Difficult Stones

The most difficult circumstances encountered during endoscopic stone removal result from technical difficulties with achieving deep biliary cannulation or in performing ES. These difficulties are sometimes because of an inaccessible papilla related to aberrant anatomy or unfavorable duodenal or papillary structures, such as a periampullary diverticulum, or prior surgery, such as Billroth II or Roux-en-Y reconstruction. Techniques have been described for the unique challenge of selective bile duct cannulation in a patient with a Billroth II partial gastrectomy (Lin et al, 1999). The performance of ES is also a challenge, because the visualized anatomy is inverted. In especially difficult cases, needle-knife sphincterotomy with a stent, nasobiliary drain, or guidewire used as a guide for cutting may be an option, or specially designed reverse-direction accessories. The literature on techniques of cannulation and sphincterotomy in Roux-en-Y reconstructions is limited (Wright et al, 2002), but some success has been reported with balloon enterosocopy or overtube-assisted ERCP (Koornstra et al, 2008, Kikuyama et al, 2009).

When ES has been successfully performed, extraction may be hindered by a variety of stone factors—including size, number, consistency, shape, and location of stones—ductal factors such as contour and diameter at the level of and distal to the stones, and the presence of coexisting pathology (e.g., stricture or tumor). Stones that are likely to be more difficult to extract and may require adjuvant techniques to remove them are those that appear larger than the endoscope on radiographic imaging (usually >15 mm); stones that are numerous or hard in consistency; stones that are square, piston shaped, or faceted that tightly fit the bile duct or that are packed against each other; intrahepatic stones; or stones located proximal to a stricture or narrowed distal bile duct or in a sigmoid-shaped duct. Techniques that have been developed to reduce stone size and facilitate endoscopic removal include mechanical lithotripsy, extracorporeal shock-wave lithotripsy (ESWL), intracorporeal lithotripsy with laser or electrohydraulic probes, and chemical contact dissolution therapy. Treatment options must be discussed jointly by the endoscopist, surgeon, and interventional radiologist when difficulties are encountered (Fig. 36.6; see Chapter 27).

FIGURE 36.6 Endoscopic retrograde cholangiopancreatography showing a dilated bile duct containing multiple faceted stones in a postcholecystectomy patient positioned such that standard extraction techniques might be difficult.

Mechanical Lithotripsy

Removal of large CBD stones is a challenge for the most skilled endoscopist. Mechanical lithotripsy (see Chapter 27; Leung & Tu, 2004; Leung et al, 2001) remains the best initial option for stones that cannot be removed by conventional techniques, because it can be applied safely and effectively during the initial endoscopic procedure. Standard retrieval baskets contained within a polytef (Teflon) sheath may fail to crush large stones, because most exceed the breaking strength of the catheter. Standard-type, lithotripsy-compatible, and lithotripsy-convertible baskets are commercially available. Mechanical lithotripters are modifications of standard Dormia baskets and possess great tensile strength (Fig. 36.7). The reinforced basket is opened in the CBD, and the stone is entrapped within the braided wires. This procedure can be performed through the endoscope instrumentation channel, or it can be done after the endoscope has been removed from the patient and a metal sheath has been extended over the inner Teflon catheter. The end of the metal sheath is attached to a winding mechanism, which retracts the basket when cranked and impales the stone against the rigid distal end of the metal sheath leading to stone fracturing. The stone fragments can be removed with the same basket or a standard retrieval basket or balloon. In experienced centers, this technique allows removal of more than 90% of difficult bile stones that are refractory to standard extraction techniques (Chang et al, 2005; Shaw et al, 1993; Van Dam & Sivak, 1993).

FIGURE 36.7 Endoscopic retrograde cholangiopancreatography sequence showing use of transendoscopic mechanical lithotripsy. A, Positioning of the mechanical lithotripsy basket with its metallic sheath in the proximal bile duct (left) and its slow withdrawal toward the distally placed stone to entrap it (right). B, Process of lithotripsy after stone entrapment within the basket (left) as the basket wires cut through the stone (right) to produce stone fragmentation.

Other Lithotripsy Modalities

For the 5% of patients with biliary stones resistant to ES and mechanical lithotripsy, other methods are available, including intracorporeal techniques (laser or electrohydraulic probes) and ESWL (Adamek et al, 1996). The choice between these methods or surgery depends largely on availability and local expertise.

