Appendix 1: MeSH search terms and data collect form used to collect

Appendix 1: MeSH search terms and data collect form used to collect data: CENTRAL on 31 March 2015 was searched using the following exploded MeSH headings and keywords: #1 MeSH descriptor Vancomycin explode all trees #2 MeSH descriptor Cephalosporins explode all trees #3 MeSH descriptor Ciprofloxacin explode all trees #4 MeSH descriptor Ofloxacin explode all trees #5 MeSH descriptor Aztreonam explode all trees #6 MeSH descriptor Trimethoprim-Sulfamethoxazole Combination explode all trees #7 MeSH descriptor Oxazolidinones explode all trees #8 (antibiotic* or antibacterial* or antimicrobial* or cefazolin or cefepime or vancomycin or aztreonam or ciprofloxacin or levaquin or trimethoprim or linezolid):ti,ab,kw #9 MeSH descriptor Anti-Infective Agents, Local explode all trees #10 antiseptic*:ti,ab,kw #11 MeSH descriptor Iodophors explode all trees #12 MeSH descriptor Chlorhexidine explode all trees #13 MeSH descriptor Povidone-Iodine explode all trees #14 MeSH descriptor Alcohols explode all trees #15 (iodophor* or povidone or iodine or chlorhexidine or betadine or alcohol*):ti,ab,kw #16 (#1 OR #2 OR #3 OR #4 OR #5 OR #6 OR #7 OR #8 OR #9 OR #10 OR #11 OR #12 OR #13 OR #14 OR #15) #17 MeSH descriptor Surgical Wound Infection explode all trees #18 MeSH descriptor Surgical Wound Dehiscence explode all trees #19 (surg* NEAR/5 infect*):ti,ab,kw #20 (surg* NEAR/5 wound*):ti,ab,kw #21 (surg* NEAR/5 site*):ti,ab,kw #22 (surg* NEAR/5 incision*):ti,ab,kw #23 ((postoperative post-operative or postoperative) NEXT (wound NEXT infection*)):ti,ab,kw #24 (#17 OR #18 OR #19 OR #20 OR #21 OR #22 OR #23) #25 MeSH descriptor Replacement, Arthroplasty explode all trees The above strategy was also used to search Ovid MEDLINE, Ovid EMBASE and EBSCO CINAHL. We also combined the Ovid MEDLINE search with the Cochrane Highly Sensitive Search Strategy for identifying randomized trials in MEDLINE: sensitivity- and precisionmaximizing version 1. The EMBASE and CINAHL searches were also combined with the trial filters developed by the Scottish Intercollegiate Guidelines Network 2. There were no restrictions on the basis of date or language of publication.

1

Data extraction form: Name of person/reviewer extracting data:

Author of article: Title: Source (e.g. Journal title): Date of study: Study location (geographical): Care setting (e.g. hospital): Inclusion/exclusion criteria (list of patient inclusion and exclusion criteria) Inclusion: Exclusion:

Sample size: number in each arm of trial a priori power calculation?

YES

NO

NOT STATED

trial powered adequately?

Patient baseline characteristics: • • •

age range: gender: medical condition(s):

TRIAL DESIGN DETAILS: single-centre/multicentre trial?

Study type randomized controlled trial/matched control/unmatched concurrent control/historic control:

Allocation 2

• • •

was it random? method of randomisation: was it concealed?

YES

NO

NOT STATED

YES

NO

NOT STATED

Intervention details • • •

care setting: treatment group(s): control(s):

co-interventions: duration of intervention: • • •

who delivered intervention? was the provider blinded? was the patient blinded?

YES NO YES NO

NOT STATED NOT STATED

Outcome measures • • • • • • •

what were they? methods of assessing outcome measures: blind assessment? YES when were they measured? validity of assessment: inter-assessor reliability: length of follow-up:

NO

NOT STATED

Costs • • •

considered? cost-effectiveness details: Results:

YES

NO

NOT STATED

(note: if the study is an economic evaluation then a separate assessment procedure, used by the NHS Centre for Reviews & Dissemination for such economic evaluations, will be followed)

Analysis: • • • •

description of analysis employed: statistical methods: comparisons made: intention to treat analysis? 3

• • •

adjustment for confounding? subgroups considered: exploration of heterogeneity:

Results: Missing data:

length of follow-up:

withdrawals/drop outs - are proportion and characteristics of participants lost to follow-up comparable for the study groups at the end of the trial?

reasons for withdrawal:

lost to follow-up: Number of infections (primary outcome): • • • •

Intervention arm (1): Intervention (or control) arm (2): Intervention arm (if more than 2 intervention arms are included in the trial): Intervention are (if more than 2 intervention arms are included in the trial):

Number of adverse events: Intervention arm (1): • • •

Intervention (or control) arm (2): Intervention arm (if more than 2 intervention arms are included in the trial): Intervention are (if more than 2 intervention arms are included in the trial):

Conclusions: Implications (e.g. for practice): Other comments: Methodological quality of study:

comparability of intervention: 4

baseline comparability:

5

Criteria for assessing potential sources of bias 1. Was the allocation sequence randomly generated? Low risk of bias

The investigators describe a random component in the sequence generation process such as: referring to a random number table; using a computer random number generator; coin tossing; shuffling cards or envelopes; throwing dice; drawing of lots. High risk of bias

The investigators describe a non-random component in the sequence generation process. Usually, the description would involve some systematic, non-random approach, for example: sequence generated by odd or even date of birth; sequence generated by some rule based on date (or day) of admission; sequence generated by some rule based on hospital or clinic record number. Unclear

Insufficient information about the sequence generation process provided to permit judgement of low or high risk of bias. 2. Was the treatment allocation adequately concealed? Low risk of bias

Participants and investigators enrolling participants could not foresee assignment because one of the following, or an equivalent method, was used to conceal allocation: central allocation (including telephone, web-based and pharmacy-controlled randomisation); sequentially numbered drug containers of identical appearance; sequentially numbered, opaque, sealed envelopes. High risk of bias

Participants or investigators enrolling participants could possibly foresee assignments and thus introduce selection bias, such as allocation based on: using an open random allocation schedule (e.g. a list of random numbers); assignment envelopes used without appropriate safeguards (e.g. if envelopes were unsealed or nonopaque or not sequentially numbered); alternation or rotation; date of birth; case record number; any other explicitly unconcealed procedure. Unclear

Insufficient information provided to permit judgement of low or high risk of bias. This is usually the case if the method of concealment is not described or not described in sufficient detail to allow a definite judgement, for example if the use of assignment envelopes is described, but it remains unclear whether envelopes were sequentially numbered, opaque and sealed. 6

3. Blinding - was knowledge of the allocated interventions adequately prevented during the study? Low risk of bias

Any one of the following. • • •

No blinding, but the review authors judge that the outcome and the outcome measurement are not likely to be influenced by lack of blinding. Blinding of participants and key study personnel ensured, and unlikely that the blinding could have been broken. Either participants or some key study personnel were not blinded, but outcome assessment was blinded and the non-blinding of others was unlikely to introduce bias.

High risk of bias

Any one of the following. • • •

No blinding or incomplete blinding, and the outcome or outcome measurement is likely to be influenced by lack of blinding. Blinding of key study participants and personnel attempted, but likely that the blinding could have been broken. Either participants or some key study personnel were not blinded, and the non-blinding of others was likely to introduce bias.

Unclear

Either of the following. • •

Insufficient information provided to permit judgement of low or high risk of bias. The study did not address this outcome.

4. Were incomplete outcome data adequately addressed? Low risk of bias

Any one of the following. • • • •

No missing outcome data. Reasons for missing outcome data unlikely to be related to true outcome (for survival data, censoring unlikely to be introducing bias). Missing outcome data balanced in numbers across intervention groups, with similar reasons for missing data across groups. For dichotomous outcome data, the proportion of missing outcomes compared with observed event risk not enough to have a clinically relevant impact on the intervention effect estimate. 7

•

•

For continuous outcome data, plausible effect size (difference in means or standardised difference in means) among missing outcomes not enough to have a clinically relevant impact on observed effect size. Missing data have been imputed using appropriate methods.

High risk of bias

Any one of the following. • •

•

• •

Reason for missing outcome data likely to be related to true outcome, with either imbalance in numbers or reasons for missing data across intervention groups. For dichotomous outcome data, the proportion of missing outcomes compared with observed event risk is enough to induce clinically relevant bias in intervention effect estimate. For continuous outcome data, plausible effect size (difference in means or standardised difference in means) among missing outcomes is enough to induce clinically relevant bias in observed effect size. ‘As-treated’ analysis done with substantial departure in the intervention received from that assigned at randomisation. Potentially inappropriate application of simple imputation.

Unclear

Either of the following. • •

Insufficient reporting of attrition or exclusions to permit judgement of low or high risk of bias (e.g. number randomised not stated, no reasons for missing data provided). The study did not address this outcome.

5. Are reports of the study free of suggestion of selective outcome reporting? Low risk of bias

Either of the following •

•

The study protocol is available and all of the study’s pre-specified (primary and secondary) outcomes that are of interest in the review have been reported in the prespecified way. The study protocol is not available but it is clear that the published reports include all expected outcomes, including those that were pre-specified (convincing text of this nature may be uncommon).

