Ascites in Animals With Right Heart Failure: Correlation With Lymphatic Dysfunction

Authors: Yoav Dori, Jeremy Mazurek, Edo Birati, Christopher Smith

Abstract

Background

Congestive heart failure is a leading cause of morbidity and mortality worldwide. One of the signs of congestive heart failure is fluid overload including pulmonary edema, peripheral edema, and ascites. The cause of fluid overload remains incompletely understood, and management of these patients continues to be a challenge. The role of lymphatic circulation abnormalities in the cause and pathophysiology of fluid overload also remains unclear. Here we report on a study in a large animal model of right heart failure caused by severe tricuspid regurgitation comparing cardiovascular and lymphatic findings in a group of animals that did not develop ascites with a group of animals that developed ascites.

Methods and Results

Thirteen Yorkshire pigs were included in this study divided into 2 groups. Group 1 included 6 animals that did not develop ascites, and Group 2 included 7 animals that had developed ascites. The groups were compared on hemodynamic parameters as well as comparison of the animal's lymphatic anatomy and function. There was no difference between the groups in degree of tricuspid regurgitation and central venous pressure, with inferior vena cava pressure measuring 11.6±1.6 versus 13.2±3.7 (P=0.534) and superior vena cava pressure measuring 12.0±2.3 versus 13.7±3.2 (P=0.366). There was also no difference between the groups in all measured hemodynamic parameters, including right ventricular pressure, pulmonary artery pressure, and left ventricular function. The weighted liver size in the ascites group was significantly larger than in the nonascites group (30.3±12.4 versus 63.3±14.0 mL/kg, respectively; P=0.001). The 2 groups also differed in the number of animals with regurgitant thoracic duct flow (Group 1/6,17% versus Group 6/7, 86%; P=0.029) and the minimal thoracic duct diameter (Group 2.3±0.3 versus Group 4.2±2.2; P=0.035).

Conclusions

In animals with right heart failure caused by severe tricuspid regurgitation, fluid overload did not correlate with hemodynamic parameters but rather with changes in the lymphatic system, including regurgitant lymphatic flow, minimal thoracic duct diameter, and liver size. This study is consistent with lymphatic dysfunction and not cardiovascular function playing a significant role in the cause of fluid overload. Further studies are needed to confirm these findings.

Clinical PerspectiveWhat Is New? In an animal model of right heart failure, no significant association was demonstrated between measures of heart function or hemodynamic parameters and the likelihood of developing ascites.The factors associated with development of ascites were markers of lymphatic function.Lymphatic valve incompetence and regurgitant lymphatic flow appear to play a key role in development of lymphatic failure leading to fluid overload. What Are the Clinical Implications? Lymphatic dysfunction in patients with heart failure is possibly the key driver that determines which patients will develop symptoms of fluid overload, including edema and ascites.Treatments are needed to address lymphatic dysfunction in these patients.

There are an estimated 26 million patients with heart failure (HF) globally, with about 670 000 new cases a year in the United States alone. ^1^ More than 50% of patients with HF caused by left ventricular dysfunctions are asymptomatic. ^2^ , ^3^ , ^4^ However, tissue congestion resulting from HF is one of the leading causes of morbidity in this patient population and a large burden on our health care system. Between 2006 and 2010, there were about 960 000 annual emergency department visits because of acute decompensated congestive heart failure in the United States. ^5^ A great majority of these patients (84%) were admitted. Hospitalization because of acute decompensated congestive heart failure is mainly caused by symptoms rather than low cardiac output. ^6^

The reason for the prevalence of right heart failure (RHF) is not as clear. This is partially because of a lack of a consistent definition of RHF. ^7^ , ^8^ RHF is common in HF with reduced ejection fraction (48%). ^9^ It is less prevalent in patients with HF with preserved ejection fraction (≈30%), such as in patients with right ventricular dysfunction caused by pulmonary hypertension. In patients with HF and right ventricle dysfunction, even without left ventricular dysfunction, a high mortality is predicted. ^10^ Furthermore, in patients with RHF, the severity of tricuspid regurgitation (TR) was found to be a predictor of mortality independent of left ventricular ejection fraction. ^11^

