INTRODUCTION
Endoscopic retrograde cholangiopancreatography (ERCP), which is widely available for the diagnosis and treatment of pancreaticobiliary diseases, is one of the most complex endoscopic procedures. A reported complication rate after ERCP varies between 5% and 10% (1). Precise identification of risk factors for ERCP complications is important for improving the safety of ERCP.
Acute pancreatitis remains the most common and serious complication after ERCP with the reported incidence ranging from 1.3% to 15.1% in most prospective series (2,3,4,5,6,7,8). Both patient- and procedure-related factors may be responsible for the pathogenesis of post-ERCP pancreatitis (1). Patient-related factors that have been reported include younger age, female sex, suspected sphincter of Oddi dysfunction (SOD), and previous post-ERCP pancreatitis (2,3,5,7,8). In addition, some procedure-related factors such as difficult cannulation, needle-knife precut, frequent injections into the pancreatic duct, and pancreatic sphincterotomy have been identified (2,3,5,7). However, the reported risk factors vary widely from study to study. These discrepancies may be attributable to heterogeneous patient populations, different levels of endoscopic expertise, different cannulation techniques, and disparate definition of post-ERCP pancreatitis among the studies.
Post-ERCP hyperamylasemia is associated with injury of the pancreatic parenchyma after ERCP, and manifests asymptomatically or with acute pancreatitis. Asymptomatic hyperamylasemia is more common than acute pancreatitis after ERCP, occurring in 7.7–70% of procedures (9,10,11,12). Previous studies have suggested that factors leading to either acute pancreatitis or asymptomatic hyperamylasemia after ERCP are similar except for age and gender (9,10,11). As acute pancreatitis and asymptomatic hyperamylasemia after ERCP exhibit completely different clinical manifestation and prognosis, the etiology of and the factors predisposed to these two clinical courses may be different. Identification of the difference in the risk factors between the two clinical courses may give a clue to further understanding the underlying mechanisms.
The aim of this study was to identify the independent risk factors for ERCP complications and to analyze whether the risk factors are different for post-ERCP pancreatitis and asymptomatic hyperamylasemia.
METHODS
This was a prospective, multicenter study conducted at 14 centers in China, of which 8 were teaching hospitals affiliated with universities, 2 were municipal hospitals, and 4 were provincial hospitals. The protocol was approved by the ethics committees at all participating centers.
Patients
During the periods between May 2006 and April 2007 (the period varied from 3 to 12 months among different centers), participating centers enrolled consecutive ERCP procedures, which included all procedures where an endoscope was inserted with the intent of cannulating the bile, the pancreatic duct, or both. Patients in whom the papilla of Vater was not reached and those who underwent stent removal procedures without ductal cannulation were excluded. Patients with high serum amylase levels before ERCP were also excluded.
Study protocol and data collection
Baseline values for liver function, serum amylase, and a complete blood count were determined before the procedure. Serum amylase levels were evaluated at 3 h and 24 h and a complete blood count was evaluated at 24 h after the procedure. All enrolled patients required at least 2 days of hospitalization after ERCP, even if complications were not present. Prolongation of hospitalization was at the discretion of the physician. Patient characteristics, including gender, age, a history of surgical intervention, and a history of complications after ERCP, were collected before ERCP. Details of ERCP, including ERCP indications and findings, cannulation techniques, duct opacification, sphincterotomy and other therapeutic procedures were recorded at the time of the procedure. All patients were followed up until hospital discharge to assess their clinical conditions including symptoms such as abdominal pain and to monitor the development of complications after ERCP. Data were recorded prospectively by the endoscopists and physicians in charge of the patients on a detailed data collection sheet.
Definitions
In this study, "suspected sphincter of Oddi dysfunction" referred to the pre-ERCP suspicion of a functional or structural abnormality of the sphincter of Oddi, independent of any manometric findings, thought to be the cause of recurrent abdominal pain or pancreatitis (13). Chronic pancreatitis was defined as the presence of mild, moderate, or severe pancreatographic abnormalities according to Cambridge criteria (14). Bile duct diameter was the measured maximal duct diameter within 2 cm of the papilla adjusted for magnification. If any stricture was present, bile duct diameter was measured distal to the stricture. Bile duct was defined to be dilated when its diameter is more than 1 cm. Cannulation time was measured between the time when the cannula or papillotome was advanced out of the endoscope channel, in front of the papilla, and the time when successful deep cannulation was evidenced by injection of contrast or the time when the procedure was abandoned because of unsatisfactory cannulation. Cannulation was judged difficult when an attempt took place for up to 10 min. Number of injections with the pancreatic duct contrast was the total number of times when any volume of the contrast was injected into a pancreatic duct. Acinarization was defined as any fluoroscopically observed focal or diffuse pancreatic parenchymal blush of contrast. Transpancreatic precut is one of precut techniques that a standard traction papillotome is wedged into the pancreatic orifice, with a cutting wire aimed in the biliary direction. Whether transpancreatic precut underwent pancreatic stenting was at the discretion of the endoscopists.
