Preoperative combined mechanical and oral antibiotic bowel preparation for preventing complications in elective colorectal surgery

Objectives

This is a protocol for a Cochrane Review (intervention). The objectives are as follows:

The aim of this review is to assess the evidence for the use of combined mechanical and oral antibiotic bowel preparation for preventing complications in elective colorectal surgery.

Background

Description of the condition

Colorectal operations are amongst the most frequently performed surgical procedures worldwide. In addition to emergency surgery (e.g. bowel perforations, diverticulitis, lower gastrointestinal bleeding), colon resections are performed for treating inflammatory diseases (e.g. ulcerative colitis or Crohn's disease) and colorectal cancer (Kuhry 2008; Spanjersberg 2011).

The treatment outcome is significantly influenced by the surgical procedure itself along with the occurrence of postoperative complications. Amongst the most serious complications are anastomotic leaks and surgical site infections (SSIs), which can lead to a prolonged and severe clinical course with impaired long‐term outcome. In oncological cases, prolonged postoperative recovery can lead to a delayed start of adjuvant oncological treatment and thus to an increased risk of metastasis or local recurrence (Young 2012; Kulu 2015; Nachiappan 2016; Beck 2020). The incidence of complications is, amongst other factors, related to the type of surgery performed. Whilst anastomotic leaks are infrequent for small bowel and elective colon resections (1% to 3%), they occur far more frequently after rectal resections (10% to 23%) (Walker 2004; Weidenhagen 2007; Kulu 2015; ISOS 2016; Toh 2018). Compared to other abdominal operations, colorectal resections present an eight‐fold higher risk for the occurrence of adverse events, in particular the rate of SSIs, which is estimated to be between 9% and 24%, whereas it is reported as 2% to 9% in non‐colorectal surgery (Anjum 2017; Migaly 2019). One reason for the increased complication rate in colorectal surgery is the high bacterial colonisation by the physiological intestinal flora, the so‐called microbiome.

Overall, there are numerous recommendations to reduce the risk of surgical complications. These range from specified preoperative preparation of the surgical field (e.g. preoperative whole‐body bathing or showering, hair removal, and disinfection methods) to intravenous antibiotic prophylaxis (Nelson 2014; WHO 2018; Ling 2019; NICE 2019). However, there is disagreement about what might be the best strategy for bowel preparation before elective colon and rectal surgery. 

Description of the intervention

The underlying idea of preoperative bowel preparation is to clean the bowel from faeces, thereby reducing the bacterial load, which could lead to a lower rate of postoperative complications, especially SSIs and anastomotic leakage. Furthermore, removing the faeces makes it easier to manipulate the bowel in laparoscopic surgeries and lowers the risk of unwanted faecal spillage into the abdominal cavity.

In summary, there are four widespread interventions for bowel preparation before colorectal surgery in everyday clinical practice:

• no bowel preparation;

• mechanic bowel preparation (MBP);

• oral antibiotics (oAB);

• combination of oral antibiotics and mechanical bowel preparation (MBP + oAB).

For MBP, osmotically active solutions are mainly used nowadays (Kumar 2013; WHO 2018). By translocating fluid into the intestine, these lead to the development of diarrhoea and as a result the emptying of the bowel. Given the constant development of the applied solutions in the last decade, complications such as dehydration, electrolyte imbalance, or cardiovascular dysfunction rarely occur (Kumar 2013).

Regarding the antibiotic regimen, a combination of active substances that are effective against both aerobic and anaerobic bacteria are used (WHO 2018; Migaly 2019). Numerous different protocols exist regarding the active substance, dosage, and duration of treatment (WHO 2018). Two meta‐analyses from 2018 recommend a combination of an aminoglycoside (kanamycin or neomycin) and metronidazole or erythromycin administered one to three days before surgery (McSorley 2018; Toh 2018).

There is some evidence that luminal faeces may lead to an inactivation of the topically acting antibiotics. Consequently, oAB should be administered after or at least in combination with MBP (Schardey 2017). However, the evidence for this hypothesis is ambiguous. Recent meta‐analyses failed to demonstrate a significant difference in SSI or anastomotic leak rates after combined or sole oral antibiotic bowel preparation, calling into question the need for combination therapy (Rollins 2019; Nelson 2020). 

