J Am Board Fam Pract: first published as 10.3122/jabfm.7.1.74 on 1 January 1994. Downloaded from Recurrent Pancreatitis After Treatment With Hydrochlorothiazide

Raymond]. Rion, MD

The cause of acute pancreatitis is usually readily the emergency department with acute epigastric apparent after a basic examination. The most pain and recurrent vomiting. common causes of pancreatitis are cholelithiasis The patient had a history of mild hypertri­ and alcoholism, which account for 75 to 80 per­ glyceridemia, consumed three alcoholic drinks a cent of all cases. I In approximately 10 percent of month, and had no history of biliary tract disease. cases, no cause is found.2,3 Other causes of pancrea­ On examination the patient was a large muscular titis include trauma, the postpartum and post­ man in extreme distress. His blood pressure was operative states, toxins, hyperlipidemias (types I, 172/113 mmHg, respirations were 24/min, tem­ Iv, V), hypercalcemia, hyperparathyroidism, pep­ perature was 98. 8°F (37.1 0c), and weight was 242 lb tic ulcer disease, endoscopic retrograde cholangio­ (110 kg). His abdomen had no bowel sounds and pancreatography (ERCP), and certain infections was slightly distended. He had extreme tender­ (mumps and coxsackievirus).I,3,4 Many drugs have ness in the epigastric area with marked guarding. been shown or presumed to cause pancreatitis No masses were palpable. Laboratory studies re­ (Table 1). Chlorothiazide was introduced in 1958 vealed the following values: an amylase of 1131 UIL, and is an effective, widely prescribed, and gener­ lipase 11,063 UIL, glucose 204 mg/dL, lactate ally well tolerated antihypertensive agent. Soon dehydrogenase 483 UIL, and a white cell count of after the introduction of chlorothiazide, an asso­ 18,900//J.V. His calcium level was 10.1 mg/dL, ciation with pancreatitis was observed.s The phosphorus was 3.3 mg/dL, and triglycerides were patient presented here had two discrete episodes 276 mg/dL. of pancreatitis attributable to thiazide diuretics. He was admitted to a community hospital, where a nasogastric tube was inserted, intra­ Case Report venous fluids were initiated, and parenteral The patient was a 43-year-old man who had a meperidine was administered. An ultrasonic ex­ history of hypertension, gout, and pancreatitis of amination of his gallbladder was normal. On the http://www.jabfm.org/ unknown cause in 1985. At that time he was 2nd hospital day a computerized tomogram (CT) taking hydrochlorothiazide, triamterene, and showed severe necrotizing pancreatitis. His tem­ pindolol for hypertension. Findings on a gall perature was 101.1°F (38.4°C); at this time the bladder sonogram were normal, and no cause for laboratory values were white cell count 20,400//J.L, the pancreatitis was determined. The patient was amylase 298 UIL, lactate dehydrogenase 1776 prescribed only pindolol at discharge. In 1989 UIL, glucose 137 mg/dL, and calcium 7.8 mg/dL. on 27 September 2021 by guest. Protected copyright. the patient's antihypertensive medication was Blood and urine specimens were cultured, em­ changed to 400 mg of labetalol and 50 mg of piric cefoxitin therapy was initiated, and a clonid­ hydrochlorothiazide. In July 1991 the patient ine patch was applied. changed physicians, and the thiazide diuretic was On the 7th hospital day total parenteral nutri­ discontinued. In early 1992 the patient's blood tion was begun; the patient did not require insu­ pressure was not adequately controlled by lin. All cultures were negative. A second CT scan labetalol alone, and 50 mg of hydrochlorothiazide showed phelgmous changes in the peri pancreatic was again prescribed. WIthin 1 month he came to fat, with necrosis of the body of the pancreas and development of a 5 -cm fluid collection in the

Submitted, revised, 15 September 1993. lesser sac. Clinically, the patient improved, and From the Department of Family Practice, University of Michi­ his amylase level returned to normal. On the 10th gan Medical School, Ann Arbor. Address reprint requests to hospital day he had a severe episode of abdominal Raymond Rion, MD, University of Michigan Medical School, pain. A third CT scan showed further pancreatic Department of Family Practice, 1018 Fuller Street, Ann Arbor, MI 48109-0708. destruction and several cystic areas in the head

74 JABFP Jan.-Feb. 1994 Vol. 7 No.1 J Am Board Fam Pract: first published as 10.3122/jabfm.7.1.74 on 1 January 1994. Downloaded from

