Dynamic Narrowing of the Diaphragmatic Vena Cava in Ovis Aries

Summary

Dorset sheep (ovis aries) are common models in translational cardiovascular research due to physiologic and anatomic similarities to humans. While employing ovine subjects to study single-ventricle physiology, we repeatedly observed position-based changes in central venous pressure (CVP) which could not be explained by hydrostatic (gravitational) effects. Inferior vena cava (IVC) narrowing or compression has been demonstrated in numerous species, and we hypothesized that this phenomenon might explain our observations in ovis aries. This study aimed to characterize position-dependent morphology of the IVC in ovis aries using catheter-based hemodynamic and dimensional measurements, three-dimensional MRI reconstruction, and histological analysis. Baseline measurements revealed a significant reduction in IVC dimensions at the level of the diaphragm (dVC) compared to the abdominal vena cava (aVC) and thoracic vena cava (tVC). We also observed a transdiaphragmatic pressure gradient along the IVC, with higher pressures in the aVC compared to the tVC. We found that variation of position and fluid status altered IVC hemodynamics. Histological data showed variable muscularity along the length of the IVC, with greater smooth muscle content in the aVC than the tVC. These findings will improve understanding of baseline ovine physiology, help refine experimental protocols, and facilitate the translation of findings to the clinic.

Introduction

Cardiovascular disease (CVD) is the leading cause of premature death and lifelong morbidity worldwide. Global incidence continues to rise despite significant advancements in prevention, diagnosis, and treatment over the last century (Braunwald, 2012; Martin et al., 2024). Animal models are an indispensable medium for discerning the broad physiologic implications of CVD, and an improved understanding of baseline physiology of these models will enable more accurate interpretation of findings (DiVincenti et al., 2014; Quinn, 2013; Zaragoza et al., 2011).

Ovis aries are a common model in translational cardiovascular research because they exhibit similarities to humans in several properties, including blood volume, heart rate, intracardiac pressure, distribution of blood components, and platelet aggregation profiles (DiVincenti et al., 2014; Quinn, 2013). Ovine ability to withstand anesthesia and cardiopulmonary bypass makes them excellent surgical candidates (Quinn, 2013). Ovine models have proven invaluable to our group and others in the assessment of single-ventricle physiology and development of cardiovascular prosthetics (Blum et al., 2022; DiVincenti et al., 2014; Park et al., 2022; Schleimer et al., 2018; Shinoka et al., 1995).

During hemodynamic evaluation of ovine subjects in various unrelated studies, we repeatedly observed a 10–15 mmHg shift in central venous pressure (CVP) upon switching from left to right lateral recumbency and vice versa, which appeared to be independent of hydrostatic (gravitational) effects. Existing literature suggests that this phenomenon may be attributable to diaphragmatic and/or visceral organ compression of the inferior vena cava (IVC), which has been previously shown to induce vessel narrowing in numerous species (Bauman et al., 2015; Doppman et al., 1966; Gasthuys et al., 1991; Rosenthal, 1998; Rubinson, 1967; Terada & Takeuchi, 1993; Wachsberg, 2000; Youngblood et al., 2020). Intraabdominal pressure (IAP) and animal position have been identified as modulators of this phenomenon (Bauman et al., 2015; Doppman et al., 1966; Gasthuys et al., 1991; Klein & Sherman, 1977; Masey et al., 1985; Meyer et al., 2010; Rosenthal, 1998; Rubinson, 1967; Terada & Takeuchi, 1993; Trimmel, Hierweger, et al., 2022; Trimmel, Podgoršak, et al., 2022a, 2022b; Vesal & Karimi, 2006; Wachsberg, 2000; Wagner et al., 1990; Youngblood et al., 2020).

Ovine IVC narrowing has been suggested to occur but never shown (Masey et al., 1985; Trimmel, Podgoršak, et al., 2022b). One study has assessed the influence of IAP on central venous hemodynamics in ovis aries (Masey et al., 1985), and two studies anecdotally evaluated the effect of position on CVP (Trimmel, Podgoršak, et al., 2022b; Vesal & Karimi, 2006). However, to date there is no unified understanding of how animal position influences IVC dimensions and central venous hemodynamics in ovis aries.

The presence of a transient, physiologic narrowing of the IVC would have important implications for the use of ovine models in cardiovascular research as this phenomenon could lead to unpredictable or incorrect interpretation of various clinical parameters, such as CVP (Hutchinson & Shaw, 2016). In this report, we characterize a physiologic narrowing of the diaphragmatic vena cava in ovis aries. Our results will better inform researchers’ baseline understanding of central venous hemodynamics in ovine subjects and may prove useful in evaluating data from preclinical studies.

Materials and Methods

Animal Subjects

Animal work was performed according to the ARRIVE guidelines, approved by the Institutional Animal Care and Use Committee (IACUC) at the Abigail Wexner Research Institute, and conducted in accordance with the Guide for the Care and Use of Laboratory Animals, 8th edition. The research was conducted at a USDA-licensed and AAALAC-International-accredited facility. Animals were kept in a 12-hour light:dark cycle and were housed in groups except during the immediate post-operative recovery period.

