Introduction
Globally, the incidence of post-operative bile leakage after cholecystectomy ranges from 0.5% to 3.14%. It accounts for up to 47.8% of the post-operative biliary referrals in Indian tertiary surgical centers [1][2]. Anatomical misidentification, severe inflammatory distortion, and dense subhepatic fibrosis or unrecognized anatomical variants may compromise the biliary tree integrity and lead to clinical sequelae ranging from low-output cystic stump leaks to major common bile duct transections [3]. If persistent biliary extravasation is not treated, biloma, peritonitis, abdominal sepsis, secondary biliary strictures, and a significant prolongation of hospital stay are the consequences [3].
The implementation of Strasberg’s Critical View of Safety (CVS) and therapeutic endoscopic retrograde cholangiopancreatography (ERCP) have improved outcomes, but much of the published literature is based on univariate comparisons that do not adjust for baseline demographic confounders or structural event sparseness [1,2,4].s
To overcome these limitations, an institutional dataset of 317 cholecystectomy procedures was analyzed in this study [2]. Twelve confirmed post-operative bile leak cases were strictly matched to 48 non-leak controls in a 1:4 ratio for age and gender [2][3]. We examined the multivariable contributions of pre-operative liver function tests, chronicity of disease, intraoperative fibrosis of Calot’s triangle, duration of the operation, dissection technique (fundus-first or classical), and operator experience in predisposing to post-operative biliary disruption, removing baseline demographic confounding [2][4].
Methods and Materials
Study Design
It was a quantitative observational study, with a retrospective case-control design and 1:4 matching. The study was conducted at IMS and SUM Hospital, Bhubaneswar, Odisha, India, a tertiary care teaching hospital having separate general surgery and surgical gastroenterology units over a 2-year archival window.
Population and Sampling Procedure
Total population sampling (observed incidence rate: 3.78%) was used to identify all n=12 patients meeting verified post-operative bile leak criteria from a baseline institutional dataset of N=317 cholecystectomies (laparoscopic, converted to open, and primary open). Controls were selected by individual density sampling without replacement from the non-leak pool (N=305). Each case was strictly matched 1:4 to 4 control patients by gender and age (±2 years) using computer-generated block randomization, resulting in a final cohort of N=60 (12 cases, 48 controls).
Eligibility and case definition
Case Definition: Confirmed post-operative bile leakage according to International Study Group of Liver Surgery (ISGLS) criteria: fluid drained from the abdominal cavity or localized collection with bilirubin concentration at least 3 times higher than concurrent serum total bilirubin measured on or after post-operative day 3, or direct visualization of active leakage during ERCP, re-laparoscopy, or percutaneous drainage.
Adults (≥18 years) undergoing elective or emergency cholecystectomy within the 2-year window, with complete archival medical records including pre-operative LFTs, standardized operative notes, and 30-day post-operative follow-up notes.
Exclusion criteria: 1. Concomitant primary radical or extended hepato-biliary resections (radical cholecystectomy, hepatectomy, Whipple procedure, etc.); 2. Pre-existing uncorrected biliary strictures; 3. Active choledocholithiasis requiring open CBD exploration in the same setting; 4. Incomplete medical records.
Definition of Key Variables & Standardization of Surgery
-
Fibrosis of Dense Calot’s Triangle: Standardized post-hoc using documented operative notes cross-referenced with Parkland Grading Scale (Grade 4–5: severe inflammation/frozen Calot's) and Nassar Difficulty Scale (Grade 3–4: dense scarring obliterating anatomy)
-
Operator Experience & Supervision: Senior Consultant (>5 years post-residency experience) vs. Junior Trainee/ Resident. The institutional policy required direct scrubbed consultant supervision of each procedure performed by residents.
-
Dissection Technique: Classical (isolation of cystic duct/artery within Calot's triangle first) vs. Fundus-First (top-down retrograde mobilization of gallbladder off hepatic bed before hilar clearance)
Preoperative laboratory values were extracted from archived electronic health records. Pre-operative liver function tests (LFTs) were performed on automated clinical chemistry analyzers (Roche Cobas 6000 / c501 module) using standard enzymatic and spectrophotometric methodologies. Total bilirubin was assayed by the diazotized sulfanilic acid method (normal reference ≤1.2 mg/dL) and alkaline phosphatase (ALP) by the standard IFCC 37°C p-nitrophenyl phosphate kinetic cleavage assay (normal reference cutoff defined as ≤120 U/L).
Statistical analysis
Continuous data were tested for normality and expressed as mean ± standard deviation (SD) or median with interquartile range (IQR) and compared using Student's paired t-test or Mann-Whitney U test, respectively. Categorical variables were compared using Pearson’s Chi-square or Fisher’s exact test.
