Antibiotics versus placebo for prevention of postoperative infection after appendicectomy

Betina Ristorp Andersen; Finn Lasse Kallehave; Henning Keinke Andersen

doi:10.1002/14651858.CD001439.pub2

. 2005 Jul 20;2005(3):CD001439. doi: 10.1002/14651858.CD001439.pub2

Antibiotics versus placebo for prevention of postoperative infection after appendicectomy.

Betina Ristorp Andersen ¹, Finn Lasse Kallehave ², Henning Keinke Andersen ^3,^✉

Editor: Cochrane Colorectal Cancer Group

PMCID: PMC8407323 PMID: 16034862

Abstract

Background

Appendicitis is the most common cause of acute abdominal pain requiring surgical intervention. The cause of appendicitis is unclear and the mechanism of pathogenesis continues to be debated. Despite improved asepsis and surgical techniques, postoperative complications, such as wound infection and intraabdominal abscess, still account for a significant morbidity. Several studies implicate that postoperative infections are reduced by administration of antimicrobial regimes.

Objectives

This review evaluated the use of antibiotics compared to placebo or no treatment in patients undergoing appendectomy. Will these patients benefit from antimicrobial prophylaxis? The outcomes were described according to the nature of the appendix, as either simple appendicitis (including the non‐infectious stage) and complicated appendicitis. The efficacy of different antibiotic regimens were not evaluated.

Search methods

We searched The Cochrane Central Register of Controlled Trials (Cochrane Library 2005 issue 1); Pubmed ; EMBASE; and the Cochrane Colorectal Cancer Group Specialised Register (April 2005). In addition, we manually searched the reference lists of the primary identified trials.

Selection criteria

We evaluated Randomised Controlled Trials (RCTs) and Controlled Clinical Trials (CCTs) in which any antibiotic regime were compared to placebo in patients suspected of having appendicitis, and undergoing appendectomy. Both studies on children and adults were reviewed. The outcome measures of the studies were: Wound infection, intra abdominal abscess, length of stay in hospital, and mortality.

Data collection and analysis

Eligibility and trial quality were assessed, recorded and cross‐checked by two reviewers.

Main results

Forty‐five studies including 9576 patients were included in this review. The overall result is that the use of antibiotics is superior to placebo for preventing wound infection and intraabdominal abscess, with no apparant difference in the nature of the removed appendix. Studies exclusively on children and studies examining topical application reported results in favour to the above, although the results were not significant.

Authors' conclusions

Antibiotic prophylaxis is effective in the prevention of postoperative complications in appendectomised patients, whether the administration is given pre‐, peri‐ or post‐operatively, and could be considered for routine in emergency appendectomies.

Plain language summary

Antibiotic prophylaxis could be considered for routine in emergency appendectomies.

Appendicitis is the most common cause of acute abdominal pain requiring surgical intervention. This is associated with increased risk of postoperative complications, wound infection being the most commonly reported. Standard prophylaxis is an anti‐bacterial treatment. In order to reduce cost, toxicity and the risk of developing bacterial resistance, it is desirable to establish the shortest and most effective prophylaxis for postoperative complications.   This review reports that antibiotic prophylaxis is effective in the prevention of postoperative complications in people who had the appendix removed. Regardless whether the antibiotic was given before, during or after the surgery.

Background

Appendicitis is the most common cause of acute abdominal pain requiring surgical intervention, and affects all age groups. On suspicion of acute appendicitis the standard procedure is surgery.   The number of patients undergoing appendectomy (removal of the appendix) in the population is significantly higher than the number of diagnosed acute appendicitis. Fenyö (Fenyö1995) reported that 15‐30% of all appendectomies are unnecessary due to a non‐inflamed appendix, and Luckmann (Luckmann 1989) suggests even higher numbers. The actual incidence of acute appendicitis varies, the overall lifetime risk for acute appendicitis is 6‐20% (Blewett 1995; Addiss 1990). Addiss reported 8.6% for males and 6.7% for females, compared to an overall life time risk for appendectomy of 12% (males) and 23.1% (females) in US.   Removal of the appendix is most frequently reported in adolescents (10‐20 years of age). The incidence in children under the age of 5, and in adults over 70 years of age, is small (Graffeo 1996; Addiss 1990; Luckmann 1989).

The cause of acute appendicitis is unclear. The mechanism of pathogenesis could involve obstruction of the appendiceal lumen, preventing the escape of intraluminal secretion (Arnbjörnsson 1983). This could result in increased intraluminal pressure, leading to transmural tissue necrosis . However, the mechanisms of pathogenesis continue to be debated.   The predominant microbial flora associated with acute appendicitis are E.Coli, Kleibciella, Proteus and Bacteroides (Altemeier 1938; Leigh 1974; Bennion 1990; Blewett 1995). These microbes may cause postoperative infection depending on the degree of inflammation of appendix, surgical technique and duration of operation. Postoperative complications parallel that found from perforated viscus of any cause. The most common being intra abdominal abscess and wound infection was reported in as high as 40% of all appendectomy cases (Almqvist 1995). Despite improved asepsis and surgical techniques, postoperative infections still accounts for morbidity and increased length of stay in hospital as major consequences.   Several studies have implicated that postoperative infections are reduced by the administration of antimicrobial regimes. The general assumption is that it is better to treat patients with complicated appendicitis with antibiotics than with placebo, in contrast to patients with simple appendicitis. Yet other studies have questioned the use of antibiotics at all in appendectomised patients. In order to reduce cost, toxicity and the risk of developing bacterial resistance, it is desirable to establish the shortest and most effective prophylaxis for postoperative complications.

Bearing the above points in mind, and given that the use of antibiotics in clinical practice is frequently questioned, in this systematic review we examined the effect of antimicrobial treatment versus placebo treatment in appendectomised patients. To reflect the clinical decision‐making process we wanted to explore whether patients operated on suspicion of appendicitis could prevent postoperative infection if they received antimicrobial therapy prior to surgery or during surgery. This implicates that appendectomy is carried out for reasons other than acute appendicitis (negative appendectomy), and is included in this study as well, due to exposure of the colonic microbial flora.

A description of the alternatives to antimicrobial prophylaxis, such as delayed closure and surgical lavage, will not be covered in this review.

The clinical diagnosis of appendicitis is generally divided in two subgroups:

Simple appendicitis include terms like 'non‐inflamed' (normal appendix), 'acutely inflamed', 'phlegmonous', 'acute', 'suppurative', 'mildly inflamed with or without peritonitis'. This condition also covers minimal appendicitis, early appendicitis, and uncomplicated appendicitis.   Complicated appendicitis includes 'gangrenous appendicitis', 'perforated appendicitis', 'local pus collection at operation', 'general peritonitis', and 'intra abdominal abscess'.

The analyses are stratified to include subgroups of patients with simple appendicitis and complicated appendicitis.

