Portosystemic Shunts: Diagnosis, Prognosis, and Treatment of 64 Cases (1993–2001)

Introduction

A portosystemic shunt is any vascular anomaly that allows blood from the portal circulation to bypass the liver and be delivered directly into the systemic circulation. Portosystemic shunts are suspected when animals present with signs of hepatic encephalopathy (e.g., head pressing, ataxia, wandering, seizures), depression, vomiting, and long anesthetic recovery times, or if urate calculi are present in nonpredisposed breeds.

Bile acids and blood ammonia tests are frequently performed as noninvasive screening tests. Paired bile acids are performed by drawing a fasted blood sample, providing a small meal, and then drawing another blood sample 2 hours later.¹ The advantage of systemic bile acid quantification is the stability of bile acids in the blood sample. The blood may be sent to another laboratory with little concern for alteration of bile acids in transit. The disadvantage of this test is that dual samples must be drawn 2 hours apart; therefore, a short appointment for clients is not possible. Resting ammonia levels require only one blood sample, but due to instability of ammonia in vitro, the test must be run immediately, or the blood must be separated and the plasma must be frozen at −20°C to avoid false results.¹ Some clinicians prefer one test over the other, and some perform both tests. One factor that should be taken into consideration is the sensitivity of each test. There is evidence that the sensitivity of paired bile acids is greater than for resting ammonia levels.¹ However, one paper suggests equal sensitivity.² Specificity is of less concern, since both tests are specific for liver disease but do not differentiate portosystemic shunts from other hepatic pathologies.

If screening tests support a diagnosis of a portosystemic shunt, further diagnostics such as abdominal ultrasonography, contrast portography, rectal scintigraphy, or exploratory surgery are warranted to confirm this diagnosis. Abdominal ultrasonography is frequently used because it is noninvasive; however, the limitations of this procedure for shunt detection must be understood by the clinician to avoid a misdiagnosis. Studies have reported a range of 80.5% to 95% for sensitivity of this imaging modality.³⁴

Treatment for portosystemic shunts include medical therapy, surgical attenuation of shunt vessels, or both. Medical therapy is aimed at decreasing clinical signs such as hepatic encephalopathy and urinary stones, but it does not resolve the underlying cause and does nothing to prevent further deterioration of liver function. Due to these limitations, most veterinarians recommend surgical attenuation of the shunt vessel if possible.

Complete ligation of a shunt vessel is only possible in certain circumstances. In many animals, the liver cannot accommodate a large and sudden increase in portal blood flow, and if the shunt vessel is suddenly occluded, the result is portal hypertension. Acute, severe portal hypertension can be fatal, and chronic, moderate portal hypertension may result in multiple acquired shunts. Guidelines for safe occlusion have been published.⁵

Cases in which complete shunt ligation cannot be safely performed during one surgical procedure are treated by gradual shunt attenuation. One technique for gradual occlusion is partial ligation of the shunt vessel. The vessel is incompletely ligated so that portal pressures increase but remain within a safe limit. Some partially ligated vessels gradually close over time without any further surgical intervention. In these cases, local inflammation eventually leads to complete thrombosis. A second surgery to attempt complete ligation may be performed if persistent shunting remains.

Other methods for gradual occlusion of portosystemic shunts include ameroid ring placement, cellophane band placement, and thrombogenic intravascular coils.^6–9 The advantage of these techniques is gradual occlusion of the shunt vessel without the need for further surgery. Ameroid ring placement is the only alternative gradual occlusion technique currently used at Auburn University. The ameroid ring is a compressed, dehydrated, casein ring surrounded by a metal band. It contains a central channel so that it can be placed around a vessel. When it becomes hydrated by normal body fluids, it expands inward, causing gradual occlusion of the shunt vessel.

There have been various studies evaluating the success of partial ligation and complete ligation of portosystemic shunts.^10–13 Studies have shown that complete ligation leads to a better long-term prognosis than partial ligation. At least 29% to 41% of animals with partial ligation have recurrent signs.¹⁰¹¹ Less information is available concerning the success of ameroid ring placement. Even though ameroid constrictors are designed to gradually occlude shunt vessels, they still have the potential to create moderate or severe portal hypertension, which may lead to death or the development of multiple acquired shunts. The authors are aware of only one study comparing outcome after ameroid constrictor placement to that of partial ligation.⁹ That study indicated that ameroid constrictor placement led to less postoperative complications than partial ligation. However, the result was not statistically significant.

