Health Technology Assessment 1998; Vol. 2: No. 7 Review

Antimicrobial prophylaxis in colorectal surgery: a systematic review of randomised controlled trials

F Song AM Glenny

Health Technology Assessment NHS R&D HTA Programme HTA Standing Group on Health Technology Chair: Professor Sir Miles Irving, Professor of Surgery, University of Manchester, Hope Hospital, Salford †

Dr Sheila Adam, Professor Sir John Grimley Evans, Professor Ian Russell, Department of Health Department of Geriatric Medicine, Department of Health, Sciences & Professor Martin Buxton, Radcliffe Infirmary, Oxford † Clinical Evaluation, University of York Professor of Economics, Brunel University † Mr John H James, Professor Trevor Sheldon, Professor Angela Coulter, Chief Executive, Kensington, Chelsea & Director, NHS Centre for Reviews & Westminster Health Authority † Director, King’s Fund, London Dissemination, University of York Professor Richard Lilford, Professor Mike Smith, Professor Anthony Culyer, Regional Director, R&D, West Midlands † Deputy Vice-Chancellor, University of York Director, The Research School Professor Michael Maisey, of Medicine, University of Leeds † Dr Peter Doyle, Professor of Radiological Sciences, Dr Charles Swan, Executive Director, Zeneca Ltd, UMDS, London ACOST Committee on Medical Research Consultant Gastroenterologist, & Health Dr Jeremy Metters, North Staffordshire Royal Infirmary Deputy Chief Medical Officer, Professor John Farndon, Department of Health † Dr John Tripp, Professor of Surgery, University of Bristol † Department of Child Health, Royal Devon Mrs Gloria Oates, & Exeter Healthcare NHS Trust † Professor Charles Florey, Chief Executive, Oldham NHS Trust Professor Tom Walley, Department of Epidemiology & Dr George Poste, Department of Pharmacological Public Health, Ninewells Hospital & Chief Science & Technology Officer, † Therapeutics, University of Liverpool † Medical School, University of Dundee SmithKline Beecham † Professor John Gabbay, Professor Michael Rawlins, Dr Julie Woodin, Director, Wessex Institute for Health Wolfson Unit of Clinical Pharmacology, Chief Executive, † Research & Development † University of Newcastle-upon-Tyne Nottingham Health Authority Dr Tony Hope, Professor Martin Roland, The Medical School, University of Oxford † Professor of General Practice, Professor Howard Glennester, University of Manchester Professor of Social Science & Mr Hugh Ross, Administration, London School of Chief Executive, The United Bristol Economics & Political Science Healthcare NHS Trust † † Current members

HTA Commissioning Board Chair: Professor Charles Florey, Department of Epidemiology & Public Health, Ninewells Hospital & Medical School, University of Dundee †

Professor Ian Russell, Professor Sir Miles Irving Dr Tim Peters, Department of Health, Sciences & (Programme Director), Professor of Department of Social Medicine, Clinical Evaluation, University of York * Surgery, University of Manchester, University of Bristol † † Dr Doug Altman, Hope Hospital, Salford Professor David Sackett, Director, Institute of Health Sciences, Professor Alison Kitson, Centre for Evidence Based Medicine, Oxford † Director, Royal College of Oxford Nursing Institute † Mr Peter Bower, Professor Martin Severs, Independent Management Consultant, Professor Martin Knapp, Professor in Elderly Health Care, Newcastle-upon-Tyne † Director, Personal Social Services Portsmouth University † Ms Christine Clark, Research Unit, London School of Economics & Political Science Dr David Spiegelhalter, Hon. Research Pharmacist, Hope Hospital, MRC Biostatistics Unit, Institute of † Salford Dr Donna Lamping, Public Health, Cambridge Professor David Cohen, London School of Hygiene & † Dr Ala Szczepura, Professor of Health Economics, Tropical Medicine Director, Centre for Health Services Studies, University of Glamorgan Professor Theresa Marteau, University of Warwick † Mr Barrie Dowdeswell, Director, Psychology & Genetics Chief Executive, Royal Victoria Infirmary, Research Group, UMDS, London Professor Graham Watt, Department of General Practice, Newcastle-upon-Tyne Professor Alan Maynard, Woodside Health Centre, Glasgow † Professor Martin Eccles, Professor of Economics, Professor of Clinical Effectiveness, University of York † Professor David Williams, University of Newcastle-upon-Tyne † Department of Clinical Engineering, Professor Sally McIntyre, University of Liverpool Dr Mike Gill, MRC Medical Sociology Unit, Brent & Harrow Health Authority † Glasgow Dr Mark Williams, Public Health Physician, Bristol Dr Jenny Hewison, Professor Jon Nicholl, Senior Lecturer, Department of Psychology, Director, Medical Care Research Unit, Dr Jeremy Wyatt, University of Leeds † University of Sheffield † Institute for Health Sciences, † Dr Michael Horlington, Professor Gillian Parker, University College London Head of Corporate Licensing, Smith & Nuffield Professor of Community Care, * Previous Chair Nephew Group Research Centre University of Leicester † † Current members HTA

How to obtain copies of this and other HTA Programme reports. An electronic version of this publication, in Adobe Acrobat format, is available for downloading free of charge for personal use from the HTA website (http://www.hta.ac.uk). A fully searchable CD-ROM is also available (see below). Printed copies of HTA monographs cost £20 each (post and packing free in the UK) to both public and private sector purchasers from our Despatch Agents. Non-UK purchasers will have to pay a small fee for post and packing. For European countries the cost is £2 per monograph and for the rest of the world £3 per monograph. You can order HTA monographs from our Despatch Agents: – fax (with credit card or official purchase order) – post (with credit card or official purchase order or cheque) – phone during office hours (credit card only). Additionally the HTA website allows you either to pay securely by credit card or to print out your order and then post or fax it.

Contact details are as follows: HTA Despatch Email: [email protected] c/o Direct Mail Works Ltd Tel: 02392 492 000 4 Oakwood Business Centre Fax: 02392 478 555 Downley, HAVANT PO9 2NP, UK Fax from outside the UK: +44 2392 478 555 NHS libraries can subscribe free of charge. Public libraries can subscribe at a very reduced cost of £100 for each volume (normally comprising 30–40 titles). The commercial subscription rate is £300 per volume. Please see our website for details. Subscriptions can only be purchased for the current or forthcoming volume.

Payment methods Paying by cheque If you pay by cheque, the cheque must be in pounds sterling, made payable to Direct Mail Works Ltd and drawn on a bank with a UK address. Paying by credit card The following cards are accepted by phone, fax, post or via the website ordering pages: Delta, Eurocard, Mastercard, Solo, Switch and Visa. We advise against sending credit card details in a plain email. Paying by official purchase order You can post or fax these, but they must be from public bodies (i.e. NHS or universities) within the UK. We cannot at present accept purchase orders from commercial companies or from outside the UK.

How do I get a copy of HTA on CD? Please use the form on the HTA website (www.hta.ac.uk/htacd.htm). Or contact Direct Mail Works (see contact details above) by email, post, fax or phone. HTA on CD is currently free of charge worldwide.

The website also provides information about the HTA Programme and lists the membership of the various committees.

Antimicrobial prophylaxis in colorectal surgery: a systematic review of randomised controlled trials

F Song AM Glenny

NHS Centre for Reviews and Dissemination University of York UK

Published May 1998

This report should be referenced as follows:

Song F,Glenny AM. Antimicrobial prophylaxis in colorectal surgery: a systematic review of randomised controlled trials. Health Technol Assessment 1998;2(7).

Health Technology Assessment is indexed in Index Medicus/MEDLINE and Excerpta Medica/ EMBASE. Copies of the Executive Summaries are available from the NCCHTA web site (see overleaf). NHS R&D HTA Programme

he overall aim of the NHS R&D Health Technology Assessment (HTA) programme T is to ensure that high-quality research information on the costs, effectiveness and broader impact of health technologies is produced in the most efficient way for those who use, manage and work in the NHS. Research is undertaken in those areas where the evidence will lead to the greatest benefits to patients, either through improved patient outcomes or the most efficient use of NHS resources. The Standing Group on Health Technology advises on national priorities for health technology assessment. Six advisory panels assist the Standing Group in identifying and prioritising projects. These priorities are then considered by the HTA Commissioning Board supported by the National Coordinating Centre for HTA (NCCHTA). This report is one of a series covering acute care, diagnostics and imaging, methodology, pharmaceuticals, population screening, and primary and community care. It was identified as a priority by the Pharmaceutical Panel. The views expressed in this publication are those of the authors and not necessarily those of the Standing Group, the Commissioning Board, the Panel members or the Department of Health. The editors wish to emphasise that funding and publication of this research by the NHS should not be taken as implicit support for the recommendations for policy contained herein. In particular, policy options in the area of screening will, in England, be considered by the National Screening Committee. This Committee, chaired by the Chief Medical Officer, will take into account the views expressed here, further available evidence and other relevant considerations. Reviews in Health Technology Assessment are termed ‘systematic’ when the account of the search, appraisal and synthesis methods (to minimise biases and random errors) would, in theory, permit the replication of the review by others.

Series Editors: Andrew Stevens, Ruairidh Milne and Ken Stein Assistant Editors: Jane Robertson and Jane Royle

The editors have tried to ensure the accuracy of this report but cannot accept responsibility for any errors or omissions. They would like to thank the referees for their constructive comments on the draft document.

ISSN 1366-5278

© Crown copyright 1998

Enquiries relating to copyright should be addressed to the NCCHTA (see address given below).

Published by Core Research, Alton, on behalf of the NCCHTA. Printed on acid-free paper in the UK by The Basingstoke Press, Basingstoke.

Copies of this report can be obtained from:

The National Coordinating Centre for Health Technology Assessment, Mailpoint 728, Boldrewood, University of Southampton, Southampton, SO16 7PX, UK. Fax: +44 (0) 1703 595 639 Email: [email protected] http://www.soton.ac.uk/~hta Health Technology Assessment 1998; Vol. 2: No. 7

Contents

List of abbreviations ...... i Additional antibiotics ...... 39 New-generation cephalosporins ...... 39 Executive summary ...... iii Route of administration ...... 40 Adverse events with antibiotic prophylaxis ... 40 1 Background ...... 1 Bacteriology and antibiotic resistance ...... 40 Cost and cost-effectiveness ...... 40 2 Methods ...... 3 Antibiotic prophylaxis in surgical practice ... 43 Inclusion criteria ...... 3 Literature search strategy ...... 3 5 Conclusions ...... 45 Validity assessment ...... 3 Implications for policy ...... 45 Data extraction ...... 3 Recommendations for research ...... 45 Data synthesis ...... 4 Acknowledgements ...... 47 3 Results ...... 5 Quality of the studies ...... 5 References ...... 49 Antibiotic prophylaxis versus Appendix 1 Medline search strategies ...... 57 no-antibiotic controls ...... 6 Different antibiotics and surgical Appendix 2 Data extraction sheet and wound infection ...... 6 field definitions ...... 59 Other antibiotics ...... 25 Additional antibiotics ...... 29 Appendix 3 RCTs of antimicrobial Single-dose versus multiple-dose regimens ... 31 prophylaxis in colorectal surgery Timing and duration of treatment ...... 31 (1984–95) ...... 63 Other outcomes ...... 31 Appendix 4 Number of comparisons Cost and cost-effectiveness ...... 33 between antibiotics in included trials ...... 105

4 Discussion ...... 35 Health Technology Assessment reports Risk factors ...... 35 published to date ...... 107 Methodological limitations of the study ...... 36 Timing and duration ...... 37 Health Technology Assessment Inadequate antibiotic regimens ...... 38 panel membership ...... 109

Health Technology Assessment 1998; Vol. 2: No. 7

List of abbreviations

CI confidence interval RCT randomised controlled trial SD standard deviation* SWI surgical wound infection UTI urinary tract infection* WI wound infection* i.v. intravenous administration* i.m. intramuscular administration* p.o. oral administration* b.d. twice daily* t.d.s three-times daily*

* Used only in tables

i

Health Technology Assessment 1998; Vol. 2: No. 7

Executive summary

Background studies assessed, 147 RCTs, including a total of Wound infections are the most frequent 23,049 patients, met the review inclusion criteria. nosocomial infections among surgical patients and are related to an increase in morbidity and The principal outcome assessed in the review was the mortality, a prolongation of hospital stay and an incidence of SWI. Where possible, abdominal wound increase in the cost of medical care. Colorectal infections were recorded separately from perineal surgery is associated with a high risk of infection wound infections. Data on other postoperative due to endogenous contamination by bacteria infections and adverse events were also collected. in the contents of the large bowel. Data synthesis It is now generally accepted that antimicrobial Studies were grouped according to the antibiotic prophylaxis is one of many important measures used, route of administration, and number of doses that should be taken to prevent postoperative administered (i.e. single versus multiple doses). wound infections. To achieve the most cost-effective Where appropriate, formal meta-analysis and use of antimicrobial prophylaxis, consideration of investigation of heterogeneity among trials the choice, delivery and regimen of antimicrobial were conducted. agents is necessary. Results Objectives The quality of the RCTs has improved over the last This review evaluates the relative efficacy of 12 years, though there are still many methodolog- antimicrobial prophylaxis in patients undergoing ical problems, including inappropriate method of colorectal surgery where there is a high risk of patient allocation, lack of blinding during outcome surgical wound infection (SWI). assessment, and insufficient sample size. The criteria for inclusion and exclusion were described in 87% of the included trials. The exclusion criteria Methods most frequently used were allergy to study drugs, preoperative use of other antibiotics, impaired Data sources renal or liver function, children or very old Literature searches of Medline, Embase and the patients, pregnancy or lactation, and certain Cochrane Controlled Trials Register were con- types of colorectal operations. ducted to identify randomised controlled trials (RCTs) published between 1984 and 1995, which More than 70 different antibiotic regimens were investigated antimicrobial prophylaxis in the tested in 147 trials. The overall rate of SWI across prevention of postoperative wound infection in all the included trials of antimicrobials prophylaxis patients who had undergone colorectal surgery. (excluding four non-antibiotic groups) was 11.1% Bibliographies of reviews and all identified trials (n = 22,927). were examined to locate additional studies. A sample of key journals was also handsearched. The results of this review confirm that the use of All languages were considered. antimicrobial prophylaxis is generally effective for the prevention of SWIs in colorectal surgery. Some Validity assessment and data extraction antimicrobial regimens appear to be less effective The identified studies were assessed for both than others in this indication. For example, mono- relevance and validity by one reviewer and checked therapy with either metronidazole, doxycycline or by another. Data extraction was carried out by one piperacillin are inadequate for prophylactic reviewer using an electronic data extraction form. treatment in colorectal surgery. This process was again checked by a second reviewer. For articles containing insufficient detail, authors The review found that a single dose or short-term were contacted for clarification. Of all the use of an antimicrobial agent is as efficacious as iii Executive summary

long-term, postoperative use. Pooled results from efficacy, as a reduction in infection rates is associ- 17 trials that compared a single-dose regimen with ated with a shorter hospital stay, the ‘hotel’ costs of a multiple-dose regimen, showed no significant which account for the highest proportion of overall difference in the rate of SWI (odds ratio = 1.17; cost during treatment. 95% confidence interval [CI]: 0.89, 1.54). There is a lack of convincing evidence concerning the importance of a second-dose regimen when Conclusions surgical procedures are longer than 2 hours. The use of antimicrobial prophylaxis is efficacious There is no convincing evidence to suggest that the in the prevention of SWI in colorectal surgery. second- and third-generation cephalosporins are With the exception of a few inadequate regimens, more efficacious than the first-generation cephalo- there is no significant difference in the rate of SWI sporins in this indication (6% versus 6.4%; odds between many regimens. The use of a multiple- ratio = 0.93; 95% CI: 0.46, 1.86). dose regimen may be unnecessary for the preven- tion of SWI, as single-dose regimens have been Establishing the efficacy of different routes demonstrated to be as efficacious as multiple of administration of antibiotic prophylaxis was dosing and in addition, may be associated with complicated by the use of different antibiotics less toxicity, fewer adverse events, less risk of or use of extra antibiotics. No additional benefit developing bacterial resistance and lower costs. was observed in six trials that compared parenteral Similarly, no convincing evidence supports the alone, with parenteral plus topical use of antibiotic idea that the new-generation cephalosporins are prophylaxis. Several trials, showing extra benefit of more efficacious than first-generation cephalo- oral antibiotics, used inadequate parenteral anti- sporins in preventing SWI in colorectal surgery. biotics such as metronidazole alone, or piperacillin alone. Oral or topical application of antibiotics in addition to the parenteral administration of appro- Implications for policy priate antibiotics seem to be of limited value in most cases. Two principles are important to follow when selecting an antimicrobial prophylactic regimen In general, the estimates of efficacy of many of in colorectal surgery: the different regimens included are similar and it is very difficult, if not impossible, to identify the • antibiotics or antibiotic combinations best one. However, the Type-II error or lack of should be active against both aerobic and statistical power cannot be ruled out as a potential anaerobic bacteria; reason for statistically non-significant findings in • the administration of antibiotics should be many small trials. timed to ensure that the tissue concentration of antibiotics around the wound area is sufficiently A total of 74 of the 134 trials published in English high when bacterial contamination occurs. reported adverse events following antibiotic pro- phylaxis in colorectal surgery. Skin rash, diarrhoea, Universal acceptance and use of a regimen should and nausea were commonly mentioned adverse be avoided in order to minimise the development events that may be attributable to the use of some of antibiotic-resistant bacteria. Based on the antibiotic treatments. No serious toxicity or adverse research evidence, guidelines should be developed events were reported except in one trial that locally in order to achieve a more cost-effective use reported postoperative bleeding in some patients of antimicrobial prophylaxis in colorectal surgery. treated with latamoxef.

The costs associated with SWI are high in terms of Recommendations for research both antibiotic treatment and prolonged hospital- isation with some studies reporting an additional Further studies of efficacy may be of little value 12 days in hospital as a result of SWI. Three trials and would require large numbers of patients in that included cost data in comparisons of mono- order to demonstrate a statistically significant therapy and combination therapy showed that difference. Future research should focus on the monotherapy was as effective as the combination understanding of the practical use of antimicrobial regimens but less expensive. The overall cost data prophylaxis in colorectal surgery in the UK and available from the RCTs suggest that drug acquis- the cost-effectiveness of different regimens of iv ition costs need to be viewed in terms of their antibiotic prophylaxis. Health Technology Assessment 1998; Vol. 2: No. 7

Chapter 1 Background

ound infections are the most frequent should not be used in further trials of W nosocomial infections among surgical colorectal surgery.6 patients and are related to an increase in morbidity and mortality, a prolongation of hospital stay and It is now generally accepted that antimicrobial an increase in the cost of medical care.1,2 prophylaxis is one of many important measures that should be taken to prevent postoperative Exogenous or endogenous contamination of the wound infections in contaminated or clean- surgical wound by pathogenic organisms is one of contaminated surgery. Many different antimicrobial the most important factors related to the risk of agents are available with different pharmacokinetic surgical wound infection (SWI), though it does not and pharmacodynamic features, spectra of activity, necessarily mean that a SWI will be inevitable.3 and toxicities. To achieve the most cost-effective According to the risk of bacterial contamination, use of antimicrobial prophylaxis, consideration surgical wounds have conventionally been classified of the choice, delivery and regimen of anti- as clean, clean-contaminated, contaminated, or microbial agents is necessary.1,7,8 However, it is dirty.4 Colorectal surgery is associated with a high often argued that questions regarding the choice risk of infection due to endogenous contamination of antibiotics, and the timing and duration of by bacteria in the contents of the large bowel. treatment have yet to be answered satisfactorily.9 Postoperative wound infections have been shown Other important issues include the ease of admin- to occur in about 40% of patients after colorectal istration and cost-effectiveness. The occurrence surgery in whom antibiotic prophylaxis was of antibiotic-resistant bacteria should also be not used.5 considered when assessing the appropriate use of antimicrobial prophylaxis. A review by Baum and co-workers6 compared the wound infection and mortality rates of patients The main objective of this review, which was given antimicrobial prophylaxis for colorectal commissioned by the NHS HTA programme, is to surgery with those of no-antibiotic controls. By evaluate the relative effectiveness of antimicrobial pooling results from 26 randomised controlled prophylaxis in patients undergoing colorectal trials (RCTs), it was found that the wound infection surgery. Cost-effectiveness of antimicrobial rate was significantly higher in the no-antibiotic prophylaxis is also summarised from the RCTs that control group with 36% of patients developing examined this aspect of treatment. (It should be infections compared with 22% in the antibiotic noted that the following agents are not available prophylaxis groups. They also found that mortality in the UK: cefmatazole, cefonicid, cefoperazone, was significantly reduced by using antibiotic cefotetan, cefotiam, cephalothin, chlorchinaldole, prophylaxis (11.2% versus 4.5%, p < 0.01). It mezlocillin, latamoxef, ornidazole, sulbactam, and was concluded that the no-antibiotic control thymostimulin.)

1

Health Technology Assessment 1998; Vol. 2: No. 7

Chapter 2 Methods

Inclusion criteria method of randomisation, blind outcome assess- This review includes RCTs, published between ment, written definition of SWI, withdrawals, and 1984 and 1995, that evaluate antimicrobial prophy- a priori calculation of sample size. True random- laxis in the prevention of postoperative wound isation is defined as concealed patient allocation infection in patients who have undergone color- until the point of allocation.10 According to ectal surgery. The year 1984 was chosen because examples listed in the Cochrane Collaboration activity in this field at this time included the intro- Handbook,10 the following methods were duction of many new antibiotics, changes in the considered to be true randomisation: clinical use of antibiotic prophylaxis, improvement in surgical procedures, possible emergence of • centralised or pharmacy-controlled randomisation antibiotic-resistant micro-organisms, and the • pre-numbered or coded identical containers large volume of literature on this topic. which are administered serially to participants • on-site computer system combined with group assignments in a locked unreadable computer Literature search strategy file that can be accessed only after entering characteristics of an enrolled subjects A search of Medline from 1984, Embase/Embase • sequentially numbered, sealed, opaque envelopes. Alert (1983–present) and the Cochrane Controlled Trials Register was conducted with the assistance of the NHS Centre for Reviews and Dissemination Data extraction information services (appendix 1). The titles and abstracts were assessed and copies of relevant The wound itself has the greatest risk of infection studies collected. The references of retrieved from endogenous bacterial contamination during articles were checked to locate other relevant trials. surgical procedures.11 In this review, therefore, Studies in all languages were considered for the the rate of SWI is used as the principal outcome evaluation, surgical wound outcome and study measure to assess the relative effectiveness of anti- quality. For reasons of difficulty and cost of microbial prophylaxis in colorectal surgery. The translation, studies published in non-English definition and diagnosis of SWIs often vary in pub- languages were not used in the discussion of lished studies. If possible, only abdominal wound adverse events, bacteriological test, and cost issues. infections are included because of their reliability. However, in some trials it was not clear whether peri- A sample of several key journals (Acta Chirurgica neal wound infections or other infections related to Scandinavica, British Journal of Surgery, Journal of the surgical procedures were included. In these Antimicrobial Chemotherapy, Archives of Surgery, Annals cases, the reported results are presented. of Surgery) were handsearched in addition to the electronic database searches. No additional trials Data on other postoperative infections (systemic were located during handsearching. infections) were also collected from each trial. Remote infection (such as respiratory tract infections, urinary tract infections) may be Validity assessment important but the definition and diagnosis is more problematic. Anastomotic leakage is not The relevance and validity of each study were considered in this review because it is mainly assessed and data were extracted by one reviewer associated with the surgical technique used and and checked by another. Data from individual complex patient characteristics. Other outcomes studies were extracted using a pre-defined form that are important include prolonged hospital- (appendix 2) and managed using an Idealist data- isation and increased cost of medical care. A base. When there was insufficient data on outcome data extraction sheet is detailed in appendix 2. of SWI in the article, authors were contacted for clarification (three authors were contacted, two Mortality data were recorded where available, replied). Indicators of study quality included as mortality is a reliable and sensitive outcome 3 Methods

measure to compare the results of antibiotic Review Manager Software (RevMan, version 3.0 prophylaxis with no-antibiotic controls in colorectal for Windows). The Mantel-Haenszel method was surgery.6 However, when different regimens of chosen for calculating overall odds ratios and the antibiotic prophylaxis are compared, the mortality chi-squared method for testing heterogeneity rate is too low for comparing relative efficacy.12 between individual studies.

It is often inappropriate to pool results from trials Data synthesis quantitatively. For example, trials may compare an agent with other regimens in many different ways, The studies were grouped according to the such as alone or in combination with other agents. antibiotic used. Where appropriate (e.g. when Therefore, results from individual studies are the same regimen of antibiotic prophylaxis was described in narrative form. To facilitate discussion, assessed in several trials), formal meta-analysis figures are used to present the results of SWI for and investigation of heterogeneity among trials antimicrobial prophylaxis regimens that have was conducted using MetaView in Cochrane been assessed in four or more trials.

4 Health Technology Assessment 1998; Vol. 2: No. 7

Chapter 3 Results

total of 147 RCTs met the inclusion criteria 27 trials that included both elective and emergency A for this review. Details of each trial, including colorectal surgery. the quality assessment, type of colorectal surgery, antibiotic prophylaxis used, clinical results, and conclusions are shown in appendix 3.13–156 Among Quality of the studies these RCTs, 134 were published in English and 13 in other languages. The trials were conducted Table 1 summarises the results of the quality in 25 different countries, mainly in Europe and assessment of RCTs. The proportion of the trials North America. A quarter of the trials were carried that used truly concealed randomisation was 37% out in the UK, 12% in the USA, 10% in Germany, and the proportion of trials that assessed outcome 10% in Denmark, 6% in Italy, Sweden and blindly was 40%. The majority of the trials did not Australia. The review included 24 trials of patients calculate the required sample size (84%) and many who underwent abdominal surgery, and from included fewer than 100 patients (39%). The with- which the results for colorectal surgery could drawals were not reported in 22% of the trials and be separated from other procedures. follow-up period was not clear for 33% of the trials. The proportion of the trials that did not give a The overall rate of SWI with various regimens written definition of SWI was 21%. of prophylactic antimicrobial agents in colorectal surgery was 11.1% (2540/22,927) When the studies were grouped according to after excluding the four no-antibiotic arms. the year of publication, it can be seen that the The rate of SWI was 10.6% in the 120 trials quality of trials has improved over the last 12 years that included patients undergoing elective (Table 1). The proportion of trials in which the colorectal surgery only, and 13.4% in the method of randomisation was truly concealed

TABLE 1 Summary of quality assessment of RCTs of antibiotic prophylaxis in colorectal surgery (1984–95)

Year of publication 1984–87 1988–91 1992–95 All trials

True randomisation 30% 32.8% 65.2% 36.7%

Blinding assessment 44.4% 34.4% 43.5% 40.1%

A priori calculation of sample size 6.3% 19.7% 34.8% 16.3%

Definition of SWI 76.2% 80.3% 82.6% 78.9%

Follow-up, 28 days or longer 44.4% 62.3% 73.9% 56.5%

Withdrawal rate < 10% 28.6% 45.9% 34.8% 36.7% Unsure 38.1% 13.1% 4.3% 22.4%

Sample size of the trials < 100 55.6% 34.4% 4.3% 38.8% 100–299 42.9% 52.4% 56.5% 49.0% > 300 1.6% 13.1% 39.1% 12.2%

No. of centres in the trials Single centre 77.8% 70.5% 11.7% 64.4% 2–4 centres 11.1% 14.8% 15.0% 13.3% 5 centres or more 11.1% 14.8% 30.4% 15.7%

Total number of trials 63 61 23 147 5 Results

was 30%, 33%, and 65%, respectively, for the three separately for each antibiotic agent that has been time periods, 1984–87, 1988–91, and 1992–95. assessed in at least four trials. Trials published between 1992 and 1995 were also of better quality in terms of a priori calculation of Ampicillin sample size, definition of SWI, appropriate follow- Ampicillin is a broad-spectrum penicillin and is up period, and the reporting of withdrawals. active against certain Gram-positive and Gram- negative organisms. It may be inactivated by penicil- More multicentre studies have also been published linases produced by bacteria such as Staphylococcus more recently and, in addition, patient numbers aureus and Escherichia coli. When it is administered have risen; the sample size was smaller than 100 in orally, less than half is absorbed. more than 50% of the trials published between 1984 and 1987 compared with only 4.3% of trials Five trials compared ampicillin plus metronidazole published since 1992. with other regimens of prophylactic antibiotics in colorectal surgery (Table 3).56,69,90,133,134 Two trials showed that ampicillin plus metronidazole was Antibiotic prophylaxis versus more effective than metronidazole alone in the 56,133 no-antibiotic controls prevention of SWI. In the trial by Jensen and co-workers,69 it was found that ampicillin plus The review identified four trials published since metronidazole was associated with a significantly 1984 that compared patients receiving antibiotic higher SWI rate than cefuroxime plus metronid- prophylaxis for colorectal surgery with a control azole in 39 patients who underwent emergency group not given antibiotics (Table 2).53,56,137,150 colorectal surgery, but there was no difference in The antibiotics used prophylactically in these four 272 patients who underwent elective colorectal trials were gentamicin plus metronidazole,53 metro- surgery. Two other trials did not find a difference nidazole alone or metronidazole plus ampicillin,56 between ampicillin plus metronidazole and mezlocillin plus oxacillin,137 and cefoxitin.150 The ceftriaxone alone or cefoxitin alone.90,134 results from the individual studies showed consist- ently that the SWI rate was much lower in the No statistically significant differences were found antibiotic groups than that in the control groups when ampicillin plus sulbactam was compared with (12.9% versus 40.2%; pooled odds ratio = 0.24; gentamicin plus metronidazole15 and cefoxitin.41 95% CI: 0.13, 0.43). The use of topical ampicillin in addition to oral antibiotics or in addition to intravenous cefotaxime was not found to have an additional benefit.128,129 Different antibiotics and surgical wound infections Aztreonam Aztreonam is a monocyclic beta-lactam antibiotic Few regimens were directly compared, each with an antibacterial spectrum limited to Gram- comparison being evaluated by a limited number negative aerobic bacteria. Therefore, it is often of trials, in which the sample size was often too combined with clindamycin (active against Gram- small to have sufficient statistical power (appendix positive and anaerobic bacteria) in trials of 4). Below, the results of SWI rates are discussed antibiotic prophylaxis in colorectal surgery.

TABLE 2 Antibiotic prophylaxis versus no-antibiotic control

Study [reference] Prophylaxis Control Odds ratio Odds ratio n/N n/N (95% CI fixed) [95% CI fixed]

Gomez-Alonso [53] 6/35 13/31 0.29 [0.09, 0.89] Gottrup [56] 11/94 13/41 0.29 [0.11, 0.71] Schiessel [137] 2/29 12/31 0.12 [0.02, 0.59] Utley [150] 3/13 11/19 0.13 [0.04, 1.06]

Total (95% CI) 22/171 49/122 0.24 [0.13, 0.43]

0.01 0.1 1 10 100 (log scale) Favours prophylaxis Favours control 6 Health Technology Assessment 1998; Vol. 2: No. 7

TABLE 3 Ampicillin versus other antibiotics

Regimen and study Incidence of SWI Odds ratio Odds ratio [reference] (95% CI fixed) [95% CI fixed] Ampicillin Other n/N antibiotics n/N

(Ampicillin + metronidazole) × 10 vs. 1/48 10/46 0.8 [0.01, 0.63] metronidazole × 10 Gottrup [56]

(Ampicillin + metronidazole) × 1 vs. 1/188 6/170 0.15 [0.02, 1.23] (metronidazole × 1) Roland [133]

(Ampicillin × 12) + (metronidazole × 9) 14/175 13/177 1.10 [0.50,2.41] vs. (cefoxitin × 3) Rorbaek-Madsen [134]

(Ampicillin + metronidazole) × 1 vs. 5/36 4/32 1.13 [0.28, 4.63] (ceftriaxone × 1) Luke [90]

(Ampicillin + metronidazole) × 3 vs. 4/12 1/16 7.50 [0.71, 78.91] (cefuroxime + metronidazole) × 1 Jensen [69]

(Ampicillin + metronidazole) × 1 vs. 6/11 1/16 18.00 [1.72, 188.09] (cefuroxime + metronidazole) × 1 Jensen [69]

(Ampicillin + metronidazole) × 3 vs. 8/92 7/91 1.14 [0.40, 3.29] (cefuroxime + metronidazole) × 1 Jensen [69]*

(Ampicillin + metronidazole) × 1 vs. 8/89 7/91 1.19 [0.41, 3.42] (cefuroxime + metronidazole) × 1 Jensen [69]*

(Ampicillin + sulbactam) × 3 vs. 6/63 7/65 0.87 [0.28, 2.75] (gentamicin + metronidazole) × 3 AhChong [15]

(Ampicillin + sulbactam) × 4 vs. 7/44 9/48 0.82 [0.28, 2.43] (cefoxitin × 4) De La Hunt [41]

(Ampicillin + neomycin + erythromycin) 6/50 2/50 3.27 [0.63, 17.07] × 3 vs. (cefotaxime × 3) Raahave [128]

(Ampicillin × 1) + (cefotaxime × 3) 5/81 6/89 0.91 [0.27,3.10] vs. (cefotaxime × 3) Raahave [129]

0.01 0.02 1 50 1000 (log scale) Favours ampicillin Favours other antibiotics

*Estimated number of SWI 7 Results

Aztreonam plus metronidazole was a less mycin is associated with an increased risk of pseudo- efficacious regimen than cefotaxime plus membranous colitis which may be fatal. Clinda- metronidazole (32.4% versus 12.9%) for the mycin plus gentamicin was less effective than prevention of SWIs in colorectal surgery.106 The clindamycin plus aztreonam.109,132 The difference authors claimed that this was because aztreonam was not statistically significant when clindamycin plus metronidazole provided good Gram-negative plus gentamicin were compared with cefoxitin, and anaerobic cover but little activity against gentamicin plus lincomycin, or gentamicin Gram-positive bacteria. plus metronidazole.32,89

Four trials compared aztreonam plus clindamycin Cefotaxime with other antibiotics (Table 4).25,50,109,132 Two very Cefotaxime is a ‘third-generation’ cephalosporin small trials compared aztreonam plus clindamycin with great activity against Gram-negative bacteria. with cefotetan25 or ceftriaxone50 and no significant differences were found. By pooling the results from Most trials showed no significant difference in SWI two trials,109,132 it was found that the rate of SWI in rate when cefotaxime alone was compared with the aztreonam plus clindamycin group was signifi- other regimens (Table 5), such as ticarcillin plus cantly lower than that of the gentamicin plus clavulanic acid72 (Timentin®, SmithKline Beecham clindamycin group (7.2% versus 13.6%, overall Pharmaceuticals, Welwyn, UK), cefoperazone,70 odds ratio = 0.51; 95% CI: 0.3, 0.87). penicillin plus streptomycin,84 mezlocillin,98 cefoxitin,139 and oral neomycin plus erythromycin Clindamycin plus topical ampicillin.128 Jones and co-workers71 Clindamycin is active against Gram-positive cocci, found that cefotaxime as a single dose (second including penicillin-resistant staphylococci and dose if operation continued over 2 hours) was less also against many anaerobes, particularly Bacteroides effective than four doses of cephazolin or cefoxitin fragilis, and is often combined with other antibiotics when all patients received intraluminal neomycin such as aztreonam or gentamicin. However, clinda- and erythromycin (14% versus 3%, p = 0.057).

TABLE 4 Aztreonam versus other antibiotics

Regimen and study Incidence of SWI Odds ratio Odds ratio [reference] (95% CI fixed) [95% CI fixed] Aztreonam + Other clindamycin antibiotics n/N n/N

(Aztreonam + metronidazole) × 3 vs. 23/71 9/70 3.25 [1.38, 7.66] (cefotaxime + metronidazole) × 3 Morris [106]

(Aztreonam + clindamycin) × 3 vs. 0/26 1/32 0.40 [0.02, 10.14] (cefotetan × 3) Bellantone [25]

(Aztreonam + clindamycin) × 5 vs. 1/18 0/14 2.49 [0.09, 65.76] (ceftriaxone × 1) Franceshini [50]

(Aztreonam + clindamycin) × 3 vs. 8/66 12/72 0.69 [0.26, 1.81] (gentamicin + clindamycin) × 3 Rodolico [132]

(Aztreonam + clindamycin) × 3 vs. 13/224 29/230 0.43 [0.22, 0.84] gentamicin + clindamycin) × 3 Mozzillo [109]

0.01 0.1 1 10 100 (log scale) Favours aztreonam + clindamycin Favours other antibiotics

8 Health Technology Assessment 1998; Vol. 2: No. 7

Five trials compared cefotaxime plus metronidazole appeared to be as efficacious as three doses of with other regimens (Table 5). It was found that a cefoxitin with or without metronidazole.110 regimen of three doses of cefotaxime plus metro- nidazole was significantly more effective than Ceftriaxone aztreonam plus metronidazole.106 Other trials did Ceftriaxone is a third-generation cephalosporin not find significant differences between cefotaxime with a longer half-life than other cephalosporins. plus metronidazole and other regimens such as co-amoxiclav,82 cefoxitin,81 ceftriaxone plus ornida- The difference in the SWI rate was not significant zole,49 and cefuroxime plus metronidazole.135 when ceftriaxone alone was compared with aztre- onam plus clindamycin,50 ampicillin plus metro- When cefotaxime plus metronidazole in a single nidazole,90 cephazolin plus metronidazole,91 cefoxi- dose was compared with multiple doses, no signifi- tin,110 or cefotiam plus gentamicin and metronida- cant difference was found.81,130 Hakansson and co- zole155 (Table 8). Matikainen and Hiltunen92 found workers59 showed that a single dose of cefotaxime that ceftriaxone plus tinidazole was more effi- plus metronidazole was significantly more effective cacious than netilmicin plus tinidazole. Ceftriaxone than three doses of cefotaxime alone (6.6% versus plus ornidazole was as efficacious as amikacin plus 15.7%, p = 0.0006). In other trials, however, cefo- metronidazole.147 Ceftriaxone plus metronidazole taxime alone was as efficacious as cefotaxime plus was as efficacious as ceftazidime plus metronida- metronidazole,47 or plus topical ampicillin.129 zole,51 cefamandole plus metronidazole,62 and co-amoxiclav,113 and more efficacious than genta- Cefotetan micin plus metronidazole.30 Two trials compared Cefotetan is a long-acting, broad-spectrum, third- ceftriaxone plus metronidazole with oral neomycin generation cephalosporin. plus erythromycin. In the trial by Weaver and co- workers,152 ceftriaxone plus metronidazole was Seven trials compared cefotetan alone with other significantly more efficacious than oral neomycin regimens (Table 6).19,25,64,68,107,117,118,140,149 No signifi- plus erythro-mycin (9.7% versus 41.4%, p = 0.005). cant difference was observed when cefotetan alone In the trial by Kling and Dahlgren80 there were no was compared with cefoxitin,68,117 gentamicin plus abdominal wound infections and the rate of peri- metronidazole,149 co-amoxiclav,19 cefuroxime plus neal wound infection was not significantly different metronidazole,140 piperacillin,64 or clindamycin between the two groups. plus aztreonam.25 It was found that cefotetan alone was as effective as cefotetan plus metronidazole,107 Cefuroxime and cefotetan plus thymostimulin.118 Greig and Cefuroxime is a second-generation cephalosporin, co-workers57 found no extra benefit when cefotetan which are less susceptible than the earlier cephalo- was used as a topical application in addition to sporins to inactivation by penicillinases. intravenous gentamicin plus metronidazole. In the RCTs reported in this review, cefuroxime Cefoxitin was almost always combined with metronidazole Cefoxitin is a second-generation cephamycin that (Table 9). Cefuroxime plus metronidazole was is active against bowel flora including B. fragilis. more efficacious than metronidazole alone,35,39,63,101 piperacillin alone,131 and co-amoxiclav.116 Three trials There was no significant difference in SWI rate compared cefuroxime plus metronidazole with mez- when cefoxitin alone was compared with other locillin alone.33,44,143 A statistically significant differ- antibiotic regimens in 15 trials (Table 7). Cefoxitin ence was found in only one of these trials.33 When alone was at least as efficacious as cefoxitin plus oral the results of these three trials were pooled, the rate neomycin (or tinidazole) and erythromycin.37,120,141 of SWI in the cefuroxime plus metronidazole group was significantly lower than that in the mezlocillin In addition to oral neomycin and erythromycin, group (odds ratio = 0.55; 95% CI: 0.32, 0.93). three doses of cefoxitin, (2 g) seemed to be less efficacious compared with a single dose The results with cefuroxime plus metronidazole of cefmetazole (2 g) in the prevention of post- were not significantly different from those with operative wound infection.45,126 Cefoxitin plus ampicillin plus metronidazole,69 cefoxitin,38 netilmi- oral neomycin and erythromycin was more cin plus metronidazole,63 imipenem,75 latamoxef,77 efficacious than oral neomycin and erythro- imipenem plus cilastatin,115 cefotaxime plus metro- mycin alone138 and was of similar efficacy as oral nidazole,135 or cefotetan.140 No significant difference neomycin and erythromycin plus cefonicid46 was found in a small trial that compared a single or plus cephazolin.71 Single-dose ceftriaxone dose of cefuroxime plus metronidazole with three 9 Results

TABLE 5 Cefotaxime versus other antibiotics

Regimen and study Incidence of SWI Odds ratio Odds ratio [reference] (95% CI fixed) [95% CI fixed] Cefotaxime Other n/N antibiotics n/N

(Cefotaxime × 3) vs. (neomycin + erythromycin) 2/50 6/50 0.31 [0.06, 1.59] × 3 + (ampicillin × 1) Raahave [128] (Cefotaxime × 1) vs. (cefoperazone × 1) 1/24 2/23 0.46 [0.04, 5.41] Jones [70] (Cefotaxime × 1) vs. (mezlocillin × 1) 3/52 2/48 1.41 [0.23, 8.81] Mendel [98] (Cefotaxime × 1) vs. (cefoxitin × 5) 7/36 3/34 2.49 [0.59, 10.57] Shatney [139] (Cefotaxime × 1) vs. (cephazolin or cefoxitin) × 4 4/29 2/71 5.52 [0.95, 32.02] Jones [71] (Cefotaxime × 4) vs. (ticarcillin/clavulanic acid × 1) 1/8 0/11 4.60 [0.16, 128.55] Jones [72] (Cefotaxime + metronidazole) × 3 vs. 9/70 23/71 0.31 [0.13, 0.73] (aztreonam + metronidazole) × 3 Morris [106] (Cefotaxime × 4) + (metronidazole × 3) vs. 4/102 6/122 0.79 [0.22, 2.88] (ceftriaxone + ornidazole) × 1 Fingerhut [49] (Cefotaxime + metronidazole) × 3 vs. (co-amoxiclav × 3) 8/88 7/76 0.99 [0.34, 2.86] Kwok [82] (Cefotaxime + metronidazole) × 1 vs. (cefuroxime + 33/454 32/453 1.03 [0.62, 1.71] metronidazole) × 3 Rowe-Jones [135] (Cefotaxime + metronidazole) × 1 vs. cefoxitin × 1 7/84 10/73 0.57 [0.21, 1.59] Kow [81] (Cefotaxime × 3) + (metronidazole × 1) vs. (cefoxitin × 3) 9/81 8/81 1.14 [0.42, 3.12] Kow [81] (Cefotaxime × 3) + (metronidazole × 1) × 3 vs. 9/81 7/84 1.37 [0.49, 3.83] (cefotaxime + metronidazole) × 1 Kow [81] [All: Metronidazole x 6] (Cefotaxime × 3) vs. (cefotaxime × 1) 1/17 2/17 0.47 [0.04, 5.72] Rangabashyam [130] (Cefotaxime × 3) vs. (cefotaxime +metronidazole) × 1 44/280 19/287 2.63 [1.49, 4.63] Hakansson [59] (Cefotaxime × 3) vs. (cefotaxime × 3) + ampicillin 6/89 5/81 1.10 [0.32, 3.75] Raahave [129] (Cefotaxime × 4) vs. (cefotaxime × 4) + metronidazole 4/74 10/143 0.76 [0.23, 2.51] Favre [47] (Cefotaxime × 3) vs. (penicillin × 19) + (streptomycin × 9) 0/52 1/48 0.30 [0.01, 7.58] Lauridsen [84] (Cefotaxime + metronidazole) × 10 + cefotaxime vs. 2/21 4/19 0.39 [0.06, 2.45] (cefotaxime + metronidazole) × 10 Moesgaard [104] 0.01 0.1 1 10 100 (log scale) Favours cefotaxime Favours other antibiotics 10 Health Technology Assessment 1998; Vol. 2: No. 7

TABLE 6 Cefotetan versus other antibiotics

Regimen and study Incidence of SWI Odds ratio Odds ratio [reference] (95% CI fixed) [95% CI fixed] Cefotetan Other n/N antibiotics n/N

(Cefotetan × 2) vs. (gentamicin × 15) 10/75 17/75 0.52 [0.22,1.24] + metronidazole × 1 Tudor [149]

(Cefotetan × 2) vs. (cefuroxime + 10/68 5/36 1.07 [0.34, 3.41] metronidazole) × 3 Skipper [140]

(Cefotetan × 1) vs. (piperacillin × 3) 14/75 13/78 1.15 [0.50, 2.64] Hershman [64]

(Cefotetan × 1) vs. (co-amoxiclav × 1) 4/103 2/105 2.08 [0.37, 11.62] Arnaud [19]

(Cefotetan × 3) vs. (clindamycin + 1/32 0/26 2.52 [0.10, 64.58] aztreonam) × 3 Bellantone [25]

(Cefotetan × 1) vs. (cefoxitin × 5) 25/164 11/75 1.05 [0.49, 2.26] Jagelman [68]

(Cefotetan × 1) vs. (cefoxitin × 4) 18/197 23/206 0.80 [0.42, 1.53] Periti [117]

(Cefotetan × 2) vs. (cefotetan × 2) + 39/278 34/253 1.05 [0.64, 1.72] (metronidazole × 3) Morton [107]

(Cefotetan × 1) vs. (cefotetan × 1) + (thymostimulin × 7) 72/434 56/425 1.31 [0.90, 1.91] Periti [118]

[All: Gentamicin + metronidazole] 15/64 18/65 0.80 [0.36, 1.77] Cefotetan in saline lavage vs. saline only Greig [57] 0.01 0.1 1 10 100 (log scale) Favours cefotetan Favours other antibiotics doses of the same antibiotics (10.5% versus 3.4%, Single or multiple doses of co-amoxiclav have p = 0.554).14 Cefuroxime alone appeared to be less been compared with other regimens of prophy- efficacious than cefuroxime plus metronidazole in lactic antibiotics in eight trials of colorectal a small trial but the difference was not statistically surgery (Table 10). In one trial it was found that significant (9.4% versus 0%, p = 0.24).38 co-amoxiclav was significantly less efficacious than cefuroxime plus metronidazole (p = 0.045).116 Co-amoxiclav However, the significant difference observed in Co-amoxiclav (Augmentin®, SmithKline Beecham this trial was associated with a higher rate of re- Pharmaceuticals, Welwyn, UK), is a combination of operation for postoperative complications such amoxycillin and clavulanic acid. Amoxycillin has a as anastomotic leak or haematoma. similar antibacterial spectrum to ampicillin but is better absorbed than ampicillin when given orally. Seven trials did not find differences in the SWI Clavulanic acid is used to inactivate penicillinases. rate between co-amoxiclav arms and other anti- Therefore, it is believed that co-amoxiclav is active biotics such as cephradine plus metronidazole,145 against penicillinase-producing bacteria that are gentamicin plus metronidazole,61 ceftriaxone plus resistant to amoxycillin. metronidazole,113 cefotaxime plus metronidazole,82 11 Results

TABLE 7 Cefoxitin versus other antibiotics

Regimen and study Incidence of SWI Odds ratio Odds ratio [reference] (95% CI fixed) [95% CI fixed] Cefoxitin Other n/N antibiotics n/N

(Cefoxitin × 10) vs. (cephalothin × 18) 0/30 3/30 0.13 [0.01, 2.61] Antonelli [17]

(Cefoxitin × 5) vs. (cefotaxime × 1) 3/34 7/36 0.40 [0.09, 1.70] Shatney [139]

