Congenital Extrahepatic Abdominal Arteriovenous Fistula and Apparent Patent Ductus Venosus in a Dog

Introduction

An arteriovenous (AV) fistula is an abnormal channel or passage between an artery and a vein. AV fistulae are reported infrequently in small animals.¹ They are classified as either congenital or acquired and as either central (cardiac) or peripheral.¹ In the dog, peripheral AV fistulae have been reported in various locations, including the prepuce, inguinal region after castration, metatarsal pad, subcutaneous tissue on the flank, spinal cord, intestine, and liver.^2–8 They can have congenital, traumatic, or neoplastic etiologies.⁹

Congenital portosystemic shunts may be classified as either intrahepatic or extrahepatic.¹⁰ Intrahepatic portosystemic shunts (IHPSS) may be further classified as left-, central-, or right-divisional.¹¹ Many left-divisional IHPSSs have a consistent morphology compatible with a patent ductus venosus (PDV).¹²

This report describes the investigation and treatment of an extrahepatic abdominal AV fistula and apparent PDV in a German shepherd dog.

Case Report

A 3 mo old male German shepherd dog was presented with a 2 wk history of diarrhea with possible melena followed by inappetence and progressive abdominal distension. Physical examination revealed the presence of a grade II/VI right basal heart murmur and marked abdominal distension with a fluid wave on ballottement. Hepatojugular reflux was not demonstrated.

Initial investigations included a minimum database and abdominal ultrasound. Hematology revealed thrombocytosis but was otherwise normal. Serum biochemical analysis revealed a slightly elevated blood urea nitrogen (BUN, 7.6 mmol/L reference range, 3.0–10.0 mmol/L) and alanine aminotransferase (1.44 μkat/L; reference range, 0.28–1.35 μkat/L). A moderate hypoalbuminemia (18 g/L; reference range, 21–43 g/L) and hypercholesterolemia (13.92 mmol/L; reference range, 3.20–6.20 mmol/L) were also noted. Elevations in Ca (2.66 mmol/L; reference range, 1.98–2.38 mmol/L) and phosphorus (2.52 mmol/L; reference range, 0.74–2.10 mmol/L) were considered unremarkable for a 3 mo old dog. Fasting bile acids were normal (9 μmol/L; reference range, 0–15 μmol/L). Fasting ammonia and postprandial bile acids were elevated at 156.6 μmol/L(reference range, 0–60 μmol/L) and 155 μmol/L (reference range, 0–15 μmol/L), respectively.

Abdominocentesis was performed to obtain diagnostic samples and to relieve the dog's discomfort. Fluid analysis confirmed the fluid to be a pure transudate (total protein was 1.0 g/L; reference range <25 g/L, nucleated cell count was 0.1 × 10⁹/L; reference range <1-1.5 × 10⁹/L). A urine sample was obtained and showed significant proteinuria with urine protein/creatinine ratio of 8.35 (reference range, <0.5). The sediment analysis showed the presence of WBCs (5–10/high-power field [HPF]; reference range, <5/HPF), scant red blood cells (5/HPF; reference range, <5/HPF), scant epithelial cells (5/HPF; reference range, <5/HPF), and struvite crystals. Bacterial culture was negative. After draining 200 L of fluid, abdominal palpation revealed the presence of fremitus in the caudodorsal abdomen, and auscultation of this area revealed a flow murmur.

Abdominal ultrasonography confirmed the presence of peritoneal effusion. The study also confirmed the presence of a distended aneurysmal ileocolic vein, measuring 3 cm in diameter at its widest point in the caudal abdomen (Figures 1, 2). That vessel could be followed tapering cranially to the porta hepatis. Color and pulsed-wave Doppler ultrasonography showed the flow within the aneurysm to be markedly turbulent. Parallel to that vein, a large vessel with unidirectional pulsatile flow (showing systolic peaks with a high resistance waveform) was also identified (Figures 3, 4). That vessel could be followed cranially to the cranial mesenteric artery and the aorta and caudally to enter the portal vein aneurysm at a level approximately ventral to the fifth lumbar vertebrae. An extrahepatic arterioportal fistula was diagnosed. Multiple abnormal vessels in the mesentery were thought to represent acquired portosystemic shunts. The flow within the portal vein was initially turbulent at the level of the fistula, which then became hepatopetal but slow (mean flow was 13 cm/sec, slowing to 8 cm/sec) towards the porta hepatis, where the vein narrowed to a more normal size. Those findings were considered consistent with portal hypertension.¹³