Many different factors affecting outcome after surgery have been evaluated. Presence of hepatic encephalopathy prior to surgery has been shown to negatively affect outcome.¹⁴ In addition, intrahepatic shunts are more difficult to ligate than extrahepatic shunts; therefore, intrahepatic shunts may have a more guarded prognosis.¹⁵ Pre- and postligation portal pressures within the published safe limits do not appear to influence outcome.¹¹ There is controversy concerning the effect of age at presentation on outcome.¹¹¹²

Animals should be monitored for continued shunting after surgery. Continued shunting may occur because of incomplete vessel occlusion or after the development of multiple acquired shunts. The ideal method of postoperative monitoring for continued shunting is unknown. Portal scintigraphy¹⁶¹⁷ and contrast portograms¹⁸ may be used to monitor for continued shunting. Contrast portograms are invasive and require general anesthesia. Portal scintigraphy requires heavy sedation and increased hospitalization because of the excretion of radioactive waste. Therefore, both of these methods are less convenient than bile acid measurements. Studies show that decreasing bile acid levels are associated with decreased shunting, but studies do not agree on whether this test is highly correlated with clinical outcome.¹⁰¹¹

The objectives for this study were to compare the sensitivity of blood ammonia levels to paired pre- and postprandial bile acid tests for detecting portosystemic shunts; to determine the sensitivity of ultrasound examination for the detection of portosystemic shunts; to compare the age at presentation between animals with congenital shunts and animals with acquired shunts; to report short- and long-term success rates obtained by various surgical treatments (i.e., partial ligation, complete ligation, and ameroid ring placement); to evaluate age at presentation as a prognostic indicator of surgical success; to evaluate pre- and postoperative bile acid values to determine if a difference existed between treatment groups and to determine the percentage that were within the reference range after surgery; and to report on the outcome of surgically treated intrahepatic shunts, medically managed congenital shunts, medically managed acquired shunts, and two cases of congenital, multiple, extrahepatic shunts.

Materials and Methods

Records of dogs and cats with a diagnosis of portosystemic shunting between the years 1993 and 2001 at the Auburn University Veterinary Teaching Hospital were reviewed. Cases without confirmation of portosystemic shunting by surgical exploration, portal scintigraphy, intravenous (IV) portography, abdominal ultrasonography, or necropsy were excluded.

Data recorded included species, breed, sex, age at presentation, clinical signs at presentation, type of shunt (i.e., intra- versus extrahepatic), preoperative bile acid values, preoperative ammonia levels, type of surgery performed (if surgery was performed), early postoperative mortality, and postoperative bile acid and ammonia blood levels. Early postoperative mortality included animals that died or were euthanized within 1 week of surgery. Some bile acid values recorded were from laboratories other than Auburn University Clinical Pathology Laboratory. To help standardize all bile acid and ammonia values between different laboratories, bile acids and ammonia were recorded as multiples of the high end of the reference range for the laboratory at which the sample was evaluated. All nonpaired bile acid samples were recorded as multiples of the high normal value for the postprandial bile acid sample.

For long-term follow-up, owners or referring veterinarians were contacted and questioned concerning animal survival and the presence of any abnormal clinical signs. For cases in which the owner could not be contacted and the local veterinarian had no follow-up information, the clinical signs recorded at the last visit at Auburn University were used for long-term follow-up information. The outcomes of two cases that had two extrahepatic shunts are described. These two cases were not included in the statistical comparison of long-term outcome between treatment groups. There were not enough cases of multiple congenital shunts to include them as a separate group in the statistical analysis, and the authors felt that long-term prognosis of these shunts may differ from single extrahepatic shunts. Long-term complications, which included recurrence of clinical signs or the postoperative development of multiple acquired shunts, were recorded.

The sensitivity for detection of portosystemic shunting of paired bile acids, blood ammonia level, and abdominal ultrasonography were calculated. All paired bile acids and blood ammonia levels performed when animals had confirmed portosystemic shunting either before or after surgery were included in the calculation of the sensitivities of these two tests. Bile acid values were considered abnormal when either the preoperative or postoperative value was above the high end of the reference range for the laboratory at which the sample was evaluated. Values were considered abnormal for blood ammonia levels if the value was above the high end of the reference range for the laboratory where the sample was evaluated.

Radiology reports concerning abdominal ultrasonography were evaluated. The percentage was calculated for abdominal ultrasonographic examinations that were positive for a portosystemic shunt when a portosystemic shunt was confirmed. Abdominal ultrasonograms were considered positive when the report listed identification or suspicion of a portosystemic shunt. Specificity of abdominal ultrasonographic examinations for detection of portosystemic shunts was beyond the scope of this study, since no cases without portosystemic shunts were evaluated. Only abdominal ultrasonographic examinations performed by a board-certified radiologist were included.

