Materials and Methods

Case selection—All cats that had surgical attenuation of a single intrahepatic or extrahepatic CPSS and underwent IMP before and after temporary occlusion of their CPSS at the Royal Veterinary College from August 2000 through December 2006 were eligible for inclusion.

Medical records review—Preoperative data recorded for each cat included age, breed, sex, body weight, duration of clinical signs prior to surgery, medical treatments received prior to surgery, duration of and clinical response to medications, and results of bile acids stimulation testing. Clinical signs were grouped into 5 categories that included neurologic, gastrointestinal, urinary, growth, and other. Cats were graded according to the existence or lack of each category of clinical signs. Clinical response to medical management was subjectively graded as poor when there was no clinical improvement, partial when there was improvement but clinical signs were still evident, or complete when there were no remaining clinical signs.

Anesthesia records and surgical reports were also reviewed to determine whether perioperative drugs were used, type of shunt used, type of CPSS attenuation (complete vs partial) performed, visual signs of portal hypertension during surgery (present vs absent), blood pressures in the portal vein before and after temporary complete shunt occlusion, and nature of any surgical complications.

Kennel sheets and discharge notes were reviewed for evidence of short-term postoperative complications, treatments administered for postoperative complications, and outcome. Postattenuation neurologic complications were considered to be any neurologic signs (including ataxia, blindness, muscle tremors, abnormal behavior, and seizures) that developed within the immediate postoperative period (0 to 72 hours, with 0 hours representing the time upon which surgery concluded) in a cat with a plasma ammonia concentration that was below or near the reference limit and blood glucose and serum electroyte concentrations that were within respective reference ranges.

Records of follow-up examinations performed after the initial CPSS surgery were reviewed. A second surgery, approximately 12 weeks after the first, was the recommended treatment of choice for all cats that initially underwent partial attenuation of a CPSS. The aim of the second surgery was to try to achieve additional (ideally complete) attenuation of the CPSS. The same data recorded for the initial surgery were recorded for the follow-up examination, except that clinical response to first surgery was recorded instead of clinical response to medical management. Clinical response to first surgery was subjectively graded as poor when there was no clinical improvement, partial when there was improvement but clinical signs were still evident, or complete when there were no remaining clinical signs.

CPSS attenuation—Partial or complete CPSS attenuation was performed via a ventral midline coeliotomy by use of 2-0 polypropylene suture with or without 2-0 silk suture, according to published techniques.^6,14,17 A mesenteric vein was catheterized for measurement of portal venous blood pressure and IMP before and after temporary complete occlusion of the CPSS with a Rummel tourniquet. The end point for attenuation of the shunt was decided on the basis of portal venous blood pressure, visual assessment of the color of the pancreas and intestines, central venous pressure, and arterial blood pressure, as described elsewhere.^6,17 When partial CPSS attenuation was performed, a second 2-0 polypropylene ligature was tied loosely around the shunt and left in situ to facilitate additional attentuation of the shunt at a later date.

Portovenography—Digital subtraction IMP was performed by use of a mobile C-arm^a fluoroscope, according to published techniques.^6,9 The mesenteric vein was catheterized with the largest bore catheter possible (20 or 22 gauge), and a bolus of 1 mL of iodinated contrast mediumVkg (1.0 mg/kg [0.45 mg/lb]) was injected as quickly as possible into a mesenteric vein for each portovenogram. Images were acquired at a rate of 1/s from the start of contrast medium injection until portal opacification faded. For each series, the image revealing maximal opacification of the intrahepatic portal vasculature was selected for analysis.

Analysis of portovenograms—Four series of portovenograms were evaluated: first surgery before shunt occlusion, first surgery after temporary complete occlusion with a Rummel tourniquet, second surgery before occlusion, and second surgery after temporary complete occlusion with a Rummel tourniquet. Prints of all portovenograms were shuffled, assigned an arbitrary number, and ranked according to the number of generations of intrahepatic portal vessels that were visible by 2 observers, who reached a consensus without reference to any clinical data. One observer was a board-certified radiologist (CRL), and the other was a faculty surgeon (VJL).

