2020 WSES guidelines for the detection and management of bile duct injury during cholecystectomy

LC is the preferable approach also for AC and is associated with lower mortality and morbidity rates [28‐34]. The risk of conversion to open surgery appears to be higher with male sex, age > 60 years, obesity, cirrhosis, previous upper abdominal surgery, presence of comorbidity, large bile stones, fever, elevated serum bilirubin levels, gangrenous cholecystitis, severe acute, and chronic cholecystitis, contracted gallbladder on imaging, duration of complaints > 48 h, and emergency LC [35‐37]. Conversion to open surgery may be considered for patient safety if the operating surgeon cannot manage a difficult LC; however, there is no evidence to support that conversion to open per se will avoid or reduce the risk of BDI [5, 38, 39]. The critical view of safety (CVS) technique was introduced in 1995 to guarantee the safest approach to LC by promoting the recognition of gallbladder elements, particularly the hepatocystic triangle (composed of the cystic duct, common bile duct, and liver) [40, 41], a crucial step to reduce the risk of BDI associated with mistakes in visual perception. The literature has demonstrated that when the CVS is identified, the risk of iatrogenic intraoperative complications is minimized [42‐44]. Thus, routine use of CVS is recommended over other techniques, such as the infundibular approach [5, 42, 45, 46]. However, achieving a complete CVS is easily obtained in only 50% of cases. The component most commonly incomplete is clearance of the lower third of the gallbladder from the liver bed, and CVS cannot always be applied if the hepatocystic angle is affected by advanced inflammation or contracting fibrosis due to preceding episodes of inflammation.

It has been reported that injuries of the common bile duct are more common during the early learning curve in laparoscopic cholecystectomy [47]. Thus, the use of CVS could be of greater importance for trainees and residents; in this scenario, the trainee or resident must secure the CVS, and the supervising surgeon must confirm the CVS before the cystic duct and cystic artery are ligated.

Whenever a CVS cannot be achieved and the biliary anatomy cannot be clearly defined, alternative techniques such as the “fundus-first (top-down)” approach or subtotal cholecystectomy (STC) should be considered [5, 48]. Several studies have shown how the “fundus-first” technique is associated with reduced rates conversion rate and iatrogenic complications (including BDIs) during difficult operations, such as in cases of severe AC [49‐52], although the risk of vascular and biliary injuries cannot be completely eliminated [10, 53]. It is essential to recognize approaching areas of danger during LC and, in response, stop dissection and change to a bailout procedure (STC or cholecystostomy) to minimize the need for conversion and to reduce the risk of BDI [48, 54‐56]. However, STC is associated with significantly more surgical site infections, a need for re-interventions, and a longer hospital stay than total cholecystectomy [57]. STC showed advantages over a converted cholecystectomy in which conversion will not solve the difficulty of an inflamed hepatocystic triangle [58].

Intraoperative cholangiography (IOC) is an imaging technique that may be used during LC to recognize choledocholithiasis and define the biliary anatomy [59]. However, its routine use is not currently advisable since it is not associated with a significant reduction in rates of complications and BDIs during LC [60, 61]. Indeed, BDI may also occur after IOC because of misinterpretation of the IOC findings. IOC may be recommended in cases of intraoperative suspicion of BDI, misunderstanding of the biliary anatomy, or even inability to see the CVS, as well as in patients with AC or a history of AC, for whom intraoperative imaging, although associated with longer operative time, could be of greatest benefit [5]. Importantly, identification of a BDI using IOC can lead to earlier diagnosis and treatment.

Alternatively, the use of indocyanine green fluorescence cholangiography (ICG-C) [62] as an intravenous infusion before surgery can be a useful technique to visualize the structures of the biliary tree, particularly the cystic duct, without the need for X-ray imaging. The usefulness of ICG-C to prevent BDIs has been suggested in several studies and has also proven useful for acute and chronic gallbladder diseases and in those situations in which IOCs cannot be used [63‐65].

Systematic reviews analyzing data from RCTs [66] and population-based studies [67, 68] showed higher BDI rates in acute conditions, supporting the hypothesis of increasing BDI risk with increasing severity of local inflammation, such as in the case of acute cholecystitis (AC) [67]. Different time frames for operating on patients presenting with symptomatic AC have been proposed, ranging from no more than 72 h up to 10 days. Further delays are associated with disease progression, and despite medical treatments, unfavorable conditions for safe surgical interventions exist [39]. Indeed, an increase in the complication rate and need of conversion to open cholecystectomy has been reported when the time from symptom appearance to surgery was prolonged [69‐71], with the latest timepoint to safely operate on AC patients being 10 days from symptoms appearance [39].

