Time to interval cholecystectomy and associated outcomes in a population aged 50 and above with mild gallstone pancreatitis

Background Cholecystectomy on index admission for mild gallstone pancreatitis (GSP) is recommended, although not always feasible. This study examined rates and outcomes of people aged ≥ 50 years who underwent interval (delayed) cholecystectomy at increasing time points. Methods Hospitalisation and death data were linked for individuals aged ≥ 50 years admitted to hospital in New South Wales, Australia with mild GSP between 2008–2018. Primary outcome was interval cholecystectomy timing. Secondary outcomes included mortality, emergency readmission for gallstone-related disease (GSRD) (28 and 180-day), and length of stay (LOS) (index admission and total six-month GSRD). Results 3,003 patients underwent interval cholecystectomy: 861 (28.6%) at 1–30, 1,221 (40.7%) at 31–90 and 921 (30.7%) at 91–365 days from index admission. There was no difference in 365-day mortality between groups. Longer delay to cholecystectomy was associated with increased 180-day emergency GSRD readmission (17.5% vs 15.8% vs 19.9%, p < 0.001) and total six-month LOS (5.9 vs 8.4 vs 8.3, p < 0.001). Endoscopic retrograde cholangiopancreatography (ERCP) was increasingly required with cholecystectomy delay (14.5% vs 16.9% vs 20.4%, p < 0.001), as were open cholecystectomy procedures (4.8% vs 7.6% vs 11.3%, p < 0.001). Extended delay was associated with patients of lower socioeconomic status, regional/rural backgrounds or who presented to a low volume or non-tertiary hospital (p < 0.001). Conclusion Delay to interval cholecystectomy results in increased rates of emergency readmission, overall LOS, risks of conversion to open surgery and need for ERCP. Index admission cholecystectomy is still recommended, however when not possible, interval cholecystectomy should be performed within 30 days to minimise patient risk and healthcare burden.


Background
Gallstone disease is a leading cause of acute pancreatitis comprising 40-70% of all cases [1].Laparoscopic cholecystectomy is recommended for patients presenting with gallstone pancreatitis (GSP) to prevent recurrence [2].Optimal timing of cholecystectomy is dependent on episode severity as classified by the Revised Atlanta Criteria [1,2].In cases of mild GSP, characterised by the absence of organ failure and local or systemic complication, international guidelines recommend laparoscopic cholecystectomy be performed during the same (index) admission [2][3][4][5].Delay in cholecystectomy beyond index admission is associated with increased risk of recurrent gallstone-related disease (GSRD) including pancreatitis, cholecystitis, and cholangitis [5][6][7].A systematic review by Van Baal et al [8] identified an 18% greater GSRD readmission rate within 90 days of index admission for mild GSP for those in whom cholecystectomy was delayed.
Index cholecystectomy for mild GSP is not always feasible [5], this may be due to patient, service, or surgeon factors.These include provision and access to emergency surgical services [9], surgeon availability and to allow for medical optimisation of patients or for medications such as antithrombotic agents to be withheld.This is seen more 380 Page 2 of 8 commonly in the older population, with poor adherence to guideline care demonstrated internationally [10] despite evidence supporting safe laparoscopic surgery in older people [11][12][13][14].A recent ten year population-based study revealed a 38% adherence to index cholecystectomy in a cohort aged 50 and above in New South Wales, Australia [5].Compared to index cholecystectomy, patients undergoing interval cholecystectomy were at greater risk of emergency readmission secondary to GSRD and open cholecystectomy procedures [5].Investigation into the timing of interval cholecystectomy may provide further insight into the variation in healthcare provision and impact on outcomes.This population-based retrospective cohort study aims to 1) quantify timing of interval cholecystectomy and 2) compare mortality, emergency readmission and length of stay (LOS) associated with increasing delay in interval cholecystectomy for mild GSP in individuals aged ≥ 50 years across all hospitals in New South Wales (NSW), Australia over a 10-year period.

Study design
The study cohort was identified from the Surgical Care of Older People (SCOPE) cohort; a linked dataset comprising linked hospital (Admitted Patient Data Collection (APDC)) and death (Register of Births, Deaths, and Marriages (RBDM) data for all adults aged ≥ 50 years admitted to all NSW hospitals under a surgical specialty between 1 January 2007 and 1 July 2019.The APDC is a census of all inpatient services provided in NSW public, private and day-stay facilities and contains information on patients' demographics, diagnoses, and procedures, coded using the International Classification of Diseases and Related Health Problems, Tenth Revision, Australian Modification (ICD-10-AM) and Australian Classification of Health Interventions (ACHI) [15,16].Ethics approval for this study was obtained from the NSW Population and Health Services Research Ethics Committee (2018HRE0201).Study design, data abstraction and findings were reported in accordance with The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines [17].

