A comparison of nonionic versus ionic contrast media: Results of a prospective, multicenter study

0016-5107/95/4204-031255.00 + 0 GASTROINTESTINAL ENDOSCOPY Copyright © 1995 by the American Society for Gastrointestinal Endoscopy

A comparison of nonionic versus ionic contrast media: results of a prospective, multicenter study G. Kenneth Johnson, MD, Joseph E. Geenen, MD, Rudolph A. Bedford, MD, John Johanson, MD Oliver Cass, MD, Stuart Sherman, MD, Walter J. Hogan, MD, Michael Ryan, MD William Silverman, MD, Steven Edmundowicz, MD, Mark Payne, MD, and other members of Midwest Pancreaticobiliary Study Group Background: Pancreatitis is one of the most common complications associated with ERCP. Multiple factors have been implicated for this potentially serious complication. Numerous suggestions for minimizing risks at ERCP have been offered, one of which is to use nonionic, low osmolarity contrast agents for pancreatic injection. Results of previous studies comparing different contrast media have been inconclusive. Methods: To evaluate the role contrast material plays in the development of post-ERCP pancreatitis, the Midwest Pancreaticobiliary Group performed a prospective double-blind controlled study. A total of 1,979 consecutive ERCP patients were enrolled, and 1,659 patients with pancreatic duct injections were divided into subgroups according to the complexity of the ERCP. Post-ERCP pancreatitis was compared between similar groups. Patients were randomized to receive injections of nonionic, low osmolarity contrast or standard ionic contrast media. Results: The overall incidence of post-procedural pancreatitis was 10.2%. Those with diagnostic ERCP had the lowest incidence at 5.6%. Therapeutic procedures (12.3%) and sphincter of Oddi manometry (15.2%) had higher rates. Those injected with standard (ionic) contrast had an incidence of 10.4% and after injection with lower osmolar (nonionic) contrast, there was a 10% postprocedural pancreatitis rate. Conclusions: Patients with more complex procedures develop pancreatitis more frequently. The use of low osmolar (nonionic) contrast media does not decrease the incidence of post-ERCP pancreatitis. (Gastrointest Endosc 1995; 42:312-6.) Since the initial reports of successful endoscopic cholangiopancreatography, different concentrations of radiocontrast media have been evaluated for optimal visualization of the pancreatic duct. 1-3 As physicians expanded the clinical role of endoscopic cholanReceived May 20, 1994. For revision July 18, 1994. Accepted November 14, 1994. From the Midwest Pancreaticobiliary Study Group: Racine, Milwaukee, Marshfield, Wisconsin; Rockford, Illinois; Minneapolis, Minnesota; Indianapolis, Indiana; Pittsburgh, Pennsylvania; St. Louis, Missouri; Kansas City, Kansas. Reprint requests: Joseph E. Geenen, MD, 1333 College Ave., Racine, WI 53403. 37/1/62572 312

GASTROINTESTINAL ENDOSCOPY

giopancreatography, the potential for invasive therapeutic procedures became apparent; however, the chances of an increased complication rate also raised concerns. 4-s Acute pancreatitis after ERCP was believed to be "the most feared complication. "6 Recommendations regarding techniques for increasing the safety of endoscopic pancreatography were suggested by Galvan and Klotz s with very little information to support them. One recommendation was to avoid injection of contrast material with concentrations above 30%. Concern about avoiding high concentrations of contrast media and hyperosmolar contrast materials in the pancreatic duct was not shared by a reviewing panel of experts in 1974. 9 VOLUME 42, NO. 4, 1995

Table 1. Results of controlled studies comparing incidence of pancreatic enzyme elevation and pancreatitis after ERCP with high-osmolarity (ionic) and Iow-osmolarity (nonionic) contrast agents 2s Source

Enzyme elevation, incidence

Pancreatitis, incidence

Author/year/ref.

