MRI Findings in a Dog with Kernicterus

Introduction

Jaundice, or icterus, is a common clinical finding in a variety of veterinary diseases. Kernicterus is a term used to describe varying degrees of brain damage secondary to deposition of bilirubin in the gray matter of the brain (especially the basal ganglia) and spinal cord that is also accompanied by nerve cell degeneration. The term kernicterus is most appropriately used in reference to a postmortem anatomic diagnosis made by visualizing yellow pigmentation in the basal ganglia. The terms bilirubin encephalopathy and bilirubin-induced neurologic dysfunction are more accurately used to refer to clinical signs that result. In practice, the terms are often used interchangeably, and kernicterus is often presumed antemortem in the presence of specific historical, physical examination, laboratory, and diagnostic imaging findings.^1,2

Most reports of clinical kernicterus involve human neonates. It is now considered to be a relatively rare condition due to screening protocols in place to help with early recognition and treatment of significant jaundice. Although jaundice is relatively common, occurring in up to 60% of human neonates, severe jaundice (typically defined as total serum bilirubin > 342–513 μmol/l) is rare and often preventable. However, kernicterus has been reported to occur in up to 12–20% of cases of severe jaundice.^3,4 Kernicterus is rarely diagnosed in adults and only sparsely referenced in the veterinary literature, with reports limited to single cases in a foal, a kitten, and an adult dog.^5–7

Although kernicterus is rarely reported in animals, the true incidence of this condition in dogs and cats with severe hyperbilirubinemia remains unknown. It is often very difficult to assess either subtle or mild neurologic defects or intellectual defects in veterinary patients. It is common for veterinary patients with high bilirubin concentrations to be weak, lethargic, and depressed, but determining if those clinical signs are secondary to the underlying disease process, anemia, medications, or true neurologic disease is challenging. There is very limited information available about the histopathologic findings within the central nervous system (CNS) of icteric animals, and the authors of this report are not aware of any prior reports of cross-sectional neurologic imaging in canine patients with either confirmed or suspected kernicterus.

Case Report

A 9 yr old spayed female mixed-breed dog presented to an emergency clinic with a 12 hr history of inappetence, vomiting, and lethargy. There was no history of prior illness, and the most recent vaccinations had been administered 3 mo prior to presentation. Point-of-care testing revealed a packed cell volume (PCV) of 25%, plasma total solids of 60 g/l, and macroscopic red blood cell (RBC) agglutination on a glass slide. Thoracic radiographs revealed a mild, diffuse, interstitial pulmonary pattern. Abdominal radiographs revealed diffuse splenomegaly. A presumptive diagnosis of immune-mediated hemolytic anemia was made, and treatment with IV fluids, famotidine, dexamethasone, and doxycycline was initiated. Twelve hours later, the dog was referred to the University of Florida Small Animal Hospital for further evaluation and treatment.

On presentation to the emergency service at the University of Florida Small Animal Hospital, the dog was bright, alert, responsive, and ambulatory. She weighed 22 kg and had a body condition score of 5/9. Physical examination was unremarkable, except for pale, icteric mucous membranes, pyrexia (39.6C), and splenomegaly. Point-of-care testing revealed a PCV of 18%, total solids of 70 g/l, marked hemolysis, and a positive saline agglutination test. Hematologic abnormalities included lymphopenia (0.28 × 10⁹/L; reference range, 1.0–4.8 × 10⁹/L), eosinopenia (0.3 × 10⁹/L; reference range, 0.1–1.25 × 10⁹/L), a left shift (band neutrophils, 1.27 × 10⁹/L; reference range, 0–0.3 × 10⁹/L) with metamyelocytes (0.06 × 10⁹/L, reference range, 0 × 10⁹/L) and toxic changes, and a normal number of mature neutrophils, some of which displayed dark gray, amorphous, cytoplasmic granules suggestive of hemosiderin and hemotoidin crystals. There was also a severe anemia (hematocrit was 18.2%) with 2+ anisocytosis, 1+ polychromasia, 3+ spherocytes, 1+ ghost cells, 0.4 × 10⁹/L nucleated RBCs, and evidence of regeneration (reticulocytes, 85.9 × 10⁹/L). Serum biochemical abnormalities included an elevated alkaline phosphatase (3.34 μkat/L, reference range, 0.27–1.85 μkat/L), elevated aspartate aminotransferase (4.86 μkat/L; reference range, 0.17–0.77 μkat/L), elevated total bilirubin (123.1 μmol/l; reference range, 0–6.8 μmol/l), decreased calcium (2.18 mmol/L; 2.38–2.9 mmol/L), decreased creatinine (17.7 μmol/l; 70.7–150.3 μmol/l), elevated blood urea nitrogen (13.2 mmol/L; 2.5–9.6 mmol/L), and elevated glucose (8.55 mmol/L; 4.83–6.93 mmol/L). Urinalysis revealed a specific gravity of 1.022, pigmenturia, and proteinuria (4+). Urine Culture was negative. Prothrombin time and activated partial thromboplastin time were within normal limits. An ELISA^a for detection of Dirofilaria immitis antigen and antibodies to Ehrlichia canis, Anaplasma phagocytophilum, and Borrelia burgdorferi was negative, as was immunoflourescent antibody testing for Rickettsia ricketsii. Thoracic radiographs were considered to be within normal limits, and an abdominal ultrasound showed moderate, diffuse splenomegaly with normal echotexture.

