Portosystemic Vascular Anomalies & Hepatic MVD: Histological, Clinical, & Treatment Options
ACVIM 2008
Sharon A. Center, DVM, DACVIM
Ithaca, NY, USA


Extrahepatic portosystemic vascular anomalies (PSVA) in small breed dogs, originally described in the 1970s, have received notorious attention in the veterinary literature. We estimate that this congenital malformation comprises approximately 0.6 to 2.0% of a specialty hospital patient population. However, microvascular dysplasia (MVD), a genetically related hepatic vascular disorder is far more common; 15 to 30:1 compared to the prevalence of PSVA in affected kindreds based on investigations by the author in an expansive genotyping initiative. MVD occurs as a single entity or coexists with PSVA. Unfortunately, discovery of high serum bile acid concentrations in dogs with MVD often leads to expensive and invasive diagnostic assessments for PSVA in pet dogs. Dogs affected only with MVD are typically asymptomatic and their hepatic vascular abnormalities non-progressive. The majority of dogs affected only with MVD do not need medical treatments as assigned for dogs with symptomatic PSVA and have a normal life expectancy. Many MVD affected dogs are discovered on the basis of high serum bile acid concentrations revealed upon screening kindreds for PSVA or serendipitously at the time of non-related illnesses. Misconceptions regarding management of MVD has caused confusion among clinicians, breeders of commonly affected small pure breed dogs, and pet dog owners.

Histological Features

The normal liver receives blood from both the portal vein and hepatic artery; these circulations are regulated independently. Portal flow, normally provides 60-70% of hepatic blood flow. The liver has no inherent ability to control portal blood flow and the portal vein lacks valves. Rather, portal flow is regulated by resistance vessels in the splanchnic viscera (pre-hepatic) with circulation determined by their net outflow. The liver indirectly influences portal perfusion by regulating hepatic arterial flow (hepatic artery buffer response, HABR). The HABR is mediated by adenosine wash out1-3, i.e., in the circumstance of low portal blood flow, less adenosine is washed out leading to hepatic arterial dilation and increased flow from this relatively high pressure system. Thus, following acute interruption of portal venous perfusion, the HABR maintains hepatic blood flow. While the hepatic artery provides approximately 20-30% of the total hepatic blood flow in health, in the presence of portal hypoperfusion, this flow increases by up to 4 fold. Recent evidence confirms that arteriolar responses also compensate for chronic portal hypoperfusion. In the dog, hepatic outflow also is adjusted by throttling musculature surrounding hepatic venules.

The histological features of PSVA have been frequently described and largely reflect portal hypoperfusion. Typical microscopic changes include: increased hepatic arteriole (small arteries) cross sections, obvious small non-perfused vessels (lymphatics) within the adventitia of the portal triad, small, juvenile or under-developed portal triads (very small structures) randomly located in the hepatic parenchyma, hepatic lobular atrophy, an increased number of small binucleated hepatocytes, prominent hepatic venule throttling musculature (hypertrophied) that appears variably contracted, and in some dogs, multifocal lipogranulomas and lipogranulomas perivascular to the hepatic venule. Hepatic venules inappropriately located adjacent to or within portal triads are apparent on careful inspection of large hepatic biopsy specimens from some dogs with MVD. These likely enable direct intrahepatic shunting. In some dogs with PSVA as well as some dogs with MVD lacking PSVA, lipogranulomatous and inflammatory (non-suppurative, may involve eosinophils) zone 3 lesions impose a veno-occlusive effect (occluding acinar egress of blood). Dogs with the zone 3 lesions are significantly at risk for poor response to surgical shunt attenuation (formation of acquired shunts, chronic recurrent abdominal effusion, post-operative death). In general, microscopic abnormalities associated with PSVA are indistinguishable from those associated with MVD, particularly the most obvious feature of arteriolar duplication. The increased number of arteriolar cross sections (referred to as arteriolar duplication) reflects cross sections of tortuous or coiled arterioles (as demonstrated on arteriograms of dogs with portal hypoperfusion). Arteriolar conformational changes reflect a compensatory response to portal hypoperfusion (increased arterial flow and subsequently blood pressure due to acquired dominance of the intrahepatic arterial circulation). Experimental work (rodents with chronic [6 week] portal vein attenuation) documented a three-fold increase in hepatic arteriolar blood flow in the circumstance of chronic portal hypoperfusion.4 In this model, tortuous, clustered, and sometimes disorganized arterioles surrounded intrahepatic portal veins creating a histologic appearance similar to that affiliated with PSVA and MVD. It remains unclear if this is a remodeling response (hypertrophy, expansion) involving pre-existent arterioles or de novo angiogenesis. In this model, arterioles develop connections with more than a single acinus and branches directly drain into sinusoids as well as terminal hepatic venules. The latter communications establish conduits of intrahepatic shunting.