(Cefoxitin × 3) vs. (metronidazole + tobramycin) × 3 4/27 7/26 0.47 [0.12,1.86] + (penicillin × 5) Perrott [119]

(Cefoxitin × 3) vs. (ampicillin × 12) + (metronidazole × 9) 13/177 14/175 0.91 [0.42, 2.00] Rorbaek-Madsen [134]

(Cefoxitin × 5) vs. (cefotetan × 1) 11/75 25/164 0.96 [0.44, 2.06] Jagelman [68]

(Cefoxitin × 4) vs. (cephazolin × 4) 1/36 1/35 0.97 [0.06, 16.16] Jones [71]

(Cefoxitin × 2) vs. (gentamicin + clindamycin) × 3 2/37 3/35 0.61 [0.10, 3.89] Mosimann [108]

(Cefoxitin × 3) vs. (gentamicin + clindamycin) × 2 7/30 6/30 1.22 [0.36, 4.17] Cainzos [32]

(Cefoxitin × 4) vs. (ampicillin + sulbactam) × 4 9/48 7/44 1.22 [0.41, 3.61] De La Hunt [41]

(Cefoxitin × 1) vs. (cefotaxime + metronidazole) × 1 10/73 7/84 1.75 [0.63, 4.85] Kow [81]

(Cefoxitin × 3) vs. (cefotaxime × 3) + (metronidazole × 1) 8/81 9/81 0.88 [0.32, 2.40] Kow [81]

(Cefoxitin × 4) vs. (cefotetan × 1) 23/206 18/197 1.25 [0.65, 2.39] Periti [117]

(Cefoxitin × 1) vs. (cefuroxime + metronidazole) × 1 2/31 0/30 5.17 [0.24, 112.29] Corman [38]

(Cefoxitin × 1) vs. (cefuroxime × 1) 2/31 3/32 0.67 [0.10, 4.29] Corman [38]

(Cefoxitin × 3) vs. (neomycin + erythromycin) × 3 2/51 3/44 0.56 [0.09, 3.50] Stellato [141]

(Cefoxitin × 1) vs. (piperacillin × 1) 5/30 2/30 2.80 [0.50, 15.73] Armengaud [18] 0.01 0.1 1 10 100 (log scale) Favours cefoxitin Favours other antibiotics

continued 12 Health Technology Assessment 1998; Vol. 2: No. 7

TABLE 7 contd Cefoxitin versus other antibiotics

Regimen and study Incidence of SWI Odds ratio Odds ratio [reference] (95% CI fixed) [95% CI fixed] Cefoxitin Other n/N antibiotics n/N

(Cefoxitin × 3) vs. (ampicillin + sulbactam) × 3 12/126 10/142 1.39 [0.58, 3.34] Menzel [99]

(Cefoxitin × 3) vs. (piperacillin + 12/126 12/137 1.10 [0.47, 2.54] metronidazole) × 3 Menzel [99]

(Cefoxitin × 5) vs. (cefoxitin × 5) + (neomycin 15/141 9/169 2.12 [0.90, 5.00] + erythromycin) × 3 Coppa [37]

(Cefoxitin × 3) vs. (cefoxitin × 3) + (tinidazole × 1) + 0/39 4/41 0.11 [0.01, 2.03] (neomycin × 3) Peruzzo [120]

(Cefoxitin × 3) vs. (cefoxitin + neomycin + 2/51 3/51 0.65 [0.10, 4.08] erythromycin) × 3 Stellato [141]

(Cefoxitin × 3) vs. (ceftriaxone × 1) 3/45 6/45 0.46 [0.11, 1.98] Nel [110]

[All: (Neomycin + erythromycin)] (Cefoxitin × 3) vs. 3/51 3/44 0.85 [0.16, 4.46] no additional antibiotics Stellato [141]

[All: (Neomycin + erythromycin)] (Cefoxitin × 3) vs. 5/101 14/96 0.31 [0.11, 0.88] no additional antibiotics Schoetz [138]

[All: (Neomycin + erythromycin)] (Cefoxitin × 3) 2/12 0/28 13.57 [0.60,306.66] vs. (cefmetazole × 3) Plouffe [126]

[All: (Neomycin + erythromycin) × 3] (Cefoxitin × 3) 4/32 1/63 8.86 [0.95,82.89] vs. (cefmetazole × 1) DiPiro [45]

[All: (Neomycin + erythromycin) × 3] (Cefoxitin × 5) 1/17 2/23 0.66 [0.05,7.89] vs. (cefonicid × 1) Fabian [46]

[All: (Neomycin + erythromycin) × 3] (Cefoxitin × 5) 1/36 1/35 0.97 [0.06,16.16] vs. (cephazolin × 4) Jones [71]

0.01 0.1 1 10 100 (log scale) Favours cefoxitin Favours other antibiotics

13 Results

TABLE 8 Ceftriaxone versus other antibiotics

Regimen and study Incidence of SWI Odds ratio Odds ratio [reference] (95% CI fixed) [95% CI fixed] Ceftriaxone Other n/N antibiotics n/N

(Ceftriaxone × 1) vs. (aztreonam + 0/14 1/18 0.40 [0.02, 10.64] clindamycin) × 5 Franceshini [50]

(Ceftriaxone × 1) vs. (ampicillin + 4/32 5/36 0.89 [0.22, 3.63] metronidazole) × 1 Luke [90]

(Ceftriaxone + glycerol suppository) × 1 vs. 8/94 7/96 1.18 [0.41, 3.40] (cephazolin × 1) + metronidazole suppository Lumley [91]

(Ceftriaxone × 1) vs. (cefoxitin × 3) 6/45 3/45 2.15 [0.50, 9.21] Nel [110]

(Ceftriaxone + metronidazole) × 1 vs. 1/30 2/30 0.48 [0.04, 5.63] (ceftazidime + metronidazole) × 4 Garcia [51]

(Ceftriaxone + metronidazole) × 1 vs. 4/56 6/75 0.88 [0.24, 3.30] (cefamandole + metronidazole) × 1 Hall [62]

(Ceftriaxone + metronidazole) × 2 vs. 4/100 4/100 1.00 [0.24, 4.11] (co-amoxiclav) × 2 Nyam [113]

(Ceftriaxone + metronidazole) × 1 vs. 5/59 14/61 0.31 [0.10, 0.93] (gentamicin + metronidazole) × 1 Burdon [30]

(Ceftriaxone × 1) vs. (cefotiam + gentamicin 1/30 2/30 0.48 [0.04, 5.63] + metronidazole) × 3 Wohlfart [155]

(Ceftriaxone + tinidazole) × 1 vs. 8/315 38/313 0.19 [0.09, 0.41] (netilmicin + tinidazole) × 1 Matikainen [92]

(Ceftriaxone + ornidazole) × 3 vs. 2/25 1/25 2.09 [0.18, 24.62] (amikacin × 5) + metronidazole × 7 Tsimoyiannis [147]

(Ceftriaxone + metronidazole) × 1 vs. 2/27 1/27 2.08 [0.18, 24.41] (neomycin + erythromycin) × 3 Kling [80]

(Ceftriaxone + metronidazole) × 1 vs. 3/31 12/29 0.15 [0.04, 0.62] (neomycin + erythromycin) × 3 Weaver [152] 0.01 0.1 1 10 100 (log scale) Favours ceftriaxone Favours other antibiotics

14 Health Technology Assessment 1998; Vol. 2: No. 7

TABLE 9 Cefuroxime plus metronidazole versus other antibiotics

Regimen and study Incidence of SWI Odds ratio Odds ratio [reference] (95% CI fixed) [95% CI fixed] Cefuroxime + Other metronidazole antibiotics n/N n/N

(Cefuroxime + metronidazole) × 1 vs. 1/16 6/11 0.06 [0.01, 0.58] (ampicillin + metronidazole) × 1 Jensen [69]

(Cefuroxime + metronidazole) × 1 vs. 1/16 4/12 0.13 [0.01, 1.40] (ampicillin + metronidazole) × 3 Jensen [69]

(Cefuroxime + metronidazole) × 1 vs. 7/91 8/89 0.84 [0.29, 2.43] (ampicillin + metronidazole) × 1 Jensen [69]

(Cefuroxime + metronidazole) × 1 vs. 7/91 8/92 0.88 [0.30, 2.52] (ampicillin + metronidazole) × 3 Jensen [69]

(Cefuroxime + metronidazole) × 3 vs. 0/13 0/12 Not estimable (metronidazole + netilmicin) × 3 Haverkorn [63]

(Cefuroxime + metronidazole) × 1 vs. 0/30 3/32 0.14 [0.01, 2.79] (cefuroxime × 1) Corman [38]

(Cefuroxime + metronidazole) × 1 vs. 0/30 2/31 0.19 [0.01, 4.20] (cefoxitin) × 1 Corman [38]

(Cefuroxime + metronidazole) × 1 vs. 0/30 0/27 Not estimable cefoxitin × 4 Corman [38]

(Cefuroxime + metronidazole) × 3 vs. 2/79 8/69 0.20 [0.04, 0.97] (co-amoxiclav × 3) Palmer [116]

(Cefuroxime × 3) + (metronidazole × 4) vs. 8/119 18/104 0.34 [0.14, 0.83] (piperacillin × 4) Reynolds [131]

(Cefuroxime × 3) + (metronidazole × 4) vs. 8/119 9/107 0.78 [0.29,2.11] (piperacillin × 4) + (metronidazole × 6) + (neomycin × 8) Reynolds [131]

(Cefuroxime + metronidazole) × 1 vs. 3/27 7/33 0.46 [0.11, 2.00] (metronidazole × 1) Mittermayer [101]

(Cefuroxime + metronidazole) × 1 vs. 6/40 10/40 0.53 [0.17, 1.63] (metronidazole × 1) Cunliffe [39] 0.001 0.02 1 50 1000 (log scale) Favours cefuroxime Favours other antibiotics + metronidazole

continued 15 Results

TABLE 9 contd Cefuroxime plus metronidazole versus other antibiotics

Regimen and study Incidence of SWI Odds ratio Odds ratio [reference] (95% CI fixed) [95% CI fixed] Cefuroxime + Other metronidazole antibiotics n/N n/N

(Cefuroxime + metronidazole) × 3 vs. 0/13 3/11 0.09 [0.00, 1.97] (metronidazole × 3) Haverkorn [65]

(Cefuroxime × 6) + (metronidazole × 2) vs. 1/36 8/35 0.10 [0.01, 0.82] (metronidazole × 2) Claesson [35]

(Cefuroxime + metronidazole) × 10 vs. 6/52 13/43 0.30 [0.10, 0.88] (mezlocillin × 7) Cann [33]

(Cefuroxime + metronidazole) × 3 vs. 10/53 15/51 0.56 [0.22, 1.39] (mezlocillin × 3) Diamond [44]

(Cefuroxime + metronidazole) × 3 vs. 14/56 16/54 0.79 [0.34, 1.84] (mezlocillin × 1) Stubbs [143]

(Cefuroxime + metronidazole) × 1 vs. 27/108 32/121 0.93 [0.51, 1.68] (latamoxef × 1) Kingston [77]

(Cefuroxime + metronidazole) × 3 vs. 5/36 10/68 0.94 [0.29, 2.98] (cefotetan × 2) Skipper [140]

(Cefuroxime + metronidazole) × 3 vs. 33/454 32/453 1.03 [0.62, 1.71] (cefotaxime + metronidazole) × 1 Rowe-Jones [135]

(Cefuroxime + metronidazole) × 3 vs. 6/30 5/31 1.30 [0.35, 4.82] (imipenem + cilastatin) × 1 Pacelli [115]

(Cefuroxime + metronidazole) × 3 vs. 17/122 22/114 0.68 [0.34, 1.35] (imipenem × 4) Karran [75]

0.001 0.02 1 50 1000 (log scale) Favours cefuroxime Favours other antibiotics + metronidazole

mezlocillin alone or plus metronidazole,87,100 of SWI was not statistically significant (odds ratio = and cefotetan.19 1.41; 95% CI: 0.71, 2.82). There was no significant difference when topical use of co-amoxiclav was Bates and co-workers23 compared a single compared with parenteral use in the study by dose of co-amoxiclav with three doses of Pollock and co-workers,127 in which the patients 16 co-amoxiclav. The difference in the rate also received metronidazole. Health Technology Assessment 1998; Vol. 2: No. 7

TABLE 10 Co-amoxiclav* versus other antibiotics

Regimen and study Incidence of SWI Odds ratio Odds ratio [reference] (95% CI fixed) [95% CI fixed] Co-amoxiclav Other n/N antibiotics n/N

(Co-amoxiclav × 1) vs. (cefotetan × 1) 2/105 4/103 0.48 [0.09, 2.68] Arnaud [19]

(Co-amoxiclav × 3) vs. (cephradine + 17/84 9/36 0.76 [0.30, 1.92] metronidazole) × 3 Tehan [145]

(Co-amoxiclav × 2) vs. (gentamicin × 2) + 16/116 18/121 0.92 [0.44, 1.90] (metronidazole × 1) Hall [61]

(Co-amoxiclav × 2) vs. (ceftriaxone + 4/100 4/100 1.00 [0.24, 4.11] metronidazole) × 2 Nyam [113]

(Co-amoxiclav × 3) vs. (cefotaxime + 7/76 8/88 1.01 [0.35, 2.94] metronidazole) × 3 Kwok [82]

(Co-amoxiclav × 3) vs. (mezlocillin × 3) 4/30 4/39 1.35 [0.31, 5.89] Menzies [100]

(Co-amoxiclav × 1) vs. (mezlocillin + 6/52 9/59 0.72 [0.24, 2.19] metronidazole) × 1 Lohde [87]

(Co-amoxiclav × 3) vs. (cefuroxime + 8/69 2/79 5.05 [1.03, 24.65] metronidazole) × 3 Palmer [116]

(Co-amoxiclav × 1) vs. (co-amoxiclav × 3) 23/113 17/111 1.41 [0.71, 2.82] Bates [23]

(Co-amoxiclav + metronidazole) × 1 vs. (co-amoxiclav local injection + 1 11/47 9/40 1.05 [0.39, 2.87] metronidazole) × Pollock [127] 0.01 0.1 1 10 100 (log scale) Favours co-amoxiclav Favours other antibiotics

* Amoxycillin plus clavulanic acid

Doxycycline was administered 3–4 hours before Doxycycline is a tetracycline antibiotic with the operation. a broad spectrum of activity. Doxycycline was assessed in eight trials for the prevention of Doxycycline alone was less efficacious than SWIs after colorectal surgery (Table 11). doxycycline plus metronidazole26 or doxycycline plus tinidazole.52 Doxycycline plus metronidazole Doxycycline alone was as efficacious as appeared to be less efficacious than fosfomycin metronidazole plus netilmicin.29 Doxycycline plus metronidazole (p = 0.054).16 A large trial was as efficacious as metronidazole in one found that doxycycline added to tinidazole trial13 but was less efficacious in another reduced the SWI rate significantly compared trial78 in which the single dose of antibiotics with tinidazole alone (p = 0.017).112 17 Results

Gentamicin oral ciprofloxacin plus intravenous metronida- Gentamicin is a broad-spectrum aminoglycoside. zole,93 aztreonam plus clindamycin,109 and It is inactive against anaerobic bacteria and not ceftriaxone plus metronidazole (p < 0.05).30 absorbed from the gut. Two trials comparing gentamicin plus metro- Gentamicin alone was less efficacious in the pre- nidazole with oral neomycin plus erythromycin vention of SWI in colorectal surgery when compared on the day before surgery showed conflicting with gentamicin plus ticarcillin.42 Gentamicin was results. One trial found that the rate of SWI often combined with metronidazole, or clindamycin was higher in the group of gentamicin plus when used for the prevention of SWI in colorectal metronidazole compared with that of oral surgery (Table 12). Gentamicin plus metronidazole neomycin plus erythromycin, though the differ- was as efficacious as ampicillin plus sulbactam,15 ence was not statistically significant (p = 0.095).48 co-amoxiclav,61 latamoxef,94 cefotetan,149 and The other trial showed, however, that a single metronidazole.153 Gentamicin plus metronidazole dose of gentamicin plus metronidazole was was as efficacious as gentamicin plus lincomycin significantly more efficacious than oral neomycin or gentamicin plus clindamycin.89 plus erythromycin (p = 0.03).83 It seems that these latter results were more reliable because of the The rate of SWI in the group of gentamicin plus superior study quality in terms of blind outcome metronidazole was significantly higher than that of measurement, longer follow-up, low withdrawal

TABLE 11 Doxycycline versus other antibiotics

Regimen and study Incidence of SWI Odds ratio Odds ratio [reference] (95% CI fixed) [95% CI fixed] Doxycycline Other n/N antibiotics n/N

(Doxycycline × 2) vs. (metronidazole × 8) 16/76 21/81 0.76 [0.36, 1.60] Aberg [13]

(Doxycycline × 1) vs. (metronidazole × 1) 12/67 2/52 5.45 [1.16, 25.58] Kling [78]

(Doxycycline × 6) vs. (netilmicin + 5/50 4/50 1.28 [0.32, 5.07] metronidazole) × 3 Brolin [29]

(Doxycycline + metronidazole) × 1 vs. 16/258 7/259 2.38 [0.96, 5.89] (fosfomycin × 2) + (metronidazole × 1) Andaker [16]

(Doxycycline + tinidazole) × 1 vs. 4/132 14/135 0.27 [0.09, 0.84] (tinidazole × 1) Norwegian study [112]

(Doxycycline × 1) vs. (doxycycline + 12/126 2/135 7.00 [1.53, 31.93] metronidazole) × 1 Bergman [26]

(Doxycycline × 1) vs. (doxycycline + 10/107 3/116 3.88 [1.04, 14.51] tinidazole) × 1 Gerner [52]

(Doxycycline × 1) vs. (doxycycline × 4) 1/53 2/49 0.45 [0.04, 5.15] Goransson [54]

0.01 0.1 1 10 100 (log scale) Favours doxycycline Favours other antibiotics 18 Health Technology Assessment 1998; Vol. 2: No. 7

TABLE 12 Gentamicin versus other antibiotics

Regimen and study Incidence of SWI Odds ratio Odds ratio [reference] (95% CI fixed) [95% CI fixed] Gentamicin Other n/N antibiotics n/N

(Gentamicin + metronidazole) × 3 vs. 7/65 6/63 1.15 [0.36, 3.62] (ampicillin + sulbactam) × 3 AhChong [15]

(Gentamicin × 2) + (metronidazole × 1) vs. 18/121 16/116 1.09 [0.53, 2.26] (co-amoxiclav × 2) Hall [61]

(Gentamicin + metronidazole) × 3 vs. 7/48 2/45 3.67 [0.72, 18.71] (neomycin + erythromycin) × 3 Figueras-Felip [48]

(Gentamicin + metronidazole) × 1 vs. 5/67 14/62 0.28 [0.09, 0.82] (neomycin + erythromycin) × 3 Lau [83]

(Gentamicin + metronidazole) × 3 vs. 13/45 4/40 3.66 [1.08, 12.36] (ciprofloxacin × 1) + (metronidazole × 3) McArdle [93]

(Gentamicin + metronidazole) × 10 vs. 7/42 4/42 1.90 [0.51, 7.05] (ciprofloxacin × 7) + (metronidazole × 10) McArdle [93]

(Gentamicin + metronidazole) × 3 vs. 6/45 5/41 1.11 [0.31, 3.95] (latamoxef × 3) McCulloch [94]

(Gentamicin + metronidazole) × 1 vs. 14/61 5/59 3.22 [1.08, 9.60] (ceftriaxone + metronidazole) × 1 Burdon [30]

(Gentamicin × 15) + (metronidazole × 1) vs. 17/75 10/75 1.91 [0.81, 4.49] (cefotetan × 2) Tudor [149]

(Gentamicin + metronidazole) × 4 vs. 4/21 5/20 0.71 [0.16, 3.12] (metronidazole × 4) Weidema [153]

(Gentamicin + metronidazole) × 1 vs. 18/65 15/64 1.25 [0.57, 2.77] (gentamicin + metronidazole + cefotetan) Greig [57]

(Gentamicin + metronidazole) × 10 vs. 1/30 3/30 0.31 [0.03, 3.17] (gentamicin + lincomycin) × 10 Lozano [89]

(Gentamicin + metronidazole) × 10 vs. 1/30 4/30 0.22 [0.02, 2.14] (gentamicin + clindamycin) × 10 Lozano [89] 0.01 0.1 1 10 100 (log scale) Favours gentamicin Favours other antibiotics

continued 19 Results

TABLE 12 contd Gentamicin versus other antibiotics

Regimen and study Incidence of SWI Odds ratio Odds ratio [reference] (95% CI fixed) [95% CI fixed] Gentamicin Other n/N antibiotics n/N

(Gentamicin + metronidazole) × 2 vs. 22/209 23/219 1.00 [0.54, 1.86] (gentamicin + metronidazole) × 7 Moesgaard [103]

(Gentamicin + clindamycin) × 3 vs. 29/230 13/224 2.34 [1.18, 4.63] (aztreonam + clindamycin) × 3 Mozzillo [109]

(Gentamicin + clindamycin) × 3 vs. 12/72 8/66 1.45 [0.55, 3.80] (aztreonam + clindamycin) × 3 Rodolico [132]

(Gentamicin + clindamycin) × 3 vs. 3/35 2/37 1.64 [0.26, 10.46] (cefoxitin × 2) Mosimann [108]

(Gentamicin + clindamycin) × 2 vs. 6/30 7/30 0.82 [0.24, 2.81] (cefoxitin × 3) Cainzos [32]

(Gentamicin × 7) vs. (gentamicin + 14/26 3/29 10.11 [2.44, 41.93] ticarcillin) × 7 Desaive [42]

0.01 0.1 1 10 100 (log scale) Favours gentamicin Favours other antibiotics

rate, and appropriate definition of SWI (see Table 14 shows the results from trials that investi- details in appendix 3).48,83 gated the effect of adding metronidazole to other antibiotics. Three trials found significant benefit Metronidazole of additional metronidazole to neomycin,67 doxy- Metronidazole is active against anaerobic cycline,26 and cefotaxime.59 The addition of metro- bacteria. It seems that metronidazole alone nidazole to cefuroxime was beneficial but not statis- was inadequate in the prevention of SWI in tically significant in a small trial (p = 0.24).38 No colorectal surgery when compared with other significant benefit was found when metronidazole antibiotics (Table 13). Metronidazole alone was was used in addition to cefotetan,107 ceftriaxone,91 significantly less efficacious than metronidazole latamoxef,105 and ceftizoxime.66 plus ampicillin,56 metronidazole plus doxycycline (p = 0.056),133 metronidazole plus cefuroxime or Mezlocillin netilmicin,35,63 metronidazole plus fosfomycin,86 Mezlocillin is an acylureidopenicillin with in vitro cephazolin plus neomycin and erythromycin.76 activity against a wide range of Gram-positive and Six other trials did not find a significant differ- Gram-negative organisms, including anaerobes. ence between metronidazole alone and other antibiotics such as doxycycline alone,13 oral neo- The efficacy of mezlocillin in preventing SWI mycin plus erythromycin,21 metronidazole plus was investigated in 11 trials (Table 15). Three trials nalidixic acid,79 or gentamicin plus metronida- compared mezlocillin alone with cefuroxime plus zole.153 Kling and co-workers78 found that a single metronidazole33,44,143 and one found a significant dose of metronidazole was significantly more difference in favour of cefuroxime plus metronida- efficacious than a single dose of doxycycline zole.33 The rate of SWI in the mezlocillin group 20 (p = 0.02). was not significantly different from that with Health Technology Assessment 1998; Vol. 2: No. 7

TABLE 13 Metronidazole alone versus other antibiotics

Regimen and study Incidence of SWI Odds ratio Odds ratio [reference] (95% CI fixed) [95% CI fixed] Metronidazole Other n/N antibiotics n/N

(Metronidazole × 3) vs. (neomycin + 8/78 4/55 1.46 [0.42, 5.10] erythromycin) × 3 Auger [21]

(Metronidazole × 3) vs. (cephazolin + 14/47 4/55 5.41 [1.64, 17.86] neomycin + erythromycin) × 3 Khubchadani [76]

(Metronidazole × 10) vs. (ampicillin + 10/46 1/48 13.06 [1.60, 106.72] metronidazole) × 10 Gottrup [56]

(Metronidazole × 3) vs. (netilmicin or 3/11 0/25 21.00 [0.98, 449.33] cefuroxime + metronidazole) × 3 Haverkorn [63]

(Metronidazole × 1) vs. (nalidixic acid + 23/70 13/33 0.75 [0.32, 1.78] metronidazole) × 1 Kling [79]

(Metronidazole × 4) vs. (fosfomycin + 6/23 0/26 19.69 [1.04, 372.06] metronidazole) × 4 Lindhagen [86]

(Metronidazole × 1) vs. (ampicillin or 6/170 1/188 6.84 [0.82, 57.42] doxycycline + metronidazole) × 1 Roland [133]

(Metronidazole × 4) vs. (gentamicin + 5/20 4/21 1.42 [0.32, 6.27] metronidazole) × 4 Weidema [153]

(Metronidazole × 2) vs. (cefuroxime × 6) 8/35 1/36 10.37 [1.22,88.02] + (metronidazole × 2) Claesson [35]

(Metronidazole × 1) vs. (cefuroxime + 7/33 3/27 2.15 [0.50, 9.29] metronidazole) × 1 Mittermayer [101]

(Metronidazole × 1) vs. (cefuroxime + 10/40 6/40 1.89 [0.61, 5.82] metronidazole) × 1 Cunliffe [39]

(Metronidazole × 8) vs. (doxycycline × 3) 21/81 16/76 1.31 [0.62, 2.76] Aberg [13]

(Metronidazole × 1) vs. (doxycycline × 1) 2/52 12/67 0.18 [0.04, 0.86] Kling [78]

0.01 0.1 1 10 100 (log scale) Favours metronidazole Favours other antibiotics 21 Results

ticarcillin/clavulanic acid,97 cefotaxime,98 Latamoxef alone was as efficacious as other ceftizoxime,98 latamoxef,98 and co-amoxiclav.100 antibiotic regimens in the prevention of SWI in colorectal surgery (Table 16). The antibiotic Latamoxef regimens that were compared with latamoxef Latamoxef (Moxalactam®, Shinogi, Japan), is a alone included ciprofloxacin plus metronidazole55 cephalosporin-like beta-lactam antibiotic with a cefuroxime plus metronidazole,77 gentamicin plus broad spectrum of activity against gastrointestinal metronidazole,94 mezlocillin plus metronidazole,98 pathogens including strains of Enterobacteriaceae and co-amoxiclav,123 cephazolin,125 and cephazolin plus B. fragilis that are resistant to older cephalosporins metronidazole.146 Adding metronidazole to lata- such as cephazolin. moxef was not found to be beneficial compared

TABLE 14 Metronidazole as an additional antibiotic

Regimen and study Incidence of SWI Odds ratio Odds ratio [reference] (95% CI fixed) [95% CI fixed] Additional Other metronidazole antibiotics n/N n/N

(Doxycycline + metronidazole) × 1 vs. 2/135 12/126 0.14 [0.03, 0.65] (doxycycline × 1) Bergman [26]

(Cefuroxime + metronidazole) × 1 vs. 0/30 3/32 0.14 [0.01, 2.79] cefuroxime × 1 Corman [38]

(Cefotaxime + metronidazole) × 1 vs. 19/287 44/280 0.38 [0.22, 0.67] (cefotaxime × 3) Hakansson [59]

(Ceftizoxime × 2) + (metronidazole × 1) vs. 7/89 9/85 0.72 [0.26, 2.03] (ceftizoxime × 2) Hosie [66]

(Latamoxef × 2) + (metronidazole × 1) vs. 13/56 11/53 1.15 [0.47, 2.86] (latamoxef × 2) Morris [105]

(Cefotetan × 2) + (metronidazole × 3) vs. 34/253 39/278 0.95 [0.58, 1.56] (cefotetan × 2) Morton [107]

(Neomycin × 3) + (metronidazole × 3) vs. 0/31 8/37 0.06 [0.00, 1.00] (neomycin × 3) Jagelman [67]

(Ceftriaxone × 1) + (metronidazole 7/90 8/94 0.91 [0.31, 2.61] suppository) vs. (ceftriaxone × 1) Lumley [91]

(Cefotaxime × 4) + (metronidazole or 5/72 4/74 1.31 [0.34, 5.07] ornidazole × 2) vs. (cefotaxime × 4) Favre [47]

0.01 0.1 1 10 100 (log scale) Favours with metronidazole Favours without metronidazole

22 Health Technology Assessment 1998; Vol. 2: No. 7

TABLE 15 Mezlocillin versus other antibiotics

Regimen and study Incidence of SWI Odds ratio Odds ratio [reference] (95% CI fixed) [95% CI fixed] Mezlocillin Other n/N antibiotics n/N

(Mezlocillin × 7) vs. (cefuroxime + 13/43 6/52 3.32 [1.14, 9.70] metronidazole) × 10 Cann [33]

(Mezlocillin × 3) vs. (cefuroxime + 15/51 10/53 1.79 [0.72, 4.47] metronidazole) × 3 Diamond [44]

(Mezlocillin × 1) vs. (cefuroxime + 16/54 14/56 1.26 [0.54, 2.93] metronidazole) × 3 Stubbs [143]

(Mezlocillin × 1) vs. (ticarcillin/clavulanic 9/93 9/85 0.90 [0.34, 2.40] acid × 1) Melbourne study [97]

(Mezlocillin × 1) vs. (cefotaxime × 1) 2/48 3/52 0.71 [0.11, 4.44] Mendel [98]

(Mezlocillin × 1) vs. (ceftizoxime × 1) 1/53 2/47 0.43 [0.04, 4.93] Mendel [98]

(Mezlocillin + metronidazole) × 1 vs. 6/46 3/44 2.05 [0.48, 8.76] (mezlocillin + metronidazole) × 7 Bittner [27]

(Mezlocillin × 3) vs. (co-amoxiclav × 3) 4/39 4/30 0.74 [0.17, 3.25] Menzies [100]

(Mezlocillin + metronidazole) × 1 vs. 9/59 6/52 1.38 [0.46, 4.18] (co-amoxiclav × 1) Lohde [87]

(Mezlocillin + metronidazole) × 1 vs. 2/54 1/46 1.73 [0.15, 19.73] (mezlocillin + metronidazole) × 9 Mendel [98]

(Mezlocillin + metronidazole) × 1 vs. 4/77 4/77 1.00 [0.24, 4.15] (mezlocillin + metronidazole) × 3 Grundmann [58]

(Mezlocillin + metronidazole) × 9 vs. 2/57 2/63 1.11 [0.15, 8.14] (latamoxef × 1) Mendel [98]

(Mezlocillin + oxacillin) × 3 vs. 2/29 1/30 2.15 [0.18, 25.07] (neomycin + bacitracin + clindamycin) Schiessel [137]

0.01 0.1 1 10 100 (log scale) Favours mezlocillin Favours other antibiotics

23 Results

with latamoxef alone.105 In a trial that included It cannot be absorbed from the gut. Erythromycin more than 100 patients in each treatment group, has a similar but not identical antibacterial spec- it was found that the single dose of latamoxef was trum to that of penicillin, and it is active against as efficacious as multiple doses of latamoxef.60 anaerobic organisms. Neomycin plus erythromycin Latamoxef plus metronidazole suppository was is often used orally (1 g of each) on the day before more efficacious than oral tetracycline plus surgery (1 p.m., 2 p.m., and 11 p.m.). This regimen metronidazole suppository.136 was compared with other antibiotic regimens in nine trials (Table 17). Neomycin plus erythromycin Neomycin is an aminoglycoside, active against some The overall rate of SWI was significantly higher in Gram-positive and many Gram-negative organisms. the oral neomycin plus erythromycin group than

TABLE 16 Latamoxef versus other antibiotics

Regimen and study Incidence of SWI Odds ratio Odds ratio [reference] (95% CI fixed) [95% CI fixed] Latamoxef Other n/N antibiotics n/N

(Latamoxef × 1) vs. (ciprofloxacin + 1/57 4/54 0.22 [0.02, 2.06] metronidazole) × 1 Gortz [55]

(Latamoxef × 3) vs. (neomycin + 5/64 3/67 1.81 [0.41, 7.09] metronidazole) × 3 Hinchey [65]

(Latamoxef × 1) vs. (cefuroxime + 32/121 27/108 1.08 [0.60, 1.95] metronidazole) × 1 Kingston [77]

(Latamoxef × 3) vs. (gentamicin + 5/41 6/45 0.90 [0.25, 3.22] metronidazole) × 3 McCulloch [94]

(Latamoxef × 1) vs. (mezlocillin + 2/63 2/57 0.90 [0.12, 6.62] metronidazole) × 9 Mendel [98]

[All: (Neomycin + metronidazole)] 10/36 15/56 1.05 [0.41, 2.69] (Latamoxef × 1) vs. (co-amoxiclav × 1) Playforth [123]

(Latamoxef × 3) vs. (cephazolin × 3) 1/26 1/24 0.92 [0.05, 15.58] Plouffe [125]

(Latamoxef × 3) vs. (cephazolin + 3/60 5/60 0.58 [0.13, 2.54] metronidazole) × 3 Thomas [146]

[All: Metronidazole] (Latamoxef × 1) 6/39 15/43 0.34 [0.12,0.99] vs. tetracycline Sauven [136]

(Latamoxef × 1) vs. (latamoxef × 8) 12/119 10/126 1.30 [0.54, 3.13] Hall [60]

(Latamoxef × 2) vs. (latamoxef × 2) 11/53 13/56 0.87 [0.35,2.15] + metronidazole Morris [105] 0.01 0.1 1 10 100 (log scale) Favours latamoxef Favours other antibiotics 24 Health Technology Assessment 1998; Vol. 2: No. 7

that in the group of neomycin plus erythromycin including Proteus spp. and B. fragilis. Clavulanic with additional antibiotics (overall odds ratio = acid inactivates penicillinase produced by 3.34; 95% CI: 1.66, 6.72).83,121,138,141 Weaver and bacteria resistant to ticarcillin. co-workers152 found that oral neomycin plus erythromycin was significantly less efficacious than Ticarcillin/clavulanic acid (Timentin) was found to ceftriaxone plus metronidazole. In another trial, be as efficacious as metronidazole plus netilmicin,28 no abdominal wound infection was observed in cefotaxime,72 and mezlocillin97 (Table 20). Ticarcillin/ patients receiving oral neomycin plus erythromycin clavulanic acid was more efficacious than oral tinida- or ceftriaxone plus metronidazole.80 zole.96 A single dose of ticarcillin/clavulanic acid was found to be as efficacious as two doses.40 The difference was not statistically significant when neomycin plus erythromycin was compared with Tinidazole metronidazole,21 cefoxitin alone,141 and erythromycin Tinidazole is similar to metronidazole in activity plus metronidazole.85 Coppa and Eng37 found that but has a longer half-life and can therefore be the rate of SWI was lower in patients receiving par- given less frequently. enteral cefoxitin plus oral neomycin plus erythro- mycin compared with cefoxitin alone, but the Tinidazole alone was significantly less efficacious difference was not statistically significant (p > 0.1). than tinidazole plus doxycycline,112 and less effi- cacious than ticarcillin95 and ticarcillin/clavulanic Two trials compared oral neomycin plus erythromy- acid96 (Table 21). Tinidazole plus doxycycline was cin with gentamicin plus metronidazole.48,83 One significantly better than doxycycline alone.52 Tin- found that oral neomycin plus erythromycin was idazole plus ceftriaxone was significantly more significantly less efficacious than gentamicin plus efficacious than tinidazole plus netilmicin.92 metronidazole (p = 0.016),83 whereas the other found a non-significant benefit of neomycin plus erythromycin (p = 0.10).48 The former results are Other antibiotics likely to be more reliable, however, because of the superior quality of the study. Some agents were assessed in fewer than four trials and almost all of these agents have already been Netilmicin mentioned when they were compared with other Netilmicin is an aminoglycoside. It is used in agents that were more frequently tested. serious Gram-negative infections that are resistant to gentamicin. A single dose of cefamandole plus metronidazole was as efficacious as a single dose of ceftriaxone Matikainen and Hiltunen92 found that netilmicin plus metronidazole.62 When parenteral cefaman- plus tinidazole was significantly less efficacious than dole was added to oral neomycin plus erythro- ceftriaxone plus tinidazole (Table 18). There was no mycin, no significant benefit was observed in significant difference in the SWI rates when a small trial.121 netilmicin plus metronidazole was compared with ticarcillin/clavulanic acid,28 doxycycline,29 cefuro- When all patients received oral neomycin plus xime plus metronidazole,63 and piperacillin.151 erythromycin, additional parenteral cefmetazole was associated with a lower rate of SWI compared Piperacillin with additional cefoxitin but the difference was It has been suggested that piperacillin has a broad not statistically significant.45,126 antibacterial spectrum covering both aerobic and anaerobic bacteria.43 However, results of included Cefonicid in addition to oral neomycin plus erythro- trials (Table 19) showed that piperacillin alone was mycin was as efficacious as cefoxitin used in combi- significantly less efficacious than piperacillin plus nation with oral neomycin plus erythromycin.46 metronidazole, cefuroxime plus metronidazole,131 Cefoperazone was as efficacious as cefotaxime.70 and piperacillin plus oral ciprofloxacin.144 There Multiple doses of cefotiam plus gentamicin was of was no significant difference when piperacillin similar efficacy as a single dose of ceftriaxone.155 was compared with cefotetan,64 piperacillin plus The difference between cephalothin and cefoxitin sulbactam,142 or netilmicin plus metronidazole.151 was not statistically significant.17 Cephradine plus metronidazole was as effective as co-amoxiclav.145 Ticarcillin/clavulanic acid Ticarcillin is active against Pseudomonas Three trials tested ciprofloxacin plus metronida- aeruginosa, certain other Gram-negative bacilli zole. Ciprofloxacin plus metronidazole was as 25 Results

TABLE 17 Neomycin plus erythromycin

Regimen and study Incidence of SWI Odds ratio Odds ratio [reference] (95% CI fixed) [95% CI fixed] Neomycin + Other erythromycin antibiotics n/N n/N

(Neomycin + erythromycin) × 3 vs. 4/55 8/78 0.69 [0.20, 2.40] (metronidazole × 3) Auger [21]

(Neomycin + erythromycin) × 3 vs. 2/45 7/48 0.27 [0.05, 1.39] (gentamicin + metronidazole) × 3 Figueras-Felip [48]

(Neomycin + erythromycin) × 3 vs. 14/62 5/67 3.62 [1.22, 10.74] (gentamicin + metronidazole) × 1 Lau [83]

(Neomycin + erythromycin) × 3 vs. 3/44 2/51 1.79 [0.29, 11.25] (cefoxitin × 3) Stellato [141]

(Neomycin + erythromycin) × 3 vs. 12/29 3/31 6.59 [1.62, 26.75] (ceftriaxone + metronidazole) × 1 Weaver [152]

(Neomycin + erythromycin) × 3 vs. 1/27 2/27 0.48 [0.04, 5.64] (ceftriaxone + metronidazole) × 1 Kling [80]

(Neomycin + erythromycin) × 3 vs. 7/61 2/64 4.02 [0.80, 20.17] (erythromycin × 3) + (metronidazole × 6) Lewis [85]

(Neomycin + erythromycin) × 3 + 4/55 14/47 0.18 [0.06, 0.61] (cephazolin × 3) vs. (metronidazole × 3) Khubchandani [76]

(Neomycin + erythromycin) × 3 + 3/65 5/67 0.60 [0.14, 2.62] (gentamicin + metronidazole) × 1 vs. (gentamicin + metronidazole) × 1 Lau [83]

(Neomycin + erythromycin) × 3 + 9/169 15/141 0.47 [0.20, 1.12] (cefoxitin × 5) vs. (cefoxitin × 6) Coppa [37]

(Neomycin + erythromycin) × 3 + 6/50 2/50 3.27 [0.63, 17.07] (ampicillin × 1) vs. (cefotaxime × 3) Raahave [128]

(Neomycin + erythromycin) × 3 + 3/51 2/51 1.53 [0.24, 9.57] (cefoxitin × 3) vs. (cefoxitin × 3) Stellato [141]

0.01 0.1 1 10 100 (log scale) Favours neomycin + Favours other antibiotics erythromycin

26 continued Health Technology Assessment 1998; Vol. 2: No. 7

TABLE 17 contd Neomycin plus erythromycin

Regimen and study Incidence of SWI Odds ratio Odds ratio [reference] (95% CI fixed) [95% CI fixed] Neomycin + Other erythromycin antibiotics n/N n/N

(Neomycin + erythromycin) × 3 vs. 14/62 3/65 6.03 [1.64, 22.18] (neomycin + erythromycin) × 3 + (gentamicin + metronidazole) × 1 Lau [83]

(Neomycin + erythromycin) × 3 vs. (neomycin + 1/36 0/34 2.92 [0.11, 74.06] erythromycin) × 3 + (cefamandole × 4) Petrelli [121]

(Neomycin + erythromycin) × 3 vs. 14/96 5/101 3.28 [1.13, 9.49] (neomycin + erythromycin) × 3 + (cefoxitin × 3) Schoetz [138]

(Neomycin + erythromycin) × 3 vs. 3/44 3/51 1.17 [0.22, 6.12] (neomycin + erythromycin) × 3 + (cefoxitin × 3) Stellato [141]

0.01 0.1 1 10 100 (log scale) Favours neomycin + Favours other antibiotics erythromycin

TABLE 18 Netilmicin versus other antibiotics

Regimen and study Incidence of SWI Odds ratio Odds ratio [reference] (95% CI fixed) [95% CI fixed] Netilmicin Other n/N antibiotics n/N

(Netilmicin + metronidazole) × 3 vs. 5/46 3/46 1.75 [0.39, 7.79] (ticarcillin/clavulanic acid × 1) Blair [28]

(Netilmicin + metronidazole) × 3 vs. 4/50 5/50 0.78 [0.20, 3.10] (doxycycline × 6) Brolin [29]

(Netilmicin + metronidazole) × 3 vs. 0/12 0/13 Not estimable (cefuroxime + metronidazole) × 3 Haverkorn [63]

(Netilmicin × 3) + (metronidazole × 2) vs. 13/105 17/108 0.76 [0.35, 1.65] (piperacillin × 3) Walker [151]

(Netilmicin or tobramycin + tinidazole) × 38/313 8/315 5.30 [2.43, 11.56] 1 vs. (ceftriaxone + tinidazole) × 1 Matikainen [92]

0.01 0.1 1 10 100 (log scale) Favours netilmicin Favours other antibiotics 27 Results

TABLE 19 Piperacillin versus other antibiotics

Regimen and study Incidence of SWI Odds ratio Odds ratio [reference] (95% CI fixed) [95% CI fixed] Piperacillin Other n/N antibiotics n/N

(Piperacillin × 3) vs. (cefotetan × 1) 13/78 14/75 0.87 [0.38, 2.00] Hershman [64]

(Piperacillin × 4) vs. (cefuroxime × 3) + 18/104 8/119 2.90 [1.21, 7.00] (metronidazole × 4) Reynolds [131]

(Piperacillin × 3) vs. (netilmicin × 3) + 17/108 13/105 1.32 [0.61, 2.88] (metronidazole × 2) Walker [151]

(Piperacillin × 4) vs. (piperacillin × 4) + 18/104 9/107 2.28 [0.97, 5.34] (metronidazole × 6) + (neomycin × 8) Reynolds [131]

(Piperacillin × 1) vs. (piperacillin + 39/168 32/158 1.19 [0.70, 2.02] sulbactam) × 1 Stewart [142]

(Piperacillin × 1) vs. (piperacillin × 1) + 39/168 18/159 2.37 [1.29, 4.35] (ciprofloxacin × 2) Taylor [144]

(Piperacillin × 1) vs. (cefoxitin × 1) 2/30 5/30 0.36 [0.06, 2.01] Armengaud [18]

(Piperacillin × 1) vs. (piperacillin × 4) 3/25 2/25 1.57 [0.24, 10.30] Devecioglu [43]

0.01 0.1 1 10 100 (log scale) Favours piperacillin Favours other antibiotics

TABLE 20 Ticarcillin/clavulanic acid

Regimen and study Incidence of SWI Odds ratio Odds ratio [reference] (95% CI fixed) [95% CI fixed] Ticarcillin/ Other clavulanic acid antibiotics n/N n/N

(Ticarcillin/clavulanic acid × 1) vs. 3/46 5/46 0.57 [0.13, 2.55] (netilmicin + metronidazole) × 3 Blair [28]

(Ticarcillin/clavulanic acid × 1) vs. (cefotaxime × 4) 0/11 1/8 0.22 [0.01, 6.08] Jones [72]

(Ticarcillin/clavulanic acid × 1) vs. (mezlocillin × 1) 9/85 9/93 1.11 [0.42, 2.93] Melbourne study [97]

(Ticarcillin/clavulanic acid × 2) vs. (tinidazole × 1) 2/83 12/84 0.15 [0.03, 0.68] Melbourne study [96] 0.01 0.1 1 10 100 (log scale) Favours ticarcillin/clavulanic acid Favours other antibiotics 28 Health Technology Assessment 1998; Vol. 2: No. 7

TABLE 21 Tinidazole versus other antibiotics

Regimen and study Incidence of SWI Odds ratio Odds ratio [reference] (95% CI fixed) [95% CI fixed] Tinidazole Other n/N antibiotics n/N

(Tinidazole × 1) vs. (ticarcillin × 2) 24/121 10/125 2.85 [1.30, 6.24] Melbourne study [95]

(Tinidazole × 1) vs. (ticarcillin/ 12/84 2/83 6.75 [1.46, 31.18] clavulanic acid × 2) Melbourne study [96]

(Tinidazole × 1) vs. (metronidazole × 10) 0/50 2/50 0.19 [0.01, 4.10] Athanasiadis [20]

(Tinidazole × 1) vs. (tinidazole + 14/135 4/132 3.70 [1.19, 11.56] doxycycline) × 1 Norwegian study [112]

(Tinidazole + ceftriaxone) × 1 vs. 8/315 38/313 0.19 [0.09, 0.41] (tinidazole + netilmicin or tobramycin) × 1 Matikainen [92]

(Tinidazole + doxycycline) × 1 vs. 3/116 10/107 0.26 [0.07, 0.96] (doxycycline × 1) Gerner [52] 0.01 0.1 1 10 100 (log scale) Favours tinidazole Favours other antibiotics

efficacious as latamoxef,55 cephazolin plus metro- antibiotic agents (excluding trials that have nidazole,114 and gentamicin plus metronidazole.93 been included in Tables 13 and 14). It appears that piperacillin alone was less efficacious than The difference was not significant when fosfomycin piperacillin plus oral ciprofloxacin,144 or plus plus metronidazole was compared with doxycycline oral neomycin and metronidazole.131 Doxycycline plus metronidazole16 or ampicillin plus metronida- alone was less effective than doxycycline plus zole plus bacitracin plus neomycin.111 Fosfomycin metronidazole,26 or tinidazole.52 Tinidazole plus metronidazole was significantly more effi- alone was less efficacious than tinidazole cacious than metronidazole alone (p = 0.007).86 plus doxycycline.112

Imipenem plus cilastatin was as efficacious as Oral neomycin plus erythromycin was cefuroxime plus metronidazole.115 Imipenem alone similarly efficacious with or without the addi- was associated with a higher rate of SWI compared tion of cefoxitin.141 Cefotaxime alone was as with cefuroxime plus metronidazole, though the efficacious as cefotaxime plus topical ampi- difference was not statistically significant (p = 0.27).75 cillin.129 The addition of metronidazole to cefotaxime was found to be beneficial59 No statistically significant difference was found (Table 14). Cefoxitin alone was as effective when penicillin plus streptomycin or penicillin plus as cefoxitin plus oral neomycin and erythro- metronidazole and tobramycin were compared with mycin.37,141 However, in a small trial cefoxitin cefotaxime84 or cefoxitin.119 alone was more effective than cefoxitin plus oral tinidazole, though the difference was not significant (p = 0.12).120 The outcome was not Additional antibiotics significantly improved when additional anti- biotics were added to cefotetan,107 ceftriaxone,91 Table 22 shows the results of trials that compared ceftizoxime,66 gentamicin plus metronidazole,57 a regimen with the same regimen plus other and latamoxef.105 29 Results