Objectives

A systematic search for relevant literature from controlled clinical trials was performed to find evidence relating to the use of antibiotics in patients undergoing appendectomy on the suspicion of appendicitis.   Relevant data were extracted from these reports of clinical trials and analysis performed which reflects the clinical decision‐making process. Given the fact, that patients are operated upon suspicion of appendicitis, we wished to explore whether postoperative infection and prolonged stay in hospital could be prevented if the patients were given antimicrobial therapy prior to, under, or after the surgery.

Methods

Criteria for considering studies for this review

Types of studies

This review evaluated randomised controlled trials (RCTs) and controlled clinical trials (CCTs) in which treatment with any antimicrobial regime was compared to placebo in patients with suspected appendicitis undergoing appendectomy. The authors did not discriminate between trials that describe appendices in various stages, that is if they are non inflamed or perforated.   Studies in which the antimicrobial regime was compared to a control group receiving no treatment were excluded from this review, unless perioperative treatment was stated and subject blinding was found adequate.

Types of participants

Children and adults with suspected appendicitis based on either clinical conditions or intra operatively diagnosed by the surgeon. No restrictions to age or gender.

Appendectomised patients were ideally reported in two subgroups: Simple appendicitis and complicated appendicitis. If pathology was not stated in a trial, the data were admitted statistically into the subgroup 'appendicitis', which covers all aspects of pathology.

The definition of 'simple appendicitis' includes the appendix which is normal at operation (non‐inflamed); acutely inflamed; phlegmonous; acute; suppurative; mildly inflamed with or without peritonitis. This condition also covers minimal appendicitis; early appendicitis; and uncomplicated appendicitis.   The definition of 'complicated appendicitis' includes gangrenous appendicitis; perforated appendicitis; local pus collection at operation; general peritonitis; and intra abdominal abscess.   Trials reporting patients with periappendicular abscess not operated upon are excluded from this review.

Types of interventions

Trials comparing any antimicrobial regime versus placebo administered before, during, or after appendicectomy, and reporting postoperative infection, length of stay in hospital, and mortality, were considered for this review.

The primary topic was to examine whether antimicrobial therapy given to patients operated on suspicion of appendicitis could prevent postoperative infection.

The types of intervention are specified in the table of comparisons.

The term 'clinical diagnosis' refers to the condition pre‐ or perioperatively diagnosis, usually performed by the clinician. The use of paraclinical diagnosis (bloodtest, urintest e.g.) is included in this term.

The term 'pathoanatomical diagnosis' refers to the more accurate postoperative diagnosis in which the nature of the removed appendix have been established either macroscopically or microscopically.

Trials reporting patients treated topically with antimicrobial agents disolved in aqueous solution were considered for inclusion if the comparison group was treated with the aqueous solution. Otherwise we excluded the study.

Types of outcome measures

Trials were considered if one or more of the following clinical outcomes were reported:

1) Wound infection (discharge of pus from the wounds).

2) Postoperative intra abdominal abscess (persistent pyrexia without any other focus, after operation, palpable mass in the abdomen or discharge of pus from the rectum).

3) Length of stay in hospital

4) Mortality

In this review we analysed patients with a normal appendix and patients with perforated appendix. This approach considered the pathoanatomical distinction of appendicitis, using the terms for subgroup analysis as described above.

Search methods for identification of studies

The following bibliographic databases were searched in order to identify relevant primary studies:

Cochrane Central Register of Controlled Trials (CENTRAL), Cochrane Library 2005 issue 1   PubMed, from 1966 to April 2005   EMBASE, from 1980 to April 2005   Cochrane Colorectal Cancer Group Specialised Register, April 2005

The following strategy was used to search the Pubmed database, combined with the Cochrane Collaboration highly sensitive search strategy for identifying randomised controlled trials and controlled clinical trials:

#1 appendic*   #2 appendec*   #3 appendek*   #4 #1 or #2 or #3   #5 apendic*   #6 apendec*   #7 apendek*   #8 #5 or #6 or #7   #9 appendix*   #10 apendix*   #11 #9 or #10   #12 #4 or #8 or #11

From the studies identified with the search term appendix* we have excluded Boon*, Spong*, Neuro* and Psyc*.

The Cochrane Central Register of Controlled Trials (CENTRAL) and CCCG specialised register were searched using the above search strategy, but without the Cochrane Collaboration highly sensitive search strategy for identifying randomised controlled trials and controlled clinical trials.

The search strategies were then combined with free text words (truncation as indicated):

#1 antibiot*   #2 antimic*   #3 metronida*   #4 #1 or #2 or #3

All identified trials from the search were evaluated for inclusion. Identified and included studies were further examined for additional studies from the respective reference list. .

The EMBASE database was searched using the following strategy. Search string from #22 to #33 describes the filter used for identification of randomised controlled trials and controlled clinical trials:

#35 #29 or #34   #34 #21 and #33   #33 random*   #32 #29 or #31   #31 #21 and #28   #30 case*   #29 #21 and #27   #28 random* or clin*   #27 #22 or #24 or #26   #26 #23 not #25   #25 'case‐control‐study' / all subheadings   #24 'randomization‐' / all subheadings   #23 explode 'controlled‐study' / all subheadings   #22 explode 'clinical‐trial' / all subheadings   #21 #19 not #20   #20 boon* or spong*   #19 #17 not #18   #18 neuro* or psyc*   #17 #8 and #16   #16 #11 or #15   #15 #12 or #13 or #14   #14 metronida*   #13 antimic*   #12 antibiot*   #11 #9 or #10   #10 explode 'antibiotics‐and‐their‐derivatives' / all subheadings   # 9 explode 'antibiotic‐agent' / all subheadings   # 8 #1 or #7   # 7 #5 or #6   # 6 #2 or #3 or #4   # 5 'appendectomy‐' / all subheadings   # 4 'appendix‐perforation' / all subheadings   # 3 'acute‐appendicitis' / all subheadings   # 2 'appendicitis‐' / all subheadings   # 1 append* or apend*

The final number was manually searched for trials fulfilling the inclusion criteria.

This update revealed 11 new eligible studies, of which five were considered for inclusion (Harahsheh 2002; McGreal 2002; Taylor 2000; Taylor 2004; Simpson 2002). The remaining six trials described interventions outside the scope of this review. None of the five studies fulfilled the inclusion criteria for various reasons. One study was an abstract (Simpson 2002) and is awaiting assessment, and the four studies were excluded as described in the table of characteristics of excluded studies

Data collection and analysis

Appropriate criteria for assessing the quality of the included studies are methods (either qualitative or quantitative) that combine searches, outcomes, and meta‐analysis.

Location and selecting studies:   All identified trials were reviewed independently by two authors (HKA and BRA) in order to evaluate, whether the trial should be included or should be excluded. A data extraction form was prepared prior to the search for literature. Disagreements were solved in an consensus meeting eventually with help from a third author.   Excluded studies were detailed in the table of excluded studies.

Critical appraisal of studies:   The methodological quality of each trial was assessed by the same two authors. Details of the randomisation method, blinding, and the number of patients lost during the study were recorded. When necessary information was missing from a published trial, the primary author were contacted for clarification.

Collecting data:   Data was independently extracted using a standard data extraction form. Recorded data was cross‐checked. The data was entered into Metaview in RevMan 4.2.6 for analysis. The following parameters were extracted: wound infections; intra abdominal abscesses; number of days in hospital; and mortality.