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

A portosystemic shunt was identified in the left division of the liver, consistent with a PDV.¹¹ The vessel appeared to be a continuation of an enlarged left branch of the intrahepatic portal vein. The shunting vessel made a distinct bend to the left (towards the transducer when using a left ventral intercostal window, and away from the transducer when using a right dorsal intercostal window) prior to entering an ampulla on the left side of the caudal vena cava at the level of the diaphragm.¹²

Echocardiography was performed and those findings were considered unremarkable. The heart murmur was considered innocent in nature.¹⁴

The dog was treated with amoxicillin^a (10 mg/kg per os [PO] q 12 hr), lactulose syrup^b (3 mL PO q 12 hr), and a prescription dog food^c. Systolic blood pressure was normal (120 mm Hg), and benazepril^d (0.25 mg/kg PO q 24 hr) was prescribed to reduce glomerular protein loss. Repeat abdominocentesis performed daily was required to maintain the dog's comfort. Diuresis was attempted with a combination of furosemide^e (2 mg/kg PO q 8 hr) and spironolactone^f (2.5 mg/kg PO q 24 hr) but was considered ineffective.

The dog was treated medically for 15 days prior to surgical intervention. Under general anesthesia, the abdomen was approached via a midline celiotomy. Approximately 400–500 mL of ascitic fluid was removed from the peritoneal cavity with surgical suction. Examination of the abdominal contents confirmed the presence of a single AV fistula in the region of the ileocecocolic junction (Figures 5, 6). The adjacent ileocolic and cranial mesenteric veins were massively distended. The liver was considered subjectively large for a dog with an IHPSS, but grossly, the external appearance of the liver lobes was considered relatively normal. There was gross evidence of multiple acquired portosystemic shunt formation between the portal vein and the caudal vena cava in the region of the root of the mesentery. The vessels involved in the fistula appeared to be a branch of the cranial mesenteric artery and the ileocolic vein. A jejunal vein was cannulated and mesenteric venous (portal) pressure was 13 cm water (H₂O) using manometry. Manual occlusion of the artery at the site of the junction with the vein resulted in cessation of fremitus, elevation in diastolic blood pressure, and the formation of a dicrotic notch on the blood pressure waveform. Manual occlusion also resulted in a transient bradycardia (resting heart rate was 110 beats/min, which decreased to 80 beats/min during occlusion). The portal pressure dropped to 3 cm H₂O, and the liver was observed to decrease in size. No significant deleterious changes in the cardiovascular parameters were noted, and the decision was made to fully close the AV communication by double ligation^g.

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

Visual examination of the left cranial aspect of the liver demonstrated the presence of a left-divisional IHPSS consistent with a PDV. The PDV was isolated as previously described and was encircled using a single ligature of 2-0 polypropylene (Figure 7).¹⁵ Following the temporary complete ligation of the shunt, the portal pressure remained unchanged at 3 cm H₂O. There was no other clinical evidence of portal hypertension, and the vessel was fully attenuated using a single ligature of polypropylene.

View Figure

• Download as Powerpoint
• Download Figure

A liver biopsy was obtained from the left lateral lobe of the liver, and a core biopsy was obtained from the left kidney. Histopathology of the liver biopsy showed primary hypoplasia of the portal veins with centrilobular degeneration of hepatocytes. Findings from the renal biopsy showed a chronic glomerulopathy with mesangial expansion by pink hyaline-like proteinaceous material. The material was not definitively identified but was thought to be a proteinaceous deposit caused by either a congenital anomaly of the glomerular basement membrane or amyloidosis.

