Antibiotic prophylaxis for patients undergoing elective laparoscopic cholecystectomy

Abstract

Background

Cholecystectomy is a common surgical procedure. In the open cholecystectomy area, antibiotic prophylaxis showed beneficial effects, but it is not known if its benefits and harms are similar in laparoscopic cholecystectomy. Some clinical trials suggest that antibiotic prophylaxis may not be necessary in laparoscopic cholecystectomy.

Objectives

To assess the beneficial and harmful effects of antibiotic prophylaxis versus placebo or no prophylaxis for patients undergoing elective laparoscopic cholecystectomy.

Search methods

We searched the The Cochrane Hepato‐Biliary Group Controlled Trials Register, the Cochrane Central Register of Controlled Trials (CENTRAL) in The Cochrane Library (Issue 3, 2010), MEDLINE (1985 to August 2010), EMBASE (1985 to August 2010), SCI‐EXPANDED (1985 to August 2010), LILACS (1988 to August 2010) as well as reference lists of relevant articles.

Selection criteria

Randomised clinical trials comparing antibiotic prophylaxis versus placebo or no prophylaxis in patients undergoing elective laparoscopic cholecystectomy.

Data collection and analysis

Our outcome measures were all‐cause mortality, surgical site infections, extra‐abdominal infections, adverse events, and quality of life. All outcome measures were confined to within hospitalisation or 30 days after discharge. We summarised the outcome measures by reporting odds ratios and 95% confidence intervals (CI), using both the fixed‐effect and the random‐effects models.

Main results

We included eleven randomised clinical trials with 1664 participants who were mostly at low anaesthetic risk, low frequency of co‐morbidities, low risk of conversion to open surgery, and low risk of infectious complications. None of the trials had low risk of bias. We found no statistically significant differences between antibiotic prophylaxis and no prophylaxis in the proportion of surgical site infections (odds ratio (OR) 0.87, 95% CI 0.49 to 1.54) or extra‐abdominal infections (OR 0.77, 95% CI 0.41 to 1.46). Heterogeneity was not statistically significant.

Authors' conclusions

This systematic review shows that there is not sufficient evidence to support or refute the use of antibiotic prophylaxis to reduce surgical site infection and global infections in patients with low risk of anaesthetic complications, co‐morbidities, conversion to open surgery, and infectious complications, and undergoing elective laparoscopic cholecystectomy. Larger randomised clinical trials with intention‐to‐treat analysis and patients also at high risk of conversion to open surgery are needed.

Plain language summary

Antibiotic prophylaxis for elective gallbladder surgery

Elective gallbladder surgery is the most common elective surgical procedure in the abdomen. Antibiotic prophylaxis is a common conduct in open cholecystectomy, and it is also applied to patients undergoing laparoscopic cholecystectomy without any proof of effectiveness. Laparoscopic surgery offers some advantages related to less manipulation and shorter length of surgical wound, so antibiotic prophylaxis effect could be lower than in open surgery. This meta‐analysis of eleven randomised clinical trials could not find sufficient evidence to support or refute the use of antibiotic prophylaxis to reduce surgical site infection or global infections in patients with low anaesthetic risk, low co‐morbidities, and low‐risk of conversion to open surgery, and undergoing elective laparoscopic cholecystectomy. This is why large and well‐designed randomised trials including patients with high‐risk of conversion to open surgery should be conducted in order to define the beneficial or harmful effects of the antibiotics when given as a prophylaxis.

Background

Cholecystectomy is the universally accepted method to manage symptomatic uncomplicated cholelithiasis and other benign gallbladder diseases because it can cure the disease and has low morbidity and mortality (Keus 2010). Cholelithiasis is one of the most common abdominal diseases of adults (Bowen 1992; Angelico 1997). Although many patients are not symptomatic and gallbladder stones are an incidental finding during ultrasonography (Wenckert 1966; Gracie 1982), 1% to 4% of patients have digestive complaints that compel a treatment (NIH 1993). Furthermore, biliary dyskinesia and gallbladder polyps, although less frequent, are also indications of elective cholecystectomy (Schwesinger 1996; Mainprize 2000).

The most frequent complication in patients undergoing cholecystectomy is surgical site infection (Henry 1983). Surgical site infection was reported in 10% to 23% of the patients who had been operated on before the routine use of antibiotic prophylaxis was introduced in 1960 (Kaufman 1986; Chandrashekhar 1996). The surgical site infection increases length of stay and costs and decreases the quality of life (Strachan 1977; Morran 1978; Lykkegaard 1981; Strubbs 1983).

Since 1960, antibiotic prophylaxis has been considered as the best intervention to prevent surgical site infection in elective surgery (Strachan 1977; Morran 1978; Morran 1984; Sykes 1984; Harnoss 1985; Kaufman 1986). Antibiotic prophylaxis includes preoperative administration of wide‐spectrum antibiotics against the most frequent bacteria involved in surgical site infection, trying to get high tissue levels of the antibiotic at the surgical wound in order to avoid colonisation and growth of microorganisms (Barie 2000; Weed 2003). It is accepted that antibiotic prophylaxis must be administered in all surgical procedures classified as clean‐contaminated or in selected patients undergoing clean procedures (Bratzler 2004). Cholecystectomy is considered clean‐contaminated on the basis that the biliary tract is entered without significant spillage during the procedure. Some randomised clinical trials have confirmed that antibiotic prophylaxis in open cholecystectomy is decreasing the risk of surgical site infection (Jewesson 1996; Lippert 1998; Agrawal 1999). 
 
 Since 1986, with the original description of laparoscopic cholecystectomy, surgical treatment of benign gallbladder diseases, especially cholelithiasis, has undergone a dramatic change. Development of minimally invasive surgery for cholelithiasis showed a decrease in length of stay (Berggren 1994), costs (Srivastava 2001), postoperative pain (McMahon 1994), and an increase in quality of life of patients operated on (Barkun 1992; Keus 2006). Although this surgical method diminished the length and manipulation of the surgical wound, antibiotic prophylaxis use did not change. Recent clinical trials have re‐evaluated the usefulness of antibiotic prophylaxis in laparoscopic cholecystectomy (Illig 1997; Higgins 1999; Harling 2000; Tocchi 2000; Mahatharadol 2001). In laparoscopic cholecystectomy the incisions are smaller than in open cholecystectomy, and the laparoscopic technique is likely associated with less risk of microbiological contamination because all manipulations of the excised organ are made through a trocar that isolates the surgical wound from the gallbladder and decreases the contact of the wound with the external environment. Nonetheless, there is no consensus about these considerations, and many surgeons still use antibiotic prophylaxis in laparoscopic cholecystectomy.

We found two meta‐analyses investigating the beneficial and harmful effects of antibiotics when used as prophylaxis of surgical site infection in patients undergoing elective laparoscopic cholecystectomy (Al‐Ghnaniem 2003; Catarci 2004). The authors concluded that antibiotic prophylaxis do not seem to offer advantages, but the meta‐analyses have methodological flaws that make their conclusions prone to bias such as poor reporting of methods and inclusion of retrospective studies.

