• Open Access

Splenosystemic Shunts in Cats: A Retrospective of 33 Cases (2004–2011)

Authors

  • J-S. Palerme,

    1. Department of Clinical Sciences, College of Veterinary Medicine, North Carolina State University, Raleigh, NC
    Search for more papers by this author
  • J.C. Brown,

    1. Department of Molecular Biomedical Sciences , College of Veterinary Medicine, North Carolina State University, Raleigh, NC
    Search for more papers by this author
  • S.L. Marks,

    1. Department of Clinical Sciences, College of Veterinary Medicine, North Carolina State University, Raleigh, NC
    Search for more papers by this author
  • A.J. Birkenheuer

    Corresponding author
    1. Department of Clinical Sciences, College of Veterinary Medicine, North Carolina State University, Raleigh, NC
    • Corresponding author: Adam J. Birkenheuer, Department of Clinical Sciences, College of Veterinary Medicine, North Carolina State University, 1060 William Moore Drive, Raleigh, NC 27606; e-mail: ajbirken@ncsu.edu.

    Search for more papers by this author

Abstract

Background

Portosystemic shunts are uncommonly reported in cats. The majority of reports describe congenital shunts in young cats originating from the left gastric vein. Although they are only rarely reported, acquired portosystemic shunts in cats appear to be more variable in their anatomic location.

Hypothesis/Objective

To describe the signalment and disease conditions found in cats with splenosystemic shunts.

Animals

Thirty-three client-owned cats with documented splenosystemic shunts.

Materials and Methods

Retrospective study. All cats with vascular communications between the splenic and left renal veins or the splenic vein and caudal vena cava diagnosed ultrasonographically between 2004 and 2011 were included. Collected data included age, breed, sex, presenting complaints, clinicopathologic data, as well as clinical diagnosis when available.

Results

Splenosystemic shunts were identified in 1.3% of the cats that had an abdominal ultrasound performed during the study period. Older, spayed female cats were found to be significantly overrepresented when compared with the total population of cats having undergone ultrasound over the same time period. A large proportion of cats (42%) had a hepatopathy with the potential for associated portal hypertension.

Conclusions and Clinical Importance

Neither the signalment of cats in this report nor the anatomy of their portovascular anomalies shared similarities with those cats previously identified with single-vessel shunts. The relevance and etiology of these newly described splenosystemic shunts remain elusive and warrantsfurther investigation.

Abbreviations
CT

computed tomography

PH

portal hypertension

PSS

portosystemic shunt

Portosystemic shunts are reportedly less common in cats than in dogs.[1-6] Congenital portosystemic shunts (PSS) are the most commonly identified PSS in cats and are the result of an anomalous embryonic vascular development and can occur without any identifiable underlying physical cause.[2] In contrast, acquired PSS develop in response to prehepatic or hepatic portal hypertension (PH). Acquired PSS are not encountered with posthepatic PH, such as right-sided heart failure, because of the lack of any pressure gradient between the portal and systemic systems. Canine and feline experimental models of PH have demonstrated that acquired PSS result from the opening of pre-existing embryonic vessels between the portal and systemic circulation, creating a lower resistance pathway that diverts a portion of the portal blood flow into the systemic circulation.[7-10] Identified causes of PH resulting in acquired PSS are limited to 5 cats with hepatic fibrosis,[4, 11, 12] 2 with arteriovenous fistulas,[13, 14] 2 with portal vein occlusion by a thrombus or intraluminal mass,[13, 15] one with portal vein hypoplasia[16] and one that developed acquired shunts after attenuation of a congenital PSS.[11]

The main tributaries of the portal system are composed of the caudal and cranial mesenteric veins, the splenic vein, and the gastroduodenal vein. These tributaries are in turn formed by the convergence of smaller branches of the portal system. Acquired PSS may occur between any level of the portal system and the vena cava or its tributaries. The majority of feline models and clinical cases of PH in cats have shown that the development of a prominent vessel between the splenic vein and left renal or gonadal veins is the most constant feature.[9, 12, 13, 17, 18] Although they are commonly referred to as splenorenal shunts, splenosystemic may be a more appropriate term that takes into account the shunting between not only the main vessels but also their tributaries.