The dog made a good recovery after surgery and was discharged to the owner's care 36 hr postoperatively with oral amoxicillin and lactulose for 3 wk at doses previously prescribed. The dog was bright and appetent for 1 mo postoperatively eating a conventional adult diet with no clinical signs or evidence of ascites. One week after stopping antibiotic treatment, the owner reported that the dog's appetite had decreased. Because there were acquired shunts seen at the time of surgery, management of hepatic encephalopathy was reinstigated, and blood and urine were submitted for analyses. The results showed persistence of the high urine/protein creatinine ratio (13.14), marked elevation of BUN (33.9 mmol/L), and mild elevation of creatinine (168 μmol/L; reference range, 44–141 μmol/L). There was persistence of the hypoalbuminemia (1.8 g/dL), and fasting bile acids were elevated (75.9 μmol/L). The urine specific gravity was 1.013, which (in conjunction with the elevation in BUN) was consistent with renal dysfunction. A urine sample was submitted for bacterial culture, which proved negative. Administration of benazepril was reintroduced in an attempt to reduce the proteinuria. Despite supportive treatment and IV fluid therapy, the dog deteriorated. One week later, the dog had significant changes in BUN (55 mmol/L), creatinine (353.6 μmol/L), and severe hyperphosphatemia (8.4 mmol/L)and hyperkalemia (7.88 mmol/L; reference range, 3.5–4.5 mmol/L). The owners elected for the dog to be euthanatized and declined a postmortem examination.

Discussion

AV fistulae are reported infrequently in domestic animals.¹ To the authors' knowledge, this is the first report of a congenital extrahepatic abdominal AV fistula and concurrent PDV IHPSS in a dog. The cause of the AV fistula was not established conclusively, but due to the dog's young age at presentation and the lack of both previous abdominal trauma and surgical intervention, a congenital etiology was considered most likely.

Peripheral AV fistulae result in reduced peripheral vascular resistance and arterial hypotension. Reflexes to maintain arterial blood pressure result in Na and H₂O retention, tachycardia, and increased cardiac output. As blood volume increases, collateral circulation develops as pre-existing arteries dilate. As the volume of blood being shunted through the fistula increases, there is a corresponding increase in venous return to the heart, ultimately leading to high output cardiac failure.¹ Under normal circumstances, an AV fistula between a branch of the cranial mesenteric artery and the ileocolic vein would not result in such changes because drainage of the ileocolic vein into the hepatic portal capillary network would prevent the AV fistula from having direct access to a systemic vein. On the contrary, the presence of a large IHPSS (as seen in this dog) would allow blood in the ileocolic vein to bypass the portal capillary network, providing a direct connection between the ileocolic vein and the caudal vena cava. Under those circumstances, the fistula would behave as a peripheral AV fistula. Despite the AV fistula in this case having a direct connection with the systemic venous circulation, there was no evidence of high output cardiac failure at the time of presentation. Further, the dog had no signs of cardiac failure and echocardiographic findings were considered unremarkable.

The incidence of congenital portosystemic shunts in German shepherd dogs is considered low in the United Kingdom. Out of 367 dogs with congenital portosystemic shunts seen by the one of the authors over the last 10 yr, only 6 shunts (1.6%) were found in German shepherd dogs (R.W., unpublished data, December 2012). In all 6 of those dogs, the shunt observed was categorized as left divisional, and in 5 dogs it was confirmed to be a PDV.

The causal mechanism for closure of the PDV remains unclear in domestic species. In the dog, it is considered to have closed between 2 and 10 days postpartum.^16–18 In humans, there has been an assumption that most unusual, nonsurgical, portacaval anastomoses have been found in patients with both raised portal pressure and liver cirrhosis.¹⁹ It has been postulated that under such circumstances recanalization of the ductus venosus might occur in response to the rise in portal pressure. In the dog described in this report, the presence of a congenital AV fistula between a branch of the cranial mesenteric artery and the ileocolic vein induced a rise in portal pressure. The congenital nature of the fistula would suggest that the elevation in portal pressure would have been present from the time of the dog's birth. As such, rather than causing a recanalization of the ductus, the elevation in portal pressure might have prevented the closure of the ductus venosus in the first instance.