Objectives

To assess the beneficial or harmful effects of antibiotic prophylaxis in patients undergoing elective laparoscopic cholecystectomy.

Methods

Criteria for considering studies for this review

Types of studies

All randomised clinical trials comparing prophylactic antibiotics versus no antibiotics or placebo before laparoscopic cholecystectomy. Trials were included, irrespective of number of patients randomised, database registry, and language of the publication. Quasi‐randomised trials were excluded, but data on adverse events, if available, were to be recorded.

Types of participants

Adult patients (more than 17 years old) undergoing laparoscopic cholecystectomy with preoperative clinical diagnosis of cholelithiasis without acute cholecystitis or other benign non‐acute inflammatory disease of the gallbladder. Jaundiced patients were excluded.

Types of interventions

Antibiotic prophylaxis versus placebo or no antibiotic, administered intravenously or orally, prior to elective laparoscopic surgery.

Types of outcome measures

Primary outcomes

1. All‐cause mortality. 
 2. Surgical site infection defined according to the Centre of Disease Control (CDC) classification (Mangram 1999), recorded as 'yes' or 'no'. 
 3. Extra‐abdominal infections defined according to the CDC classification (Garner 1996), recorded as 'yes' or 'no'. 
 4. Adverse events, defined as allergic reactions to antibiotics (ICH‐GCP 1997). 
 5. Quality of life.

All outcome measures were confined to within hospitalisation or 30 days after discharge.

Search methods for identification of studies

Electronic searches

We searched The Cochrane Hepato‐Biliary Group Controlled Trials Register (Gluud 2010), the Cochrane Central Register of Controlled Trials (CENTRAL) in The Cochrane Library, MEDLINE, EMBASE, Science Citation Index Expanded (SCI‐EXPANDED), and LILACS (Royle 2003). Searches were not performed before 1985, as laparoscopic cholecystectomy was not used before 1985. Search strategies with the time span of the searches are given in Appendix 1.

Searching other resources

We wrote to authors of included trials for information on any published, unpublished, and ongoing trials. We also checked the reference lists of all the identified trials for relevant trials. Published abstracts from the conference proceedings from the American Digestive Disease Week (DDW) published in Gastroenterology and the United European Gastroenterology Week (UEGW) published in Gut were also handsearched. In addition, we contacted experts in the field and pharmaceutical companies and asked them to provide details of pertinent clinical trials or any relevant unpublished materials. The international societies of minimally invasive surgery (European and American Laparoscopic Surgery Society) and the international gastrointestinal surgery societies (Gastrointestinal Surgery Society) were also contacted and asked to provide information on any unpublished studies.

Data collection and analysis

We conducted the review according to our protocol (Sanabria 2004), the recommendations of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2009), and The Cochrane Hepato‐Biliary Group Module (Gluud 2010).

Selection of studies

Alvaro Sanabria (AS), Eduardo Valdivieso (EV), and Luis Dominguez (LD) extracted the data, and a consensus was sought in case of disagreement. AS and Gabriel Gomez (GG) performed the statistical analysis. We listed identified trials and made an evaluation whether the trials fulfilled the inclusion criteria. We also listed the excluded trials with the reasons for exclusion. We resolved differences among us through discussion until consensus was reached.

Data extraction and management

The following information was extracted for each trial included in this review:

1. Inclusion and exclusion criteria for patients being considered.

2. Use of prophylactic antibiotics: prophylaxis is defined as the use of antibiotic before the infection occurs and prior to surgery.

3. Number of randomised patients.

4. Number and reason of patients not randomised.

5. Exclusion after randomisation.

6. Drop‐outs.

7. Intention‐to‐treat analysis.

8. Number of laparoscopic cholecystectomies converted to open cholecystectomies.

9. Adverse events.

Data on septic and non‐septic abdominal complications (surgical site infections, biliary tract injury, biliary leakage), extra‐abdominal complications, and number of deaths were extracted when present.

The statistical package Review Manager 5 (RevMan 2008) provided by The Cochrane Collaboration was used. For dichotomous outcomes, the impact of the intervention was expressed as an odds ratio (OR) together with 95% confidence interval (CI). We did not have continuous variables in this review. We used a random‐effects (DerSimonian 1986) model and a fixed‐effect model (DeMets 1987). In case of discrepancy between the two models (eg, one giving a significant intervention effect and the other giving no significant intervention effect), we reported both results; otherwise, we reported only the results from the fixed‐effect model.

Assessment of risk of bias in included studies

Methodological quality is defined as the confidence that the design and report will restrict bias in the intervention comparison (Moher 1998). According to empirical evidence (Schulz 1995; Moher 1998; Jüni 2001; Kjaergard 2001; Wood 2008), the evaluation of the risk of bias could be achieved through assessing the following domains:

Allocation sequence generation

• Low risk of bias: sequence generation was achieved using computer random number generation or a random number table. Drawing lots, tossing a coin, shuffling cards, and throwing dice are adequate if performed by an independent adjudicator.

• Uncertain risk of bias: the trial is described as randomised, but the method of sequence generation was not specified.

• High risk of bias: the sequence generation method is not, or may not be, random. Quasi‐randomised studies, those using dates, names, or admittance numbers in order to allocate patients are inadequate and will be excluded for the assessment of benefits but not for harms.

Allocation concealment

• Low risk of bias: allocation was controlled by a central and independent randomisation unit, sequentially numbered, opaque and sealed envelopes or similar, so that intervention allocations could not have been foreseen in advance of, or during, enrolment.

• Uncertain risk of bias: the trial was described as randomised but the method used to conceal the allocation was not described, so that intervention allocations may have been foreseen in advance of, or during, enrolment.

• High risk of bias: if the allocation sequence was known to the investigators who assigned participants or if the study was quasi‐randomised. Quasi‐randomised studies will be excluded for the assessment of benefits but not for harms.

Blinding

• Low risk of bias: the trial was described as blinded, the parties that were blinded, and the method of blinding was described, so that knowledge of allocation was adequately prevented during the trial.

• Uncertain risk of bias: the trial was described as blind, but the method of blinding was not described, so that knowledge of allocation was possible during the trial.

• High risk of bias, the trial was not blinded, so that the allocation was known during the trial.

Incomplete outcome data

• Low risk of bias: the numbers and reasons for dropouts and withdrawals in all intervention groups were described or if it was specified that there were no dropouts or withdrawals.

• Uncertain risk of bias: the report gave the impression that there had been no dropouts or withdrawals, but this was not specifically stated.

• High risk of bias: the number or reasons for dropouts and withdrawals were not described.

Selective outcome reporting

• Low risk of bias: pre‐defined, or clinically relevant and reasonably expected outcomes are reported on.

• Uncertain risk of bias: not all pre‐defined, or clinically relevant and reasonably expected outcomes are reported on or are not reported fully, or it is unclear whether data on these outcomes were recorded or not.

• High risk of bias: one or more clinically relevant and reasonably expected outcomes were not reported on; data on these outcomes were likely to have been recorded.

