Ammonia is a normal constituent of body fluids and is mainly produced in the gastrointestinal tract by bacterial metabolism of urea and glutamine.1 Ammonia is lipid soluble and easily crosses cell membranes.1,2 In the gastrointestinal tract, it diffuses through the intestinal mucosa into the portal venous circulation, where it is converted to ammonium; then the blood is detoxified by the urea cycle in the liver.1,2 Plasma ammonia concentration becomes abnormally increased in patients with PSS, liver failure, and urea cycle disorders.3–5
It is widely accepted that ammonia is a key factor in the pathogenesis of HE.6 Ammonia initiates HE by altering astrocyte function. Astrocytes are the main cells that metabolize ammonia in the brain. The conversion of glutamate and ammonia to glutamine causes osmotic stress, which results in astrocyte swelling, cerebral edema, and intracranial hypertension.7
Ammonia diffuses from the blood and crosses the BBB into the brain.8 Additional ammonia is formed within the brain by the metabolism of endogenous nitrogen–containing substances.8–10 Ammonia diffuses more freely into the brain of human patients with severe liver disease than healthy control subjects.11,12 Diffusion of ammonia from the blood to the brain increases as the arterial ammonia concentration increases.11 Interestingly, an increase in BBB permeability results in an increase in the diffusion of ammonia into the brain11,12 and can result in toxic ammonia concentrations in the brain even when arterial ammonia concentrations are near reference limits.11
Dogs with PSS often have clinical signs of HE,13 but research regarding the association between PSS and HE is limited. In 2 studies,14,15 dogs with PSS had abnormally increased CSF concentrations of several amino acids suggestive of an abnormally increased ammonia concentration, but CSF ammonia concentrations were not formally measured in either study. In an experimental study,16 ammonia concentrations in the CSF and blood were significantly greater in dogs that underwent a partial hepatectomy and creation of a portocaval shunt than in control dogs that underwent a sham operation. The dogs that underwent partial hepatectomy and creation of a portocaval shunt also developed signs of HE.16
In human patients with HE, arterial ammonia concentration is significantly greater than venous ammonia concentration,11,12 and an arterial ammonia concentration of approximately 150 μmol/L is significantly associated with the development of intracranial hypertension and cerebral edema.17,18 Likewise, in dogs with HE, arterial ammonia concentration is significantly greater than venous ammonia concentration, and there is a strong correlation between HE and the presence of portosystemic collateral circulation.19
The objective of the study reported here was to compare ammonia concentrations in arterial blood, venous blood, and CSF samples between dogs with and without congenital EHPSS in an attempt to elucidate the pathogenesis of HE. We hypothesized that the arterial ammonia concentration would be greater than the venous ammonia concentration in dogs with EHPSS and that there would be a positive correlation between ammonia concentrations in the blood and CSF.
Materials and Methods
Animals
All study procedures were approved by the Ethical Committee of the Faculty of Veterinary Medicine of Ghent University (EC2012/164 and EC2013/33) and by the Belgian Deontological Committee. Six healthy adult Beagles from a research colony (controls) and 19 client-owned dogs with congenital EHPSS were prospectively evaluated between July 2012 and October 2015. Owner consent was obtained for all dogs with a congenital EHPSS prior to study enrollment.
Study procedures and sample collection
Food but not water was withheld from all dogs > 4 months old for 12 hours before examination; for dogs ≤ 4 months old, food but not water was withheld for ≤ 4 hours. Control dogs were considered healthy on the basis of results of a physical examination, CBC, serum biochemical analysis, urinalysis, and abdominal ultrasonographic examination. Dogs with EHPSS likewise underwent a physical examination, CBC, serum biochemical analysis, and an abdominal ultrasonographic examination to assess the abdominal organs and screen for vascular abnormalities. Clinical signs of HE were graded on a 5-point scale as described,20 where 0 = clinically normal; 1 = abnormally decreased mobility or mild apathy; 2 = severe apathy or mild ataxia; 3 = salivation, severe ataxia, head pressing, apparent blindness, or circling; and 4 = seizures or coma.
Blood samples (1.0 mL) for preprandial and postprandial serum bile acid concentration analysis were collected by jugular venipuncture from each dog. Following collection of the preprandial blood sample, each dog was fed 2 teaspoons of a commercial highly digestible protein and fat diet,a and the postprandial blood sample was collected 2 hours later.