Accumulation of excess interstitial fluid (IF) is a well‐known phenomenon of HF. The production and volume of IF are governed by the arterial, venous, and tissue oncotic pressures, together with the capillary permeability and reflection properties, as described by the Starling equation. ^12^ , ^13^ Elevated central venous pressure (CVP) is one of the main manifestations of progressive HF and is common in patients with RHF. Heart failure causes elevated CVP as a result of increased blood volume, backflow, and neurohormonal venous vasoconstriction. ^14^ The kidneys play a key role in this process. ^15^ Elevated CVP both increases the production of IF by reducing the arterial–venous pressure gradient and at the same time impedes the removal of excess IF by increasing the afterload on the central lymphatics that have to empty into the higher pressure veins. ^16^ IF production rates also depend on the permeabilities of the endothelia that line the capillary vessels in addition to the plasma and tissue hydrostatic and oncotic pressures (Starling equation). ^12^ The permeability properties differ between organs and regions of the body, giving rise to variability of IF production rate and its protein content. ^17^ , ^18^ , ^19^ Excess IF in organs and tissue that border an internal cavity can lead to leak of fluid into the cavity and fluid accumulation, as is the case with ascites or pleural effusions.

One of the main functions of the lymphatic system is to collect excess IF and return it to the venous system. ^20^ IF is absorbed by the lymphatic capillaries, which are a network of permeable vessels. ^21^ From there the fluid flows into the collecting lymphatic channels that are surrounded by a smooth muscular layer. Contraction of the muscle propels the fluid forward and into the thoracic duct (TD). Unidirectional flow in this system is maintained by valves that are abundant throughout the lymphatic ducts. Daily, about 8 L of fluid is filtered out of the blood and into the interstitial space. ^13^ Of this fluid, 2 to 3 L/d are returned to the venous system via the TD and the rest via proximal lymphovenous communications.

Lymphatic abnormalities in patients with HF have been well documented. For example, Blalock and Burwell created an animal model of elevated CVP in dogs by inducing Pick disease (constrictive pericarditis), which resulted in elevated TD pressure, TD dilation, and ascites in the animals. ^22^ Using ultrasound, Seeger et al described TD dilation and regurgitant TD outlet flow in patients with HF. ^23^ Witte et al also described TD dilation and increased TD flow and pressure in patients with severe HF. ^16^ Decompression of the TD in these patients resulted in reduction in edema, ascites, lower CVP, and improved urine output.

It is estimated that peripheral congestion occurs in 60% of patients with acute HF and in about 20% of patients with chronic HF. ^24^ Similarly, not all patients with HF develop ascites or pleural effusions. ^25^ , ^26^ , ^27^ The risk factors for development of tissue congestion and ascites are poorly understood. Breidthardt et al did not find a correlation between edema and cardiac function, renal function, elevated B‐type natriuretic peptide, and elevated CVP. ^28^ Meyer et al found similar results. ^29^ Complicating these studies is the heterogeneity of patient populations, which makes comparing patients with and without fluid overload difficult. Here we report on a prospective study that compared lymphatic and cardiovascular parameters in a group of animals with elevated CVP and isolated RHF secondary to severe TR that developed ascites and a group with similar heart failure that did not develop ascites. We hypothesized that changes in the lymphatic system, rather than cardiac function or hemodynamic parameters, will correlate with ascites development.

Methods

The data that support the findings of this study are available from the corresponding author upon reasonable request. For the animal model, institutional animal care and use committee approval was obtained from the University of Pennsylvania for this study. All animals were housed and treated within established guidelines.

In the absence of preliminary data about the likelihood of our primary outcome and variability, no formal pre hoc power calculations were performed. We sought to compare 2 groups of animals, 1 with and 1 without ascites, with at least 5 animals in each group. Recognizing a failure rate of ≈50%, we predicted that we would need to use 10 to 15 pigs for this study to reach our evaluable population goal. Animals were premedicated with intramuscular ketamine 22 mg/kg, telazol 2 mg/kg, and xylazine 20 mg/kg. Animals were then intubated, anesthetized with isoflurane 0.5% to 2.0%, and mechanically ventilated with 100% oxygen. During all procedures, blood pressure, temperature, and ECG was continuously monitored. Isotonic saline was used to maintain adequate blood pressure. Using the Seldinger technique, a sheath was placed in the jugular vein, and a catheter was inserted into the right ventricle. The catheter was used to avulse ≥1 leaflet of the tricuspid valve, leading to severe TR. Immediately following the valve evulsion, acute RHF with hypotension occurred and was treated with a bolus of Ringer solution (25–50 mL/kg) and vasopressors as needed. RHF was verified by right ventricular dilation, as assessed using cardiovascular magnetic resonance or ultrasound, and by the presence of elevated CVP and severe TR by echocardiography. Left ventricular function was assessed by calculating the ejection fraction from echocardiography or magnetic resonance imaging.