Classification systems that are essentially based on the biliary injury location include the first classification published by Bismuth in 1982 [92, 93], followed by the Strasberg’s one proposed in 1995 [90], and other classification systems published by McMahon [94], Bergman [95], Neuhaus [96], and Csendes [97]. Conversely, classification systems that integrate vascular injuries into the description of BDIs are the Stewart-Way classification published in 2007 [98], the Hannover classification [17], the one proposed by Lau et al. [99], and more recently the ATOM (Anatomic, Time Of detection, Mechanism) classification published by the European Association for Endoscopic Surgery (EAES) in 2013 [100]. The ATOM integrates the Bismuth, Strasberg, Neuhaus, McMahon, Connor, and Lau classifications into a composite, all-inclusive, nominal system (Tables 2 and 3), which combines bile tract anatomical damage, vascular injury, timing of detection, and mechanism of damage in an exhaustive classification system covering all possible injuries. The EAES intended the ATOM classification as a specific effort toward standardization and transformation of BDI definitions into a unified language. Moreover, the ATOM system was thought to facilitate the collection of data for epidemiological and comparative studies, ultimately leading to a more precise determination of the true incidence of BDI incurred during LC and consequently favoring the development of preventive measures [101]. The main drawback is that it may be too complex and time-consuming to be used in routine clinical practice.

Clinically, BDIs are often grouped into minor or major injuries. Minor BDIs include injuries caused by electrocautery burns or a partial cut from sharp dissection with shears and are not associated with tissue loss. These injuries can typically be repaired primarily with sutures and placement of abdominal drains in the area [102]. Conversely, major BDIs (i.e., Strasberg E) are associated with tissue loss (e.g., the common bile duct is clipped and transected) and require complex reconstruction with a Roux-en-Y hepaticojejunostomy.

Once BDI has occurred and been recognized during LC (approximately 25% of cases [99, 103]), a detailed and precise surgical report will be critically important to guide BDI management. The surgical report must include the indication for the surgery, the patient’s comorbidities, the operative time, the amount of blood loss, the type of injury that occurred (in detail), and the use of drainage.

The ideal report should follow the CVS schema described by Strasberg in 1995 [90]. The CVS is composed of three critical steps: (1) the visualization of the hepatocystic triangle with no exposure of the common bile duct; (2) the exposure of the lower part of the gallbladder and its separation from the liver bed; and (3) the visualization of only 2 structures that enter the gallbladder: the cystic duct and the cystic artery. The surgeon must report during which of the CVS steps difficulties were encountered and BDI occurred.

Whether a timeout during LC, prior to transecting any ductal structure, is performed should be reported. Similarly, it is important to report whether another surgeon was consulted at the time of the dissection or during the difficult steps. Any anatomical abnormality or unusual findings should be described, including:

Whenever an intraoperative biliary tract imaging technique (IOC or ICG-C [104]) is performed, these findings must also be reported, and cholangiography images should be included in the report [105, 106].

Particularly, the following should be specified:

If possible, a drawing of the BDI with biliary drainage positioning (if used) could be helpful. If a videotape of the surgical procedure is available, it should be added to the report [107‐109].

The presence of an unexplained source of bile in the operative field must raise the suspicion of a BDI. In these cases, the use of IOC is helpful to detect BDI, although it requires additional training and longer operative times [7, 115]. A meta-analysis on 860 patients showed that the selective versus the routine use of IOC is associated with a comparable chance of detecting BDI (odds ratio, OR: 0.36; 95% CI: 0.01–8.92; z = 0.63; p = 0.53) [116]. On the contrary, its use appeared to be helpful in terms of BDI risk reduction in patients with AC (moderate or severe) [67, 78].

IOUS could be useful to evaluate vascular injuries associated with BDI and should be preferred to hilar dissection during intraoperative staging to avoid further damage [117].

ICG-C provides real-time imaging of the extrahepatic biliary tract during LC and represents a noninvasive, quick, safe, and easy-to-apply tool [64]. A recent meta-analysis of 19 studies including 772 patients explored the potential of ICG-C to identify biliary structures during LC [118]. Four studies compared the use of ICG-C to IOC in 215 patients and found no significant differences for cystic duct, common bile duct, or common hepatic duct visualization. A recent survey involving 3411 surgeons (with an average of 16.1 years of practice) highlighted how the use of adjuncts such IOC, ICG-C, or intraoperative ultrasound, either routinely or selectively during difficult cholecystectomies, is not significantly associated with a lower risk of BDIs [9]. It is important to emphasize that factors such as geographic distance between facilities, equipment, expertise, and logistics, vary significantly between institutions. Some authors proposed that in cases of suspected BDI, asking the opinion of another surgeon (physically or virtually) may be an easy, effective, and inexpensive alternative to IOC [119].

