Antibiotic prophylaxis for elective hysterectomy

Reuben Olugbenga Ayeleke; Selma M Mourad; Jane Marjoribanks; Karim A Calis; Vanessa Jordan

doi:10.1002/14651858.CD004637.pub2

. 2017 Jun 18;2017(6):CD004637. doi: 10.1002/14651858.CD004637.pub2

Antibiotic prophylaxis for elective hysterectomy

Reuben Olugbenga Ayeleke ¹, Selma M Mourad ², Jane Marjoribanks ¹, Karim A Calis ³, Vanessa Jordan ^1,^✉

Editor: Cochrane Gynaecology and Fertility Group

PMCID: PMC6441670 PMID: 28625021

Abstract

Background

Elective hysterectomy is commonly performed for benign gynaecological conditions. Hysterectomy can be performed abdominally, laparoscopically, or vaginally, with or without laparoscopic assistance. Antibiotic prophylaxis consists of administration of antibiotics to reduce the rate of postoperative infection, which otherwise affects 40%‐50% of women after vaginal hysterectomy, and more than 20% after abdominal hysterectomy. No Cochrane review has systematically assessed evidence on this topic.

Objectives

To determine the effectiveness and safety of antibiotic prophylaxis in women undergoing elective hysterectomy.

Search methods

We searched electronic databases to November 2016 (including the Cochrane Gynaecology and Fertility Group Specialised Register, the Cochrane Central Register of Studies (CRSO), MEDLINE, Embase, PsycINFO, and the Cumulative Index to Nursing and Allied Health Literature (CINAHL), as well as clinical trials registers, conference abstracts, and reference lists of relevant articles.

Selection criteria

All randomised controlled trials (RCTs) comparing use of antibiotics versus placebo or other antibiotics as prophylaxis in women undergoing elective hysterectomy.

Data collection and analysis

We used Cochrane standard methodological procedures.

Main results

We included in this review 37 RCTs, which performed 20 comparisons of various antibiotics versus placebo and versus one another (6079 women). The quality of the evidence ranged from very low to moderate. The main limitations of study findings were risk of bias due to poor reporting of methods, imprecision due to small samples and low event rates, and inadequate reporting of adverse effects.

Any antibiotic versus placebo

Vaginal hysterectomy

Low‐quality evidence shows that women who received antibiotic prophylaxis had fewer total postoperative infections (risk ratio (RR) 0.28, 95% confidence interval (CI) 0.19 to 0.40; four RCTs, N = 293; I² = 85%), less urinary tract infection (UTI) (RR 0.58, 95% CI 0.43 to 0.77; eight RCTs, N = 1473; I² = 44%), fewer pelvic infections (RR 0.28, 95% CI 0.20 to 0.39; 11 RCTs, N = 1693; I² = 57%), and fewer postoperative fevers (RR 0.43, 95% CI 0.34 to 0.54; nine RCTs, N = 1562; I² = 47%) than women who did not receive such prophylaxis. This suggests that antibiotic prophylaxis reduces the average risk of postoperative infection from about 34% to 7% to 14%. Whether this treatment has led to differences in rates of other serious infection remains unclear (RR 0.20, 95% CI 0.01 to 4.10; one RCT, N = 146; very low‐quality evidence).

Data were insufficient for comparison of adverse effects.

Abdominal hysterectomy

Women who received antibiotic prophylaxis of any class had fewer total postoperative infections (RR 0.38, 95% CI 0.21 to 0.67; one RCT, N = 158; low‐quality evidence), abdominal wound infections (RR 0.51, 95% CI 0.36 to 0.73; 11 RCTs, N = 2247; I² = 6%; moderate‐quality evidence), UTIs (RR 0.41, 95% CI 0.31 to 0.53; 11 RCTs, N = 2705; I² = 28%; moderate‐quality evidence), pelvic infections (RR 0.50, 95% CI 0.35 to 0.71; 11 RCTs, N = 1883; I² = 11%; moderate‐quality evidence), and postoperative fevers (RR 0.59, 95% CI 0.50 to 0.70; 11 RCTs, N = 2394; I² = 55%; moderate‐quality evidence) than women who did not receive prophylaxis, suggesting that antibiotic prophylaxis reduces the average risk of postoperative infection from about 16% to 1% to 6%. Whether this treatment has led to differences in rates of other serious infection remains unclear (RR 0.44, 95% CI 0.12 to 1.69; two RCTs, N = 476; I² = 29%; very low‐quality evidence).

It is unclear whether rates of adverse effects differed between groups (RR 1.80, 95% CI 0.62 to 5.18; two RCTs, N = 430; I² = 0%; very low‐quality evidence).

Head‐to‐head comparisons between antibiotics

Vaginal hysterectomy

We identified four comparisons: cephalosporin versus penicillin (two RCTs, N = 470), cephalosporin versus tetracycline (one RCT, N = 51), antiprotozoal versus lincosamide (one RCT, N = 80), and cephalosporin versus antiprotozoal (one RCT, N = 78). Data show no evidence of differences between groups for any of the primary outcomes, except that fewer cases of total postoperative infection and postoperative fever were reported in women who received cephalosporin than in those who received antiprotozoal.

Only one comparison (cephalosporin vs penicillin; two RCTs, N = 451) yielded data on adverse effects and showed no differences between groups.

Abdominal hysterectomy

We identified only one comparison: cephalosporin versus penicillin (N = 220). Data show no evidence of differences between groups for any of the primary outcomes. Adverse effects were not reported.

Combined antibiotics versus single antibiotics

Vaginal hysterectomy

We identified three comparisons: cephalosporin plus antiprotozoal versus cephalosporin (one RCT, N = 78), cephalosporin plus antiprotozoal versus antiprotozoal (one RCT, N = 78), and penicillin plus antiprotozoal versus penicillin (one RCT, N = 18). Data were unavailable for most outcomes, including adverse effects. We found no evidence of differences between groups, except that fewer women receiving cephalosporin with antiprotozoal received a diagnosis of total postoperative infection, UTI, or postoperative fever compared with women receiving antiprotozoal.

Abdominal hysterectomy

We identified one comparison (penicillin plus antiprotozoal vs penicillin only; two RCT, N = 155). Whether differences between groups occurred was unclear. Adverse effects were not reported.

Comparison of cephalosporins in different regimens

Single small trials addressed dose comparisons and provided no data for most outcomes, including adverse effects. Whether differences between groups occurred was unclear. No trials compared route of administration.

The quality of evidence for all head‐to‐head and dose comparisons was very low owing to very serious imprecision and serious risk of bias related to poor reporting of methods.

Authors' conclusions

Antibiotic prophylaxis appears to be effective in preventing postoperative infection in women undergoing elective vaginal or abdominal hysterectomy, regardless of the dose regimen. However, evidence is insufficient to show whether use of prophylactic antibiotics influences rates of adverse effects. Similarly, evidence is insufficient to show which (if any) individual antibiotic, dose regimen, or route of administration is safest and most effective. The most recent studies included in this review were 14 years old at the time of our search. Thus findings from included studies may not reflect current practice in perioperative and postoperative care and may not show locoregional antimicrobial resistance patterns.

Plain language summary

Antibiotic prophylaxis for elective hysterectomy

Review question

Are antibiotics effective and safe for preventing postoperative infection in women undergoing elective (non‐urgent) hysterectomy?

Background

Surgical operation carried out to remove the uterus (hysterectomy) is commonly performed. Most cases are performed as non‐urgent (elective) procedures for non‐cancerous (benign) conditions affecting the uterus, such as menstrual pain or abnormal bleeding patterns. Antibiotics are usually given before the operation is performed (prophylactic antibiotics, or antibiotic prophylaxis) to prevent or reduce the occurrence of infection after the procedure. Researchers in the Cochrane Collaboration reviewed the evidence on effectiveness and safety of antibiotics used to prevent infection after non‐urgent surgical operation to remove the uterus. Evidence is current to November 2016.

Study characteristics

We identified 37 randomised controlled trials (RCTs), which included a total of 6079 women and compared 20 different types of antibiotics versus placebo (an inactive pill) or versus one another.

Key results

This review found moderate‐quality evidence showing that antibiotics appear to be effective in preventing infection in women undergoing non‐urgent surgical removal of the uterus through the vagina or abdomen. This suggests that antibiotic prophylaxis reduces the average risk of postoperative infection after vaginal hysterectomy from about 62% to 12% to 25%, and after abdominal hysterectomy from about 39% to 8% to 26%.

However, evidence is insufficient to show whether use of prophylactic antibiotics influences rates of adverse effects (side effects), or whether any one antibiotic is more effective or safer than the others.

When antibiotics are compared head‐to‐head or in combination versus single antibiotics, it is unclear which individual antibiotic was more effective and safer, or whether combined antibiotics were more effective and safer than single antibiotics. The quality of the evidence for these comparisons is very low.

It is also unclear which dose regimen or route of administration of antibiotics is safest or most effective in women undergoing elective hysterectomy.

The most recent of the studies included in this review was published 14 years ago, at the time of our search. Thus findings from the included studies may not reflect current practice in perioperative and postoperative care and may not show locoregional antimicrobial resistance patterns.

Quality of the evidence

The quality of evidence for our main comparisons ranged from very low to moderate. The main limitations of this evidence are risk of bias due to poor reporting of randomisation methods, imprecision due to small sample sizes and low event rates, and inadequate reporting of adverse effects.

Summary of findings

Background

Description of the condition

Hysterectomy is one of the most commonly performed operations, particularly in the United States, where the lifetime risk of a hysterectomy is 45% (Merrill 2013). Most hysterectomies are elective (non‐urgent) procedures for benign gynaecological conditions; the most common of these in the United States is leiomyoma (fibroids). Other common indications include endometriosis, heavy menstrual bleeding, and uterovaginal prolapse. This surgery can be performed abdominally, laparoscopically, or vaginally, with or without laparoscopic assistance (Farquhar 2002). The incidence of postoperative infection after hysterectomy was found to be 2% in a recent large cohort from the United States, in which women had surgery between 2012 and 2015 (Upall 2016). In older cohorts, this percentage is likely to be higher owing to factors such as longer hospital stay and prolonged postoperative urinary catheterisation. Some types of hysterectomy may be more susceptible to infectious complications than others, depending on the extent of the breach in body tissues and in the genital tract.

Even with the best surgical and postoperative care, hysterectomy is unavoidably associated with high risk of infection because the procedure breaches the genital tract ‐ an area commonly colonised by a wide variety and large numbers of micro‐organisms. In addition, most women undergoing hysterectomy require an indwelling urinary catheter for the first 24 hours, which increases the risk of urinary tract infection. Common sites of infection after hysterectomy include bladder, pelvic floor, the cuff of tissue at the top of the vagina (vaginal vault), and the abdominal wound; related complications include pelvic abscess, infected haematoma (accumulation of blood from the wound), septicaemia (infection of the blood), and pneumonia (Duff 1980; Faro 2001). Such infections are usually caused by a mixture of bacteria from the woman's own vaginal or urethral tissues ‐ both Gram‐positive and Gram‐negative, and both aerobic and anaerobic (these terms refer to the staining techniques used in identification, and whether the bacteria are oxygen dependent). The individual woman's susceptibility to infection depends upon the effectiveness of her immune system, the virulence of the bacteria present, and the extent of tissue trauma and fluid collection resulting from surgery (Duff 1980).