Results

Description of studies

Forty‐five studies, including 9576 patients, were included in the meta‐analyses of this review. Twelve studies had more than one arm; seven studies had two arms and five studies had three arms. Age‐range of the included patients was three months to 94 years of age. Sixteen studies reported the male/female ratio. The morphologic descriptions of the pathoanatomical distinction of appendicitis were very weakly described. Twenty‐one of the studies did not divide the appendix into pathological subgroups. Twenty‐one of the studies reported a division into simple or complicated appendicitis. No attempt was made to elucidate age‐ or gender related correlations.   In summary, including the arms of all identified studies we examined 63 trials; 46 trials examined single drug therapy and 17 trials examined a combination of two or more drugs. The antibiotics used in the studies were divided into nine major antibiotic groups. The most common antibiotics used were Cephalosporin and imidazole derivatives.   All forty‐five studies reported wound infection. Twelve studies reported intra abdominal abscess and ten studies reported length of stay in hospital. Only two studies reported mortality.

Thirty‐two of the identified studies were excluded for various reasons, and one study is awaiting assesment (Simpson 2002). Allocation concealment being the major factor, followed by insufficient information on treatment strategies. Sixteen of those were potentially eligible for the meta‐analysis, but any doubt on subject blinding led to their exclusion. In eleven of the studies the patients in the placebo group received no treatment. In nine studies, the antibiotics in the treatment group was administrated in a way, that made it possible for the patient as well as the clinician to know whether the patient was in the treatment group or in the control group. These studies were excluded because of inadequate blinding (Amgwerd 1981, Bates 1974, Bröte 1976, Evans 1973, Feltis 1967, Gómez‐Alonso 1984, Kling 1985, McLean 1983, Okubadejo 1976). Two trials (Crosfill 1969, Gilmore 1975) were excluded because it was unclear which patients received systemic antibiotics. Yet another three studies described interventions that could make it difficult to establish the effect of antibiotics alone (Harahsheh 2002; Taylor 2000; Taylor 2004). One trial (Birkigt 1989) was excluded due to discord in the table of the results. Another trial (Herrera‐Garcia 1985) did not distinguish between wound infection and intra abdominal abscess in the result section. Yet another study was excluded due to irrigation with physiological serum in the control group (Badia 1994).   Five studies did not fulfil the inclusion criteria for any meta‐analyses. Two of the trials did not have results on appendicitis exclusively (Dixon 1984, Pollock 1972) and one (Rambo 1972) only reported one case of appendicitis. One trial was an abstract which did not obtain all the necessary details and it was impossible to get the full study. One trial were based on prior published trials (Leigh 1978). A detailed description is available in the reference list under "excluded studies".

Risk of bias in included studies

The methodological quality of each trial were assessed by the same two authors (HKA and BRA). Details of the randomisation method, blinding, and the number of patients lost during the study were recorded.   Missing information from reports of trials was sought by contacting the authors.   As a common denominator studies were excluded if the placebo group was not stated. However, if subject blinding was adequately described, studies reporting "control" or "no treatment" were considered for peri‐operatively treatment only .

The randomisation method of included studies was adequately defined in 17 of the 45 studies, either by computerised random numbers or sealed envelopes. In twenty‐three studies the randomisation method was not stated. The last five studies were indexed as quasi‐randomised (CCT), by using birth date as concealment method. These studies were all categorised as having inadequate randomisation. Thirty of the studies were double or triple blinded. Ten of the studies had at least blinding of the patient, but unclear blinding of surgeon and outcome assessor. In five of the studies the blinding was found unclear.   Six studies displayed odd numbers in the comparison groups. They were individually evaluated for drop‐outs or other reasons to clarify the imbalance and were finally included.   Wound infection was defined in 34 of the studies. The definition differed, but the major description was discharge of pus from the wounds. Three out of eleven studies defined intra abdominal abscess.   We did not observe any time related variations regarding the methodological quality of the studies. Older studies such as Stoller 1965 were found to be adequately randomised and fully comparable to newer studies.

Effects of interventions

Study validation in meta‐analysis   In the current updated systematic review we have identified a total of 78 RCT's and CCT's for inclusion, of which 32 studies were excluded for various reasons. One trial is awaiting assessment.

In the first published version of this systematic review, we identified 70 RCT's and CCT's, of which 27 were excluded. In the first updated version (2003) we identified additional three studies, one was included and two were excluded. In this updated version we identified additional five studies, four excluded and one awaiting assessment.   Forty‐five controlled trials (9576 patients), of which five trials were dealing with topical use of antibiotics, were included in the meta‐analysis of this review. Outcome measures were sought divided in either simple or complicated appendicitis according to the definition in "types of participants". If no attempt was made to sub‐divide the patients, results were presented as "appendicitis". Results are pooled data from all studies and focus on appendectomied patients without specification of the clinical severity of disease or pato‐anatomic characterisation of the removed appendix. Analyses of pato‐anatomic clinical relevant entities (normal removed appendix, simple appendicitis (phlegmonuos) or complicated appendicitis (gangrenous or perforated) confirmed the overall result: An effect of antibiotics when compared to placebo. A subgroup analysis of different treatment strategies (see table 1) confirmed this overall result.

The clinical heterogeneity   amongst the trials was reflected in parameters such as study population; diagnosis; strategy of treatment; type of antibiotics; the outcome analysis and length of follow up. The majority of included studies did not distinguish between adults and children. Six trials exclusively on children (a total of 776 patients, aged 3 months ‐ 15 years of age) were reported. Included studies were different in strategy and use of antibiotics but all compared to placebo or control treatment (subject blinded).

The statistical heterogeneity depends on the clinical and methodological differences within the trials. We used a validated quality scale (Jadad 1996) in order to identify systematic differences among the included trials. This scale produces scores from 0‐5 points. Trials with less than 3 point considered "poor quality studies". Each of the included trials was quality assessed independently by two of the three authors and the scores were compared.   Visual inspection of trials sorted by either size or quality score did not affect the outcome. Therefore, it was decided not to exclude low score studies.   To assess publication bias we used the regression test for asymmetry in the funnel plot a specified cut‐off number of five trials in the meta‐analysis was chosen. In total we performed visual inspection for asymmetry in a total of nine different outcomes, without detecting any discrepancies. This finding further supports the notion, that small trials show the same results as greater trials.   Sub‐group analysis after randomisation may introduce bias due to significantly different population sizes. In addition, we observed that some of the included studies were unequally randomised as 3:2 or 2:1 and even 2:1:1 in multiple arm studies. Observed imbalance in the comparisons groups were evaluated to ensure an effect of the study design or any drop‐outs on the overall results. Reasons for odd numbers in comparisons group were either multiple arm studies (Donovan 1979 A, Foster 1981 A, Viitanen 1984 A), unequally numbers of post‐operative drop‐outs or other reasons (Azabache 1987 A, Gledhill 1983, Keiser 1983, Stoller 1965).   In summary, we found that none of the included 45 studies revealed any significant discrepancy from the overall result. Conclusively, the statistical and clinical heterogeneity of the trials included in this review still make comparison of the individual trials possible with respect to the overall question ‐ antibiotics versus placebo (no treatment).