All dogs were anesthetized for collection of arterial and venous blood samples and a CSF sample for measurement of ammonia concentration and trans-splenic portal scintigraphy. The same anesthesia protocol was used for all dogs. Briefly, a 22-gauge catheter was aseptically placed in a cephalic vein, and each dog was premedicated with butorphanol tartrateb (0.2 mg/kg, IV). Ten minutes later, anesthesia was induced with propofolc (2 to 4 mg/kg, IV to effect) and maintained with a constant rate infusion of propofol (0.2 to 0.4 mg/kg/min, IV). Each dog was intubated, and 100% oxygen (flow rate, 1 L/min) was supplied through the endotracheal tube for the duration of the anesthetic session.
Prior to the transsplenic portal scintigraphy procedure, an arterial blood sample (700 μL) was collected from a femoral artery by use of a 25-gauge needle attached to a 1-mL syringe. Then, a venous blood sample (700 μL) was collected from a jugular vein by use of a 21-gauge needle attached to a 2.5-mL syringe. Each blood sample was transferred to a specialized heparinized whole blood separator tubed immediately after collection, and the tubes were placed on melting ice and immediately submitted to an in-house laboratory for determination of ammonia concentration. To collect the CSF sample, a small area (3 × 3 cm) of skin over the atlanto-occipital region was clipped and aseptically prepared. A CSF sample (0.5 mL) was aseptically collected via a cisternal puncture with a 21-gauge needle. The CSF sample was collected directly into a sterile tube without additives. If blood contamination was evident, another sample was collected. Similar to the arterial and venous blood samples, the tube containing the CSF sample was immediately placed on melting ice and submitted to an in-house laboratory for determination of ammonia concentration.
Transsplenic portal scintigraphy was performed as described21 to determine the presence (dogs with EHPSS) or absence (control dogs) of PSS. Briefly, intrasplenic injection of sodium pertechnetatee was performed with ultrasound guidance, and a dynamic scan was simultaneously initiated with a nuclear γ cameraf equipped with a low-energy, high-resolution collimator.
Sample processing
Blood samples obtained for preprandial and postprandial bile acid analysis were centrifuged, and the serum was harvested. The serum samples were then sent to an external laboratory for bile acid analysis.
Ammonia concentrations in arterial and venous blood samples and CSF samples were determined as soon as possible after collection by use of a portable blood ammonia analyzerg (device A) and a nonportable biochemical analyzerh (device B). Both devices measure the amount of gaseous ammonia after the ammonium ions in the sample have been converted. The blood samples analyzed by device A were obtained from whole blood separator tubes by use of a capillary tube immediately before the separator tube was inserted into device B. For each CSF sample, 200 μL of the sample was transferred to a sample cup,i and the sample analyzed by device A was obtained from the sample cup by use of a capillary tube immediately before the sample cup was inserted into device B.
For each remaining CSF sample, a WBC count was performed manually by microscopic examination.
Statistical analyses
Statistical analyses were performed by use of a commercial software package.j Data distributions were checked for normality by use of the Kolmogorov-Smirnov test with Lilliefors significance correction. The mean ± SD was reported for data that were normally distributed, and the median (range) was reported for data that were not normally distributed. Comparisons between dogs with EHPSS and control dogs were performed with an unpaired t test or Mann-Whitney U test for independent continuous variables that were and were not normally distributed, respectively, and a paired t test or Wilcoxon signed rank test for paired data (eg, preprandial and postprandial serum bile acid concentrations) that were and were not normally distributed, respectively. Correlation was assessed with the Pearson product-moment coefficient (r) or Spearman rank coefficient (p) for variables that were and were not normally distributed, respectively. Values of P ≤ 0.05 were considered significant for all analyses.
Results
Dogs
The control Beagles consisted of 3 spayed females and 3 castrated males and ranged in age from 36 to 54 months and in body weight from 9.1 to 16.0 kg. The dogs with EHPSS consisted of 5 sexually intact females, 5 spayed females, 7 sexually intact males, and 2 castrated males and ranged in age from 3 to 65 months and in body weight from 1.5 to 13.4 kg. Dogs with EHPSS included Yorkshire Terriers (n = 4), Bichons Frises (2), Chihuahuas (2), and Dachshunds (2) as well as a Beagle, Jack Russell Terrier, Maltese, Miniature Pincher, Miniature Schnauzer, mixed-breed dog, Norwich Terrier, Russian Tsvetnaya Bolonka, and Scottish Collie (1 each).
