A total of 10,329 nerves (8,030 nerves were located proximal to the bifurcation [220 sections] and 2,299 nerves distal to the bifurcation [80 sections]) located around the renal arteries were identified from 300 sections. Representative images of perfusion-fixed renal arteries are shown in Fig. 1. The distribution of nerves, stratified by the distance from the arterial lumen to the nerve within the proximal, middle, and distal renal artery sections, is reported in Table 2. There was no significant correlation between the mean number of nerves per section and vessel diameter (Spearman’s r = –0.004, p = 0.95) nor between the mean number of nerves per section and vessel length (Spearman’s r = –0.31, p = 0.06).
Table 3 shows the distribution of nerves stratified according to the distance from the arterial lumen relative to their circumferential location. The distribution of nerves distal to the bifurcation of the renal artery is shown in Online Table 1. From the first 20 patients (40 renal arteries), 6 kidneys had accessory renal arteries. The mean lumen diameter of the accessory artery was 2.3 ± 0.3 mm. The number of nerves and mean distance from the accessory artery lumen to the nerves are shown in Online Table 2. The cumulative incidence of the distance of the nerves from the arterial lumen was calculated on the basis of a total of 8,030 nerves before the bifurcation (Fig. 2).
For nerve size measurements, 8 renal arteries (4 from hypertensive subjects and 4 from nonhypertensive subjects) were selected to delineate nerve size by using NFP staining, and 4 sections (proximal, middle, distal, and post-bifurcation) were selected from each renal artery. The representative images of nerve size measurements are shown in Online Figure 1. A total of 1,517 peri-arterial nerves from 32 sections were measured (Online Table 3). The correlation between the number of nerves identified by using NFP and H&E staining was excellent, involving a total of 32 co-registered sections (Online Fig. 2). Dual marker for immunofluorescence staining using tyrosine hydroxylase (TH) and calcitonin gene–related peptide (CGRP) was performed on 15 sections from 5 patients (mean age 38 ± 10 years; 4 male and 1 female subject; 3 with hypertension and 2 with no hypertension; mean length of renal artery 2.8 ± 0.8 cm).
Overall, the TH-positive area was far greater than the CGRP-positive area (mean individual TH-positive area/CGRP-positive area ratio 25.1 ± 33.4 [p < 0.0001] for CGRP-positive area versus TH-positive area). The ratio of TH-positive/CGRP-positive area was not significantly different between the proximal, middle, and distal segments of the renal arteries (p = 0.25) (Fig. 3). The presence of sympathetic nerve fibers in kidney tissue is shown in Online Figure 3.
The comparison of the mean number of individual nerves, mean individual nerve distance, and mean individual nerve size between hypertensive and nonhypertensive cases is shown in Table 4.
Human renal nerve location was investigated in 40 renal arteries from 20 subjects with an emphasis on the distance from the lumen of renal artery to each nerve. The maximal mean number of nerves was observed in the proximal and middle segments of the renal artery, whereas the least average number of nerves was seen in the distal segment. The circumferential distribution was greatest in the ventral and least in the dorsal regions. In the main renal artery, the 50th, 75th, and 90th percentile of nerve distance from renal artery lumen was 2.44, 4.28, and 6.39 mm, respectively (Central Illustration). In 8 selected renal arteries, nerve size was similar among the proximal, middle, and distal segments. Compared with the TH-positive fibers, the CGRP-positive fibers were rarely observed, whereas the TH/CGRP ratio was similar between the proximal, middle, and distal segments. Nerve anatomy in hypertensive subjects did not differ compared with that of nonhypertensive patients.
Although precise knowledge of peri-arterial renal nerve distribution is key to understanding renal sympathetic denervation procedures, the information has not been fully elucidated in humans. Recently, Atherton et al. (6) reported anatomy of peri-arterial renal nerves in 9 renal arteries from 5 patients. One of the shortcomings of this study is the inadequate method used for the histopathological sectioning; the authors only examined 2.5 mm of the perirenal tissue around the renal artery, therefore introducing a bias with respect to assessing the true distance of peri-arterial nerves from the lumen. In this regard, the results of the present study are more comprehensive owing to the detailed analysis of nerve distance, including circumferential orientation of distances measured in a well-powered number of hypertensive and nonhypertensive patients. Another difference between the present study and that by Atherton et al. relates to the substantially greater number of nerves counted along the peri-arterial renal circumference in our study, which may partly explain the different methodological approaches by including a larger area of the perirenal tissue in our sections to quantify nerve fibers. Furthermore, we used immunohistochemistry to investigate the ratio of afferent to efferent nerve fibers within nerve fascicles, which is an important step toward understanding how disrupted neuronal cross-talk may also affect arterial hypertension.