Electrohydraulic Lithotripsy

Since its development during the 1950s in the former Soviet Union as a method to fragment rocks during mining, electrohydraulic lithotripsy has been adapted for medical use in the treatment of nephrolithiasis and, more recently, biliary tract calculi. The electrohydraulic probe consists of two coaxially isolated electrodes at the tip of a flexible catheter, which is capable of delivering electric sparks in short, rapid pulses leading to sudden expansion of the surrounding liquid environment and generating pressure waves that result in stone fragmentation (Picus, 1990). Stone contact with the electrode is achieved through the use of basket or balloon catheter systems (Siegel et al, 1990; Arya et al, 2004) or under direct visual targeting (Fig. 36.8) through the working channel of a daughter endoscope (Hixson et al, 1992), minimizing the risk of CBD injury and perforation. Reports document complete stone clearance after multiple sessions in 86% of patients (Adamek et al, 1996; Hixson et al, 1992; Siegel et al, 1990; Arya et al, 2004), and in a prospective nonrandomized trial, electrohydraulic lithotripsy was comparable to ESWL in stone clearance (Adamek et al, 1996). Its main advantages over laser lithotripsy are its lower cost and increased portability.

FIGURE 36.8 Use of carrier duodenoscope and direct cholangioscope (arrow) inserted within the bile duct with protruding electrohydraulic lithotripsy probe (curved arrow) in close proximity to the stone to be fragmented (not visible).

Dissolution Therapy

Contact chemical dissolution of stones (see Chapter 28) has been attempted by perfusing the CBD with solvents administered via an indwelling nasobiliary tube, percutaneous transhepatic catheter, cholecystostomy tube, or an existing T-tube. The initial results with these agents were disappointing because of incomplete stone dissolution and complications. A semisynthetic vegetable oil, monooctanoin, composed of 70% glycerol-1-monooctanoate and 30% glycerol-1,2-dioctanoate, was used experimentally for the dissolution of CBD stones beginning in 1977. Results collected from 222 clinicians treating 343 patients with CBD stones between 1977 and 1983 reported a success rate for complete stone dissolution of only 25.6% and an additional partial success rate of 28% (Palmer & Hoffmann, 1986). Serious adverse effects leading to discontinuation of treatment occurred in 5% of patients, including hemorrhage from duodenal ulceration, acute pancreatitis, jaundice, pulmonary edema, acidosis, anaphylaxis, septicemia, and leukopenia, but no deaths were reported. The use of organic solvents, such as the aliphatic ether methyl tert-butyl ether (MTBE) (Allen et al, 1985), also has been disappointing, with complete stone dissolution achieved in only 30% to 45% and an unacceptable complication rate related to systemic absorption from spillover of solvent into the duodenum and intrahepatic bile ducts (Brandon et al, 1988; Diaz et al, 1992; Kaye et al, 1990; Murray et al, 1988; Neoptolemos et al, 1990). Sophisticated computer-controlled two-way pump systems may reduce complications and improve efficacy, but total success for large CBD stones using cholesterol solvents is unlikely, because these stones are composed primarily of bile pigments with small concentrations of cholesterol. Expectations of developing a solvent-chelating agent (ethylenediaminetetraacetic acid) for pigment stones have not been realized. As a result of its low efficacy and morbidity, contact dissolution therapy has not assumed an important role in patients with refractory CBD stones, and newer agents with better methods for instillation are awaited.

Endoprosthesis Placement

In the few situations in which stone extraction is incomplete or impossible because of stone size, local anatomy, bleeding, or technical difficulty leading to incomplete ES, a nasobiliary tube (Fig. 36.9) or endoprosthesis (Fig. 36.10) must be inserted to provide biliary decompression and prevent stone impaction in the distal CBD. This is a temporizing therapy to allow the patient’s clinical condition to improve, until complete stone clearance is achieved via additional endoscopic maneuvers or surgery. In patients with uncertain duct clearance, a follow-up cholangiogram can be obtained via a nasobiliary tube without the need for repeat ERCP.

FIGURE 36.9 Radiograph showing placement of a nasobiliary drainage tube (small arrows) for the initial treatment of severe cholangitis caused by a single bile duct stone (large arrow).

FIGURE 36.10 Endoscopic retrograde cholangiopancreatography showing placement of a 10-Fr diameter endoprosthesis for the temporary management of cholangitis caused by several bile duct stones.

