Study population

Client-owned dogs that were diagnosed with IHPSS and treated with PTCE using a Venovo stent at The University of Georgia (UGA) Veterinary Teaching Hospital from 2020 through 2024 were included in the study. Cases were identified by electronic medical record review and were included if the dog was diagnosed with an IHPSS and underwent PCTE with a Venovo stent. Dogs that had stents other than the Venovo implanted during the procedure were excluded from the study. All dogs underwent CT angiography under anesthesia for definitive diagnosis of IHPSS and planning for PTCE.

Medical records review

Preoperative data collected from medical records of dogs in this study included presenting clinical signs, CBCs and serum biochemistry, available liver function tests (pre- and postprandial bile acids, ammonia levels), diagnostic imaging (ultrasonography and CT angiography), division (right, left, or central) and number (single or multiple) of IHPSSs, medications, and diet. Intraoperative data collected included signalment and weight at the time of procedure, vascular access site, pre- and postembolization caudal vena cava and shunt vessel pressures, Venovo stent size (diameter X length, millimeters), number of coils placed, size of coils placed (diameter X length, millimeters), and adverse events. Fluoroscopy-guided angiography was performed for further assessment of the IHPSS and to facilitate stent selection. Measurements, including the diameter of caudal vena cava cranial and caudal to the shunt orifice and the length of shunt defect, were recorded. Postoperative data included hospitalization time, adverse events, resolution or persistence of clinical signs, any ongoing medications or diet for management of IHPSS at last follow-up, available persistent CBC and serum biochemistry abnormalities, immediate and late postoperative imaging results, and patient outcomes, including survival time (date of procedure until date of death). Intraoperative and postoperative adverse events were categorized as major or minor. Major adverse events were defined as life threatening and required intensive medical intervention. Minor adverse events were self-limiting and responsive to minimal or no intervention. Postoperative data were collected through communication with primary veterinarians, owners, records if recheck appointments were performed at the UGA, or telemedicine consults if recheck appointments were performed with referring veterinarians.