Description of the intervention

"Antibiotic prophylaxis" refers to administration of antibiotics to prevent infection: It has been used in surgery since antibiotics were introduced in the 1950s, in an attempt to reduce the rate of postoperative infection. Such infection not only causes patient morbidity but may result in additional costs, extended hospital stay, and increased antibiotic use, which promotes the emergence of antimicrobial resistant organisms (Dellinger 1994). Antibiotic prophylaxis for hysterectomy has been extensively studied, and it has been estimated that such prophylaxis has reduced the rate of postoperative infection by more than half; otherwise, about 40% to 50% of women would develop infection after vaginal hysterectomy, and more than 20% after abdominal hysterectomy (Duff 1980; Mittendorf 1993). National guidelines now recommend this practice for all types of hysterectomy (ACOG 2009; Bratzler 2013; Dellinger 1994; Nelson 2016; RCOG 1999; SIGN 2008; Van Eyk N, van Schalkwyk J 2012), although in reality, application of such guidelines is variable (Gorecki 1999).

Although various antibiotic regimens and routes of delivery have been used, the most frequent current practice consists of a single dose of antibiotic given intravenously within two hours of the surgical incision, to facilitate optimum serum antibiotic levels during the operation (Classen 1992; DiPiro 1984; Nelson 2016). A single dose has been reported to be as effective as multiple doses, although some researchers have suggested repeat dosing if surgery is long or blood loss is high (DiPiro 1986; Tanos 1994). If prophylaxis is continued postoperatively, it is recommended that the duration of therapy should not exceed 24 hours (Dellinger 1994).

The type of antibiotic most commonly used is active against a wide range of bacteria (broad spectrum); this type includes amoxicillin/clavulanic acid (Augmentin) or a cephalosporin. Cephalosporins are grouped into generations according to their antimicrobial properties, with the oldest type referred to as "first generation". Subsequent generations of these drugs have progressively widened their antibacterial coverage against Gram‐negative organisms while showing a concurrent reduction in effectiveness against Gram‐positive organisms; moreover, wide use of very broad‐spectrum antibiotics greatly increases the risk of emergence of drug‐resistant bacteria (BNF 2002). It is generally recommended that first‐ or second‐generation cephalosporins should be used for prophylaxis, as they appear to be equally effective for this purpose, less expensive than other treatments, and less likely to favour drug resistance (Fukatsu 1997; Tanos 1994; Weed 2003).

How the intervention might work

Prophylaxis works by briefly bolstering tissue defence mechanisms to promote rapid restoration of normal immune responses after the trauma of surgery.

Why it is important to do this review

A very large body of evidence on prophylactic antibiotics for hysterectomy involves hundreds of clinical trials. However, review authors have not systematically assessed this evidence in recent times. Existing meta‐analyses conducted some years back focused mainly on abdominal hysterectomy. No meta‐analysis has focused on trials involving other routes of hysterectomy.

Several Cochrane reviews of prophylactic antibiotics for elective surgery have reported mixed findings. Two of these examined the topic of caesarean section (Gyte 2014; Nabhan 2016). Gyte 2014 evaluated different classes of prophylactic antibiotics for women undergoing caesarean section and found that cephalosporins and penicillins had similar efficacy for preventing immediate postoperative infection. Investigators provided no data on late infection, nor on outcomes for the baby. Nabhan 2016 compared routes of administration of prophylactic antibiotics and concluded that data show no clear difference between irrigation and intravenous routes in rates of post‐caesarean endometritis. A review on elective endoscopic retrograde cholangiopancreatography (Brand 2010) reported that antibiotic prophylaxis appeared to reduce rates of bacteraemia, cholangitis, and septicaemia. A review of different regimens of antibiotic prophylaxis for people undergoing orthognathic surgery (Brignardello‐Petersen 2015) found that long‐term antibiotic prophylaxis decreased the risk of skin and skin structure infection compared with short‐term prophylaxis, but comparisons between short‐term prophylaxis and a single preoperative dose were inconclusive. Reviews of antibiotic prophylaxis for elective open inguinal hernia repair (Sanchez‐Manuel 2012) or for elective laparoscopic cholecystectomy (Sanabria 2010) provided no clear evidence of benefit for the intervention group.

Objectives

To determine the effectiveness and safety of antibiotic prophylaxis in women undergoing elective hysterectomy.

Methods

Criteria for considering studies for this review

Types of studies

Randomised, controlled trials (RCTs) of women having an elective total or subtotal hysterectomy by any route and comparing prophylactic antibiotics versus placebo or versus a different type, route, or timing of antibiotic. Trials were at least double‐blinded (i.e. with participants and clinicians blinded). We did not include quasi‐randomised trials (e.g. trials that allocated treatment by date of birth, day of the week, medical record number, month of the year, or the order in which participants were enrolled in the study).

We excluded from the review studies that did not analyse at least 80% of women randomised for at least one outcome. When trials analysed at least 80% of participants for some outcomes but analysed less than 80% of participants for other outcomes, we included only those outcomes analysed for at least 80% of participants. The rationale for excluding trials with high numbers of withdrawals is that attrition was unlikely to be equally distributed between trial arms: Women who did not develop infection were more likely to be lost to follow‐up than those who did develop infection.

Types of participants

Women of any age without serious comorbidity (such as cancer) undergoing an elective total or subtotal abdominal, vaginal, laparoscopic, or laparoscopically assisted hysterectomy, with or without oophorectomy, for a benign gynaecological condition such as fibroids, endometriosis, uterovaginal prolapse, or heavy menstrual bleeding.

Types of interventions

Prophylactic antibiotics versus placebo or a different type or regimen of antibiotics.

The term "prophylactic" was defined as follows. Prophylactic: antibiotic(s) given when an individual had no signs or symptoms of infection, when no antibiotics had been taken within the previous 48 hours, and when the first dose was given up to 12 hours preoperatively and the last dose was given not more than 24 hours postoperatively.

Types of antibiotics

Antibiotics were classified into the following types.

Cephalosporins.
1. First‐generation (e.g. cefazolin, cephradine, cephazolin, cephalexin, cefadroxil).
2. Second‐generation (e.g. cefoxitin, cefuroxime, cephamandole, cefaclor, cefprozil, loracarbef).
3. Third‐generation (e.g. cefotaxime, cefotetan, ceftazidime, ceftriaxone, cefixime, cefpodoxime proxetil, ceftibuten, cefdinir, cephoperazone, ceftizoxime).
4. Fourth‐generation (e.g. cefepime).
Penicillins (e.g. penicillin, amoxicillin).
Macrolides (e.g. erythromycin, clarithromycin, azithromycin).
Fluoroquinolones (e.g. ciprofloxacin, levofloxacin, oxfloxacin).
Sulfonamides (e.g. co‐trimoxazole, trimethoprim).
Tetracyclines (e.g. tetracycline, doxycycline).
Aminogylocosides (e.g. gentamycin, tobramycin).
Glycopeptides (e.g. vancomycin).
Antiprotozoals (e.g. metronidazole, anitroimidazole).
Combination drugs.
1. Augmentin (amoxicillin and clavulanic acid).
2. Other combinations of drugs (will be considered individually).

Antibiotic regimens include the following.

Route: Any systemic regimen was included, irrespective of the route of administration (e.g. intravenous, intramuscular, oral, rectal).
Number of doses (e.g. single vs repeated doses).

Types of outcome measures

We considered trials if they reported any of the following clinical outcomes.

Primary outcomes

Infection: measured as the proportion of women who within eight weeks of surgery developed one of the following as defined by the study.
1. Total postoperative infection.
2. Abdominal wound infection (e.g. wound cellulitis, abscess, dehiscence).
3. Pelvic infection (including vaginal cuff (vault) infection, pelvic inflammatory disease, pelvic abscess, infected haematoma).
4. Urinary tract infection.
5. Other serious infection or infectious complication, such as septicaemia, septic shock, distant infection (e.g. pneumonia).
Postoperative fever of > 38° on at least two occasions more than four hours apart, excluding the day of surgery.
Total adverse effects: morbidity (e.g. allergic reaction, diarrhoea, bacterial resistance, or as defined by the study) and mortality (infection‐related and all‐cause).

We classified primary outcomes as early (before discharge from hospital or within seven days of surgery), late (at follow‐up: within eight weeks of surgery), or total (early + late).

Secondary outcomes

Need for therapeutic antibiotics ‐ early (before discharge from hospital or within seven days of surgery), late (at follow‐up: within eight weeks of surgery), or total (early + late).
Length of hospital stay.
Quality of life.

Search methods for identification of studies

In consultation with the Gynaecology and Fertility Group Information Specialist, we searched the following databases for all published and unpublished RCTs.

Electronic searches

We searched the following electronic databases, trial registers, and websites up to 29 November 2016.

Gynaecology and Fertility Group (CGF) Specialised Register of Controlled Trials.
Cochrane Central Register of Studies Online (CRSO).
MEDLINE.
Embase.
PsycINFO.
Cumulative Index to Nursing Allied Health and Literature (CINAHL).
1. We combined the MEDLINE search with the Cochrane highly sensitive search strategy for identifying randomised trials, which appears in the Cochrane Handbook for Systematic Reviews of Interventions (Version 5.0.2, Chapter 6, 6.4.11). We combined Embase, PsycINFO, and CINAHL searches using trial filters developed by the Scottish Intercollegiate Guidelines Network (SIGN) (http://www.sign.ac.uk/methodology/filters.html#random).
Other electronic sources of trials included:
1. trial registers for ongoing and registered trials;
2. http://www.clinicaltrials.gov (a service of the US National Institutes of Health);
3. http://www.who.int/trialsearch/Default.aspx (World Health Organization International Clinical Trials Registry Platform search portal) (Note: it is now mandatory for Cochrane reviews to include searches of trial registers);
4. DARE (Database of Abstracts of Reviews of Effects) in the Cochrane Library (http://onlinelibrary.wiley.com/o/cochrane/cochrane_cldare_articles_fs.html) (for reference lists from relevant non‐Cochrane reviews);
5. Web of Knowledge (http://wokinfo.com/ ‐ another source of trials and conference abstracts);
6. OpenGrey (http://www.opengrey.eu/ ‐ for unpublished literature from Europe);
7. Latin American Caribbean Health Sciences Literature (LILACS database) (http://regional.bvsalud.org/php/index.php?lang=en ‐ for trials from the Portuguese‐ and Spanish‐speaking world); and
8. PubMed and Google Scholar (for recent trials not yet indexed in MEDLINE).

For details of search strategies, see Appendix 1,Appendix 2,Appendix 3,Appendix 4,Appendix 5, and Appendix 6.

Searching other resources

We handsearched the reference lists of articles retrieved by the search and contacted experts in the field to request additional data. We also handsearched relevant journals and conference abstracts not included in the CGF register, in liaison with the Information Specialist from the CGF Group.

Data collection and analysis

Selection of studies

After an initial screen of titles and abstracts retrieved by the search, we retrieved the full texts of all potentially eligible studies. At least two review authors (of VJ, JM, and RA) independently examined these full‐text articles for compliance with the inclusion criteria and selected studies that were eligible for inclusion in the review. We contacted study investigators as required to clarify study eligibility. We resolved disagreements regarding study eligibility by discussion or by consultation with a third review author. We documented the selection process using a PRISMA (Preferred Reporting Items for Systematic Reviews and Meta‐Analyses) flow chart.

Data extraction and management

Two review authors independently extracted data from eligible studies using a data extraction form that they had designed and pilot‐tested. We resolved disagreements by discussion or by consultation with a third review author. Data extracted included study characteristics and outcome data. When studies had multiple publications, review authors collated multiple reports of the same study, so that each study rather than each report was the unit of interest in the review, and assigned such studies a single study ID with multiple references.

We contacted study investigators to request additional data on methods and/or results, as required.

Assessment of risk of bias in included studies

Two review authors independently examined included studies for risk of bias using the Cochrane "Risk of bias" assessment tool (Higgins 2011) to assess selection (random sequence generation and allocation concealment); performance (blinding of participants and personnel); detection (blinding of outcome assessors); attrition (incomplete outcome data); reporting (selective reporting); and other bias such as differences in demographic characteristics of participants. We took care to search for within‐trial selective reporting, as seen in trials failing to report obvious outcomes, or reporting them in insufficient detail to allow inclusion. We sought published protocols and compared outcomes between the protocol and the final published study.