High osm

Low osm

High osm

Low osm

Osnes et al., 197715 Hamilton et al., 198216 Hannigan eta]., 198517 Reimer-Jensen et al., 1985 is Makela et al., 198619 Banerjee et al., 198620 Cunliffe et al., 198721 O'Connor et al., 198823 Barkin et al., 199124 Silverman et al., 199125 Rodriguez et al., 199126 Total

ND 15/33 28/32 9/28 10/44 ?/25 16/48 23/25 2/15 ND 11/25 114/275

ND 8/17 25/28 4/26 7/44 ?/24 8/46 17/25 0/22 ND 5/23 74/255

0/12 0/33 0/32 0/28 0/44 5/25 8/48 0/25 3/15 9/81 2/25 27/368

0/12 0/17 1/28 0/26 0/44 0/24 2/46 0/25 1/22 11/75 1/23 16/342

ND, Not done.

Post-ERCP pancreatitis was described thoroughly in 1974, and a recommendation was made to avoid acinarization of the pancreatic parenchyma by the pioneer group in Germany. 10 This recommendation was based on a retrospective review of 569 cases undergoing pancreatic injections during a period of 2 years. The incidence ofpost-ERCP pancreatitis was variably reported during the 1970s with a wide range, from 1% 11 to 39.5% 12. However, many of these reports were based on retrospective data, and definitions of pancreatitis differed. Contrast media have been thought to be a source of irritation of the pancreatic ductal epithelium. Animal experiments were undertaken in an attempt to evaluate the possible association of contrast media with damage to ductal epithelia. Injury was believed to be less severe after the injection of nonionic low-osmolarity contrast media in the cat. 13 In contrast, Saari et al. showed acinar destruction and injury to be least apparent in the ducts of piglets after injection of the standard high-osmolarity contrast media. 14 Clinical studies from a large number of centers have attempted to compare a variety of contrast a g e n t s 15-26 (Table I). However, present data from animal and human studies have been conflicting and inconclusive. Ionic media of high osmolarity are the standard contrast agents injected for pancreatography. In other radiologic procedures, many studies have shown that nonionic, low-osmolarity contrast media produce less vascular and tissue irritation than ionic, high-osmolarity contrast materials. In an attempt to define further the role of contrast media in the development of post-ERCP pancreatitis, the Midwest Pancreaticobiliary Group undertook a prospective, double-blinded, randomized, controlled trial with a large number of consecutive ERCP patients. VOLUME 42, NO. 4, 1995

PATIENTS AND METHODS Physicians from 12 participating endoscopy centers affiliated with the Midwest Pancreaticobiliary Group studied a total of 1979 p a t i e n t s (Table 2). All patients h a d clinical indications for evaluation with ERCP. P a t i e n t s with histories of suspected allergic reaction to contrast m e d i a and 192 patients who were believed to be uncooperative were excluded. Informed consent was obtained before enrollment in this s t u d y as required by each center's Institutional Research Review Board. P a t i e n t s were assigned to control or study groups with cards n u m b e r e d consecutively a n d pulled in sequence by the endoscopy nurse at each institution before beginning the ERCP procedure. The physician performing the s t u d y was blinded to the type of contrast injected. The study group consisted of patients receiving a nonionic, lowosmolarity agent (iopamidol), which contains 300 mg/mL of nonionic, organically bound iodine (616 mOsm/L). The control group received an ionic high-osmolarity agent (diatrizoate meglumine) containing 282 mg/mL of ionic iodine (1500 mOsm/L). ERCP procedures were performed in the s t a n d a r d fashion using side-viewing duodenoscopes. Conscious sedation most frequently included meperifline and/or benzodiazepine. Clinical pancreatitis was defined before the study as a n increase in pain after the procedure, an increase in s e r u m a m y l a s e to more t h a n twice n o r m a l values, and a prolonged hospital stay of more t h a n 48 hours. The severity of pancreatitis was graded as mild for those patients with up to 3 additional hospital days, m o d e r a t e for those staying between 4 a n d 7 additional days, and severe for those who spent more t h a n 7 additional hospital days. All p a t i e n t s with prolonged hospital stays of more t h a n 7 days were evaluated with CT for complications of pancreatitis: abscess, pseudocyst, or pancreatic necrosis "phlegmon." S e r u m amylase levels were obtained from all patients 3 to 4 hours after completion of the procedure. P a t i e n t groups were subdivided as follows: those undergoing diagnostic p a n c r e a t o g r a p h y (Group I); those who h a d diagnostic p a n c r e a t o g r a p h y plus sphincter of Oddi m a n o m e t r y (Group II); those undergoing diagnostic p a n c r e a t o g r a p h y with or without cholangiograGASTROINTESTINAL ENDOSCOPY