Surgical wedge and laparoscopic liver biopsies collected from multiple liver lobes in dogs with MVD and PSVA confirm that microscopic lesions are inconsistent among liver lobes. The caudate lobe notoriously has the most normal architecture in affected dogs (receives first portal vein branch and often is adequately perfused in dogs with MVD and in some dogs with PSVA). Thus, needle biopsy and single liver lobe biopsies can miss lesions in MVD affected dogs. This circumstance likely influenced outcome in a study comparing total serum and HPLC measured bile acids in Maltese dogs.5 Portovenograms, colorectal scintigraphy, and MRI contrast studies in dogs with MVD corroborate histologic evidence that liver lobes have variable portal perfusion. It is probable that variable liver lobe perfusion in MVD affected dogs explains apparent "portal streamlining" described in one study where definitive diagnosis of PSVA was determined by scintigraphy is suspected patients.3 Considered collectively, histologic and imaging features corroborate that MVD and PSVA represent complex disorders of hepatic angiogenesis / vasculogenesis that has a spectrum of severities. (See accompanying table) Because dogs with surgically created portosystemic shunts develop histologic features identical with those in PSVA, liver biopsy cannot confirm portal hypoplasia (lack of portal vein development) without knowledge of patient signalment. Rather, the histologic lesion is appropriately termed portal hypoperfusion.

Clinical Signs

The author has studied large kindreds of dogs as well as a large population of unrelated purebred dogs in our clinical patient population with PSVA as well as dogs with MVD.It is clear that MVD is more common than PSVA and that most of these dogs do not demonstrate clinical signs. While PSVA affected dogs represent the most severe vascular malformations and usually manifest clinical signs, we estimate that approximately 15-20% of these are asymptomatic, reflecting their relatively smaller degree of portosystemic shunting compared to symptomatic dogs. Dogs with portoazygous shunts are generally least symptomatic and often present as adults with ammonium biurate calculi or have their disorder serendipitously discovered. Generally, dogs discovered to have a PSVA later in life have been asymptomatic and often have a good response to PSVA ligation (see below). The clinical signs associated with symptomatic PSVA in dogs are well described in the veterinary literature and variably include: meal related neuroencephalopathic signs or somnolence, nausea or vomiting, diarrhea or constipation, polyuria and polydipsia, intermittent fever, ptyalism, maniacal or aggressive behavior, seizures, coma, signs attributable to ammonium biurate urolithiasis (ureteral, cystic, urethral locations) and chronic urinary tract infections. PSVA dogs also have increased susceptibility to infections owing to reduced function of their hepatic mononuclear phagocytes (Kupffer cells). Minor bite wounds, tick bites, subcutaneous infections, lacerations, and even vaccinations may lead to illness requiring hospitalization (fluid therapy, antimicrobials). The severity of clinical signs in symptomatic PSVA patients is highly variable and is largely modified by feeding an appropriately formulated diet. Seemingly, the relative risk for seizures in dogs with PSVA is high in the author's hospital compared to dogs with other organ system disorders.