We resolved disagreements by discussion or by consultation with a third review author. We described all judgements fully and presented conclusions in the "Risk of bias" table; we incorporated these into the interpretation of review findings by performing sensitivity analyses (see below).

Measures of treatment effect

For dichotomous data (e.g. infection rates), we used numbers of events in control and intervention groups of each study to calculate risk ratios (RRs). For continuous data (e.g. length of hospital stay), when studies reported exactly the same outcomes, we calculated mean differences (MDs) between treatment groups. We reversed the direction of effect of individual studies, if required, to ensure consistency across trials. We intended to treat ordinal data (e.g. quality of life scores) as continuous data if any included studies reported ordinal data. We presented 95% confidence intervals (CIs) for all outcomes. We compared the magnitude and direction of effects reported by studies versus how they were presented in the review, while taking account of legitimate differences.

Unit of analysis issues

The primary analysis was per woman randomised.

Dealing with missing data

We analysed the data on an intention‐to‐treat basis as far as possible and attempted to obtain missing data from the original trialists. When these were unobtainable, we analysed only available data.

When studies reported sufficient detail for calculation of mean differences but no information on associated standard deviation (SD), we assumed the outcome to have a standard deviation equal to the highest SD from other studies within the same analysis.

Assessment of heterogeneity

We considered whether clinical and methodological characteristics of included studies were sufficiently similar for meta‐analysis to provide a clinically meaningful summary. We assessed statistical heterogeneity by using the I² measurement. We took an I² measurement greater than 50% to indicate substantial heterogeneity (Higgins 2003; Higgins 2011).

Assessment of reporting biases

In view of the difficulty of detecting and correcting for publication bias and other reporting biases, review authors aimed to minimise their potential impact by ensuring a comprehensive search for eligible studies and by staying alert for duplication of data. When we included 10 or more studies in an analysis, we used a funnel plot to explore the possibility of small‐study effects (the tendency for estimates of the intervention effect to be more beneficial in smaller studies).

Data synthesis

When studies were sufficiently similar, we combined the data using a fixed‐effect model.

We graphically displayed an increase in risk of a particular outcome within meta‐analyses to the right of the centre‐line, and a decrease in risk of a particular outcome to the left of the centre‐line.

We made the following comparisons.

Any antibiotic versus placebo.
Specific antibiotics versus placebo.
Head‐to‐head comparisons of antibiotics.
Comparisons of antibiotic regimens.

We subgrouped all analyses by surgical route: vaginal or abdominal. We did not pool these subgroups.

Subgroup analysis and investigation of heterogeneity

We subgrouped our main analysis according to the surgical route used (vaginal or abdominal). We did not undertake other prespecified subgroup analyses.

When we detected substantial heterogeneity (I² > 50%), we explored possible explanations by performing sensitivity analyses. We took any statistical heterogeneity into account when interpreting results, especially if we noted any variation in the direction of effect estimates.

Sensitivity analysis

When heterogeneity was substantial (I² > 50%), we conducted sensitivity analysis by choosing a statistical model (fixed‐effect vs random‐effects) and an effect estimate (risk ratio vs odds ratio), regardless of the number of trials included in an analysis. We planned to explore other clinical or methodological differences between studies only if data showed variation in the direction of effect.

Overall quality of the body of evidence: "Summary of findings" table

We prepared two separate "Summary of findings" tables for vaginal hysterectomy and abdominal hysterectomy based on the review's main comparison, that is, any antibiotics versus placebo. We used GRADEPRO (GRADEPro GDT 2014) and Cochrane methods (Higgins 2011) and used these tables to evaluate the overall quality of the body of evidence for main review outcomes (total postoperative infections, abdominal wound infection, urinary tract infection, pelvic infection, other serious infection, postoperative fever, and total adverse effects) by applying GRADE criteria (study limitations (i.e. risk of bias), consistency of effect, imprecision, indirectness, and publication bias). Two review authors working independently made judgements about evidence quality (high, moderate, low, or very low) and resolved disagreements by discussion. We justified, documented, and incorporated our judgements into reporting of results for each outcome.

Results

Description of studies

Results of the search

The search produced a total of 940 titles and abstracts after duplicates were removed; we considered 149 full‐text articles for further assessment. Thirty‐seven trials in 42 reports met the eligibility criteria for inclusion, and we excluded 107 full‐text articles. See Characteristics of included studies and Characteristics of excluded studies tables. The PRISMA flow chart in Figure 1 illustrates the flow of literature throughout the search and assessment process.

Included studies

Study design and setting

We included 37 studies in this review (Benigno 1986; Boodt 1990; Chongsomchai 2002; Crosthwaite 1985; Davi 1985; Dhar 1993; Dhar 1993a; Duff 1982; Egarter 1988; Eron 1989; Faro 1988; Gall 1983; Hager 1989; Hedican 1976; Hemsell 1980; Hemsell 1983; Hemsell 1984; Hemsell 1985; Hemsell 1985a; Hemsell 1987; Hemsell 1989; Henriksson 1998; Holman 1978; Houang 1984; Houang 1984a; Jaffe 1985; Janssens 1982; Kauer 1990; Ledger 1973; Mathews 1977; Mathews 1979; Mendelson 1979; Polk 1980; Schepers 1981; Smith 1984; Stage 1982; Vincelette 1983).

The most recent study was Chongsomchai 2002, which was already 14 years old at the time of our search.

All included studies were parallel, double‐blinded, randomised controlled trials (RCTs). Twenty‐nine studies were two‐arm RCTs (Boodt 1990; Crosthwaite 1985; Davi 1985; Dhar 1993; Dhar 1993a; Duff 1982; Faro 1988; Gall 1983; Hager 1989; Hedican 1976; Hemsell 1980; Hemsell 1983; Hemsell 1984; Hemsell 1985a; Hemsell 1989; Henriksson 1998; Holman 1978; Houang 1984a; Jaffe 1985; Janssens 1982; Ledger 1973; Mathews 1977; Mathews 1979; Mendelson 1979; Polk 1980; Schepers 1981; Smith 1984; Stage 1982; Vincelette 1983). Eight studies were three‐arm RCTs (Benigno 1986; Chongsomchai 2002; Egarter 1988; Eron 1989; Hemsell 1985; Hemsell 1987; Houang 1984; Kauer 1990).

Seventeen studies were conducted in the United States (Benigno 1986; Duff 1982; Eron 1989; Gall 1983; Hager 1989; Hedican 1976; Hemsell 1980; Hemsell 1983; Hemsell 1984; Hemsell 1985; Hemsell 1985a; Hemsell 1987; Hemsell 1989; Holman 1978; Ledger 1973; Polk 1980; Stage 1982); five studies were conducted in the United Kingdom (Houang 1984; Houang 1984a; Mathews 1977; Mathews 1979; Smith 1984); two were conducted in Canada (Mendelson 1979; Vincelette 1983); and three in the Netherlands (Boodt 1990; Kauer 1990; Schepers 1981). Two studies each were conducted in Australia (Crosthwaite 1985; Egarter 1988) and India (Chandigarth) (Dhar 1993; Dhar 1993a); one study each was conducted in Belgium (Janssens 1982), Israel (Jaffe 1985), Sweden (Henriksson 1998), and Thailand (Chongsomchai 2002). The remaining two studies did not provide information on the countries in which they were conducted (Davi 1985; Faro 1988).

Six of the included studies were conducted at more than one centre: 14 centres (Stage 1982), four centres (Benigno 1986), three centres (Hager 1989; Henriksson 1998), and two centres (Chongsomchai 2002; Eron 1989); five studies did not report the number of centres (Davi 1985; Egarter 1988; Faro 1988; Hemsell 1985; Schepers 1981); and each of the remaining 26 studies was conducted at a single centre (Boodt 1990; Crosthwaite 1985; Dhar 1993; Dhar 1993a; Duff 1982; Gall 1983; Hedican 1976; Hemsell 1980; Hemsell 1983; Hemsell 1984; Hemsell 1985a; Hemsell 1987; Hemsell 1989; Holman 1978; Houang 1984; Houang 1984a; Jaffe 1985; Janssens 1982; Kauer 1990; Ledger 1973; Mathews 1977; Mathews 1979; Mendelson 1979; Polk 1980; Smith 1984; Vincelette 1983).

Participants

The 37 included studies enrolled a total of 6079 women. Seventeen studies randomised or analysed a total of 100 or fewer women (Crosthwaite 1985; Dhar 1993; Dhar 1993a; Duff 1982; Gall 1983; Hager 1989; Hedican 1976; Hemsell 1980; Hemsell 1985a; Houang 1984a; Jaffe 1985; Kauer 1990; Ledger 1973; Mathews 1977; Mathews 1979; Mendelson 1979; Smith 1984); eight studies randomised or analysed a total of 101 to 200 women (Egarter 1988; Faro 1988; Hemsell 1983; Hemsell 1984; Hemsell 1985; Janssens 1982; Schepers 1981; Vincelette 1983); five studies randomised or analysed a total of 201 to 300 women (Eron 1989; Hemsell 1987; Hemsell 1989; Holman 1978; Stage 1982); five studies randomised or analysed a total of 301 to 400 women (Benigno 1986; Chongsomchai 2002; Davi 1985; Henriksson 1998; Houang 1984); one study randomised a total of 403 women (Boodt 1990); and another randomised a total of 557 women (Polk 1980).

A common inclusion criterion was that women had to be scheduled for elective abdominal hysterectomy, vaginal hysterectomy, or both types of hysterectomy for a benign condition. Thirteeen studies included women scheduled for abdominal hysterectomy (Boodt 1990; Chongsomchai 2002; Davi 1985; Dhar 1993a; Duff 1982; Gall 1983; Hemsell 1983; Hemsell 1985; Houang 1984a; Jaffe 1985; Mathews 1977; Schepers 1981; Smith 1984); 14 studies included women scheduled for elective vaginal hysterectomy (Benigno 1986; Dhar 1993; Egarter 1988; Faro 1988; Hager 1989; Hedican 1976; Hemsell 1980; Hemsell 1984; Hemsell 1985a; Hemsell 1987; Kauer 1990; Ledger 1973; Mathews 1979; Mendelson 1979); nine studies included women scheduled for either abdominal or vaginal hysterectomy (Crosthwaite 1985; Eron 1989; Hemsell 1989; Holman 1978; Houang 1984; Janssens 1982; Polk 1980; Stage 1982; Vincelette 1983); and one study did not report the type of hysterectomy for which women were scheduled (Henriksson 1998).

No included studies focused on antibiotic prophylaxis in participants undergoing laparoscopically performed hysterectomy.

Common exclusion criteria were emergency hysterectomy; pregnancy‐related hysterectomy; hypersensitivity to antibiotics such as cephalosporin, penicillin, amoxicillin, etc.; and use of antibiotics within two to seven days before surgery.

Interventions

Included studies compared different classes of antibiotics with placebo or with each other. Included studies identified the following treatment groups.