Percutaneous transvenous coil embolization procedure

The PTCE procedure was performed similar to previous reports.^1–3,7,8 Premedication and induction agents varied and were chosen based on anesthesiologist preference. Combinations of premedication drugs generally included an opioid, such as hydromorphone (0.1 mg/kg), methadone (0.3 mg/kg), butorphanol (0.2 to 0.4 mg/kg), or buprenorphine (0.01 to 0.02 mg/kg), along with another agent, such as midazolam (0.2 mg/kg), dexmedetomidine (1 to 5 μg/kg), alfaxalone (2 mg/kg), or lidocaine (2 mg/kg) at standard doses. Maropitant and metoclopramide were included in some anesthetic protocols at 1 mg/kg each. Induction agents included the use of single-agent propofol (3 to 6 mg/kg) or a combination of propofol and ketamine (1 to 2 mg/kg). Some dogs received a lidocaine continuous rate infusion at 2 to 3 mg/kg/h, and 1 dog received a ketamine continuous rate infusion at 1 mg/kg/h during the procedure. All dogs were maintained under general anesthesia with isoflurane or sevoflurane, positioned in dorsal or left lateral recumbency, and the ventral cervical region was clipped and aseptically prepared. A stab incision was made over the external jugular vein using a #11 scalpel blade. An 18-gauge X 2-inch or 22-gauge X 1-inch over-the-needle catheter was used to obtain percutaneous access. A 0.018-inch X 40-cm exchange wire was advanced into the jugular vein, and the catheter was exchanged for a micropuncture introducer set. A J-tipped wire was advanced through the needle into the vena cava. An introducer sheath (Pinnacle Introducer Sheath; Terumo Interventional Systems) with a dilator was advanced over the wire, and the dilator and wire were removed. The introducer was sutured in place to the skin. An angled-tip hydrophilic guidewire and a 4-French (Fr) X 65-cm or 4-Fr X 100-cm non–taper-angled catheter (Glidecath Catheter; Terumo Interventional Systems) were advanced into the shunt vessel. Fluoroscopic guidance confirmed catheter location in the shunt vessel, and the guidewire was removed. A shunt vessel pressure was obtained through the Glidecath. A 5-Fr X 65-cm or 5-Fr X 100-cm pigtail marker catheter was passed alongside the Glidecath into the caudal vena cava with the tip positioned caudal to the shunt communication to the vena cava. Pressure was assessed in the caudal vena cava using the pigtail catheter. Simultaneous iohexol contrast injections (Omnipaque, GE Healthcare; 350 mg I/mL) were given through the catheters to perform dual portocavography using fluoroscopy. The Glidecath was then removed. A Bentson guidewire was passed through the pigtail catheter, and the catheter was removed over the wire. A Venovo venous stent system (Becton Dickinson and Co; Figure 1) was prepared and passed over the wire into the caudal vena cava. The center of the stent was positioned to span the shunt ostium at least 20 mm above and 20 mm below the shunt orifice. The stent was deployed under fluoroscopic guidance using the track wheel (Figure 2; Supplementary Video S1), and the delivery system was removed, maintaining wire position. The Glidecath was passed over the Bentson wire and into the shunt vessel under fluoroscopic guidance, and then the wire was removed. Pressures from the shunt were obtained following the deployment of the stent. A second Glidecath was advanced into the distal aspect of the shunt vessel over a hydrophilic guidewire. The wire was removed, and iohexol contrast was administered through the second Glidecath to evaluate the shunt vessel flow immediately before coil placement. The Glidecath was positioned in the shunt vessel to allow pressure measurement during coil deployment. If the pressures were appropriate, an embolization coil (MReye Embolization Coil; Cook Medical Inc) was advanced through the Glidecath and deployed into the shunt vessel under fluoroscopic guidance. Angiogram via hand injection of iohexol contrast was performed through the straight Glidecath within the shunt vessel to evaluate flow and obtain pressures. This process was repeated for subsequent coils until adequate flow attenuation was achieved or portal pressures had increased to the point where acute portal hypertension was a concern. To avoid acute portal hypertension, previous guidelines recommend a portal-caval pressure gradient increase of no more than 5 to 6 mm Hg, which was achieved in the current study.³ The total amount of iohexol contrast used during each of these procedures ranged from 1 to 5 mL/kg. Various coil sizes and numbers used were based on clinician experience, monitoring of pressures within the shunt or portal vein and vena cava, and repeated shunt angiography to determine flow from shunt vessel to vena cava. The catheters were removed, and the access site was closed in a Z-stitch pattern.¹⁶ A bandage was placed over the site, and a wrap was placed around the neck using cast padding and Vetwrap. Lateral and ventrodorsal radiographs centered over the diaphragm were taken to document the stent and coil placement (Figure 3).

View Full Size

Top panel: lateral view of an 18 X 100-mm Venovo venous stent. A—Lateral close-up view of the 3-mm flared ends and nitinol markers. B—Intraluminal view of the stent.

Citation: American Journal of Veterinary Research 2025; 10.2460/ajvr.25.02.0061

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Close-up view of the Venovo stent thumb wheel on the delivery system of a Venovo venous stent. The large thumbwheel, shown here, is used for slow deployment of the stent.

Citation: American Journal of Veterinary Research 2025; 10.2460/ajvr.25.02.0061

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Postoperative radiographic views demonstrating the position of the Venovo stent and coils. A—Left lateral radiograph from case 6. B—Ventrodorsal radiograph from case 6.

Citation: American Journal of Veterinary Research 2025; 10.2460/ajvr.25.02.0061

• Download Figure
• Download figure as PowerPoint slide

Statistical analysis

Descriptive statistics were calculated using Excel’s Analysis ToolPak, version 16.86 (Microsoft Corp). Normally distributed variables were expressed by the mean and SD, and non-normally distributed variables were expressed by the median and IQR.

Study population

Fourteen dogs with IHPSS that underwent PTCE with the Venovo stent were included. There were 10 mixed-breed dogs, with 3 Goldendoodles, 2 Labrador Retriever mixes, 1 Labradoodle, 1 Golden Retriever mix, 1 German Shepherd Dog mix, 1 Plott Hound mix, and 1 Miniature Australian Shepherd mix. The remaining breeds consisted of Golden Retriever (2), Labrador Retriever (1), and Doberman Pinscher (1). There were 8 intact male, 1 neutered male, 4 intact female, and 1 spayed female dogs. The mean age at the time of the procedure was 13.4 months (SD, 5.5 months), and the mean weight was 19.5 kg (SD, 6.9 kg).