Definitions of individual complications after ERCP were similar to those reported by Cotton et al. (15). Post-ERCP pancreatitis was diagnosed when new-onset or worsened abdominal pain lasted for more than 24 h, and was associated with an increase in serum amylase level of at least three times greater than the normal upper limit at 24 h after the procedure. The severity of the pancreatitis was graded mild when hospitalization was prolonged by 2–3 days, moderate by 4–10 days, and severe by more than 10 days; however, the severity was graded as severe when a necrosis, abscess, or pseudocyst lesion occurred, or there was a need for percutaneous drainage or surgery. Hemorrhage was considered as clinical, as well as endoscopic, evidence of hemorrhage when it was associated with a decrease in the hemoglobin level by greater than 2 g/dl. Perforation referred to perforation of retroperitoneum or bowel walls documented by any of radiographic techniques. Cholangitis was defined as an elevation in the temperature to more than 38 °C because of a biliary cause without evidence of other concomitant infections. Acute cholecystitis was considered a complication when emergency cholecystectomy was required after ERCP but there were no suggestive clinical signs of cholecystitis before the procedure. Hyperamylasemia was defined as an increase in serum amylase level of at least three times greater than the normal upper limit at 24 h after the procedure, irrespective of whether symptoms were present.
Statistical analysis
The data were entered into a database program (Access; Microsoft, Redmond, WA, USA), and then tabulated with spreadsheet software (Excel; Microsoft) and analyzed with a statistical package (SPSS 15.0, SPSS, Chicago, IL, USA). Because those multiple ERCP procedures performed on the same patients could not be considered to be independent of one another, analyses were based only on the first recorded procedure for each patient, to control for potential correlations. The association between the successful deep cannulation rate and case volumes of individual endoscopists was tested using 
2-test or Fisher's exact test when necessary. A total of 22 potentially relevant risk factors, based on previously published prospective studies (16), were assessed using univariate analysis with 2-test, with Yates' correction or Fisher's exact test when necessary.
All variables were further included in a forward stepwise regression multivariate analysis. A binary logistic regression was used to identify the independent risk factors of overall complications. A multinomial logistic regressions model was used to identify the predictors of post-ERCP pancreatitis or asymptomatic hyperamylasemia when the uneventful procedures were defined as the reference category, and also to determine the different risk factors between pancreatitis and asymptomatic hyperamylasemia when the procedures with asymptomatic hyperamylasemia were defined as the reference category. Statistical significance was indicated by a P value of less than 0.05.
RESULTS
The mean number of ERCP procedures performed at each center ranged from 2.1 to 10.0 per endoscopist per week. Overall, 3,178 procedures were performed on 2,691 patients. Of the 2,691 consecutive patients scheduled to undergo ERCP, 1,238 (46.00%) were women and 1,453 (54.00%) were men, with a mean (
s.d.) age of 57.88 (17.43) years. Indications for ERCP are shown in Table 1. Within 2,691 diagnostic (n=62, 2.30%) or therapeutic (n=2,629, 97.70%) procedures represented the patients' first-only ERCP, 2,597 (96.51%) achieved deep cannulation of the desired duct. The relationship between the biliary or pancreatic duct deep cannulation rate and case volumes of individual endoscopists is shown in Table 2.
Table 2 - The successful deep cannulation rate of the first-only ERCP in relation to different case volumes of individual endoscopists.