TABLE 22 Additional antibiotics*

Regimen and study Incidence of SWI Odds ratio Odds ratio [reference] (95% CI fixed) [95% CI fixed] Additional No additional antibiotics antibiotics n/N n/N

(Cefoxitin × 3) vs. (cefoxitin × 3) + 4/41 0/39 0.11 [0.01, 2.03] (tinidazole × 1) + (neomycin × 3) Peruzzo [120] (Neomycin + erythromycin) × 3 vs. 3/51 3/44 1.17 [0.22, 6.12] (neomycin + erythromycin + cefotaxime) × 3 Stellato [141] (Cefoxitin × 3) vs. (cefoxitin + neomycin + 3/51 2/51 0.65 [0.10, 4.08] erythromycin) × 3 Stellato [141] (Cefotaxime × 3) vs. (cefotaxime × 3) + (ampicillin × 1) 5/81 6/89 1.10 [0.32, 3.75] Raahave [129] (Piperacillin × 1) vs. (piperacillin + sulbactam) × 1 32/158 39/168 1.19 [0.70, 2.02] Stewart [142] (Gentamicin + metronidazole) × 1 vs. 15/64 18/65 1.25 [0.57, 2.77] (gentamicin + metronidazole + cefotetan) Greig [57] (Cefotetan × 1) vs. (cefotetan × 1) + (thymostimulin × 7) 56/425 72/434 1.31 [0.90, 1.91] Periti [118] (Cefoxitin × 5) vs. (cefoxitin × 5) + (neomycin + 9/169 15/141 2.12 [0.90, 5.00] erythromycin) × 3 Coppa [37] (Piperacillin × 4) vs. (piperacillin × 4) + 9/107 18/104 2.28 [0.97, 5.34] (metronidazole × 6) + (neomycin × 8) Reynolds [131] (Piperacillin × 1) vs. (piperacillin × 1) + 18/159 39/168 2.37 [1.29, 4.35] (ciprofloxacin × 2) Taylor [144] (Neomycin + erythromycin) × 3 vs. (metronidazole 3/65 14/62 6.03 [1.64, 22.18] + gentamicin) × 1 + (neomycin + erythromycin) × 3 Lau [83] (Metronidazole + gentamicin) × 1 vs. (metronidazole 3/65 5/67 1.67 [0.38, 7.28] + gentamicin) × 1 + (neomycin + erythromycin) × 3 Lau [83] (Neomycin + erythromycin) × 3 vs. 5/101 14/96 3.28 [1.13, 9.49] (neomycin + erythromycin + cefoxitin) × 3 Schoetz [138] (Tinidazole × 1) vs. (tinidazole + doxycycline) × 1 4/132 14/135 3.70 [1.19, 11.56] Norwegian study [112] (Doxycycline × 1) vs. (doxycycline + tinidazole) × 1 3/116 10/107 3.88 [1.04, 14.51] Gerner [52] (Neomycin + erythromycin) × 3 vs. (neomycin + 0/34 1/36 2.92 [0.11, 74.06] erythromycin) × 3 + (cefamandole × 4) Petrelli [121] 0.001 0.02 1 50 1000 (log scale) Favours no additional Favours additional

* 30 Excluding metronidazole Health Technology Assessment 1998; Vol. 2: No. 7

Single-dose versus multiple-dose The rate of total mortality ranged from 0% to regimens 16.4% according to the results reported in 42 trials. This gives an average death rate of 4.1%. However, Seventeen trials compared a single-dose regimen this is likely to be an overestimation of the average with a multiple-dose regimen (two or more doses) death rate for all patients undergoing colorectal using the same antibiotic or the same combination surgery because of reporting bias (i.e. trials that of antibiotics (Table 23). None of these trials found involve deaths are more likely to report mortality a significant difference in postoperative SWI data than those in which no deaths occur). between single-dose and multiple-dose regimens. The antibiotics that were investigated in these trials With the exception of two trials,91,156 no significant were cefoxitin,38,81 cefuroxime plus metronida- difference was found in the number of deaths zole,14 metronidazole,34 ampicillin plus metronida- between different antibiotic groups. Both elective zole,69,74 co-amoxiclav,23 latamoxef,60 mezlocillin and emergency colorectal surgery were included in plus metronidazole,27,58,98 ticarcillin/clavulanic the trial by Lumley and co-workers.91 There were acid,40 cefotaxime plus metronidazole,81 gentamicin five deaths in the cephazolin group and no deaths plus ornidazole,154 piperacillin plus tinidazole,148 in the two ceftriaxone groups. However, more doxycycline,54 and cefotaxime.88 When the results patients underwent emergency operations in the from these 17 trials were pooled, there was no cephazolin group (12% versus 5%) and there were significant difference between the single- and also more surgical wounds that were classified as multiple-dose groups (10.6% versus 9.7%; contaminated or dirty (31% versus 21%) in this odds ratio = 1.17; 95% CI: 0.89, 1.54). group. In addition, all five deaths were infection related but not because of wound infection. Zuber and co-workers156 recorded five deaths in the Timing and duration of treatment cephazolin group (n = 50) and no deaths in the cephazolin plus ornidazole group (n = 50). (This The duration of operation and the half-life of an paper was in German and details related to cause antibiotic may be related to the effectiveness of a of death were not explored.) single dose or short-term use of antibiotic prophy- laxis. It was claimed that an extended duration of Intraabdominal abscess operation is associated with a higher rate of SWI.157 Results of intraabdominal abscesses were available Several trials that compared a single dose with from 46 trials. The average rate of intraabdominal multiple doses reported the duration of operation. abscess in these 46 trials was about 3% and no The average duration of operation was 92 minutes statistically significant difference was observed in the trial by Carr and co-workers,34 135 minutes between the different antibiotic groups. in Cuthbertson’s trial40 and 116 minutes (range: 20–270) in Wenzel’s trial.154 In the trial by Jensen Respiratory and urinary tract infection and co-workers, the proportion of patients whose The average rate of reported respiratory operation lasted longer than 180 minutes was infection after colorectal surgery was 5.4%, 29%.69 It was also found that the rate of SWI was according to 37 trials. Among these, several similar in patients whose operations lasted less found significant differences between the than 3 hours and in those whose operations were antibiotic groups.16,55,91,93,106,118 longer than 3 hours’ duration.69 Thirty-nine trials reported the results of urinary tract infection after colorectal surgery. The average Other outcomes rate was 9.9% and seven of these trials found statistically significant differences.27,78,80,91,96,98,115 The results of deep wound infections, remote infections (such as urinary tract infection, respir- Adverse events atory tract infection) and mortality, if reported, are Among 134 trials that were published in English, also presented in appendix 3. However, these results 74 (55%) measured and reported results of adverse have not been reported consistently across trials and events after antibiotic prophylaxis in colorectal therefore their reliability may suffer from the surgery. This proportion was 58%, 56% and 45% potential reporting and detection bias. during 1984–87, 1988–91 and 1992–95, respec- tively. Patients with a history of drug allergy were Mortality not included in the trials. Skin rash, diarrhoea, and The rate of total mortality ranged from 0% to nausea were commonly reported adverse events that 16.4% according to the results reported in 42 trials. may be attributable to the antibiotic agents used. 31 Results

TABLE 23 Single versus multiple doses of antibiotic

Regimen and study Incidence of SWI Odds ratio Odds ratio [reference] (95% CI fixed) [95% CI fixed] Single Other n/N antibiotics n/N

(Cefoxitin × 1) vs. (cefoxitin × 3) 10/73 8/81 1.45 [0.54, 3.89] Kow [81] (Cefoxitin × 1) vs. (cefoxitin × 4) 2/31 0/27 4.66 [0.21, 101.47] Corman [38] (Cefotaxime × 1) vs. (cefotaxime × 3) 4/30 3/30 1.38 [0.28, 6.80] Lohr [88] (Co-amoxiclav × 1) vs. (co-amoxiclav × 3) 23/113 17/111 1.41 [0.71, 2.82] Bates [23] (Ticarcillin/clavulanic acid × 1) vs. 16/146 17/132 0.83 [0.40, 1.72] (ticarcillin/clavulanic acid × 2) Cuthbertson [40] (Latamoxef × 1) vs. (latamoxef × 8) 12/119 10/126 1.30 [0.54, 3.13] Hall [60] (Ampicillin + metronidazole) × 1 vs. 6/11 4/12 2.40 [0.44, 12.98] (ampicillin + metronidazole) × 3 Jensen [69] (Ampicillin + metronidazole) × 1 vs. 8/89 8/92 1.04 [0.37, 2.89] (ampicillin + metronidazole) × 3 Jensen [69] (Ampicillin + metronidazole) × 1 vs. 9/149 8/145 1.10 [0.41, 2.94] (ampicillin + metronidazole) × 4 Juul [74] (Cefotaxime + metronidazole) × 1 vs. 7/84 9/81 0.73 [0.26, 2.05] (cefotaxime + metronidazole) × 3 Kow [81] (Mezlocillin + metronidazole) × 1 vs. 4/77 4/77 1.00 [0.24, 4.15] (mezlocillin + metronidazole) × 3 Grundmann [58] (Mezlocillin + metronidazole) × 1 vs. 6/46 3/44 2.05 [0.48, 8.76] (mezlocillin + metronidazole) × 7 Bittner [27] (Mezlocillin + metronidazole) × 1 vs. 2/54 1/46 1.73 [0.15, 19.73] (mezlocillin + metronidazole) × 9 Mendel [98] (Cefuroxime + metronidazole) × 1 vs. 2/19 1/29 3.29 [0.28, 39.14] (cefuroxime + metronidazole) × 3 Aberg [14] (Metronidazole × 1) vs. (metronidazole × 2–4) 7/22 11/68 2.42 [0.80, 7.30] Carr [34] (Doxycycline × 1) vs. (doxycycline × 4) 1/53 2/49 0.45 [0.04, 5.15] Goransson [54] (Gentamicin + ornidazole) × 1 vs. 6/30 10/30 0.50 [0.15, 1.62] (gentamicin + ornidazole) × 4 Wenzel [154] (Piperacillin + tinidazole) × 1 vs. 4/61 5/63 0.81 [0.21, 3.19] (piperacillin × 4) + (tinidazole × 3) Tuchmann [148] 0.01 0.1 1 10 100 (log scale) Favours single Favours multiple 32 Health Technology Assessment 1998; Vol. 2: No. 7

No serious toxicity or adverse events were reported Monotherapy except in the trial by Morris and co-workers.105 Three trials compared the efficacy and cost Morris observed postoperative bleeding in 12 of the of monotherapy with combination therapy. One 97 patients treated with latamoxef. The trial was trial concluded that the monotherapy regimen discontinued, and the investigators concluded that (ampicillin/sulbactam, HK $10.11 [£0.85]) was even short-term treatment with latamoxef could not cheaper than the combination of cephalosporins be advised.105 These results were not confirmed, and metronidazole in Hong Kong.15 Another found however, by other trials that compared latamoxef that the efficacy of co-amoxiclav was comparable with tetracycline,136 cephazolin,125 cephazolin plus with cefotaxime plus metronidazole, while mono- metronidazole,146 co-amoxiclav,123 gentamicin plus therapy of co-amoxiclav (three doses) was cheaper, metronidazole,94 cefuroxime plus metronidazole,77 simpler and easier to administer than cefotaxime and ciprofloxacin,55 where no excessive bleeding plus metronidazole (three doses) (US $14.70 versus was observed. US $25.80 [£9.79 versus £17.18]).82 A third trial concluded that monotherapy with co-amoxiclav was effective and cheaper than the combination of Cost and cost-effectiveness cefuroxime plus metronidazole (£8.10 versus £24.91) in abdominal surgery.116 However, this Among the trials reviewed, 18 included some conclusion was based on the results of all patients quantitative cost information. The proportion with abdominal surgery. When only colorectal of trials that reported quantitative cost information surgery was considered, the SWI was significantly was 10% for those published during 1984–91, and higher in the co-amoxiclav group than in the 26% for trials published during 1992–95. The cost cefuroxime plus metronidazole group (11.6% information reported in these trials should be versus 2.5%), and therefore the reduction in interpreted with caution because of small sample comparative efficacy will have a direct effect on sizes, incomplete measures of cost and benefit, cost-effectiveness data. and data from many different countries in differ- ent years. (Equivalent costs in £ sterling are calcu- Single dose or short-term use lated from the average exchange rate for the year Costs of single dose or short-term use were of publication, according to the financial statistics compared with those of multiple doses or long- from the Office for National Statistics.) Only three term use of antimicrobial prophylaxis in four UK studies reported cost information.116,123,144 trials. Jones and Wojeski71 concluded that over US $200,000 (£122,000) could be saved in their Cost of surgical wound infection hospital if single-dose cefotaxime prophylaxis was A study has reported that superficial SWI prolonged used rather than multiple-dose cephazolin or hospital stay by an average of 12.6 days.158 Another cefoxitin. Becker and Alexander24 compared study estimated that on average the hospital stay was cefoxitin for 5 days with cefoxitin for 16 hours. The 12 days longer for patients with wound infection or cost of the two regimens was US $586 and US $66 suture line leak, and the treatment cost increased (£328 and £37), respectively, and no SWI was by about US $2000 (£1120).55 A study in Scotland reported in either group. Corman and co-workers38 found that the median cost to the hospital of a compared single-dose cefuroxime plus metronida- wound infection after colorectal surgery in 1990 was zole with single-dose cefuroxime alone, four-dose £978 (95% CI: £484.04–£1521.22), including £858 cefoxitin, and single-dose cefoxitin. There were no hotel costs, £83.02 dressing costs, and £37.02 drug SWIs in the cefuroxime plus metronidazole group costs.174 In another Scottish study, Taylor and co- or the four-dose cefoxitin group. It was reported workers144 estimated that the financial consequence that the costs of these two regimens were US $30 was £1000–1300 for each infected patient (accord- and US $68 (£20 and £45), respectively. Single-dose ing to an estimate of 8.8 more days of hospital stay cefoxitin or single-dose cefotaxime plus metronida- for each patient with SWI, and the cost of £120–150 zole were compared with a three-dose regimen of for each hospital bed per day). the same antibiotics by Kow and co-workers.81 They concluded that the single-dose regimen was asso- In a study where more than 300 patients were ciated with significant cost savings, particularly if included and more than 24 patients developed the combination of cefotaxime plus metronidazole postoperative wound infection, it was estimated was used. that each wound infection cost about US $8400 (£4720) per patient, while the total cost of Less expensive drugs parenteral antibiotic use for the entire study The rates of SWI after two regimens of prophylactic was about US $14,880 (£8361).37 latamoxef plus metronidazole and co-amoxiclav 33 Results

plus metronidazole have been found to be com- treatment because of improved efficacy. One trial parable, but the cost of co-amoxiclav (£2.62 for compared ten doses of gentamicin plus lincomycin 1.2 g) was lower than that of latamoxef (£6.11 for (US $24.7, £19.04), gentamicin plus clindamycin 1 g).123 The cost of ceftriaxone was higher than (US $51.6, £39.77) and gentamicin plus metronida- cefamandole (AUS $22.78 versus AUS $8.22 per zole (US $117.7, £90.71).89 Although gentamicin dose [£10.04 versus £3.62]), but they resulted in plus metronidazole was the most expensive regi- similar SWI rates.62 In a trial in which amikacin plus men, it was associated with the lowest postoperative metronidazole started just before surgery for 48 wound infection rate and the lowest total hospital hours (seven doses) was compared with ceftriaxone cost (US $1587.7, £1223.57), compared with plus ornidazole (three doses) for 48 hours, the gentamicin plus lincomycin (US $1629.7, £1255.93) efficacies and costs were similar.147 and gentamicin plus clindamycin (US $1611.6, £1241.99). A study in Scotland compared piper- More efficacious regimens acillin alone (one dose) with piperacillin (one Gomez-Alonso and co-workers53 found that, dose) plus oral ciprofloxacin (two doses).144 Higher compared with a control group, prophylactic SWI rates in the piperacillin group (23.2% versus gentamicin plus metronidazole resulted in a lower 11.3%) resulted in patients remaining in hospital SWI rate and saved US $406 (£304) per patient. on average 2.7 days more than patients in the com- Two trials observed that more expensive regimens bination group. The addition of oral ciprofloxacin may be associated with a lower total cost of reduced the overall cost of hospital treatment.144

34 Health Technology Assessment 1998; Vol. 2: No. 7

Chapter 4 Discussion

Risk factors the wound because it is very difficult (if not impossible) for this to be complete. Patient’s natural The occurrence of SWI depends on two necessary resistance (systemic and local) is essential to prevent conditions: contamination of the wound by patho- contaminated wounds becoming infected. The risk genic bacteria, and a defective host resistance to of SWI increases if patient’s resistance is compro- infection.159 All factors associated with bacterial mised because of, for example, radiotherapy, contamination and patient’s resistance will also corticotherapy, chemotherapy, previous transplant- be associated with the risk of SWI (see Box). ation, diabetes, old age, obesity or weight loss.7 In addition, patient’s local resistance may be impaired because of ischaemia at the operation site.159 Determinants of SWI in colorectal surgery

Colorectal surgery National Nosocomial Infection Surveillance Type of operation (NNIS) is a risk index that can be used to estimate the risk of developing SWI.157 It includes three Bowel preparation risk factors: Surgical techniques • a patient having an American Society of Anesthesiologists (ASA) preoperative assessment Volume and score of 3, 4, or 5; (the ASA score is a measure virulence of of patient’s general condition; it ranges from 1, contaminated bacteria when a patient is normally healthy, to 5, when Natural systemic resistance a patient is very likely to die within 24 hours) Natural local resistance • an operation classified as either contaminated or dirty-infected Prophylactic antibiotics • an operation of more than T75% hours’ duration, where T75% depends on the operative procedure being performed. T75% is 3 hours in colon SWI surgery because the 75th percentile of the distribution for duration of colon surgery was found to be 3 hours. Colorectal surgery is classified as clean- contaminated or contaminated, depending on For patients undergoing surgery of the colon, it was the type of operation. For example, the risk of found that the rate of SWI was 3.2%, 8.5%, 16.1%, contamination may be higher for rectal resections and 22.2% for those with 0, 1, 2, and 3 risk factors, (abdominal–perineal operations) than other types respectively.157 Therefore, patients who undergo of colorectal surgery.160 Measures should be taken colorectal surgery may have different underlying to reduce the exogenous and endogenous con- risks of postoperative wound infection because of tamination of the surgical wound by pathogenic differences in patient characteristics, diagnoses, bacteria. Because a large volume of bacterial flora and level of complexity of the operations. is contained in the large bowel, mechanical cleans- ing is often used to reduce the contents of the As wound contamination by pathogenic bacteria colon before the operation. Seventy-eight per cent is often inevitable and host resistance is often of included trials (105 of 134 trials published in defective, the use of antibiotic prophylaxis becomes English) stated that preoperative bowel prepar- important in preventing infection following color- ations was carried out, though this does not neces- ectal surgery. In 1981, Baum and co-workers6 con- sarily mean that mechanical cleansing was not cluded that no-antibiotic controls should not be performed in the remaining 29 trials. considered in further trials of colorectal surgery. However, three trials that included no-antibiotic It is clear that mechanical cleansing alone is controls were published in 1984 and one in insufficient in preventing bacterial contamination of 1985.53,56,137,150 In these four trials, the SWI rate 35 Discussion

ranged from 32% to 58% in the no-antibiotic significant difference in patients’ characteristics groups, considerably higher than that in the and operation procedures between intervention antibiotic prophylaxis groups (Table 2). groups, the potential selection bias is still possible when the patient assignment is not appropriate. In the trials included in this review, some factors were identified as being associated with an increas- Follow-up and withdrawals ed risk of SWI following colorectal surgery, such as Incisional SWI is defined as an infection that duration of operation, obesity, the presence of occurs at an incision site within 30 days of drains, left-sided colonic resection, and inflam- surgery.165 A considerable number of SWIs may matory bowel disease. Two trials reported that the be detected after discharge from the hospital.166 surgeon’s experience might be a significant pre- Patients were usually followed up for 4 weeks dictor of postoperative wound infection.40,97 or longer in many trials. In some trials, only Perioperative blood transfusion was also found to infections occurring during hospital stay were be associated with an increased risk of SWI in two recorded. About 30% of the trials did not report trials.69,142 However, a reliable analysis of risk factors the duration of follow-up. The observed variation is impossible because potential risk factors were in the rate of SWI may be partially due to the inconsistently measured and findings might have different length of follow-up between the trials. been selectively reported in the trials. Patients may withdraw or be excluded from evaluation after randomisation for various reasons, Methodological limitations of for example, protocol violation, cancellation of the studies scheduled surgery, or death within 48 hours of surgery. The withdrawal rate ranged from 0% to Evans and Pollock161 used a score system to 42% in those trials that reported the number of assess the quality of published RCTs (1979–86) of withdrawals. About 23% of the trials did not record antibacterial prophylaxis in colorectal surgery. The withdrawals. If withdrawals from the trial are non- score system included design and conduct, analysis, random, and there are systematic differences and presentation. They found that only 23% of the between the groups, the different groups in a trial trials reached a score of more than 70 (maximum may no longer be comparable and attrition bias score 100). The most frequent error in design was a may be introduced. Intention-to-treat analysis is a faulty method of randomisation (36%). They also method of including all participants in an analysis observed that 80% (45/56) of the papers did not according to their initially assigned intervention report the fate of withdrawals. In 32 ‘negative’ trials group, regardless of whether or not they are with- only two considered the Type-II error. They found drawn, fully comply with treatment, or cross over no evidence of improvement in the quality and receive the alternative treatment. However, the standard of these reports over the 7 years.161 clinical outcome of participants who withdrew was not available for many trials. In the present review, In the present review, internal validity was the the rate of SWI was calculated by using number main concern. The quality of included studies of patients for whom SWI outcomes were evaluated. was assessed using a checklist which includes study components such as design and analysis. An overall Definition of surgical wound infection score is not calculated because different scales A written definition of SWI was available in 79% of may result in important difference in quality the included trials. SWI can be classified differently, assessment and none has been developed for example, primary versus secondary, superficial with sufficient rigour.162 versus deep, minor versus major, and early versus late. The SWI was defined in most cases as “the Patient allocation presence of purulent discharge” in the surgical Empirical evidence has shown that an inappro- wound, with or without positive bacteriological priate method of patient allocation is related to evidence. Swelling and reddening at the wound biased estimation of relative efficacy of healthcare site and temperature exceeding 38 °C were also interventions.163,164 In order to avoid the selection included as criteria for SWI.13,38 If possible, we bias, patients should be allocated to each group used abdominal wound infection as a principal randomly and the procedure should be concealed outcome measure to compare efficacy of antibiotic until allocation (true randomisation).11 The appro- prophylaxis. However, in some trials, it was not priateness of the method of randomisation used clear whether the wound infection included was judged to be uncertain in about 63% of the perineal wound infection or deep intraabdominal 36 included trials. Though many trials reported no infections. As the main concern is the internal Health Technology Assessment 1998; Vol. 2: No. 7

validity, the differences in the definition of SWI frequently in the earlier trials but were excluded across trials may have little impact on the evalu- from recent trials. ation of relative efficacy within trials. However, detection bias may occur when the diagnosis Inadequate direct comparison criteria are less objective, such as swelling or More than 70 different regimens of antibiotic reddening of surgical wound. In the diagnosis prophylaxis were compared in 147 trials (excluding of SWI, blind assessment of the outcome is a different doses or duration of treatment or anti- method to avoid detection bias. Only 40% of biotic combinations). Only a limited number of the included trials assessed outcome blindly. regimens were compared directly and each com- parison was evaluated by a limited number of trials, Magnitude of relative efficacy and in which the sample size was often too small to have sample size sufficient statistical power (see appendix 4). The The majority of included trials compared different quantitative pooling of results from individual regimens of antibiotic prophylaxis. The difference studies was not appropriate in most cases because in the mortality rate is generally expected to be of the diverse regimens used. very small. Therefore, SWI is used as the principal outcome measure in the review. Even when SWI Publication and related bias is measured as the main outcome, the difference The majority of the included trials reported results between comparison groups is still small, or which showed no statistically significant difference moderate in most cases. A meta-analysis concluded between the groups. Therefore, the risk that treat- that “comparisons of new therapies with standard ment effects have been over-estimated due to therapies will become prohibitively expensive publication bias seems to be remote. It is not clear because of the large number of patients required”.6 whether bias exists due to unrepresentative Thus, it is not surprising to find that about 80% inclusion of non-English language publications. of the trials that compared different regimens did not find a statistically significant difference Generalisability of results from between groups. included trials It may not be possible to generalise the results from When no significant difference was found, it may RCTs if the trials’ participants were different from be because both regimens being compared were patients seen in the ordinary practice. The criteria similarly effective. However, another reason for for inclusion and exclusion were described in 87% non-significant findings may be the low power due of the trials reviewed here. The exclusion criteria to inadequate sample size, particularly when the most frequently used were allergy to study drugs, difference between antibiotics is small or moderate. preoperative use of other antibiotics, impaired Suppose the rate of SWI can be lowered by a renal or liver function, children or very old regimen of antibiotic prophylaxis from 10% to 5%, patients, pregnancy or lactation, and certain types to detect such a difference at the 5% significance of colorectal operations. Lack of informed consent level and 80% power, more than 400 patients are may be an important reason for exclusion but required in each group.167 Such a large number of only a few trials stated this explicitly.60,62,67,84 Data patients was unusual in the trials included. Eighty- reported in the trials were insufficient for a reliable two per cent of the trials had not undertaken an comparison between trial participants and patients a priori calculation of required sample size. It is who were excluded. encouraging that the sample sizes of studies pub- lished more recently (1992–95) are larger than studies published previously (1984–87). The num- Timing and duration ber of patients included in each group was on aver- age 51, 77 and 124 for the trials published during According to the definition of prophylaxis, 1984–87, 1988–91, and 1992–95, respectively. antimicrobial agents should be administered before the onset of SWI. Once the infection The increase in the number of patients in the occurs antibiotic administration should be clinical trials does not correspond with an increase considered as therapeutic, rather than prophy- in the number of trials that found significant differ- lactic, for wound infection. Bartlett and Burton12 ence. The proportion of trials finding significant classified the timing of the antibiotic treatment difference in SWI was 20.6%, 18.0%, and 17.4% into three categories: for trials published during 1984–87, 1988–91, and 1992–95, respectively. This may also be because • preoperative regimens, in which antibiotics are inadequate antibiotic regimens were tested more given before the day of surgery 37 Discussion

• perioperative regimens, in which antibiotics are the operation. For example, oral neomycin plus given before or during the surgical procedure erythromycin is given from 9–20 hours before the • postoperative regimens, in which antibiotics are operation. The main aim is to reduce the risk of given after surgery. bacterial contamination of the surgical wound by reducing the bacteria in the large bowel prior to A different approach was used by Classen and co- surgery. Some trials showed that oral neomycin workers168 to define the timing of prophylactic and erythromycin on the day before the surgery is administration of antibiotics. It was defined as: effective at reducing SWI, but may not be sufficient on its own. Further lowering of the rate of SWI may • early, if prophylactic antibiotics were be achieved by adding parenteral antibiotics administered 2–24 hours before the immediately before surgery.83,138 surgical incision • preoperative, if administered 2 hours before the incision Inadequate antibiotic regimens • perioperative, if administered within 3 hours after the incision Some regimens appear to be poor at preventing • postoperative, if administered 3–24 hours SWI in colorectal surgery because of inadequate after the incision. antimicrobial coverage, or inappropriate timing and dosing. Due to the divergent (polymicrobial) In order to be effective against postoperative wound nature of the micro-organisms that are potentially infection, it is crucial that the concentration of anti- responsible for infection following colorectal biotics in the tissue surrounding the surgical wound surgery, the antibiotics used should be of broad- should be sufficient at the time of bacterial contam- spectrum activity and active against both aerobic ination. Therefore, prophylactic antibiotics should and anaerobic bacteria. Metronidazole is active be administered 2 hours before surgery.168 However, against anaerobic bacteria but inadequate against it is not clear for how long antibiotics should be aerobic bacteria and should therefore be combined given after the operation. In some trials, antibiotic with other antibiotics. The following regimens were prophylaxis was used for several days after the oper- also shown to be inadequate in some trials: ation. It has been argued that antibiotic prophylaxis neomycin alone,67 gentamicin alone,42 doxycycline should only be used as short-term treatment, in alone,26,52 cefotaxime alone,59 tinidazole order to reduce toxicity, costs and the possibility alone,95,96,112 and piperacillin alone.131,144 of developing bacteria resistance.169 In the overview by Baum and co-workers,6 the Seventeen trials compared a single-dose regimen SWI rate after colorectal surgery in patients who with a multiple-dose regimen. None of these trials received antibiotic prophylaxis was, on average, found a significant difference in postoperative SWI 22% in 26 trials published between 1965 and 1980. rate between the two regimens. Because the single In another review by Bartlett and Burton,12 the dose of antibiotics is as efficacious as a multiple- overall rate of SWI in colorectal surgery with anti- dose regimen, the use of single dose or short-term biotic prophylaxis was 18% according to 23 trials antibiotics (e.g. less than 24 hours) is justified in published between 1960 and 1980. In our review, the prevention of SWI in colorectal surgery. the overall rate of SWI in antibiotic groups was 11%, which is lower than rates reported previously. Several trials reported the duration of operation. This difference may be due to different definitions These trials did not provide any convincing evi- of SWI, different methods of bowel preparation in dence of the relation between the efficacy of single- trials, variation in patient selection, different length dose regimens and the duration of operation. How- of follow-up, or different intensity of surveillance. It ever, it has been recommended that a second dose may also be due to a more effective or appropriate is needed if the operation is over 2 hours’ duration, use of antibiotic prophylaxis in trials published particularly if the half-life of the antibiotic is more recently. short.1,4,5,7 Due to a lack of convincing evidence, clinicians need to consider other factors such as In Baum’s review, the antibiotics that were used perioperative blood loss to decide whether a included metronidazole, gentamicin, cephalothin, second-dose regimen is required. doxycycline, and neomycin plus erythromycin.6 In Bartlett and Burton’s review, the trials used a wider When an antibiotic is administered more than a range of antibiotics including neomycin, kanamy- day before the operation, the tissue concentration cin, erythromycin, penicillin, gentamicin, cephalo- 38 of the antibiotic agent may not be adequate during thin, doxycycline, metronidazole, tinidazole, Health Technology Assessment 1998; Vol. 2: No. 7

tetracycline, ampicillin, bacitracin, chlorchinaldole, of some agents could be improved by the addition clindamycin, and cefamandole.12 According to the of metronidazole (Table 14). results from recently published trials, other regi- mens may be more effective than monotherapy In summary, it appears that additional antibiotics with metronidazole, gentamicin, and doxycycline, were beneficial when they were added to pipera- and the combination of neomycin plus erythro- cillin, doxycycline, tinidazole, and metronidazole, mycin (Tables 11–13 and 17). Significant advances but no additional benefit was observed when anti- in therapy since 1975 have resulted in a more biotics were added to cefoxitin, cefotetan, ceftri- effective range of antibiotic regimens in the axone, ceftizoxime, latamoxef, and gentamicin prevention of SWIs.170 plus metronidazole.

There is no obvious difference, however, in the rate of SWI when the RCTs included in this review are New-generation cephalosporins grouped according to the year of publication. The overall rate of SWI was 11% (6195 patients), 12% There is no consistent evidence to suggest that (10,297 patients), and 11% (6262 patients) for the new-generation cephalosporins are more trials published during 1984–87, 1988–91, and effective than old-generation cephalosporins in 1992–95, respectively. the prevention of SWI in colorectal surgery. Six trials compared a first-generation cephalosporin with a second- or third-generation cephalosporin Additional antibiotics and the differences in rates of SWI between groups were not statistically significant (Table 24). Similarly, Many trials tested whether combination therapy pooling results from these six trials to increase the was more effective than monotherapy. For example, power did not show a statistically significant differ- metronidazole alone was found to be less effica- ence between the new-generation and the first- cious than a combination of metronidazole with generation cephalosporins (overall rate of SWI: other antibiotics (such as cefuroxime, fosfomycin, 6.0% versus 6.4%; odds ratio = 0.93; 95% CI: and ampicillin) (Table 13). Similarly, the efficacy 0.46, 1.86).

TABLE 24 Second- and third-generation cephalosporins versus first-generation cephalosporins

Regimen and study Incidence of SWI Odds ratio Odds ratio [reference] (95% CI fixed) [95% CI fixed] Second/third First generation generation n/N n/N

(Cefoxitin × 11) vs. (cephalothin × 18) 0/30 3/30 0.13 [0.01, 2.61] Antonelli [17]

(Cefoxitin × 5) vs. (cephazolin × 4) 1/36 1/35 0.97 [0.06, 16.16] Jones [71]

(Cefotaxime × 1) vs. (cephazolin × 4) 4/29 1/35 5.44 [0.57, 51.68] Jones [71]

[All: Metronidazole suppository] 7/90 7/96 1.07 [0.36, 3.19] (Ceftriaxone × 1) vs. (cephazolin × 1) Lumley [91]

(Latamoxef × 3) vs. (cephazolin × 3) 1/26 1/24 0.92 [0.05, 15.58] Plouffe [125]

(Latamoxef × 3) vs. (cephazolin + 3/60 5/60 0.58 [0.13, 2.54] metronidazole) × 3 Thomas [146] 0.01 0.1 1 10 100 (log scale) Favours second/third generation Favours first generation 39 Discussion

Route of administration Bacteriology and antibiotic resistance Prophylactic antibiotics can be administered in three ways.12 Most trials reported results of bacteriological testing (110/134). Bacteria isolated from infected • Oral administration antibiotics are given by wounds in colorectal surgery were often a mixture mouth, rectum, colostomy, or nasogastric tube. of aerobic and anaerobic organisms, among which • Parenteral route antibiotics are administered E. coli and B. fragilis were most common. S. aureus subcutaneously, intramuscularly, was also frequently isolated. The type of bacteria or intravenously. isolated from surgical wounds might be altered • Topical administration antibiotics are applied after use of antibiotic prophylaxis. For example, directly to the surgical wound at the time bacteria isolated from wound infections were pre- of operation. dominantly aerobic bacteria when prophylactic metronidazole had been administered.35 One trial The intravenous route is the most reliable method found that anaerobes were isolated from only one of achieving a rapid increase in plasma level and of 15 wound infections in the metronidazole group tissue concentration of the antibiotic, because the but from three of six wound infections in patients intramuscular route may be limited by pharmaco- receiving cephozolin plus oral neomycin kinetic features of a drug and the absorption of plus erythromycin.76 antibiotics by oral route may be unreliable.7 A regimen of antibiotic prophylaxis in surgery may No trial compared different routes of the same become ineffective if antibiotic-resistant bacteria antibiotic. No extra benefit was observed in develop. The type and extent of antibiotic resist- six trials that compared parenteral alone with ance may vary from country to country and within parenteral plus topical use of antibiotic prophy- a country.171 Evidence has shown that increased laxis (Table 25). However, the regimens that includ- usage of antibiotics might be related to a rise in ed oral antibiotics were associated with a lower rate the prevalence of resistant bacteria and vice versa.172 of SWI in nine out of 12 trials, though the differ- It was suggested that the development of antibiotic- ence was statistically significant in only three trials. resistant bacteria may be reduced if hospital Some of these trials used inadequate parenteral infections could be prevented and if the use antibiotics such as metronidazole alone or pipera- of antibiotics could be reduced.172 cillin alone. In general, oral or topical application of antibiotics in addition to the parenteral adminis- By preventing postoperative wound infection, a tration of appropriate antibiotics seem to be of single dose or short-term antibiotic prophylaxis can limited value. reduce the need for long-term antibiotic therapy and therefore may contribute to reducing selection In the USA, the most commonly used oral regimen of antibiotic-resistant bacteria. To be effective in is a combination of neomycin and erythromycin on preventing SWIs, prophylactic antibiotics should the day before operation.160 However, this regimen be selected according to the local presence and alone was less efficacious than intravenous regi- prevalence of antibiotic-resistant bacteria.172 For mens of ceftriaxone plus metronidazole,152 addi- these reasons, the search for the ideal prophylactic tional cefamandole,121 additional cefoxitin,138 and regimen must be a continuous process and univer- gentamicin plus metronidazole.83 sal acceptance and use of a regimen should be avoided.173 Adverse events with antibiotic prophylaxis Cost and cost-effectiveness Consideration of toxicity profiles and adverse Both costs and benefits of antimicrobial events is important when selecting prophylactic prophylaxis in surgery may be direct or indirect, antimicrobial agents. Reports of adverse events and a number of components may be included.2 appear to have declined over the years and conse- Costs of antibiotic prophylaxis include costs of quently, it seems that concern about adverse events antibiotic drugs, equipment, and staff time. These associated with antimicrobial prophylaxis may also may be offset to some extent by reductions in the have diminished. However, skin rash, diarrhoea, length of the hospital stay. Although this report and nausea are still commonly reported and may has not attempted a systematic review of the cost- 40 be attributable to the use of some antibiotic agents. effectiveness of antimicrobial prophylaxis in Health Technology Assessment 1998; Vol. 2: No. 7

TABLE 25 Route of administration: parenteral versus parenteral plus oral or topical

Regimen and study Incidence of SWI Odds ratio Odds ratio [reference] (95% CI fixed) [95% CI fixed] Parenteral Parenteral only plus oral/topical n/N n/N

(Cefoxitin × 3) vs. (cefoxitin × 3) + 0/39 4/41 0.11 [0.01, 2.03] (tinidazole × 1) + (metronidazole oral × 3) Peruzzo [120]

(Cefoxitin × 3) vs. (cefoxitin × 3) + 2/51 3/51 0.65 [0.10, 4.08] (neomycin + erythromycin oral) × 3 Stellato [141]

(Cefoxitin × 5) vs. (cefoxitin × 5) + 15/141 9/169 2.12 [0.90, 5.00] (neomycin + erythromycin oral) × 3 Coppa [37]

(Gentamicin + metronidazole) × 1 vs. 5/67 3/65 1.67 [0.38, 7.28] (gentamicin + metronidazole) × 1 + (neomycin + erythromycin oral) × 3 Lau [83]

(Gentamicin + metronidazole) × 3 vs. 13/45 4/40 3.66 [1.08, 12.36] (metronidazole × 3) + (ciprofloxacin oral) McArdle [93]

(Gentamicin + metronidazole) × 10 vs. 7/42 4/42 1.90 [0.51, 7.05] (metronidazole × 10) + (ciprofloxacin oral × 1) McArdle [93]

(Piperacillin × 1) vs. (piperacillin × 1) + 39/168 18/159 2.37 [1.29, 4.35] (ciprofloxacin oral × 2) Taylor [144]

(Piperacillin × 4) vs. (piperacillin × 4) + 26/223 9/107 1.44 [0.65, 3.18] (metronidazole × 6) + (neomycin oral × 8) Reynolds [131]

(Metronidazole × 3) vs. (cephazolin + 14/47 4/55 5.41 [1.64, 17.86] neomycin + erythromycin oral) × 3 Khubchandani [76]

(Metronidazole × 1) vs. (metronidazole × 1) + 16/58 9/61 2.20 [0.88, 5.48] (metronidazole × 3) + (neomycin oral × 4) Playforth [124]

(Latamoxef × 3) vs. (metronidazole × 3) + 5/64 3/67 1.81 [0.41, 7.90] (neomycin oral × 3) Hinchey [65]

(Fosfomycin + metronidazole) × 1 vs. 6/77 7/72 0.78 [0.25, 2.46] (ampicillin × 1) + (metronidazole + bacitracin + neomycin) × 6 Nohr [111]

0.01 0.1 1 10 100 (log scale) Favours parenteral Favours parenteral + oral/topical

continued 41 Discussion

TABLE 25 contd Route of administration: parenteral versus parenteral plus oral or topical

Regimen and study Incidence of SWI Odds ratio Odds ratio [reference] (95% CI fixed) [95% CI fixed] Parenteral Parenteral only plus oral/topical n/N n/N

(Cefotaxime × 3) vs. (neomycin + 2/50 6/50 0.31 [0.06, 1.59] erythromycin oral) × 3 + (ampicillin topical × 1) Raahave [128]

(Cefotaxime × 3) vs. (cefotaxime × 3) + 6/89 5/81 1.10 [0.32, 3.75] (ampicillin topical × 1) Raahave [129]

(Cefotaxime + metronidazole) × 10 + 4/19 2/21 2.53 [0.41, 15.75] (cefotaxime + metronidazole) × 10 + (cefotaxime topical × 1) Moesgaard [104]

(Gentamicin + metronidazole) × 7 vs. 4/38 5/41 0.85 [0.21, 3.42] (gentamicin + metronidazole) × 7 + (gentamicin + metronidazole) topical Moesgaard [102]

(Gentamicin + metronidazole) × 1 vs. 18/65 15/64 1.25 [0.57, 2.77] (gentamicin + metronidazole) + cefotetan topical Greig [57]

(Ampicillin + metronidazole) × 9 vs. 5/98 5/105 1.08 [0.30, 3.83] (ampicillin + metronidazole) × 9 + ampicillin topical Juul [73] 0.01 0.1 1 10 100 (log scale) Favours parenteral Favours parenteral + oral/topical

colorectal surgery, some general issues about cost- infection following colorectal surgery was £978 in effectiveness of antibiotic prophylaxis are discussed, 1990.174 These are clearly all costs that could be as well as cost information that was reported in reduced or avoided with an appropriate some RCTs. prophylactic regimen.

Cost of surgical wound infection The relative cost-effectiveness of different regimens The risk of postoperative wound infection is high need to be compared to decide which regimen (> 50%) in colorectal surgery if antimicrobial should be selected in the prevention of SWI in prophylaxis is not employed, and SWI is associated colorectal surgery. A cost-effectiveness ratio is with a prolonged hospitalisation and costly determined by the net cost and the effect of a antibiotic treatment. healthcare intervention.175 The net cost of a regimen depends not only on the cost of a regimen Relative cost-effectiveness (including the cost of drug purchase, preparation The costs associated with SWI are high, in terms of and administration of the drug) but also the savings both antibiotic treatment and prolonged hospital- associated with the use of antibiotic prophylaxis, isation. Two studies showed that patients with such as a reduction in hospital stay. When there is SWIs remained in hospital for approximately no difference in the efficacy and safety of prophy- 12 days longer than those patients without infec- lactic antibiotics, the cost and ease of use become 42 tions,55,158 and another study calculated that wound criteria for the selection of regimens.173,176 Health Technology Assessment 1998; Vol. 2: No. 7

It is possible to reduce the cost of antibiotic In an audit of surgical prophylaxis in three prophylaxis without adversely affecting SWI Hospital Trusts in Tayside, Scotland, it was found rate.177–181 For example, where appropriate: that approximately 28% of patients undergoing colorectal surgery received antibiotic prophylaxis • a single dose or short-term regimen (less than for more than 24 hours.185 Against the standard 24 hours after surgery) can be adopted instead recommended by the three hospitals, a second of inappropriate long-term use of antibiotics dose of antibiotic prophylaxis was seldom given • more effective and less costly drugs can be used for operations lasting more than 2 hours. instead of expensive drugs • monotherapy can be used instead of Janknegt and co-workers186 studied 33 guidelines combination therapy. for antibiotic prophylaxis in surgery used in 89 Dutch hospitals. Since 1991, more guidelines have recommended the use of single-dose prophy- Antibiotic prophylaxis in laxis in colorectal surgery compared with before surgical practice this date (50% versus 8%). The antibiotics recom- mended generally are active against anaerobic Evidence from hospital surveys suggests that and aerobic bacteria. The most commonly recom- inappropriate use of antimicrobial prophylaxis in mended regimens for colorectal surgery include surgery is common in many countries. In seven co-amoxiclav (8% before 1991 versus 25% since Malaysian hospitals, the use of surgical antibiotic 1991), cefuroxime plus metronidazole (25% versus prophylaxis was investigated and it was found that 25%), and gentamicin plus metronidazole (33% it was often prescribed for an unnecessarily long versus 20%).186 period.182 A Spanish study (n = 714) showed that prophylaxis was given for a mean of 8.4 days, and Oral neomycin plus erythromycin on the day that the first choice antibiotic was selected in before or 9–20 hours prior to surgery were often only 20% of the cases.183 A survey on antibiotic recommended by authors in North America.4,5,8 prophylaxis in 889 surgical departments in Results from trials that compared this regimen German hospitals found that inappropriate alone with other regimens were inconsistent. antibiotics were selected 70.5% of the time, Oral neomycin plus erythromycin were found and that the duration of prophylaxis was not to be significantly less efficacious than parenteral optimal in 57.1% of cases.184 ceftriaxone plus metronidazole,152 gentamicin plus metronidazole,83 or with additional cefoxitin.138 In a District General Hospital in England, Dobrzanski and co-workers179 identified the major In the British National Formulary (BNF No. 34, problems associated with the use of antimicrobial September 1997), a single dose of gentamicin plus prophylaxis in abdominal and arterial surgery. metronidazole or cefuroxime plus metronidazole These included no antibiotics at induction (35%), are recommended for preventing infection in questionable antibiotics at induction (22%), colorectal surgery. A survey of guidelines for anti- questionable postoperative antibiotics (25%), microbial prophylaxis in surgery in 392 hospitals in and unnecessarily long postoperative treatment the UK found that formal guidelines were available courses (70%). The investigators developed the in 47% of the 160 hospitals that responded.187 surgical antibiotic prophylaxis guidelines.179 Recommended regimens for colorectal surgery According to these guidelines, one perioperative included co-amoxiclav (6%), second-generation and two postoperative doses of cefotaxime (1 g) cephalosporin (2%), metronidazole alone (4%), plus metronidazole (500 mg) should be intraven- aminoglycoside plus metronidazole (19%), first- ously administered, in addition to oral neomycin generation cephalosporin plus metronidazole (1 g) plus metronidazole (400 mg) for bowel (15%), second-generation cephalosporin plus preparation on the day before abdominal surgery. metronidazole (42%), and other antibiotics The antibiotic regimen recommended in these (12%). It was claimed that 94% of these regimens guidelines may not be optimal; for example, included antibiotics with activity against both a single dose may be enough in many patients. aerobic and anaerobic bacteria.187 However, by introducing the guidelines, the use of surgical antibiotic prophylaxis became more In a recent survey of antibiotic policies, Bloxham appropriate and the cost of antibiotic prophylaxis sent a total of 172 letters to consultant micro- per surgical patient was reduced from £38.13 to biologists in hospitals in the UK.188 Recommenda- £16.93. The rate of SWI was not followed in tions on antibiotic prophylaxis in colorectal surgery this study.179 was available in 49 of the 68 antibiotic policies 43 Discussion

returned. Cefuroxime plus metronidazole was the plus metronidazole in five policies. Other anti- most frequently recommended (28 policies) as the biotics recommended for antimicrobial prophylaxis first choice. Co-amoxiclav was recommended in in colorectal surgery included cephradine and seven hospital antibiotic policies and gentamicin cefotaxime plus metronidazole.

44 Health Technology Assessment 1998; Vol. 2: No. 7

Chapter 5 Conclusions

ntibiotic prophylaxis is effective in the area is sufficiently high when bacterial A prevention of SWI in colorectal surgery. It contamination occurs. appears that some regimens may be inadequate, such as metronidazole alone, and piperacillin alone In addition, universal acceptance and use of a compared with other regimens. The efficacy of regimen should be prevented in order to minimise many different regimens may be similar and it is the development of antibiotic-resistant bacteria. very difficult to identify the most effective. There is Based on available research evidence, guidelines no convincing evidence to suggest that the new- should be developed locally in order to achieve generation cephalosporins are more efficacious more cost-effective use of antimicrobial prophylaxis than first-generation cephalosporins in the in colorectal surgery. prevention of SWI in colorectal surgery.

