Probable sepsis‐associated encephalopathy as a component of multiple organ dysfunction syndrome in a cat with pyothorax

Abstract

A 2‐year‐old male neutered domestic shorthair cat with bacterial pyothorax was referred to a tertiary hospital for treatment of sepsis. He met criteria for multiple organ dysfunction syndrome and developed new‐onset neurological dysfunction subsequent to the development of sepsis. Neurological changes included generalised seizures, coma and proprioceptive and cranial nerve deficits. The cat additionally had pulmonary, hepatobiliary and gastrointestinal dysfunction. Diagnostic findings included a normal brain MRI and a normal cerebrospinal fluid (CSF) total cell count (0.6 cells/μL, reference interval 0‐8 cells/μL) with a relative increase in neutrophils (35%) and mast cells (31%). Exclusion of other aetiologies and a consistent clinical progression prompted a diagnosis of probable sepsis‐associated encephalopathy (SAE). Treatment followed guideline recommendations for sepsis, including prompt broad‐spectrum intravenous antimicrobial therapy, source control, intravenous fluid therapy and vasoactive agents to normalise perfusion, mechanical ventilation, and support for other organ dysfunctions. Treatment for seizures included multiple anticonvulsant medications (midazolam, levetiracetam, propofol and phenobarbital). The cat had a good long‐term functional outcome; however, generalised seizures persisted up to 2 years post‐discharge. This is the first report of a syndrome of neurological dysfunction in a cat consistent with SAE in humans.

Sepsis‐associated encephalopathy (SAE) is a common organ dysfunction in human sepsis and is well described in the human literature. SAE is defined as diffuse cerebral dysfunction induced by the systemic response to a primary infection outside of the central nervous system (CNS), without evidence of secondary brain infection or other causes of encephalopathy. ¹ , ²

The pathophysiology of SAE is multifactorial and not fully understood. Reported mechanisms in humans include excessive microglial and astrocyte activation, endothelial cell degeneration, cerebral dysautoregulation, blood–brain barrier (BBB) dysfunction and tight‐junction protein loss. ³ , ⁴ , ⁵ , ⁶ The clinical presentation of SAE can involve cognitive dysfunction, consciousness impairment and seizures. ² , ⁶ , ⁷ , ⁸ The development of SAE is associated with increased mortality and may also result in long‐term cognitive impairment. ⁹ , ¹⁰

While multiple organ dysfunction syndrome (MODS) is known to occur in cats with sepsis, SAE is not clinically described. ¹¹ This report describes probable SAE in a cat with pyothorax.

Case

A 2‐year‐old indoor‐outdoor male neutered domestic shorthair cat weighing 4.2 kg was referred to an emergency clinic for evaluation of acute vomiting, and a 2‐week history of progressive lethargy. The cat had no history of seizures.

On physical examination, the cat was obtunded, tachypnoeic (respiratory rate 56/min) with increased respiratory effort, and approximately 10% dehydrated. In‐house haematology (Table 1) identified a marked neutropenia, monocytosis, lymphopenia and eosinopenia; however, the leukocyte differentials were not confirmed with a blood smear. In‐house biochemistry (Table 1) revealed mild hyperbilirubinaemia (24 μmol/L). Thoracic radiographs revealed pleural effusion. A unilateral thoracostomy tube was placed under local anaesthesia, and 135 mL of grossly purulent material was drained from the pleural space. A diagnosis of pyothorax was made. Sepsis was diagnosed based on evidence of infection and fulfilment of 2/4 criteria (tachypnoea and leukopenia) for the systemic inflammatory response syndrome (SIRS). ¹²

Table 1.

