The prevalence of gallstone disease is one of the most common gastrointestinal pathologies, causing a considerable burden on both clinical and economic resources of the health system. According to the results of a recently conducted systematic review and meta-analysis of data from 115 studies and over 32 million patients, the current prevalence rate of gallstones globally is approximately 6.1%, differing significantly by sex, geographical location and economic background of the population; being higher in females compared to males and in upper-middle-income countries (1). Similar estimates for the years from 2000 to 2024 show that one person out of twenty globally suffers from gallbladder stones (2). Symptomatic and complicated forms of gallstone disease, such as biliary colic, acute cholecystitis, choledocholithiasis and gallstone pancreatitis often require surgical intervention, and there are hundreds of thousands of cholecystectomies performed annually in high-, middle- and low-income countries, which implies billions of dollars' worth of expenditures each year (3).

Laparoscopic cholecystectomy (LC) is currently considered as the gold standard in surgical treatment of benign gallbladder disease due to its benefits, such as a lower level of postoperative pain, shorter hospital stay, faster recovery, and improved cosmetic outcome(4). Despite all above-mentioned benefits and wide use of minimally invasive surgical technique, LC is still not free of risks. The current rates of postoperative complications vary from about 2.5% to 12%, and although relatively rare, there are some serious complications that may occur after surgery, such as bile duct injury, major bleeding, bowel injury and bile leakage(4, 5). Conversion from LC to open cholecystectomy in up to 1% to 13% of procedures is an important indicator of surgical difficulty and also carries a higher risk of postoperative complications and delayed recovery(6, 7).

Various pre-, intra-operative, and patient-related risk factors have been suggested as predictors of adverse perioperative outcomes after LC, such as advanced age, male gender, obesity, diabetes mellitus, ASA classification, acute or chronic presentation, thickened gallbladder wall, adhesions in Calot's triangle, anatomic variability of the biliary tree, timing of surgery in relation to the onset of symptoms, as well as experience of the surgeon and center (8). However, a large body of available literature is based on single-center retrospective or prospective studies with heterogeneous case-mix, different definitions of complications, and poor generalizability. There is currently lack of large multicenter data assessing perioperative risk factors, complications, and predictors of early functional recovery, especially in South Asian healthcare systems, which show significant differences from the ones in developed countries regarding presentation of disease, referrals, and available resources (9-11).

The knowledge about which risk factors play an important role in prediction of complications, conversion, and delayed recovery has important practical value, because it would allow for preoperative risk stratification and counseling, help to make decision about timing of surgery and necessity of senior surgeon participation, facilitate ERAS pathway design, as well as optimize resource use (12-15). The aim of the present multicenter study was thus the systematic evaluation of perioperative risk factors for postoperative complications and independent predictors of early recovery following elective and emergency laparoscopic cholecystectomy.

2. Literature Review

2.1 Epidemiology and Global Burden of Gallstone Disease

Gallstones have been found to exhibit considerable geographic variation, with a higher incidence reported among Western and South Americans than among Asians and Africans, as well as having a preference for females, older patients, and those who have obesity or metabolic syndrome(16). Gallstone disease has been identified as a developing problem around the world in relation to increasing obesity, dietary changes, and aging population by the World Gastroenterology Organisation, with hundreds of thousands of gallbladder surgeries conducted per year in the US, UK, and other high-income countries(17). In low- and middle-income countries, including Pakistan, where laparoscopy is now becoming available, the absolute number of LCs being done is likely to increase(18).

2.2 Risk Factors for Postoperative Complications

There have been several recently published cohort studies on predictors of LC complications. One Chinese cohort included 478 patients where, after multivariate analysis, independent risk factors for postoperative complications were found to be adhesion to Calot's triangle (OR 3.04, 95% CI 1.42-6.51), anatomical variation of the biliary tree (OR 4.37, 95% CI 1.76-10.81), and thickening of the gallbladder wall of more than 5 mm. The postoperative complications occurred in 7.5% of the patients (19). Also, one Korean multicenter prospective cohort study included 2,514 patients from 18 academic centers and developed a web-based surgical risk calculator including 56 preoperative and intraoperative variables, of which sex, age, smoking, emergency operation, hypertension, diabetes mellitus, chronic obstructive pulmonary disease, preoperative endoscopic bile duct stone removal and ASA classification were all associated with postoperative complications, and the obtained prediction models had an area under the curve from 0.70 to 0.86 depending on the category of complication Similar results were obtained in the Japanese study with 423 patients who underwent LC in acute cholecystitis, and a Physiological and Operative Severity Score for the enumeration of Mortality and morbidity (POSSUM) of 48.3 or higher, and moderate to severe grade of cholecystitis, independently predicted the postoperative complications. Thus, composite physiological scoring might be helpful in addition to individual risk factors (20).

The role of obesity and diabetes in the unfavourable LC outcomes is discussed repeatedly. Multicenter audit revealed that the postoperative complications were significantly correlated with obesity (OR 3.5, 95% CI 1.5-8.2), emergency presentation (OR 2.1, 95% CI 1.2-3.6), and diabetes (OR 2.3, 95% CI 1.1-4.8), and the conversion rate to open surgery was independently correlated with acute cholecystitis, previous abdominal surgery, obesity and emergency operation(21). Another study among the obese patients undergoing emergency LC found greater technical difficulty, higher conversion rates and higher rates of surgical site infection as compared with non-obese patients (22, 23). In the elderly patients, Romanian multicenter analysis revealed that the degree of systemic inflammatory response was the major determinant of the poor outcome, diabetes was specifically associated with surgical and systemic morbidity and cardiovascular complications in the presence of cerebrovascular or renal comorbidity, but laparoscopic surgery proved superior to the open approach in all age groups(24).