In the event of BDI detected during LC, surgeons must promptly analyze the injury and choose between an intraoperative repair or “drain now and fix later” strategy [120]. For minor BDIs (i.e., Strasberg A–D and conditionally E2), a direct repair, with or without the placement of a T-tube, and the placement of abdominal drains in the area is considered safe and appropriate [121]. This strategy is reported in 5–58% of BDI cases in the literature [102, 111, 122, 123].

If available on site, endoscopic decompression might be considered in cases of Strasberg A injury [124]. However, the recur to this strategy is blunted by the high rate of repair failure (up to 64%) [111].

For major BDIs (i.e., Strasberg E) associated with tissue loss and whenever an ischemic injury is suspected, a Roux-en-Y hepaticojejunostomy is the recommended method of reconstruction [9, 111, 122, 125‐128], with the placement of a T-tube at a healthy region of the common bile duct, either proximal or distal to the injury, to decrease the incidence of future stricture formation [129]. Any dissection in the hilum may make subsequent reconstruction more difficult or cause further biliary or vascular injury. Thus, in case of insufficient experience in hepato-pancreato-biliary (HPB) surgery, it is recommended to place a drain in the right upper quadrant and transfer the patient to a center with experienced HPB surgeons [129]. Conversion to an open surgery to solely confirm diagnosis or perform injury staging is not recommended.

A recent systematic review and meta-analysis [130] considering 10 low-quality studies showed that on-table repair (by direct suture or bilioenteric anastomosis) is associated with a higher incidence, although non statistically significant, of failure than postoperative repair (60% vs. 34.1%; OR: 2.06; 95% CI: 0.89–4.73; p = 0.09). Moreover, non-expert immediate repair attempts are associated with worse outcomes than expert repair potentially compromising later revisions of the injury by a specialist [130]. As supported by another single-center cohort study on 200 patients with BDIs, on-table repair by non-HPB specialists appears to be an independent risk factor for recurrent cholangitis, biliary strictures, revision surgery, and overall morbidity [126]. On the contrary, an early referral to an HPB center can significantly decrease the rate of postoperative complications (OR: 0.24; 95% CI: 0.09–0.68; p = 0.007) and biliary strictures (OR: 0.28; 95% CI: 0.17–0.47; p < 0.001) compared to delayed referral [130]. Whenever a sufficient HPB experience is locally available, some data suggest that the earlier the repair, the better the results [79, 102, 122, 126, 131], whereas other studies support that similar and good outcomes are to be expected for on-table / early (within 72 h [128]) repair vs. postoperative repair (within 1 week [131]) when the BDI is managed by HPB surgeons or in HPB referral centers [130].

However, it must be considered that in some countries or regions, a tertiary/specialist care center may be too distant, and the “traveling surgeon” practice may be inappropriate [20, 127]. In these specific cases, it is of utmost importance to assure an optimal local management before referral, especially when, due to logistic and geographical constraints, the time prior to transport may be prolonged [20].

Because the hepatic blood supply is mainly carried by the portal vein, the interruption of the right branch of the hepatic artery alone is usually well tolerated [117, 132, 133]. Whenever an injury is recognized, the immediate repair of the right hepatic artery is not the most frequent option even in tertiary care centers, being good results only occasionally reported (i.e., no occurrence of liver infarction and uneventful follow-up) [132, 133]. Indeed, opportunities for immediate arterial repair are limited due to the low rate of injury recognition, the low number of patients affected by symptomatic liver ischemia, and the high level of technical expertise required. Moreover, an extensive imaging workup with a contrast-enhanced CT scan is mandatory prior to attempting the vascular repair. Thus, given the low clinical impact and the technical complexity of the procedure, the efficacy of arterial reconstruction remains questionable [132, 134].

Vasculobiliary injuries, defined by the presence of both biliary (bile duct obstruction or hilar plate division) and vascular injuries (hepatic artery and/or portal vein injury), lead to liver ischemia in 10% of cases [132, 133]. Their management depends on the evidence and extent of the liver injury (e.g., ischemia, necrosis, or atrophy). Their stabilization may require few weeks or months. In general, the surgical management should be delayed to allow for an accurate imaging workup and strategic planning, which involves HPB surgeons.