The same criteria were used for portovenogram grading as described by the authors in another report¹³ involving dogs with CPSSs (Appendix). Briefly, a grade of 1 was assigned when no intrahepatic portal vasculature was visible, a grade of 2 was assigned when faint opacification of vestigial portal vessels was evident, a grade of 3 was assigned when faint opacification of a few second- or third-generation portal vessels was evident, and a grade of 4 was assigned when there was good opacification of third- and fourth-generation portal vessels.

Statistical analysis—The Wilcoxon signed rank test was used to determine the significance of differences in portovenogram grade before and after temporary complete occlusion and between first and second surgeries. A 1-tailed Spearman rank correlation (ρ) test was used to examine correlations between portovenogram grade (before and after shunt occlusion) at first surgery and age, body weight, duration of clinical signs prior to surgery, duration of medical treatment prior to surgery, clinical response to medical treatment prior to surgery (subjective measurement), preprandial and postprandial serum bile acids concentrations, plasma ammonia concentrations, portal venous blood pressures before and after shunt occlusion, and response to first surgery (subjective measurement). Correlation coefficients with values < 0.4, 0.4 to 0.7, and > 0.7 were interpreted as representing weak, moderate, and strong correlations, respectively. Correlation coefficients were deemed to be significant at a value of P < 0.05.

The χ² test for trend was used to evaluate the relationships between portovenogram grade (before and after shunt occlusion) at first surgery and the following binary variables: clinical signs at first surgery and follow-up examination (presence vs absence), nature of the CPSS (intrahepatic vs extrahepatic), type of shunt (partial vs complete attenuation at first surgery), visual signs of portal hypertension at first surgery (presence vs absence), and postoperative complications including postattenuation neurologic complications (presence vs absence).

The Wilcoxon test was used to determine the significance of differences between preprandial and postprandial serum bile acids concentrations, plasma ammonia concentration, and portal venous blood pressures before and after shunt occlusion at first surgery and second surgery. For all comparisons, differences were deemed significant at a value of P < 0.05. All statistical analyses were carried out by use of a standard statistics software package.^c

Results

Animals—Twenty-five cats met the inclusion criteria for the study, including 21 cats with 1 extrahepatic CPSS and 4 cats with 1 intrahepatic CPSS. Median age of the cats was 10 months (range, 4 to 116 months), and median body weight was 2.5 kg (5.50 lb; range, 1.2 to 4.5 kg [2.64 to 9.90 lb]). There were 8 intact males, 7 castrated males, 3 intact females, and 7 spayed females. Breeds of cats included domestic shorthair (n = 9), British shorthair (5), Persian (4), Ragdoll (3), Siamese (2), Burmilla (1), and domestic longhair (1).

CPSS attenuation—Intraoperative mesenteric portovenography was performed before and after temporary complete CPSS occlusion in all 25 cats, followed by complete attenuation of the shunt in 9 (36%) cats and partial attenuation of the shunt in 16 (64%) cats. Three to 12 months later (median, 3.8 months), 12 of the 16 cats that received partial shunt attenuation at the first surgery underwent a second surgery. At the second surgery, results of IMP revealed the shunt was still partially patent in 10 of the 12 cats; 9 of these cats then underwent complete shunt attenuation, and 1 cat underwent an additional partial attenuation (this cat had an extrahepatic left gastric portocaval shunt, which was the most common type of CPSS in our sample of cats). In 2 of the 12 cats, results of IMP indicated that the shunt was completely closed, but multiple acquired portosystemic shunts were also evident. Four cats that underwent partial attenuation at the first surgery were not returned for a second surgery Two of the 4 cats were doing well clinically 3 months after surgery, and their owners declined additional surgery despite our recommendations. One of the 4 cats died in a road traffic accident 2 months after the first surgery, and another cat died of refractory postoperative seizures following the first surgery.