After the tricuspid valve evulsion procedure, animals were assessed by imaging every 2 to 4 weeks for development of ascites. The date of first occurrence of ascites was taken as an end point, at which point the animals underwent complete hemodynamic evaluation, evaluation of ventricular function, and TD size; the animals were then euthanized. Animals that did not develop ascites were given sufficient time to develop symptoms and then were euthanized if symptoms did not develop within 4 to 5 months because of unmanageable size and because this time point was much longer than time given to the animals that did develop ascites. Before euthanizing, these animals also underwent complete hemodynamic evaluation, assessment of ventricular function, and TD size.

Magnetic resonance imaging was performed on a Siemens TIM Trio magnetic resonance imaging scanner (Siemens Health Care AG, Erlangen, Germany). Quantification of ascites volume was done by T2‐weighted 3‐dimensional spin‐echo sequence (1.6×1.6×1.6 mm voxels, repetition time, 1600 ms, echo time [TE] 208 ms, flip angle 125°, bandwidth 446 Hz, 2 averages). The ascites volumes were segmented using neighborhood connected thresholds analysis using OsiriX software (Pixmeo SARL, Bernex, Switzerland). Cardiac dimensions and function were measured using retrospectively gated fast low‐angle shot sequence (repetition time, 49.9 ms; TE, 2.43 ms; matrix, 192×125; field of view, 276×340 mm; slice thickness, 6 mm; spacing, 0 mm; flip angle, 15°). For each slice of the short‐axis stack, cine containing 25 phases were recorded. OsiriX software with the plugin MRHeart was used to calculate left ventricular ejection fraction. Liver volumes were calculated by manual segmentation of the T2‐weighted spin‐echo images or of a stack of respiratory gated, T2‐weighted, half‐Fourier single‐shot turbo spin‐echo (repetition time, 2000 ms; TE, 90 ms, matrix, 320×183; field of view, 289×380 mm; slice thickness, 5 mm; spacing, 1 mm; flip angle, 180°). The TD was imaged using contrast‐enhanced, 3‐dimensional fast low‐angle shot sequence after contrast (Gadavist; Bayer AG) was injected into both inguinal lymph nodes percutaneously (voxels, 1.04×1.04×1.2 mm; repetition time, 3.45 ms; TE, 1.2 ms; matrix, 264×384; field of view 274×399 mm; slice thickness, 1.2 mm; flip angle, 15°). Alternatively, a 3‐dimensional fast low‐angle shot inversion recovery prepped sequence was used (voxels, 1.58×1.58×1.6 mm; repetition time, 400.48 ms; TE, 1.5 ms; inversion recovery, 250 ms; matrix, 272×304; field of view, 429×480 mm; slice thickness, 1.6 mm; flip angle, 18°).

Ultrasound (EPIQ 5; Philips Medical Systems, Best, the Netherlands) was used to measure cardiac function, tricuspid valve regurgitation, presence of ascites, and to guide needle placement for lymphangiograms. Transthoracic echo using s‐5 or s‐8 probes was used to obtain left ventricular short‐axis and subxiphoid 4‐chamber views. These images were analyzed using OsiriX software and the plugin Ejection Fraction to calculate left ventricular ejection fraction. Alternatively, a transesophageal echo was used to obtain left ventricular short‐axis view.

Pressure measurements were performed using fluid filled catheters connected to manometers (Transpac IV; Utah Medical, Midvale, UT) and micromanometer tipped catheters (Mikro‐Tip; Millar, Houston, TX) using LabChart version 8.1.13 (ADInstruments, Colorado Springs, CO).

Lymphatics vessels size (TD and liver lymph node efferents) was measured offline from acquired images. When measured from fluoroscopy images, pixel size was calibrated using the size of a catheter in the image. Lymph flow patterns were determined by tracking a droplet of ethiodized oil (Lipidol; Guerbet, Princeton, NJ) (Figure 1). A location in the TD was chosen (anterograde of ethiodized oil injection), and if the tracked droplet crossed that location retrogradely during >1 respiratory cycle, the flow pattern was judged to be bidirectional (regurgitant). Phase maps of tracked droplets illustrating the difference in paths are shown in Figure 1B and 1C.

