A 2-year-old female neutered domestic shorthair cat was presented to our multidisciplinary specialist referral hospital for investigation of acute onset of tachypnoea and dyspnoea. On physical examination, the patient was depressed, 5% dehydrated, tachycardic (180 beats per minute) and tachypnoeic (60 breaths per minute). Mild respiratory effort was present and lung sounds were attenuated, especially in the left hemithorax. The remainder of the physical examination was normal, including temperature (38.9ºC) and body condition score (2/5; 3.2 kg). Abnormalities on haematological and serum biochemical analysis performed at our hospital were confined to a neutrophilia with a left shift (neutrophils 22.7 × 109/l, reference interval [RI] 5.5–19.5 × 109/l; band neutrophils 4.54 × 109/l, RI 2.5–12.5 × 109/l) and mild hypoproteinaemia (49 g/l; RI 54–78 g/l). Thoracic radiography showed bilateral pleural effusion, which was more marked on the left, and a small number of isolated gas bubbles in the left hemithorax (). Thoracostomy tubes (12 Fr G; Portex, Smiths Medical International) were placed bilaterally with the Seldinger technique and the pleural effusion drained. The cytological analysis of the fluid was compatible with septic inflammation, confirming the diagnosis of pyothorax (). Actinobacillus ureae was cultured. Following drainage of the pleural effusion, intravenous fluid therapy (IVFT) was initiated with Hartmann’s solution (Aqupharm 11; Animalcare) at 4 ml/kg/h. Multimodal analgesia consisting of meloxicam (Metacam; Boehringer Ingelheim) 0.1 mg/kg subcutaneously continued orally (0.05 mg/kg PO q24h), methadone (0.2 mg/kg IV q6h [Comfortan; Dechra]) and levobupivacaine (1mg/kg intrapleurally q8h [Chirocaine; AbbVie]) was initiated. Antibiotic therapy was instituted with metronidazole (10 mg/kg IV q12h [Metronidazole; Baxter Healthcare]) and cefuroxime sodium (20 mg/kg IV q8h [Zinacef; GSK]), prior to receiving the bacterial culture results. Once bacterial culture results became available, the antibiotic choice was confirmed to be suitable. The thoracostomy tubes were drained every 4 h initially and flushed every 12 h with 40 ml (12.5 ml/kg) of saline in each side. The fluid was then re-aspirated after 2 mins in order to decrease the bacterial burden. Full fluid analysis, including cytology, was performed periodically in our in-house laboratory (). Pain was assessed every 2 h, according to the Glasgow Feline Composite Measure Pain Scale and analgesic requirements were adjusted accordingly. Respiratory rate and effort improved within 2 days of hospitalisation. Pleural fluid analysis over the first 12 days revealed reducing cellularity but persistent septic inflammation (). Thoracic radiographs were acquired on day 7 and day 10 following admission. On day 7, thoracic radiographs were acquired as the patient improved clinically and only small volumes were drained from the thoracostomy tube. The thoracic imaging was performed in order to identify an underlying cause and any residual pleural effusion. No underlying cause was found. The pleural effusion was reduced but unresolved, and lucent areas superimposed over the left lung were still present. These areas were likely gas pockets, although a cavitary lesion such as an abscess with gas content could not be ruled out. Owing to this possibility and the lack of resolution of effusion, the left thoracostomy drain was repositioned. Correct positioning of this drain was confirmed with further thoracic radiography. On day 10, thoracic radiographs were repeated to re-assess the thorax and identify an underlying cause for the pyothorax. No underlying cause was identified on these radiographs. The left thoracic drain was in the subcutaneous tissues and no longer intrathoracic, while the right remained correctly positioned. The volume of pleural effusion from the left hemithorax had increased when compared with the volumes previously drained. The right drain was removed due to minimal production and the left drain was repositioned, confirming its correct location with further thoracic radiography. Thoracic ultrasound was performed, which revealed evidence of persistent pleural free fluid, mainly in the left hemithorax, suggesting further pockets of infection or abscessation. Fluid analysis confirmed worsening septic inflammation (). Given the deterioration of the infection after 12 days of antibiotic therapy, drainage and lavage, the patient underwent exploratory thoracotomy in the light of the persistent infection as assessed by cytology. The patient was premedicated with dexmedetomidine hydrochloride (Dexdomitor; Orion Pharma) 5 µg/kg and methadone (0.2 mg/kg IV). General anaesthesia was induced with alfaxalone (3 mg/kg IV [Alfaxan; Jurox]) and maintained with total IV anaesthesia with alfaxalone 7 mg/kg/h. After endotracheal intubation, oxygen was supplemented and lungs were mechanically ventilated using a volume-controlled mode. Intraoperative monitoring included electrocardiogram, pulse oximetry, capnography, spirometry, oesophageal temperature, and oscillometric and invasive arterial blood pressures. Intraoperative analgesia consisted of 1 mg/kg IV ketamine boluses (Anaestamine; Animalcare) and 10 µg/kg/h remifentanil hydrochloride (Ultiva; Aspen Pharma) infusion. Hartmann’s solution was administered intraoperatively at a rate of 5 ml/kg/h. Metronidazole and cefuroxime sodium were continued as previously prescribed with no additional antibiotics given. A standard median sternotomy was performed with an oscillating saw to osteotomise the sternum and allow access to both hemithoraces. Surgical exploration identified thickened, fibrotic mediastinum covered with purulent material. The mediastinum was debrided and areas of encapsulated abscesses were found in the thoracic cavity and, of