PEDIATRICS Vol. 118 No. 4 October 2006, pp. 1660-1663 (doi:10.1542/peds.2006-1182)
ARTICLE |
The Spectrum of Valproic Acid–Associated Pancreatitis
Department of Pediatrics, Division of Gastroenterology, Medical College of Wisconsin, Children's Hospital of Wisconsin, Milwaukee, Wisconsin
 |
ABSTRACT |
|---|
 TOP ABSTRACT PATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
OBJECTIVE. Our goal was to characterize valproic acid–associated pancreatitis in children.
PATIENTS AND METHODS. The charts of all patients with pancreatitis (diagnosed by using strict criteria) associated with valproic acid during a 10-year period were reviewed. Clinical and laboratory results were abstracted.
RESULTS. Twenty-two patients with valproic acid–associated pancreatitis were seen during the study period. Symptoms were similar to those of patients with pancreatitis from other etiologies and included abdominal pain/tenderness (83%), vomiting/retching (74%), abdominal distention (30%), and fever/chills (26%). Valproic acid levels were in the therapeutic range in all but 1 patient. The mean duration of therapy before the onset of pancreatitis was 32 months. The serum lipase level was >3 times the reference value in all patients, but the serum amylase level was not significantly elevated in 31% of the patients tested. Imaging studies altered clinical management in only 1 patient. The length of stay was generally brief (mean: 8 days). Two patients died. Of the 5 patients who were rechallenged, 4 had relapses.
CONCLUSIONS. Valproic acid–associated pancreatitis does not depend on valproic acid serum level and may occur any time after the onset of therapy. The serum lipase level is more sensitive than the serum amylase level and should be obtained when pancreatitis is suspected. Early imaging studies did not change clinical management. Rechallenge with valproic acid is dangerous and should be avoided.
Key Words: pancreatitis • valproic acid
Abbreviations: VPA—valproic acid • CT—computed tomography
Since the first report of valproic acid (VPA)–associated pancreatitis by Batalden et al1 in 1979, there have been a number of case reports and small series of 
5 patients describing this entity in mixed populations of children and adults (reviewed in refs 2–4). Many of these cases were fatal or severe, suggesting the possibility of reporting bias. The largest single-institution report includes only 11 patients seen over a 6-year period.5
In a recent review of pancreatitis at the Children's Hospital of Wisconsin we found 14 cases of VPA-associated pancreatitis, including several children who were rechallenged with the drug.6 VPA was the most common cause of drug-induced pancreatitis and represented 7.6% of all cases of pancreatitis seen at our institution during the study period. To better characterize VPA-associated pancreatitis in children we reviewed these cases and an additional 8 cases seen at our institution and report our experience.
|
PATIENTS AND METHODS |
|---|
|
TOPABSTRACTPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The charts of all patients with pancreatitis at Children's Hospital of Wisconsin between January 1, 1996, and December 31, 2005, who were receiving VPA therapy were reviewed. As previously reported,6 the diagnosis of pancreatitis was accepted if the patient had symptoms compatible with the diagnosis of pancreatitis and in whom the serum amylase or lipase level was >3 times the upper reference limit or if there were imaging or operative findings consistent with the diagnosis of pancreatitis. Other etiologies such as trauma, choledocholithiasis, viral infections, genetic, and anatomic and metabolic defects were excluded clinically or by laboratory or radiographic studies.
Data collected included duration of VPA therapy, concurrent medications, length of illness before hospitalization or emergency department visit, and previous history of pancreatitis. Pertinent symptoms, laboratory studies including aspartate aminotransferase, alanine aminotransferase, and peak amylase and lipase levels, time until pancreatic enzymes normalized, comorbid conditions, length of stay, outcome, and recurrences were recorded. Laboratory testing was performed by the clinical laboratory of Children's Hospital of Wisconsin.
The Children's Hospital of Wisconsin Research and Publications Committee/Human Rights Review Board approved this study.
|
RESULTS |
|---|
|
TOPABSTRACTPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Twenty-two patients were identified as having had pancreatitis during the study period and were taking VPA. VPA was prescribed for seizures for 17 patients and as a mood stabilizer for 5 patients. Significant comorbid conditions were present or developed in 10 patients, including chronic renal failure (2), mononucleosis (1), Fanconi syndrome (1), hepatoblastoma (1), diabetes mellitus (1), cerebral palsy (1), panhypopituitarism (1), hydrocephalus (1), Asperger syndrome (1), and bipolar disorder (1).