The finding of a severe, regenerative anemia with evidence of visibly pigmented serum, macroscopic and microscopic RBC agglutination, and spherocytosis was considered to support the diagnosis of an immune-mediated hemolytic anemia, and treatment was targeted for that condition, in addition to providing supportive care as needed. IV fluid therapy was initiated with lactated Ringer’s solution^b with 20 mEq/L potassium chloride^c, famotidine^d (0.5 mg/kg IV q 12 hr), dexamethasone^e (0.11 mg/kg IV q 12 hr), metoclopramide^f (1 mg/kg/24 hr IV continuous rate infusion), doxycycline^g (10 mg/kg IV q 24 hr), and maropitant^h (1 mg/kg subcutaneously q 24 hr).

Twelve hours after hospitalization, the dog developed a head bob and generalized ataxia and her neurologic status continued to decline to comatose over the next 4 hr. Several additional diagnostic tests, including serial PCVs and venous blood gases, plasma ammonia (fasted), coagulation profile, and serum bilirubin, were performed as the dog’s neurologic status worsened. The venous blood gas analysis performed when neurologic signs were first noticed revealed a metabolic acidosis with a pH of 7.202, bicarbonate of 11.7 mmol/L, and lactate of 3.44 mmol/L. Measured electrolytes (including ionized calcium), glucose, venous partial pressure of CO₂, indirect blood pressure, and O₂ saturation were within normal limits. The PCV had declined to 14%. The rate of crystalloid fluid administration was increased, and packed RBCsⁱ (12.7 mL/kg) were transfused. The PCV increased to 27% posttransfusion, but venous blood gas abnormalities persisted. Ammonia concentration was abnormal (51 mmol/L; reference range, 0–35 mmol/L). The prothrombin time (9.4 sec) and activated partial thromboplastin time (20.4 sec) were also abnormal (reference ranges, 5.1–8.8 sec and 7.7–13.7 sec, respectively). The patient was significantly more icteric, and total bilirubin had increased from 123.1 μmol/l to 940.5 μmol/l (reference range, 0–6.8 μmol/l). Another new laboratory abnormality noted at that time was a decreased albumin concentration (16 g/L; reference range, 29–38 g/L). Additional treatments with lactulose^j enemas (1 mg/kg rectally q 6 hr) and phytonadione^k (3 mg/kg subcutaneously q 24 hr) were initiated. An infusion of IV human immunoglobulin^l (136 mg/kg IV over 6 hr) was started after the transfusion.

Several hours after becoming comatose, the dog began having cluster seizures that were not responsive to multiple doses of diazepam^m (0.5 mg/kg IV). Treatment with phenobarbitalⁿ (3 mg/kg IV q 12 hr) was initiated, and one dose of mannitol^o (0.5 mg/kg IV) was administered. The cluster seizures ceased, and further individual breakthrough seizures were treated successfully with doses of diazepam (0.5 mg/kg IV).

The following day, approximately 24 hr after the first seizure was noted, the dog was transferred to the internal medicine service and an MRI examination of the brain was performed that morning. There was marked bilaterally symmetrical increase in signal intensity on T2-weighted and fluid-attenuated inversion recovery (FLAIR) sequences (Figures 1, 2) within regions corresponding to thalamus, lateral and medial geniculate nuclei, caudate nuclei, regions of other basal nuclei (especially the globus pallidus), deep cerebellar nuclei, and throughout cerebral cortical gray matter, along with subjectively slight increases in signal intensity in regions of the subthalamic nuclei, substantia nigra, and hippocampus. Changes within cortical gray matter were most marked at the level of the cingulate gyri and lateral rhinal sulci. On T1-weighted images, the caudate nuclei, region of the globus pallidus, and deep cerebellar nuclei were slightly hyperintense with respect to cortical gray matter, but no other abnormalities were seen (Figures 1, 2). There was no evidence of enhancement on T1-weighted images after contrast^p administration and no evidence of hemorrhage on T2*-weighted images. Cerebellar cortical gray matter was normal in appearance. Given those changes, a metabolic disease, storage disease, or toxic condition was suspected.