Analysis of portovenograms—Portovenogram grades for cats at the first and second surgeries were summarized (Table 1). Portovenogram grade increased significantly following temporary complete occlusion at the first (P < 0.001) and second (P = 0.046) surgeries. At the first surgery, the median portovenogram grade assigned before shunt occlusion (1; range, 1 to 4) was significantly (P < 0.001) different from that assigned after shunt occlusion (4; range, 2 to 4). At the second surgery, the median portovenogram grade assigned before shunt occlusion (2.5; range, 1 to 4) was also significantly (P = 0.046) different from that assigned after shunt occlusion (4; range, 2 to 4). Furthermore, there was a significant (P = 0.006) increase in portovenogram grade between portovenograms obtained before shunt occlusion at the first and second surgeries (Figure 1).

View Full Size

Representative portovenograms obtained from a cat with an extrahepatic CPSS before and after temporary complete occlusion at first (A and B) and second (C and D) surgeries. At first surgery at 6 months of age, portovenogram grade increased from 1 to 3 Three months later, at the second surgery, portovenogram grade increased from 2 to 4.

Citation: Journal of the American Veterinary Medical Association 234, 2; 10.2460/javma.234.2.221

• Download Figure
• Download figure as PowerPoint slide

Anatomic characteristics of CPSSs—Twenty-one cats had a single extrahepatic shunt, including 16 with left gastric portocaval shunts and 5 with other portocaval shunts. At the first surgery, 17 of these 21 cats had a portovenogram grade of 1 before shunt occlusion and 12 out of 21 had a portovenogram grade of 4 after shunt occlusion. Four cats had a single intrahepatic shunt, including 3 with a left divisional shunt (patent ductus venosus) and 1 with a central divisional shunt. At the first surgery, all 4 cats with an intrahepatic shunt had a portovenogram grade of 1 before shunt occlusion. Two of these cats had a portovenogram grade of 3 and the other 2 had a grade of 4 after shunt occlusion. No significant relationship was found between portovenogram grade at first surgery and whether the CPSS was intrahepatic or extrahepatic.

At the first surgery, the most common grade for portovenograms obtained before shunt occlusion was 1 and the most common grade for portovenograms obtained after shunt occlusion was 4. Of the 16 cats that underwent partial attenuation of the CPSS, 14 had a portovenogram grade of 1 before shunt occlusion and 7 had a portovenogram grade of 4 after shunt occlusion. Of the 9 cats that underwent complete attenuation of the CPSS, 7 had a portovenogram grade of 1 before shunt occlusion and 7 had a portovenogram grade of 4 after shunt occlusion. No significant relationship was detected between degree of shunt attenuation and portovenogram grade.

The most common portovenogram grade among the 12 cats that returned for a second surgery before and after shunt occlusion was 4. In 2 of the cats evaluated for a second surgery the original CPSS had closed and multiple acquired portosystemic shunts had developed. Portovenogram grades were 2 and 4 for these cats.

Cat characteristics and correlations among variables—At the first surgery a significant but weak, positive correlation was detected between increasing age and the portovenogram grade assigned after shunt occlusion (ρ = 0.359; P = 0.04; Figure 2). Graphic depiction of the distribution of age among the various portovenogram grades assigned after shunt occlusion suggested that this relationship may have been strongly influenced by the finding that 4 of 5 cats > 14 months of age had a grade of 4. No significant correlation was evident between portovenogram grade and body weight.

View Full Size

Graph indicating a weak (ρ = 0.359) but significant (P = 0.04) positive correlation between age and grade for the portovenogram obtained at surgery after shunt occlusion in 25 cats with a CPSS. A grade of 1 was assigned when no intrahepatic portal vasculature was visible, a grade of 2 was assigned when faint opacification of vestigial portal vessels was evident, a grade of 3 was assigned when faint opacification of a few second- or third-generation portal vessels was evident, and a grade of 4 was assigned when there was good opacification of third- and fourth-generation portal vessels.

