Effectiveness of conservative management versus laparoscopic cholecystectomy in the prevention of recurrent symptoms and complications in adults with uncomplicated symptomatic gallstone disease (C-GALL trial): pragmatic, multicentre randomised controlled trial

Abstract Objective To assess the clinical and cost effectiveness of conservative management compared with laparoscopic cholecystectomy for the prevention of symptoms and complications in adults with uncomplicated symptomatic gallstone disease. Design Parallel group, pragmatic randomised, superiority trial. Setting 20 secondary care centres in the UK. Participants 434 adults (>18 years) with uncomplicated symptomatic gallstone disease referred to secondary care, assessed for eligibility between August 2016 and November 2019, and randomly assigned (1:1) to receive conservative management or laparoscopic cholecystectomy. Interventions Conservative management or surgical removal of the gallbladder. Main outcome measures The primary patient outcome was quality of life, measured by area under the curve, over 18 months using the short form 36 (SF-36) bodily pain domain, with higher scores (range 0-100) indicating better quality of life. Other outcomes included costs to the NHS, quality adjusted life years (QALYs), and incremental cost effectiveness ratio. Results Of 2667 patients assessed for eligibility, 434 were randomised: 217 to the conservative management group and 217 to the laparoscopic cholecystectomy group. By 18 months, 54 (25%) participants in the conservative management arm and 146 (67%) in the cholecystectomy arm had received surgery. The mean SF-36 norm based bodily pain score was 49.4 (standard deviation 11.7) in the conservative management arm and 50.4 (11.6) in the cholecystectomy arm. The SF-36 bodily pain area under the curve up to 18 months did not differ (mean difference 0.0, 95% confidence interval −1.7 to 1.7; P=1.00). Conservative management was less costly (mean difference −£1033, (−$1334; −€1205), 95% credible interval −£1413 to −£632) and QALYs did not differ (mean difference −0.019, 95% credible interval −0.06 to 0.02). Conclusions In the short term (≤18 months), laparoscopic surgery is no more effective than conservative management for adults with uncomplicated symptomatic gallstone disease, and as such conservative management should be considered as an alternative to surgery. From an NHS perspective, conservative management may be cost effective for uncomplicated symptomatic gallstone disease. As costs, complications, and benefits will continue to be incurred in both groups beyond 18 months, future research should focus on longer term follow-up to establish effectiveness and lifetime cost effectiveness and to identify the cohort of patients who should be routinely offered surgery. Trial registration ISRCTN registry ISRCTN55215960.


Introduction
This statistical analysis plan (SAP) documents the analysis for the C-GALL trial, a randomised controlled trial comparing laparoscopic cholecystectomy with observation/conservative management for preventing recurrent symptoms and complications in adults with uncomplicated symptomatic gallstones.The SAP is based on the protocol version 11 and any deviations from the plan will be described.

Study Aims and Objectives
The primary aim of the study is to assess the clinical and cost-effectiveness of observation/conservative management compared with cholecystectomy for preventing recurrent symptoms and complications in adults presenting with uncomplicated symptomatic gallstones in a secondary care setting.
The primary patient objective is to compare medical management with cholecystectomy in terms of participants' quality of life using the SF-36 short-form health survey bodily pain domain at up to 18 months after randomisation.
The secondary objectives are to compare medical management with surgical treatment (cholecystectomy) in terms of condition specific quality of life; SF-36 domains (excluding bodily pain domain); complications; need for further treatment; persistent symptoms; health care resource use; costs.Secondary outcomes are at 18 and 24 months after randomisation.SF-36 short-form health survey bodily pain domain at up to 24 months after randomisation will also be reported.

General Study Design
A pragmatic, multi-centre parallel group patient randomised superiority trial (with internal pilot phase) to test if the strategy of standard cholecystectomy is more (cost-) effective than observation/conservative management at 18 months post-randomisation.
The null hypothesis being tested is that there is no difference between medical management and cholecystectomy.The alternative hypothesis is that cholecystectomy is superior.

Interventions to be evaluated
Laparoscopic cholecystectomy: this is the current standard surgical procedure for the management of symptomatic gallstone disease.The gall bladder is removed with the stones within it using keyhole techniques (laparoscopy).
Medical management: this involves the prescription of analgesics to relieve the biliary pain along with generic lifestyle advice.In the long-term, it may also include advice to eat a healthy diet with regular meals.