Haematological and biochemical results from a 2‐year‐old male neutered domestic shorthair septic cat with pyothorax prior to referral from the primary care veterinarian, at the emergency clinic, and after presentation to The Animal Hospital at Murdoch University (TAHMU)

Parameter	Units	Measured value, day relative to TAHMU presentation					Reference range
		−3 days	−1 day	Day 1	Day 7	Day 14
Haematocrit	L/L	0.37	0.33	0.25	0.28	N/A	0.24–0.45
White blood cells	109/L	8.06	26.60	13.4	26.5	N/A	5.5–19.5
Neutrophils	109/L	0.41	25.5 (granulocytes)	1.72	23.58	N/A	2.5–12.5
Band neutrophils	109/L	N/A	N/A	10.75	0	N/A	0.00–0.30
Lymphocytes	109/L	0.43	N/A	0.8	1.86	N/A	1.5–7.0
Monocytes	109/L	7.19	N/A	0	0.53	N/A	0.00–0.85
Eosinophils	109/L	0.01	N/A	0	0.53	N/A	0.00–1.50
Platelets	109/L	260	231	N/A* adequate	N/A* adequate	N/A	300–700
Creatine kinase (CK)	U/L	N/A	N/A	16,305	380	77	50–200
Aspartate aminotransferase (AST)	U/L	N/A	N/A	632	162	53	26–43
Alanine transaminase (ALT)	U/L	34	45	112	161	66	6–83
Alkaline phosphatase (ALP)	U/L	11	<10	11	21	13	25–93
Gamma‐glutamyl transferase (GGT)	U/L	16	N/A	1	1	<1	1–5
Total bilirubin (TBil)	μmol/L	24	85	84	6	6	2–17
Cholesterol	mmol/L	3.20	3.74	4.3	4.4	4.9	2.50–3.90
Glucose	mmol/L	7.44	7.83	6.9	11.2	4.7	3.8–6.1
Urea	mmol/L	11.7	5.1	4.6	3.7	4.0	7.0–10.7
Creatinine	μmol/L	63	56	49	54	81	70–159
Sodium (Na)	mmol/L	153	N/A	146	146	150	145–156
Potassium (K)	mmol/L	4.8	N/A	3.8	5.2	3.8	3.8–5.2
Total protein	g/L	74	61	40	62	77	54–78
Albumin	g/L	22	17	12	17	24	21–39
Globulin	g/L	52	45	28	45	53	25–50

He was transferred between the emergency clinic and primary care veterinarian for overnight and daytime care, respectively, for the next 3 days. Treatment included piperacillin‐tazobactam (48 mg/kg IV q8h), enrofloxacin (3 mg/kg SC q24h), methadone (0.1 mg/kg IV or IM q4h), supplemental oxygen, and intravenous fluid therapy (lactated Ringer's solution; LRS) for rehydration and maintenance. Fluid was drained from the thoracostomy tube q2h, with volumes ranging from 3 to 27 mL. Culture of the pleural effusion yielded growth of a Pasteurella sp. that was pansusceptible (ampicillin, amoxicillin‐clavulanic acid, doxycycline, enrofloxacin, marbofloxacin, sulphatrimethoprim).

Twenty‐four hours after presenting to the emergency clinic, he was noted to be twitching (further information not recorded), but this resolved over 5 min. Thirty‐six hours after presentation, he had a generalised tonic–clonic seizure which was treated with diazepam (1.2 mg/kg IV). Euglycaemia at the time of seizure was confirmed with a portable glucometer (5.8 mmol/L). Throughout the next 2 days, he had approximately 10 further generalised seizures, consistent with cluster seizures, treated with propofol or diazepam IV to effect. Abbreviated bloodwork (Table 1) was repeated on day 3 with a total bilirubin of 85 μmol/L, concurrent with the development of icteric mucous membranes.

Approximately 84 h after initial presentation he was referred to The Animal Hospital at Murdoch University (TAHMU). On presentation to TAHMU he weighed 3.7 kg, was comatose and hypothermic (36.7°C), with pale pink and icteric mucous membranes and a normal capillary refill time. His heart rate was 180/min, with muffled cardiac sounds, and normokinetic peripheral pulses. His respiratory rate was 24/min with bilaterally dull lung sounds. He was considered to have ongoing sepsis based on fulfilling 2/4 SIRS criteria (hypothermia, leukopenia) due to infection. An abbreviated neurological exam identified mydriasis, absent pupillary light reflexes, palpebral reflexes, menace reflexes, dazzle reflexes and facial sensation bilaterally. No resting nystagmus or strabismus was evident. The oculocephalic and gag reflexes were weak. This neurologic examination was performed more than 2 h after the last documented administration of propofol and more than 10 h since the last documented administration of diazepam. Owner consent was provided for treatment and hospital admission, which included consent for patient data use in research and publication.