2.3 Predictors of Difficult Laparoscopic Cholecystectomy and Conversion to Open Surgery

Conversion to open cholecystectomy is a well-documented surrogate marker of operative difficulty and is consistently associated with adverse perioperative outcome. A study conducted in Eastern India among 93 patients noted a conversion rate of 10.75% where acute cholecystitis, history of multiple attacks of biliary colic, and ultrasonographic features of a contracted gallbladder have been found to be important predictor(25). A single center study from Pakistan, Peshawar noted a lower conversion rate of 3.7% among 349 patients undergoing elective laparoscopic cholecystectomy but equally validated the predictive potential of various clinical, laboratory and radiological variables evaluated pre-operatively(16). A more recent prospective study from India suggested the use of a simpler preoperative and intra-operative scoring system where the appearance of gall bladder wall, adhesions in Calot's triangle, presence of common bile duct stones, formation of cholecystenterostomy fistula and achievement of critical view of safety were included to categorize risk of conversion, with 10.4% of 318 patients converting to open technique (26). A cohort study in Finland among 373 patients with acute calculous cholecystitis found elevated serum level of C-reactive protein above 150 mg/L, age greater than 65 years, diabetes mellitus, gall bladder gangrene and pericholecystic abscess as independent risk factors for conversion while conversion was a predictor of postoperative complications(27). Another study from Egypt tried to develop a pre-operative prediction model for patients who would undergo technically difficult early LC for acute calculous cholecystitis (28). Similarly an earlier clinic-radiological scoring system developed in India for predicting difficult procedures based on ultrasonographic parameters like gall bladder wall thickness and stone characteristics showed good predictability(29). Taken together these studies suggest that a relatively consistent number of preoperative variables consisting of inflammatory markers, imaging findings, burden of comorbidities and disease duration can be conveniently gathered pre-operatively to predict difficulty in the procedure.

2.4 Timing of Surgery in Acute Cholecystitis

The ideal time of conducting cholecystectomy among patients with acute cholecystitis is well documented. In the landmark multicenter ACDC study, patients were randomized to undergo immediate laparoscopic cholecystectomy within 24 hours of admission or to first be treated with antibiotics followed by delayed cholecystectomy between 7 to 45 days and showed significantly reduced morbidity in the immediate surgery group (11.8% versus 34.4%) without an increased conversion rate and mortality; thus, making early laparoscopic cholecystectomy the preferred modality in operable patients (24). In a systematic review and meta-analysis of randomized controlled trials which compared early and delayed laparoscopic cholecystectomy in patients with acute cholecystitis, early cholecystectomy, usually done within 72 to 96 hours after the appearance of symptoms, showed reduced total length of hospital stay without a higher rate of bile duct injury, bile leakage, wound infection or conversion rate than delayed surgery after conservative treatment (16). More recent regional evidence from Pakistan, however, has shown that early laparoscopic cholecystectomy reduces hospital stay although with significantly increased operating times and intraoperative bleeding compared to delayed surgery(6). Another randomized trial has similarly reported longer stay and higher costs for delayed surgery [also 20]. Preoperative waiting time has been independently shown to influence the length of stay in an acute care surgical setting(16).

2.5 Enhanced Recovery After Surgery and Predictors of Early Recovery

Intra-operative and post-operative evidence-based care practices bundled together under ERAS, including multimodal analgesia, early oral feeding, early mobilization and judicious use of surgical drains, have been increasingly adopted in laparoscopic cholecystectomy procedures aiming to facilitate functional recovery and shortened hospital stay(24, 28). Retrospective analysis of elderly patients with acute cholecystitis reported an ERAS pathway associated with earlier return of bowel function, reduced postoperative pain score, and shortened hospital stay, without increased incidence of complications when compared to standard perioperative care [14]. Similar benefits of implementation of an ERAS protocol in terms of reduced utilization of postoperative surgical drains, shorter time to resume oral intake, and faster hospital discharge have been shown by randomized and quasi-experimental studies conducted in Pakistan at a tertiary level(23, 29). Application of an ERAS pathway in an ambulatory surgical setting in Mexico resulted in an average postoperative stay of just 7.4 hours with a 96% rate of same-day discharge and without increased morbidity or reoperation rate(16). The importance of early structured mobilization, one of the central components of ERAS, has been proven in its positive correlation with the restoration of intestinal motility and pain reduction in elective LC [26]. Multimodal analgesia techniques also play an important role in early recovery: recently published retrospective cohort study has found a single-shot ultrasound-guided transversus abdominis plane block reducing opioid consumption and facilitating early mobilization compared to standard intravenous patient-controlled analgesia (28). Despite the above-mentioned evidence, multicenter data on the influence of preoperative risk factors, intra-operative findings, and perioperative care protocols on the results of early recovery have not been extensively studied yet, especially in resource-poor countries such as Pakistan.

2.6 Classification of Surgical Complications and Outcome Reporting

It is crucial for any research work to be reported in an unambiguous manner that facilitates comparisons between different studies and centers. The Claven-Dindo Classification system classifies complications based on the level of therapeutic intervention needed to manage them and is now the gold standard of reporting surgery outcomes. This system will be used in the current study to facilitate international comparisons (30).

2.7 Summary of the Literature and Identified Gap

The current literature suggests that laparoscopic cholecystectomy is safe overall, but a specific subgroup of patients does develop complications, needs conversion to open surgery, or experience delayed functional recovery. The known risk factors include demographic factors (age, sex, obesity), clinical factors (diabetes, ASA classification, acute vs chronic presentation, duration of symptoms), radiological factors (thickening of the gallbladder wall, pericholecystic collection), biochemical factors, and perioperative management factors (time of operation, ERAS pathways, pain management). Nevertheless, most of the research conducted so far has been conducted in a single center, retrospectively, using a small number of patients. Multicenter prospective data that take into account perioperative risk factors, a complication taxonomy, and objectively-defined outcomes of early recovery are rare, especially in a South Asian setting(31).

