• Open Access

Reversible Encephalopathy Secondary to Thiamine Deficiency in 3 Cats Ingesting Commercial Diets


Corresponding author: S. Marks, Department of Medicine and Epidemiology, School of Veterinary Medicine, University of California, Davis, One Shields Avenue, Davis, CA 95616; e-mail: slmarks@ucdavis.edu.


Association of American Feed Control Officials


body condition score


fluid-attenuated inversion recovery


high-performance liquid chromatography






thiamine diphosphate


Wernicke's encephalopathy

Case 1

A 14-year-old male castrated domestic shorthair cat was examined because of an acute onset of vomiting, inappetence, and ataxia in all 4 limbs of 2 days duration. There was no history of any exposure to known toxins or trauma and the cat had no medical problems before this examination. The cat was fed a diet of commercially available but unbalanced canned cat food (beef, turkey, or chicken based) mixed with a dry cat food formulated for feline maintenance (chicken based) for the previous 9 months. The CBC and serum biochemistry profile did not reveal important abnormalities. The cat was 5% dehydrated, in poor body condition score (BCS 2/9), and the rectal temperature was 97.0°F (36.1°C). The cat was moderately obtunded, weakly ambulatory, tetraparetic with ataxia in all 4 limbs, and had a wide-based stance. Proprioceptive positioning was absent in all 4 limbs. Menace response was inconsistent in both eyes. Fundic examination, thoracic radiographs, and abdominal ultrasound did not reveal abnormalities. The neuroanatomical localization was multifocal intracranial disease.

The cat had a generalized seizure the day after initial examination and was treated with diazepama (1.6 mg/kg IV once) followed by phenobarbitalb (5 mg/kg IV q12h). The cat became severely obtunded, had an absent menace in both eyes, and was nonambulatory; the phenobarbital dosage was reduced to 3 mg/kg IV q12h and dexamethasone sodium phosphatec was administered at 0.1 mg/kg IV q24h. Magnetic resonance (MR) imagingd of the brain followed by cisternal cerebrospinal fluid (CSF) collection and analysis was done on the 3rd day of hospitalization. MR imaging revealed bilaterally symmetrical well-demarcated hyperintense lesions on T2-weighted (T2W), T2*, and fluid-attenuated inversion recovery (FLAIR) images involving the cerebral cortex (parietal, occipital, hippocampal lobe), the subcortical white matter, thalamus, tectum, dorsal, and ventral medulla. There was enhancement of the cerebral cortical lesions as well as faint enhancement of the thalamic and dorsal medullary lesions after IV administration of contraste (Fig 1). Analysis of CSF collected from the cisterna magna revealed a mildly increased protein concentration (29 mg/dL; N < 25 mg/dL). The MR and CSF findings were most consistent with a metabolic or toxic encephalopathy. Another potential but much less likely differential for the bilateral symmetrical lesions on MRI involving the cerebrum and brain stem was a neurodegenerative disorder such as a mitochondrial encephalopathy.

Figure 1.

 Transverse magnetic resonance images, case 1. (A) Hyperintensity in the cerebrum and caudal colliculi (arrows). (B, C) Mild cerebral contrast enhancement (arrow). T1W, T1-weighted; T2W, T2-weighted; A, T2W; B, precontrast T1W; C, postcontrast T1W.

Approximately 3 mL of whole blood was collected in sodium heparin tubes and then immediately transferred to an amber transport tube to protect the sample from light for thiamine analysis. The specimen was frozen at −20° overnight and then dispatched by courier on ice packs to keep the specimens frozen.

