An 8 yr old spayed female domestic shorthair presented with a 1 wk history of anorexia and lethargy. Cytological examination of the liver using fine-needle aspiration revealed numerous adipocytes, confirming hepatic lipidosis. The cat was hospitalized for treatment; subsequently, the clinical signs and blood analysis results improved. On day 17, the nasoesophageal tube was removed because of recovery of appetite and vitality. However, owing to insufficient voluntary food intake, partial parenteral nutrition (PPN) was initiated using an amino acid solution with a low branched-chain amino acid (BCAA)/aromatic amino acid (AAA) ratio and arginine deficiency. Within 24 hr of PPN administration, neurological signs were observed, including vocalization, panting, hyperactivity, and a subsequent comatose state. Although hepatic encephalopathy was considered, there was limited evidence of hepatic failure. The clinical course was most consistent with acute hyperammonemia secondary to arginine deficiency, although a contributing role of the low BCAA/AAA ratio cannot be excluded. This case suggests that amino acid PPN with arginine deficiency and a low BCAA/AAA ratio may contribute to neurological signs in cats, underscoring the importance of careful amino acid selection in nutritional formulations.

Cushing's syndrome (CS) is a common endocrine disorder in dogs that can significantly impair their quality of life. Diagnosis is often challenging because of its variable clinical presentation, making it difficult to identify suitable candidates for further diagnostic tests. This study employed machine learning algorithms to assist in CS diagnosis using routinely available screening diagnostics, including complete blood count, serum chemistry panel, and urinalysis parameters such as urine specific gravity and urine protein-to-creatinine ratio. Data were collected from 153 control dogs initially suspected of CS but later excluded and 152 dogs with confirmed CS. A boosted tree algorithm (gradient boosting) was trained on 80% of the collected data, with the remaining 20% reserved for testing. The developed model demonstrated an accuracy of 88.5% [95% confidence interval (CI): 80.5-96.5%], a sensitivity of 83.3% (95% CI: 70.7-96.7%), a specificity of 93.5% (95% CI: 84.9-100%), and an area under the receiver operating characteristic curve of 0.912 (95% CI: 0.835-0.988), indicating excellent discriminatory ability. A user-friendly graphical interface was also developed to facilitate clinical implementation, potentially improving diagnostic efficiency and owner satisfaction.