The mean age at diagnosis was 9.9 years (median: 9.5; range: 2–18 years). Males and females were equally affected. Symptoms at presentation included abdominal pain/tenderness (19 of 23 [83%]), vomiting/retching (17 of 23 [74%]), abdominal distention (7 of 23 [30%]), and fever/chills (6 of 23 [26%]).
All 22 patients had elevation of their serum amylase and/or lipase levels of >3 times the reference values. The mean serum amylase level was 373 IU/L (range: 14–713 IU/L; reference: <81 IU/L). The serum amylase level was within the reference range or <3 times the reference value in 9 (39%) patients; the level was not obtained for 1 patient. The mean serum lipase level was 6580 IU/L (range: 947–17874 IU/L; reference: <300 IU/L). The serum lipase level was elevated in all patients. In the patients with renal failure, the serum lipase levels were above the range typically seen in those with renal failure without pancreatitis (2663 and 6789 IU/L, respectively). Because the pancreatic enzymes had not normalized at the time of discharge in many of the patients, the length of time until normalization could not be determined. Although either the aspartate aminotransferase or alanine aminotransferase level was elevated in 7 (30%) of the 22 patients, it was >3 times the reference value in only 1 patient.
Nineteen patients had 
1 imaging study. Findings on the 15 patients who had computed tomography (CT) scans were acute pancreatitis (8), free fluid (2), pseudocyst (1), ruptured appendix with abscess (1), buried bumper syndrome with normal pancreas (1), normal (1), and pancreatitis followed by pseudocyst on follow-up examination (1). Findings on the 9 children who had ultrasonographic evaluation included pancreatitis (4), normal (2), free fluid (1), prominent common bile duct (1), and normal after pancreatitis seen on a previous CT scan (1). Four patients had both CT and ultrasonographic evaluations, but they were performed within 24 hours in only 1 patient; thus, the comparative utility of each study could not be assessed. Magnetic resonance cholangiopancreatography and endoscopic retrograde cholangiopancreatography demonstrated acute pancreatitis in 1 patient each. Three patients had no imaging studies performed on them. Pancreatic imaging results were abnormal for both patients with renal failure. Except in the patient with a pseudocyst seen on a delayed imaging study, treatment was not altered by the results of the radiologic studies.
The mean VPA dosage at the time of diagnosis was 45 mg/kg per day (range: 14–88 mg/kg per day). The mean duration of VPA therapy before the onset of pancreatitis was 32 months (median: 26 months; range: 2–84 months). The mean VPA level at diagnosis was 73 µg/mL (range: 5–207 µg/mL; therapeutic range: 50–150 µg/mL). The VPA level was above the therapeutic range in only 1 patient. Of the 5 patients who received VPA after recovery from the acute event, 3 experienced recurrences and 1 had 3 subsequent admissions for vomiting, but pancreatic enzyme measurements were not performed. One patient died during a subsequent episode of pancreatitis. Two (9%) had a recurrence even though VPA was discontinued; 1 of these 2 patients had polycystic kidney disease and was on dialysis.
The length of hospitalization typically was brief (median: 6 days; range: 1 day to 5.5 months). Fourteen patients were discharged from the hospital in 8 days. Of 22 patients, 21 were discharged in 34 days. One patient was hospitalized for 5.5 months. Prolonged hospitalization was typically for comorbid conditions such as sepsis. Two patients (9%) died, one as a result of necrotizing pancreatitis and the other as a result of septic shock 6 days after admission. One patient who died had panhypopituitarism.
|
DISCUSSION |
|---|
|
TOPABSTRACTPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
VPA, one of the most widely used anticonvulsants in children, is possibly the most common cause of drug-induced pancreatitis in North American children.6 Since the first report by Batalden et al1 in 1979, multiple case reports and small series, including 3 reviews, have been published, with considerable overlap in cases.2–4 Many of the publications included both children and adults. These case reports were undoubtedly highly selective, which introduced the possibility of reporting bias.