Any antibiotic versus placebo (Boodt 1990; Chongsomchai 2002; Crosthwaite 1985; Davi 1985; Dhar 1993; Dhar 1993a; Duff 1982; Egarter 1988; Gall 1983; Hedican 1976; Hemsell 1980; Hemsell 1983; Henriksson 1998; Holman 1978; Houang 1984; Jaffe 1985; Janssens 1982; Ledger 1973; Mathews 1977; Mathews 1979; Mendelson 1979; Polk 1980; Smith 1984; Vincelette 1983).
Cephalosporin versus placebo (Chongsomchai 2002; Davi 1985; Duff 1982; Gall 1983; Hedican 1976; Hemsell 1980; Hemsell 1983; Holman 1978; Ledger 1973; Mendelson 1979; Polk 1980; Stage 1982).
Penicillin versus placebo (Chongsomchai 2002; Houang 1984).
Antiprotozoal versus placebo (Crosthwaite 1985; Dhar 1993; Dhar 1993a; Egarter 1988; Hemsell 1983; Henriksson 1998; Janssens 1982; Vincelette 1983).
Sulphonamides versus placebo (Jaffe 1985; Mathews 1977; Mathews 1979; Smith 1984).
Cephalosporin plus antiprotozoal versus placebo (Boodt 1990).
Penicillin plus antiprotozoal versus placebo (Houang 1984).
Lincosamide versus placebo (Egarter 1988).
Cephalosporin versus penicillin (Benigno 1986; Chongsomchai 2002; Faro 1988; Hager 1989).
Cephalosporin versus tetracycline (Hemsell 1985a).
Cephalosporin versus antiprotozoal (Kauer 1990).
Antiprotozoal versus lincosamide (Egarter 1988).
Cephalosporin plus antiprotozoal versus cephalosporin only (Kauer 1990).
Cephalosporin plus antiprotozoal versus antiprotozoal only (Kauer 1990).
Penicillin plus antiprotozoal versus penicillin only (Houang 1984; Houang 1984a).
Cephalosporin early administration versus usual timing (both single dose) (Eron 1989).
Cephalosporin one dose versus two doses (Hemsell 1985).
Cephalosporin one dose versus three doses (Hemsell 1984; Hemsell 1985).
Cephalosporin one dose versus multiple doses (Mendelson 1979).
Cephalosporin one gram versus two grams (Hemsell 1987).

Included studies administered antibiotics through the following routes.

Intravenous (IV) (Benigno 1986; Boodt 1990; Chongsomchai 2002; Duff 1982; Egarter 1988; Faro 1988; Gall 1983; Hager 1989; Hemsell 1985; Hemsell 1985a; Hemsell 1989; Henriksson 1998; Jaffe 1985; Kauer 1990; Mathews 1979; Mendelson 1979; Polk 1980; Schepers 1981; Stage 1982; Vincelette 1983).
Intramuscular (IM) (Davi 1985; Hemsell 1980; Hemsell 1983; Hemsell 1987; Smith 1984).
IV and IM (Eron 1989; Hedican 1976; Hemsell 1984; Holman 1978).
Oral (Crosthwaite 1985; Dhar 1993; Dhar 1993a; Janssens 1982).
IV and rectal (Houang 1984; Houang 1984a).

One of the included studies did not state the route used for administration of antibiotics (Ledger 1973).

Investigators administered antibiotics as a single dose, as multiple doses, or as single versus multiple doses in the following studies.

Single dose (Boodt 1990; Chongsomchai 2002; Crosthwaite 1985; Dhar 1993; Dhar 1993a; Duff 1982; Hager 1989; Hemsell 1987; Janssens 1982; Ledger 1973; Mathews 1977; Mathews 1979).
Multiple doses (Boodt 1990; Davi 1985; Egarter 1988; Faro 1988; Gall 1983; Hedican 1976; Hemsell 1980; Hemsell 1983; Hemsell 1984; Henriksson 1998; Holman 1978; Houang 1984; Houang 1984a; Ledger 1973; Polk 1980; Schepers 1981; Stage 1982; Vincelette 1983).
Single dose versus multiple doses (Eron 1989; Hemsell 1985; Hemsell 1985a; Hemsell 1989; Janssens 1982; Mendelson 1979).

Timing and duration of administration varied in the included studies. However, none of the included studies administered the first dose of antibiotics more than 12 hours before surgery and the last dose more than 24 hours after surgery.

Outcomes

Primary outcome measures of this review were presence of postoperative infection (total postoperative infections, abdominal wound infection, pelvic infection, urinary tract infection (UTI), other serious infection (such as pneumonia, septicaemia, septic shock), and postoperative fever), total adverse effects such as morbidity (e.g. diarrhoea, allergic reactions), and mortality. Thirty‐six included studies reported data on at least one of the review's primary outcome measures (Benigno 1986; Boodt 1990; Chongsomchai 2002; Crosthwaite 1985; Davi 1985; Dhar 1993; Dhar 1993a; Duff 1982; Egarter 1988; Eron 1989; Faro 1988; Gall 1983; Hager 1989; Hedican 1976; Hemsell 1980; Hemsell 1983; Hemsell 1984; Hemsell 1985; Hemsell 1985a; Hemsell 1987; Hemsell 1989; Henriksson 1998; Holman 1978; Houang 1984; Houang 1984a; Jaffe 1985; Janssens 1982; Kauer 1990; Ledger 1973; Mathews 1977; Mathews 1979; Polk 1980; Schepers 1981; Smith 1984; Stage 1982; Vincelette 1983); and one of the included studies did not report data on any of the review's primary outcomes (Mendelson 1979). Twenty‐five included studies reported data on adverse effects, most in narrative form (Benigno 1986; Chongsomchai 2002; Crosthwaite 1985; Davi 1985; Dhar 1993; Dhar 1993a; Duff 1982; Eron 1989; Gall 1983; Hager 1989; Hemsell 1980; Hemsell 1984; Hemsell 1985a; Hemsell 1987; Hemsell 1989; Henriksson 1998; Jaffe 1985; Kauer 1990; Mathews 1977; Mathews 1979; Polk 1980; Schepers 1981; Smith 1984; Stage 1982; Vincelette 1983). Common adverse effects included allergy reactions and diarrhoea. None of the included studies reported any incident of mortality.

Secondary outcome measures included any requirement for therapeutic antibiotics, length of hospital stay, and quality of life following surgery. Twenty‐seven included studies reported on at least one of the secondary outcome measures (Benigno 1986; Boodt 1990; Chongsomchai 2002; Dhar 1993; Dhar 1993a; Duff 1982; Egarter 1988; Eron 1989; Faro 1988; Gall 1983; Hager 1989; Hemsell 1980; Hemsell 1983; Hemsell 1984; Hemsell 1985; Hemsell 1985a; Hemsell 1987; Hemsell 1989; Holman 1978; Jaffe 1985; Kauer 1990; Ledger 1973; Mathews 1977; Mathews 1979; Polk 1980; Stage 1982; Vincelette 1983). Secondary outcome measures commonly reported were need for therapeutic antibiotics and length of hospital stay; no studies provided data on quality of life. The remaining 10 studies did not report on any of the secondary outcome measures (Crosthwaite 1985; Davi 1985; Hedican 1976; Henriksson 1998; Houang 1984; Houang 1984a; Janssens 1982; Mendelson 1979; Schepers 1981; Smith 1984).

Excluded studies

Review authors determined that 107 studies were not eligible for inclusion in this review. Common reasons for exclusion were administration of antibiotics more than 12 hours before surgery or for more than 24 hours after surgery and non‐blinding of participants and personnel. For further details on reasons for exclusion of studies, see Characteristics of excluded studies table.

Risk of bias in included studies

See Figure 2 and Figure 3.

Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.

Risk of bias summary: review authors' judgements about each risk of bias item for each included study.

Allocation

Random sequence generation

We considered processes used in sequence generation to be adequate in 10 of the included studies because they involved the use of computers (Benigno 1986; Chongsomchai 2002; Faro 1988; Hemsell 1984; Hemsell 1987; Hemsell 1989) or random number tables (Hemsell 1983; Holman 1978; Kauer 1990; Ledger 1973). We therefore rated these studies as having low risk of bias with respect to random sequence generation. The remaining 27 studies provided insufficient information to permit conclusive judgements on the process involved in sequence generation; thus we rated them as having unclear risk of bias.

Allocation concealment

We rated 17 studies as having low risk of bias with respect to allocation concealment (Benigno 1986; Chongsomchai 2002; Dhar 1993a; Duff 1982; Hager 1989; Hedican 1976; Hemsell 1980; Hemsell 1983; Hemsell 1987; Henriksson 1998; Holman 1978; Kauer 1990; Ledger 1973; Mathews 1977; Mathews 1979; Smith 1984; Stage 1982). We considered the processes involved in concealing allocations in these studies to be adequate; these included remote or central allocation through the hospital pharmacy and use of sealed opaque envelopes. We assessed the remaining 20 studies as having unclear risk because information was insufficient to allow conclusive judgements with respect to allocation concealment.

Blinding

We considered that blinding was likely to influence findings for both primary and secondary review outcomes. Although we considered all included studies to be adequate with regard to blinding of both participants and physicians, most did not provide adequate information on how participants were evaluated postoperatively. Only 16 studies reported sufficient information on outcome assessment and/or participant follow‐up; we thus rated these studies as having low risk with respect to performance and detection bias (Benigno 1986; Boodt 1990; Chongsomchai 2002; Eron 1989; Hemsell 1980; Hemsell 1983; Hemsell 1985a; Henriksson 1998; Holman 1978; Houang 1984; Houang 1984a; Janssens 1982; Kauer 1990; Ledger 1973; Polk 1980; Stage 1982). The remaining 21 studies did not provide sufficient information on whether outcome assessors were blinded; we therefore rated these studies as having unclear risk with respect to performance and detection bias (Crosthwaite 1985; Davi 1985; Dhar 1993; Dhar 1993a; Duff 1982; Egarter 1988; Faro 1988; Gall 1983; Hager 1989; Hedican 1976; Hemsell 1984; Hemsell 1985; Hemsell 1987; Hemsell 1989; Jaffe 1985; Mathews 1977; Mathews 1979; Mendelson 1979; Schepers 1981; Smith 1984; Vincelette 1983).

Incomplete outcome data

We judged 16 studies as having low risk with respect to incomplete outcome data or attrition bias (Chongsomchai 2002; Dhar 1993; Dhar 1993a; Duff 1982; Egarter 1988; Hager 1989; Hemsell 1985a; Hemsell 1987; Henriksson 1998; Kauer 1990; Ledger 1973; Mathews 1977; Polk 1980; Smith 1984; Stage 1982; Vincelette 1983). Proportions of withdrawals/losses to follow‐up and reasons for withdrawal in these studies were fairly well balanced or similar across treatment groups, or outcome data were analysed on an intention‐to‐treat (ITT) basis by including all randomised women in data analyses. Nineteen studies provided insufficient information on the number of withdrawals/losses to follow‐up and/or on reasons for withdrawal, and data were not analysed on the basis of ITT (Benigno 1986; Boodt 1990; Crosthwaite 1985; Davi 1985; Faro 1988; Gall 1983; Hedican 1976; Hemsell 1980; Hemsell 1983; Hemsell 1984; Hemsell 1985; Hemsell 1989; Houang 1984a; Janssens 1982; Mathews 1979; Mendelson 1979; Schepers 1981). We thus rated these studies as having unclear risk with respect to attrition bias. We rated the remaining two studies as having high risk of bias: In one of these studies, proportions of withdrawals were not balanced between groups and data were not analysed on the basis of ITT (Eron 1989); in the other study, proportions of withdrawals and reasons for withdrawal were not balanced across treatment groups (Jaffe 1985).

