Controversy surrounds the diagnosis of corticosteroid-induced pancreatitis. Despite many initial reports, subsequent reviews questioned whether corticosteroids caused pancreatitis in any of these cases [1, 2]. Only one report of corticosteroid-induced pancreatitis in which recurrence was shown with rechallenge has been published [3]. We report a second such case.

A 51-year-old man with the acquired immunodeficiency syndrome (AIDS) had an episode of pancreatitis in March 1993 attributed to didanosine treatment. The pancreatitis resolved after didanosine therapy was discontinued. Ultrasonography of the abdomen at that time showed no gallstones or sludge. The patient had normal serum albumin, calcium, and triglyceride levels.

In June 1994, the patient began receiving prednisone (40 mg/d) for immune thrombocytopenic purpura. The patient's other medications were zidovudine, metoprolol, zalcitabine, gemfibrozil, and sucralfate. Six weeks later, he developed epigastric pain similar to that experienced with the previous episode of pancreatitis. The patient stopped taking prednisone, and the symptoms resolved.

On 23 August 1994, prednisone therapy was resumed for treatment of immune thrombocytopenic purpura. The patient had normal serum amylase and lipase levels and was receiving no other new medications. Two days later, he developed epigastric pain with serum amylase levels of 384 U/dL (normal, 40 to 150 U/dL) and serum lipase levels of 16.4 U/mL (normal, less than 1.5 U/mL). Prednisone therapy was discontinued; within 2 days, the symptoms resolved, and the amylase and lipase levels returned to normal.

On 11 September 1994, the patient began receiving dexamethasone (3 mg intravenously every 6 hours) for treatment of possible adrenal insufficiency. The following day, he again developed epigastric pain and tenderness. Serum amylase and lipase levels increased from normal to 1570 U/dL and 59.1 U/mL, respectively. Two days after dexamethasone therapy was discontinued, the symptoms resolved, and the amylase and lipase levels returned to normal.

Chronic asymptomatic elevations in amylase and lipase levels are common in patients with AIDS [4]; acute clinical pancreatitis in these patients is often caused by medications. Corticosteroid treatment may be followed by cytomegalovirus disease in patients with AIDS, but symptoms usually develop more than a week after discontinuation of steroid therapy [5]. Our patient (who had no concurrent opportunistic infections) developed symptoms of pancreatitis each time he received corticosteroids, and he rapidly improved after therapy with the corticosteroids was discontinued. We believe that corticosteroids should again be added to the list of medications capable of causing pancreatitis.