Vested interest bias

• Low risk of bias: if there was no risk of vested interests on the side of researchers, manufacturers, or funding bodies; or any personal conflicts by the authors of the trial publication that might have unduly influenced judgements were disclosed in an honest and upright manner.

• Uncertain risk of bias: if it is not possible to say that there were or there were not any financial interests on the side of researchers, manufacturers, or funding bodies reported in the trial publications.

• High risk of bias: if there was risk for vested interests, eg, the trial was funded by a drug manufacturer, or researches had received money for the performance of the trial, and interests like these could have influenced the results of the trial report.

Trials with low risk of bias were considered trials that were judged as trials having 'low risk of bias' in all of the above specified individual domains. Trials with high risk of bias were considered those trials that were judged as trials having 'uncertain risk of bias' or 'high risk of bias' in one or more of the above specified individual domains.

Measures of treatment effect

Dichotomous data were analysed calculating the odds ratio (OR) for each trial, expressing the uncertainty with 95% confidence intervals (CI). Comparisons were made between trials evaluating antibiotic prophylaxis against no intervention or placebo administered intravenously or orally, prior to elective laparoscopic surgery.

Unit of analysis issues

Groups within the randomised clinical trial.

Dealing with missing data

Regarding the surgical site infection, we included patients with incomplete or missing data in the sensitivity analyses by imputing them according to the following scenario.

• Available case analysis: data on only those whose results are known, using as denominator the total number of patients who completed the trial. This meta‐analysis was done per‐protocol.

• Sensitivity analysis using worst‐best case analysis (Analysis 1.3) and best‐worst case analysis (Analysis 1.4).

1.3. Analysis.

Comparison 1 Antibiotic prophylaxis versus placebo or no‐prophylaxis, Outcome 3 Surgical site infection (Intention‐to‐treat analysis worst‐best case scenario).

1.4. Analysis.

Comparison 1 Antibiotic prophylaxis versus placebo or no‐prophylaxis, Outcome 4 Surgical site infection (Intention‐to‐treat analysis best‐worst case scenario).

Assessment of heterogeneity

The chi‐squared test for heterogeneity was used to provide an indication of between‐study heterogeneity (Higgins 2002). In addition, the degree of heterogeneity observed in the results was quantified using the I² statistic, which can be interpreted as the percentage of variation observed between the studies attributable to between‐study differences rather than sampling error (chance).

Assessment of reporting biases

We used a funnel plot to provide a visual assessment of whether treatment estimates are associated with trial size. The detection of biases varies with the magnitude of the treatment effect (Egger 1997; Macaskill 2001), the distribution of study size, and whether a one‐ or two‐tailed test is used (Macaskill 2001). We used two tests to assess funnel plot asymmetry, adjusted rank correlation test (Begg 1994), and regression asymmetry test (Egger 1997).

Subgroup analysis and investigation of heterogeneity

We had planned a subgroup analysis in order to compare the intervention effect in trials with low risk of bias to that of trials with high risk of bias. Furthermore, we had planned to explore causes of heterogeneity (defined as the presence of statistical heterogeneity by chi‐squared test with significance set at P value less than 0.10 and measure the quantities of heterogeneity by I² (Higgins 2002)) by comparing different groups of trials stratified according to patient risk factors, level of experience of the surgeon, and other factors that may explain heterogeneity. Values of I² were categorised according to Higgins (Higgins 2003): low, up to 25%; moderate, from 25 to 50% and high, higher than 75%.

Funnel plot 
 Funnel plot for selected trials is shown in Figure 1 and Figure 2. The adjusted rank correlation test and the regression asymmetry test were not statistically significant (intercept P value 0.42), which does not suggest bias.

1.

Funnel plot of comparison: 1 Antibiotic prophylaxis versus no‐prophylaxis, outcome: 1.1 Surgical site infection.

2.

Funnel plot of comparison: 1 Antibiotic prophylaxis versus no‐prophylaxis, outcome: 1.2 Global infections.

Sensitivity analysis

• A sensitivity analysis using worst‐best case and best‐worst case analyses.

Results

Description of studies

Results of the search

Our searches provided fifteen studies, four of which we had to exclude. Three did not fulfil our inclusion criteria: Anselmi 1995 was a quasi‐randomised trial since the generation of the allocation sequence was based on dates; Uchiyama 2003 was not a randomised trial; and Chang 2006 included patients with acute cholecystitis. The fourth, Gonzales 1996, was a thesis that we were unable to obtain. All the remaining eleven trials were randomised clinical trials with parallel designs. Ten of these trials were identified in MEDLINE and one in LILACS. Those from MEDLINE were in English, and the one from LILACS in Spanish. Authors and other sources of information did not provide further trials.

The inclusion criteria were similar for the selected trials: adult patients scheduled for elective laparoscopic cholecystectomy, with preoperative diagnosis of benign condition (gallbladder stones, chronic cholecystitis, cholesterolosis, or gallbladder polyps).

The exclusion criteria were also similar for the included trials: (a) pregnant or lactating women; (b) antibiotic allergy; (c) antibiotic therapy within 48 hours to 7 days prior to surgery; (d) clinically active infection at the moment of surgery; (e) evidence or suspicion of common bile duct stones; (f) contraindications for laparoscopic cholecystectomy; (g) indication of obligatory antibiotic prophylaxis because the medical condition was different from biliary disease (immunosuppression, corticoid use, etc). Csendes 1995, Tocchi 2000, Kuthe 2006, Uludag 2009, Yildiz 2009, and Sharma 2010 excluded patients converted to open cholecystectomy; Harling 2000 excluded those who had an operative cholangiogram performed; and Uludag 2009 excluded patients older than 60 years and with American Society of Anaesthesiology (ASA) classification higher than II. 
 
 At the entry into the trials, participants were not reported to be different regarding age, sex, severity (measured with ASA classification), and concomitant diseases. Csendes 1995, Illig 1997, Tocchi 2000, Mahatharadol 2001, Koc 2003, Uludag 2009, and Sharma 2010 reported bile spillage or intraoperative gallbladder rupture without statistically significant differences, but Yildiz 2009 found differences in this variable. Trials that reported ASA classification included mainly low‐risk patients. 
 
 Interventions were prophylactic antibiotics with one preoperative dose in eight trials (Csendes 1995; Higgins 1999; Harling 2000; Mahatharadol 2001; Kuthe 2006; Uludag 2009; Yildiz 2009; Sharma 2010), or preoperative and postoperative administration less than 24 hours in three trials (Illig 1997; Tocchi 2000; Koc 2003). We considered these differences to be clinically not relevant, because there is a policy accepting that antibiotic prophylaxis is extended up to 24 hours. The control group was given either placebo of saline solution (Higgins 1999; Tocchi 2000; Koc 2003; Kuthe 2006; Uludag 2009; Yildiz 2009; Sharma 2010) or no antibiotic (Csendes 1995; Illig 1997; Harling 2000; Mahatharadol 2001). Harling 2000 used a plastic bag to remove the gallbladder from the abdomen in all patients in the 'no antibiotic' group, Tocchi 2000 and Uludag 2009 in all patients, and Sharma 2010 used it only in patients with perforated gallbladder. The antibiotics used in the trials were cefuroxime (Harling 2000; Kuthe 2006), cefazolin (Csendes 1995; Illig 1997; Higgins 1999; Mahatharadol 2001; Uludag 2009; Yildiz 2009), cefotetan (Higgins 1999), cefotaxime (Tocchi 2000; Koc 2003) and ceftriaxone (Sharma 2010). In the Illig 1997 trial, seven patients in the control group received antibiotics against the protocol, and antibiotics were provided to patients with higher risk of infection (inflammation and conversion to open surgery). This violation could influence the results of the trial, with a bias acting as a measure that artificially decreases the difference of surgical site infection. The Higgins 1999 trial had two antibiotic groups, one with cefazolin and one with cefotetan, which were pooled together for analysis, because both antibiotics belong to the same pharmacological family. 
 