Outcomes:   The major outcome "wound infection" was described in all included trials. However it must be emphasised, that the definition of wound infection is subject to a broad variety and in 9 of the 45 included trials no description was stated. In addition, clinical test for wound infection was also subject to broad variety in the studies describing this parameter.   The outcome "intra abdominal abscess" was described in 16 of the 45 included trials.   The outcome "length of hospital stay" was described in nine of the 45 included trials. Despite a marked heterogeneity among the trials, we decided to present the meta‐analysis. The overall tendency in the included trials was no intention of describing the length of follow‐up.   The outcome "mortality" was described in two trials with a total of three deaths. One patient died from lung emboli after three months, another from cardiac infarction and a third from peritonitis due to a perforated appendix. Therefore, no direct comparison was made possible, and no meta‐analysis was made.

Study groups and results:   Thirty‐four of the included trials include patients with all clinical aspects of appendicitis. The meta‐analysis of this review are presented according to the clinical distinguishable diagnosis:   normal removed appendix, simple appendicitis (phlegmonuos) or complicated appendicitis (gangrenous or perforated).   The overall result is that the use of antibiotics is superior to placebo for the outcomes "wound infection" and "intra abdominal abscess".

None of the 45 trials described "exclusion of children", rather a variety in age for included patients. seven trials exclusively on children (776 patients, aged 3 months ‐ 15 years) were reported, and the resulting meta‐analysis showed non‐significant effects.

Four trials examined topical application of antibiotics compared to placebo ( 679 patients). The meta‐analysis did not show a significant effect.

Treatment strategies and subgroup results are presented in 'comparisons and data'. The drugs are administered either systemic or topical, without distinguishing between antibiotic regimens. Time of administration was divided in either pre‐ or peri‐operatively, as the term post‐operatively from a clinical point of view is irrelevant and not reported in the studies. However, the later may be relevant with one of the two former.

Number of drugs used is characterised as either 'single agent' or 'multiple agents'. Frequency of administration is divided in either 'single dose' or 'repetitive dose'.

The results are presented summarically as Peto/odds ratio (95% CI), fixed‐effect model (log scale ranging from 0.01 to 10.00 shown in the meta‐analysis graphs). Values less than 1.0 favour treatment with antibiotics. In addition the numerical values are notified as incidences/included patients for each Group (antibiotics versus placebo). The overall result shows a significant advantage for the use of antibiotics compared to administration of placebo.

Discussion

Recent studies have produced significant evidence to support the use of antibiotic therapy in patients undergoing appendectomy. We therefore tested the hypothesis:

"Does antibiotic treatment reduce post‐operative infection rates in appendectomised patients"?

Antibiotic prophylaxis is one of many measures that should be taken into account in order to reduce postoperative morbidity (primarily wound infection). Many have reviewed the value of antimicrobial prophylaxis in clean‐contaminated, contaminated and in clean operations, but even though benefits are well established, no general recommendations on antibiotic prophylaxis have been settled. This systematic review focus on appendectomised patients, randomly assigned either antibiotic therapy or placebo (control) treatment. Recommended duration of antibiotic administration is not possible from the presented meta‐analysis, as we did not compare single dose versus multiple dose. In addition, the efficacy of the therapeutic use of different antibiotic regimens efficacy was not the scope of this review.   From a clinical point of view, the main concern on administration of antibiotics is how to reduce post surgical complications and the intention to shorten the length of the hospital stay. We did not address the use of antibiotics in the treatment of appendicitis without surgery.

The results confirm an overall effect of antibiotics, regardless administered as either prophylactics (single dose administration) in case of "normal removed appendix" or "simple appendicitis"(phlegmonous), or as repetitive treatment in case of "complicated appendicitis" (gangrenous or perforated).

Wound infection (superficial) or intra abdominal abscess (deep infection) are the most common post surgical complications and reflect the severity of disease and/or quality of surgical procedure. Long‐term complications such as intestinal obstruction and infertility are not addressed in the included controlled studies with short follow up period, but need analysis from large epidemiological trials.   The outcome "length of hospital stay" is closely associated with the nature of post surgical infections and indirectly to the efficacy of antibiotics. A limited number of placebo controlled studies reviewed the length of hospital stay (Bates 1980, Bauer 1989, Busuttil 1981 A, Gurry 1976, Keiser 1983, Paakkonen 1982, Raahave 1970, Winslow 1983). From a historical point of view, we have to emphasise radical changes in "hospital stay policy", presently favouring early hospital discharge.   In summary, the presented meta‐analysis on this parameter is based on studies from the past three decades and therefore difficult to interpret.

Mortality is a rare complication to appendectomy and only reported in two trials. Since the main outcome is short term investigation on the effect of antibiotic versus placebo on wound infection and intra abdominal abscess, this could explain the low number of reported deaths due to exclusion from study. Only one trial reported short term mortality as a reason for exclusion of patients from the study. An overview on actual mortality rate should be sought in epidemiological studies rather than randomised studies.

Seven trials exclusively on children (age: 3 months ‐15 year) were identified. Both nature of disease and treatment were comparable to the ones for adults. Interestingly, subgroup analysis of this population showed a non‐significant reduction in infection rate, which can be explained by a limited number of patients without heterogeneity between single study results. Another factor is the reported pathogenesis of appendicitis in children, which shows higher incidence of complicated appendicitis. However, although not divided in specific sub‐populations, the majority of studies include children in their analysis. It is only reasonable to assume that the overall antibiotic efficacy also is representative for children. Several non‐randomised studies have revealed differences in the pathology of the diagnosed appendicitis which could explain the non‐significant reduction in the infection rate.

The meta‐analysis on topical use of antibiotics shows clinical heterogeneity, primarily explained in a reported diversity of experimental set‐up. Eklund 1987 exclude all patients with perforated appendicitis and Gurry 1976 only include those patients. Eklund 1987 exclude patients treated with additional antibiotics before randomisation and Stoller 1965 exclude those patients after randomisation. Gurry 1976 and Foster 1981 A do not describe their procedure. In addition, the trials use four types of antibiotics and different ways of administration. Eklund 1987 and Gurry 1976 use a solution, Foster 1981 A uses powder and Stoller 1965 uses an aerosol. Regarding statistical heterogeneity, the trials express some methodological differences as well. Three of the trials are randomised‐controlled trials and one of the trials is a controlled clinical trial. Foster 1981 A uses single blinding (patients), Eklund 1987 double blinding, and Gurry 1976 and Stoller 1965 triple‐blinding.   Therefore, non‐significant results could be explained by clinical and methodological differences. Historically, topical use of antibiotics is now generally replaced with a higher hygienic standard.

Wound infection is significantly reduced in pooled data as well as subgroup analysis, without any indication of statistical heterogeneity, confirmed by quality assessment of single studies, where high quality studies were compared to low quality studies without discrepancies in the effect parameter. Only deviance from this pattern is the non‐significant result from topical use of antibiotics.