High risk of bias

Any one of the following. •

Not all of the study’s pre-specified primary outcomes have been reported. 8

• • • •

One or more primary outcomes are reported using measurements, analysis methods or subsets of the data (e.g. subscales) that were not pre-specified. One or more reported primary outcomes were not pre-specified (unless clear justification for their reporting is provided, such as an unexpected adverse effect). One or more outcomes of interest in the review are reported incompletely so that they cannot be entered in a meta-analysis. The study report fails to include results for a key outcome that would be expected to have been reported for such a study.

Unclear

Insufficient information to permit judgement of low or high risk of bias. It is likely that the majority of studies will fall into this category. 6. Other sources of potential bias Low risk of bias

The study appears to be free of other sources of bias. High risk of bias

There is at least one important risk of bias. For example, the study: • • • •

had a potential source of bias related to the specific study design used; or had extreme baseline imbalance; or has been claimed to have been fraudulent; or had some other problem.

Unclear

There may be a risk of bias, but there is either: • •

insufficient information to assess whether an important risk of bias exists; or insufficient rationale or evidence that an identified problem will introduce bias.

9

Appendix 2: Characteristics and risk of bias in included studies. Characteristics of excluded studies:

Characteristics of included studies:

Chareancholvanich 2012 3 Methods Participants

RCT single center. The dates for the study were not reported. 112 participants undergoing primary TKA. Exclusion criteria: immune compromised patients, including diabetes mellitus, gouty arthritis and autoimmune diseases, patient who might have any technical error during the operation, patients who could not be followed up on per the schedule, and patients who had a history of allergy to cephalosporins or fosfomycin. Average age group I (fosfomycin): 68.02±8.54 years; male;female =8:48 Average age group II (cefuroxime): 70.7±8.51; male;female = 6:50

Interventions

Group I (n=56) (fosfomycin): 2 g IV one hour before skin incision and 2 g IV 12 hours after first dose Group II (n=56) (cefuroxime): 1.5 g IV one hour before skin incision, 750 mg IV 8 hours after first dose, and 750 mg IV 16 hours after first dose.

Outcomes

Infection at 6 months defined as: • •

• • •

Pain and tenderness at the surgical site which patient did not experience before; Skin temperature of the operated knee that was 5 degrees Celsius higher than the skin temperature of the nonoperated knee; Significantly high C-reactive protein at the 3rd and 6th month follow-up; Significantly high or sustained high ESR at the 3rd and 6th month follow-up; Osteolytic area around the prosthesis wider than 2mm

10

Notes

Study was performed in Thailand. Study was totally supported by Thai Meiji Company.

Risk of bias table

Bias

Authors' judgement

Support for judgement

Random sequence generation low (selection bias)

Quote: "Enrolled patients were randomized by using a computer program with blocks of tow so that the numbers of the patients in each group were similar."

Allocation concealment (selection bias)

Comment: Allocation concealment was not reported on in the study

unclear

Blinding of participants and unclear personnel (performance bias)

Comment: Blinding of participants and personnel performing the procedure was not reported on.

Blinding of outcome assessment (detection bias)

unclear

Comment: Blinding of clinicians assessing for the outcome was not reported on.

Incomplete outcome data (attrition bias)

low

Comment: All 112 patients were reported on for the infection outcome at 6 month follow-up.

Selective reporting (reporting low bias)

Comment: Infection as defined in the methods section was reported on in the results section. However, the protocol for this trial was not obtained.

Other bias

Quote: "Potential conflicts of interest: The study was supported in its entirety by Thai Meiji Company."

high

Davis 1987 4 Methods Participants

Double-blind, multicenter (4 centers) RCT. The dates for the study were not reported 70 participants undergoing primary THA and 39 undergoing primary TKA Exclusion criteria: < 18 years of age, significant underlying disease, known allergy to cephalosporin antibiotics, signs/symptoms suggesting preoperative infection, or known renal impairment (CCr < 50 ml/min, serum CR > 2.0 mg %, or 11

BUN > upper limit of normal range); pregnancy, patients requiring concurrent use of a corticosteroid or second antimicrobial agent, and patients having undergone successful or suppressive antimicrobial therapy (i.e. a prior treatment course of antibiotic therapy) within 3 days prior to entry into study Group 1 (n = 55): mean age = 66.3 years; male:female = 11:47. Group 2 (n = 54): mean age of = 64.5 years; male:female = 17:42 No clear breakdown of sex and age according to type of implant received Interventions

Group 1 (n = 55 ): cefonicid (2nd generation cephalosporin, broad spectrum antibiotic resembling penicillin) 1 g preoperatively IV followed by 1 g once daily for 3 days Group 2 (n = 54): cefazolin (1st generation cephalosporin, broad spectrum antibiotic resembling penicillin) 1 g preoperatively followed by 1 g every 8 h for up to 72 h

Outcomes

Infection for up to 30 days postoperatively, assessed by clinical observation 'Adverse experiences' evaluated throughout the study period (no definition of 'adverse experiences' provided)

Notes

Informed consent obtained Trial performed in the USA Sample sizes of approximately 50 participants in each prophylactic group would find a difference of 5 or 6 infections to be statistically significant (P < 0.05) While 117 participants were entered into the trial, 8 were excluded as they did not receive primary total knee or hip implants, making the total number of participants = 109

Risk of bias table Bias

Authors' judgement

Random sequence generation unclear (selection bias)

Support for judgement Quote: "patients were randomly assigned" Comment: the randomization method was not stated 12

in study Allocation concealment (selection bias)

unclear

Comment: allocation concealment not reported

Blinding of participants and high personnel (performance bias)

Quote: "Study drugs were reconstituted by a pharmacist or study assistant who did not participate in the patient evaluations and were provided to the investigator as blinded medication. In order to maintain a blind study throughout the drug administration period, additional placebo injections were provided for those patients assigned to once-daily treatment with cefonicid." Comment: assumed that participants were also blinded to the treatment arms


unclear

Comment: not stated whether clinicians evaluating infection were aware of treatment allocation


low

Comment: Out of 117 participants: 4 participants in cefonicid group and 7 in cefazolin group were considered 'unevaluable' due to protocol violations. 2 of the 4 in the cefonicid group received concomitant postoperative antibiotics and 2 died of non-drug related causes; 2 of the 7 in the cefazolin group had dosage violations; 2 received concomitant postoperative antibiotics; 2 had UTIs; and 1 had a UTI and refused medication


Comment: infection was defined in methods section (clinical observation) and reported on in results section. However, the protocol for this trial was not obtained

Other bias

Comment: we identified no other biases

low

De Lalla 1993 5 Methods Participants

Single-center RCT conducted from January 1991-February 1992 People receiving a primary TKA Group 1 (n = 11): average age = 70 ± 6 years; average weight = 71 ± 9.5 kg; mean height = 165 ± 5 cm; male:female = 3:8 Group 2 (n = 13): average age = 68.5 ± 5.6 years; average weight = 72.6 ± 7 kg; mean height = 164.5 ± 4.5 cm; male:female = 13

1:12. No exclusion criteria reported Interventions

Group 1: systemic teicoplanin (a semisynthetic glycopeptide antibiotic used in the prophylaxis and treatment of serious infections caused by Gram-positive bacteria, including MRSA), 800 mg in 100 ml saline, administered into the forearm, infused over 5 min, 2.5 h prior to surgery, in a single dose Group 2: regional teicoplanin (administered into the foot vein), 400 mg in 100 ml saline infused over 5 min in a single dose immediately after the tourniquet was inflated

Outcomes

Evaluation of teicoplanin levels (mg/L) in serum, bone, skin, synovia and subcutaneous tissues at 20, 40, 60, 80 and 100 min after tourniquet initiation for TKA surgery Adverse events (not defined) Prosthetic or wound infections evaluated for up to 1 year follow up

Notes

Trial conducted in Italy. The teicoplanin concentrations obtained with regional prophylaxis were to found to be 2-10 times higher than those achieved with systemic prophylaxis. Peak levels in different tissues were also significantly higher (P value < 0.05) with regional prophylaxis. Quote: "The intraoperative levels of teicoplanin in subcutaneous, skin, and bone tissues (1 to 3 mg/kg) obtained by systemic administration might not have been adequate to inhibit some strains of coagulase-negative staphylococci, whose minimum inhibitory concentrations (MICs) ranged from 2 to 4 mg/Liter." (MIC defined as the minimum dose of antibiotics required to inhibit the growth of bacteria)


Authors' judgement



Comment: method of randomization not reported


unclear


Blinding of participants and

high

Comment: not stated whether participants were 14

personnel (performance bias)

blinded. Assumed that clinicians performing the procedure with systemic vs regional antibiotic prophylaxis were not blinded


unclear

Comment: not stated whether clinicians assessing the outcomes of acute adverse events, serum concentrations of teicoplanin, or infections were blinded for these outcomes


low

Comment: all participants were reported on for each outcome


Comment: all outcomes identified in the methods section were reported on in the results section. However, the protocol for this trial was not obtained

Other bias

Comment: we identified no other biases in the study

low

DeBenedictis 1984 6 Methods Participants

Single-center RCT conducted from June 1981-January 1982 Patients receiving either a TKA or THA Group 1 (n = 37): mean age = 66 years; male:female = 15:22 Group 2 (n = 39): mean age = 69 years; male:female =16:23 . Exclusion criteria: evidence of joint infection, a history of sensitivity to penicillin or cephalosporins, significant renal impairment, or interfering antimicrobial therapy