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

The patient later developed a respiratory acidosis due to hypoventilation that further complicated the pre-existing metabolic acidosis. Additional changes in IV fluid type^q, administration rate, and treatment with Na bicarbonate^r were not successful in improving pH. Ventilatory support was offered, but declined by the client. Approximately 72 hr after presentation and 60 hr after development of neurologic signs, the patient arrested and cardiopulmonary resuscitation was unsuccessful.

Gross examination of formalin-fixed, sectioned brain revealed multifocal, distinct yellow foci (Figure 3) that corresponded to thalamus (particularly rostral, ventral rostral, ventral lateral, and ventral caudal thalamic nuclei), interthalamic adhesion, lateral and medial geniculate nuclei, putamen, subthalamic nuclei, substantia nigra, cuneate nuclei, cerebellar nuclei (particularly dentate and interposed nuclei), and cortical gray matter. Histologically, affected regions exhibited multifocal, moderate to severe neuronal necrosis with moderate astrocytosis. The hippocampus, Purkinje cells, and brainstem also exhibited multifocal, moderate neuronal necrosis. Although neuronal necrosis and astrocystosis were predominantly localized to sites of gross yellow discoloration, bilirubin was not specifically identified in the cytoplasm of necrotic cells. Mild to moderate multifocal thrombi were identified in the lungs. In the liver, severe diffuse canalicular and interlobular bile casts, mild hepatocellular degeneration, biliary hyperplasia, hemosiderosis, and lipogranuloma formation were noted. Marked diffuse bilirubinuric nephrosis with glomerular and tubular proteinosis, tubular degeneration, and intratubular casts were noted in both kidneys. In the spleen, there was moderate erythrophagocytosis and hemosiderosis with extramedullary hematopoiesis.

View Figure

• Download as Powerpoint
• Download Figure

Discussion

Kernicterus develops when unconjugated bilirubin (B_u) accumulates within neurons in specific areas of the brain. Although B_u has potent antioxidant properties and may protect brain cells from oxidative damage (even when present at modestly increased levels), at higher concentrations, B_u is toxic to astrocytes and neurons, damaging mitochondrial, endoplasmic reticulum, and plasma membranes.^8,9 In healthy animals, most B_u in plasma is bound to albumin. Albumin-bound unconjugated bilirubin (Alb-B_u) does not readily cross the blood-brain barrier (BBB). However, a very small fraction of bilirubin may be present in a free (unbound) bilirubin (B_f) form which can readily enter CNS tissues and cerebrospinal fluid. Accumulation of B_u in the CNS and cerebrospinal fluid is also limited by active exportation, probably mediated by the multidrug resistance-associated proteins (MRPs) and multidrug resistance P-glycoproteins (MDRs) in choroid plexus epithelia, capillary endothelia, astrocytes, and neurons.⁹ Risk factors for kernicterus therefore include increased BBB permeability, an increased concentration of B_f, diminished capacity for exportation of bilirubin, or some combination of these.

The BBB is a dynamic and regulated interface between circulating blood and the CNS based on unique features of cerebral microvascular endothelium, which, together with astrocytes, pericytes, neurons, and extracellular matrix, constitutes a neurovascular unit essential for maintaining the health and function of the CNS. Tight junctions between endothelial cells restrict paracellular diffusion of water soluble substances from the blood to the brain. Permeability of the BBB can be altered by hypoxia, ischemia, inflammation, infection, trauma, drugs, and systemic acid/base disturbances.¹⁰ In this patient, although there was neither gross nor histologic evidence of vasculitis, meningitis, or meningoencephalitis and no history of trauma, other risk factors including anemia (as a cause of hypoxia), acidosis, and numerous medications may have influenced BBB permeability to B_u, including the Alb-B_u form.

Although the BBB is not typically permeable to Alb-B_u, B_f is able to diffuse passively across cell membranes. B_u can bind to albumin with very high affinity and can also bind to other proteins in the plasma; thus, B_f typically represents < 0.1% of the total serum B_u.^9,11 However, if B_u concentration increases or albumin concentration deceases, the proportion of B_f can increase significantly. There are several methods reported for measurement of B_f in patient samples, but that value is not routinely measured in most veterinary laboratories. As a very general rule, if the molar ratio between total serum bilirubin concentration and serum albumin concentration is < 1.0, the concentration of B_f is likely to be low, as is the risk of neurotoxicity.¹¹ Along with high total B_u concentration and low albumin concentration, several other factors, such as either acidosis or medications that compete for albumin binding sites, can significantly influence the concentration of B_f.⁹ High B_f concentrations are thought to be the primary risk factor for kernicterus in humans.