A decision tree for the management of intraoperatively detected BDIs is displayed in Fig. 1.

In patients with previous biliary infections (e.g., cholecystitis, cholangitis) and patients with preoperative instrumentation such as endoscopic stenting at ERCP/sphincterotomy, endoscopic nasobiliary drainage (ENBD), or percutaneous transhepatic biliary drainage/cholangiography (PTBD/PTC), there may be preexisting bactobilia. Consequently, bile flow into the peritoneal cavity may lead to local or systemic sepsis. Thus, intraoperative antibiotic coverage must be initiated or continued in case an antibiotic prophylaxis has been already administered. Bile culture is mandatory to narrow the coverage spectrum and prevent antibiotic resistance. Treatment should last no more than 24h [136, 137]. The recommended antibiotics include cefazolin, cefamandole, or cefuroxime (to be substituted by gentamicin and clindamycin in case of allergy) [136, 137]. In case of infection and ongoing drainage, the following antibiotics can be considered: piperacillin/tazobactam, ceftriaxone, or other 4^th-generation cephalosporins [144], for a minimum of 5 days of treatment.

In the case of biliary obstruction without bile leak or signs of sepsis, antibiotic therapy may not be required. However, the majority of patients with biliary obstruction have infected bile and grow bacteria from cultures even when clinical cholangitis is not yet present. Sepsis may occur after biliary instrumentation and drainage using endoscopic stenting, ENBD, or PTBD. Antibiotic prophylaxis is appropriate and recommended to prevent the occurrence of healthcare-associated acute cholangitis, especially in the setting of predictable incomplete drainage [145].

The first priority in case of bile leakage is “source control” and early “goal-directed therapy” [140]. Antibiotic therapy should be initiated as soon as evidence of cholangitis or infected fluid collections appears [146]. In patients without shock, radiological and bacteriological sampling can be performed to obtain definitive diagnostic studies before starting parenteral antibiotic therapy. A 6-h delay period might be tolerated. In the presence of severe sepsis or shock, the investigation window should be substantially shortened, and broad-spectrum antibiotics should be started within 1 h of the initiation of signs and symptoms. Treatment should be adapted according to bile culture findings [136, 137]. In the worst cases of severe complicated intra-abdominal sepsis, open abdomen (OA) therapy for optimal source control may be considered [147, 148], although the biological basis for OA in such cases is currently being subjected to rigorous scientific scrutiny [149].

In the case of external biliary fistula without intraperitoneal collection, antimicrobial therapy might not be necessary if infectious signs are absent. The natural history of an external fistula depends on the anatomical subtype of injuries. In complex BDIs requiring delayed surgical repair, complete healing of the fistula is an absolute prerequisite for surgery. During the waiting period, several patients may experience cholangitis. The Tokyo guidelines published in 2018 for the severity grading and management of cholangitis may be applicable [144]. Biliary drainage, most often using PTBD, should be placed in cases of uncontrolled or recurrent cholangitis. Parenteral broad-spectrum antibiotics should be started and subsequently adapted to bile and blood cultures [150, 151]. Management of biloma and peritonitis requires percutaneous drainage and surgery, respectively. In the case of cholangiolytic abscesses, which are usually small and multiple, parenteral antibiotics and biliary drainage (endoscopic or percutaneous) may be indicated. A large cholangiolytic abscess not responding to parenteral antibiotics within 48–72 h may require imaging and US- or CT-guided percutaneous needle aspiration or catheter drainage. The antibiotics most often used in cases of biliary peritonitis are piperacillin/tazobactam, imipenem/cilastatin, meropenem, ertapenem, or aztreonam associated with amikacin in cases of associated shock and fluconazole in cases of fragility or delayed diagnosis.

The optimum duration of antibiotic therapy in the setting of biliary infection is a matter of debate. According to the Tokyo Guidelines [144], an additional 4 days of antibiotic therapy is required after source control of cholangitis by decompression of the biliary tree. Treatment should be continued for 2 weeks in the presence of Enterococcus or Streptococcus to prevent the risk of infectious endocarditis. Frailty and comorbid factors must also be accounted for in the titration of therapy. However, other studies showed that only 3 additional days are sufficient to reduce the risk of recurrence [152, 153]. For biloma and generalized peritonitis, a treatment of 5–7 days should be considered [140].