How the intervention might work

It is well known that the intestinal microbiome is important for myriads of physiological processes like metabolism of drugs and nutrition degradation, biosynthesis of neurotransmitters and hormones, and influences immune maturation, host cell proliferation, and neurological signalling – to name but a few examples. However, the microbiome also plays an important role in disease development, for example autoimmune and gastrointestinal disease, but also neuropsychiatric illnesses (Lynch 2016). There is growing evidence that the intestinal microbiome is also involved in wound‐healing processes, especially in the healing of bowel anastomosis or the development of an anastomotic leakage (Schardey 2017). In addition to the surgical technique, bacterial colonisation of the intestinal mucosa of the anastomotic region also influences the occurrence of an insufficiency. Due to the surgical trauma and resulting ischaemia, mucosal bacteria such as Enterococcus faecalis or Pseudomonas aeruginosa develop the ability to express collagenases and activate matrix metalloproteinase 9 (MMP9) in the patient's intestinal tissue. This mechanism promotes the degradation of synthesised tissue leading to vulnerability of the newly created anastomosis (Shogan 2015; Anjum 2017; Schardey 2017).

In order to prevent these wound‐healing disturbances, preoperative preparation of the bowel is intended to create a clean working environment by reducing bacterial contamination of the intestine and respectively of the surgical field.

Why it is important to do this review

In several publications, as well as a Cochrane Review published in 2011, the use of MBP compared to no preparation or rectal enemas did not demonstrate an improved outcome for patients, which led to the recommendation to refrain from preoperative bowel preparation (Güenaga 2011).

In contrast, a current large registry study with more than 8000 patients demonstrated a significantly lower rate of postoperative SSIs as well as a shorter length of hospital stay with combined therapy of oAB and MBP compared to no bowel preparation or monotherapy with MBP or oAB. In addition, the combination therapy group also had the lowest readmission rate (Klinger 2019). Based on these data, the American Society of Colon and Rectal Surgeons recommended a combined mechanical and oral antibiotic bowel preparation prior to elective colon and rectal resections (Migaly 2019).

Despite the evidence for a beneficial effect of supplementing oAB to MBP, a current survey amongst members of the German Society of General and Visceral Surgery revealed that MBP alone is performed in over 50% of colon and over 75% of rectal operations. Additional oAB was only performed in about 10% of these operations (Buia 2019). A comparable survey amongst members of the European Society of Coloproctology revealed similar results. Whilst the majority of respondents reported to regularly use MBP, less than 10% prescribe oral antibiotic therapy (Devane 2017). In the United States, the rate of usage of combination therapies is much higher, although it has decreased over the last few years. In a survey conducted by the American Society of Colon and Rectal Surgeons in 1990, 88% of the participants reported using a combined bowel preparation (Solla 1990). In a more recent survey from 2010, only 36% of the colorectal surgeons still prescribe oral antibiotic therapy (Markell 2010).

The results of these surveys reveal that there is currently no uniform approach. Although there are already several meta‐analyses on this topic, they differ considerably in their conclusions: whilst some meta‐analyses report a benefit of combined oral and intravenous antibiotic therapy, with an unclear effect of concurrent mechanical bowel preparation (Rollins 2019; Nelson 2020), other meta‐analyses have shown superiority of the combination of oral antibiotic prophylaxis and mechanical bowel preparation (Toh 2018). However, these differences are not due to more recent findings, but rather to differences in the literature search, study selection criteria, and data extraction management. In order to establish an evidence‐based therapy, a structured and high‐quality meta‐analysis of the available evidence is necessary to provide optimal guidance on preoperative bowel preparation aiming to reduce the postoperative complication rate as well as overall mortality.

Objectives

The aim of this review is to assess the evidence for the use of combined mechanical and oral antibiotic bowel preparation for preventing complications in elective colorectal surgery.

Methods

Criteria for considering studies for this review

Types of studies

We will include all published, unpublished, and ongoing randomised controlled trials (RCTs) or quasi‐RCTs (trials in which randomisation is attempted but potentially predictable, such as allocating participants by day of the week or sequence of entry into trial) comparing preoperative bowel preparation using oAB and MBP prior to elective colon or rectal surgery. We will consider all identified studies for inclusion regardless of date, location, or language of publication. 