In human neonates, high B_f concentration may occur due to a combination of factors that include a diminished capacity for conjugation of bilirubin by the liver resulting in high levels of B_u, decreased affinity for B_u binding with α-fetoprotein compared with adult albumin, and inhibitors of protein binding found in neonatal plasma that can include a variety of commonly used medications.⁹ Many early publications also report that the neonatal BBB is not fully developed making it more permeable to bilirubin. More recent studies suggest that BBB permeability is not significantly different in healthy newborns compared with adults, although there are still questions about whether the BBB might be more easily damaged in newborns. In most case reports of kernicterus in adult humans, multiple factors are involved, such as congenital disorders that result in high baseline B_u levels, disorders or procedures that decrease serum albumin levels, and administration of medications that displace B_u from albumin.¹¹ Similarly, in the dog described in this report, several factors may have contributed to a rise in B_f concentration, including high B_u levels resulting from acute severe hemolysis, low serum albumin levels, acidosis, and the use of multiple medications with unknown potential for affecting bilirubin-albumin binding. At presentation, the patient had a total bilirubin/albumin ratio of 0.3, but by the time the neurologic signs progressed to seizures and coma, the ratio had increased to 4.1. Unfortunately, B_f concentration was not measured in this case.

Upregulation of MRPs by B_u may represent an important adaptive mechanism that protects the CNS from toxicity.¹² There are reports of mutations of MDR1 resulting in abnormal expression of MRPs in canines associated with a variety of neurologic side effects to various medications.^13–17 The dog reported in this case was a shepherd crossbreed, and the potential for a MDR1 mutation was not investigated.

MRI is a useful diagnostic tool to aid in presumptive antemortem diagnosis of kernicterus in humans.^18–21 A MRI was performed in this case to help rule out other potential causes of rapid decrease in neurologic state and coma, including tumors, meningitis, meningoencephalitis, thromboembolism, and hemorrhage. Abnormal signal was identified in areas that corresponded to bilirubin deposition, and neuronal damage observed during the postmortem examination.

Signal changes seen on T2-weighted scans were consistent with those described in humans exhibiting kernicterus.¹⁸ The specific cause of the high signal intensity of the lesions is unknown, but areas affected have been shown to be sites of either preferential neuronal damage or necrosis secondary to deposition of B_u. The localization of changes in the dog reported here is similar, but not identical, to that seen in humans and also differs slightly from changes previously described in one dog on postmortem examination.⁷ In humans, the most characteristic imaging finding in cases of kernicterus is hyperintensity of the globus pallidus on T2 and FLAIR sequences. Similar signal hyperintensity may also be seen in the subthalamic nuclei, thalamus, putamen, hippocampus, and cranial nerve nuclei (especially the third, fourth, and sixth nuclei).¹⁸ The cerebral cortex and white matter are most often relatively spared. In this case, significant changes were seen in cerebral cortical gray matter, but there were apparently either minimal or no changes in the hippocampus and cranial nerve nuclei. Signal hyperintensity on T1 sequences, such as that identified in the caudate nuclei in the dog described in this report, has been reported as a feature of acute kernicterus in human neonates and would be unusual for several of the other initial differential diagnoses, such as hypoxic injury and seizure-induced injury.^18–20 Although characteristic MRI findings of kernicterus in humans have been described, it should be noted that there are still questions about specificity and sensitivity of MRI for diagnosis.¹⁹

In one previously reported case of kernicterus in the dog, bilaterally symmetrical yellow foci were identified grossly only in the thalamic and subthalamic nuclei.⁷ Changes involving cerebral cortical gray matter and basal nuclei, such as those seen in this case, were not mentioned. The reason for different predilection sites for bilirubin deposition in the case reported compared with both the previously reported canine case and the human literature is not known.

It is important to note that although the dog described here did have kernicterus, the authors cannot be certain that this was the proximate cause of either the observed neurologic signs or the MRI abnormalities. It is not possible to rule out the possibility that other known or potential problems such as hypoxia, thromboembolism, acidosis, seizure-induced injury, or hyperammonemia may have either caused or contributed to the clinical signs in this dog. Likewise, although the histopathologic and MRI findings are similar to those described in humans, the presence of other systemic abnormalities makes it impossible to prove that the observed kernicterus was the primary cause of those abnormalities.

Conclusion

This case provides additional evidence that kernicterus can occur in adult dogs with increased serum bilirubin concentrations. Although the incidence of kernicterus in hyperbilirubinemic dogs remains unknown, it is possible that it has been underreported due to the difficulty in recognizing more subtle signs, attribution of more obvious signs to other concurrent disease processes (e.g., hepatic encephalopathy, hypoxia, hemostatic disorders, etc.), and the lack of antemortem diagnostic testing capable of diagnosing the condition. The MRI finding of bilaterally symmetrical T2-weighted and FLAIR hyperintense regions within the thalamus, cerebral cortical gray matter, and basal nuclei correlated well with gross and histopathologic lesions of kernicterus identified at necropsy in this dog. There may be a future role for the use of MRI in the diagnosis of kernicterus in canine patients, although further studies are required to define sensitivity and specificity.