Citation: Journal of the American Veterinary Medical Association 234, 2; 10.2460/javma.234.2.221

• Download Figure
• Download figure as PowerPoint slide

All 25 cats had received medical treatment appropriate for their clinical signs for a median of 3 weeks (range, 0 to 30 weeks). Treatment strategies included 1 or more of the following: antibiosis, a low-protein diet, lactulose, or antiseizure medication (eg, phenobarbitone, diazepam, or both). There was no clinical response to medical treatment in 3 (12%) cats, a partial clinical response in 14 (56%) cats, and complete resolution of clinical signs in 8 (32%) cats. Subjective grade of response to medical treatment was positively but mildly correlated (ρ = 0.380; P = 0.03) with grade assigned at the first surgery for the portovenogram obtained after shunt occlusion. A significant (P = 0.01), moderate, positive correlation (ρ = 0.472) was detected between increased duration of medical treatment prior to surgery and the grade assigned at the first surgery for the portovenogram obtained after shunt occlusion (Figure 3).

View Full Size

Graph indicating a moderate (ρ = 0.472) but significant (P= 0.01) positive correlation between duration of medical treatment received prior to surgery and grade for the portovenogram obtained at surgery after shunt occlusion in 25 cats with a CPSS. See Figure 2 for remainder of key.

Citation: Journal of the American Veterinary Medical Association 234, 2; 10.2460/javma.234.2.221

• Download Figure
• Download figure as PowerPoint slide

At the first surgery, clinical signs relating to a CPSS were reported for all 25 cats, with a median duration of signs of 3 months (range, 0.5 to 50 months). There was no correlation between duration of clinical signs prior to surgery and portovenogram grade. All 25 cats had neurologic signs that included evidence of depression, abnormal behavior (eg, hysteria and violent mood swings), head pressing, circling, ataxia, loss of balance and walking into objects, unresponsive staring into space, muscle tremors, seizures, blindness, and ptyalism. Eighteen (72%) cats were considered small for their age. Ten (40%) cats had gastrointestinal signs, including vomiting, diarrhea, and anorexia. Ten (40%) cats had other clinical signs, including prolonged recovery from anesthesia, pyrexia, polyuria, polydipsia, and nasal discharge. Three (12%) cats had urinary indications of recurrent cystitis. No significant relationships were evident between portovenogram grades (before and after shunt occlusion) at first surgery and the existence or lack of each of these groups of clinical signs.

Clinical signs were recorded for 12 cats after a median interval of 3.8 months after first surgery. Five cats still had neurologic signs. All 12 cats remained small for their age and still had gastrointestinal signs. One cat had urinary signs, and 1 cat had ptyalism. The χ² test for trend revealed that the probability of neurologic signs when the second surgery was performed decreased significantly (P = 0.006) with increasing grades for portovenograms obtained at first surgery after shunt occlusion. All 5 cats with neurologic signs still had elevated preprandial and postprandial serum bile acids concentrations, were receiving ongoing medical management for hepatic encephalopathy, and had neurologic signs similar to or milder than their original encephalopathic neurologic signs at the first surgery.

Subjective evaluation of clinical response to the first surgery revealed that response was partial in 5 cats and complete in 7 cats. A significant (P < 0.001), strong, positive correlation (ρ = 0.828) was detected between degree of clinical response to the first surgery and portovenogram grade obtained after shunt occlusion during the first surgery.

Serum bile acids and plasma ammonia concentrations—Preprandial and postprandial serum bile acids concentrations at the first surgery were available for 23 and 22 cats, respectively. Median preprandial serum bile acids concentration was 44 μmol/L (range, 3 to 263 μmol/L; reference range, 0.1 to 5 μmol/L). Median postprandial serum bile acids concentration was 136 μmol/L (range, 37 to 303 μmol/L).