 Laparoscopic surgery included skin preparation with povidone‐iodine solution, use of three or four ports, creation of pneumoperitoneum, and extraction of gallbladder through the operating port, always using a bag (Tocchi 2000; Uludag 2009) or at the surgeon's preference as Sharma 2010 in cases of perforated gallbladder. Incisions were closed at the end of the operation with absorbable or non‐absorbable suture. Illig 1997, Harling 2000, Tocchi 2000, Koc 2003, Kuthe 2006, Uludag 2009, Yildiz 2009, and Sharma 2010 took a sample of bile for culture. Tocchi 2000 drained always the subhepatic region through the opening of a 5 mm trocar with a silicone tube, which was removed on the first postoperative day and Sharma 2010 only used a drain in cases of perforated gallbladder.

Infectious complications were defined clinically as fever with a clinical focus and culture positives from infectious sites, as recommended by the The Centers for Disease Control and Prevention (CDC) criteria for surgical site infection (Mangram 1999). Illig 1997, Harling 2000 and Uludag 2009 reported isolated microorganisms from each surgical site infection (SSI), without differences between groups.

Risk of bias in included studies

Generation of the allocation sequence was considered of low risk of bias in three trials (Tocchi 2000; Kuthe 2006; Uludag 2009) and high in the remaining eight. Allocation concealment was of low risk of bias in four trials (Higgins 1999; Harling 2000; Uludag 2009; Sharma 2010) and high in the remaining seven trials. Seven trials (Higgins 1999; Tocchi 2000; Koc 2003; Kuthe 2006; Uludag 2009; Yildiz 2009; Sharma 2010) used placebo. Blinded assessment was performed in seven trials classified as low risk of bias (Higgins 1999; Harling 2000; Tocchi 2000; Koc 2003; Kuthe 2006; Yildiz 2009; Sharma 2010) and three trials had high risk of bias regarding this component (Illig 1997; Mahatharadol 2001; Uludag 2009). Csendes 1995 did not perform blinding. All trials had a low risk of bias regarding incomplete outcome data and selective outcome reporting. All trials had a high risk of bias regarding vested interest bias.

Follow‐up was adequate in all trials, except in Illig 1997, who reported six losses to follow‐up in 250 patients, and Mahatharadol 2001, who reported one loss to follow‐up in 101 patients (Figure 3; Figure 4).

3.

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

4.

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

Intention‐to‐treat analysis for primary outcomes was made only in Illig 1997. Other trials excluded patients after randomisation, especially patients with non‐specified protocol violations and patients who underwent conversion to open surgery, which has a higher risk of surgical site infection. We have tried to contact the authors of original trials to get information about reasons for exclusion, number and distribution of patients excluded, and outcomes of these patients. As we could not obtain more information, we made a sensitivity analysis for surgical site infection, using worst‐best and best‐worst scenarios for those trials reporting number of patients excluded by group. This analysis showed a high sensitivity to post‐randomisation exclusions (Analysis 1.3 and Analysis 1.4). Finally, all trials were of high risk of bias. 
 
 Sample‐size estimation was only clearly described in Higgins 1999 and Illig 1997. The other trials did not report on any form of sample size calculation.

Effects of interventions

In total, 1664 patients were recruited in the eleven trials; 900 in the prophylaxis group and 764 in the no‐prophylaxis group. The rate of conversion to open surgery was 2.4% in Higgins 1999, 8.4% in Illig 1997, 3.8% in Harling 2000, and 4.1% in Uludag 2009. Koc 2003 excluded 20 of the 112 (17.8%) patients originally included in the trial because of protocol violations. Higgins 1999 excluded 38 of 450 (8.4%) patients because of protocol violations. Harling 2000 excluded 30 of 106 (28.3%) patients because of the need of cholangiography, concomitant antibiotic usage, spillage of bile from gallbladder, and conversion to open surgery. Mahatharadol 2001 excluded two of 102 (1.9%) patients because of lost to follow‐up and protocol violation. Kuthe 2006 excluded 11 of 104 (1.5%) patients because of intra‐operative spillage, conversion to open cholecystectomy, and refusal to continue in the trial, and Uludag 2009 excluded 6 of 144 (4.1%) patients because of conversion to open surgery. Tocchi 2000 excluded 6 of 90 (6.6%) patients, but he did not state the reasons. The other trials did not report the number of patients that were lost to follow‐up. As only one trial had an intention‐to‐treat analysis, we performed two sensitivity analyses, using a worst‐best case (a poor outcome is assigned to each treated patient for whom data were missing) and a best‐worst case (a good outcome is assigned to each treated patient for whom data were missing) scenarios. 
 
 All‐cause mortality 
 Mortality was not reported in any of the trials. 
 
 Surgical site infection 
 The number of surgical site infections was similar in the two groups: 24 of 900 (2.7%) patients in the prophylaxis group had a surgical site infection against 25 of 764 (3.3%) in the no‐prophylaxis group. The OR was 0.87, 95% CI (0.49 to 1.54). No statistically significant differences or heterogeneity were observed (Analysis 1.1). As I² values were lower than 2%, the planned analyses for heterogeneity were not necessary. The sensitivity analysis with the worst‐best case scenario yields an OR of 2.80, 95% CI 1.75, 4.46, and the best‐worst case scenario yields an OR of 0.34, 95% CI (0.21 to 0.54), showing a high sensitivity to post‐randomisation exclusions. 
 
 Extra‐abdominal infections 
 Seven trials reported infectious events that are different from surgical site infections (Csendes 1995; Illig 1997; Higgins 1999; Tocchi 2000; Kuthe 2006; Uludag 2009; Sharma 2010), and these occurred in 7 of 657 (1.0%) patients in the prophylaxis group and 22 of 531 (1.8%) patients in the no‐prophylaxis group. For the Tocchi 2000 trial, we counted only four complications, although the trial refers to five events. This is because a subhepatic collection secondary to technical defect cannot be assumed as a direct outcome for antibiotic prophylaxis. The OR was 0.66 (95% CI 0.25 to 1.74). No statistically significant differences or heterogeneity were observed (Analysis 1.2). As I² values were lower than 2%, the planned analyses for heterogeneity were not necessary. 
 
 Adverse events 
 Information about adverse events was not given in any of the trials. 
 