313

Table 2. E R C P stratification

Groups (I.III) 1000 900 B00 700 600 500 400 300 200 100

86/830

83/829*

•

No pancreatitis

[ ] Pancreatitis

Ionic

Non-ionic *p > 0.05

Figure 1. Comparative incidence of post-ERCP pancreatitis in patients injected with ionic or nonionic contrast agent (Groups I, II, and III). Group IV patients, who received no pancreatic duct injection, are excluded.

phy plus therapeutic ERCP (Group III). The majority of patients in this group received biliary therapeutics. Patients who did not have pancreatography but had cholangiography with or without therapeutic procedures were classified as Group IV. Because the effect of contrast media upon the pancreas could not be evaluated in this group, they will be analyzed separately and reported in the future. Standard report forms for data were submitted by the 12 centers for review to one central location. In 25 patients, injections of either the pancreatic duct or the biliary ducts were unsuccessful; they were subsequently classified as "failed cases" and were eliminated from analysis (Table 2). The frequency of postprocedure pancreatitis was compared between patients receiving nonionic and those receiving ionic contrast media. Statistical comparison using the chi-squared test and t test of population proportions were used to determine significance. RESULTS

A total of 1979 patients were randomized to the study, and 25 cases were eliminated because of failure to inject either duct. The remaining 1954 cases were subdivided into groups and analyzed. Post-ERCP pancreatitis, as previously defined, occurred in 178 of 1954 patients, for an overall incidence of 9.1%. This includes the patients who did not have pancreatograms but who underwent biliary diagnostic study and/or biliary therapeutic procedures (Group IV). Because this group did not receive pancreatic duct injection, it will be analyzed and reported separately. The 1659 patients studied with pancreatograms were divided into subgroups I through III and are portrayed in Table 3. Of 308 patients undergoing diagnostic pancreatograms (Group I) who were injected with standard contrast (diatrizoate meglumine), pancreatitis developed in 13 (4.2%), and of 314 Group I patients receiving nonionic contrast (iopamidol), pancreatitis developed in 22 (7.0%). No statistical differ314 G A S T R O I N T E S T I N A L E N D O S C O P Y

Group

Type of procedure

No. of patients

I II III IV Failed Total

Diagnostic Sphincter of Oddi manometry Therapeutic Pancreatic duct not injected Neitherduct injected

622 211 826 295 25 1979

ence was noted in Group I, although a trend actually favoring the high-osmolarity, ionic contrast medium was seen (p = .07). Of 108 patients undergoing diagnostic study plus sphincter of Oddi manometry (Group II) who were injected with standard ionic contrast pancreatitis developed in 17, for a 15.7% incidence. In the 103 Group II patients injected with nonionic contrast pancreatitis developed in 15, for a 14.6% incidence. No statistical difference was demonstrated between different contrast media in this group. Group III comprised 826 patients who underwent diagnostic pancreatograms plus therapeutic procedures, such as endoscopic sphincterotomy (ES), stone extraction, stent placement, balloon dilation, or a combination of procedures. Most of these patients also had cholangiograms and therapeutic biliary procedures. Of 414 patients receiving standard ionic contrast, pancreatitis developed in 56, for a 13.5% incidence. Of 412 receiving nonionic contrast, pancreatitis developed in 46, for an 11.2% incidence. Again, no statistical difference was noted. Of the 1659 patients in Groups I, II, and III, postERCP pancreatitis developed in 169, for a combined incidence of 10.2%. Of 830 patients receiving standard, high-osmolarity, ionic contrast, pancreatitis developed in 86, for an incidence of 10.4%. Of 829 patients receiving low-osmolarity, nonionic contrast, pancreatitis developed in 83, for a 10% incidence. No statistical difference was noted between the two contrast agents in all patients studied with pancreatic duct injections (Groups I, II, and III; Fig. 1). The variation in the incidence of post-ERCP pancreatitis was statistically significant between Group I (5.6%) and Group II (15.2%) and between Group I (5.6%) and Group III (12.3%). Group I patients had only diagnostic procedures, which included injections into the pancreatic duct. Groups II and III had greater manipulation, with manometry study and/or therapeutic procedures. It appears t h a t the higher incidence of pancreatitis in Groups II and III is likely the consequence of the added manipulation t h a t occurs at or near the orifice of the pancreatic duct during the diagnostic manometric procedure and the more complex therapeutic procedures. The incidence of severe pancreatitis in patients VOLUME 42, NO. 4, 1995

