Abstract
Three adult Chihuahuas were presented for evaluation after smoke inhalation during a house fire. All three dogs received supportive care and supplemental oxygen. After initial improvement, the dogs developed seizures. Despite anticonvulsant therapy and supportive care, the dogs died. The brains of two dogs were examined. Lesions were identified that were compatible with acute carbon monoxide (CO) toxicity. Lesions were confined to the caudate nucleus, the globus pallidus, and the substantia nigra bilaterally, as well as the cerebellum, cerebral cortex, and dorsal thalamus. This case report describes the clinicopathological sequelae in acute CO toxicity.
Introduction
A variety of toxic inhalants are produced in house fires. Among these are carbon monoxide (CO), carbon dioxide, and hydrogen cyanide. While all of these toxins are likely involved in the pathogenesis of smoke inhalation, CO toxicity plays a prominent role. Carbon monoxide is produced by the combustion of organic materials. It readily crosses the respiratory epithelium where it preferentially and competitively binds to hemoglobin (Hb), forming carboxyhemoglobin (COHb). Carbon monoxide has a 200- to 230-fold greater affinity for Hb than oxygen (O2).1 As a result of CO binding, the O2-carrying capacity of Hb is reduced, leading to hypoxemia and hypoxia. Not only is O2 displaced when CO is bound to Hb, but also the Hb molecule undergoes a conformational change that results in a greater affinity for O2 at the remaining heme sites, leading to a shift to the left of the O2-dissociation curve.2 This O2-dissociation curve’s left shift further impairs O2 delivery to the tissue.2
Although smoke inhalation commonly affects the respiratory system, central nervous system (CNS) disturbance can also develop. The CNS signs can be divided into acute and delayed toxicity. The veterinary literature contains few reports regarding the neurological consequences of smoke inhalation.3–6 The following case report details the clinicopathological sequelae of acute CO toxicity secondary to smoke inhalation in three Chihuahuas and the histopathological findings in two of these dogs.
Case Report
Three adult Chihuahuas were presented to The University of Georgia for smoke inhalation approximately 12 hours after exposure to smoke during a house fire. The dogs were removed from the house after 30 minutes of exposure in a smoke-filled room. Except for case no. 3, the other dogs had no prior health problems. The owner stated that case no. 3 had an abnormal gait throughout its life.
Immediately after removal from the house fire, the dogs were taken to an emergency veterinary clinic. At the emergency clinic, only one dog (case no. 1) was ambulatory. The other two dogs were presented in lateral recumbency. All three dogs were described as depressed with cyanotic mucous membranes. No abnormalities were noted on auscultation of the thoraces, and none of the dogs exhibited evidence of cutaneous burns. All three dogs were treated identically. Each received approximately 90 mL/kg of crystalloidsa intravenously (IV) over approximately 1 hour, followed by 2.5 mL/kg per hour of crystalloids. All three dogs were given supplemental O2 support via an O2 cage. Within hours of initial therapy, all three dogs demonstrated improvements in their mental states. The mentation of case no. 2 was normal, while the other two dogs remained depressed. The two dogs that were presented in lateral recumbency had regained the ability to ambulate. Approximately 9 hours after presentation, all three dogs were transferred to the teaching hospital at The University of Georgia’s College of Veterinary Medicine.
Case No. 1
On admission to The University of Georgia, case no. 1 (1.9-kg, intact male) was obtunded but responsive and able to walk unassisted. On physical examination, abnormalities were limited to obtundation and ocular lesions. Respiratory rate was considered slightly elevated at 40 breaths per minute (BPM). Corneal ulcerations, conjunctival hyperemia, blepharospasm, and miosis were noted bilaterally. Neurological examination was normal except for obtundation. A complete blood count (CBC) and serum biochemical profile were performed. Abnormalities observed on the CBC included mild thrombocytosis (787 × 103 cells/μL, reference range 211 to 621 × 103 cells/μL) and neutrophilic leukocytosis (white blood cell [WBC] count 21.5 × 103 cells/μL, reference× range 5.1 to 13.0 × 103 cells/μL; neutrophils 19.350 103 cells/μL, reference range 2.9 to 12.0 × 103 cells/μL). Abnormalities observed on the serum biochemical profile included elevated alanine aminotransferase (246 U/L, reference range 12 to 108 U/L), hyperglycemia (149 mg/dL, reference range 77 to 120 mg/dL), and hypocholesterolemia (125 mg/dL, reference range 129 to 264 mg/dL). Thoracic radiographs were normal.