A single dose or short-term use (less than 24 hours Recommendations for research after operation) is as efficacious as long-term postoperative use of antibiotic prophylaxis, and The review of RCTs presented here has highlighted may be associated with less toxicity, fewer adverse some important aspects of investigation and of anti- events, a reduced risk of developing bacterial biotic use. The quality of the included trials has resistance, and lower overall costs. improved over the last 12 years, though there still appear to be many methodological problems such as inappropriate methods of patient allocation, lack Implications for policy of blinding during assessment of outcomes, and insufficient sample sizes. Two principles are important to follow when selecting an appropriate regimen. Further studies of efficacy may be of little value and would require large numbers of patients in • Antibiotics or antibiotic combinations order to demonstrate a statistically significant should be active against both aerobic and difference between regimens. Future research anaerobic bacteria. should focus on the understanding of the practical • The administration of antibiotics should use of antimicrobial prophylaxis in colorectal be timed to ensure that the tissue concen- surgery in the UK and the cost-effectiveness of tration of antibiotics around the wound the different regimens.

45

Health Technology Assessment 1998; Vol. 2: No. 7

Acknowledgements

his review was commissioned by the NHS R&D helpful comments on the review protocol and T HTA programme. The authors would like to draft report, and Professor MRB Keighley, Mr P thank Julie Glanville and Janette Boynton for assist- McCulloch, and Professor A Pollock for responding ing with the search and location of the literature. We to our inquiries about their studies. The report also are most grateful to Professor R Freeman, Professor benefited greatly from the HTA referees’ comments. DJ Leaper, Mr EW Taylor, Professor RJC Steele, Dr We would like to thank Professor Trevor Sheldon for PG Davey and Professor CS McArdle for their his supervision and support during the review.

47

Health Technology Assessment 1998; Vol. 2: No. 7

References

1. Dellinger EP, Gross PA, Barrett TL, Krause PJ, 14. Aberg C, Thore M. Single versus triple dose Martone WJ, McGowan JE, et al. Quality standard for antimicrobial prophylaxis in elective abdominal antimicrobial prophylaxis in surgical procedures. surgery and the impact on bacterial ecology. Clin Infect Dis 1994;18:422–7. J Hosp Infect 1991;18:149–54. 2. McGowan JE. Cost and benefit of perioperative 15. AhChong K, Yip AWC, Lee FCW, Chiu KM. antimicrobial prophylaxis: methods for economic Comparison of prophylactic ampicillin/sulbactam anaysis. Rev Infect Dis 1991;13:879–89. with gentamicin and metronidazole in elective colorectal surgery: a randomized clinical study. 3. McDonald P, Finlay-Jones J. Microbial evolution J Hosp Infect 1994;27:149–54. of surgical infection. In: Watts JM, McDonald P, O’Brien PE, Marshall VR, Finlay-Jones J, editors. 16. Andaker L, Burman LG, Eklund A, Graffner H, Infection in surgery: basic and clinical aspects. Hansson J, Hellberg R, et al. Fosfomycin/metro- Edinburgh: Churchill Livingstone, 1981:3–9. nidazole compared with doxycycline/metronidazole for the prophylaxis of infection after elective 4. Waddell TK, Rotstein OD. Antimicrobial colorectal surgery. Eur J Surg 1992;158:181–5. prophylaxis in surgery. CMAJ 1994;151:925–31. 17. Antonelli W, Borgani A, Machella C, Morri F, 5. Ludwig KA, Carlson MA, Condon RE. Parrino A, Poloni M, et al. Comparison of two Prophylactic antibiotics in surgery. Annu systematic antibiotics for the prevention of Rev Med 1993;44:385–93. complications in elective colorectal surgery. 6. Baum ML, Anish DS, Chambers TC, Sacks HS, Ital J Surg Sci 1985;15:225–58. Smith H, Fagerstrom RM. A survey of clinical 18. Armengaud F, Jobard J, Bernard E, Cordero C, trials of antibiotic prophylaxis in colon surgery: Mouiel J, Sicard D, et al. Single dose antibiotic evidence against further use of no-treatment prophylaxis in colorectal surgery: cefoxitin versus controls. New Engl J Med 1981;305:795–9. piperacillin [French]. Presse Med 1986;15:2351–2. 7. Martin C. Antimicrobial prophylaxis in surgery: 19. Arnaud JP, Bellissant E, Boissel P, Carlet J, Chastang general concepts and clinical guidelines. Infect C, LaFaix C, et al. Single-dose amoxycillin-clavulanic Control Hosp Epidemiol 1994;15:463–71. acid versus cefotetan for prophylaxis in elective 8. Page CP, Bohnen JM, Fletcher R, McManus colorectal surgery: a multicentre prospective, AT, Solomkin JS, Wittman DH. Antimicrobial randomized study. J Hosp Infect 1992;22:23–32. prophylaxis for surgical wounds. Arch Surg 20. Athanasiadis VS, Khulgatz C, Hanel E. Peri- 1993;128:79–88. operative antibiotikapropylaxw in der kolon – 9. Gorbach SL, Condon RE, Conte JE, Kaiser AB, und rektumchirurgie [German]. Zentralbl Chir Ledger WJ, Nichols RL. Evaluation of new anti- 1985;110:532–8. infective drugs for surgical prophylaxis. Clin Infect 21. Auger P, Legros G, Girard R, Laverdiere M, Dis 1992;15:S313–38. Bergeron M, Bourgouin J, et al. Intravenous metro- 10. Mulrow CD, Oxman AD. Cochrane Collaboration nidazole V’s oral erythromycin base plus neomycin Handbook. [Update 9 December 1996]. Available in the prevention of infection following elective in The Cochrane Library [database on disk and colorectal surgery. Curr Ther Res 1987;42:922–31. CD-ROM]. The Cochrane Collaboration; Issue 1. 22. Bates T, Crathern BC, Bradley SP, Zlotnik RD, Oxford: Update Software, 1997. Crouch C, James RDG, et al. Timing of prophylactic 11. Pollock A. Surgical infection. London: Edward antibiotics in abdominal surgery: trial of a pre- Arnold Ltd., 1987. operative versus an intraoperative first dose. Br J Surg 1989;76:52–6. 12. Bartlett SP, Burton RC. Effects of prophylactic antibiotics on wound infection after elective 23. Bates T, Roberts JV, Smith K, German KA. A colon and rectal surgery. Am J Surg randomised trial of one versus three doses of 1983;145:300–9. Augmentin as wound prophylaxis in at-risk abdominal surgery. Postgrad Med J 1992;68:811–16. 13. Aberg C, Olin B, Oresland T, Lundholm C, Bernander S, Stromberg A, et al. Comparison of 24. Becker JM, Alexander DP. Colectomy, mucosal metronidazole with doxycycline prophylaxis in proctectomy and ileal pouch-anal anastomosis: a elective colorectal surgery. Acta Chir Scand prospective trial of optimal antibiotic management. 1984;150:79–83. Ann Surg 1991;213:242–47. 49 References

25. Bellantone R, Pacelli F, Sofo L, Doglietto GB, 38. Corman ML, Robertson WG, Lewis TH, Oden- Bossola M, Ratto C, et al. Systematic preoperative heimer DB, Zegarra P, Prager ED. A controlled prophylaxis in elective oncological colorectal clinical trial: cefuroxime, metronidazole and surgery: cefotetan versus clindamicin plus cefoxitin as prophylactic therapy for colorectal aztreonam. Drugs Exp Clin Res 1988;XIV:763–6. surgery. Complications in Surgery 1993;20:37–40.

26. Bergman L, Solhaug JH. Single-dose chemo- 39. Cuncliffe WJ, Carr N, Schofield PF. Prophylactic prophylaxis in elective colorectal surgery: a com- metronidazole with and without cefuroxime in parison between doxycycline plus metronidazole elective colorectal surgery. J R Coll Surg Edinb and doxycycline. Ann Surg 1987;205:77–81. 1985;30:123–5.

27. Bittner R, Butters M, Rampf W, Kapfer X. Duration 40. Cuthbertson AM, McLeish AR, Penfold JCB, of antibiotic prophylaxis in colorectal surgery – Ross H. A comparison between single and double one-shot dose vs short-term prophylaxis [German]. dose intravenous Timentin for the prophylaxis Langenbecks Arch Chir 1989;374:272–9. of wound infection in elective colorectal surgery. 28. Blair JE, McLeod RS, Cohen Z, Devlin HR. Dis Colon Rectum 1991;34:151–5. Ticarcillin/clavulanic acid (Timetin) compared 41. De La Hunt MN, Karran SJ. Sulbactam/ampicillin to metronidazole/netilmicin in preventing post- compared with cefoxitin for chemoprophylaxis in operative infection after elective colorectal surgery. elective colorectal surgery. Dis Colon Rectum Can J Surg 1987;30:120–3. 1986;29:157–9. 29. Brolin J, Lahnborg G, Ljung A, Rietz KA. 42. Desaive C. Utilisation de la ticarcilline et/ou de la A comparison between netilmicin with metro- gentamicine dans la propylaxie de l’infection en nidazole and doxycycline as prophylaxis in chirurgie recto-colique: etude randomisee elective colorectal surgery. Ann Chir Gynaecol [French]. Acta Therapeutica 1985;11:405–15. 1986;75:219–24. 30. Burdon DW, Keighley MRB. Ceftriaxone and 43. Devecioglu S, Tuylu Y, Zissis NP. Piperacillin metronidazole as single-dose prophylaxis in prophylaxis in colorectal surgery. A randomized colorectal surgery. S Afr Med J 1987;(suppl):15–18. comparative evaluation of two dosage schedules. Saudi Medical Journal 1990;11:385–8. 31. Cai CJ. Clinical study of prophylactic use of gentamicin and metronidazole in the surgery of 44. Diamond T, Mulholland CK, Hanna WA, Parks TG. colorectal carcinoma [Chinese]. Chinese J Surg A prospective randomized trial to compare triple 1992;30:237–40. dose mezlocillin with triple dose cefuroxime plus metronidazole as prophylaxis in colorectal surgery. 32. Cainzos M, Potel J, Puente JL. Short-term antibiotic J Hosp Infect 1988;12:215–19. prophylaxis in colorectal surgery: a comparative study of gentamicin plus clindamycin vs cefoxitin. 45. DiPiro JT, Welage LS, Levine BA, Wing PE, Acta Therapeutica 1986;12:399–412. Stanfield JA, Gaskill HV, et al. Single-dose cefmetazole versus multiple dose cefoxitin for 33. Cann KJ, Watkins RM, George C, Rayne-James J, prophylaxis in abdominal surgery. J Antimicrob Crawfurd E, Rogers TR. A trial of mezlocillin Chemother 1989;23:71–-7. versus cefuroxime with or without metronidazole for the prevention of wound sepsis after biliary 46. Fabian TC, Mangiante EC, Boldreghini SJ. and gastrointestinal surgery. J Hosp Infect Prophylactic antibiotics for elective colorectal 1988;12:207–14. surgery or operation for obstruction of the small bowel: a comparison of cefonicid and cefoxitin. 34. Carr ND, Hobbiss J, Cade D, Schofield PF. Rev Infect Dis 1984;6:S896–S900. Metronidazole in the prevention of wound sepsis after elective colorectal surgery. J R Coll Surg Edinb 47. Favre JP, Bouchet Y, Clotteau JE, Hypousteguy L, 1984;29:139–42. Marchal G, Mercier R, et al. Prophylactic use of 35. Claesson BEB, Filipsson S, Holmlund DEW, cefotaxime in colonic and rectal surgery. Matzsch TW, Wahlby L. Selective cefuroxime J Antimicrob Chemother 1984;14:247–53. prophylaxis following colorectal surgery based 48. Figueras-Felip J, Basilio-Bonet E, Lara-Eisman F, on intra-operative dipslide culture. Br J Surg Fava-Bargallo P, Isamat-Baro E, Rosell-Abaurrea F. 1986;73:953–7. Oral is superior to systematic to systematic 36. Colizza S, Fazio D, Addari A, Grande R, Cucchiara antibiotic prophylaxis in operations upon G. Short-term prophylaxis with cefuroxime in the colon and rectum. Surg Gynecol Obstetr colorectal surgery for cancer. J Surg Oncol 1984;158:359–62. 1987;35:266–8. 49. Fingerhut A, Hay J-M, and the French Association 37. Coppa GF, Eng K. Factors involved in antibiotic for Surgical Research. Single-dose ceftriaxone, selection in elective colon and rectal surgery. ornidazole and povidone-iodine enema in elective 50 Surgery 1988;104:853–8. left colectomy. Arch Surg 1993;128:228–32. Health Technology Assessment 1998; Vol. 2: No. 7

50. Franceshini FD, Mastio A, Manzoli D, Mancini P, 63. Haverkorn MJ. Peroperative systematic prophylaxis Rubbo G, Crescioli R. Short-term antimicrobial in colorectal surgery. Drugs Exp Clin Res prophylaxis in colorectal surgery: ceftriaxone vs 1985;X1:111–14. aztreonam-clindamicin association in randomized studies [Italian]. Chirurgia 1989;XLII:229. 64. Hershman MJ, Swift RI, Reilly DT, Logan WA, Sackier JM, Gompertz H, et al. Prospective 51. Garcia JP, Pedroso JC. Ceftriaxone single dose comparative study of cefotetan with piperacillin versus ceftazidime multiple doses in the prophylaxis for prophylaxis against infection in elective if infection in colorectal surgery. Eur Surg Res colorectal surgery. J R Coll Surg Edinb 1989;2:14–18. 1990;35:29–32. 52. Gerner T, Nygaard K, Kaaresen R, Mjoerud J, 65. Hinchey EJ, Richards GK, Lewis R, Echave V, Larsen S, et al. Antibiotic prophylaxis in colorectal Biron JS, Weissglass I. Moxalactam as single-agent surgery. Acta Chir Scand 1989;155:121–4. prophylaxis in the prevention of wound infection following colon surgery. Surgery 1987;101:15–19. 53. Gomez-Alonso A, Lozano F, Perez A, Almazan A, Abdel-Lah A, Cuadrado F. Systematic prophylaxis 66. Hosie KB, Fielding JWL, Alexander-Williams J, with gentamicin-metronidazole in appendicectomy Temple JG, Keighley MRB. Ceftizoxime alone and colorectal surgery: a prospective controlled or in combination with metronidazole as clinical study. Int Surg 1984;69:17–20. prophylaxis in elective colorectal surgery. Drug Invest 1992;4:13–16. 54. Goransson G, Nilsson-Ehle I, Olsson S-A, Petersson BG, Bengmark S. Single versus multiple dose 67. Jagelman DG, Fazio VW, Lavery IC, Weakley FL. doxycycline prophylaxis in elective colorectal A prospective, randomised, double-blind study surgery. Acta Chir Scand 1984;150:245–9. of 10% mannitol mechanical bowel preparation combined with oral neomycin and short-term, 55. Gortz G, Boese-Landgraf J, Hopfenmuller W, perioperative, intravenous Flagyl as prophylaxis Rodloff A, Kotwas J. Ciprofloxacin as single-dose in elective colorectal resections. Surgery antibiotic prophylaxis in colorectal surgery. Diagn 1985;98:861–5. Microbiol Infect Dis 1990;13:181–5. 68. Jagelman DJ, Fabian TC, Nichols RL, Stone HH, 56. Gottrup F, Diederich P, Sorensen K, Nielson SV, Wilson SE, Zellner SR. Single-dose cefotetan versus Ornsholt J, Brandsborg O. Prophylaxis with whole multiple-dose cefoxitin as prophylaxis in colorectal gut irrigation and antimicrobials in colorectal surgery. Am J Surg 1988;155:71–6. surgery: a prospective, randomised double-blind clinical trial. Am J Surg 1985;149:317–22. 69. Jensen LS, Anderson A, Fristrup SC, Holme JB, Hvid HM, Kraglund K, et al. Comparison of 57. Greig J, Morran C, Gunn R, Mason B, Sleigh D, one dose versus three doses of prophylactic McArdle C. Wound sepsis after colorectal surgery: antibiotics and the influence of blood trans- the effect of cefotetan lavage. Chemioterapion fusion, on infectious complications in acute 1987;6:595–6. and elective colorectal surgery. Br J Surg 58. Grundmann R, Burkardt F, Scholl H, Koelschbach 1990;77:513–18. D. One versus three doses of metronidazole/ 70. Jones RN, Wojeski WV. Single-dose cephalosporin mezlocillin for antibiotic prophylaxis in colon prophylaxis of 929 surgical procedures in a prepaid surgery. Chemioterapia 1987;6:604–5. group practice: a prospective, randomised com- 59. Hakansson T, Raahave D, Hansen OH, Pedersen parison of cefoperazone and cefotaxime. Diagn T. Effectiveness of single dose prophylaxis with Microbiol Infect Dis 1987;6:323–34. cefotaxime and metronidazole compared with three doses of cefotaxime alone in elective 71. Jones RN, Wojeski W, Bakke J, Porter C, Searles M. colorectal surgery. Eur J Surg 1993;159:177–80. Antibiotic prophylaxis of 1036 patients undergoing elective surgical procedures: a prospective 60. Hall JC, Watts JM, O’Brien P, Turnbridge J, randomized comparative trial of cephazolin, McDonald P. Single-dose antibiotic prophylaxis cefoxitin and cefotaxime in a prepaid medical in contaminated abdominal surgery. Arch Surg practice. Am J Surg 1987;153:341–6. 1989;124:224–47. 72. Jones RN, Wojeski WV. Single-dose surgical 61. Hall C, Curran F, Burdon DW, Keighley MRB. prophylaxis using ticarcillin/clavulanic acid A randomized trial to compare amoxycillin/ (Timentin): a prospective, randomised com- clavulanate with metronidazole plus gentamicin parison with cefotaxime. Diagn Microbiol Infect in prophylaxis in elective colorectal surgery. Dis 1987;7:219–23. J Antimicrob Chemother 1989;24:195–202. 73. Juul P, Merrild U, Kronborg O. Topical ampicillin 62. Hall JC, Hall JL, Christiansen K. A comparison in addition to a systematic antibiotic prophylaxis of the roles of cefamandole and ceftriaxone in in elective colorectal surgery: a prospective abdominal surgery. Arch Surg 1991;126:512–16. randomised study. Dis Colon Rectum 1985;28:804–6. 51 References

74. Juul P, Klaaborg KE. Single or multiple doses of 87. Lohde E, Scholz L, Gemperle A, Langmark H, metronidazole and ampicillin in elective colorectal Hopfenmuller W, Abri O, et al. Comparative surgery: a randomised trial. Dis Colon Rectum analysis of mezlocillin/metronidazole and 1987;30:526–8. amoxicillin/clavulanic acid as ‘one-shot’ antibiotic prophylaxis in colorectal surgery 75. Karran SJ, Sutton G, Gartell P, Karran SE, Finnis D, [German]. Zentralbl Chir 1992;117:325–30. Blenkinsop J. Imipenem prophylaxis in elective colorectal surgery. Br J Surg 1993;80:1196–8. 88. Lohr J, Wagner PK, Rothmund M. Perioperative antibioticaprophylaxe (Einmal-oder Mehrfachgabe) 76. Khubchandani IT, Karamchandani MC, Sheets bei elektiven colorectalen eingriffen [German]. JA, Stasik JJ, Rosen L, Riether RD. Metronidazole Chirurg 1984;55:512–14. versus erythromycin, neomycin and cefazolin in prophylaxis for colonic surgery. Dis Colon Rectum 89. Lozano F, Alonso AG, Almazan A, Garcia JG, 1989;32:17–20. Cuadrado F, Moran MR. A comparison of three different prophylactic parenteral antibiotic 77. Kingston RD, Kiff RS, Duthie JS, Walsh S, Spicer regimens of colorectal surgery: a prospective A, Jeacock J. Comparison of two prophylactic study. Int Surgery 1985;70:227–31. single-dose intravenous antibiotic regimens in the treatment of patients undergoing elective 90. Luke M, Iversen J, Sondergaard J, Kvist E, Lund P, colorectal surgery in a district general hospital. Andersen F, et al. Ceftriaxone versus ampicillin J R Coll Surg Edinb 1989;34:208–11. and metronidazole as prophylaxis against infections after clean-contaminated abdominal surgery. 78. Kling PA, Homlund D, Burman L. Single-dose Eur J Surg 1991;157:45–9. intravenous metronidazole v. doxycycline prophylaxis in colorectal surgery. Acta Chir 91. Lumley JW, Siu SK, Pillay SP, Stitz R, Kemp RJ, Scand 1985;151:163–8. Faoagali J, et al. Single dose ceftriaxone as prophylaxis for sepsis in colorectal surgery. 79. Kling P-A, Burman LG. Failure of single-dose Aust N Z J Surg 1992;62:292–6. metronidazole prophylaxis in colorectal surgery. Acta Chir Scand 1988;154:305–9. 92. Matikainen M, Hiltunen KM. Parenteral single dose ceftriaxone with tinidazole versus aminoglycoside 80. Kling P-A, Dahlgren S. Oral prophylaxis with with tinidazole in colorectal surgery: a prospective neomycin and erythromycin in colorectal surgery. single-blind randomized multicentre study. Int J Arch Surg 1989;124:705–7. Colorect Dis 1993;8:148–50. 81. Kow L, Toouli J, Brookman J, McDonald PJ. 93. McArdle CS, Moran CG, Pettit L, Gemmell CG, Comparison of cefotaxime plus metronidazole Sleigh JD, Tillotson GS. Value of oral antibiotic versus cefoxitin for prevention of wound prophylaxis in colorectal surgery. Br J Surg infection after abdominal surgery. World J 1995;82:1046–8. Surg 1995;19:680–6. 94. McCulloch PG, Blamey SL, Finlay IG, Baird A, 82. Kwok. Amoxycillin and clavulanic acid versus Sleigh D, Gardner E, et al. A prospective com- cefotaxime and metronidazole as antibiotic parison of gentamicin and metronidazole and prophylaxis in elective colorectal resectional moxalactam in the prevention of septic compli- surgery. Chemotherapy 1993;39:135–9. cations associated with elective operations of the colon and rectum. Surg Gynecol Obstet 83. Lau WY, Chu KW, Poon GP, Ho K. Prophylactic 1986;162:521–4. antibiotics in elective colorectal surgery. Br J Surg 1988;75:782–5. 95. University of Melbourne Colorectal Group. Clinical trial of prophylaxis of wound sepsis in 84. Lauridsen F, Bjoernsen K, Nielsen SAD, Hansen elective colorectal surgery comparing ticarcillin OH. Short-term prophylaxis with cefotaxime in with tinidazole. Aust N Z J Surg 1986;56:209–13. colorectal surgery: a prospective, randomized trial. Dis Colon Rectum 1988;31:25–7. 96. University of Melbourne Colorectal Group. Systematic Timentin is superior to oral tinidazole 85. Lewis RT, Goodall RG, Marien B, Lloyd-Smith W, for antibiotic prophylaxis in elective colorectal Park M, Wiegand FM. Is neomycin necessary surgery. Dis Colon Rectum 1987;30:786–9. for bowel preparation in surgery of the colon? Oral neomycin plus erythromycin versus 97. University of Melbourne Colorectal Group. erythromycin-metronidazole. Can J Surg A comparison of single-dose systematic Timentin 1989;32:265–70. with mezlocillin for prophylaxis of wound infection in elective colorectal surgery. Dis Colon Rectum 86. Lindhagen J, Andaker L, Hojer H. Comparison 1989;32:940–3. of systematic prophylaxis with metronidazole/ placebo and metronidazole/fosfomycin in 98. Mendel V, Jung D, Heymann H. Single-shot colorectal surgery. Acta Chir Scand antibiotic prophylaxis in colon surgery. 52 1984;150:317–23. Chemioterapia 1987;6:597–600. Health Technology Assessment 1998; Vol. 2: No. 7

99. Menzel J, Bauer J, Pritzbuer E, Klempa I. 110. Nel CJC. Ceftriaxone as prophylaxis for elective Periperative anwendung von ampicillin/sulbactam, abdominal and colorectal surgery. S Afr J Surg cefoxitin und piperacillin/metronidazol in der 1989;(suppl):6. elektiven colon- und rectumchirurgie [German]. 111. Nohr M, Andersen JC, Juul-Jensen KE. Prophylactic Chirurg 1993;64:649–52. single-dose fosfomycin and metronidazole com- 100. Menzies D, Gilbert JM, Shepherd MJ, Rogers TR. pared with neomycin, bacitracin, metronidazole A comparison between amoxycillin/clavulanate and ampicillin in elective colorectal operations. and mezlocillin in abdominal surgical prophylaxis. Acta Chir Scand 1990;156:223–30. J Antimicrob Chemother 1989;24:203–8. 112. The Norwegian Study Group for Colorectal Surgery. Should antimicrobial prophylaxis in 101. Mittermayer H, Gross C, Brucke P. Single dose colorectal surgery include agents effective against cefuroxime/metronidazole versus metronidazole both anaerobic and aerobic microorganisms? alone in elective colorectal surgery. Am Surg A double-blind, multicenter study. Surgery 1984;50:418–23. 1985;97:402–7. 102. Moesgaard F, Nielsen ML. Failure of topically 113. Nyam DCNK, Yeo M, Cheong D, Goth HS. applied antibiotics, added to systematic prophylaxis, Antibiotic prophylaxis in colorectal surgery: a to reduce perineal wound infection in abdomino- randomised, double-blind, controlled trial of perineal excision of the rectum. Acta Chir Scand amoxycillin-clavulanic acid versus ceftriaxone and 1988;154:589–92. metronidazole. Asian J Surg 1995;18:227–30. 103. Moesgaard F, Lykkegaard-Nielsen M. Preoperative 114. Offer C, Weuta H, Bodner E. Efficacy of cell-meditated immunity and duration of anti- preoperative with ciprofloxacin or cefazolin in biotic prophylaxis in relation to postoperative colorectal surgery. Infection 1988;16:46–7. infectious complications. Acta Chir Scand 115. Pacelli F, Brisinda G, Bellantone R, Doglietto 1989;155:281–6. GB, Crucitti F. Single dose imipenem-cilastatin 104. Moesgaard F, Nielsen ML, Hjortrup A, Kjersgaard compared with three doses of cefuroxime and P, Sorensen C, Larsen PN, et al. Intraincisional metronidazole as prophylaxis in elective antibiotic in addition to systematic antibiotic colorectal surgery: a prospective randomized treatment fails to reduce wound infection rates study. J Chemother 1991;3:372–5. in contaminated abdominal surgery. Dis Colon 116. Palmer BV, Mannur KR, Ross WB. An observer Rectum 1989;32:36–8. blind trial of co-amoxiclav versus cefuroxime plus metronidazole in the prevention of postoperative 105. Morris DL, Fabricius PJ, Ambrose NS, Scammell B, wound infection after general surgery. J Hosp Infect Burdon DW, Keighley MRB. et al. A high incidence 1994;26:287–92. of bleeding is observed in a trial to determine whether addition of metronidazole is needed with 117. Periti P, Mazzei T, Tonelli F. Single-dose cefotetan latamoxef for prophylaxis in colorectal surgery. versus multiple-dose cefoxitin antimicrobial J Hosp Infect 1984;5:398–408. prophylaxis in colorectal surgery: results of a prospective, multicenter, randomised study. 106. Morris DL, Wilson SR, Pain J, Edwardson KF, Dis Colon Rectum 1989;32:121–7. Jones J, Strachan C, et al. A comparison of aztreonam/metronidazole and cefotaxime/ 118. Periti P, Tonelli F, Mazzei T, Ficari F, and the metronidazole in elective colorectal surgery: Italian Study Group on Antimicrobial Prophylaxis antimicrobial prophylaxis must include in Abdominal Surgery. Antimicrobial chemo- Gram-positive cover. J Antimicrob Chemother immunoprophylaxis in colorectal surgery with 1990;25:673–8. cefotetan and thymostimulin: prospective controlled multicenter study. J Chemother 107. Morton AL, Taylor EW, Lindsay G, Wells GR. 1993;5:37–42. A multicenter study to compare cefotetan alone 119. Perrott CAV, Hinder RA, Cassel R, Koornhof HJ, with cefotetan and metronidazole as prophylaxis Naude G, Kleinman M, et al. Prophylactic against infection in elective colorectal operations. antimicrobials in elective colorectal and biliary Surg Gynecol Obstet 1989;169:41–5. surgery. S Afr Med J 1985;68:387–91. 108. Mosimann F, Chamero J. Preventive preoperative 120. Peruzzo L, Savio S, Lalla FD. Systematic versus antibiotic therapy in elective colon surgery. systematic plus oral chemoprophylaxis in elective A controlled prospective randomized study. colorectal surgery. Chemioterapia 1987;6:601–2. Schweiz Med Wochenschr 1987;117:570–3. 121. Petrelli NJ, Conte CC, Herrera L, Stulc J, 109. Mozzillo N, Dionigi R, Ventriglia L. Multicenter O’Neill P. A prospective trial of preoperative study of aztreonam in the prophylaxis of colorectal, prophylactic cefamandole in elective colorectal gynecologic and urologic surgery. Azttre: New Dev surgery for malignancy. Dis Colon Rectum Curr Perspect Chemother 1989;35:58–71. 1988;31:427–9. 53 References

122. Petropoulos P, Dietrich PY, Ammann J, Ayer G, 135. Rowe-Jones DC, Peel ALG, Kingston RD, Shaw Buchmann P, Martinoli S. Prophylaxie antibiotique JFL, Teasdale C, Cole DS. Single dose cefotaxime par dose unique pour la chirurgie colo-rectale plus metronidazole versus three dose cefuroxime elective [French]. Helv Chir Acta 1985;52:703–6. plus metronidazole as prophylaxis against wound infection in colorectal surgery: multi- 123. Playforth MJ, Smith GMR, Evans M, Pollock centre prospective randomised study. BMJ AV. Single-dose intravenous antibiotics for the 1990;300:18–22. prophylaxis of abdominal surgical wound infection: a trial of amoxycillin/clavulanate 136. Sauven P, Playforth MJ, Smith GMR, Evans M, against latamoxef. Surg Res Comm 1987;1:173–80. Pollock AV. Single-dose antibiotic prophylaxis of abdominal surgical wound infection: a trial of 124. Playforth MJ, Smith GMR, Evans M, Pollock AV. preoperative latamoxef against preoperative Antimicrobial bowel preparation: oral, parenteral tetracycline lavage. J R Soc Med 1986;79:137–41. or both? Dis Colon Rectum 1988;31:90–3. 137. Schiessel R, Huk I, Wunderlich M, Rotter M, 125. Plouffe JF, Perkins RL, Fass RJ, Carey LC, Wewalka G, Schemper M. Postoperative infections Macynski ME. et al. Comparison of the effec- in colonic surgery after enteral bacitracin-neomycin- tiveness of moxalactam and cefazolin in the clindamycin or parenteral mezlocillin-oxacillin prevention of infection in patients undergoing prophylaxis. J Hosp Infect 1984;5:289–97. abdominal operations. Diagn Microbiol Infect Dis 1985;3:25–31. 138. Schoetz DJ, Roberts PL, Murray JJ, Coller JA, 126. Plouffe JF. Cefmetazole versus cefoxitin in Veidenheimer MC. Addition of parenteral cefoxitin prevention of infections after abdominal surgery. to regimen of oral antibiotics for elective colorectal J Antimicrob Chemother 1989;23:85–8. operations. Ann Surg 1990;212:209–12. 127. Pollock AV, Evans M, Smith GMR. Preincisional 139. Shatney CH. Antibiotic prophylaxis in elective intraparietal Augmentin in abdominal operations. gastro-intestinal tract surgery: a comparison Ann R Coll Surg Engl 1989;71:97–100. of single-dose pre-operative cefotaxime and multiple-dose cefoxitin. J Antimicrob Chemother 128. Raahave D, Hesselfeldt P, Pedersen TB. Cefotaxime 1984;14:241–5. i.v. versus oral neomycin-erythromycin for prophy- laxis of infections after colorectal operations. 140. Skipper D, Karran SJ. A randomized prospective World J Surg 1988;12:369–73. study to compare cefotetan with cefuroxime plus metronidazole as prophylaxis in elective colorectal 129. Raahave D, Hesselfeldt P, Pedersen T, Zachariassen surgery. J Hosp Infect 1992;21:72–7. A, Kann D, Hansen OH. No effect of topical ampicillin prophylaxis in elective operations 141. Stellato TA, Danziger LH, Gordon N, Hau T, Hull of the colon or rectum. Surg Gynecol Obstet CC, Zollinger RM, et al. Antibiotics in elective colon 1989;168:112–14. surgery. Am Surg 1990;56:251–4. 130. Rangabashyam N, Rathnasami A. Prophylaxis of 142. Stewart M, Taylor EW, Lindsay G, and the West of infection following colorectal surgery. Infection Scotland Surgical Infection Study Group. Infection 1991;19:459–61. after colorectal surgery: a randomised trial of prophylaxis with piperacillin versus sulbactam/ 131. Reynolds JR, Jones JA, Evans DF, Hardcastle JD. Do piperacillin. J Hosp Infect 1995;29:135–42. preoperative oral antibiotics influence sepsis rates following elective colorectal surgery in patients 143. Stubbs RS, Griggs NJ, Kelleher JP, Dickinson IK, receiving perioperative intravenous prophylaxis? Moat N, Rimmer DMD. Single dose mezlocillin Surg Res Comm 1989;7:71–7. versus three dose cefuroxime plus metronidazole for the prophylaxis of wound infection after large 132. Rodolico G, Puelo S, Blandono G, Russello D, bowel surgery. J Hosp Infect 1987;9:285–90. Amoedo C, Latteri F, et al. Colorectal surgery: short-term prophylaxis with clindamycin plus 144. Taylor EW, Lindsay G, and the West of Scotland aztreonam or gentamicin. Rev Infect Dis Surgical Infection Study Group. Selective 1991;13:612–15. decontamination of the colon before elective colorectal surgery. World J Surg 1994;18:926–32. 133. Roland M. Prophylactic regimens in colorectal surgery: an open, randomised, consecutive trial 145. Tehan S, Whittaker J. A multi-centre double-blind on metronidazole used alone or in combination prospective study comparing the efficacy and with ampicillin or doxycycline. World J Surg tolerance of Augmentin with the combination of 1986;10:1003–8. cephradine plus metronidazole as surgical prophylaxis. Surg Res Comm 1989;6:97–105. 134. Rorbaek-Madsen M, Toftgaard C, Graversen HP, Kristiansen JD, Lauesen N, Randberg FA, et al. 146. Thomas WEG, Cooper MJ, Holt A, Reeves D. Cefoxitin for one day versus ampicillin and Latamoxef: single agent prophylaxis in metronidazole for three days in elective colorectal colorectal surgery. J Antimicrob Chemother 54 surgery. Dis Colon Rectum 1988;31:774–7. 1985;16:121–8. Health Technology Assessment 1998; Vol. 2: No. 7

147. Tsimoyiannis EC, Paizis JB, Kabbani K, Lekkas 159. Burke J. Current perspective of surgical infection. ET, Floras GA, Boulis SA. Short-term antibiotic In: Watts JM, McDonaldson PJ, O’Brien PE, prophylaxis in elective colorectal surgery. Marshall VR, Finlay-Jones JJ, editors. Infection in Chemotherapy 1991;91:66–9. surgery: basic and clinical aspects. Edinburgh: Churchill Livingstone, 1981:14–26. 148. Tuchmann V, Breyer S, Ganzinger U. Antibiotic prophylaxis in colorectal surgery, short-term vs one- 160. Gorbach SL. Antimicrobial prophylaxis for shot prophylaxis – a multicentre study [German]. appendectomy and colorectal surgery. Rev Fortchr Med 1988;106:537–60. Infect Dis 1991;13:S815–S819. 149. Tudor RG, Haynes I, Youngs DJ, Burdon DW, 161. Evans M, Pollock AV. The inadequacy of published Keighley MRB. Comparison of short-term antibiotic random control trials of antibacterial prophylaxis cover with a third-generation cephalosporin against in colorectal surgery. Dis Colon Rectum conventional five-day therapy using metronidazole 1987;30:743–6. with an aminoglycoside in emergency and complicated colorectal surgery. Dis Colon 162. Moher D, Jadad AR, Tugwell R. Assessing the quality Rectum 1988;31:28–32. of randomised controlled trials. Int J Technol Assess Health Care 1996;12:195–208. 150. Utley RJ, Macbeth WAAG. Preoperative cefoxitin: a double-blind prospective study in the prevention 163. Chalmers TC, Celano P, Sacks HS, Smith H. Bias in of wound infection. J R Coll Surg Edinb treatment assignment in controlled clinical trials. 1984;29:143–6. New Engl J Med 1983;309:1358–61. 151. Walker AJ, Taylor EW, Lindsay G, Dewar EP, and 164. Schulz KF, Chalmers I, Hayes RJ, Altman D. the West of Scotland Surgical Infection Study Empirical evidence of bias. JAMA 1995;273:408–12. Group. A multicentre study to compare pipera- 165. Garner JS, Jarvis WR, Emori TG, Horan TC, cillin with the combination of netilmicin and Hughes JM. CDC definitions for nosocomial metronidazole for prophylaxis in elective colorectal infections. Am J Infect Control 1988;16:128–40. surgery undertaken in district general hospitals. J Hosp Infect 1988;11:340–8. 166. Hyryla MLJ, Sintonen H. The use of the health services in the management of wound infection. 152. Weaver M, Burdon DW, Youngs DJ, Keighley MRB. J Hosp Infect 1994;26:1–14. Oral neomycin and erythromycin compared with single-dose systematic metronidazole and ceftri- 167. Altman DG. Practical statistics for medical axone prophylaxis in elective colorectal surgery. research. London: Chapman and Hall, Am J Surg 1986;15:437–41. 1991:455–60. 153. Weidema WF, Van den Boogaard AEJM, Wesdorp 168. Classen DC, Evans S, Pestotnik SL, Horn SD, RIC, Van Boven CPA, Greep JM. 24-Hour systematic Menlove RL, Burke JP. The timing of prophylactic prophylaxis with gentamicin and metronidazole, or administration of antibiotics and the risk of metronidazole alone, in elective colorectal surgery surgical wound infection. New Engl J Med after mechanical bowel preparation with mannitol 1992;326:281–6. and whole gut irrigation. Acta Chir Belg 1985;85:349–53. 169. Danielsen S, Midtvedt T, Giercksky KE. Preventive antibiotics in elective colorectal surgery. Nord Med 154. Wenzel M, Heinrich M, Schmidt C. Peri-operative 1989;104:247–55. infection prophylaxis with ornidazole and gentami- cin in elective colonic surgery. Pharmatherapeutica 170. Nichols RL. Surgical infections: prevention and 1985;4:351–5. treatment. Am J Surg 1996;172:68–74. 155. Wohlfahrt R, Siedek M. Perioperative infection 171. Gold HS, Moellering RC. Antimicrobial drug prophylaxis in colon surgery. Chemioterapia resistance. New Engl J Med 1996;335:1445–54. 1987;6:603–5. 172. Office of Technology Assessment. Impact 156. Zuber M, Durig M, Neff U, Laffer U. Antibiotic of antibiotic-resistant bacteria. OTA-H-629. prophylaxis in colorectal surgery: cefazolin- Washington DC: US Government Printing ornidazole versus cefazolin-placebo [German]. Office, 1995. Helv Chir Acta 1989;56:211–15. 173. Norrby SR. Cost effective prophylaxis of surgical 157. Culver DH, Horan TC, Gaynes RP, Martone WJ, infections. Pharmacoeconomics 1996;10:129–39. Jarvis WR, Emori TG, et al. Surgical wound infection 174. Davey P, Lynch B, Malek M, Byrne D, Thomas P. rates by wound class, operative procedure and Cost-effectiveness of single dose cefotaxime patient risk index. Am J Med 1991;91:152S–157S. plus metronidazole compared with three doses 158. Vegas AA, Jodra VM, Garcia ML. Nosocomial each of cefuroxime plus metronidazole for infection in surgery wards: a controlled study of the prevention of wound infection after increased duration of hospital stays and direct cost colorectal surgery. J Antimicrob Chemother of hospitalisation. Eur J Epidemiol 1993;9:504–10. 1992;30:855–64. 55 References

175. Eisenberg J. Clinical economics. A guide to 182. Lim VK, Cheong YM, Suleiman AB. The use of the economic analysis of clinical practices. surgical antibiotic prophylaxis in seven Malaysian JAMA 1989;262:2879–86. hospitals. Southeast Asian J Trop Med Public Health 1994;25:698–701. 176. Fry DE. Antibiotics in surgery. An overview. Am J Surg 1988;155:11–15. 183. Delgadillo L, Ramirez R, Cebrecos J, Arnau JM, Laporte JR. The use of antibiotics in surgical 177. Scalley RD, Irwin AD, Poduska PJ, Wolff AJ, prophylaxis. The characteristics and consequences. Cochran RS. Surgical antibiotic prophylaxis, patient Med Clin (Barc) 1993;100:404–6. morbidity, and cost reduction: a three year study. Drug Intell Clin Pharm 1987;21:648–52. 184. Kappstein I, Daschner FD. Use of perioperative antibiotic prophylaxis in selected surgical 178. Scher KS, Bernstein JM, Arenstein GL, Sorensen C. procedures – results of a survey in 889 surgical Reducing the cost of surgical prophylaxis. Am J Surg departments in German hospitals. Infection 1990;56:32–5. 1991;19:391–4.

179. Dobrzanski S, Lawley DI, McDermott I, Selby M, 185. Davey P, Napier A, Barrie H, Dodd T, McMillan J, Ausobsky JR. The impact of guidelines on peri- Ruta D. Audit of surgical prophylaxis in three operative antibiotic administration. J Clin Pharm hospital trusts in Tayside. 1997, unpublished. Ther 1991;16:19–24. 186. Janknegt R, Wijnands WJA, Stobberingh E. 180. Evans RS, Pestotnik SL, Burke JP, Gardner RM, Antimicrobial prophylaxis in bowel surgery Larsen RA, Classen DC. Reducing the duration in The Netherlands. Eur J Clin Microbial Infect of prophylactic antibiotic use through computer 1994;13:596–600. monitoring of surgical patients. Drug Intell Clin 187. Widdison AL, Pope NRJ, Brown EM. Survey of Pharm 1990;24:351–4. guidelines for antimicrobial prophylaxis in surgery. J Hosp Infect 1993;25:199–205. 181. Davey PG, Vacani P, Parker SE, Malek MM. Assessing cost effectiveness of antimicrobial 188. Bloxham CA. Towards evidence-based antibiotic treatment: monotherapy compared with prescribing: a national survey of antibiotic policies combination therapy. Eur J Surg 1994;573:67–72. [dissertation]. Durham: University of Durham; 1996.

56 Health Technology Assessment 1998; Vol. 2: No. 7

Appendix 1 Medline search strategies

Set Search 047 32 or 40 or 46 048 26 and 47 001 Surgical Wound Infection/ 049 (animal not (human and animal)).sh. 002 Postoperative Complications/dt,ec,pc,su 050 exp HIV Infections/ 003 exp preoperative care/ 051 bone marrow transplantation/ 004 exp intraoperative care/ 052 or/49-51 005 Premedication/ 053 48 not 52 006 premedication.tw. 054 [exp surgery, digestive system (non mesh)/] 007 or/2-6 055 exp digestive system/su 008 Bacterial Infections/dt,ec,pc,su 056 abdomen/su 009 Infection/dt,ec,pc,su 057 “colon and rectal surgery (specialty)”/ 010 Sepsis/dt,ec,pc,su 058 anastomosis, surgical/ 011 (bacteri$ adj2 infect$).tw. 059 appendectomy.tw. 012 (wound$ adj2 infect$).tw. 060 cholecystectomy.tw. 013 (wound$ adj2 contamin$).tw. 061 choledochostomy.tw. 014 sepsis.tw. 062 portoenterostomy.tw. 015 or/8-14 063 sphincterotomy.tw. 016 exp Anti-Infective Agents/ 064 colectomy.tw. 017 Antibiotics/tu 065 protocolectomy.tw. 018 antibiotic$.tw. 066 enterostomy.tw. 019 antimicro$.tw. 067 cecostomy.tw. 020 (antimicro$ adj3 prophyla$).tw. 068 colostomy.tw. 021 (anti$ adj2 infect$).tw. 069 duodenostomy.tw. 022 or/16-21 070 ileostomy.tw. 023 15 or 22 071 jejunostomy.tw. 024 7 and 23 072 esophagectomy.tw. 025 1 and 22 073 esophagoplasty.tw. 026 24 or 25 074 esophagostomy.tw. 027 randomized controlled trial.pt. 075 gastrectomy.tw. 028 (randomized controlled trial$ or 076 gastroenterostomy.tw. randomised controlled trial 077 hepatectomy.tw. 029 random allocation/ 078 fundoplication.tw. 030 double-blind method/ 079 peritoneovenous shunt$.tw. 031 (single adj blind adj method).tw. 080 (transplant$ and (liver or pancreas)).tw. 032 or/27-31 081 or/54-80 033 clinical trial.pt. 082 surg$.tw. 034 exp clinical trials/ 083 (digesti$ adj2 system$).tw. 035 (clinical$ adj5 trial$).tw. 084 biliary$.tw. 036 ((singl$ or doubl$ or trebl$ or tripl$) adj5 085 bilio$.tw. (blind$ or mask$)) 086 (colon$ or rectum$ or rectal$ or cecum$).tw. 037 placebos/ 087 gastric$.tw. 038 (placebo$ or random$).tw. 088 gastro$.tw. 039 research design/ 089 jejunoileal.tw. 040 or/33-39 090 esophag$.tw. 041 comparative study/ 091 pancrea$.tw. 042 exp evaluation studies/ 092 anastomosis$.tw. 043 follow-up studies/ 093 intestin$.tw. 044 prospective studies/ 094 abdomen$.tw. 045 (control$ or prospectiv$ or volunteer$).tw. 095 oesophageal$.tw. 046 or/41-45 096 (stomach$ or liver$).tw. 57 Appendix 1

097 or/83-96 098 82 and 97 099 81 or 98 100 53 and 99 101 100

58 Health Technology Assessment 1998; Vol. 2: No. 7

Appendix 2 Data extraction sheet and field definitions

Field Data to be entered

Reviewers Initials of reviewer carrying out data extraction, plus those of reviewer checking Refs First author’s surname, publication year, e.g. Ahmed, 1987 Copies State how the article was located, e.g. name the database, handsearching. If an article is to have more than one record, only the first record should list the method of location, for any further records enter Repeat. Author All authors Title Article title Journal Full journal title PYear Publication year Volume Journal volume Chapter Chapter Page Page numbers Place_of_publication Type of publication (journal article, book, unpublished) Country Country of origin Aim_of_study Enter the author’s aim of the study Surgical_procedure Enter colorectal, appendicectomy or biliary Surgery Give further details of the procedures used, e.g. elective, emergency Bowel_prep Description of bowel preparation procedure Duration_of _operation Enter the average (preferably mean) length of operation in minutes Study_design Describe the method of randomisation, outcome measurement, etc. Randomisation Code for truly randomised (TR), pseudo randomised (PR) or not stated (NS) Blinding_assessor Was the person measuring the outcome blind to the antibiotic used? Enter Yes, No or Unsure Inclusion_exclusion Enter the inclusion and exclusion criteria used for entering subjects to the trial Interventions Give a clear description of the regimens used in each group Purpose State the purpose, e.g. single vs. multiple dose same antibiotic, different doses different antibiotics, same doses different antibiotics, different doses extra antibiotics other description

59 Appendix 2

Antibiotic_A Antibiotic_B State only the names of the antibiotic used in each regimen, e.g. Antibiotic_C ampicillin + sulbactam; gentamicin + metronizadole Antibiotic_D }

Dose_A Dose_B State the dosage used for each treatment group, e.g. Dose_C 100 mg × 5; 1.5 g × 1 + 750 mg × 2; (1.5 mg/kg + 500 mg) × 3 Dose_D }

Time_A Time_B Enter preop, periop or postop Time_C Time_D }

Route_A Route_B Enter the route of administration, e.g. i.v., p.o. Route_C Route_D }

Duration_A Duration_B Enter the duration of administration in hours Duration_C Duration_D }

Number_A Number_B State the number of patients evaluated for SWI in each treatment group Number_C Number_D }

Blood_tissue_conc Give details of any blood and tissue concentrations reported Definition_SWI Enter the authors definition of SWI Other_outcomes Enter other outcomes measured and their definitions if stated Dur_of_follow-up Enter duration of follow-up postoperatively in days Age Enter the average age of all the subjects undergoing surgery (colorectal, biliary or appendicectomy) Sex Enter total male to female (M/F) ratio for the subjects Underlying_disease Give details on the underlying diseases of the subjects as a whole Other_char Give details of any other characteristics reported 60 Comparable State whether the groups were comparable or not Health Technology Assessment 1998; Vol. 2: No. 7

SWI_A SWI_B Give the total number of patients with SWIs SWI_C SWI_D } p_value Enter p-value if available Adverse_events Describe adverse events for each treatment group Other_results Give results of other outcomes measured including different subgroups of SWIs and other infections, hospital stay, mortality. Enter total number of patients with any infection. Withdrawal_rate Give the numbers withdrawing after randomisation Withdrawals Give reasons and group for withdrawals Cost State whether cost was discussed in the article, Yes or No Cost_effect Give details of costs reported (costs of drugs, or cost implications, e.g. nursing workload, hospital stay) Conclusion_au State author’s conclusions Review_comments Enter any comments on the trial design, statistics used, etc.