Table 3. Incidence of post-ERCP pancreatitis associated with pancreatograms* High osmolarity (ionic)

Patient group No. % No. I-Dx 13/308 4.2 22/314 II-SOM 17/108 15.7 15/103 III-Therapeutic 56/414 13.5 46/412 Total 86/830 10.4 83/829 *Group IV not included because pancreatic duct not injected. tP > .05. studied with pancreatograms (Group I, II, and III) was as follows: Group I, 2/622 (0.3%); Group II, 3/211 (1.4%); Group IlI, 7/826 (0.8%). Within these groups, no significant difference was noted between the two types of contrast media. There were no deaths. Two patients with pancreatitis required surgery, one with severe hemorrhagic pancreatitis and one with mild pancreatitis for repair of a bile duct leak. Those with severe pancreatitis were hospitalized between 7 and 20 extra days (mean, 9).

DISCUSSION Since the early reports of post-ERCP pancreatitis and the recommendation to avoid acinarization by Classen's group in 1974,1° physicians have sought other explanations for the source of post-ERCP pancreatitis. A variety of endoscopic and clinical approaches have been tried in an attempt to decrease the incidence of this post-ERCP complication. 27-29 Irritation of the pancreatic duct epithelium by contrast media has been suggested as one potential source of pancreatic duct damage. H u m a n studies attempting to clarify the role played by radiocontrast within the pancreatic duct have been conflicting and inconclusive. All of these studies were compromised because the groups were not large enough to prevent statistical type II errors. 3° Previous reports from multiple different investigators also suffer from a lack of uniform protocol and data compilation (Table 1). Our study was designed to include a large number of patients in order to determine whether post-ERCP pancreatitis could be decreased or prevented by using nonionic, relatively low-osmolarity contrast media. Data from each of 12 centers in the Midwest Pancreaticobiliary Group revealed a total incidence of postERCP pancreatitis of 9.1%. The vast majority of these patients had mild or moderate degrees of pancreatitis. Only 13 of 1954 (0.66%), a small percentage of our patients, had severe pancreatitis31; 6 of these were injected with nonionic contrast and 7 with ionic contrast media. The total incidence of pancreatitis in all degrees of severity was similar between those receiving standard ionic contrast (9.4%) and those receiving nonionic contrast (8.9%). Of patients studied with V O L U M E 42, NO. 4, 1995