Full table On the basis of the first-only recorded ERCP, complications developed in 213 (7.92%) patients. ERCP-induced pancreatitis occurred in 116 (4.31%); mild in 95 (3.53%), moderate in 20 (0.74%), and severe in 1 (0.04%). The case with severe pancreatitis occurred in a patient with suspected SOD who underwent major papilla pancreatic sphincterotomy, and required 30 days of hospitalization with medical treatments including fluid replacement, gastrointestinal decompression, and nasojejunal feeding (Table 3). Post-ERCP hyperamylasemia occurred in 512 (19.03%) patients, 116 with post-ERCP pancreatitis and 396 asymptomatic (Table 3). ERCP-related deaths occurred in seven patients with malignant biliary obstruction, three died from cholangitis, two from hepatic encephalopathy with cirrhosis, one from pulmonary infection, and one from acute renal failure (Table 3). All other complications after ERCP are presented in Table 3.
Analyses of risk factors were performed only for overall post-ERCP complications, pancreatitis and asymptomatic hyperamylasemia; the numbers of other complications were too small to be further analyzed in our series.
Univariate analysis
Sphincter of Oddi manometry was not performed in any case in the present series. Biliary orifice balloon dilatation was performed in 17 cases, in which post-ERCP pancreatitis or hyperamylasemia did not occurred. Thus, these two factors were excluded from the analysis. Precut sphincterotomy was used to attempt to access to the bile duct, and all specific precut sphincterotomy techniques in this study were performed including needle-knife precut and transpancreatic precut. A total of 22 variables including 10 patient-related factors and 12 procedure-related factors were evaluated.
Of the 22 factors, 8 were found to be significantly associated with overall complications after ERCP in the univariate analysis; 2 patient-related factors and 6 procedure-related factors (Table 4). Significant patient-related risk factors were female gender and periampullary diverticulum. Significant procedure-related risk factors included cannulation time>10 min, 
1 pancreatic deep wire pass, needle-knife precut, transpancreatic precut, 2 pancreatic contrast injections, and major papilla pancreatic sphincterotomy.
Table 4 - Risk factors for overall complications after ERCP in the univariate and multivariate analyses.
Full table Of the 22 factors, 12 were found to be significantly associated with post-ERCP pancreatitis in the univariate analysis; 4 patient-related factors and 8 procedure-related factors (Table 5). Significant patient-related risk factors included female gender, age 
60 year, suspected SOD, and distal common bile duct diameter 1 cm. Significant procedure-related risk factors included cannulation time>10 min, transpancreatic precut, needle-knife precut, acinarization or opacification of pancreatic branch duct, 2 pancreatic contrast injections, 1 pancreatic deep wire pass, major papilla pancreatic sphincterotomy, and pancreatic stent placement.
Table 5 - Univariate analysis of potential risk factors for post-ERCP pancreatitis or asymptomatic hyperamylasemia.
Full table Of the 22 factors, 10 were found to be significantly associated with asymptomatic hyperamylasemia in the univariate analysis; 3 patient-related factors and 7 procedure-related factors (Table 5). Significant patient-related risk factors included age 60 year, suspected SOD, and distal common bile duct diameter 1 cm. Significant procedure-related risk factors included cannulation time>10 min, transpancreatic precut, needle-knife precut, 2 pancreatic contrast injections, 1 pancreatic deep wire pass, major papilla pancreatic sphincterotomy, and pancreatic stent placement.
Multivariate analysis
The results of the forward stepwise binary logistic regression from the pool of 22 potential risk factors for overall complications are shown in Table 4. Five risk factors were identified to be independently associated with overall complications by multivariate analysis, two were patient-related factors (i.e., female gender and periampullary diverticulum), and three were procedure-related factors (i.e., cannulation time>10 min, 1 pancreatic deep wire pass, and needle-knife precut).
The results of the forward stepwise multinomial logistic regression from the pool of 22 potential risk factors for pancreatitis or asymptomatic hyperamylasemia compared with the uneventful procedures are shown in Table 6. Five risk factors were identified to be independently associated with pancreatitis by multivariate analysis, two were patient-related factors (i.e., female gender and age 60 year), and three were procedure-related factors (i.e., cannulation time>10 min, 1 pancreatic deep wire pass, and needle-knife precut). Four risk factors were identified to be independently associated with asymptomatic hyperamylasemia, all were procedure-related factors (i.e., cannulation time>10 min, 1 pancreatic deep wire pass, needle-knife precut, and major papilla pancreatic sphincterotomy).
Table 6 - Multinomial logistic regression for the predictors of post-ERCP pancreatitis or asymptomatic hyperamylasemia compared with the uneventful procedures.
Full table The multinomial logistic regression model from a pool of 22 potential factors for pancreatitis compared with asymptomatic hyperamylasemia demonstrated that only female gender (adjusted odds ratios (ORs): 1.61, 95% confidence interval: 1.06–2.45, P=0.026) was the different risk factor between pancreatitis and asymptomatic hyperamylasemia.
