Types of studies

We included all randomized controlled trials, regardless of sample size, publication status, language, or publication date, that compared fibrin sealant with control for the prevention of postoperative pancreatic fistula (POPF) in people undergoing pancreatic resections. We excluded quasi‐randomized trials, 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), and non‐randomized studies because of the potential for bias (Reeves 2011).

Types of participants

We included people, regardless of age, sex, or race, who underwent elective pancreatic resections (open or laparoscopic) for any pancreatic or extra‐pancreatic disease.

We excluded people who underwent total pancreatectomy as this eliminates any source for a POPF.

Types of interventions

Intervention: fibrin sealant, regardless of glue or patch, which was applied to stump closure reinforcement, pancreatic anastomosis reinforcement, or main pancreatic duct occlusion.

Comparison: placebo or no fibrin sealant treatment.

Types of outcome measures

Electronic searches

For the initial version of this review, we searched the Cochrane Central Register of Controlled Trials (CENTRAL) (2015, Issue 7), MEDLINE (1946 to 26 August 2015), Embase (1980 to 26 August 2015), Science Citation Index Expanded (1900 to 26 August 2015), and Chinese Biomedical Literature Database (CBM) (1978 to 26 August 2015) (New Reference). For this updated review, we searched the following electronic databases from 2015 to 12 April 2018, with no restrictions on language or date of publication:

1. CENTRAL (Appendix 2);

2. MEDLINE (Ovid) (Appendix 3);

3. Embase (Ovid) (Appendix 4);

4. Science Citation Index Expanded (Web of Science) (Appendix 5); and

5. Chinese Biomedical Literature Database (CBM) (Appendix 6).

Searching other resources

We checked reference lists of all primary studies and review articles for additional references. We contacted authors of identified studies and asked them to identify other published and unpublished studies.

We searched PubMed for errata or retractions from eligible studies and reported the date this was done within the review (www.ncbi.nlm.nih.gov/pubmed). We also searched the meeting abstracts via the HPB journal (onlinelibrary.wiley.com/journal/10.1111/(ISSN)1477‐2574; accessed 12 April 2018) and Conference Proceedings Citation Index to explore further relevant clinical studies.

Selection of studies

Two review authors (Junhua Gong, YC) independently screened the titles and abstracts of all the studies we identified as a result of the search and coded them as 'retrieve' (eligible or potentially eligible/unclear) or 'do not retrieve'. We retrieved the full‐text study reports/publications and two review authors (Junhua Gong, YC) independently screened the full text and identified studies for inclusion, and identified and recorded reasons for exclusion of the ineligible studies. We resolved any disagreements through discussion or, if required, we consulted a third review author (ZZ). We identified and excluded duplicates and collated multiple reports of the same study, so that each study, rather than each report, was the unit of interest in the review. We recorded the selection process in sufficient detail to complete a PRISMA flow diagram and a Characteristics of excluded studies table (Moher 2009).

Data extraction and management

We used a standard data collection form for study characteristics and outcome data, which had been piloted on at least one study in the review. Two review authors (NC, Jianping Gong) extracted the following study characteristics from included studies.

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

2. Participants: number (N), mean age, age range, gender, severity of condition, diagnostic criteria, inclusion criteria, and exclusion criteria.

3. Interventions: intervention, comparison.

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

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

Two review authors (NC, Jianping Gong) independently extracted outcome data from included studies. We noted in the Characteristics of included studies table if outcome data were not reported in a usable way. We resolved disagreements by consensus or by involving a third review author (ZZ). One review author (YC) copied across the data from the data collection form into Review Manager 5. We double checked that the data were entered correctly by comparing the study reports with how the data were presented in the systematic review. A second review author spot‐checked study characteristics for accuracy against the study report.

Assessment of risk of bias in included studies

Two review authors (Junhua Gong, YC) independently assessed risk of bias for each study using the criteria outlined in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011b). We resolved any disagreements by discussion or by involving a third review author (ZZ). We assessed the 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 outcome reporting;

7. other bias.