Incidence of pancreatitis

Low osmolarity (nonionic) % 7.0 14.6 11.2 10.0

p value .07 .41 .15 N.S.t

Combinedno. 35/622 32/211 102/826 169/1659

% 5.6 15.2 12.3 10.2

pancreatograms (Groups I, II, and III), post-ERCP pancreatitis developed in 10.2%. In these groups, no significant difference was noted between patients receiving standard ionic contrast (diatrizoate meglumine) and those injected with nonionic contrast (iopamidol) in regard to degree of severity or percentage with post-ERCP pancreatitis. The patients in the diagnostic group (Group I) showed the most variation in results. Those receiving lower-osmo]arity, nonionic contrast (iopamidol) showed a higher incidence of pancreatitis than those receiving the standard, ionic contrast (diatrizoate meglumine), although this difference was not statistically significant (Table 3). Considerable variance was observed in the occurrence of complicating pancreatitis between Groups I and II and between Groups I and III. In Group I, the incidence of pancreatitis was less than that in Groups II and III. It appears that the incidence ofpost-ERCP pancreatitis increases with additional diagnostic and therapeutic manipulations. The higher incidence of pancreatitis may be related to the number of injections and manipulations into and around the pancreatic duct. 32 As suggested by Wolf in 1986, 33 we have attempted to provide a clear medical basis with research directed at clarifying the choices of contrast agents currently available. Changes in the physical and chemical properties of contrast media may offer some advantage in the performance of ERCP in the future; however, current use of the nonionic, lower-osmolarity contrast media for routine ERCP study cannot be recommended. The added expense is not cost-effective for this procedure at this time. Continued evaluation of factors believed to predispose to the development of post-ERCP pancreatitis should be studied prospectively in the future with large groups of patients.

ACKNOWLEDGMENTS Members of the Midwest Pancreaticobiliary Study Group: G. Aliperti, St. Louis, Mo.; H. Parker, MD, and M. Schmalz, MD, Milwaukee, Wis.; J. Frakes, MD, and R. Greenlaw, MD, Rockford, Ill.; J. Goff, MD, Golden, Col.; R. Hawes, MD, and G. Lehman, MD, Indianapolis, Ind.; G. Mayeux, MD, Marshfield, Wis.; GASTROINTESTINAL ENDOSCOPY

315

J. V e n n e s , MD, a n d P. Y a k s h e , MD, M i n n e a p o l i s , Minn.

18.

REFERENCES 1. McCune W, Shorb P, Moscovitz H. Endoscopic cannulation of the ampulla of Vater: a preliminary report. Ann Surg 1968; 167:752-5. 2. Takagi K, Ikeda S, Nakagawa Y, et al. Retrograde pancreatography and cholangiography by fiber duodenoscope. Gastroenterology 1970;59:445-52. 3. Oi I, Kobayashi S, Kondo T. Endoscopic pancreatocholangiography. Endoscopy 1970;2:103-6. 4. Classen M, Koch H, Demling L. Duodenoscopy: methods and findings. Gastrointest Endosc 1971;18:78-80. 5. Classen M, Demling L. Endoskopische sphinkterotomie der papilla Vateri und stein extraktion aus dem ductus choledochus. Dtsch Med Wochenschr 1974;99:496-7. 6. Cotton P. Cannulation of the papilla of Vater by endoscopy and retrograde cholangiopancreatography (ERCP). Gut 1972;13: 1014-25. 7. Kawai K, Akasaka Y, Murakami K, Tada M, Kohli Y, Nakajima M. Endoscopic sphincterotomy of the ampulla of Vater. Gastrointest Endosc 1974;20:148-51. 8. Galvan A, Klotz A. Is transduodenal pancreatography ever contraindicated? Gastrointest Endosc 1973;20:28-30. 9. Shapiro R. Do's and don'ts at pancreatography. Gastrointest Endosc 1974;20:129. 10. Ruppin H, Amon R, Ettl W, Classen M, Demling L. Acute pancreatitis after endoscopic/radiological pancreatography (ERP). Endoscopy 1974;6:94-8. 11. Bilbao M, Dotter C, Lee T, Katon R. Complications ofendoscopic retrograde cholangiopancreatography (ERCP). A study of 10,000 cases. Gastroenterology 1976;70:314-20. 12. Brust R, Thomson A, Wensel R, Sherbaniuk R, Costopoulos L. Pancreatic injury following ERCP: failure of prophylactic benefit of TrasyloL Gastrointest Endosc 1977;24:77-9. 13. Bub H, Burner W, Riemann J, Stolte M. Morphology of the pancreatic duct epithelium after traumatization of the papilla of Vater or endoscopic retrograde pm,creatograph with various contrast media in cats. Scand J Gastroenterol 1983;18:581-92. 14. Saari A, Kivisaari C, Standertskjold-Nordenstam C, Brackett K, Schroder T. Experimental pancreatography: a comparison of three contrast media. Scand J Gastroenterol 1988;23:53-8. 15. Osnes M, Skjennald A, Larsen S. A comparison of a new nonionic (metrizamide) and a dissociable (metrizoate) contrast medium in endoscopic retrograde pancreatography (ERP). Scand J Gastroenterol 1977;12:821-5. 16. Hamilton I, Lintott D, Rothwell J, Axon A. Metrizamide as contrast medium in endoscopic retrograde cholangio-pancreatography. Clin Radial 1982;33:293-5. 17. Hannigan B, Keeling P, Slavin B, Thompson R. Hyperamy-