3. Objectives

3.1 General Objective

To assess perioperative risk factors, postoperative complication rates and their pattern, as well as independent determinants for early recovery in laparoscopic cholecystectomy patients at various tertiary care institution.

3.2 Specific Objectives

• To assess the frequency and profile of intra- and post-operative complications (intra-operative bile duct injury, haemorrhage, bile leakage, surgical wound infection, and conversion to open cholecystectomy) after laparoscopic cholecystectomy using the Claven-Dindo classification system.
• To assess the independent predictors of post-operative complications and conversion to open surgery from demographic, clinical, biochemical, and radiological parameters at the time of presentation preoperatively.
• To assess the independent intra-operative predictors of complications from operating time, intra-operative blood loss, difficult dissection of Calot’s triangle, and achievement of critical view of safety.
• To assess the predictors of early recovery from time to mobilize, oral intake, and length of hospital stay after surgery, as well as 30-day readmission rate.
• To formulate and validate an internal risk prediction model for postoperative complications and delayed recovery.

4. Research Hypothesis

Research Hypothesis (H1): Certain preoperative (age, obesity, diabetes mellitus, ASA status, acute cholecystitis, raised inflammatory markers) and intraoperative (longer operation time, difficulty in Calot’s triangle dissection, inability to obtain the critical view of safety) parameters are individually and significantly correlated with the occurrence of postoperative complications, conversion to open cholecystectomy, and delayed early recovery after laparoscopic cholecystectomy.

Null Hypothesis (H0): There is no significant correlation between the tested perioperative parameters and postoperative complications, conversion to open cholecystectomy, or delayed early recovery after laparoscopic cholecystectomy.

