Thiazolidinedione agents decrease blood sugar by lowering insulin resistance. The thiazolidinedione agent troglitazone has been removed from the market because of its association with severe liver injury. Rosiglitazone, another compound in this class, has been associated with a few cases of liver injury. Pioglitazone, the newest of these agents, has been associated with liver damage in only one previous case report. May and associates describe a patient who developed acute, reversible, mixed hepatocellular–cholestatic liver damage while taking pioglitazone.

This 49-year-old man with type 2 diabetes was given pioglitazone in addition to his previous regimen of glyburide and metformin. His starting dosage of pioglitazone was 15 mg per day, which was increased to 30 mg per day after four months. He developed anorexia, nausea, and upper abdominal discomfort, and was evaluated for these symptoms six months after he started taking pioglitazone. His liver enzyme levels were normal and, because of hyperglycemia, the dosage of pioglitazone was increased to 45 mg per day. Several days later, the patient developed icteric sclera, and blood tests revealed the following: total bilirubin level of 5.7 mg per dL (98 μmol per L); aspartate aminotransferase level of 114 U per L; ala-nine aminotransferase level of 218 U per L; and alkaline phosphatase was more than twice the normal level. Sonography and viral hepatitis test results were normal.

The patient stopped using pioglitazone but continued all other medications. His liver chemistries continued to worsen, and a liver biopsy revealed mild centrilobular cholestasis in hepatocytes and bile canaliculi, accompanied by lymphocytic infiltrates in the portal tract. This finding represented a mixed hepatocellular–cholestatic type of liver damage most likely caused by drug toxicity. After several weeks, the liver chemistry tests began to improve and returned to normal about six weeks later, along with resolution of all symptoms.

The authors conclude that pioglitazone is potentially hepatotoxic and, especially during the first year of therapy, patients should be closely monitored for liver dysfunction. As with troglitazone and rosiglitazone, damage to the liver is probably idiosyncratic, although many more cases were reported with troglitazone. It remains unclear whether the concomitant use of other glucose-lowering or antihypertensive medications exacerbates the potential liver toxicity of pioglitazone.

In an editorial in the same journal, Nierenberg notes that although pioglitazone probably causes hepatic injury, the reporting of only two cases makes a definite conclusion difficult to reach. He notes that infrequent drug toxicities may not be identified until a very large number of people have used the drug. Monitoring of liver function tests as recommended, both before drug administration and every two months during the first year of use, will allow early drug termination if liver enzyme levels increase to more than three times the upper limit of normal or if clinical hepatitis develops. Physicians can help by reporting all acute drug reactions to the manufacturer and/or the U.S. Food and Drug Administration (FDA). (The MedWatch program [available online atwww.fda.gov/med-watch/safety.htm] allows physicians to complete a simple form that relays information directly to the FDA.) Nierenberg concludes that pioglitazone appears to be significantly safer than troglitazone, but the frequency of severe liver side effects of pioglitazone may not be known until more patients have used the drug.