316 GASTROINTESTINAL E N D O S C O P Y

19. 20. 21. 22. 23.

24. 25.

26.

27. 28. 29. 30.

31.

32.

33.

lasemia after ERCP with ionic and non-ionic contrast media. Gastrointest Endosc 1985;31:109-10. Reimer-Jensen A, Malchow-Moller A, Matzen P, et al. A randomized trial of iohexol versus amidotrizoate in endoscopic retrograde pancreatography. Scand J Gastroenterol 1985;1: 83-6. Makela P, Dean P. The frequency of hyperamylasemia after ERCP with diatrizoate and iohexol. Eur J Radiol 1986;6:303-4. Banerjee A, Grainger S, Manners R, Thompson R. Safer endoscopic retrograde pancreatography? [Abstract] Gut 1986;27: A601. Cunliffe W, Cobden I, Lavelle M, Lendrum R, Tait N, Venables C. A randomized prospective study comparing two contrast media in ERCP. Endoscopy (suppl) 1987;19:201-2. Rambow A, Staritz M, Meyer zum Buschenfelde K. Contrast media for ERCP. Endoscopy 1988;20:126-7. O'Connor H, Ellis W, ManningA, Lintott D, McMahon M, Axon A. Iopamidol as contrast medium in endoscopic retrograde pancreatography: a prospective randomized comparison with diatrizoate. Endoscopy 1988;20:244-7. Barkin J, Casal G, Reiner D, Goldberg R, Phillips R, Kaplan S. A comparative study of contrast agents for endoscopic pancreatography. Am J Gastroenterol 1991;86:1437-41. Silverman W, Ruffolo T, Rogers D, Hawes R, Earle D, Lehman G. Post ERCP pancreatitis (PEP); frequency with ionic and non-ioniccontrast media in high risk patients [Abstract]. Am J Gastroenterol 1991;86:192. Rodriguez J, Stoltenberg P, Avots A, et al. A prospective randomized comparison of the safety and efficacy ofiohexo1180 and diatrizoate 60% as contrast agents for ERCP [Abstract]. Gastroenterology 1991;100:297. Cotton P, Lehman G, Vennes J, et al. Endoscopic sphincterotomy, complications and their management. An attempt at consensus. Gastrointest Endosc 1991;37:383-93. Sherman S, Lehman G. ERCP and endoscopic sphincterotomyinduced pancreatitis. Pancreas 1991;6:350-67. Thornton J, Axon A. Towards safer endoscopic retrograde cholangiopancreatography. Gut 1993;34:721-4. Freiman J, Chalmers T, Smith H, Kuebler R. The importance of beta type II error and sample size in the design and interpretation of the randomized control trial. N Engl J Med 1978;299: 690-4. Bedford R, Johnson G, Geenen J, et al. Does the use ofnon-ionic contrast prevent post-ERCP pancreatitis: results of a prospective multicenter study [Abstract]. Gastrointest Endosc 1993; 39:270. Johnson G, Cass O, Geenen J, et al. Post ERCP pancreatitis: a comparative study of patients with and without pancreatic duct injection at the time of ERCP [Abstract]. Gastrointest Endosc 1993;39:284. Wolf G. Safer, more expensive iodinated contrast agents: how do we decide? Radiology 1986;159:557-8.

VOLUME 42, NO. 4, 1995