Statistical Analysis

All measurements in this study including TD dimensions and echo ejection fraction were done by cardiac and lymphatic specialists who were blinded to the data. Data were analyzed using IBM SPSS version 27. Variables are reported as mean±SD and were compared using the Mann‐Whitney U test. The frequencies of categorical variables were compared using the Fisher exact test. The analysis was exploratory and hypothesis generating, and there was no primary association of interest; therefore, no adjustments were made for multiple comparisons.

Results

Data from 13 animals, 7 with ascites (WA) and 6 without ascites (WOA), were analyzed. The study's end points were appearance of ascites or the last study of the animals that did not develop ascites. Time to the end points was measured from the tricuspid valve injury procedure and was statistically significantly longer for the WOA animals than the WA animals (135.5±47.1 versus 59.4±39.5, P=0.022) (Table 1). Correspondingly, the mean weight at termination of the WOA animals was larger than the WA animals (70.5±15.3 versus 45.6±13.7 kg, P=0.014).

The animal's cardiac function and hemodynamic parameters are summarized in Table 2. None of the hemodynamic parameters were found to be statistically significantly different between the 2 groups. Pulmonary pressure and wedge pressures, as well as the left ventricular function, were similar between the groups. Central venous pressures were elevated in both groups, but there was no difference between the inferior vena cava pressure (11.6±1.6 versus 13.2±3.7, P=0.534) and superior vena cava pressure (12.0±2.3 versus 13.7±3.2, P=0.366) between the 2 groups. There was also no difference between right ventricular systolic and end‐diastolic pressure between the groups.

In the presence of elevated CVP, the TD became dilated (Figure 2). There was no significant difference between the mean TD pressures in WA animals and WOA animals (11.1±3.4 versus 11.3±3.0 mm Hg, P=0.731). There was also no statistical difference between the mean maximal TD cross‐section size (4.7±1.5 versus 6.7±2.3, P=0.295), but there was a difference between the mean minimal size of the TD of the WOA group, which was significantly smaller than that of the WA group (2.3±0.3 versus 4.2±2.2 mm, P=0.035). The fractions of WA animals with bidirectional flow (6/7) were significantly larger than in the WOA (1/6) animals (P=0.029).

The weighted sizes (volume per kilogram) of the liver were compared and found to be significantly smaller in WOA animals compared with WA animals (30.3±12.4 versus 63.3±14.0 mL/kg, P=0.001).

Discussion

The lymphatic system plays a major role in immune regulation, transport of macromolecules, and fluid balance. Here we report on a study that compared animals with RHF who developed ascites to a control group that did not develop ascites. The main finding of this study is that the differences between animals with RHF that developed ascites and those that did not was not in their cardiovascular and hemodynamic parameters, but rather in the structure and function of their central lymphatic system and liver size.

The circulation in the liver is known to be sensitive to elevated CVP, and congestive hepatopathy was first described in detail by Sherlock. ^30^ However, the risk factors for the development of fluid overload in patients with heart failure are poorly understood. Renal sodium and fluid retention, as well as shifts in the Starling equation caused by elevated CVP, have been implicated as the cause of fluid overload, but multiple studies are not consistent with this assumption. Some patients with relatively preserved renal and cardiac function have severe fluid overload symptoms, whereas other patients with poor renal and cardiac function do not. Similar to the patient population and previous studies, we found that ascites did not correlate with the degree of heart failure or hemodynamic parameters such as elevated CVP. Both animal groups in this study were similar in all measured cardiac parameters. In contrast, our results found that WA animals differed from WOA animals in 3 (1) liver size, (2) minimal TD diameter, and (3) the TD flow pattern. The left ventricular function of our animal remains intact, and thus, the observed liver enlargement is purely passive congestive and appears to correlate with the development of ascites.