Extra1 Extra2

If a paper examines more than one type of surgery, the data should be extracted for colorectal only. If relevant data are not available for a particular field enter NA. If, however, a field is not applicable for an article mark it with ‘~’.

61

Health Technology Assessment 1998; Vol. 2: No. 7

Appendix 3 RCTs of antimicrobial prophylaxis in colorectal surgery (1984–95)

63 Appendix 3 continued scess: 1/81 colorectal surgery. colorectal ): 25/259 vs. 27/258. Death: 5/259 vs. 5/258. Mild infection: 21/81 vs. 11/76. infection, Severe Deep surgical sepsis: 4/29 vs. 2/19. No intraabdominal abscess in either group. Median Prophylactic ampicillin/sulbactam is effective in reducing in reducing ampicillin/sulbactam is effective Prophylactic No significant differences in infection rates were seen in rates were in infection No significant differences Fosfomycin plus metronidazole is as well tolerated and effective tolerated and effective is as well plus metronidazole Fosfomycin Doxycycline was significantly more effective in preventing in preventing effective more was significantly Doxycycline Conclusion: Conclusion: Conclusion: SWI: 1/29 vs. 2/19. SWI: 21/81 vs. 16/76. SWI: 6/63 vs. 7/65. Authors’ conclusions Conclusion: data rated out from other procedures) out from rated Gentamicin, 1.5 mg/kg body Doxycycline, 400 mg plus Cefuroxime, 1.5 g i.v. at skin plus metronidazole. either triple- or single-dose cefuroxime Doxycycline, 200 mg i.v. at against anaerobes. Cefuroxime, 1.5 g i.v. at skin Metronidazole, 1 g intravenous Ampicillin plus sulbactam, 1.5 g i.v. Fosfomycin, 8 g plus WI:7/259 Abdominal vs. 16/258.WI:3/259 Deep vs. 4/258. Group B: Group Group B: Group morning of the first 2 postoperative days. morning of the first 2 postoperative A: Group B: Group start of anaesthesia. the skin incision, and placebo 8 h later. surgery. colorectal complications after elective infective Group A: Group A: Group A: Group Group B: Group * Cancer patients: NA. incision and metronidazole, 500 mg i.v. at result: event [Adverse NA. Cost information: NA.] Cancer patients: 93%. postoperative preoperatively, two by followed hospital stay: 12 vs. 11 days. Cancer patients: NA. metronidazole, 1 g i.v. half an hour given Septicaemia: 6/259 vs. 3/258. Pneumonia: 13/259 vs. 5/258. Elective colorectal surgery. colorectal Elective Cancer patients: 84%. 40 min at over given administration (i.v.) abdominal surgery. Elective purulent lesion: 18/81 vs. 11/76. Suspected abscess: 2/81 vs. 0/76. Ab surgery. colorectal Elective bacteriological culture. surgery. colorectal Elective 8 h. doses every postoperative from the wound;from secretion or serous yielding pathogenic bacteria on culture. in the and 100 mg given premedication * 15 16 13 14 Indicates data for all surgical procedures included in a trial (i.e where data for elective colorectal operation cannot be sepa included operation elective colorectal in a trial all surgical procedures data for (i.e where Indicates data for Randomisation: Not clear.Blind outcome assessment: No. Age: NA. anaesthesia, doses of both with repeated antibiotics after 8 h and 16 h. Randomisation:True.Blind outcome assessment: No. Follow-up: 42 days. Age: 63 (22–92). SWI: of pus or purulent discharge Presence 8 h. doses of 750 mg every result: event [Adverse Yes.Cost information: Yes.] Withdrawal: 10%. in the wound, cellulitis or or marked plus metronidazole, weight 500 mg (WI) following infection the risk of wound 1-centre. discharge with positive serous i.v. the operation, before and two Follow-up: 30 days.Withdrawal: 4%. SWI: Discharge of pus. jejunal, patients undergoing operation who received ileal or colorectal Randomisation: Not clear. Age: [SD] 14). deviation 67 (standard skin incision, before and a further 8 g Urinary (UTI tract infection Andaker, 1992 Blind outcome assessment: No.Follow-up: 28 days.Withdrawal: 7.5%. 8-centres. SWI: Purulent discharge. 8 h later. of fosfomycin result: event [Adverse Yes.Cost information: NA.] metronidazole, 1 g i.v. half an hour before of in the prevention plus metronidazole as the combination of doxycycline Aberg, 1984 Study and quality Surgery and definition of SWI Antibiotic regimensA vs. Results (Group B vs. Group C vs. Group D) Group * Note: calculation, in the table is a rough (in years) presented The age in any one trial, all groups across based on the available NA = not available or applicable Aberg, 1991 1-centre.AhChong, 1994 incision, plus metronidazole, 500 mg at Randomisation:True.Blind outcome assessment: Yes.Follow-up: 10–15 days.Withdrawal: 21%. Age: 66. SWI: with and reddening Swelling > 38 °C and no other temperature the patients received thereafter 2 days. 8 h for 500 mg every premedication. and in the evening 500 mg was given dose of A postoperative vs. 5/76. Bacteraemia: 2/81 vs. 0/76. result: event [Adverse Yes.Cost information: NA.] 64 2-centres. cause; known or purulent secretion complications in general, better protection gave but metronidazole continued

Study and quality Surgery and definition of SWI Antibiotic regimens Results (Group A vs. Group B vs. Group C vs. Group D) Authors’ conclusions

Antonelli, 1985 17 Elective colorectal surgery. Group A: Cefoxitin, 2 g i.v. 30 min Overall SWI: 2/40 vs. 9/37. Abdominal WI: 0/30 vs. 3/30. Cancer patients: NA. preoperatively, and intraoperatively Perineal WI:1/6 vs. 2/4. Peritonitis: 0 vs. 2. Septicaemia: 0 vs. 0. Randomisation:True. Age: 66 (SD 12). (operation over 2 h) and then Anastomotic leak: 0 vs. 1. Dehiscence: 0 vs. 2. Fistula: 1 vs. 1. Blind outcome assessment: No. every 8 h for 72 h. [Adverse event result: NA. Cost information: NA.] Follow-up: 42 days. SWI: Purulent discharge or dehiscence. Withdrawal: NA. Group B: Cephalothin, 2 g i.v. 2 h Conclusion: Cefoxitin used during the perioperative and postoperative 1-centre. preoperatively, 2 g during surgery, and period is a well-tolerated, simple and effective method of reducing

every 6 h postoperatively for 4 days. postoperative sepsis.

Armengaud, 1986 18 (French) Emergency or elective colorectal surgery. Group A: Cefoxitin, 2 g i.v. SWI: 5/30 vs. 2/30 (p = 0.42). Cancer patients: NA from translation. (time not stated). [Adverse event result: Yes.Cost information: NA from translation] Randomisation: Not clear. Age: 65. Blind outcome assessment: Yes. Group B: Piperacillin, 4 g i.v. Conclusion: The two antibiotics have equivalent prophylactic effectiveness Follow-up: 21 days. SWI: Not defined. (time not stated). in colorectal surgery. Withdrawal: 0%.

Arnaud, 199219 Elective colorectal surgery. Group A: Amoxycillin/clavulanic acid, Surgical wound cellulitis: 0/105 vs. 1/103. Surgical wound abscess: Cancer patients: 73%. 2 g/200 mg as a single 30-min infusion 2/105 vs. 4/103. Intraabdominal abscess: 6/105 vs. 3/103. Peritonitis: 1/105 Randomisation:True. Age: 66 (SD 12). at induction of anaesthesia. If operation vs. 1/103. Septicaemia: 1/105 vs. 2/103. Bacteraemia: 1/105 vs. 1/103. Blind outcome assessment: Yes. lasted over 4 h, a perioperative top-up [Adverse event result: NA. Cost information: NA.] Follow-up: 30 days. SWI: Primary infections subclassified as dose (2 g/200 mg) was given during Withdrawal: 6%. minor infections in cases of stitch the 3 h after incision. Conclusion: Co-amoxiclav and cefotetan have very similar efficacy Health TechnologyAssessment 19-centres. abscess or wound abscess, and major when used for the prevention of postoperative infection in elective infections in the case of intraperitoneal Group B: Cefotetan, 2 g as a single 30-min colorectal surgery. abscess, peritonitis, bacteraemia, or infusion on the induction of anaesthesia. septicaemia of intestinal origin. No further dose when operation lasted

over 4 h.

Athanasiadis, 1985 20 (German) Colorectal surgery. Group A: Tinidazole, 160 mg i.v. 80 min SWI: 0/50 vs. 2/50. Perineal WI: 1/50 vs. 4/50. Cancer patients: NA from translation. prior to operation, during anaesthesia. [Adverse event result: Yes.Cost information: NA from translation.] Randomisation: Not clear. Age: 21–86. Blind outcome assessment: No. Group B: Metronidazole, 500 mg i.v. 2 h Conclusion: Prophylactic treatment with tinidazole as a preoperative Follow-up: 6-12 days. SWI: Secondary healing or abscess in the preoperatively and then every 8 h over high and single dose is important for preventing postoperative infection

Withdrawal: 1%. area of the wound, or fever. the next 3 days. in patients undergoing bowel surgery. 7 No. 2: Vol. 1998; 1-centre.

NA = not available or not applicable Note: The age (in years) presented in the table is a rough calculation, across all groups in any one trial, based on the available data

continued 65 Appendix 3 continued preoperative preoperative * /18 vs. 0/22. for the prevention of the prevention for One dose of a suitable i.v. at antibiotic is at least as effective The overall and antibiotic-related complication rates were similar complication rates were and antibiotic-related The overall This study failed to show any advantage to starting antibiotics any This study failed to show The parenteral metronidazole,The parenteral 2 g in divided doses on the day SWI: 23/113 vs. 17/111. Conclusion: Authors’ conclusions SWI: 25/116 vs. 24/101. Conclusion: significantly greater in patients who received a prolonged postoperative course postoperative a prolonged in patients who received greater significantly Conclusion: Conclusion: data rated out from other procedures) out from rated Oral neomycin, 1 g plus WI: 0/18 vs. 0/22. Intraabdominal infection: 0/18 vs. 0/22. after Febrile Placebo. antibiotics. of intravenous Amoxycillin/clavulanic acid Amoxycillin/clavulanic doses. SWI as multiple preventing be a risk of over- may However, there Metronidazole, 500 mg i.v. plus preoperatively. Erythromycin base,Erythromycin 1 g plus to the a satisfactory alternative and cost-effective of operation represents Cefoxitin, 1 g i.v. 6 h every Amoxycillin/clavulanic acid Amoxycillin/clavulanic Metronidazole, 1 g i.v. prior Definite WI: 8/78 vs. 4/55. Possible WI: 5/78 vs. 3/55. Bacteraemia: 3/78 vs. 3/55. Metronidazole, 500 mg i.v. plus Group A: Group and 12 h later. and 12 h later. cephazolin, 1 g i.v. preoperatively, and 6 h A: Group (co-amoxiclav), 250 mg/125 mg i.v. at induction of anaesthesia, and at 8 h and 16 h later. emergency surgery. whelming systemic sepsis in verypatients having elderly B: Group Group B: Group Group A: Group A: Group Both groups: postoperative dose.postoperative in both groups, ileal pouchitis was though the incidence of late postoperative Group B: Group Group B: Group * * Cancer patients: NA. (co-amoxiclav), 250 mg/125 mg i.v. at result: event [Adverse NA. Cost information: NA.] Elective colorectal surgery. colorectal Elective of a purulent discharge).(presence abdominal surgery. Emergency and elective operation. At risk abdominal surgery. surgery. colorectal Elective Cancer patients: 0%. erythromycin, day. 1 g on preoperative 2: day postoperative 5/18 vs. 9/22. white blood cells 2 Elevated Cancer patients: NA. cephazolin, 1 g i.v. intraoperatively, and 6 h result: event [Adverse NA. Cost information: NA.] Cancer patients: 76%. to the operation, further with two doses UTI: 8/78 vs. 5/55. Chest infection: 4/78 vs. 3/55. discharge); infected definitely 2 p.m. and 11 p.m. before on the day operation of the colon. in elective infections postoperative * * 24 21 22 23 Indicates data for all surgical procedures included in a trial (i.e where data for elective colorectal operation cannot be sepa included operation elective colorectal in a trial all surgical procedures data for (i.e where Indicates data for Randomisation:True.Blind outcome assessment: Yes. Age: 55 (SD 21.7). induction of anaesthesia. Auger, 1987 Bates, 1989 Randomisation: Not clear. Age: 54. Randomisation: Not clear.Blind outcome assessment: Yes.Follow-up: 56 days.Withdrawal: 2.5%.1-centre. Age: 33. SWI: Purulent drainage, of regardless or if non-purulent material results culture contained pathogenic bacteria. Cefoxitin, 2 g i.v. operation and before at 6 h and 12 after the initial dose. 5 days, for beginning 6 h after the fixed 9/18 vs. 4/22. nitrogen: or blood urea creatinine Elevated 1 Mean hospital stay: 8.4 vs. 8.7. result: event [Adverse Yes.Cost information: Yes.] Study and quality Surgery and definition of SWI Antibiotic regimensA vs. Results (Group B vs. Group 1-centre. C vs. Group D) Group * Note: calculation, in the table is a rough (in years) presented The age in any one trial, all groups across based on the available Bates, 1992 NA = not available or applicable Blind outcome assessment: Yes. Withdrawal: 9.5%. 1-centre. Becker, 1991 Follow-up: 30 days. SWI: Discharge of pus. Randomisation: Not clear.Blind outcome assessment: No.Follow-up: NA. Age: 62. SWI: (wound infected Possibly at 8 h and of 500 mg being administered result: event [Adverse Yes.Cost information: 16 h after the first dose. NA.] Withdrawal: 24%. Follow-up: 30 days. SWI: Discharge of pus. 3-centre. serous positive draining culture sulphate, neomycin 1 g p.o. at 1 p.m., plus erythromycin administration of oral neomycin 66 Withdrawal: 14%. inflamed without discharge or continued Health Technology Assessment 1998; Vol. 2: No. 7 46. continued al surgery. on.] the short-term dose of 48 h. UTI: 40.9% vs. 18.6%. Intraabdominal abscess: 1/46 vs. 0/46. UTI: 1/46 vs. 5/46. Subphrenic abscess:Subphrenic 1/32 vs. 0/26. UTI: 0/32 vs. 1/26. Cefotetan appears to be as effective as clindamycin plus as clindamycin to be as effective appears Cefotetan Single-dose administration of a long-acting, broad-spectrum The one-shot dose of the antibiotic combination Timentin and metronidazole plus netilmicin are equally equally plus netilmicin are Timentin and metronidazole Authors’ conclusions easy to administer, cheap, relatively events. and does not cause major adverse SWI: 6/46 vs. 3/44. Conclusion: [Adverse event result: event [Adverse NA. Cost information: NA.] Conclusion: SWI:1/32 vs. 0/26. SWI: 3/46 vs. 5/46. Conclusion: effective antianaerobic agent, antianaerobic effective septic rate in elective the postoperative lowers is also plus metronidazole surgery.The combination of doxycycline colorectal Conclusion: [Adverse event result: event [Adverse Yes.Cost information: NA.] data Clindamycin, 600 mg i.v. Doxycycline, 400 mg plus Mezlocillin, 4 g plus against wound the prophylaxis for of mezlocillin plus metronidazole Metronidazole, 500 mg plus Mezlocillin, 4 g plus Cefotetan, 2 g i.v. at the time Doxycycline, 400 mg plus WI:2/135 Superficial vs. 12/126.WI:2/135 Deep vs. 4/126. Septicaemia: acid Ticarcillin/clavulanic Group A: Group 16 h later. Group B: Group Group B: Group Group A: Group A: Group A: Group Group B: Group Group B: Group of induction anaesthesia, then at 8 h and surgery. colorectal septic complications after elective in preventing effective Elective colorectal surgery. colorectal Elective surgery. colorectal Elective surgery. colorectal Elective Cancer patients: NA. further of anaesthesia and two doses Respiratory tract infection: 0/32 vs. 2/26. hospital stay: Postoperative Cancer patients: 62%. metronidazole, 1.5 g i.v. 0.2–6 h 0/135 vs. septic complications:4/135 4/126.Total vs. 20/126. Cancer patients: translation. NA from metronidazole, 500 mg i.v. 20–30 min information: result/cost event [Adverse translati NA from Cancer patients: 34%. (Timentin), 3 g/100 mg i.v. infused at aetiology: unknown Fever 1/46 vs. 1/46.Anastomotic dehiscence: 1/46 vs. 1/ 25 (German) surgery. colorectal Elective 26 27 28 Bellantone, 1988 Bittner, 1989 Bergman, 1987 Study and quality Surgery and definition of SWI1-centre. Antibiotic regimensA vs. Results (Group B vs. Group C vs. Group D) Group micro-organism.1-centre. anaesthesia, further 8 h. two doses every Blair, 1987 colorect in elective against infection in prophylaxis aztreonam and those with abscess or peritonitis.Note: calculation, in the table is a rough (in years) presented The age in any one trial, all groups across 2 days. 8 h for every and repeated based on the available NA = not available or applicable Withdrawal: 11%. of pathogenic without culture plus aztreonam, 1 g i.v. at induction of Randomisation: Not clear.Blind outcome assessment: No. Follow-up: NA. Age: 65. SWI: Purulent discharge with or 12 h. every 14.7 vs. 16.2 days. Randomisation: Not clear.Blind outcome assessment: Yes. Follow-up: 1 month. Age: 62. SWI: culture. Discharge and positive preoperatively. result: event [Adverse Yes.Cost information: NA.] Randomisation: Not clear.Blind outcome assessment: No. Age: NA preoperatively. Withdrawal: 14%.2-centres. placebo. antibiotic, on aerobes, mainly with an effect in combination with an Withdrawal: 0%. with little secretion superficial infections metronidazole, 500 mg i.v. preoperatively, is equivalent to infections Follow-up: > 30 days. SWI: was made between Differentiation Withdrawal: 7%. or peritoneal cavity. surgical wound netilmicin, 80 mg i.v. infused within 15 min Randomisation:True.Blind outcome assessment: Yes. Follow-up: 14 days. Age: 43 (15–86). SWI: the Purulent discharge from induction of anaesthesia. obstruction: Small bowel 1/46 vs. 2/46. 1-centre. 67 continued Appendix 3 continued < 0.05). -life. p serum and tissue levels throughout the entire the entire throughout serum and tissue levels Abscess: 4/59 vs. 4/61. Patients with sepsis: 8/59 vs. 14/61. Remote infections: 3/50 vs. 3/50. doxycycline prophylaxis administered for 5 days. for administered prophylaxis doxycycline Cefoxitin had a similar infection rate to that of the combination had a similar infection Cefoxitin Abdominal wound sepsis occurred in 8.5% of patients the sepsis occurred Abdominal wound A combination of netilmicin plus metronidazole administered administered A combination of netilmicin plus metronidazole Effective metronidazole concentrations were found in serum found concentrations were metronidazole Effective Conclusion: Conclusion: SWI: 4/50 vs. 5/50. SWI: 5/59 vs. 14/61. Conclusion: SWI: 0/8 vs. 0/8. Conclusion: Authors’ conclusions and tissue in both groups but gentamicin was detected only in Group B from B from in Group but gentamicin was detected only and tissue in both groups data Cefoxitin, 2 g i.v. prior to Gentamicin, 120 mg plus ( with 23% in the gentamicin group compared ceftriaxone group Doxycycline, 200 mg i.v. 4 h operation WI in colorectal the incidence of postoperative reduces 1 day for Gentamicin, 80 mg plus tissue specimens and blood samples.The intraoperative short-term Metronidazole, 500 mg plus Ceftriaxone, 2 g plus Gentamicin, 80 mg plus WI:6/30 Abdominal vs. 7/30.WI:3/6 Perineal vs. 10/12. Intraabdominal Gentamicin, 80 mg skin incision, then 2 h and 6 after the of gentamicin plus clindamycin, with but short-term antibiotic prophylaxis Group B: Group anaesthesia and at 6 h 12 later. first dose. systemic antibiotics alone has inadequate efficacy. Group B: Group Group A: Group B: Group A: Group A: Group metronidazole, 400 mg p.o. every the operation, before 2 days 8 h for and gentamicin, by followed 80 mg plus to be rational. appears prophylaxis and metronidazole metronidazole, 500 mg i.v. at induction of i.v. oral medication combined with perioperative preoperative gentamicin Group A: Group B: Group Cancer patients: NA. metronidazole, 1.5 g i.v. at start of result: event [Adverse NA. Cost information: NA.] Elective colorectal surgery. colorectal Elective Cancer patients: 58%. was suspected. when an infection surgery. colorectal Emergency and elective netilmicin, 100 mg i.v. on induction result: event [Adverse Yes.Cost information: NA.] and anorexia. length with fever surgery. colorectal Elective Cancer patients: 72%. clindamycin, 600 mg intramuscular abscess: 2/30 vs. 2/30. Faecal fistula: 0/30 vs. 1/30. Septicaemia: 0/30 vs. 0/30. Cancer patients: 100%. plus metronidazole, 400 mg oral result: event [Adverse NA. Cost information: NA]. 32 30 29 (Chinese) cancer. of colorectal radical resection Elective 31 1-centre. Randomisation: Not clear.Blind outcome assessment: Yes. Age: NA. operation. Randomisation:True.Blind outcome assessment: No. Age: 67 (18–86).Burdon, 1987 Cai, 1992 of anaesthesia and 8 h 16 later. Randomisation:True.Blind outcome assessment: Yes.Follow-up: 28 days.Withdrawal: 0%. Age: 55 (19–82). SWI: Not defined. administration (i.m.), 1 h and 8 h later. preoperatively Death: 0/30 vs. 0/30. result: event [Adverse NA. Cost information: NA.] Randomisation: Not clear.Blind outcome assessment: No. Age: 55 (SD 13). 8 h for every administration (p.o.) operation. before 2 days Brolin, 1986 Study and quality Surgery and definition of SWI Antibiotic regimensA vs. Results (Group B vs. Group C vs. Group D) Group Note: calculation, in the table is a rough (in years) presented The age in any one trial, all groups across based on the available Cainzos, 1986 NA = not available or applicable Follow-up: 21–28 days.1-centre sepsis was graded as minor or major. SWI:Wound opening at least 2 cm in a wound of pus from operation. contamination because of its long half period of potential wound Withdrawal: 10%.1-centre. which a wound from was positive biological culture for and then 100 mg daily preoperatively as at least as well Withdrawal: 2.5%. a surgeon or was opened by opened spontaneously 5 days. sepsis was defined as the discharge Major wound metronidazole, 1.5 g i.v. at start of in high Ceftriaxone results Follow-up: 30 days. SWI: or if a micro- in the wound Pus present 1-centre 68 Follow-up: NA.Withdrawal: 0%. SWI: Not defined. continued Health Technology Assessment 1998; Vol. 2: No. 7 continued on: NA.] gery.The possible reasons for failure failure for gery.The possible reasons th the potential anaerobic pathogens in colonic th the potential anaerobic Primary major WI:5/43Primary major vs. 2/52. Primary minor Deep surgical infections: 2/36 vs. 6/35. Mezlocillin may be an effective sole prophylactic agent in sole prophylactic be an effective Mezlocillin may Although the number of patients in the present study is of patients in the present Although the number sepsis of wound better control produced A treatment Group This study supports the concept that, although metronidazole Conclusion: Deaths: 0/22 vs. 3/21 vs. 2/24 vs. 3/23. mean (SD): Hospital stay 18(7) vs. Conclusion: Conclusion: B,than Group better seemed to afford B treatment but Group Authors’ conclusions [Adverse event result: event [Adverse Yes.Cost information: NA.] SWI:13/43 vs. 6/52. SWI: 1/36 vs. 8/35. Conclusion: data rated out from other procedures) out from rated Metronidazole, 1 g i.v. at Metronidazole, 500 mg i.v. preoper- Metronidazole, 500 mg i.v. doses of i.v. of postoperative indicate that the addition the results Metronidazole, 500 mg i.v. 21(14) vs. 19(9) vs. 19(15). antibiotics. No additional Cefuroxime, 750 mg i.v. at the end complications. other infectious from protection Cefuroxime, 750 mg plus Mezlocillin, 5 g at induction of Metronidazole, 500 mg i.v.WI:7/22 Abdominal vs. 2/21 vs. 4/24 vs. 5/23.WI:1/3 Perineal vs. 2/6 vs. 2/3 Cefuroxime, 750 mg i.m. sepsis: Wound 3/26 vs. 6/26. Other sepsis (peritoneal, urinary, Additional antibiotic of Additional Group A: Group A: Group B: Group All patients: B: Group A: Group Group C: Group preoperatively, and at 8 h 16 postoperatively, plus placebo. D: Group small, sepsis achieved the rate of wound does not reduce metronidazole agent. dose of this antimicrobial a single preoperative by B: Group atively, and at 8 h, 16 h and 24 postoperatively. of the operation, 6 h. six times every repeated Group B: Group Group A: Group Emergency and elective abdominal surgery. Emergency and elective being pus; wound sutured secondary – discharging metronidazole, 500 mg at induction of becoming contaminated with micro-organisms. surgery. colorectal Elective 72 h. for Cancer patients: 77%. preoperatively, plus placebo. surgery. colorectal Elective infection, prevent to adequately surgery, colorectal following discussed. are vs. 3/4. Anastomotic dehiscence: 1/22 vs. 3/21 vs. 1/24 vs. 1/23. was recorded. surgery. Elective Cancer patients: 88.5%. preoperatively, the 750 mg liquid over pulmonary): 9/26 vs. 3/26. mortality: Operative 2/26 vs. 1/26. Cancer patients: NA. anaesthesia, plus mezlocillin 2 g every WI: 8/43 vs. 3/52.WI:0/43 Secondary major vs. 0/52.WI: Secondary minor Cancer patients: 65%. induction of anaesthesia and 12 h infection: patients with postoperative 3/36 vs. Total 14/35. * 35 36 33 34 Indicates data for all surgical procedures included in a trial (i.e where data for elective colorectal operation cannot be sepa included operation elective colorectal in a trial all surgical procedures data for (i.e where Indicates data for 1-centre. serum, bile or intestinal contents subsequently anaesthesia, 8 h and then every sur but not in colorectal appendicectomy Carr, 1984 Randomisation: Not clear.Blind outcome assessment: Yes. Follow-up: 52 days. Age: 65. SWI: the main suture Purulent discharge from and at 8 h preoperatively result: event [Adverse NA. Cost informati Colizza, 1987 Randomisation: Not clear.Blind outcome assessment: No.Follow-up: 30 days. Age: median 62. SWI: Not defined. skin closure, fascia before and 750 mg 6 h. times every four result: event [Adverse Yes.Cost imformation: i.v. at the end of operation, NA] repeated Withdrawal: NA.1-centre. was negative. if culture line even postoperatively, plus placebo. Withdrawal: NA. 1-centre. Claesson, 1986 Study and quality Surgery and definition of SWI Antibiotic regimensA vs. Results (Group B vs. Group C vs. Group D) Group * Note: calculation, in the table is a rough (in years) presented The age in any one trial, all groups across based on the available Cann, 1988 NA = not available or applicable Randomisation: Not clear.Blind outcome assessment: Yes. Follow-up: 30 days.Withdrawal: Age: 7%. NA. SWI: Primary a dry – the first discharge from 48 h. 8 h for 0/43 vs. 1/52. Randomisation: Not clear.Blind outcome assessment: No. Follow-up: 28 days. Age: 63 (20–86). SWI: Discharge of pus. when Classified as early postoperatively. result: event [Adverse NA. Cost information: NA.] Withdrawal: 1%.1-centre. after the operation, within 10 days found and cefuroxime, 1.5 g i.v. 2 days. 8 h for every otherwise as late. primary sepsis Only wound copes wi almost invariably 69 surgery, drug. this agent should be combined with an antiaerobic continued Appendix 3 e continued hout s. had If the trend in. Death: 0/30 vs. 2/32 vs. 0/27 vs. 2/31. < 0.05). p nfections associated with colorectal associated with colorectal nfections WI by type of surgery:WI by Rectal resection: 14/62 vs. 7/61; Deaths: 1/40 vs. 3/40. (SD): Hospital stay 17(8) vs. 19(11). The single-dose combination regimen is as effective in the is as effective The single-dose combination regimen This study was terminated with 40 patients in each group with no This study was terminated with 40 patients in each group Postoperative WI is associated with length of operation and Postoperative Conclusion: SWI:15/141 vs. 9/169. SWI: 0/30 vs. 3/32 vs. 0/27 vs. 2/31. SWI: 6/40 vs. 10/40. Conclusion: Authors’ conclusions in their hospital. bacteriological environment [Adverse event result: event [Adverse NA. Cost information: Yes.] Conclusion: the decision for antibiotic prophylaxis based on his own experience within the based on his own antibiotic prophylaxis the decision for data Cefoxitin, 2 g 30–60 min prior Cefoxitin, 2 g 30–60 min prior Metronidazole, 500 mg i.v. Cefuroxime, 1.5 g 30–60 min Cefoxitin,A, as in Group plus Cefuroxime, 1.5 g plus Cefoxitin, 1–2 g i.v. at induction Metronidazole, 500 mg Group A: Group Group A: Group A: Group B: Group 24 h preoperatively. antibiotics. of oral and parenteral to incision. Group B: Group C: Group three 6 h for to incision and every doses. additional D: Group Group B: Group base, p.o. 4 g/day in divided doses for a combination by lowered and can be significantly location of colon resection Cancer patients: 53%. surgery. before the day administered result: event [Adverse NA. Cost information: Yes.] Cancer patients: 67%. i.v. plus cefuroxime, 1.5 g i.v. preoperatively. result: event [Adverse NA. Cost information: NA.] Elective colorectal surgery. colorectal Elective Cancer patients: 82%. surgery. colorectal Elective of anaesthesia, intraoperatively, and every surgery: intraperitoneal 1/79 vs. 2/108. of hospitalisation: Days wit were plus neomycin Oral erythromycin surgery. colorectal Elective sweating, > 38 °C, temperature or purulent drainage, swelling, or erythema at the incision site. prior to incision. 38 39 37 Randomisation: Not clear.Blind outcome assessment: No. Age: 66. Randomisation: Not clear.Blind outcome assessment: No. Age: 61. Coppa, 1988 NA = not available or applicable Corman, 1993 Cunliffe, 1985 Study and quality Surgery and definition of SWI Antibiotic regimensA vs. Results (Group B vs. Group C vs. Group D) Group Note: calculation, in the table is a rough (in years) presented The age in any one trial, all groups across based on the available 1-centre. or systemic infections, including chills, Follow-up: 30 days.Withdrawal: 12%. SWI:An unsatisfactory was defined response of wound in terms of patients with evidence metronidazole, 500 mg 30–60 min prior to incision. i of postoperative prevention cefoxit with multiple-dose therapy operation as standard Randomisation: Not clear.Blind outcome assessment: No. Follow-up: NA. Age: 63 (21–94). SWI: cultures. Suppurated and had positive day. the first postoperative 6 h for 132; infection 27.2 ( with infection Withdrawal: 0%. neomycin, p.o. 8 g/day plus erythromycin Follow-up: 10 days. SWI: the main Purulent discharge from preoperatively. group the two demonstrated between significant difference Withdrawal: NA.1-centre. of negative line irrespective suture bacteriological culture. persisted, statistical significanc to achieve of patients required the number of 250–300. be in the order would be that each surgeon should make It may 70 1-centre. continued Health Technology Assessment 1998; Vol. 2: No. 7 al continued 32. al I: 3 vs. 1. ction rates, with the compared < 0.001, exact test). Fisher’s Minor SWI: 7/26 vs. 2/29. p ( Major WI:3/44Major vs. 4/48.WI:4/44 Minor vs. 5/48. Intraabdominal infection: 1/25 vs. 1/25. UTI: 1/25 vs. 2/25. The combination of gentamicin plus ticarcillin is recommended in is recommended The combination of gentamicin plus ticarcillin A single dose of intravenous Timentin was as effective as two as two Timentin was as effective A single dose of intravenous in reducing Sulbactam combined with ampicillin was effective a lower schedule showed the four-dose Patients receiving Septicaemia: 1/26 vs. 1/29. Deaths: 5/26 vs. 4/29. Conclusion: SWI: 14/26 vs. 3/29 Deaths: 4/146 vs. 5/132. SWI: 7/44 vs. 9/48. Deaths: 1/44 vs. 3/48. SWI:2/25 vs. 3/25. Conclusion: Conclusion: Conclusion: Authors’ conclusions emergency or as a complement to oral antibiotics in elective colorectal surgery. colorectal emergency or as a complement to oral antibiotics in elective data Gentamicin, 80 mg plus result: event [Adverse Yes.Cost information: translation.] NA from Ticarcillin/clavulanic acid Ticarcillin/clavulanic result: event [Adverse NA. Cost information: NA.] Cefoxitin, 2 g i.v. doses, four result: event [Adverse Yes.Cost information: NA.] Piperacillin, 2 g at induction Gentamicin, 80 mg i.v. at Ticarcillin/clavulanic acid Ticarcillin/clavulanic WI:16/146 All vs.WI:11/146 17/132 (excluding minor vs. 13/132). Sulbactam, 1 g plus ampicillin, Piperacillin, 2 g at the induction Group B: Group Group A: Group Group A: Group B: Group A: Group B: Group A: Group B: Group Cancer patients: translation. NA from 48 h. anaesthesia and then t.d.s for Major SWI: 7/26 vs. 1/29.WI:2 Perineal vs. 2. Intraperitone Cancer patients: 72%. 1 g i.v. 6 h, doses every four Intraabdominal: 2/44 vs. 0/48. Respiratory tract infection: 3/44 vs. 4/48. Cancer patients: 70%. of anaesthesia then perioperatively, Pneumonia: 0/25 vs. 1/25. Cancer patients: 82%. (Timentin), 3.1 g i.v. commenced before WI only: Minor 6/146 vs. 10/132. Septicaemia: 4/146 vs. 5/132. Intraabdomin Elective colorectal surgery. colorectal Elective discharge. a serous inflammation with only surgery. colorectal Elective after start of operation. hospital stay.alter management or prolong against surgical infection. prophylaxis doses for surgery. colorectal Emergency and elective septic complications. postoperative risk from 40 41 43 (French) surgery. colorectal Emergency and elective 42 Randomisation:True.Blind outcome assessment: Yes. Age: 60. Follow-up: NA.Withdrawal: 15%. SWI: translation. NA from ticarcillin, 5 g i.v. 1 h preoperatively and then t.d.s. 48 h. for Randomisation:True.Blind outcome assessment: Yes. Follow-up: 30 days. Age: 63. SWI: the suture Purulent discharge from Randomisation: Not clear.Blind outcome assessment: No. Follow-up: NA. Age: 66. SWI: sepsis – discharge of pus from Major wound 30 min. over skin incision and given abscess: 8/146 vs. 10/132.WI:4/9 Perineal vs. 4/9. UTI: 24/146 vs. 24/1 commencing at induction of anaesthesia.Randomisation: Not clear.Blind outcome assessment: UTI: 14/44 vs. Yes. 16/48.Anastomotic leak/fistula: 2/44 vs. 7/48. Follow-up: 28 days. Age: 45 (20–81). SWI: Not defined. site at the operative (Infection doses. more two 6 h for and every result: event [Adverse Yes.Cost information: NA.] Withdrawal: 17%.1-centre.De La Hunt, 1986 line or a non-purulent discharge that contained (Timentin), 3.1 g i.v. skin before pathogenic bacteria.WI – localised Minor Withdrawal: 12%.2-centres. incision and a second dose 2 h Desaive, 1985 the wound; minor – non-purulent, bacteriologically 6 h, every commencing at induction discharge which did not significantly wound positive of anaesthesia. Cuthbertson, 1991 Devecioglu, 1990 Study and quality Surgery and definition of SWI Antibiotic regimensA vs. Results (Group B vs. Group C vs. Group D) Group 1-centre. 1-centre.NA = not available or applicable Note: calculation, in the table is a rough (in years) presented The age in any one trial, all groups across based on the available morbidity [> 38 °C > 24 h].) and febrile results piperacillin group. two-dose Withdrawal: NA. clinical evidence, was determined by culture of anaesthesia and perioperatively.71 infe incidence of local and remote continued Appendix 3 continued the prevention the prevention h combination therapy of cefuroxime of cefuroxime h combination therapy Severe WI:5/51 vs.Severe 4/10.WI:10/51 Minor vs. 6/53. = 0.042). One intraabdominal abscess, no anastomotic p ( Cefonicid offers the advantages associated with offers Cefonicid There was no significant difference in antibiotic prophylaxis with in antibiotic prophylaxis was no significant difference There Administration of single-dose cefmetazole is as effective Administration of single-dose cefmetazole is as effective SWI:15/51 vs. 10/53. SWI:1/63 vs. 4/32 SWI: 2/23 vs. 1/17. Conclusion: Conclusion: Authors’ conclusions Conclusion: addressed when cefonicid enters the market place) and the freeing place) and the freeing enters the market when cefonicid addressed data rated out from other procedures) out from rated Cefuroxime, 1.5 g plus Cefoxitin, 2 g i.v. preoperatively, Cefoxitin, 2 g i.v. 0.5–1 h patient care. hours for of nursing Mezlocillin, 5 g at induction Cefmetazole, 2 g i.v. Cefonicid, 1 g i.v. 0.5–1 h Group A: Group B: Group A: Group B: Group A: Group preoperatively. an Patients received than 24 h. no more for thereafter thereafter for no more than 24 h. no more for thereafter B: Group 6 h and 2 g every preoperatively be cost can only pharmacy costs (the final consideration of overall * Emergency and elective colorectal surgery. colorectal Emergency and elective Cancer patients: 70%. hospital was delayed. from abdominal surgery.Elective of anaesthesia, then 2 g at 8 h and Cancer patients: 41%. Septicaemia: 2/51 vs. 0/53. Chest infection: 8% vs. 5%. UTI: 27% vs. 25%. of incision.or as a result drainage If wound isolated on culture.Aorganisms were deep occurred to have was considered space infection patients: All colorectal Neomycin, 1 g plus of intraabdominal abscess was evidence if there or peritonitis. preoperatively. erythromycin, 1 g p.o. at 19 h, 18 h, and 9 h surgery. colorectal Elective result: event [Adverse Yes.Cost information: NA.] Cancer patients: 50%. All patients: Neomycin, 1 g plus doses. two 6 h for and then every abdominal operations. of SWIs following base, erythromycin 1 g p.o. times three dehiscence and no deaths. pus discharged, the patient suffered constitutional symptoms and discharge postoperatively. metronidazole, 500 mg at 8 h and 16 * 44 45 46 Indicates data for all surgical procedures included in a trial (i.e where data for elective colorectal operation cannot be sepa included operation elective colorectal in a trial all surgical procedures data for (i.e where Indicates data for Randomisation: Not clear.Blind outcome assessment: No. Follow-up: 42 days. Age: 60 (18–92). SWI: with purulent Minor – swelling DiPiro, 1989 Randomisation: Not clear. Age: 54. 16 h postoperatively.3-centres.Fabian, 1984 result: event [Adverse NA. Cost information: if infected it was considered was serous NA.] Randomisation: Not clear.Blind outcome assessment: Yes. Follow-up: NA. Age: 50 (19–81). SWI:The discharge of pus. operation. before on the day result: event [Adverse Yes.Cost information: NA.] Diamond, 1988 Blind outcome assessment: No. dose of study drug if operation additional Study and quality Surgery and definition of SWI Antibiotic regimensA vs. Results (Group B vs. Group C vs. Group D) Group * Note: calculation, in the table is a rough (in years) presented The age in any one trial, all groups across based on the available NA = not available or applicable Withdrawal: 14%.1-centre. 6 h and placebo every preoperatively dose, administration of a single daily of reduction among which are 1-centre. but no constitutional upset; – severe anaesthesia, then cefuroxime, 750 mg plus plus metronidazole. Withdrawal: 7%. discharge positive or bacteriologically metronidazole, 500 mg at induction of mezlocillin monotherapy, wit compared Withdrawal: 21%. 72 Follow-up: 28 days. SWI: Drained purulent material spontaneously exceeded 2–4 h. for of cefoxitin regimen three-dose as a standard continued Health Technology Assessment 1998; Vol. 2: No. 7 . : continued sing staff. plications talisation ge: 3/74 vs. 2/72 olon flora used in combination with ormation: NA.] Single-dose ceftriaxone plus ornidazole combined with Cefotaxime alone 4 g intraoperatively is useful for the prevention the prevention is useful for alone 4 g intraoperatively Cefotaxime plus erythromycin efficacy observed with neomycin The greater Conclusion: Authors’ conclusions Conclusion: Conclusion: data Metronidazole, 750 mg/day Metronidazole, 500 mg plus Cefotaxime,A, as in Group plus and colonic surgery.The incidence of postoperative of SWIs during rectal Ceftriaxone sodium, 1 g plus or diverticular disease. carcinoma left colonic operation for in elective Cefotaxime, 1 g i.v., at complications, Infection major: 6/74 vs. 4/72 vs. 4/71; minor: 21/74 vs. 23/72 vs. Neomycin, 1 g plus Erythema, not opened: pain or tenderness in wound or excessive swelling sodium,Cefotaxime 1 g abscess: Wound 4/102 vs. 6/122.Anastomotic leakage: 4/102 vs. 4/122. cefotaxime,A. as in Group and 8 h 16 after the first dose. of antibiotics during the operation. and the adequate serum levels anaesthesia. Group A: Group A: Group A: Group Group B: Group Group B: Group C: Group p.o. operation, before 3 days for plus Group B: Group Elective colorectal surgery. colorectal Elective surgery. colorectal Elective surgery. colorectal Elective Cancer patients: 73%. premedication, and 1 g with the abdomen 18/71. Local: 8/74 vs. 11/72 vs. abscess: 7/71.Wound 2/74 vs. 2/72 vs. 5/71 Cancer patients: 78%. erythromycin,Cancer patients: 1 g p.o. mannitol 2 h before 74%. 1/45 vs. opened, 0/48.Wound no pus: 1/45 vs. 6/48. Pus visible in wound injections, (four total 4 g), plus Peritonitis: 3/102 vs. 5/122. Deaths: 4/102 vs. 1/122. septicaemia or death related with infection. septicaemia or death related Minor as local (wound abscess,Minor as local (wound fistula, purulent and pus visible in wound. gentamicin, 80 mg i.v. 2 h preoperatively, in the c – the decrease effects of two an addition is probably discharge at drain site) or extraabdominal (urinary or pulmonary infection, lymphangitis,unexplained fever). or ornidazole, metronidazole 1 g i.v. in two and injections (500 mg at premedication sepsis is not changed, is when cefotaxime imidazole derivatives. 500 mg with the final injection of cefotaxime). reoperation or autopsy.reoperation ornidazole, 1 g i.v. at induction of the nur for costs and work reduced Single-dose antibiotic prophylaxis 48 49 47 Figueras-Felip, 1984 Fingerhut, 1993 NA = not available or applicable Randomisation:True.Blind outcome assessment: Yes.Follow-up: 30 days. Age: 65 (SD 12). SWI: Discharge of pus or inflammation serous starting at induction of anaesthesia. metronidazole, 500 mg (three injections, total 1.5 g) i.v. 24 h for result: event [Adverse Yes.Cost information: NA.] Favre, 1984 Study and quality Surgery and definition of SWI Antibiotic regimensA vs. Results (Group B vs. Group C vs. Group D) Group Note: calculation, in the table is a rough (in years) presented The age in any one trial, all groups across based on the available Randomisation:True.Blind outcome assessment: No.Follow-up: mean 20 days.Withdrawal: NA.8-centres. Age: 65. SWI: combining local wound Estimated by than 3 h more operation lasted for abscess with abdominal wall discharge. Major as intraabdominal abscess, peritonitis, (days): 21 vs. 20 vs. 21. injection of 1 g was given an additional 3 h after the second injection. result: event [Adverse Yes.Cost inf Randomisation:True.Blind outcome assessment: No. visceral procedure, any open but before 4 h. 1 g doses every more two If the Follow-up: > 7 days.Withdrawal: NA. Abdominal wall discharge: 2/74 vs.1-centre. 3/72 vs. 0/73. dischar Age: Perineal 70. vs. 0/71. Antibiotic treatment: 20/74 vs. 19/72 vs. 16/71. Mean hospi SWI: Classified as no signs of sepsis; erythema, pain or tenderness in or excessive swelling preoperatively. not opened; wound opened, wound no pus; 22 hours before preparation (bowel operation), and at 18 h 10 result: event [Adverse NA. Cost information: NA.] 2/45 vs. 7/48. Pus visible in perineal wound: 7/12 vs. 8/12. Deaths: 1/45 vs. 4/48. Withdrawal: 15%. discharge. Peritonitis, anastomotic leakage – com against infective regimen enemas is an effective povidone/iodine 20-centres. fistulography, diagnosed by at or found 73 continued Appendix 3 l