We graded each potential source of bias as high, low, or unclear risk, and provided a quote from the study report together with a justification for our judgement in the 'Risk of bias' table. We summarized the 'Risk of bias' judgements across different studies for each of the domains listed. We considered blinding separately for different key outcomes where necessary (e.g. for unblinded outcome assessment, risk of bias for all‐cause mortality may be very different than for a participant‐reported pain scale). Where information on risk of bias related to unpublished data or correspondence with a trialist, we noted this in the 'Risk of bias' table.

When considering treatment effects, we took into account the risk of bias for the studies that contributed to that outcome.

Measures of treatment effect

We analysed dichotomous data as risk ratio (RR) and continuous data as mean difference (MD) with 95% confidence intervals (CIs). In the case of rare events (e.g. mortality), we calculated the Peto odds ratio (Peto OR) (Deeks 2011). We ensured that higher scores for continuous outcomes had the same meaning for the particular outcome, explained the direction to the reader and reported where the directions were reversed if this was necessary.

We undertook meta‐analyses only where this was meaningful, that is, if the treatments, participants, and underlying clinical question were similar enough for pooling to make sense.

A common way that trialists indicated when they had skewed data was by reporting medians and interquartile ranges. When we encountered this, we noted that the data were skewed and considered the implication of this.

Where multiple study arms were reported in a single study, we included only the relevant arms. If two comparisons (e.g. drug A versus placebo and drug B versus placebo) were entered into the same meta‐analysis, we halved the control group to avoid double counting.

Unit of analysis issues

The unit of analysis was the individual participant. We did not find any cross‐over or cluster‐randomized trials.

Dealing with missing data

We contacted investigators or study sponsors in order to verify key study characteristics, and obtained missing numerical outcome data where possible (e.g. when a study was identified as abstract only). However, there was no reply in most of the cases. Thus, we used only the available data in the analyses.

Assessment of heterogeneity

We used the I² statistic to measure heterogeneity among the studies in each analysis (Higgins 2003). When we identified substantial heterogeneity (greater than 50%), we explored it by prespecified subgroup analysis, and we interpreted summary effect measures with caution.

Assessment of reporting biases

We did not perform funnel plots to assess reporting biases because the number of studies included under each comparison was less than 10 (Sterne 2011).

Data synthesis

We performed the meta‐analyses using Review Manager 5 software (RevMan 2014). For all analyses, we employed the random‐effects model for conservative estimation, except for the Peto odds ratio (OR) which only has a fixed method.

Subgroup analysis and investigation of heterogeneity

We intended to perform the following subgroup analyses.

1. Randomized controlled trials with low risk of bias versus randomized controlled trials with high risk of bias

2. The type of fibrin sealants (glue and patch)

3. Different etiologies (pancreatic cancer, chronic pancreatitis, and others)

4. High‐risk people (e.g. fatty pancreas, soft pancreas, small pancreatic duct) versus low‐risk people

In previous versions of this review, we performed a meta‐analysis of fibrin sealants versus no fibrin sealants for overall pancreatic surgery, with subgroups of sealing location and type of operation. We have restructured this version of the review and the changes are detailed in the Differences between protocol and review section.

Sensitivity analysis

We performed sensitivity analyses to determine whether the conclusions were robust according to the decisions made during the review process, as follows.

1. Changing between a fixed‐effect model and a random‐effects model

2. Changing statistics between risk ratios (RR), risk differences (RD), and odds ratios (OR) for dichotomous outcomes

3. Changing statistics between mean difference (MD) and standardized mean differences (SMD) for continuous outcomes

4. Excluding randomized controlled trials with low quality

5. Excluding studies published in languages other than English

If the results did not change, they were considered to have low sensitivity. If the results changed, they were considered to have high sensitivity.