 Quality of life 
 None of the trials reported on the quality of life.

1.1. Analysis.

Comparison 1 Antibiotic prophylaxis versus placebo or no‐prophylaxis, Outcome 1 Surgical site infection.

1.2. Analysis.

Comparison 1 Antibiotic prophylaxis versus placebo or no‐prophylaxis, Outcome 2 Extra‐abdominal infections.

Discussion

Our systematic review could not find evidence to support or refute the use of antibiotic prophylaxis for reduction of surgical site infections in patients at low risk of acquiring infections or complications when undergoing elective laparoscopic cholecystectomy when prophylactic antibiotics were administered.

Preliminary observations showed a decrease in surgical site infection after beginning of laparoscopic cholecystectomy when antibiotic prophylaxis was applied. So, some randomised clinical trials were designed to assess the effectiveness of antibiotic prophylaxis (Anselmi 1995; Csendes 1995; Illig 1997; Higgins 1999; Harling 2000; Tocchi 2000; Mahatharadol 2001; Koc 2003). However, because of the low frequency of surgical site infections and complications, each one of the trials had a small sample size to definitely probe a positive effect. Previously, Al‐Ghnaniem 2003 and Catarci 2004 reported meta‐analyses on this topic, but the first was published only as an abstract and did not allow an adequate quality assessment to be made, and the second included a retrospective study and did not include other relevant trials.

Our systematic review that includes eleven randomised clinical trials was not able to find significant statistical differences in the occurrence of surgical site infection and global infections between patients undergoing laparoscopic cholecystectomy who received and those who did not receive prophylactic antibiotics. However, it must be highlighted that the number of participants included in this systematic review is lower than those needed to show a real equivalence between treatments because of the low rate of outcomes. Another fact that must be realised is the lack of intention‐to‐treat analysis in all but one of the analysed trials. This is important to stress because trials with no intention‐to‐ treat analyses tend to produce misleading results (Moher 1998). Most of the trials analysed included only patients with no protocol violations and without conversion to open surgery. This has created a selection of patients with a very low risk of surgical site infection ‐ a reason that can explain the results for no difference between treatments. The sensitivity analysis showed a high dependence of results related with post‐randomisation exclusions, which supports even more the potential risk of bias of our finding. If one shall apply these findings to all patients who are going to undergo elective laparoscopic cholecystectomy is still a matter of personal judgement until a definite recommendation could be given.

Conductance of more clinical trials with low risk of bias must be considered in order to increase the sample size and to surpass methodological weaknesses found in this review. A future randomised trial will need a sample size of approximately 3500 patients by intervention group. Neither clinical heterogeneity nor statistical heterogeneity were identified in the relevant clinical variables of the trials included, so it could be that the compared patients were similar. Since many of the risk of bias criteria were not clearly reported, we cannot be sure that bias did not have an effect in the trials. We believe that lost to follow‐up, found in Illig 1997 and Mahatharadol 2001, was minimal and do not introduce important bias to the results, but as previously commented, the lack of intention‐to‐treat analysis makes this conclusion difficult to apply to all candidates to laparoscopic cholecystectomy.

Another important point to discuss is the individual risk of the patients. The included trials had selected mainly patients at low risk of acquiring an infection, making its management reasonably simple, and the life of the patient was not compromised. However, this statement cannot be applied to high‐risk patients as those prone to conversion to open surgery, those with co‐morbidities and those with high anaesthetic risk scores as ASA, for whom these conditions are not accomplished, and an infection could cause the death of the patient. In relation to the generalisability of the results, it is important to highlight this concern. New trials including patients at high risk of acquiring infections from the laparoscopic cholecystectomy intervention are necessary to confirm the finding of no difference in effect. The existing clinical trials suggest that antibiotic prophylaxis in laparoscopic cholecystectomy may not be needed for patients at low risk of acquiring infections or complications and with low risk of conversion to open surgery, and that each patient undergoing surgery must be assessed carefully when selecting the therapy.

Authors' conclusions

Implications for practice.

With the present data, the evidence does not support or refute the use of antibiotic prophylaxis to reduce surgical site infection in low‐risk patients undergoing laparoscopic cholecystectomy. The use of prophylactic antibiotics administration must be selected individually in those patients with medium or high risk.

Implications for research.

It is necessary to conduct randomised clinical trials with large sample sizes to get enough number of patients and outcomes to probe the equivalence between treatments. Moreover, it is also necessary to asses this assumption in high‐risk patients. Reporting requirements of randomised clinical trials as suggested by the CONSORT statement (www.consort‐statement.org) should be strictly applied, since lack of good reporting could wrongly decrease the methodological quality of surgical randomised clinical trials.

Appendices

Appendix 1. Search Strategies

Database	Period	Search strategy
The Cochrane Hepato‐Biliary Group Controlled Trials Register	August 2010	((antibiotic* OR prophyla*) AND (laparoscop* AND cholecystectom*)) AND #20 >=1985
The Cochrane Central Register of Controlled Trials in The Cochrane Library	Issue 3, 2010	#1 MeSH descriptor Antibiotic Prophylaxis explode all trees #2 antibiotic* OR prophyla* #3 (#1 OR #2) #4 MeSH descriptor Laparoscopy explode all trees #5 laparoscop* AND cholecystectom* #6 (#4 OR #5) #7 (#3 AND #6), from 1985 to 2010
MEDLINE (Ovid SP)	1985 to August 2010	1. exp Antibiotic Prophylaxis/ 2. (antibiotic* or prophyla*).mp. [mp=title, original title, abstract, name of substance word, subject heading word, unique identifier] 3. 1 or 2 4. exp Laparoscopy/ 5. (laparoscop* and cholecystectom*).mp. [mp=title, original title, abstract, name of substance word, subject heading word, unique identifier] 6. 4 or 5 7. 3 and 6 8. (random* or blind* or placebo* or meta‐analysis).mp. [mp=title, original title, abstract, name of substance word, subject heading word, unique identifier] 9. 7 and 8 10. limit 9 to yr="1985 ‐Current"
EMBASE (Ovid SP)	1985 to August 2010	1. exp antibiotic prophylaxis/ 2. (antibiotic* or prophyla*).mp. [mp=title, abstract, subject headings, heading word, drug trade name, original title, device manufacturer, drug manufacturer] 3. 1 or 2 4. exp LAPAROSCOPY/ 5. (laparoscop* and cholecystectom*).mp. [mp=title, abstract, subject headings, heading word, drug trade name, original title, device manufacturer, drug manufacturer] 6. 4 or 5 7. 3 and 6 8. (random* or blind* or placebo* or meta‐analysis).mp. [mp=title, abstract, subject headings, heading word, drug trade name, original title, device manufacturer, drug manufacturer] 9. 7 and 8 10. limit 9 to yr="1985 ‐Current"
LILACS	1985 to August 2010	#1 "COLECISTECTOMIA" #2 "LAPAROSCOPIA" #3 "ANTIBIOTICOS" #4 "PROFILAXIS" #5 #1 OR #2 #6 #3 OR #4 #7 #5 AND #6
Science Citation Index Expanded (http://apps.isiknowledge.com)	1985 to August 2010	# 5 #4 AND #3 # 4 TS=(random* or blind* or placebo* or meta‐analysis) # 3 #2 AND #1 # 2 TS=(laparoscop* and cholecystectom*) # 1 TS=(antibiotic* or prophyla*)