Initial diagnostics included venous electrolytes and blood gas, haematology and serum biochemistry (Table 1). He had a mild nonregenerative anaemia, a moderate mature neutropenia, and a marked bandemia with moderate neutrophil toxic changes. His moderate hyperbilirubinaemia had persisted, and he had a marked hypoalbuminaemia resulting in a moderate hypoproteinaemia. Moderately increased muscle enzyme activities were also evident. Piperacillin‐tazobactam and enrofloxacin were discontinued and antibiotic therapy was changed to trimethoprim‐sulfadiazine (30 mg/kg IV q12h) (Trivetrin®, Jurox) for ongoing coverage of the Pasteurella sp., and clindamycin (11 mg/kg IV q12h) (Dalacin C® injection, Pfizer) for empirical coverage of Toxoplasma gondii.

Due to the comatose mentation and poor gag, endotracheal intubation was performed without need for general anaesthesia. Computed tomography (CT) of the thorax and head with IV contrast (Omnipaque, GE Healthcare) was performed shortly thereafter. This identified bilateral mild pleural effusion, bilateral mild pneumothorax, an alveolar pattern in the ventral lung fields most marked in the left cranial lobe, a bulla on the dorsal aspect of the left caudal lung lobe and diffuse pleural thickening with nodular regions. The sternal, cranial‐mediastinal and tracheobronchial lymph nodes were moderately enlarged and hypoattenuating, suggestive of abscessation or necrosis. The brain was normal.

Overnight he remained comatose and intubated but was spontaneously ventilating with normoxia (SpO₂ 95–99%, FiO₂ 0.4). Intermittent focal seizure activity continued despite levetiracetam (20 mg/kg IV q8h) (Levetiracetam IV, Apotex) and midazolam (0.3 mg/kg IV PRN) (Midazolam Viatris, Viatris). Analgesia was provided with buprenorphine (0.02 mg/kg IV q8h) (Ilium Temvet injection, Troy Animal Healthcare). Intermittent saline flushing and suction of the thoracostomy tube was commenced; however, worsening pneumothorax prompted conversion to continuous suction (Argyle Thora‐Seal III Chest Drainage Unit, Cardinal Health, Dublin, OH). Routine nursing care for a recumbent, intubated patient was performed.

On day 2, the cat was instrumented with a right jugular central venous catheter (MILA peel away small animal IV catheter 4Fr x 8 cm double lumen, MILA International, Florence, KY.), medial saphenous peripherally inserted central catheter (MILA guidewire small animal IV catheter 20ga x 12 cm single lumen, MILA International, Florence, KY.) and indwelling urinary catheter with closed collection system (SurgiVet Slippery Sam tomcat urethral catheter 3.5Fr x 11 cm, SurgiVet, Minneapolis, MN.). A brain MRI (SIGNA 1.5 T MRI scanner, GE HealthCare, Chicago, IL) was performed under IV sedation (medetomidine 0.5 μg/kg/h and midazolam 0.4 mg/kg/h) (Ilium medetomidine, Troy Animal Healthcare). The MRI sequences obtained included T1W transverse, T2W transverse and sagittal, T2* transverse and fluid attenuation inversion recovery (FLAIR) transverse. Additional T1W transverse, dorsal and sagittal images were obtained following IV gadolinium contrast (Gadovist, Bayer, Leverkusen, Germany). No abnormalities were detected (Figure 1).

Figure 1.

Midcortex T1W post IV contrast (top left), T2W (top right), T2*W (bottom left) and midsagittal T2W (bottom right) MRI images of a normal brain in a cat with sepsis and multiple organ dysfunction syndrome secondary to pyothorax.