Clinical presentation

All dogs had clinical signs before medical management was initiated, including lethargy (n = 9), polydipsia (8), polyuria (7), hyporexia (7), diarrhea (7), vomiting (6), ptyalism (6), altered mentation or behavior (6), head pressing (4), running into walls (4), ataxia (4), tremors or other seizure-like activity (3), weight loss (2), weakness (2), staring into space (2), pacing (2), acting blind (2), circling (2), abdominal distention (2), panting (1), epistaxis (1), and excessive vocalization (1). All dogs were on varying forms of medical management before the procedure for a median of 4.0 months (IQR, 3.0 to 7.3 months). Medications included lactulose (14), metronidazole (9), omeprazole (8), levetiracetam (7), pantoprazole (1), amoxicillin–clavulanic acid (1), and/or ursodiol (1). Twelve dogs were fed a low-protein diet at the time of the procedure, and 4 dogs received supplements, including probiotics, psyllium powder, folic acid, or omega 3 fatty acids, with their diet. Clinical signs improved in all dogs once medical management was initiated.

Four dogs had a normal physical examination on presentation to the UGA Veterinary Teaching Hospital. Physical examination abnormalities before the procedure included low body or muscle condition score (6), unilateral cryptorchid (2), grade 1/6 left apical systolic heart murmur (2), scrotal pyoderma (1), bilateral entropion (1), ear debris (1), miosis (1), and forelimb lameness (1).

Preoperative laboratory work and imaging

Data were collected from the most recently available CBC and serum biochemistry panel in dogs prior to the procedure after medical management. Complete blood count abnormalities included microcytosis (12 of 14 [86%]), anemia (8 of 14 [57%]), hypochromia (7 of 13 [54%]), thrombocytopenia (7 of 14 [50%]), leukocytosis (5 of 13 [38%]), monocytosis (3 of 13 [23%]), eosinophilia (3 of 12 [25%]), neutrophilia (2 of 14 [14%]), and lymphocytosis (2 of 13 [15%]). Serum biochemistry abnormalities included decreased BUN (9 of 14 [64%]), hypoproteinemia (9 of 14 [64%]), hypoalbuminemia (9 of 14 [64%]), elevated ALT (9 of 13 [69%]), elevated ALP (7 of 13 [54%]), hypocholesterolemia (7 of 13 [54%]), decreased creatinine (7 of 14 [50%]), hypomagnesemia (4 of 8 [50%]), hypoglobulinemia (6 of 14 [43%]), hyperphosphatemia (3 of 13 [23%]), hypocalcemia (3 of 14 [21%]), hypoglycemia (3 of 14 [21%]), hyperchloremia (3 of 13 [23%]), hypochloremia (1 of 13 [8%]), hyponatremia (1 of 13 [8%]), and hyperbilirubinemia (1 of 13 [8%]).

Ammonia levels prior to the procedure were available for 10 dogs and were elevated in 7 dogs (7 of 10 [70%]). Preprandial and postprandial bile acid testing was performed in 10 dogs prior to the procedure and were both elevated in 8 of those dogs (8 of 10 [80%]). One dog had normal preprandial and postprandial bile acids (1 of 10 [10%]), and another dog had normal preprandial but elevated postprandial bile acids (1 of 10 [10%]). Three dogs had only resting bile acids performed, and those values were elevated. One dog did not undergo bile acid testing. Overall, of the 13 dogs where some form of bile acid testing was performed, there were elevations in 12 of 13 (92%).

An abdominal ultrasound was performed in 8 of 14 dogs prior to referral. All dogs underwent CT angiography under general anesthesia with breath hold to evaluate the shunt for procedural planning. All dogs had a single IHPSS identified. These were further classified into a right divisional IHPSS in 7 dogs (7 of 14 [50%]), left divisional IHPSS in 5 dogs (5 of 14 [36%]), and central divisional IHPSS in 2 dogs (2 of 14 [14%]).

Before the procedure, 7 of 14 dogs had a urinalysis performed. These results revealed no abnormal findings in 2 of 7 dogs (29%), bacteriuria in 3 of 7 dogs (43%), and crystalluria in 4 of 7 dogs (57%). Between abdominal ultrasound, CT imaging, and urinalysis findings, there was evidence of crystalluria or urolithiasis in 12 dogs (12 of 14 [86%]), suspected or confirmed to be ammonium urate secondary to the shunt. Preoperative case details are summarized in Supplementary Table S1.