Case inclusion criteria
We identified all admitted patients aged ≥ 50 years with a primary diagnosis of acute GSP from 1 July 2008 to 30 June 2018, who underwent cholecystectomy within 12 months of index admission.Index admission with acute GSP was identified using the primary diagnosis ICD-10-AM codes K85.1 -'biliary acute pancreatitis' and K85.9-'acute pancreatitis, unspecified' [15].Interval admission cholecystectomy was defined as cholecystectomy delayed to a subsequent admission at least 24 h after index pancreatitis admission.An eighteen-month exclusion period prior to 1 July 2008 was implemented to preclude recurrent cases, ensuring only first presentations of acute GSP were captured.Patients admitted after 30 June 2018 were excluded to allow a one-year follow-up for all included patients.

Demographic characteristics
Patient demographic characteristics examined included: sex, age, Charlson Comorbidity Index (CCI), private health insurance status recorded at time of admission, Socio-Economic Index for Areas-Index of Relative Socio-economic Disadvantage (SEIFA-IRDS) and Accessibility/Remoteness Index of Australia (ARIA +).The 17 comorbidities contributing to the CCI were identified using the Quan ICD-10-AM coding algorithm [19], which has been shown to provide good prediction for 30 and 365-day mortality in the older surgical population [20].A one-year lookback period was used to optimise identification of comorbidities [21].Comorbidity was categorised into three groups (0, 1-2 and ≥ 3 comorbidities).CCI ≥ 3 was considered multimorbid.SEIFA-IRSD enabled ranking of geographic areas of residence into quintiles (1 = most disadvantaged, 5 = most advantaged) based on socioeconomic status [22].ARIA + provided classification of remoteness into major city, inner regional, outer regional, remote and very remote based on physical distance from nearest service centres [23].

Statistical analysis
Statistical analysis was undertaken using SAS Enterprise Guide v7.1 (SAS Institute Inc, Cary NC, USA).Descriptive statistics were used to characterise cohort demographics and assess differences in secondary outcomes between 1-30, 31-90 and 91-365-day interval cholecystectomies.Cochran-Armitage test for trend (1-sided) was used to test for increased proportions across the increasing time periods, and ANOVAs to compare differences in means for continuous variables (age and LOS).Post-hoc tests using Tukey's p-value adjustment were used to determine where the between group difference occurred.Univariate binomial regression was used to test the association of individual demographic and hospital characteristics on delayed interval cholecystectomies (defined as > 30 days between index admission and cholecystectomy) and reported as relative risk (RR) with 95% confidence intervals.

Results
A total of 3,003 patients were identified as meeting the inclusion criteria of age ≥ 50 years, admitted to a NSW hospital with acute GSP who received interval cholecystectomy within a one year period.Of these, 861 (28.6%) occurred 1-30 days from index admission, 1,221 (40.7%) occurred 31-90 days from index admission and 921 (30.7%) occurred 91-365 days from index admission.Median time from index admission to interval cholecystectomy was 56 (IQR 27-107) days.
There was no significant difference in 365-day mortality following index presentation with GSP between the interval cholecystectomy groups (Table 2).Twenty eight day readmission was highest in those undergoing surgery within 30 days post index admission (15.7%), whilst 180 day readmission was highest in those with the longest delay to surgery (19.9%).