Plasma ammonia concentrations were available for 19 cats at the first surgery. Median serum ammonia concentration was 289 μmol/L (range, 78 to 506 μmol/L; reference limit, < 70 μmol/L). There was no significant association between preprandial and postprandial serum bile acids or plasma ammonia concentrations at the first surgery and portovenogram grade.

Preprandial and postprandial serum bile acids concentrations at follow-up were available for 10 and 8 cats, respectively. Median preprandial serum bile acids concentration was 11 μmol/L (range, 1 to 83 μmol/L). Median postprandial serum bile acids concentration was 35.5 μmol/L (range, 3 to 143 μmol/L). Between the first and second surgeries, there was a significant decrease in preprandial (P = 0.01) and postprandial (P = 0.046) serum bile acids concentrations.

First surgery grade for the portovenogram obtained after shunt occlusion was significantly, negatively correlated with preprandial (ρ = −0.568; P = 0.04) and postprandial (ρ = −0.756; P = 0.02) serum bile acids concentrations at follow-up. At that time, median preprandial serum bile acids concentrations among cats with first surgery grades for portovenograms obtained after shunt occlusion of 2, 3, and 4 were 37, 4, and 2 μmol/L, respectively. In addition, median postprandial bile acids concentrations among cats with first surgery grades for portovenograms obtained after shunt occlusion of 2, 3, and 4 were 51, 66.5, and 4 μmol/L, respectively. The strengths of these relationships were compromised by the wide overlapping distributions of values for serum bile acids concentrations, particularly among cats with portovenogram grades 2 and 3.

Serum ammonia concentrations were available for 6 cats at the second surgery. Median plasma ammonia concentration was 143.5 μmol/L (range, 88 to 548 μmol/L). There was no significant association between serum ammonia concentration at follow-up and portovenogram grade (before or after shunt occlusion).

Portal venous blood pressures—Records of visual evidence of portal hypertension during complete temporary occlusion were available for 22 cats at the first surgery. Fourteen of 14 partial attenuations were associated with visual evidence of portal hypertension, compared to none of 8 complete attenuation procedures. Visual evidence of portal hypertension at the first surgery was not associated with portovenogram grade (before or after shunt occlusion).

Portal venous blood pressures before and after complete shunt occlusion were available for 20 and 16 cats, respectively. Median portal venous blood pressure before shunt occlusion was 8 mm Hg (range, 4 to 13 mm Hg). Median temporary complete portal venous blood pressure after shunt occlusion was 19 mm Hg (range, 9 to 70 mm Hg). There was no significant association between portovenogram grade (before or after occlusion) at the first surgery and portal venous blood pressure (before or after occlusion).

At the second surgery, visual evidence of portal hypertension was recorded solely for the 1 cat that could only tolerate an additional partial attenuation. The 9 cats that underwent complete attenuation and the 2 cats with complete shunt closure and multiple acquired shunts did not have visual evidence of portal hypertension. Portal venous blood pressure was measured in all 10 cats that underwent additional or complete attenuation. Median portal venous blood pressure was 8.5 mm Hg (range, 3 to 15 mm Hg) before shunt occlusion; that after shunt occlusion was 13 mm Hg (range, 3 to 21 mm Hg).

Intraoperative and postoperative complications—At the first surgery, there were no intraoperative complications. Between 12 and 72 hours after the first surgery, the only major complication was the development of postattenuation neurologic complications in 15 (60%) cats. Postattenuation neurologic complications included ataxia (n = 9), blindness (8), muscle tremors (8), abnormal behavior (6), and seizures (8). These complications varied in severity from mild ataxia to refractory seizures. All cats with postattenuation neurologic complications had plasma ammonia concentrations that were below or near the reference limit, as well as blood glucose and serum electrolyte concentrations that were within respective reference ranges.