DISCUSSION
ERCP has evolved to become an essential therapeutic modality for a variety of pancreatic and biliary diseases since the introduction of endoscopic sphincterotomy (17). With the development and increased availability of alternative imaging techniques, such as multidetector helical CT scanners, magnetic resonance cholangiopancreatography and endoscopic ultrasound, diagnostic ERCP have become less commonly performed in recent years. The National Institutes of Health in 2002 recommended that diagnostic ERCP should be avoided whenever possible (18). In our study, the proportion of diagnostic procedures was only 2.30%, which is lower than previous prospective large-scale risk factor studies where the diagnostic procedures ranged from 22.9% to 48.1% (3,4,5,6,7,8).
A number of prospective multicenter studies have helped us to understand patient- and procedure-related risk factors (2,3,4,5,6,7), thus diagnostic ERCP in high-risk patients and high-risk ERCP techniques were avoided as possible in our study, but the incidences of overall post-ERCP complications and pancreatitis are still largely in line with that reported in other series (2,3,6,19). Transient hyperamylasemia, which is a manifestation of injury of the pancreas and sometimes not associated with any clinical signs, is not usually recognized in previous studies (20). Thus, our study is thought to be different from other studies in two respects: first, some unrecognized risk factors, such as pancreatic deep wire pass and transpancreatic precut, were estimated, and second, risk factors for asymptomatic hyperamylasemia were examined.
Earlier studies have shown that patients with suspected SOD carry a substantial risk for pancreatitis after any type of ERCP, whether diagnostic, manometric, or therapeutic (21). In a meta-analysis, a suspicion of SOD quadrupled the risk of post-ERCP pancreatitis (OR: 4.09) (22). However, suspected SOD was a significant risk factor for pancreatitis only by univariate but not by multivariate analysis in the present study. The discrepancy in the results would be related to the two facts that the proportion of patients with suspected SOD in our study population was only 5.4% and that the procedures for suspected SOD were mostly performed by endoscopists with higher case volumes in university hospitals.
Younger age was significant risk factor for pancreatitis in our study and in four other studies using multivariable analysis (4,5,7,8). However, a different age cutoff was used in these studies with per 5 years in one and 70 years in one, 60 years in three studies, respectively. The progressive decline in pancreatic exocrine function with aging may protect older patients from pancreatic injury (23). Female patients undergoing ERCP were found to be at 1.84 times higher risk of developing pancreatitis, compared with men when evaluated by multivariate analysis in the present study, which confirms the findings in two other prospective studies where the rates of pancreatitis were significantly higher in women than in men (3,8). In addition, we found that female patients were at 1.52 times higher risk of overall complications, which was not identified in previous studies (16). Periampullary diverticulum was also found to be an independent risk factor for overall complications, but not for pancreatitis, probably because this factor was related to other ERCP complications such as hemorrhage and perforation, but not pancreatitis.
Case volume of endoscopists was also evaluated in our study and it was found not to be a significant factor affecting successful deep cannulation, overall complications, and pancreatitis. These results may be explained by the fact that all procedures were performed by endoscopists with a higher case volume (2 per week). The case volume of endoscopists varied significantly across studies. Freeman et al. (2) reported the case volume ranged from 0.2 to 16 per endoscopist per week and found that a lower case volume (<1 per week) was a significant risk factor for complications only by univariate but not by multivariate analysis. Williams et al. (8) classified the case volume with <50, 50–100, 100–150, 150–200, and>200 per year, but failed to observed any association between the case volumes and complications including pancreatitis by multivariate analysis.
Technical variables are of obvious importance in overall post-ERCP complications and pancreatitis. Difficult cannulation, which can produce papillary trauma, has been shown to be an independent risk factor for complications and pancreatitis (2,3,6,19). Although the difficulty of cannulation was all judged by the number of cannulation attempts in previous studies, the cutoff numbers vary widely, with six attempts in two studies, eight attempts in one study and twenty attempts in one study (2,3,6,7). Of the four prospective studies, three have shown that difficult cannulation was an independent risk factor for pancreatitis in the multivariate analysis. A cannulation attempt is not easily quantifiable and time for cannulation may be recorded more exactly. In this study, the cutoff time of difficult cannulation was defined with 10 min, and cannulation time above this cutoff time did increase the risk of complications and pancreatitis by multivariate analysis (OR: 1.51 and 1.76, respectively).