Compared with proximal and middle segments, there were fewer nerves in the distal segments, and the distance from the lumen to nerves was shorter. One potential explanation for this anatomic variability is the fact that paravertebral aortorenal ganglia are located near the ostia of the renal arteries, and they receive substantial rami from thoracic splanchnic nerves (greater, lesser, and least) (7). These rami terminate in the aortorenal ganglia, and a large number of fibers gives rise to the renal plexus. Therefore, the number of nerve fibers is likely to be maximum near the ganglia (e.g., near the ostia of the renal arteries or in the proximal regions of the renal arteries). Moreover, the aortorenal ganglia are located in either the superior, inferior, or ventral segments of the renal artery, which support our finding that the density of nerves is less in the dorsal segments of the renal artery.
To our knowledge, the density of efferent or afferent nerves has not been investigated in the human autonomic peri-arterial renal nervous system to date. We demonstrated a greater abundance of efferent compared with afferent nerve fibers and that the proportion of afferent nerve fibers is not different between the proximal, middle, and distal segments. Recently, Tellez et al. (8) reported that efferent fibers are far greater than afferent fibers in pigs, which aligns with our findings in humans. The renal efferent sympathetic nerves innervate the 3 major neuroeffectors in the kidney (9). The stimulation of beta1-adrenoreceptors on juxtaglomerular granular cells increases renin secretion rate, the stimulation of alpha1B-adrenoreceptors on renal tubular epithelial cells increases renal tubular sodium reabsorption, and the stimulation of alpha1A-adrenoreceptors on the renal arterial resistance vessels decreases renal blood flow.
Although renal afferent sensory nerves are not adrenergic in nature, they play an important role in the central sympathetic regulation (10). Radiofrequency reduction of peri-arterial renal autonomic nerves has been an important therapeutic innovation for the treatment of resistant hypertension. Our results suggest that the nonselective partial renal denervation of both efferent and afferent nerve fibers is likely to play a causative role for the effective treatment of resistant hypertension, which has been achieved in the Symplicity HTN-1 and -2 and the EnligHTN I trials (2,11) and other non–BP-related effects (12). However, a population of 10% to 15% nonresponders remains, which may suggest that a certain critical threshold of nerve injury is required or that sympathetic overdrive may not be the only mechanism of resistant hypertension (13).
Our results could have a substantial clinical impact on refinement of renal denervation strategies on multiple levels. First, application of thermal energy could be focused on proximal and middle arterial segments in which peri-arterial sympathetic nerve fibers are concentrated. However, the distance from the arterial lumen to nerve is shorter in distal arterial segments, which suggests less thermal energy may be required to achieve nerve damage. Further studies will be required to address these questions.
With respect to the current guideline (13), >40% of hypertensive patients are anatomically ineligible for renal denervation therapy (14). The prevalence of duplicated renal arteries (main artery + accessory or polar [aberrant] artery) is as high as 20% of hypertensive patients. Therefore, it is debatable whether radiofrequency ablation of renal sympathetic nerves should be extended to accessory/polar arteries (15). In addition, it has been reported that the efficacy of renal sympathetic denervation procedures is less pronounced in patients with accessory/polar arteries compared with those without. However, greater BP reduction was observed in the denervated accessory artery group compared with the incompletely denervated accessory group. This novel finding may likely be explained by the fact that the effective nerve area targeted by radiofrequency ablation is probably smaller in patients with accessory/polar arteries because these have not been deemed suitable for this treatment to date. Although accessory artery size is relatively small (diameter 2.3 ± 0.3 mm), we demonstrated the existence of nerves around accessory renal arteries, which could be targeted for renal denervation in the future. The cumulative distribution of nerve distance from renal arterial lumen provided in the present study shows that >75% of sympathetic nerves are located within a distance of 4.28 mm, which provides meaningful insights into the treatment zone that should be targeted by renal denervation procedures. Our study further shows that renal nerves are least concentrated in the dorsal region compared with the ventral, superior, and inferior regions.
