Types of studies

We included randomised controlled trials (RCTs). We did not include quasi‐randomised trials as these may have been subjected to increased bias.

Types of participants

Women of any age undergoing excision of the cervical transformation zone for cervical intraepithelial neoplasia (CIN), irrespective of surgical techniques.

Types of interventions

Prophylactic antibiotics (irrespective of regimens) versus placebo or no treatment.

Classification of antibiotics were as follows (Marjoribanks 2004).

1. Cephalosporins

2. Penicillins

3. Macrolides

4. Fluoroquinolones

5. Sulfonamides

6. Tetracyclines

7. Aminogylocosides

8. Glycopeptides

9. Antiprotozoals

10. Combination drugs

Antibiotic regimen includes the following.

1. Administration route (for example intravenous, intramuscular, oral, vaginal)

2. Number of doses (for example single versus multiple doses)

Types of outcome measures

Electronic searches

We searched the following electronic databases: Cochrane Central Register of Controlled Trials (CENTRAL; 2016, Issue 4) in the Cochrane Library, MEDLINE (1946 to May week 1, 2016), Embase (1980 to 2016, week 19), and LILACS (1970 to May, 2016).

Appendix 1, Appendix 2, Appendix 3, and Appendix 4 display the search strategies for CENTRAL, MEDLINE, Embase and LILACS, respectively.

Searching other resources

Unpublished and grey literature

We searched the following sources for ongoing trials.

1. ISRCTN registry, metaRegister of Controllled Trials (mRCT http://www.isrctn.com/page/mrct).

2. Physicians Data Query (https://www.cancer.gov/publications/pdq).

3. Clinical trials.gov http://www.clinicaltrials.gov.

4. International Clinical Trials Registry Platform (ICTRP) (http://www.who.int/ictrp/en/).

5. World Health Organization (WHO) International Clinical Trial Registry (http://apps.who.int/trialsearch/Default.aspx).

We searched electronic databases including Greynet.org (http:// www.greynet.org), the Ohio College Library Center (OCLC) WorldCat Dissertations and Theses (WorldCatDissertations; https://www.oclc.org/support/services/firstsearch/documentation/dbdetails/details/WorldCatDissertations.en.html) and Index to theses (ProQuest Dissertations & Theses: UK & Ireland) to identify the possible relevant conference abstracts and proceedings.

Handsearching

We handsearched reports of conferences from the following sources.

1. Gynecologic Oncology (Annual Meeting of the Society of Gynecologic Oncology)

2. International Journal of Gynecological Cancer (Annual Meeting of the International Gynecologic Cancer Society)

3. Annual Meeting of the European Society of Medical Oncology (ESMO)

4. Annual Meeting of the British Gynaecological Cancer Society (BGCS)

5. Biennial Meeting of the Asian Society of Gynecologic Oncology (ASGO)

6. Biennial Meeting of Asia and Oceania Federation of Obstetrics and Gynaecology (AOFOG)

7. Biennial Meeting of the European Society of Gynaecologic Oncology (ESGO)

We also checked the citation lists of the included studies and key textbooks for potentially relevant references. We searched for papers in all languages and would have had them translated, if necessary.

Selection of studies

We downloaded all titles and abstracts retrieved by electronic searching to Endnote, and duplicates were removed. Two review authors (CK and BC) independently examined the remaining references. We excluded those studies that did not meet the inclusion criteria and obtained copies of the full text of potentially relevant references. We (CK and BC) independently assessed the eligibility of retrieved papers. We resolved disagreements by discussion and, if there was no consensus, by appeal to a third review author (JT or PL). We documented the reasons for exclusion (see Excluded studies and Characteristics of excluded studies).

Data extraction and management

We (CK and BC) abstracted data independently onto a data abstraction form specifically designed for the review. We resolved differences between review authors by discussion or by appeal to a third reviewer (JT or PL), if necessary. The same two review authors independently assessed the quality of the trials and abstracted data using forms specifically designed for the review, according to Cochrane guidelines (Higgins 2011).

Assessment of risk of bias in included studies

We assessed and reported on the methodological risk of bias of the included studies according to Chapter 8 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011), which recommends the explicit reporting of the following individual elements for RCTs.

