We report a case of biopsy confirmed chronic active hepatitis that was probably attributable to paroxetine. So far as we are aware, no other similar cases have been published, although there are a few reported cases of hepatitis associated with fluoxetine.1 2

A 54 year old woman was treated for depression. She had been receiving thyroxine 100 μg daily, isosorbide mononitrate 10 mg three times a day, atenolol 100 mg daily, and aspirin 300 mg daily for 22 months, drugs she continued to take over subsequent years. After two months of treatment with amitriptyline 50 mg daily, paroxetine 20 mg daily was substituted, a drug that she had not taken before. Results of liver function tests 10 months later were abnormal: aspartate transaminase concentration was 256 IU/l (normal 13-42 IU/l), with normal bilirubin and alkaline phosphatase concentrations. Liver function tests had given normal results two months before the start of paroxetine treatment. Results reached a maximum six months later (aspartate transaminase concentration 299 IU/l and alkaline phosphatase concentration 254 IU/l (normal 30-130 IU/l)). Subsequently, she received ethamsylate for six weeks, and she was diagnosed as having diabetes mellitus (treated by dietary intervention) 18 months after she started taking paroxetine.

Liver ultrasonography yielded normal results, and serological tests for hepatitis A, B, or C and for autoantibodies gave negative results. A liver biopsy specimen taken after 17 months of paroxetine treatment showed chronic active hepatitis with eosinophilic infiltration, suggesting a drug related cause. Paroxetine treatment was therefore stopped. Aspartate transaminase concentrations fell five weeks later to 63 IU/l, returning to normal after a further 13 weeks, although alkaline phosphatase concentration remained minimally increased at 150 IU/l. Alkaline phosphatase concentration subsequently returned to normal, and the results of liver function tests were within normal limits 15 months after paroxetine was stopped. A repeat liver biopsy could not be justified on ethical grounds.

The Committee on Safety of Medicines has been notified of six possible cases of hepatitis associated with paroxetine, although only two of them were confirmed by biopsy (personal communication). Transient abnormalities in liver function tests have, however, been recognised before in association with paroxetine,3 and in 54 cases abnormal results without other evidence of hepatic disease have been reported to the committee. In our patient the temporal relation between paroxetine and confirmed hepatitis and its reversal after stopping the drug suggests a causal association.