All 15 cats that developed postattenuation neurologic complications received medical treatment, which initially consisted of phenobarbitone (between 1 mg/kg [0.45 mg/ lb], IY q 12 h, and 4 mg/kg [1.8 mg/lb], IV, q 4 h), depending on the severity of the signs. Treatment with orally administered phenobarbitone at a dosage of 1 to 2 mg/kg (0.45 to 0.91 mg/lb) every 12 hours was initiated if and when cats became stable. The dosage of phenobarbitone was increased incrementally to effect to treat so-called breakthrough seizures. Three cats with seizures refractory to phenobarbitone were also treated with an IV propofol infusion to effect (0.1 to 0.5 mg/kg/h [0.04 to 0.23 mg/lb/h] for 24 to 48 hours), which was then tapered and discontinued over 24 to 48 hours when the seizures resolved. Two additional cats with seizures refractory to phenobarbitone were euthanatized at the owner's request on days 2 and 5 following surgery In the surviving cats in which neurologic signs resolved by the time of discharge or at a subsequent reexamination, administration of phenobarbitone was gradually tapered and discontinued over 2 to 3 weeks.

A χ² test for trend revealed that the probability of postattenuation neurologic complications decreased significantly (P = 0.03) with increasing grade for the portovenogram obtained after shunt occlusion at first surgery According to results of portovenograms obtained after shunt occlusion at first surgery, 4 of 4 cats with portovenogram grade 2 had postattenuation neurologic complications, as did 5 of 7 cats with portovenogram grade 3 and 6 of 14 cats with portovenogram grade 4.

Minor (nonneurologic) complications following the first surgery included mild ascites (n = 3) and pyrexia (1) that resolved without any specific treatment. There was no significant association between portovenogram grade (before or after shunt occlusion) and the probability of nonneurologic postoperative complications.

At the second surgery, there were no intraoperative complications. Mild postattenuation neurologic complications in the form of ataxia and disorientation developed in 1 cat that had developed and was treated for postattenuation neurologic complications after the first surgery. In this cat, the signs resolved following phenobarbitone administration. According to results of portovenograms obtained after shunt occlusion at the second surgery, 0 of 12 cats with portovenogram grade 2 had postattenuation neurologic complications, as did 0 of 12 cats with portovenogram grade 3 and 1 of 12 cats with portovenogram grade 4.

Discussion

In the study reported here, IMP revealed a gradual improvement in intrahepatic portal venous branching in the months following CPSS attenuation. Furthermore, the grade for the portovenogram obtained at first surgery after shunt occlusion was significantly associated with age, duration of and response to medical treatment prior to surgery, probability of postattenuation neurologic complications and clinical response following the first surgery, existence of neurologic signs at follow-up, and decrease in preprandial and postprandial serum bile acids concentrations following the first surgery. Therefore, IMP may help identify cats that will respond well (clinically and biochemically) to attenuation of a single CPSS.

The portovenogram grading system used in the present study has been published.¹³ Researchers in that study (using dogs) and our study detected an immediate (following temporary complete occlusion) and sustained (at the second surgery) increase in opacification of the intrahepatic portal vasculature. The consistency of results between the 2 studies confirms the reliability of the portovenogram grading system and suggests that a similar maturation of hypoplastic portal vasculature occurs in dogs and cats following CPSS attenuation and exposure to increased portal venous blood flow. However, in contrast to results of 2 studies^12,13 in dogs, the degree of opacification of the intrahepatic portal vasculature in cats was not associated with partial or complete shunt attenuation at the first surgery. The reasons for this difference were unclear, but availability of a smaller number of cats in our study and variations in the degree of partial attenuation (ie, partial can represent any attenuation < 100%) could have affected this result. Certainly, there was no difference in the approximate percentage of cats that underwent complete versus partial attenuation in our study, compared with the percentages of dogs and cats that underwent CPSS attenuation in other studies.^1,4,6,17

Age of cats was positively correlated with the postocclusion portovenogram grade, suggesting that, as in dogs,¹³ older animals may have better developed intrahepatic portal vasculature at first surgery than younger animals. The strength of this relationship was weak, suggesting that this relationship may not be direct. Indeed, we would suggest that cats scheduled for surgery at an older age had tolerated portosystemic shunting for a longer period before evaluation for CPSS, compared with the period for younger cats, because of greater intrahepatic portal blood flow.