A retrospective cohort of 21 Dorset ovine subjects (Ovis aries) ranging from 5–7 years of age, and prospective cohort of 3 surgically naïve Dorset ovine subjects ranging from 0.3–1 years of age were used for this study (Table 1). Of the retrospective cohort, 2 subjects were surgically naïve and the remaining 19 subjects had received intrathoracic IVC interposition conduit tissue-engineered vascular graft (TEVG) implants (Table 1). Our group previously showed that TEVG scaffolds remodel into autologous neovessels by one-year post-implantation in this surgical model (Blum et al., 2022). The 19 TEVG recipients had received interventional imaging at least 1-year from operation and met strict inclusion criteria (i.e., <25% TEVG narrowing) to ensure geometry of the TEVG should have no consequential influence on vena cava hemodynamics. The 19 TEVG recipients that met inclusion criteria for this study were selected from an original cohort of 114 TEVG recipients.

Table 1: Cohort Sizes for Baseline Measurements, Positional Studies, and Bolus Studies.

Age ranges and sample sizes for hemodynamic and dimensional characterization studies. All cohorts were comprised of a mixture of subjects with naïve anatomy and subjects with tissue-engineered vascular graft (TEVG) implants. All TEVG subjects were at least 1-year from operation and demonstrated >75% graft patency at the time of measurement acquisition.

Cohort	Total Subjects	Naïve Anatomy	TEVG Implant	Age Range
Baseline Measurements	21	2	19	1.8–7 years
Positional Characterization	3	3	0	0.3–1 year
Fluid Status Characterization	5	2	3	1.8–7 years

Pressure Catheterization and Intravascular Ultrasound

19 subjects with TEVG implants and 2 with naïve anatomy underwent catheterization studies which were retrospectively analyzed for this study. For subjects with TEVG implants, only catheterization data from studies ≥1 year post-operation was utilized, per inclusion criteria. In these studies, subjects were sedated using diazepam (Hospira Pharmaceuticals, Lake Forest, IL), butorphanol (Zoetis, Parsippany, NJ), and ketamine (Hospira), and then intubated and placed on left lateral recumbency. A 9-French sheath (Terumo Medical Corporation, Somerset, NJ) was inserted via the right internal jugular vein and an intravenous bolus of heparin (Hospira) was administered (150 U/kg). A 5-French Judkins right (JR) 2.5 catheter (Cook Medical, Bloomington, IN) was guided into the internal jugular vein through the superior vena cava (SVC) into the right atrium. Using an angled guidewire (Glidewire, Terumo Medical Corporation, Ann Arbor, MI), the JR catheter was then passed through the thoracic vena cava (tVC) into the abdominal IVC (aVC) where a Rosen exchange guidewire (Cook Medical) was placed. A digital angiogram was obtained by injecting contrast agent 68% through a multitrack angiographic catheter (B. Braun Medical Inc, Bethlehem, PA) positioned in the aVC. Pressure measurements were obtained in the aVC, tVC, and right atrium (RA) or SVC (Figure 1). For catheter-based pressure measurements, the zero-point was set at the level of the subject’s heart while on left lateral recumbency (Figure 2A). For dimensional measurements, an intravascular ultrasound (IVUS) catheter (Volcano Corporation, San Diego, CA) was advanced over the Rosen guidewire to obtain images within the aVC, diaphragmatic vena cava (dVC), and tVC (Figure 2B). Images were analyzed using Volcano software to obtain a cross-sectional area (CSA) and circumference as described previously (Pepper et al., 2017).

Figure 1.

Diagram depicting basic ovine anatomical structures discussed in the study. Regions of the vena cava are labeled as abdominal vena cava (aVC), diaphragmatic vena cava (dVC), thoracic vena cava (tVC), and superior vena cava (SVC). The dVC, where maximum vessel narrowing occurs, is located at the point where the vena cava passes through the diaphragm and is concealed in this image. The right atrium (RA) of the heart is also abbreviated.

Figure 2.

A.) An angiogram with labels defining the abdominal (aVC), diaphragmatic (dVC), and intrathoracic (tVC) regions of the inferior vena cava. The diaphragm is marked by the orange dotted line. Directionality of flow is indicated by an arrow pointing toward the right atrium. B.) Representative image of an intravascular ultrasound (IVUS). The vessel wall is indicated using a yellow dotted line. The probe used for dimensional measurements is indicated with a red dotted line. C.) Three-dimensional reconstruction of the vena cava from the aVC below the hepatic veins through the right atrium junction. Regions of the vena cava are marked with brackets and arrows point to the hepatic veins and point of maximal narrowing. D.) Baseline aVC and tVC pressure measurements in subjects (n=21) on left lateral recumbency. A Wilcoxon matched-pairs signed rank test was used for comparison (p<0.001). E.) Baseline measurements of aVC, dVC, and tVC cross-sectional area in subjects (n=21) on left lateral recumbency. Friedman’s Test with Dunn’s multiple comparisons was used for comparison (p>0.999, p<0.001, p<0.001). F.) Baseline measurements of aVC, dVC, and tVC circumference in subjects (n=21) on left lateral recumbency. Friedman’s Test with Dunn’s multiple comparisons was used for comparison (p>0.999, p<0.001, p<0.001).