Subsequently, a left lateral thoracotomy was performed under total intravenous anaesthesia (propofol 0.1–0.25 mg/kg/min, midazolam 0.2 mg/kg/h and medetomidine 0.5–1 μg/kg/h) (Provive 1%, Baxter) with analgesia consisting of 0.6 mg/kg of 0.5% bupivacaine (Bupivacaine Seriluer 0.5%, Pfizer) intercostal (IC) blocks between IC spaces 4–8 and a 2.5–5 μg/kg/min remifentanil IV CRI (Remifentanil Alphapharm, Mylan). A partial left cranial lung lobectomy and complete left caudal lung lobectomy were performed with stapling devices (TA 60–3.5 mm and TA30‐2.5 mm reloadable stapler, Covidien, Dublin, Ireland) due to consolidation and black discoloration of the parenchyma of the cranial lobe and air leakage from the caudal lung lobe. Additionally, debridement of necrotic mediastinal tissue was performed. The previous thoracostomy tube was removed and a left trocar type thoracostomy tube was placed (Argyle chest trochar 16 g, Cardinal Health, Dublin, OH). Lung was submitted for histopathology and aerobic culture and susceptibility. Postoperatively, minimal air and fluid were evacuated from the thoracostomy tube.

At the completion of surgery, a cisternal cerebrospinal fluid (CSF) sample was collected. Cytology was consistent with blood contamination and a relative increase in neutrophils and mast cells. CSF fluid analysis was as follows: cell count (0.6 cells/μL [0‐8 cells/μL]), protein (11.70 mg/dL [<30 mg/dL]), erythrocytes (162/μL [none]), differential 35% neutrophils, 31% mast cells, 17% lymphocytes, 14% large mononuclear cells and 3% eosinophils. Serum cryptococcal antigen lateral flow assay (CrAg® LFA, Immy, Norman, OK.) on day 3 of hospitalisation was negative.

Pressure controlled synchronous intermittent mechanical ventilation that had been commenced during surgery was continued postoperatively in the ICU given inadequate respiratory drive and hypoxemia. Initially, peak inspiratory pressures (PIP) of 16–22 cm H₂O, and positive end‐expiratory pressure (PEEP) of 5–8 cm H₂O were required, with a mandatory frequency (f) of 20–30 breaths/min and FiO2 of 40%–60% to achieve SpO₂ >94% and ETCO₂ of 36–51 mmHg. Analgesia was provided with a fentanyl CRI (2 μg/kg/h IV) (Sublimaze fentanyl citrate injection, Akorn) and midazolam was continued (0.1–0.3 mg/kg/h CRI). Ventilator settings were adjusted to target normoxemia and normocapnia. Spontaneous breaths were evident throughout the period of mechanical ventilation and pressure support of 7–10 cm H₂O was used as a component of the ventilation strategy.

Postoperative transfusion of one unit of type‐specific (LabTest BT A + B, Alvedia, Limonest, France) crossmatch compatible packed red blood cells (pRBCs) was required due to moderate anaemia (PCV 21%), in combination with hypotension and impaired oxygen delivery (central venous oxygen saturation [ScvO₂] 47%–59%). Additionally, a synthetic colloid fluid (2.6 mL/kg IV) (Voluven 6% hydrozyethyl starch 130/0.4, Fresenius Kabi) was administered. After these interventions, PCV increased to 29% and ScvO₂ increased to 76%. Thereafter, the PCV remained above 24% throughout hospitalisation without further transfusion (Table 1).

Serial venous electrolyte panels (ABL800 FLEX blood gas analyser, Radiometer, Copenhagen, Denmark) were monitored throughout hospitalisation, which included measurement of sodium, potassium, chloride, ionised calcium and glucose. The only significant abnormality was a severe hypokalaemia (2.7 mmol/L [3.9–5.2]). Intravenous fluid therapy was provided with lactated Ringer's Solution, Baxter (LRS) at 2 − 3 mL/kg/h supplemented with 30‐60 mmol/L of potassium chloride (Potassium chloride Juno, Juno Pharmaceuticals) as required until day 8. Other treatments during the period of mechanical ventilation included ongoing antimicrobials (trimethoprim‐sulfadiazine and clindamycin as previously), levetiracetam (20 mg/kg IV q8h) and ondansetron (0.3 mg/kg IV q 8 h) (Ondansetron Kabi, Fresenius Kabi).