In 1993 Asconape et al2 reported 3 cases and reviewed 24 from a survey of physicians "with a special interest in epilepsy" and 12 cases from the literature. The mean age of their patients was 16.4 years; 9 patients were adults. The definition of pancreatitis was unclear, because the required degree of elevation of amylase and/or lipase levels was not specified. Most patients developed pancreatitis within the first 6 months of treatment, although 18.8% of the patients developed pancreatitis after >2 years of VPA therapy. Seventy-six percent of the patients were receiving multiple drugs. There was no relationship between VPA dose and the development of pancreatitis. Six patients had hemorrhagic pancreatitis. Of 9 patients who were rechallenged, 6 relapsed. There were 3 deaths (12.5%).
Chapman3 reviewed 45 adult and pediatric patients from 31 publications. According to their table, 20 of these patients were children. The diagnosis of pancreatitis was not standardized or reported in this review. Some patients did not have measurements of the serum amylase or lipase level; others had values within the reference range, making the diagnosis of pancreatitis questionable. Fifty-eight percent of the patients developed pancreatitis within 1 year of starting VPA, and 13 patients (29%) died.
Yazdani et al4 reported 2 cases of fatal acute pancreatitis in adults receiving VPA therapy and reviewed 36 cases of VPA-induced pancreatitis from the literature including 9 fatal cases (25%). There is considerable overlap between these cases and those reviewed by Asconape et al2 and Chapman et al.3 Not all patients had serum levels of pancreatic enzymes measured, and in some patients they were within the reference range again, raising the question of whether they all had pancreatitis. Ten patients were rechallenged, and 6 developed pancreatitis after the rechallenge. There was no relationship between VPA dose and the development of pancreatitis.
Sinclair et al5 reported 11 children, aged 4 to 16 years, with VPA-induced pancreatitis seen in a 6-year period. The level of elevation of the serum lipase level that was required for the diagnosis of pancreatitis was not provided. In the 3 patients who did not have hyperlipasemia (>3 times the reference value), 2 had an abnormal pancreas on imaging. In the third patient, imaging was not obtained. The latter patient had abdominal pain for 2 weeks and may not have had pancreatitis. There was no association between dose of VPA, serum level, length of treatment, use of other medications, and the development of pancreatitis. All patients recovered after withdrawal of VPA. No patients were rechallenged. There were no deaths.
We reviewed pancreatitis in 22 children receiving VPA during a 10-year period in our institution. Because we used accepted, strict inclusion criteria, we may have not included all cases.6 Although the serum lipase levels were >3 times the reference value in all patients, the serum amylase levels was <3 times the reference value in 9 patients. Thus, serum lipase was more sensitive than serum amylase in our patients.
Significant comorbid conditions occurred in 10 of our patients, including renal failure and diabetes, conditions known to be associated with benign hyperamylasemia and with pancreatitis. There was no relationship between the occurrence of pancreatitis and duration of VPA therapy, dosage, serum level, and the use of other anticonvulsants. VPA-associated pancreatitis was not benign, because 2 patients died. Of the 5 patients rechallenged with VPA, 3 developed documented pancreatitis. A fourth patient had symptoms compatible with recurrent pancreatitis, but serum levels of pancreatic enzymes were not measured.
The serum amylase level may be elevated in some patients receiving VPA without the development of pancreatitis. For example, Bale et al7 reported that 12 (20%) of 61 patients receiving VPA developed elevated serum amylase levels during a 1-year period of monitoring, only 1 of whom developed pancreatitis. In another study, 32 (23%) of 134 patients receiving VPA developed hyperamylasemia, none of whom developed pancreatitis. Of their patients not receiving VPA, 5.6% developed hyperamylasemia.8 There is no published information on serum lipase levels in patients receiving VPA. Thus, it is important that in a patient receiving VPA, the diagnosis of pancreatitis not be based on the finding of an elevated serum amylase level without a compatible clinical picture.
Although the fatality rate in our series was higher (9%) than in our total pancreatitis population (5%), it was lower than that of previous studies. Our mortality rate was lower than that found in most other series, suggesting the possibility of publication bias in the case reports and small series in favor of more severe cases. Alternatively, these numbers may not be meaningful because of the small number of patients.
The serum VPA levels in 21 of 22 patients at our institution were in the therapeutic range. The duration of therapy had been 2 to 84 months, similar to previous publications. Thus, the diagnosis of pancreatitis must be considered in any patient receiving VPA regardless of the duration of therapy.
