We had the impression we were seeing a high incidence of pancreatitis in our inflammatory bowel disease (IBD) patients taking a pH-dependent, delayed-release form of 5-aminosalicylate (pHDDR-5-ASA). We wondered if 5-aminosalicylate (5-ASA) was causing the pancreatitis. We therefore elected to review our experience with pancreatitis in our patients with IBD.
From a listing of all elevated amylase and lipase values measured in our hospital laboratory between 1993 and 1995, we identified 11 patients with inflammatory bowel disease who had at least a twofold increase in one of these values. (During this period, amylase and lipase levels were not part of any routine panel in this hospital, nor were they part of any IBD protocols. Hence levels were probably drawn because of a symptom.) These patients' charts were reviewed for evidence that suggested clinical pancreatitis and an etiology for that pancreatitis.
Of the 11 patients with elevated amylase or lipase values, 10 had records available for evaluation. In this group of patients, there were 19 apparent bouts of pancreatitis. The clinical features are summarized briefly below:
An 8.6-year-old girl with Crohn's colitis developed nausea, vomiting, abdominal pain, bloody stools, and a lipase of 1,504 U/l (normal, <239 U/l; not age-related) while taking prednisone, amoxicillin, and pHDDR-5-ASA. She was treated with intravenous ranitidine and methylprednisolone, and the pHDDR-5-ASA was discontinued. Her symptoms resolved and she was discharged within 48 hours. The clinical impression was that she had 5-ASA-induced pancreatitis.
At 8.7 years of age, while taking metronidazole and prednisone because of continued bloody diarrhea, the patient was rechallenged with pHDDR-5-ASA. Within 48 hours she developed fever, thirst, and marked anorexia. Her amylase level was 409 U/l (normal, <100 U/l; not age-related) and her lipase level was 3,061 U/l. The pHDDR-5-ASA was discontinued, and the symptoms resolved within 48 hours. This confirmed she had 5-ASA-induced pancreatitis.
At 8.8 years of age she was admitted with a continued flare of her colitis and was administered azathioprine and methylprednisolone intravenously. She rapidly developed fever and abdominal pain (amylase 254 U/l; lipase 782 U/l). The azathioprine was discontinued and the symptoms promptly resolved. Azathioprine was assigned as the probable cause of the pancreatitis.
At 9.3 years of age during a flare of colitis, while taking cyclosporin A, co-trimoxazole, and methotrexate, she developed fatigue, anorexia and vomiting (amylase 110 U/l; lipase 877 U/l). Findings in an abdominal computed tomographic (CT) scan were normal. Her duodenal folds were prominent on an upper gastrointestinal series. A nonobstructing pancreatic duct stricture was seen during investigation with endoscopic retrograde cholangio-pancreatography (ERCP). The symptoms of pancreatitis and the enzyme levels gradually improved without specific therapy. The colitis remained active, and she eventually required colectomy with Hartman pouch and end ileostomy. In the year since surgery, she has discontinued all medication and has remained well. The etiology of this bout of pancreatitis was not assigned.
A 9.9-year-old girl with ulcerative colitis was in remission while taking prednisone and pHDDR-5-ASA. In anticipation of discontinuing treatment with steroids, the pHDDR-5-ASA dose was increased by 70%. Two weeks later, after a reduction in the prednisone dose from 15 mg alternating with 5 mg to 15 mg every other day, she developed loose bloody stools followed 12 hours later by nausea and vomiting (amylase 172 U/l; lipase 1,505 U/l). Findings in an abdominal sonogram were normal. The pHDDR-5-ASA was discontinued. Her symptoms and enzyme levels gradually improved. She tolerated sulfasalazine at 10 years of age. The initial impression of 5-ASA-induced pancreatitis was rejected after the successful rechallenge with sulfasalazine.
A 12.7-year-old boy was seen with vomiting, diarrhea, and abdominal pain. During appendectomy, small-bowel Crohn's disease was discovered. After surgery, while taking pHDDR-5-ASA, gentamicin, ampicillin, morphine, metoclopramide, and acetaminophen, he developed abdominal pain and vomiting (amylase 118 U/l; lipase 373 U/l). The pHDDR-5-ASA was discontinued. An abdominal sonogram and a CT scan of the abdomen produced no abnormal findings. His condition gradually improved. He eventually tolerated pHDDR-5-ASA at 13.3 years of age. The initial impression of 5-ASA-induced pancreatitis was disproved by rechallenge with the drug.