An oesophagostomy tube (Oesophagostomy tube 14 g × 23 cm, SurgiVet, Minneapolis, MN.) had been placed on day 1, and enteral nutrition was commenced at 1/3 resting energy requirement (RER). During the period of mechanical ventilation, nutrition was provided by placement of a gastric‐through‐oesophagostomy tube (MILA nasogastric feeding tube weighted 6Fr x 55 cm, MILA International, Florence, KY.) and administration of a liquid diet (Ensure, Abbott Nutrition, Chicago, IL.) was continued at 1/3 RER. Feeding was gradually increased throughout his care, reaching full RER on day 7. A metoclopramide CRI (2 mg/kg/day) (Metomide, Ceva Sante Animale) was commenced on day 4 of hospitalisation as feeding volume was increased.

During sedation vacations on each day of mechanical ventilation, PLRs were noted to return bilaterally. As his pulmonary compliance improved, peak pressures and PEEP were gradually reduced, and he was transitioned to continuous positive airway pressure (CPAP) with pressure support after 46 hours of mechanical ventilation. During this period, the fentanyl CRI was titrated between 2 and 4 μg/kg/h, midazolam between 0.1 and 0.3 mg/kg/h and ketamine was also used (0.5 mg/kg/h IV CRI) (Ceva ketamine injection, Ceva Sante Animale) for 16 hours. Extubation was achieved on day 5, after 70 hours of mechanical ventilation in the ICU. Fentanyl and midazolam infusions were discontinued concurrently. Mask oxygen was continued for a further 2 days.

Following extubation, the cat remained markedly obtunded; he was recumbent but able to lift his head and had ongoing improvement in cranial nerve function. Pupillary light responses were present, as were facial sensation, palpebral and gag reflexes. Menace was absent. On day 6 he had a focal facial seizure. The following day he had two focal facial seizures that were treated with midazolam 0.1 mg/kg IV and an additional dose of levetiracetam 20 mg/kg IV. No further seizure activity was seen for the remainder of hospitalisation.

From day 6, the severity of obtundation improved. He regained the ability to ambulate with an ataxic gait on day 7. Proprioceptive placing was absent in all limbs until day 8, from which point it gradually returned. Menace response was absent OU until day 15, when it returned OD. The OS menace remained absent throughout hospitalisation. On day 9, blepharospasm OS with a pinpoint corneal defect was noted, prompting commencement of chloramphenicol ophthalmic ointment (Chlorsig eye ointment 1%, Aspen Pharmacare Australia) that was continued until day 13 when fluorescein staining was negative.

On day 7, the urinary catheter was removed. His mature neutropenia and bandemia had resolved, and a moderate mature neutrophilia was present (Table 1). Oxygen therapy was discontinued as he was normoxic (SpO₂ 100%, FiO2 21%). Overnight, he regurgitated and erythromycin at 1 mg/kg IV q8h (Erythrocin IV, Amdipharm) was added. On day 9, the central venous catheter was removed, and medications were given via the oesophagostomy tube. This included clindamycin 12.5 mg/kg q12h (Clincain, Intervet), metoclopramide 0.6 mg/kg q8h (Maxolon, Inova Pharmaceuticals) and levetiracetam 25 mg/kg q8h (Levetiracetam‐GH, Generic Health). Trimethoprim‐sulfadiazine was discontinued. He began to voluntarily eat soft food on day 7 and by day 9 had an excellent appetite. On day 15 the oesophagostomy tube was removed, medications were discontinued, and he was discharged from the hospital on day 16.

His neurological examination prior to discharge was as follows: mentation quiet, alert and responsive, intact direct and consensual PLRs OU, intact palpebral OU, menace present OD but absent OS, normal tongue movement, symmetry and gag and normal proprioception in all limbs. He was noted to have generalised muscle atrophy, but much less ataxia, and good strength.

In the 2 months postdischarge, he had three further generalised tonic–clonic seizures. Levetiracetam was restarted (30 mg/kg PO q8h). Four additional seizures within the next 3 months prompted transition to phenobarbitone monotherapy at 1 mg/kg PO q12h (Phenomav, Mavlab), which was gradually increased to 4 mg/kg q12h due to ongoing seizure activity. Thereafter, seizure frequency was approximately one isolated seizure every 4 months for the next year. At last report, 8 years after discharge, there had been no reported seizure activity for 6.5 years whilst maintained on phenobarbitone at 3‐4 mg/kg q12h. Other than the seizures, he returned to a normal neurologic status by 1 month after hospital discharge.