Selective reporting

Protocols were not available for any of the included studies, and review authors could not determine whether outcomes were selectively reported. Therefore, the process of detecting selective reporting bias in included studies involved careful assessment of methods sections to determine which outcomes were prespecified and whether data were reported on all prespecified outcomes. Thirteen studies provided data on all outcomes prespecified in the methods sections; we rated these as having low risk with respect to selective reporting (within‐trial selective reporting) (Benigno 1986; Chongsomchai 2002; Duff 1982; Hager 1989; Hemsell 1980; Henriksson 1998; Jaffe 1985; Mathews 1977; Mathews 1979; Polk 1980; Smith 1984; Stage 1982; Vincelette 1983). Twenty‐three studies provided insufficient information to allow conclusive judgements with respect to selective reporting; therefore, we rated these studies as having unclear risk of selective reporting bias (Boodt 1990; Crosthwaite 1985; Davi 1985; Dhar 1993; Dhar 1993a; Egarter 1988; Eron 1989; Faro 1988; Gall 1983; Hedican 1976; Hemsell 1983; Hemsell 1984; Hemsell 1985; Hemsell 1985a; Hemsell 1987; Hemsell 1989; Holman 1978; Houang 1984; Houang 1984a; Janssens 1982; Kauer 1990; Ledger 1973; Mendelson 1979). We rated the only remaining study as having high risk of selective reporting because evidence showed selective reporting, with no data reported on some of the outcomes prespecified in the methods section (Schepers 1981).

Other potential sources of bias

We assessed other potential sources of bias with respect to whether data showed significant differences between treatment groups in terms of baseline demographic characteristics of participants, such as age and body mass index (BMI). In 28 studies, baseline demographic characteristics were similar between treatment groups; thus we rated these studies as having low risk with respect to other potential sources of bias (Boodt 1990; Chongsomchai 2002; Crosthwaite 1985; Dhar 1993; Dhar 1993a; Duff 1982; Egarter 1988; Eron 1989; Gall 1983; Hager 1989; Hemsell 1980; Hemsell 1983; Hemsell 1984; Hemsell 1987; Henriksson 1998; Holman 1978; Houang 1984; Houang 1984a; Jaffe 1985; Kauer 1990; Ledger 1973; Mathews 1977; Mathews 1979; Mendelson 1979; Polk 1980; Smith 1984; Stage 1982; Vincelette 1983). The remaining nine studies provided insufficient information to allow conclusive judgements with respect to whether significant differences in baseline demographic characteristics were evident between treatment groups; we thus rated these studies as having unclear risk with respect to other sources of bias (Benigno 1986; Davi 1985; Faro 1988; Hedican 1976; Hemsell 1985; Hemsell 1985a; Hemsell 1989; Janssens 1982; Schepers 1981).

Effects of interventions

See: Table 1; Table 2; Table 3; Table 4

Summary of findings for the main comparison. Antibiotics compared with placebo for prophylaxis in elective vaginal hysterectomy.

Antibiotics compared with placebo for prophylaxis in elective vaginal hysterectomy
Population: women having elective vaginal hysterectomy  Settings: hospital  Intervention: antibiotics  Comparison: placebo
Outcomes	*Illustrative comparative risks (95% CI)**		Relative effect  (95% CI)	Number of participants  (studies)	Quality of the evidence  (GRADE)	Comments
	Assumed risk	Corresponding risk
	Placebo	Antibiotics
Total postoperative infections ‐ early and late	Moderate^a		RR 0.28   (0.19 to 0.4)	293  (4 studies)	⊕⊕⊕⊝  low^b,c,f
	618 per 1000	173 per 1000  (117 to 247)
Urinary tract infection	Moderate^a		RR 0.58   (0.43 to 0.77)	1473  (8 studies)	⊕⊕⊕⊝  moderate^b
	127 per 1000	74 per 1000  (55 to 98)
Pelvic infection	Moderate^a		RR 0.28   (0.20 to 0.39)	1693  (11 studies)	⊕⊕⊕⊝  moderate^b,d
	134 per 1000	38 per 1000  (27 to 52)
Other serious infections	Moderate^a		RR 0.20   (0.01 to 4.10)	146  (1 study)	⊕⊝⊝⊝  very low^b,e
	27 per 1000	5 per 1000  (0 to 111)
Postoperative fever	Moderate^a		RR 0.43   (0.34 to 0.54)	1562  (9 studies)	⊕⊕⊕⊝  moderate^b
	219 per 1000	94 per 1000  (74 to 118)
Total adverse effects ‐ not reported	This outcome was not reported
The basis for the assumed risk* (e.g. median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on assumed risk in the comparison group and relative effect of the intervention (and its 95% CI)  CI: confidence interval; RR: risk ratio
GRADE Working Group grades of evidence  High quality: Further research is very unlikely to change our confidence in the estimate of effect  Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate  Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate  Very low quality: We are very uncertain about the estimate

Antibiotics compared with placebo for prophylaxis in elective abdominal hysterectomy
Population: women having elective abdominal hysterectomy  Settings: hospital  Intervention: antibiotics  Comparison: placebo
Outcomes	*Illustrative comparative risks (95% CI)**		Relative effect  (95% CI)	Number of participants  (studies)	Quality of the evidence  (GRADE)	Comments
	Assumed risk	Corresponding risk
	Placebo	Antibiotics
Total postoperative infections ‐ early and late	Moderate^a		RR 0.38   (0.21 to 0.67)	158  (1 study)	⊕⊕⊝⊝  low^b,c
	388 per 1000	147 per 1000  (82 to 260)
Abdominal wound infection	Moderate^a		RR 0.51   (0.36 to 0.73)	2247  (11 studies)	⊕⊕⊕⊝  moderate^b
	65 per 1000	33 per 1000  (23 to 47)
Urinary tract infection	Moderate^a		RR 0.41   (0.31 to 0.53)	2705  (11 studies)	⊕⊕⊕⊝  moderate^b
	132 per 1000	54 per 1000  (41 to 70)
Pelvic infection	Moderate^a		RR 0.50   (0.35 to 0.71)	1883  (11 studies)	⊕⊕⊕⊝  moderate^b
	83 per 1000	42 per 1000  (29 to 59)
Other serious infections	Moderate^a		RR 0.44   (0.12 to 1.69)	476  (2 studies)	⊕⊝⊝⊝  very low^b,d,e
	27 per 1000	12 per 1000  (3 to 46)
Postoperative fever	Moderate^a		RR 0.59   (0.50 to 0.70)	2394  (11 studies)	⊕⊕⊕⊝  moderate^b
	242 per 1000	143 per 1000  (121 to 169)
Total adverse effects	Moderate^a		RR 1.80   (0.62 to 5.18)	430  (2 studies)	⊕⊝⊝⊝  very low^b,e
	23 per 1000	41 per 1000  (14 to 119)
The basis for assumed risk* (e.g. median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on assumed risk in the comparison group and relative effect of the intervention (and its 95% CI)  CI: confidence interval; RR: risk ratio
GRADE Working Group grades of evidence  High quality: Further research is very unlikely to change our confidence in the estimate of effect  Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate  Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate  Very low quality: We are very uncertain about the estimate

Antibiotics compared with alternative antibiotics for prophylaxis in elective vaginal hysterectomy
Population: women having elective vaginal hysterectomy  Settings: hospital  Intervention: antibiotics  Comparison: alternative antibiotics
Outcomes	Illustrative comparative risks	Relative effect  (95% CI)	Number of participants  (studies)	Quality of the evidence  (GRADE)	Comments
	Antibiotics vs alternative antibiotics
Total postoperative infections ‐ early and late Urinary tract infection Pelvic infection Other serious infections Postoperative fever	When data were available, no evidence showed a difference between any groups compared for any of our primary outcomes, except: fewer cases of total postoperative infection and postoperative fever in women who received cephalosporin than in those who received antiprotozoal fewer cases of total postoperative infection, UTI, or postoperative fever in women receiving cephalosporin with antiprotozoal than in those receiving antiprotozoal only		cephalosporin vs penicillin (2 RCTs, 470 women) cephalosporin vs tetracycline (1 RCT, 51 women) cephalosporin vs antiprotozoal (1 RCT, 78 women) antiprotozoal vs lincosamide (1 RCT, 80 women) cephalosporin plus antiprotozoal vs cephalosporin only (1 RCT, 78 women) cephalosporin plus antiprotozoal vs antiprotozoal only (1 RCT, 78 women) penicillin plus antiprotozoal vs penicillin only (1 RCT, 18 women)	⊕⊝⊝⊝  very low^a,b
Total adverse effects	No evidence of a difference between cephalosporin and penicillin. No data available for other comparisons		cephalosporin vs penicillin (2 RCTs, 451 women)	⊕⊝⊝⊝  very low^a,b
CI: confidence interval; RCT: randomised controlled trial GRADE Working Group grades of evidence  High quality: Further research is very unlikely to change our confidence in the estimate of effect  Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate  Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate  Very low quality: We are very uncertain about the estimate

Head‐to‐head comparisons of antibiotics for prophylaxis in elective abdominal hysterectomy
Population: women having elective abdominal hysterectomy  Settings: hospital  Intervention: antibiotics  Comparison: alternative antibiotics
Outcomes	Illustrative comparative risks	Relative effect  (95% CI)	Number of participants  (studies)	Quality of the evidence  (GRADE)	Comments
	Antibiotics vs alternative antibiotics
Total postoperative infections ‐ early and late Abdominal wound infection Urinary tract infection Pelvic infection Other serious infections Postoperative fever	No clear evidence of differences between groups		cephalosporin vs penicillin (1 RCT, 220 women) penicillin plus antiprotozoal vs penicillin only (2 RCT, 155 women)	⊕⊝⊝⊝  very low^1,2
Total adverse effects	No data reported on adverse effects
CI: confidence interval; RCT: randomised controlled trial GRADE Working Group grades of evidence  High quality: Further research is very unlikely to change our confidence in the estimate of effect  Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate  Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate  Very low quality: We are very uncertain about the estimate

Date	Event	Description
1 April 2019	Amended	Correction of data errors in one included study
18 March 2008	Amended	Converted to new review format
8 April 2003	New citation required and conclusions have changed	Made substantive amendments

