The Role of Hepatobiliary Scintigraphy as the Initial Investigative Modality for Significant Bile Leak following Laparoscopic Cholecystectomy

The growing experience in laparoscopic surgery and its associated technology has led to a significant decrease in the rate of bile duct injury (BDI) as related to laparoscopic cholecystectomy (LC) of up to 0.22%, but still higher in comparison with the “classical” open surgery (0.06%) [1‒5]. Main reasons include loss of tactile sensations during laparoscopic surgery and inability to define the exact anatomical morphology in conjunction with inflammatory and fibrotic changes that interfere with proper anatomical discriminations. Consequently, rates of post-operative bile leaks are also elevated and at a greater percentage following the laparoscopic approach [4‒7]. BDI following LC is considered to be more severe and is defined by several well-known classifications, to help proper definition, treatment approach, and prognosis [6‒9]. The clinical presentation of BDI, based on anatomical classification (Strasberg and the integrated ATOM classifications for example) can be divided into two main groups, whether the dominant part relates to bile leak or to biliary stricture [8, 9]. Most bile leaks are discovered post-operatively and have diverse clinical manifestations, depending on whether a sub-hepatic drain was inserted at the conclusion of surgery [6, 7, 10‒13]. The investigative algorithm regarding post-LC bile leak includes invasive as well as non-invasive means, like abdominal ultrasonography, computed tomography, hepatobiliary scintigraphy (HBS, HIDA), endoscopic retrograde cholangio-pancreatography (ERCP), and magnetic resonance imaging (MRCP) [11, 12, 14‒19]. Usually, surgeons prefer to leave a drain in the gallbladder fossa whenever they face some technical difficulties (and suspect possible post-operative complications) due to diverse intraoperative causes. Consequently, most leaks are manifested as bile secretion in the drain, usually without peritoneal irritation or bile accumulation or elevated liver enzyme levels. Relating to the group that exhibits asymptomatic (controlled) post-operative leaks in the drains (our study group), we aimed to study the possibility of decreasing the rate of using invasive means like ERCP as the first investigative modality (without any intention to lessen its valuable advantages). Our intention was to start off with evaluation by HBS (most readily available test in our institution, while ERCP sometimes takes more time and specific preparations and while MRCP is not always available) in an attempt to avoid invasive means that might carry additional well-known adverse effects, such as bleeding, perforation, and acute pancreatitis. HBS is a sensitive radionuclide diagnostic study in detecting bile leaks by providing functional information regarding the presence of an active leak, in spite of its poor spatial resolution. It enables tracing the production and flow of bile from the formative phase in the liver and its passage through the biliary tree into the small intestine [9, 18, 20]. It should be noted that by no means, we did not intend to contradict the superiority of MRCP as an investigative tool, and we decided to reserve it to selected and uncertain cases (due to lack of its immediate availability and higher expenses related to its use). We hypothesized that invasive means like ERCP should be avoided whenever the biliary scan demonstrates clear passage of the radionuclide material from the liver through bile ducts into the small bowel, without significant leakage. In 2006, we began conducting a prospective non-randomized study of 11 years duration, implementing this policy, to be evaluated following its conclusion. The purpose of our study was to analyze our approach regarding post-LC bile leaks in the drains, and to study whether relying on biliary nuclear scan as the first investigative modality could avoid further invasive means in selected cases.

This study was conducted over a period of 11 years (June 2006–September 2017), consequential to the policy in our medical center mandating to start with HBS (HIDA) for each asymptomatic biliary leakage of at least 50 mL/24 h following LC that persisted in the drain (a bulb suction drain) beyond the first post-op day. We stopped enrolling patients for the study at the end of 2017, for the assessment of the beneficial consequences regarding our policy and to enable prolonged follow-up. We did not change our policy afterward.

Technical data regarding the use of HBS in our study: the amount of intravenous radiotracer used was 5 mCi of technetium (^99mTc). We used 4 views of static images (every 5 min) in combination with dynamic images. We used planar imaging, adding SPECT/computed tomography whenever bile leak was suspected, in order to validate its source and extent. Most of the leaks occurred early during the HBS, and by 90 min the vast majority of bile leaks could have been identified in the later images. As the patients who were included in our study had drain located postoperatively in the gallbladder fossa, the leaks were contained (not spreading in the peritoneal cavity). It should be mentioned that we never perform the injection of contrast medium from the drainage tube to detect the site of duct injury because our opinion is (based on past experience) that the injected medium would fill the post-operative sub-hepatic space, masking and hiding crucial anatomic and functional details, thus avoiding the timing of correct approach in diverse scenarios. Regarding the definitions of the normal and abnormal findings of hepatobiliary scintigraphy: in general, HBS evaluates the function of hepatocytes, patency, and integrity of the biliary ducts, gallbladder contractility, and function of the sphincter of Oddi. The issue of the visualization of the gallbladder (when considering inflammatory and other pathologic phenomena) is not relevant to our current study, dealing with post-cholecystectomy complications. Normal HBS (in studying the possibility of bile leak) involves the immediate demonstration of hepatic parenchyma followed sequentially tracer appearance in the intra- and extra-hepatic biliary ductal system, and the upper small bowel, to be seen within 1 h. In cases of bile leak (an abnormal scan), the tracer will be found in a location other than the liver, bile ducts, bowel, or urine [9, 18, 20].