Frequencies of the following characteristics were recorded for the cases: extrahepatic shunts that could be completely ligated during one surgical procedure; intrahepatic shunts that could be completely ligated during one surgical procedure; and partially ligated extrahepatic or intrahepatic shunts that could be completely ligated during a second procedure.

Surgical treatments of congenital shunts were divided into three groups: complete ligation, partial ligation, or ameroid ring placement. Early postoperative mortality and the development of long-term complications for each surgical treatment group were determined. If an animal had partial ligation during one procedure, followed by complete ligation during another procedure, the partial ligation surgery was counted in the partial ligation group, and the complete ligation surgery was counted in the complete ligation group for the determination of early postoperative mortality. Animals that had partial ligation followed by complete ligation were considered to be part of the complete ligation group for comparison of long-term outcome. Since intrahepatic shunts are technically more challenging, postoperative mortality and long-term complications for animals with intrahepatic shunts were calculated separately from those with extrahepatic shunts.

Short- and long-term outcomes were further compared between two age groups: those ≤24 months and those >24 months at presentation for surgery. Early postoperative mortality for the original surgery and the percentage that developed long-term complications were calculated for each group. The median ages for animals with multiple acquired shunts and animals with congenital shunts were calculated. All animals with many (i.e., more than two small extrahepatic shunts) were considered to have multiple acquired shunts.

Median pre- and postprandial bile acids were calculated preoperatively for each group (i.e., partial ligation, complete ligation, and ameroid ring placement groups, as well as animals with multiple acquired shunts). Median bile acids were also calculated postoperatively for the partial ligation, complete ligation, and ameroid groups. Postoperative bile acids used for statistical comparison between treatment groups were collected from 2 days to 1 year after surgery. Only one case had bile acids performed 2 days postoperatively. This case was treated by complete ligation. All other bile acid samples evaluated were taken at least 6 weeks postoperatively.

Results

Sixty-four cases of portosystemic shunts met the inclusion criteria. Sixty-three cases were canine, and one was feline. Yorkshire terrier was the most common breed identified in this study, accounting for 14 cases. Other breeds included miniature schnauzer (n=9), shih tzu (n=7), pug (n=6), Maltese (n=3), golden retriever (n=3), Pomeranian (n=2), Labrador retriever (n=2), Shetland sheepdog (n=2), and one each of the following: miniature dachshund, bichon frise, German shepherd dog, rottweiler, basset hound, Silky terrier, Australian cattle dog, Dalmatian, Jack Russell terrier, Skye terrier, English setter, Chihuahua, Siberian husky, toy poodle, and small mixed-breed dog. The one cat included in the study was a domestic shorthair. Twenty animals were male, 27 animals were female, four were castrated male, and 13 were spayed female. Age at presentation ranged from 2 to 84 months; 41 (64%) were ≤1 year of age; 11 (17%) were from 13 to 36 months of age; and 12 (19%) were >36 months of age. The median ages at presentation for animals with multiple acquired portosystemic shunts and congenital shunts were 9 months (range, 7 to 48 months) and 8.5 months (range, 2 to 84 months), respectively; this was not statistically different (P=0.612).

Clinical signs were recorded for 61 patients. Clinical signs included lethargy (25/61; 41%), hepatic encephalopathy without seizures (e.g., incoordination, head pressing, circling, disorientation, pacing) (25/61; 41%), seizures (10/61; 16%), weight loss (7/61; 11%), vomiting (3/61; 5%), and blindness (1/61; 1.5%). Urinary calculi were also diagnosed in 19 (30%) cases.

All results of bile acids and ammonia levels are listed in the Table. Paired pre- and postprandial bile acids were used as a screening test in 57 instances in which portosystemic shunting was confirmed. Fifty of these were preoperative values or were taken from animals that had no surgical treatment. Seven were postoperative values in animals with continued shunting after surgery (due to either the development of multiple acquired shunts or the inability to completely close the shunt vessel). Blood ammonia levels were used as a screening test in 37 instances in which portosystemic shunting was confirmed. Thirty-one of these were preoperative and six were postoperative values in animals with continued shunting after surgery. When comparing paired bile acid measurements to blood ammonia levels for dogs with confirmed portosystemic shunts, all 57 bile acid measurements were abnormal, whereas 32 out of 37 blood ammonia levels were abnormal. This difference in sensitivity was statistically significant (P=0.008). It is important to note that there were three instances (case nos. 8, 28, and 49) in which either the pre- or postprandial bile acid sample was normal when a shunt was confirmed. Case no. 28 had a preprandial bile acid sample that was only two points above the reference range and a postprandial sample within the reference range. In addition, there were seven instances in which only one bile acid sample was measured in cases of confirmed portosystemic shunting. It was not specified in the records if these were preprandial or postprandial samples. All were abnormal.