Although we investigated a large number of nerves, the number of autopsy cases was limited (n = 20 [with only 4 women]). Thus, there is a chance of not being able to detect significant differences due to sample size of hypertensive versus nonhypertensive patients, and male versus female subjects. However, because this is the first human study comparing nerve anatomy between patients with and without hypertension, valuable information can be retrieved from our findings. In addition, the number of accessory arteries was limited (n = 6), and there was no polar (aberrant) arteries in the study samples. Immunofluorescence for efferent and afferent nerves was performed in a limited number of sections (n = 15). Despite these limitations, we observed a significant difference in staining intensity between efferent versus afferent nerves. Nerve identification and measurements were performed on H&E-stained sections, which is not a specific stain for nerve fibers. However, the correlation between the number of nerves identified on H&E-stained sections and those identified by using immunohistochemistry was excellent in selected sections. Although pressure-perfusion fixation using 10% formalin was performed in the present set of samples, tissue shrinkage must have occurred, which is estimated to be close to 20% (4). However, these adjustments were not added to our data because of age dependency of tissue shrinkage and ex vivo assessment.
Compared with proximal and middle renal segments, distal segments had significantly fewer mean number of peri-arterial renal nerves. The total number of nerves in the dorsal arterial region was less than in the ventral region. The distribution of the distance of nerves from the renal arterial lumen varied considerably, from <1 mm to >10 mm; however, the 75th percentile of the distance was 4.28 mm. Peri-arterial renal nerves were dominantly composed of efferent rather than afferent fibers, and the relative proportion of afferent fibers was not different between the proximal, middle, and distal segments. Peri-arterial renal nerve anatomy among hypertensive and nonhypertensive patients was not different. Understanding these anatomic characteristics is important for the development of renal denervation devices.
COMPETENCY IN MEDICAL KNOWLEDGE: Hypertension is resistant to antihypertensive treatment in up to 15% of hypertensive patients, and renal sympathetic denervation is an investigational therapy for patients with resistant hypertension.
TRANSLATIONAL OUTLOOK: The relationship between the anatomic location of sympathetic nerves within the renal arteries (proximal, middle, or distal and ventral or dorsal) may influence the amount of energy required to achieve catheter-based renal denervation. Further research is needed to determine the impact of these anatomic factors on the antihypertensive efficacy of the procedure.
The authors thank Kai Shen and Torie Samuda (CVPath) for their excellent technical assistance.
For a supplemental methods section, tables, and figures, please see the online version of this article.
This work was supported in part by Medtronic Cardiovascular (Santa Rosa, California), but the manuscript was prepared independently by CVPath Institute, Inc., a private nonprofit research organization. Dr. Sakakura is supported by a research fellowship from the Banyu Life Science Foundation International; and has received speaking honoraria from Abbott Vascular, Boston Scientific, and Medtronic Cardiovascular. Dr. Virmani has received research support from 480 Biomedical, Abbott Vascular, Atrium, Biosensors International, Biotronik, Boston Scientific, Cordis J&J, GlaxoSmithKlinehttp://dx.doi.org/10.13039/100004330, Kona, Medtronichttp://dx.doi.org/10.13039/100004374, Microport Medical, OrbusNeich Medical, ReCor, SINO Medical Technology, Terumo Corporation, and W.L. Gore; has speaking engagements with Merck; receives honoraria from 480 Biomedical, Abbott Vascular, Biosensors International, Boston Scientific, CeloNova, Claret Medical, Cordis J&J, Lutonix, Medtronic, Terumo Corporation, and W.L. Gore; and is a consultant for 480 Biomedical, Abbott Vascular, Medtronic, and W.L. Gore. Dr. Joner is a consultant for Biotronik and Cardionovum; and has received speaking honoraria from Abbott Vascular, Biotronik, Cordis J&J, Medtronic, and St. Jude. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- blood pressure
- calcitonin gene–related peptide
- hematoxylin and eosin
- neurofilament protein
- tyrosine hydroxylase
- Received December 19, 2013.
- Revision received March 1, 2014.
- Accepted March 11, 2014.
- American College of Cardiology Foundation
- Rocha-Singh K.
- Maleszewski J.,
- Lu J.,
- Fox-Talbot K.,
- Halushka M.K.
- DiBona G.F.
- Krum H.,
- Sobotka P.,
- Mahfoud F.,
- Bohm M.,
- Esler M.,
- Schlaich M.
- Worthley S.G.,
- Tsioufis C.P.,
- Worthley M.I.,
- et al.
- Mahfoud F.,
- Luscher T.F.,
- Andersson B.,
- et al.
- Rimoldi S.F.,
- Scheidegger N.,
- Scherrer U.,
- et al.
- Id D.,
- Kaltenbach B.,
- Bertog S.C.,
- et al.