Types of participants

Adult participants (18 years of age and older) undergoing elective colorectal surgery. 

There are no exclusion criteria regarding the indication for surgery (both benign and malignant diseases are eligible); the type of colorectal surgery performed ((extended) right/left hemicolectomy, transverse colectomy, sigmoid resection, rectal resections, proctocolectomies or reanastomoses (provided a colorectal anastomosis is performed)); any previous treatment of the patient (e.g. neoadjuvant therapy); patient comorbidities; or the timing and location of the surgery.

For studies that include both eligible participants and others (e.g. children), we will attempt to extract data only from the eligible participants. We will request this information from the study authors if necessary. If it is not possible to separate eligible data, the study will be excluded from statistical analysis and only described narratively.

Types of interventions

We will include any combination therapy of preoperative oAB and MBP. 

We anticipate that there will be different protocols in terms of the timing, duration, and frequency of administration, as well as the dosage of substances used (both in terms of antibiotics and mechanical bowel preparation). Any type of preoperative oral antibiotic prophylaxis, as well as any method of mechanical bowel cleansing, is eligible regardless of the mode of delivery, dose, duration, intensity, and co‐interventions. However, combination therapy and parenteral antibiotic prophylaxis (in both the control and intervention group) is mandatory. Regarding the effect of MBP or oAB alone, analyses already exist in further Cochrane Reviews (Güenaga 2011; Nelson 2014). We will include studies using standard of care regarding parenteral antibiotic (AB) prophylaxis in both participant groups (intervention and control).

We will include the following comparisons:

• combination of oAB and MBP versus oAB alone;

• combination of oAB and MBP versus MBP alone;

• combination of oAB and MBP versus no preoperative preparation (or placebo control).

Types of outcome measures

We will analyse the occurrence of the predefined primary and secondary outcome parameters listed below within 30 days postoperatively. We will assess only the incidence of complications, and not the time of their occurrence. If a study provides data at different time points, we will use only the latest date in the time frame for synthesis. If a study also reports follow‐up data after day 30, we will also consider these data and report them narratively. In addition, we will investigate the impact of these events on the duration of postoperative inpatient therapy. The outcome parameters listed below can be differentiated into efficacy and safety outcomes, whereby all listed outcome parameters, with the exception of length of hospital stay, refer to patient safety.

Reporting one or more of the listed outcomes is not a study inclusion criterion of the review. If we include a study that does not address any of the mentioned outcomes, we will contact the authors and request data on the outcomes of interest. If data on these outcomes are not available, we will report the outcomes of that study narratively.

Primary outcomes

If an outcome occurs in a participant at multiple sites (e.g. wound infection) or at different time points during the 30‐day postoperative observation interval, we will count such an outcome measure only once.

We will consider the following primary outcomes:

• number of participants with SSIs (infection of the incision site involving the skin and subcutaneous tissue (superficial), deep soft tissue (deep), or a part of the body deeper than the fascia/muscle layers that was opened or manipulated during the surgical procedure (organ/space) and occurs within 30 days after the surgical procedure);

• number of participants with anastomotic leakage.

For the above outcome parameters, an absolute risk reduction of 5% can be considered clinically relevant. Based on the literature risk for SSIs of 9% to 24% (median 16.5%) (Anjum 2017), a reduction to 4% to 19% (median 11.5%) would have to be achieved through the intervention. For anastomotic leaks, the intervention would need to reduce the risk from 1% to 23% (median 11%) to 0% to 18% (median 9%) (Kulu 2015).

The use of the median of the desired risk values to calculate the risk ratio (RR) results in a value of about 0.7 for SSIs and 0.8 for anastomotic leaks. This means that the RR and 95% confidence interval (CI) should be equal to or smaller than 0.7 for SSIs and equal to or smaller than 0.8 for anastomotic leaks to be considered a clinically relevant difference.

Secondary outcomes

We will record the number of participants with the following adverse events (as defined by the primary study author) within 30 days postoperatively.

• Mortality.

• Postoperative complications (Clavien‐Dindo grade I and II).