Nasobiliary tubes are rarely tolerated beyond a few days. The primary problem with tube placement has been accidental dislodgment, and this has led to the alternative therapy of temporary biliary endoprosthesis placement (Kiil et al, 1989; Rustgi & Schapiro, 1991). In a poor-risk surgical patient, ES and long-term placement of a biliary endoprosthesis has been proposed as a nonsurgical alternative (Cotton et al, 1987; Foutch et al, 1989; Nordback, 1989; Soomers et al, 1990). Stent patency is not of major clinical importance in this situation, because the sphincterotomy typically provides an adequate conduit for bile flow around the prosthesis, which then serves principally to prevent stone impaction. The use of oral ursodeoxycholic acid in these patients has resulted in CBD stone disappearance, but additional studies are necessary to determine the true efficacy in this setting (Johnson et al, 1991). Of 84 patients intentionally treated with permanent stents for endoscopically irretrievable stones and followed for a mean of 3 years, 49 (58%) developed biliary complications, and nine died as a result of complications. Most of the patients had a long, symptom-free interval, however, before complications developed, supporting stenting as effective short-term treatment (Bergman et al, 1995; Maxton et al, 1995). In a randomized study with a shorter mean follow-up (1.5 years) that compared placement of a plastic biliary stent with ES as definitive therapy with stone clearance by means of the basket, balloon, or mechanical lithotripter, the patients with stents had a significantly greater rate of cholangitis (36%) than the patients managed by a conventional endoscopic duct clearance approach (14%). The high risk of long-term complications does not support the concept of permanent stent therapy except for patients with severe comorbidity and a short life expectancy.

Complications of Endoscopic Therapy

Early complications of ES have been well documented, and despite the disparate indications and selection of patients among centers, the incidence seems to be remarkably consistent at 5% to 10% (Cotton, 1984; Cotton & Vallon, 1981; Freeman et al, 1996; Geenen et al, 1981; Leese et al, 1985a; Masci et al, 2001; Seifert et al, 1982; Siegel, 1981; Vandervoort et al, 1996). The expected higher complication rate during early experience with ES is reflected in a single-center series comparing the results of the first 394 procedures (Leese et al, 1985a), which carried an overall morbidity rate of 10.4%, with a subsequent consecutive group of 300 sphincterotomies, in which this rate decreased to less than 6%. However, the respective proportions of individual complications remain similar in most reports: acute hemorrhage from the sphincterotomy site, 2% to 2.9%; acute pancreatitis, 1.5% to 5.4%; cholangitis, 1% to 2.7%; and retroperitoneal perforation, 0.3% to 1%, with small numbers of other problems, such as impacted basket, gallstone ileus, and acute cholecystitis overall accounting for 1.1%. Emergency surgery was required in 1% to 2.5% of cases for bleeding, cholangitis, perforation, and pancreatitis in descending order of frequency. Complication rates must be interpreted with caution, because definitions of hemorrhage, acute pancreatitis, cholangitis, and perforation often differ and are not always given.

A consensus conference provided a set of standards for defining complications and their severity (Cotton et al, 1991a). We have always included any episode of overt bleeding (hematemesis or melena) and a decrease in hemoglobin of 2 g/dL or more after ES as significant hemorrhage, although some authors have included only episodes requiring transfusion. Use of aspirin or related drugs in the usual doses does not seem to increase the incidence of hemorrhage (Freeman et al, 1996). The diagnosis of pancreatitis and cholangitis must depend on the presence of clinically recognizable syndromes rather than asymptomatic hyperamylasemia or transient elevation of temperature alone. There does not seem to be a significant influence on the rate or type of complication based on the initial presentation of the CBD stone (pain alone, jaundice alone, pancreatitis, a combination of these), except that cholangitis and cholecystitis are more likely after ES if cholangitis is preexisting. Present evidence does not suggest that complications are more likely in older patients or after previous biliary surgery (Clarke et al, 2001).

Statistically significant risk factors for complications include difficulty in cannulation, access (precut) papillotomy, suspected sphincter of Oddi dysfunction, more than two pancreatic cannulations, acinarization of the pancreas, failed biliary access or drainage, and technical skill of the endoscopist; the age and general medical condition of the patient were not statistically significant (Freeman et al, 1996; Neoptolemos et al, 1989; Vandervoort et al, 2002). Duodenal diverticula, although sometimes rendering ERCP and ES technically more difficult, do not seem to add any further risk. Many series do not include nonendoscopic complications occurring after ES, such as cardiovascular, cerebrovascular, or respiratory events. In addition, although surgery for the treatment of complications is usually documented, surgery for failed endoscopic therapy is often not documented. These factors are important if comparative data from the surgical literature are to be interpreted correctly.

