Urgent endoscopic retrograde cholangiopancreatography for acute biliary pancreatitis: Few answers and more questions

SUMMARY

In this multicentre randomized study from the Netherlands, patients with acute biliary pancreatitis (ABP) predicted to have a severe course were randomized to receive urgent endoscopic retrograde cholangiopancreatography (ERCP) with sphincterotomy or conservative therapy. The study randomized these patients within 24 hours of presentation and the patients with cholangitis were excluded from the study. The primary end-point was a combination of death or major complications (post-ERCP persistent organ failure, cholangitis, other infectious complications, pancreatic necrosis or insufficiency). Of the 232 patients in the intention-to-treat analysis, 118 were randomized to the urgent ERCP group. The primary end-point was noted to occur in 38% of patients in the ERCP group and in 44% of those in the conservative group (p=not significant). Except for cholangitis, which was less frequent in the ERCP group, no differences in individual end-points were noted between the two groups. The authors suggest that a conservative strategy is appropriate to predict severe acute gallstone pancreatitis in the absence of cholangitis.

COMMENT

Acute gallstone pancreatitis is the most common cause of acute pancreatitis globally and the incidence is increasing because of the rising prevalence of obesity and gallstone disease.¹ Impacted stones in the common bile duct or ampulla are believed to be the initiating and propagating event in biliary pancreatitis.² Hence, it has been presumed that urgent biliary decompression may ameliorate and modify the disease course, especially vis-à-vis the severity of the disease. The literature has shown variable results of biliary decompression, especially in unselected patients of ABP (Table I).^3–10 Pooled data of these randomized controlled trials show a clear advantage of ERCP in patients with concomitant cholangitis and persistent cholestasis but not in the case of unselected patients with ABP.¹¹ Various guidelines suggest urgent (<24 hours) ERCP in the presence of concomitant cholangitis and early (<72 hours) ERCP in those with persistent cholestasis in ABP. Guidelines also suggest a lack of benefit of ERCP in case of mild disease.^12–15 In patients with predicted severe pancreatitis, limited literature is available. Previous studies had many lacunae such as small sample size, heterogeneous inclusion, variable timing of ERCP and sphincterotomy rates in this group of patients.^3,6,8

The present randomized trial by the Dutch pancreatitis group has attempted to address these lacunae. The trial has certain strengths as it is the largest trial in patients with predicted severe ABP, a well-designed study with clearly defined criteria for cholestasis. All the ERCPs were done early in the course of the disease (<24 hours) and endoscopic sphincterotomy was done in all the patients who underwent ERCP. While this recent trial on the role of urgent ERCP in ABP provides evidence for a conservative approach in the absence of cholangitis, there are concerns regarding the generalizability of these findings. The primary outcome used in this study was a composite of mortality and complications at 6 months as in other trials by the group. This end-point has benefits as it assesses the effect of intervention with respect to multiple parameters and the fact that assessment is up to 6 months since acute pancreatitis can be associated with late mortality. However, the trend to use this composite outcome at a time gap from the actual intervention and this ‘one size fits all’ approach is of concern.¹⁶ Benefits from early ERCP would be derived from a change in the course of disease and should accrue early in the course of the disease. Over 6 months, other factors such as infection of the local complications could impact outcomes and cloud the actual benefit from the intervention. Assessment at shorter duration (at 1–2 weeks) with parameters such as development of severe pancreatitis (persistent organ failure), worsening systemic inflammatory response syndrome (SIRS) and organ failure, as used in previous trials, may have been more meaningful.^4,9 Nevertheless, given the null hypothesis of this trial (i.e. early sphincterotomy may reduce disease severity in predicted severe acute necrotizing pancreatitis), the use of composite measure at 6 months is acceptable as it provides a complete assessment of severity during the entire course of acute pancreatitis.

The use of elevated alanine transaminase as the sole criterion to diagnose biliary pancreatitis in a considerable number of patients is also of concern. Previous studies included these criteria along with the presence of other evidence of gallstone disease including either elevated serum bilirubin or alkaline phosphatase when imaging failed to show evidence of gallstone/ sludge.^7,8 Further, patients in the ERCP group had higher SIRS scores and C-reactive protein (CRP) at admission. These baseline differences could have masked the possible benefit of ERCP as sicker patients were included in the intervention arm, which might impact the results. This is especially so because in a substantial number of patients (19%) in the ERCP group, the procedure could not be done. The reason for technical failure was large periampullary diverticulum in 3 patients and complications of pancreatitis including periampullary oedema and respiratory failure in 7 patients. Eventually, in the ERCP arm, there was imbalance for these two reasons: failure to complete the procedure and higher SIRS rates that could have impacted the primary outcome. In such a situation, a per-protocol analysis should also have been done. The authors had committed this in the statistical plan, but this analysis was not provided.

In view of these concerns, we believe that the final word on the role of early ERCP in acute pancreatitis has not yet been said. Future studies should address these concerns about patient selection and selection of appropriate outcome measures before the role of ERCP in this group of patients can be defined.