The meta‐analysis of intra abdominal abscess show identical patterns, but only the analyse of the pooled data is statistical significant. Interestingly, the subgroup analysis show non‐significant effect of antibiotics, which can be explained by the limited number of patients / events rather than heterogeneity or study quality.

We analysed the efficacy of antibiotics, administered either prophylactic or post‐surgically, notified as "short term versus long term treatment". Both strategies reduces wound infection significantly and intra abdominal abscess non‐significantly (again due to a low number of events). Although significant results we are not in a position of comparing these strategies, which needs to be adressed in a review comparing the different regimens.

Authors' conclusions

Implications for practice.

It seems reasonable to conclude that antibiotic prophylaxis is effective in the prevention of post‐operative complications in appendectomied patients, whether the administration is given pre‐, per‐ or post‐operatively, and should be considered for routine use in emergency appendicectomy. The overall strategy (type of antibiotics and/or preferred time of administration) needs to be evaluated in another systematic review.

Our results indicate that single doses have the same impact as multiple doses. In order to reduce cost, toxicity and the risk of developing bacterial resistance, it is desirable to establish the shortest, effective prophylaxis for post‐operative complications, and from the meta‐analysis it seem that single doses have the same impact as multiple doses.

The general assumption is that it is better to treat patients with complicated appendicitis with antibiotics than with placebo, in contrast to patients with simple appendicitis. Yet other studies have questioned the use of antibiotics at all in appendectomised patients.

Implications for research.

Despite the presented meta‐analysis indicate similar results on single dose efficacy and multiple dose efficacy, we did not compare these outcome parameters. In addition, the time for administration did not seem to have any significant impact at all, but weren't compared either. Conclusively these parameters should therefore be reviewed. Other strategies such as irrigation and delayed closure could alternatively have positive effects on reduction of postoperative complications, but need further examination in another systematic review.   This review confirms the effect of antibiotics in reducing wound infection and abscess formation in all pato‐anatomic subgroup and after removal of a normal appendix. Recommendations to the use of antibiotics in all cases of appendectomies patients is still controversy regarding the general accepted recommendations prescribing only use of antibiotics in case of complicated appendicitis. On the other hand the literature stated superior effects following preoperatively administration incompatible to per operatively verifying (by visualisation) of the disease severity as condition for initialisation of antibiotic treatment.

What's new

Date	Event	Description
5 September 2008	Amended	Converted to new review format.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Wound infection	47	8812	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.33 [0.29, 0.38]
1.1 Appendicitis	21	2343	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.31 [0.24, 0.42]
1.2 Simple appendicitis	26	5317	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.37 [0.30, 0.46]
1.3 Complicated appendicitis	24	1152	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.28 [0.21, 0.38]
2 Postoperative intra abdominal abscess	16	4468	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.43 [0.25, 0.73]
2.1 Appendicitis	8	1033	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.35 [0.13, 0.91]
2.2 Simple appendicitis	8	2968	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.46 [0.23, 0.94]
2.3 Complicated appendicitis	4	467	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.54 [0.12, 2.43]
3 Length of stay in hospital	8	1200	Mean Difference (IV, Fixed, 95% CI)	‐1.69 [‐1.78, ‐1.61]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Wound infection	21	1704	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.32 [0.22, 0.47]
1.1 Normal appendix	21	1555	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.28 [0.18, 0.44]
1.2 Perforated appendicitis	6	149	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.47 [0.22, 1.00]
2 Postoperative intra abdominal abscess	3	741	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.28 [0.08, 0.91]
2.1 Normal appendix	3	673	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.45 [0.10, 1.98]
2.2 Perforated appendicitis	1	68	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.12 [0.02, 0.86]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Wound infection	4	679	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.77 [0.49, 1.23]
1.1 Appendicitis	2	113	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.65 [0.27, 1.54]
1.2 Simple appendicitis	2	399	Peto Odds Ratio (Peto, Fixed, 95% CI)	1.30 [0.64, 2.64]
1.3 Complicated appendicitis	2	167	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.43 [0.18, 1.01]
2 Postoperative intra abdominal abscess	0	0	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.0 [0.0, 0.0]
2.1 Appendicitis	0	0	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.0 [0.0, 0.0]
2.2 Simple appendicitis	0	0	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.0 [0.0, 0.0]
2.3 Complicated appendicitis	0	0	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.0 [0.0, 0.0]
3 Lenght of stay in hospital	1	29	Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Wound infection	11	2191	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.34 [0.25, 0.45]
1.1 Appendicitis	2	234	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.34 [0.14, 0.80]
1.2 Simple appendicitis	9	1514	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.37 [0.25, 0.54]
1.3 Complicated appendicitis	8	443	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.29 [0.18, 0.47]
2 Postoperative intra abdominal abscess	2	446	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.34 [0.05, 2.45]
2.1 Appendicitis	0	0	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.0 [0.0, 0.0]
2.2 Simple appendicitis	2	382	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.34 [0.05, 2.45]
2.3 Complicated appendicitis	1	64	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.0 [0.0, 0.0]
3 Length of stay in hospital	1	121	Mean Difference (IV, Fixed, 95% CI)	0.40 [‐0.01, 0.81]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Wound infection	2	215	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.14 [0.05, 0.39]
1.1 Appendicitis	1	133	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.12 [0.02, 0.61]
1.2 Simple appendicitis	1	82	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.16 [0.04, 0.59]
1.3 Complicated appendicitis	0	0	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.0 [0.0, 0.0]
2 Postoperative intra abdominal abscess	0	0	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.0 [0.0, 0.0]
2.1 Appendicitis	0	0	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.0 [0.0, 0.0]
2.2 Simple appendicitis	0	0	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.0 [0.0, 0.0]
2.3 Complicated appendicitis	0	0	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.0 [0.0, 0.0]
3 Length of stay in hospital	0	0	Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Wound infection	12	3358	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.43 [0.34, 0.55]
1.1 Appendicitis	6	585	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.37 [0.22, 0.60]
1.2 Simple appendicitis	6	2347	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.52 [0.37, 0.73]
1.3 Complicated appendicitis	5	426	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.35 [0.21, 0.58]
2 Postoperative intra abdominal abscess	3	2115	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.67 [0.32, 1.42]
2.1 Appendicitis	2	380	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.69 [0.12, 3.99]
2.2 Simple appendicitis	1	1554	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.74 [0.31, 1.79]
2.3 Complicated appendicitis	1	181	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.32 [0.03, 3.29]
3 Length of stay in hospital	2	209	Mean Difference (IV, Fixed, 95% CI)	‐0.80 [‐1.34, ‐0.26]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Wound infection	4	216	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.23 [0.12, 0.45]
1.1 Appendicitis	3	161	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.24 [0.11, 0.52]
1.2 Simple appendicitis	1	43	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.19 [0.04, 0.89]
1.3 Complicated appendicitis	1	12	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.25 [0.03, 2.52]
2 Postoperative intra abdominal abscess	0	0	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.0 [0.0, 0.0]
2.1 Appendicitis	0	0	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.0 [0.0, 0.0]
2.2 Simple appendicitis	0	0	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.0 [0.0, 0.0]
2.3 Complicated appendicitis	0	0	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.0 [0.0, 0.0]
3 Length of stay in hospital	0	0	Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Wound infection	3	256	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.16 [0.07, 0.36]
1.1 Appendicitis	3	256	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.16 [0.07, 0.36]
1.2 Simple appendicitis	0	0	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.0 [0.0, 0.0]
1.3 Complicated appendicitis	0	0	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.0 [0.0, 0.0]
2 Postoperative intra abdominal abscess	2	161	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.12 [0.02, 0.89]
2.1 Appendicitis	2	161	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.12 [0.02, 0.89]
2.2 Simple appendicitis	0	0	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.0 [0.0, 0.0]
2.3 Complicated appendicitis	0	0	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.0 [0.0, 0.0]
3 Length of stay in hospital	0	0	Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Wound infection	14	2097	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.46 [0.35, 0.60]
1.1 Appendicitis	8	1039	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.45 [0.30, 0.68]
1.2 Simple appendicitis	6	875	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.46 [0.30, 0.70]
1.3 Complicated appendicitis	6	183	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.47 [0.24, 0.90]
2 Postoperative intra abdominal abscess	3	448	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.14 [0.01, 1.30]
2.1 Appendicitis	2	269	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.13 [0.01, 2.08]
2.2 Simple appendicitis	1	179	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.15 [0.00, 7.38]
2.3 Complicated appendicitis	0	0	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.0 [0.0, 0.0]
3 Length of stay in hospital	4	434	Mean Difference (IV, Fixed, 95% CI)	‐1.0 [‐1.81, ‐0.19]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Wound infection	6	851	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.18 [0.11, 0.27]
1.1 Appendicitis	1	200	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.25 [0.10, 0.62]
1.2 Simple appendicitis	4	562	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.20 [0.11, 0.35]
1.3 Complicated appendicitis	5	89	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.08 [0.03, 0.22]
2 Postoperative intra abdominal abscess	3	488	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.38 [0.05, 2.72]
2.1 Appendicitis	1	133	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.98 [0.06, 15.92]
2.2 Simple appendicitis	2	355	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.15 [0.01, 2.38]
2.3 Complicated appendicitis	0	0	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.0 [0.0, 0.0]
3 Length of stay in hospital	1	90	Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]