#	Query	Results
S55	S42 AND S54	37
S54	S43 OR S44 OR S45 OR S46 OR S47 OR S48 OR S49 OR S50 OR S51 OR S52 OR S53	1,093,627
S53	TX allocat* random*	5,654
S52	(MH "Quantitative Studies")	15,044
S51	(MH "Placebos")	9,902
S50	TX placebo*	41,786
S49	TX random* allocat*	5,654
S48	(MH "Random Assignment")	41,925
S47	TX randomi* control* trial*	114,733
S46	TX ( (singl* n1 blind) or (singl n1 mask) ) or TX ( (doubl n1 blind) or (doubl n1 mask) ) or TX ( (tripl n1 blind) or (tripl n1 mask) ) or TX ( (trebl n1 blind) or (trebl n1 mask*) )	863,204
S45	TX clinic* n1 trial*	195,043
S44	PT Clinical trial	79,858
S43	(MH "Clinical Trials+")	206,820
S42	S3 AND S41	172
S41	S4 OR S5 OR S6 OR S7 OR S8 OR S9 OR S10 OR S11 OR S12 OR S13 OR S14 OR S15 OR S16 OR S17 OR S18 OR S19 OR S20 OR S21 OR S22 OR S23 OR S24 OR S25 OR S26 OR S27 OR S28 OR S29 OR S30 OR S31 OR S32 OR S33 OR S34 OR S35 OR S36 OR S37 OR S38 OR S39	72,958
S40	TX trovafloxacin	63
S39	TX augmentin	29
S38	TX amoxicillin‐clavulanic	142
S37	TX cefotetan	41
S36	TX(cefpirome or cefazolin)	468
S35	TX(cephalosporin or piperacillin)	2,612
S34	TX(doxycycline or cefamandole)	1,877
S33	TX ceftriaxone	1,425
S32	TX(mezlocillin or moxalactam)	23
S31	TX(cefonicid or cefoperazone)	71
S30	TX(cefoxitin or clindamycin)	1,369
S29	TX (cephradine or mezlocillin)	22
S28	TX (antibiotic* N5 prophyla*)	5,303
S27	TX(cefotaxime or tinidazole)	583
S26	(MM "Antibiotic Prophylaxis")	2,178
S25	TX antimicrobial*	16,309
S24	TX anti‐bacterial agent*	12
S23	TX(antibiotic* or anti biotic*)	53,541
S22	TX penicillin	3,530
S21	TX metronidazole	1,864
S20	(MM "Antiprotozoal Agents")	500
S19	(MM "Vancomycin")	1,013
S18	(MH "Tobramycin")	318
S17	(MM "Tetracycline") OR (MH "Tetracyclines") OR (MH "Minocycline")	1,063
S16	(MM "Penicillin G+") OR (MH "Penicillins") OR (MH "Dicloxacillin")	2,251
S15	(MH "Ofloxacin")	602
S14	(MM "Gentamicins")	626
S13	(MH "Antiinfective Agents, Fluoroquinolone") OR (MH "Trovafloxacin") OR (MH "Ofloxacin") OR (MH "Alatrofloxacin")	1,345
S12	(MM "Erythromycin+") OR (MM "Clarithromycin")	1,250
S11	(MM "Doxycycline")	503
S10	(MM "Clavulanic Acid")	41
S9	(MM "Cefoxitin")	22
S8	(MM "Cefazolin")	120
S7	(MM "Cefaclor")	9
S6	(MM "Azithromycin")	614
S5	(MM "Amoxicillin")	461
S4	(MM "Antibiotics+") OR (MH "Antibiotics, Combined") OR (MH "Antibiotic Prophylaxis") OR (MH "Antibiotics, Lactam") OR (MH "Antibiotics, Antineoplastic") OR (MH "Antibiotics, Peptide") OR (MH "Antibiotics, Macrolide") OR (MH "Antibiotics, Antitubercular") OR (MH "Antibiotics, Antifungal") OR (MH "Aminoglycosides")	31,924
S3	S1 OR S2	6,646
S2	TX hysterectom*	6,646
S1	(MM "Hysterectomy+") OR (MM "Hysterectomy, Vaginal")	2,727

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Total postoperative infections ‐ early and late	4		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only
1.1 Vaginal hysterectomy	4	293	Risk Ratio (M‐H, Fixed, 95% CI)	0.28 [0.19, 0.40]
1.2 Abdominal hysterectomy	1	158	Risk Ratio (M‐H, Fixed, 95% CI)	0.38 [0.21, 0.67]
2 Abdominal wound infection	11		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only
2.1 Abdominal hysterectomy	11	2247	Risk Ratio (M‐H, Fixed, 95% CI)	0.51 [0.36, 0.73]
3 Urinary tract infection	16		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only
3.1 Vaginal hysterectomy	8	1473	Risk Ratio (M‐H, Fixed, 95% CI)	0.58 [0.43, 0.77]
3.2 Abdominal hysterectomy	12	2705	Risk Ratio (M‐H, Fixed, 95% CI)	0.41 [0.31, 0.53]
4 Pelvic infection	19		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only
4.1 Vaginal hysterectomy	11	1693	Risk Ratio (M‐H, Fixed, 95% CI)	0.28 [0.20, 0.39]
4.2 Abdominal hysterectomy	11	1883	Risk Ratio (M‐H, Fixed, 95% CI)	0.50 [0.35, 0.71]
5 Other serious infections	2		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only
5.1 Vaginal hysterectomy	1	146	Risk Ratio (M‐H, Fixed, 95% CI)	0.2 [0.01, 4.10]
5.2 Abdominal hysterectomy	2	476	Risk Ratio (M‐H, Fixed, 95% CI)	0.44 [0.12, 1.69]
6 Postoperative fever	16		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only
6.1 Vaginal hysterectomy	9	1562	Risk Ratio (M‐H, Fixed, 95% CI)	0.43 [0.34, 0.54]
6.2 Abdominal hysterectomy	11	2394	Risk Ratio (M‐H, Fixed, 95% CI)	0.59 [0.50, 0.70]
7 Total adverse effects	2		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only
7.1 Abdominal hysterectomy	2	430	Risk Ratio (M‐H, Fixed, 95% CI)	1.8 [0.62, 5.18]
8 Need for therapeutic antibiotics	9		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only
8.1 Vaginal hysterectomy	6	1309	Risk Ratio (M‐H, Fixed, 95% CI)	0.51 [0.37, 0.68]
8.2 Abdominal hysterectomy	6	1359	Risk Ratio (M‐H, Fixed, 95% CI)	0.74 [0.59, 0.93]
9 Length of hospital stay	9		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
9.1 Vaginal hysterectomy	4	853	Mean Difference (IV, Fixed, 95% CI)	‐1.35 [‐1.78, ‐0.92]
9.2 Abdominal hysterectomy	7	1510	Mean Difference (IV, Fixed, 95% CI)	‐0.59 [‐0.76, ‐0.43]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Abdominal wound infection	2		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only
1.1 Abdominal hysterectomy	2	462	Risk Ratio (M‐H, Fixed, 95% CI)	0.71 [0.32, 1.57]
2 Urinary tract infection	2		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only
2.1 Vaginal hysterectomy	2	226	Risk Ratio (M‐H, Fixed, 95% CI)	1.25 [0.51, 3.04]
2.2 Abdominal hysterectomy	1	146	Risk Ratio (M‐H, Fixed, 95% CI)	1.0 [0.34, 2.96]
3 Pelvic infection	6		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only
3.1 Vaginal hysterectomy	4	375	Risk Ratio (M‐H, Fixed, 95% CI)	0.36 [0.17, 0.75]
3.2 Abdominal hysterectomy	4	662	Risk Ratio (M‐H, Fixed, 95% CI)	0.42 [0.22, 0.83]
4 Other serious infections	2		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only
4.1 Vaginal hysterectomy	2	246	Risk Ratio (M‐H, Fixed, 95% CI)	0.25 [0.03, 2.21]
4.2 Abdominal hysterectomy	1	146	Risk Ratio (M‐H, Fixed, 95% CI)	1.0 [0.14, 6.91]
5 Postoperative fever	3		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only
5.1 Vaginal hysterectomy	2	130	Risk Ratio (M‐H, Fixed, 95% CI)	0.45 [0.21, 0.97]
5.2 Abdominal hysterectomy	1	100	Risk Ratio (M‐H, Fixed, 95% CI)	0.39 [0.18, 0.85]
6 Total adverse effects	1		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only
6.1 Abdominal hysterectomy	1	146	Risk Ratio (M‐H, Fixed, 95% CI)	2.0 [0.63, 6.35]
7 Need for therapeutic antibiotics	3		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
7.1 Vaginal hysterectomy	2	196	Risk Ratio (M‐H, Random, 95% CI)	0.55 [0.15, 1.95]
7.2 Abdominal hysterectomy	2	246	Risk Ratio (M‐H, Random, 95% CI)	0.55 [0.15, 2.02]
8 Length of hospital stay	5		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
8.1 Vaginal hysterectomy	3	276	Mean Difference (IV, Fixed, 95% CI)	‐0.86 [‐1.22, ‐0.49]
8.2 Abdominal hysterectomy	3	358	Mean Difference (IV, Fixed, 95% CI)	‐1.33 [‐1.68, ‐0.97]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Urinary tract infection	1		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only
1.1 Vaginal hysterectomy	1	80	Risk Ratio (M‐H, Fixed, 95% CI)	0.71 [0.25, 2.06]
2 Postoperative fever	1		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only
2.1 Vaginal hysterectomy	1	80	Risk Ratio (M‐H, Fixed, 95% CI)	1.0 [0.06, 15.44]
3 Length of hospital stay	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
3.1 Vaginal hysterectomy	1	80	Mean Difference (IV, Fixed, 95% CI)	‐0.40 [‐0.77, ‐0.03]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Postoperative fever	1		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only
1.1 Vaginal hysterectomy	1	78	Risk Ratio (M‐H, Fixed, 95% CI)	0.5 [0.03, 7.68]
2 Length of hospital stay	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
2.1 Vaginal hysterectomy	1	78	Mean Difference (IV, Fixed, 95% CI)	0.30 [‐0.43, 1.03]

Methods	Design: randomised double‐blinded No. eligible: not stated No. randomised: 356 No. analysed:298 Drop‐outs/withdrawals: 58 (27 piperacillin, 23 cephalothin, and 8 cefoxitin groups) were excluded for the following reasons: pre‐study administration of antibiotics (3), preoperative infection (15), dosage violation (11), total abdominal hysterectomy performed (8), failure to attend clinical follow‐up examination (21) Years of recruitment: not stated Setting: 7 study centres, United States
Participants	Inclusion criteria: scheduled to undergo vaginal hysterectomy  Exclusion criteria: receipt of antimicrobial therapy within 7 days before entrance into study, history of hypersensitivity to cephalosporin or penicillin, renal or hepatic or both, test results significantly outside normal limits, infection at time of screening for enrolment of study  Age: 19 to 80 years  Type of hysterectomy: vaginal, some with associated procedures
Interventions	Two protocols: piperacillin vs cephalothin, piperacillin vs cefoxitin Treatment 1: piperacillin (penicillin)  Treatment 2: cephalothin (first‐generation cephalosporin)  Treatment 3: cefoxitin (second‐generation cephalosporin)  Dose: 3 doses of 2 grams, same regimen for all treatment groups  Route: IV  Single/multiple doses: multiple  Duration of course of antibiotics: approx. 13 hours  Timing of doses: 2 grams in first hour, then 2 grams 6‐hourly
Outcomes	Total postoperative infections Pelvic infection Postoperative fever Adverse effects Need for therapeutic antibiotics Length of hospital stay Follow up: 3 to 10 weeks
Funding	Not stated
Notes	No SDs for LOS
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"Computer‐generated randomization schedule"
Allocation concealment (selection bias)	Low risk	"Schedule maintained by hospital pharmacy"
Blinding (performance bias and detection bias)   All outcomes	Low risk	"The investigator and staff were unaware of the antibiotic assignment"
Incomplete outcome data (attrition bias)   All outcomes	Unclear risk	Insufficient information on proportions of withdrawals and reasons for withdrawals per treatment group
Selective reporting (reporting bias)	Low risk	Outcome data available on all prespecified outcomes
Other bias	Unclear risk	"The data were identical with regard to patient selection and criteria used for evaluation..."