Interventions

Group 1: 1 g cefonicid administered IM or IV 30 min prior to incision and once daily for 3 days. Placebo infusions of physiologic saline were used to bring the total daily administration up to 3 doses (n: THA = 12; TKA = 21)) Group 2: 1 g cefazolin 30 min prior to incision and every 8 h for 72 h post surgery (n: THA = 14; TKA = 21)

Outcomes

Signs of adverse drug reactions and infection (defined as sepsis) up to 12 months post surgery

Notes

Trial obtained Informed consent and followed guidelines for human experimentation of the Reading Hospital (Reading, PA, USA) The study received financial assistance from Smith Kline & French Laboratories 15


Authors' judgement






unclear

Blinding of participants and low personnel (performance bias)

Quote: "We conducted a double blind study that compared cefonicid to cefazolin in patients undergoing total hip or knee replacement." Comment: participants and clinicians were blinded to the treatment arms (double-blind study)


unclear

Comment: not clear from the trial report whether the outcome assessor was blinded to treatment allocation


high

Comment: 3/76 participants that entered the study were not reported on in the results section (1 participant in the cefazolin group and 2 in the cefonicid group), but the reason for this was not clear

Selective reporting (reporting high bias)

Comment: adverse drug reactions mentioned in methods section but not reported on in the results section. However, infection was reported on although it was not mentioned in methods section

Other bias

Comment: the study received financial assistance from Smith Kline & French Laboratories, the manufacturer of the cefonicid

high

Ericson 1973 7 Methods Participants

Multicenter (2 centers) double-blind prospective RCT conducted from 1 November 1970-30 May 1972 Patients receiving either a primary THA (Charnley or Moore endoprosthesis) or a Thornton nail for femoral fracture 171 participants evaluated. 118 s had THA and 53 had surgery for pertrochanteric fracture or Moore endoprosthesis. Mean age = 62.4-62.7 years; health status and sex unclear 16

Interventions

Group 1: cloxacillin (a type of penicillin) 1 g IM 1 h prior to operation and thereafter 3 times at 6 h intervals followed by oral administration of 2 x 0.5 g cloxacillin tablets every 6 h until day 14 plus 2 x 0.5 g probenecid tablets (which make antibiotics more effective by preventing body from passing them in urine) orally twice a day for 14 days (n = 60) Group 2: placebo administered in the same manner along with probenecid (n = 58)

Outcomes

Notes

Considered only deep infections. Superficial infections that healed without developing into deep infections were not followed. Follow-up occurred at 1-2.5 years and up to 5-6.5 years The study was conducted in Sweden 59 participants were eliminated/excluded from trial: 31 (19 from cloxacillin;12 from placebo) because of side effects Only participants receiving the hip implants were evaluated. The Thornton nail for femoral fracture was not evaluated. It could not be determined from the Carlsson 1977 8 data whether any participants reported on in the RCT trial were revised/reoperated on due to infection


Authors' judgement



unclear


Support for judgement Quote: "The double blind technique was used and the patients were consecutively and randomly assigned to antibiotic or placebo group." Comment: method of randomization not reported Comment: allocation concealment not reported Quote: "The double blind technique was used . . . " Comment: given the treatment regimens described, it is likely that participants and clinicians were blinded to the treatment groups


unclear

Comment: not clear whether clinical outcome assessors were blinded to treatment groups 17


unclear

Comment: 59/230 participants excluded - unclear how many of these had a THA


Comment: infection was listed as an outcome in the methods section and reported on in the results section

Other bias


low

Evard 1988 9 Methods Participants

Multicenter (11 sites) RCT, conducted from March 1982-July 1983 955 people receiving THA Average age = 65.8 years (SD 11); male:female = 45:55 Exclusion criteria: malignant hip tumors, antibiotic treatment indicated for other reasons, an allergy to cephalosporins

Interventions

Group 1 (n = 488): cefamandole 1.5 g IV given in 1st injection, and 0.75 g in the following 8 injections over a 2-day period.(2nd generation cephalosporin with broad spectrum resembling penicillin) for 2 days. Group 2 (n = 477): cephazolin (1st generation cephalosporin with broad spectrum resembling penicillin) 1 g given every 6 h for 5 days

Outcomes

Deep infections followed up for 1 year. Deep infections defined as: • • •

Notes

subacute sepsis with isolation of pathogenic organisms obtained by direct aspirate; acute infection with clinical signs of sepsis (pain, fever, redness of the wound); and discharging sinus

The study was conducted in France A placebo was given to the cefamandole group for an additional 3 days to ensure that the dosing regimens were over the same 5 day period.

Risk of bias table 18

Bias

Authors' judgement



Quote: "Each patient was numbered in sequence just prior to the operation. Numbered boxes contained either cephazolin and cefamandole placebo, or cefamandole and cephazolin placebo, according to a random list which was balanced every 10 numbers. Randomization was stratified in each center. The code was available to the attending physician in a sealed envelope in each box. The reason for opening this envelope had to be noted on the envelope before sending it to the trial secretary."


Comment: allocation occurred just before the operation, however, it was not clear whether allocation was concealed

unclear


Quote: "A double placebo was used to ensure that the treatment was administered double blind." Comment: we assumed this meant that both clinician and participant were blinded to the intervention


low

Comment: cultures for deep infection were evaluated by an outside laboratory, so it is likely that laboratory personnel were not aware of the intervention to which participants had been allocated


low

Comment: intention-to-treat analysis was used; 18 participants in each treatment arm had prophylactic treatment discontinued or altered; reasons for discontinuation or alteration were reported and the reasons were balanced between groups


Comment: deep infection was identified as an outcome in the methods section and reported on in the results section. However, the protocol for this trial was not obtained

Other bias


low

Friedman 1990 Methods Participants

Single-center RCT. The dates for the study were not reported, however, investigation occurred over a 1 year period 24 people receiving a primary TKA. Participants were 18-75 years old. 18 participants had osteoarthritis and 6 had rheumatoid 19

arthritis Exclusion criteria: an allergy to penicillins or cephalosporins, requiring any type of dialysis, a history of alcoholic cirrhosis of the liver, or having revision TKA Interventions

Participants were randomly assigned to 1 of 3 groups with all receiving 1 g cefazolin in 10 ml of 5% dextrose systemic infusion over a 2 minute period. Group 1 (n = 8): tourniquet inflation at 1 min after cefazolin infusion Group 2 (n = 8): tourniquet inflation at 2 min after cefazolin infusion Group 3 (n = 8): tourniquet inflation at 5 min after cefazolin infusion

Outcomes

Notes

Cefazolin penetration (concentration of cefazolin) into serum, soft tissue, and bone as measured in blood samples obtained before cefazolin administration (baseline), at 10, 30, and 60 min after tourniquet inflation, just prior to tourniquet release and 5 min post tourniquet release. Adequate cefazolin concentrations for soft tissue and bone were defined as ≥ to 4 x minimum inhibitory concentration 90 (MIC 90 = 1 μg/ml) of cefazolin to Staphylococcus aureus and coagulase-negative staphylococci. Study conducted in the USA


Authors' judgement



Comment: randomization method not reported



unclear


Comment: anesthesiologists who administered the antibiotic knew which type of antibiotic was given. Whether participants were blinded to their intervention was not reported


Comment: not stated whether clinicians assessing cefazolin concentration were blinded to the

unclear

20

treatment arms Incomplete outcome data (attrition bias)

low

Comment: reported on all 24 participants that entered the trial


Comment: cefazolin concentrations defined in the methods section were reported on in the results section. However, the protocol for the trial was not obtained

Other bias


low

Gunst 1984 10 Methods Participants

RCT, the dates for the study were not reported People receiving a total hip implant - 93 hip implants in 84 participants (9 bilateral replacements). Group 1 (n = 46): average age = 63.9 years (SD 13.8); male:female = 21:25 Group 2(n = 47): average age = 65.11 years (SD 11.3): male:female = 21:26

Interventions

Group 1: IV cefamandole 1.5 g before incision followed by 1.5 g every 4 h up to 24 h Group 2: no antibiotic

Outcomes

Minor complications: defined as superficial and infections occurring outside operative site (e.g. urinary tract infection) Deep infection- according to culture result, evidence of radiologic resorption of the implant, and local inflammation evaluated at 6 days, 3 months and 1 year

Notes

Study conducted in France and published in French. A clean air system was used in the operating theatre. Superficial infections at the operative site were not evaluated.

Risk of bias table

21

Bias

Authors' judgement



Comment: randomization scheme not reported



unclear

Blinding of participants and unclear personnel (performance bias)

Comment: not reported whether participants or clinicians performing procedure were blinded to intervention


unclear

Comment: not reported whether clinicians assessing infections were blinded to intervention


low

Comment: all participants were reported on at 1 year


Comment: infections were defined in methods section and reported on in results section. However, a trial protocol was not obtained.

Other bias

Comment: we identified no other bias were identified

low

Hill 1981 11 Methods

Multicenter (9 sites) RCT conducted from October 1975-June 1978

Participants

2137 participants included: peoples undergoing primary THA; average age = 64.5 years (SD = 10): male:female = 42%:58% Exclusion criteria: malignant hip tumors, indications for antibiotic treatment (e.g. multiple infections, replacement hip, on immunosuppressive therapy).

