Discussion

Acute cholangitis is a common consequence of biliary obstruction and may present with a wide spectrum of symptoms and signs from mild fever to septic shock and multi-organ failure necessitating admission to the critical care unit for organ support. The severity of cholangitis is defined as mild (responsive to antibiotics and supportive therapy), moderate (neither responsive to medical therapy nor associated with organ dysfunction), or severe (associated with at least one sign of organ dysfunction).19 Approximately 80% of patients respond to antibiotic therapy alone while the remainder require early effective biliary drainage in addition to antibiotics and supportive therapy.20 The reported mortality varies from 10% to 30%; however, this is in all patients with acute cholangitis including the mild form of cholangitis. The major cause of death is multiple organ failure with irreversible shock, and mortality rates have not significantly improved over the years.21–25 It is important to note that unless early and appropriate biliary drainage is performed and systemic antibiotics are administered, mortality may be up to 100%.26 The data on the outcome of emergency ERCP in this cohort of critically ill patients receiving organ support are limited. Available data suggest that mortality is variable between 25% and 55%, despite emergency ERCP.27–29 The timing of ERCP in these patients would appear to be important to achieve a favourable outcome, but there is no compelling existing data to support emergency ERCP in these patients. The definition of what constitutes an emergency ERCP has not been clearly reported. It could be defined as ERCP in a patient who is critically unwell and is unable to wait due to the severity of the illness until the next available elective ERCP session, assuming that it is within 24 h. It has also been defined as an ERCP performed within few hours of presentation.30 The Surviving Sepsis Campaign, however, recommends source control with attention to the balance of risks and benefits of the chosen intervention as soon as the diagnosis of severe sepsis is made to improve the outcome.31 In our patient cohort we performed ERCP within 6–8 h of diagnosis of severe sepsis due to acute cholangitis not responding to standard therapy, with critical physiological deterioration. Our unit provides four scheduled ERCP lists a week and therefore most patients presenting during weekdays with mild to moderate acute cholangitis not requiring critical care support can undergo an ERCP on lists already in place. These were not considered as emergency ERCP in this study as these patients did not require ventilation or critical care support.

After agreeing to provide a tertiary emergency ERCP service, our endoscopic services underwent major reorganisation with collaboration from our critical care colleagues. A major component of this restructuring was the early involvement of critical care physicians facilitating the transfer of patients both from within the hospital and neighbouring referral hospitals to the critical care unit after assessment. Those who were deemed too unwell to undergo an ERCP in our unit with conscious sedation and monitoring had an ERCP performed in the emergency operating rooms. These facilities have appropriate fluoroscopic equipment with mobile digital C arm and support staff including on-call endoscopy nurses. Although the quality of cholangiogram is inferior to a fixed fluoroscopy unit, the quality was more than adequate for the type of procedure undertaken in these circumstances. All our procedures were performed by consultants who are specialists in hepatobiliary endoscopy. The critical care unit was considered inadequate to perform high quality ERCP because of the limitation of fluoroscopy, patient positioning and radiation risk to others.

The predominant underlying aetiology of severe cholangitis in both our cohort and the available literature is choledocholithiasis. Malignant biliary obstruction was the cause of cholangitis in two patients who both died within 30 days. Both had biliary stents in situ and presented with obstruction, which indicates that severe sepsis in patients with malignant stricture is often a secondary process after previous biliary interventions. In patients who required a repeat ERCP, it could be argued that a biliary stent placement to ensure biliary drainage would have prevented a subsequent repeat procedure. On the other hand, it is equally possible that placing a biliary stent in all patients undergoing emergency ERCP would subject all patients to an unnecessary ERCP at a later date. Therefore, we recommend that this decision should be left at the discretion of the endoscopists who can judge the adequacy of biliary drainage in an individual case. Regarding the non-survivors, despite prompt and successful biliary decompression the mortality was considerable, with almost all patients dying within 24 h from overwhelming sepsis.

Our overall mortality is lower than 55%, as described by Saleem and colleagues,28 although in their study there was a failure to achieve biliary cannulation in 18% of cases, which most likely would have caused delay in biliary decompression. Ramirez and colleagues27 reported a similar mortality rate to ourselves of 25% in 32 critically unwell patients undergoing ERCP in the radiology department. The wide variability of mortality in both papers may reflect a heterogeneity of the patients as the number of ventilated patients was similar (47% vs 60%), but Saleem and colleagues28 proposed that they had more severely ill patients and ERCP was performed in the intensive care unit where there were no adequate facilities to perform ERCP under fluoroscopy. The critical step to attain biliary drainage at ERCP is achieving deep biliary cannulation, which was 100% in our series. This was attained despite the majority of patients being naive to ERCP. Once biliary cannulation is achieved, biliary drainage can almost always be achieved, which is the primary aim. Stone therapy and duct clearance, although desirable, are of secondary importance in this cohort of critically ill patients.

A measure of successful ERCP in this situation is a drop in bilirubin levels, which has been demonstrated in our study within 24 h in those who survived. Failure to show a decrease in bilirubin levels within 24–48 h may be an early indicator of unsuccessful biliary decompression and correlate with increased mortality.

In the UK, the National Confidential Enquiry into Patient Outcome and Death (NCEPOD), which operates under the umbrella of the National Institute of Health and Clinical Excellence as an independent confidential enquiry and whose main aim is to improve the quality and safety of patient care, published their report on endoscopic services ‘Scoping our practice’. It was reported that many patients undergoing emergency endoscopy are severely ill, elderly and often poorly prepared for an interventional procedure. Furthermore, in 68% of cases the NCEPOD advisors considered the procedure futile.

We demonstrate a high degree of technical success with a good survival rate of 75% at 30 days after emergency ERCP in ventilated, critically ill and often elderly patients. It is likely that mortality would be extremely high and close to 100% in this patient cohort if emergency ERCP had not been carried out. It is difficult to predict in which patient an emergency ERCP would be successful and in which it would be futile. Our data suggest an aggressive approach is justified. Emergency ERCP should therefore not be considered futile. The type of service required to deliver emergency ERCP is not likely to be available in smaller institutions. This study supports the concept of the concentration of ERCP services into fewer, higher volume units with appropriate critical care involvement facilitating the transfer and support of these patients. We believe that further studies in this area will support our findings and will help make this intervention in appropriate settings a standard practice.