PDF icon Download PDF
Endoscopy 2004; 36(12): 1104-1106
DOI: 10.1055/s-2004-826036
Editorial
© Georg Thieme Verlag KG Stuttgart · New York

ERCP: Targeting the Stone

S.  Adamsen1
  • 1Dept. of Gastrointestinal Surgery, Herlev Hospital, Copenhagen University, Herlev, Denmark
Further Information

Publication History

Publication Date:
01 December 2004 (online)

Also available ateRef
   Permissions and Reprints
Preview

The risks of complications of endoscopic retrograde cholangiopancreatoscopy (ERCP), including fatalities, increases with a decreasing probability of bile duct stones [1] [2]. A useful phrase to bear in mind is that ERCP is most dangerous for people who need it least [3]. This is the most important reason why the goal of patient selection for ERCP is to ensure that an absolute minimum of patients are subjected to a futile and risky procedure without a need for therapeutic intervention. On the other hand, a necessary ERCP must not be delayed, since this exposes the patient to the risks of complications such as cholangitis and pancreatitis.

Patients with gallbladder stones have concurrent bile duct stones in 8 - 18 % of cases, depending on the population. The probability ranges from less than 1 % in patients without symptoms or signs suggestive of bile duct stones [4] [5] to 100 % in patients with cholangitis in whom ultrasonography or magnetic resonance imaging documents a stone [4] [6]. Overlooked bile duct stones are found in less than 4 % of patients after cholecystectomy in Western populations [7] [8] [9].

With gallstone prevalences of 8 - 22 % [10] and cholecystectomy rates in Europe and the USA of 69 - 200 per 100 000 [10] [11] [12], the quantitative problem is significant. The endoscopic method is unchallenged as the cornerstone for bile duct stone removal, since laparoscopic bile duct stone removal has not gained widespread general use even though it was first introduced 15 years ago [13] and several groups have published series including more than 200 successful cases [5]. Laparoscopic bile duct clearance adds 1 h to the procedure time in laparoscopic cholecystectomy [5] - but so does intraoperative ERCP, without prolonging or complicating the postoperative course [5]. Randomized studies comparing preoperative [14] [15] or postoperative [16] endoscopic duct clearance with a single-stage laparoscopic procedure have found comparable clearance rates in the ERCP and laparoscopy groups.

Recently, a randomized study comparing a wait-and-see approach with cholecystectomy after sphincterotomy and stone removal, including all patients regardless of age (which ranged from 18 to 80 years) found that 37 % needed cholecystectomy, although those needing surgery were younger than those who did not [17]. In the middle-aged and elderly (who are at greater risk for complications after laparoscopic cholecystectomy than younger patients [18]), it appears that no more that one-fifth will have symptoms from the gallbladder stones later, and almost 90 % of these patients have symptoms leading to cholecystectomy within 2 years of ERCP [19]. This makes endoscopic treatment with clinical follow-up, rather than cholecystectomy with concurrent bile duct intervention, the preferred primary choice for these patients.

Several groups have taken up the challenge of developing and evaluating methods of calculating the probability that the patient may have bile duct stones, in order to fine-tune the selection of patients for ERCP, and have applied these methods clinically. The methods range from relatively simple risk-factor identification [7] [20] to mathematical models made simple to use as nomograms [21] or formulas leading to a discriminant factor [22], and multivariate analysis using logistic regression to identify independent predictors of stones [23] or a need for therapy [24].

When only a single criterion is used for selection, the number of negative ERCPs increases, as shown by Katz et al. in the present issue of Endoscopy [7]. Methods using combined criteria will result in higher predictive values. Everyday use of mathematical formulas is relatively simple, since it can be integrated into the laboratory’s computers, with a numerical value being printed out, or can be used in personal digital assistant-based calculators. The use of a neural network for prediction has been evaluated and found useful in a small series [25].