A 17.9-year-old girl with gastroduodenal and colonic Crohn's disease developed anorexia, vomiting, diarrhea, and abdominal pain 2 months after beginning treatment with azathioprine (amylase 157 U/l; lipase 448 U/l). The azathioprine was discontinued and the symptoms resolved over a 2-week period. Probable cause of the pancreatitis was assigned to azathioprine.
An 11.6-year-old girl with ileocolonic Crohn's disease developed nausea, vomiting, and headache 2 months after beginning treatment with azathioprine (amylase 156 U/l). The azathioprine was discontinued and her symptoms resolved. The impression was that she had azathioprine-associated pancreatitis.
At 17.2 years of age, while taking prednisone, ranitidine, amitriptyline, vivactil, and pHDDR-5-ASA she began to feel ill with vague abdominal symptoms (amylase 118 U/l; lipase 238 U/l). Her treatment with pHDDR-5-ASA was discontinued, but she continued to feel ill. Her amylase level remained mildly elevated but her lipase values were mostly in the normal range. Investigation with ERCP showed cholesterol crystals in the bile. She underwent papillotomy and laparoscopic cholecystectomy, which resulted in resolution of her symptoms. Her condition was assessed as biliary pancreatitis.
A 12.3-year-old girl with ileocolonic Crohn's disease developed abdominal pain, vomiting, and diarrhea while taking azathioprine, prednisone, and zinc (amylase 328 U/l; lipase 4,545 U/l). An abdominal sonogram showed a mildly enlarged pancreas. A CT scan of the abdomen showed no other abnormalities. The azathioprine was discontinued, and the symptoms resolved within 2 weeks. Her condition was classified as azathioprine-induced pancreatitis.
A 13.7-year-old girl with ileocolonic Crohn's disease developed abdominal pain and dehydration while receiving parenteral nutrition, prednisone, and metronidazole (amylase 119 U/l; lipase 493 U/l). Findings in an abdominal CT scan showed bowel-wall edema, and she was treated with cyclosporin A. An abdominal sonogram showed gallbladder sludge. She began taking ursodiol, and her symptoms resolved and enzyme levels returned to normal. The impression was that she had pancreatitis induced by biliary disease.
A 14.9-year-old girl with ulcerative colitis who had undergone colectomy, with J-pouch pullthrough, developed diarrhea. She was treated with 5-ASA enemas and metronidazole. An endoscopy of the pouch resulted in findings of inflammation in biopsy. Metronidazole was discontinued and ciprofloxacin begun. The patient developed nausea and vomiting and was admitted to the hospital (amylase 262 U/l; lipase 23 U/l). A repeat determination showed an amylase level of 143 U/l. An abdominal sonogram was normal. Test results for C. difficile toxin were positive. The symptoms resolved when the patient was treated with vancomycin alone. The etiology of this apparent pancreatitis was not assigned.
A 10.8-year-old girl with Crohn's colitis was admitted to the hospital with frequent bloody stools, abdominal pain, and upper respiratory symptoms, while taking sulfasalazine and prednisone (amylase 168 U/l; lipase 229 U/l). The symptoms resolved and the enzyme levels returned to normal after treatment with intravenous methylprednisolone.
At 10.9 years of age, while taking prednisone, pHDDR-5-ASA, and rectal steroid foam, she developed epigastric abdominal pain and diarrhea (amylase 205 U/l; lipase 266 U/l). The 5-ASA was discontinued. Abdominal sonogram and esophagogastroduodenoscopy (EGD) were normal. The symptoms lessened after treatment with metronidazole and intravenous methylprednisolone.
At 11.5 years of age she again developed bloody diarrhea and epigastric abdominal pain. Pancreatic enzyme levels were mildly elevated with an amylase in the 95 to 159 U/l range and lipase in the 90 to 294 U/l range. Medications included prednisone, metronidazole, and 5-ASA enemas. An EGD and an abdominal sonogram were normal. A test for clostridium difficile toxin yielded positive results. The symptoms lessened after treatment with oral vancomycin.