Bolus Studies

Five subjects, including two with naïve anatomy and three with TEVG implants, underwent 20mL/kg 0.9% NaCl intravenous fluid bolus studies to assess the impact of hydration status on IVC hemodynamic performance. Subjects ranged in weight from 51.5–102 kg. Subjects underwent baseline (pre-bolus) catheterization and IVUS imaging on left lateral recumbency. Measurements were repeated immediately after complete bolus administration. Results were retrospectively analyzed.

Positional Catheterization

Three surgically-naïve subjects underwent prospective positional catheterization studies using the procedures described above. Pressure measurements were obtained in the SVC, tVC, and aVC with subjects on left lateral and then dorsal recumbency, allowing two minutes for hemodynamic equilibration following position changes. The catheter zero-point was not reset between positions.

MRI

One surgically-naïve subject underwent a cardiac MRI while on right lateral recumbency using a 3-Tesla magnet (Siemens Medical Solutions, Erlangen, Germany) with a 32-channel phased-array cardiac coil. Contrast (Ferumoxytol 3 mg/kg body weight, Cambridge, MA) was injected intravenously for magnetic resonance angiography (MRA). Flash IR sequence was run, and MRI files were imported into Elucis software (Realized Medical, Ottowa, Canada) and segmented to provide a three-dimensional representation of the IVC (Figure 2C).

Histological Characterization of the IVC

Tissue samples containing the IVC with liver were harvested and fixed with 4% formalin for 1 week, then transferred to 70% ethanol for long-term storage. Upon removal from ethanol, explants were processed through xylene and paraffin embedded. 4 μm transverse histological sections of the tVC and aVC and a longitudinal section of the dVC were mounted on slides and heat-fixed. Standard techniques for hematoxylin and eosin (H&E) were used to visualize cell density and location. A modified Russell-Movat’s pentachrome procedure was used to determine constituents of the vessel (StatLab, McKinney, TX). Immunohistochemistry (IHC) was used to identify α-smooth muscle actin (α-SMA)-positive cells. Slides were deparaffinized in xylene and rehydrated in graded alcohol. Anti-α-SMA antibody (Agilent Dako, Santa Clara, CA; 1:2000) was applied overnight at 4ºC. Primary antibody binding was detected using goat anti-mouse IgG biotinylated antibody (Vector Laboratories, CA, USA, 1:1500) followed by avidin-horseradish peroxidase (Vector). After antibody staining, slides underwent chromogenic development and counterstaining as previously described (Blum et al., 2022). 5x images were obtained with a Zeiss Axio Imager.A2 microscope (Carl Zeiss, Oberkochen, Germany). 10x images were obtained from Histowiz (Brooklyn, NY).

Statistical Analysis

Data was analyzed using GraphPad Prism 9.4.1 (GraphPad Software, Boston, MA). A p-value of < 0.05 was considered significant. Data were tested for normality using a Shapiro-Wilk test to select the appropriate statistical methods. A Wilcoxon matched-pairs signed rank test was used to assess variation in baseline pressure measurements in the aVC and tVC, position-based change in transdiaphragmatic pressure gradient, and distensibility of the aVC versus tVC. Friedman’s Test with Dunn’s multiple comparisons was used to compare baseline CSA and circumference of the aVC, dVC, and tVC. Multiple Wilcoxon matched-pairs signed rank tests were used to compare position-based pressure changes in the aVC, dVC, and tVC. Multiple paired t-tests were used to compare pressure, CSA, and circumference changes in specific regions of central venous circulation as a function of fluid status. A paired t-test was used to assess variation of transdiaphragmatic pressure gradient as a function of fluid status as well as compliance and absolute pressure changes in the aVC versus tVC. An unpaired t-test was used to compare α-SMA positive area in the aVC and tVC samples.

Results

Baseline Physiology

Baseline pressure measurements conducted on left lateral recumbency demonstrated significantly higher blood pressures in the aVC than tVC (p<0.001, Wilcoxon test) (Figure 2D). Average aVC pressure was 11.7±2.7 mmHg while average tVC and RA pressures were 0.9±1.3 mmHg and 0.8±1.3 mmHg respectively (Table 3). The transdiaphragmatic pressure gradient was calculated as the difference between aVC pressure and tVC pressure and was 10.8±3.0 mmHg (Range: 6.0–16.5 mmHg). Further evaluation via IVUS allowed for characterization of vena cava dimensions within the abdominal, diaphragmatic, and thoracic regions (Figure 2E–F, Table 4). Both CSA and circumference (p<0.001, Friedman’s Test with Dunn’s multiple comparisons) of the dVC were significantly reduced compared to both the aVC and tVC. There was no difference between the aVC and tVC for either CSA or circumference (p>0.999, Friedman’s Test with Dunn’s multiple comparisons) (Figure 2E–F, Table 4). This indicates, for ovine subjects on left lateral recumbency, a reduction in IVC dimensions as the vessel exits the abdomen and crosses the diaphragm before it reopens in the thoracic cavity. This experimental data is supported by a 3D reconstruction of an ovine subject that underwent MRI scanning on right lateral recumbency, which demonstrates a narrowing of the IVC as it crosses the diaphragm (Figure 2C, Supplemental Movie). We next characterized physiologic variability of this narrowing under altered position and fluid status conditions.

