Acute cholecystitis

Background: In comparison to delayed laparoscopic cholecystectomy, emergency laparoscopic chole- cystectomy has a shorter length of stay and eliminates the risk of recurrent episodes of acute cholecystitis. Nevertheless, there is concern that emergency laparoscopic cholecystectomy is associated with higher morbidity in acute cholecystitis patients. The present large cohort study compares operation-related adverse outcomes between emergency and delayed laparoscopic cholecystectomy and de- termines the risk of readmission before delayed laparoscopic cholecystectomy to guide surgical decision-making. Methods: Patients diagnosed with acute cholecystitis who underwent emergency or delayed laparoscopic cholecystectomy between 2015 and 2019 were included. Perioperative outcomes were compared using univariate and multivariate analysis, adjusting for preoperative variables. The rate of readmission before delayed laparoscopic cholecystectomy was determined. Results: In total, 811 patients were included (median age, 58 years; male:female, 1:1.5): 227 emergency laparoscopic cholecystectomies (28.0%), 555 delayed laparoscopic cholecystectomies (68.4%), and 29 emergency laparoscopic cholecystectomies whilst awaiting delayed laparoscopic cholecystectomy (3.6%). Emergency laparoscopic cholecystectomy was associated with increased incidences of subtotal chole- cystectomy (OR 1.94; P ¼ .011), bile leak (OR 2.38; P ¼ .013), intraoperative drains (OR 2.54; P < .001), prolonged postoperative length of stay (OR 7.26; P < .001), postoperative imaging (OR 1.83, P ¼ .006), and postoperative readmission (OR 1.90; P ¼ .005). There was a 13.5% risk of readmission over 2 months while waiting delayed laparoscopic cholecystectomy and a 22.5% risk over the median waiting time (5 months, 9 days). Conclusion: Emergency laparoscopic cholecystectomy is positively associated with a multitude of operation-related adverse outcomes in acute cholecystitis, compared to delayed laparoscopic cholecystectomy. The bene ﬁ t of delayed laparoscopic cholecystectomy should be balanced against the signi ﬁ cant readmission risk before delayed laparoscopic cholecystectomy. Emergency laparoscopic cholecystectomy may be the most pragmatic strategy for centers dealing with high volumes of biliary admissions and long elective-surgery waiting times. When opting for delayed laparoscopic cholecystectomy, con ﬁ rming the date of surgery before discharge may ensure timely intervention and avoid the morbidity and expense of readmission.


Introduction
Patients admitted with acute cholecystitis (AC) can be managed by emergency laparoscopic cholecystectomy (ELC) or initial conservative management followed by delayed laparoscopic cholecystectomy (DLC). ELC is generally the preferred approach, at least for mild acute cholecystitis, due to a shorter length of stay, a similar conversion rate, and the elimination of recurrent biliary symptoms. 1e3 However, studies supporting ELC are typically small randomized controlled trials not powered according to rates of overall morbidity; therefore, the relative rates of many perioperative outcomes between ELC and DLC are overlooked. 4e7 Large population-level studies report significantly higher morbidity after ELC; however, these are not specific to AC. 8e10 The traditional surgical literature also has a strong emphasis on conversion-to-open as an endpoint, an increasingly uncommon event, particularly as the laparoscopic subtotal approach gains increasing popularity. 11 A large contemporary cohort study of AC patients is indicated to compare perioperative morbidity between the 2 approaches. Proponents of ELC argue that readmission before DLC with further episodes of AC increases the difficulty of cholecystectomy due to repeated episodes of inflammation. 11 Currently, the risk of readmission before DLC is unreported in AC patients. Quantifying this risk would help surgeons and institutions decide whether to aggressively pursue ELC versus DLC, especially in conjunction with acknowledgment of the relative risks of outcomes between the 2 groups.
To better inform the surgical decision-making and consent process, the present study uses a large cohort to compare outcomes of ELC versus DLC in patients admitted with AC. The risk of readmission before DLC and risk factors for readmission are also determined.