Medical data were collected from the medical charts of all patients who underwent LC during those 11 years and included demographic, clinical, laboratory, and imaging data, as well as intraoperative information (type of surgery and intraoperative findings) and the histopathological reports of the resected gallbladder. Data regarding the nature and timing of the leak were collected. We screened the reports of all the HBS performed in the Institute of Nuclear Medicine, and the data regarding post-operative follow-up (till May 2022) that were stored in the outpatient clinic charts. We were assisted by the computerized medical data that were kept in the hospital archive. All the ERCP reports during this period were carefully screened to study those patients in whom ERCP was done following LC whenever it was suspected that BDI had occurred.

Regarding inclusion criteria: the study population included those patients (adults above 18 years old) in whom significant post-operative bile leak occurred (>50 mL/24 h) that had not been stopped spontaneously and who had been referred to HBS examination using a nuclear tracer (^99mTc-IDA, HIDA) as the first (and sometimes even the only) investigative modality, to assess proper passage of bile to the duodenum (excluding complete interruption of the biliary tree), and roughly estimate the magnitude of leakage. It should be noted that this modality is easily accessible upon request in our institution, and without delaying any additional examinations whenever necessary.

Regarding exclusion criteria: A. patients who died during hospitalization due to reasons unrelated to surgery and biliary disease; B. patients who underwent HBS for other indications, such as acute cholecystitis; C. patients in whom bile leak was suspected, due to abdominal peritoneal signs and elevated liver function tests and peritoneal irritation, fever, leukocytosis; D. patients having LC without leaving post-operative drainage (uncontrolled leak).

The study was approved by our medical center Institutional Ethical Committee (Helsinki, No. 0068 - 17). Statistical analysis was done using IBM SPSS Statistics, version 25.0. Quantitative data were shown by parameters of mean and median values, standard deviation, and range. Comparison of quantitative data between subgroups was done by using Mann-Whitney test. Qualitative data were presented by frequencies and percentage. Comparison of qualitative data between subgroups was done by using χ² test or Fisher’s exact test. We have calculated the values of sensitivity and specificity of the nuclear scan, as well as the positive and negative predictive values of bile scan against ERCP modality. p value equal or less than 0.05 was considered to be statistically significant.

During the period time of 11 years, 3,124 laparoscopic cholecystectomies were done in our department of surgery, of these, 67 patients were suspected to have post-operative BDI, according to bile leak that appeared in a drain placed under the gallbladder fossa at the conclusion of surgery. It should be mentioned that differential diagnosis of such leak is wide and not necessarily mandates interventions. Our study population includes only those for whom we began investigation by using HBS (^99mTc-IDA), when the amount of daily secreted bile was greater than 50 mL on the first and second post-operative days and did not resolve spontaneously. This does not mean that all those cases necessarily represented BDI. Our study group included 50 of the 67 patients, of whom, 27 (54%) patients underwent biliary scan as the only necessary investigative means, and 23 patients had an initial biliary scan followed by ERCP. For the remaining of the original 67 patients (17 patients), ERCP had been ordered (with no biliary scan al all) as the first investigative and curative means following asymptomatic bile leak in the drain, and with no details in the medical reports for the reason that biliary scan was not the initial modality. We presume the physicians who ordered ERCP were not aware of the study protocol. We therefore decided not to omit the important information that might be related to this subgroup, and to relate to those 17 patients as a control group, considering that this subgroup was not a part of a planned randomized study. Thus, all 67 subjects were divided into 3 subgroups: subgroup a: 27 patients (40.3%) with HBS only, subgroup b: 23 patients (34.3%) with HBS followed by ERCP, and subgroup c: a control group of 17 patients (25.4%) having ERCP without preceding HBS (Table 1).