Methods	Randomised controlled trial. Randomisation method not stated.   Double‐blinded. Unclear blinding of outcome assessor.
Participants	190 patients enrolled in the study.   Age‐range: Treatment group: 10‐45, Placebo group: 10‐40   Mean: Treatment group: 19,2 +/‐ 6,7 (s.d.), Placebo group: 21,3 +/‐ 8,7 (s.d.)   Patients with normal, gangrenous or perforated appendix and antibiotics prior to surgery were excluded before randomisation.
Interventions	Treatment group: Metronidazole 500 mg=100 ml i.v inter‐operatively.   Placebo group: Saline 100 ml i.v peroperatively
Outcomes	Wound infection (Purulent discharge from the incision).
Notes	No drain treatment.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Unclear risk	B ‐ Unclear

Methods	Randomised controlled trial. Randomisation method stated as simple. Patient and surgeon blinded. Unclear blinding of outcome assesor.
Participants	258 eligible patients. 92 excluded before randomisation. 166, 88 males and 78 females enrolled in the study.   Age‐range: 6‐71   Mean: 23   12 patients excluded after randomisation.   The study has 2 subgroups (see Azabache 1987 B).   In Azabache 1987 A 110 patients enrolled in study.
Interventions	Treatment group: Gentamycin 4 mg/kg and Clindamycin 15 mg/kg one hour inter‐operatively hereafter Gentamycin eighthourly and Clindamycin sixhourly for 24 hours if appendix was suppurativ and 72 hours if appendix was gangrenous/perforated.   Placebo group: Not stated. Regime as above.
Outcomes	Wound infection (Pus in the cellular space under or above the rectusmuscle 3 weeks postoperatively)
Notes	No drain treatment.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Unclear risk	B ‐ Unclear

Methods	Randomised controlled trial. Randomisation method stated
Participants	258 eligible patients. 92 excluded before randomisation. 166, 88 males and 78 females enrolled in the study.   Age‐range: 6‐71   Mean: 23   12 patients excluded after randomisation.   The study has 2 subgroups (see Azabache 1987 A).   In Azabache 1987 B 110 patients enrolled in study.
Interventions	Treatment group:   Penicilin G 400.000 UI/kg and Chloramphnicol 50 mg/kg one hour intra operatively. Hereafter penicillin G each four hour and chloramphenicol each six hour for 24 hours, if appendix was suppurative and 22 hours if appendix was gangrenous/perforated.   Placebo: Not stated. Regimen as above.
Outcomes	Wound infection (Pus in the cellular space under or above the rectusmuscle 3 weeks postoperatively)
Notes	No drain treatment.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Unclear risk	B ‐ Unclear

Methods	Controlled clinical trial. Random allocation by the admittiong house surgeon. No blinding of the patient. Adequate blinding of surgeon and unclear blinding of outcome assesor.
Participants	200 eligible patients. 30 excluded after randomisation. 170, 90 males and 80 females enrolled in study.   Age‐range: 0‐90
Interventions	Treatment group: 1 g metronidazole supp 1 hour preoperatively and 1 g supp or orally 8 hourly postoperatively for 7 days. children below 12 received half dose.   Placebo group: No treatment
Outcomes	Woundinfection (A frank discharge of pus)   Lenght of stay in hospital
Notes	No drain treatment
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Allocation concealment?	High risk	C ‐ Inadequate

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Wound infection	7	1090	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.64 [0.37, 1.10]
1.1 Appendicitis	1	133	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.98 [0.39, 2.44]
1.2 Simple appendicitis	6	704	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.92 [0.33, 2.57]
1.3 Complicated appendicitis	3	253	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.31 [0.12, 0.77]
2 Postoperative intra abdominal abscess	6	1003	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.29 [0.10, 0.83]
2.1 Appendicitis	1	133	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.25 [0.05, 1.26]
2.2 Simple appendicitis	5	648	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.14 [0.02, 0.98]
2.3 Complicated appendicitis	2	222	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.78 [0.11, 5.70]
3 Length of stay in hospital	0	0	Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Wound infection	5	953	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.52 [0.29, 0.93]
1.1 Appendicitis	1	133	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.98 [0.39, 2.44]
1.2 Simple appendicitis	4	598	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.85 [0.26, 2.80]
1.3 Complicated appendicitis	2	222	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.16 [0.06, 0.44]
2 Postoperative intra abdominal abscess	6	955	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.29 [0.10, 0.83]
2.1 Appendicitis	1	133	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.25 [0.05, 1.26]
2.2 Simple appendicitis	5	600	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.14 [0.02, 0.98]
2.3 Complicated appendicitis	2	222	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.78 [0.11, 5.70]
3 Length of stay in hospital	0	0	Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]