continued oup with Perineal wound:Perineal 2 vs. 7. Intraabdominal abscess: Deaths: 2/31 vs. 1/35. Intraabdominal abscess: Perineal infection:Perineal 0 vs. 2. Intraperitoneal: 1 vs. 1 Single-dose ceftriaxone plus metronidazole seems to be the ideal Single-dose ceftriaxone plus metronidazole in combination with a nitroamidazole Single-dose doxycycline The results show the greater effectiveness of the combination effectiveness the greater show The results There are no significant differences between the two regimens. the two between no significant differences are There SWI: 1/30 vs. 2/30. SWI: 3/116 vs. 10/107. SWI:13/31 vs. 6/35. Conclusion: Conclusion: result: event [Adverse NA. Cost information: Yes.] Conclusion: Authors’ conclusions surgery. rectal among the patients who underwent marked SWI: 0/14 vs. 1/18. Conclusion: no antibiotics. operation more than doxycycline monotherapy.The difference was most difference monotherapy.The than doxycycline operation more data Ceftazidime, 2 g plus Doxycycline, 400 mg plus Gentamicin, 80 mg i.m. plus Clindamycin, 0.6 g plus Ceftriaxone, 2 g plus Doxycycline, 400 mg plus did not receive group Control Ceftriaxone, 1 g before Group A: Group B: Group A: Group B: Group A: Group Group B: Group first dose. 24 h.for inhibitory minimum 14–24 h. concentration for Group A: Group Group B: Group to induction of anaesthesia, 8 h and every concentrations exceeding the that provides it has a long half-life Cancer patients: 74%. tinidazole, 1.6 g i.v. 2 h infusion over 3 vs. patients: infected 4.Total 8 vs. 20. Elective colorectal. Elective Cancer patients: 50%. surgery. colorectal Elective metronidazole, 500 mg i.v. at 30 min UTI: 1/30 vs. 2/30. Respiratory tract infection: 1/30 vs. 2/30. appendicectomy Emergency and elective surgery.and colorectal antibiotics. any 2/31 vs. 0/35. Cancer patients: translation. NA from anaesthetic. information: result/cost event [Adverse translation.] NA from 53 (Italian) surgery. colorectal Elective 50 52 51 Randomisation: Not clear.Blind outcome assessment: Yes. Follow-up: 30 days. Age: 68 (16–89). SWI: Purulent discharge, or fluid discharge yielding operation. shortly before result: event [Adverse Yes.Cost information: NA.] Randomisation: Not clear.Blind outcome assessment: No. Follow-up: 30 days. Age: 65 (14–89). SWI: Not defined. prior to induction of anaesthesia. result:Randomisation: event [Adverse NA. Not clear. Cost information: NA.] Cancer patients: 79%. Blind outcome assessment: No. Age: 61. Follow-up: 21–29 days. metronidazole, 500 mg i.v. 2 h Franceshini, 1989 Gomez-Alonso, 1984 Garcia, 1989 Gerner, 1989 Randomisation: Not clear.Blind outcome assessment: No. Age: NA NA = not available or applicable Study and quality Surgery and definition of SWI Antibiotic regimensA vs. Results (Group B vs. Group C vs. Group D) Group 1-centre. Note: calculation, in the table is a rough (in years) presented The age in any one trial, all groups across based on the available 1-centre. 2-centres. Withdrawal: NA.Withdrawal: 6%. bacteriological culture. positive placebo. metronidazole, 500 mg i.v. 30 min prior operation because in colorectal of infection combination in the prophylaxis septic complication rate in colorecta the postoperative compound lowered Withdrawal: NA. SWI: Not defined. preoperatively, 72 h. 8 h for then every gr with the control compared gentamicin plus metronidazole 74 Follow-up: NA.Withdrawal: 0%.1-centre. SWI: Suppuration of wounds, dehiscence of anastomosis, > 38 °C beyond fever postoperative. 6 days aztreonam, 1 g in the morning and operation, before evening and at anaesthesia and at 8 h 16 after continued Health Technology Assessment 1998; Vol. 2: No. 7 continued mpared with mpared as perioperative prophylaxis. as perioperative 0/48. tive infections,tive with the Deep abscess: 0/53 vs. 2/49.Anastomotic leakage: A single dose of doxycycline preoperatively results results preoperatively A single dose of doxycycline no benefit confers lavage that cefotetan show The results Economic reasons justify the single application justify the single application Economic reasons is a rapid,Whole gut irrigation and easily well-tolerated SWI: 1/53 vs. 2/49. Conclusion: result: event [Adverse Yes.Cost information: Yes.] vs. 1/48.WI:2 Perineal vs. 2 vs. 3.Anastomotic leakage: 5/41 vs. 2/46 vs. 2/48. SWI: 18/65 vs. 15/64. Conclusion: Conclusion: Conclusion: Authors’ conclusions organisms. and aerobic against anaerobic agents effective antimicrobial undergoing colorectal surgery. colorectal undergoing data Whole gut irrigation plus Whole gut irrigation Doxycycline, 0.2 g preoperatively Ciprofloxacin, 200 mg i.v. at plus Whole gut irrigation Pneumonia: 3/41 vs. 4/46 vs. 1/48. UTI: 1/41 vs. 1/46 vs. 1/48. Deaths 1 litre of saline containing 1 litre Doxycycline, 0.4 g in 1 litre (Moxalactam),Latamoxef 2 g SWI:1/57 vs. 4/54. UTI: 3/57 vs. 3/54. Pneumonia: 5/57 vs. 1/54. Deaths: 4/57 alone.Whole gut irrigation WI:13/41 Abdominal vs. 10/46 vs. 1/48. Intraabdominal abscess: 5/41 vs. 4/46 1 litre of saline lavage.1 litre sepsis in patients wound in preventing with saline lavage compared Group A: Group B: Group A: Group A: Group Group B: Group B: Group Group C: Group ampicillin, 1 g plus metronidazole, 500 mg i.v. 30 min preoperatively, 8 h then every 3 days.for colorectal in elective preparation bowel method of preoperative performed consisting of prophylaxis operation if combined with systemic antimicrobial operative procedure. operative Group A: Group B: Group cefotetan, 1 g at the end of Cancer patients: 86%. of saline as an infusion completed 2 h 4/53 vs. of infections: number 1/49.Total 5/53 vs. 5/49. Cancer patients: 70%.Cancer patients: 90%. i.v. at induction of anaesthesia. vs. 1/54. Elective colorectal surgery. colorectal Elective or anastomotic leakage. surgery. colorectal Elective surgery. colorectal Elective surgery. colorectal Emergency and elective Cancer patients: NA. practical and economical standpoints. from treatment four-dose All patients: Gentamicin, 120 mg plus metronidazole, 500 mg i.v., on induction result: event [Adverse NA. Cost information: NA.] 54 56 55 57 Randomisation:True.Blind outcome assessment: No. Follow-up: < 30 days. Age: NA. SWI: Pus or fluid emptied spontaneously preoperatively. result: event [Adverse Yes.Cost information: NA.] Randomisation: Not clear. Age: NA. Gortz, 1990 Blind outcome assessment: Yes. Randomisation: Not clear. Age: NA. Randomisation: Not clear.Blind outcome assessment: No. Age: NA. of anaesthesia. Follow-up: NA. SWI: Not defined. (The assessment was based Blind outcome assessment: Yes.Follow-up: NA.Withdrawal: 8%. induction of anaesthesia plus metronidazole, Multi-centre. SWI: of pus Superficial accumulation surgical drainage. requiring preoperatively, 3 days. 8 h for then every causes within 4 days: 2/41 vs. 0/46 vs. metronidazole, 500 mg i.v. 30 min due to infection: 1/41 vs. 1/46 vs. 1/48. noninfectious Deaths from result: event [Adverse NA. Cost information: NA.] NA = not available or applicable Goransson, 1984 Greig, 1987 Study and quality Surgery and definition of SWI Antibiotic regimens1-centre.A vs. Results (Group B vs. Group C vs. Group D) Group Gottrup, 1985 of pathogens was not required.) Note: calculation, in the table is a rough (in years) presented The age in any one trial, all groups across based on the available Withdrawal: 12%. the wound, or after incision from or pus A, Group as for then doxycycline, 0.1 g in an acceptable rate of postopera Follow-up: 30 days.Withdrawal: NA. 1-centre. SWI: the wound. Discharge of pus from 1-centre. the abdomen at laparotomy, from recovered i.v. postoperatively. 3 days for co associated advantages of one-dose treatment Withdrawal: 7.5%. on the clinical appearance; demonstration 500 mg i.v. 2 h preoperatively. of ciprofloxacin, 200 mg or latamoxef, 2 g 75 continued Appendix 3 are are continued surgery. nst a wide nificantly fewer wound wound fewer nificantly de insufficient protection against the obligate de insufficient protection SWI-elective: 11/102 vs. 6/105. SWI-emergency: Intraabdominal abscesses: 2.6% vs. 1.3%. Peritonitis: A single perioperative dose of mezlocillin plus metronidazole is dose of mezlocillin plus metronidazole A single perioperative is active and metronidazole One dose of cefotaxime demonstrated that cephalosporins studies have Our previous A single dose of latamoxef is as effective as a 48-hour course is as effective A single dose of latamoxef [Adverse event result: event [Adverse NA. Cost information: NA.] result: event [Adverse NA. Cost information: NA.] result: event [Adverse Yes.Cost information: NA.] Conclusion: Conclusion: Conclusion: Authors’ conclusions preparation. bowel and all patients with insufficient preoperative SWI: 4/77 vs. 4/77. SWI:12/119 vs. 10/126. [Adverse event result: event [Adverse Yes.Cost information: NA.] Conclusion: data rated out from other procedures) out from rated Three times the same Three Cefotaxime, 2 g plus Gentamicin, 120 mg Latamoxef, 1 g i.v. at induction Metronidazole, 500 mg plus Cefotaxime, 2 g i.v. at induction WI: 44/280 vs. 19/287. Anastomotic leakage: 22/241 vs. 8/239. (Moxalactam),Latamoxef 1 g acid,Amoxycillin/clavulanic WI:16/116 Abdominal vs. 18/121.WI:13/48 Perineal vs. 8/44.An abscess: Group A: Group A: Group A: Group A: Group Group B: Group B: Group B: Group and 6 h later. is justified, treatment multiple where circumstances including long operations metronidazole, 1.5 g. anaerobes. acid. amoxycillin/clavulanic also be true for It may Group B: Group collection of pus detected by either ultrasound-collection of pus detected by guided aspiration or spontaneous discharge of and anaerobes, range of aerobes undesirable effects, and has few cost. administrated and at reasonable as being easily as well Elective colorectal surgery. colorectal Elective Cancer patients: NA. surgery. colorectal Elective Cancer patients: 71%. mezlocillin, 5 g preoperatively. 2.6% vs. 2.6%. Pneumonia: 2.6% vs. 2.6%. Deaths: 2.6% vs. 5.2%. the peritoneum. from pus directly abdominal surgery. Emergency and elective of anaesthesia, and 3 h 9 later. Intraabdominal abscess: 7 vs. dehiscence: 7.Wound 3 vs. 5. of pathogenic organisms.Adischarge with culture doses. seven than 7 days. more at home for dressings surgery. colorectal Elective Cancer patients: 37%. after contaminated abdominal infection when attempting to prevent 1 g/200 mg (co-amoxiclav, 1.2 g) 15/182 vs. 10/177. Septicaemia: 2/182 vs. 2/177. site: Patients with infected in an extension of the hospital stay or required or required in an extension of the hospital stay Cancer patients: NA. i.v. at induction of anaesthesia. 1/17 vs. 4/21. * 58 59 60 61 Indicates data for all surgical procedures included in a trial (i.e where data for elective colorectal operation cannot be sepa included operation elective colorectal in a trial all surgical procedures data for (i.e where Indicates data for Randomisation: Not clear. Age: 59. Randomisation:True. Age: 70 (20–97). Hall, 1989 Hall, 1989 Randomisation: Not clear.Blind outcome assessment: No. Age: 57 (18–92). operation, before immediately and 2 h later. 41/182 vs. 32/177. Blind outcome assessment: No. Blind outcome assessment: No. Follow-up: > 30 days. SWI: with or without Pus in a surgical wound Follow-up: NA. SWI: Not defined.Follow-up: 30 days. SWI: the wound. Discharge of pus from combination of antibiotics, at metronidazole, 1.5 g i.v. at induction of 2-centres. WI was classified as major if it resulted Withdrawal: 10%. constitutional disturbance such as fever, operation and 2 h before immediately Withdrawal: NA.1-centre. Withdrawal: 14%.2-centres. Anastomosic leakage – air or faeces discharged anaesthesia. a drain or fistula. through Intraabdominal abscess – premedication, 1.5 h after skin incision WIs after colon resection. the rate of sufficient to reduce However, there in sig against a wide range of organisms and resulted 1-centre. alone. doses of cefotaxime than three infections agai It is active dehiscence, hospital stay. smell or prolonged foul later, dose of plus one preoperative alone, provi without metronidazole Study and quality Surgery and definition of SWI Antibiotic regimensA vs. Results (Group B vs. Group C vs. Group D) Group * Note: calculation, in the table is a rough (in years) presented The age in any one trial, all groups across based on the available Grundmann, 1987 Hakansson, 1993 NA = not available or applicable Randomisation:True. Age: NA. Blind outcome assessment: No. Follow-up: 35 days. SWI: wound discharge or a serous Purulent wound of anaesthesia, a further 6 h for then every 76 Withdrawal: 0.7%. continued Health Technology Assessment 1998; Vol. 2: No. 7 continued was Local infectious complications:Local infectious 1/12 vs. 0/13 A short course of preoperative systemic prophylaxis with systemic prophylaxis A short course of preoperative Single-dose antibiotic prophylaxis is as effective as multiple-dose as multiple-dose is as effective Single-dose antibiotic prophylaxis Single-dose prophylaxis with 1 g of cefamandole naftate Single-dose prophylaxis Conclusion: SWI: 0/12 vs. 0/13 vs. 3/11. studied. centres UTI: 11/75 vs. 9/78. Other infections: 0/75 vs. 1/78. Septic deaths: 1/75 vs. 2/78. Conclusion: SWI:4/56 vs. 6/75. Conclusion: Authors’ conclusions treatment. regimens. antibiotic and effective to be a well-tolerated appears Cefotetan data rated out from other procedures) out from rated Metronidazole, 500 mg i.v. Metronidazole, 500 mg i.v. Piperacillin, 2 g i.v. doses three result: event [Adverse Yes.Cost information: NA.] Cefamandole, 1 g plus a cost and provided inexpensive and sodium carbonate was relatively Metronidazole, 500 mg i.v. Cefotetan, 2 g i.v. at induction WI:14/75 Abdominal vs. 13/78.WI:4 Perineal vs. 2. Drain site infection: 2/75 Ceftriaxone, 1 g plus Group A: Group Group A: Group Group B: Group cefuroxime, by immediately followed 1.5 g i.v. doses. three C: Group saline i.v. by immediately followed B: Group Group A: Group induction of anaesthesia. doses. three induction of anaesthesia. dose of 1 g ceftriaxone sodium. associated with the use of a once-daily Group B: Group * * Cancer patients: NA. induction of anaesthesia, then at 8 h vs. 4/11. complication: Distant infectious 2/12 vs. 1/13 vs. 4/11. Cancer patients: 77%. of anaesthesia. vs. 3/78. Intraabdominal infection: 4/74 vs. 5/78. Chest infection: 12/75 vs. 15/78. Elective abdominal operation (laparotomy). Elective of pathogenic organisms.with culture surgery. colorectal Elective metronidazole, 500 mg i.v. after of 64%.When was required, saving treatment a 23% cost saving All patients: Startedafter the treatment surgery. colorectal Elective Cancer patients: NA. metronidazole, 500 mg i.v. after result: event [Adverse Yes.Cost information: Yes.] 63 64 * 62 Indicates data for all surgical procedures included in a trial (i.e where data for elective colorectal operation cannot be sepa included operation elective colorectal in a trial all surgical procedures data for (i.e where Indicates data for Randomisation: Not clear.Blind outcome assessment: No. Follow-up: > 10 days. Age: NA. SWI: Not defined. and 16 h later. result: event [Adverse NA. Cost information: NA.] Randomisation: Not clear. Age: 61 (SD 17). Withdrawal: 28%.Multi-centre.Blind outcome assessment: Yes. netilmicin, by immediately followed 100 mg i.v. doses. three plus netilmicin or cefuroxime, metronidazole without the use of oral non-absorbable antimicrobials, in the acceptable results gave Follow-up: 42 days.Withdrawal: 9%. SWI: of an abscess or discharging pus Presence the wound.This from definition excluded erythema commencing at induction of anaesthesia. Randomisation:True.Blind outcome assessment: No. Age: 56 (14–98). Hershman, 1990 Hall, 1991 Haverkorn, 1985 Study and quality Surgery and definition of SWI1-centre. Antibiotic regimensA vs. Results (Group B vs. Group C vs. Group D) Group * Note: calculation, in the table is a rough (in years) presented The age in any one trial, all groups across based on the available NA = not available or applicable 2-centre. discharge. or serous Follow-up: 30 days. SWI: discharge wound Purulent or serous Withdrawal: 4%. 77 continued Appendix 3 continued e used without metronidazole. dergoing colon operation was associated with dergoing rated, effective, method for and inexpensive A single agent systemic antibiotic (latamoxef) when used A single agent systemic antibiotic (latamoxef) in prophylaxis effective offers ceftizoxime Intravenous with 10% mannitol preparation mechanical bowel One-day Conclusion: Conclusion: Conclusion: Authors’ conclusions colon surgery. elective content of the colon alone but due to the prophylactic agents used for agents used for content of the colon alone but due to prophylactic 1 g) and short-term, combined with oral neomycin perioperative, intravenous ≈ data Neomycin, t.d.s. 1 g orally Ceftizoxime, 2 g plus Placebo three times as for times as for Placebo three Latamoxef (Moxalactam),Latamoxef 2 g WI:5/64 Abdominal vs. 3/67.WI:0/64 Perineal vs. 1/67. UTI: 7/64 vs. 8/67. Ceftizoxime, 2 g plus WI: 7/89 vs. 9/85. UTI: 8/89 vs. 7/85. Chest infection: 7/89 vs. 4/85. Pyrexia: Metronidazole, 15 mg/kg ( 0.5 g) 6 h and 12 later. resections. colorectal the septic complications of elective reducing ≈ Group A: Group B: Group A: Group B: Group ceftizoxime 2 h later. ceftizoxime A. Group Both groups received a second dose of received Both groups A: Group B: Group metronidazole, 1 g i.v. 1 h preoperatively 500 mg at 8 h and 16 by followed a combination of oral neomycin by WI rate comparable to that achieved a after the first dose. metronidazole.The data obtained demonstrated that the plus intravenous in the bacterial of a reduction rate was not the result infection reduced ( Elective colorectal surgery. colorectal Elective Cancer patients: NA. i.v. and at 4 h 1 h preoperatively surgery. colorectal Elective Cancer patients: NA. Pneumonia: 0/64 vs. 2/67. Anastomotic leak: 1/64 vs. 1/67. metronidazole, surgery. colorectal Elective 500 mg i.v. infusion at Cancer patients: 43%. 3/89 vs. 5/85. Hospital stay: 12/89 vs. 11/85 days. 1 g oral neomycin received Both groups Septic complication: 0/31 vs. 8/37. 3 h with a total of g given every result: event [Adverse Yes.Cost information: NA.] systemic disturbance. 67 65 66 Randomisation:True.Blind outcome assessment: Yes. Follow-up: > 21 days. Age: 67. SWI: Inflammation and induration with the Randomisation:True.Blind outcome assessment: Yes. Follow-up: 42 days. Age: 59 (16–96). 8 h after the first dose. SWI: Discharged pus. Minor – superficial pus in the Randomisation:True.Blind outcome assessment: No. result: event [Adverse NA. Cost information: NA.] Age: 50 (14–89). induction of anaesthesia. result: event [Adverse Yes.Cost information: NA.] to all patients. Hinchey, 1987 Hosie, 1992 Jagelman, 1985 NA = not available or applicable Study and quality Surgery and definition of SWI Antibiotic regimensA vs. Results (Group B vs. Group C vs. Group D) Group Note: calculation, in the table is a rough (in years) presented The age in any one trial, all groups across based on the available Withdrawal: 4%. discharge of pus, results. of culture irrespective operation plus before on the day in patients un prophylactically Withdrawal: 13%. incision; dehiscence, major – wound or associated placebo i.v. infusion at induction Follow-up: 28 days. b operation and may colorectal SWI: Passage or isolation of pus with a positive 3-centres. 1-centre. with intraabdominal or pelvic abscess and/or of anaesthesia. 78 Withdrawal: 22%.1-centre. period. follow-up in the postoperative culture i.v. 1 h preoperatively, then 7.5 mg/kg is a well-tole metronidazole continued Health Technology Assessment 1998; Vol. 2: No. 7 continued ormation: NA.] . Length of hospital stay: 9.6 vs. 9.2 days. Cefoperazone can also be used as a single-dose prophylactic can also be used as a single-dose prophylactic Cefoperazone A single preoperative dose of cefotetan,A single preoperative 2 g was as dose of antibiotic prophylaxis One single preoperative SWI: 2/23 vs. 1/24. Conclusion: Intraabdominal abscess: 2 vs. 0. infection: Peritoneal 1 vs. 0. Stitch abscess: groups) between that no difference Conclusion: Conclusion: well-tolerated and effective as multiple doses of cefoxitin in the reduction of in the reduction doses of cefoxitin as multiple and effective well-tolerated operation is sufficient.The colorectal of in acute and elective likelihood Authors’ conclusions surgery. WIs after colorectal postoperative data rated out from other procedures) out from rated Ampicillin, 1 g plus operation complications after colorectal infectious developing Cefoxitin, 2 g i.v. 30–60 min 2 vs. 0. UTI: 4 vs. 2. Pneumonia: 1 vs. 1. morbidity: Febrile 2 vs. 0. Ampicillin, 3 g plus result: event [Adverse Yes.Cost information: NA.] Cefotaxime 1 g as slow i.v. 1 g as slow Cefotaxime bolus surgical procedures, all elective nearly antibiotic for though its application Cefotetan, 2 g i.v. 30–60 min WI, grade 0: wound 139/164 vs. 63/75; grade 1: wound 17 vs. 10; wound Cefuroxime, 3 g plus Emergency patients: SW 1/16 vs. 6/11 vs. 4/12. Cefoperazone, 1 g as a slow Group A: Group A: Group A: Group Group B: Group B: Group Group C: Group metronidazole, 500 mg i.v. after induction to the use of blood transfusion. is related of anaesthesia and 8 h 16 later. B: Group injection after induction of anaesthesia. operation is uncertain. colorectal as a single dose for 2 h. approximately * prophylactic antibiotics (abdominal,prophylactic i.v. bolus injection after induction of result: event [Adverse Yes.Cost information: Yes.] Cancer patients: 32%. initial incision. before Cancer patients: 77%. grade 2: 5 vs. 1; grade 3: wound 3 vs. 0.WI:14 Major vs. 6.WI:9 Minor vs. 2. metronidazole, 1.5 g i.v. after induction patients: Elective SWI 7/91 vs. 8/89 vs. 8/92. Elective colorectal surgery. colorectal Elective surgery. colorectal Emergency and elective requiring surgical procedures Elective 3 – infection throughout wound or intraabdominal wound throughout 3 – infection abscess. the surgical incision or peritoneal cavity, of bacteriological or laboratory regardless investigation. 1 g was administered An additional lasted if the procedure intraoperatively or drug serum half-lives longer than two * 68 69 70 Indicates data for all surgical procedures included in a trial (i.e where data for elective colorectal operation cannot be sepa included operation elective colorectal in a trial all surgical procedures data for (i.e where Indicates data for Randomisation: Not clear.Blind outcome assessment: Yes. Cancer patients: NA. obstetrics/gynaecology, orthopaedic). anaesthesia. Randomisation: Not clear. Age: NA. Randomisation:True.Blind outcome assessment: No. Age: 65 (18–90). of anaesthesia. (Note: patients, elective for to authors’ statement SWI was estimated according Blind outcome assessment: No. Follow-up: 30 days. SWI:Accumulation of pus, either with Follow-up: 30 days.Withdrawal: 17%.6-centres. SWI: 0 – no erythema or discharge; 1 – cellulitis with or without minimal purulent exudate; initial incision, before 6 h then 2 g every Atelectasis: 1 vs. 1 2 – cellutlitis with moderate purulent exudate; and than 24 h postoperatvely. no more for Withdrawal: 8%. result: event [Adverse Yes.Cost inf 4-centres. surgical spontaneous discharge or requiring drainage. metronidazole, 1.5 g i.v. after induction of anaesthesia. Jones, 1987 1-centre. SWI: of purulent material drained from Presence Withdrawal: 13%. Jagelman, 1988 Jensen, 1990 Study and quality Surgery and definition of SWI Antibiotic regimensA vs. Results (Group B vs. Group C vs. Group D) Group * Note: calculation, in the table is a rough presented (in yeras) The age in any one trial, all groups across based on the available NA = not available or applicable Follow-up: 30 days. Age: 52 (18–92). 79 continued Appendix 3 continued 5 vs. 3/98. = 0.057). p prophylaxis with ampicillin and prophylaxis ion: Yes.] (A + B vs. C, Topical ampicillin should be omitted in elective colorectal colorectal ampicillin should be omitted in elective Topical The beta-lactam inhibitor combination reported herein should herein The beta-lactam inhibitor combination reported A single-dose prophylactic cephalosporin can be used for all cephalosporin can be used for A single-dose prophylactic SWI: 1/35 vs. 1/36 vs. 4/29 SWI:0/11 vs. 1/8. Conclusion: Authors’ conclusions Conclusion: Conclusion: data rated out from other procedures) out from rated Cefotaxime, 1 g i.v. on arrival Cefoxitin, 2 g i.v. on arrival Cefotaxime, 1 g i.v. at induction use as a beta-lactamases for along with other newer be considered No further prophylactic No further prophylactic Ticarcillin/clavulanic acid Ticarcillin/clavulanic Subcutaneous and subfascial Cephazolin, 1 g i.v. on arrival in our practice with the exception indicated surgical procedures Group A: Group A: Group Group B: Group 24 h. 6 h for C: Group in the operating room. If operation lasted longer than 2 h, one dose of 1 g was administered additional patients receiving for intraoperatively Group A: Group in the operating room, and 1 g every resections. of colorectal cephazolin or cefotaxime. B: Group of anaesthesia, further then three doses cost-effective, agent on a wide variety of single-dose surgical prophylactic antibiotic treatment. saline in each of the surgical wounds. B: Group is used. metronidazole * * Elective surgical procedures requiring prophylactic prophylactic requiring surgical procedures Elective such as Intraluminal antimicrobials prophylactic requiring surgical procedures Elective antibiotics (abdominal, obstetrics/gynaecology, (Timentin), 3.1 g i.v. bolus injection slow result: event [Adverse Yes.Cost informat surgery. colorectal Elective Cancer patients: 83%. All patients:Ampicillin, 1 g plus metro- WI:5/105 Deep vs. 5/98. Dehiscence: 7/105 vs. nidazole, 4/98. 500 mg i.v. Hernia: t.d.s. induction 10/10 from Deaths (without infection): 12/105 vs. 8/98. the surgical incision or peritoneal cavity, of bacteriological or laboratory regardless investigation. in the operating room, and 2 g every antibiotics (abdominal, obstetrics/gynaecology, were or erythromycin neomycin result: event [Adverse Yes.Cost information: Yes.] * * 71 72 73 Indicates data for all surgical procedures included in a trial (i.e where data for elective colorectal operation cannot be sepa included operation elective colorectal in a trial all surgical procedures data for (i.e where Indicates data for Randomisation: Not clear.Blind outcome assessment: Yes. Follow-up: 3–6 months. Age: 69. SWI: of pus WI defined as accumulation Deep at least 3 days. of anaesthesia for result: event [Adverse Yes.Cost information: NA.] Jones, 1987 Withdrawal: 33%. 2-centres. SWI: of purulent material drained from Presence 24 h. 8 h for Jones, 1987 Withdrawal: 12%. Study and quality Surgery and definition of SWI Antibiotic regimensA vs. Results (Group B vs. Group C vs. Group D) Group 1-centre.Juul, 1985 SWI: site. Purulence at the wound * Note: calculation, in the table is a rough (in years) presented The age in any one trial, all groups across based on the available postoperatively. procedures. operative NA = not available or applicable Withdrawal: 7%. surgical drainage. requiring of 1 g ampicillin in 10 ml application systemic operation when a perioperative Randomisation: Not clear.Blind outcome assessment: Yes. Cancer patients: NA. orthopaedic).1-centre. upon induction of anaesthesia. Randomisation: Not clear.Blind outcome assessment: Yes. Cancer patients: NA. orthopaedic). operations. used in all colorectal Follow-up: 30 days. Age: 51 (18–92). 80 Follow-up: 30 days. Age: 54 (18–96). continued Health Technology Assessment 1998; Vol. 2: No. 7 /145. continued minal abscess: Wound cellulitis:Wound 1/55 vs. 1/47.Anastomotic leak: Single-dose antibiotic prophylaxis with metronidazole is not with metronidazole Single-dose antibiotic prophylaxis Both regimens of imipenem provided equivalent protection to equivalent protection of imipenem provided Both regimens Single preoperative dose of metronidazole and ampicillin dose of metronidazole Single preoperative Conclusion: Peritonitis: 0/113 vs. 1/114 vs. 1/122. Septicaemia: 2/113 vs. 2/114 vs. 6/122. SWI: 4/55 vs. 14/47. Conclusion: plus metronidazole. of cefuroxime the triple-dose therapy by that afforded Authors’ conclusions effective in reducing WIs.A against both aerobes combination of drugs directed in reducing effective [Adverse event result: event [Adverse Yes.Cost information: NA.] Conclusion: is at least as effective as a 3-day course for reducing septic complications reducing course for as a 3-day is at least as effective data Cefuroxime, 1.5 g plus in the imipenem regimens the two between no obvious differences were There Imipenen, 1 g i.v. at induction of result: event [Adverse Yes.Cost information: NA.] Metronidazole, 1 g i.v. 1 h rate to the infection reduce to effectively in required and anaerobes Metronidazole, 500 mg plus surgery. colorectal elective following Imipenem, 1 g i.v. at induction WI:23/113 Abdominal vs. 22/114 vs. 17/122.WI:7/113 Perineal vs. 3/114 Neomycin, 1 g plus Received no further Received Group A: Group B: Group A: Group Group C: Group metronidazole, 500 mg i.v. at induction of anaesthesia, then further doses of cefuroxime, 0.75 g plus metronidazole, infection, a postoperative of patients developing proportion or type of infection duration of hospital stay, thus supporting the use of a short-course prophylaxis surgery. colorectal elective in non-contaminated patients undergoing third postoperative days. postoperative third 500 mg at 8 h and 16 postoperatively. 12 h postoperatively. Group B: Group preoperatively, and 500 mg i.v. at 6 h and acceptable range. Group A: Group Group B: Group ampicillin 1 g i.v. t.d.s. the second and for Cancer patients: NA. erythromycin, 1 g p.o. at 1 p.m. 2 p.m. and 1/55 vs. 1/47. Cancer patients: 77%. of anaesthesia with a further dose 3 h after vs. 6/122. abscess: Intrapelvic 2/113 vs. 2/114 vs. 4/122. Intraabdo Elective colorectal surgery. colorectal Elective Cancer patients: 68%. All patients: Metronidazole, 1.5 g plus WI:9/149 Deep vs. ampicillin, dehiscence: surgery. colorectal 8/145.Wound Elective 3 g i.v. 4/149 vs. during induction of 3/145. Anastomotic leakage: 4/149 vs. 5/145. Intraabdominal abscess: 1/149 vs. 2 colon surgery. and non-obstructive Elective 76 75 74 Randomisation:True.Blind outcome assessment: No.Follow-up: 7 days. Age: NA. SWI: Exhibited redness, induration, or discharge and at 6 h 12 postoperatively. 11 p.m. operation, before the day plus cephazolin, 1 g i.v. 1 h preoperatively. result: event [Adverse NA. Cost information: NA.] Randomisation: Not clear.Blind outcome assessment: No. Follow-up: 42–56 days. Age: 68. SWI: occurred, the wound If purulent discharge from operation. 2/113 vs. 1/114 vs. 3/122. Discharging drain: 8/113 vs. 8/114 vs. 10/122. Withdrawal: 15%.1-centre. was obtained, bacteriological culture a positive or a anaesthesia and 3 h postoperatively, then a at the site of operation. deep abscess developed furtherdoses (500 mg) at 8 h and 16 h. two Karran, 1993 Khubchandani, 1989 NA = not available or applicable Juul, 1987 Study and quality Surgery and definition of SWI Antibiotic regimensA vs. Results (Group B vs. Group C vs. Group D) Group Note: calculation, in the table is a rough (in years) presented The age in any one trial, all groups across based on the available Withdrawal: 34%.1-centre. of pus. Randomisation: Not clear.Blind outcome assessment: No. Follow-up: 30 days. Age:Withdrawal: 65. 5%. SWI:Accumulation of pus either with spontaneous surgical drainage. discharge or requiring antibiotic treatment. anaesthesia and operation. Septicaemia: 1/149 vs. 0/145. Pneumonia: 16/149 vs. 8/145. 1-centre. 81 continued Appendix 3 continued lactics against major UTI: 7/39 vs. 3/31 vs. 2/33. Minor WI:26/121Minor vs. 19/108.WI:6/121 Major vs. No signficant difference was observed between regimens. was observed between No signficant difference Both current regimens evaluated are simple, are evaluated regimens Both current tolerated, well dose (1 g) of of a single intravenous effect The prophylactic neither the reduced of nalidixic acid to metronidazole Addition SWI:32/121 vs. 27/108. SWI: 14/39 vs. 9/31 vs. 13/33. [Adverse event result: event [Adverse Yes.Cost information: NA.] Pneumonia: 1/52 vs. 0/67. UTI: 4/52 vs. 0/67. result: event [Adverse Yes.Cost information: NA.] Conclusion: Conclusion: Conclusion: Conclusion: Authors’ conclusions WI in over 96% of cases. WI in over incidence of Gram-negative aerobic infection nor the total surgical infection nor the total surgical infection infection aerobic incidence of Gram-negative data Metronidazole, 1 g plus rate.A was not superior to a 1 g dose of 3 g dose of metronidazole Cefuroxime, 1 g plus Doxycycline, 0.2 g i.v. over result: event [Adverse Yes.Cost information: NA]. Metronidazole, 3 g i.v. at Metronidazole, 1.5 g plus prophylaxis. Latamoxef (Moxalactam),Latamoxef 1 g Metronidazole, 1 g i.v. over WI:for Superficial all patients 2/52 vs. 12/67; operation 2/50 vs. elective for Metronidazole, 1 g i.v. at sulfate,Neomycin 1 g plus WI:1/27 Perineal vs. 2/27. No septicemia. UTI: 5/27 vs. 0/27. Group A: Group A: Group A: Group A: Group Group B: Group B: Group B: Group antibiotics for prolonged periods.) prolonged antibiotics for (0.2 g) of doxycycline. induction of anaesthesia. anesthesia. Group B: Group ceftriaxone, 2 g i.v. at induction of (Note: received Nine patients in each group loading dose intravenous was better than that of a standard metronidazole C: Group nalidixic acid, infusion, 3 g slow at metronidazole. Elective colorectal surgery. colorectal Elective Cancer patients: 63%. and emergency surgery. Elective Cancer patients: 50%. i.v. at induction of anaesthesia. 8/108. Deaths: 9/121 vs. 12/108. (median) 13 vs. Hospital stay 15 days. colorectal. Elective Cancer patients: 57%. 30 min, 3–4 h preoperatively. 11/64. Deaths: 3/52 vs. 2/67.Anastomotic dehiscence: 0/52 vs. 1/67. surgery. Elective Cancer patients: 100%. induction of anaesthesia. Chest infection: 1/39 vs. 1/31 vs. 1/33. base, erythromycin 1 g p.o. at 1 p.m., result: event [Adverse NA. Cost information: NA] 77 78 79 80 Randomisation: Not clear. Age: 65 (SD 13.5). Randomisation: Not clear. Age: mean 59. Randomisation: Not clear. Age: 62 (21–95). Randomisation: Not clear.Blind outcome assessment: No. Age: Mean 67 (44–81). 2 p.m. and 11 p.m. on the day operation. before Blind outcome assessment: No. Blind outcome assessment: No. Blind outcome assessment: No. Follow-up: 6 weeks. SWI: Purulent discharge.A was a major infection metronidazole, 500 mg i.v. at induction Follow-up: > 28 days.Withdrawal: 11%. SWI:Visible pus.Follow-up: 42 days. SWI: or Discharge of pus either spontaneously 30 min, 3–4 h preoperatively. induction of anaesthesia. Kingston, 1989 Kling, 1988 Kling, 1989 NA = not available or applicable Kling, 1985 Study and quality Surgery and definition of SWI1-centre. Antibiotic regimens bacteriology. positive A vs. Results (Group B vs. Group C vs. Group D) Group Note: calculation, in the table is a rough (in years) presented The age in any one trial, all groups across based on the available Follow-up: NA.Withdrawal: 14%. SWI:The discharge of pus. needed to determine the efficacy of antimicrobial Better and larger trials are 1-centre. 82 Withdrawal: 5%. and discharge associated with pain and/or pyrexia of anaesthesia.1-centre. prophy effective time and are no nursing require Withdrawal: 14%.1-centre. debridement. by continued Health Technology Assessment 1998; Vol. 2: No. 7 continued is as good is as good with complications: oxiclav is recommended because of its easier is recommended oxiclav same degree of protection against postoperative against postoperative of protection same degree A single dose of cefotaxime plus metronidazole administered administered plus metronidazole A single dose of cefotaxime Both co-amoxiclav and the combination of cefotaxime plus and the combination of cefotaxime Both co-amoxiclav advantage of combining oral and systemic is no additional There [Adverse event result: event [Adverse Yes.Cost information: Yes.] Conclusion: Authors’ conclusions 17/62 vs. 8/67 vs. 8/65. (mean): Hospital stay 14 vs. 12 vs. 12 days. Conclusion: Conclusion: data rated out from other procedures) out from rated Cefotaxime, 1 g plus Cefoxitin, 2 g at induction changes in surgical technique and require complications may infective Both p.o. and i.v. antibiotics surgery. colorectal in elective preparation used with mechanical bowel Cefotaxime, 1 g plus Cefotaxime, 500 mg plus Metronidazole, 500 mg plus result: event [Adverse Yes.Cost information: NA.] Cefoxitin, 2 g i.v. at induction surgery:WI: colorectal Elective 8/65 vs. 5/71 vs. 8/70 vs. 7/67. Co-amoxiclav, 1.2 g i.v. on call WI only: 6/76 vs. 7/88. only: Deep infection 1/76 vs. 1/88. Neomycin, 1 g plus WI,major: Abdominal 13/62 vs. 4/67 vs. 3/65; minor: 1/62 vs. 1/67 vs. 0/65. Group A: Group B: Group postoperatively. Group A: Group B: Group A: Group B: Group Group C: Group of anaesthesia then at 6 h and 12 postoperatively. D: Group the potential bacterial contamination. to reduce other measures metronidazole, 500 mg at induction more two by of anaesthesia followed at 6 h and 12 doses of cefotaxime C: Group * Cancer patients: NA. of anaesthesia. surgery:WI: Emergency colorectal 2/8 vs. 2/13 vs. 0/11 vs. 2/14. Cancer patients: 94%.Cancer patients: 100%. to the operating theatre, more and two and deep infection: Wound 1/76 vs. 1/88. erythromycin, 1 g p.o. at 1 p.m., 2 p.m.,WI,major: Perineal 4/62 vs. 2/67 vs. 2/65; minor: 0/62 vs. 0/67 vs. 1/65. Emergency and elective abdominal surgery. Emergency and elective discharge with a positive or a serous wound of anaesthesia. surgery. colorectal Elective and infection postoperative in reducing is effective preoperatively as definite infection. also taken were biotic treatment 8 h. every postoperatively surgery. colorectal Elective group.infected B. A and Group as in Group and marked cellulitis that warranted systemic anti- cellulitis that warranted and marked theatre, doses more and two cost. use and lower the patients had returned home so that cultures home so that cultures the patients had returned also included in the were could not be taken and gentamicin to be systemic metronidazole recommend antibiotics.We * 82 81 83 Indicates data for all surgical procedures included in a trial (i.e where data for elective colorectal operation cannot be sepa included operation elective colorectal in a trial all surgical procedures data for (i.e where Indicates data for Blind outcome assessment: No. Kwok, 1993 Randomisation:True. Age: 54 (15–93). Follow-up: 30 days. SWI: the of purulent discharge from Presence metronidazole, 500 mg at induction Kow, 1995 Blind outcome assessment: Yes. Follow-up: > 30 days. SWI: with serous Purulent discharge.Wounds Randomisation:True.Blind outcome assessment: Yes. Follow-up: > 28. Age: 61 (25–94). SWI: or pus the surgical wound Pus discharge from Randomisation:True. Age: 64 (SD 14). 8 h. every doses postoperatively result: event [Adverse NA. Cost information: Yes.] and 11 p.m. operation. before on the day abscess: Intraperitoneal 3/62 vs. 2/67 vs. 2/65. Patients Lau, 1988 Withdrawal: 4%.1-centre. bacteriological positive discharge which gave discharge after with serous and wounds cultures gentamicin, i.v. 2 mg/kg body weight over operation. half an hour before Study and quality Surgery and definition of SWI Antibiotic regimensA vs. Results (Group B vs. Group C vs. Group D) Group * Note: calculation, in the table is a rough (in years) presented The age in any one trial, all groups across based on the available NA = not available or applicable Withdrawal: 32%. 2-centres. of pathogenic organisms. culture used regimens. as other commonly in reducing Further improvements Withdrawal: 7%.1-centre. drainage. that required in the wound accumulation metronidazole, 500 mg as an i.v. infusion bacteriological culture discharge with positive Serous the offer metronidazole 15 min on call to the operating over infection. The use of co-am 83 continued Appendix 3 continued vs. 3/64. xis against anaerobic bacteria is provided. xis against anaerobic = 0.007, exact test).WI:2/23 Fisher’s Perineal p Anastomotic leak: 1/61 vs. 2/64. Intraabdominal abscess: Short-term prophylaxis with cefotaxime is as effective as is as effective with cefotaxime Short-term prophylaxis is not necessary antibiotic preparation for Oral neomycin Monoprophylaxis with systemic metronidazole is with systemic metronidazole Monoprophylaxis SWI: 7/61 vs. 2/64. Conclusion: Conclusion: Authors’ conclusions long-term prophylaxis with penicillin plus streptomycin. long-term prophylaxis Conclusion: data Erythromycin base,Erythromycin as in Cefotaxime, 2 g i.v. at the Fosfomycin, 2 g in 100 ml 5.5% Penicillin, 2 ml, IU, i.m. plus WI: 1/48 vs. rupture: 0/52.Wound 1/48 vs. 1/52.Anastomosis: 4/48 vs. 4/52. base,Erythromycin 1 g plus 5.5% glucose solution, 100 ml septic complications after elective aerobic inadequate in preventing Group A: Group A: Group B: Group in both groups. doses. metronidazole t.d.s. operation. before 2 days for Group B: Group induction of anaesthesia and at 3 h 6 h later. contaminated, was heavily If the wound t.d.s. was given metronidazole 3 days for Group A: Group after each immediately was given infusion.metronidazole B: Group surgery, colorectal therapy but short-term combined prophylactic i.v.glucose was given after each of the four is an efficacious regimen. plus fosfomycin with systemic metronidazole Elective colorectal surgery. colorectal Elective Cancer patients: 82%. streptomycin, 500 mg i.m. in the evening metronidazole: Additional 11/48 vs. 8/52. surgery. colorectal Elective Cancer patients: 69%. surgery. colorectal Elective sulfate, neomycin 1 g p.o. at 1 p.m., 2 p.m. 0/61 vs. 1/64. infection: Bloodstream 1/61 vs. sepsis:7/61 0/64.Total began treatment Prophylactic WI:6/23 Abdominal vs. 0/26 ( Cancer patients: 84%. operation. 1–1.5 h before All patients: vs. 0/26. Intraabdominal abscess: 1/23 vs. 0/26. Patients with septic 86 84 85 Randomisation: Not clear.Blind outcome assessment: Yes. Follow-up: 30 days. Age: 68. SWI: the wound; Pus drained from a sample of the and 11 p.m. operation. before on the day result: event [Adverse NA. Cost information: NA.] Randomisation:True.Blind outcome assessment: Yes.Follow-up: NA. Age: 73 (44–91). SWI: of pus requiring Superficial accumulation 4 days. for (b.d.) surgery, before then penicillin t.d.s. for result: event [Adverse Yes.Cost information: twice daily and streptomycin NA.] 6 days Lauridsen, 1988 Lewis, 1989 NA = not available or applicable Study and quality Surgery and definition of SWI Antibiotic regimensA vs. Results (Group B vs. Group C vs. Group D) Group 1-centre. Note: calculation, in the table is a rough (in years) presented The age in any one trial, all groups across based on the available Lindhagen, 1984 Randomisation:True.Blind outcome assessment: Yes.Follow-up: > 28 days.Withdrawal: NA. Age: NA. SWI: Discharge of pus, classified as major if the patient was ill and as minor when symptoms 8 h. times every three metronidazole, 500 mg i.v. infused over 20 min. After operation it was repeated complications: result: event [Adverse 8/23 vs. Yes.Cost information: 0/26. NA.] Withdrawal: 9%.1-centre. surgical drainage. Withdrawal: 5%. culture. discharge was then obtained for A, Group plus metronidazole, 750 mg p.o. of the colon when adequate prophyla 84 1-centre. trivial and local. only were continued Health Technology Assessment 1998; Vol. 2: No. 7 continued < 0.05, p 3/30. urgery. Perineal WI:1/30 Perineal vs. 1/30. UTI: 10/30 vs. 7/30. ), supplemented with an antianaerobic by but should preferably * Although not statistically significant the present study showed study showed significant the present Although not statistically This study is not able to prove a significant difference between between a significant difference This study is not able to prove Single-dose prophylaxis is as efficient as multiple dose application dose application is as efficient multiple Single-dose prophylaxis Ceftriaxone seems to be more efficient than ampicillin plus Ceftriaxone seems to be more agent to prevent deep WIs.(For deep surgery, colorectal patients undergoing agent to prevent no the ceftriaxone and ampicillin was observed in SWIs between difference Conclusion: the combination, both C) to have (Group gentamicin plus metronidazole Anastomotic fistulas: 1/17 vs. 2/15 vs. 0/14. Authors’ conclusions group.) plus metronidazole SWI: 9/59 vs. 6/52. Conclusion: Conclusion: SWI: 4/30 vs. 3/30. Conclusion: [Adverse event result: event [Adverse NA. Cost information: Yes.] data rated out from other procedures) out from rated Gentamicin, 80 mg plus clinical and socioeconomic advantages, with the other combinations compared Gentamicin, 80 mg plus clindamycin, result: event [Adverse Yes.Cost information: Yes.] Cefotaxime, 3 g i.v. with Ampicillin, 2 g plus Amoxycillin/clavulanic acid,Amoxycillin/clavulanic groups. the two Gentamicin, 80 mg plus WI:3/30 vs.4/30 Abdominal vs. 1/30.WI:1/4 Perineal vs. 0/2 vs. 2/3. Mezlocillin, 5 g plus Cefotaxime, 3 g i.v. with Ceftriaxone, 1 g i.v. at induction infection: related colon surgery:Wound Elective 4/32 vs. 5/36. Group B: Group C: Group metronidazole, 500 mg 2 h preoperatively, and its effectiveness and this superiority was associated with metronidazole Group A: Group Group A: Group Group B: Group Group A: Group Group B: Group and every 8 h postoperatively for 3 days. for 8 h postoperatively and every A: Group bacteria. against anaerobic Group B: Group * Emergency and elective colorectal surgery. colorectal Emergency and elective Cancer patients: 82%. lincomycin, 600 mg i.m. 2 h preoperatively, abdominal surgeryEmergency and elective Intraabdominal abscess: 0/30 vs. 1/30 vs. 0/30. Sepsis: None. SWI: Cicatricial infection, or cleavage rupture of anaesthesia.WI (1.4% vs.4.5%, against incisional as prophylaxis metronidazole Cancer patients: translation. NA from metronidazole, 500 mg one short information: result/cost event [Adverse translation.] NA from Cancer patients: translation. NA from anaesthesia. information: result/cost event [Adverse translation.] NA from (clean-contaminated laparotomy). of anaesthesia. infection: related Acute colon surgery:Wound 1/6 vs. 1/5. * (German) surgery. Colorectal 89 (German) surgery. colorectal Elective 87 90 88 Indicates data for all surgical procedures included in a trial (i.e where data for elective colorectal operation cannot be sepa included operation elective colorectal in a trial all surgical procedures data for (i.e where Indicates data for Withdrawal: NA.1-centre. the wound. 600 mg i.m. 2 h preoperatively, and every 3 days. for 8 h postoperatively Blind outcome assessment: Yes. Follow-up: 28 days. SWI:A discharge from of any culture positive Randomisation: Not clear. Age: 60. 3 days. for 8 h postoperatively and every Pneumonia: None. Patients with infection: 4/30 vs. 5/30 vs. Lohde, 1992 Randomisation:True.Blind outcome assessment: No. Age: NA. i.v. during anaesthesia. Randomisation: Not clear.Blind outcome assessment: No. Age: 62. Study and quality Surgery and definition of SWIWithdrawal: 31%.Lohr, Antibiotic regimens 1984 1-centre. Lozano, 1985 A vs. Results (Group B vs. Group C vs. Group D) Group 2 g/0.2 g one short i.v. during anaesthesia. * Note: calculation, in the table is a rough (in years) presented The age in any one trial, all groups across based on the available Follow-up: 28 days. metronidazole, 1.5 g i.v. at induction Randomisation: Not clear.Blind outcome assessment: Yes. Age: 51 (1–96). Cancer patients: NA. Luke, 1991 NA = not available or applicable 1-centre. of the skin with discharge pus. all patients for Follow-up: NA. SWI: Not defined. Follow-up: NA. SWI: Not defined. anaesthesia, and 2 g 8 h 16 later. s colorectal in elective infections postoperative to reduce Withdrawal: NA. Withdrawal: 7%. 85 continued Appendix 3 continued 2/45 vs. 2/40 ths: 3/45 vs. 1/40 vs. 2/42 h tinidazole. e study also demonstrated a h parenteral antibiotics. h parenteral e in elective colorectal operation compared with operation compared colorectal e in elective Ceftriaxone is as effective as cephazolin and metronidazole in as cephazolin and metronidazole Ceftriaxone is as effective more to be significantly Single-dose ceftriaxone was shown Oral ciprofloxacin for prophylaxis may offer advantages in efficacy offer may prophylaxis for Oral ciprofloxacin Conclusion: Conclusion: Authors’ conclusions reduced rate of postoperative urinary rate of postoperative and respiratoryin those tract infections reduced Conclusion: data Ciprofloxacin, 1 g p.o. 1 h Cephazolin, 1 g i.v. plus metronidazole, ceftriaxone in comparison with cephazolin. patients who received Gentamicin, 120 mg plus metro- Ceftriaxone, 2 g i.v. plus Netilmicin, 150 mg or Ciprofloxacin, 1 g p.o. 1 h Ceftriaxone, 2 g i.v. plus WI: 7/90 vs. 8/94 vs. 7/96. Respiratory tract infection: 4/90 vs. 1/94 vs. Ceftriaxone, 2 g plus tinidazole, WI: 8/315 vs. 38/313. Other infection: 24/315 vs. 33/313. Gentamicin, 120 mg plus metronidazole,WI:13/45 vs. Total 4/40 vs. 7/42 vs. 4/42.WI:4/45 Major vs. 2/40 vs. 3/42 vs. 1 g suppository on call to theatre. 500 mg i.v. t.d.s. 3 days. for Group A: Group B: Group A: Group B: Group A: Group Group C: Group Group B: Group Group C: Group nidazole, 500 mg i.v. at induction of anaesthesia. Further doses of gentamicin, 80 mg t.d.s. and metronidazole, 500 mg t.d.s. 3 days. for D: Group plus metronidazole,preoperatively 500 mg i.v. at the induction of anaesthesia.Then p.o. ciprofloxacin, 750 mg p.o. b.d. and metronidazole, Emergency and elective colorectal surgery. colorectal Emergency and elective surgery. colorectal Elective surgery. colorectal Emergency and elective Cancer patients: 50%. 1 g suppository metronidazole on call Cancer patients: 65%. 9/96. UTI: 4/90 vs. 5/94 vs. 14/96. Deaths: 0/90 vs. 0/94 vs. 5/96. Cancer patients: 500 mg i.v. 43%. at induction of anaesthesia. result: event [Adverse NA. Cost information: NA.] 500 mg i.v. at induction of anaesthesia. Further 0/42;WI:9/45 minor vs. 2/40 vs. 4/42 vs. 4/42. Chest infection: 1 constitutional disturbance; minor wound sepsis – patients with cellulitis and a induction of anaesthesia. Further doses of metro- culture. wound positive wit and ease of administration compared nidazole, 500 mg i.v. at 8 h and 16 postoperatively. 92 93 91 Matikainen, 1993 NA = not available or applicable Randomisation:True.Blind outcome assessment: Yes. Follow-up: 30. Age: 56. SWI: of erythema Presence accompanied Randomisation:True.Blind outcome assessment: Yes. Age: 63 (15–92). to operating theatre. result: event [Adverse NA. Cost information: NA.] Lumley, 1992 Study and quality Surgery and definition of SWI Antibiotic regimensA vs. Results (Group B vs. Group 7-centres. C vs. Group D) Group McArdle, 1995 induction of anaesthesia. Note: calculation, in the table is a rough (in years) presented The age in any one trial, all groups across based on the available 1-centre. Withdrawal: 21%. of the wound. bacterial culture plus tinidazole, 500 mg i.v. during combined wit when both were the aminoglycosides Withdrawal: 0%. deep or superficial pus. by Follow-up: Mean 28 days. SWI: or positive Suppuration of the wound suppository on call to theatre. glycerol tobramycin, 80 mg (aminoglycosides) surgery.Th colorectal WIs following preventing WI rat efficient in reducing Randomisation: Not clear.Blind outcome assessment: No.Follow-up: 28 days. Age:Withdrawal: 62. 4%. SWI: of pus either discharging Presence drainage. or requiring spontaneously Major doses of gentamicin, 80 mg and metronidazole, 500 mg at 8 h and 16 postoperatively. vs. 8/42 vs. 3/42. UTI: 2/45 vs. 1/40 vs. 1/42 vs. 2/42. Dea vs. 1/42. result: event [Adverse Yes.Cost information: NA.] 86 1-centre. sepsis – the discharge of pus with wound plus metronidazole, preoperatively 500 mg i.v. at continued Health Technology Assessment 1998; Vol. 2: No. 7 continued within < 0.05).The mortality was p y to determine whether ions after colorectal operations. ions after colorectal Mortality: 1.5% vs. 9.2%.Anastomotic leakage: Anastomotic breakdown: 2/73 vs. 4/80. < 0.05), respectively. p There were no clinical or statistically significant differences significant differences no clinical or statistically were There Latamoxef is as effective as conventional two-drug antibiotic two-drug as conventional is as effective Latamoxef The WI rate was 8% in those patients receiving ticarcillin ticarcillin WI rate was 8% in those patients receiving The Because of the statistically significant increase in postoperative in postoperative significant increase Because of the statistically Conclusion: Conclusion: SWI: 10/125 vs. 24/121. SWI:9/85 vs. 9/93. antibiotics: 8.6% vs. 22%. Authors’ conclusions 1.5% and 9.2% ( flora. bowel and anaerobic against both aerobic effective prophylaxis 30 days: 3/87 vs. 8/90. Conclusion: Conclusion: prophylaxis and 20% in those receiving tinidazole ( and 20% in those receiving prophylaxis WI in patients receiving tinidazole prophylaxis alone, tinidazole prophylaxis WI in patients receiving it is no longer operation antibiotic colorectal elective patients undergoing ethical to deny data Latamoxef (Moxalactam),Latamoxef 1 g Tinidazole, 2 g p.o. at 10 p.m. result: event [Adverse Yes.Cost information: NA.] Mezlocillin, 2 g i.v. skin incision before regimens.A the two trial comparing single-dose between controlled Tinidazole, 2 g p.o. at 10 p.m. on the result: event [Adverse NA. Cost information: NA.] Metronidazole, 500 mg plus WI: 6/45 vs. 5/41. Intraabdominal abscess: 1/45 vs. 1/41. Fistula: 1/45 vs. 3/41. Ticarcillin, 3 g i.v. skin before acid (Timentin),Ticarcillin/clavulanic WI:2/83 Abdominal vs. 12/84.WI:2/6 Perineal vs. 1/8. Intraabdominal abscess: acid (Timentin), Ticarcillin/clavulanic Group A: Group B: Group A: Group A: Group A: Group Group B: Group weighing less than 50 kg. weighing Group B: Group Group B: Group Dosages were reduced by one-third in patients one-third by reduced Dosages were old and in those than 80 years more who were Cancer patients: 81%. 3.1 g i.v. skin incision and completed before result: event [Adverse NA. Cost information: NA.] Elective colorectal surgery. colorectal Elective Cancer patients: 67%. gentamicin, 120 mg i.v. at induction of surgery. colorectal Elective Septicaemia: 2/45 vs. 1/41. Chest infection: 6/45 vs. 7/41. surgery. colorectal Elective Cancer patients: NA. 3.1 g i.v. skin incision and 2 h later. before surgery. colorectal Elective 0/87 vs. 4/90. Septicaemia: 0/87 vs. 3/90. UTI: 11/87 vs. 23/90. Death Cancer patients: NA. incision, and at 2 h after the first dose. 8.6% vs. 7.3%. Intraabdominal abscess: 3.8% vs. 1.5%. Postoperative 95 96 97 94 Randomisation:True.Blind outcome assessment: Yes. Age: 65. after 30 min. Randomisation:True.Blind outcome assessment: Yes. Follow-up: NA. Age: 60. SWI: Discharge of pus, either anaesthesia, and at 8 h 16 after operation. result: event [Adverse Yes.Cost information: NA.] Randomisation: Not clear.Blind outcome assessment: Yes. Age: 66. Melbourne study, 1986 Melbourne study, 1987 McCulloch, 1986 Melbourne study, 1989 Study and quality Surgery and definition of SWI Antibiotic regimensA vs. Results (Group B vs. Group C vs. Group D) Group 7-centres.NA = not available or applicable Note: calculation, in the table is a rough (in years) presented The age in any one trial, all groups across based on the available contained pathogenic bacteria. when this agent is used. prophylaxis is an advantage of prolonged there Follow-up: 30 days. SWI: the suture Purulent discharge from Randomisation: Not clear.Blind outcome assessment: Yes. Age: 64. Withdrawal: 8%. or on expression spontaneously i.v. at induction of anaesthesia, and at septic complicat combinations in preventing 1-centre. the wound. from Follow-up: 30 days.Withdrawal: 15%. SWI: of of pus or recovery Presence bacteria of pathogenic potential from 8 h and 16 after surgery. operation. on the night before Withdrawal: 15%. line or non-purulent discharge that and completed after 30 min. is necessar Timentin prophylaxis with double-dose Follow-up: 30 days.Withdrawal: 13%. SWI: Purulent discharge that contained pathogenic bacteria. the operation. night before 7-centres. exudate. a wound 7-centres. 87 continued Appendix 3 continued Serious WI:3/52Serious vs. 2/48; Pulmonary infection: 1/48; Single prophylactic dose is as efficient as multiple applications. dose is as efficient multiple Single prophylactic applications. dose is as efficient multiple Single prophylactic applications. dose is as efficient multiple Single prophylactic Single prophylactic dose is as efficient as multiple applications. dose is as efficient multiple Single prophylactic 2/53; UTI: 7/47 vs. 1/53. Mortality: 0/47 vs. 0/53; Relaparotomy: 1/47 vs. 0/53; Anastomotic leak: none. 3/63 vs. 2/57. UTI: 9/63 vs. 8/57. Relaparotomy: 0/63 vs. 1/57. 1/52 vs. 2/48; UTI: 6/52 vs. 5/48; Relaparotomy: 2/52 vs. 0/48;Anastomotic Authors’ conclusions of a high standard. and the surgical technique are preparation mechanical bowel of a high standard. and the surgical technique are preparation mechanical bowel of a high standard. and the surgical technique are preparation mechanical bowel of a high standard. and the surgical technique are preparation mechanical bowel [Adverse event result: event [Adverse NA. Cost information: NA.] Conclusion: Conclusion: Conclusion: [Adverse event result: event [Adverse NA. Cost information: NA.] Conclusion: However, a mechanical bowel preparation is likely to succeed only if the to succeed only is likely preparation However, a mechanical bowel if the to succeed only is likely preparation However, a mechanical bowel if the to succeed only is likely preparation However, a mechanical bowel However, a mechanical bowel preparation is likely to succeed only if the to succeed only is likely preparation However, a mechanical bowel data Mezlocillin, 5 g. Anastomotic leak: 0/47 vs. 1/53. Mezlocillin, 5 g plus metronidazole, result: event [Adverse NA. Cost information: NA.] Mezlocillin, 5 g plus Anastomotic leak: 0/63 vs. 1/57. Mezlocillin, 5 g. leak: 1/52 vs. 0/48. Ceftizoxime, 2 g.WI:0/47 Minor vs. 1/53;WI:0/47 Serious vs. 1/53;Mezlocillin, Pulmonary infection: 5 g plus metronidazole, 4/47 vs. WI:0/54 Minor vs. 0/46;WI:2/54 Serious vs. 1/46; Pulmonary infection: 1/54 vs. Cefotaxime, 2 g.WI:0/52 Minor vs. Latamoxef, 2 g.WI:0/63 Minor vs. 1/57.WI:2/63 Serious vs. 2/57. Pulmonary infection: Group A: Group B: Group A: Group B: Group A: Group A: Group B: Group Group B: Group Elective colorectal surgery. colorectal Elective surgery. colorectal Elective surgery. colorectal Elective surgery. colorectal Elective Cancer patients: NA. Cancer patients: NA. 500 mg.Cancer patients: NA. 0/46; UTI: 9/54 vs. 7/46. Mortality: 0/54 vs. 0/46; Relaparotomy: 3/54 vs. 1/46; Cancer patients: NA. 98 98 98 98 Randomisation: Not clear. Age: 63. Randomisation: Not clear.Blind outcome assessment: No. Age: 63. Randomisation: Not clear. Age: 63. Randomisation: Not clear. Age: 63. Mendel, 1987 Mendel, 1987 Mendel, 1987 Mendel, 1987 NA = not available or applicable Study and quality Surgery and definition of SWI Antibiotic regimensA vs. Results (Group B vs. Group C vs. Group D) Group Note: calculation, in the table is a rough (in years) presented The age in any one trial, all groups across based on the available Blind outcome assessment: No. Follow-up: NA.Withdrawal: NA. SWI: Not defined. Follow-up: NA.Withdrawal: NA. SWI: Not defined. 500 mg t.d.s. 3 days. for Blind outcome assessment: No.Follow-up: NA.Withdrawal: NA. SWI: Not defined. metronidazole, 500 mg t.d.s. 3 days. for result: event [Adverse NA. Cost information: NA.] Blind outcome assessment: No. Follow-up: NA.Withdrawal: NA. SWI: Not defined. 1-centre. 1-centre. 1-centre. 88 1-centre. continued Health Technology Assessment 1998; Vol. 2: No. 7 continued furoxime furoxime iven. 41. ry WI: 2/27 vs. 1/33. /38. actic administration, is cause NA from translation.] NA from The risk of perineal WI cannot be further reduced by topical WI cannot be furtherby reduced The risk of perineal The efficacy of co-amoxiclav is comparable with that of The efficacy of co-amoxiclav All three treatment regimens have equal efficacy. have regimens treatment All three Single-dose antibiotic prophylaxis is effective in reducing primary in reducing is effective Single-dose antibiotic prophylaxis [Adverse event result: event [Adverse NA. Cost information: NA.] Conclusion: Authors’ conclusions Conclusion: SWI: 10/142 vs. 12/126 vs. 12/137. Conclusion: Conclusion: SWI: 4/30 vs. 4/39. data rated out from other procedures) out from rated Piperacillin, 4 g plus metronidazole, Mezlocillin, 5 g at induction of organisms mezlocillin,WIs caused by deep of several though the appearance Cefoxitin, 2 g i.v. on anaesthesia, and a No locally injected antibiotics. No locally Metronidazole, 500 mg. infection. of postoperative in the prevention effective regimen Local injection of 80 ml metronidazole Local injection of 80 ml metronidazole Ampicillin, 2 g plus sulbactam, 1 g i.v. on Cefuroxime, 1.5 g plus metronidazole, be must antibiotic regimen surgery.An colorectal appropriate WIs in elective Co-amoxiclav, 1.2 g at Group A: Group wound, of that wound. during and after closure Group B: Group Group A: Group B: Group C: Group 500 mg on anaesthesia, and a further doses two Group A: Group every 6–8 h. every as an i.v. was given 20 min and cefuroxime over injection. B: Group Group A: Group B: Group * Emergency and elective abdominal Emergency and elective SWI: WI, no infection; stitch infection; anaesthesia, and 8 h 16 postoperatively. surgery. colorectal Elective Cancer patients: 68%. to co-amoxiclav, sensitive despite its prophyl as a single dose at the administered Drugs were surgical interventions. active requiring WI:3/27 All vs. induction of anaesthesia. 7/33.WI:1/27 was infused Metronidazole Primary vs. 6/33. Seconda surgery. colorectal Elective Cancer patients: 63%. result: event [Adverse NA. Cost information: NA.] All patients: Metronidazole, 500 mg plus gentamicin,WI:5/41 Abdominal vs. 4/38.WI:19/41 Perineal vs. 18 80 mg at induction of anaesthesia, 8 h then every Intraabdominal abscess: 1/41 vs. 1/41. Septicaemia: 2/41 vs. 0/ Cancer patients: the translation. NA from anaesthesia, 6–8 h. and a further doses every two information: result/cost event [Adverse surgery. induction of anaesthesia, and 8 h 16 result: event [Adverse Yes.Cost information: NA.] 101 * 102 (German) surgery. colorectal Elective 100 99 Indicates data for all surgical procedures included in a trial (i.e where data for elective colorectal operation cannot be sepa included operation elective colorectal in a trial all surgical procedures data for (i.e where Indicates data for 1-centre. of the perineal and subcutaneous layer the muscular is g use of antibiotics if optimal systemic antibiotic prophylaxis Randomisation:True.Blind outcome assessment: No. Follow-up: 90 days.Withdrawal: 17%. Age: 58 (17–83). SWI: of pus, Presence either discharging drainage. or requiring spontaneously (500 mg/100 ml) and 2 ml gentamicin (80 mg/ml) into 2 days. for septic: Death from 1/41 vs. 0/41. Randomisation: Not clear. Age: 59. Follow-up: 30 days. Withdrawal: 14%. Menzel, 1993 Randomisation: Not clear.Blind outcome assessment: No. Follow-up: NA. Age:Withdrawal: NA. 4%. SWI: for Centre Definition set out by (details NA). Disease Control 6–8 h. furtherdoses every two Follow-up: 21 days.Withdrawal: 9%.1-centre. SWI: Purulent discharge only. Moderate – pus with constitutional upset (fever 500 mg. > 37.4 °C and/or leucocytosis); – severe organisms.The and anaerobic against both aerobic use of ce directed to be an appropriate in combination can be considered plus metronidazole Study and quality Surgery and definition of SWI Antibiotic regimensA vs. 2-centres. Results (Group B vs. Group C vs. Group D) Group Mittermayer, 1984 Blind outcome assessment: Yes. superficial infection; deep infection.