5.1 Study Design A prospective, multicenter, observational cohort study. 5.2 Study Setting The study would be performed at different departments of General Surgery of various tertiary care teaching hospitals in Pakistan (to be proposed and finalized at the time of protocol development), organized through the Health Services Academy, Islamabad, to ensure appropriate representation and generalizability of results. 5.3 Study Duration The proposed study will take between 12 to 18 months; this will include 3 months of protocol development and approval, 9 to 12 months of data collection through patients’ recruitment, and 2 to 3 months of data analysis and writing of the paper. 5.4 Study Population All consecutive adults undergoing elective or urgent laparoscopic cholecystectomy for benign gallbladder pathology (cholelithiasis, chronic cholecystitis, acute cholecystitis, biliary colic) in the study centers during the study period. 5.5 Sample Size Estimation The sample size will be calculated using the following formula to determine population proportion with a required precision: n = Z² × p(1-p) / d², where Z refers to the standard normal deviate for 95% confidence interval (1.96), p refers to the expected proportion of postoperative complications following laparoscopic cholecystectomy surgery (10%, based on pooled estimates of similar multicenter series(31-33), while d is the required precision (0.03). Thus, a minimum estimated sample size of about 384 participants will be obtained. Anticipating an incomplete data or follow-up rate of 10-15% and considering the requirement of 10 events per predictor variable for multivariable logistic regression analysis, a target sample size of about 500-600 participants will be recruited. 5.6 Sampling Technique Sampling of all eligible patients who attend the participating centres in the study period. 5.7 Inclusion Criteria ● Patients who are 18 years old and above. ● Patients who undergo either elective or urgent laparoscopic surgery for gallbladder disease, chronic cholecystitis or acute cholecystitis. ● Patients who sign written consent forms. 5.8 Exclusion Criteria ● Patients being planned for open cholecystectomy right from the beginning. ● Patients having gallbladder cancer. ● Patients having major abdominal surgery at the same time. ● Pregnant women (evaluated separately because of different physiology and aesthetic reasons). ● Patients not giving consent or getting lost in follow up before reaching the primary early recovery endpoint. 5.9 Data Collection Procedure After obtaining ethics committee approval, well-trained research associates from the various participating centers will identify and enrol patients meeting eligibility criteria. Preoperative data will be obtained using structured proformas, including demographic profile, comorbid diseases, symptom duration, ASA score, preoperative biochemical and haematological indices (leukocyte count, C-reactive protein, liver function tests, and bilirubin level), and preoperative ultrasound characteristics (wall thickness of gall bladder, pericholecystic fluid, size of stones and degree of impacted, and Murphy's sign). Intraoperative details will be recorded by the operating surgeon or an associated assistant and will include operating time, intraoperative blood loss, difficulty in dissection of Calot's triangle, achievement of critical view of safety, necessity for subtotal cholecystectomy, drain placement, and conversion into open operation due to documented indication. Postoperative complications will be graded according to Claven-Dindo system(34); other data will include time of first mobilization, return of oral intake, length of postoperative hospital stay, and any readmission or reoperation within 30 days. All these details will be captured in a uniform case report form and entered password-protected electronic database with study number. 5.10 Operational Definitions ● Postoperative complication: any abnormality in the postoperative course, which occurs within 30 days after surgery, and is categorized I to V according to the Claven-Dindo classification system (28). ● Early recovery: multi-endpoint that entails postoperative mobilization within 24 hours after surgery, postoperative dietary intake within 24 hours after surgery, and duration of hospital stay after surgery not exceeding 48 hours; all according to the ERAS criteria that are commonly used(16, 34). ● Conversion: any laparoscopic cholecystectomy that necessitates a long incision or a change into an open operation to accomplish the procedure. ● Difficult cholecystectomy: time of surgery more than 90 minutes, and/or failure in achieving critical view of safety within a reasonable dissection period, and/or need for subtotal cholecystectomy. 5.11 Study Variables The independent variables that will be considered are age, sex, body mass index, presence of comorbid conditions (diabetes mellitus, hypertension, chronic obstructive pulmonary disease), ASA classification, acute or chronic onset, symptom duration, pre-operative inflammatory status, sonography findings for gall bladder, surgery timing, surgical duration, intra-operative blood loss, and implementation of perioperative ERAS approach. The outcome variables will be postoperative complication occurrence and grades, conversion rate, and the outcome of early recovery endpoint. 6. Statistical Analysis Plan All analyses will be done using IBM SPSS Statistics (latest available version) and/or R software and statistical significance defined at a two-tailed p-value of <0.05. ● Descriptive statistics: The continuous variables will be presented as mean ± SD or median (IQR) based on their distribution, checked using the Shapiro-Wilk test while categorical variables will be presented as frequency and percentages. ● Bivariate analysis: Association between potential risk factors and postoperative complications, conversion, and early recovery will be performed using the independent samples t-test or Mann-Whitney U test for continuous variables and the Chi-square or Fisher's exact test for categorical variables where applicable. ● Multivariable analysis: Variables which are significantly associated (p<0.20) in bivariate analysis will be included in the multivariable binary logistic regression models for (a) predicting postoperative complications, (b) predicting conversion to open surgery, and (c) predicting delayed early recovery, with results reported as adjusted odds ratios (amor) with 95% confidence interval. ● Model performance: Discrimination of the predictive model will be evaluated using the area under the receiver operating characteristic curve (AUC), while the calibration will be checked using the Hosmer-Lemeshow goodness-of-fit test. ● Subgroup and center-level analysis: Mixed effects analysis or generalized estimating equations might be considered for handling patient-level clustering by center. ● Missing data: Patterns of missingness will be assessed, and multiple imputations will be considered for variables with missing data less than 10%; cases with missing primary outcome data will be excluded from analysis. 