It would be unusual for a dog with a congenital IHPSS to demonstrate the concurrent presence of multiple acquired portosystemic shunts. The study authors hypothesize that the cause of the multiple acquired shunts in this dog was related to the extrahepatic abdominal AV fistula. The presence of the AV fistula subjected the portal vein to arterial blood flow, leading to portal hypertension and the development of compensatory multiple acquired shunts. It might be argued that the presence of the IHPSS should have allowed for sufficient portal venous pressure run off to prevent the formation of portal hypertension. The study authors postulate that the presence of the IHPSS permitted significant pressure run off, but despite the reduction in portal pressure it was insufficient to lower the portal pressure sufficiently to prevent the development of portal hypertension and multiple acquired portosystemic shunts. Reasons for this might include the pulsatile nature of the AV fistula, the physical distance of the AV fistula from the IHPSS, and the relative size of the IHPSS compared to the portal vein and the caudal vena cava.

The histopathology results for this dog confirmed the presence of primary hypoplasia of the hepatic parenchymal portal veins. With such a finding, it might be considered unusual that the IHPSS could be completely closed without causing the subsequent development of portal hypertension. The study authors believe that complete closure of the IHPSS was possible in this dog due to the presence of pre-existing multiple acquired shunts. Angiography/venography are often considered the optimal technique for planning the treatment of AV communications and for the classification and treatment planning for IHPSSs.²⁰ Facilities for intraoperative portography were unavailable in this case. Despite the lack of those facilities, it proved possible to recognize and manage both the AV fistula and the apparent PDV with abdominal ultrasound and an exploratory abdominal surgery.

Hypoalbuminemia can be seen with hepatic synthetic failure and is common in patients with hepatic vascular anomalies. In this case, however, the hypoalbuminemia was considered most likely to be secondary to the severe protein-losing nephropathy. Of note, for a dog with a large IHPSS, although the postprandial bile acids and fasting ammonia concentrations were both elevated the fasting bile acids concentration was normal. The sensitivity of fasting bile acids for detection of a portosystemic shunt is reported to be 93% with a specificity of 67%; therefore, it is not that uncommon for this parameter to be normal in individuals with a portosystemic shunt.²¹

Further diagnostic evaluation of the kidney biopsy was not available. It would have been helpful to have further diagnostic testing performed to identify the material that was causing the mesangial expansion. Disease affecting the mesangial cells alone does not usually affect the filtration barrier. As such, the significant proteinuria seen in this dog would not be expected without the concurrent glomerulopathy. Damage to glomerular cells can result in either increased production of mesangial matrix or proliferation of mesangial cells. The description of the light microscopy in this case was more consistent with expansion of the mesangial matrix than proliferation of the mesangial cells. Further staining of the renal biopsy with either Congo red or standard toluidine blue could have been used to assess amyloid deposition. In sections that were negative for amyloid, other stains such as periodic acid-Schiff and hematoxylin could have been used to investigate other causes of mesangial expansion.²² In this dog, the etiology of the renal disease remained unknown. Despite the lack of a definitive diagnosis for the renal disease, nephrotic syndrome, in general, carries a poor prognosis. In this case, it was the progression of the renal disease, rather than the vascular anomalies, which necessitated the dog to be euthanatized.

Conclusion

To the best of the study authors' knowledge, this is the first report of a congenital extrahepatic abdominal AV fistula and concurrent IHPSS in a dog. Although not established conclusively, a congenital etiology for both vascular anomalies was considered most likely. The presence of the apparent PDV bypassed the portal capillary network, producing a direct connection between the portal vein and the caudal vena cava. As such, portal blood flow was in direct communication with the systemic venous circulation. Under those circumstances, the fistula involving the cranial mesenteric artery and the ileocolic vein behaved as a peripheral AV fistula. In addition, an idiopathic glomerulonephropathy was noted in this case.