Methods	RCT, Computer‐generated double‐blind trial
Participants	175 eligible   122 enrolled
Interventions	Ceftizoxime, Cefamandole versus placebo given IV pre‐op and 6‐12 hours after surgery   Three arm study
Outcomes	Wound infection (not defined)
Notes	Only Pt's with non‐perforative appendicitis were enrolled
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Low risk	A ‐ Adequate

Methods	As above
Participants	As above
Interventions	As above
Outcomes	As above
Notes	As above
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Unclear risk	D ‐ Not used

Methods	Randomised cotrolled trial. Randomisation by sealed envelopes. Blinding of patients and surgeon. Unclear blinding of outcome assesor.
Participants	13 patients enrolled in the study.   Age‐range: 18‐76
Interventions	Treatment group: Gentamycin 80 mg i.v. inter‐operatively.   Placebo group: No treatment
Outcomes	Wound infection (presense of pus which either spontaneously discharged or required evacuation)
Notes	No comments on drain treatment.   The results are a part of a study including other kind of gastrointestinal sugery.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Low risk	A ‐ Adequate

Methods	Clinical controlled trial. Allocation according to the month of their day of birth.   Blinding unclear.
Participants	102 enrolled in the study.   mean: Treatment group: 24,9, placebo group: 26,2   Patients with weight less than 25 kg, pregnant women or steroid treatment were excluded before randomsation.
Interventions	Treatment group: Cefamandole 2 g i.v inter‐operatively.   Placebo group: Nothing.
Outcomes	Wound infection (Not defined).
Notes	Alle patients received drain treatment.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Allocation concealment?	High risk	C ‐ Inadequate

Methods	Clinical controlled trial. Randomisation method not stated. Tripleblinded
Participants	29 patients enrolled in study. All patients had perforated appendix
Interventions	Active group: Noxytioline 2,5 g in 100 ml saline instilled in the peritoneal cavity and the wound immediately before closure.   Placebo group: 100 ml saline instilled into the peritoneal cavity and the wound immediately before closure.
Outcomes	Woundinfection (Not defined).   Length of stay in hospital
Notes	No drainage in either group.   Postoperative antibiotics were used in two control patients and in one patient in the treatment group. (Persisting peritonitis, chest infection, septicaemia)
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Unclear risk	B ‐ Unclear

Methods	RCT double blind
Participants	175 eligible Pts   145 enrolled   No children under the age of 14 years of age
Interventions	One shot 500 mg Metronidazole peri‐operative, IV   versus placebo (not stated)
Outcomes	Wound infection in simple appendicitis
Notes	No perforated appendicitis are included.   An overview of other studies are described in table 4
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Unclear risk	D ‐ Not used

Methods	Randomised controlled trail. Randomisation by random numbers.   Double‐blinded. Unclear blinding of outecome assessor.
Participants	98 eligibel patients, 26 excluded after randomisation. 98 enrolled in the study.   Age‐range: 15‐60. Mean: 27,5   Patients with active neurologic diseasis, blodd dyscrasia, hypothyroidism, hypoadrenalism, were pregnant, and patients who refused to enter the stydy were excluded.
Interventions	Treatment group: Metronidazole 1 g IV preoperatively and 500 mg eight‐hourly 5 times.   Placebo treatment: Not specified. Regime as above.
Outcomes	Wound infection (Wound erythema, definitely infected or purulent discharge).   Length of stay in hospital
Notes	No comments on drain treatment.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Low risk	A ‐ Adequate

Methods	Randomised clinical controlled trial on children
Participants	115 eligible   91 enrolled   24 excluded due to various reasons
Interventions	metronidazole versus placebo rectal suppositorium prior to surgery and 4 days follow up
Outcomes	Wound infection (not defined)
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Unclear risk	B ‐ Unclear

PERMALINK

Antibiotics versus placebo for prevention of postoperative infection after appendicectomy.

Betina Ristorp Andersen

Finn Lasse Kallehave

Henning Keinke Andersen

Abstract

Background

Objectives

Search methods

Selection criteria

Data collection and analysis

Main results

Authors' conclusions

Plain language summary

Background

Objectives

Methods

Criteria for considering studies for this review

Types of studies

Types of participants

Types of interventions

Types of outcome measures

Search methods for identification of studies

Data collection and analysis

Results

Description of studies

Risk of bias in included studies

Effects of interventions

Discussion

Authors' conclusions

Implications for practice.

Implications for research.

What's new

History

Notes

Acknowledgements

Data and analyses

Comparison 1. Systemic Antibiotics vs Placebo (Clinical).

1.1. Analysis.

1.2. Analysis.

1.3. Analysis.

Comparison 2. Systemic antibiotics vs placebo (Pathoanatomic).

2.1. Analysis.

2.2. Analysis.

Comparison 3. Topical Antibiotics vs Placebo.

3.1. Analysis.

3.3. Analysis.

Comparison 4. Pre‐operatively administered single agent, single dose Antibiotics vs Placebo.

4.1. Analysis.

4.2. Analysis.

4.3. Analysis.

Comparison 5. Pre‐operatively administered multiple agent, single dose Antibiotics vs Placebo.

5.1. Analysis.

Comparison 6. Per‐operatively administered single agent, single dose Antibiotics vs Placebo.

6.1. Analysis.

6.2. Analysis.

6.3. Analysis.

Comparison 7. Per‐operatively administered multiple agent, single dose Antibiotics vs Placebo.

7.1. Analysis.

Comparison 8. Operatively single agent and post‐operatively single agent, single dose Antibiotics vs placebo.

8.1. Analysis.

8.2. Analysis.

Comparison 9. Operatively single agent and post‐operatively single agent, multiple dose Antibiotic vs Placebo.

9.1. Analysis.

9.2. Analysis.

9.3. Analysis.

Comparison 10. Operatively multiple agent and post operatively multiple agent, single dose Antibiotic vs Placebo.

Comparison 11. Operatively multiple agent and post‐operatively multiple agent, multiple dose Antibiotics vs placebo.

11.1. Analysis.

11.2. Analysis.

11.3. Analysis.

Comparison 12. Systemic antibiotics vs Placebo in children.

12.1. Analysis.

12.2. Analysis.

Comparison 13. Operatively single agent and post‐operatively single agent, multiple dose Antibiotic vs Placebo in children.

13.1. Analysis.

13.2. Analysis.

Comparison 14. Operatively multiple agent and post‐operatively multiple agent, multiple dose Antibiotics vs placebo, children.