The patient has not had a recurrence of pancreatitis in more than a year, despite subsequent therapy with sulfasalazine. The etiology of her mild episodes of pancreatitis is entirely unclear.
An 11.9-year-old boy had abdominal pain, vomiting, and bloody diarrhea. He was diagnosed with ulcerative colitis, and his symptoms were lessened by treatment with intravenous methylprednisolone. He began taking sulfasalazine. During his hospitalization he was noted to have elevated enzyme levels (amylase 178 U/l; lipase 811 U/l). The sulfasalazine was discontinued and the enzyme levels returned to normal.
Over the next several months he developed an elevated alanine aminotransferase (ALT). A liver sonogram was normal and hepatitis panel was negative. Results from endoscopic retrograde cholangiopancreatography (ERCP) showed normal ducts, and liver biopsy findings showed a neutrophilic infiltrate in periportal areas suggestive of early sclerosing cholangitis. Levels observed in results of his liver function tests returned to normal after treatment with ursodiol. A biliary etiology for his pancreatitis was suspected.
At 12.5 years of age the patient was struck in the abdomen and immediately developed severe abdominal pain (amylase 256 U/l; lipase 1,586 U/l). A sonogram showed a slightly prominent tail of the pancreas. He was admitted, tolerated food, and was discharged. He was readmitted shortly thereafter with persistent symptoms and elevated enzyme levels. Treatment with pHDDR-5-ASA was discontinued without a change in symptoms. Findings in an abdominal computed tomography scan were normal, as were those in an ERCP. No crystals were seen in a sample of bile. This episode was assessed to be caused by abdominal trauma.
At 12.8 years of age, with a background of chronic symptoms (amylase 337 U/l; lipase 2,573 U/l) the patient began treatment with on azathioprine. Within 72 hours his amylase level had risen to 733 U/l, his lipase level to 8,290 U/l. The azathioprine was discontinued and 24 hours later his amylase level was 193 U/l and his lipase level 1,593 U/l. The enzyme levels gradually returned to normal during the next 2 months, when he received parenteral nutrition. This episode was assessed as azathioprine-induced with a background of traumatic pancreatitis.
At 13 years of age, after undergoing colectomy, he had “pancreas pain,” without vomiting (amylase 505 U/l; lipase 6,213 U/l). Medications at that time included rectal steroids, cromolyn sodium, inhaled steroids, and ursodiol. Results of a sonogram showed an echogenic pancreas and normal bile ducts. The enzyme levels promptly returned to normal. One month later, when he underwent endorectal pullthrough and cholecystectomy, stones were found in the gallbladder. He was assessed then to have had pancreatitis induced by biliary disease.
The patient subsequently had fatigue and diarrhea and a diagnosis of probable Crohn's disease was made. He has not had further evidence of pancreatitis while taking pHDDR-5-ASA, ursodiol, prednisone, and steroid enemas.
Table 1 lists the probable cause for each bout of pancreatitis. Probable cause could be assigned in only 11 of 19 bouts. Table 2 summarizes the frequency of the causes.
Three patients (5, 7, and 10) had evidence of biliary disease. Patient 7 responded to ursodiol therapy, and the other two required cholecystectomy. The patients who underwent treatment for biliary disease have had no further relapses of pancreatitis.
Five patients (1, 4, 5, 6, and 10) developed pancreatitis while taking azathioprine, and in each case the pancreatitis resolved when the medication was discontinued. Because of the very high rate of relapse with rechallenge reported for azathioprine-induced pancreatitis (see discussion), none of these patients were rechallenged, and azathioprine was accepted as probable cause.
Seven patients (1, 2, 3, 5, 8, 9, and 10) were taking a 5-ASA-containing drug when they developed pancreatitis. Only one (see above) relapsed with rechallenge, and pHDDR-5-ASA was then assessed to have caused two bouts of pancreatitis in that patient. Four patients (2, 3, 9, and 10) subsequently tolerated pHDDR-5-ASA or sulfasalazine. Patients 5 and 8 were not rechallenged.