1. Selection bias: random sequence generation and allocation concealment.

2. Performance bias: blinding of participants and personnel (participants and treatment providers).

3. Detection bias: blinding of outcome assessment.

4. Attrition bias: incomplete outcome data.

5. Reporting bias: selective reporting of outcomes.

6. Other bias

Two review authors (CK, BC) applied the ’Risk of bias’ tool independently, and we resolved differences by discussion or by appeal to a third review author (JT or PL). We judged each item as being at high, low or unclear risk of bias as set out in the criteria shown in Appendix 5 (Higgins 2011). We provided a quote from the study report or a statement as justification for our judgement. We summarised results in a ’Risk of bias’ summary figure.

Measures of treatment effect

We used the following measures of the effect of treatment.

1. For dichotomous data (e.g. adverse events or infections), we extracted the number of participants in each treatment arm who experienced the outcome of interest and the number of participants assessed at endpoint, in order to estimate a risk ratio (RR) with 95% confidence intervals (CIs).

2. For continuous data, we planned to use the mean difference (MD) or standardized mean difference (SMD) with 95% CIs.

Where possible RRs of individual studies were combined for meta‐analysis using RevMan 2014 software.

Dealing with missing data

We did not impute missing outcome data for the primary outcome and we attempted to contact study authors to obtain missing data. If data were missing to the extent that we could not have included the study in the analysis, we had planned to present the results in a narrative way.

Assessment of heterogeneity

We assessed heterogeneity using visual inspection of the forest plots. We also assessed statistical heterogeneity in each meta‐analysis using the I² statistic and Chi² test. We regarded heterogeneity as substantial if the I² statistic value was greater than 50%, or there was a low P value (< 0.10) in the Chi² test for heterogeneity (Deeks 2001; Higgins 2011). If there was substantial statistical heterogeneity, we planned to carry out subgroup analyses to assess differences between the included studies. However, if there had been clinical, methodological or considerable statistical heterogeneity (I² greater than 75%) across included studies, we had planned to use a narrative approach to data synthesis.

Assessment of reporting biases

We had planned to examine funnel plots corresponding to meta‐analysis of the primary outcome to assess the potential for small‐study effects, such as publication bias, if more than 10 studies were included. If asymmetry had been suggested by a visual assessment, we had planned to perform sensitivity analysis to investigate whether it affected the pooled results. (Sterne 2011).

Data synthesis

Where feasible, the results were pooled in a meta‐analyses. We used the random‐effects model with inverse variance weighting for all meta‐analyses (DerSimonian 1986). We performed statistical analysis using RevMan 2014.

1. For dichotomous outcomes, we calculated the RR for each study as well as for the pooled outcome.

2. For continuous outcomes, we calculated and pooled the mean difference between the treatment arms if all trials measured the outcome on the same scale; otherwise we pooled the standardised mean differences.

We prepared a ’Summary of findings’ table to present the results of the meta‐analysis, based on the methods described in Chapter 11 of the Cochrane Handbook for Systematic Reviews of Interventions (Schunemann 2011). For assessments of the overall certainty of evidence for each outcome that includes pooled data from RCTs only, we downgraded the evidence from 'high certainty' by one level for serious (or by two for very serious) study limitations (risk of bias), indirectness of evidence, serious inconsistency, imprecision of effect estimates or potential publication bias. The following outcomes were included in the 'Summary of findings' table:

1. prolonged vaginal discharge;

2. postoperative bleeding;

3. fever;

4. abdominal pain;

5. adverse effect related to antibiotic;

6. additional medical support;

7. additional self‐medication.

We presented the results of the meta‐analysis for the relevant outcomes and adverse events as outlined in the ’Types of outcome measures’ section.

Subgroup analysis and investigation of heterogeneity

As we found only a few trials, we did not perform any subgroup analyses. However, we considered factors such as type of excisional techniques (e.g. CKC, LC, or LLETZ/LEEP) and antibiotic regimen prescribed in the interpretation of review findings. In future updates, we will perform subgroup analysis based on these factors, if feasible.

Sensitivity analysis

As we found only three trials, we did not perform any sensitivity analysis. In future updates, if statistical heterogeneity is detected and there are sufficient trials included (greater than 10), we will conduct sensitivity analyses to examine the possible contribution of other clinical or methodological differences between the trials, specifically:

1. trials with adequate methodology versus those with poor methodology;

2. trials which seem to differ from the others in their clinical criteria for defining genital infections.