Table 3: Vena Cava Pressure Measurements Pre- and Post-Bolus Administration.

Pre- and post-20 mL/kg 0.9% NaCl bolus measurements of pressure in the abdominal vena cava, thoracic vena cava, and right atrium. Transdiaphragmatic pressure gradient was calculated as the difference between abdominal vena cava and thoracic vena cava pressures. P-value indicates the result of a comparison between pressure measured pre-bolus and post-bolus at each anatomic region. Multiple paired t-tests were used to analyze aVC, tVC and RA pressure variability. A single paired t-test was used to analyze transdiaphragmatic gradient pressure variability.

Location	Pre-Bolus Pressure (mmHg) (N=21)	Post-Bolus Pressure (mmHg) (N=5)	P-value
Abdominal Vena Cava	11.7 ± 2.7	18.5 ± 4.0	0.004
Thoracic Vena Cava	0.9 ± 1.3	12.8 ± 5.2	0.003
Right Atrium	0.8 ± 1.3	10.5 ± 3.9	0.003
Transdiaphragmatic Gradient	10.8 ± 3.0	5.7 ± 4.6	0.003

Table 4: Vena Cava Dimensions Pre- and Post-Bolus Administration.

Pre- and post-20 mL/kg 0.9% NaCl bolus measurements of cross-sectional area (CSA) and circumference within the abdominal vena cava, diaphragmatic vena cava, and thoracic vena cava. P-value indicates the result of a comparison between dimensions (CSA or circumference) measured pre-bolus and post-bolus at each anatomic region (multiple paired t-tests).

Location	Pre-Bolus CSA (mm2) (N=21)	Post-Bolus CSA (mm2) (N=5)	P-value
Abdominal Vena Cava	216.7 ± 65.3	271.4 ± 65.7	0.064
Diaphragmatic Vena Cava	73.3 ± 34.8	152.6 ± 93.9	0.145
Thoracic Vena Cava	233.4 ± 88.4	388.6 ± 176.4	0.079
Location	Pre-Bolus Circumference (mm) (N=21)	Post-Bolus Circumference (mm) (N=5)	P-value
Abdominal Vena Cava	58.0 ± 7.6	63.1 ± 7.0	0.091
Diaphragmatic Vena Cava	40.3 ± 7.0	51.0 ± 17.3	0.227
Thoracic Vena Cava	59.4 ± 10.2	74.2 ± 13.5	0.091

Effect of Position on IVC Anatomy

The effect of position on the transdiaphragmatic narrowing of the IVC was prospectively characterized by sequential catheterizations of three subjects on left lateral and then dorsal recumbency (Figure 3, Table 2). As subjects were moved from the left lateral to dorsal recumbency, marked pressure changes were seen in all evaluated regions of the vena cava (aVC, tVC, SVC). These did not reach statistical significance (all p=0.253, Wilcoxon test), but this would be expected with a small sample size. There appeared to be consistent trends with a decrease in aVC pressure and concomitant increase in tVC and SVC pressures as the subject was moved from the left lateral recumbency to dorsal recumbency. There was also a notable trend in pressure gradient that did not reach significance (p=0.250, Wilcoxon test), with a reduction in gradient magnitude as subjects were moved from the left lateral recumbency to the dorsal recumbency (Figure 3B). In these three subjects, the transdiaphragmatic pressure gradient decreased by 6–10 mmHg as the position changed from the left lateral recumbency to dorsal recumbency. Average transdiaphragmatic pressure gradients measured 10.0±2.0 mmHg and 2.3±0.6 mmHg in the left lateral and dorsal recumbencies respectively. Although limited by cohort size, these results suggested that position influences dVC narrowing, with improved pressure equilibration across the IVC on dorsal recumbency.

Figure 3.

A.) Transdiaphragmatic pressure gradient measured sequentially in subjects (n=3) on left lateral and dorsal recumbencies. Gradient is calculated as the difference between abdominal vena cava pressure (aVC) and thoracic vena cava (tVC) pressure. The gradients measured in the two positions were compared using a Wilcoxon matched-pairs signed rank test (p=0.250). B.) Blood pressure in the aVC, tVC, and superior vena cava (SVC) sequentially measured in subjects (n=3) on left lateral and dorsal recumbencies. Pressures in each region were compared across positions using multiple Wilcoxon matched-pairs signed rank tests (p=0.253 for all).

