For healthcare professionals only


Fatal liver failure associated with pioglitazone

BMJ 2004; 329 doi: (Published 19 August 2004) Cite this as: BMJ 2004;329:429
  1. Ed Farley-Hills, consultant in anaesthetics and intensive care (edward.farley-hills{at},
  2. R Sivasankar, specialist registrar in anaesthetics1,
  3. M Martin, medicines information pharmacist1
  1. 1 Ysbyty Gwynedd Hospital, Penrhosgarnedd, Bangor LL57 2PW
  1. Correspondence to: E Farley-Hills

    Thiazolidinediones are peroxisomal proliferator activated receptor γ agonists. Troglitazone is associated with idiosyncratic hepatic reaction, liver failure, and death and is withdrawn.1 2 The toxicity of troglitazone is unlikely to be a class effect of thiazolidinediones since rosiglitazone and pioglitazone have shown little evidence of hepatic toxicity.3 Some patients taking pioglitazone, however, have had liver failure, but no deaths are associated with it.4

    A 63 year old white man with no history of alcohol misuse was admitted to hospital with jaundice after feeling unwell for three weeks. Three months before, doctors changed his gliclazide to pioglitazone. He had also taken lercanidipine for some years and a cephalosporin antibiotic for a few days. Blood investigations found concentrations of 522 μmol/l bilirubin, 472 IU/l alkaline phosphatase, 1053 IU/l aspartate aminotransferase, 1984 IU/l alanine aminotransferase, 455 μmol/l creatinine, and 28 g/l albumin. His prothrombin time was 56 seconds. He developed encephalopathy and acidosis 36 hours after admission and doctors transferred him to intensive care.

    He had no stigmata of chronic liver disease, and hepatitis surface antigen, hepatitis A IgM, and hepatitis C antibody were negative. Ultrasound images showed normal parenchymal reflectivity with patent vessels and no biliary dilatation. When stabilised, doctors transferred him to the regional liver unit. He died nine days later.

    The histopathology report describes parenchymal damage with steatohepatitis including Mallory bodies superimposed on a severely fibrotic liver. The cause is not certain, but the degree of fibrosis suggests a chronic process, and the type of necroinflammatory activity raises the possibility of alcohol related liver injury. Alternatively the changes could be drug induced damage superimposed on chronic liver disease related to diabetes, and the time scale indicates that pioglitazone is the likely cause.

    We know of no previous cases of death associated with pioglitazone, although liver failure has been reported.


    • Funding: None.

    • Competing interests None declared.


    1. 1.
    2. 2.
    3. 3.
    4. 4.
    View Abstract