• Severe postoperative complications (Clavien‐Dindo grade III and IV).

• Incidence of postoperative ileus.

• Length of hospital stay (LOS) (days).

• Treatment‐related adverse events:

  • incidence of adverse effects of MBP such as dehydration, electrolyte imbalance, renal failure, or cardiac dysfunction (as defined by the primary study author);

  • incidence of adverse effects of oAB such as diarrhoea or pseudomembranous enterocolitis (as defined by the primary study author);

  • number of participants for whom the intervention was discontinued due to side effects, as well as the number of participants for whom therapy was initiated to treat the complications.

Our prespecified outcome measures are not independent of each other. For example, people with SSIs, anastomotic leakages, or other adverse events are additionally classified according to the severity of the complications that occurred using the Clavien‐Dindo classification (Clavien 2009).

Search methods for identification of studies

Electronic searches

We will search the following electronic databases from inception to present:

• Cochrane Central Register of Controlled Trials (CENTRAL);

• MEDLINE (via PubMed);

• Embase.

The search strategy for MEDLINE is given in Appendix 1. We will adapt the query accordingly for searches in other databases.

Searching other resources

We will also search trial registries such as ClinicalTrials.gov (www.clinicaltrials.gov) and the World Health Organization (WHO) International Clinical Trials Registry Platform (ICTRP) search portal (www.who.int/clinical-trials-registry-platform). Furthermore, we will search the reference lists of all relevant publications identified for further studies. We will contact organisations (e.g. regional colorectal surgery societies) to ask if they have knowledge of ongoing or completed studies to complement our database searches. We will seek both published and unpublished trials. We will not limit the search by language or date.

Data collection and analysis

Selection of studies

Two review authors (MW and TV) will screen the title and abstract of each study identified in the electronic search for potential relevance. Each review author will independently decide on trial inclusion using the following predetermined eligibility criteria.

• Study design: RCT or quasi‐RCT.

• Population: adults undergoing elective colorectal surgery.

• Intervention: preoperative bowel preparation using oAB and MBP.

• Comparison: no preoperative bowel preparation or sole treatment with either MBP or oAB.

Any disagreements regarding study eligibility during title and abstract screening will be resolved by a third review author or by contacting the study authors for clarification. 

Afterwards, two review authors (MW and TV) will independently carry out full‐text screening using the same inclusion criteria.

Data extraction and management

Two review authors (MW and TV) will independently abstract the following information from the included studies using a standardised form. The data extraction form will be piloted on two studies and adjusted if necessary.

• General information:

  • study ID, study title, corresponding  author and contact details, publication date, country where the study was conducted, aim of study (short description), any conflicts of interest stated by the investigators, source of funding.

• Number of participants:

  • total/intervention/control/2. intervention (if existing).

• Population:

  • age, gender, comorbidities, surgery indication, type of operation, subgroups reported, subgroups measure.

• Intervention:

  • description of the intervention, agent and dose used for MBP, agent and dose used for oAB, timing of preparation (MBP), timing of preparation (oAB), modification of intervention, co‐interventions.

• Control:

  • description of control, agent and dose used, timing of preparation.

• Outcomes:

  • incidence of wound infections, incidence of peritonitis, incidence of anastomotic leakage, spectrum of organisms detected, incidence of reoperation, incidence of pneumonia, incidence of cystitis, incidence of thrombosis, incidence of another infectious complication, incidence and duration (days) of postoperative ileus according to the definition of the primary study author, LOS (days), mortality, side effects of the intervention (e.g. electrolyte imbalances, renal failure, incidence of Clostridium difficile infection, termination of the intervention due to side effects).

If the majority of the required data are not reported in an identified publication, we will search for the associated study report or attempt to obtain the required data through correspondence with the primary study author. We will include multiple publications of one study only once, but will use all versions to collect as complete information as possible. Data extraction will be based on the main primary publication of the study, but secondary publications will also be considered for additional information. All publications will be listed in the references of included studies. If data from multiple publications are inconsistent, we will contact the primary study author for clarification.

Any disagreements will be resolved by discussion or by consultation with a third review author if required.