Population cohort
All patients who were admitted with AC and then underwent either ELC or DLC between January 2015 and December 2019 in a UK health board were included in the study. The health board covers a defined geographical region with a stable population of approximately 493,000 people. Operations were performed in 1 tertiary center and 2 day-surgical units by a total of 25 general surgical consultants. Patients who had AC diagnosed in the outpatient setting not requiring admission and those who had an admission for AC but subsequently underwent an ELC for a different indication (eg, choledocholithiasis) were excluded ( Figure 1). The definition for AC used in the present study was that described by the Tokyo 2018 Guidelines (TG18). 12

Data collection
Data were collected retrospectively from multiple databases using a deterministic records-linkage methodology. Patients were tracked between databases using a unique 10-digit patient identifier. Data were collected for all biliary-related admissions until operation. Patients were then followed up for 100 days after the operation for all readmissions and outpatient reviews. These data yielded preoperative data, operative data, significant complications (Clavien-Dindo classification !2), postoperative imaging, postoperative intervention, postoperative length of stay (PLOS), related readmission data, total length of stay (TLOS), and mortality.

Analysis
Patients were divided into ELC and DLC groups. ELC patients were further divided into those undergoing ELC on first cholecystitis admission and those undergoing ELC on second cholecystitis admission. Perioperative and postoperative outcomes were compared between ELC and DLC groups using univariate and multivariate logistic analysis. Multivariate logistic regressions were conducted for the following outcomes measures: subtotal cholecystectomy; conversion to open; intraoperative complication; drain insertion; prolonged PLOS (!3 days); postoperative complication/ imaging/intervention; and readmission. PLOS of !3 days has been demonstrated to have a significant association with perioperative mortality and therefore was classed as the cut-off for the "prolonged PLOS" outcome measure. 13 Multivariate linear regression was used to identify variables associated with TLOS.
A cholecystectomy grade score such as that outlined in TG18 was not used because it is has been shown to be poor predictor of operative-related adverse outcomes and does not account for particular preoperative variables (eg, age, comorbidities, number of previous admissions, CRP) that have demonstrated good diagnostic ability for operation-related adverse outcomes. 11,14e21 ELC patients were divided into subgroups based on time from admission until surgery (<48 hours; 48e96 hours; >96 hours), and a subgroup analysis was performed. Outcome measures were compared between subgroups using c 2 , Fisher exact, and Mann-Whitney U tests.
In patients who did not undergo ELC on first admission, the risk of readmission was determined. This was calculating by calculating the time from discharge after conservative management until readmission and was displayed using a Kaplan-Meier graph. Cox proportional hazards model was used to identify factors associated with readmission. All statistical tests were performed using the STATA/IC 16.1 software package.

Results
Overall, 811 patients were included in the analysis (median age, 58 years; M:F, 1:1.5; median ASA, 2) (Table I); 227 patients underwent first cholecystitis admission ELC, 29 patients underwent ELC on subsequent admission whilst awaiting DLC, and the remaining 555 patients underwent DLC. ELC patients were younger than DLC patients (P ¼ .004) and more likely to be female (P < .001). ELC patients were less likely to have undergone a preoperative CT abdomen/pelvis (20.3% vs 33.5; P ¼ .008) and magnetic resonance cholangiopancreatography (35.1% vs 58.0%; P < .001). ELC patients were less likely to have choledocholithiasis during index admission and had lower admission CRP, creatinine, bilirubin, and alkaline phosphatase levels (P < .001). DLC patients were more likely to have a preoperative ERCP (21.8% vs 10.9%) and cholecystostomy (6.4% vs 1.6%) when compared to ELC patients.
In the subgroup analysis of the ELC group, rates of subtotal cholecystectomy were higher in the 48-to 96-hour group and >96hour group compared to the 0-to 48-hour group (P ¼ .004). Other than this finding, there were no other significant differences in adverse outcomes between groups, including operative time, intraoperative complication, intraoperative drains, conversion to open, rates of prolonged PLOS, postoperative complication/imaging/intervention, and readmission (P > .05).

Readmission before DLC
The overall risk of readmission before DLC in our cohort was 22.3% (130/584). In total, 61.5% (80/130) were readmitted with another episode of cholecystitis, 29 (36.3%) of which proceeded to ELC on second AC admission, and the remaining 51 (63.7%) proceeded to DLC.
The Kaplan-Meier graph demonstrated a 13.5% risk of readmission before DLC over 2 months after discharge (Figure 2). The median waiting time for DLC was 5 months and 9 days, at which point the readmission risk is estimated as 22.5%. In the present study, only 11.3% and 17.4% of patients had a DLC before 6 and 8 weeks, respectively. Cox proportional hazard model was used to identify factors associated with early readmission before DLC. Variables included in the model included age<40, age 40 to 60, age >60, male sex, 2 previous biliary admissions, !3 biliary related admissions, choledocholithiasis, and gallstone pancreatitis. This model did not identify any factors significantly associated with earlier readmission (P > .05).