Interventions

Group 1 (n = 1070): cefazolin at induction of anesthesia, and every 6 h post surgery for 5 days Group 2: (n = 1067): placebo given at induction of anesthesia and every 6 h post surgery for 5 days

Outcomes

Rate of infectious complications in the hip within 2 years of THA. Infection defined as abscess, septicemia, lethal infection, and assumed to be deep in nature. The Doyon 1987 follow-up study evaluated this outcome at 3-5 years Adverse events were not noted as an outcome in the methods 22

section and were not evaluated at either follow-up Notes

The trial was performed in 9 centers in France Informed consent was not reported A power calculation was performed - based on retrospective analysis: 2500 participants needed in trial to show a difference in rates of infection of 1%-3% with a type 1 error of 0.05 and a type 2 error of 0.10, 2-sided test Appears no intention-to-treat analysis was performed Infection was not defined


Authors' judgement



Quote: "Each eligible patient was numbered in sequence just before surgery. Numbered boxes contained either 20 vials of placebo or 20 vials of 1 g of cefazolin according to a random list balanced every 10 numbers."


Quote: "Code available to attending physician in a sealed envelope in each box" - just prior to surgery. However, it was unclear whether the allocation to each treatment group was concealed

unclear


Comment: for 169 participants (99 in the placebo group, 70 in the cefazolin group), concealment of allocation was broken by the clinician after 5 days of treatment due to signs of infection. Thereafter clinicians who had previously been blinded to treatment arm partially lost their blinding


unclear

Comment: participants were recalled 6 months, 1 year, and 2 years after surgery and assessed clinically, radiologically, and biologically (erythrocyte sedimentation rate). However, it was not reported whether clinicians assessing participants were aware of treatment allocation


unclear

Comment: study conducted at 10 sites, but 1 did not send follow-up forms and was excluded from the 23

analysis. Consequently the data for evaluation came from 9 study sites. It was not clear how many participants were excluded as a result of this Selective reporting (reporting low bias)

Comment: infection listed as an outcome in introduction section and evaluated in results section. However, a trial protocol was not obtained

Other bias


low

Hinarejos 2013 12 Methods

Single-center RCT, conducted from September 2005-April 2010

Participants

People undergoing a primary TKA. Trial analyzed a total of 2948 knees with a minimum of 12 month follow-up (52 lost to followup of 3000 that were entered) Group 1 (n = 1483): average age = 75.84 years (SD 7.44); male:female = 23.3%:76.7% Group 2 (n = 1465): average age = 76.06 years (SD 7.22); male:female = 24.1%:75.9%

Interventions

Group 1: 2 g cefazolin in a 10-15 min infusion 30-60 min before incision, or, if participant had a beta-lactam allergy, 1 g vancomycin (a glycopeptide antibiotic used for Gram-positive bacteria) in a 1-h infusion 60-90 min before incision. Implant was cemented with Simplex P cement loaded with 0.5 g erythromycin (antimicrobial spectrum similar to penicillin; is often prescribed for people with allergies to penicillins) and 3 million units of colistin in 40 g of cement (Stryker). This was followed by 1 g cefazolin every 8 h or 1 g vancomycin every 12 h for the first 24 h after surgery. Group 2: 2 g cefazolin in a 10-15 min infusion 30-60 min before incision, or, if participant had a beta-lactam allergy, 1 g vancomycin in a 1-h infusion 60-90 min before incision. Implant was cemented with Simplex cement without antibiotic (Stryker, Mahwah, New Jersey, US). This was followed by 1 g cefazolin every 8 h or 1 g vancomycin every 12 h for the first 24 h after surgery

Outcomes

Incidence of infection (superficial or deep) after primary TKA according to CDC criteria. Average duration of follow-up was 38 24

months Notes

Study conducted at the Hospital de la Esperanza, Barcelona, Spain Participants signed informed consent forms A power study was performed to evaluate the size of the sample. Accepting an alpha risk of 0.05 and a beta risk in a 1-sided test for 2 independent proportions, 1370 participants were needed in each group in order to recognize a decrease in the infection from 2.2% in the control group to 1% in the study group as statistically significant Adverse events were not evaluated


Authors' judgement



Quote: "Randomization was done intraoperatively according to a list of computer generated random numbers."


Quote: "Randomization was done intraoperatively according to a list of computer generated random numbers." However, it was not clear whether allocation to each treatment group was concealed

unclear


Comment: physicians involved in the trial knew whether or not antibiotic was being loaded into the cement for each participant


low

Comment: email reply from Dr Hinarejos, 4 July 2013, the "clinical assessors for the endpoint of infection were blinded to the treatment arms"


low

Comment: Of the 3000 knees entered into the trial, 52 were lost to follow-up over a minimum 12 months of follow-up. It was not clear why this happened. Therefore a total of 2948 knees were analyzed


Comment: infections defined in the methods section were evaluated in the results section. However, a trial protocol was not obtained

Other bias

Quote: "None of the authors received payments or

low

25

services either directly or indirectly from a third party in support of any aspect of their work. No author has a relationship that could be perceived to influence what is written in this work."

Jacobson 2005 13 Methods Participants

Single-center RCT, conducted from July 2000-July 2002 Participants enrolled in the trial received a primary or revision THA or TKA Group 1 (THA; n = 36): average age = 66.3 years (SD 10.7) Group 2 (THA; n = 43): average age = 64.3 years (SD 13.7) Groups 1 and 2: male:female = 58%/42%; 4 had diabetes mellitus; 4 were on steroids or immunosuppressants; and 4 smoked Group 3 (TKA; n = 39): average age = 68.2 years (SD 10.6) Group 4 (TKA; n = 41): average age = 70.0 years (SD11.2) Groups 3 and 4: male:female = 40%/60%; 8 participants had diabetes mellitus; 4 were on steroids or immunosuppressants; and 2 smoked Exclusion criteria: known allergy or hypersensitivity to iodine, acrylate adhesives or adhesive tape; need for bilateral TKA procedures; any infection at time of surgery; previous enrolment in the study, or participation in another study involving an investigational medication within 30 days of enrolment in this study

Interventions

Group 1 (THA; n = 36): preoperative preparation with DuraPrepTM solution plus LobanTM antimicrobial incise drapes along with systemic IV antimicrobial prophylaxis with 34 receiving cefazolin, 1 vancomycin and 1 receiving no antibiotic prophylaxis (amounts of antibiotics administered not clear, but they were administered pre-operatively and for 24 h postoperatively in Groups 1-4) Group 2 (THA; n = 43): preoperative preparation with povidone 26

iodine plus Loban antimicrobial incise drapes along with systemic IV antimicrobial prophylaxis with 42 receiving cefazolin, and 1 receiving vancomycin Group 3 (TKA; n = 39): preoperative preparation with DuraPrep solution plus Loban antimicrobial incise drapes along with systemic IV antimicrobial prophylaxis with 38 receiving cefazolin and 1 receiving vancomycin Group 4 (TKA; n = 41): preoperative preparation with povidone iodine plus Loban antimicrobial incise drapes along with systemic IV antimicrobial prophylaxis with 41 receiving cefazolin Outcomes

Primary: wound contamination identified by wound culture obtained at the wound edges just prior to wound closure Secondary: drape lift during the procedure (an indicator that drape adhesion was maintained and thus a sterile surface was maintained throughout the procedure); costs of skin preparation, and 30 day postoperative wound complications (including SSI)

Notes

Study conducted at the Mayo Clinic, Rochester, MN, USA 163 of the 176 evaluable participant received a primary THA or TKA. Only these 163 were reported on in this Cochrane Review IRB approval was obtained and all participants gave written consent


Authors' judgement



unclear


Support for judgement Quote from corresponding author, Doug Osmon, MD, via email dated 5 June 2104, "The contract statistician developed the randomization scheme using SAS. The randomization scheme was administered using sealed envelopes." Comment: allocation concealment not reported Quote from corresponding author, Doug Osmon, MD, "The patients were blinded to the treatment arms. The physician's assistant (PA), who conducted 27

the randomization in the operating room, prepped and draped the patient prior to the procedure starting." Blinding of outcome assessment (detection bias)

unclear

Comment: not clear whether those assessing for cost and wound healing complications were aware of the treatment arm to which participants were allocated. For drape lift, the assessors (PA) of this outcome were aware of the treatment arm to which participants had been allocated. For wound contamination, quote: "Members of the microbiology laboratory staff were blinded to treatment."


high

Comment: according to data obtained from the corresponding author, Doug Osmon, MD, all participants entered into the trial were reported on for the outcomes of wound contamination, adverse events (including infection) and drape lift. However only 42 participants in total were evaluated for cost of skin preparation


Comment: outcomes mentioned in the methods section were reported on in the results section. However, a trial protocol was not obtained

Other bias

Quote from the corresponding author, Doug Osmon, MD, "3M (the manufacturer of DuraPrepTM and LobanTM) provided all the funding for the study as well was participated in the study design process, study monitoring, and manuscript writing." Furthermore, trial report stated: "One of the authors (Doug Osmon) received funding from 3M company through the Mayo/3M Infection Control Alliance."

high

Johnson 1987 14 Methods

Single-center RCT. Unclear when study was undertaken

Participants

22 people receiving TKA. Sex, age, weight, and diagnosis of participants unclear

Interventions

Group 1 (n = 7): 1.5 g cefuroxime by intravenous injection at 5 min before inflation of the tourniquet Group 2 (n = 6): 1.5 g cefuroxime by intravenous injection at 10 min before inflation of the tourniquet Group 3 (n = 5): 1.5 g cefuroxime by intravenous injection at 15 28

min before inflation of the tourniquet Group 4 (n = 4): 1.5 g cefuroxime by intravenous injection at 20 min before inflation of the tourniquet No exclusion criteria reported Outcomes

Penetration of cefuroxime (2nd generation cephalosporin) into the bone and subcutaneous fat surrounding the knee. Samples of bone and subcutaneous fat were extracted at regular intervals for evaluation during the procedure Quote: "For the purpose of this study an average tissue concentration of 4.0 μg/g was considered to constitute adequate prophylaxis."