Discussion

Our case report is unique as SAE has not been previously reported in cats. The cat reported herein displayed multifocal CNS signs including coma, seizures, abnormal gait, proprioceptive deficits and cranial nerve deficits, consistent with the broad spectrum of SAE manifestations described in humans. The reported multimodal CNS imaging findings, management and long‐term sequelae are also consistent with the reported spectrum of SAE in humans.

Diagnosis of SAE first requires a definitive diagnosis of sepsis and is expected to occur in cases with other organ dysfunctions. ¹³ In this case, sepsis was diagnosed based on meeting ≥2 of 4 SIRS criteria in addition to confirmation of infection with culture. ¹² The isolation of Pasteurella sp., an oropharyngeal isolate, and pulmonary consolidation in this case is consistent with a parapneumonic origin of pyothorax. ¹⁴ The cat also had strong evidence of organ dysfunction with concurrent, severe pulmonary dysfunction, hepatobiliary dysfunction and gastrointestinal dysfunction. ¹⁵ While consensus definitions of sepsis and the role of organ dysfunctions in small animals will soon be forthcoming, ¹⁶ there is currently no feline equivalent to the Sequential Organ Failure Assessment (SOFA) score used to quantify organ dysfunction in the diagnosis of sepsis in humans. ¹⁷

Diagnosis of SAE is challenging due to ill‐defined diagnostic criteria and variable clinical presentation. ¹⁸ The diagnosis of probable SAE in our case was founded on combined clinical timeline and exclusion of other differential diagnoses including metabolic, toxic, primary CNS infection, immune‐mediated and neoplastic diseases. We designed a conceptual framework to illustrate the level of evidence that was present for SAE in this cat, relative to the breadth of diagnostics described for SAE in humans (Figure 2). Notably, this is not proposed as a firm diagnostic algorithm supported by evidence in cats but rather extrapolated from human medicine and our experience. Based on this framework, this cat fulfilled the criteria for probable SAE. Although the cat did have brief periods of systemic hypotension, hypoxemia and mild anaemia, these rapidly resolved with therapy such that ischaemic brain injury from cerebral hypoperfusion does not explain the severity of the neurologic signs. Hypoglycaemia, dysnatraemia and dyscalcemia were also excluded with appropriate testing. Additionally, evidence of encephalopathy persisted after removal of sedative drugs. Furthermore, there was no MRI evidence of global hypoxic brain injury ¹⁹ or structural brain disease. As additional cases of SAE are reported in the veterinary literature this framework could be refined to improve its generalisability and validity.

Figure 2.

Conceptual framework in the diagnosis of sepsis‐associated encephalopathy (SAE) in cats. *Discontinuation of antiepileptic drugs in patients with status epilepticus or cluster seizures should not be performed, but neurologic examination should be interpreted in light of medication effects.

Cluster seizures were an early and persistent manifestation of probable SAE in our case. Human studies have shown that 10–20% of septic patients develop seizures, which are predominantly nonconvulsive but documented via electroencephalogram (EEG). ² , ²⁰ While cluster seizures appeared to respond to anticonvulsant medications in this cat, the absence of EEG is a major limitation in quantifying true seizure burden, and nonconvulsive seizures may have gone unrecognised. In addition to the use of EEG to provide objective confirmation of seizure burden, it may have also aided in neurological characterisation and mechanistic interpretation of epileptogenesis in this cat. Further investigation of EEG to assist in the diagnosis of SAE in cats with sepsis, especially to document nonconvulsive seizures, is warranted.

Neurologic examination in human patients with SAE is challenging and usually limited due to the concurrent administration of sedatives. The Glasgow Coma Scale score is incorporated into the SOFA score to quantify the severity of neurologic dysfunction as a component of MODS in humans with sepsis. ¹⁷ , ²¹ In cats, the modified Glasgow Coma Scale (mGCS) score has been primarily reported in the assessment of traumatic brain injury, ²² , ²³ , ²⁴ but is clinically used for assessment of CNS dysfunction of other causes. ²⁵ Scoring with the mGCS score was not performed in this case but would likely be a useful tool for serial assessment and prognostication of veterinary patients suspected of having SAE. Given the potential value of mGCS scoring, we attempted to retrospectively assign mGCS scores to this cat, but the medical records lacked adequate information to do so serially. At first presentation to TAHMU, the retrospectively assigned mGCS was 3/18 after levetiracetam.