Normally, the pressure difference between the portal and hepatic veins is only a few millimeters of mercury. When that difference becomes even smaller or reverses, more fluid is directed to the space of Disse (the liver's interstitial space) and from there to the collecting lymphatics. ^31^ If the liver lymphatic capacity is insufficient to drain this fluid, swelling of the liver ^27^ and dilation of the lymphatic vessels would occur. ^31^ In this study we observed that although the CVP of all animals was similar, some animals developed enlarged livers (hepatomegaly), suggesting that the difference between these groups is in the ability of the lymphatics to remove the excess fluid from the liver.

The highly permeable hepatic sinusoid epithelium transmits the hepatic vein pressure to the space of Disse and possibly higher, collecting lymphatic pressure. Furthermore, the elevated CVP also leads to elevated TD outlet pressure. In this study we did not find a difference between the TD pressure in WA and WOA animals, because the CVP was the same in both groups, suggesting that the TD absolute pressure is not by itself a risk factor for the development of ascites. However, the increased TD pressure and flow likely leads to TD dilation. Dilation of lymphatic vessels has been described in obstructed peripheral lymphatic vessels. ^32^ , ^33^ , ^34^ , ^35^ , ^36^ , ^37^ These dilated peripheral vessels were shown to lose their ability to pump lymph, because the dilation leads to impaired valves that are responsible for maintaining unidirectional flow in the lymphatic ducts. Davis et al demonstrated that the distention of the valve impedes its closing by changing the pressure gradient that is required to close it. ^38^ In this study, we found that the TD dilated in both WOA and WA animals, as evident by the similar maximal TD diameter consistent with increased flow and pressure in the TD in both groups. However, in the WOA animals, the minimal diameter of the TD was smaller than in the WA animals. When the TD dilates, lymphatic valves become incompetent. Consequently, a smaller TD minimal diameter in the WOA group supports the notion that at least some of the TD valves were functional in this group and therefore maintained some of the TD lymph pumping functionality in these animals. This is also supported by the fact that WOA maintained unidirectional lymphatic flow in contrast to the WA animals, where bidirectional flow was more common.

A cardinal factor in the propulsion of lymph in the TD is the intrathoracic pressure fluctuation during the breathing cycle. ^39^ , ^40^ During inspiration, the negative intrathoracic pressure reduces the thoracic TD pressure and creates a pressure gradient between the abdominal and thoracic parts of the TD. This pressure gradient propels the lymph along the TD. During expiration, when the thoracic TD pressure is higher, TD valve(s) prevent backflow to the abdominal TD. In this study, we analyzed the lymph flow pattern in WOA and WA animals and found them to be different. Animals with ascites exhibit regurgitant flow consistent with their larger minimal diameter of the TD. Regurgitant TD flow potentially also represents failure of the respiratory pump and is a possible contributing reason why these animals develop ascites.

The data presented here support the notion that lymphatic failure occurs when elevated CVP causes elevated hepatic sinusoid pressure and increased hepatic lymph production. The higher lymph production and elevated pressure cause TD dilation. The high volume of lymph and the elevated pressure are compensated for by increased TD drainage as long as the TD valves are functional. When the TD valves become dysfunctional, lymphatic pump failure results in a reduction in forward lymph flow, leading to an imbalance between lymphatic production and drainage. As a result, ascites and hepatomegaly occur (Figure 3).

Why some animals, like people, develop ascites, whereas others do not, is a question that remains to be answered. Perhaps this is because of an underlying susceptibility of the lymphatic system or because of a difference in the liver lymphatic production and lymphatic flow. Further studies are needed to better answer this question.

Limitations

This is a small study in animals with RHF caused by severe TR. As such, further studies need to be conducted to confirm these findings in animals with other forms of congestive heart failure. Moreover, studies need to be conducted in humans to see if similar results are found.

As noted, we did not control for increase in familywise error rate across the reported statistical analysis because we consider this research preliminary and a pilot study. Further studies will need to be conducted to confirm these findings and validate these results.

Conclusions

In this study we have shown that ascites development in animals with severe TR does not correlate with hemodynamic parameters but, rather, correlates with changes in the lymphatic system, including regurgitant lymphatic flow, minimal TD diameter, and also correlates with liver size. These results seem to be consistent with what has been observed in people with fluid overload and potentially point to a lymphatic dysfunction cause for fluid overload. Further studies are needed to determine whether these results are also seen in people and to determine the causes of lymphatic failure in these animals.

Sources of Funding

This study was supported by the Chappell Culpeper Foundation and the Jill and Mark Fishman Foundation.

Disclosures

None.