Randomisation, Allocation and Blinding
All participants who agree to enter the study will be logged with the central trial office and given a unique Study Number.Randomisation will utilise the existing proven remote automated computer randomisation application in the central trial office in the Centre for Healthcare Randomised Trials (CHaRT, a fully registered UK CRN clinical trials unit) in the Health Services Research Unit (HSRU), University of Aberdeen.
Participants are randomised 1:1 to receive either laparoscopic cholecystectomy or observation/conservative management.A random element will be incorporated into the randomisation algorithm.Treatment allocation is minimised by centre, gender and age (<35; 35-64; ≥ 65).
The trial statistician will not be blinded to treatment allocation.

Primary Outcome
The primary patient outcome measure will be quality of life as measured by area under the curve (AUC) up to 18 months post-randomisation using the SF-36 bodily pain domain (AUC measures at 3, 9, 12 and 18 months).The SF-36 1 is a generic measure obtained from a multi-purpose, short-form health survey with 36 questions.It has 36 items that is combined into 8 health domain scales (including Bodily Pain) and then further into two Summary measures.

Secondary Outcomes
The secondary outcome measures will include: Otago Gallstones Condition-Specific Questionnaire (CSQ) SF-36 health domains scales (excluding bodily pain domain)

Complications
Need for further treatment Persistent symptoms AUC up to 24 months post-randomisation for SF-36 bodily pain The CSQ 2 is a condition specific quality of life measure with four domains: Physical Functioning (pain, dyspepsia and diet changes), Systemic Functioning (fatigue), Social Functioning (daily duties, leisure, relationships) and Emotional Functioning (mood).It contains 12 items, each with a 5-point Likert response scale.
Complications will be defined as any intra-operative and post-operative complications.
Need for further treatment will be patient reported.
Persistent symptoms will be patient reported and will consist of two sections (pain and dyspepsia) of the CSQ.

Timing of Outcome Measurements
Baseline Surgery 1 3 months 9 months 12 months 18 months 6 monthly thereafter 1 Surgery (and post-surgery) CRF completed where appropriate.
CRF Case Report Form.
All times points are after randomisation.

Adverse Events
Each initial Adverse event (AE) will be considered for severity, causality or expectedness.A serious adverse event (SAE) is any AE that:

Results in death
Is life-threatening

Requires hospitalisation, or prolongation of existing inpatients' hospitalisation
Results in persistent or significant disability or incapacity Is a congenital anomaly or birth defect is otherwise considered medically significant by the investigator.
Please see the Study Protocol for more details on AEs.The number of Adverse events (AEs) and serious adverse events (SAEs) and the proportion of participants with an event will be presented.These will be tabulated and not analysed and will be summarised by Intention-to-Treat (ITT) as well as, as treated.

Sample Size and Power Calculation
The primary outcome is AUC of the SF-36 bodily pain domain up to 18 months post-randomisation.A study with 194 participants per group (388 in total) would have a 90% power at a 5% significance level to detect a difference of 0.33 standard deviation (SD).A total sample size of 430 participants to allow for 10% of complete missing outcome data.

General Methods
All the main analysis will be based on the ITT principle.The results of the trial will be presented following the standard CONSORT recommendations 3 .Baseline and follow-up data will be summarised using appropriate descriptive statistics and graphical summaries.Treatment effects will be presented with 95% confidence intervals (CIs) (apart from subgroup analysis where a stricter level will be used).There will be no adjustment to secondary outcomes CIs for multiple testing.All eligible participants who provided consent will be included in the analysis.Any post-randomisation exclusions will be removed.See Appendix (section 0) for how patient reported outcome measures (PROMs) are derived along with the sample Stata code for the analysis of the outcomes.

Interim Analysis
There are no planned interim analyses for efficacy or futility but an independent Data Monitoring Committee (DMC) will monitor trial progress and specifically any safety issues.Analysis will take place after full recruitment and follow-up.

Primary Outcome
The primary outcome, AUC SF-36 bodily pain domain up to 18 months will be analysed using linear regression with adjustment for the minimisation covariates (gender, age and including site as a random effect).The AUC for each participant will be generated by the trapezium rule.An SF-36 bodily pain profile will be presented graphically and in tabular form.If there is missing baseline data, then the centre specific baseline mean will be used.
Our primary analysis will include all participants that have at least one time point up to 30 months post-randomisation.For those participants with data missing at the 18 months time point, multiple imputation will be used to impute their score.Secondary analysis of the primary outcome will be performed on those participants that have an 18 months score.A sensitivity multiple imputation analysis will be performed using data only up to 18 months.