Management of Complications

The management of complications by centers performing ES is well standardized, but many patients are referred from other hospitals to which they are returned after the procedure, where experience in managing complications may be limited. Of all complications, hemorrhage usually requires surgical intervention, but this occurs only in a few cases. In the rare major arterial hemorrhage, endoscopic view of the papillary area is often obscured completely by blood, and further endoscopic therapy may be impossible. The sphincterotomy usually is converted to a formal surgical sphincteroplasty, which includes the likely bleeding artery. Immediate methods for hemostasis include balloon tamponade, direct bipolar electrocautery, injecting or washing the area with 1 : 10,000 epinephrine solution, application of hemostatic clips, laser coagulation, superselective arterial catheterization and embolization, and infiltration with sclerosant (Grimm & Soehendra, 1983; Leung et al, 1995; Wilcox et al, 2004).

Acute pancreatitis is managed in the standard fashion (see Chapters 53 and 54), and although many attacks are mild and self-limited, clinicians should not be complacent, because some attacks are more severe and should be graded and treated appropriately. Wire-guided cannulation decreases the risk of post-ERCP pancreatitis when compared to the contrast-injection method (Cennamo et al, 2009). There is no evidence that pre-ES or post-ES administration of aprotinin (Trasylol) or glucagon influences the incidence or severity of pancreatitis. In contrast to hemorrhage, the onset of pancreatitis may be delayed for several hours or, rarely, for 1 or 2 days. Gabexate, a synthetic protease inhibitor, and somatostatin have been shown in randomized, double-blind controlled trials to reduce the extent of enzyme elevations and the incidence of pancreatitis after ERCP (Cavallini et al, 1996; Masci et al, 2001; Poon et al, 2003). Gabexate was infused for 30 to 90 minutes before and for 6.5 to 12 hours after ERCP, whereas somatostatin was administered either as a bolus infusion or in a manner similar to gabexate (Bordas et al, 1998; Poon et al, 1999, 2003). A trial suggested diclofenac administered as a rectal suppository after ERCP reduced the incidence of post-ERCP pancreatitis (Murray et al, 2003). Trials using secretin are in progress.

Cholangitis is confined almost completely to patients in whom CBD clearance has not been achieved, and measures should be directed at providing adequate bile drainage (e.g., by nasobiliary catheter or endoprosthesis) and administering parenteral antibiotics (see Chapter 43). Emergency surgery for cholangitis carries high risk but is indicated in patients who do not improve within 24 hours.

Perforation may be asymptomatic and noticed only as retroperitoneal gas (Fig. 36.11) or extravasation of radiographic contrast material, but even in a symptomatic patient, conservative treatment is often effective, with spontaneous resolution and avoidance of potentially difficult surgery. Occasionally, this complication presents late after ES with a retroperitoneal collection of bile or pus in the flank or inguinal region (Neoptolemos et al, 1984a; Leese et al, 1985a) and requires percutaneous or surgical drainage.

FIGURE 36.11 Computed tomographic scan of the stomach with oral contrast administration showing extensive retroperitoneal, intraperitoneal, and subcutaneous air caused by a perforation from endoscopic sphincterotomy. The patient made an uneventful recovery on conservative treatment.

Gallstone ileus is a rare complication, but its recognition needs to be emphasized, because symptoms may be obscure in elderly patients, and they occur many days after ES and stone release; treatment is along standard surgical lines.

The impaction of an extraction basket occurs rarely in experienced hands, because many endoscopic maneuvers have been learned to prevent or salvage this situation. These maneuvers include 1) avoiding basket closure during initial attempts to extract a large stone to prevent impaling the basket wires in the stone surface; 2) converting the standard basket into a crushing type by replacing the handle with a mechanical lithotripter; and 3) extending the ES incision by removing the duodenoscope over the impacted basket catheter and reintroducing it alongside the catheter, introducing a second duodenoscope, or passing a sphincterotome along the same instrument channel as the impacted basket catheter when using large channel (3.7 mm or 4.2 mm) endoscopes.

Mortality

Death after ES has not been reported in a standardized way. Deaths directly attributable to the procedure are fairly constant at 0.8% to 1.5% (Cotton, 1984; Cotton & Vallon, 1981; Geenen et al, 1981; Leese et al, 1985a; Nakajima et al, 1979; Reiter et al, 1978; Safrany, 1978; Seifert et al, 1982; Siegel, 1981; Sivak, 1989), with almost equal distribution of causes between hemorrhage, pancreatitis, cholangitis, and perforation. Most of the deaths are postprocedural. Mortality rate as a proportion of complications ranges from 7% to 17%, which presumably reflects the comprehensive reporting of all complications by some authors, but reporting of only more severe complications by others. Using the accepted method of reporting 30-day surgical mortality, mortality within 1 month of ES was 3.3% compared with 0.8% immediate postprocedural mortality (Leese et al, 1985a). The additional deaths were due to a variety of vascular, respiratory, renal, infectious, and malignant conditions in a group whose mean age was 79 years. Another group of 59 elderly patients considered unfit for surgery by surgeons underwent ES (Neoptolemos et al, 1984b) with only one death (1.7%).