7. Ethical Considerations ● Approval shall be sought from the Institutional Review Board/Ethics Review Committee of the Health Services Academy and that of each participating hospital prior to commencing the recruitment process. ● The study will conform to the ethical standards set out in the Declaration of Helsinki. ● Informed consent will be obtained from all the participants through proper explanation of the aims, methods, possible hazards and benefits, as well as their rights to leave the study at any time without incurring any consequences on their medical treatment. ● The identity of the patients will always remain confidential; all the data collected will be anonymized and stored in password-secured devices which only the research team has access to. ● No other procedures or deviating from normal practice shall be done on the patient merely to fulfill the requirements of this research; the study is strictly observational in design. ● Incidental findings and adverse effects which may come up during data collection shall be handled following normal clinical procedure and reported to the relevant ethics committee. ● There is no known conflict of interest; the research will be performed independent of any commercial funding. Variable Shapiro–Wilk p Distribution Descriptive result Age (years) 0.0488 Non-normal 49.00 (38.00–57.00) Height (cm) 0.167 Normal 163.10 ± 7.81 Weight (kg) 0.0042 Non-normal 71.45 (63.08–81.83) BMI (kg/m²) 0.0623 Normal 27.15 ± 4.64 Symptom duration (days) <0.001 Non-normal 63.25 (6.50–122.47) Previous biliary attacks <0.001 Non-normal 2.00 (1.00–3.00) Preoperative waiting time (hours) <0.001 Non-normal 17.85 (9.70–26.62) Total leukocyte count <0.001 Non-normal 8.65 (6.90–10.90) CRP (mg/L) <0.001 Non-normal 10.35 (3.00–20.07) Haemoglobin (g/dL) 0.1484 Normal 12.71 ± 1.61 Platelets 0.9256 Normal 272.30 ± 68.47 Total bilirubin (mg/dL) <0.001 Non-normal 1.01 (0.73–1.36) Direct bilirubin (mg/dL) <0.001 Non-normal 0.41 (0.31–0.55) AST (U/L) <0.001 Non-normal 33.55 (24.65–45.82) ALT (U/L) <0.001 Non-normal 39.20 (24.18–55.23) ALP (U/L) 0.001 Non-normal 129.35 (90.80–160.50) Gallbladder wall thickness (mm) <0.001 Non-normal 3.50 (2.90–4.20) Largest stone (mm) <0.001 Non-normal 7.15 (5.30–9.10) Operative time (minutes) <0.001 Non-normal 71.00 (58.00–86.00) Estimated blood loss (mL) <0.001 Non-normal 44.00 (26.75–61.00) Pain score (0–10) 0.199 Normal 4.35 ± 1.45 Time to oral diet (hours) <0.001 Non-normal 12.80 (8.47–17.10) Time to first mobilization (hours) <0.001 Non-normal 10.15 (5.40–14.33) Postoperative hospital stays (hours) <0.001 Non-normal 37.80 (23.95–47.12) Ultrasound and operative characteristics Variable Category n (%) Wall thickness >5 mm No 246 (82.0%) Yes 54 (18.0%) Stone impaction No 260 (86.7%) Yes 40 (13.3%) Contracted gallbladder No 246 (82.0%) Yes 54 (18.0%) Pericholecystic fluid/collection Absent 261 (87.0%) Present 39 (13.0%) Calot's triangle difficulty Easy 182 (60.7%) Moderate 89 (29.7%) Difficult 29 (9.7%) Calot's triangle adhesions Absent 236 (78.7%) Present 64 (21.3%) Critical View of Safety achieved Yes 285 (95.0%) No 15 (5.0%) Biliary anatomical variation No 274 (91.3%) Yes 26 (8.7%) CBD stones No 285 (95.0%) Yes 15 (5.0%) Subtotal cholecystectomy No 294 (98.0%) Yes 6 (2.0%) Conversion to open surgery No 284 (94.7%) Yes 16 (5.3%) Perioperative and postoperative variables Variable Category n (%) Analgesic strategy Multimodal analgesia 144 (48.0%) Conventional analgesia 117 (39.0%) TAP block 39 (13.0%) ERAS pathway Yes 183 (61.0%) No 117 (39.0%) Postoperative complication within 30 days No 288 (96.0%) Yes 12 (4.0%) Clavien-Dindo grade* Grade I 4 (1.3%) Grade II 4 (1.3%) Grade III 2 (0.7%) Grade IV 2 (0.7%) Mobilized within 24 hours Yes 294 (98.0%) No 6 (2.0%) Hospital stay ≤48 hours Yes 232 (77.3%) No 68 (22.7%) Composite early recovery Yes 224 (74.7%) No 76 (25.3%) Readmission within 30 days No 287 (95.7%) Yes 13 (4.3%) Reoperation within 30 days No 298 (99.3%) Yes 2 (0.7%) Out of 300 patients who underwent laparoscopic cholecystectomy, the median age was 49.0 years (IQR 38.0-57.0), while the mean height and BMI were 163.10 ± 7.81cm and 27.15 ± 4.64 kg/m2, respectively. Female patients comprised 57.0% of the study group. Most of the procedures were elective (69.7%), with the commonest primary diagnosis being symptomatic cholelithiasis (35.7%), followed by chronic cholecystitis (29.7%) and acute cholecystitis (24.3%). Presence of diabetes mellitus, hypertension and COPD was seen in 17.0%, 21.3% and 5.3% of the patients, respectively. Median duration of symptoms was 63.25 days (IQR 6.50-122.47), while the median pre-operative waiting time was 17.85 hours (IQR 9.70-26.62). Median operative time was 71.0 minutes (IQR 58.0-86.0), while the median estimated blood loss was 44.0 mL (IQR. Table 1. Bivariate analysis of factors associated with postoperative complications Risk factor Test Statistic p-value Gallbladder wall thickness (mm) Mann–Whitney U 2304 0.05 Biliary anatomical variation Fisher's exact — 0.075 Hypertension Fisher's exact — 0.077 Critical View of Safety achieved Fisher's exact — 0.115 CBD stones Fisher's exact — 0.115 ALT (U/L) Mann–Whitney U 1278 0.127 Preoperative waiting time (hours) Mann–Whitney U 1284.5 0.132 Haemoglobin (g/dL) Independent t-test 1.564 0.144 Postoperative hospital stays (hours) Mann–Whitney U 2116 0.188 BMI category Chi-square 4.682 0.197 Sex Chi-square 1.652 0.199 Table 2. Bivariate analysis of factors associated with conversion to open surgery Risk factor Test Statistic p-value Postoperative hospital stays (hours) Mann–Whitney U 4378.5 <0.001 Time to first mobilization (hours) Mann–Whitney U 4105 <0.001 Time to oral diet (hours) Mann–Whitney U 3916.5 <0.001 Symptom duration (days) Mann–Whitney U 1246.5 0.002 Pain score Independent t-test 3.063 0.007 Direct bilirubin (mg/dL) Mann–Whitney U 1445.5 0.014 COPD Fisher's exact — 0.046 Analgesic strategy Chi-square 5.328 0.07 Primary diagnosis Chi-square 6.985 0.072 Weight (kg) Mann–Whitney U 2877 0.073 Total bilirubin (mg/dL) Mann–Whitney U 1692.5 0.086 AST (U/L) Mann–Whitney U 1699 0.09 Surgery type Fisher's exact — 0.095 Gallbladder wall thickness (mm) Mann–Whitney U 2833 0.097 BMI (kg/m²) Independent t-test 1.733 0.102 Pericholecystic fluid/collection Fisher's exact — 0.139 Largest stone (mm) Mann–Whitney U 2761.5 0.147 Wall thickness >5 mm Fisher's exact — 0.178 BMI category Chi-square 4.68 0.197 Table 3. Bivariate analysis of factors associated with delayed early recovery Risk factor Test Statistic p-value Postoperative hospital stays (hours) Mann–Whitney U 16141 <0.001 Primary diagnosis Chi-square 26.587 <0.001 Symptom duration (days) Mann–Whitney U 5607 <0.001 Surgery type Chi-square 18.63 <0.001 Gallbladder wall thickness (mm) Mann–Whitney U 11178.5 <0.001 Wall thickness >5 mm Chi-square 15.299 <0.001 Total leukocyte count Mann–Whitney U 11025 <0.001 CRP (mg/L) Mann–Whitney U 10792.5 <0.001 