The most frequent complaints of patients with BDI are persistent abdominal pain, abdominal distension, nausea and/or vomiting, fever, and jaundice [18]. The BDI clinical presentations are related to the type of injury. The two most frequent clinical scenarios are bile leakage and bile duct obstruction [154]. In patients with a bile leak, an early visible sign is the presence of bile from the drain or surgical incision. If the subhepatic region is not drained, a perihepatic bile collection (biloma), abscess, or biliary peritonitis may develop with corresponding clinical signs. Generally, jaundice is not observed or is mild in these cases because cholestasis does not occur [18, 154‐156]. In patients with biliary strictures, symptoms are often delayed. Cholestatic jaundice with choluria, fecal acholia, and pruritus are the most common clinical signs and symptoms. If cholangitis develops, fever with chills is typically associated with jaundice [154‐157]. Recurrent cholangitis is the main consequence of bile duct stricture, hepatic injury and dysfunction from complete bile duct occlusion. Sepsis and multiorgan failure may develop in both clinical settings.

When BDI is not identified intraoperatively or during the first postoperative week, patients may have an insidious evolution with relapsing abdominal pain, cholangitis, and bile collections. A late diagnosis, which sometimes is made years after surgery following multiple ineffective attempted repairs or inappropriate management, may result in increased complexity of bile duct repair. Moreover, even if successfully managed, the patient’s quality of life and survival may be impaired [158]. Indeed, the clinical course of undiagnosed or unrepaired BDI can evolve to secondary biliary cirrhosis with portal hypertension, liver failure, and, ultimately, death [18].

After elective LC, laboratory tests are not routinely required because mild to moderate elevations in hepatocellular enzymes are frequently observed during the postoperative period but have no pathological meaning; CO₂ pneumoperitoneum seems to be the main reason for these changes [159, 160].

In clinical practice, surgeons should consider postoperative biochemical investigations whenever difficulties were encountered during the intervention or in the presence of postoperative clinical signs suggestive of complications. These are performed to aid in the diagnosis [154, 157, 161]. Ben-Ishay et al. [161] evaluated the utility of post-LC blood examinations by retrospectively analyzing the chart data of approximately 340 patients undergoing LC and confirmed that they may be useful to make diagnoses and lead to early interventions in complicated cases. Blood tests were most often obtained in elderly patients and those who had prolonged surgery, multiple drains, and longer hospital stays.

Serum levels of direct and indirect bilirubin, aspartate aminotransferase (AST), alanine aminotransferase (ALT), alkaline phosphatase (ALP), gamma-glutamyl transpeptidase (GGT), and albumin, as well as a complete blood count (CBC), are usually measured to diagnose iatrogenic BDI [154, 157]. In BDI patients, liver function tests and cholestatic enzymes may either be elevated, supporting the clinical suspicion, or remain within the normal ranges. In the case of stenosis or complete occlusion of the bile duct, bilirubin values increase, whereas no elevation or only a slight elevation may be observed as a result of peritoneal bile absorption in the presence of bile leakage [99]. In the very early stages, cholestasis markers are increased, but there is no significant hepatic damage; therefore, aminotransferases are not increased. Early in the initial postoperative course, the determination of ALP and total bilirubin is not sensitive [162].

Biomarkers, such as C-reactive protein (CRP), procalcitonin (PCT), and serum lactate, can help to evaluate the severity of the inflammation or sepsis and provide a baseline to follow the therapeutic response [163, 164]. PCT, CRP, and lactate levels can also be used to predict fatal progression in septic patients and are associated with poor outcomes and increased mortality [163, 164].

The role of imaging is to establish the BDI diagnosis, delineate the type and extent of the injury, and plan the appropriate intervention.

Ultrasonography (US) represents the primary noninvasive and easily available diagnostic tool that allows for the detection of intra-abdominal fluid collections, dilation of the biliary ducts, and possibly associated vascular lesions by using Doppler evaluation [154, 157, 165]. Abdominal triphasic CT scanning is useful to identify the possible presence of focal intra- or perihepatic fluid collections, ascites, biliary obstruction with upstream dilation, and long-term sequelae of a long-standing bile stricture, such as lobar hepatic atrophy or signs of secondary biliary cirrhosis. CT can also identify associated vascular lesions, such as injury to the right hepatic artery [154, 157]. The sensitivity of CT is superior to that of US, especially for the detection of small fluid collections and associated vascular complications [166‐168]. US provides good anatomic and contrast resolution, but CT has higher spatial resolution and better identification of fluid collection morphology and site; CT is also essential to define collections that require percutaneous or surgical drainage. However, neither US nor CT examinations can reliably distinguish bile leaks from other postoperative fluid collections, such as blood, pus, or serous fluid, because of their similar densities. Neither can establish the precise location or the active state of a bile leak because the bile collection site may not be separate from the leak site and occasionally may even be intrahepatic [169].