With selection tests resulting in a probability figure [21] [24], the clinician needs to decide on which cut-off levels to use. In this context, the classification levels usually applied are a low probability (< 10 %), intermediate probability (11 - 55 %), and high probability (> 55 %) of common bile duct stones [26], but 75 - 80 % is probably a more appropriate level for high probability if nontherapeutic ERCP is to be avoided [24].

How good are the methods of detecting those patients who harbor bile duct stones and correctly identifying those who do not? This issue has been analyzed both by the originators of such methods themselves [21] [23] [24] [25] [27] [28] and by other groups, thus independently validating the proposed models [29]. For a selection method to be a useful tool for the clinician, its positive and negative predictive values should be high - ideally 100 %, indicating that all patients with a positive selection test have bile duct stones, and none of those with a negative selection test have stones. This is, of course, never the case, but the task is to get as close to this ideal as possible.

What is a dilated common bile duct? In patients without disorders affecting the bile ducts, the diameters of normal common bile ducts range up to 10 mm, depending on where the duct is measured [30]. The diameter increases somewhat with age [30] [31] [32], and slightly, but significantly, after cholecystectomy [33] [34]. There is a discrepancy between ultrasound measurement and measurement on endoscopic cholangiograms, depending on the way in which the ultrasonography is performed (transverse or anteroposterior scanning), since the duct may be oval in shape, especially when dilated [35]. In addition, there is significant interobserver variation between trainees and trained ultrasonographers in measuring the bile duct diameter [36]. These limitations imply that a high cut-off level should be applied, in order to minimize the number of patients subjected to further investigations. Suggested levels are 8.5 mm in patients older than 50 [30], or, probably better, 10 mm for the extrahepatic ducts on transverse measurement [31] [32] and 5 mm for the intrahepatic ducts. Future studies should include the use of transverse scanning, and observer variation must be taken into account.

A thorough meta-analysis of studies describing predictive factors for common bile duct stones, ranked these [20]. Predictors with likelihood ratios above 10 were cholangitis, preoperative jaundice, and ultrasound visualization of common bile duct stones. Positive likelihood ratios for a dilated common bile duct on ultrasound, hyperbilirubinemia, and jaundice, ranged from almost 4 to almost 7. Elevated levels of alkaline phosphatase, pancreatitis, cholecystitis, and hyperamylasemia showed positive likelihood ratios of less than 3.

The diagnostic methods for visualizing the ducts and bile duct stones, including endoscopic ultrasonography (EUS), magnetic resonance cholangiopancreatography (MRCP), helical computed tomographic cholangiography, intraoperative ultrasonography, and intraoperative cholangiography have been extensively reviewed in a recent technology review [37]. The present evidence supports the view that abdominal ultrasonography is the most cost-effective initial imaging test in the initial work-up of patients with suspected common bile duct stones (level of evidence 2A, grade of recommendation B [38]), and that predictive models are useful in stratifying the risks of patients who have common bile duct stones (1B, A), in order to identify those with a low risk who may proceed to laparoscopic cholecystectomy (2B, B), with preoperative ERCP being reserved for those with a high probability of common bile duct stones (2B, B), while patients with an intermediate risk may be further evaluated with MRCP, EUS, or intraoperative cholangiography (2B, B) [37].

Thus, patients with suspected common bile duct stones who have cholangitis, persistent jaundice, or stones visualized on ultrasonography, MRCP, or CT and patients with a calculated probability of common bile duct stones higher than 75 - 80 %, or a positive discriminant factor, can proceed directly to ERCP [20] [29] [37]. It should be borne in mind that a negative discriminant factor does not exclude the possibility that endoscopic therapy may be needed [29]. This strategy will ensure that the number of patients exposed to unnecessary ERCPs is kept as low as possible.

References

S. Adamsen, M. D.

Dept. of Gastrointestinal Surgery D113 · Herlev Hospital

DK-2730 Herlev · Denmark

Fax: + 45-4488 4009

Email: sven.adamsen@dadlnet.dk