The dog was treated with IV fluid therapya with supplemental potassium chloride (KCl, 16 mEq/L) at 4.2 mL/kg per hour. Terbutalineb (0.01 mg/kg) was administered subcutaneously q 6 hours. The dog’s corneas were treated presumptively for exposure keratitis. Ophthalmic therapy consisted of atropine ointmentc (1 drop [1 gtt] q 24 hours) and sodium hyaluronated (1 gtt q 6 hours) applied topically to both eyes. Additionally, cyclosporine ointmente was applied topically q 12 hours to both eyes. The dog was placed in an O2 cage that supplied 40% fraction of inspired oxygen (FiO2). After 24 hours, the dog was alert and behaving normally. Respiratory rate had decreased to 24 BPM, which prompted discontinuation of the supplemental O2. The dog began to eat and drink normally.
Approximately 48 hours after admission, the dog had an acute episode of status epilepticus. Treatment was initiated with diazepamf (0.25 mg/kg IV once, followed by 0.75 mg/kg IV once). The seizure activity persisted, so phenobarbitalg was administered (20 mg/kg IV given in 5 mg/kg boluses q 1 to 2 minutes for a total of four doses). Despite phenobarbital administration, the seizure activity continued. Consequently, propofolh was administered (one 3.5 mg/kg IV bolus, followed by 0.1 mg/kg per minute IV as a constant-rate infusion [CRI]), which resulted in cessation of status epilepticus. Respiratory rate was 80 BPM after the episode of status epilepticus; therefore, the dog was returned to the O2 cage, which provided supplemental O2 to a level of 40% FiO2. While in the O2 cage, the dog’s respiratory rate returned to normal. Generalized seizures continued to occur approximately q 6 hours. Given the possibility of cerebral edema, mannitoli (1 g/kg IV over 10 minutes once) followed by furosemidej (3 mg/kg IV q 4 hours for three doses) was administered. Generalized seizures continued to occur throughout the next 24 hours. Each episode was controlled with propofol boluses (2.0 to 3.5 mg/kg IV bolus) that were administered in addition to the CRI of propofol. Phenobarbital (2.5 mg/kg IV q 12 hours) administration was also continued.
Approximately 4 days after presentation, a nasoesophageal feeding tube (3.5 French) was placed to provide enteral nutrition.k Thoracic radiographs were performed to ensure the correct placement of the feeding tube. A mild increased interstitial pattern was observed diffusely in the lungs compared to the initial radiographs obtained at the time of admission. Over the course of the day, generalized seizures continued to occur, and the frequency increased to approximately one seizure per hour. The CRI of propofol was increased to 0.15 mg/kg per minute. Additionally, each seizure was treated with a bolus of propofol at dosages ranging from 2 to 4 mg/kg IV. The dog’s respiratory rate and effort began to increase. At this time, packed cell volume was 70% (reference range 35% to 55%) with normal total solids at 7.6 g/dL (reference range 5.2 to 7.8 g/dL), which was consistent with hemoconcentration presumably because of diuretic therapy. The rate of IV fluid administration was increased to 5.7 mL/kg per hour to correct the deficit. Measurement of blood glucose documented hypoglycemia (45 mg/dL, reference range 80 to 120 mg/dL). Although this value may have been artifactually low because of hemoconcentration, supplemental dextrosel was added to the IV fluids to attain a 2.5% concentration. On day 5, the dog suffered cardiopulmonary arrest and died. Only the brain was available for postmortem evaluation.
Case No. 2
Case no. 2 (1.75-kg, spayed female) had a similar presentation and clinical course as case no. 1. On admission, the physical examination was normal, except the dog held its left thoracic limb in a flexed position. This dog also had ocular abnormalities similar to case no. 1. Neurological examination was normal except for a depressed mental state and the continually flexed left thoracic limb. A CBC and serum biochemical profile were performed. Abnormalities observed on the CBC included a neutrophilic leukocytosis, left shift, and monocytosis (WBC 28.5 × 103 cells/μL, reference range 5.1 to 13.0 × 103 cells/μL; neutrophils 24.510 × 103 cells/μL, reference range 2.9 to 12.0 × 103 cells/μL; band 1.140 × 103 cells/μL, reference range 0.0 to 0.45 × 103 cells/μL; monocytes 2.850 × 103 cells/μL, reference range 0.1 to 1.4 × 103 cells/μL). Abnormalities observed on the serum biochemical profile included hyperproteinemia (8.1 g/dL, reference range 5.2 to 7.3 g/dL) and hyperalbuminemia (4.7 g/dL, reference range 2.5 to 4.2 g/dL). The dog was also hypernatremic (169 mmol/L, reference range 146 to 154 mmol/L), hyperchloremic (126 mmol/L, reference range 107 to 125 mmol/L), and hypokalemic (3.8 mmol/L, reference range 3.9 to 5.0 mmol/L). Alkaline phosphatase was also increased (207 U/L, reference range 13 to 122 U/L). Thoracic radiographs were normal.