Prior description of naturally occurring splenosystemic shunts in cats has been limited to a handful of isolated cases.[12, 13, 18] The clinical significance of splenosystemic shunts in cats is unknown, and some clinicians at our institution have suspected them to be incidental findings in older cats. The purpose of this retrospective study was to describe the signalment, clinical signs, and clinicopathologic changes in cats with splenosystemic shunts.

Materials and Methods

Case Selection

All cats that had a splenosystemic shunt documented via abdominal ultrasound were identified in the Radiology Information System of the North Carolina State University Veterinary Health Complex between 2004 and 2011 for inclusion in the study. Cases were included if the shunts were defined as an anomalous vessel or plexus of vessels with a single origin from the splenic vein and a single termination in the left renal vein or caudal vena cava. All ultrasound reports and images were reviewed by a board-certified radiologist (J.C.B.). When available, other ultrasonographic data, such as portal vein velocity and flow direction, were also included. One cat had computed tomography (CT) examination of the abdomen performed in addition to ultrasound (Fig 1).

Figure 1.

Computed-tomography reconstruction of a splenorenal shunt. Ao, Aorta; PV, portal vein; SV, splenic vein; Sh, Shunt vessel; RV, left renal vein; VC, caudal vena cava.

Data obtained from the medical records included age, breed, sex, and presenting clinical signs of cats. When available, results of CBC, serum biochemistry profile, urinalysis, serum bile acid concentrations, and fasting ammonia concentrations that were submitted at the time of the initial ultrasound examination were included. Presence of cavitary effusion, assessed by radiographs, abdominal ultrasound, or echocardiogram was recorded. Cytologic analysis of effusion fluid was also included. When available, liver histopathologic results were also included.

A cohort group of cats was included in the study for age and sex comparison. This group was composed of the total population of cats that had an abdominal ultrasound performed by the North Carolina State University Veterinary Health Complex radiology service during the same time period, regardless of the underlying disease. Cats included in the shunt group were excluded from the cohort group.

Clinical Diagnosis

Cases were reviewed and categorized based on the potential of their underlying disease condition to cause prehepatic or hepatic PH. Classification in this category was based on the patient having a diagnosis of a condition with the potential for associated PH. Cases that either did not have a clear clinical diagnosis or that were diagnosed with a condition not known to have the potential for associated PH were classified as having a pathology with undetermined potential for associated PH.

Statistics

Sex differences were compared between the study and cohort groups using fisher's exact test. Age differences between the study and cohort groups were determined by converting age into a proportion of cases aged either 9 years or less and 10 years or older and performing a Fisher's exact test. P values <.05 were considered statistically significant.

Results

Signalment

Thirty-three cases of cats with splenosystemic shunts were identified (Table S1). The incidence of detection of these shunts in cats by abdominal ultrasound was 1.3% during the study period. Our study group was composed of 30 (91%) spayed females and 3 (9%) castrated males. In comparison, the cohort group had a total of 2415 cats, comprising 62 (2.6%) intact females, 1035 (42.9%) spayed females, 56 (2.3%) intact males, 1258 (52.1%) castrated males, and 4 (0.1%) cases in which the sex was not reported. The proportion of spayed females in the study group was significantly overrepresented compared with the cohort group (P < .001).

The age of the study group was 11.7 ± 4.4 years (mean ± SD). More specifically, our study population included 11 cats aged 9 years or less and 22 cats that were 10 years or older. The age of our cohort group was 9.6 ± 4.6 years (mean ± SD), composed of 1144 cats of 9 years or less and 1271 cats aged 10 years or older. The cohort and study group failed to differ significantly with respect to age (P = .117).

Domestic shorthairs were the most common breed in our study group, accounting for 25/33 cases, followed by 3 domestic longhairs and one each of the following breeds: Himalayan, Siamese, Tonkinese, Persian, and Oriental Shorthair. The most common presenting clinical signs of cats with shunts were vomiting (33%), anorexia/inappetance (30%), and weight loss (27%).

