When evaluating a patient with suspected choledocholithiasis, the fundamental question is whether or not to recommend an endoscopic retrograde cholangiopancreatographyEUS (ERCP). Given its risks, and with the more widespread availability of endoscopic ultrasound (EUS) and magnetic resonance cholangiopancreatography (MRCP), ERCP has become (or at least should be) a nearly exclusively therapeutic tool. In order to restrict ERCP to patients with the highest probability of ongoing choledocholithiasis — in whom the risk-benefit ratio is most favorable — accurate and reproducible risk stratification strategies are necessary.
The most widely used strategy is an algorithm proposed by ASGE that selects patients for ERCP versus less invasive tests (EUS, MRCP or intraoperative cholangiography [IOC]) based upon readily available clinical, laboratory and radiographic predictors.1 We and others have conducted six validation studies to date (comprising 2,052 patients), which demonstrate that the accuracy of the ASGE guidelines is in the range of 55 to 70 percent, potentially exposing a sizeable fraction of patients to unnecessary ERCP.2–7 However, the acceptable rate of “negative” ERCP for suspected choledocholithiasis is a value judgment based on multiple factors, including local ERCP expertise and the availability, cost and performance characteristics of alternative diagnostic modalities. Indeed, while we concluded from these data that the guidelines are not adequately sensitive and specific to drive clinical management, the authors of a recent prospective validation study concluded that the 70 percent accuracy they observed in their cohort justifies the routine use of the guidelines in clinical practice.5 Similarly, surveyed practitioners caring for patients with biliary disease considered 2.5 (95 percent CI, 2.2–2.83) negative ERCPs per every 10 procedures an acceptable rate.8
Given this subjectivity and the absence of a highly accurate and widely accepted risk-stratification strategy, at our institution we generally begin the conversation with the ASGE guidelines but then adjust the plan in response to factors that are not accounted for in existing algorithms, having a low threshold to pivot toward less invasive tests in patients deemed to be at high probability for choledocholithiasis by the ASGE guidelines. For example, we may opt for less invasive testing in a high-probability patient with improving abdominal pain and an increased appetite on the day after admission. When this pivot occurs in a high-probability patient, we generally favor EUS over MRCP because it is more sensitive for small stones and does not delay care when performed in an ERCP-capable endoscopy suite. In patients at intermediate probability, we select MRCP, EUS or IOC based on test availability and the expertise of the involved surgeon.
However, given the possible disadvantages of overusing EUS and MRCP, but acknowledging the morbidity and costs associated with ERCP-related adverse events, it remains reasonable to strive toward a highly accurate stratification system that eliminates most (greater than 90 percent) diagnostic ERCPs while remaining cost and resource neutral. To this end, several alternative strategies have been proposed:
1) A pilot randomized trial comparing an EUS-guided strategy to up-front ERCP in all comers with suspected choledocholithiasis (but no concern for cholangitis) revealed that initial EUS reduced procedure time and adverse events without affecting stone-related outcomes.9 On this basis, and recognizing the limitations of existing guidelines, some experts have adopted a strategy of routine EUS at the time of possible ERCP in all patients with suspected choledocholithiasis but no evidence of ascending cholangitis. Additional research on the EUS-first strategy is warranted since endosonographic evaluation of the biliary tree is safe, accurate and can be accomplished quickly at the time of possible ERCP.
2) Several alternative scoring systems based on basic biochemical testing and bile duct size have demonstrated better performance characteristics than existing guidelines, albeit in small patient cohorts. One simple scoring strategy aims to risk-stratify patients to undergo cholecystectomy alone (score equals 0), IOC (score equals 1 or 2), MRCP (score equals three or 4) or ERCP (score equals 5). This scoring is based on initial biochemical laboratory tests (gamma glutamyl transferase greater than or equal to 350 U/L, alkaline phosphatase greater than or equal to 250 U/L, total bilirubin greater than or equal to 3 mg/dL, and direct bilirubin greater than or equal to 2 mg/dL) and common bile duct size (greater than or equal to 9 mm) on transabdominal ultrasound. This strategy eliminated negative ERCPs entirely in a small number of patients with suspected gallstone pancreatitis.10 Such strategies certainly merit validation in larger, diverse patient populations.