The dog received similar treatments as case no. 1. Intravenous fluid therapym (0.45% sodium chloride [NaCl] and 2.5% dextrose) was administered and supplemented with 16 mEq KCl/L at 4.5 mL/kg per hour. The dog was placed in an O2 cage that provided supplemental O2 to a level of 40% FiO2. Additionally, the dog received the same ocular therapy as case no. 1, with the exception of cyclosporine. Forty-eight hours after admission, O2 supplementation was discontinued. The dog was alert and active and began to eat and drink. A repeat serum biochemical profile was normal.
Three days after admission, the dog began to have generalized seizures on an hourly basis. Each seizure was treated with diazepam (0.8 mg/kg IV). After four generalized seizures, phenobarbital (4.5 mg/kg IV q 12 hours) was administered, and a CRI of diazepam (0.5 mg/kg per hour) was initiated. Generalized seizures continued to occur over the subsequent 4 hours. Consequently, the diazepam CRI was discontinued, and a propofol CRI (0.1 mg/kg per minute IV) was instituted. Generalized seizures continued to occur, and the dog was also treated with a bolus of propofol (1.7 mg/kg IV) for each seizure. The propofol CRI was increased to 0.15 mg/kg per minute, and the seizure activity stopped.
Four days after admission, pulmonary crackles were auscultated bilaterally. Radiographs of the thoracic cavity revealed a diffuse alveolar pattern in all lung fields. Given the diffuse distribution, the radiographic pattern was most consistent with noncardiogenic pulmonary edema; however, cardiogenic pulmonary edema or aspiration pneumonia could not be excluded from consideration. Consequently, furosemide (2 mg/kg IV q 4 hours) was administered for possible cardiogenic pulmonary edema. The dog was returned to the O2 cage for supplemental O2 (40% FiO2); IV fluid therapy was discontinued; and furosemide (1 mg/kg IV) was administered on an hourly basis for 3 hours. Despite aggressive diuretic therapy, the dog died. Only the brain was available for postmortem evaluation.
Case No. 3
On admission, case no. 3 (1.12-kg, intact male) was presented in a generalized seizure. Blood glucose was 66 mg/dL (reference range 80 to 120 mg/dL). The dog was administered 0.5 mL of 50% dextrose IV and diazepam (1 mg/kg IV) to stop the seizure. Afterward the dog was severely obtunded, laterally recumbent, hypothermic (35.3°C), cyanotic, and dyspneic. Despite supportive treatment that included IV fluid therapy (supplemented with 16 mEq/L KCl and 2.5% dextrose) at 4.6 mL/kg per hour and supplemental O2 (40% FiO2), the dog remained tachypneic. Approximately 36 hours after admission, the dog suffered cardiopulmonary arrest and died. Necropsy was not performed.
On gross examination of the brain of case no. 1, the cerebrum had a slightly swollen appearance on the left side in the area of the parietal/temporal lobe that involved portions of the marginal, ectomarginal, and suprasylvian gyri. On transverse sections, a mild dilatation of the lateral ventricles in their curve through the occipital lobes was seen. The left lateral ventricle was about 10 mm wide, and the right lateral ventricle was about 4 mm wide. The asymmetry observed grossly in the size of cerebral hemispheres was likely related to the difference in size of the lateral ventricles. Although asymmetric, the differences in size of the ventricles were considered normal.7,8 The septum pellucidum was absent at the level of the middle of the diencephalon. The remainder of the ventricular system was normal.
No gross lesions were seen in the brain of case no. 2. On microscopic examination, the lesions were the same in both dogs but were more developed in case no. 1. The lesions in both cases were limited to specific gray matter areas. The neuronal sites that were bilaterally affected included the caudate nucleus, globus pallidus, and substantia nigra. A small dorsal area in the thalamus was also affected. The putamen and claustrum were bilaterally normal. The lesions primarily consisted of reactive capillaries based on the prominence of their endothelial cells [Figure 1⇓]. The endothelial cells were more numerous and larger than normal [Figure 2⇓]. Occasionally a few lymphocytes were in the perivascular space surrounding the blood vessels. The adjacent neuronal cell bodies occasionally showed an acute, ischemic-type degeneration. Cytotoxic edema was minimal. Similar neuronal lesions were scattered through the neocortex of all cerebral lobes in a patchy distribution. No lesions were in the pons or medulla. In the cerebellum were numerous folia with severe, acute, ischemic-type degeneration of the Purkinje neurons. Many Purkinje neurons were absent. The same reactive blood vessel lesion was seen in the molecular layer of the cortex of the cerebellum.