Discussion
In conclusion, these data suggest that DLC for AC results in lower rates of operation-related adverse outcomes (need for subtotal, intraoperative drain placement, postoperative complication/imaging/intervention, and readmission [P < .05]). The findings were corroborated in the multivariate analysis, after adjusting for key preoperative variables (eg, degree of inflammation, number of admissions, and comorbidities) that have demonstrated predictability for a difficult perioperative course after cholecystectomy. Adjusting for these variables offsets a significant degree of bias and indicates that the risk of ELC in AC patients is considerable relative to DLC. This should be used to inform surgical decision-making and the consent process.
Interestingly, it was noted that the patients in the ELC group were younger and had lower inflammatory markers on admission (P < .05). Despite these advantages, the ELC group had worse outcomes. This further supports the benefit of performing a delayed laparoscopic cholecystectomy after a "cooling off" period. 22,23 The readmission risk before DLC is significant (13.5% over 2 months; 22.5% over the median waiting time [5 months, 9 days]) and must be minimized. Recurrent admissions ultimately result in increased morbidity that may actually outweigh the benefit of DLC. 11 Prompt DLC 6 to 8 weeks after discharge would ensure timely intervention and may avoid the morbidity, inconvenience, and expense of readmission. This should be emphasized and may be achieved by confirming the date of surgery before discharge.
The total length of stay is an important consideration with significant financial implications. Although ELC is associated with a shorter total length of stay compared to DLC (mean difference, -1.2 days), the cost related to the additional morbidity of ELC must be recognized. These patients require more imaging, intervention, and emergency readmissions, all of which have resource implications. A cost analysis that incorporates these aspects is necessary before a valid financial comparison can be made between ELC and DLC.
Previous studies that have compared ELC and DLC are often limited to the comparison of bile duct injury, conversion rates, and total length of stay. It could be argued that bile duct injury is relatively rare in contemporary practice and therefore may have somewhat limited utility as an endpoint in all but the largest studies. Likewise, rates of conversion to open surgery are decreasing, suggesting this factor may not be quite so relevant as it once was. The present study and other contemporary publications demonstrate that rates of other operation-related adverse outcomes are surprisingly common after laparoscopic cholecystectomy. 2-7,24,25 These include, but are not limited to, rates of lesser severe postoperative complications, the utilization of drains, and postoperative readmissions. Future studies addressing the issue of emergency or delayed cholecystectomy should report the multitude of problems that can occur  after cholecystectomy (Table II) to provide a detailed comparative analysis.
Although the TG18 guidelines are generally seen as being supportive of ELC for AC, it must be recognized that most of the nonrandomized studies included in the guidelines suffer from low sample sizes and do not conduct multivariate analysis to adjust for preoperative variables, thus subjecting them to significant bias. 3,11 The randomized studies reported in TG18 also depend on small sample sizes, short follow-up times and are largely outdated. 4e7 It must be acknowledged that some of the randomized studies listed in the TG18 guidelines such as Ozkardeş et al and Kolla et al actually found higher rates of operation-related adverse outcomes in the ELC group, a finding replicated in this study. 4 Large cohort studies comparing ELC and DLC have been conducted by Giger et al, the CholeS group, and the Swedish Registry and have all found significantly higher rates of perioperative adverse outcomes in the emergency group. 8e10 Although the above studies were not conducted solely on patients with AC, they strongly imply the increased risk of performing a laparoscopic cholecystectomy at the time of acute inflammation and are consistent with the present article.
In conclusion, the benefit of reduced morbidity by opting for DLC should be balanced against the significant risk of readmission before DLC. In this regard, the most pragmatic strategy is dependent on regional waiting list times for elective surgery; centers dealing with high volumes of biliary admissions and long waiting times may benefit from routine ELC despite the increased morbidity. Where DLC is chosen, every effort should be made to ensure timely intervention to avoid the morbidity, inconvenience, and expense of readmission.

Funding/Support
No funding or financial support.

Conflict of interest/Disclosure
There are no conflicts of interest to declare by any of the authors. CRP, c-reactive protein; ELC, emergency laparoscopic cholecystectomy; WBC, white blood cell count. Figure 2. Risk of being readmitted before DLC after first emergency admission (red vertical line represents 2 months after discharge).