Hepatobiliary scintigraphy (HS) has two potential advantages over US and CT. It seems to be more sensitive and specific than US or CT in detecting bile leaks [170], and in addition to confirming the presence of a bile leak, it can identify the relationship between the leak and any fluid collection as well as show the primary route of bile flow [171]. Despite this, it is frequently necessary to complete HS with additional investigations. In fact, HS can provide functional information demonstrating the presence of an active leak, but its spatial resolution is poor, and the identification of the leak site can be challenging [169, 172]. Other pitfalls of HS are that extrabiliary structures are not visualized, so no information about them can be obtained, it has poor sensitivity in patients with hepatic dysfunction and large bile duct defects with preferential bile flow in a path of least resistance, and it may not show activity in the duodenum and thus a bile leak may be misinterpreted as a complete bile duct obstruction [169].

The use of ERCP and PTC can identify a continuing bile leak, provide exact anatomical diagnosis, and allow, at the same time, the treatment of the injury by appropriately decompressing or dilating the biliary tree. ERCP can be applied to treat bile leaks using internal stents. Success using this technique may be more likely if the injury to the duct is < 5 mm, if the injury is extrahepatic, and when there is no associated abscess or biloma [173]. In the case of ERCP failure, PTC is a valuable option to accurately depict the location and nature of BDI and to perform an extraluminal percutaneous endoscopic rendezvous procedure with stent placement to restore continuity of the bile duct [174‐176].

On the other hand, ERCP and PTC are invasive techniques that are associated with a nonnegligible risk of complications, including severe acute pancreatitis (mainly after ERCP), bleeding, and cholangitis (after PTC) [177, 178]. Other disadvantages are the lack of detection of extrabiliary abnormalities and the non-visualization of ducts upstream or downstream from an obstructing lesion (e.g., stricture, stone). Moreover, PTC can be technically difficult because intrahepatic bile ducts are usually not dilated [170].

Magnetic resonance cholangiopancreatography (MRCP) represents the “gold standard” for a complete morphological evaluation of the biliary tree, as it is noninvasive, does not use ionizing radiation, and provides excellent anatomical information regarding the biliary tree anatomy proximal and distal to the level of injury [154, 157, 169]. MRCP combined with dynamic contrast-enhanced magnetic resonance using a hepatocyte-selective contrast agent with biliary excretion allows for the functional assessment of the biliary tree, and thus, the detection and localization of bile leaks with an accuracy close to 100% [179]. In the past, the use of mangafodipir trisodium as a contrast agent primarily excreted via bile — now withdrawn from the EU Market — was shown to be useful for both diagnosing a bile leak and identifying the source of the leak by direct visualization of contrast material extravasation into fluid collections [179, 180].

Several authors [181‐185] confirmed that the additional use of contrast-enhanced MRCP (CE-MRCP) using 3D and 2D T1-weighted images acquired at the hepatobiliary phase after hepato-specific contrast agent injection improves the accuracy of bile anatomy depiction and bile leak detection. In a series of 99 patients — including 24 followed after cholecystectomy, 20 after surgical reconstruction of traumatic BDI, and 16 after hydatid cystectomy — the use of CE-MRCP increased the sensitivity, specificity, and accuracy, with respective ranges (depending on the bile leak etiology) of 76–82%, 100%, and 75–91% compared to 53–63%, 51–66%, and 55–63% observed with conventional MRCP [183]. The optimal timing for hepatobiliary phase acquisitions with CE-MRCP appears to range between 60 and 90 min when looking for bile leaks [182, 186].

In the post-liver transplant setting, Boraschi et al. [187] studied 384 MRCP examinations in 232 patients. The reported sensitivity, specificity, positive predictive value, and negative predictive value for the detection of BDI were 99%, 96%, 99%, and 97%, respectively. One considerable MRCP limitation is the poor opacification of bile ducts in the presence of obstruction and unreliable depiction of the more peripheral intrahepatic bile ducts [179].

Minor BDIs (e.g., Strasberg A-D [90, 92]) require a step-up approach once diagnosed. Common symptoms (e.g., abdominal pain or distension, fever, nausea), when noted in the postoperative period, may herald postoperative complications. In the presence of bile leakage from the drain, observation and non-operative management are advisable during the first hours [191]. If no drain was placed after surgery and imaging reveals a bile collection with suspicion of minor BDI (such as a cystic duct leak or duct of Luschka), percutaneous drainage of the collection may be the definitive treatment [126]. Several case series have reported the feasibility of drainage under endoscopic ultrasound guidance [192], but more data are needed before this approach may be recommended in this specific clinical situation.