Data and analyses

Comparison 1. Antibiotic prophylaxis versus placebo or no‐prophylaxis.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Surgical site infection	11	1664	Odds Ratio (M‐H, Fixed, 95% CI)	0.87 [0.49, 1.54]
1.1 Per protocol analysis (Intention‐to‐treat analysis not performed)	11	1664	Odds Ratio (M‐H, Fixed, 95% CI)	0.87 [0.49, 1.54]
2 Extra‐abdominal infections	7	1188	Odds Ratio (M‐H, Fixed, 95% CI)	0.66 [0.25, 1.74]
3 Surgical site infection (Intention‐to‐treat analysis worst‐best case scenario)	11	1756	Odds Ratio (M‐H, Fixed, 95% CI)	2.80 [1.75, 4.46]
4 Surgical site infection (Intention‐to‐treat analysis best‐worst case scenario)	11	1756	Odds Ratio (M‐H, Fixed, 95% CI)	0.34 [0.21, 0.54]

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Csendes 1995.

Methods	Randomised clinical trial with a parallel design. Generation of the allocation sequence: unclear. Allocation concealment: unclear. Blinding: not performed. Outcome assessment was made by surgeons of the treatment group, but it was not stated if they were blinded. Follow‐up: adequate; without lost to follow‐up at 15 days. Intention‐to‐treat analysis: no. Sample size calculations: no.
Participants	105 patients randomised (50 for antibiotic group and 55 for no‐antibiotic group). Sex: 32 men/73 women. Mean age 48.1 to 49.3. Age of inclusion not reported. Inclusion criteria: clinical diagnosis of chronic cholecystitis undergoing elective cholecystectomy. Exclusions criteria: conversion to open surgery and acute cholecystitis. Trial duration: 14 months.
Interventions	Intravenous single cefazolin dose, 1 g given at the time of induction compared with no antibiotic in the control group.
Outcomes	Surgical site infection determined by pus drainage and extra‐abdominal infections.
Notes	Contacted authors April 9/2007. It was impossible to get more information about excluded patients because data were not available.
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Unclear risk	Unclear. "Randomly distributed into two groups", "Se dividieron en dos grupos al azar".
Allocation concealment?	Unclear risk	Unclear. "Randomly distributed into two groups", "Se dividieron en dos grupos al azar".
Blinding? All outcomes	High risk	"Patients were controlled during hospital stay and then, 15 days after discharge in an out‐patient setting, recording data about infection", "Los pacientes fueron controlados en su estadia intrahospitalaria y luego hasta 15 dias en forma ambulatoria consignando la presencia de cualquier infeccion"
Incomplete outcome data addressed? All outcomes	High risk	Per protocol analysis.
Vested interest bias	Unclear risk	It is not possible to say if there were or there were not any financial interests.

Harling 2000.

Methods	Randomised clinical trial with a parallel design. Generation of the allocation sequence: unclear. Allocation concealment: adequate; sealed envelopes. Blinding: single‐blind. Outcome assessment was made by surgeons of the treatment group blind to intervention. Follow‐up: adequate: without lost to follow‐up (time not stated). Intention‐to‐treat analysis: no. Sample size calculations: no.
Participants	76 patients (39 for antibiotic group and 37 for no‐antibiotic group). Sex distribution not reported. Mean age not reported. Age of inclusion not reported. Inclusion criteria: patients undergoing elective cholecystectomy. Exclusions criteria: need to carry out cholangiography, concomitant usage of antibiotics, spillage of gallbladder content, and conversion to open surgery. Trial duration: 14 months.
Interventions	Intravenous single cefuroxime dose, 750 mg given at the time of induction compared with removing of the gallbladder with a plastic bag.
Outcomes	Surgical site infection determined by pus drainage.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Unclear risk	Unclear."Patients ... were randomised to receive...".
Allocation concealment?	Low risk	"Randomisation ... was done using a sealed envelope technique".
Blinding? All outcomes	Low risk	"An infection control sister followed up the patients".
Incomplete outcome data addressed? All outcomes	High risk	Per protocol analysis.
Vested interest bias	Unclear risk	It is not possible to say if there were or there were not financial interests.

Higgins 1999.

Methods	Randomised clinical trial with a parallel design. Generation of the allocation sequence: unclear. Allocation concealment: adequate; sealed envelopes. Blinding: double‐blind. Use of placebo. Outcome assessment was made by surgeons of the treatment group blind to intervention. Follow‐up: adequate: without lost to follow‐up at 30 days. Intention‐to‐treat analysis: no. Sample size calculations: yes. Power of 80%, alfa error of 5%.
Participants	412 patients (277 for antibiotics groups and 135 for placebo). Sex distribution not reported. Mean age: 47.1 to 48.2. Age of inclusion 18 to 80. Inclusion criteria: patients scheduled to elective cholecystectomy. Exclusions criteria: pregnant or lactating women, antibiotic allergy, previous antibiotic therapy, acute cholecystitis or cholangitis, jaundice, previous biliary surgery, choledocholithiasis, prosthetic valves, contraindication to laparoscopy. Trial duration: 25 months.
Interventions	Intravenous single cefotetan and cefazolin dose, 1 g each one given at the time of induction compared with placebo.
Outcomes	Surgical site infection determined by pus drainage and extra‐abdominal infections "any infection remote to the surgical site".
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Unclear risk	Unclear.
Allocation concealment?	Low risk	Sealed envelopes.
Blinding? All outcomes	Low risk	Outcome assessment was made by surgeons of the treatment group blind to intervention.
Incomplete outcome data addressed? All outcomes	Low risk	Per protocol analysis.
Vested interest bias	Unclear risk	It is not possible to say if there were or there were not financial interests.

Illig 1997.

Methods	Randomised clinical trial with a parallel design. Generation of the allocation sequence: unclear. Allocation concealment: unclear. Blinding: not performed. Outcome assessment was made by surgeons of the treatment group, but it was not stated whether they were blinded. Follow‐up: adequate. Lost to follow‐up of 2.4% at 30 days. Intention‐to‐treat analysis: yes. Sample size calculations: yes. Power of 80%, alfa error of 5%, absolute difference to identify of 4%.
Participants	250 patients (128 for antibiotic group and 122 for no‐antibiotic group). Sex: 47 men/203 women. Mean age: 46.5 to 47.3. Age of inclusion not reported. Inclusion criteria: patients undergoing elective cholecystectomy. Exclusions criteria: acute cholecystitis, obstructive jaundice, immunosuppression, pregnancy, artificial device or graft in place, use of antibiotics 7 days prior to surgery. Trial duration: 25 months.
Interventions	Intravenous preoperative dose of cefazolin, 1 g given at the time of induction and followed by two doses at 8 and 16 hours postoperatively compared with no antibiotic in the control group.
Outcomes	Surgical site infection determined by pus drainage and extra‐abdominal infections "Lower urinary tract infection (UTIs, with symptomatic bacteriuria)...and any other problem with local effects only ...".
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Unclear risk	Patients were randomised to receive prophylactic antibiotics.
Allocation concealment?	Unclear risk	Patients were randomised to receive prophylactic antibiotics.
Blinding? All outcomes	High risk	Complications were reported to the investigators as they occurred.
Incomplete outcome data addressed? All outcomes	Low risk	Intention to treat analysis.
Vested interest bias	Unclear risk	It is not possible to say if there were or there were not financial interests.

