Types of studies

We included all randomised controlled trials (RCTs) (irrespective of sample size, language, or publication status) that compared (1) drain use versus no drain use, (2) open drain versus closed drain, or (3) different schedules for drain removal in participants undergoing appendectomy (irrespective of open or laparoscopic) for complicated appendicitis. We also included quasi‐RCTs (in which the allocation was performed on the basis of a pseudo‐random sequence, e.g. odd/even hospital number or date of birth, alternation), as described in Chapter 24 of the Cochrane Handbook for Systematic Reviews of Interventions (Reeves 2021).

Types of participants

We included all people (irrespective of age, sex, or race) who underwent emergency open or laparoscopic appendectomy for complicated appendicitis (irrespective of gangrenous appendicitis, perforated appendix without phlegmon or abscess, or perforated appendicitis with phlegmon or abscess), and receiving antibiotic regimens after open appendectomy.

Types of interventions

1. Use of drain (irrespective of type or material) versus no drain.

2. Open drain versus closed drain.

3. Early versus late drain removal (no more than two days versus more than two days).

Types of outcome measures

. 

Electronic searches

We searched the following electronic databases with no language or date of publication restrictions:

1. Cochrane Central Register of Controlled Trials (CENTRAL) (the Cochrane Library) (2020, Issue 2) (Appendix 1);

2. MEDLINE (Ovid) (1946 to 24 February 2020) (Appendix 2);

3. Embase (Ovid) (1974 to 24 February 2020) (Appendix 3);

4. Science Citation Index Expanded (Web of Science) (1900 to 24 February 2020) (Appendix 4);

5. World Health Organization International Trials Registry Platform (apps.who.int/trialsearch/) (24 February 2020);

6. US National Institutes of Health Ongoing Trials Register ClinicalTrials.gov (www.clinicaltrials.gov/) (24 February 2020);

7. Chinese Biomedical Literature Database (CBM) (1978 to 24 February 2020).

Searching other resources

We also searched the following databases on 24 February 2020:

1. Current Controlled Trials (www.controlled-trials.com/);

2. Chinese Clinical Trial Register (www.chictr.org/);

3. EU Clinical Trials Register (www.clinicaltrialsregister.eu/).

We also searched the reference lists and citations of identified studies and meeting abstracts via the Society of American Gastrointestinal and Endoscopic Surgeons (SAGES) (www.sages.org/) and Conference Proceedings Citation Index to explore further relevant clinical trials. We planned to communicate with the authors of included RCTs for additional information.

Selection of studies

After completing the searches, we merged the search results using the software package Endnote X7 (reference management software; Endnote X7) and removed any duplicate records. Two review authors (CY, CN) independently scanned the title and abstract of every record identified by the search for potential inclusion in the review. We retrieved the full text for further assessment if the inclusion criteria were unclear from the abstract. We included eligible studies irrespective of whether they reported the measured outcome data. We detected duplicate publications by identifying common authors, centres, details of the interventions, numbers of participants, and baseline data, according to Chapter 4 of the Cochrane Handbook for Systematic Reviews of Interventions (Lefebvre 2021). We intended to correspond with study authors to confirm whether the trial results had been duplicated, if necessary. We excluded papers that did not meet the inclusion criteria and listed the reasons for their exclusion. A third review author (DY) resolved any discrepancies between the two review authors by discussion.

Data extraction and management

We used a standard data collection form for study characteristics and outcome data that had been piloted on at least one study in the review. Two review authors (Li Zhe, Zhao L) extracted the following study characteristics from the included studies.

1. Methods: study design, total duration of study and run‐in period, number of study centres and location, study setting, withdrawals, date of study.

2. Participants: number of participants, mean age, age range, gender, severity of condition, diagnostic criteria, inclusion criteria, exclusion criteria.

3. Interventions: intervention, comparison.

4. Outcomes: primary and secondary outcomes specified and collected, and time points reported.

5. Notes: funding for trial, notable conflicts of interest of trial authors.

Two review authors (Li Zhe, Zhao L) independently extracted outcome data from the included studies. Any disagreements were resolved by consensus or by involving a third review author (DY). One review author (Li Zhe) copied across the data from the data collection form into Review Manager 5 (Review Manager 2020). We double‐checked that the data were entered correctly by comparing the study reports with presentation of data in the systematic review.

A second review author (Zhao L) cross‐checked study characteristics for accuracy against the trial reports.