Clinical Diagnosis

Fourteen of the 33 cases (42%) had diseases affecting the hepatic parenchyma or biliary system that could potentially cause hepatic portal hypertension. Ten of the 33 cases (30%) were diagnosed with neoplastic diseases affecting the gastrointestinal tract, respiratory system, or hematopoietic system. Two patients (6%) were diagnosed with chronic kidney disease. One cat each was diagnosed with left-sided heart failure and diabetic ketoacidosis. Finally, 5 (15%) patients had no clear clinical diagnosis.

Clinicopathologic Findings

Serum biochemistry values and CBC data were available for 31 (92%) cases (Table 1 and 2). Urinalysis reports were available for 27 (82%) cats. Urine was isosthenuric to minimally concentrated (urine specific gravity 1.008–1.020) in 12 (44%) cases. However, no information about whether urine was collected before administration of intravenous fluids was available for any of the cases. Bilirubinuria was noted in 4 (15%) cats. Three of these cats were from the group of cats categorized as having a hepatopathy with the potential for associated PH. The remaining cat was diagnosed with diabetic ketoacidosis and was euthanized within 24 hours of admission, precluding further investigation into the presence or absence a hepatopathy.

Table 1. Biochemical abnormalities identified in cats with splenosystemic shunts
 Cats with a Hepatopathy with the Potential for PH: Number (%)Cats with an Undetermined Potential for PH: Number (%)
n↓RIWRI↑RIn↓RIWRI↑RI
  1. BUN, blood urea nitrogen; ALP, alkaline phosphatase; ALT, alanine aminotransferase; GGT, gamma-glutamyl transferase; CK, creatine kinase; ↓RI, value lower than reference interval; WRI, value within reference interval; ↑RI, value higher than reference interval.

Bilirubin130 (0)7 (54)6 (46)180 (0)16 (89)2 (11)
Albumin138 (62)5 (38)0 (0)185 (28)12 (67)1 (5)
Cholesterol110 (0)11 (100)0 (0)150 (0)13 (87)2 (13)
Glucose130 (0)11 (85)2 (15)180 (0)12 (67)6 (37)
BUN133 (23)5 (38)5 (38)181 (5)12 (67)5 (28)
ALP131 (8)6 (46)6 (46)181 (5)11 (61)6 (37)
ALT131 (8)5 (38)7 (54)182 (11)12 (67)4 (22)
GGT110 (0)9 (82)2 (18)160 (0)16 (100)0 (0)
Creatinine132 (15)7 (54)4 (31)182 (11)11 (61)5 (28)
Total protein136 (46)7 (54)0 (0)185 (28)9 (50)4 (22)
Globulin131 (8)12 (92)0 (0)160 (0)15 (94)1 (6)
Phosphorus121 (8)9 (75)2 (17)181 (5)14 (78)3 (17)
Calcium135 (38)6 (46)2 (15)182 (11)16 (89)0 (0)
Magnesium100 (0)3 (30)7 (70)150 (0)14 (93)1 (7)
CK100 (0)5 (50)5 (50)141 (7)11 (79)2 (14)
Sodium133 (23)10 (77)0 (0)183 (17)15 (83)0 (0)
Potassium130 (0)12 (92)1 (8)183 (17)14 (78)1 (5)
Chloride116 (54)3 (27)2 (18)165 (31)10 (63)1 (6)
Amylase131 (8)8 (61)4 (31)181 (5)13 (72)4 (22)
Lipase100 (0)9 (90)1 (10)160 (0)15 (94)1 (6)
Table 2. Hematologic abnormalities identified in cats with splenosystemic shunts
 Cats with a Hepatopathy with the Potential for PH: Number (%)Cats with an Undetermined Potential for PH: Number (%)
n↓RIWRI↑RIn↓RIWRI↑RI
  1. WBC, white blood cells; RBC, red blood cells; HGB, hemoglobin; HCT, hematocrit; MCV, mean cell volume; MCH, mean corpuscular hemoglobin; ↓RI, value lower than reference interval; WRI, value within reference interval; ↑RI, value higher than reference interval.

  2. a

    Cases that exhibited platelet clumping were excluded.

