Ascites develops secondary to portal hypertension, with or without hypoalbuminemia. Liver disorders associated with ascites trigger physiologic responses that sustain euvolemia and splanchnic perfusion; these drive sodium and water conservation.
Portal hypertension (PH) reflects circulatory dynamics thwarting craniad flow of the portal circulation to the liver (see Anatomy of the Hepatobiliary System).
Portal hypertension is best considered, firstly, in terms of hepatic involvement. Causes can be categorized as prehepatic, hepatic, and posthepatic:
Prehepatic causes of PH involve pathologies thwarting portal venous perfusion before portal vein dissemination within the liver. Causes of prehepatic PH include portal vein inflammation, stenosis, stricture, or compression, or intraluminal masses or thrombi.
Posthepatic causes of PH include pathologies external to the liver that thwart sinusoidal circulatory egress via centrilobular sinusoids and hepatic veins. Causes include obstruction of blood flow from the liver via the hepatic vein; this can begin at the level of the heart (eg, right heart failure, cor triatriatum dexter, hemangiosarcoma involving the right atrium), pericardium (eg, restrictive pericarditis, pericardial tamponade), or vena cava (eg, thrombi, congenital web or acquired "kink," or heartworm-associated vena caval syndrome).
Hepatic causes of PH include processes thwarting craniad flow of portal circulation within the liver. These disorders are subclassified relative to hepatic sinusoids: presinusoidal, sinusoidal, and postsinusoidal causes.
Presinusoidal causes include diffuse flow occlusion of intrahepatic portal veins by thrombi (rare), portal vein phlebitis or sclerosis, and diffuse severe ischemic cholangiopathy due to arterial thrombi (impair perfusion to peribiliary vascular plexus resulting in loss of nutrient vasa vasorum of portal veins). The latter injury leads to portal vein atrophy and disappearance. However, the most common cause of severe presinusoidal PH is diffuse portal tract fibrosis associated with a proliferative-like ductal plate malformation (DPM).
This congenital disorder is characterized by deposition of exuberant extracellular matrix that expands portal tracts into fibroductal partitions populated by myriads of malformed bile duct profiles. A dominating arterial perfusion (numerous serpiginous arterioles with thick muscular walls) and a paucity of portal vein silhouettes is typically observed. Noncompliant fibrillar collagen expanding portal tracts imposes severe resistance to the low-pressure portal system, leading to development of APSSs and ascites. Portal hypertension secondary to chronic extrahepatic bile duct obstruction (EHBDO) involves presinusoidal and sinusoidal resistance factors.
Sinusoidal causes of hepatic PH include sequela of chronic hepatitis that can result in collagenization and capillarization of hepatic sinusoids and variable connective tissue impacting portal triad structure. It can also impair circulation to hepatic venules (centrilobular area), hepatic cord and sinusoidal remodeling (ie, formation of regenerative nodules with transition to cirrhosis), vascular occlusion of hepatic or portal veins or sinusoids (eg, thrombi, neoplasia, vasculitis), or diffuse sinusoidal dissemination of neoplastic cells or storage materials thwarting transhepatic perfusion (inborn errors of metabolism-glycogen storage disorders, but not typical glycogen vacuolation and not acquired hepatic lipidosis).
Rarely, arterialization of sinusoidal perfusion by an intrahepatic arteriovenous malformation causes PH and ascites as a result of hypertensive sinusoidal perfusion. This leads to hepatofugal portal venous circulation, development of APSSs, and increased ultralymph formation causing ascites.
Postsinusoidal hepatic PH may reflect intrahepatic injury thwarting circulatory egress via hepatic venules and veins or damage or occlusion of these vessels. Most common is severe regional sinusoidal collapse resulting in a sinusoidal occlusion syndrome (SOS)—ie, severe copper-associated, NSAID-associated, or toxin-mediated centrilobular hepatic injury. Primary vascular causes include a venoocclusive lesion reflecting endothelial injury or venous luminal occlusion due to degeneration, thrombi, or compression by lipogranulomas or severe occlusive myeloid metaplasia (ie, sometimes referred to as agnogenic extramedullary hematopoiesis).
In chronic remodeling liver disease causal to PH, intrahepatic resistance to sinusoidal flow initiates splanchnic hypertension.
Intrahepatic vascular resistance reflects structural remodeling and loss of vascular channels that restrict and deviate sinusoidal perfusion. Hepatic remodeling and dissecting sinusoidal fibrosis greatly compromise hepatocyte ultralymph access; this restricts hepatic cleansing functions and systemic access to hepatic synthesized proteins and metabolites. Chronic inflammation and fibrosis transform sinusoidal endothelium into one with a less fenestrated phenotype, characterized as sinusoidal collagenization and sinusoidal capillarization.
Dysfunctional sinusoidal endothelium with imbalanced response to vasoconstrictors versus vasodilators leads to a dynamic increase in sinusoidal tone. Along with this physiologic response, narrowed sinusoidal dimensions (remodeling and fibrosis) increase vascular resistance. This process escalates in necroinflammatory disorders by sinusoidal damage, deviation, disorganization, microthrombi formation, and mechanical occlusion by activated scavenging Kupffer cells.