Reactive capillaries of the globus pallidus (arrows) adjacent to the internal capsule×(asterisk). Hematoxylin and eosin (H&E) stain, 400 .
Higher magnification of the reactive capillaries in Figure 1⇑. The panel on the left is a capillary in cross-section. The panel on the right is a capillary in longitudinal section. The reactive capillaries demonstrate the increased number and × larger size of the endothelial cells. Hematoxylin and eosin (H&E) stain, 500 oil magnification.
Discussion
Neurological signs of CO toxicity reported in dogs include ataxia, absent menace response, dull sensorium, loss of consciousness, head tremors, twitching, seizures, and dementia.3–8 The neurological sequelae of smoke inhalation can be divided into acute and delayed toxicity. Delayed toxicity involves a myelinopathy affecting the cerebral deep white matter, such as the centrum semiovale, periventricular white matter, and the corpus callosum symmetrically, in which demyelination with sparing of axons is present.9,10 The exact pathogenesis of these lesions is unknown. The pathogenesis has been suggested to involve excitatory amino acid neurotoxicity caused by activation of N-methyl-d-aspartate receptors on neurons, leading to the production of nitric oxide, which causes perivascular changes that result in neutrophil activation and sequestration.11 Additionally, the generation of reactive O2 species results in lipid peroxidation that may alter the conformation of myelin basic protein, conferring immunoreactivity that leads to an inflammatory response.12
Like delayed toxicity, acute CO toxicity results in specific lesions in the nervous system. In humans, the most characteristic lesions involve neuronal necrosis of the pallidum, substantia nigra, cerebellum, and the hippocampus, as well as demyelination and necrosis of the deep white matter of the cerebral hemispheres, with relative sparing of the U-fibers.9 Similar lesions have been reported experimentally in monkeys, cats, and dogs.10,13,14 The lesions in case nos. 1 and 2 described in this report affected the pallidum, substantia nigra, and cerebellum, which suggested acute CO toxicity as the underlying etiology. While hypothetical, the lack of deep white matter pathology in the dogs of this report does not exclude acute CO toxicity. While speculative, the rapid decline of these dogs may not have allowed for development of white matter lesions. Had the dogs survived longer, perhaps white matter lesions would have developed. Given the same exposure to the house fire and similar clinical course, case no. 3 may have had similar lesions; however, without necropsy, this remains speculative.
In acute CO toxicity, neurological dysfunction typically occurs at the onset of intoxication; however, neurological dysfunction may develop after an initial recovery, as was observed in these dogs.10
Although widely acknowledged as the main effect of CO toxicity, hypoxia cannot entirely explain the pathological findings, because experimentally, CO can produce toxicity independent of COHb concentration.15 Acute CO toxicity has been related to a combination of hypoxia, hypotension, and cellular asphyxia from CO binding of mitochondrial cytochrome c and reperfusion injury, leading to lipid peroxidation by free radicals.16,17
The typical constellation of affected neuroanatomical regions does not seem to be related to an inherent sensitivity of these structures to CO, referred to as pathoclisis, but rather it reflects their blood supply.10 In experimental acute CO toxicity, lesions in the cerebral white matter and globus pallidus vary in occurrence and severity. These variations are based on a combination of CO level and degree of decreased cerebral blood flow as a result of systemic hypotension, rather than COHb concentration alone.15 Furthermore, in the face of hypotension, the globus pallidus has a greater decline in local blood flow than adjacent structures, such as the putamen and claustrum.10 This may be a reflection of the end arterial blood supply of the globus pallidus and the deep cerebral white matter, contrasted with a greater amount of vascular anastomosis supplying the U-fibers of the cerebral cortex, which ensures adequate perfusion, even in a hypotensive state.10 Consequently, the selective vulnerability of neurons of the globus pallidus, substantia nigra, and the deep cerebral white matter is likely related to the additive effect of hypoxia and hypotension.18
While not documented in the dogs of this report, hypotension might have played a role in the pathology. Carbon monoxide does have a direct effect on vascular tone, leading to vasodilatation and hypotension.19 In addition, CO toxicity is associated with cardiotoxicity, in which myocardial dysfunction leads to further reduction in systemic arterial pressure.20 In the dogs presented here, case no. 2 was hemoconcentrated on presentation, and case no. 1 developed hemoconcentration after diuretic administration, which may have contributed to hypotension despite the administration of fluid therapy.