Koc 2003.

Methods	Randomised clinical trial with a parallel design. Generation of the allocation sequence: unclear. Allocation concealment: unclear. Blinding: yes. Use of placebo. Outcome assessment was made by surgeons of the treatment group, blinded to the interventions. Follow‐up: adequate. No losses to follow up. Intention‐to‐treat analysis: no. Sample size calculations: no.
Participants	92 patients (49 for antibiotic group and 43 for no‐antibiotic group). Sex: 33 men/59 women. Mean age: 47.5 to 50.1. Age of inclusion reported as adults. Inclusion criteria: patients undergoing elective cholecystectomy. Exclusions criteria: antibiotic allergy, acute cholecystitis, previous biliary surgery, obstructive jaundice, immunosuppression, artificial device or graft in place, use of antibiotics 7 days prior to surgery. Trial duration: 12 months.
Interventions	Intravenous preoperative dose of cefotaxime, 2 g given at the time of induction and followed by a dose at 24 hours postoperatively compared with placebo in the control group.
Outcomes	Surgical site infection determined by pus drainage.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Unclear risk	After the patients were confirmed for the study, they were randomised....
Allocation concealment?	Unclear risk	After the patients were confirmed for the study, they were randomised... .
Blinding? All outcomes	Low risk	Both the surgical team and the patients were blinded to the groups.
Incomplete outcome data addressed? All outcomes	High risk	Per protocol analysis.
Vested interest bias	Unclear risk	It is not possible to say if there were or there were not financial interests.

Kuthe 2006.

Methods	Randomised clinical trial with a parallel design. Generation of the allocation sequence: Random number table. Allocation concealment: unclear. Blinding: yes. Use of placebo. Outcome assessment was made by residents of the treatment group, blinded to the interventions. Follow‐up: adequate. No losses to follow up. Intention‐to‐treat analysis: No. Sample size calculations: reported, but not clear.
Participants	93 patients (40 for antibiotic group and 53 for no‐antibiotic group). Sex: 25 men/68 women. Mean age: 42.04 to 42.38. Age of inclusion reported as adults. Inclusion criteria: patients undergoing elective cholecystectomy. Exclusions criteria: antibiotic allergy, acute cholecystitis, previous biliary surgery, obstructive jaundice, immunosuppression, artificial device or graft in place, use of antibiotics 7 days prior to surgery, conversion to open surgery. Trial duration: 12 months.
Interventions	Intravenous preoperative dose of cefotaxime, 1.5 g given at the time of induction compared with placebo in the control group.
Outcomes	Surgical site infection determined by pus drainage and extra‐abdominal infections (pyrexia of more than 38 C (excluding the first postoperative day), positive bacteriological culture from possible infection sites such as wounds, the urinary or respiratory tract..."
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Low risk	Random number table.
Allocation concealment?	Unclear risk	Unclear.
Blinding? All outcomes	Low risk	Outcome assessment was made by residents of the treatment group, blinded to the interventions.
Incomplete outcome data addressed? All outcomes	High risk	Per protocol analysis.
Vested interest bias	Unclear risk	It is not possible to say that there were or there were not financial interests.

Mahatharadol 2001.

Methods	Randomised clinical trial with a parallel design. Generation of the allocation sequence: adequate. Allocation concealment: unclear. Blinding: no. Outcome assessment was made by surgeons of the treatment group, but it was not stated if they were blinded. Follow‐up: adequate. Lost to follow‐up of 0.9%. Intention‐to‐treat analysis: no. One patient who was lost to follow‐up was excluded from the analysis. Sample size calculations: no.
Participants	100 patients (50 for antibiotic group and 50 for no‐antibiotic group). Sex: 23 men/77 women. Mean age: 51 to 52.2. Age of inclusion: 15 to 80 years old. Inclusion criteria: patients undergoing elective cholecystectomy. Exclusions criteria: antibiotic allergy, acute cholecystitis, obstructive jaundice, immunosuppression, artificial device or graft in place, use of antibiotics 48 hours prior to surgery. Duration of treatment: 7 months.
Interventions	Intravenous preoperative dose of cefazolin, 1 g given at the time of induction compared with no antibiotic in the control group.
Outcomes	Surgical site infection determined by pus drainage.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Unclear risk	By block randomisation.
Allocation concealment?	Unclear risk	By block randomisation.
Blinding? All outcomes	High risk	All patients were followed‐up for 30 days after the procedure at the out‐patient clinic or by telephone contact.
Incomplete outcome data addressed? All outcomes	High risk	Per protocol analysis.
Vested interest bias	Unclear risk	It is not possible to say if there were or there were not financial interests.

Sharma 2010.

Methods	Randomised clinical trial with a parallel design. Generation of the allocation sequence: adequate. Allocation concealment: adequate. Blinding: yes. Use of placebo. Outcome assessment was made by residents of the treatment group, blinded to the interventions. Follow‐up: adequate. Lost to follow‐up of 0%. Intention‐to‐treat analysis: no. Sample size calculations: no.
Participants	100 patients (50 for antibiotic group and 50 for no‐antibiotic group). Sex: 10 men/90 women. Mean age: 39 to 39. Age of inclusion: Older than 18 years old. Inclusion criteria: patients undergoing elective cholecystectomy for symptomatic gallstone disease. Exclusions criteria: jaundice, acute cholecystitis, cholangitis, acute pancreatitis or other acute inflammation, conversion to open cholecystectomy, immunosuppressive therapy, cardiac disorders mandating prophylactic use of antibiotics, or antibiotic use in the preceding seven days. Duration of treatment: not reported.
Interventions	Intravenous preoperative dose of ceftriaxone, 1 g given at the time of induction compared with physiologic saline as placebo in the control group.
Outcomes	"Superficial SSI was defined as erythema or purulent discharge at the surgical site above the deep fascia. A deep infection was defined as purulent material deep to the fascia or near the gallbladder fossa. A distant infection was defined as any infection remote from the surgical site."
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Unclear risk	"Anesthetist randomly opened one of a collection of sealed envelopes"
Allocation concealment?	Low risk	"Anesthetist randomly opened one of a collection of sealed envelopes", "The surgeon and medical staff were unaware of which treatment the patient received."
Blinding? All outcomes	Low risk	"The surgeon and medical staff were unaware of which treatment the patient received."
Incomplete outcome data addressed? All outcomes	High risk	Per protocol analysis.
Vested interest bias	Unclear risk	It is not possible to say if there were or there were not financial interests.

Tocchi 2000.