Table 1. Drugs Associated with Pancreatitis. discontinuation of the drug, and recurs when the 6 Definite Association Possible Association drug is re-1ntroduced. During treatment of his hypertension this patient was unintentionally Sulfonamides L-Asparaginase rechallenged with a thiazide diuretic and had a Thiazide diuretics Chlorthalidone second episode of pancreatitis. Positive challenge Furosemide Corticosteroids tests have been reported for the loop diuretic Estrogens Ethacrynic acid furosemide.7 No positive challenge tests with thi­ Tetracycline Phenformin azide diuretics could be found on the MEDLINE Pentamidine Procainamide data base. The intentional rechallenge of a patient Valproic acid Nonsteroidal anti-inflammatory should occur in a controlled research setting and agents has no diagnostic or therapeutic role in clinical Didanosine (DDI) Nitrofurantoin practice. Mercaptopurine Metronidawle There is experimental and clinical evidence Azathioprine that thiazide diuretics cause pancreatitis. In 1961 Cornish, et al. 8 studied amylase levels in 20 pa­ and body of the pancreas with no air fluid levels. tients who were taking 1 to 2 g of chlorothiazide Additional analgesics were administered and the a day and found elevated levels in 50 percent of pain subsided. the study subjects. The amylase levels increased On the 18th hospital day the patient developed more than 100 percent in the thiazide-treated increasing abdominal pain and his serum amylase patients. No patients exhibited signs or symptoms increased to 640 UIL. He was transferred to a of pancreatitis. Amylase levels returned to normal tertiary care center. A CT scan showed a 20-cm when the drug was discontinued. Chlorothiazide fluid collection around the pancreas. His white was administered to 300 mice 6 days a week at a cell count had increased to 31,0001 f.1L. The pa­ dosage of 0.02 gld (equivalent to 1 gld for a 70-kg tient underwent surgical drainage of the pseudo­ human). Twenty-one of the mice had inflamma­ cyst and debridement of the necrotic pancreas and tory pancreatic lesions, 15 animals had mild pan­ peripancreatic necrotic tissues. Forty-eight hours creatitis, and 7 had severe pancreatitis. The most later a second debridement was performed, and severe lesions were found in those mice taking the the patient was found to have a necrotic posterior drug the longest.8 More recently the medications gastric wall. He underwent a midgastrectomy and of 100 patients admitted for acute pancreatitis were cholecystostomy. A third debridement was per­ compared with medications of 100 matched patients http://www.jabfm.org/ formed, and a jejunostomy tube and a gastrostomy admitted for abdominal pain who had normal tube were inserted. amylase levels. Diuretic therapy was more com­ The patient improved and was discharged after mon in the pancreatitis group, and the difference a total hospitalization of 6 weeks. He was dis­ achieved statistical significance. The difference was charged with a gastrostomy tube, a jejunostomy due almost entirely to greater numbers of patients tube, a cholecystostomy, and three]ackson Pratt with pancreatitis taking cyclopenthiazide.9 on 27 September 2021 by guest. Protected copyright. drains. Five months after admission he underwent The first cases of pancreatitis associated with an antrectomy, reconstruction of gastrointestinal thiazide use were reported in 1959.5 More re­ continuity with Roux-en-Y gastrojejunostomy, cently a total of 13 cases of pancreatitis in nol,1- cholecystectomy, and closure of tube jejunos­ pregnant patients have been reported or re­ tomy. Two 4-cm pseudocysts were discovered and viewed. The average age in one series· was 68 drained. Six months after admission the patient years, and all patients had hypertension and had lost 25 percent of his body weight and had arteriosclerotic heart disease, and two-thirds had developed a pancreaticocutaneous fistula. congestive heart failure. 10 In the second series, the patients had hypertension but were free of other Discussion important illnesses. I I Patients developed pancrea­ A challenge test is the most reliable evidence that titis anywhere from 21 days to 5 years after a specific medication causes pancreatitis. The test beginning therapy with a thiazide. The mean dose involves documenting that pancreatitis develops was 250 to 1000 mg of chlorothiazide or the during treatment with the drug, disappears after equivalent of hydrochlorothiazide. Five of the 13