All 3 subgroups were similar regarding age, ethnic origin (Arabs/Jews), gender (M/F), medical comorbidities, and the presence of cholecystitis (insignificant p value >0.05, χ² test or Fisher’s exact test and Mann-Whitney test, depending on qualitative or quantitative parameters).

Apart from our subjects, 3 cases of post-operative bile leak were discovered in patients not having a post-operative drain. In the first case, bile secretion was noted from the surgical wounds 4 weeks following surgery. ERCP was done (following imaging modalities and percutaneous drainage) to reveal bile leak originating from the stump of the cystic duct (type A, Strasberg classification), together with choledocholithiasis. Sphincterotomy and stent insertion were performed appropriately. In the second case, bile emerged from the surgical wound 3 weeks post-operatively and ERCP revealed bile leak from the gallbladder bed (type A). In the third case, reoperation was conducted for infected bile collections and bile leak that could not be found by ERCP, so only a precut was done.

Regarding our study group (50 patients), the mean age was 57 years (range 21–86), and the male to female ratio was 6:4. Thirty nine (77%) patients have not undergone abdominal surgery previously, and the remaining had undergone a range of operations. Four patients had previously undergone surgery in proximity to the liver and bile ducts. Elective LC was recorded in 26 (52%) patients, either following biliary colic or mild cholecystitis that responded well to conservative treatment including antibiotics and who were scheduled to elective surgery several weeks afterward. Emergency LC for acute cholecystitis was recorded in 24 patients. According to the intraoperative and histopathological reports, inflammatory changes were found in the gallbladder in 36 subjects, severe inflammation was found in nine, and absence of inflammatory changes in 5 patients (total 50 subjects).

Following bile scan in the study group (50 pt), bile leak ceased after a mean of 6.3 days (median 6 days, S.D. 4.1 days, and range 1–17 days). Time from operation to scan has a mean of 3.2 days (median 2.0 days, S.D. 2.7, and range 1–14 days).

Of the group of 27 patients who underwent HBS as the only modality, a fair passage of radioactive substrate to the duodenum was seen, and without any demonstration of bile leak in the majority (Fig. 1). In 2 cases, minimal leak was noted (suspected to originate from the common bile duct (Strasberg type D) in 1 patient and from the right hepatic duct (Strasberg type D, bismuth type V) in the other patient). Consequently, ERCP was not ordered during hospital stay, and bile leak ceased spontaneously within a few days (mean 3.6 days, median 3 days, and range 0–10 days, S.D. ±2.35). All the patients were discharged in a good medical condition, and all laboratory analyses were within normal limits. Outpatient clinic follow-up (up to 15 years, median 9.0 ± 2.6 years) was uneventful without the need for further invasive procedures. As our medical center is a regional hospital serving villages and urban settlements in its vicinity, we can be certain that most patients were not lost to follow-up.

In the second sub-group of our study that included 23 patients, HBS was followed by ERCP (mean time 4.6 days after bile scan, median 3 days, S.D. ±5.9). The decision to do ERCP was based on (1) the demonstration of significant leak suspected to originate from the extra-hepatic bile ducts as noted by the biliary scan (7 patients); (2) decreased passage of nuclear substance to the duodenum (4 pt); (3) significant and consistent bile secretion in the drain that could not be explained by the nuclear scan (7 pt), and (4) suspected stones and/or injury of the main bile duct during surgery for cholecystitis with severe inflammatory changes (5 pt). The ERCP findings showed absence of bile leak in 13 patients (normal ERCP) and existing leak in 10 patients (Fig. 1). Analysis of those 10 subjects revealed fair passage to the duodenum and minor leak in 4 cases (one leak originated from Rt. hepatic duct [Strasberg type D, bismuth type V]) and 3 from the cystic duct (type A). In the remaining 6 patients, significant bile leak was seen by the HBS, and ERCP revealed minor leak in 4 patients, and significant leak from the main duct in 2 patients (type D). Both of these needed sphincterotomy and insertion of an endoscopic stent, but in one, for whom no duodenal transit was noted on HBS, additional surgery was needed to repair significant BDI (type C2, Hannover classification) by hepatico-jejunostomy.

Regarding patients who underwent ERCP only (control group, 17 patients): visualization of the bile ducts failed in 3 patients, so percutaneous trans-hepatic cholangiography/drainage (PTC) was done in 1 patient to enable rendezvous through the ampulla of Vater. In the remaining 2 cases, there was no need to proceed with endoscopic manipulation as fair bile secretion was noted to pass into the duodenum. In one case, mild leak was noted from the main bile duct (type D). As precut and stent insertion was unsuccessful, MRCP was ordered to proceed management with conservative treatment. A leak originating from the ducts of the left hepatic lobe (type D) was successfully treated by sphincterotomy and main bile duct clearance in 1 patient. In nine subjects, the leak was seen to originate from the cystic duct (type A); in seven of these nine, it was successfully treated by an endo-stent without sphincterotomy, sphincterotomy with stent in one and sphincterotomy alone in one case. Leak from the gallbladder bed (type A) was inspected in three subjects and was treated by stent and sphincterotomy (1 patient), sphincterotomy alone (1 patient), and stent alone (1 patient).