WBC131 (8)7 (53)5 (38)180 (0)10 (56)8 (44)
RBC136 (46)7 (53)0 (0)188 (44)9 (50)1 (6)
HGB138 (62)5 (38)0 (0)189 (50)9 (50)0 (0)
HCT1310 (77)3 (23)0 (0)189 (50)8 (44)1 (6)
MCV134 (31)9 (69)0 (0)181 (6)17 (94)0 (0)
MCH133 (23)9 (69)1 (8)182 (11)16 (89)0 (0)
Plateletsa93 (33)4 (44)2 (22)101 (10)7 (70)2 (20)
Packed cell volume127 (58)5 (42)0 (0)178 (47)8 (47)1 (6)
Plasma protein109 (90)1 (10)0 (0)172 (12)9 (53)6 (35)
Neutrophils130 (0)2 (16)11 (85)180 (0)5 (28)13 (72)
Bands130 (0)10 (77)3 (23)180 (0)15 (83)3 (17)
Lymphocytes138 (62)5 (38)0 (0)187 (39)11 (61)0 (0)
Monocytes131 (8)10 (77)2 (16)181 (6)12 (67)5 (28)
Eosinophils136 (46)6 (46)1 (8)183 (17)13 (72)2 (11)

Of the 2 cats that had only preprandial bile acid concentrations measured, 1 had an abnormal value (65.3 μmol/L, reference range: 0–10). Of the 2 cats that had both pre- and postprandial bile acid concentrations measured, 1 had an abnormal value (preprandial 22.7 μmol/L, reference range: 0–10 μmol/L and postprandial 35.4 μmol/L, reference range: 0–30 μmol/L). No cats categorized as having a hepatopathy with the potential for associated PH had bile acid measurements performed. Fasting ammonia concentrations were increased in 2 of the 3 cats in which they were measured (60 and 66 μmol/L, reference 8–52 μmol/L). One of these cats was categorized as having a hepatopathy with the potential for associated PH.

Histopathology

Histopathology of the liver was available for 8 cases, 7 of which were obtained at necropsy and 1 as a surgical biopsy. The most common finding was hepatic fibrosis, noted in 3 cases. Hepatocellular necrosis and hepatic lipidosis were noted in 2 cases. Although multifocal hemangiosarcoma was noted in 1 cat, histopathology was performed 34 months after the initial finding of the splenosystemic shunt. Consequently, as this cat had no definitive evidence of hepatic disease at the time the shunt was identified, it was not categorized as having a hepatopathy with the potential for associated PH.

Cavitary Effusions

Cavitary effusions were present in 10 cases categorized as having a hepatopathy with the potential for associated PH and in 8 cases categorized as having a pathology with undetermined potential for associated PH (Table 3). The presence of effusion was confirmed with radiographs in 9 cases, abdominal ultrasound in 18 cases, echocardiogram in 1 case, and necropsy in 1 case. Single cavitary effusion was present in 9 cases, all of which involved the peritoneal cavity. Five of these were subjectively noted as having trace quantities of effusion, 3 were of moderate quantity, and 1 was of marked quantity. Of the 8 cases in which bicavitary effusions were present, 7 involved the pleural and peritoneal spaces and 1 involved the pericardial and peritoneal spaces. Tricavitary effusion was noted in 1 case. Fluid analysis was available in 7 cases. Pure transudates with protein contents ranging from 2.1 to 2.5 g/dL and cell counts from 350 to 1000 cells/μL were present in 2 cases of peritoneal (one of which had a hepatopathy with the potential for associated PH) and 1 case of pleural effusions. One case of mixed inflammatory peritoneal effusion (2.5 g/dL protein, 880 × 103 cells/μL) and 2 cases of suppurative inflammatory peritoneal effusion (4.7 g/dL protein, 21.99 × 103 cells/μL and 2.7 g/dL, 7.28 × 103 cells/μL) were found in 3 cats having a hepatopathy with the potential for associated PH. In addition, 1 case of chylous pleural effusion (5 g/dL protein, 11.94 × 103 cells/μL) was also found.