Whole blood thiamine concentrationf determined via high-performance liquid chromatography (HPLC) was 1.0 μg/dL, whereas thiamine concentrations in 3 healthy control cats on a complete and balanced diet were 4.2, 4.3, and 5.6 μg/dL. Thiamine concentrations were determined in the 3 healthy cats because reference ranges are currently not available for healthy cats. The cat was treated with 100 mg thiamineg administered IM q24h. Phenobarbitalb was continued and lactated Ringer's solution (LRS) was administered IV. The cat was neurologically stable after 2 days, but remained obtunded. The phenobarbitalb dose was reduced to 1.5 mg/kg IV q12h for an additional 3 days and the cat was administered thiamineh (100 mg PO q24h) when it was eating and drinking on day 5. The cat was discharged on day 8 on PO administered thiamineh and a balanced commercial diet that had passed Association of American Feed Control Officials (AAFCO) feeding trials for all life stages. Repeat examination 10 days after discharge revealed mild obtundation and mild generalized ataxia with decreased proprioceptive positioning in all 4 limbs. Thiamine supplementation was discontinued and the cat was weaned off phenobarbitalb over 2 weeks. The cat was clinically normal at repeat examination 4 months after initial presentation, and whole blood thiamine concentration at that time was similar to healthy control cat values at 4.5 μg/dL.

Case 2

An 11-year-old FS domestic medium hair cat was presented to the VMTH with an acute onset of seizure activity over the preceding 24 hours. The medical history was otherwise unremarkable. The cat's diet for an unspecified period of time was 2 flavors of a commercially available canned diet (fish or beef/liver based) that had passed AAFCO feeding trials for feline growth and maintenance. The cat's physical examination was unremarkable except that it was obtunded and circled to the right with a right head tilt. Mild generalized ataxia with truncal sway, mild tetraparesis, decreased menace in both eyes, and slow physiologic nystagmus and vertical nystagmus in dorsal recumbency were present. Proprioceptive positioning was absent in all 4 limbs and segmental reflexes were all normal. Fundic examination, CBC, serum biochemistry profile, total T4 concentration, urinalysis, and thoracic radiographs did not reveal important abnormalities, with the exception of leukocytosis of 18,870/μL (reference range, 4,500–14,000/μL) secondary to lymphocytosis of 10,567/μL (reference range, 1,000–7,000/μL). In addition, indirect systolic and diastolic blood pressure measurements were normal.i Neuroanatomical localization was multifocal intracranial disease.

The cat was treated with phenobarbital sodiumb (3 mg/kg IV q12h) and LRS IV; neurological status was stable for 2 days, when she had 2 generalized seizures and became severely obtunded. Mannitolj (0.5 g/kg IV over 20 minutes) was administered with no change in mentation. The phenobarbitalb dose was increased to 5 mg/kg IV q12h. On day 3 of hospitalization MR imagingd of the brain was done, followed by cisternal CSF collection and analysis. T2W and FLAIR MR images revealed bilaterally well-demarcated symmetric hyperintense lesions in the cerebral cortex (parietal, occipital, hippocampal lobe), lateral geniculate nucleus, tectum, dorsal, and ventral medulla (Figs 2 and 3). The dorsal and ventral medullary T2W lesions were mildly hyperintense on T1-weighted (T1W) images. The hyperintense lesions on T1W and T2W images enhanced with IV administration of contraste (Fig 3). CSF analysis was within reference ranges except for a protein concentration of 39 mg/dL (N < 25 mg/dL). The MR abnormalities were most consistent with a metabolic or toxic disorder.

Figure 2.

 (A) Hyperintensity in the lateral geniculate nucleus (arrow), and cerebral cortex (arrowhead). (B) Hyperintensity in the hippocampus (arrowhead) and cerebral cortex (arrow). T2W, T2-weighted; A and B, Transverse T2W images, case 2.

Figure 3.

 Transverse magnetic resonance images postcontrast from case 2. (A) Hyperintensity in the dorsal medulla (region of the vestibular nuclei) and ventral medulla (region of the facial nuclei) (arrow). (B, C) Contrast enhancement of the dorsal and ventral brainstem lesions (arrow). T1W, T1-weighted; T2W, T2-weighted; A, T2W; B, precontrast T1W; C, postcontrast T1W.