Obviously, we were unable to determine the exact location and origin of the bile leak in the majority of cases included in our study, whenever HBS was the sole investigative modality to be used for spontaneously and uneventfully resolved leaks. The explanation is that HBS cannot provide an exact spatial anatomical location of biliary leaks. Basically, it was not our aim to locate the exact location of the leak as long as the leak stopped spontaneously and the patient uneventfully recovered, without the need for invasive investigative means.

In order to conclude which of our patients could avoid invasive procedures (like ERCP) while investigating significant post-operative bile leak through the drain, we should consider our three subgroups (Fig. 1): (1) twenty seven patients, for whom nuclear scan provided sufficient information and bile leak stopped spontaneously; (2) thirteen patients from the second sub-group of 23 patients for whom ERCP followed biliary scan, and in whom ERCP did not demonstrate any leak; (3) probably 7 patients belonging to the control group (ERCP as the first modality), who had mild leak or leak that should stop spontaneously, such as leak that originate from the gallbladder bed, and those who had fair bile passage to the duodenum. Altogether, we retrospectively speculate that in 47 of 67 patients (70%), invasive modalities to explore asymptomatic bile leak in the drain could have been avoided. Normal HBS demonstrated good passage of the nucleic tracer into the duodenum without evidence of bile leak, and eventually invasive investigative modalities were unnecessary. In 20/27 and 13/23 subjects among the first and second subgroups, respectively, bile scan was normal, namely, a total of 33 subjects (Fig. 1, 2). In 30 of whom, eventually no other diagnostic or management modalities were needed, and postoperative convalescence was uneventful. In other words, a positive correlation between normal bile scan (as noted in 33 patients) and the possibility to avoid invasive means (ERCP without pathology as noted in 10 patients, together with uneventful clinical convalescence in 20 patients) was defined as a negative predictive value and was valued 90.9%. According to Fisher’s exact test (two sided), it was considered significant (90.9%, vs. 58.8% in the case of abnormal scan, p = 0.021). Regarding the statistical analysis, we considered only the first and second subgroups. As such, evaluation of the subgroup that underwent ERCP following biliary scan (normal and abnormal scan) enabled the calculation of scan sensitivity, which relates to those patients having abnormal scan (7 patients) among 10 patients presenting abnormal ERCP, namely, 70%. Scan specificity was defined as patients who had fair biliary scan (10 patients), among those that demonstrated normal ERCP examination (13 patients), 76.9%.

Abdominal surgery using the laparoscopic approach is widely implemented today and for years has been considered the gold standard for LC. Although the cumulative surgical experience is vast and the post-operative complications are constantly decreasing, the rate of bile duct injuries following LC is still higher than in open-approach surgery [1–5, 21, 22]. Many studies were done regarding BDI during gallbladder surgery, including classifications that relate to the location and degree of BDI, offering investigative modalities and treatment options, considering the timing of recognition of injury as intraoperative, early, or delayed [6‒8, 21, 23]. BDI can be suspected almost immediate following a procedure in which the surgeon has left a drain near the gallbladder fossa (whenever he faces a difficult operation due to inflammatory and fibrotic changes), presenting as significant bile leak in the drain. Bile duct injuries can also be manifested by changes in liver function tests, pathological findings on physical examination, etc. Evidence to iatrogenic injury can also be suspected days following discharge from hospital, due to bile duct obstruction, intra-abdominal bile leak, and peritonitis [6, 7, 11, 13, 21, 23, 24].

In our study conducted prospectively over 11 years (associated with a prolonged follow-up), we planned to relate only to those asymptomatic subjects in whom a drain was left post-operatively in the gallbladder fossa, excreting bile of at least 50 mL per day (a controlled leak) that did not stop spontaneously by the second post-op day. We suspected that these criteria probably represented the majority of patients suffering from bile leak in the study. When bile scan is readily available, it does not delay the use of other valuable modalities when necessary and frequently tells the surgeon that the common bile duct is intact or not completely interrupted. According to our study, this may in some cases support the decision not to proceed to more invasive means. We are aware that in many medical centers the MRCP is readily available, and the higher cost is not of a concern (it should be stressed that we have MRCP in our medical center and use it whenever indicated). Hypothetically, we doubt that by doing MRCP as the first modality, we could have raised the number of unnecessary ERCP performances, whenever exact leakage could have been demonstrated by such definitive modality. We still suggest considering to start with HBS for such cases.