Two patients had strong evidence to support different causes for different bouts of pancreatitis. Patient 1 had pHDDR-5-ASA-induced pancreatitis confirmed by rechallenge and a separate bout of azathioprine-induced pancreatitis. Patient 10 had episodes of traumatic, azathioprine-induced, and biliary pancreatitis. Both of these patients also had episodes that were unexplained.
There are many accepted triggers for pancreatitis, including drugs, alcohol, gallstones, viruses, metabolic derangements, and inflammatory bowel disease. The mechanism by which these induce pancreatitis is not understood. Assigning cause without understanding mechanism is dangerous, but helpful points can be gleaned from the literature.
Inflammatory bowel disease-related abnormalities that may predispose a patient to develop pancreatitis include: stricture at the ampulla of Vater secondary to duodenal inflammation (1), spontaneous reflux of duodenal contents into the pancreatic duct secondary to duodenal inflammation (2), ductular abnormalities related to IBD-associated sclerosing cholangitis (3), direct involvement of the pancreas with a granulomatous inflammation (4), and autoantibodies to pancreatic juice in patients with Crohn's disease (5). Results of a prospective study demonstrated asymptomatic elevated levels of amylase or lipase or both in 15% to 25% of a series of 161 patients with IBD (6). In this group, acute clinical pancreatitis was found in only 3% of patients, a fraction similar to the 4.5% seen in our patient population.
Inflammatory bowel disease was not clearly identified as a cause of pancreatitis in any of our patients. Only patients 1 and 4 had evidence from biopsy results of duodenal inflammation, but four of their five combined bouts of pancreatitis appeared to be drug-related.
There is a clear-cut association between 5-ASA-containing drugs and pancreatitis. Pancreatitis has been described associated with treatments with sulfasalazine (7), pHDDR-5-ASA (8), olsalazine (9), and mesalamine enemas (10), all confirmed by rechallenge. In addition, a patient has been described who developed pancreatitis while taking olsalazine and had it recur while taking sulfasalazine (9). Positive results in a lymphocyte-stimulation test during treatment with sulfasalazine in a patient who developed pancreatitis on the drug, implies that at least some of these cases are immune-mediated (7). It is apparent that both the sulfa moiety and the 5-ASA moiety can induce pancreatitis (11).
The low rate of recurrence during rechallenge with 5-ASA in the current series suggests that other causes (IBD, for example) are contributing to the pancreatitis seen in patients who are taking these drugs. Four of the five patients, who developed pancreatitis while taking 5-ASA-containing drugs and who were rechallenged with the drug at a later date, tolerated the drug on rechallenge. It should be noted that all of them developed pancreatitis while their IBD was active, and the pancreatitis gradually resolved with withdrawal of the medications. Their IBD was in remission when they were rechallenged. It seems possible that the IBD in some way enhanced susceptibility to 5-ASA-induced pancreatitis, or that the IBD was responsible for their pancreatitis. A prospective study of pancreatitis in IBD might help to make this distinction.
Azathioprine has been associated with pancreatitis in 3% to 5% of patients with inflammatory bowel disease who are treated with the drug (12,13). The pancreatitis typically occurs after approximately 3 weeks of therapy and, in the vast majority of patients, promptly recurs with rechallenge. There is only one patient reported who continued on the drug after resolution of the pancreatitis (13). Because of the delayed onset and rapid relapse with rechallenge in virtually all patients reported, an immune mechanism has been postulated (12). It is also possible that the drug has a direct effect on pancreatic secretions in susceptible individuals (14). Because of the reported high recurrence rate, none of the patients in the current study were rechallenged with azathioprine.
Corticosteroids are frequently mentioned as potential causes for pancreatitis, though the evidence supporting this is weak (11). Eight of the ten patients in this group were receiving steroids. The steroids were not discontinued for any of the bouts of pancreatitis, and all episodes resolved, suggesting that the corticosteroids were not responsible for the pancreatitis.
We conclude that there are multiple potential causes for pancreatitis in patients with inflammatory bowel disease. Thorough evaluation of each episode is warranted. Carefully monitored rechallenge of patients with suspected 5-ASA-induced pancreatitis should be pursued before excluding these important medications from therapy. A prospective study is needed to evaluate the role of inflammatory bowel disease in causing pancreatitis.