In sparse-event data (n=12 outcome events across 7 candidate variables), standard unconditional binary logistic regression can result in biased estimates. Therefore, statistical estimation was performed using Firth’s Penalized Conditional Logistic Regression. Firth’s penalization adds a Jeffreys prior penalty function to the score equation, correcting finite-sample bias, small-sample bias, and non-convergence of parameters (separation). Akaike Information Criterion (AIC) was minimized to optimize variable selection and prevent overfitting. Crude odds ratios (cOR), adjusted odds ratios (aOR), and 95% confidence intervals (CI) were calculated. All analyzes were done using IBM SPSS Statistics (Version 27.0) and R (Version 4.3.2, package logistf). Two-sided p < 0.05 was considered statistically significant.
Results
The study examined a 1:4 matched cohort of 60 subjects (12 cases and 48 controls). Age (52.4 ± 10.2 vs. 51.8 ± 9.8 years; p = 0.842) and gender (33.3% male, 66.7% female; p = 1.000) were identical between groups, indicating that all demographic confounding was completely neutralized. Bile leak cases had significantly higher rates of elevated total bilirubin >1.2 mg/dL (58.3% vs. 16.7%; p = 0.004), elevated alkaline phosphatase >120 U/L (66.7% vs. 20.8%; p = 0.002), and acute cholecystitis presentation (75.0% vs. 25.0%; p = 0.001) compared to controls (Table 1).
| Parameter / Variable | Subcategory | Cases (n=12) | Matched Controls (n=48) | p-value |
| Age (Years) | Mean ± SD | 52.4 ± 10.2 | 51.8 ± 9.8 | 0.842† |
| Gender | Male (n, %) | 4 (33.3%) | 16 (33.3%) | 1.000 |
| Female (n, %) | 8 (66.7%) | 32 (66.7%) | ||
| Pre-op Total Bilirubin | Normal (≤1.2 mg/dL) | 5 (41.7%) | 40 (83.3%) | 0.004* |
| Elevated (>1.2 mg/dL) | 7 (58.3%) | 8 (16.7%) | ||
| Pre-op ALP | Normal (≤120 U/L) | 4 (33.3%) | 38 (79.2%) | 0.002* |
| Elevated (>120 U/L) | 8 (66.7%) | 10 (20.8%) | ||
| Clinical Presentation | Chronic Cholecystitis | 3 (25.0%) | 36 (75.0%) | 0.001* |
| Acute Cholecystitis | 9 (75.0%) | 12 (25.0%) |
Dense fibrosis of Calot’s triangle was associated with a significantly risk of bile leak by univariate analysis (83.3% vs 14.6%, cOR = 29.28, 95% CI: 5.12–167.45, p < 0.001). Similar associations were found between bile leak and fundus-first dissection (66.7% vs 10.4%; cOR = 17.20; 95% CI: 3.45–85.72; p < 0.001), operative duration ≥90 minutes (91.7% vs 29.2%; cOR = 26.14; 95% CI: 3.11–219.78; p = 0.003), and supervised junior trainee primary operator status (75.0% vs 31.3%; cOR = 6.60; 95% CI: 1.53–28.51; p = 0.011) (Table 2).
| Intra-operative Variable | Subcategory | Cases (n=12) | Matched Controls (n=48) | cOR | 95% CI | p-value |
| Dense Calot’s Fibrosis | Absent / Soft Tissues | 2 (16.7%) | 41 (85.4%) | 1.00 | Reference | — |
| Present (Dense/Frozen) | 10 (83.3%) | 7 (14.6%) | 29.28 | [5.12–167.45] | <0.001 | |
| Dissection Technique | Classical (Cystic-First) | 4 (33.3%) | 43 (89.6%) | 1.00 | Reference | — |
| Fundus-First (Top-Down) | 8 (66.7%) | 5 (10.4%) | 17.20 | [3.45–85.72] | <0.001 | |
| Operative Duration | <90 Minutes | 1 (8.3%) | 34 (70.8%) | 1.00 | Reference | — |
| ≥90 Minutes | 11 (91.7%) | 14 (29.2%) | 26.14 | [3.11–219.78] | 0.003 | |
| Primary Operator | Senior Consultant (>5 yrs) | 3 (25.0%) | 33 (68.8%) | 1.00 | Reference | — |
| Junior Trainee / Resident | 9 (75.0%) | 15 (31.3%) | 6.60 | [1.53–28.51] | 0.011 |
Firth’s penalized multivariable logistic regression identified four independent predictors of post-operative bile leakage: intraoperative dense Calot’s triangle fibrosis (aOR = 14.82, 95% CI: 2.14–102.65; p = 0.006), total operative duration ≥90 minutes (aOR = 6.75, 95% CI: 1.18–38.64; p = 0.032), fundus-first dissection technique (aOR = 5.24, 95% CI: 1.09–25.18; p = 0.038), and elevated pre-operative alkaline phosphatase >120 U/L (aOR = 4.15, 95% CI: 1.02–16.89; p = 0.046). Elevated total bilirubin (aOR = 3.42, p = 0.091), acute cholecystitis presentation (aOR = 3.89, p = 0.062), and junior trainee operator status (aOR = 2.18, p = 0.312) lost statistical significance after multivariable adjustment (Table 3).