Table 2: Vena Cava Pressure Measurements on Left Lateral and Dorsal Recumbencies.

Pressure measurements in the abdominal vena cava, thoracic vena cava, and superior vena cava on left lateral and dorsal recumbencies. Transdiaphragmatic pressure gradient was calculated as the difference between abdominal vena cava and thoracic vena cava pressures. P-value indicates the result of a comparison between pressure measured on left lateral recumbency and dorsal recumbency at each anatomic region (Wilcoxon test).

Location	Left Lateral Recumbency Pressure (mmHg) (N=3)	Dorsal Recumbency Pressure (mmHg) (N=3)	P-value
Abdominal Vena Cava	10.3 ± 2.1	7.0 ± 2.6	0.253
Thoracic Vena Cava	0.3 ± 0.6	4.7 ± 2.3	0.253
Superior Vena Cava	1.0 ± 1.0	4.3 ± 2.1	0.253
Transdiaphragmatic Gradient	10.0 ± 2.0	2.3 ± 0.6	0.250

Effect of Fluid Status on IVC Anatomy

Fluid bolus studies were retrospectively analyzed to characterize the effect of fluid status on the diaphragmatic narrowing of the IVC. Following 20mL/kg bolus administration, there was a significant increase in pressure of the aVC (p=0.004), tVC (p=0.003), and RA (p=0.003) (multiple paired t-tests) (Figure 4A, Table 3). The average transdiaphragmatic pressure gradient decreased significantly (p=0.003, paired t-test) after bolus injection, moving from 10.8±3.0 mmHg to 5.7±4.6 mmHg in the pre-bolus and post-bolus states respectively, indicating better pressure equilibration between the aVC and tVC post-bolus (Figure 4B, Table 3). Notably, pressure was increased by 4.6 ± 2.4 mmHg in the aVC and 11.2±4.0 in the tVC, a significantly larger increase in the tVC (p=0.003, paired t-test) (Figure 4C). Dimensional measurements of the aVC, dVC, and tVC consistently increased post-bolus, but these variations did not achieve significance for CSA (p=0.064, p=0.145, p=0.079) or circumference (p=0.091, p=0.227, p=0.091) (Figure 4D–E, Table 4) (multiple paired t-tests). Cross-sectional compliance and distensibility were calculated and compared for the aVC and tVC using Equations 1 and 2, where A = cross-sectional area, P = pressure, and C = circumference (Figure 4F–G).

Figure 4.

A.) Blood pressure in the abdominal vena cava (aVC), thoracic vena cava (tVC), and right atrium (RA) sequentially measured in subjects before and after administration of a 20mL/kg 0.9% NaCl saline bolus (n=21 for pre-bolus baseline, n=5 for post-bolus). Pressures in each region were compared across bolus administration using multiple paired t-tests (p=0.004, p=0.003, p=0.003). B.) Transdiaphragmatic pressure gradient sequentially measured in subjects before and after bolus administration (n=5). Gradient is calculated as the difference between aVC and tVC pressure. Gradients were compared across bolus administration using a paired t-test (p=0.003). C.) Magnitude of absolute pressure change in the aVC and tVC as a result of bolus administration (n=5). Comparisons were done using a paired t-test (p=0.003). D.) Cross-sectional area (mm²) of the aVC, diaphragmatic vena cava (dVC), and tVC sequentially measured in subjects before and after bolus administration (n=5). Dimensions in each region were compared across bolus administration using multiple paired t-tests (p=0.064, p=0.145, p=0.079). E.) Circumference (mm) of the aVC, dVC, and tVC sequentially measured in subjects before and after bolus administration (n=5). Dimensions in each region were compared across bolus administration using multiple paired t-tests (p=0.091, p=0.227, p=0.091). F.) Cross-sectional compliance (mm²/mmHg) of the aVC and tVC calculated as a function of pressure and dimension changes in these regions after bolus administration (n=5). A paired t-test was used for comparison (p=0.815). G.) Distensibility (mmHg⁻¹) of the aVC and tVC calculated as a function of pressure and dimension changes in these regions after bolus administration (n=5). A Wilcoxon matched-pairs signed rank test was used for comparison (p=0.625).

$$\text{Cross–Sectional}\text{Compliance}=\frac{\text{ΔA}}{\text{ΔP}}$$ (Equation 1)

$$\mathrm{D}\mathrm{i}\mathrm{s}\mathrm{t}\mathrm{e}\mathrm{n}\mathrm{s}\mathrm{i}\mathrm{b}\mathrm{i}\mathrm{l}\mathrm{i}\mathrm{t}\mathrm{y}=\frac{(\frac{\Delta \mathrm{C}}{{\mathrm{C}}_{\mathrm{m}\mathrm{i}\mathrm{n}}})}{\Delta \mathrm{P}}$$ (Equation 2)

Pressure measurements were not available for the dVC, so compliance and distensibility could not be calculated for this region. Despite the significant difference in absolute pressure change between the aVC and tVC after bolus administration, there was no difference in compliance (p=0.815, paired t-test) or distensibility (p=0.625, Wilcoxon test) measurements between these two regions (Figure 4F–G).