* Note: calculation, in the table is a rough (in years) presented The age in any one trial, all groups across based on the available concern. for Menzies, 1989 NA = not available or applicable Randomisation: Not clear.Blind outcome assessment: No. Age: 67 (18–88). Cancer patients: NA. postoperatively. Moesgaard, 1988 89 continued Appendix 3 continued 70. UTI: colostomy site:colostomy 1/53 vs. 6/53 vs. 12/56. Major: r against Gram-positive r against Gram-positive Intraabdominal abscess: 6/209 vs. 8/219. Intraincisional antibiotics as an addition to systemic Intraincisional antibiotics as an addition There is no significant difference in septic complications is no significant difference There in prophylaxis antibacterial combination chosen for Any Sepsis rates with latamoxef alone are the same as when it is alone are Sepsis rates with latamoxef SWI: 2/21 vs. 4/19. 0/56. sepsis: Perineal 7/16 vs. 5/15. Major: 6/16 vs. 4/15. Minor: 1/16 vs. 1/15. SWI: 22/209 vs. 23/219. result: event [Adverse NA. Cost information: NA.] Conclusion: [Adverse event result: event [Adverse NA. Cost information: NA.] Authors’ conclusions surgery. in colorectal administration do not reduce WI rates in contaminated abdominal surgery. administration do not reduce Conclusion: Conclusion: Conclusion: surgery. colorectal in elective due to combined with metronidazole However bleeding,treatment-related can no longer be advised short-term latamoxef even data rated out from other procedures) out from rated Gentamicin, 80 mg plus metronidazole, Cefotaxime, 1 g plus metronidazole, No intraincisional antibiotic. Metronidazole, either 500 mg or 1.5 g, result: event [Adverse Yes.Cost information: NA.] Gentamicin, 80 mg plus metronidazole, antibiotics.No additional Abscess: 0/53 vs. 2/56. Septicaemia: 0/53 vs. 1/56.Anastomotic dehiscence: Aztreonam, 1 g i.v. plus metronidazole, SWI: 23/71 vs. 9/70. Intraabdominal abscess: 6/71 vs. 1/70.Anastomotic leak: Cefotaxime, to the topically 2 g applied Group A: Group B: Group closure. at the time of wound subcutaneous layer A: Group A: Group Group B: Group Group B: Group Group A: Group B: Group Elective abdominal surgery. Elective drainage. or requiring spontaneously abdominal surgery.Elective 2 days. for Cancer patients: 28%. or after opening the spontaneously All patients: Cefotaxime, 2 g i.v. plus metronidazole, surgery. colorectal Elective 500 mg i.v. or intraoperatively, pre- then every result: event [Adverse NA. Cost information: NA.] All patients: sodium, Latamoxef 2 g i.v. at the start or mortality short-term rates following vs. long-term prophylaxis. N: 53 vs.WI:11/53 56.Abdominal vs. 13/56.WI: Minor surgery. colorectal Elective antibiotic therapy. Cancer patients: NA. 500 mg i.v. at induction of anaesthesia 6/71 vs. 1/70. Drain infection: 8/71 vs. 6/70. Chest infection: 17/71 vs. 6/ Cancer patients: 47%. 500 mg i.v. at start of operation and 6 h later. Septicaemia: 10/209 vs. 9/219. Death: 10/209 vs. 9/219. Cancer patients: NA. 4 h after first dose. of the operation and repeated 5/53 vs. 1/56. Sepsis at drain site: 1/53 vs. 0/56. Sepsis at culture; 3 – abnormal erythema and drainage and induration requiring organisms. slough or constitutional disturbance. 103 104 * * 105 106 Indicates data for all surgical procedures included in a trial (i.e where data for elective colorectal operation cannot be sepa included operation elective colorectal in a trial all surgical procedures data for (i.e where Indicates data for Randomisation:True.Blind outcome assessment: No. Age: NA. Randomisation: Not clear.Blind outcome assessment: Yes. Age: 57 (13–95). 3 days. 8 h for Blind outcome assessment: No. Randomisation:True. Age: 63 (18–86). Randomisation: Not clear.Blind outcome assessment: No. Follow-up: NA. Age: NA. SWI: 1 – the discharge of pus; 2 – the 8 h. doses every more two by followed 17/71 vs. 11/70. Deaths: 5/71 vs. 6/70. Moesgaard, 1989 Morris, 1990 Study and quality Surgery and definition of SWI Antibiotic regimens1-centre. Moesgaard, 1989 3-centres.A vs. Results (Group B vs. Group C vs. Group D) Group wound. * Note: calculation, in the table is a rough (in years) presented The age in any one trial, all groups across based on the available NA = not available or applicable Follow-up: 30 days.Withdrawal: 2%. SWI:Accumulation of pus, draining Follow-up: NA.Withdrawal: 7%. SWI: Major – fever, purulent discharge from with erythema the wound and necrotic 4/24 vs. 6/25. Follow-up: 30 days. SWI: of pus, Presence either discharging 500 mg i.v. at start 8 h of operation and every Withdrawal: 8%. discharge with the of a serous presence 500 mg i.v. at induction of anaesthesia followed Morris, 1984 1-centre. slough; erythema, minor – no surrounding in single or divided doses. 90 Withdrawal: 9%. 3-centres. the isolation of pathogenic bacteria from 8 h. doses every more two by include adequate cove operation must colorectal continued Health Technology Assessment 1998; Vol. 2: No. 7 continued infections in infections rapy: or prophylaxis against operation-related against operation-related or prophylaxis Wound dehiscence:Wound 7/278 vs. 12/253. UTI: 1/45 vs. 2/45. Chest infection: 1/45 vs. 1/45. This small sample trial showed that prophylaxis with cefoxitin with cefoxitin that prophylaxis This small sample trial showed Cefotetan given alone is as effective as the combination of alone is as effective given Cefotetan (when combined with use of aztreonam The prophylactic in surgical dose of ceftriaxone was as effective A single low [Adverse event result: event [Adverse Yes.Cost information: translation]. NA from SWI: 39/278 vs. 34/253. SWI: 6/45 vs. 3/45. Deaths: 8/278 vs. 3/253. result: event [Adverse Yes.Cost information: NA.] Intraabdominal abscess 1/45 vs. 2/45. Death-septic: 0/45 vs. 4/45. Authors’ conclusions surgery. colorectal clindamycin plus gentamicin. clindamycin Conclusion: Conclusion: Conclusion: Conclusion: prophylaxis as our conventional method of three doses of cefoxitin. method of three as our conventional prophylaxis data Clindamycin, 600 mg plus gentamicin, Cefotetan, 2 g at induction of anaesthesia result: event [Adverse NA. Cost information: NA.] Gentamicin, 80 mg plus clindamycin Cefoxitin, doses with or 4–6 g three result: event [Adverse Yes.Cost information: NA.] Cefoxitin, 2 g i.v. 0.5–1 h before Abdominal SWI: 2/37 vs. 3/35.Anastomic failure: 1/37 vs. 1/35. Fistule: 1/37 Cefotetan, 2 g i.v. at induction of Aztreonam, 1 g plus clindamycin,WI:13/224 Abdominal vs. 29/230.WI:10/28 Perineal vs. 22/45. Intraabdominal Ceftriaxone, 1 g. Group A: Group Group B: Group Group A: Group A: Group A: Group Group B: Group B: Group B: Group Cancer patients: 50%. incision, 2 h later. repeated vs. 1/35. Intraabdominal abscess: 0/37 vs. 0/35. Elective colorectal surgery. colorectal Elective operation. colorectal surgery. colorectal Elective Cancer patients: 87%. 900 mg at induction of anaesthesia, abdominal surgery. Elective and 8 h abscess: 6/224 vs. 8/230. Fever: 83/224 vs. 91/230.Antibiotic the Cancer patients: NA. anaesthesia and at 12 h postoperatively. Septicaemia: 3/278 vs. 3/253. Intraabdominal sepsis: 7/278 vs. 3/253. Cancer patients: NA. isolated on bacterial culture. Operation- was defined as wound, infection related intraabdominal or pelvic sepsis, septicaemia to the related specifically infection or any operations. colorectal in elective infection (French) colonic surgery. Elective 108 109 107 110 Randomisation:True.Blind outcome assessment: No. Age (mean): 66.2. Mosimann, 1987 Randomisation:True.Blind outcome assessment: No. Age: 64. Randomisation: Not clear.Blind outcome assessment: No. Follow-up: 39–46 days. Age: 62 (21–91). SWI: Discharge of pus; discharge serous Randomisation: Not clear. Age: 49. 16 h later. 36/224 vs. 43/230. stay: Postoperative 15 vs. 16 days. Morton, 1989 Nel, 1989 Mozzillo, 1989 Study and quality Surgery and definition of SWI Antibiotic regimensA vs. Results (Group B vs. Group C vs. Group D) Group 1-centre. 1-centre. NA = not available or applicable Note: calculation, in the table is a rough (in years) presented The age in any one trial, all groups across based on the available Follow-up: > 28 days. SWI: Not defined. 80 mg i.v. operation, 1 h before repeated Follow-up: 42 days.Withdrawal: 14%. SWI: the wound, Discharge of pus from dehiscence or the discharge of wound plus metronidazole, and at 12 h postoperatively i.v. slow 500 mg by infusion at induction of Withdrawal: 8%. or erythema culture and/or with a positive 900 mg at induction of anaesthesia and 8 h Blind outcome assessment: No.Follow-up: NA. SWI: Not defined. without metronidazole. Withdrawal: 9%. 8 h and 16 later. with as double-drug prophylaxis to be as effective alone appears 8-centres. which organisms were fluid from serous anaesthesia and at 8 h 16 postoperatively. f plus metronidazole cefotetan 26-centres. antibiotic therapy. induration requiring Withdrawal: NA. 16 h later.91 the incidence of postoperative significantly reduces clindamycin) continued Appendix 3 continued . Death: none. abdominal ost in terms t induction of anaesthesia. ery. is easy to plus metronidazole Fosfomycin Deep infection: 3/72 vs. 3/77. Septicaemia: 1/72 vs. 2/77. Both regimens can be recommended for perioperative use to perioperative for can be recommended Both regimens Both regimens are effective in preventing infections after infections in preventing effective are Both regimens agents effective dose of antimicrobial A single preoperative as ceftriaxone plus metronidazole. is as effective Co-amoxiclav 200 mg) and short(1 day). duration of treatment × SWI: 7/72 vs. 6/77. result: event [Adverse Yes.Cost information: NA.] 0/132 vs. 2/135. Pneumonia: 4/132 vs. 9/135. UTI: 7/132 vs. 12/135.Wound SWI: 2/34 vs. 3/36. Conclusion: Authors’ conclusions surgery. colorectal patients undergoing for prophylaxis preferred Conclusion: Conclusion: Conclusion: (2 data Ceftriaxone, 1 g plus metronidazole, Cephazolin, 2 g i.v. b.d. 3 days. for Tinidazole 1600 mg plus placebo Fosfomycin, 8 g plus metronidazole, Bacitracin, 250 mg plus neomycin, 250 mg p.o. Tinidazole, 1600 mg plus doxycycline, dehiscence: 1/132 vs. 3/135. Lung embolism: 2/132 vs. 0/135. Deaths: 2/132 acid,Amoxycillin/clavulanic 1 g/200 mg WI: 4/100 vs. 4/100. Chest infection: 1/100 vs. 2/100. UTI: 0/100 vs. 1/100. Ciprofloxacin, 200 mg i.v. 12 h. doses every two bacterial infections, prevent dose the advantage of low having ciprofloxacin 1 g infusion 1 h before operation. 1 g infusion h before Group A: Group A: Group A: Group B: Group A: Group Group B: Group Group B: Group Group B: Group Cancer patients: 91%. at induction of anesthesia, (1.2 g co-amoxiclav) Anastomotic leak: 0/100 vs. 1/100. Intraabdominal abscess: none Elective colorectal surgery. colorectal Elective Cancer patients: 78%. All patients: Placebo was used in both groups, the operation. before starting 2 days surgery. colorectal Elective Cancer patients: 87%. Pneumonia: 2/72 vs. 2/77. i.v. administered All drugs were within 2 h UTI: 7/72 vs. 5/77. Deaths: 5/72 vs. 2/77. preoperatively. WI: 4/132 vs. 14/135. abscess: Perineal 1/132 vs. surgery. colorectal Elective 7/135. Intra surgery.Colorectal Cancer patients: 81%. abscess: 0/132 vs. 3/135.Anastomosis dehiscence: 0/132 vs. 3/135. Septicaemia: with the start The first dose was given of anaesthesia. metronidazole, received Both groups 500 mg i.v. result: event [Adverse Yes.Cost information: NA.] as discharge of pus from the perineal as discharge of pus from or drain channel.wound (vitamin solution). organisms seems to be the bowel and aerobic against both anaerobic were included in the assessment. were 112 113 111 114 Norwegian study,Norwegian 1985 Blind outcome assessment: Yes. Blind outcome assessment: Yes. Randomisation:True.Blind outcome assessment: Yes. Follow-up: NA. Age: 62 (23–93). SWI:Abscess or wound Follow-up: NA. and at 8 h postoperatively. SWI: Not defined. result: event [Adverse NA. Cost information: NA.] Randomisation: Not clear. Age: 66 (56–74). Randomisation:True. Age: 65. Randomisation: Not clear.Blind outcome assessment: No. Age: NA. 24 h. 8 h for every Study and quality Surgery and definition of SWI Antibiotic regimensA vs. Results (Group B vs. Group C vs. Group D) Group 1-centre.Offer, 1988 1-centre. Note: calculation, in the table is a rough (in years) presented The age in any one trial, all groups across based on the available postoperatively. time. of pharmacy and nursing Nohr, 1990 Nyam, 1995 Follow-up: 28 days.Withdrawal: 13%. SWI: of pus in the wound Presence that either discharged spontaneously t.d.s. operation, before 2 days for plus metronidazole, 500 mg p.o. operation, before times on the day three surg colorectal elective Follow-up: 30 days.Withdrawal: 5%.10-centres. SWI: of pus or discharge Presence bacteriological in a positive resulting 400 mg. culture. sepsis was defined Perineal Withdrawal: NA. discharging pus.Withdrawal: 1.4%. vs. 4/135. Duration of hospital stay: 11.3 vs. 15.4. 500 mg at induction of anaesthesia, and at 12 h result: event [Adverse Yes.Cost information: NA.] Being a single antibiotic regimen, the c it reduced 1-centre. drainage. or required wounds Perineal plus ampicillin, 1 g i.v. 1 h preoperatively. administer as a single-dose infusion 92 NA = not available or applicable continued Health Technology Assessment 1998; Vol. 2: No. 7 continued 34. e: 25 vs. 15/434. Peritonitis: pital stay: 15 vs. 15 days. cefuroxime plus metronidazole for the for plus metronidazole cefuroxime or abdominal operation and that potential savings may may or abdominal operation and that potential savings UTI: 0/30 vs. 4/30. Intraabdominal abscess: 1/30 vs. A single intravenous dose of imipenem plus cilastatin appears dose of imipenem plus cilastatin appears A single intravenous Pro-host immunotherapy with thymostimulin administered administered with thymostimulin immunotherapy Pro-host It is sufficient for prophylactic purposes for an antimicrobial an antimicrobial purposes for prophylactic It is sufficient for This study demonstrates that co-amoxiclav is an effective is an effective This study demonstrates that co-amoxiclav [Adverse event result: event [Adverse Yes.Cost information: NA.] prevention of infection in elective colorectal surgery. colorectal in elective of infection prevention 7/425 vs. patients: infected 9/434.Total 63/425 vs. 67/434. Respiratory tract 5/206 vs. 2/197. diarrhoea: Postoperative 22/206 vs. 19/197. Febrile SWI: 5/30 vs. 6. Conclusion: SWI: 8/69 vs. 2/79. Authors’ conclusions this type of prophylaxis. be of value for the duration of operation would Conclusion: agent to cover the critical moment of bacterial colonisation during agent to cover procedure.A longer than a half-life operative single dose of a drug having [Adverse event result: event [Adverse NA. Cost information: NA.] Conclusion: Conclusion: data rated out from other procedures) out from rated Cefuroxime, 1.5 g i.v. plus Cefotetan, i.v. 2 g rapid at induction infection: 8/425 vs. 20/434. UTI: 58/425 vs. 58/434.Thrombophlebitis: Cefotetan, 2 g i.v. at start of morbidity: 53/206 vs. 39/197. Deaths: 4/206 vs. 4/197. Postoperative Cefuroxime, 1.5 g plus metronidazole, Imipenem plus cilastatin, 1 g at (Augmentin),Co-amoxiclav 1.2 g i.v. Cefoxitin, 1 g i.v. at the start of WI:23/206 Abdominal vs. 18/197.WI:5/13 Perineal vs. 2/10. Cefotetan, i.v. 2 g rapid at induction N:WI:56/425 425 vs. abdominal vs. 434.Total 72/434.Abdominal wound Group B: Group cefuroxime, 750 mg plus metronidazole, be made with its use in the field of general operation without sacrificing 500 mg i.v. at 8 h and 16 after initial dose. quality of care. Group B: Group Group B: Group Group A: Group A: Group A: Group A: Group Group B: Group Cancer patients: 100%. anaesthesia. 1/30.Anastomotic leakage: 1/30 vs. 1/30. Hospital stay: 13.6 vs. 15.7 days. Elective colorectal surgery. colorectal Elective abdominal surgery. Emergency and elective surgery. colorectal Elective Cancer patients: 86%. surgery. colorectal Elective Cancer patients: 87%. operation and 3 h, 6 h, and 12 h after first dose. Intraabdominal abscess: 1/206 vs. 0/197.Anastomotic dehiscenc criteria.) in the absence of above chemoprophylaxis of anaesthesia, plus thymostimulin, 70 mg i.m. serous: 34/425 vs. 45/434.Abdominal purulent: wound 22/425 vs. 27/4 Cancer patients: NA. at induction of anaesthesia, and 2 h 8 later. result: event [Adverse Yes.Cost information: Yes.] > 100,000 CFU/ml; the need of administering after completion of chemotherapy antimicrobial surgery. in colorectal can potentiate the efficacy of antimicrobial chemoprophylaxis 7 days for intraoperatively less than 6 h; SWI, in other sites [abdom- and infection inal abscess, bronchopneumonia, etc.]. Bacteriuria * 116 115 117 118 Indicates data for all surgical procedures included in a trial (i.e where data for elective colorectal operation cannot be sepa included operation elective colorectal in a trial all surgical procedures data for (i.e where Indicates data for Randomisation: Not clear. Age: 67 (34–88). Blind outcome assessment: No. NA = not available or applicable Palmer, 1994 Periti, 1989 Study and quality Surgery and definition of SWI1-centre. Antibiotic regimens1-centre. A vs. Results (Group B vs. Group C vs. Group D) Group * Note: calculation, in the table is a rough (in years) presented The age in any one trial, all groups across based on the available Randomisation: Not clear.Blind outcome assessment: No. Follow-up: NA. Age: 64 (SD 13). SWI: Not defined. of the (Criteria defining failure starting 48 h preoperatively. 7 days for WI:6/425 Perineal vs. 6/434.Abdominal abscess: 5/4 Randomisation: Not clear.Blind outcome assessment: No. Age: 63 (13–85). Follow-up: NA.Withdrawal: 0%. SWI: Purulent discharge with or without of pathogenic microorganism. culture metronidazole, 500 mg i.v. at induction of further 8 h. anaesthesia two doses every doses of as three to be as effective Randomisation:True.Blind outcome assessment: Yes. Age: 53 (12–96). Pacelli, 1991 Periti, 1993 Withdrawal: 11%. chemoprophylaxis: > 38 °C [excluding the first fever of anaesthesia. 2/425 vs. 0/434. Septicaemia: 1/425 vs. 3/434. Hos 19-centres. occasions with an interval 24 h] on at least two of not result: event [Adverse NA. Cost information: NA.] Follow-up: NA.Withdrawal: 11%. > 10-centres. SWI: material. Drained purulent or serous operation. (Mean days): stay 15 vs. 16. Follow-up: 21–35 days.Withdrawal: 11%. SWI: Erythema, discharge or dehiscence. 500 mg i.v. at induction of anaesthesia, then antibiotic f prophylactic 93 continued Appendix 3 continued ion of penicillin, ctive than the single systemic ctive UTI: 6/34 vs. 3/36. Anastomotic leaks: 2/80 vs. 3/80. = 0.12, exact test). Fisher’s Chest infection: 2/39 p The addition of perioperative intravenous cefamandole to a intravenous of perioperative The addition The single agent, cefoxitin, in prophylaxis is as effective Concomitant use of parenteral and oral combination in elective and oral combination in elective Concomitant use of parenteral A single dose of ceftriaxone provides adequate prophylactic adequate prophylactic A single dose of ceftriaxone provides SWI: 0/34 vs. 1/36. Conclusion: Authors’ conclusions Conclusion: [Adverse event result: event [Adverse Yes.Cost information: NA.] Conclusion: SWI: 4/80 vs. 3/80. Conclusion: data rated out from other procedures) out from rated Cefoxitin, 2 g at start of operation prophylactic for it preferable advantages which make the single agent are No i.v. antibiotics. Cefoxitin, (same doses and timing Cefotaxime, 2 g plus ornidazole, Metronidazole, 500 mg plus WI: 7/26 vs. 4/27. Patients with sepsis: 10/26 vs. 6/27. Cefoxitin, as i.v. bolus 0.5 h WI: 0/39 vs. 4/41 ( Cefamandole, 1 g i.v. 1 h neoplasias in this of colorectal resection WIs following benefit in decreasing Cefotaxime, 2 g i.v. at induction Group B: Group and two more doses, more and two 6 h. one every and biliary use in colorectal surgery. and 9 h preoperatively. with cefoxitin. haemoprophylaxis B: Group Group A: Group A: Group Group B: Group A: Group doses. of four Group A: Group Group B: Group Elective colorectal and biliary colorectal surgery. Elective to the expected, out of proportion swelling surgery. colorectal Elective surgery. colorectal Elective Cancer patients: 100%. and metronidazole.The an aminoglycoside cost and ease of administering mechanical a 3-day All patients underwent erythromycin and received preparation bowel result: event [Adverse NA. Cost information: NA.] Cancer patients: NA. tobramycin, 80 mg at start of operation, then result: event [Adverse NA. Cost information: NA.] Cancer patients: 39%. and then 6 h 12 preoperatively vs. 0/41. UTI: 2/39 vs. 4/41. morbidity: Febrile 13/39 vs. 13/41. Diarrhoea: Cancer patients: translation. NA from of anaesthesia. result: event [Adverse Yes.Cost information: translation] NA from (French) surgery. colorectal Elective 122 * 120 119 121 Indicates data for all surgical procedures included in a trial (i.e where data for elective colorectal operation cannot be sepa included operation elective colorectal in a trial all surgical procedures data for (i.e where Indicates data for 1-centre? or if frank suppuration occurred. Randomisation: Quasi.Blind outcome assessment: No. Age: 59. base, 1 g plus neomycin, 1 g p.o. at 1 p.m., 2 p.m., and 11 p.m. operation. before on the day Perrott, 1985 Study and quality Surgery and definition of SWI Antibiotic regimensA vs. Results (Group B vs. Group C vs. Group D) Group 5-centres.* Note: calculation, in the table is a rough (in years) presented The age in any one trial, all groups across based on the available plus a further dose of ornidazole 12 h later. Peruzzo, 1987 Petrelli, 1988 Petropoulos, 1985 NA = not available or applicable Randomisation: Not clear.Blind outcome assessment: No. Age: 64. at 8 h and 16 later, plus penicillin, 2 milliunits a further by at start of operation followed Randomisation: Not clear.Blind outcome assessment: No. Age: 67 (47–87). postoperatively. 1/39 vs. 3/41. hospital stay: Postop 11.5 (9–22) vs. 11.0 (9–18). Follow-up: NA.Withdrawal: NA. SWI: material, Drainage of serous-purulent with of pathogen micro-organisms or without culture A) plus tinidazole, as in Group 2 g p.o. 2 h Follow-up: 90 days.Withdrawal: NA. 1-centre. SWI: Discharge of pus. preoperatively, a total 6 h for then every of patients. small group no further with oral antibiotics provides preparation mechanical bowel good Randomisation: Not clear.Blind outcome assessment: No. Age: 67. Withdrawal: 5%. Follow-up: 28 days. SWI: and abnormal reddening Showing doses, four 4 h. one every and biliary in colorectal operation as a combinat 1-centre. a discharging abdominal wound. from 1 g p.o. plus neomycin preoperatively 19 h, 18 h effe operation is no more colorectal 94 Follow-up: NA.Withdrawal: 0%. SWI: Not defined. 500 mg at induction of anaesthesia, and two surgery. colorectal in elective antibiotic cover further 8 h and 16 later doses of cefotaxime continued Health Technology Assessment 1998; Vol. 2: No. 7 continued ity that cephalosporins can cause Minor WI: 11/58 vs. 8/61. Major WI: 5/58 vs. 1/61. A single preoperative dose of co-amoxiclav is as effective as one is as effective dose of co-amoxiclav A single preoperative of adequate tissue levels to ensure It is advisable not only than cephazolin in preventing effective is no more Latamoxef Both regimens are well tolerated and provide effective effective tolerated and provide well are Both regimens Authors’ conclusions patients in the cefoxitin three in colon operation where noted were differences SWI: 10/36 vs. 15/56. SWI: 0/9 vs. 7/18. SWI: 16/58 vs. 9/61. Conclusion: SWI: 1/24 vs. 1/26. SWI: 0/28 vs. 2/12. Conclusion: Conclusion: Conclusion: data rated out from other procedures) out from rated Cefoxitin, 2 g preoperatively, and WIs. had infections, group minor of which were two Co-amoxiclav, 1.2 g in 20 ml of solution complications in abdominal surgery. of infective the prophylaxis for of latamoxef Neomycin, 1 g p.o. 6 h plus every operations. in colorectal partially by decontaminating the bowel infection (Moxalactam),Latamoxef 1 g i.v in this study population. infections postoperative Latamoxef, 1 g in 20 ml of solution result: event [Adverse Yes.Cost information: Yes.] Metronidazole, i.v. at premedication. Cephazolin, 1 g i.v. on call to Cefmetazole, 2 g preoperatively, Group B: Group Group B: Group Group A: Group A: Group A: Group metronidazole, or 500 mg i.v. 1 g rectal 6 h. doses every two Group A: Group Group B: Group metronidazole, 8 h, 200 mg every 24 h, for dose of preoperative plus parenteral B: Group * discharge from hospital.discharge from Minor infections did neither. and haemoserous Serous as single i.v. injection at induction of anaesthesia.Cancer patients: 74%. speculation about the possibil Continued dose of an antibiotic intraoperative Anastomotic leak/abs: 4/58 vs. 7/61. Deaths: 5/58 vs. 6/61. Emergency and elective abdominal surgery.Emergency and elective patients: All colorectal Either 24-h p.o.Cancer patients: neomycin NA. using metronidazole: Ileocolorectal caused constitutional disturbances as single i.v. injection at induction of anaesthesia. plus metronidazole, or single suppository not counted as infections. of discharges were surgery. colorectal Elective All patients: or a single preoperative abdominal surgery. Elective Cancer patients: NA.pathogen. this indication. for encourage the use of co-amoxiclav bleeding complications may abdominal surgery.Elective Cancer patients: NA. and twice the operating theatre operation also colorectal Patients undergoing result: event [Adverse Yes.Cost information: NA.] and twice on call to the operating theatre preparation. oral antibiotic bowel received result: event [Adverse Yes.Cost information: NA.] * * * 123 124 125 126 Indicates data for all surgical procedures included in a trial (i.e where data for elective colorectal operation cannot be sepa included operation elective colorectal in a trial all surgical procedures data for (i.e where Indicates data for Playforth, 1987 1-centre. Randomisation: Not clear.Blind outcome assessment: No. Follow-up: NA. Age: NA. SWI: Discharge of pus. Major infections Randomisation: Not clear.Blind outcome assessment: No. Follow-up: NA. Age: 20–89. metronidazole, 1 g approx. 2 h preoperatively. (emergency), Ileocolorectal no metronidazole: SWI: and delaying Major – causing pyrexia bacteria. against faecal aerobic effective result: event [Adverse NA. Cost information: NA.] 1-centre.Playforth, 1988 patients’ and delayed including pyrexia Randomisation: Not clear.Blind outcome assessment: Yes. Age: NA.Randomisation: Not clear. Age: 52. 6 h. every postoperatively Plouffe, 1985 Study and quality Surgery and definition of SWI Antibiotic regimensA vs. Results (Group B vs. Group C vs. Group D) Group 1-centre.Plouffe, 1989 Blind outcome assessment: No. * Note: calculation, in the table is a rough (in years) presented The age in any one trial, all groups across based on the available 6 h. every postoperatively NA = not available or applicable Withdrawal: 1%. Withdrawal: NA.1-centre. hospital; discharge from patient’s or minor. Follow-up: NA. SWI: Purulent discharge and isolated the risk of endogenous bacterial but also to reduce antimicrobials Follow-up: NA.Withdrawal: 31%. SWI: Not defined. 8 h. doses every and two intraabdominal surgery.The only elective for prophylaxis perioperative Withdrawal: 20%. 95 continued Appendix 3 e: continued or the prophylaxis of SWI. or the prophylaxis y in preventing postoperative infective infective postoperative y in preventing Wound dehiscence:Wound 2/50 vs. 2/50.Anastomotic leak: dehiscence:Wound 1/89 vs. 2/81. Intraabdominal The incidence of WI was considerably lower lower WI was considerably The incidence of Prophylactic cefotaxime and metronidazole are effective in effective are and metronidazole cefotaxime Prophylactic A three-dose intravenous course with cefotaxime is comparable course with cefotaxime intravenous A three-dose Topical ampicillin could not lower the WI rate further when the ampicillin could not lower Topical = 0.005). than effective injection is more intraparietal Preincisional SWI: 2/50 vs. 6/50. SWI: 6/89 vs. 5/81. Conclusion: SWI: 2/17 vs. 1/17. [Adverse event result: event [Adverse Yes.Cost information: NA.] Conclusion: Conclusion: Authors’ conclusions p 6 (8.5%) vs. 3 (4.5%). Conclusion: Nevertheless, a of patients is needed to reach a study with larger number data Neomycin, 1 g plus erythromycin, Co-amoxiclav, 1.2 g injected the antibiotic into abdominal wall (8.4% vs. given in the group 15.9%, Cefotaxime, 2 g i.v. at induction result: event [Adverse Yes.Cost information: NA.] Cefotaxime, 1 g h preoperatively conclusion. definitive Amoxycillin/clavulanic acid Amoxycillin/clavulanic SWI: with metronidazole: 11/47 vs. 9/40; without metronidazole: 6/15 vs. 3/20. Cefotaxime, 2 g i.v. at induction of Cefotaxime, 2 g i.v. at induction Cefotaxime, 1 g h preoperatively. surgery. colorectal the incidence of septic complications following reducing Group A: Group was used in the majority of Metronidazole A: Group A: Group Group B: Group colorectal patients. colorectal and 8 h 16 postoperatively. Group B: Group Group B: Group Group A: Group Group B: Group Cancer patients: NA. 1 g/200 mg (co-amoxiclav, 1.2 g) i.v. at result: event [Adverse Yes.Cost information: NA.] Elective abdominal surgery. Elective minor and occurred late – all these were after patients had left hospital. spinal needle. surgery. colorectal Elective Cancer patients: 81%. surgery. colorectal Elective Cancer patients: of all surgical procedures.) 84%. (This conclusion was based on the results anaesthesia, and at 6 h 12 after first dose. 5/50 vs. 2/50. Intraabdominal abscess: 1/50 vs. 1/50. surgery. colorectal Elective Cancer patients: 85%. of anaesthesia, and at 6 h 12 later. abscess: 2/89 vs. 3/81.Anastomoses: 71/89 vs. 66/81.Anastomotic leakag All patients: Metronidazole, 500 mg b.d. 3 days, for with the first dose administered result: event [Adverse NA. Cost information: NA.] 130 128 129 127 Randomisation:True.Blind outcome assessment: Yes. Age: NA. induction of anaesthesia. Randomisation:True. Age: 73 (39–90). Randomisation: Not clear.Blind outcome assessment: No. Age: 53 (15–80). prior to operation. immediately Blind outcome assessment: No. Follow-up: NA. SWI: Obvious collection of pus, which operation, before times on the day 1 g three Raahave, 1988 NA = not available or applicable Pollock, 1989 Study and quality Surgery and definition of SWI Antibiotic regimensA vs. Results (Group B vs. Group C vs. Group D) Group 1-centre.Raahave, 1989 2-centres.Rangabashyam, 1991 Note: calculation, in the table is a rough (in years) presented The age in any one trial, all groups across based on the available at closure. during closure. and subcutaneously subfascially was used. a 12-h systemic antibiotic cover complications. Follow-up: 30 days. SWI: Discharge of pus. Major – causing fever Withdrawal: 0%.1-centre. discharge from the patient’s and delaying along the line of subcutaneously hospital); only; dressings minor – requiring abdominal incision, proposed using a f injection of co-amoxiclav intravenous Randomisation: Not clear.Blind outcome assessment: No. Age: 71. Follow-up: NA.Withdrawal: 12%. SWI: Collection of pus, or at the site of incision. spontaneously draining either Follow-up: NA. plus ampicillin, in the wound 2 g powdered of anaesthesia, and at 6 h 12 later, SWI: Not defined. Withdrawal: 9%. or on incision. drained spontaneously plus ampicillin, in the wound 2 g powered orall plus erythromycin to neomycin 96 Withdrawal: 21% 1-centre. continued Health Technology Assessment 1998; Vol. 2: No. 7 continued s. 1. 88. ained with the combination Intraperitoneal infection:Intraperitoneal 6 vs. 2. infection: Perineal The efficacy of piperacillin can be improved by preoperative oral preoperative by The efficacy of piperacillin can be improved to be a valid alternative appears plus clindamycin Aztreonam A single preoperative intravenous infusion of metronidazole in infusion of metronidazole intravenous A single preoperative [Adverse event result: event [Adverse Yes.Cost information: NA.] Conclusion: SWI: 8/66 vs. 12/72. Conclusion: SWI: 6/170 vs. 1/188. Authors’ conclusions bacteria. selection of tetracycline-resistant due to the increasing Conclusion: of cefuroxime plus metronidazole and it remains to be seen whether and it remains plus metronidazole of cefuroxime ampicillin and doxycycline in this investigation,ampicillin and doxycycline choice ampicillin is the preferred data Metronidazole, 500 mg plus cefuroxime, Piperacillin, 2 g i.v. at induction of the addition by could be improved with combination therapy these results Clindamycin, 0.6 g plus gentamicin, following of infections the prophylaxis for to gentamicin plus clindamycin Metronidazole, 1.5 g i.v. plus prophylactic is a simple and effective combination with ampicillin or doxycycline Metronidazole, 1.5 g i.v.. Metronidazole, 8 h 400 mg every WI:9/107 Abdominal vs. 18/104 vs. 8/119.WI:7/16 Perineal vs. 8/18 vs. Clindamycin, 0.6 g plus aztreonam, Group A: Group Group A: Group A: Group cefuroxime, 750 mg. Group B: Group further 8 h. doses every anaesthesia and three C: Group oral antibiotic prophylaxis. of preoperative three by 1.5 g at induction of anaesthesia followed doses of and two further doses of metronidazole B: Group Group B: Group ampicillin, 6.0 g i.v. plus doxycycline, 400 mg i.v. between was found surgery.Although in colorectal no difference regimen Elective colorectal surgery. colorectal Elective Cancer patients: 75%. plus neomycin, 1 g p.o. 48 h prior 6 h for every 5/18. Chest infection: 6/107 vs. 7/104 vs. 9/119. surgery. colorectal Elective Cancer patients: 84%. 1 g i.v. surgery. colorectal 30 min preoperatively,Elective and 8 h Cancer patients: NA. result: event [Adverse NA. Cost information: NA.] agents were antimicrobial Different but as separate simultaneously administered 8 vs. 2.WI plus septicaemia: 1 vs. 0.WI plus perineal infection: 0 v 131 132 133 Randomisation: Not clear.Blind outcome assessment: Yes. Follow-up: 15 days. Age: 95% > 51. SWI: Inflammation of the wound, a serous infusions, 1 h preoperatively. approximately Chest infection: 3/170 vs. 3/188. UTI: 8/170 vs. 4/1 Randomisation:True.Blind outcome assessment: No.Follow-up: 30 days. Age: 67 (20–89). SWI: of pus purulent fluid in Presence 2 g i.v. by at induction of anaesthesia followed 8 h and to operation with the last dose given Randomisation: Not clear. 12 h preoperatively, result: respectively, event [Adverse plus piperacillin, NA. Cost information:Blind outcome assessment: NA.] No. Age: 64 (SD 8). 16 h postoperatively. Rodolico, 1991 Reynolds, 1989 NA = not available or applicable Study and quality Surgery and definition of SWI Antibiotic regimensA vs. Results (Group B vs. Group C vs. Group D) Group 1-centre.Roland, 1986 culture. microbiological positive 16 h postoperatively. Note: calculation, in the table is a rough (in years) presented The age in any one trial, all groups across based on the available Withdrawal: 15%. or purulent discharge, and a positive 1-centre. upon incision. Withdrawal: 16%. or spontaneously appearing the wound further 8 h. three doses every Follow-up: NA. obt antibiotic therapy. were Excellent results SWI: the surgical Purulent discharge from 17-centres. the discharge. of a pathogen from culture Withdrawal: NA. discharge with a or a serous wound 80 mg i.v. and 8 h 30 min preoperatively surgery. colorectal 97 continued Appendix 3 continued s. 2/175. ally pneumonia): ally s other than SWI vs. 15/43. perforated Ileocolorectal Overall incidence of complications:Overall 130/470 vs. A single dose of latamoxef at induction of anaesthesia provides at induction of anaesthesia provides A single dose of latamoxef Short-term prophylactic treatment with cefoxitin is at with cefoxitin treatment Short-term prophylactic plus dose of cefotaxime A single preoperative Conclusion: tetracycline lavage. SWI: 32/453 vs. 33/454. Pneumonia: 17/177 vs. 19/175. Deaths: 6/177 vs. 12/175. (abscess, septicaemia, chest, urinary tract): 71/470 vs. 91/471. Authors’ conclusions surgery. colorectal antibiotics. least as efficient as 3-day treatment with ampicillin plus metronidazole with ampicillin plus metronidazole treatment least as efficient 3-day septic complications after elective of postoperative in the prevention postoperative and has practical advantages in eliminating the need for Conclusion: Conclusion: data rated out from other procedures) out from rated Peritoneal and parietal irrigation with and parietal irrigation Peritoneal Ampicillin, 1 g plus metronidazole, 500 mg result: event [Adverse Yes.Cost information: NA.] Cefuroxime, 1.5 g plus metronidazole, result: event [Adverse Yes.Cost information: NA.] Latamoxef, 1 g i.v. at induction of Cefoxitin, 2 g i.v. 15 min sepsis: Wound 13/177 vs. 14/175. Patients with septic complications: 20/177 vs. Cefotaxime, 1 g plus metronidazole, suppositories, with or without oral neomycin. A: Group result: event [Adverse anaesthesia. Yes.Cost information: NA.] B: Group of saline containing 1 g tetracycline at the one litre conclusion of the operation. WI than peritoneal and parietal against abdominal better prophylaxis Group A: Group A: Group Group B: Group B: Group doses of cefuroxime, 750 mg plus metronidazole,500 mg at 8 h and 16 postoperatively. of cefuroxime regimen is as efficacious a three-dose metronidazole in preventing plus metronidazole operation WI after colorectal * Elective colorectal surgery. colorectal Elective surgery. colorectal Emergency and elective abdominal surgery.Emergency and elective operations, ileocolorectal elective For patients Cancer patients: NA. not perforated. Ileocolorectal SWI: 6/39 metronidazole preoperative given were SWI: 3/7 vs. 4/8. Cancer patients: 76%. preoperatively, and at 4 h 10 after the 25/175.WI:2/23 Perineal vs. 5/33. Intraabdominal abscess: 1/177 v Cancer patients: 66%. 500 mg i.v. infusion after induction of anaesthesia 150/471. Deaths: 26/470 vs. 31/471. Death due to sepsis (usu 134 135 * 136 Indicates data for all surgical procedures included in a trial (i.e where data for elective colorectal operation cannot be sepa included operation elective colorectal in a trial all surgical procedures data for (i.e where Indicates data for Blind outcome assessment: Yes. Follow-up: 28 days.1-centre. SWI: Discharge of pus. Randomisation:True. Age: 63 (12–95). Randomisation:True.Blind outcome assessment: No. Follow-up: 30 days. Age: 68. SWI:Abscess or discharge of pus. Randomisation:True.Blind outcome assessment: Yes. Follow-up: 4-8 weeks. Age: 68 (16–92). first dose. SWI:Abdominal incision discharged pus. opening of peritoneum. and 5–10 min before 3/470 vs. 4/471. UTI: 30/470 vs. infection 45/471.Total Peritonitis: 1/177 vs. 3/175. Septicaemia: 4/177 vs. 4/175. UTI: 14/177 vs. 18/175. Rorbaek-Madsen, 1988 Rowe-Jones, 1990 Sauven, 1986 Study and quality Surgery and definition of SWI Antibiotic regimensA vs. Results (Group B vs. Group C vs. Group D) Group * Note: calculation, in the table is a rough (in years) presented The age in any one trial, all groups across based on the available NA = not available or applicable Withdrawal: 21%. 98 Withdrawal: 11%.6-centres.Withdrawal: 11%.14-centres. i.v. 6 h (ampicillin) and every 15 min preoperatively a total of 72 h. for and 8 h (metronidazole) 500 mg i.v. infusion after induction of anaesthesia peritoneal opening, and before further and two i.v. continued Health Technology Assessment 1998; Vol. 2: No. 7 continued o cost savings o cost savings Fistulae: 1/31 vs. 1/30 vs. 3/30. Pneumonia: 3/31 nical bowel cleansing and oral antimicrobial cleansing and oral antimicrobial nical bowel ts undergoing elective colorectal operations who colorectal elective ts undergoing s prophylaxis in elective colorectal surgery. colorectal in elective s prophylaxis ortant, the its potential to reduce both for an are that it is a single agent given in two doses, in two that it is a single agent given an are Anastomotic leakage: 4/96 vs. 3/101. Intraabdominal The addition of perioperative parenteral cefoxitin greatly reduced reduced greatly cefoxitin parenteral of perioperative The addition for of a single-dose antibiotic regimen The effectiveness plus as a combination of cefuroxime is as effective Cefotetan Perioperative administration of antibiotics dramatically and administration of antibiotics dramatically Perioperative vs. 1/30 vs. 0/29. Intraabdominal abscess: 1/31 vs. 1/30 vs. 0/30. Peritonitis: SWI: 12/31 vs. 1/30 vs. 2/29. SWI: 14/96 vs. 5/101. Conclusion: SWI: 7/36 vs. 3/34. Conclusion: Conclusion: Authors’ conclusions Conclusion: data rated out from other procedures) out from rated Parenteral antibiotics;Parenteral mezlocillin, 4 g or intravenously. of administration – intraluminally of the route regardless Intraluminal antibiotics; neomycin, 2/31 vs. 0/30 vs. 0/29. UTI: 2/31 vs. 3/30 vs. 1/29. Deaths: 2/31 vs. 0/30 vs. 0/29. Neomycin, base, 1 g plus erythromycin Cefoxitin, 2 g i.v. or i.m. 30–90min Cefuroxime, 1.5 g plus metronidazole, No antibiotics. Neomycin, base, 1 g plus erythromycin Cefotaxime, 1 g i.v. or i.m. 30–90min Cefotetan, 2 g at induction of WI (grade 3 and 4):10/68 vs. Total 5/36. Death: 7/68 vs. 3/36. Group A: Group B: Group A: Group B: Group A: Group B: Group A: Group B: Group then at 8 h and 16 postoperatively. Group C: Group Cancer patients: 90%. Elective colorectal surgery. colorectal Elective surgery. colorectal Elective Cancer patients: 46%. of the wound. suspect area a clinically over incision, 60 min before 6 h and 12 thereafter. abdominal surgery. Elective 1 g p.o. at 1 p.m., 11 p.m. 2 p.m.and on the day Cancer patients: NA. agents. abscess: 3/96 vs. 2/101. drained purulent material, it was considered than 24 h postoperatively.laboratory testing. surgery. colorectal Elective preoperatively.Cancer patients: NA. its contribution t strains and for emergence of resistant with or without dehiscence. anaesthesia and at 12 h postoperatively. result: event [Adverse Yes.Cost information: NA.] result: event [Adverse Yes.Cost information: NA.] doses each. three suppuration of the stab wound was suppuration of the stab wound WI. to be a not considered plus oxacillin, 2 g i.v. on induction of anaesthesia * 137 138 139 140 Indicates data for all surgical procedures included in a trial (i.e where data for elective colorectal operation cannot be sepa included operation elective colorectal in a trial all surgical procedures data for (i.e where Indicates data for Randomisation: Not clear. Age: 63. 5-centres. infected, of bacteriological or regardless the quality of patient care. without compromising Randomisation:True.Blind outcome assessment: Yes. Follow-up: 120 days. Age: 55 (18–96).Shatney, 1984 SWI: drainage; Purulent wound the exudate Randomisation: Not clear.Blind outcome assessment: No. Age: 44 (18–75). the operation. before Skipper, 1992 Randomisation: Not clear.Blind outcome assessment: No. Age: NA. result: event [Adverse Yes.Cost information: NA.] Schoetz, 1990 Schiessel, 1984 Study and quality Surgery and definition of SWI Antibiotic regimensA vs. Results (Group B vs. Group C vs. Group D) Group * Note: calculation, in the table is a rough (in years) presented The age in any one trial, all groups across based on the available NA = not available or applicable Blind outcome assessment: Yes.Follow-up: 28 days.Withdrawal: 18%. SWI: with Inflammation of the wound within discharge of pus occurring of to the last 3 litres added 900 mg/l were fluid. the irrigation 1 g/l plus bacitracin, 50,000 IU/l plus clindamycin, result: event [Adverse Yes.Cost information: NA.] Withdrawal: 14%. Withdrawal: 12%.1-centre. drained or was either spontaneously Follow-up: 30 days. 1 g p.o. at 1 p.m., 2 p.m. and 11 p.m. on the day of skin staples after removal expressed the operation, before WIs in patien and cefoxitin, the incidence of 2 g i.v. within SWI: If the incision or peritoneal cavity with mecha had been prepared no more 6 h for and every preoperatively Follow-up: 30 days.Withdrawal: 17%.1-centre. is imp surgical prophylaxis SWI: 1 – no infection; 2 – erythema; positive 3 – purulent or bacteriologically 750 mg plus metronidazole, 500 mg at 8 h and 500 mg at induction of anaesthesia, and cefuroxime, when used a metronidazole WI discharge; or 4 – severe The advantages of cefotet 16 h postoperatively. in agents given two are plus metronidazole cefuroxime whereas 1-centre. surgery. from 4 weeks By definition, surgery, colorectal WIs following the incidence of reduces significantly 99 continued Appendix 3 continued vs. 1/51. 1 vs. 0/51. vs. 36/158. Deaths within 42 days: 8 vs. 9. ion as the commonly used regimen of three of three used regimen ion as the commonly Abscess/peritonitis: 1/44 vs. 5/51 vs. 1/51. No advantage could be identified in the combination of oral Sulbactam improves the efficacy of piperacillin as prophylaxis the efficacy of piperacillin as prophylaxis Sulbactam improves A single dose of mezlocillin is as effective for the prophylaxis the prophylaxis for A single dose of mezlocillin is as effective SWI: 3/44 vs. 2/51 vs. 3/51. 5/168 vs. 1/158.WI plus septicaemia: 3/168 vs. 4/158. Intraabdominal abscess SWI: 16/54 vs. 14/56. result: event [Adverse NA. Cost information: NA.] Conclusion: Authors’ conclusions Conclusion: Conclusion: data Both p.o. and i.v. antibiotics. Piperacillin, 4 g plus sulbactam, 2 g i.v. as plus septicaemia: 1/168 vs. 0/158.WI plus septicaemia intraabscess: g plus metronidazole, Cefuroxime,1.5 Cefoxitin, 1 g i.v. at induction of surgery. colorectal antibiotics in elective and intravenous Neomycin, 1 g plus erythromycin Piperacillin, 4 g i.v. as single bolus dose WI alone: 30/168 vs. 27/158. Intraabdominal abscess alone: 5/168 vs. 1/158. Mezlocillin, 5.0 g i.v. at start of operation. Group C: Group Group A: Group A: Group B: Group A: Group B: Group Group B: Group the anaesthesia and at 6 h 12 following first dose. Plus p.o. placebo. Elective colorectal surgery. colorectal Elective Cancer patients: 84%. base, 1 g p.o. at 1 p.m., 2 p.m. and 11 p.m. Anastomotic major: 1/44 vs. 2/51 vs. 0/51. Leak minor: 0/44 vs. surgery. colorectal Elective 1/51 possible. whenever surgery. colorectal Elective Cancer patients: 80%. at induction of anaesthesia. Septicaemia alone: 10/168 vs. plus intraabdominal abscess: 3/158.Wound Cancer patients: 67%. made on clinical criteria with confirmation microbiological against operation but does little to protect colorectal in elective WI. staphylococcal to be evidence of failure of antibiotic of failure to be evidence diagnoses were prophylaxis.These 142 141 143 Stubbs, 1987 Randomisation:True.Blind outcome assessment: No. Age: 67 (SD 13). Randomisation: Not clear. Age: 66 (26–87). Randomisation:True.Blind outcome assessment: Yes.Follow-up: NA.Withdrawal: 14%. Age: 68 (16–91). SWI: with visible pus or Wound culture. a positive Plus i.v. placebo. patients prior to operation for on the day whose operation was scheduled at 8 a.m. Septicaemia: 1/44 vs. 2/51 vs. 0/51. Death: 0/44 vs. 2/5 result: event [Adverse NA. Cost information: NA.] Stellato, 1990 Study and quality Surgery and definition of SWI Antibiotic regimensA vs. Results (Group B vs. Group C vs. Group D) Group 1-centre.NA = not available or applicable Note: calculation, in the table is a rough (in years) presented The age in any one trial, all groups across based on the available discharge. was wound Stewart, 1995 Follow-up: 42 days.Withdrawal: 12%.13-centres. SWI: WI, intraabdominal infection within or septicaemia occurring bolus dose at induction of anaesthesia. considered of operation were 42 days Blind outcome assessment: No. 1/168 vs. infections: 0/158.Total 55/168 1994) Taylor, A was the same as in (Group 500 mg i.v. result: event [Adverse at start of operation, NA. two by followed Cost information: NA.] 1-centre. Follow-up: NA. SWI: of primary Failure healing in any doses of cefuroxime, 750 mg plus metronidazole, operat WI after large bowel of 100 Withdrawal: 5%. portionor when there of the wound 500 mg i.v. 8 h and 16 later. plus metronidazole. doses of cefuroxime continued Health Technology Assessment 1998; Vol. 2: No. 7 vs. continued 59 vs. 5/168. nge of surgical dditional dditional bidity: 23/159 vs. 55/168. Death Intraabdominal abscess: 2 vs. 5.Anastomotic leak: Co-amoxiclav has been shown to be comparable to the regimen to be comparable the regimen has been shown Co-amoxiclav The intravenous administration of latamoxef at induction of administration of latamoxef The intravenous The administration of ciprofloxacin, with the 500 mg twice daily SWI: 17/84 vs. 9/36. SWI: 3/60 vs. 5/60. postoperative complications:postoperative 7 vs. 21. Deaths within 30 days: 1 vs. 5. 5/168.Wound plus septicaemia:5/168.Wound 1/159 vs. 3/168. Intraabdominal abscess plus Authors’ conclusions such operations. for preparations septic complications. preventing Conclusion: Conclusion: [Adverse event result: event [Adverse NA. Cost information: Yes.] Conclusion: after the incidence of infection cathartic reduces significantly preoperative part operations and should form of the preoperative colorectal elective anaesthesia provides good serum and tissue levels during colorectal operations. during colorectal serum and tissue levels good anaesthesia provides in as a combination of cephazolin plus metronidazole It is at least as effective data rated out from other procedures) out from rated Cephradine, 1 g i.v. plus metronidazole, single agent, the combination regimen. unlike Cephazolin, 1 g plus metronidazole, result: event [Adverse Yes.Cost information: NA.] No oral antibiotic. B was the same as in Stewart, (Group 1995) Latamoxef, 1 g i.v. at induction of Ciprofloxacin, 500 mg b.d. p.o. on septicaemia: 0/159 vs. 1/168.WI plus septicaemia intraabdominal abscess: Co-amoxiclav, 1.2 g i.v. plus placebo Group B: Group A: Group Group B: Group Group A: Group Group A: Group Group B: Group * Elective colorectal surgery. colorectal Elective Cancer patients: 79%. All patients: Single dose of piperacillin, 4 g i.v. at induction of anaesthesia. WI alone: 17/159 vs. 30/168. Intraabdominal abscess alone: 2/1 general or gynaecological Elective surgery.Cancer patients: at induction of anaesthesia, First dose was given NA. Septicaemia alone: 3/158 vs. 10/168.WI plus intraabdominal abscess: 0/159 (in one centre, mild erythema was classed which an organism was discharge from isolated; discharge associated 4 – serous and then 8 h 16 after the first dose. 1 g suppository. with either erythema or other factors e.g. of infection indicative pyrexia. surgery. colorectal Elective result: event [Adverse Yes.Cost information: NA.] procedures. of administration a the convenience offers Co-amoxiclav Cancer patients: 75%. anaesthesia, and at 6 h 12 postoperatively. 1/49 vs. 3/52. antibiotics: postoperative Additional 4 vs. 12. A * 146 144 145 Indicates data for all surgical procedures included in a trial (i.e where data for elective colorectal operation cannot be sepa included operation elective colorectal in a trial all surgical procedures data for (i.e where Indicates data for Randomisation: Not clear. Age: 51. 13-centres.Thomas, 1985 WI);2 as – purulent discharge; 3 – serous Randomisation: Not clear.Blind outcome assessment: Yes. Age: 64 (32–92). Randomisation:True.Blind outcome assessment: No. Age: 67 (SD 13). alr 1994 Taylor, Tehan,1989 Study and quality Surgery and definition of SWI Antibiotic regimensA vs. Results (Group B vs. Group C vs. Group D) Group Blind outcome assessment: Yes. * Note: calculation, in the table is a rough (in years) presented The age in any one trial, all groups across based on the available NA = not available or applicable Follow-up: 42 days.Withdrawal: NA. SWI: Not defined. 500 mg i.v. at induction of anaesthesia, and at 6 h and 12 postoperatively. Withdrawal: 14%.13-centres. clinical criteria with microbiological it was available.) confirmation whenever within 42 days: 7/189 vs. 13/192. Follow-up: 42 days. SWI: Not defined. (Diagnosis based on operation. before the day 0/159 vs. mor patients with infective 1/168.Total Follow-up: 28 days. SWI: erythema 1 – moderate or severe suppository. a wide ra for as prophylaxis of cephradine plus metronidazole Withdrawal: 16%. 1-centre. 101 continued Appendix 3 continued 5. iotic m drain: 0/75 vs. 1/75. Stomal and it might be wise to combine tion rate further. in vivo = 0.055, exact test). Fisher’s p ( Both prophylactic antibiotic schemes were comparably effective. comparably antibiotic schemes were Both prophylactic Cefotetan was a useful drug in colorectal surgery, was a useful drug in colorectal Cefotetan but should be in in a significant reduction has resulted cefoxitin Preoperative The antibiotic cover in elective colorectal operation with colorectal in elective The antibiotic cover spp. is rather disappointing Conclusion: sepsis: 4/75 vs. 5/75.WI:3/6 Perineal vs. 4/9. Intraabdominal abscess: SWI: 3/13 vs. 11/19 Bacteroides Bacteroides Conclusion: Conclusion: Authors’ conclusions surgery. colorectal elective patients undergoing for events adverse with a nitroimidazole. cefotetan Conclusion: metronidazole plus amikacin.Themetronidazole short-term ornidazole plus ceftriaxone constitutes an effective, antibiotic prophylaxis from free therapy inexpensive data Piperacillin, 4 g plus tinidazole, Metronidazole, 1.5 g i.v. infusion at 8/75 vs. 7/75.Anastomotic dehiscence: 4/75 vs. 3/75. Superficial wound Placebo. Ornidazole, 1 g plus ornidazole plus ceftriaxone, preoperatively both administered Piperacillin, 4 g plus tinidazole, 0.8 g i.v. SWI: 4/61 vs. 5/63. infection: postoperative Overall 23% vs. 19%. Metronidazole, 500 mg plus with visible pus: Wound 1/25 vs. 2/25.Anastomotic leakage: 1/25 vs. 0/25. Cefotetan, doses. 2–3 g two WI:10/75 Abdominal vs. 17/75. Pus in incision: 3/75 vs. 8/75. Open-wound Cefoxitin, 2 g i.v. as three 6 g given 5/3 mg) every 8 h for 5 days. 8 h for 5/3 mg) every sepsis: 2/75 vs. 2/75. Chest/upper respiratory tract infection: 3/75 vs. 5/75. × Group A: Group Group B: Group Group A: Group A: Group B: Group A: Group Group B: Group Group B: Group every 24 h postoperatively for 2 days. for 24 h postoperatively every with a classic antibiotic cover from significantly that do not differ results Cancer patients: translation. NA from operation. before information: result/cost event [Adverse translation.] NA from Elective colorectal surgery. colorectal Elective Emergency and complicated colorectal operation. Colorectal obvious inflammatory and pus. reaction Cancer patients: 80%. amikacin, 500 mg i.v. and 2 h preoperatively Pulmonary embolism: 0/25 vs. 1/25. thrombosis: Deep vein 0/25 vs. 1/2 surgery. Cancer patients: NA. doses, skin incision. the first before result: event [Adverse NA. Cost information: NA.] and no prolongation in hospital stay.and no prolongation used in combination with vitamin K malnourished patients. Its activity against dehiscence: minor – no serious morbidity (German) surgery. colorectal Elective 147 148 149 150 Randomisation: Not clear.Blind outcome assessment: No. Age: 41% > 71 years. NA = not available or applicable Randomisation: Not clear. Cancer patients: 0.0%. Randomisation: Not clear.Blind outcome assessment: Yes. Age: 64. Withdrawal: 1.3%.1-centre. SWI: Pus in the incision. Major – associated with constitutional disturbance and result: event [Adverse Yes.Cost information: NA.] Tsimoyiannis, 1991 Tuchmann,1988 Utley, 1984 Study and quality Surgery and definition of SWI Antibiotic regimensA vs. Results (Group B vs. Group C vs. Group D) Group Withdrawal: 0%. 1988 Tudor, 24 h postoperatively. 12 h (tinidazole) for and every Note: calculation, in the table is a rough (in years) presented The age in any one trial, all groups across based on the available Follow-up: NA. SWI: Not defined. 0.8 g i.v. operation, before 8 h (piperacillin) and every the infec could decrease Intestinal lavage Randomisation: Not clear.Blind outcome assessment: No.Follow-up: NA. Age: 61 (SD 15). SWI: Obvious collection of pus, 12 h or every 8 h (metronidazole) every postoperatively. 2 days (amikacin) for result: event [Adverse Yes.Cost information: Yes.] Blind outcome assessment: No.Follow-up: 28 days. Age: NA. startin of operation plus gentamicin (weight Withdrawal: NA.1-centre. dehiscence: 2/75 vs. 0/75. Purulent discharge fro with purulent discharge; of wound severe kg – infection: with an deep purulent infection use in general surgical procedures. prophylactic for Follow-up: 42 days. SWI: Moderate – superficial inflammation antib broad-spectrum to be an effective WIs in this study and appears Withdrawal: NA. or which drained spontaneously 102 1-centre. on incision ceftriaxone, 2 g i.v. and 2 h preoperatively postoperatively, clinical 2 days and per 24 h for provides continued Health Technology Assessment 1998; Vol. 2: No. 7 s. 7/30. continued ncidence epsis: 14/29 a beta-lactamase inhibitor Fever: 6/30 vs. 4/30. Intraabdominal abscess: 2/30 vs. 2/30. From this small study it cannot be concluded that the From No significant differences were found in the results.All found were the No significant differences The incidence of infection was similar in the two groups.The was similar in the two The incidence of infection antibiotics is with appropriate Single-dose systemic prophylaxis Authors’ conclusions pathogens. aerobic bacterial contamination by the normal microflora of the colon. the normal microflora elective colorectal operation can be omitted. colorectal elective SWI: 6/30 vs. 10/30. Respiratory tract infection: 1/30 vs. 1/30. Sepsis: 1/30 vs. 0/30. Deaths: 0/30 vs. 1/30. [Adverse event result: event [Adverse NA. Cost information: NA.] Chest infection: 5/108 vs. 8/105. UTI: 13/108 vs. 7/105. Conclusion: Conclusion: Conclusion: Conclusion: data Ornidazole plus gentamicin as for Ornidazole plus gentamicin as for result: event [Adverse Yes.Cost information: NA.] Netilmicin, 120 mg i.v. at induction result: event [Adverse NA. Cost information: NA.] Ceftriaxone, 2 g plus metronidazole, Metronidazole, 500 mg i.v. 8 h every Piperacillin, 2 g i.v. at induction of WI:17/108 Abdominal vs. 13/105. Major: 3/108 vs. 1/105. Minor: 14/108 vs. sulfate,Neomycin 1 g plus WI,major: Abdominal 11/29 vs. 2/31, minor: 1/29 vs. 1/31.WI:4 Perineal vs. 0. Gentamicin, 80 mg plus metronidazole,WI:4/21 Superficial vs. 5/20. Intraabdominal abscess: 1/21 vs. 0/20. Ornidazole, 1 g i.v. plus gentamicin, Group A: Group A: Group A: Group A: Group Group B: Group Group B: Group B: Group B: Group ). 3 Elective colorectal surgery. colorectal Elective hospital. discharge from or delayed surgery. colorectal Elective 20,000 cells/mm surgery. colorectal Elective surgery. colorectal Elective Cancer patients: 62%. 80 mg i.v. 45 mins preoperatively. Anastomosis insufficiency: 1/30 vs. 5/30. Peritonitis: 1/30 vs. 3/30. UTI: 6/30 v Cancer patients: 70%. anaesthesia, and at 8 h 16 Cancer patients: 70%. 12/105.WI:1 Perineal vs. 5. abscess: Subphrenic 1/108 vs. 0/105. base, erythromycin 1 g p.o. at 1 p.m., 2 p.m.Cancer patients: NA. Abscess: 1/29 vs. 0/31. Septicaemia: 2/29 vs. with s number 0/39.Total 500 mg i.v. starting at premedication, and then Symptomatic anastomotic dehiscence: 1/21 vs. 0/20. dehiscence; then greater a hospital stay 14 days; than (more or leukocytosis a single-dose, high serum agents that provide long-acting antimicrobial period without disturbing concentrations during the postoperative on bacteriological culture.WIs were threatened major when they considered life, further surgical intervention required postoperatively. induction of anaesthesia and at 12 h of bacteriological findings suggest that the addition effective, it more render to piperacillin would the i in reducing not only its activity against Enterobacteriaceae. WIs but also improving of staphylococcal 153 154 152 151 akr 1988 Walker, NA = not available or applicable Weidema, 1985 Wenzel, 1985 Study and quality Surgery and definition of SWI Antibiotic regimensA vs. Results (Group B vs. Group C vs. Group D) Group 1-centre. Note: calculation, in the table is a rough (in years) presented The age in any one trial, all groups across based on the available evr 1986 Weaver, Randomisation: Not clear.Blind outcome assessment: No. Age: 70 (45–84). Randomisation: Not clear.Blind outcome assessment: No. Follow-up: NA. Age: NA. SWI: Discharge of pus. was Major infection and 11 p.m. the operation. before on the day vs. 3/31. with major sepsis: number Total 13/29 vs. 2/31. Randomisation:True.Blind outcome assessment: No. Age: 64 (23–89). postoperatively.Randomisation:True.Blind outcome assessment: No. Follow-up: 28 days. abscess: Pelvic 3/108 vs. 0/105. plus septicaemia: Peritonitis 1/108 vs. Age: 0/105. 59 (SD 14). SWI: of pus in the wound. Presence 24 h. 8 h for every result: event [Adverse NA. Cost information: NA.] Follow-up: NA. SWI: the wound, Discharge of frank pus from Follow-up: NA.Withdrawal: NA. SWI: wound Oedematous and/or red with a purulent secretion.A, further Group three doses of by followed ornidazole, doses 12 h and three 500 mg every Withdrawal: 11%.7-centres. dehiscence, wound or the discharge of serous of anaesthesia and at 8 h 16 postoperatively, isolated which organisms were fluid from plus metronidazole, 500 mg i.v. infusion at Withdrawal: NA. defined as purulent discharge associated i.v. 1.5 g slow infusion at the start of the Withdrawal: NA. 24 h, for starting at premedication. before administration of gentamicin in the antibiotic prophylaxis 1-centre. of gentamicin, 8 h. 80 mg every due to were postoperatively complications which occurred infectious 1-centre. with: fever; discharge; foul-smelling operation. and erythromycin. superior to oral neomycin to use effective It is more 103 continued Appendix 3 < 0.05). p = 0.28). Hospitalisation days: 16.7 vs. 20.3 ( p Combination of cephazolin plus ornidazole shows better results better results Combination of cephazolin plus ornidazole shows Ceftriaxone proved its efficiency in infection prophylaxis prophylaxis its efficiency in infection Ceftriaxone proved [Adverse event result: event [Adverse Yes.Cost information: translation.] NA from SWI: 1/30 vs. 2/30. result: event [Adverse Yes.Cost information: NA.] Authors’ conclusions dosage.practical advantages of a once-daily difference was not statistically significant. was not statistically difference Conclusion: Conclusion: when compared with cephazolin alone in reducing SWI rates. with cephazolin alone in reducing when compared However, the gentamicin, in the clinical response and metronidazole.There was no difference antibiotic regimens, the two between but ceftriaxone has the considerable data Cephazolin, 2 g i.v. on anaesthesia. Cefotiam, doses, 1 g three gentamicin, Cephazolin, 2 g plus ornidazole, 1 g i.v. SWI: 4/50 vs. 10/50 ( Ceftriaxone, 2 g short i.v. infusion. Group B: Group Group A: Group Group A: Group B: Group Cancer patients: translation. NA from on anaesthesia. B. 5 deaths in group Elective colorectal surgery. colorectal Elective Cancer patients: NA. (German) surgery. colorectal Elective 155 156 Randomisation:True.Blind outcome assessment: Yes. Age: 68. Randomisation: Not clear. Age: 63. Wohlfart, 1987 Zuber, 1989 Study and quality Surgery and definition of SWI Antibiotic regimens1-centre. A vs. Results (Group B vs. Group C vs. Group D) Group NA = not available or applicable Note: calculation, in the table is a rough (in years) presented The age in any one trial, all groups across based on the available Follow-up: NA. SWI: of pus or abscess. Presence Blind outcome assessment: No. doses, 80 mg two plus metronidazole, 500 mg Withdrawal: 5%. 104 Follow-up: NA.Withdrawal: NA. 1-centre. SWI: Not defined. doses. two colon operation in comparison with a potent combination dose of cefotiam, for continued Health Technology Assessment 1998; Vol. 2: No. 7