HE
None of the control dogs had clinical signs of HE (HE grade = 0). The median HE grade was 3 (range, 2 to 4) for dogs with EHPSS. Of the 19 dogs with EHPSS, 15 had signs of apathy and some degree of head pressing, ataxia, or circling, and 7 had seizures, but none were comatose.
Serum bile acid concentrations
Preprandial and postprandial serum bile acid concentrations for the controls and dogs with EHPSS were summarized (Table 1). For controls, all serum bile acid concentrations were well within the reference range (< 19 μmol/L). The median preprandial (166 μmol/L; range, 27 to 381 μmol/L) and postprandial (218 μmol/L; range, 49 to 656 μmol/L) serum bile acid concentrations for the dogs with EHPSS were significantly (P < 0.001) greater than those for controls.
Descriptive data for HE grade, preprandial and postprandial serum bile acid concentrations, and arterial and venous blood and CSF ammonia concentrations as determined by 2 devices for 6 healthy adult Beagles (controls) and 19 dogs with EHPSS.
Variable | Controls | Dogs with EHPSS |
---|---|---|
HE grade | 0 | 3 (2–4) |
Preprandial serum bile acid (μmol/L) | 6 (6–11) | 166 (27–381) |
Postprandial serum bile acid (μmol/L) | 8 (1–12) | 218 (49–656) |
Arterial ammonia (μmol/L) | ||
Device A | 18.2 ± 8.0 | 173.1 ± 57.5 |
Device B | 16.8 ± 8.1 | 119.7 ± 69.4* |
Venous ammonia (μmol/L) | ||
Device A | 19.3 ± 7.7 | 158.0 ± 53.8† |
Device B | 23.0 ± 9.5 | 103.8 ± 58.3* |
CSF ammonia (μmol/L) | ||
Device A | 6.0 ± 5.3 | 106.6 ± 55.6 |
Device B | 3.7 ± 2.0 | 60.1 ± 42.9* |
Values represent the median (range) or mean ± SD. Clinical signs of HE were graded on a 5-point scale, where 0 = clinically normal; 1 = abnormally decreased mobility or mild apathy; 2 = severe apathy or mild ataxia; 3 = salivation, severe ataxia, head pressing, apparent blindness, or circling; and 4 = seizures or coma. Ammonia concentrations in arterial and venous blood samples and CSF samples were measured in parallel by a portable blood ammonia analyzer (device A) and nonportable biochemical analyzer (device B). All values for dogs with EHPSS were significantly (P < 0.001) greater than the corresponding values for the control dogs.
Value differs significantly (P < 0.001) from the corresponding value measured by device A.
Value differs significantly (P < 0.05) from the arterial ammonia concentration measured by the same device.
Ammonia concentrations
Ammonia concentrations in arterial and venous blood samples and CSF samples were summarized (Table 1). The mean arterial and venous ammonia concentrations for dogs with EHPSS were significantly (P < 0.001) greater than those for controls. The arterial and venous ammonia concentrations measured by device A were significantly greater (P < 0.001) than those measured by device B for dogs with EHPSS; nevertheless, there was a strong positive correlation between the ammonia concentrations measured by devices A and B for both arterial (r = 0.884) and venous (r = 0.819) blood samples. Although arterial ammonia concentrations were greater than venous ammonia concentrations, only those measured by device A differed significantly (P < 0.05). There was a strong positive correlation between arterial and venous ammonia concentrations measured by device A (r = 0.960) and device B (r = 0.900). There was a weak, albeit significant, positive correlation between preprandial serum bile acid concentration and venous ammonia concentration measured by device A (ρ = 0.461) and device B (ρ = 0.486) as well as between postprandial serum bile acid concentration and venous ammonia concentration measured by device A (ρ = 0.414) and device B (ρ = 0.394). Likewise, there was only a weak, albeit significant, positive correlation between HE grade and arterial ammonia concentration measured by device A (ρ = 0.445) and device B (ρ = 0.461).