Histology

For tissue staining and analysis, the IVC was segmented into three regions corresponding with hemodynamic and dimensional analysis: abdominal, diaphragmatic, and thoracic (Figure 5A). Pentachrome staining revealed a difference in muscle composition along the length of the IVC. α-SMA staining revealed greater smooth muscle cell content in the aVC compared to tVC (Figure 5B) (p<0.0001, unpaired t-test). Longitudinal sections of the transdiaphragmatic vena cava showed the change in smooth muscle composition from aVC to tVC (Supplemental Figure 1).

Figure 5.

Histological characterization of the ovine inferior vena cava (IVC). A.) Hematoxylin and eosin (H&E), Movat’s pentachrome, and alpha smooth muscle actin (αSMA) staining of thoracic (tVC), diaphragmatic (dVC), and abdominal (aVC) sections. The IVC was segmented into the tVC, dVC, and aVC components. Adventitia, media, and intima of the transverse sections of tVC and aVC are indicated using arrows. The dVC sample was taken from a longitudinal section where D-dVC is diaphragm-facing and the HV-dVC is hepatic vein-facing. H&E staining reveals nuclei (purplish-blue) and extracellular matrix and cytoplasm (pink). Pentachrome staining reveals muscle (red), collagen and reticular fibers (yellow), and nuclear and elastic fibers (black). Sections taken from same subject were used for representative images of the entire cohort. B.) Quantitation of α-smooth muscle actin (αSMA) is expressed as fractional area staining of the aVC (n=4) and tVC (n=3). Differences in αSMA staining between sections were evaluated using an unpaired t-test (p<0.0001).

Discussion

This study aimed to characterize a narrowing of the ovine diaphragmatic vena cava (dVC) and evaluate its variability in response to position and fluid status. We employed a comprehensive approach, utilizing catheter-based hemodynamic and dimensional measurements, 3D MRI reconstruction, and histological analysis.

At baseline, on left lateral recumbency we observed significantly reduced CSA and circumference of the IVC at the level of the diaphragm. This finding was supported by 3D MRI reconstructions of the IVC in subjects positioned on right lateral recumbency. Dimensional findings were corroborated by hemodynamic measurements, as the transdiaphragmatic pressure gradient displayed significantly higher pressure in the aVC than the tVC. Narrowing of the dVC was influenced by position and fluid status.

Positional studies revealed a notable decrease in the transdiaphragmatic pressure gradient as subjects were transitioned from left lateral recumbency to dorsal recumbency. Pressure characterization in discrete regions of the vena cava also showed marked differences, with an apparent decrease in aVC pressure and increases in tVC and SVC pressure as subjects transitioned from left lateral to dorsal recumbency. These findings did not reach statistical significance, however based on trend consistency across a limited cohort (n=3), we anticipate that analysis of a larger cohort would demonstrate similar, significant findings. Our results demonstrate a high pressure (>10 mmHg) vena cava region caudal to the diaphragm and low-pressure (0–1 mmHg) vena cava region cranial to the diaphragm for subjects on left lateral recumbency. Although this portion of our study did not yield significant findings, we propose that the pressure contrast between high- and low-pressure regions is diminished on dorsal recumbency. In this position, the high-pressure region caudal to the diaphragm shifts down to 6–8 mmHg and the low-pressure region cranial to the diaphragm shifts up to 4–5 mmHg, indicating better equilibration. This supports the conclusion that dVC narrowing is more pronounced on left lateral recumbency and thus sensitive to position.

Position-based CVP alterations unattributable to hydrostatic pressure have been noted in numerous species, including ovine subjects (Klein & Sherman, 1977; Meyer et al., 2010; Rosenthal, 1998; Terada & Takeuchi, 1993; Trimmel, Podgoršak, et al., 2022b). These effects were partly attributed to variable IVC compression by visceral organs and the central tendon of the diaphragm (Klein & Sherman, 1977; Meyer et al., 2010; Rosenthal, 1998; Terada & Takeuchi, 1993; Trimmel, Hierweger, et al., 2022; Trimmel, Podgoršak, et al., 2022b). Notably, in most studies, subjects were anesthetized; the influence of anesthesia on positional CVP effects is unclear, with studies in non-anesthetized bovine and ovine subjects reporting conflicting results (Meyer et al., 2010; Tagawa et al., 1994; Vesal & Karimi, 2006; Wagner et al., 1990). The only ovine study examining positional CVP variability was conducted in non-anesthetized subjects, and CVP was 3–4 mmHg both when on right lateral recumbency and standing (Vesal & Karimi, 2006). The authors mention that after anesthetizing one subject, CVP in the right lateral recumbency rose to 7–8mmHg (Vesal & Karimi, 2006). While this finding was anecdotal (n=1) and the directionality does not agree with our measurements, it indicates that anesthesia may influence or enable positional effects on CVP. Anesthesia can drive ruminal gas accumulation and intraabdominal pressure (IAP) increases in ruminants (Trimmel, Hierweger, et al., 2022). This could lead to compression of the diaphragm and IVC (Meyer et al., 2010; Rosenthal, 1998; Trimmel, Hierweger, et al., 2022). During terminal anesthetization (>30 hours) of ovine subjects on right lateral recumbency, CVP measured 10.8±2.6 mmHg at baseline and increased significantly over time, despite orogastric tube placement for ruminal decompression (Trimmel, Hierweger, et al., 2022). This suggests that prolonged anesthesia exacerbates dVC narrowing via ruminal compression. Indeed, a study in canines associated increased IAP with dVC collapse (Doppman et al., 1966; Rubinson, 1967). A similar study in dorsally positioned ovine subjects showed that as IAP increased from 0–25mmHg, aVC pressure rose from 3–13mmHg and tVC pressure rose from 0–7mmHg (Masey et al., 1985). In this ovine study, the transdiaphragmatic pressure gradient increased alongside IAP (Masey et al., 1985). The baseline transdiaphragmatic pressure gradient in this IAP study agrees with our findings on dorsal positioning, and their findings support a role for IAP in ovine central venous hemodynamics (Masey et al., 1985). Overall, our results align with the existing body of literature, suggesting that positional variation impacts central venous hemodynamics in anesthetized ruminants. Future studies should assess how anesthesia duration affects IAP and CVP measurements for ruminants in various positions.