Appendix 4 Number of comparisons between antibiotics in included trials

105

Appendix 4

All others All 2 1 2 1

Control

Tinidazole

Ticarcillin/clavulanic acid Ticarcillin/clavulanic

11

Piperacillin

Netilmicin + metronidazole + Netilmicin 11 3

Neomycin + erythromycin + Neomycin

23

Latamoxef

Mezlocillin

1111 114

Mezlocillin + metronidazole + Mezlocillin

Metronidazole

21 112 Gentamicin + metronidazole + Gentamicin

21 11 1 Gentamicin + clindamycin + Gentamicin

2 21118

Doxycycline 111 12211 1 1 2 115 Co-amoxiclav

12 35 142 3 21 1 111431115 21111 1 2 131 1 11 1 2

Ciprofloxacin + metronidazole + Ciprofloxacin 4 1111 11 12 1 21 11 1 1 2 1 3

Cefuroxime + metronidazole + Cefuroxime

4 1

Ceftriaxone

Ceftriaxone + metronidazole + Ceftriaxone

Cefoxitin

Cefoxitin + neomycin + erythromycin + neomycin + Cefoxitin

Cefotetan

Cefotaxime

Cefotaxime + metronidazole + Cefotaxime

Cephazolin + metronidazole + Cephazolin

Aztreonam + clindamycin + Aztreonam Ampicillin + metronidazole + Ampicillin 21111

106 Ampicillin + metronidazole + clindamycinAztreonam Cephazolin + metronidazole + metronidazoleCefotaxime CefotaximeCefotetan + erythromycin + neomycin Cefoxitin CefoxitinCeftriaxone + metronidazole 1Ceftriaxone + metronidazoleCefuroxime 1 3 1 + metronidazoleCiprofloxacin 2Co-amoxiclav 1 1 1Doxycycline 1 1 Gentamicin + clindamycin 1 2 1Gentamicin + metronidazoleMetronidazole 1 1 2 1 Mezlocillin + metronidazole 1 1 1 2Mezlocillin 2 1 1 Latamoxef 2 1 + erythromycinNeomycin 1 1 2Netilmicin + metronidazole 1 2 2Piperacillin 2 1 acidTicarcillin/clavulanic Tinidazole Control 1 1 1All others 1 1 1 2 1 1 1 1 1 2 1 1 1 2 1 2 4 1 2 3 1 3 8 1 1 2 1 5 Number of comparisons between antibiotics in included trials Number of comparisons between Health Technology Assessment 1998; Vol. 2: No. 7

HTA panel membership

This report was identified as a priority by the Pharmaceutical Panel.

Acute Sector Panel Chair: Professor John Farndon, University of Bristol †

Professor Senga Bond, Professor Richard Ellis, Mr Ian Hammond, Professor John Norman, University of Newcastle- St James’s University Hospital, Bedford & Shires Health & University of Southampton upon-Tyne † Leeds Care NHS Trust Dr John Pounsford, Professor Ian Cameron, Mr Leonard Fenwick, Professor Adrian Harris, Frenchay Hospital, Bristol † Southeast Thames Regional Churchill Hospital, Oxford Freeman Group of Hospitals, Professor Gordon Stirrat, Health Authority † Professor Robert Hawkins, Newcastle-upon-Tyne St Michael’s Hospital, University of Bristol † Ms Lynne Clemence, Dr David Field, Bristol † Dr Chris McCall, Mid-Kent Health Care Trust Leicester Royal Infirmary † General Practitioner, Dorset † Professor Michael Sheppard, Professor Francis Creed, Queen Elizabeth Hospital, Ms Grace Gibbs, Professor Alan McGregor, University of Manchester † Birmingham † West Middlesex University St Thomas’s Hospital, London † Professor Cam Donaldson, Hospital NHS Trust Mrs Wilma MacPherson, Dr William Tarnow-Mordi, University of Aberdeen University of Dundee Dr Neville Goodman, St Thomas’s & Guy’s Hospitals, Mr John Dunning, Southmead Hospital London Professor Kenneth Taylor, Papworth Hospital, Services Trust, Professor Jon Nicoll, Hammersmith Hospital, Cambridge † Bristol † University of Sheffield † London

Diagnostics and Imaging Panel Chair: Professor Mike Smith, University of Leeds †

Professor Michael Maisey, Dr Mansel Hacney, Professor Chris Price, Mr Stephen Thornton, Guy’s & St Thomas’s Hospitals, University of Manchester London Hospital Medical Cambridge & Huntingdon London * School † Health Commission Professor Sean Hilton, Professor Andrew Adam, St George’s Hospital Dr Ian Reynolds, Dr Gillian Vivian, † UMDS, London Medical School, Nottingham Health Authority Royal Cornwall Hospitals Trust † Dr Pat Cooke, London Professor Colin Roberts, Dr Jo Walsworth-Bell, RDRD, Trent Regional University of Wales College South Staffordshire Mr John Hutton, Health Authority of Medicine † Health Authority † MEDTAP International Inc., Ms Julia Davison, London † Miss Annette Sergeant, Dr Greg Warner, St Bartholomew’s Hospital, Chase Farm Hospital, General Practitioner, † London Professor Donald Jeffries, Enfield Hampshire † Professor Adrian Dixon, St Bartholomew’s Hospital, Professor John Stuart, † University of Cambridge † London University of Birmingham Professor MA Ferguson-Smith, Dr Andrew Moore, Dr Ala Szczepura, University of Cambridge † Editor, Bandolier † University of Warwick †

Methodology Panel Chair: Professor Martin Buxton, Brunel University †

Professor Anthony Culyer, Dr Rory Collins, Mr Philip Hewitson, Dr Maurice Slevin, University of York * University of Oxford Leeds FHSA St Bartholomew’s Hospital, Dr Doug Altman, Institute of Professor George Davey-Smith, Professor Richard Lilford, London Health Sciences, Oxford † University of Bristol Regional Director, R&D, Dr David Spiegelhalter, † Professor Michael Baum, Dr Vikki Entwistle, West Midlands Institute of Public Health, † Royal Marsden Hospital University of Aberdeen † Mr Nick Mays, King’s Fund, Cambridge † Professor Nick Black, Professor Ray Fitzpatrick, London Professor Charles Warlow, London School of Hygiene University of Oxford † Professor Ian Russell, Western General Hospital, † † Edinburgh † & Tropical Medicine Professor Stephen Frankel, University of York Professor Ann Bowling, University of Bristol Professor David Sackett, * University College London Dr Stephen Harrison, Centre for Evidence Based Previous Chair † † † Medical School University of Leeds Medicine, Oxford Current members 109 continued HTA panel membership

continued Pharmaceutical Panel Chair: Professor Tom Walley, University of Liverpool †

Professor Michael Rawlins, Mr Barrie Dowdeswell, Ms Sally Knight, Dr Frances Rotblat, University of Newcastle- Royal Victoria Infirmary, Lister Hospital, Stevenage † Medicines Control Agency † * upon-Tyne Newcastle-upon-Tyne Dr Andrew Mortimore, Mrs Katrina Simister, Dr Colin Bradley, Dr Desmond Fitzgerald, Southampton & SW Hants Liverpool Health Authority † † University of Birmingham Mere, Bucklow Hill, Health Authority Dr Ross Taylor, Cheshire Professor Alasdair Mr Nigel Offen, Essex Rivers University of Aberdeen † Dr Alistair Gray, Healthcare, Colchester † Breckenridge, RDRD, Dr Tim van Zwanenberg, Health Economics Research Northwest Regional Dr John Posnett, Northern Regional Health Unit, University of Oxford † Health Authority University of York Authority Professor Keith Gull, Ms Christine Clark, Mrs Marianne Rigge, Dr Kent Woods, RDRD, † University of Manchester Hope Hospital, Salford The College of Health, Trent RO, Sheffield † † Mrs Julie Dent, Dr Keith Jones, London Ealing, Hammersmith Medicines Control Agency Mr Simon Robbins, & Hounslow Health Authority, Professor Trevor Jones, Camden & Islington London ABPI, London † Health Authority, London †

Population Screening Panel Chair: Professor Sir John Grimley Evans, Radcliffe Infirmary, Oxford †

Dr Sheila Adam, Dr Tom Fahey, Professor Alexander Markham, Dr Sarah Stewart-Brown, Department of Health* University of Bristol † St James’s University Hospital, University of Oxford † † Ms Stella Burnside, Mrs Gillian Fletcher, Leeds Ms Polly Toynbee, Altnagelvin Hospitals Trust, National Childbirth Trust † Professor Theresa Marteau, Journalist † Londonderry † UMDS, London Professor George Freeman, Professor Nick Wald, Dr Carol Dezateux, Institute of Charing Cross & Westminster Dr Ann McPherson, University of London † Child Health, London † Medical School, London General Practitioner, † Professor Ciaran Woodman, Dr Anne Dixon Brown, Dr Mike Gill, Brent & Harrow Oxford † Centre for Cancer NHS Executive, Health Authority Professor Catherine Peckham, Epidemiology, Manchester Anglia & Oxford † Dr JA Muir Gray, RDRD, Institute of Child Health, † Professor Dian Donnai, Anglia & Oxford RO London St Mary’s Hospital, Dr Ann Ludbrook, Dr Connie Smith, Manchester † University of Aberdeen † Parkside NHS Trust, London

Primary and Community Care Panel Chair: Dr John Tripp, Royal Devon & Exeter Healthcare NHS Trust †

Professor Angela Coulter, Professor Shah Ebrahim, Mr Edward Jones, Professor Dianne Newham, King’s Fund, London * Royal Free Hospital, London Rochdale FHSA King’s College London Professor Martin Roland, Mr Andrew Farmer, Professor Roger Jones, Professor Gillian Parker, University of Manchester * Institute of Health Sciences, UMDS, London University of Leicester † Oxford † Dr Simon Allison, Mr Lionel Joyce, Dr Robert Peveler, † University of Nottingham Ms Cathy Gritzner, Chief Executive, Newcastle City University of Southampton The Patients’ Association † Health NHS Trust Mr Kevin Barton, Dr Mary Renfrew, Professor Martin Knapp, East London & City Professor Andrew Haines, University of Oxford London School of Economics Health Authority † RDRD, North Thames Ms Hilary Scott, & Political Science Regional Health Authority Tower Hamlets Healthcare Professor John Bond, Professor Karen Luker, NHS Trust, London † University of Newcastle- Dr Nicholas Hicks, University of Liverpool upon-Tyne † Oxfordshire Health Authority † Professor David Mant, Ms Judith Brodie, Professor Richard Hobbs, NHS Executive South & West † Age Concern, London † University of Birmingham † Dr Fiona Moss, North Thames Dr Nicky Cullum, Professor Allen Hutchinson, British Postgraduate * Previous Chair University of York † University of Sheffield † Medical Federation † † Current members

110 National Coordinating Centre for Health Technology Assessment, Advisory Group Chair: Professor John Gabbay, Wessex Institute for Health Research & Development †

Professor Mike Drummond, Professor James Raftery, Professor Andrew Stevens, Centre for Health Economics, Health Economics Unit, Department of Public Health University of York † University of Birmingham † & Epidemiology, † Ms Lynn Kerridge, Dr Paul Roderick, University of Birmingham Wessex Institute for Health Research Wessex Institute for Health Research & Development † & Development Dr Ruairidh Milne, Professor Ian Russell, Wessex Institute for Health Research Department of Health, Sciences & Clinical & Development † Evaluation, University of York † Ms Kay Pattison, Dr Ken Stein, Research & Development Directorate, Wessex Institute for Health Research NHS Executive † & Development † † Current members Health TechnologyAssessment 98 o.2 o 7 No. 2: Vol. 1998;

Copies of this report can be obtained from:

The National Coordinating Centre for Health Technology Assessment, Mailpoint 728, Boldrewood, University of Southampton, Southampton, SO16 7PX, UK. Fax: +44 (0) 1703 595 639 Email: [email protected] http://www.soton.ac.uk/~hta ISSN 1366-5278