An acute hepatitis resembling autoimmune hepatitis occurring during imatinib therapy in a gastrointestinal stromal tumor patient.

A 65-year-old man was admitted to our hospital with a 6-month history of increasing abdominal size. His medical history was unremarkable. On physical examination, a mass was felt in the right abdomen. Laboratory tests were within normal range. In particular, aspartate aminotransferase (AST) and alanine aminotransferase (ALT) were 21 and 22 U/dL, respectively (normal range <=40 U/dL). Computed tomography scan showed a 16-cm peritoneal mass without evidence of hepatic involvement. A gastrointestinal stromal tumor (GIST) was diagnosed and imatinib (IM) therapy (400 mg/d) was started. One month later, abdominal computed tomography scan showed initial shrinkage of this mass, but an asymptomatic progressive increase of AST and ALT values was observed. IM was withdrawn on the hypothesis of a rare toxicity due to IM. Nevertheless, AST and ALT went up to 418 and 1056 U/dL, respectively (Fig. 1). All other liver function tests were within normal range, and no viral infection could be detected. Alcohol and other drug abuse were excluded. Antinuclear antibodies and M2 fraction of antimitochondrial antibodies were positive. A liver biopsy showed no tumoral cells but interface hepatitis with piecemeal necrosis (Fig. 2). This picture was consistent with an autoimmune hepatitis (AH).1 Ursodeoxycholic acid (1200 mg/d) and prednisone (1 mg/kg/d) were begun. Six weeks later, imatinib was given at the dose of 400 mg/d. The aminotransferase serum levels decreased progressively (Fig. 1) and prednisone, as well as ursodeoxycholic acid, was suspended after 6 months. One month later, a liver biopsy showed a complete resolution of the inflammatory disease (Fig. 3). IM therapy was continued, and the patient was well at his last follow-up visit. Figure 1 Figure 2 Figure 3 Hepatotoxicity (mild transaminitis) is observed in 10% of patients treated with imatinib, and a grade 3 or 4 elevation in liver function tests affects fewer patients.2 Imatinib has been occasionally reported to cause hepatotoxicity 3 with a common histologic signature: a centrolobular hepatic necrosis with lymphoplasmacytic infiltration and interface hepatitis. In one case, antinuclear antibodies were elevated while autoantibodies were either normal or not evaluated in other reports. None of the largest published series examine the 3% to 5% reported G3-4 hepatotoxicity,2 and this is the first case, to our knowledge, of AH described during imatinib therapy in GIST. The association of GIST, IM, and AH may have been coincidental, but the chronological sequence raises the suspicion of a causal relationship. Less clear is the pathogenesis of this phenomenon. A first hypothesis is that IM itself could have been the triggering event. In a relatively “young” drug, rare side effects become evident in the long run. However, we suggest this is an unlikely mechanism. Indeed, autoimmune drug (eg, phenytoin) side effects require drug withdrawal, and readministration causes flare up of disease. A second hypothesis 4 is that massive tumor cell lysis elicited by IM might have been the precipitating cause. This seems more reasonable and consistent with the described evolution of the disease. The awareness of this complication allows us to manage it without stopping imatinib.