Common clinicopathologic features of PSVA include a borderline non-regenerative anemia, low MCV, target cells, low BUN, low creatinine, low cholesterol, and slightly low albumin concentrations, normal to variable increases in liver enzymes (mild to modest), low normal glucose concentration (except for ill young toy breed dogs: e.g., Yorkshire Terriers, Maltese that may have symptomatic hypoglycemia), and ammonium biurate crystalluria (examine a minimum of 3 urine specimens). High serum bile acid concentration are typically found unless enteric malabsorption is a co-existent problem. Bile acid quantification should be done with paired samples (pre-meal and 2-hours post meal). I no longer fast dogs for 12-hour to collect a pre-meal sample as the point of the endogenous serum bile acid challenge is to detect values > 25 uM/L. Use of the fasting serum bile acid "normal range" should be discarded as 15-20% of all dogs tested demonstrate a higher pre-meal than post-meal bile acid concentration due to physiologic variables influencing the test; the fasting "normal range' has confused test interpretation. Blood samples for quantification of bile acids should be free of hemolysis and the laboratory should remove lipemia before testing the sample. I collect bile acid samples into lithium heparin, remove the needle and vacutainer stopper before gently placing blood in the tube (avoids hemolysis). Blood ammonia concentrations continue to remain controversial in diagnosis of PSVA because environmental contaminants can result in false positive tests values, the test has low reproducibility, and is inconvenient because ammonia is too labile for routine handling. Finding ammonium biurate crystalluria is nearly pathognomonic for PSVA when coupled with patient signalment, clinicopathologic and imaging findings. The recently validated Protein C test is useful for predicting the degree of shunting and monitoring patient response to surgical shunt attenuation. Values < 70% are useful for differentiating PSVA from MVD (most MVD dogs have values > 70%). We use this cutoff for prioritizing expensive ultrasound and colorectal imaging studies in dogs with high total serum bile acids suspected of having a PSVA.


Surgical PSVA Attenuation

Current textbook dogma predicts that all dogs with PSVA have a shortened life span if they are not provided surgical shunt attenuation. This is clearly not true. We have diagnosed dogs with PSVA as old as 13 yrs and have medically managed asymptomatic and symptomatic dogs effectively with dietary therapy for years. Surprisingly, even dogs with extrahepatic portal atresia can be effectively managed with strict dietary control (the author has maintained 3 such dogs in her home for 3 years with a prescription diet formulated for dogs with hepatic insufficiency without additional lactulose or metronidazole). The best method of surgical attenuation of PSVA remains controversial among surgeons. Ameroid constrictors can considerably reduce the table time in surgery, but they present a hazard for dogs that cannot tolerate complete shunt attenuation. Selecting a "large" ameroid to avoid complete shunt attenuation is a strategy used by some surgeons. Early use of ameroid constrictors in some Yorkshire Terriers resulted development of acquired portosystemic shunts because some dogs were unable to accommodate complete shunt occlusion (severe intrahepatic hypoplasia). While some contend that slow constriction of the PSVA by an ameroid occluder allows the hepatic vasculature to accommodate increased (forced) portal flow, physiologic mechanisms consistent with this hypothesis are undefined in the experimental literature. We and others approximate that 15 to 20% of the dogs presenting for PSVA cannot accommodate even partial PSVA occlusion. The best review of ameroid constrictor performance described findings in 168 dogs (7.1% mortality, 10% postoperative complications: seizures, hemoperitoneum, acute ascites, sudden death [thrombi]; 58/162 dogs had a portoazygous shunt [less symptomatic and more amenable to attenuation]).6 Authors reported 86/108 (80%) dogs with excellent outcome (no treatments or protein restriction), and 22/108 (20%) requiring continued medical management; some demonstrated unremitting HE. Depending upon how data was analyzed, up to 28% of dogs had unsuccessful outcomes (required treatment, had unremitting HE, or died); 48/168 (31%) dogs were lost to follow-up. Persistent shunting (scintigraphy) was documented in 21% of dogs (6-10 weeks after surgery). Some of these did not require dietary or medical treatments, consistent with observations in asymptomatic PSVA dogs. Silk ligation to individual tolerance is still maintained as a useful method of PSVA attenuation by some surgeon with the extent of occlusion determined by changes in vital signs, portal pressure gradient, appearance of visceral perfusion and arterial response. Gauging the degree of shunt ligation using intra-operative assessments of change in portal blood pressure is complicated by the cardiovascular influence of anesthetic agents. While normal or tolerated pressures (< 10 Cm water, < 7.6 mm Hg) may be observed at surgery, these may radically change upon recovery such that a large margin of error exists. This can result in acquired shunting in some dogs judged able to accommodate complete ligation or ameroid constrictor application. Very slow onset shunt attenuation as possibly achieved with loose cellophane shunt banding, may provide the best long term outcome.7 However, chronic follow-up data has not yet been published for this procedure. Postoperative seizure complications in dogs with PSVA range between 4% to 6%. Treatment with KBr is not proven to be protective for postoperative seizures, propofol CRI is not a proven rescue maneuver, and flumazenil benzodiazepine reversal is inconsistent at best.