Diet samples provided by the owner were submitted for thiamine analysis,k and whole blood was collected in a sodium heparin tube and shipped in an amber transport tube for measurement of blood thiamine concentration.f Whole blood thiamine concentration was 0.6 μg/dL, well below the concentrations documented in 3 previously sampled healthy control cats. The thiamine content of the 4 samples of the 2 diet varieties ranged from 0.45 to 1.34 mg/kg dry matter (DM), which is 9–27% of the AAFCO minimum (5 mg/kg DM).1 The cat was treated with thiamineg at 100 mg IM q12h and continued on phenobarbitalb (5 mg/kg PO q12h for 4 days, then 2.5 mg PO q12h), and switched to a balanced commercially available diet formulated for feline maintenance. No further seizures were noted during hospitalization, and the cat's mentation and tetraparesis progressively improved. The cat's thiamineh dose was reduced after 4 days to 20 mg PO q24h for 14 days after which the cat was lost to further follow-up.

Case 3

An 11-year-old FS domestic shorthair cat was examined because of acute onset of abnormal behavior for 3 days before presentation. The cat had been vocalizing at night, and on the day of presentation became nonambulatory and agitated. The cat's diet for the previous 2 years was a commercially available canned fish-based diet formulated for all life stages of cats. A CBC, serum biochemistry profile, total T4 concentration, ECG, fecal flotation, FeLV/FIV serology, and thoracic radiographs performed by the referring veterinarian did not reveal important abnormalities with the exception of hyperglycemia (blood glucose 297 mg/dL [reference range, 70–150 mg/dL]). The cat had been administered butorphanoll (0.5 mg/kg SC once) and ampicillinm (22 mg/kg IV once) before referral.

On examination the cat was hypothermic at 96.5°F (35.8°C) and had a BCS of 2/9. She was anxious, alert, and hyperesthetic to stimuli. She remained nonambulatory but sternal with a right head tilt and would fall to the right with severe generalized ataxia when supported. She had an absent menace response in the right eye with normal pupillary light reflexes in both eyes. Postural responses and spinal reflexes were normal. Fundic examination was normal. Neuroanatomical localization was multifocal intracranial disease. The cat was treated with buprenorphinen (0.01 mg/kg IV q6h), mannitolj (1.0 g/kg IV q8h given over 20 minutes), dexamethasone sodium phosphatec (0.15 mg/kg IV once), LRS IV, and maintained on a heating pad.

The following day the cat's cranial nerve exam was normal, and she was ambulatory with severe generalized ataxia. MR imagingo revealed well defined, bilaterally symmetric, round, hyperintense foci involving the rostral colliculi, lateral geniculate nuclei, and thalamic nuclei on the T2W images. There was no contrast enhancement.p CSF was not collected. The abnormalities on MR imaging were most consistent with a metabolic or toxic etiology. Whole blood was collected in a sodium heparin tube, centrifuged, and the plasma was submitted for thiamine analysis in an amber transport tube.f Diet samples were not available. The plasma thiamine concentration on presentation was <0.06 μg/dL. A 6-year-old spayed female healthy cat's plasma thiamine concentration evaluated at the same time was 0.92 μg/dL. The cat was treated with thiamineg (100 mg SC once, then 25 mg IM q12h for 5 days, then 25 mg PO q12h). Other therapies were discontinued. The cat's diet was changed to a commercially available canned diet formulated for all life stages. The cat achieved a complete clinical recovery within 3 weeks and the plasma thiamine concentration at that time was 1.34 μg/dL.


Thiamine deficiency in humans results in brain lesions within 2–3 weeks that are restricted to selective, vulnerable regions with a high thiamine content and turnover such as the mesencephalic tegmentum, mamillary bodies, and medial thalami.2 Cats are susceptible to thiamine deficiency because of their high-dietary requirement for thiamine, and because fish-based diets that contain thiaminases before processing are often fed to cats.3 In people, thiamine requirement is directly related to both total caloric intake and the proportion of calories provided as carbohydrates.4