Murphy's sign Chi-square 10.603 0.001 Time to first mobilization (hours) Mann–Whitney U 10565.5 0.002 Time to oral diet (hours) Mann–Whitney U 10469 0.003 Direct bilirubin (mg/dL) Mann–Whitney U 6805 0.009 Pericholecystic fluid/collection Chi-square 5.836 0.016 Calot's triangle adhesions Chi-square 4.836 0.028 ALP (U/L) Mann–Whitney U 7370 0.081 Height (cm) Independent t-test 1.719 0.088 Total bilirubin (mg/dL) Mann–Whitney U 7417.5 0.094 Biliary anatomical variation Chi-square 2.594 0.107 Weight (kg) Mann–Whitney U 9521.5 0.123 Platelets Independent t-test 1.541 0.126 Diabetes mellitus Chi-square 2.079 0.149 Previous biliary attacks Mann–Whitney U 7629 0.164 Calot's triangle difficulty Chi-square 3.579 0.167 Subtotal cholecystectomy Fisher's exact — 0.173 Age (years) Mann–Whitney U 9399.5 0.175 BMI category Chi-square 4.872 0.181 Pain score Independent t-test 1.305 0.194 In the bivariate analyses, the occurrence of postoperative complications was not associated with any of the studied risk factors at the p<0.05 level of significance, although multiple factors passed the predefined p<0.20 level of significance to enter the multivariable model. Conversion to open surgery was significantly associated with various clinical and biochemical factors, such as symptom duration, pain score, direct bilirubin, and COPD. Delayed early recovery had the highest number of significant associations with factors, such as emergency surgery, primary diagnosis, symptom duration, gallbladder wall thickness, inflammation markers, and Murphy's sign, among others. This suggests that the clinical and inflammatory nature of gallbladder disease may have an impact on recovery following the surgery. However, as postoperative complications and conversion were relatively rare outcomes (4.0% and 5.3%, respectively), the bivariate results need to be treated as exploratory and confirmed by a multivariable logistic regression model. Table 1. Multivariable logistic regression for postoperative complications within 30 days Predictor amor 95% CI p-value Biliary anatomical variation (Yes vs No) 3.59 0.898–14.354 0.071 Gallbladder wall thickness (per 1 mm increase) 1.229 0.845–1.788 0.28 Table 2. Multivariable logistic regression for conversion to open surgery Outcome: conversion to open surgery = Yes Reference: No conversion Predictor amor 95% CI p-value Symptom duration (per 1-day increase) 0.981 0.968–0.994 0.004 Direct bilirubin (per 1 mg/dL increase) 0.018 0.001–0.459 0.015 Table 3. Multivariable logistic regression for delayed early recovery Outcome: delayed early recovery = Composite early recovery "No" Reference: early recovery achieved Predictor amor 95% CI p-value Primary diagnosis Biliary colic vs acute cholecystitis 0.316 0.039–2.579 0.282 Chronic cholecystitis vs acute cholecystitis 0.503 0.064–3.955 0.514 Symptomatic cholelithiasis vs acute cholecystitis 0.483 0.068–3.442 0.467 Emergency vs elective surgery 1.094 0.281–4.263 0.897 Murphy's sign: positive vs negative 1.007 0.468–2.164 0.987 Symptom duration (per 1-day increase) 0.996 0.990–1.003 0.278 Gallbladder wall thickness (per 1 mm increase) 1.067 0.786–1.448 0.679 Total leukocyte count (per unit increase) 1.045 0.927–1.178 0.471 CRP (per 1 mg/L increase) 0.998 0.989–1.008 0.748 Model performance Evaluation of the discrimination of the final predictive models was done using the Area Under Receiver Operating Characteristic Curve (AUC), while the calibration was performed using Hosmer-Lemeshow Goodness-of-Fit Test. Table. Discrimination and calibration of the final logistic regression models Outcome AUC Hosmer–Lemeshow χ² df p-value Interpretation Postoperative complications 0.697 6.125 8 0.633 Modest discrimination; good calibration Conversion to open surgery 0.793 4.446 8 0.815 Good discrimination; good calibration Delayed early recovery 0.683 6.137 8 0.632 Modest discrimination; good calibration The model had an AUC value of 0.697, which is a fair discriminative capacity in identifying postoperative complications from the absence of postoperative complications in the patients. The Hosmer-Lemeshow goodness-of-fit test for the model is not significant (p-value = 0.633). Conversion to open surgery: For this model, there was good discriminative capability with an AUC of 0.793. The Hosmer-Lemeshow test for this model is also not significant (p-value = 0.815). Delayed early recovery: AUC value is 0.683, which shows a fair discriminative ability. The Hosmer-Lemeshow test for this model is also not significant (p-value = 0.632). Of the three models, the conversion to open surgery model had the best discrimination (AUC=0.793). The post-operative complications and delayed early recovery models had good discrimination with Acs of 0.697 and 0.683, respectively. All three models had non-significant Hosmer-Lemeshow goodness-of-fit tests (p>0.05), meaning that there is no statistical evidence of lack of fit in the models. Thus, the models had good calibration although their discrimination, especially regarding post-operative complications and delayed recovery, was only moderate. Sub-group and center level analysis In this dataset, there are patients from five participating centers. Since it is possible that patients from the same center might have correlated responses, a generalized estimating equation (GEE) approach might be considered where the center would serve as the clustering variable. Table 1. Center distribution Participating center n % Center A 68 22.7 Center B 45 15 Center C 50 16.7 Center D 73 24.3 Center E 64 21.3 Total 300 100 Table 2. Center-adjusted GEE analysis Outcome Predictor Adjusted OR 95% CI p-value Postoperative complications Biliary anatomical variation 3.487 0.894–13.610 0.072 Gallbladder wall thickness 1.243 1.089–1.419 0.001 Conversion to open surgery Symptom duration 0.98 0.973–0.988 <0.001 Direct bilirubin 0.015 0.0001–2.379 0.104 Delayed early recovery Biliary colic vs acute cholecystitis 0.315 0.048–2.084 0.231 Chronic cholecystitis vs acute cholecystitis 0.503 0.299–0.847 0.01 Symptomatic cholelithiasis vs acute cholecystitis 0.478 0.191–1.197 0.115 Emergency vs elective surgery 1.099 0.474–2.546 0.826 Murphy's sign positive vs negative 1.009 0.793–1.284 0.939 Symptom duration 0.996 0.989–1.003 0.288 Gallbladder wall thickness 1.066 0.848–1.340 0.584 Total leukocyte count 1.044 0.943–1.155 0.405 CRP 0.998 0.988–1.008 0.745 After accounting for clustering by participating center: • Postoperative complications: gallbladder wall thickness was an independent risk factor for complications (OR 1.243, 95% CI 1.089-1.419, p=0.001). Biliary anatomy variant had a borderline association (OR 3.487, p=0.072). • Converting to open surgery: symptom duration was an independent risk factor for converting (OR 0.980, 95% CI 0.973-0.988, p<0.001). Direct bilirubin no longer had a statistically significant association when adjusted for center. • Delayed early recovery: Chronic cholecystitis, compared with acute cholecystitis, was significantly associated with the outcome (OR 0.503)