Assessment of risk of bias in included studies

Two review authors (CY, CN) independently assessed risk of bias in the included trials using the Cochrane 'Risk of bias' tool as described in Chapter 8 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2017). We assessed risk of bias for the following domains.

1. Random sequence generation

2. Allocation concealment

3. Blinding of participants and personnel

4. Blinding of outcome assessment

5. Incomplete outcome data

6. Selective reporting bias

7. Other sources of bias (baseline imbalances)

We judged each domain as at low risk, high risk, or unclear risk of bias according to the criteria used in the Cochrane 'Risk of bias' tool (see Appendix 5) (Higgins 2017). We considered a trial to be at low risk of bias if the trial was assessed as at low risk of bias across all domains. Otherwise, we considered trials at unclear or high risk of bias regarding one or more domains as at high risk of bias overall. Any differences in opinion were resolved by discussion or by consulting a third review author (DY) to reach consensus if necessary.

The results of the 'Risk of bias' assessment are presented in a 'Risk of bias' graph and 'Risk of bias' summary, generated by Review Manager 5 (Review Manager 2020).

Measures of treatment effect

We performed the meta‐analyses using the software package Review Manager 5 (Review Manager 2020). For dichotomous outcomes, we calculated the risk ratio with 95% confidence interval (CI) (Deeks 2021). In the case of rare events (e.g. mortality), we calculated the Peto odds ratio (Deeks 2021). For continuous outcomes, we calculated the mean difference with 95% CI (Deeks 2021). For continuous outcomes with different measurement scales in different randomised clinical trials, we would calculate standardised mean differences with 95% CI.

Unit of analysis issues

The unit of analysis was the individual participant. We intended to analyse data using the generic inverse‐variance method in Review Manager 5 for cluster randomised trials (Higgins 2021). We intended to analyse only data from the first period of treatment for cross‐over trials (Higgins 2021). We intended to combine groups to create a single pair‐wise comparison for trials with multiple intervention groups. We did not encounter any cluster‐randomised trials, cross‐over trials, or trials with multiple intervention groups in this update.

Dealing with missing data

We contacted the original investigators to request further information in the case of missing data; however, we received no reply and thus used only the available data in the analyses.

Assessment of heterogeneity

We described heterogeneity in the data using the Chi² test (Deeks 2021). We considered a P value of less than 0.05 to be statistically significant heterogeneity (Deeks 2021). We also used the I² statistic to measure the quantity of heterogeneity. In case of statistical heterogeneity or clinical heterogeneity (or both), we performed meta‐analysis but interpreted the result cautiously and investigated potential sources of the heterogeneity. We explored clinical heterogeneity by comparing the characteristics of participants, interventions, controls, outcome measures, and study designs in the included studies. We planned to undertake the following approaches for explanation and solution:

1. check again that the data were correct;

2. change the effect measure;

3. perform analysis using the random‐effects model;

4. perform sensitivity analysis by excluding potentially biased trials;

5. perform subgroup analysis or meta‐regression;

6. present all trials and provide a narrative discussion.

Assessment of reporting biases

We planned to perform and examine a funnel plot to explore possible publication bias (Doleman 2020). However, as the number of included trials was less than 10, we did not produce any funnel plots, as recommended in Chapter 10 of the Cochrane Handbook for Systematic Reviews of Interventions (Sterne 2017).

Data synthesis

We performed the meta‐analyses using Review Manager 5 software (Review Manager 2020). For all analyses, we employed the random‐effects model for conservative estimation, except for the Peto odds ratio, which only has a fixed‐effect method.

Subgroup analysis and investigation of heterogeneity

We intended to perform the following subgroup analyses.

1. Trials at low risk of bias versus trials at high risk of bias.

2. Laparoscopic appendectomy versus open appendectomy.

3. Adults versus children.

4. Gangrenous appendicitis versus perforated appendicitis.

5. One type of appendix stump closure versus another.

We did not perform any of our planned subgroup analyses as there was an insufficient number of trials included in the review.

Sensitivity analysis

We performed the planned sensitivity analyses, as follows.

1. Changing statistics amongst risk ratios, risk differences, and odds ratios/Peto odds ratios for dichotomous outcomes.

2. Changing statistics between mean difference and standardised mean differences for continuous outcomes.

3. Excluding RCTs at high risk of bias.

We did not perform the planned sensitivity analysis by excluding non‐English literature because none of the studies were published in languages other than English. We performed a post hoc sensitivity analysis in response to peer‐reviewer for the primary outcome (intraperitoneal abscess) by excluding quasi‐RCTs to determine whether the conclusions were robust to the decisions made during the review process.