Notes

Study conducted in the United Kingdom


Authors' judgement






unclear

Blinding of participants and high personnel (performance bias) Blinding of outcome assessment (detection bias)

high


low

Comment: allocation concealment not reported Comment: since antibiotic concentrations were evaluated by pathology laboratories and were reported back to the clinicians, it is likely that the assessors were aware of the treatment arms Comment: all participants were reported on


Comment: antibiotic concentration mentioned in the methods section and reported in the results section. However, a trial protocol was not obtained

Other bias


low

Josefsson 1981 15 29

Methods

Multicenter RCT (9 sites), study conducted from March 1976May 1978

Participants

1685 total hip arthroplasties performed on 1596 participants. Of these participants, 814 were women with an average age of 70 years and 782 were men with an average age of 68. Osteoarthritis was the primary diagnosis (1438 hips). The vast majority received a primary THA (97.7% with 2.3% for failed hemiarthroplasty or osteotomy). No exclusion criteria reported

Interventions

Group 1 (n = 821): 0.5 g gentamicin (a broad-spectrum antibiotic used chiefly for severe systemic infections) added to 40 g cement package of Palacos cement Group 2 (n = 812): 1 IV or IM dose 24 h prior to surgery along with a course of antibiotic treatment of one of the following: • • • • •

• • • •

Outcomes

cloxacillin (n = 64): 1 g x 4; probenecid 0.5 g x 2 for 9 days cloxacillin (n = 261): 1 g x 4 for 7 days; cephalexine (1st generation cephalosporin antibiotic; n = 132): 1 g x 4 for 11 days; given IV for first 24 h; cloxacillin (n = 116); 1 g X 4 for 7 days dicloxacillin (a penicillin-like antibiotic used to treat certain infections caused by bacteria; n = 35) 0.5 g x 3 for 8-10 days; dicloxacillin (n = 142): 1 g x 4 for 14 days; cloxacillin (n = 28): 1 g x 4 for 11 days; phenoxymethylpenicillin (commonly known as penicillin V; n = 73) 0.65 g x 4 for 10 days; cephalexine (n = 77): 1 g x 4 (9 days); IV first 24 h (

Superficial and deep infections were evaluated at 3, 6, 12, 24 months, 5 years 16 and 10 17 years postoperatively Superficial infection defined as: abnormal redness of wound; presence of secretion and firm diagnosis that precluded antibiotic treatment Deep infection defined as: pain, elevated erythrocyte sedimentation rate (more than 35 mm/h) and progressive radiographic resorption of bone stock

Notes

Study conducted at 9 hospitals in Sweden Infection data were not broken down according to sex or treatment group 30


Authors' judgement






unclear


Comment: surgeons probably knew to which intervention participants were allocated as needed to mix gentamicin into the cement when appropriate. Unclear whether participants knew to which group they had been allocated


unclear

Comment: not reported whether outcome assessors were blinded


high

Comment: At 2 year follow-up 451/812 arthroplasties in Group 1 were evaluated, and 488/821 in Group 2 were evaluated


Comment: superficial and deep infections listed in the methods section and reported on in results section. However, a trial protocol was not obtained

Other bias


low

Kalmeijer 2002 18 Methods Participants

Single-center double-blind placebo-controlled RCT, conducted from January 1997-July 1999 People receiving either a primary THA (n = 251) or primary TKA (n = 178) Exclusion criteria: an active infection at time of inclusion, or antibiotic treatment in the previous 24 h

Interventions

Group 1: nasal swabs of both nares taken for culture on day before surgery. Mupirosin ointment (Glaxo-SmithKline) applied to both nares twice/day. At least 2 doses of mupirocin administered before surgery. 2 g cefamandole administered 30-60 min before surgery, followed by 2 more doses of 1 g each at 8 h and 16 h after surgery. If participants were allergic to cephalosporins, then given 600 mg clindamycin (another antibiotic) 30-60 min before surgery, followed by 2 more doses 31

of 600 mg each at 8 h and 16 h after surgery. Group 2: nasal swabs of both nares taken for culture on day before surgery. Placebo ointment (appearance identical to mupirocin) applied to both nares twice per day. At least 2 doses of placebo administered before surgery. The same cefamandole antibiotic treatment regimen as described in group 1 was also provided in group 2. Outcomes

Primary outcome: SSI (superficial or deep) rate up to 1 month after surgery. After discharge from the hospital participants were followed up by phone using a standardized questionnaire based on the CDC criteria for infection 19. If one of the criteria was met, the participant was evaluated at an outpatient department by an orthopedic surgeon Secondary outcome: rate of SSI due to S aureus, the rate of endogenous (originating from the person) S aureus SSIs, and the duration of hospital stay. An endogenous SSI was defined as an SSI with a S aureus isolate that was identical on pulsed-field gel electrophoresis analysis to the nasal isolate of that participant

Notes

Intention-to-treat analysis performed Informed consent received from all participants Trial conducted at the Ignatius Hospital, Breda, The Netherlands Above analysis is a subset of all participants entered into the trial which included: primary THA and TKA, revision (second or later) THA (n = 47) or TKA (n = 12), back procedures (n = 100) or other orthopedic procedure (n = 16). In total, 315 participants were entered into the mupirocin group and 299 were entered into the placebo group The reviewers attempted to contact lead author MD Kalmeijer and co-authors via email on 26 March 2014 and 31 March 2014. Co-author (JAJW Kluytmans) replied on 2 April 2014 that data were not available for evaluation


Authors' judgement

Random sequence generation low

Support for judgement Quote: "Patients were randomized according to an 32

(selection bias) Allocation concealment (selection bias)

individual randomization list by an independent pharmacist." low

Quote: "Patients were administered a nasal ointment (either mupirocin or placebo both identical to each other), according to the next number on the randomization list."


Comment: trial was double-blind in nature, so we assumed that clinicians applying the ointment and participants receiving the ointment would be unaware of treatment group


unclear

Comment: not clear whether outcome assessors were aware of treatment group to which each participant had been allocated


unclear

Quote: "Of the 627 patients included [all procedures and not just primary THA and TKA], 13 could not be evaluated." Comment: not clear whether any of the THA and TKA participants could not be evaluated


Comment: all outcomes identified in the methods section of the study were reported on in the results section. However, a study protocol was not obtained

Other bias


low

Mauerhan 1994 20 Methods Participants

Multicenter (15 centers) double-blind RCT, conducted from November 1989-April 1991 Patients receiving either primary THA or TKA or a revision (second) hip or knee implant (revision implants were excluded from this review and made up 131 hip and 62 knee implants) 546 THA participants enrolled, and 615 TKA participants: 285 in Group 1 and 261 in Group 2 Average age for both groups combined = 65; male:female = 533:821 The primary diagnosis was osteoarthritis (1006 participants; 75% of all participants enrolled) Exclusion criteria: allergy to cephalosporins, renal impairment (a serum creatinine level of more than 229 micromoles/l), and 33

neutropenia (< 1000 granulocytes/mm3) Interventions

Group 1: cefuroxime 1.5 g IV administered 15-60 min prior to surgery followed by 750 mg at 8 h and 16 h after surgery Group 2: cefazolin 1 g IV administered 15-60 min prior to surgery and every 8 h for 9 doses (3 days total)

Outcomes

Classified wound infections as superficial or deep, depending upon whether they had developed above or below the fascia Deep wound infection, identified from culture specimens of any purulent drainage from inflamed wound, assessed up to 1 year post procedure Drug-related adverse events

Notes

Study conducted in the USA A power calculation was performed at a 5% difference in the over-all rate of infection between the 2 groups in the evaluable population at a power of 85% An intention-to-treat analysis was performed Protocol approved by the institutional review board for each center and written informed consent was signed by each participant who entered the study


Authors' judgement



unclear


Support for judgement Quote: "Randomization schedule was generated by computer for each investigator before study and was administered on the basis of the chronological order in which their patients had been in the study." Comment: allocation concealment not reported Quote: "Eligible patients were stratified according to the operative procedure and then were randomly assigned in a double blind fashion . . . " Comment: we therefore assumed that both 34

participants and clinicians were blinded to the antibiotic treatment arm Blinding of outcome assessment (detection bias)

unclear

Comment: not clear whether outcome assessors were blinded


high

Comment: 520/1354 participants (38%; 259 in Group 1and 261 in Group 2) were excluded due to having at least 1 major protocol violation, such as administration of additional antibiotics during the perioperative period or incorrect timing of the first dose. This left 834 participants who were evaluated.