All CSF samples were macroscopically normal, and the CSF WBC count was within the reference range (< 8 cells/μL) for all controls and dogs with EHPSS. The mean CSF ammonia concentration for dogs with EHPSS was significantly (P < 0.001) greater than that for controls. The mean CSF ammonia concentration measured by device A was significantly (P < 0.001) greater than that measured by device B. There was a strong significant (P < 0.001) positive correlation between arterial and CSF ammonia concentrations measured by device A (r = 0.884) and device B (r = 0.870) as well as between venous and CSF ammonia concentrations measured by device A (r = 0.892) and device B (r = 0.725). The HE grade was not significantly (P = 0.09) correlated with CSF ammonia concentration measured by device A, but there was a significant (P = 0.05) weak positive correlation (ρ = 0.396) between HE grade and CSF ammonia concentration measured by device B.
Discussion
Results of the present study indicated that ammonia concentrations in arterial and venous blood samples and CSF samples of dogs with EHPSS were significantly greater than those for healthy control dogs. There was also a strong positive correlation between ammonia concentrations in the CSF and blood regardless of whether it was arterial or venous.
The pathogenesis of HE is complex, and although ammonia is not the sole agent involved, historically it has been the only measurable neurotoxin.22 The distribution and presence of ammonia and ammonium are dependent on both intracellular and extracellular pH.1,11,12,23 At a physiologic pH, most ammonia is in the form of the positively charged ammonium ion (NH4+) rather than in its gaseous form (NH3).1,2,11,12 In general, ammonia passively diffuses toward a lower concentration gradient in the brain and is retained in the brain parenchyma owing to a difference between systemic and brain pH. In humans, the normal systemic pH is 7.4, and the brain pH is somewhat lower.23 Because the brain pH is lower than the systemic pH, ammonia retained in the brain is converted to the ammonium ion, which does not readily diffuse across cell membranes.11 Ammonia diffuses more freely into the brain of human patients with severe liver disease than into the brain of healthy control subjects.11,12 Toxicosis results from an increase in plasma ammonia concentration in conjunction with increases in intracellular and extracellular pH.23 The toxic effect of ammonia on the brain results from direct interaction with both excitatory and inhibitory neurotransmission.23 The consequences of hyperammonemia in the CNS include amino acid, cerebral energy, and neurotransmission disturbances; axonal and dendritic growth impairment; and alterations in nitric oxide synthesis, oxidative stress, and signal transduction and channel transporter activities,24–28 which lead to the swelling and death of astrocytes.24,29,30
In human patients with HE, disease severity is positively correlated with blood ammonia concentration,31 and an arterial ammonia concentration ≥ 150 μmol/L is associated with a negative prognosis.17,18 There is also a strong positive correlation between disease severity and blood ammonia concentration in dogs with HE.19 In the present study, the blood ammonia concentration for dogs with HE (regardless of disease severity) was significantly greater than that for healthy dogs; however, the correlations between blood ammonia concentrations and disease severity (ie, HE grade) were rather weak for dogs with EHPSS.
Traditionally, in human medicine, the ammonia concentration is measured in arterial rather than venous blood samples,11 whereas in dogs, determination of venous ammonia concentration is part of the diagnostic workup for dogs with suspected PSS.1,2 To our knowledge, prior to the present study, there was only 1 study19 in which ammonia concentrations were compared between arterial and venous blood samples for dogs with HE caused by various etiologies. Unfortunately, detailed data regarding the blood ammonia concentrations for dogs with PSS (a subgroup of the dogs with HE) were not provided in that study,19 although the arterial ammonia concentration was significantly greater than the venous ammonia concentration for all dogs with HE. Similarly, mean arterial ammonia concentration was greater than the mean venous ammonia concentration for the dogs with EHPSS in the present study; however, that difference was statistically significant only when the ammonia concentration was measured by device A.