Discussion
This population-based study provides insight into the management of people ≥ 50 years of age with mild GSP, in NSW, over 10 years to 2018 and the impact of delayed surgical intervention.
These data give an indication of the 'current state of play' in NSW hospitals in respect to cholecystectomy for this disease.Index cholecystectomy is regarded 'best practice' for mild GSP.It is recognised that this is not always feasible for a variety of reasons.However the results from this study show that increasing delay is associated with negative outcomes for both patients and the health system.The increasing utilisation of acute surgical units and protected emergency operating times has been recommended and implemented in larger centres [9] but may be impossible to implement in smaller centres due to funding and staffing constraints.Where surgery on index admission is not possible, this study provides data to support a recommendation for admission within 30 days as opposed to 90 and 365 days.
In this population, interval cholecystectomy was performed with considerable delay; 71.3% of interval cholecystectomies were deferred beyond 30 days of index admission and on average performed at 56 days.Despite no difference in mortality, increasing delay in cholecystectomy was significantly associated with recurrent emergency readmission, greater LOS and unnecessarily complex or additional procedures.91-365 day cholecystectomy posed the greatest overall risk of emergency GSRD readmission.This was concordant with the significantly greater requirement for ERCP associated with cholecystectomy delay beyond 30 days, likely characterising the recurrence of biliary obstruction, cholangitis or the finding of choledocholithiasis.
High emergency readmission at 28 days was observed in the 1-30 day interval cholecystectomy group.We hypothesise this may reflect those with recurrent symptoms requiring readmission who, in most cases, then undergo a second admission emergency cholecystectomy; rather than elective.This is consistent with the average time to GSRD readmission of within 28 days of discharge in previous studies, and the comparatively low 28-day emergency GSRD readmission rates found in the 31-90 and 91-365 day cholecystectomy groups [6,24].This highlights a cohort of patients who would have clearly benefited from index or early interval cholecystectomy.
Older patients with GSP are at higher risk of complicated biliary tract disease necessitating open cholecystectomy procedures [25,26].Classical surgical teaching dictates delaying cholecystectomy to allow resolution of inflammation and improve chances of a laparoscopic procedure.However, delay in cholecystectomy increases the risk of both inability to perform a cholangiogram and conversion to an open procedure, as shown in this population and described in previous series of the older cohort [10,27].Furthermore, delay results in a greater ERCP requirement, which is associated with a more challenging subsequent cholecystectomy with longer operative times and risks of open conversion [28][29][30].Collectively, this may result from recurrent acute/ chronic inflammation, bacterial colonisation and scarring of the hepatoduodenal ligament, intra-abdominal adhesions or patient factors not definable by this study [28].
Overall, interval cholecystectomy performed within 30 days of discharge was superior, and on average at least a 1.3 day shorter index admission and 1.4 day shorter total LOS at six months than further delayed counterparts.This increase in LOS likely reflects the culmination of increased emergency GSRD readmissions, ERCPs and open cholecystectomy procedures associated with increasing operative delay.
Our findings highlight clear variation in the scheduling of interval cholecystectomy.Scheduling cholecystectomy within 30 days results in a decreased total six-month LOS and reduces emergency GSRD readmissions, the need for ERCP with sphincterotomy and conversion to open procedures.Accordingly, interval cholecystectomy performed as early as possible and within 30 days of index admission, may provide the best comparable outcomes to gold standard care with index cholecystectomy [5].Given the increased patient risk and health service burden associated with delay, interval cholecystectomy for mild GSP needs to be prioritised through national standardised recommendation for admission as urgent (Category 1), and ideally no later than 30 days from index admission.
Demographic and hospitalisation characteristics identified in this study represent risk factors for excessive delay in cholecystectomy and warrant further exploration.Lower socioeconomic status, regional and remote residence and admission to hospitals of low surgical volume and non-tertiary status were significantly associated with delay in interval cholecystectomy.Whilst this may in part reflect surgeon perception and availability, it highlights inequitable provision of health care.Regional centres carry a significant burden of emergency general surgery admissions [31] and insufficient access to emergency operating time acts to promote deferral of definitive management and further  [9].This model of care has been further validated in the Australian rural setting; the area for which our findings would suggest it is most needed [31].These findings precede the COVID-19 pandemic which has led to a substantial increase in time to surgery for elective procedures [32].Although private health insurance may provide an alternate avenue for early access to definitive management, this should not be seen as a replacement for early access through the public healthcare system.Additionally, whilst delay in favour of optimisation of pre-existing comorbidities is reasonable assuming that optimisation is achievable, this should be limited to within 30 days of discharge from index admission where possible given the increased risks associated with further delay.The strengths of this study include its size and population-based design which allows exploration of important patient and system level confounding factors.There was no loss to follow up based on study design, and readmissions, excluding those interstate, were accurately accounted for as the administrative hospital data includes all presentations to all NSW hospitals.There are however inherent limitations to administrative data.The databases supporting this study design are not established primarily for research and reliance is placed on accurate diagnosis and procedural coding.Values for private health insurance, SEIFA-IRDS, ARIA + and urgency of admission were not assigned in some individuals and therefore excluded from respective characteristic analyses.Furthermore, direct measures of acute pancreatitis severity such as end organ failure and local or systemic complications stipulated in the Revised Atlanta Classification, were not available.Proxy variables representative of moderately severe/severe cases of GSP such as major pancreatic resection, cholecystostomy and ICU admission during index admission were employed to address this.Complications such as bile duct injury and bile duct reconstruction, that potentially necessitate ICU admission are at risk of exclusion and underrepresentation in this data set.Differentiation of ICU admission into planned or emergency and pre-or post-cholecystectomy may have provided further insight into pancreatitis severity, however these data were not available.

Conclusion
In this population aged 50 and above, only one third of patients underwent interval cholecystectomy within 30 days of index admission.Delay in surgery beyond 30 days is associated with increasing risk of emergency readmission, ERCP and conversion to open cholecystectomy.Index cholecystectomy remains the treatment of choice, but when this is not possible, interval cholecystectomy should be recommended for admission within 30 days of discharge from hospital to minimise these risks and prevent undue health service burden.

Abbreviations
Abbreviations GSP: Gallstone pancreatitis; GSRD: Gallstone-related disease; ERCP: Endoscopic retrograde cholangiopancreatography; LOS: Length of stay; NSW: New South Wales; SCOPE: Surgical Care of Older People; APDC: Admitted Patient Data Collection; RBDM: Register of Births, Deaths, and Marriages; ICD-10-AM: International Classification of Diseases and Related Health Problems, Tenth Revision, Australian Modification; ACHI: Australian Classification of Health Interventions; STROBE: Strengthening the Reporting of Observational Studies in Epidemiology; ICU: intensive care unit; CCI: Charlson Comorbidity Index; SEIFA-IRDS: Socio-Economic Index for Areas-Index of Relative Socio-economic Disadvantage; ARIA +: Accessibility/Remoteness Index of Australia; SD: Standard deviation; IQR: Interquartile range