We report the case of an 8-year-old boy with pyridoxamine 5'-phosphate oxidase (PNPO) deficiency. He developed seizures at 24 h of age that were refractory to standard anticonvulsant therapy and a trial of pyridoxine but responded to pyridoxal phosphate (PLP) at 28 days of life. Genetic testing identified compound heterozygous mutations in the PNPO gene. Management of encephalopathic episodes required escalation of PLP dose to 100 mg/kg/day by 2 years of age. Routine blood tests at this time showed significantly deranged liver function tests (LFTs). A wedge liver biopsy showed early cirrhosis with marked elevation of pyridoxal and pyridoxic acid levels in the liver sample. Despite extensive investigation, no cause other than PLP therapy could be identified for the cirrhosis. The PLP dose was weaned to 50 mg/kg/day before episodes of encephalopathy recurred. Concurrent with the reduction of his PLP dose, LFTs showed improvement. However, at 8 years of age, there is persistent evidence of hepatic fibrosis and early portal hypertension. We hypothesise that hepatic toxicity due to PLP or its degradation products is the cause of cirrhosis in this boy. Until further evidence becomes available, we would suggest that people with PNPO deficiency are treated with the minimum dose of PLP required to prevent episodes of encephalopathy.