The neurological assessment of critically ill hospitalised patients requires interpretation in the context of neuroactive medications. Use of medications such as benzodiazepines, alpha‐2 agonists, opioids, ketamine, and other anticonvulsants is prone to confounding neurological assessment. Where safe, it is recommended that neurological assessment be performed with interruption of sedative and analgesic medications, but antiepileptic drugs should not be stopped in cases with status epilepticus or cluster seizures. ²⁶ In this cat, we used sedation vacations in an attempt to minimise the confounding effects of sedatives on serial neurologic examination while concurrently assessing respiratory status.

MRI is an important diagnostic modality in the exclusion of other causes of encephalopathy and diagnosis of SAE in humans. Up to 50% of humans with SAE have a normal brain MRI; other findings include cerebral infarction, leukoencephalopathy and cytotoxic and vasogenic oedema. ⁷ , ²⁷ , ²⁸ , ²⁹ , ³⁰ Although our case had a normal brain MRI further investigation into the spectrum of MRI manifestations of SAE in cats is required. The role of specific MRI sequences, particularly diffusion weighted imaging (DWI) and FLAIR in assessment of SAE in humans is important and thus have been included in the requirements to achieve a definitive diagnosis of SAE in our proposed conceptual framework. These techniques allow for more detailed examination for oedema and leukoencephalopathy respectively, are noninvasive, and more readily available. ³¹ Absence of DWI is a limitation in this cat however performing high sensitivity DWI in very small animals requires high magnetic field strength. Additional related techniques such as magnetic resonance spectroscopy and molecular MRI are described in assessment of human SAE; however, are not yet accessible in veterinary medicine. ³² Given the cost and complexity of performing MRI in a septic cat, further investigation into CT imaging in cats with suspected SAE is also warranted. Some brain CT abnormalities have been reported in human patients with SAE, ³³ and concurrent CT of additional body regions has utility in diagnostic workup for other aspects of sepsis ³⁴ .

CSF analysis in humans with SAE is usually associated with minimal cytological change and normal to mildly increased protein. ³⁵ The role of CSF and plasma biomarkers, such as C‐type natriuretic peptide, S100B protein and neuron‐specific enolase, in diagnosis of SAE is under investigation and their application in veterinary species is unknown. ³⁶ In this cat, there was an increased proportion of neutrophils and mast cells in the CSF, albeit in the presence of a low nucleated cell count. The significance of this is challenging to interpret. Mast cell activation is implicated in blood–brain barrier impairment in a rodent model of SAE and may reflect neuroinflammation. ³⁷ Blood contamination is also another possibility, although there was no evidence of peripheral mastocytosis. Since CSF analysis is nonspecific for SAE, its value is primarily to rule out primary CNS infection and immune‐mediated meningoencephalitis.

Whilst there is no consensus regarding treatment of SAE in humans, the Surviving Sepsis Campaign (SSC) guidelines ³⁸ provide direction for treating sepsis more broadly. Treatment of the cat was aligned to the SSC recommendations with comprehensive source control, broad‐spectrum intravenous bactericidal antimicrobials and specific treatments for organ dysfunctions. Despite percutaneous source control, evidence of sepsis persisted, prompting surgical source control at our hospital. A significant burden of infected tissue prior to surgery may have resulted in progression of septic organ dysfunctions and contributed to the severity of the neurologic manifestations.