Comment: reported on infection as outlined in the methods section. However, a trial protocol was not obtained

Other bias

Comment: we identified other biases

low

McQueen 1990 21 Methods Participants

Multicenter RCT (2 study sites), probably conducted around January 1982-January 1985 387 people receiving either a THA (380 hips) or TKA (25 knees). Note: some participants received both. While not explicitly stated, it was assumed all participants received primary implants. The average age for both groups was 67 years; male:female ratio: Group 1 = 33:67; Group 2 = 42:58 No exclusion criteria reported

Interventions

Group 1: IV 1.5 g cefuroxime at induction of anesthesia followed by 2 x 750 mg doses IM at 6 h and 12 h after surgery Group 2: 1.5 g cefuroxime powder mixed by surgeon with each pack of CMW type 1 cement powder; barium sulfate added as a marker

Outcomes

Superficial or deep wound infections up to 2 years after surgery Superficial infection defined as infection superficial to the deep fascia; diagnosis made on the finding of a discharge from the wound with erythema and either positive or negative bacteriological cultures, but with no delay in wound healing Deep infection defined as infection extending beneath the deep fascia, with persistent wound discharge or joint pain and positive or negative cultures from deep tissues 35

Notes

2 centers in Scotland performed the procedures. The McQueen 1987 International Orthopedics study 22 listed was a subset of the McQueen 1990 study. All data in the McQueen 1987 study were maximally extracted.


Authors' judgement






unclear


Comment: clinicians performing procedure were aware of whether participant received IV or local antibiotic as responsible for mixing cefuroxime powder into the cement. Trial described as singleblind; we assumed from language in the paper that the participants were blinded


high

Comment: we assumed that outcome assessors were not blinded


low

Comment: 401/405 operations (201 for intravenous cefuroxime and 204 for cement impregnated cefuroxime) on 387 participants evaluated over a period of 2 years


Comment: infection outcomes listed in the methods section (deep or superficial; early or late) were reported on in the results section. However, a trial protocol was not obtained

Other bias

Comment: Glaxo Group Research supplied the antibiotic (IV or local) free of charge

high

Morrison 2014 23 Methods Participants

Single-center RCT, conducted from March 2010-November 2011 Patients receiving a primary THA and TKA THA participant characteristics (n = 292): average age = 58.8 years (SD 9.43); BMI = 29.6 (SD 6); male:female = 132/160; 36

diabetes = 10.6% TKA participant characteristics (n = 285): average age = 63.5 ± 9.1 years; BMI = 32 (SD 6.2); male:female = 129:156 ; diabetes = 20.7% Interventions

Group 1 (THA = 149; TKA = 140): standard skin preparation with chlorhexidine (preoperative shower), and alcohol and betadine (intra-operative skin preparation). After application of the incise drape, additional skin preparation with application of iodine povacrylex (iodophor)/alcohol combination to the surgical site If participant was ≤ 70 kg: 1 g preoperative IV cefazolin If participant was > 70 kg: 2 g IV cefazolin Participants allergic to penicillin or to cephalosporins were given 1 g vancomycin Cefazolin was administered within 1 h prior to surgery and vancomycin within 2 h prior to surgery Group 2 (THA = 143; TKA = 145): standard skin preparation with chlorhexidine (preoperative shower), and alcohol and betadine (intra-operative skin preparation) followed by surgical incise draping once the skin was dry Same antibiotic prophylaxis as administered in group 1 based on weight and allergic reactions.

Outcomes

Primary outcome: persistence of wound drainage (> 48 h) following THA and TKA until patient discharge, used as a surrogate for wound infection 24 Secondary outcomes: SSI as defined by CDC at 6 weeks; length of hospital stay and the need for readmission or reoperation within 6 weeks following operation

Notes

Trial was funded by 3M Trial conducted in the USA IRB approval and informed consent obtained Persistent wound drainage beyond 48 h was studied as an endpoint due to a correlation between it and infection24,25,26

37

Clinical trial registered on ClinicalTrials.gov website: NCT01097135


Authors' judgement



Quote: "Use of a random number generator in Excel employed by the statistician."


high

According to discussion with lead author on 13 March 2014,"Patients were prepped with antisepsis in an adjoining room to the operating room. The physicians were not aware of the treatment groups, but the patients were, prior to the procedure starting."


Physicians performing the procedure were not made aware of treatment groups but participants were aware because they received an additional application of antisepsis


low

According to discussion with lead author on 13 March 2014, "Nurses assessing for deep infection were not aware of the treatment groups."


high

A total of 600 participants entered the study. 578 received Group 1 or Group 2 therapy. 22 were excluded (Group 1: 15 withdrew consent and 2 cancelled surgery; Group 2: 2 withdrew consent; 2 cancelled surgery and 1 died during postoperative period)


SSI listed as outcome in the methods section and reported on in the results section. A copy of the trial protocol was reviewed to confirm this

Other bias

3M funded the study antiseptics and paid the expenses of the study coordinator. However, the lead author said on 13 March 2014: "3M . . . they had no say in the findings, the study write-up and presentation of the findings."

high

Phillips 2014 27 38

Methods Participants

Single-center RCT, conducted from March 2011-March 2012 People undergoing either primary or revision THA (n = 591) or TKA (n = 596) or spine fusion Demographics of primary THA group: • •

Group 1 (n = 298): median age = 63 years (21-89); male:female = 123/175; Group 2 (n = 293); median age = 63 years (30-84); male:female = 120/173

Demographics of primary TKA group: • •

Group 1 (n = 299); median age = 63 years (36-93); male:female = 100/199; Group 2 (n = 297); median age = 63 years (41-92); male:female = 102/195

Exclusion criteria: pregnancy, breastfeeding, allergy to mupirocin or povidone-iodine, interval from pre-surgical assessment clinic visit to surgery of less than 7 days and an infectious indication for surgery. The need for nasal intubation was added as an exclusion criterion shortly after study initiation (typically for cervical spine surgery) Interventions

Group 1: twice daily application of nasal mupirocin ointment (Bactroban, Nasal®, mupirocin calcium ointment 2%, GlaxoSmithKline) on the intranasal mucosal surfaces of each nostril for 5 days prior to surgery (THA = 298 participants; TKA = 299 participants) Group 2: 2 applications of povidone-iodine solution (3MTM Skin and Nasal Antiseptic, povidone-iodine solution 5% w/w, 3M Corporation) on the intranasal mucosal surfaces of each nostril (using a cotton swab) within 2 h of surgical incision (THA = 293; TKA = 297) Both groups also received: 6 chlorhexidine wipes (2% Chlorhexidine Gluconate Cloth Patient Preoperative Skin Preparation, Sage Products) used on the skin the evening before and again on the morning of surgery. 1 wipe was used for each of the following 6 skin areas: neck/chest/arms, abdomen/groin, right leg/foot, left leg/foot, back and buttocks. At the time of surgery all participants received routine antimicrobial prophylaxis which consisted of cefazolin 1 g or clindamycin 600 mg (if participant reported a ß-lactum allergy or vancomycin 1 g for those 39

colonized with MRSA. Antibiotic infusion was started within 1 h of incision or 2 h for vancomycin and additional doses given during surgery according to accepted guidelines. Standard preoperative surgical site skin preparation also consisted of 2% chlorhexidine gluconate/70% isopropyl alcohol solution Outcomes

Deep surgical site infection within 3 months after surgery according to CDC definition Adverse events related to mupirocin and povidone-iodine application that included: headache, rhinorrhoea (nasal cavity filling with fluid), nasal irritation, congestion, cough, and pharyngeal irritation

Notes

Trial conducted at New York University Trial was registered on ClinicalTrials.gov website under clinical trial number:NCT01313182 3M Corporation provided a research grant for the study Preoperative S aureus antibiotic susceptibility testing and strain typing was performed including testing for MRSA.


Authors' judgement



unclear

Support for judgement Quote: "Subjects were stratified by arthroplasty or spine fusion surgery, and then randomized 50:50 to either mupirocin or povidone-iodine treatment groups in blocks of 100." Comment: allocation concealment not reported


Comment: physicians and participants knew the group to which they had been allocated


low

Quote: "Infection Prevention and Control practitioners reviewing the records were blinded to study participation and receipt of study treatment."


low

Comment: all participants were reported on via an intention-to-treat analysis

Selective reporting (reporting high bias)

Comment: protocol defined the following outcomes: deep and superficial infection at 12 months; length 40

of hospital stay due to re-admission; readmission within 12 months of the procedure; adverse events to mupirocin and povidone-iodine. The study evaluated participants at 3 months only for infection and adverse events Other bias

high

Comment: 3M Corporation provided a research grant for this study

Pollard 1979 28 Methods

Multicenter (3 centers) prospective RCT conducted from January 1976-June 1977

Participants

297 participants received 310 THAs (3 bilateral cases). 7 participants excluded due to death, leaving 303 total hip operations. Of these 303 operations; 178 performed on women and 125 on men. 53 participants aged 50-59 years; 133 aged 6069 years; 103 aged 70-79 years and 14 aged < 50 years. 279 primary THAs, 24 secondary or subsequent procedures. Exclusion criteria: those with a history of hypersensitivity to cephalosporins or penicillin or with an appreciable degree of renal impairment

Interventions

Group 1 (n = 146): cephaloridine (1st generation cephalosporin with a broad spectrum antibiotic resembling penicillin) 1 g IV at induction of anesthesia and 2 further IM doses of 1 g each 6 h and 12 h later Group 2 (n = 157): flucloxacillin (a narrow-spectrum beta-lactam antibiotic of the penicillin class used to treat infections caused by susceptible Gram-positive bacteria) 500 mg IM 1 h prior to surgery and then 500 mg 4 times daily for 14 days. For the first 24 h flucloxacillin was given IM, with subsequent doses given by mouth