Abdominal ultrasonography was used as a diagnostic imaging test in 49 cases in which portosystemic shunting was confirmed with surgery, nuclear scintigraphy, contrast portography, or necropsy. In 40/49 (81%) cases, a shunt was diagnosed or highly suspected. There were nine cases in which no remark concerning portosystemic shunting was made in the ultrasonography report.

There were 48 (75%) animals with single extrahepatic shunts, two (3%) with two extrahepatic shunts, five (8%) with intrahepatic shunts, eight (12.5%) with multiple acquired shunts, and one with a shunt that was identified only with abdominal ultrasonography but was not definitively classified. Fifty-one cases were taken to surgery for either partial ligation (n=21), ligation (n=10), partial ligation followed by complete ligation at a later date (n=3), or ameroid ring placement (n=17). Three cases of single extrahepatic shunts and one unclassified shunt were not attenuated due to owner wishes. Two of these were euthanized at the time of diagnosis. One case diagnosed by portography was not identified during surgery, and therefore no attenuation was performed. Five cases diagnosed with multiple acquired shunts were euthanized at the time of diagnosis.

There were a total of 38 cases, including extra- and intrahepatic shunts, in which complete ligation was attempted during the first surgical procedure. However, only 10 (26.3%) of these could be completely ligated during one surgical procedure. When evaluating extrahepatic shunts alone, the success rate of complete ligation during one procedure was 30.3%. There were eight cases (i.e., two intrahepatic and six extrahepatic) in which complete ligation was attempted after partial ligation. Only three (37.5%) of these could be completely ligated during the second surgical procedure. When evaluating extrahepatic portosystemic shunts alone, three of six (50%) could be completely ligated during a second surgical procedure. Case nos. 10, 12, and 17 were completely ligated 4, 6, and 12 months, respectively, after the first surgery. Case nos. 22, 39, and 64 were taken to surgery 4 years, 1.4 years, and 3 months after the first surgery for further ligation, which was not possible in each case.

The overall postoperative early mortality rate for all animals was 5/51 (9.8%). The postoperative mortality rate for animals with extrahepatic shunts alone was 4/46 (8.7%). Three of 19 (15.8%) animals with partial ligations for single extrahepatic shunts died during the first 5 days after surgery. Case no. 63 died from status epilepticus, case no. 14 died from postoperative coma, and case no. 62 died at home within 5 days of discharge for unknown reasons. One of 15 (6.7%) animals with ameroid placement for single extrahepatic shunts died 5 days postoperatively due to status epilepticus (case no. 56). The difference in postoperative early mortality rates between partial ligation and ameroid ring placement groups was not statistically significant (P=0.6). None of the animals with complete ligation died in the early postoperative period.

Long-term follow-up of at least 2 months was available for 15 animals with partial ligation (12 extrahepatic shunts and three intrahepatic shunts), 11 animals with complete ligation, 13 animals with ameroid placement for one shunt, one animal with multiple acquired shunts, and three animals with single shunts without surgical management. Two animals with partial ligation developed multiple acquired shunts, and three developed clinical signs that could be attributed to further shunting. Case no. 9 was 5 months of age at the time of the initial surgery, and multiple shunts were noted 6 months later when the animal presented for depression. Case no. 28 was 6 months old at the time of presentation, and multiple acquired shunts were noted 3 years after surgery when the animal presented for urethral calculi. Case no. 39 presented 1 year postoperatively for depression and ascites and died 1.5 years after surgery due to progression of disease. Case no. 22 presented for dark, tarry stools 4 years after partial ligation. Complete ligation was not possible at that time. Case no. 16 had episodes of intermittent depression. Continued shunting was confirmed by ultrasonography 2 months postoperatively.

Two animals with ameroid ring placement developed long-term complications. Case no. 42 was 5 months old at original presentation, and multiple acquired shunts were noted 15 months after surgery when the animal presented for elective ovariohysterectomy. Case no. 38 was 4 years old at presentation, and abdominal ultrasonography revealed ascites and evidence of shunting 3 months after surgery when the animal represented for seizures. That animal died 22 months after surgery.