Percutaneous transvenous coil embolization procedure

Percutaneous vascular access was obtained through the right jugular vein in 12 dogs (12 of 14 [86%]) and the left jugular vein in 2 dogs (2 of 14 [14%]). Fluoroscopic angiography was performed in all dogs during the operation, and measurements are included in Table 1. The mean vena cava diameter cranial to the shunt orifice was 15.4 mm (SD, 2.8 mm). The mean vena cava diameter caudal to the shunt orifice was 17.5 mm (SD, 3.7 mm). The approximate length that spanned these cranial and caudal measurements was available for 13 dogs, with a mean length of 78.3 mm (SD, 12.4 mm). The most commonly selected Venovo stent size (diameter X length) was 18 X 100 mm (n = 3), followed by 14 X 60 mm (2), 20 X 100 mm (2), 12 X 80 mm (1), 14 X 80 mm (1), 16 X 80 mm (1), 18 X 80 mm (1), 20 X 80 mm (1), 18 X 120 mm (1), and 20 X 140 mm (1). The mean ratio of stent diameter to caval diameter cranial to the shunt orifice was 0.9 (SD, 0.1). The mean ratio of stent diameter to caval diameter caudal to the shunt orifice was 1.0 (SD, 0.1). The mean number of coils placed was 4.6 (SD, 3.2 coils). Coils were placed in 12 of 14 dogs, and coil sizes (diameter X length) included 8 mm X 5 cm (31), 5 mm X 5 cm (21), 10 mm X 8 cm (10), 10 mm X 10 cm (1), and 3 mm X 2 cm (1). Median shunt vessel pressure preattenuation was 4.5 mm Hg (IQR, 3.0 to 5.1 mm Hg). Mean vena cava pressure preattenuation was 4.8 mm Hg (SD, 2.6 mm Hg). Mean shunt vessel pressure post-PTCE was 6.6 mm Hg (SD, 3.2 mm Hg). Mean vena cava pressure post-PTCE was 4.4 mm Hg (SD, 2.6 mm Hg).

Early postoperative results and adverse events

Six minor adverse events occurred in 4 dogs total (4 of 14 [29%]). Two dogs had a minor intraoperative and a minor postoperative adverse event. Three minor intraoperative adverse events occurred in 3 of 14 dogs (21%), including intermittent ventricular premature complexes (2) and blood loss from the jugular access site (1). The 2 dogs that had ventricular premature complexes had received dexmedetomidine and ketamine as premedication and induction agents. The intermittent arrhythmias did not require intervention, and hemostasis was achieved with manual pressure to control minor hemorrhage from the access site. No major intraoperative adverse events were reported. Immediate postoperative abdominal radiographs including the entire stent were performed in all dogs. One dog had a coil that was suspected to have moved ventrally within the intrahepatic shunt vessel postoperatively (1 of 14 [7%]), and the remainder of the radiographs revealed appropriate stent and coil positions. Early postoperative adverse events occurred in 3 of 14 dogs (21%), which were all minor and did not require significant intervention. The adverse events included regurgitation during recovery (1), mild anemia secondary to intraoperative bleeding and subsequent dilution from IV fluid administration (1), and incisional bleeding approximately 3 hours after the procedure after the dog became excited when it started to eat (1). One of the dogs that had intermittent ventricular premature complexes intraoperatively also had postoperative incisional bleeding, which occurred as the patient became excited when offered food, and strikethrough from the bandage was noted. Hemostasis was achieved with manual pressure and sedation. The dog that had blood loss from the jugular access site intraoperatively also had mild anemia postoperatively. No major postoperative adverse events were reported. The hospitalization time from day of procedure to time of discharge for all dogs was 1 day.