Long-Term Morbidity

Long-term follow-up of 1 to 15 years after ES in postcholecystectomy patients (Cotton, 1984; Escourrou et al, 1984; Hawes et al, 1990; Ikeda et al, 1988; Rosch et al, 1981; Seifert et al, 1982; Sivak, 1989) shows more than 90% of patients to be well on symptomatic review and 7% to 11% have significant symptoms secondary to recurrent stones (5%), with or without stenosis of the ES site (1.5% to 3%) and cholangitis (2%). One study with long-term follow-up of 15 years after ES found an early complication rate of 15% with a high incidence of late complications, including recurrent stones and pancreatitis in 24% (Bergman et al, 1996); however, this study included only procedures performed between 1976 and 1980. A higher rate of asymptomatic stones and stenosis might result from the use of plain abdominal radiography for detecting pneumobilia, barium studies for detecting duodenobiliary reflux, radionuclide scanning, and repeat ERCP, which do not seem justified outside clinical trials. Most of these long-term complications are amenable to further endoscopic treatment.

Percutaneous Approach

In the 5% to 10% of patients for whom the endoscopic approach is unsuccessful at clearing the CBD of stones, two nonsurgical approaches are available: a rendezvous procedure and a complete percutaneous procedure (see Chapters 27 and 28). If the papillary region can be reached endoscopically, the rendezvous procedure may be used. This involves a two-stage procedure with the introduction of a percutaneous guidewire through the bile ducts and papilla into the duodenum in the first stage followed by an ERCP (Dowsett et al, 1989; Martin, 1994; Verstandig et al, 1993). In one report of rendezvous procedures for choledocholithiasis, 0.9% (15 of 1753) of ERCPs were unsuccessful at biliary cannulation, usually owing to duodenal diverticula or Billroth II anatomy (Calvo et al, 2001). Three patients underwent surgery, and 93% of the remaining patients underwent a successful rendezvous procedure. There was one complication with a retroperitoneal perforation that required surgical management, and during follow-up, only one patient developed recurrent choledocholithiasis requiring a repeat rendezvous procedure. In the multicenter prospective trial of endoscopic biliary sphincterotomy complications (Freeman et al, 1996), the combined endoscopic-percutaneous approach was a risk factor for the development of complications, with a high rate of complications at 22.6% with 6.5% classified as severe.

The complete percutaneous approach is labor intensive and typically requires multiple sessions. It has been established as treatment for hepatolithiasis (Yeh et al, 1995; see Chapters 39 and 44), but it may be used for choledocholithiasis. The procedure involves initial establishment of a transhepatic fistula, followed by stone extraction under fluoroscopy or cholangioscopy 7 to 8 days after the fistula forms. In a series of 31 patients with failed endoscopic procedures, percutaneous biliary access was achieved in all patients, and stone clearance was complete in 87% after a mean of 5.6 sessions (Van der Velden et al, 2000). All patients underwent balloon dilation of the papilla, and most patients required additional means of stone fragmentation, including mechanical lithotripsy, electrohydraulic lithotripsy, and ESWL. Complications including pancreatitis and bacteremia occurred in 9.7% and did not require surgical intervention. The four patients who did not respond to percutaneous management underwent surgery. Bile duct access by endoscopic ultrasonography has largely removed the need for percutaneous approaches (see Chapter 14).

Open Bile Duct Exploration (See Chapters 29 and 35)