Guidewires can be used to obtain deep cannulation of the desired duct under fluoroscopic guidance, and to perform wire-guided sphincterotomy, stent placement, and other therapeutic techniques (24). Wire-guided cannulation techniques were not widely utilized until more flexible and atraumatic guidewires have been developed recently. In previous two multicenter observational ERCP complication studies (2,6), guidewire used for cannulation was not wide relatively. The two studies showed that cannulation with guidewires was associated with a higher risk of pancreatitis by univariate (but not multivariate) analysis. However, in most cases, the technique might be used as a crossover method after failed catheter and contrast injection, thus inheriting an increased risk as a result of excessive papillary trauma or pancreatic duct injections. In contrast, a recent prospective randomized study on 300 patients found that guidewire cannulation significantly decreased the likelihood of post-ERCP pancreatitis, compared with a sphincterotome alone (8.6% vs. 16.6%, P=0.02), reportedly by facilitating cannulation and reducing need for precut sphincterotomy (25). An earlier similar study by Lella et al. (26) comparing papillotome with contrast injection to wire-guided cannulation also demonstrated that the rate of pancreatitis in the guidewire group was significantly lower (0% vs. 4%, P<0.05), presumably because of less pancreatic contrast injection. In our study, soft-tipped hydrophilic guidewires were the mostly used ones for selective cannulation, and pancreatic deep wire pass was found to be an independent risk factor for overall complications and pancreatitis by multivariate analysis (OR: 1.80 and 2.77, respectively); however, this finding was not obtained in a previous study (3). Although guidewire cannulation is easier, less likely to require precut sphincterotomy and can be used to avoid contrast injection of the pancreatic duct as described above, the method may be accompanied with wire cannulation of the pancreatic duct, which is a risk for complications and pancreatitis, according the present study. Pancreatic contrast injection was an independent risk factor for post-ERCP pancreatitis in many previous studies (2,3,4,7,11,19,27). However, our results showed that this factor was significant only by univariate but not by multivariate analysis. A possible explanation for these findings is that unlike other similar studies the frequent use of guidewires cannulation in our procedures usually minimized unintentional injections into the pancreatic duct and thus made pancreatic contrast injection to be associated with pancreatic deep wire pass. Like other recent studies (2,5), our study showed that acinarization or opacification of the pancreatic branch duct was found to be a risk factor by univariate analysis only, probably because of the low prevalence (1.67%, 45 of 2,691) of this factor in our study.
In our series, we evaluated major papilla and minor papilla pancreatic sphincterotomy and found that minor papilla sphincterotomy was not a significant risk factor for overall complications and pancreatitis by univariate analysis, probably because of the low prevalence (0.41%, 11 of 2,691) of this factor, as demonstrated in previous reports (28). Major papilla pancreatic sphincterotomy was a significant risk factor for overall complications and pancreatitis only by univariate but not by multivariate analysis, which was different from the finding in a previous large multivariate analysis where pancreatic sphincterotomy was an independent risk of pancreatitis (3). A possible reason for the discrepancy is that pancreatic sphincterotomy was frequently used for patients with advanced chronic pancreatitis in China, which is protective in general, rather than for pancreas divisum or SOD, as often done in the United States, in which pancreatic sphincterotomy is much riskier. Pancreatic stenting was a significant risk factor for pancreatitis only by univariate but not by multivariate analysis in our study, suggesting it may not increase the risk of pancreatitis. The significantly higher post-ERCP pancreatitis rate in this maneuver (by univariate analysis) is almost certainly related to the fact that the pancreatic stents were placed in high-risk cases. Previous randomized controlled trials and a meta-analysis also demonstrated that pancreatic stents were beneficial in reducing the incidence of pancreatitis (29,30,31).