In the present study, dogs with EHPSS had high CSF ammonia concentrations that were strongly and positively correlated with blood ammonia concentrations. Investigators of other studies7,14 have presumptively stated that the ammonia concentration is abnormally increased in the CSF of dogs with congenital EHPSS without actually measuring the CSF ammonia concentration. Data regarding CSF ammonia concentration are lacking, likely because of a paucity of validated techniques to measure the ammonia concentration in CSF. In the present study, 2 commercial devices (device A and device B) were used to measure ammonia concentration in blood samples as well as CSF samples. Both devices have the same mode of action. The respective test slides for each device contain a buffer in the top layer that converts ammonium ions in the sample into gaseous ammonia, which passes through a selectively permeable membrane and reacts with a pH indicator (bromocresol green for device A and bromophenol blue for device B). For both devices, color development is proportional to the amount of ammonia in the sample. To our knowledge, no device has been validated to measure the ammonia concentration in CSF. For dogs with EHPSS, the CSF ammonia concentration measured by device A was consistently greater than that measured by device B. Nevertheless, for each device, there was a strong positive correlation between the ammonia concentration measured in blood (arterial or venous) and that measured in CSF, which suggested that either device can be used in clinical practice to provide an estimate of the CSF ammonia concentration. However, the CSF ammonia concentrations measured in the present study should be interpreted cautiously in a comparative rather than absolute manner until the devices have been validated for measurement of the ammonia concentration in CSF samples.
In humans, the ammonia concentration in blood does not optimally reflect that in the brain, and the ammonia uptake by the brain increases as arterial ammonia concentration increases, albeit in a nonlinear manner (ie, the rate of ammonia uptake by the brain is greater than the rate of increase in arterial ammonia concentration).11 The permeability of the BBB to ammonia is a critical determinant in the rate of ammonia uptake by the brain; thus, the effect of ammonia on the brain may be greater than that predicted solely on the basis of blood ammonia concentration.11,12,32 Similar studies have not been performed in dogs, and data regarding the permeability of the BBB to ammonia are unavailable for dogs with PSS. The metabolic uptake of ammonia in the brain was not measured for the dogs of the present study, and the assumption that permeability of the BBB to ammonia is greater in dogs with EHPSS than healthy dogs requires further investigation. In human patients with subclinical or clinical HE, the BBB is particularly permeable to the gaseous form of ammonia, but measureable amounts of the ammonium ion also cross the BBB.11,33 In the present study, the ammonia concentrations in both the blood and CSF of dogs with EHPSS were significantly greater than the corresponding concentrations in the healthy controls, which suggested that an increasing concentration of ammonia in the brain can lead to toxicosis and severe clinical signs of HE. However, we could not definitively rule out the possibility that seizures and other clinical signs of HE in some of the dogs with EHPSS were caused or triggered by another neurodegenerative disease process. If the clinical signs of HE were indeed induced by an increase in cerebral ammonia concentration, administration of a medical treatment regimen that alters the ratio between the ionic and gaseous forms of ammonia might decrease the influx of ammonia into the CSF and brain prior to surgery to correct the EHPSS, but this requires investigation.
Currently, measurement of preprandial and postprandial serum bile acid concentrations is the most commonly used liver function test to diagnose PSS in dogs. The diagnostic sensitivity and specificity of preprandial serum bile acid concentration for identification of dogs with PSS range from 88% to 93% and from 68% to 87%, respectively.2,34 Similar to results of other studies,2,34 the dogs with EHPSS in the present study had preprandial and postprandial serum bile acid concentrations that were markedly increased from the reference range. However, preprandial and postprandial serum bile acid concentrations were only weakly correlated with venous ammonia concentration.
The present study had a few limitations. Only a limited number of dogs with various degrees of HE were evaluated. Also, ammonia is unstable ex vivo, and samples need to be processed with special care so as not to influence the test results. Precautions such as the use of melting ice for sample transport and the nearly immediate processing of samples (time from sample collection to measurement of ammonia concentration was < 120 seconds for all samples) and discarding of CSF samples that were grossly contaminated with blood should have minimized the risk for preanalytic errors, although they can never be completely excluded.