This multicenter study investigated the perioperative profile, postoperative complications, conversion rates, and early postoperative recovery in 300 patients who underwent laparoscopic cholecystectomy. In general, it was found that there was relatively low prevalence of postoperative complications and conversions to open surgery, as well as a higher proportion of cases of delayed early postoperative recovery. It should be emphasized that factors that predict each of the considered surgical outcomes were different, which reflects the fact that postoperative complications, conversion, and functional recovery are different aspects of surgical outcome.

The study group was a median age of 49 years old and consisted of 57.0% females. Elective procedures were done in most cases (69.7%), and the most common diagnosis was symptomatic cholelithiasis, followed by chronic and acute cholecystitis. Overall, median operation time was 71 minutes, and median estimated blood loss was 44 malls. These results imply that most of the procedures were done rather quickly and with minimal bleeding. Relatively high proportion of elective operations might contribute to positive perioperative profile of the cohort as well. Despite this, clinical characteristics had significant heterogeneity, including cases of acute inflammatory process, presence of concomitant diseases, gallbladder wall thickening, Calot's triangle adhesions and other features that make the procedure difficult.

Postoperative complications within 30 days occurred in 12 out of 300 patients (4.0%). According to Clavien-Dindo classification, four patients developed grade I complications, four patients developed grade II complications, and two patients developed grade III and grade IV complications correspondingly. Grade V complications were not identified in the cohort. Thus, although the overall frequency of complications was relatively low, there was some proportion of patients that had complications that required rather intensive treatment.

Low frequency of complications is clinically important since it shows that laparoscopic cholecystectomy was safe in the present patient cohort. However, low number of complications severely restricts the power of the study to reveal predictors of adverse outcomes. As per bivariate analysis, no variable reached the conventional statistical significance level (p<0.05). Gallbladder wall thickness had the strongest association with the complications (p=0.050), while biliary anatomical variation, hypertension, critical view of safety failure, CBD stones, ALT, and preoperative waiting time also reached prespecified p-value of 0.20 to be included into multivariable analysis.

In the multivariable logistic regression model, symptom duration was still a statistically significant predictor with an odds ratio of 0.981 (95% CI 0.968–0.994; p=0.004) per each additional day of symptoms. Bilirubin also stayed a statistically significant predictor with an adjusted odds ratio of 0.018 (95% CI 0.001–0.459; p=0.015). However, it is essential to interpret these results with care especially considering the non-intuitive direction and magnitude of the estimated predictors' effect. The adjusted odds ratio reported to be below 1 for the symptom duration predictor should be interpreted not as increased risk but as the decreased modelled odds of conversion with an increase in symptom duration. Consequently, the statement that increased symptom duration predicts higher risk of conversion would be incorrect.

After adjusting for clustering at the center level, symptom duration stays independent predictor with an odds ratio of 0.980 (95% CI 0.973–0.988; p<0.001), while the association between the direct bilirubin level and the conversion becomes statistically insignificant with an odds ratio of 0.015 (95% CI 0.0001–2.379; p=0.104). It can be assumed that the relationship between the symptom duration predictor and the conversion outcome is relatively stable with the center-level adjustment, while the effect of direct bilirubin predictor appears to be unstable.

Among all three models, the conversion outcome model demonstrates the best predictive power with an AUC of 0.793. This can be considered a good discrimination. This means that the model was able to differentiate between the cases of conversion and the cases when procedure was performed laparoscopically relatively well. The Hosmer-Lemeshow test is not significant (p=0.815).

However, in multivariable logistic regression, duration of symptoms remained statistically significant (adjusted OR 0.981 per day increase, 95% CI 0.968–0.994; p=0.004). Direct bilirubin also proved statistically significant (adjusted OR 0.018, 95% CI 0.001–0.459; p=0.015). It is noteworthy to mention that both findings need to be analyzed with care since neither the direction, nor the magnitude of the effect seems to be obvious. Thus, according to the results, there was a negative association between duration of symptoms and conversion probability. In other words, the reported odds ratio is below 1, and hence, one can speak about decreased conversion risk associated with increased duration of symptoms. That is why it would not be correct to claim that there is a positive relationship between these two variables.

After considering the clustering by center, the independent association between duration of symptoms and conversion was revealed (OR 0.980, 95% CI 0.973–0.988; p<0.001), while the relationship between direct bilirubin and conversion became non-significant (OR 0.015, 95% CI 0.0001–2.379; p=0.104). Therefore, the relationship between duration of symptoms and conversion seems to be more stable in comparison with that between direct bilirubin and conversion probability.

The conversion model proved to have the best predictive ability compared to other two models of outcome with the highest area under ROC curve equal 0.793. Thus, there was good discrimination which means that the model successfully discriminated patients undergoing conversion from laparoscopic surgery. The Hosmer-Lemeshow test turned out to be non-significant (p=0.815).

Early recovery was documented in 224 patients (74.7%), while 76 patients (25.3%) failed to satisfy the composite early-recovery endpoint. Despite 98.0% mobilization rate within 24 hours, 77.3% patients were hospitalized for no more than 48 hours. This implies that hospital length of stay played a significant role in the lack of achievement of the composite early-recovery endpoint. The median time to first mobilization was 10.15 hours, median time to oral diet was 12.80 hours, and median postoperative hospital stay was 37.80 hours.