Precut papillotomy refers to a variety of endoscopic techniques used to gain access to the bile duct. It has been controversial whether higher rates of complications and pancreatitis after precut sphincterotomy are because of the precut itself or of the antecedent repeated cannulation attempts (32). A randomized controlled trial by Tang et al. (33) comparing precut papillotomy to persistent cannulation showed that the rates of complications and pancreatitis were similar (4.2% vs. 9.1% and 0% vs. 4.5%, respectively). However, in two multicenter studies that used multivariate analysis to adjust for confounding variables related to difficult cannulation, precutting independently and significantly increased the risk of com-plications and pancreatitis (2,5). A meta-analysis also showed that precut sphincterotomy was a highly significant risk for post-ERCP pancreatitis even after adjusting for other variables (22). In our study, we assessed all specific precut techniques including needle-knife precut and transpancreatic precut and found that only needle-knife precut was an independent risk factor for overall complications and pancreatitis (OR: 2.70 and 4.34, respectively). Transpancreatic precut was a significant risk factor for overall complications and pancreatitis only by univariate, but not by multivariate analysis, although the incidence of overall complications and pancreatitis in transpancreatic precut was similar with needle-knife precut (14.71% (20 of 136) vs. 18.42% (14 of 76), and 11.76% (16 of 136) vs. 11.84% (9 of 76), respectively). Thus, the risk associated with transpancreatic precut may be more related to techniques that involve repeated cannulation attempts and pancreatic deep wire pass. Our data suggest that the risk after transpancreatic precut might be lower than needle-knife precut. A randomized study by Catalano et al. (34) also reported that transpancreatic precut had lower rates of complications and pancreatitis, compared with needle-knife precut (3.5% vs. 17.7% and 3.5% vs. 11.7%, respectively), but this result was not statistically significant. Our observation still needs to be confirmed by more randomized controlled studies comparing specific precut techniques.
Both Post-ERCP pancreatitis and asymptomatic hyperamylasemia occur because of injury of the pancreatic tissue induced by ERCP techniques, but the reason why some patients eventually develop pancreatitis and others asymptomatic hyperamylasemia remains unknown. The underlying mechanisms of the two vastly different clinical courses may include two respects (35): one may be attributable to the difference in the severity of the injury to the pancreas, and the other to the difference in the magnitude of inflammatory response to the injury of pancreas. Asymptomatic hyperamylasemia is associated with mild injury of pancreas, perhaps without inflammatory response to pancreas. Pancreatitis may be associated with more severe injury of pancreas, meanwhile with inflammatory response to pancreas.
In this study, only female gender seemed to be the patient-related risk factor that differed between pancreatitis and asymptomatic hyperamylasemia, whereas the procedure-related risk factors were similar between pancreatitis and asymptomatic hyperamylasemia. A single-center prospective study by Dickinson and Davies (9) also showed that the factors leading to either pancreatitis or hyperamylasemia are similar except with regard to the age and gender of the patient, although the sample size of the study is relatively too small to allow multivariate analysis. However, the other single-center study by Christoforidis et al. (10) found that younger age (<50) and relapsing pancreatitis were risk factors for hyperamylasemia or pancreatitis significantly, and more women developed either hyperamylasemia or pancreatitis without statistical significance. The difference in patient-related factors between asymptomatic hyperamylasemia and pancreatitis still needs to be verified by other large-scale prospective multicenter studies in future.
According to the present study, we suppose that female gender is more prone to the development of pancreatitis, but not of asymptomatic hyperamylasemia. The underlying reasons may be that women are more susceptible than men to inflammatory response to the similar degree of pancreatic injury, when undergoing similar techniques of cannulation, sphincterotomy, or pancreatic manipulation. In future studies, other host factors at the cellular level, including genetic susceptibility to pancreatitis, hormonal or inflammatory mediators should be examined, which will help further understand the different magnitude of inflammatory response between pancreatitis and asymptomatic hyperamylasemia, and the underlying mechanisms of the two vastly clinical courses.
In conclusion, patient-related risk factors are as important as procedure-related ones in determining the risk for overall post-ERCP complications and pancreatitis. These findings emphasize the importance in the avoidance of high-risk procedures such as unintentional pancreatic deep wire pass, prolonged cannulation time, and needle-knife precut, especially in high-risk patients such as younger women and patients with periampullary diverticulum. However, the risk factors for asymptomatic hyperamylasemia may be mostly procedure related. The difference in patient-related factors between pancreatitis and asymptomatic hyperamylasemia suggests that the host susceptibility represents an important factor in the mechanism by which post-ERCP pancreatitis develops.
CONFLICT OF INTEREST
Guarantor of the article: Zhao-Shen Li, MD.
Specific author contributions: None.
Financial support: This study was supported by a grant from Boston Scientific Corporation, China.
Potential competing interests: None.