Saline bolus administration significantly decreased the transdiaphragmatic pressure gradient and significantly increased aVC, tVC, and RA pressures in subjects on left lateral recumbency. The decreased transdiaphragmatic gradient indicates better equilibration of aVC and tVC pressures than under standard conditions. The Hagen-Poiseuille equation (Equation 3) suggests the decrease in pressure gradient could be driven by changes in any or all of three variables: decreased blood viscosity, decreased flow, or increased dVC dimensions (Hirshfeld & Nathan, 2020). In this equation, ΔP = pressure change across the blood vessel length, μ = blood viscosity, L = blood vessel length, Q = volumetric flow rate, and A = cross-sectional area of vessel (Hirshfeld & Nathan, 2020). Blood viscosity is proportional to hematocrit (i.e. red blood cell concentration), which is decreased by bolus administration (Trejo-Soto & Hernández-Machado, 2022). Hematocrit values were not obtained pre- and post-bolus; therefore, we cannot infer how changes to blood viscosity might explain the decreased transdiaphragmatic pressure gradient. Next, decreased blood flow in the vena cava could drive the decreased pressure gradient – we did not directly measure flow so we cannot spectulate on this. Lastly, an increase in dVC dimensions could lead to the decreased post-bolus pressure gradient. However, dimensional characterization showed no significant changes in CSA or circumference in the aVC, dVC, or tVC. Most notably, dVC dimensions did not increase post-bolus. The narrowed dVC separates the aVC and tVC and should dictate the transdiaphragmatic pressure gradient (ΔP) under constant flow and viscosity. This apparent discrepancy could be explained by the Hagen-Poiseuille equation – an increase in vessel area can drive exponentially larger increases in flow. Therefore, insignificant dimensional findings do not rule out the influence of altered vessel morphology when interpreting the reduced transdiaphragmatic pressure gradient.

$$\text{Hagen-Poiseuille}\text{Equation:}\Delta P=\frac{8\mathrm{\mu }\pi LQ}{A^2}$$ (Equation 3)

No studies have characterized how fluid status influences the transdiaphragmatic pressure gradient of the ovine vena cava. However, three studies evaluated RA pressure and aVC CSA in the context of bolus administration in ovis aries; where comparable, our results are in general agreement (Ivey-Miranda et al., 2022; Manavi et al., 2020; Sheridan et al., 2023). In these studies, after administrating 1–2 liters of blood-hydroxyethyl starch mixture, RA pressure increased from 2 to ~7mmHg, and aVC CSA increased from 214mm² to ~315mm² (Ivey-Miranda et al., 2022; Sheridan et al., 2023). In our study, after 1–2 liters of 0.9% saline was administered, RA pressure increased from 2 to 10.5mmHg and aVC CSA increased from 213mm² to 271mm². The slightly larger RA pressure increase and slightly smaller aVC dimension increase in our study might be ascribed to differences in subject position, which these studies do not report. Fluid administration differences may also explain these variations, as in other studies, progressively larger boluses (50mL, 100mL, etc) were given with 2-minute equilibration intervals (Ivey-Miranda et al., 2022; Sheridan et al., 2023). We administered the entire bolus at once, giving vessels less time to compensatorily dilate, potentially leading to smaller dimensional changes and larger hemodynamic effects.

