Second-generation thiazolidinediones and hepatotoxicity

Ann Pharmacother. 2004 Sep;38(9):1419-23. doi: 10.1345/aph.1E072. Epub 2004 Jul 20.


Objective: To report a case of hepatotoxicity probably caused by pioglitazone, summarize case reports of hepatotoxicity induced by rosiglitazone or pioglitazone, and make recommendations regarding routine liver enzyme measurement in patients taking these agents.

Case summary: A 39-year-old black woman with type 2 diabetes mellitus, hypertension, and congestive heart failure presented to a pharmacist-staffed diabetes comanagement service. She reported fatigue, dark brown urine, nausea, itching, and loss of appetite. Pioglitazone was promptly discontinued because her symptoms were consistent with those of hepatic dysfunction and pioglitazone was identified as a potential cause. The patient was referred to her physician. Liver enzyme levels were checked 13 days after initial presentation and found to be abnormal: alanine aminotransferase 490 U/L, aspartate aminotransferase 360 U/L, alkaline phosphatase 851 U/L, total bilirubin 3.1 mg/dL, direct bilirubin 2.0 mg/dL, and indirect bilirubin 1.1 mg/dL. Within 2(1/2) months of discontinuing pioglitazone, the patient's symptoms resolved and liver enzyme levels returned to normal.

Discussion: Troglitazone, a thiazolidinedione (TZD), was removed from the market because of hepatotoxicity. Reported cases involving the newer TZDs, rosiglitazone and pioglitazone, have been few in number and less severe in consequence. Six cases of rosiglitazone-induced hepatotoxicity and 5 of pioglitazone-induced hepatotoxicity have been reported. Most patients improved symptomatically 2-4 weeks following discontinuation of the offending TZD, with normalization of liver enzyme levels in 2 weeks to 6 months following TZD discontinuation.

Conclusions: Although the timeline and extent of liver enzyme elevation in this case are unclear, the Naranjo probability scale suggests that a causal relationship between pioglitazone and liver disease is probable. Patients with previous TZD-induced hepatotoxicity should not be rechallenged. Cases of hepatotoxicity with second generation TZDs, although clearly linked, have been few in number and less severe in consequence when compared to troglitazone. We agree with current package labeling that requires baseline and then periodic measurement of liver enzymes in patients taking pioglitazone or rosiglitazone.

Publication types

  • Case Reports

MeSH terms

  • Adult
  • Chemical and Drug Induced Liver Injury / enzymology
  • Chemical and Drug Induced Liver Injury / etiology*
  • Diabetes Mellitus, Type 2 / drug therapy
  • Female
  • Humans
  • Hypoglycemic Agents / adverse effects*
  • Hypoglycemic Agents / therapeutic use
  • Liver / enzymology
  • Liver Function Tests
  • Pioglitazone
  • Thiazolidinediones / adverse effects*
  • Thiazolidinediones / therapeutic use


  • Hypoglycemic Agents
  • Thiazolidinediones
  • Pioglitazone