Methods	Design: randomised double‐blinded No. randomised: not explicitly stated No. analysed:310 Drop‐outs/withdrawals: not stated Years of recruitment: not stated Setting: not stated
Participants	Inclusion criteria: not stated  Exclusion criteria: not stated  Age: not stated  Type of hysterectomy: abdominal
Interventions	Treatment: 2 grams cefoxitin (second‐generation cephalosporin) Control: placebo Route: IM Single/multiple doses: multiple Timing of doses: 20 minutes preoperatively, then 6 and 12 hours later
Outcomes	Postoperative infection, early Urinary tract infection Postoperative fever Adverse effects (narrative data only)
Funding	Not stated
Notes	Abstract only
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Insuficient information on random sequence generation
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding (performance bias and detection bias)   All outcomes	Unclear risk	States "double‐blind"; no additional details given
Incomplete outcome data (attrition bias)   All outcomes	Unclear risk	Insufficient information on withdrawals and reasons for withdrawals
Selective reporting (reporting bias)	Unclear risk	Insufficient information to make a conclusive judgement
Other bias	Unclear risk	Insufficient information to make a conclusive judgement

Methods	Design: randomised blinded No. eligible: not stated No. randomised: not explicitly stated No. analysed: 150 Drop‐outs/withdrawals: not mentioned Years of recruitment: not stated Setting: United States (details not reported)
Participants	Inclusion criteria: women having elective abdominal hysterectomy "without standard exclusions"  Exclusion criteria: not stated  Age: not stated  Type of hysterectomy: abdominal
Interventions	Treatment 1: 2 grams cefoxitin (second‐generation cephalosporin) × 1 dose Treatment 2: 2 grams cefoxitin (second‐generation cephalosporin) × 2 doses Treatment 3: 2 grams cefoxitin (second‐generation cephalosporin) × 2 doses Route: IV Single/multiple doses: single vs multiple regimens Timing of doses: not stated
Outcomes	Postoperative infection, early and early + late Postoperative fever Need for therapeutic antibiotics Hospital length of stay ‐ data for each group not extractable Follow‐up: not stated, but states "no late infections observed for 149 women seen following surgery"
Funding	Drugs supplied by Merck, Sharp & Dohme
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Described as random ‐ no additional details
Allocation concealment (selection bias)	Unclear risk	Method not reported
Blinding (performance bias and detection bias)   All outcomes	Unclear risk	"With placebo blinding" ‐ probably double‐blinded; no additional information on outcome assessor
Incomplete outcome data (attrition bias)   All outcomes	Unclear risk	No information reported on withdrawals and reasons for withdrawals
Selective reporting (reporting bias)	Unclear risk	Insufficient information to make a conclusive judgement
Other bias	Unclear risk	Little information about eligibility criteria ‐ study applicability unclear

Methods	Design: randomised double blinded No. eligible: not stated No. randomised: 316 No. analysed: 291 primary analysis, 258 secondary (per protocol) analysis Drop‐outs/withdrawals: 25 from primary analysis (15 case records not traceable, 4 hysterectomy not performed, 5 given wrong prophylaxis) Years of recruitment: not stated Setting: 3 tertiary centres, Sweden
Participants	Inclusion criteria: women scheduled for hysterectomy  Exclusion criteria: antibiotics in previous 2 weeks, allergy to study drugs, taking anticoagulants or disulfiram, habitual alcohol abuse, breastfeeding  Age: not stated  Type of hysterectomy: abdominal
Interventions	Treatment: 500 mg metronidazole (antiprotozoal) Control: placebo Route: IV Single/multiple doses: multiple Timing of doses: during induction of anaesthesia, then 8 hours later
Outcomes	Postoperative infection, early Pelvic infection Wound infection Adverse effects (narrative data only) Follow‐up: to 6 days postoperative
Funding	Not stated
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Described as randomised; no additional details
Allocation concealment (selection bias)	Low risk	Bottles for infusion "labeled identically and only distinguished by a code number"
Blinding (performance bias and detection bias)   All outcomes	Low risk	Reported that "none of the investigators knew if the patient had got metronidazole or placebo"
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Primary analysis "based on all randomised patients from whom information was available"
Selective reporting (reporting bias)	Low risk	Outcome data available on all prespecified outcomes
Other bias	Low risk	Baseline demographic characteristics similar between treatment groups

Methods	Design:described as randomised; medical and nursing personnel not aware of study groups No. eligible: not stated No. randomised: 345 (to four different operation groups of which only two groups (abdominal hysterectomy and vaginal hysterectomy) were relevant. Number randomised to each group at baseline not reported No. analysed: 295 (Abdominal hysterectomy group = 158; vaginal hysterectomy group = 28) Drop‐outs/withdrawals: 50 (14 in treatment group 1, 18 in the other 2 groups (see below); reasons given included required antibiotics owing to intraoperative findings, did not have planned type of surgery, data missing at follow‐up) Years of recruitment: 1982 to 1983 Setting: Chelsea Hospital for Women, London
Participants	Inclusion criteria: patients for elective vaginal or abdominal hysterectomy  Exclusion criteria: not stated (2 patients with preop UTI excluded from analysis)  Age: not stated  Type of hysterectomy: abdominal or vaginal
Interventions	Treatment 1: 500 mg ampicillin + 500 mg penicillanic acid sulphone (penicillin) (with placebo suppository) Treatment 2: 500 mg ampicillin + 1 gram metronidazole (antiprotozoal) Control: placebo suppository Route: penicillin IV, metronidazole by rectal suppository Single/multiple doses: multiple Timing of doses: suppository 2 hours preoperatively, IV penicillin(s) immediately after induction of anaesthesia
Outcomes	Postoperative infections: early and early+ late Abdominal wound infection Urinary tract infection (2 participants with preop UTI excluded from analysis) Postoperative fever Pelvic/vaginal infection Follow‐up: 6 weeks
Funding	Not stated
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Described as randomised; no additional details
Allocation concealment (selection bias)	Unclear risk	Method not reported
Blinding (performance bias and detection bias)   All outcomes	Low risk	States that "the study was so designed that the medical and nursing personnel would not be aware of the group allocation of the patients studied"
Incomplete outcome data (attrition bias)   All outcomes	Unclear risk	Number of withdrawals per treatment group stated but reasons for withdrawals not reported by treatment groups
Selective reporting (reporting bias)	Unclear risk	Insufficient information to make a conclusive judgement
Other bias	Low risk	Baseline demographic characteristics similar between treatment groups

Methods	Design: randomised placebo‐controlled No. eligible: not stated No. randomised: 98 No. analysed:90 Drop‐outs/withdrawals: 8 (2 for positive preoperative urine culture in treatment group; 3 for positive preoperative urine culture, 2 for malignancy, and 1 for protocol mistake in placebo group) Years of recruitment: not stated Setting: Meir General Hospital, Israel
Participants	Inclusion criteria: women admitted for elective abdominal hysterectomy for benign condition  Exclusion criteria: antibiotics in previous 2 weeks, allergy to study drugs  Age: 46 to 48 years  Type of hysterectomy: abdominal
Interventions	Treatment 1: 15 mL co‐trimoxazole (antiprotozoal): 12000 mg sulphamethoxazole, 240 mg trimethoprim Control: placebo Route: IV Single/multiple doses: single Timing of doses: infused during last 30 minutes before surgery
Outcomes	Urinary tract infection Postoperative fever Adverse effects (narrative data only) Hospital length of stay
Funding	Not stated
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	States "randomly assigned" ‐ no other details
Allocation concealment (selection bias)	Unclear risk	Method not described
Blinding (performance bias and detection bias)   All outcomes	Unclear risk	States that "the placebo group received the placebo with the saline in the same manner"; no details reported on outcome assessor or evaluation of participants
Incomplete outcome data (attrition bias)   All outcomes	High risk	Proportions of withdrawals and reasons for withdrawals not balanced across treatment groups
Selective reporting (reporting bias)	Low risk	Data available on all prespecified outcomes
Other bias	Low risk	Baseline demographic characteristics similar between treatment groups

Methods	Design: randomised double blinded. Publication reports 2 separate studies (1 and 2), for which placebo data were pooled No. eligible: not stated No. randomised: not stated No. analysed: study 1: n = 53; study 2: n = 92 Drop‐outs/withdrawals: not reported Years of recruitment: not stated Setting: St Elisabeth Hospital, Turnhout, Belgium
Participants	Inclusion criteria: "abdominal or vaginal hysterectomy patients" ‐ but also states that patients with shaving culdotomy were eligible  Exclusion criteria: not stated  Age: not stated  Type of hysterectomy: abdominal or vaginal
Interventions	Treatment 1: 1 to 2 grams tinidazole (antiprotozoal) Control: placebo Route: oral Single/multiple doses: study 1 = multiple, study 2 = single Study 1: first dose approximately 18 hours preoperatively, second dose 6 hours later, postoperative days 3, 4, and 5: 1 dose of 1 gram daily Study 2: single preoperative 2 gram dose given 6 to 8 hours preoperatively
Outcomes	Postoperative infection, early Study reports outcomes as "wound infection morbidity" (WIM). In this review, WIM grades 2 and 3 reported (i.e. those defined in review as "clinically relevant")
Funding	Not stated
Notes	Publication also describes third study ‐ described as randomised with no mention of blinding
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	States that studies were randomised ‐ no additional details
Allocation concealment (selection bias)	Unclear risk	Method not reported
Blinding (performance bias and detection bias)   All outcomes	Low risk	States that "the double‐blind code was broken only after completion in each of the two studies..."
Incomplete outcome data (attrition bias)   All outcomes	Unclear risk	Proportions of withdrawals and reasons for withdrawals not reported across treatment groups
Selective reporting (reporting bias)	Unclear risk	Insufficient information to make a conclusive judgement
Other bias	Unclear risk	Insufficient information to make a conclusive judgement

Methods	Design: randomised double‐blinded No. women eligible: 164 No. women randomised: 100 No. women analysed: 100 Drop‐outs/withdrawals: none Years of recruitment: 1970 to 1972 Setting: University of Michigan Medical Centre
Participants	Inclusion criteria: premenopausal women having vaginal hysterectomy Exclusion criteria: allergy to study drugs, high preoperative blood urea, already receiving prophylactic antibiotics, "vaginal approach was decided upon in the operating room" Age: mean 35 years Type of hysterectomy: vaginal
Interventions	Treatment: 1 gram cephaloridine (first‐generation cephalosporin) Control: placebo Route: not stated Single/multiple doses: multiple Timing of doses: first dose on call to operating room, second dose on return from recovery room, third dose at bedtime night of operation
Outcomes	Postoperative infection, early Urinary tract infection Pelvic infection Postoperative fever Need for therapeutic antibiotics Hospital length of stay Follow‐up: to hospital discharge *Also reports "other morbidity" ‐ no separate data for "other serious infections"
Funding	Eli Lilly Company
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Random numbers table
Allocation concealment (selection bias)	Low risk	States that allocation was "assigned by the pharmacy service" ‐ probably remote allocation
Blinding (performance bias and detection bias)   All outcomes	Low risk	States that "the code identifying placebo or active drug was broken only after the patient had been discharged and the clinical summary sheets...completed"
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Reported no drop‐outs
Selective reporting (reporting bias)	Unclear risk	Insufficient information to make a conclusive judgement
Other bias	Low risk	Baseline demographic characteristics similar between treatment groups

Methods	Design: randomised double‐blinded; stratified by menopausal status No. eligible: 1511 underwent non‐radical elective hysterectomy: reasons for non‐participation stated No. randomised: 557 No. analysed: 515 Drop‐outs/withdrawals: 52 (26 in each group started on therapeutic antibiotics by surgeon) Years of recruitment: 1976 to 1978 Setting: Boston Hospital for Women, Massachusetts, USA
Participants	Inclusion criteria: all women booked for elective, non‐radical, abdominal or vaginal hysterectomy  Exclusion criteria: active infection, use of antibiotics within past 2 weeks, pelvic surgery within 2 weeks, sensitivity to study drugs  Age: mean 41 to 42 years  Type of hysterectomy: abdominal or vaginal
Interventions	Treatment: cephazolin (first‐generation cephalosporin) Control: placebo Route: IM Single/multiple doses: multiple Duration of course of antibiotics Timing of doses: first dose 1 to 2 hours preoperatively, 2 more doses at 6‐hour intervals
Outcomes	Abdominal wound infection Urinary tract infection Pelvic infection Postoperative fever Need for therapeutic antibiotics Adverse effects (narrative data only) Hospital length of stay Follow‐up: 6 weeks
Funding	Eli Lilly and Company
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Described as randomised; stratified by menopausal status: no additional details
Allocation concealment (selection bias)	Unclear risk	Method not reported
Blinding (performance bias and detection bias)   All outcomes	Low risk	States that "participants, their physicians and all investigators were blind to the allocation throughout the study"
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Proportions of withdrawals and reasons for withdrawal balanced across treatment groups
Selective reporting (reporting bias)	Low risk	Data available on all prespecified outcomes
Other bias	Low risk	Baseline demographic characteristics similar between treatment groups