Comparing our positional and bolus studies reveal important differences. First, in positional studies, although all trends are non-significant, an ostensible decrease in transdiaphragmatic pressure is driven by an apparent decrease in aVC pressure and concomitant increase in tVC pressure, a relatively low-pressure system. In fluid bolus studies, the decreased transdiaphragmatic pressure gradient is driven by an increase in tVC pressure that is larger than the increase in aVC pressure, a relatively high-pressure system. Next, position changes appeared to minimize the transdiaphragmatic pressure gradient more effectively than fluid status changes. In positional studies, the smallest transdiaphragmatic pressure gradients were seen in subjects on dorsal recumbency, with an average of 2.33 mmHg. In bolus studies, the transdiaphragmatic pressure gradient was smallest post-bolus, with an average of 5.7 mmHg; two subjects had gradients of 1 mmHg and the remaining three had gradients ranging from 6–11 mmHg. Therefore, for subjects on left lateral recumbency, the transdiaphragmatic pressure gradient seemed to be more effectively reduced by repositioning to dorsal recumbency than administration of a 20mL/kg bolus. Positional changes likely cause significant transdiaphragmatic IVC narrowing, creating a large physiologic gradient in left lateral recumbency that is nearly absent in dorsal recumbency. In comparison, changes in fluid status do not alter vessel narrowing as significantly. Although the dVC narrowing may be lessened after a large fluid bolus, the pressure gradient persists, though reduced. We did not measure vessel dimensions in positional studies, so we cannot assess the validity of this interpretation.

Histologically, the dVC acts as a “transition zone” with decreasing smooth muscle content toward the cranial direction, highlighting the interplay between anatomical structures and the ability of the vena cava to adapt to mechanical constraints within the abdominal cavity. The tVC was primarily composed of collagen with infrequent smooth muscle bundles, with a thin layer of smooth muscle cells in the intima. In contrast, the aVC was very muscular and lacked a distinct border between the intima and the media. There were numerous smooth muscle bundles separated by collagen. Our findings indicate that the ovine vena cava smooth muscle layer is discontinuous and uneven. This observation is consistent with descriptions of the canine inferior vena cava in which the distribution of vascular smooth muscle cells varied across sections, with some areas almost completely lacking a smooth muscle layer (Isayama et al., 2013). These trends are also consistent with findings that, cranial to the diaphragm, the human IVC narrows and changes in muscular and collagen composition; however, these groups do not examine segments far below the liver (Ferraz-de-Carvalho et al., 1994; Fiennes & Du Boulay, 1967; Hashizume et al., 1995). While the precise reasoning for the increased muscularity of the aVC is unknown, there are a few possible explanations. The aVC is surrounded by large, mobile organs including the liver and rumen, so the vessel likely experiences frequent compression, which may explain increased muscle content as an adaptation to retain vessel patency. One observational study suggested that the longitudinal arrangement of the aVC smooth muscle bundles may serve to widen the lumen and help to direct blood flow (Ferraz-de-Carvalho et al., 1994). Alternatively, the aVC muscularity may reflect a need to accommodate higher internal pressures compared to the tVC. Our hemodynamic findings showed that aVC pressures were consistently higher, aligning with the need for a more muscular vascular wall compared to the highly collagenous tVC. Literature describing the functional anatomy of ovine and human segments of the IVC is scarce, and the findings presented herein are observational. Further morphofunctional studies are necessary to understand the nuanced histological differences across the cranial-caudal axis of the vena cava. Lastly, in our histological investigation, we identified a single unicuspid valve-like structure in the IVC at the location of the diaphragm proximal to the intersection of large hepatic veins (Supplemental Figure 2). This was not an isolated finding but was observed in all adult and juvenile subjects. Further characterization was outside the scope of the current investigation but to the best of our knowledge, valve-like structures within the IVC have not been previously reported.

The transdiaphragmatic narrowing of the ovine IVC has significant implications, especially as ovine subjects are a key model organism for cardiovascular research and CVP measurements are critical for clinical management. (DiVincenti et al., 2014; Hutchinson & Shaw, 2016). This study suggests that varying position and fluid status can alter IVC hemodynamics and importantly, downstream central venous pressure (CVP) measurements, potentially affecting clinical management decisions. Alterations in CVP that can be explained by normal physiology may be falsely ascribed to underlying pathology, leading to misinformed treatment algorithms. By characterizing this anatomic feature and its dependence on both position and fluid status, we hope to provide clinicians and researchers with improved context for interpretating data from preclinical studies using ovine subjects.

Supplementary Material

Supinfo1

Supplemental Figure 1. Representative image of a longitudinal section of the dVC stained for αSMA. The intensity of the smooth muscle stain increases from the tVC (left) to aVC (right).

Supinfo2

Supplemental Figure 2. A.) Gross image of vena cava at the level of the liver. Blue arrow indicates the unicuspid valve-like structure. B.) Smooth muscle-stained section of the vena cava indicating thoracic and abdominal sections for directionality. Blue arrow indicates the valve-like structure.

Video

Supplemental Movie: Three-dimensional reconstruction of the vena cava from the aVC below the hepatic veins through the right atrium junction.

Abbreviations/Terminology:

(aVC)

Abdominal vena cava

(dVC)

Diaphragmatic vena cava

(tVC)

Thoracic vena cava

(SVC)

Superior vena cava