Methods	Randomised clinical trial with a parallel design. Generation of the allocation sequence: adequate. Computer randomisation. Allocation concealment: unclear. Blinding: yes. Use of placebo. Outcome assessment was made by surgeons of the treatment group who were blinded to the intervention. Follow‐up: adequate. No losses to follow up. Intention‐to‐treat analysis: no. One patient who was lost to follow‐up was excluded from the analysis. Sample size calculations: no.
Participants	84 patients (44 for antibiotic group and 40 for no‐antibiotic group). Sex: 33 men/51 women. Mean age: 49.5 to 53.6. Age of inclusion not stated. Inclusion criteria: patients undergoing elective cholecystectomy. Exclusions criteria: antibiotic allergy, acute cholecystitis, obstructive jaundice or choledocholithiasis, previous ERCP, corticosteroid therapy, use of antibiotics 7 days prior to surgery, conversion to open surgery. Trial duration: 2 years.
Interventions	Intravenous preoperative dose of cefotaxime, 2 g given at the time of induction and followed by a dose at 24 hours postoperatively compared with placebo in the control group.
Outcomes	Surgical site infection determined by pus drainage and extra‐abdominal infections "infectious complications were defined as pyrexia with a body temperature higher than 380 C twice a day (excluding the first postoperative day) and culture findings positive for pathogens from infectious sites such as wounds, the urinary or respiratory tract...".
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Low risk	Compute‐matrix randomisation.
Allocation concealment?	Unclear risk	Compute‐matrix randomisation.
Blinding? All outcomes	Low risk	Patients were followed postoperatively.
Incomplete outcome data addressed? All outcomes	High risk	Per protocol analysis.
Vested interest bias	Unclear risk	It is not possible to say if there were or there were not financial interests.

Uludag 2009.

Methods	Randomised clinical trial with a parallel design. Generation of the allocation sequence: adequate. Random selection. Allocation concealment: Closed envelopes. Blinding: No. Use of placebo. Outcome assessment was made by surgeons of the treatment group who were not blinded to the intervention. Follow‐up: adequate. No losses to follow up. Intention‐to‐treat analysis: no. Patients converted to open cholecystectomy were excluded Sample size calculations: no.
Participants	144 patients (68 for antibiotic group and 76 for no‐antibiotic group). Sex: 22 men/122 women. Mean age: 42.5 to 44.6. Age of inclusion not stated. Excluded patients older than 60 years Inclusion criteria: patients undergoing elective cholecystectomy. Exclusions criteria: patients older than 60 years; antibiotic intake in the 7 days prior to surgery; acute cholecystitis in the 6 months prior to the procedure; evidence of cholangitis and/or obstructive jaundice and biliary pancreatitis; regular corticosteroid therapy; pregnancy or lactation; previous biliary tract surgery or previous endoscopic retrograde cholangiopancreatography; presence of American Society of Anesthesiologists classification (ASA) higher than score II; evidence of diabetes mellitus; body mass index higher than 30; and conversion to open cholecystectomy. Trial duration: 3 years.
Interventions	Intravenous preoperative dose of cefazolin, 1 g given at the time of induction compared with placebo in the control group.
Outcomes	Surgical site infection determined by pus drainage and extra‐abdominal infections "number of septic complications".
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Low risk	Random selection "One of 2 packages containing either cefazolin or placebo was chosen randomly by the anaesthesiologist for each patient."
Allocation concealment?	Low risk	Closed envelope method "The package was opened by the anaesthesiologist, and the type of solution administered (cefazolin or placebo) was recorded. The medical staff and the patient were unaware of the content of the solution."
Blinding? All outcomes	Unclear risk	Patients were followed postoperatively by the same surgical group.
Incomplete outcome data addressed? All outcomes	High risk	Per protocol analysis.
Vested interest bias	Unclear risk	It is not possible to say if there were or there were not financial interests.

Yildiz 2009.

Methods	Randomised clinical trial with a parallel design. Generation of the allocation sequence: adequate. Unclear. Allocation concealment: unclear. Blinding: yes. Use of placebo. Outcome assessment was made by surgeons of the treatment group who were blinded to the intervention. Follow‐up: adequate. No losses to follow up. Intention‐to‐treat analysis: no. One patient who was lost to follow‐up was excluded from the analysis. Sample size calculations: no.
Participants	208 patients (105 for antibiotic group and 103 for no‐antibiotic group). Sex: 58 men/150 women. Mean age: 49.9 to 51.3. Age of inclusion not stated. Inclusion criteria: patients undergoing elective cholecystectomy. Exclusions criteria: conversion to laparotomy, acute cholecystitis, evidence of obstructive jaundice, history of biliary tract surgery, prosthetic valves, chronic hepatic diseases, acute pancreatitis, immunosuppression, steroid therapy, chronic systemic infections, allergy to β‐lactam antibiotics. Trial duration: 3 years.
Interventions	Intravenous preoperative dose of cefazolin, 1 g given at the time of induction compared with placebo in the control group.
Outcomes	Surgical site infection determined by body temperature higher than 38°C, purulent discharge from the incisions, and any abdominal signs of infection.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Unclear risk	Unclear.
Allocation concealment?	Unclear risk	Unclear.
Blinding? All outcomes	Low risk	"The patients were followed up regularly until the day of discharge and on 7th and 30th days postoperatively by the same surgical team (Dr. OA and Dr. EH) who were blind to AP."
Incomplete outcome data addressed? All outcomes	High risk	Per protocol analysis.
Vested interest bias	Unclear risk	It is not possible to say if there were or there were not financial interests.

Characteristics of excluded studies [ordered by study ID]

Study	Reason for exclusion
Anselmi 1995	Quasi‐randomised trial. Randomisation was performed after date of birth.
Chang 2006	Included patients with acute cholecystitis.
Gonzales 1996	Published as a theses. We could not obtain copy of the thesis.
Uchiyama 2003	Comparative, not randomised trial.

Differences between protocol and review

As the protocol for this review was published in 2004, we had to update it following the guidelines in the Reviewer's Handbook (Higgins 2009) before starting our work on the review. The changes we have made refer to the risk of bias and order of the outcomes.

Contributions of authors

AE Sanabria: conceiving, designing, and co‐ordination of the review.

AE Sanabria, E Valdivieso, and G Gomez: developing of search strategy; undertaking searches; screening search results; management of papers retrieval; screening of retrieved papers against inclusion criteria; writing to authors of papers for additional information; collecting data for the review; providing of additional data about papers; management of data for the review; appraising quality of papers; abstracting data from papers; analysing of data; interpreting data; providing a methodological perspective; providing a clinical perspective; and writing the review.

L Dominguez: undertaking searches; screening search results; screening of retrieved papers against inclusion criteria; collecting data for the review; providing of additional data about papers; appraising quality of papers; abstracting data from papers; interpreting data; providing a clinical perspective; and writing the review.

Sources of support

Internal sources

• Pontificia Universidad Javeriana‐Hospital Universitario San Ignacio, Colombia.

• Universidad de La Sabana, Colombia.

External sources

• No sources of support supplied

Declarations of interest

None known.

New