particular concern, an abscess involving the right cranial lung lobe and heart base (). The right cranial lobe was atelectatic and the right caudal lung lobe was covered by non-restrictive fibrinous adhesions. The left lung lobes were replaced by a thick cord of fibrotic tissue. Lung recruitment manoeuvres were performed manually, achieving only a small degree of inflation of the right caudal and accessory lobes (). During debridement of the right caudal lobe, air leakage was detected from the parenchymal surface, due to an 8 mm pulmonary laceration on the ventromedial aspect of this lung lobe. There was no evidence of acute or historical haemorrhage associated with the laceration. A subtotal pericardiectomy was performed. Part of the pericardium was excised and a 2 cm × 2 cm pericardial flap was raised from the pericardium adjacent to the laceration. The pericardium used for this flap was overlying the heart’s apex and part of the right ventricle. The base of the flap was overlying the right ventricle. During this procedure a minimal amount of purulent pericardial effusion was drained. Tissue samples of pericardium and mediastinum were submitted for bacterial culture. The pericardial flap was reflected caudolaterally and apposed over the laceration of the cranial part of the right caudal lung lobe. The free edge of the patch was then sutured with 4/0 polydioxanone (PDS II; Ethicon) in a simple continuous pattern to a rim of fibrinous adhesion that was partially covering this lobe. Further debridement of the adhesions was not possible without causing further severe parenchymal lacerations. Prior to closure, the chest was lavaged with 300 ml/kg of warmed sterile saline (Aqupharm 1; Animalcare) and checked for air leakage. A thoracostomy tube was placed in the right side, and the previously placed left-sided tube was maintained. The thoracic cavity was closed routinely with 2/0 polydioxanone placed in a figure-of-eight pattern for apposition of the sternebrae and 3/0 polyglecaprone 25 (Monocryl; Ethicon) simple continuous closure for the subcutaneous and intradermal layers. The cat was hospitalised in our intensive care unit for the first 48 h postoperatively and oxygen supplemented in an oxygen cage for the first 16 h. The cat was mildly tachycardic and tachypnoeic, but no dyspnoea was observed in the immediate postoperative period. Thoracostomy tubes were drained every 2 h. Postoperative medication consisted of meloxicam (0.05 mg/kg q24h PO), levobupivacaine (1 mg/kg intrapleurally q4h), methadone (0.2 mg/kg IV q6h), cephalexin (20 mg/kg PO q12h; Therios; SOGEVAL), metronidazole (20 mg/kg PO q12h; Metronidazole, Zentiva) and marbofloxacin (2 mg/kg PO q24h; Marbocyl; Vétoquinol). IVFT was continued with Hartmann’s solution (4 ml/kg/h). Pain assessment was continued as preoperatively and analgesia was titrated accordingly. The left-sided thoracostomy tube was removed after 9 h owing to minimal fluid production. The right-sided tube was removed after 7 days and the tip sent for bacterial culture. During the postoperative period of hospitalisation, tachycardia and tachypnoea improved and no episodes of pyrexia were recorded. Samples of the pleural effusion were sent for cytological analysis periodically, showing resolving neutrophilic inflammation with no bacteria. The bacterium isolated from the surgical tissue sample was Staphylococcus epidermidis, and was sensitive to cephalexin and marbofloxacin. Metronidazole was discontinued when the second culture and sensitivity results were available. The cat was discharged 8 days postoperatively. Marbofloxacin and cephalexin were continued for 3 weeks after discharge. Thirteen weeks postoperatively, the cat was presented to the referring vet with a 24 h history of lethargy. No increase in respiratory rate or effort was detected on initial examination, but pyrexia (40.2ºC) was detected. This episode was treated with amoxicillin/clavulanic acid (20 mg/kg PO q12h) and metronidazole (20 mg/kg PO q12h) for 7 days and resolved with no complications. The cat was re-examined 19 weeks postoperatively and had received no treatment for the previous 5 weeks. The owner reported good progress since the time of surgery, no dyspnoea, an active demeanour and good appetite. The patient tolerated normal levels of activity with no episodes of dyspnoea. Only occasional episodes of a dry, non-productive cough were noted once every couple of days, which were triggered when the animal was in dusty environments. On thoracic auscultation, mild, muffled lung sounds were noted in the left hemithorax. The rest of the physical examination was unremarkable, including temperature, respiratory rate and effort. Survey radiographs obtained at this point revealed good bilateral aeration of the pulmonary parenchyma. The caudoventral border of the left caudal lung lobe demonstrated a mildly undulating margin on the right lateral projection and mild retraction on the dorsoventral projection. These changes were interpreted as focal areas of lack of full expansion of the lung. In addition, there were pleural fissure lines between the right cranial and middle lung lobes, likely secondary to fibrosis from the previous confirmed pyothorax. No pleural effusion was noted (). Owner instructions at this stage consisted of continuing normal activity and monitoring the cough. In case of deterioration or persistence of the cough, it was advised to re-examine the patient and discuss further diagnostic options. Thirty-eight weeks postoperatively, the owner still reported an occasional cough, less evident during cold weather. No episodes of dyspnoea were reported and exercise tolerance was still good with some periods of tachypnoea for less than 1 min after high levels of activity. No more episodes of lethargy or pyrexia were described.