Case no. 17, which had complete ligation, developed signs of restlessness, pacing, and collapse 6 years after surgery. Diagnostics revealed marked elevation of bile acids; however, nuclear scintigraphy, portography, or surgery was not performed to further determine the cause of these signs. The long-term complication rates of extrahepatic portosystemic shunts were 42%, 15.4%, and 9% for animals with partial ligation, ameroid ring placement, and complete ligation, respectively. The difference between animals treated with partial ligation versus ameroid ring placement was not statistically significant (P=0.2).

The effect of medical management along with surgery was not evaluated during this study because of the low numbers of animals still on medications after surgery. Case no. 29, with partial ligation, was placed on metronidazole 4 years after surgery because of a short period of disorientation noted by the owner. No diagnostics were performed, and the owner reported no clinical signs after the one episode. Case no. 41, with complete ligation, was on metronidazole 3 years after surgery but had no clinical signs. The reason the animal was still on medication was undetermined. Case no. 16, with partial ligation, was on lactulose 2 years after surgery and had intermittent depression. This animal had confirmed continued shunting by portography and abdominal ultrasonography. Case no. 50, with partial ligation, was depressed 7 months after surgery. Continued shunting was diagnosed on ultrasonographic examination, and the animal was placed on lactulose. This animal had no clinical signs 11 months after surgery. A surgical attempt at complete ligation was unsuccessful.

Case no. 43 was the only animal with multiple acquired shunts with long-term follow-up. This animal was still alive 2.5 years after surgery. Of the three animals with congenital shunts without surgical management, case no. 45 was still alive 6 months after diagnosis, case no. 24 was euthanized 3 years after diagnosis for progressive seizures, and case no. 61 was alive 5.5 years after diagnosis.

Age was evaluated as a risk factor for mortality or long-term complications after surgery. The early postoperative mortality rates for animals ≤24 months and >24 months of age that were undergoing surgery for congenital extrahepatic shunts were 8.6% (3/35) and 9.1% (1/11), respectively. This was not significantly different (P=1.000). Age was evaluated as a risk factor for long-term complications. Of the animals followed long-term (i.e., ≥2 months), 21.4% (6/28) of those aged ≤24 months and 22.2% (2/9) of those aged >24 months that were treated surgically for single extrahepatic shunts had long-term complications.

There were a total of 47 cases with preoperative paired bile acid values available for evaluation. Nineteen of these cases were in the partial ligation group, nine were in the complete ligation group, 12 were in the ameroid ring group, and seven were in the acquired shunt group. A total of 16 cases had postoperative paired bile acid samples that were available for analysis; seven of these were in the partial ligation group, five were in the ligation group, and four were in the ameroid ring group. There were no significant differences in preprandial bile acid values (P=0.096) or postpran-dial bile acid values (P=0.188) when comparing the following groups: preoperative partial ligation, preoperative complete ligation, preoperative ameroid ring placement, multiple acquired shunts, postoperative partial ligation, postoperative complete ligation, and postoperative ameroid ring placement. There was a tendency for lower postoperative pre- and postprandial bile acid values as compared to preoperative values. There was also a tendency for lower postoperative bile acid values for the complete ligation group compared to the ameroid ring or partial ligation groups. No animals had normal postoperative bile acid measurements.

Two animals that had two extrahepatic portosystemic shunts were treated with ameroid ring placement. Both cases of dual extrahepatic shunts were diagnosed by surgical exploration. For case no. 55, one shunt was portocaval and the other initiated at a gastric vein and coursed through the diaphragm, presumably to the azygous vein. In this case, pressures were measured and increased to unacceptable levels when either of the shunts was attenuated. The decision was made to place an ameroid ring on one vessel (the specific vessel was not recorded) and recheck the animal later. The owners did not return in 2 months for the scheduled visit. When contacted by telephone 5 months after surgery, the owners reported no clinical signs and did not wish to pursue ligation of the second shunt. For case no. 51, two shunts originated from the gastrosplenic vein. One communicated with the caudal vena cava, and the other coursed through the diaphragm presumably to the azygous vein. An ameroid ring was placed on each of the extrahepatic shunts in the other case. This animal died within 2 months of surgery. Necropsy was not performed, and the cause of death was unknown.

