Morphometric analysis of the human common hepatic artery reveals a rich and accessible target for sympathetic liver denervation
www.nature.com/scientificreports
OPEN
Morphometric analysis
of the human common hepatic
artery reveals a rich and accessible
target for sympathetic liver
denervation
Abraham Rami Tzafriri1*, Fernando Garcia‑Polite1, John Keating1, Raffaele Melidone1,
Jennifer Knutson1, Peter Markham1, Elazer R. Edelman2,3 & Felix Mahfoud4
This study quantified the distribution of nerves and adjacent anatomies surrounding human common
hepatic artery (CHA) as guidance for catheter based denervation. CHA collected from cadaveric human
donors (n = 20) were histologically evaluated and periarterial dimensions and distributions of nerves,
lymph nodes, pancreas and blood vessels quantified by digital morphometry. Nerve abundance
decreased significantly with distance from the aortic ostium (P < 0.0001) and was higher in the
Superior/Inferior compared to the Anterior/Posterior quadrants (P = 0.014). In each locational group,
nerves were absent from the artery wall, and starting 0.5–1.0 mm from the lumen exhibited a first
order dependence on radial distance, fully defined by the median distance. Median subject-averaged
nerve distance to the lumen was 2.75 mm, ranging from 2.1–3.1 mm in different arterial segments and
quadrants and 2.0–3.5 mm in individuals. Inter-individual variance was high, with certain individuals
exhibiting 50th and 75th nerve distances of, respectively, 3.5 and 6.5 mm The pancreas rarely
approached within 4 mm of the lumen proximally and 2.5 mm more distally. The data indicate that the
CHA is a rich and accessible target for sympathetic denervation regardless of sex and diabetes, with
efficacy and safety most optimally balanced proximally.
Treatment of type 2 diabetes mellitus includes lifestyle modifications and medication r egimens1 that require
concerted efforts from both physicians and patients to be effective owing to high rates of noncompliance. Indeed,
it has been estimated that non-adherence to antidiabetic drugs ranges from 53 to 65% and may be responsible
for uncontrolled glucose levels in about 23% of patients2. Development of alternative approaches that are less
prone to adherence issues could therefore offer a significant advancement in diabetes care. Given the association
of activation of the sympathetic nervous system with altering insulin r esistance3 and the development of type
2 diabetes m
ellitus4, catheter-based sympathetic denervation of the digestive system has been suggested as one
such alternative5.
Physiological and anatomical studies6–9 both suggest the common hepatic artery (CHA) as a promising candidate for such interventions. Such a novel approach can build on the extensive experience with catheter-based
renal denervation as therapy for hypertension10. Despite the recent string of successful clinical studies of renal
denervation using second generation d
evices11–13, it is important to note that this was preceded by the failure
of the first blinded randomized sham controlled study with the first generation single electrode radiofrequency
device14. Postmortem histopathological analysis of human tissue samples demonstrated that radiofrequency
ablation patterns failed to affect the majority of nerves in the renal artery periadventitia15. Moreover, concurrent preclinical animal and computational studies had predicted such procedural limitations based on a careful evaluation of the asymmetric periadventitial distributions of nerves16 and heat conducting and absorbing
microanatomies such as muscles, lymph nodes, and adjacent blood v essels17.
1
CBSET Inc., 500 Shire Way, Lexington, MA 02421, USA. 2IMES, MIT, 77 Massachusetts Avenue, Cambridge,
MA, USA. 3Cardiovascular Division, Brigham and Women’s Hospital, Harvard Medical School, Boston, MA,
USA. 4Department of Internal Medicine III, Saarland University, Homburg/Saar, Germany. *email: rtzafriri@
cbset.org
Scientific Reports |
(2022) 12:1413
| https://doi.org/10.1038/s41598-022-05475-6
1
Vol.:(0123456789)
www.nature.com/scientificreports/
Age, years
All (n = 20)
Diabetic (n = 11)
Nondiabetic (n = 9)
P value
65
66
65
0.8682
Height, inches
67.2
68.5
65.7
0.4074
Sex (male/female)
10/10
6/5
4/5
ns
ns
Race (White/African American)
19/1
10/1
9/0
Overweight (BMI ≥ 25 kg/m2) (yes/no)
12/8
7/4
5/4
ns
Diabetes (yes/no)
11/9
11/0
0/9
–
Hypertension (yes/no)
8/12
4/7
4/5
ns
CKD
18/2
10/1
8/1
ns
Table 1. Patient characteristics. Values are means or n. Categorical data were analyzed by using Fisher’s exact
test (ns denotes P > 0.05). BMI body mass index, CKD chronic kidney disease.
Figure 1. Longitudinal characterization of a representative common hepatic artery. Proximal, middle, and
distal segments shown with H&E stain (A–C) and corresponding TH immunohistochemistry (D–F). H&E
staining shows segments beginning just distal to splenic artery (SA) bifurcation. Prominent lymph nodes (LN)
are present in (B) and (C). Portal vein (PV) branch is indicated. TH staining is strong and diffuse (brown color)
within all nerve bundles. No ganglia are present. A 2 mm scale bar was used for all 6 panels.
The above evidence demonstrated the importance of defining the target microanatomy as a basis for optimizing the design of denervation catheters and treatment p
rotocols18,19. The current study therefore set out to
quantitatively determine the size and number distribution of nerves relative to the lumen of the human CHA, as
a function of radial distance and orientation and the distance from the aortic ostium. As context for the potential
efficacy and safety of CHA denervation, we also quantitatively defined the spatial distributions of periadventitial
lymph nodes and adjacent vessels, as well as the pancreas.
Results
Quantitative histomorphometric evaluation was performed on a total of 20 CHA to determine nerve type,
dimensions, and distribution, as well as characterization of non-target anatomical sites (Table 1). Donors were
48–78 years old (65 ± 8.6) and distributed equally between male (n = 10) and females (n = 10). Per donor medical
histories, eleven of the samples were designated diabetics, 9 non-diabetics. Diabetic and non-diabetic donors
were similarly distributed in terms of sex, age, height, body mass index (BMI), hypertension, and renal disorders.
Male donors were similar to females in age, diabetes, and BMI, but were 6 inches taller (P = 0.002).
Tissue sections across the length of the CHA exhibited abundant nerves within the loose adventitial/periarterial adipose and fibrovascular tissue that were often surrounded by lymph nodes, pancreas and vascular structures, including comparably sized portal veins and smaller arterial and venous branches (Fig. 1). Nerves were
non-myelinated, arranged in thin-walled, irregular fascicles. No associated ganglionic structures were present.
There was strong diffuse Tyrosine hydroxylase (TH) staining of periarterial nerves confirming their predominantly sympathetic nature. Staining of representative CHA sections from each (...truncated)