References
- Freeman ML. Adverse outcomes of ERCP. Gastrointest Endosc 2002;56:S273–82. | Article | PubMed |
- Freeman ML, Nelson DB, Sherman S et al. Complications of endoscopic biliary sphincterotomy. N Engl J Med 1996;335:909–18. | Article | PubMed | ChemPort |
- Freeman ML, DiSario JA, Nelson DB et al. Risk factors for post-ERCP pancreatitis: a prospective, multicenter study. Gastrointest Endosc 2001;54:425–34. | Article | PubMed | ChemPort |
- Loperfido S, Angelini G, Benedetti G et al. Major early complications from diagnostic and therapeutic ERCP: a prospective multicenter study. Gastrointest Endosc 1998;48:1–10. | Article | PubMed | ChemPort |
- Masci E, Toti G, Mariani A et al. Complications of diagnostic and therapeutic ERCP: a prospective multicenter study. Am J Gastroenterol 2001;96:417–23. | Article | PubMed | ChemPort |
- Vandervoort J, Soetikno RM, Tham TCK et al. Risk factors for complications after performance of ERCP. Gastrointest Endosc 2002;56:652–6. | Article | PubMed |
- Cheng CL, Sherman S, Watkins JL et al. Risk factors for Post-ERCP pancreatitis: a prospective multicenter study. Am J Gastroenterol 2006;101:139–47. | Article | PubMed |
- Williams EJ, Taylor S, Fairclough P et al. Risk factors for complication following ERCP; results of a large-scale, prospective multicenter study. Endoscopy 2007;39:793–801. | Article | PubMed | ChemPort |
- Dickinson RJ, Davies S. Post-ERCP pancreatitis and hyperamylasemia: the role of operative and patient factors. Eur J Gastroenterol Hepatol 1998;10:423–8. | Article | PubMed | ChemPort |
- Christoforidis E, Goulimaris I, Kanellos I et al. Post-ERCP pancreatitis and hyperamylasemia: patient-related and operative risk factors. Endoscopy 2002;34:286–292. | Article | PubMed | ChemPort |
- Andriulli A, Clemente R, Solmi L et al. Gabexate or somatostatin administration before ERCP in patients at high risk for post-ERCP pancreatitis: a multicenter, placebo-controlled, randomized clinical trial. Gastrointest Endosc 2002;56:488–95. | Article | PubMed |
- Andriulli A, Leandro G, Niro G et al. Pharmacologic treatment can prevent pancreatic injury after ERCP: a meta-analysis. Gastrointest Endosc 2000;51:1–7. | Article | PubMed | ChemPort |
- Behar J, Corazziari E, Guelrud M et al. Functional gallbladder and sphincter of Oddi disorders. Gastroenterology 2006;130:1498–509. | Article | PubMed |
- Sarner M, Cotton PB. Classification of pancreatitis. Gut 1984;25:756–9. | Article | PubMed | ISI | ChemPort |
- Cotton PB, Lehman G, Vennes JA et al. Endoscopic sphincterotomy complications and their management: an attempt at consensus. Gastrointest Endosc 1991;37:383–91. | Article | PubMed | ISI | ChemPort |
- Aronson N, Flamm CR, Bohn RL et al. Evidence based assessment: patient, procedure, or operator factors associated with ERCP complications. Gastrointest Endosc 2002;56:S294–302. | Article | PubMed |
- Kawai K, Akasaka Y, Murakami K et al. Endoscopic sphincterotomy of the ampulla of Vater. Gastrointest Endosc 1974;20:148–51. | PubMed | ChemPort |
- Cohen S, Bacon BR, Berlin JA et al. National Institutes of Health State-of-the-Science Conference Statement: ERCP for diagnosis and therapy, 14–16 January. Gastrointest Endosc 2002;56:803–9. | Article | PubMed |
- Friedland S, Soetikno RM, Vandervoort J et al. Bedside scoring system to predict the risk of developing pancreatitis following ERCP. Endoscopy 2002;34:483–8. | Article | PubMed | ChemPort |
- Haber GB. Prevention of post-ERCP pancreatitis. Gastrointest Endosc 2000;51:100–3. | Article | PubMed | ChemPort |
- Fogel EL, Eversman D, Jamidar P et al. Sphincter of Oddi dysfunction: pancreatobiliary sphincterotomy with pancreatic stent placement has a lower rate of pancreatitis than biliary sphincterotomy alone. Endoscopy 2002;34:280–5. | Article | PubMed | ChemPort |
- Masci E, Mariani A, Curioni S et al. Risk factors for pancreatitis following endoscopic retrograde cholangiopancreatography: a meta-analysis. Endoscopy 2003;35:830–4. | Article | PubMed | ChemPort |