Table 3. Presence and distribution of cavitary effusions in cats with splenosystemic shunts
Type of EffusionCats with a Hepatopathy with the Potential for PHCats with an Undetermined Potential for PH
No effusion411
Peritoneal45
Peritoneal and pericardial10
Peritoneal and pleural43
Tricavitary10

Shunt Imaging

All splenosystemic shunts were identified as an anomalous vessel arising from the splenic vein, coursing caudal to the left kidney assuming a generally tortuous path, and continuing to the left renal vein, caudal vena cava adjacent to the renal vein, or medial to the left kidney. Thirty of the cats (90%) had shunts described as being single. The 3 remaining cats, all from the group with a pathology with undetermined potential for associated PH, had shunts that originated and ended as single vessels, but formed a plexus at some point in their trajectory. Although all the shunts arose from the splenic vein, 13 (39%) cases terminated in the caudal vena cava and 14 (42%) cases terminated in the left renal vein (Table 4). In 6 (18%) cases, the vessel could only be followed medial to the left kidney, and the site of termination could not be stated definitively as renal or caval. Eight cats had more than 1 abdominal ultrasound examination performed. Of these, 3 cases had no shunt detected on initial examination, but the shunts were detected on repeat examinations performed between 17 and 105 days later (mean 72.3 days). The persistence of shunts was confirmed in 6 cases imaged between 9 and 601 days (mean 225.2 days) after initial identification. Resolution of shunts was not documented in any case. Characterization of the blood flow within the portal vein was available in a limited number of cases. The direction of the blood flow was noted in 8 (24%) cases and was found to be hepatopetal in 7 (88%) cats and hepatofugal in 1 (12%) cat. Three of the 7 cats with hepatopetal blood flow had portal velocities measured. One of these cats, categorized as having a hepatopathy with the potential for associated PH, had a portal velocity measurement below the reference range (<10 cm/s), suggesting portal hypertension.

Table 4. Site of shunt termination in cats with splenosystemic shunts
Site of TerminationCats with a Hepatopathy with the Potential for PHCats with an Undetermined Potential for PH
Left renal vein77
Caudal vena cava58
No definitive termination24

One cat had CT imaging of the abdomen performed identifying an elongated, tortuous vessel arising from the splenic vein and coursing caudal to the left kidney before terminating on the left renal vein (Fig 1).

Discussion

We describe a series of cats with single-vessel portosystemic shunts. It could not be determined whether these PSS are congenital or acquired in origin. However, these PSS do not appear to share similar clinical or anatomic features with previously described congenital PSS in cats.

Splenorenal or splenocaval shunts have not been described in any of the published retrospective studies of congenital PSS in cats.[1-3, 6, 19-21] The most common type of extrahepatic congenital PSS in cats involved the left gastric vein and the caudal vena cava. The most common clinical signs reported in cats with congenital PSS (ptyalism, neurologic signs) were rarely reported in our study group. In addition, the signalment of our study group differed in both age and sex distribution when compared with cats with congenital PSS. Although the age in previous studies ranged between 8 and 16 months, the mean age of our study group was of 11.7 years. Most strikingly, although previous studies reported either an equal sex distribution or an overrepresentation of males, spayed females were significantly (P < .001) overrepresented and accounted for 91% of the cats in our study group.

Unfortunately, the retrospective and descriptive nature of our report precludes us from making any conclusions as to the etiology of these shunts. However, we have 2 main hypotheses regarding their development: (1) they are acquired shunts secondary to past or present portal hypertension; or (2) they are congenital shunts of unknown clinical significance. It is interesting to note that a substantial portion (42%) of these cats had clinicopathologic or histopathologic changes that might suggest a hepatopathy with the potential of associated PH. Diseases of the biliary system were found in 5 cats in our study (4 cases of cholangitis and 1 case of biliary obstruction). Underscoring the link between biliary inflammation and periportal fibrosis, 1 retrospective study of cats with cholangitis found that 15/16 subjects had periportal fibrosis on histopathology.[22] Three cats in this study were diagnosed with polycystic disease. A retrospective study of feline polycystic disease demonstrated that almost 50% of cats with hepatic lesions had histopathologic evidence of periportal bridging fibrosis.[23] Furthermore, in a case report of 2 cats with polycystic disease, both cats were reported to have anomalous vessels originating from the junction of the splenic and portal veins.[11] Of the remaining cases in our study, the most common findings implicating the liver were the presence of hepatic masses. These masses may have caused the development of PH either by direct impingement on portal circulation by causing secondary inflammation and alteration of the hepatic architecture.