There is no widely accepted specific therapy for SAE beyond adherence to the general SSC guidelines and symptomatic treatment of specific manifestations such as seizures. In our case, pharmaceuticals included antimicrobial therapy, analgesia, antiepileptics, sedatives, anaesthetics, antiemetics and prokinetics. Nociception, particularly perioperatively, required a combination of systemic pure and partial mu‐agonist opioids, ketamine and locoregional anaesthesia. Antiepileptic therapy included a combination of benzodiazepines, levetiracetam and propofol, with transition to phenobarbitone at home. In retrospect, phenobarbitone loading (a second‐tier therapy) should have been performed when initial cluster seizures were observed to reduce the risk of complications of recurrent seizure activity including kindling and mirroring. Ketamine at higher than analgesic doses, a third‐tier therapy, may have also been beneficial in this case. ³⁹ There is some evidence in septic humans, that patients prescribed opioids, benzodiazepines or propofol were more likely to have SAE. ⁴⁰ Given the paucity of veterinary specific literature, multimodal analgesia should be titrated judiciously to ensure patient comfort. Medetomidine and propofol were also used in the cat to facilitate anaesthesia, mechanical ventilation and patient comfort. Alpha‐2 agonists and propofol have been investigated for their potential benefits in SAE in humans ⁴¹ , ⁴² , ⁴³ , ⁴⁴ , ⁴⁵ and are part of third‐tier therapy for status epilepticus and cluster seizures in veterinary medicine. ³⁹ Nonpharmacological neurorehabilitative interventions that are beneficial in humans with SAE, including early mobilisation, promotion of a normal circadian rhythm, and maximising sleep quality, ⁸ , ⁴⁶ may also be useful in affected veterinary patients.

Gastrointestinal support was provided in our case with ondansetron and metoclopramide, which are both serotonergic medications. Although they have not been demonstrated to be associated with SAE, they have the potential to cause extrapyramidal signs. ³² , ⁴⁷ Low‐dose erythromycin was used as an additional prokinetic and has not been associated with SAE. ⁴⁸ These therapies, targeted at ameliorating gastrointestinal dysfunction, may influence the microbiota‐brain‐gut axis and thereby modulate SAE. ⁴⁹ As the effects of specific pharmaceuticals on risk and treatment of human SAE continue to be elucidated, pharmaceutical treatment of cats with SAE should be goal directed, with consideration for neurologic sequelae.

The development and severity of SAE has been reported as a negative prognostic marker in humans. ¹ , ² In our case report, the cat survived to discharge with a good long‐term functional outcome, albeit after a protracted ICU stay. In human survivors of SAE, long‐term neuropsychological sequelae are recognised, including persistent neurocognitive impairment, psychiatric burden, and seizures. ⁶ , ⁹ , ²⁰ , ⁵⁰ In our case, epilepsy was a major component in his clinical progression with persistence of generalised seizure activity up to 2 years post‐discharge. With the frequency of convulsive seizures as well as the possibility of unobserved nonconvulsive seizures, kindling and mirroring may have contributed to epilepsy postdischarge in this cat. ⁵¹ In retrospect, more aggressive antiepileptic drug therapy consistent with the American College of Veterinary Internal Medicine consensus statement for the management of cluster seizures, ³⁹ would have been ideal in this case. It remains impossible to differentiate the contributions of kindling and mirroring from the direct consequence of SAE in the persistence of long‐term seizures in this cat.

Other limitations in the management of the case preclude the ability to both completely exclude other causes of neurologic signs and thoroughly describe the neurologic manifestations of SAE. Protozoal serologic testing and CSF bacterial culture should have been performed to definitively exclude CNS infection. Pre‐existing or genetic epilepsy also remain a differential diagnoses for the seizures but do not fully explain the neurologic abnormalities in this case.

We appreciate that although not previously reported, neurologic dysfunction likely occurs in many septic veterinary patients. Lack of recognition of SAE may occur due to attribution of neurological signs to refractory shock, or the requirement of anaesthesia for mechanical ventilation ⁵² . Additionally, SAE may not manifest or be thoroughly investigated prior to death from other complications of sepsis or euthanasia associated with poor prognosis or financial burden.

In conclusion, this case demonstrates that cats with sepsis can develop neurologic dysfunction consistent with SAE in humans. Future studies are required to investigate the incidence, spectrum of clinical presentations, ideal diagnostic and therapeutic approach, as well as the prognosis of SAE in cats. The authors propose that SAE is included in future veterinary definitions of sepsis‐associated organ dysfunction.

Conflicts of interest and sources of funding

Dr Chan's residency training program is supported by an Australian Government Research Training Program (RTP) Scholarship https://doi.org/10.82133/C42F-K220.

Data availability statement

Data sharing not applicable to this article as no datasets were generated or analysed during the current study.