Methods	Design: randomised double‐blinded (abstract only) No. eligible: not stated No. randomised: 107 No. analysed: 103 Drop‐outs/withdrawals: 4 (reasons not reported) Years of recruitment: not stated Setting: The Netherlands
Participants	Inclusion criteria: premenopausal women undergoing abdominal hysterectomy  Exclusion criteria: not stated  Age: not stated  Type of hysterectomy: abdominal
Interventions	Treatment: deposition (second‐generation cephalosporin) Control: placebo Route: IV Single/multiple doses: multiple Timing of doses: first dose 30 minutes preoperatively, second dose 6 hours later
Outcomes	Postoperative infection Adverse effects (narrative data only) Follow‐up: not stated
Funding	Not stated
Notes	No extractable data ‐ no denominators
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Described as randomised; no additional details
Allocation concealment (selection bias)	Unclear risk	Method not described
Blinding (performance bias and detection bias)   All outcomes	Unclear risk	States "double‐blind"; no additional details
Incomplete outcome data (attrition bias)   All outcomes	Unclear risk	Proportions of withdrawals and reasons for withdrawal/drop‐out not reported
Selective reporting (reporting bias)	High risk	Data not available on all prespecified outcomes; thus evidence of selective reporting
Other bias	Unclear risk	Insufficient detail to determine risk

Study	Reason for exclusion
Adno 1979	Antibiotics given > 12 hours preoperatively and 3 days postoperatively
Allen 1972	Antibiotics given for 72 hours postoperatively
Appelbaum 1978	Antibiotics given 24 hours preoperatively and 7 days postoperatively
Appelbaum 1980	Prophylaxis given for up to 48 hours postoperatively
Batres 1980	Prophylaxis were given for up to 4 days postoperatively. Only participants were blinded to treatment
Bian 1987	Prophylactic antibiotics given 48 hours before surgery
Bivens 1975	Prophylaxis given the night before surgery and postsurgical treatment continued for 48 hours
Britt 1978	Antibiotics given for 48 hours post surgery
Brouwer WK, Hoo2	Study methods did not indicate that blinding had been used
Brown 1986	Study methods did not indicate that blinding had been used. No placebo was used for the comparative group even though different regimens were provided
Brown 1988	Study not blinded for those administering treatment and for those assessing outcomes
Cartana J, Yarn2	Study methods did not indicate that blinding had been used
Chimura 1987	Postoperative antibiotics given for 5 days
Ciraru‐Vigneron 1988	Study methods did not indicate that blinding had been used
de Lalla1993	Study methods did not indicate that blinding had been used
Ferrari 1980	Study methods did not indicate that blinding had been used: 1 group received no treatment and no placebo; some participants received therapeutic antibiotics during the course of the study
Fischbach 1988	Study methods did not indicate that blinding had been used. No placebo was used for the control group
Forney 1976	Provided antibiotics before conisation > 24 hours before hysterectomy
Friese 1988	Study methods did not indicate that blinding had been used
Friese 1989	Study methods did not indicate that blinding had been used
Fujiwara 1994	Study methods did not indicate that blinding had been used. No placebo was used for the control group
Goodlin 1974	Prophylactic given the night before surgery, then for 4 days postoperatively
Gordon 1982	Study methods did not indicate that blinding had been used
Harms 1987	Study methods did not indicate that blinding had been used
Haverkorn 1987	Postoperative antibiotics given up to 6 days postoperatively
Hayashi 2000	Postoperative antibiotics given for 2 to 3 days
Hemsell 1990a	No evidence of blinding
Huang 1987	Study methods did not indicate that blinding had been used. No placebo was used for the control group
Ireland 1982	Study was not blinded. No placebo was used for the control group
Jacobson 1982	Study procedure not double‐blinded
Jennings 1978	Prophylaxis administered the night before the operation, then was carried on for > 48 hours postoperatively
Jones RN, Wojes2	Only a single‐blinded study. Treatment regimens differed between groups
Jyothi 2010	Uncertain whether this was a true randomised controlled trial or a double‐blinded study
Kauppila 1983	23% of participants not analysed
Khan 1981	More than 27% of women had repair of prolapse rather than hysterectomy
Knippenberger 1984	No dose information. Significant proportion of participants received additional postoperative antibiotics because of infectious disease
Kunz 1982	Quasi‐randomised (alternating days, allocation according to even/odd dates)
Larsson 2002	Prophylaxis given the evening before surgery and for 7 days postoperatively
Littlejohn 1985	One group received IV and the other IM; no attempt made with placebo for blinding
Luke 1999	28% of participants not analysed
Maki 1984	Comparison of cephalosporins ‐ no placebo group
Mamsen 1992	Participants with malignancy included ‐ no separate data
Mangioni 1991	Study methods did not indicate that blinding had been used
Mansani 1984	In comparative study group, antibiotics were given for 5 days postoperatively
Manthorpe 1982	Antibiotics given 1 day preoperatively and 72 hours postoperatively
Marsden 1985	Antibiotics given 16 hours preoperatively and 72 hours postoperatively
Matkaris 1991	Comparison of cephalosporins ‐ no placebo group
Mattheussens 1985	Study methods did not indicate that blinding had been used. No placebo was used for the comparative group even though different regimens were given
McDonald 1984	Study not blinded
McDonald 1988	Study not blinded; also, 1 of the treatment arms extended prophylactic antibiotics for 4 days
McGregor 1994	More than 30% of participants not analysed
Mele 1985	Prolonged antibiotic administration
Mele 1988	Prophylactic antibiotics given > 48 hours postoperatively
Mercer 1988	Study methods did not indicate that blinding had been used
Mickal 1980	Study methods did not indicate that randomisation had been used
Moroni 1979	Study methods did not indicate that blinding had been used. No placebo was used for the control group
Moroni 1984	No placebo and no blinding used
Mozzillo 1989	Study methods did not indicate that blinding had been used
Multicenter 1989	Interventions not relevant: cephalosporin vs cephalosporin (2 different generations similar in dose and route of administration)
Munck 1989	Participants included those undergoing hysterectomy for treatment of malignant disease
Ohm 1975	Antibiotic prophylaxis administered 24 hours before the operation, then for up to 5 days postoperatively
Ohm 1976	Treatment consisted of a 5‐day course of antibiotics
Ohm MJ, Galask 2	Postoperative treatment consisted of a 5‐day course of antibiotics
Ohm MJ, Galask 3	Antibiotic prophylaxis administered 24 hours before the operation, then for up to 5 days postoperatively
Olgiati 1980	Study methods did not indicate that blinding had been used
Oliva 1990	Study methods did not indicate that blinding had been used
Orr 1988	Study methods did not indicate that blinding had been used
Periti 1988	No placebo and no blinding used
Periti P, Mazze2	Treatment protocols differed for the 2 drugs; no attempt was made to blind this
Perri 1986	Antibiotic prophylaxis given up to 4 days postoperatively
Phoolcharoen 2012	Interventions not relevant: cephalosporin vs cephalosporin (2 different generations similar in dose and route of administration)
Popkin 1983	Comparison groups given prophylactic treatment the day before surgery. Blinding of treatment not attempted
Poulsen 1984	No blinding; control group given no placebo treatment
Poulsen HK, Bor2	Study methods did not indicate that blinding had been used
Queck 1991	Control group not given placebo; therefore, no attempt to blind groups
Rapp 1982	Prophylaxis administered the night before the operation, then carried on for 48 hours postoperatively
Rapp 1986	Different drug administration protocols employed. Therefore, no attempt to blind treatment groups
Regallo 1987	Study methods did not indicate that blinding had been used. No placebo mentioned even though different regimens were employed
Reggiori 1996	Study methods did not indicate that blinding had been used
Reggiori A, Rav2	In comparison group, antibiotics given for 6 days postoperatively. No attempt to blind participants or physicians
Regidor 2000	Open randomised study; therefore, not double‐blinded
Roberts 1978	Study methods did not indicate that randomisation had been used
Roy 1982	Study methods did not indicate that blinding had been used
Roy 1984	In only 1 group, antibiotics were given postoperatively. No attempt was made to blind participants or physicians by using a placebo
Roy 1988	Study methods did not indicate that blinding had been used
Roy 1989	Study methods did not indicate that blinding had been used
Roy 1990	Study methods did not indicate that blinding had been used
Roy 1998	28% of participants not analysed
Santarelli 1988	Antibiotics given for 72 hours postoperatively
Savage 1984	Antibiotics given 4 to 12 hours before surgery and 3 days after surgery
Scarpignato 1980	Antibiotic prophylaxis carried on for 5 days postoperatively in 1 group. No blinding was employed
Siekmann 1983	Blinding status unclear
Simoes 2008	Not a double‐blinded study
Stocklund 1980	Antibiotics given 12 hours before surgery and 5 days after surgery
Sutthijumroon 1990	Study not double‐blinded
Suvonnakote 1988	Antibiotics given for > 24 hours post surgery
Szalay 1996	Participants and interventions not relevant; included participants with malignancy (abdominal hysterectomy); antibiotics given to 1 group for 3 days before surgery (vaginal hysterectomy)
Tarczali 1997	Study methods did not indicate that blinding had been used. No placebo was used for the control group
Tchabo 1985	Study methods did not indicate that blinding had been used. No placebo was used for comparative group even though different regimens were employed
Turano 1992	Open randomisation
van der Linden1993	Used open randomisation technique
Vecek 1993	No information on blinding
Voss 1989	Not double‐blinded
Walker 1982	Prophylaxis given 12 to 16 hours before the operation
Wideman 1982	Blinding was not mentioned and placebo was not used
Zivny 1997	Study methods did not indicate that blinding had been used. No placebo was used for the comparative group even though different regimens were given

PERMALINK

Antibiotic prophylaxis for elective hysterectomy

Reuben Olugbenga Ayeleke

Selma M Mourad

Jane Marjoribanks

Karim A Calis

Vanessa Jordan

Abstract

Background

Objectives

Search methods

Selection criteria

Data collection and analysis

Main results

Authors' conclusions

Plain language summary

Summary of findings

Background

Description of the condition

Description of the intervention

How the intervention might work

Why it is important to do this review

Objectives

Methods

Criteria for considering studies for this review

Types of studies

Types of participants

Types of interventions

Types of antibiotics

Types of outcome measures

Primary outcomes

Secondary outcomes

Search methods for identification of studies

Electronic searches

Searching other resources

Data collection and analysis

Selection of studies

Data extraction and management

Assessment of risk of bias in included studies

Measures of treatment effect

Unit of analysis issues

Dealing with missing data

Assessment of heterogeneity

Assessment of reporting biases

Data synthesis

Subgroup analysis and investigation of heterogeneity

Sensitivity analysis

Overall quality of the body of evidence: "Summary of findings" table

Results

Description of studies

Results of the search

1.

Included studies

Study design and setting

Participants

Interventions

Outcomes

Excluded studies

Risk of bias in included studies

2.

3.

Allocation

Random sequence generation

Allocation concealment

Blinding

Incomplete outcome data

Selective reporting

Other potential sources of bias

Effects of interventions

Summary of findings for the main comparison. Antibiotics compared with placebo for prophylaxis in elective vaginal hysterectomy.

Summary of findings 2. Antibiotics compared with placebo for prophylaxis in elective abdominal hysterectomy.

Summary of findings 3. Head‐to‐head comparisons of antibiotics for prophylaxis in elective vaginal hysterectomy.

Summary of findings 4. Head‐to‐head comparisons of antibiotics for prophylaxis in elective abdominal hysterectomy.

1. Any antibiotics versus placebo

Primary outcomes

1.1 Total postoperative infections ‐ early and late

1.1. Analysis.

4.

1.1.1 Vaginal hysterectomy

1.1.2 Abdominal hysterectomy