- Laugier R, Bernard JP, Berthezene P et al. Changes in pancreatic exocrine secretion with age: pancreatic exocrine secretion does decrease in the elderly. Digestion 1991;50:202–11. | PubMed | ChemPort |
- Freeman ML, Guda NM. ERCP cannulation: a review of reported techniques. Gastrointest Endosc 2005;61:112–25. | Article | PubMed |
- Artifon ELA, Sakai P, Cunha JEM et al. Guidewire cannulation reduces risk of post-ERCP pancreatitis and facilitates bile duct cannulation. Am J Gastroenterol 2007;102:2147–53. | Article | PubMed |
- Lella F, Bagnolo F, Colombo E et al. A simple way of avoiding post-ERCP pancreatitis. Gastrointest Endosc 2004;59:830–4. | Article | PubMed |
- Mehta SN, Pavone E, Barkun JS et al. Predictors of post-ERCP complications in patients with suspected choledocholithiasis. Endoscopy 1998;30:457–63. | Article | PubMed | ChemPort |
- Rabenstein T, Schneider HT, Bulling D et al. Analysis of the risk factors associated with endoscopic sphincterotomy techniques: preliminary results of a prospective study, with emphasis on the reduced risk of acute pancreatitis with low dose anticoagulation treatment. Endoscopy 2000;32:10–9. | Article | PubMed | ChemPort |
- Tarnasky PR, Palesch YY, Cunningham JT et al. Pancreatic stenting prevents pancreatitis after biliary sphincterotomy in patients with sphincter of Oddi dysfunction. Gastroenterology 1998;115:1518–24. | Article | PubMed | ChemPort |
- Fazel A, Quadri A, Catalano MF et al. Does a pancreatic duct stent prevent post-ERCP pancreatitis? A prospective randomized study. Gastrointest Endosc 2003;57:291–4. | Article | PubMed |
- Singh P, Das A, Isenberg G et al. Does prophylactic pancreatic stent placement reduce the risk of post-ERCP acute pancreatitis? A meta-analysis of controlled trials. Gastrointest Endosc 2004;60:544–50. | Article | PubMed |
- Sriram PVJ, Rao GV, Reddy DN. The precut—when, where and how? A review. Endoscopy 2003;35:S24–30. | Article | PubMed | ChemPort |
- Tang SJ, Haber GB, Kortan P et al. Precut papillotomy vs. persistence in difficult biliary cannulation: a prospective randomized trial. Endoscopy 2005;37:58–65. | Article | PubMed |
- Catalano MF, Linder JD, Geenen JE. Endoscopic transpancreatic papillary septotomy for inaccessible obstructed bile ducts: comparison with standard pre-cut papillotomy. Gastrointest Endosc 2004;60:557–61. | Article | PubMed |
- Cooper ST, Slivka A. Incidence, risk factors, and prevention of post-ERCP pancreatitis. Gastroenterol Clin N Am 2007;36:259–76. | Article |
Appendices
Appendix
The following centers, listed in descending order of the number of patients enrolled (in parentheses), participated in the study.
Changhai Hospital, Second Military Medical University, Shanghai, China (1,047);
Heilongjiang Provincial Hospital, Harbin, China (508);
The First Affiliated Hospital of Nanchang University, Nanchang, China (356);
The Second Affiliated Hospital of Nanjing Medical University, Nanjing, China (213);
Anhui Provincial Hospital, Hefei, China (205);
The First People's Hospital of Hangzhou, Hangzhou, China (202);
Fujian Provincial Hospital, Fuzhou, China (165);
Shandong Communication Hospital, Jinan, China (108);
Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, China (97);
The First Affiliated Hospital of Nanjing Medical University, Nanjing, China (68);
The Affiliated Hospital of Qingdao University, Qingdao, China (66);
Qilu Hospital, Shandong University, Jinan, China (53);
Tianjin Union Medical Center, Tianjin, China (53);
Nanfang Hospital, Nanfang Medical University, Guangzhou, China (37).
Acknowledgments
We thank all those endoscopists and physicians who assisted in data collection. We are indebted to Professor Quan-Cai Cai, Center for Clinical Epidemiology, Changhai Hospital and Professor Jian Lu, Department of Health Statistics, Second Military Medical University, for statistical analyses.