If no improvements or worsening of symptoms occur, endoscopic management becomes mandatory [126, 193]. The same is true for low output biliary fistulas (i.e., a bile leak from the liver bed such as a Luschka’s duct) [193‐195]. Various endoscopic treatments (i.e. biliary stenting, endoscopic biliary sphincterotomy, and nasobiliary drainage) are highly effective to treat biliary leaks, except in the case of transection of the common bile duct or common hepatic duct. The time lapse between biliary injury and endoscopic treatment does not seem to significantly impact on the treatment outcomes [196].

ERCP is the key tool in BDIs management because it allows the identification of the site of bile leak and, most importantly, allows internal biliary drainage if the diagnosis of minor BDI is confirmed. Moreover, incidental diagnoses, such as choledocholithiasis or bile duct stricture, may also be treated in a single procedure. For this reason, ERCP is nowadays widely recommended as first-line therapy for postoperative biliary leaks [197].

The reported success rate of ERCP in this situation ranges between 87.1% and 100%, depending on the grade and the location of the leak [198‐204]. Bile leaks are divided into categories: (1) low grade, where the leak can only be identified after complete opacification of the intrahepatic biliary system; and (2) high grade, where the leak can be observed before intrahepatic opacification [203]. Leaks that respond more favorably to endoscopic treatment are those located at the end of a cystic duct stump or from a duct of Luschka, usually associated with low output [197].

The limits of the endoscopic diagnosis concern the lack of visualization of aberrant or sectioned bile ducts (i.e., an aberrant right hepatic biliary duct) and the difficulty in visualization of intra-hepatic proximal leaks. Endoscopic management should be preferred when there is at least partially documented continuity of the BDI (at the MRCP) or a very close proximity of the two biliary stumps (the proximal and the distal stumps); these are the conditions in which attempting endoscopic repair with the multistenting strategy [205].

The main goal of endoscopic therapy is to reduce the transpapillary pressure gradient to facilitate preferential bile flow through the papilla as opposed to the site of the leak, providing time to the biliary tree injury to heal. This is most commonly achieved by placing a transpapillary stent. Temporary naso-biliary drainage showed a similar efficacy when compared to plastic stents but has a lower patient compliance, so it should not be considered as the first choice [206]. There is little consensus on the role of sphincterotomy alone in the management of these patients [207, 208]. Avoiding sphincterotomy may minimize the risk for immediate (e.g., bleeding or perforation) and long-term complications (e.g., cholangitis or pancreatitis) [209]. The most frequent approach is the combination of biliary sphincterotomy with the placement of plastic stents or fully/partially covered metal stents, which is associated with a high success rate in low-grade biliary leaks [126, 199, 202‐204, 210, 211], and it is deemed even more effective in cases of high-grade leaks [199‐201, 203]. Although less investigated in the literature, long-term (at 10 years) outcomes of endoscopic treatment with stent placement appeared to be good and effective in patients with postoperative biliary strictures [212‐214].

Plastic stents are recommended to be placed to treat bile duct leaks [201]. For refractory bile leaks, fully covered self-expanding metal stents were demonstrated to be superior to multiple plastic stents in a non-randomized trial [215]. Stents are left in place for approximately 4 to 8 weeks in many studies and removed if retrograde cholangiography shows the resolution of the leakage.

The first-line approach to benign biliary strictures complicating cholecystectomy is endoscopic, as well. When recognized early in the post-operative period, strictures are often due to surgical trauma (e.g., energy device) and associated with bile leak. These strictures respond to endoscopic treatment more favorably than fibrotic strictures, which have a delayed diagnosis. Temporary placement of multiple plastic stents over a long period of time is the preferred treatment, with a success rate ranging from 74 to 90%, but with a recurrence rate as high as 30% within 2 years from stent removal [212, 216, 217]. In case of post-cholecystectomy bile strictures located > 2 cm from the main hepatic confluence, fully covered SEMS can be an alternative to plastic stents [201].

When ERCP is unsuccessful or not feasible, PTBD becomes an alternative. Moreover, PTBD can be useful for septic patients with a complete obstruction of the common bile duct as part of the multidisciplinary approach when ERCP fails or when surgical repair failures need to be treated (i.e., stricture of the hepaticojejunostomy). PTBD in the presence of bile leakage may be more difficult as a result of non-dilated bile ducts but still leads to a technical success of 90% and a short-term clinical success of 70–80% in expertise centers [214, 218, 219].