Results of the present study indicated that maturity of the intrahepatic portal vasculature was positively correlated with clinical response to medical treatment prior to surgery. Similar findings have not been reported, but this association appears logical. However, the strength of the correlation may have been weak because of clinician bias toward surgical versus medical treatment in the study cats, which eventually all underwent surgical intervention.

We also found that better developed intrahepatic portal vasculature was positively correlated with duration of medical treatment prior to surgery. It is unlikely that medical management itself influenced the development of the portal vasculature but, rather, that cats that responded better to medical treatment may have had a longer interval to surgical intervention.

The prevalence of postattenuation neurologic complications in cats in this study (60%) was higher than that in other reports,^6,18–22 including that of a retrospective study¹⁴ from our institution in which postattenuation neurologic complications affected 18 of 49 (37%) cats. This difference may reflect our increasing awareness of the wide spectrum of neurologic signs that can be detected following surgery, combined with our increased tendency to actively look for and treat even the mildest signs of postattenuation neurologic complications. Despite this approach, 8 out of 15 cats with postattenuation neurologic complications in the present study had seizures.

In our previous retrospective study¹⁴ of cats that underwent CPSS attenuation, the development of postattenuation neurologic complications was not associated with the existence of preoperative seizures, type of shunt, degree of shunt attenuation, or age of cat. To our knowledge, no risk factors for the development of postattenuation neurologic complications in cats that underwent CPSS attenuation have been reported elsewhere. We report for the first time that a lower IMP grade is associated with an increased risk of developing postattenuation neurologic complications. The etiology of postattenuation neurologic complications is unknown; therefore, the reason for a correlation between grade for the portovenogram obtained at the first surgery after shunt occlusion and development of postattenuation neurologic complications is only speculative.

One hypothesis to explain the development of postattenuation neurologic complications is that seizures occur because of sudden changes in the release of endogenous benzodiazepines following CPSS attenuation.²³ Alternatively, there could be sudden changes or imbalances in any number of other endogenous substances or neurotransmitters. In the present study, cats with poorly developed intrahepatic portal vasculature were more likely to develop postattenuation neurologic complications than were cats with better developed intrahepatic portal vasculature. It is possible that cats with a greater degree of intrahepatic portal development or change following attenuation were somehow more predisposed to the sudden potential changes or imbalances of endogenous substances or neurotransmitters.

A good clinical response to the first surgery, absence of neurologic signs at follow-up, and reduced preprandial and postprandial serum bile acids concentrations at follow-up were all significantly more likely in cats with high grades for portovenograms obtained at first surgery after shunt occlusion. This is in agreement with findings in dogs that IMP grade is significantly, positively correlated with resolution of clinical signs and decreased preprandial serum bile acids concentrations following CPSS attenuation.¹³ It is intuitive that animals with better developed hepatic portal vasculature would have better clinical and biochemical responses to CPSS attenuation. Interestingly, all 5 cats with neurologic signs at follow-up had postattenuation neurologic complications following the first surgery. It was therefore impossible to determine whether the neurologic signs in those cats were solely attributable to ongoing hepatic encephalopathy or whether residual postattenuation neurologic complications were also contributing to their clinical condition. The nature of any relationship between the resolution of the signs of both hepatic encephalopathy and postattenuation neurologic complications after surgery is unclear at this stage.

Findings of the present study were limited by the small number of cats involved. However, this is the only study designed to investigate the relationship between the results of portovenography and surgical success in cats with CPSS, and the results are compatible with those reported for a similar study involving dogs. Testing for associations between a large number of variables and portovenogram grade means, as occurred in our study, is vulnerable to type I and II errors. However, throughout our study, we attempted to consider only those variables that we believed had true biological relevance. To our knowledge, ours is the only study to identify a potential indicator, postocclusion IMP grade, for the development of postattenuation neurologic complications in cats.