Morbidity and mortality after open bile duct exploration have been well documented (Allen et al, 1981; Blamey et al, 1983; Dixon et al, 1983; Doyle et al, 1982; Girard & Legros, 1981; Glenn, 1974, 1975; Kune, 1972; McSherry & Glenn, 1980; Pitt et al, 1981; Vellacott & Powell, 1979; Way et al, 1972), with examination of specific risk factors (Blamey et al, 1983; Dixon et al, 1983; Pitt et al, 1981) that allow some prediction of the likelihood of complications and perhaps predict the need for preoperative biliary drainage. Direct comparison with the endoscopic data is not valid, because extremely diverse groups of patients are being treated by these modalities. This situation highlights the concept of “apples and oranges” succinctly stated by Cotton (1984), but mention must be made of published surgical figures to enable some clinical judgments to be made in the absence of randomized trials. It is immediately apparent that, in contrast to endoscopic therapy, surgical morbidity and mortality are determined by patient age, comorbidities (Glenn, 1975; McSherry & Glenn, 1980; Vellacott & Powell, 1979), hemotologic and biochemical factors (Blamey et al, 1983; Dixon et al, 1983; Pitt et al, 1981), and whether the operation is elective (Houghton & Donaldson, 1983; Sullivan et al, 1982) or emergent (Boey & Way, 1980; Thompson et al, 1982). Mortality rates range from 1% in relatively fit younger patients to 28% in unfit and elderly patients and 12% to 14% in younger patients undergoing emergency surgery for cholangitis. These figures nearly all refer to primary bile duct operations and should be compared only with results of ES when the gallbladder is in situ. Equivalent results for secondary bile duct explorations are less well recorded, but mortality rates of 2% are possible for elective surgery (Girard & Legros, 1981; McSherry & Glenn, 1980). Consideration also must be given to biliary drainage operations, which avoid problems of stone retention, although this may be at the expense of an increased postoperative morbidity and mortality. Average mortality rates are 1% to 5% for choledochoduodenostomy and transduodenal sphincteroplasty (Capper, 1961; Jones, 1978; Lygidakis, 1981; Madden et al, 1970; Speranza et al, 1982; Stuart & Hoerr, 1972). A review of 246 such operations (Baker et al, 1985) in which one of these drainage procedures was used revealed a mortality rate of 5.4% for each of the two types of operations and major morbidity rates of 12% for choledochoduodenostomy and 21% for transduodenal sphincteroplasty. The results stated here should be read in conjunction with those reported in Chapters 29 and 35.

Long-Term Follow-up of Surgically Explored and Reexplored Bile Ducts

Long-term follow-up data after open bile duct exploration are lacking. The wide variation in reported recurrent stone rates has been mentioned, but morbidity and the need for further surgery have been found in 5% after 5 years (Larson et al, 1966), 10% after 12 years (Peel et al, 1975), and 21% after 6 to 11 years (Lygidakis, 1983) when exploration without biliary drainage has been performed. After choledochoduodenostomy (see Chapter 29), there was no morbidity in a 6- to 11-year follow-up (Lygidakis, 1983) with similar results for transduodenal sphincteroplasty (Degenshein, 1964; Stuart & Hoerr, 1972). Follow-up of 90% of the survivors of one series of 246 patients 1 to 12 years postoperatively (mean, 4.4 years) revealed that complications had been treated in 3% of the choledochoduodenostomy group (mainly sump syndrome) and 6% of the transduodenal sphincteroplasty group (mainly cholangitis), with additional symptoms discovered on interview in 8% of the choledochoduodenostomy and 5% of the transduodenal sphincteroplasty groups.

Laparoscopic Common Bile Duct Exploration

Laparoscopic CBD exploration (see Chapter 34) was first performed in the early 1990s and is now used increasingly in experienced tertiary referral centers (Khoo et al, 1996; Martin et al, 1998; Rhodes et al, 1995). Ductal exploration may be accomplished through the cystic duct or directly through a choledochotomy. The transcystic route is the least invasive and generally does not require any ductal manipulation or drainage procedure, whereas choledochotomy requires either closure of the duct over a T-tube or primary closure of the choledochotomy with or without a biliary stent placed in an antegrade fashion without need for a T-tube (Huang et al, 1996; Rhodes et al, 1995). Bile duct clearance rates average 90%, with a median rate of conversion to open operation of 4% (Tranter & Thompson, 2002). Complication rate is 2.5%, with a median mortality rate of 1% (Strasberg et al, 1995).

In a randomized trial of laparoscopic CBD exploration versus postoperative ERCP for CBD stones (Rhodes et al, 1998), the clearance rates at the first intervention were the same (75%). With subsequent treatment, predominantly in the form of repeated ERCP, duct clearance rates approached 100% in both groups. In a similar study sponsored by the European Association of Endoscopic Surgeons (Cuschieri et al, 1996), patients were randomized to preoperative ERCP followed by laparoscopic cholecystectomy or laparoscopic cholecystectomy with laparoscopic CBD exploration. Duct clearance was similar in the two groups, but there was a higher rate of conversion to open surgery and decreased length of stay with single-stage surgical treatment. Our group has shown that therapeutic ERCP can be performed safely as an outpatient procedure without need for hospitalization (Tham et al, 1997). To minimize hospital stay, the outpatient ERCP would need to be coupled with laparoscopic cholecystectomy, which requires good coordination between the surgeon and the endoscopist.