It should be noted that most bile leaks resolve spontaneously when adequately drained, and subclinical bile leaks occur frequently following elective LC, approaching 7%, and may easily be detected by HBS [18, 25]. Biliary scan is a well-known sensitive modality for diagnosis of bile leaks (through and before the laparoscopic era) by reflecting non-invasively the physiological route of bile stream from the liver to the small intestine, although complementary means are needed for exact anatomical resolution [16‒19]. In our study, it seemed that when considering all subgroups, we could have probably avoided ERCP in 47 cases out of the 67 subjects who suffered from bile leak and had post-operative drains as according to accumulating experience, mild leaks originating from the gallbladder fossa usually stop spontaneously without clinical consequences when adequately drained (it was supported during prolonged follow-up). However, for the sake of calculating the negative predictive value of biliary scan to avoid invasive modalities, we took into account only the first and second subgroups, and it was found to be 90.9% (Fig. 1, 2). Studies regarding the predictive value of biliary scan have already been done but were small and not prospective in nature and the study population was not identical [18, 26]. We were convinced that there is a place to use HBS which enables fair precision of relevant data and early diagnosis of biliary leak without the ultimate need for ERCP, leaving MRCP for selected cases [18‒27]. Another point of concern is that as ERCP and MRCP modalities are not always easily available, small leaks that are manifested on biliary scan may resolve spontaneously in the meantime, thus explaining why ERCP analysis was found to be normal while the scan showed mild bile leak.

Limitations regarding our manuscript include the relatively small number of participants in our study group, in spite of more than three thousand laparoscopic cholecystectomies being operated during a period of 11 years. In addition, it might be of interest whether using MRCP first would lead to the same reduction in invasive modalities like ERCP.

In conclusion, although biliary scan has been gradually replaced by advanced modalities like MRCP, we do suggest, following the results of our prospective study that in case of immediate availability, there is a place for such modality (hepatobiliary scan) as an initial investigative approach in cases of asymptomatic biliary leak in the drain following LC, decreasing the percentage of invasive modalities (like ERCP) as the first modality, whenever possible.

It should be noted that the conduction and progression of our approach and management were done prospectively. However, the decision to analyze and study the consequences and advantage of our policy (attempting to decrease the use of invasive means such as ERCP) was made several years later. The study was done retrospectively by the analysis of the relevant medical files. The study in its current form was approved by the Institutional Ethical Committee of the Galilee Medical Center (Helsinki – 0068 – 17 - NHR). Obviously, informed consent for surgery and any relevant procedure were obtained from each patient whenever necessary. Specifically, in detail, written informed consent was obtained from each participant for surgery and additional investigative means. As the clinical policy was prospectively conducted and the data for study analysis were collected retrospectively from the medical files, a written informed consent was not required for participation in the study as it was based on publicly available data, in accordance with our Local Ethical Committee and current GCP guidelines.

The authors declare that they have no competing interests.

The authors received no funding for this work.

Prof. Amitai Bickel: substantial contributions to conception and design, acquisition of data and analysis, and interpretation of data; drafting and revision of the manuscript for important intellectual content; and final manuscript approval for submission and publication. Ron Legrissi, M.D.: substantial contributions to conception and design and acquisition of data; revision of the manuscript; and final manuscript approval for submission and publication. Jacqueline Jerushalmi, M.D: substantial contributions to conception, design, and acquisition and interpretation of data; drafting of the manuscript for important intellectual content; and final manuscript approval for submission and publication. Savit Wisam, M.D.: substantial contribution to conception and design, analysis and interpretation of data; revision of the manuscript; and final manuscript approval. Michael Weiss, M.D.: substantial contribution to conception and design; revision of the manuscript critically for important intellectual content; and final manuscript approval for submission and publication. Moshe Shiller, M.D.: substantial contribution to conception and design; revision of the manuscript; and final manuscript approval for submission and publication. Samer Ganam M.D.: substantial contribution to conception and design and acquisition of data; drafting of the manuscript for important intellectual content; and final manuscript approval for submission and publication. Eli Kakiashvilli M.D: substantial contributions to conception, analysis, and interpretation of data; drafting and revision of the manuscript critically for important intellectual content; and final manuscript approval for submission and publication.

All data generated or analyzed during this study are included in this published article. In case someone would like to request the data, they should contact the corresponding author (see the Title page).