Response to surgical PSVA attenuation is judged on the basis of clinical signs, normalization of routine clinicopathologic tests, resolution of ammonium biurate crystalluria, and decline of total serum bile acid concentrations. However, many dogs retain abnormally increased serum bile acid values. This status can be attributed to continued small flow shunting, failure of shunt ligation, presence of two PSVA (only 1 ligated), ligation of the wrong vessel, development of acquired portosystemic shunts, or concurrent MVD. Prospective comparison of pre- and post-surgical Protein C activity suggests that this parameter better substantiates improved portovenous circulation despite continued high serum bile acid concentrations. Dogs with poor surgical outcome fail to increase Protein C activity > 70%. Dogs with Protein C activity > 70% before surgery are minimally symptomatic and can be fully ligated with excellent outcome. Observations suggest that Protein C activity > 70% reflects relatively low grade portosystemic shunting for which dietary therapy alone may be effective.

Medical Management PSVA

First line medical management of PSVA involves restriction of the dietary protein allowance (2.2 to 2.5 gm/kg body weight per day protein) and modifying the source of dietary protein to dairy or soy sources preferably, or while meat chicken. All red meats, fish, and eggs are avoided. Several excellent prescription diets formulated for dogs with hepatic encephalopathy have made a remarkable improvement treatment options and the success of medical therapy. Meals are fed frequently and in small portions to maximize digestion and assimilation. Providing continuous access to a dry prescription diet formulated for dogs with hepatic insufficiency and meal feeding the canned formula twice daily, has been a successful management strategy for many PSVA dogs of small breeds. Many dogs can be maintained exclusively on a strict dietary regimen without additional supplements (such as lactulose or metronidazole). However, if dietary restriction alone does not provide optimal response (recurrent hepatic encephalopathy, persistent ammonium biurate crystalluria) then additional medical interventions are used. The second line strategy is to incorporate lactulose to modify the enteric environment. A synthetic disaccharide that undergoes microbial fermentation in the enteric canal, lactulose derived organic acids acidify the enteric environment, cripple ureases and proteases, trap the ammonium ion, and induce a colonic catharsis. Lactulose also augments bacterial nitrogen fixation further reducing availability of ammoniagenic products in the enteric canal. Careful dosing to achieve several soft stools per day is advised; individual titration starting at a low dose [0.5 ml per 5 to 10 kg, PO, BID to TID] is titrated to effect. Dosing too high results in abdominal cramping, painful borborygmi, flatulence, and liquid diarrhea that can progress to hematochezia. Metronidazole is used if response to lactulose and diet prove insufficient; 7.5 mg/kg PO SID to BID. Metronidazole can cause inappetence due to dysgeusia and rarely, granulocytopenia. Clavamox can alternatively be used as an enteric modifying antimicrobial intolerant of metronidazole. Neomycin is uncommonly used in the author's hospital; as an aminoglycoside, chronic neomycin absorption in dogs with inflammatory bowel disease has produced otoxicity (2 dogs) and renal damage (1 dog). Low liver zinc concentrations consistently found in liver tissue of PSVA dogs are thought to reflect reduced intake and increased urinary losses (surmised from human studies). We routinely increase zinc intake either through diet or as a separate supplement (zinc acetate: 1-5 mg elemental zinc per kg body weight, titrate against changes in plasma zinc concentrations) because of the importance of zinc in metalloenzymes, including some integral to the urea cycle. Dogs presented in hepatic coma, should receive cleansing enemas (lukewarm crystalloid fluids in very small patients). Thereafter, retention enemas containing lactulose can effectively deter production/absorption of colonic toxins. Crystalloid fluid therapy, avoidance of neuroglycopenia, and broad spectrum antimicrobials are usually provided. Treatment of severe hepatic encephalopathy is beyond the scope of this discussion. Dogs with PSVA do not need treatment with ursodeoxycholic acid. These dogs have high serum bile acids due portosystemic bile acid flux. Since most PSVA dogs lack necroinflammatory lesions they do not need s-adenosylmethionine nor other hepatoprotectants (e.g., milk thistle, phosphatidylcholine). Glutathione concentrations in dogs with PSVA confirm adequate antioxidant status in most dogs. Hepatic fibrosis is not a typical feature of PSVA in dogs.