| Predictor Variable | Baseline Reference Category | Adjusted OR (aOR) | 95% CI | LR p-value |
| Elevated Total Bilirubin (>1.2 mg/dL) | Normal (≤1.2 mg/dL) | 3.42 | [0.82–14.28] | 0.091 |
| Elevated Alkaline Phosphatase (>120 U/L) | Normal (≤120 U/L) | 4.15 | [1.02–16.89] | 0.046 |
| Acute Cholecystitis Presentation | Chronic Cholecystitis | 3.89 | [0.94–16.12] | 0.062 |
| Dense Calot’s Triangle Fibrosis | Absent / Soft Adhesions | 14.82 | [2.14–102.65] | 0.006 |
| Operative Duration ≥90 Minutes | Operative Duration <90 Mins | 6.75 | [1.18–38.64] | 0.032 |
| Fundus-First Dissection Technique | Classical Dissection | 5.24 | [1.09–25.18] | 0.038 |
| Junior Trainee Operator | Senior Consultant Operator | 2.18 | [0.48–9.92] | 0.312 |
Anatomic origin of leakage documented during ERCP (66.7%) or re-laparotomy (16.7%) was cystic duct stump leaks in 58.3% (n=7), superficial liver bed / subvesical ducts of Luschka in 25.0% (n=3), and major lateral common bile duct injuries in 16.7% (n=2).
The ISGLS grading was Grade A (3 cases, 25.0%), Grade B (7 cases, 58.3%), and Grade C (2 cases, 16.7%). Median hospital stay was prolonged to 11.5 days (IQR: 9.0–14.2) in the bile leakage group vs. 2.0 days (IQR: 1.0–3.0) in controls (p < 0.001). Therapeutic ERCP with stenting (66.7%, n=8) was the most common secondary procedure for leak cases, followed by percutaneous biloma drainage (33.3%, n=4) and exploratory re-laparotomy (16.7%, n=2), with 0.0% 30-day mortality (Table 4).
| Parameter / Outcome Measure | Subgroup Category | Cases (n=12) | Matched Controls (n=48) | Statistical Test |
| Anatomic Site of Bile Leakage | Cystic Duct Stump Leak | 7 (58.3%) | N/A | Subgroup Frequency |
| Duct of Luschka / Subvesical Duct | 3 (25.0%) | N/A | Subgroup Frequency | |
| Major Common Bile Duct Injury | 2 (16.7%) | N/A | Subgroup Frequency | |
| ISGLS Bile Leak Severity Grade | Grade A (Minor management change) | 3 (25.0%) | N/A | Subgroup Frequency |
| Grade B (Non-surgical intervention) | 7 (58.3%) | N/A | Subgroup Frequency | |
| Grade C (Re-laparotomy / Repair) | 2 (16.7%) | N/A | Subgroup Frequency | |
| Median Hospital Stay (Days) (IQR) | Days | 11.5 (9.0–14.2) | 2.0 (1.0–3.0) | p < 0.001 (Mann-Whitney) |
| Secondary Interventions Required | ERCP with Biliary Stenting | 8 (66.7%) | 0 (0.0%) | Fisher's Exact (p < 0.001) |
| Percutaneous Biloma Drainage | 4 (33.3%) | 0 (0.0%) | Fisher's Exact (p = 0.001) | |
| Re-laparotomy & Biliary Repair | 2 (16.7%) | 0 (0.0%) | Fisher's Exact (p = 0.048) | |
| 30-Day Mortality | Mortality Rate | 0 (0.0%) | 0 (0.0%) | Exact Equivalence |
Discussion
Bile leakage continues to be a significant iatrogenic complication of cholecystectomy, resulting in patient morbidity and utilization of resources [5]. Isolating true biological and technical predictors requires controlling for baseline demographic bias [6]. In this study, a 1:4 matched design (N=60; 12 cases, 48 controls) neutralized age and gender confounding. Following multivariable adjustment using Firth’s penalized binary logistic regression, intraoperative dense Calot’s triangle fibrosis (aOR = 14.82, p = 0.006), total operative duration ≥90 minutes (aOR = 6.75, p = 0.032), fundus-first dissection (aOR = 5.24, p = 0.038), and elevated pre-operative alkaline phosphatase (aOR = 4.15, p = 0.046) were established as independent predictors.