Evaluation of the five animals with intrahepatic shunts included a 4-month-old, female Yorkshire terrier (case no. 1); a 1.5-year-old, male Australian cattle dog (case no. 11); a 5-month-old, male Jack Russell terrier (case no. 29); a 1-year-old, male Skye terrier (case no. 31); and a 6-month-old, female golden retriever (case no. 50). Partial ligation was the form of treatment in all of these animals. Case no. 1 died postoperatively because of portal hypertension. Two of the animals were rechecked for possible complete ligation. Surgery was attempted on case no. 50, but excessively high portal pressure resulted when the shunt was occluded. Ultrasound revealed an enlarged portal vein consistent with increased portal pressure in case no. 31; therefore, no further surgery was attempted. Three animals had long-term follow-up. Case no. 29 had one episode of disorientation and was placed on metronidazole. Ultrasound revealed continued shunting 6 weeks postoperatively. No further clinical signs were reported 4 years postoperatively. Case no. 31 had no clinical signs at 4 months. Case no. 50 appeared lethargic 7 months postoperatively but was showing no clinical signs 11 months postoperatively while on lactulose.

Discussion

Many different breeds were represented in this study, with small-breed dogs comprising the majority of cases. This is consistent with previous literature that reports a predisposition of miniature and toy breeds for portosystemic shunts.⁵ Ninety-four percent of small dogs in this study had extrahepatic shunts, which is consistent with previous reports that intrahepatic portosystemic shunts in small dogs are uncommon.⁵ Seventy-one percent of large-breed dogs in this study had extrahepatic shunts, which is greater than the 35% previously reported.⁵ One possible explanation is that the small number of large-breed dogs included in this study may not provide a true representation of the large dog population. Another possibility is that the percentage of large-breed dogs with extrahepatic portosystemic shunts is larger than what has been historically reported.

Age at presentation of animals with congenital shunts was similar to previous studies, with 60% presenting in the first year of life and 75% presenting by 2 years of age.⁵ However, it was surprising that in this study, the median age for animals diagnosed with multiple acquired shunts was <1 year of age. The significance of this finding is unknown because of the small number of cases of multiple acquired shunts. Furthermore, this study was based on animals with confirmed portosystemic shunts. It is possible that many older animals presenting with signs consistent with liver failure may have multiple acquired shunts that are never confirmed. The findings of this study suggest that age alone cannot be used to differentiate multiple acquired shunts from congenital shunts.

Common presenting signs in this study were hepatic encephalopathy and lethargy, with vomiting, seizures, and weight loss being less common. The pathophysiology of hepatic encephalopathy is not completely understood; however, all theories agree that this condition is most likely caused by one or a combination of toxins released from the gastrointestinal tract into the portal system, which would be detoxified in the liver of normal animals. The reason for signs such as head pressing, circling, or dementia being more common than seizures is unknown, but a possible explanation may be the benzodiazepine-like compounds released from the gut, which are inhibitory to the central nervous system.¹⁹

Thirty percent of cases in this study also had concurrent urinary calculi. Although the type of stone was not always reported in the record, most of these stones were probably urate in composition. Animals with portosystemic shunts are predisposed to urate calculi formation because of decreased ammonia and urate metabolism by the liver, resulting in a high concentration of ammonia urates secreted in the urine.⁵ Portosystemic shunting should be considered in animals that present for ammonium biurate stones, and the urinary tract should be thoroughly evaluated in every animal with a portosystemic shunt.

In this study, paired bile acid testing was a more sensitive detector of portosystemic shunting than blood ammonia levels. A pre- or postprandial bile acid sample was abnormal in all cases in this study with portosystemic shunts, whereas 13% of cases with portosystemic shunts had normal blood ammonia levels. This is consistent with another study that revealed normal blood ammonia levels in five of 44 dogs with portosystemic shunts,² but it is in contrast to another study that revealed 100% sensitivity of ammonia for the detection of portosystemic shunts.¹⁴ Only eight cases in the study revealed 100% sensitivity of blood ammonia levels. In addition, lactulose and oral antibiotics, which may have been given prior to the animals’ visit at a referral hospital, can decrease blood ammonia levels.¹ These results suggest that a portosystemic shunt should not be ruled out on the basis of a normal blood ammonia level. Alternatively, it would be very unlikely for an animal with both normal pre-and postprandial bile acid samples to have portosystemic shunting. Since three animals with portosystemic shunting in this study had either a normal pre- or postprandial bile acid sample, one normal value should not be used as a criterion to exclude a portosystemic shunt. There is no apparent advantage in performing both blood ammonia levels and paired bile acid samples when screening for a portosystemic shunt.

Abdominal ultrasonography is a valuable tool in the diagnosis of portosystemic shunting. It is noninvasive, requires little to no sedation, and can also identify urinary calculi. Furthermore, animals require no quarantine as with rectal scintigraphy. In the authors’ study, the sensitivity of abdominal ultrasonography for the detection of portosystemic shunting was 81%, which is closer to the lower end of the range reported in the literature (80.5% to 95%).³⁴ The specificity of ultrasonographic examination was beyond the scope of this study, but it has been reported in the literature to range from 66.7% to 98%.³⁴ Factors that may affect the ultrasonographic examination include movement of the patient, a full gastrointestinal tract, and operator experience. In the hands of an experienced ultrasonographer, this test is very useful in the diagnosis of shunts, but ultrasonography alone should not rule out the presence of a portosystemic shunt.