Medical Management of MVD

Dogs with MVD do not require special diets for hepatic insufficiency, lactulose, antioxidants, ursodeoxycholate, or hepatoprotectants. Most MVD dogs live a full lifespan without any signs of hepatic insufficiency or hepatobiliary dysfunction. Examination of liver biopsies collected from elderly dogs diagnosed with MVD as puppies (early 1990s: Cairn Terriers, Tibetan Spaniels, Yorkshire Terriers) has verified a lack of progressive degenerative or inflammatory lesions. Dogs with zone 3 inflammation and veno-occlusive lesions usually have concurrent inflammatory bowel disease and are treated with hypoallergenic diets (soy based home cooked ration), anti-inflammatory dexamethasone (dosed q3days), and in some, ultra-low dose aspirin (0.5 mg/kg PO SID to EOD). Metronidazole is used adjunctively to manage the bowel inflammation. Knowledge of MVD is clinically useful when estimating dosages for drugs requiring first pass hepatic extraction. We recommend serum bile acid quantification in all puppies of pure breeds having prevalence for PSVA and MVD. Testing at is recommended at 4 to 6 months of age (paired samples, pre- & 2-hours post-meal) so that high bile acids are not discovered during assessments for non-hepatic illnesses.


1.  Lautt WW. Am J Physiol Gastrointest Liver Physiol 1985;249: G549.

2.  Lautt WW, et al. Can J Physiol Pharmacol 1988;66: 807.

3.  Lautt WW et al. Can J Physiol Pharmacol 1989; 67:1023.

4.  Yukihiro Y, et al. Am J Physiol Gastrointest Liver Physiol.2001; 280:21.

5.  Tisdall PL, et al. Aust Vet J. 1995;72:121.

6.  Mehl, et al. J Am Vet Med Assoc 2005;226:2020.

7.  Frankel, et al. J Am Vet Med Assoc.2006;228:1355.

8.  Toulza O, et al. J Am Vet Med Assoc. 2006;229:1761.

Speaker Information
(click the speaker's name to view other papers and abstracts submitted by this speaker)

Sharon Center, DVM, DACVIM
Cornell University
Ithaca, NY

MAIN : Liver Study Group : Portosystemic Vascular Anomalies
Powered By VIN