LAUSR

Sir, We would like to compliment Vaynshtein et al. with their publication: ‘Predictors for choledocholithiasis in patients undergoing endoscopic ultrasound’ [1]. They identified a high alkaline phosphatase level (ALP) of >300 IU/L as a risk factor for common bile duct stones in patients with an intermediate risk of having bile duct stones. Moreover, they show EUS to be an excellent tool for selecting patients at intermediate risk for bile duct stones for therapeutic ERC. In approximately 2/3 of patients’ bile duct stones were ruled out by EUS. These patients were therefore not exposed to the risk of complications of a therapeutic ERC [2]. These findings are in line with previous publications on this subject [3–5]. However, in our opinion there are some items that need further clarification. First, the aim of the study was to identify additional predictive factors in patients already deemed to be at intermediate risk for bile duct stones in order to further improve patient selection for immediate ERC, without prior EUS or MRCP. In this retrospective cohort study of 175 patients, univariate and subsequent multivariate logistic regression analysis identified ALP >300 IU/L as a significant predictor for CBD stones at EUS (OR 2.98, p< .001). However, the authors do not comment on whether this finding should have clinical implications, that is whether patients with suspected bile duct lithiasis and an ALP of >300 IU/l should now be referred for immediate ERC without a prior EUS. Furthermore, they do not state how their finding relates to the existing guidelines on non-invasive strategies for bile duct stone risk assessment [6,7]. Second, the results of EUS in this publication are described as either positive or negative for stones. The authors do not clearly define ‘bile duct lithiasis’. It is therefore unclear if and how biliary sludge and microlithiasis were classified in this study. Microlithiasis and sludge are detected using EUS in patients with suspected bile duct lithiasis in 20–25% of cases [5,8]. Although the clinical relevance is still not completely clear, the presence of sludge and/or microlithiasis can cause both biliary symptoms and complications of bile duct lithiasis such as pancreatitis. In our opinion, sludge and microlithiasis should therefore be regarded as bile duct lithiasis. One of the reasons for the known interobserver variance in EUS for bile duct lithiasis is the fact that clear definitions regarding bile duct stones, sludge and microlithiasis are lacking. This clearly is an opportunity for future research. Third, based on EUS findings 62 out of 175 (35%) patients in this study were referred for therapeutic ERC. In 47 out of 57 patients (82%) who underwent an ERC, bile duct stones were detected and treated. Vaynshtein et al. did not mention the time interval between EUS and ERCP. This is relevant because spontaneous passage of bile duct content from the bile duct into the duodenum is known to occur over time [9]. As the use of EUS for diagnosing bile duct lithiasis is becoming more widespread, the proficiency of endosonographers is increasing, and scientific data pointing to its merits are accumulating, it becomes more difficult to proceed to a therapeutic ERC without performing an EUS first. This has implications for clinical practice, especially in hospitals without EUS (or MRCP) facilities, since the positive predictive value of non-invasive prediction rules is at best >50%. Initiatives to improve the diagnostic accuracy of non-invasive bile duct stone prediction rules should therefore be encouraged.