The authors found the percentage (30.3%) of cases in which the shunt could be completely ligated during the first surgical procedure to be close to the low end of the range reported in the literature (34% and 52%).¹⁰¹¹ Similar to a previous study in which 66.7% (8/12) of dogs with single extrahepatic shunts underwent complete ligation² during a second surgical procedure, 50% (3/6) of cases in this study could be completely ligated during a second surgical procedure. Based on results of the previous study in which clinical signs recurred in almost half of the animals with partial ligation only,¹¹ it would seem beneficial to attempt further ligation if portosystemic shunting persists. However, owners should be informed that the second surgical procedure may be unsuccessful in completely ligating the shunt.

Postoperative mortality for extrahepatic shunts has been reported to be 2%,¹⁰ 14%,¹¹ and 21%.¹³ The overall postoperative mortality rate (8.7%) for animals with congenital extrahepatic shunts in this study was within that range. Common postoperative complications that may lead to mortality include portal hypertension and status epilepticus. Portal hypertension may be secondary to excessive occlusion of the shunt vessel or thrombosis of partially ligated shunt vessels. Portal hypertension after ameroid ring placement may occur due to an undersized ring⁸ or theoretically it may be the result of acute vessel kinking once the animal changes position. Another theoretical cause of portal hypertension is unexpected, rapid closure of the shunt vessel by an ameroid ring. The current recommendation is to use an ameroid ring that constricts the vessel ≤25% when initially placed.⁸ In this study, the postoperative mortality rate was low for all surgical treatment groups, which supports findings of another study comparing ameroid ring placement to partial ligation.⁹ At this time, there is no definitive evidence for superiority of one technique of gradual occlusion over another in preventing short-term complications.

Only 9% (1/11) of cases with complete ligation had recurrent clinical signs that were noted 6 years after surgery. This one case had incomplete ligation followed by complete ligation at a second surgery. None of the 10 animals with complete ligation at the first surgery had recurrent signs. This supports other papers that report a good prognosis for animals with complete ligation.¹⁰¹¹ The long-term complication rates for partial ligation and ameroid ring placement in the authors’ study were 42% and 15.4%, respectively, which are consistent with previous studies. One previous clinical study evaluating ameroid constrictors in 14 animals revealed a 17% long-term complication rate due to the development of multiple acquired shunts.⁸ A long-term complication rate of 41% has been reported after partial ligation.¹¹ After partial ligation, some shunt vessels may achieve complete occlusion due to vessel thrombosis. However, recurrence of clinical signs may be a result of the development of multiple acquired shunts or continued shunting from the shunt vessel. Ameroid constrictors are designed to slowly result in complete occlusion of the shunt vessel in a range of 2 weeks to 90 days.⁶⁸ Therefore, continued shunting would seem unlikely after the placement of an ameroid ring, unless multiple acquired shunts develop. Larger numbers may have revealed a statistical difference in long-term complications between animals treated with ameroid constrictors and those treated with partial ligation alone.

The long-term complication rate in the current study may have underestimated the true long-term complication rate, because follow-up of at least 2 years postsurgery was available for only eight of 15 animals with partial ligation and five of 13 with ameroid ring placement. The previous report revealed that animals with recurrent signs present an average of 23 months after surgery.¹¹ The reason certain animals develop multiple acquired shunts after surgery is unknown. It may be that the portal system or hepatic vasculature in some animals will never accommodate normal portal blood flow, or there may be an unacceptably rapid rate of vascular occlusion.

The authors’ study found a survival time of 2 months to 5 years in dogs with portosystemic shunts managed with medical treatment alone. Previously reported long-term survival was from 2 months to 2 years.¹⁵ Due to the small number of cases in the authors’ study, no firm conclusion can be made concerning long-term survival of animals with portosystemic shunts that do not undergo surgical correction. However, this study does illustrate the possibility of long-term medical management of certain animals with portosystemic shunts that do not undergo surgical attenuation.

Older animals did not appear to be at increased risk for early postoperative mortality or the development of long-term complications. This is in agreement with a previous study evaluating 56 dogs with extrahepatic shunts treated surgically;¹¹ but it is in contrast to the study by Johnson, et al.,¹³ which revealed that animals that presented at >2 years of age had a less favorable prognosis than those that presented at ≤2 years of age. Due to the conflicting information, age at the time of presentation should not be considered when deciding if a patient is a good candidate for surgical management.