In the present study, the CSF ammonia concentrations for dogs with EHPSS were significantly greater than those for healthy control dogs, and there was a strong positive correlation between the ammonia concentrations in the CSF and blood, which suggested that the permeability of the BBB to ammonia may be abnormally increased in dogs with EHPSS. Furthermore, for dogs with EHPSS, the ammonia concentration was markedly increased from the reference range in both arterial and venous blood samples. This indicated that, in dogs, venous blood samples, which are generally easy to obtain, can be substituted for arterial blood samples, which can be difficult to obtain and often require anesthetizing the patient, for measurement of blood ammonia concentration. Additionally, because ammonia passes through the BBB into the brain in a nonlinear manner relative to the blood ammonia concentration, caution is necessary to ensure that the presence or severity of HE is not underestimated when blood ammonia concentrations are interpreted. Further investigation of the relationship between blood or CSF ammonia concentration and clinical signs of HE or the surgical outcome for dogs with EHPSS is warranted.
Acknowledgments
Supported in part by a European College of Veterinary Surgeons’ Surgeon-in-Training Research Grant.
Presented in part at the 23rd Annual Scientific Meeting of the European College of Veterinary Surgeons, Copenhagen, July 2014.
The authors thank Sara Kol for language editing.
ABBREVIATIONS
BBB | Blood-brain barrier |
EHPSS | Extrahepatic portosystemic shunts |
HE | Hepatic encephalopathy |
PSS | Portosystemic shunts |
Footnotes
Prescription Diet a/d, Hill's Pet Nutrition NV, Brussels, Belgium.
Butorphanol, Dolorex, Intervet N V, Oostkamp, Belgium.
Propofol, Propovet Multidose, Abbott Laboratories Ltd, Maidenhead, Berkshire, England.
Heparinized whole blood separator, Idexx Laboratories Europe, Hoofddorp, The Netherlands.
Sodium pertechnetate, Drytec technetium 99mTc generator, GE Healthcare, Amersham, England.
GCA-7200A/DI, Toshiba Medical Systems Corp, Otawara, Japan.
PocketChem BA, A. Menarini Diagnostics Benelux, Zaventem, Belgium.
Catalyst Dx, Idexx Laboratories Inc, Westbrook, Me.
Catalyst sample cup, Idexx Laboratories Europe, Hoofddorp, The Netherlands.
SPSS Statistics, IBM Corp, Armonk, NY.
References
1. Goggs R, Serrano S, Szladovits B, et al. Clinical investigation of a point-of-care blood ammonia analyzer. Vet Clin Pathol 2008; 37: 198–206.
2. Gerritzen-Bruning MJ, van den Ingh T, Rothuizen J. Diagnostic value of fasting plasma ammonia and bile acid concentrations in the identification of portosystemic shunting in dogs. J Vet Intern Med 2006; 20: 13–19.
3. Strombeck DR, Meyer DJ, Freedland RA. Hyperammonemia due to a urea cycle enzyme deficiency in two dogs. J Am Vet Med Assoc 1975; 166: 1109–1111.
4. Zandvliet MM, Rothuizen J. Transient hyperammonemia due to urea cycle enzyme deficiency in Irish Wolfhounds. J Vet Intern Med 2007; 21: 215–218.
5. Center SA, Magne ML. Historical, physical examination, and clinicopathologic features of portosystemic vascular anomalies in the dog and cat. Semin Vet Med Surg (Small Anim) 1990; 5: 83–93.
6. Albrecht J, Dolinska M. Glutamine as a pathogenic factor in hepatic encephalopathy. J Neurosci Res 2001; 65: 1–5.
7. Prakash R, Mullen KD. Mechanisms, diagnosis and management of hepatic encephalopathy. Nat Rev Gastroenterol Hepatol 2010; 7: 515–525.
8. Lemberg A, Fernández MA. Hepatic encephalopathy, ammonia, glutamate, glutamine and oxidative stress. Ann Hepatol 2009; 8: 95–102.
9. Norenberg MD, Martinez-Hernandez A. Fine structural localization of glutamine synthetase in astrocytes of rat brain. Brain Res 1979; 161: 303–310.
10. Norenberg MD. The role of astrocytes in hepatic encephalopathy. Neurochem Pathol 1987; 6: 13–33.
11. Lockwood AH. Blood ammonia levels and hepatic encephalopathy. Metab Brain Dis 2004; 19: 345–349.
12. Lockwood AH, Yap EW, Wong WH. Cerebral ammonia metabolism in patients with severe liver disease and minimal hepatic encephalopathy. J Cereb Blood Flow Metab 1991; 11: 337–341.