Alternatively, it is possible that the inciting cause of these shunts may have resolved by the time that an ultrasound was performed. It is documented that acquired PSS do not rapidly resolve even after the portal blood flow has returned to normal. In a rat model of portal hypertension, acquired shunts were present in 33% of the animals 6 weeks after release of the portal ligature.[24] A similar phenomenon exists with human postoperative liver transplant patients. Although patients with hepatic cirrhosis-induced portosystemic shunts have quick resolution of PH after receiving a liver transplant, the presence of collateral circulation has been shown to persist for up to 4 years.[25, 26] Shunts involving the splenic vein were found to be the most persistent. In addition, some patients continued to suffer from hepatic encephalopathy in spite of successful liver transplantation, suggesting that the proportion of shunted blood remained clinically significant. We failed to document resolution of the shunts with additional abdominal ultrasound examinations performed up to 600 days after their original identification. In other species, benign hepatopathies such as lipidosis can cause alterations in portal blood flow and PH.[27-29] To our knowledge, the effect of hepatic lipidosis on portal blood flow in cats has not been thoroughly evaluated.[30-32] It is conceivable that the splenosystemic shunts in some of the cases in this study could be caused by historical PH.

The majority of our cases were ovariohysterectomized females. This contrasts with what is reported with congenital PSS. Although experimentally induced PH caused development of similar shunts in both male and female cats, the majority of clinical case reports involve spayed females.[4, 9-12] This raises the possibility that previous surgery in the region of the noted vessel(s) might contribute to their formation. Adhesions to the left ovarian pedicle in these patients might predispose to vascular communications between the portal (splenic vein) system and the ovarian vein. Ovarian pedicle adhesions are often noted as an incidental finding when performing surgery in previously ovariohysterectomized cats, and vessels within some of these adhesions might be noted (K. Matthews, personal communication). Adhesions involving the ovarian pedicles have been found in association with colonic entrapment after ovariohysterectomy in both dogs and cats.[33-35] It is possible that anatomic differences between dogs and cats may make cats particularly susceptible to adhesion and subsequent shunt formation in this area. Whether or not the presence of adhesions near the ovarian pedicle in cats increases the likelihood of acquired shunt formation is unknown and warrants investigation.

The possibility exists that these shunts represent congenital PSS detected in older animals. There are reports of middle to older aged dogs being diagnosed with congenital PSS.[36-38] As typically noted in animals with congenital PSS, these older dogs mostly presented with neurologic signs. In contrast to our study group, portoazygous shunts appeared to be overrepresented in these reports. Proposed theories to explain the delay in detection of shunts in these dogs include increased sensitivity of the aging brain to ammonia or an age-related decreased in liver function. Similarly, it is possible that the cats in our study might have been compensating before developing a systemic disease that either altered blood flow between the portal and systemic circulation or caused liver function to deteriorate. Alternatively, our findings might represent congenital PSS of little or no clinical significance and their detection was purely incidental while investigating the primary cause of the cat's clinical signs. We believe that reports of their detection will likely increase as the sensitivity of diagnostic ultrasound continues to improve.

In light of the fact that similar shunts have been reported in experimental as well as clinical cases of PH, it is tempting to conclude that the splenosystemic shunts reported here are acquired in origin. Unfortunately, the descriptive nature of our report precludes us from making any conclusions as to their etiology or whether or not these shunts require specific medical or surgical management. Studies specifically evaluating liver function and portal blood flow in cats with splenosystemic shunts and studies determining the prevalence of splenosystemic shunts in young female cats are warranted.

Acknowledgment

We thank Dr Kyle Mathews for his insight and help in manuscript preparation.

Conflict of Interest: Authors disclose no conflict of interest.

Ancillary