Acute selective neuronal necrosis (i.e., ischemic cell change) can also occur with ischemia, hypoglycemia, and seizures.21 Unlike pan-necrosis in which both neurons and glial and vascular elements are affected, selective neuronal necrosis refers to necrosis of neurons with sparing of glial and vascular elements.22 Selective neuronal vulnerability refers to the susceptibility of specific neurons to undergo necrosis in a specific anatomical pattern.22 Neurons demonstrate selective vulnerability as a result of excitatory amino acid neurotoxicity related to excessive concentrations of glutamate.23 As with acute CO toxicity, excitatory amino acid neurotoxicity has been implicated in the pathogenesis of ischemia, hypoglycemia, and seizure-related neuronal necrosis.22 Despite the potential common pathogenesis involved in ischemia, hypoglycemia, and seizures, these conditions result in different patterns of selective neuronal vulnerabilities. In general, these conditions can affect laminae III, IV, and V of the cerebral cortex as well as the hippocampus, amygdala, some basal nuclei, and cerebellar Purkinje neurons.21 However, ischemia affects the middle laminae of the cerebral cortex, while hypoglycemia affects the superficial laminae.24 In the hippocampus, ischemia affects neurons in the cornu ammonis (CA) 4 and CA1 sectors, while hypoglycemia affects neurons in the dendate gyrus.24 Seizures can result in hippocampal pathology primarily affecting the CA3 pyramidal cells and sparing dentate granule cells.25–27 Lastly, ischemia and hypoglycemia can affect the caudate nucleus and putamen.24,28 Seizures can result in neuronal necrosis involving the cingulate gyrus, dorsomedial thalamus, and amygdala.29–34
The histological changes (i.e., reactive endothelial cells and selective neuronal necrosis) observed in the brains of the two dogs in this report were similar to those found with other etiologies; however, the topography of the lesions suggests that neuronal necrosis was a consequence of acute CO toxicity.
Conclusion
This is the first case report documenting histological brain lesions secondary to smoke inhalation in dogs. The lesions in the dogs of this report have been described in humans with acute CO toxicity.9 Of particular interest is the fact that the dogs described herein demonstrated initial clinical improvement, only to develop seizures within hours to days after smoke inhalation. A similar clinical course of improvement followed by deterioration has been observed experimentally with acute CO toxicity.10
Guidelines for treatment of humans with smoke inhalation have been defined.16,17,35,36 One of the key therapeutic interventions involves O2 support, occasionally employing hyperbaric O2 chambers.36,37 Oxygen therapy decreases the elimination half-life of CO. In humans breathing room air, the elimination half-life of CO is 320 minutes, whereas the elimination half-life of CO in humans breathing 100% O2 is only 80 minutes. In humans breathing O2 at 3 atmospheres absolute, the CO elimination half-life is 23 minutes.35 In veterinary patients, treatment is largely directed at supportive care, primarily through O2 support for the respiratory system.35,38–41 Specific therapeutic interventions designed to protect the CNS have not been elucidated in animals. Experimental evidence suggests that along with hypoxia, hypotension plays a role in the pathogenesis of CO toxicity.15,18 In the future, animals suffering from smoke inhalation should undergo blood pressure monitoring and receive treatment where applicable. In addition, to help protect them from CNS damage, animals should be given supplemental O2 and supportive care.
Footnotes
-
↵ a Lactated Ringer’s injection; Hospira, Lake Forest, IL 60045
-
↵ b Brethine; Novartis Pharmaceuticals Corp., East Hanover, NJ 07936
-
↵ c Atropine 1% ointment; Bausch and Lomb, Inc., Tampa, FL 33637
-
↵ d I-Drop Vet; I-MED Pharma, Inc., Montreal, Quebec, Canada H9B 3H7
-
↵ e Optimmune; Schering-Plough Animal Health, Kenilworth, NJ 07033
-
↵ f Valium; Roche Laboratories, Inc., Nutley, NJ 07110
-
↵ g Phenobarbital injection-generic; Elkins-Sinn, Inc., Cherry Hill, NJ 08003
-
↵ h Propofol; Abbott Laboratories, North Chicago, IL 60064
-
↵ i Mannitol generic; Abbott Laboratories, North Chicago, IL 60064
-
↵ j Lasix; Butler Animal Health Supply, Dublin, OH 43017
-
↵ k Clinicare; Abbott Laboratories, North Chicago, IL 60064
-
↵ l 50% dextrose; Hospira, Lake Forest, IL 60045
-
↵ m 0.45% NaCl; Hospira, Lake Forest, IL 60045
References