Assessment of risk of bias in included studies

We will assess risk of bias in the included studies using Cochrane's RoB 2 tool (Sterne 2019; Higgins 2021). Two review authors (MW and TV) will independently assess risk of bias of the results, with any disagreements resolved through discussion with a third review author.

We are interested in the effect of allocation at baseline, regardless of whether the intervention was delivered as intended (i.e. the 'intention‐to‐treat' effect). Consistent with the RoB 2 tool (Sterne 2019; Higgins 2021), we will consider the following five domains in our assessment:

• risk of bias arising from the randomisation process;

• risk of bias due to deviations from the intended interventions (effect of assignment as well as adhering to intervention);

• missing outcome data;

• risk of bias in measurement of the outcome;

• risk of bias in selection of the reported result.

We will assess risk of bias for the following five outcomes that also contribute to the summary of findings tables of the review, measured at a time point closest to the 30‐day postoperative window (see 'Summary of findings and assessment of the certainty of the evidence' section below): 

• SSIs;

• anastomotic leakage;

• mortality;

• incidence of postoperative ileus;

• LOS.

We will assess these domains using the 'Excel tool to implement RoB 2' (available at www.riskofbias.info/welcome/rob-2-0-tool/current-version-of-rob-2), and follow the recommended algorithm of signalling questions and response options to reach one of the following risk of bias judgements for each of the five domains, and consequently for each of our prespecified outcome measures: 

• low risk;

• some concerns;

• high risk.

To assess the overall risk of bias of an outcome, we will consider judgements of the five individual domains. To be judged as 'low risk', all domains must be rated as at low risk of bias. We will assess an outcome as 'some concerns' if risk of bias has been rated as some concerns in at least one domain, and high risk of bias in no domains. We will assess an outcome as 'high risk' if high risk of bias is identified in even one domain. We will also classify an outcome as 'high risk' if we judge several domains as 'some concerns', as we consider confidence in such an outcome to be considerably reduced (Higgins 2021).

Measures of treatment effect

For continuous outcomes (e.g. LOS), we will extract the final mean value and standard deviation (SD) of each outcome of interest as well as the number of participants evaluated at the final assessment in each treatment arm. With these data, we will calculate the mean difference (MD) and the 95% CI and, if appropriate, a pooled estimate of treatment effects. If MDs and SDs are reported by the primary study authors, those data will be used. If studies report data for an outcome on different scales, we will calculate the standardised mean difference (SMD) to allow the results to be compared.

For dichotomous outcomes (SSIs, anastomotic leakage, mortality, postoperative complications, incidence of postoperative ileus, treatment‐related adverse effects), we will extract the RR including the 95% CI. If the RR is not reported, the number of affected participants will be extracted to estimate an RR and its 95% CI. 

Unit of analysis issues

We expect neither cross‐over studies nor cluster‐randomised studies in this clinical area, therefore this protocol does not specify how such studies should be handled. If such a study is unexpectedly identified, the methodology for further action will be determined subsequently. Our review will include one intervention group (oAB + MBP) and three comparison groups (comparison 1: oAB alone; comparison 2: MBP alone; comparison 3: no preoperative bowel preparation). As there are no restrictions on the methodology of the intervention, no further subdivision of the intervention or comparison groups is intended.

If we identify studies with multiple groups, we will attempt to assign all relevant study groups to our predefined intervention or comparison group, possibly by combining multiple subgroups.

An example of a possible study with multiple groups could be: Group I (MBP) versus Group II (oAB1) versus Group III (oAB2) versus Group IV (MBP + oAB1) versus Group V (MBP + oAB2).

In such a case, Group I would be assigned to comparison 2, Groups II and III to comparison 1, and Groups IV and V to the intervention.

Dealing with missing data

For relevant data missing from a trial report, we will attempt to contact the corresponding author to obtain the missing information. As missing data could lead to a relevant bias, we will examine the impact of including such studies in the overall evaluation of the results through a sensitivity analysis, in which all studies with more than 20% missing data are excluded. We will not impute missing information.