In the case of major BDIs (e.g., Strasberg E1–E5) in which there is a complete loss of common and/or hepatic bile duct continuity, carefully planned surgical treatment is required. Even when an endoscopic approach has been performed, high-grade bile leaks are difficult to manage successfully [191] and represent an independent risk factor for morbidity [199]. Early aggressive surgical repair (performed within 48 h from diagnosis) seems to guarantee good results, avoid the onset of sepsis, and provide advantages in terms of reduced costs and rate of hospital readmissions [195, 220, 221]. On the other hand, after 48–72 h, while inflammation tends to decrease, the phase of proliferation and healing begins and further complicates surgical repair. A key point is the technical contribution of the surgeon [126, 194, 222]: several studies have emphasized higher rates of postoperative failure, morbidity, and mortality when a primary surgeon without HPB expertise attempts to repair the injury [53, 189, 223‐225]. Accordingly, in the case of a lack of HPB experience, referral to a tertiary care center immediately after diagnosis is essential to ensure early surgical repair with Roux-en-Y hepaticojejunostomy [90], which showed superior outcomes at 5 years compared to late repairs [126]. An end-to-end anastomosis may be performed if the loss of continuity makes it technically possible, but this approach is associated with increased failure rates [195, 226]. Regardless of the technique used, tension-free bilioenteric anastomosis with good mucosal apposition and vascularized ducts is the mainstay of treatment [227]. Recently, robotic procedures have been suggested due to enhanced visualization, better tissue handling, and more precise surgery [228]. In the presence of increased tissue fragility, the expertise of the HPB surgeon in the tertiary care center is likely to improve the final results and consequently the long-term outcomes [194, 195, 229].

When BDIs are recognized during the early postoperative period (within 2 weeks), persistent tissue injury from ongoing inflammation and the absence of bile duct dilatation appear to negatively influence the results and often lead to late strictures [229]. When the diagnosis is not immediate or logistic constraints limit referral to a tertiary care center, the “drain now, fix later” algorithm [120] with percutaneous drainage of the biloma [193], targeted antibiotic therapy, and nutritional support seems to be the best approach. After clinical stabilization, delayed surgical treatment can be safely performed [20]. When a major leak results in progressive biliary peritonitis, urgent surgical intervention is required with laparoscopic lavage of the abdominal cavity and drain placement [229]. Once the patient has been successfully stabilized, several weeks (2–3 weeks) are usually required for the resolution of the acute inflammatory phase. Accordingly, this time will allow for lowering the risks associated with extensive reconstructive surgery by reducing inflammation and guaranteeing the assessment of the extent of ischemic injury resulting from associated vascular injuries (right hepatic artery lesions have been recognized in 25% of BDIs) [111, 126, 133, 222, 229, 230]. Similarly, a 1-week delay in the diagnosis of major BDIs suggests the need for a “timeout” of 2–3 months before intervention [105]. Combined BDI and hepatic artery injuries further increase the morbidity and mortality of BDIs and may necessitate an early referral to a specialized HPB center. Combined repair may avoid ischemic damage to the liver parenchyma and the risk of leakage or stricture of the bilioenteric anastomosis. When an injury is not promptly recognized, hepatic necrosis, atrophy, or even abscess within the ischemic parenchyma may occur, which would ultimately require liver resection with bilioenteric anastomosis [117, 231, 232]. Roux-en-Y bilioenteric anastomosis represents the gold standard treatment for major BDIs and is ideally performed during the immediate postoperative period (within 72 h). However, late repair should be considered after the resolution of acute or subacute situations and the closure of a biliary fistula on dilated bile ducts. Indeed, in these complicated clinical scenarios, mortality and morbidity rates after late repair are significantly lower than rates after immediate and early repair: 0.8% vs. 2.8% vs. 2.2% and 14.3% vs. 39.2% vs. 28.7%, respectively [111]. Furthermore, the need for a second procedure, that is, failure of the first repair, appeared to be higher after immediate and early repair than after late repair: 56.7% vs. 40.7% vs. 6.8%, respectively [111, 233]. However, the E-AHPBA multicenter study showed, after multivariate regression analyses, that the timing of biliary reconstruction with hepaticojejunostomy does not have any impact on severe postoperative complications, the need for reintervention, or liver-related mortality, leaving advisable the choice of an individualized treatment strategy after iatrogenic BDI [234].