We found 2 patients per year, which is only a slightly higher rate than that found by Sinclair et al.5 VPA-associated pancreatitis may be underdiagnosed because of lack of awareness of the condition and the fact that the serum lipase level is far more likely to be elevated than the serum amylase level and should be the diagnostic test of choice because, in our series, the serum amylase level was nondiagnostic in 39% of the patients. In many hospitals, the serum amylase level is the test of choice when pancreatitis is being considered
Although results of radiologic studies were typically abnormal in the 15 patients imaged, in only 1 patient did they alter the medical management. This patient developed a pseudocyst on a delayed study. Our findings are consistent with adult guidelines, that early imaging does not alter the management or outcome in a first episode of acute pancreatitis, unless choledocholithiasis is suspected, pancreatitis fails to resolve, or severe pancreatitis develops.9 There are no published guidelines for imaging children with pancreatitis.
VPA-associated pancreatitis, the most common cause of drug-associated pancreatitis in our population, should be considered whenever a child receiving VPA develops symptoms including abdominal pain, fever, vomiting, distension, or fever. The serum lipase level is more sensitive than the serum amylase level. A patient with VPA-associated pancreatitis should never be rechallenged, because the relapse rate is high. VPA-associated pancreatitis may have a higher fatality rate than pancreatitis of other causes in children.
|
FOOTNOTES |
|---|
Accepted May 23, 2006.
Address correspondence to Steven L. Werlin, MD, Department of Pediatrics, Medical College of Wisconsin, 8701 Watertown Plank Rd, Milwaukee, WI 53226. E-mail: swerlin{at}mcw.edu
The authors have indicated they have no financial relationships relevant to this article to disclose.
These data were presented in part in poster form at the annual meeting of the North American Society for Pediatric Gastroenterology, Hepatology, and Nutrition; October 22, 2005; Salt Lake City, UT.
Dr Fish's current address is Department of Pediatrics, Marshfield Clinic, 100 N Oak Ave, Marshfield, WI 54449.
|
REFERENCES |
|---|
|
TOPABSTRACTPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
- Batalden PB, Van Dyne BJ, Cloyd P. Pancreatitis associated with valproic acid therapy.
Pediatrics. 1979;64
:520
–522
[Abstract/Free Full Text] - Asconape JH, Kiffin Penry J, Dreifuss Fritz E, Riela A, Mirza W. Valproate-associated pancreatitis. Epilepsia. 1993;34 :177 –183[CrossRef][Web of Science][Medline]
- Chapman SA, Wacksman GP, Patterson BD. Pancreatitis associated with valproic acid: a review of the literature. Pharmacotherapy. 2001;21 :1549 –1560[CrossRef][Web of Science][Medline]
- Yazdani K, Lippmann M, Gala I. Fatal pancreatitis associated with valproic acid, review of the literature. Medicine (Baltimore). 2002;81 :305 –310[CrossRef][Medline]
- Sinclair DB, Berg M, Breault R. Valproic acid-induced pancreatitis in childhood epilepsy: case series and review.
J Child Neurol. 2004;19
:498
–5502
[Abstract/Free Full Text] - Werlin SL, Kugathasan S, Cowan Frautschy B. Pancreatitis in children. J Pediatr Gastroenterol Nutr. 2003;37 :591 –595[Web of Science][Medline]
- Bale JF Jr, Gay PE, Madsen JA. Monitoring of serum amylase levels during valproic acid therapy. Ann Neurol. 1982;11 :217 –218[CrossRef][Web of Science][Medline]
- Jimenez-Rodriguezvila M, Caro-Paton A, Conde M, et al. Side-effects of sodium valproate, mainly related to its hepatic and pancreatic toxicity. Int J Clin Pharmacol Res. 1986;6 :217 –224[Web of Science][Medline]
- Munoz-Bongrand N, Panis Y, Soyer, P, et al. Serial computed tomography is rarely necessary in patients with acute pancreatitis: a positive study in 102 patients. J Am Coll Surg. 2001;193 :146 –152[CrossRef][Web of Science][Medline]
PEDIATRICS (ISSN 1098-4275). ©2006 by the American Academy of Pediatrics
CiteULike 
Connotea 
Del.icio.us 
Digg 
Facebook 
Reddit 
Technorati 
Twitter What's this?
eLetters:
Read all eLetters
- Vaproate is an epigenetic modulator
- Heikki Savolainen
- Pediatrics Online, 5 Oct 2006 [Full text]
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