Two of 3 cats in this series were ingesting commercially available canned diets that have not been associated with any diet recalls to date, and all cats manifested acute neurological signs consistent with multifocal intracranial disease. All of the published reports to date of thiamine deficiency in cats eating commercially available diets have involved canned foods.3,5–7 Canned foods are more susceptible to thiamine loss because of the high temperature involved in the processing of these diets, in particular when the pH is above 5.8,9 Further, one cat was consuming an unknown proportion of his daily calories as a commercially available but unbalanced diet without additional thiamine. The ingredient list for all 3 varieties of this diet included beef, turkey, or chicken, and a source of soluble fiber, and water. As required by law, the product carried a Nutritional Adequacy Statement stating, “this product is intended for intermittent or supplemental feeding only.” The labeling of this unbalanced diet underscores the importance of veterinarians educating pet owners regarding commercial diet choices and the assessment of the Nutritional Adequacy Statement on a pet food label.

The primary sources of thiamine supplementation in commercially available pet foods are synthetic (thiamine mononitrate and thiamine hydrochloride), and these sources are more susceptible to destruction than the thiamine present in plant and animal tissues.1 Thiamine can be destroyed by sulfate trace minerals and sulfite preserved meats which cleave the thiamine at the methylene bridge.10 In addition, thiamine is oxidized by ultraviolet and gamma irradiation11 and degraded by thiaminase enzyme activity found predominantly in shellfish, fish viscera, and some bacteria.3,8 Thiaminase is destroyed by heat processing; however, if raw ingredients are not promptly or properly cooked, destruction could result in lower than expected concentrations of thiamine. Given there are no real safety concerns regarding increased dosages of thiamine in either the cat or dog, and processing losses can exceed 90%,1 fortification in diet premixes warrants careful consideration to ensure appropriate thiamine concentrations in the final product.

The neurological manifestations of thiamine deficiency observed in the 3 cats are consistent with those documented previously. Additionally, impaired vision, mydriasis, and ventroflexion of the neck occur in thiamine-deficient cats.3,7 Seizures can progress to coma and death if the thiamine deficiency is not treated. These signs are often preceded by anorexia and vomiting, which can exacerbate the thiamine deficiency.3,7 The progressive encephalopathy has been well documented in Wernicke's encephalopathy (WE), the acute neuropsychiatric syndrome, and human counterpart of thiamine deficiency.12 MR imaging is a powerful tool used to support the diagnosis of acute WE.13 Disruption of transmembrane osmotic gradients results in cerebral edema and ischemia.14 Similar MR imaging features have been reported in a cat and included hyperintensity of the lateral geniculate nuclei, caudal colliculi, periaqueductal gray matter, medial vestibular nuclei, cerebellar nodulus, and facial nuclei on T2W and FLAIR images bilaterally, with a subtle hyperintensity on T1W images; contrast enhancement was not noted.7 In another case report, bilateral hyperintensities occurred in the lateral geniculate nuclei, caudal colliculi, facial nuclei, and medial vestibular nuclei on T2W and FLAIR images. These lesions were hypointense on T1W images, and did not enhance with contrast.15 All of the cats in the present report had bilaterally symmetrical T2W and FLAIR hyperintensity in the brainstem, and 2 of the 3 cats also had changes in the cerebral cortex, findings that have not been documented previously in thiamine-deficient cats. Two of the 3 cats also had contrast enhancement. In people with thiamine deficiency, MR imaging has been used to follow the resolution of lesions following thiamine supplementation,16 and a recent case report documented the same phenomenon in a cat with suspected thiamine deficiency.15 Reversible MR imaging lesions have been described in people and dogs after seizures, underscoring the importance of repeating imaging after seizure control to help differentiate between seizure-induced changes and primary multifocal parenchymal abnormalities.17,18 The cases reported in this series responded favorably to appropriate supportive therapy and thiamine supplementation. Although MR was not repeated, clinical improvement was noted in all cases, and reevaluation of whole blood and plasma thiamine concentrations in 2 cats after supplementation revealed thiamine concentrations in the same range as in 3 healthy cats. This case series highlights the utilization of direct measurement of thiamine in feline blood by HPLC for confirming thiamine deficiency. This thiamine assay is now commercially available in many countries and has replaced the erythrocyte transketolase activity assay because of its superior sensitivity and specificity for thiamine status.19 Whole blood is the preferred specimen for thiamine assessment, and approximately 80% of thiamine present in whole blood is found in red blood cells. The concentration of thiamine diphosphate (TDP), the primary active form of vitamin B1, is measured in this assay. Approximately 90% of vitamin B1 present in whole blood is TDP. Thiamine and thiamine monophosphate comprise the remaining 10% and are not measured with the HPLC assay. In contrast to whole blood, thiamine and thiamine monophosphate are measured in plasma samples, and the biologically active form of vitamin B1, TDP, is not found in measurable concentrations in plasma. Plasma thiamine concentrations are reflective of recent dietary intake, and do not reflect body stores.f It would have been optimal to have measured whole blood thiamine concentrations in the 3rd cat to obtain a more accurate assessment of TDP and a better indication of body stores of thiamine in the cat. Measurement of plasma thiamine concentration is inferior for assessment of overall thiamine status, and is typically reserved for assessing dietary compliance in people who are being supplemented with thiamine following diagnosis of thiamine deficiency.f