Postcholecystectomy with T-Tube In Situ (See Chapter 29)

CBD stones of less than 10 mm in diameter detected in the early postoperative period on T-tube cholangiography may pass spontaneously or with hydrostatic pressure from flushing or perfusing the T-tube. Most of these stones require additional mechanical manipulation, however. Increased morbidity, mortality, and retained stone rate after secondary bile duct explorations stimulated the development of alternative techniques, including hydraulic T-tube irrigation with or without pharmacologic relaxation of the sphincter of Oddi with glucagon or nitrates (Tritapepe et al, 1988), T-tube infusion of cholesterol solvents (Tritapepe et al, 1988), T-tube tract choledochoscopy and lithotripsy (Josephs & Birkett, 1992), and percutaneous extraction of stones through a mature T-tube tract. Success rates with the last technique have ranged from 77% to 96%, but multiple sessions are often required and the technique may be complicated by sepsis, biliary trauma, and biliary leakage in 4% to 8% of patients (Burhenne, 1980; Caprini, 1988; Cotton, 1990; Mason, 1980; Nussinson et al, 1991; Tritapepe et al, 1988). Delays of 4 to 6 weeks are required before manipulation to allow maturation of the T-tube tract.

Alternatively, early ES can be applied safely and effectively after stone detection without the need for T-tube maturation, allowing timely discharge from hospital. The results from eight endoscopic series totaling 337 patients indicate an overall endoscopic success rate of 90% with a morbidity rate of 7% and mortality rate of 0.6% (Bickerstaff et al, 1988; Danilewitz, 1989; Lambert et al, 1988; Nussinson et al, 1991; O’Doherty et al, 1986; Simpson et al, 1985; Soehendra et al, 1981; Tandon et al, 1990). The choice between these two techniques depends on local expertise, because direct comparisons have not been made in controlled trials.

Postcholecystectomy Without T-Tube

The incidence of retained CBD stones is approximately 2% to 5% after conventional and laparoscopic cholecystectomy and 5% to 15% after CBD exploration. ES remains the treatment of choice for elderly patients seen days to years after cholecystectomy (Danilewitz, 1989), because these patients are at higher risk from further abdominal surgery (Cranley & Logan, 1980; Sheridan et al, 1987). ES also is recommended for choledocholithiasis in young, fit postcholecystectomy patients, even those with normal diameter ducts (Cotton et al, 1998).

Pregnancy

Symptomatic choledocholithiasis during pregnancy poses a diagnostic and therapeutic challenge. The frequency of choledocholithiasis in pregnancy requiring intervention may be 1 in 1200. Traditionally, intervention in these patients has been surgical. Although surgery has an overall low morbidity and mortality for the expectant mother, it carries a 4- to 6-week recovery period with a possible increased risk of fetal loss.

Endoscopic treatment can be performed safely with a complication rate comparable to that in nonpregnant patients (Jamidar et al, 1995; Kahaleh et al, 2004; Tham et al, 2003, Simmons et al, 2004). We reviewed our experience with 15 pregnant patients who underwent ERCP (Tham et al, 2003). Precautions were taken to minimize fetal exposure to radiation, which included covering the pelvis of all patients with an abdominal lead apron and minimizing fluoroscopy. No adverse outcomes occurred to the mother or the fetus, and the average fetal radiation dose was 310 mrad, which is below the 5- to 10-rad level considered teratogenic. These patients should undergo cholecystectomy on completion of pregnancy to prevent recurrence.

Patients with Gallbladder In Situ

In elderly patients or in patients with significant comorbidity, a deliberate decision often is made to leave the gallbladder in situ after ES and CBD stone removal. The short-term and long-term results and complications of ES in patients with gallbladders do not differ from those in postcholecystectomy patients (Kaw et al, 2002). One randomized study of high-risk surgical patients reported a higher rate of recurrent biliary symptoms in the ES group compared with patients who underwent open cholecystectomy with operative cholangiography (Targarona et al, 1996); however, details of the endoscopic procedure were not reported in the article, and the failure rate for ES was high at 20%. In view of their shorter life expectancy, high-risk patients can be managed expectantly, unless symptoms dictate otherwise. Careful follow-up for 5 to 10 years in patients with intact gallbladders indicates that approximately 10% develop gallbladder symptoms or complications sufficient to warrant cholecystectomy, and most occur in the first year (Davidson et al, 1988; Escourrou et al, 1984; Hill et al, 1991; Rosseland & Solhaug, 1984; Siegel et al, 1988; Tanaka et al, 1987). The risk for developing gallbladder symptoms is similar to that of patients with asymptomatic cholelithiasis (Gracie & Ransohoff, 1982). Most patients requiring cholecystectomy can be treated electively.