Heightened sinusoidal resistance increases hepatic ultralymph formation, eventually exceeding lymphatic capacitance. This distends lymphatic channels (portal, perivenous [central], and subcapsular), contributing to ascites (ie, often weeps from subcapsular lymphatics in Glisson's capsule [mesothelial capsule of the liver]). Neovascularized arterial perfusion circumvents regenerative nodules, limiting hepatocyte circulatory exposure. Capillarized and collagenized sinusoids and neovascularized pathways allow faster transhepatic circulatory throughput, causing direct intrahepatic sinusoidal shunting.
All causes of critical hepatic PH develop increased hepatic arterial perfusion. This reflects the hepatic arterial buffer response (HABR), a normal physiologic compensation responding to decreased portal venous perfusion. Conversely, loss of arterial perfusion does not result in a compensatory increase in portal perfusion and may be lethal. In PH, the HABR leads to hepatofugal (backward) flow of blood into the valveless portal system with eventual formation of acquired portosystemic shunts (APSSs).
Development of APSSs usually occurs within 4–6 weeks of critical PH, when the hepatic venous pressure gradient (the pressure gradient between the hepatic vein and portal vein pressures) is > 10 mm Hg—equivalent to a portal pressure > 12 mm Hg. While APSSs decompress splanchnic hypertension to some degree, pressure never completely normalizes. APSSs represent dilation of preexisting but closed vascular channels between the portal and systemic circulation and increased angiogenesis stimulated by vascular endothelial growth factor.
Development of PH has important systemic consequences, including development of splanchnic vasodilation (mechanistically associated with augmented nitric oxide [NO] production in splanchnic vasculature). Splanchnic vasodilation unfortunately increases blood flow into the portal venous system, provoking systemic arterial hypotension and a decline in effective or central blood volume. This engages compensatory endogenous mechanisms conserving body sodium and water (ie, neurohumoral vasoactive systems, including activation of renin, synthesis of angiotensin, and release of vasopressin). Hemodynamically this expands blood volume and cardiac output leading to a hyperdynamic circulatory state. Further physiologic compensations eventually culminate in circulatory imbalance and ascites formation.
Splanchnic PH impacts all abdominal organs, the most critical being the kidneys and alimentary canal. Renal perfusion can be adversely affected, causing kidneys to have heightened sensitivity to hypovolemia or abrupt changes in cardiac output and glomerular filtration rate. Consequences of PH include development of ascitic effusion, splanchnic vasodilation, risk of bleeding associated with APSSs, and development of a portal-enteric vasculopathy.
A syndrome of portal enteric vasculopathy, also referred to as hypertensive gastroenteropathy, can lead to insidious gastrointestinal bleeding in dogs with severe PH. This is typically accompanied by engorged intestinal lymphatics and some degree of a protein-losing enteropathy. While PH is critically associated with hypertensive gastroenteropathy, other factors also contribute to this complication (ie, cytokine activation and release of growth factors and hormones). Increased gastric and intestinal blood flow coupled with venular congestion thwarts visceral microperfusion. This results in dilated ectatic vasculature (arterioles, capillaries, submucosal and subserosal veins) with increased risk for diapedesis and spontaneous hemorrhage.
The hyperdynamic circulation characteristic of PH is also thought to inhibit gastric mucosal defense mechanisms, impairing release of growth factors and provoking release of proinflammatory mediators. These changes render gastric and duodenal mucosa more vulnerable to injury with risk for spontaneous bleeding. While histologic changes have rarely been described in affected dogs, lesions similar to those observed in humans have been noted. These include observation of mucosal or submucosal ectatic vessels without intraluminal clots or fibrinoid degeneration. Lesion distribution in dogs usually involves the stomach, duodenum, or colon.
Ascites formation reflects circulatory dynamics thwarting craniad flow of blood via the liver associated with an acquired imbalance of sodium and water homeostasis. This is variably associated with a subnormal albumin concentration. Abdominal effusion associated with hepatic disease is usually a modified or pure transudate, the later being associated with concurrent PH and hypoalbuminemia (ie, serum albumin concentration < 2 g/dL).
Drugs used to attenuate portal hypertension ideally should decrease PH without enhancing systemic vasodilation. No drugs can fully arrest hemodynamic imbalances initiated and sustained by splanchnic PH. Nonselective beta-blockers remain the cornerstone of treatment for splanchnic hypertension with APSSs. The most commonly used drug in humans for reducing splanchnic hypertension and risk of hemorrhage from APSSs (and hypertensive enteric vasculopathy) is propranolol.
Therapeutic goals include a modest decrease in resting heart rate with preservation of systemic blood pressure (mean arterial blood pressure > 90 mm Hg). Propranolol has been used in dogs with decompensated cirrhosis and diuretic-resistant ascites (low dosage of 0.05 mg/kg, PO, every 12 hours, slow upward titration to effect).