Assessment of heterogeneity

Clinical heterogeneity may be caused by several factors, such as patient age and comorbidities, indication, and treatment prior to surgery, as well as the procedures used to decontaminate the bowel. In addition, different surgical procedures, as well as the type of access route (minimally invasive versus open surgery), also contribute to heterogeneity. Methodological heterogeneity may be caused by the different risks of bias between studies. To assess the impact of clinical and methodological heterogeneity, we will perform subgroup and sensitivity analyses on these topics. We will identify statistical heterogeneity through visualisation of the forest plots as well as the Chi² test. We will use the I² statistic for the quantification of heterogeneity.

We will interpret the I² value as follows (Deeks 2021).

• < 30% to 40%: little or no heterogeneity.

• 41% to 74%: moderate heterogeneity.

• 75% to 100%: considerable heterogeneity.

Possible reasons for considerable heterogeneity will be investigated and reported.

Assessment of reporting biases

To assess whether selective reporting of outcomes might affect the review findings, we will match study protocols or registry entries, if available, with the published information on outcomes.
If we include at least 10 trials per outcome, we will use funnel plots to assess the potential for small‐study effects.

Data synthesis

Before conducting a meta‐analysis, we will check whether the participants, interventions, comparisons, and outcomes are sufficiently similar to ensure that the result of the analysis is clinically meaningful.

To calculate an overall treatment effect, we will combine the data using a random‐effects model. As different agents of MBP and a variety of oAB with different administration intervals and combination options are included, a variance between the included trials must be assumed. 

We will include all eligible studies in the meta‐analysis regardless of their risk of bias rating. In a second step, we will evaluate the influence of studies judged at 'some concerns' or 'high risk' in a sensitivity analysis. We will perform the statistical analysis with Review Manager 5 software (Review Manager 2020). To combine dichotomous outcome data, we will use the method proposed by Mantel‐Haenszel (Deeks 2021). If meta‐analysis is not appropriate due to an insufficient number of eligible studies or substantial heterogeneity between studies, we will provide a narrative description of study characteristics and results.

Subgroup analysis and investigation of heterogeneity

Provided that sufficient studies are identified to justify subgroup analyses or meta‐regressions (at least 10 studies per outcome), we will investigate the possible causes of heterogeneity by means of subgroup analyses. 

We plan to carry out subgroup analyses for the following factors.

• (Extended) right‐sided versus (extended) left‐sided versus rectal resections.

• Minimally invasive versus open surgical procedure.

• Operations indicated for the restoration of bowel continuity (primary anastomosis) with versus without creation of a protective stoma.

• Comparison of different antibiotic therapy plans:

  • depending on the protocols used (number of different regimens as well as their prevalence), both a subgroup analysis of different oAB options and an analysis of the effect of different parenteral antibiotic regimens will be considered, provided comparability is ensured.

We will assess differences between subgroups by performing a test for heterogeneity across subgroups (i.e. Cochran's Q) and analysing I². 

We will use the primary outcome parameters in subgroup analyses.

Sensitivity analysis

We will conduct sensitivity analysis to examine the robustness of our findings, investigating:

• the influence of studies judged at 'some concerns' or 'high risk' on the effect size by removing these studies for each outcome and re‐analysing the remaining studies to see if the results are influenced by these factors;

• the impact of missing data on effect size by excluding studies with missing data > 20% and re‐analysing the remaining studies;

• the influence of the type of studies included on the effect size by excluding quasi‐RCTs and re‐analysing only RCTs.

Summary of findings and assessment of the certainty of the evidence

We will create summary of findings tables using GRADEpro GDT software for each planned comparison (MBP + oAB versus oAB alone/MBP alone/no bowel preparation) for the following outcomes (GRADEpro GDT), as measured during the 30‐day postoperative duration (Schünemann 2021):

• SSIs;

• anastomotic leakage;

• mortality;

• incidence of postoperative ileus;

• LOS.

If studies collected outcome data at multiple time points, we will use the measurement from the latest time point within the 30‐day window.

Two review authors will independently assess the certainty of the evidence based on the five GRADE domains (overall RoB 2 judgement, imprecision, inconsistency, indirectness, and publication bias) (Schünemann 2013). Any disagreements will be resolved through discussion or by involving a third review author. The GRADE assessment will result in one of four degrees of certainty (high, moderate, low, or very low certainty), expressing our confidence in the estimate of impact.