In summary, thiamine deficiency is a readily reversible neurological disorder that warrants consideration in any cat manifesting compatible clinical and neurological signs. The differential diagnoses for the multifocal intracranial disease suspected in all cats based on their neurological examinations included toxic and metabolic disease. It is noteworthy that thiamine deficiency was only strongly considered once the MR images were reviewed, and these findings underscore the importance of considering thiamine deficiency in seizuring cats or those manifesting multifocal intracranial signs. This study was limited by the relatively short follow-up in 1 of the 3 cats and our inability to repeat MRIs to follow resolution of the cats' intracranial lesions. Veterinarians should be cognizant that thiamine deficiency is not necessarily associated with the ingestion of diets associated with pet food recalls or large outbreaks, as the diets that were fed to 2 of the 3 cats in this series were commercially available diets formulated for maintenance that had not been associated with any diet recalls to date. Foods should also be evaluated for their adequacy as the sole or primary diet, as many unbalanced canned products are widely available and are marketed and labeled similarly to complete diets. Further studies assessing thiamine content of canned feline diets are warranted.

The diet histories in all cats were completed once the diagnosis of thiamine deficiency was suspected based on the MR images, further illustrating the importance of obtaining a comprehensive dietary history proactively in all sick animals. Given the safety of thiamine supplementation, regardless of the route of administration, consideration for empiric therapy is warranted in cats with compatible neurological manifestations, particularly when a delay in treatment may cause irreversible brain damage and is potentially life-threatening.13


a Valium, Hospira Inc, Lake Forest, IL

b Phenobarbital sodium injection, Baxter Healthcare Corporation, Deerfield, IL

c Dexium-SP, Bimeda Inc, Riverside, MO

d 1.5-T scanner, General Electric Signa LX, Milwaukee, WI

e Magnevist, Gadopentate dimeglumine 469.01 mg/mL, Berlex Laboratories, Wayne, NJ

f ARUP Laboratories, Salt Lake City, UT

g Thiamine HCl injection, Vedco Inc, St Joseph, MO

h Thiamine monophosphate tablets (100 mg), Nature's Bounty, Bohemia, NY

i Cardell Model 9401 Veterinary Diagnostic Monitor, Midmark Corp, Versailles, OH

j Mannitol injection USP, Hospira Inc

k Eurofins Scientific Inc, Des Moines, IA

l Torbugesic, Fort Dodge, IA

m Omnipen injection, Novaplus, Princeton, NJ

n Buprenex, Reckitt Benckiser Pharmaceuticals Inc, Richmond, VA

o 1.0 T Siemens Magnetom Expert, Erlangen, Germany

p Omniscan, 238.2 mg gadodiamide, GE Healthcare, Milwaukee, WI


The authors thank Drs Liran Tzipory, Sophie Petersen, Patrick Kenny, and Daniel York for their help with two of the cases seen at the UC Davis VMTH. The authors are grateful for John Doval's assistance in editing the figures for this manuscript.