The risk of developing biliary symptoms seems to depend on the continuing presence of stones in the gallbladder. In two reports (Siegel et al, 1988; Tanaka et al, 1987), none of the patients with acalculous gallbladders at initial ES developed symptoms or complications that subsequently required cholecystectomy. In patients with gallstones, there were no reliable predictors to identify patients with cholelithiasis who were likely to develop gallbladder complications after ES. Nonfilling of the gallbladder with radiographic contrast material at ERCP (Worthley & Toouli, 1988) or with radiotracer by nuclear scintigraphy (Holbrook et al, 1991) may confer a higher likelihood of developing symptoms, although results from other series do not concur (Davidson et al, 1988; Hill et al, 1991). These results suggest that the low risk of subsequent gallbladder complications after ES and bile duct clearance, when patients are kept under supervision, outweighs the known higher mortality rate of cholecystectomy in high-risk patients and mitigates against routine cholecystectomy in this group. Few physicians would recommend this approach in a fit patient, in whom the expectation for long-term survival and development of symptoms necessitating cholecystectomy are greater. The risk of elective cholecystectomy is probably smaller than the risk of subsequently developing gallbladder symptoms in this group, so elective cholecystectomy would seem most appropriate. This idea was confirmed by a study of 120 patients with known gallstones who underwent successful ES with bile duct clearance; the study excluded patients who were not surgical candidates, and patients were randomized to laparoscopic cholecystectomy within 6 weeks after ES or a wait-and-see policy with median 2.5-year follow-up. Nearly half the patients in the wait-and-see group developed biliary-related problems, including biliary pain and cholecystitis, which led to cholecystectomy or repeat ERCP or both, whereas none of the patients in the cholecystectomy group experienced biliary-related problems. The wait-and-see group also had a higher rate of conversion to an open procedure at the time of surgery (Boerma et al, 2002).

The place of preoperative ES in young, low-risk surgical patients with suspected choledocholithiasis has been examined to determine whether preoperative ES confers an advantage in terms of morbidity, mortality, and length of hospital stay compared with a conventional approach of open cholecystectomy, operative cholangiography, and CBD exploration. Two prospective, randomized controlled trials (Neoptolemos et al, 1987; Stiegmann et al, 1992) have addressed this question with the conclusion that no advantage is conferred by initial endoscopic bile duct clearance before cholecystectomy. High-risk surgical patients were excluded from these studies, and both studies were conducted in the era of open cholecystectomy, before the use of high-quality MRCP.

Laparoscopic Cholecystectomy and Endoscopic Retrograde Cholangiopancreatography

Laparoscopic cholecystectomy (see Chapter 34) has replaced open cholecystectomy as the procedure of choice for gallbladder removal in most patients with symptomatic cholelithiasis. Optimal management of patients with suspected or documented CBD stones varies according to the expertise available in a particular institution. Laparoscopic transcystic CBD exploration (Hunter, 1992) or direct choledochotomy (Jacobs et al, 1991) can be performed at the time of laparoscopic cholecystectomy, but these procedures require special expertise and equipment. Laparoscopic transcystic choledochoscopy is hindered by the small size and tortuosity of the cystic duct, making instrument passage and stone lithotripsy and removal potentially difficult.

In patients with suspected CBD stones, MRCP has replaced ERCP as the diagnostic modality of choice to identify the presence or absence of CBD stones (see Chapter 17). A need for endoscopic management has evolved, with ERCP playing a central role in preoperative and postoperative CBD stone removal. Initially, we and others (Cotton et al, 1991b) found it useful to stratify patients preoperatively into low, intermediate, and high likelihood for CBD stones (Table 36.1) based on established clinical and biochemical indicators, including elevated liver function tests, a dilated CBD (>8 mm on ultrasound), and a history of jaundice or pancreatitis (Del Santo et al, 1985; Hauer-Jensen et al, 1985; Lacaine et al, 1980). In the absence of clinical or biochemical indicators, CBD stones are present in only 2% to 3% of individuals (Neuhaus et al, 1992b). Small asymptomatic stones may be of minor clinical relevance, because their natural history is unknown, and conceivably many if not most of these calculi would pass spontaneously without ever coming to clinical attention. We have favored selective evaluation of the CBD only for patients with risk factors for choledocholithiasis.

Table 36.1 Cholangiography Stratification for Risk of Choledocholithiasis