However, investigative studies of PH in dogs secondary to chronic biliary cirrhosis (induced by chronic EHBDO) did not demonstrate a decrease in splanchnic pressure with propranolol. Nonselective beta-blockers antagonize the beta1 adrenergic receptors in the heart and the beta2 adrenergic receptors in the periphery. The beta1 effect decreases heart rate and cardiac output and subsequently decreases flow into the splanchnic circulation. The beta2 effect leads to unopposed alpha1-adrenergic activity, which causes splanchnic vasoconstriction and further decrease of portal inflow.
Propranolol should not be trialed in animals with chronic obstructive pulmonary disease, reactive airway disease, asthma, or congestive heart failure because it may worsen bronchoconstriction and decreases output. In humans, carvedilol, another nonselective beta-blocker with antioxidant properties, provides a greater attenuation of portal pressure. At present, this drug is ranked as first-line treatment for decompensated PH based on meta-analyses of published clinical trials.
Cardiovascular experimental studies in healthy dogs, and treatment of dogs with mitral insufficiency given carvedilol 0.2–0.3 mg/kg, PO, every 12 hours, for several weeks reported no adverse effects. A dose for dogs with hepatic insufficiency is unknown; initial treatment at no more than 0.05–0.105 mg/kg, PO, every 12 hours, might be a reasonable starting dose. Other pharmacological interventions studied in humans remain controversial and have not been shown in placebo-controlled trials to provide greater benefit.
Therapeutic strategies for control of ascites include dietary sodium restriction and administration of diuretics to increase urinary sodium elimination. When needed, therapeutic abdominocentesis is also used. The first step is dietary sodium restriction to an intake of ≤ 100 mg sodium/418.4 kJ (100 kcal) diet (25 mg/kg/d; < 0.1% dry-matter basis in food). However, sodium restriction alone is usually insufficient and too slow in onset for efficient management. Thus, diuretics are usually also recommended.
Diuretic therapy should slowly decrease ascites without causing dehydration, metabolic alkalosis, or hypokalemia. Reducing ascites by ≤ 1%–1.5% of total body weight/d is recommended by initially using combined treatment with furosemide (1–2 mg/kg, PO, every 12 hours) and spironolactone (loading dosage 2–4 mg/kg for 2–3 doses, then 1–2 mg/kg, PO, every 12 hours).
Reevaluation every 7–10 days permits careful upward titration of diuretic dosages. Combining a loop diuretic (furosemide) with spironolactone (aldosterone antagonist) decreases risk of iatrogenic hypokalemia. Take care to avoid concurrent treatment with an angiotensin-converting enzyme (ACE) inhibitor (ie, telmisartan, enalapril, benazepril) concurrent with spironolactone because this may cause clinical hypotension and acute renal failure.
If ascites is slow to mobilize with sodium restriction and diuretics, measuring urinary fractional excretion of sodium can help determine whether dietary restriction and diuretic dosing are adequate.
If ascites causes tense abdominal distention compromising cardiac output, ventilation, appetite, or patient comfort, therapeutic abdominocentesis may be undertaken. In humans, 8 g of human albumin is administered for every 5 L of removed effusion to offset development of postdiuresis circulatory dysfunction ~12 hours after effusion removal.
Postdiuresis circulatory dysfunction reflects re-equilibration of body fluids, potentially worsening hypoalbuminemia (removed by abdominocentesis). This can culminate in systemic hypotension and splanchnic and renal vasoconstriction. The latter responses increase risk for hepatorenal syndrome (reversible renal vasoconstriction associated with liver failure complicated by ascites).
Because there is no access to species-specific albumin for dogs and cats, polyionic fluids are used with therapeutic abdominocentesis. Large-volume abdominocentesis should never be performed without concurrent diuretic administration. The goals of the treatment should be to remove enough volume to improve patient comfort (reducing abdominal pressure) and to improve renal perfusion, cardiac output, and response to diuretic therapy. Once ascitic effusion is mobilized, diuretics can often be used intermittently with concurrent dietary sodium restriction to control ascites accumulation.
In humans, high-dose albumin infusions (1–1.5 g albumin/kg per week for months) are highly effective in preventing new episodes of refractory ascites, hepatic encephalopathy (HE), hepatorenal syndrome, and bacterial infections and are shown to decrease hospital admissions and improve patient survival rate.
In the event of APSSs associated with spontaneous hemorrhage (usually into the alimentary canal), a combination of plasma transfusion, vasoactive drugs, antimicrobials, and, rarely, surgery to ligate involved APSSs are considerations. Surgical intervention carries high risk and is a salvage procedure. In humans with severe splanchnic hypertension, insertion of a transjugular intrahepatic portosystemic shunt (TIPS) dramatically lowers splanchnic pressure and risk of bleeding from APSSs. However, insertion of a TIPS increases risk for HE due to the enhanced portosystemic shunting that decompresses the splanchnic hypertension.