Presurgical bile acid samples (i.e., both pre- and postprandial) were not significantly different between groups (partial ligation, complete ligation, ameroid ring, or multiple acquired shunts). Therefore, bile acid samples cannot be used presurgically to help determine if a shunt can be completely ligated. Pre- and postprandial values were not significantly different between groups after surgery, even though the median pre- and postprandial bile acid values for the complete ligation group were <33% of those for the ameroid ring and partial ligation groups. In the study by Hottinger, et al.,¹¹ postsurgical bile acids were significantly less for animals with complete ligation than for those with partial ligation. The lack of a significant difference in the authors’ study may be a reflection of the small number of animals that had postsurgical bile acid samples performed (16 cases in the authors’ study versus 56 cases in the study by Hottinger, et al.). Larger numbers may have also revealed a significant difference between preoperative and postoperative bile acids values, since a trend for a decrease in bile acid values was present after surgery in this study.

In the authors’ study, no animal had normal paired bile acid samples after surgery. This is in agreement with the study by Lawrence, et al.,¹² in which none of the 20 cases had normal bile acids after surgery; but this is in contrast to a study by Hunt, et al.,¹⁰ which found that 78% of cases had normal paired bile acid samples after surgery. The failure of postsurgical values to return within reference ranges could be a reflection of continued shunting or irreversible liver pathology. Hepatic atrophy from decreased blood flow and hepatatrophic factors is consistently present in animals with portosystemic shunts, and surgical correction may not completely reverse previous liver pathology.²⁰

The authors report two cases of animals with two extrahepatic portosystemic shunts. There is a report of two cases of multiple intrahepatic portosystemic systemic shunts in the literature.²¹ In those dogs, both shunts originated from one contributary of the portal vein and were treated by attenuating the vessel at the origin of both shunts. Since only two cases of dual extrahepatic portosystemic shunts were noted in the authors’ study and the outcomes were dramatically different, no conclusions can be drawn as to the best treatment modality or the prognosis.

The authors’ study revealed a postoperative mortality rate of 20% for intrahepatic shunts, similar to that of another study (18%) in which 45 cases of intrahepatic shunts were evaluated.²² Complete ligation during one surgical procedure was not possible in any of the authors’ cases or in cases of the previous study. Further ligation during a second surgical procedure was not possible in the two cases in the authors’ study that were evaluated for further ligation. In the previous report of 45 intrahepatic shunts,²² complete ligation was possible in three of six animals that presented for a second surgery. In the authors’ study, three animals with long-term follow-up had no clinical signs at 4 months, 11 months, and 4 years postoperatively. However, two of these animals were placed on medical management because one or more episodes of clinical signs were noted postoperatively. In the previous report,²² nine of 37 animals had persistent or recurrent signs after surgery. Trends identified in the authors’ study are in agreement with previous reports that have stated the prognosis for surgically treated intrahepatic portosystemic shunts is less favorable than for extrahepatic shunts.⁵¹⁵

Conclusion

Several conclusions can be made concerning the diagnosis and prognosis of portosystemic shunts. This study supported previous studies that revealed paired bile acid samples are more sensitive as a screening test than a single blood ammonia level. Abdominal ultrasonographic examination is a sensitive tool for the diagnosis of portosystemic shunts, but it should not be used alone to rule out this disease. The prognosis for animals with surgically treated portosystemic shunts is fair. The overall early postoperative mortality rate for animals with congenital shunts was 9.8%. Although the mortality rate for animals treated with partial ligation was slightly greater than the rate for those treated with ameroid constrictors, this value did not reach statistical significance. Furthermore, there was a tendency for a lower long-term complication rate with ameroid ring placement than with partial ligation alone. The long-term prognosis was good for animals in which complete ligation could be performed. However, animals in which partial ligation alone is performed have a guarded prognosis with a 42% long-term complication rate in this study. This suggests that complete ligation should be attempted during a second surgical procedure if persistent shunting is present. Also, many animals do not have normal liver function after surgery. Paired bile acid samples did not return to the reference range in any animal in this study. Without established normal postoperative values, bile acids alone may not be adequate for postoperative monitoring. This study did not reveal age at presentation to be an important prognostic factor for outcome after surgical management of portosystemic shunts. Furthermore, age at presentation should not be used alone to predict the presence of multiple acquired shunts versus congenital shunts. No firm conclusions concerning medically managed shunts, intrahepatic shunts, and dual extrahepatic shunts can be drawn from this study.