13. Broome CJ, Walsh VP, Braddock JA. Congenital portosystemic shunts in dogs and cats. N Z Vet J 2004; 52: 154–162.
14. Butterworth J, Gregory CR, Aronson LR. Selective alterations of cerebrospinal fluid amino acids in dogs with congenital portosystemic shunts. Metab Brain Dis 1997; 12: 299–306.
15. Holt DE, Washabau RJ, Djali S, et al. Cerebrospinal fluid glutamine, tryptophan, and tryptophan metabolite concentrations in dogs with portosystemic shunts. Am J Vet Res 2002; 63: 1167–1171.
16. Meyer HP, Legemate DA, van den Brom W, et al. Improvement of chronic hepatic encephalopathy in dogs by the benzodiazepine-receptor partial inverse agonist sarmazenil, but not by the antagonist flumazenil. Metab Brain Dis 1998; 13: 241–251.
17. Bhatia V, Singh R, Acharya SK. Predictive value of arterial ammonia for complications and outcome in acute liver failure. Gut 2006; 55: 98–104.
18. Bernal W, Hall C, Karvellas CJ, et al. Arterial ammonia and clinical risk factors for encephalopathy and intracranial hypertension in acute liver failure. Hepatology 2007; 46: 1844–1852.
19. Rothuizen J, van den Ingh TS. Arterial and venous ammonia concentrations in the diagnosis of canine hepato-encephalopathy. Res Vet Sci 1982; 33: 17–21.
20. Torisu S, Washizu M, Hasegawa D, et al. Brain magnetic resonance imaging characteristics in dogs and cats with congenital portosystemic shunts. Vet Radiol Ultrasound 2005; 46: 447–451.
21. Morandi F, Cole RC, Tobias KM, et al. Use of 99mTCO4(−) trans-splenic portal scintigraphy for diagnosis of portosystemic shunts in 28 dogs. Vet Radiol Ultrasound 2005; 46: 153–161.
22. Carrera I, Kircher PR, Meier D, et al. In vivo proton magnetic resonance spectroscopy for the evaluation of hepatic encephalopathy in dogs. Am J Vet Res 2014; 75: 818–827.
23. Lockwood AH, McDonald JM, Reiman RE, et al. The dynamics of ammonia metabolism in man. Effects of liver disease and hyperammonemia. J Clin Invest 1979; 63: 449–460.
24. Brusilow SW, Koehler RC, Traystman RJ, et al. Astrocyte glutamine synthetase: importance in hyperammonemic syndromes and potential target for therapy. Neurotherapeutics 2010; 7: 452–470.
25. Felipo V, Butterworth RF. Neurobiology of ammonia. Prog Neurobiol 2002; 67: 259–279.
26. Braissant O, McLin VA, Cudalbu C. Ammonia toxicity to the brain. J Inherit Metab Dis 2013; 36: 595–612.
27. Cagnon L, Braissant O. Hyperammonemia-induced toxicity for the developing central nervous system. Brain Res Rev 2007; 56: 183–197.
28. Cudalbu C. In vivo studies of brain metabolism in animal models of hepatic encephalopathy using 1H magnetic resonance spectroscopy. Metab Brain Dis 2013; 28: 167–174.
29. Albrecht J, Zielinska M, Norenberg MD. Glutamine as a mediator of ammonia neurotoxicity: a critical appraisal. Biochem Pharmacol 2010; 80: 1303–1308.
30. Brusilow SW, Traystman R. Hepatic encephalopathy. N Engl J Med 1986; 314: 786–787.
31. Ong JP, Aggarwal A, Krieger D, et al. Correlation between ammonia levels and the severity of hepatic encephalopathy. Am J Med 2003; 114: 188–193.
32. Or M, Peremans K, Martlé V, et al. Regional cerebral blood flow assessed by single photon emission computed tomography (SPECT) in dogs with congenital portosystemic shunt and hepatic encephalopathy. Vet J 2017; 220: 40–42.
33. Ott P, Larsen FS. Blood-brain barrier permeability to ammonia in liver failure: a critical reappraisal. Neurochem Int 2004; 44: 185–198.
34. Ruland K, Fischer A, Hartmann K. Sensitivity and specificity of fasting ammonia and serum bile acids in the diagnosis of portosystemic shunts in dogs and cats. Vet Clin Pathol 2010; 39: 57–64.