Appendices

Appendix 1. Search strategy for MEDLINE

MEDLINE will be searched via PubMed using the following proposed search string:

#1           "Colorectal Surgery"[MeSH Terms] OR "Colectomy"[Mesh]

#2           colorectal [tiab] OR colon [tiab] OR rectal [tiab] OR proctolog* [tiab] OR proctocolonic [tiab]

#3           surger* [tiab] OR surgical* [tiab] OR resect* [tiab] OR incisi* [tiab] OR excisi* [tiab] OR invasive* [tiab] OR restorati* [tiab] OR operation* [tiab] OR operative* [tiab] OR perioperati* [tiab] OR peri‐operati* [tiab] OR "surgery"[Subheading] OR "Surgical Procedures, Operative"[Mesh]

#4           #2 AND #3

#5           #4 OR #1

#6           "Gastrointestinal Agents"[Mesh] OR "Laxatives" [Pharmacological Action] OR "Enema"[Mesh] OR "Cathartics"[Mesh] OR "Laxatives"[Mesh] OR "Antibiotic Prophylaxis"[Mesh] OR "Anti‐Bacterial Agents"[Mesh]

#7           antibacterial* [tiab] OR anti bacterial* [tiab] OR antibiotic* [tiab] OR neomycin [tiab] OR metronidazole [tiab] OR ciprofloxacin [tiab] OR colistin [tiab] OR tobramycin [tiab] OR paromomycin [tiab] OR erythromycin [tiab] OR levofloxacin [tiab]

#8           oral [tiab] OR orally [tiab]

#9           #7 AND #8

#10        bowel preparat* [tiab] OR intestine preparat* [tiab] OR colon preparat* [tiab] OR gut preparat* [tiab] OR bowel cleansing [tiab] OR intestine cleansing [tiab] OR colon cleansing [tiab] OR gut cleansing [tiab] OR laxative* [tiab] OR purgative [tiab] OR enema [tiab]

#11        #9 AND #10

#12        #6 OR #11

#13        #5 AND #12

#14        ("randomized controlled trial"[Publication Type] OR "controlled clinical trial"[Publication Type] OR "randomized"[Title/Abstract] OR "placebo"[Title/Abstract] OR "clinical trials as topic"[MeSH Terms:noexp] OR "randomly"[Title/Abstract] OR "trial"[Title]) NOT ("animals"[MeSH Terms] NOT "humans"[MeSH Terms])

#15        #13 AND #14

Contributions of authors

Maria A Willis (MW) drafted the protocol. The Background and all clinical aspects of the protocol were subsequently reviewed by Tim O Vilz (TV). The methodology was checked by Ingrid Toews (IT) (Institute for Evidence in Medicine) and revised with her support. The search strategy for MEDLINE was developed by MW with the support of Kathrin Grummich (Information Specialist, Cochrane Germany). The protocol was subsequently checked by Jörg J Meerpohl (JM).

Sources of support

Internal sources

• Department of General, Visceral, Thorax and Vascular Surgery, University Hospital Bonn, Bonn, Germany, Germany

  Salary (TV); Salary (MW before and after the nine‐month leave to work full time on the review)

• Institute for Evidence in Medicine, Freiburg, Germany

  Salary (JM)

External sources

• German Cancer Aid, Germany

  Scientific grant for the development of a guideline on 'Perioperative Management of Gastrointestinal Tumours (POMGAT)' including salary payment for MW during a nine‐month leave of absence from clinical practice to complete the review in collaboration with Cochrane Germany 

Declarations of interest

MW: surgical resident at the University Hospital Bonn. This planned review is intended as a partial aspect of the evidence workup for a new guideline on perioperative medicine for gastrointestinal tumours. This guideline is being developed with the AWMF (Arbeitsgemeinschaft der Wissenschaftlichen Medizinischen Fachgesellschaften e.V.) and the German Guideline Program in Oncology, and is funded by the German Cancer Aid (see Sources of support).

IT: none.

JM: none.

TV: scientific grant from the German Cancer Aid to create a guideline for 'Perioperative Management of Gastrointestinal Tumours (POMGAT) (see Sources of support); publications in the field of perioperative medicine; head of the Colorectal Surgery and Proctology section at the University Hospital Bonn.

New