Ultimate clinical outcome

Thirteen of 14 dogs had an initial follow-up appointment at the UGA Veterinary Teaching Hospital or the primary veterinarian after the procedure in a median of 36.0 days later (IQR, 29.0 to 59.5 days). The time from procedure to last available follow-up was a median of 182 days (IQR, 58 to 588 days). All dogs initially had resolution of clinical signs, with 6 dogs (6 of 14 [43%]) that did not experience recurrence of clinical signs from the time of procedure to the last available follow-up. The other 8 dogs experienced similar clinical signs (8 of 14 [57%]) during the follow-up period, with 3 dogs exhibiting a multitude of signs (gastrointestinal, neurologic, and urinary related). Three dogs had developed neurologic signs by the last available follow-up (3 of 14 [21%]). One of these dogs was transitioned back to an exclusive hepatic diet, and the clinical signs improved. Three dogs developed urocystoliths since the last available follow-up (3 of 14 [21%]), and 2 of those dogs underwent a procedure for removal (1 cystotomy 758 days postprocedure and 1 percutaneous cystolithotomy 680 days postprocedure). The dog that underwent a cystotomy also had 2 mucosal polyps on the bladder that were biopsied at another specialty hospital, but those results were lost to follow-up. The dog tested BRAF negative before the procedure. Three dogs experienced chronic recurring urinary tract infections after the procedure by the time of the last available follow-up (3 of 14 [21%]). Two of those 3 dogs had a history of urinary tract infections before the procedure as well. Five dogs developed similar gastrointestinal signs, including vomiting, diarrhea, and/or inappetence (5 of 14 [36%]). One of these dogs had a septic abdomen and underwent an exploratory laparotomy at 280 days post-PTCE where a perforated duodenal ulcer was excised. Another 1 of these 5 dogs was diagnosed with exocrine pancreatic insufficiency at 78 days postprocedure, which was believed to be the cause of the gastrointestinal symptoms.

Postprocedural radiographs at a subsequent recheck visit were available for 9 of 14 dogs (9 of 14 [64%]). For 7 of these dogs, the radiographs were performed at a 1- or 2-month follow-up appointment for the purposes of rechecking the stent and coils at a median of 29 days (IQR, 26 to 41 days) postprocedure. All radiographs revealed that the stents and coils were in a similar position, with no stent fractures. In 4 of the 9 dogs where radiographs were available, the images were taken for other diagnostic purposes, including 2 dogs with urinary signs for suspected cystoliths (589 and 680 days postprocedure), 1 dog with gagging and restlessness (278 days postprocedure), and 1 dog with gastrointestinal signs for suspected septic abdomen (280 days postprocedure). No abnormalities with the stent and coils were noted in these reports.

At the time of the last available follow-up, 7 dogs were still on medications, including lactulose (3), omeprazole (3), famotidine (1), pantoprazole (1), levetiracetam (1), metronidazole (1), and esomeprazole (1). Six dogs were still being fed a protein-restricted or hepatic diet. Postoperative laboratory results were not available for 1 dog. All dogs with available bloodwork results had persistence of some laboratory abnormalities. The most common laboratory abnormalities postoperatively included microcytosis (n = 11), elevated ALT (10), thrombocytopenia (8), decreased BUN (6), hypochromasia (4), hypoproteinemia (4), hypoalbuminemia (3), elevated ALP (3), leukocytosis (3), hypercholesterolemia (2), anemia (2), hypoglobulinemia (2), neutrophilia (1), hypoglycemia (1), hypomagnesemia (1), hyperphosphatemia (1), and decreased creatinine (1). Abnormal preoperative and postoperative bloodwork results are shown in Supplementary Table S2.

At the time of data collection, 2 of the 14 dogs were deceased (2 of 14 [14%]). One dog died 75 days postprocedure after developing seizures secondary to hypoglycemia that was unresponsive to treatment and subsequently going into cardiac arrest at an outside emergency hospital. This dog had been doing clinically well prior to the day of cardiac arrest, with complete resolution of clinical signs and improvements in bloodwork abnormalities, including an increase in albumin (1.9 g/dL preprocedure, 3.1 g/dL postprocedure) and BUN levels (4 mg/dL preprocedure, 9 mg/dL postprocedure). While unproven, the suspected cause of death for this dog was xylitol toxicity based on history, clinical signs, and severe unresponsive hypoglycemia. Another dog died 867 days postprocedure from an unknown cause. The dog also had complete resolution of clinical signs and improvements in bloodwork postoperatively, including an increase and normalization of albumin (2.3 g/dL preprocedure, 2.5 g/dL postprocedure), BUN (6 mg/dL preprocedure, 7 mg/dL postprocedure), and Hct (32.5% preprocedure, 39.4% postprocedure) levels. The dog had persistent microcytosis (52 fL preprocedure, 50 fL postprocedure) and hypoproteinemia (3.8 g/dL preprocedure, 4.4 g/dL postprocedure) after the procedure that had either remained static or improved. Postoperative case details are summarized in Supplementary Table S3.