14.1. Analysis.

14.2. Analysis.

Methods	Clinical controlled trial. Allocation based on the last digit of the patients birth year.
Participants	315 eligible patients, 278 enrolled. Age range 15‐83 years, 159 males and 119 females
Interventions	Treatment group: Metronidazole suppository (1 g) 30‐60 min preoperatively.   Control group: No antimicrobial therapy
Outcomes	Wound infection, defined as the presence of pus or a purulent exudate at the wound
Notes	Appendicitis defined as either normal, inflamed or gangrenous
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Allocation concealment?	High risk	C ‐ Inadequate

Methods	Randomised controlled trial. Both surgeon and outcome assessor are blinded
Participants	143 eligible   129 enrolled   of which 55 had appendectomy
Interventions	Gentamycin and glindamycin versu placebo
Outcomes	post‐operative infection
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Unclear risk	B ‐ Unclear

Methods	Randomised controlled trial by use of sealed envelope.
Participants	231 eligible pts 113 enrolled   withdrawals not stated
Interventions	500 mg Metronidazole one shot peri‐operatively. IV versus control
Outcomes	Wound infection (Presence of pus in the operation wound requiring
Notes	No children under the age of 12 years
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Low risk	A ‐ Adequate

Methods	Randomised controlled trial   Randomisation method not stated
Participants	857 eligible patients. 313 patients excluded after randomisation.   544 patients enrolled in study. Alle children.   Patients with perforated appendicitis were excluded. Patients allergic to antibiotics or recieving antibiotic within 72 hours before surgery were excluded. 193 patients received treatment compared to 176 in the control group
Interventions	Treatment group: metronidazole 20 mg/ kg IV 1 hour preoperatively.   Placebo group: No treatment
Outcomes	Wound infection (subcutaneus abscess with suppuration confirmed by spontaneus wound rupture debriemnt or incision)   Intrabadominal absces (absces within the intraabdominal cagvity diagnosed at operation or by rectal examination, x‐ray or ultrasound)
Notes	No comment on drain treatment
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Unclear risk	B ‐ Unclear

Methods	Clinical controlled trial. Allocation by birthday.   Patient blinded. Unclear blinding of surgeon and outcome assessor.
Participants	588 eligible patients. 123 excluded after randomisation. 465 enrolled in the study.   Age‐range: 15‐94 Mean: 33,4   315 enrolled in this part of the study.   The study has 3 subgroups (see Viitanen 1984 B).   In Viitanen 1984 A 315 patients enrolled in study.
Interventions	Treatment group: Tinidazole 500 mg in 100 saline i.v preoperatively.   Placebo group: No treatment.
Outcomes	Wound infection (Pus from the wounds).   Two patients died, 1 of peritonitis induced by perforation of the appendix, 1 died of cardiac infarction.
Notes	Treatment initiated on anaesthetized patient.   No drain treatment.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Allocation concealment?	High risk	C ‐ Inadequate

Methods	Randomised controlled trial, method not stated
Participants	105 eligible pt   95 enrolled   10 pts excluded due to prior antibiotic record or failure in randomisation procedure
Interventions	Metronidazole versus placebo as suppository for seven days
Outcomes	wound infection   (discharge from the wound)
Notes	All suppositories contain Witepsol 35 and 75
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Unclear risk	B ‐ Unclear

Study	Reason for exclusion
Amgwerd 1981	Adequate randomisation, but control group received no treatment, and treatment group received pre‐operative and post‐operative IM treatment which means inadequate blinding and the results of infection may be biased.
Andersen 1972	All patients received post‐operative antibiotic.
Badia 1994	The treatment group (antibiotics) is compared to a 'control group' receiving a solution with physiological serum.
Bates 1974	Randomisation by sealed envelopes. Control group received no treatment, and treatment group received peri‐operative topical treatment which means inadequate blinding and the results of infection may be biased.
Birkigt 1989	Randomised clinical trial. There are discord in the table of results. The number of patients differ within the same groups.
Birkigt 1991	Unknown allocation consealment and odd distribution numbers in the treatment group versus control group.
Bröte 1976	Randomisation not stated. Control group received no treatment, and treatment group received postoperative IV treatment which means inadequate blinding and the results of infection may be biased.
Calman 1971	Abstract. It has been impossible for the reviewers to get a full study.
Crosfill 1969	Randomised controlled trial. Systemic antibiotics were given as supplement to topical treatment to the most seriously ill. The author does not describe which patients that get systemic antibiotics.
Dixon 1984	The trial group is patients in elective and emergency operations. No reports on appendicitis exclusively.
Evans 1973	Randomisation by toss of a coin. Control group received no treatment, and treatment group received postoperative IM treatment which means inadequate blinding and the results of infection may be biased.
Everson 1977	Insufficient subject blinding. Patients receiving post‐operative injection three times compared to control group (no treatment)
Feltis 1967	Randomisation method not stated. Control group received no treatment, and treatment group received pre‐operative, peri‐operative and post‐operative IV treatment, which means unclear blinding, and the results of infection may be biased.
Gilmore 1973	Inadequate allocation consealment
Gilmore 1975	Randomisation by af table of random numbers. Unclear which patients received systemic antibiotic treatment within the topical treatment and placebogroup.
Gómez‐Alonso 1984	Adequate randomisation, but control group received no treatment, and treatment group received pre‐operatively and post‐operative IM treatment which means inadequate blinding and the results of infection may be biased.
Harahsheh 2002	Randomised single blin trial, in which all pts received daily antiseptic dressings with povidone‐iodine (10% in alcohol).
Herrera‐Garcia 1985	Adequate randomisation, but no differentiation between wound infection and intra‐abdominal abscess in the result section.
Kling 1985	Adequate randomisation, but control group received no treatment, and treatment group received preoperative IV treatment 1‐3 hours prior to operation, which means unclear blinding and the results of infection may be biased.
Leigh 1978	A study based on prior publised trials, of which we have included one, excluded one (not randomised), and yet another not possible to identify.
Marti 1978	Initially three arm study comparing two different antibiotics, which later becomes a two‐arm study. No indication of wound infection ‐ only abscesses are described.
McGreal 2002	All pts received antibiotic prophylaxis. Clinical question focuses on two different ways of wound closure.
McLean 1983	Adequate randomisation. Control group received no treatment, and treatment group received post‐operative IV and orally administrated treatment, which means inadequate blinding and the results of infection may be biased.
Okubadejo 1976	Randomisation by Fisher´s tabels. Control group received no treatment, and treatment group received post‐operative treatment which means inadequate blinding and the results of infection may be biased.
Pietila 1982	No statement of placebo/control and any indication of withdrawal of patients.
Pollock 1972	The treatment group is composed of abdominal surgery, potentially contaminated operations. No reports on appendicitis exklusively.
Rambo 1972	Randomisation adequate. The trial include miscellaneous emergency surgery, of which only one appendectomy was reported.
Tanphiphat 1978	Unclear what treatment is given to who post‐operatively. In addition, wound lavage is offered all enrolled pt, so it is difficult to establish a clear effect of antibiotic treatment.
Taylor 2000	Both comparisons groups received antibiotic treatment
Taylor 2004	Both comparisons groups received antibiotic treatment. The protocol seem identical to the Taylor 2000 reference
Walsh 1981	Unclear what treatment is given to who post‐operative. From table 1, there is a clear distinction between antibiotics and control group. From table 3 it is difficult to establish a clear effect of antibiotic treatment. Therefore we decided to exclude the study.
Willis 1979	A summary of metronidazole in the prevention and treatment of anaerobic sepsis, in which appendicitis is a part. Seems identical to prior publication in 1976 (which is included)