Outcomes

Superficial or deep infection over a period of at least 12 months Superficial infection defined as minor (purulent discharge without fever); moderate (purulent discharge accompanied by fever); or severe (major wound dehiscence) Deep infection defined as early (< 6 months after surgery) or late (> 6 months after surgery). Early infection diagnosed by the presence of pain, fever, redness of the wound, and a discharge containing pathogenic organisms. Late infection diagnosed by the 41

presence of 2 or more of the following criteria: pain in the hip, a discharging sinus, isolation of pathogenic organisms from a sinus, isolation of material from which no organisms could be cultured and which contained many polymorphonuclear leukocytes; erythrocyte sedimentation rate raised above the preoperative level by 30 mm or more in the first hour; or radiologic evidence of infection such as periosteal reaction, bone reabsorption, or irregular reabsorption of the calcar Adverse events were not specified as an outcome in the methods section Notes

Study conducted in the United Kingdom 7 participants were excluded from trial due to 5 deaths and 2 lost to follow-up leaving 290 participants who underwent 303 total hip replacements


Authors' judgement






unclear


Comment: anesthesiologists who administered the antibiotic knew which type of antibiotic was given, but no mention of whether participants were blinded


unclear

Comment: not clear who diagnosed instances of the infection during the follow-up period


low

Comment: of 297 participants with 310 implants; 290 participants (303 implants) were followed up for at least 12 months. 5 of the 207 participants died and 2 were lost to follow-up


Comment: infection listed as an outcome in methods section and reported on in results section. However, the trial protocol was not obtained

Other bias


low

42

Richardson 1993 29 Methods

Single-center RCT, no information available regarding date(s) or time period over which the investigation occurred

Participants

People receiving a primary cemented TKA. No exclusion criteria were reported

Interventions

Group 1 (n = 8): IV systemic 1 g cephamandole 5 min before tourniquet inflation Group 2 (n = 8): IV systemic 2 g cephamandole 5 min before tourniquet inflation Group 3 (n = 16): IV systemic 1 g cephamandole 5 min before tourniquet inflation and a 'tourniquet-release dose' of 1 g cephamandole before tourniquet release All groups received a 1 g dose of cephamandole 6 h after the procedure ended

Outcomes

Concentration of cephamandole through evaluation of blood samples and knee drain fluid. Blood samples taken at time of cephamandole IV systemic administration and at tourniquet release. Samples of drain fluid aspirated from the drain line 10 min after tourniquet release and again immediately before the 2nd dose of cephamandole 6 h after the procedure ended. Concentrations measured in mg/L Wound infection (not stated in the methods section, but included in results section

Notes

Approval for the investigation from the hospital ethical committee and informed consent obtained from the participants Study conducted in the United Kingdom (Scotland)


Authors' judgement



Quote: "Patients were allocated to the three groups by stratified randomization with a block size of four."

Allocation concealment


unclear

43

(selection bias) Blinding of participants and high personnel (performance bias)

Comment: anesthesiologists who administered the antibiotic knew which type of antibiotic was given. Whether participants were blinded was not reported


unclear

Comment: blinding of outcome assessors not reported


low

Comment: all participants evaluated


Comment: serum and knee fluid concentrations were identified as outcomes in the methods section and reported on in the results section. However the trial protocol was not obtained

Other bias


low

Ritter 1989 30 Methods Participants

Single-center RCT, dates of conduct unknown People receiving either a primary THA (n = 101 total hips) or a primary TKA (n = 95 total knees) Group 1 participants (n = 98); male:female = 31:67; diagnosis of osteoarthritis = 81% Group 2 participants (n = 98); male:female = 42:56; diagnosis of osteoarthritis = 79%. Average age of study population = 66.4 years No exclusion criteria reported

Interventions

Group 1 (45 THAs and 53 TKAs): 2 intraoperative doses of cefuroxime, 1500 mg and 750 mg, and continued IV cefuroxime for 24 h postoperatively (750 mg every 8 h) Group 2: (56 THAs and 42 TKAs): 2 intraoperative doses of cefuroxime (1500 mg and 750 mg) IV

Outcomes Notes

Latent deep wound infection or other related complications over a 1 year time frame. Deep infection was not defined Conducted at a single center in the USA


Bias

Authors' judgement





Quote: "Upon entering the operating room, patients were randomly assigned to one of two groups."

unclear

Comment: allocation concealment not reported Blinding of participants and high personnel (performance bias)

Comment: physicians knew to which group participants had been allocated. It was not clear whether participants knew to which group they had been allocated


unclear

Comment: not reported whether outcome assessors were blinded


low

Comment: 191/196 participants were evaluated for infection over a 1 year time frame (5 participants died)


Comment: latent deep wound infection listed as an outcome in the methods section and evaluated in the results section. However, a trial protocol was not obtained

Other bias


unclear

Scaglione 1997 31 Methods

Single-center RCT, dates of conduct unknown

Participants

42 participants (18 women and 24 men) admitted for routine hip arthroplasty. Median age = 66 years (60-71 years)

Interventions

Group 1 (n = 22): 2 g cefodizime (3rd generation cephalosporin) IV infused over 30 min prior to surgery Group 2 (n = 20): 2 g ceftriaxone (3rd generation cephalosporin) IV infused over 30 min prior to surgery

Outcomes

Concentration of cefodizime and ceftriaxone in the serum and bones of participants During surgery, a sample of blood from the surgical field was collected at the time of removal of the femoral head, and the time after IV administration was noted. Samples of blood from the drainage tube were collected every 2 h for 24 h after surgery. The blood specimen and the femoral head bone were sent to the laboratory for evaluation of the concentration of each antibiotic. The mean concentrations of each antibiotic 45

measured in mg/L Notes

Study conducted in Italy Supported by a grant from MURST Study approved by the local ethics committee and informed consent obtained for each participant


Authors' judgement






unclear


Comment: anesthesiologists who administered the antibiotic knew which type of antibiotic was given. No report of whether participants were blinded to treatment


unclear



low

Comment: all participants were evaluated


Serum and bone antibiotic concentrations as defined in the methods section were reported on in the results section. However, a trial protocol was not obtained

Other bias


low

Schulitz 1980 32 Methods

Single-center RCT, conducted from 1 January 1972-31 46

December 1974 Participants

259 people requiring THA due to coxarthrosis Group 1 (n = 105); male:female = 35:70 Group 2 (n = 89): male:female = 29:60 Information about the ages of participants was not provided

Interventions

Group 1: 600 mg lincomycin (for participants allergic to penicillin or where bacteria have developed resistance to penicillin) IV 1 h and 6 h post surgery and 2 further 600 mg lincomycin IV injections on 2nd day post surgery. From day 3 to day 10, 1 g lincomycin given 3 times daily Group 2: no antibiotic therapy was administered at any time

Outcomes

Infection - either superficial or deep, evaluated over a 730 day period Superficial infection defined as: presence of temperature, signs of inflammation with superficial induration of the wound and purulent wound drainage Deep infection defined as: signs of inflammation with deep induration of the wound area

Notes

Study conducted in Germany 65/259 participants were excluded due to: 18 deaths; 12 from Group 2 who received antibiotics post surgery; 16 received another antibiotic during the 2 year follow-up; 7 because additional surgery was required for reasons other than infection; and 10 had a bilateral implant within < 6 months of the first surgery. In total, 40 were excluded from Group 2 and 25 from Group 1 The study did not report whether any of the infections resulted in follow on surgery or in replacement of the implant. However, did report that 1 of the infections in Group 2 received antibiotics beginning on the 3rd postoperative day and in another participant in Group 2 did not have consent granted for surgical revision

Risk of bias table

47

Bias

Authors' judgement





Comment: concealment of allocation not reported

unclear


Comment: clinicians who administered antibiotic prophylaxis (or not) were probably aware of the treatment groups as antibiotic and cement require mixing together prior to using


unclear



high

Comment: 65/259 participants had to be excluded (reasons given above)


Comment: outcomes of superficial and deep infections listed in the methods section were reported on in the results section. However, a trial protocol was not obtained

Other bias

Comment: we identified no other bias

low

Soave 1986 33 Methods

Single-center RCT conducted from May 1980-December 1980

Participants

101 participants receiving primary THA (n = 63) or TKA (n = 40) Main indication for surgery: osteoarthritis 72% Average age for trial = 63 years Exclusion criteria: those who had received antibiotics in the week prior to admission, had impaired renal or hepatic function, had known sensitivity to penicillin or cephalosporins, or were pregnant

Interventions

Group 1: ceforanide (2nd generation cephalosporin, with broad spectrum resembling penicillin) 1 g preoperatively and 1 g 12 h later Group 2: cephalothin (1st generation cephalosporin with broad spectrum resembling penicillin) 2 g preoperatively, 2 g intraoperatively after blood and bone specimens obtained and 1 g 48

every 6 h for 3 additional doses Outcomes

Fever and operative site infection: assessed daily until hospital discharge, at 6 weeks, then at 12 and 18 months Operative site was cultured to identify type of infection

Notes

Study conducted at the hospital for special surgery, New York, New York, USA Each participant gave signed informed consent