3) Advanced predictive modeling techniques may also hold promise. A recently developed artificial neural network model predicted choledocholithiasis with an area under the receiver operating characteristic curve of 0.88 (95 percent CI, 0.831–0.938).11 Unfortunately, machine-learning algorithms did not substantially outperform the ASGE algorithm in our cohort of approximately 700 patients from two academic institutions.15
4) We hypothesized that the trend in liver function tests over time is clinically informative, with decreasing values suggesting spontaneous stone passage and prompting less invasive initial intervention. However, we conducted two studies in different geographic and socioeconomic populations, which demonstrated that the evolution of liver chemistries over time does not predict persistent choledocholithiasis.3,6
When considering ERCP for suspected choledocholithiasis, three important caveats are worth mentioning. First, in patients at intermediate risk who are not cholecystectomy candidates, upfront ERCP for biliary sphincterotomy to reduce the risk of recurrence is reasonable, and confirmatory testing is generally not needed.12 Second, in the context of suspected choledocholithiasis, the presence of a prior biliary sphincterotomy is likely highly protective against post-ERCP pancreatitis because it facilitates cannulation and reduces the risk of pancreatic injury by separating the biliary and pancreatic orifices. In such cases, I have a lower threshold to proceed with ERCP. Third, in patients undergoing cholecystectomy who require ERCP, most surgeons prefer that stones be cleared pre operatively, although intra-operative and post-operative ERCP have been shown to be safe and equally effective.13,14
In conclusion, the fundamental question of when to perform ERCP for suspected choledocholithiasis remains unanswered, although several intriguing lines of research may provide clarity. In the interim, a thoughtful case-by-case strategy founded on existing evidence, but sensitive to the principle of minimizing unnecessary risk, will serve our patients well.
Dr. Elmunzer has no conflicts to disclose.
1. Maple JT, Ben-Menachem T, Anderson MA, et al. The role of endoscopy in the evaluation of suspected choledocholithiasis. Gastrointest Endosc 2010;71:1-9.
2. Rubin MI, Thosani NC, Tanikella R, et al. Endoscopic retrograde cholangiopancreatography for suspected choledocholithiasis: testing the current guidelines. Dig Liver Dis 2013;45:744-749.
3. Adams MA, Hosmer AE, Wamsteker EJ, et al. Predicting the likelihood of a persistent bile duct stone in patients with suspected choledocholithiasis: accuracy of existing guidelines and the impact of laboratory trends. Gastrointest Endosc 2015;82:88-93.
4. Magalhães J, Rosa B, Cotter J. Endoscopic retrograde cholangiopancreatography for suspected choledocholithiasis: from guidelines to clinical practice. World J Gastrointest Endosc 2015;7:128-134.
5. Sethi S, Wang F, Korson AS, et al. Prospective assessment of consensus criteria for evaluation of patients with suspected choledocholithiasis. Dig Endosc 2016;28:75-82.
6. Suarez AL, LaBarre NT, Cotton PB, et al. An assessment of existing risk stratification guidelines for the evaluation of patients with suspected choledocholithiasis. Surg Endosc 2016 Feb 19. epub ahead of print
7. Nárvaez Rivera RM, González JA, Monreal Robles R, et al. Accuracy of ASGE criteria for the prediction of choledocholithiasis. Rev Esp Enferm Dig;108. doi: 10.17235/ reed.2016.4212/2016. epub ahead of print
8. Elmunzer BJ, Debenedet AT, Volk ML, et al. Clinical yield of diagnostic endoscopic retrograde cholangiopancreatography in orthotopic liver transplant recipients with suspected biliary complications. Liver Transpl 2012;18:1479-1484.
9. Lee YT, Chan FKL, Leung WK, et al. Comparison of EUS and ERCP in the investigation with suspected biliary obstruction caused by choledocholithiasis: a randomized study. Gastrointestinal Endoscopy 2008;67:660-668.
10. Sherman JL, Shi EW, Ranasinghe NE, et al. Validation and improvement of a proposed scoring system to detect retained common bile duct stones in gallstone pancreatitis. Surgery 2015;157:1073-109.
11. Jovanovic P, Salkic NN, Zerem E. Artificial neural network predicts the need for therapeutic ERCP in patients with suspected choledocholithiasis. Gastrointest Endosc 2014;80:260-268.
12. Unpublished data, Elmunzer, BJ and Waljee, AK.
13. da Costa DW, Schepers NJ, Römkens TE, et al. Endoscopic sphincterotomy and cholecystectomy in acute biliary pancreatitis. Surgeon 2016;14:99-108.
14. Byrne MF, McLoughlin MT, Mitchell RM, et al. For patients with predicted low risk for choledocholithiasis undergoing laparoscopic cholecystectomy, selective intraoperative cholangiography and postoperative endoscopic retrograde cholangiopancreatography is an effective strategy to limit unnecessary procedures. Surg Endosc 2009;23:1933-1937.