Pancreatitis and Hydrochlorothiazide 75 J Am Board Fam Pract: first published as 10.3122/jabfm.7.1.74 on 1 January 1994. Downloaded from

patients died from complications of pancreatitis. must be aware of the potential for thiazide di­ Two patients developed chronic pancreatitis dur­ uretics and other drugs to cause pancreatitis. It is ing therapy with a thiazide diuretic. In one case extremely important to obtain a complete medi­ the patient was also taking prednisone; which of cation history in patients seeking care for acute the two drugs was responsible for the pancreatitis pancreatitis. Physicians must be cognizant of the was not clear.12 The other case was a 9-year-old role medications can play in the cause of acute boy who died of rheumatic heart disease while pancreatitis and prescribe medications judiciously taking hydrochlorothiazide and was found at for patients who have a history of so-called idio­ autopsy to have pathologic changes of chronic pathic pancreatitis. pancreatitis. 13 As the present case illustrates, the clinical course of a patient with thiazide-induced References pancreatitis can be severe. Thiazides generally 1. Carey LC. Recurrent acute pancreatitis - rarely idio­ cause hemorrhagic or necrotic pancreatitis. The pathic: Dupontlecture. CanJ Surg 1990; 33:107-12. clinical appearance of a patient who develops pan­ 2. Grendell JH. Idiopathic acute pancreatitis. Gastro­ creatitis from thiazides is indistinguishable from enterol Clin North Am 1990; 843-8. that of patients with pancreatitis from other 3. Steer ML. Classification and pathogenesis of pancrea­ titis. Surg Clin North Am 1989; 69:467-80. causes. Other than discontinuing the diuretic, the 4. Thomson SR, Hendry WS, Mcfarlane GA, Davidson treatment of thiazide-induced pancreatitis is no AI. Epidemiology and outcome of acute pancreatitis. different from other cases of pancreatitis. BrJ Surg 1987; 74:398-401. The exact mechanism of thiazide-induced pan­ 5. Johnston DH, Cornish AL. Acute pancreatitis in creatitis is unknown. Several possibilities have patients receiving chlorothiazide. JAMA 1959; 170: been suggested. Thiazides are well known to in­ 2054-6. 6. Mallory A, Kern F Jr. Drug-induced pancreatitis: a crease serum triglyceride levels.'4 In some patients critical review. Gastroenterology 1980; 78:813-20. thiazides have been documented to increase serum 7. Stenvinkel P, Alvestrand A. Loop diuretic-induced calcium and decrease serum phosphorus levels. 14 pancreatitis with rechallenge in a patient with malig­ These alterations in calcium and phosphorus nant hypertension and renal insufficiency. Acta Med levels are also observed in hyperparathyroidism, Scand 1988; 224:89-91. 8. Cornish AL, McClellanJT, Johnston DH. Effects of and primary hyperparathyroidism is known to chlorothiazide on the pancreas. N Engl J Med 1961; increase the risk of pancreatitis. It has been found 265:673-5. at autopsy that some patients who developed pan­ 9. BourkeJB, McIllmurray MB, Mead GM, Langman creatitis during thiazide diuretic therapy have MJ. Drug-associated primary acute pancreatitis. http://www.jabfm.org/ parathyroid hyperplasia. I I Thiazides might act di­ Lancet 1978; 1:706-8. rectly on the parathyroids or might affect both the 10. Eckhauser ML, Dokler M, Imbembo AL. Diuretic­ associated pancreatitis: a collective review and illus­ parathyroid glands and the pancreas. I I trative cases. AmJ Gastroenterol1987; 82:865-70. After this patient's initial episode of pancreati­ 11. Pickle man J, Straus FH 2d, Paloyan E. Pancreatitis tis, no cause was determined. It has been reported associated with thiazide administration. A role for the that up to 20 percent of pancreatitis cases are parathyroid glands? Arch Surg 1979; 114:1013-6. on 27 September 2021 by guest. Protected copyright. idiopathic.4 Other sources have reported that 5 to 12. Weaver GA, Bordley J 4th, Guiney WB, D'Accurzio A. Chronic pancreatitis with cyst formation after 10 percent of cases are idiopathic.2,3 Careyl found prednisone and thiazide treatment. Am J Gastro­ that after a thorough examination few cases of enterol 1982; 77: 164-8. pancreatitis are, in fact, idiopathic. Before diag­ 13. Shanklin DR. Pancreatic atrophy apparently second­ nosing idiopathic pancreatitis in any patient, a ary to hydrochlorothiazide. N Engl J Med 1962; meticulous search for other possible causes must 266: 1097 -9. be conducted. 14. Weiner 1M. Diuretics and other agents employed in the mobilization of edema fluid. In: Gilman AG, Rail This case illustrates the clinical course of a TW, Nies AS, Taylor P, editors. Goodman and patient who had recurrent pancreatitis after treat­ Gilman's the pharmacological basis of therapeutics. ment with a thiazide diuretic. Family physicians 8th ed. New York: Pergamon Press, 1990:721.

76 JABFP Jan.-Feb. 1994 Vol. 7 No.1