A primary strength of this study lies in its methodological optimization. Exact matching prevented age and male sex—frequently noted univariate risk factors in registry data-from biasing parameter estimates [7][8]. Furthermore, Firth’s penalized likelihood estimation resolved small-sample issues and sparse-data bias (n=12 events), providing stable confidence intervals despite fitting 7 predictors [9].
Intraoperative technical challenges exert direct pathophysiological pressure on biliary integrity. Dense Calot’s triangle fibrosis obliterates tissue planes, hinders attainment of Strasberg’s Critical View of Safety (CVS), and masks anatomical variants or subvesical ducts of Luschka [5][11]. Elevated pre-operative alkaline phosphatase reflects periductal edema and intrabiliary hypertension, increasing mechanical shear stress on ligated cystic stumps [11]. Upstream clinical features, such as acute cholecystitis, showed high univariate risk, but were mediated downstream through severe scarring and operative complexity [6][12].
Regarding surgical technique, fundus-first (top-down) dissection carried a 5.24-fold higher multivariable risk of bile leakage (aOR = 5.24, p = 0.038). In surgical practice, fundus-first dissection is frequently deployed as a rescue maneuver when Calot's triangle is frozen [5][13]. Therefore, fundus-first mobilization acts primarily as an indication bias surrogate for extreme anatomical difficulty. Attempting top-down traction in densely fibrotic fields can cause 'tenting' of the common bile duct, leading to thermal injury or partial ductal transection unless combined with subtotal cholecystectomy [6][13].
Similarly, long operative time (≥90 minutes, aOR = 6.75, p = 0.032) is in keeping with large surgical series, including the UK CholeS study of 8,820 patients that demonstrated increased leak rates with prolonged procedures [7]. The length of time is due to tedious thermal dissection, capillary bleeding, and fatigue in difficult surgical fields.
According to a detailed anatomical classification, the most frequent causes of leaks (treated endoscopically or by drainage) were cystic duct stump insufficiency (58.3%) and extravasation from subvesical duct of Luschka (25.0%), whereas major common bile duct injuries (16.7%) necessitated re-laparotomy [1][10]. The median length of stay in the managed cases was 5.75 times longer (11.5 vs 2.0 days, p < 0.001) [6][7]. However, 30-day mortality was 0.0% for secondary therapeutic ERCP (66.7%), percutaneous drainage (33.3%), and re-laparotomy (16.7%), supporting modern clinical rescue pathways [7][10].
Junior trainee operator status was significant on univariate analysis (p = 0.011) but attenuated on multivariable adjustment (aOR = 2.18, p = 0.312) [12][14][15]. Trainee involvement in surgery, under direct consultant supervision, does not independently compromise surgical safety.
These findings emphasize a fundamental clinical dilemma: in complex cholecystectomies with severe scarring of Calot’s triangle, the risk of major bile duct injury is increased with attempts at complete gallbladder clearance using a fundus-first approach [9][13]. Safe intraoperative bailouts, such as laparoscopic subtotal cholecystectomy, are associated with a controlled, low-grade post-operative leak rate (14-18%, easily managed endoscopically) and near-zero major duct transections [9][13].
Limitations
Limitations of the study include unmeasured residual confounders (e.g., electrosurgical settings, drain volume variations) and a single-center retrospective design with 12 leak events. However, using a 1:4 exact-matching protocol and Firth’s penalized logistic regression controlled for demographic confounding and sparse-data bias, yielding stable parameter estimates.
Conclusions
Intraoperative anatomical hostility (dense Calot's fibrosis), long operation time (≥90 minutes), and fundus-first top-down dissection are independent risk factors for post-operative bile leakage. Mandatory intra-operative time-outs, early senior review, and low thresholds for subtotal cholecystectomy are practical strategies to prevent major biliary morbidity in regional health settings.
Declarations
Conflict of Interest Statement
The authors declare no financial, commercial, or personal conflicts of interest regarding the research, authorship, or publication of this study.
Ethical Approval
This study was conducted in accordance with the Declaration of Helsinki. Data extraction from archival records was fully anonymized and de-identified.
Data availability
Data Available on corresponding author.
Authors’ contributions
All authors’ equal contribution
Acknowledgements
None
Funding
None