Bivariate analysis found the maximum number of significant associations for delayed early recovery. The primary diagnosis, surgery type, symptom duration, gallbladder wall thickness, total leukocyte count, CRP, Murphy's sign, time to mobilization, time to oral intake, direct bilirubin, presence of pericholecystic fluid/collection, and Calot's triangle adhesions were significant for delayed recovery. Overall, these findings imply that recovery was significantly linked to disease severity, inflammation, and technical difficulty of surgical intervention.

It is clinically plausible to consider an association between inflammatory and radiologic parameters and delayed recovery in terms of the current data set. Patients who had thick gallbladder walls, increased inflammatory markers, positive Murphy's sign, pericholecystic collections, and Calot's triangle adhesions had higher inflammatory or disease features. According to the current dataset, these factors were significantly linked to lack of achievement of the composite early-recovery endpoint in bivariate analysis. However, it should be noted that bivariate association does not imply causation.

Importantly, many of the associations were lost in multivariable adjustment. In the final logistic regression model, none of the investigated delayed early recovery predictors was statistically significant. Primary diagnosis, emergency versus elective surgery, Murphy's sign, symptom duration, gallbladder wall thickness, total leukocyte count, and CRP had p-values above 0.05. Overall, the model failed to identify any statistically significant independent predictor of delayed early recovery in multivariable analysis.

The accuracy of predictions made with this model was quite low, with an area under ROC curve of 0.683. However, the Hosmer–Lemeshow test failed to demonstrate any significant problems with the model (p=0.632). Therefore, despite the model's poor discrimination power, it could be considered acceptable in terms of calibration.

After adjustment for center using GEE, chronic cholecystitis in comparison with acute cholecystitis was significantly associated with the delayed-recovery outcome (OR 0.503, 95% CI 0.299–0.847; p=0.010). Other variables tested in the study did not show any significant associations after adjustment for center. This result shows how important it is to account for clustering of data in multicenter studies.

The 30-day readmission rate was 4.3%, while reoperation was needed only in two cases (0.7%). These results provide further support to the favorable short-term postoperative outcomes noted in the study population. However, since the study examined postoperative outcomes over 30 days, it cannot reveal potential complications developing later.

Comparison of all three predictive models showed that conversion to open surgery is the outcome where the predictors were able to achieve the most accurate prediction. The AUC value for the conversion model was 0.793, while 0.697 and 0.683 were recorded for postoperative complications and delayed early recovery, respectively. All three models showed acceptable calibration according to nonsignificant Hosmer–Lemeshow tests.

The relatively high AUC value for the conversion model can be explained by the fact that conversion is a parameter that can be measured directly. On the contrary, postoperative complications and delayed early recovery depend on many factors, such as patient's characteristics, disease severity, operative conditions, perioperative management and center-specific practice. The low AUC values for these outcomes prove that the variables used in the present models do not explain individual differences in postoperative recovery or complications.

Furthermore, the center-adjusted analysis shows that the postoperative outcomes of surgery in multicenter settings can be influenced by different centres’ peculiarities. In this study, five centers provided their patients, and the proportion of the sample coming from center varied between 15.0% and 24.3%. Some associations changed when the clustering of data was accounted for, and Gallbladder wall thickness became significant for postoperative complications, direct bilirubin lost significance for conversion, while chronic cholecystitis became significantly associated with delayed early recovery.

The multicenter study indicates that laparoscopic cholecystectomy had generally positive short-term surgical outcomes, considering that 4.0% of patients had postoperative complications, 5.3% of patients underwent conversion to open surgery, 4.3% of patients had 30-day readmissions, and reoperations were performed in only 0.7% of patients. In general, 74.7% of patients reached the composite early-recovery endpoint, which shows that most of the patients had satisfactory postoperative recovery.

Differences in the association patterns have been found for the main outcomes in this research. Gallbladder wall thickness was an independent predictor of postoperative complications, controlling for clustering by the participating center, while symptom duration was significantly associated with conversion to open surgery. Delayed early recovery was associated with several factors reflecting inflammatory and difficult gallbladder disease in bivariate analysis; nevertheless, most of them were not significantly associated with the outcome in multivariable and center-adjusted analyses.

As far as the predictive models are concerned, the conversion-to-open-surgery model showed the best discriminatory ability (AUC=0.793), while the postoperative complication and delayed recovery models showed moderate discrimination (Acs=0.697 and 0.683, respectively). All the three models were shown to be well-calibrated according to the nonsignificant Hosmer-Lemeshow tests. The results demonstrate that prediction of conversion is more promising in terms of the available predictors, while the prediction of postoperative complications and early recovery should be improved.

The results confirm the significance of systematic preoperative and intraoperative evaluation of patients undergoing laparoscopic cholecystectomy, especially assessment of inflammatory parameters of the gallbladder and technical complexity of the surgery. Nevertheless, low numbers of postoperative complications and conversions restrict accuracy of some estimates, and observational nature of the study does not allow drawing any causal conclusions. Therefore, the predictive models should be considered as preliminary and require validation on larger prospective cohorts.

The results show that laparoscopic cholecystectomy provides favorable perioperative outcomes in the studied population, and the gallbladder wall thickness and symptom duration are the important predictors of adverse outcomes in selected cases. The findings can be considered as valuable evidence for perioperative risk assessment and can be helpful in developing patient counseling, surgical planning, and recovery strategies. However, future multicenter studies involving larger samples and validation are needed for confirmation of these associations and development of robust prediction models.