Secondary Outcomes
For the secondary outcome, SF-36 bodily pain AUC up to 24 months post-randomisation will be analysed in a similar way to the primary outcome.All other secondary outcomes will be analysed with generalised models appropriate for the distribution of the outcome with adjustment for the minimisation covariates (gender, age and including site as a random effect) up to 18 and 24 months.
SF-36 domains (excluding bodily pain), CSQ and persistent symptoms will be analysed using repeated measures including baseline score as an additional covariate.Treatment effects will be estimated from time-by-treatment interactions at each time point.P-values will only be presented at 18 and 24 months post-randomisation.Data collected after 24 months will be summarised descriptively, but no formal statistical test will be used.

Subgroup Analysis
Planned subgroup analyses are intended to explore potential effect modifications of gender (male, female), age (<35; 35-64; ≥ 65) and ethnicity using the UK census ethnic groupings (white; mixed/multiple ethnic groups; Asian/Asian British; Black/African/Caribbean/Black British; Other) or a similar grouping depending on the data.Subgroup by treatment interactions will be assessed by including interaction terms in the models outlined above.
A stricter level of statistical significance (2-sided 1% significance level) will be applied to these analyses given their exploratory nature.Corresponding 99% confidence intervals will therefore be calculated.

Compliance
We will explore the influence of compliance on the treatment effect for the primary outcome by doing a per-protocol analysis and if appropriate a complier adjusted causal estimation (CACE) will be explored.Compliance will be defined as participants who receive their allocated treatment within 24 months.For the laparoscopic cholecystectomy group, participants who receive emergency laparoscopic cholecystectomy will be defined as non-compliant.

Missing Outcome Data
The sensitivities of all treatment effect estimates to missing outcome data will be explored for the primary outcome only; these models will explore the robustness of the treatment estimates.We will follow the strategy outlines in White 4 .The analysis will use all available data that we believe are valid under the assumption of missing at random.We will then use a suite of sensitivity analysis to explore the robustness of the primary analysis to departures from assumptions, including all randomised participants.If required, sensitivity analyses will include but is not limited to multiple imputation, imputing a range of values for missing data under missing not at random assumptions as well as imputing data assuming a linear trend.

Missing Baseline Data
Data missing at baseline will reported as such.If required for models for primary and/or secondary outcomes continuous data will be imputed with the centre specific mean of that variable, missing binary/categorical data will include a missing indicator 5 .

Statistical software
All analysis will be carried out in Stata 16 6 (or the current version available).

COVID-19
The effect of COVD-19 will be explored.In the first instance, periods before and after COVID-19 will be summarised using appropriate descriptive statistics and graphical summaries.If need be, formal analysis will be carried out to explore the effect of COVID-19.The total score ranges from 0-100 with a higher score indicating higher symptom burden and therefore poor quality of life.If any question is not answered (i.e.there is missing data) then the score cannot be calculated.

Persistent symptoms
This is an unvalidated score and will be calculated as described for the CSQ measure but only using the relevant questions (see section 0) * The Otago gallstones condition-specific questionnaire Gender binary (0 for male and 1 for female) Medical binary (coded 0 for laparoscopic cholecystectomy and 1 for observation/conservative management) TimePoint is a categorical variable corresponds to the follow-up time points (coded 3 = 3 months, 9 = 9 months, 12 = 12 months, 18 = 18 months and 24 = 24 months) CentreNo categorical (corresponds to each recruiting centre) StudyNo is a unique participant identifier Complications binary (0 for no complications and 1 for had complication(s)) Medical binary (coded 0 for laparoscopic cholecystectomy and 1 for observation/conservative management) AgeBand categorical (<35 years; 35-64 years; ≥65 years) Gender binary (0 for male and 1 for female) CentreNo categorical (corresponds to each recruiting centre) irr is incidence rate ratio Treatment effect lincomest _b [1.Medical]   Need for further treatment mepoisson FurtherTreatment i.Medical i.AgeBand i.Gender || CentreNo:, irr vce(robust) FurtherTreatment binary (coded as 0 for no further treatment and 1 for had further treatment) Medical binary (coded 0 for laparoscopic cholecystectomy and 1 for observation/conservative management) AgeBand categorical (<35 years; 35-64 years; ≥65 years) Gender binary (0 for male and 1 for female) CentreNo categorical (corresponds to each recruiting centre) irr is incidence rate ratio