Authors' judgement


Support for judgement Quote: "Patients were randomized according to a schedule kept in the pharmacy." Comment: nature of schedule not clear


unclear



Comment: clinicians administering antibiotics probably not blinded, as doses were different for each treatment arm


unclear

Comment: not clear whether outcome assessors blinded


high

Comment: 81/101 participants followed up for 18 months. It was not clear why 20 participants were excluded at 18 month follow-up


Comment: infection was defined in methods section and reported on in results section. However, a trial protocol was not obtained

Other bias


low

Soriano 2008 34

49

Methods Participants

Single-center prospective double-blind RCT. conducted from September 2004-December 2005 Patients receiving a primary cemented TKA; no differences statistically between groups for primary characteristics (n = 908) Group 1 (n = 466): average age = 70.9 years (SD 7.2); male:female = 26.4%:73.6%; diagnosed with arthritis = 98.7% Group 2 (n = 442): average age = 71.2 years (SD 7.6); male:female = 23.5%:76.5%; diagnosed with arthritis = 98.8% Exclusion criteria: an allergy to penicillin

Interventions

Group 1: placebo administered 10-30 min before inflation of the tourniquet (plus incision/start of surgical procedure) and 1.5 g cefuroxime 10 min before release of the tourniquet Group 2: 1.5 g cefuroxime 10-30 min before inflation of the tourniquet (plus incision/start of surgical procedure) and placebo administered 10 min prior to release of the tourniquet Both arms of the trial received a postoperative dose of 1.5 g cefuroxime 6 h after the end of the surgical procedure

Outcomes

Deep tissue infection rate as defined by the CDC criteria assessed at 3 and 12 months after surgery. Adverse events were not noted as an outcome in the methods section

Notes

Informed consent provided Study conducted in Spain Participants received 10% povidone iodine solution as an antiseptic Adverse events, including reoperation for deep infection were not reported on in the study


Bias

Authors' judgement



Comment: email response on 25 November 2012, from Dr Soriano (lead author) stated: "The randomization was computer generated and under the control of the pharmacist at his hospital."


Comment: allocation occurred the morning of the procedure and well in advance of the procedure - per email response from Dr Soriano (above). not clear whether concealment of allocation was maintained

unclear


Comment: clinicians administering the antibiotic were aware of the allocation scheme but participants were blinded to the type of antibiotic received


low

Comment: clinicians assessing the outcome were blinded to the treatment arm according to email from Dr Soriano (above)


low

Comment: all participants were reported on in the results section


Comment: infection listed as an outcome in methods section and was reported on in results section. However, a trial protocol was not obtained

Other bias


low

Suter 1994 35 Methods Participants

Single center RCT, conducted from March 1988-February 1990 496 people requiring a THA Group 1 (n = 250): average age = 66.5 years (SD 8.8); male:female = 65:185; 222/250 (88.8%) diagnosed with osteoarthritis Group 2 (n = 246): average age = 68.2 years (SD 8.1); male:female = 75:171; 209/246 (85%) diagnosed with osteoarthritis Exclusion criteria: history of allergic reaction to cephalosporins or glycopeptides, pregnancy or lactation, renal insufficiency, local or systemic infections, and treatment with antibiotics within the previous 2 weeks

Interventions

Group 1: teicoplanin 400 mg as a single IV bolus provided 60-90 51

min before surgery Group 2: cefamandole 2 g IV provided 60-90 min prior to surgery plus 1 g cefamandole injected at end of surgery Outcomes

Notes

Infection rates over a 2 year period. Infection defined as: pain, local tenderness, abnormal erythrocyte sedimentation rate, radiographic signs of infection or positive bacterial cultures of the periprosthetic space Study conducted in Italy Obtained oral informed consent


Authors' judgement



Quote: "A random list was generated from casual numbers tables."



unclear


Comment: clinicians administering antibiotics were not blinded to allocation. While not specifically stated, it appears that participants were blinded to allocation (Introduction states that this was a "singleblind study").


unclear



high

Quote: "Ten patients in teicoplanin group and 14 in cefamandole group did not have a sufficient followup period and thus were not included in the efficacy analysis."


Comment: infection listed as an outcome in methods section and evaluated in results section. However, a trial protocol was not obtained

Other bias


low

Tyllianakis 2010 36 Methods

Single center RCT, conducted from 2003-2006 52

Participants

435 people receiving either a porous-coated primary THA or partially-cemented (tibial component only) or totally-cemented (both tibial and femoral component) TKA implant Average age of participants = 64 years Group 1 = 113 THA; 75 TKA Group 2 = 83 THA; 35 TKA Group 3 = 64 THA; 65 TKA Follow-up was for a mean period of 3.8 years Exclusion criteria: diabetes mellitus, chronic renal insufficiency, malignant tumor, receiving steroids (classified as potentially immunosuppressed), or asymptomatic bacteriuria

Interventions

Group 1: IV cefuroxime 1.5 g, 45 minutes prior to wound incision and 2 doses 750 mg cefuroxime 8 h and 16 h postoperatively Group 2: IV fusidic acid 500 mg (a bacteriostatic antibiotic useful against Gram-positive bacteria that exhibit antibiotic resistance), 1 h before wound incision and 2 doses 8 h and 16 h postoperatively Group 3: IV vancomycin 1 g (used to treat especially resistant bacteria such as MRSA), 1 h before wound incision and 12 h and 24 h postoperatively

Outcomes

Infection rate (superficial or deep) evaluated over a mean 3.8 year period. Infection was defined as: • •

superficial: erythema, local pain, and tenderness, and commonly drainage deep: radiographic changes, fever and night sweats; drainage and dehiscence of the surgical incision; fascia penetrate

While a cost-effectiveness analysis was not a pre-defined endpoint identified in the methods section, it was mentioned in the conclusion that broad spectrum cefuroxime was less expensive than specific antistaphylococcal agents Notes

Study conducted in Greece

53


Authors' judgement


Support for judgement Quote: "Randomization was performed according to the room number where patients were hospitalized. There were three different nursing rooms with five beds each. The nurse who admitted patients to each room were blinded to the protocol." Comment: The randomization scheme was explained in an email received from Dr Karageorgos on 28 December 2012


unclear


Comment: allocation concealment not reported Quote: "Physicians knew to which antibiotic group patients were allocated to and patients were blinded to treatment arm." Comment: Quote from email received from Dr Karageorgos (as above)


low

Comment: according to email from Dr Karageorgos (as above) clinical assessors were blinded to treatment arms


low

Quote: "Four hundred thirty five (435) out of 435 patients were followed for a mean time of 3.8 (2-5) years."


Comment: infection was listed as an outcome in the methods section and reported on in results section. However, a trial protocol was not obtained

Other bias


low

Vainionpää 1988 37 Methods Participants

Single-center RCT conducted between 1983-1984 58 patients undergoing primary THA (n = 40) or TKA (n = 18): 54

mean age = 67.4 years (54-79); male:female = 14:44 Exclusion criteria: revision arthroplasties Interventions

Group 1 (TKA; n = 9): 2 g cefamandole systemic IV over a 10 min period after anesthesia had been established and just before the operation began (note: tourniquet was inflated approximately 15 min after antibiotic administered), Then 1 g cefamandole every 6 h parenterally (IV) for 3 days Group 2 (TKA; n = 9): 2 g cloxacillin systemic IV over a 10 min period after anesthesia had been established and just before the operation began (note: tourniquet was inflated approximately 15 min after the antibiotic was given). Then 2 g cloxacillin every 8 h parenterally (IV) for 1 day and then 1 g dicloxacillin orally for 2 days Group 3 (THA; n = 20): 2 g cefamandole systemic IV over a 10 min period after anesthesia had been established and just before the operation began, then 1 g cefamandole every 6 h parenterally (IV) for 3 days Group 4 (THA: n = 20): 2 g cloxacillin systemic IV over a 10 min period after anesthesia had been established and just before the operation began, then 2 g cloxacillin every 8 h parenterally (IV) for 1 day and then 1 g dicloxacillin orally for 2 days

Outcomes

Concentrations of cefamandole and cloxacillin in the blood serum (in mg/L) at 16-30, 31-45, and 46-60 min after antibiotic administration in 58 participants Concentrations of cefamandole and cloxacillin in the synovial fluid of 28 participants by means of needle aspiration prior to the opening of the joint capsule Adverse events including allergic reaction (not a defined outcome in the methods section, but evaluated in the results section) Deep infection or loosening of the implant during 2-year follow up (not a defined endpoint in the methods section but evaluated in the results section). The definition for deep infection was not reported

Notes

Study conducted in Finland, with permission from the Central Medical Board of Finland. No statistical analysis was provided. 55

A grant from the Eli Lilly Corporation paid for the laboratory analysis for this study. Eli Lilly is a manufacturer of cefamandole and cloxacillin


Authors' judgement






unclear


Comment: anesthesiologists who administered the antibiotic knew which type was given. Blinding of participants was not reported


unclear

Comment: it was not mentioned whether outcome assessors were blinded


low

Comment: all 58 participants were assessed for antibiotic concentration


Comment: antibiotic concentration was identified as an outcome in the methods section and was reported on in the results section. However, a trial protocol was not obtained

Other bias

Comment: Eli Lilly provided a grant for the laboratory analysis. Eli Lilly is a manufacturer of both cloxacillin and cefamandole

high

van Rijen 2012 38 Methods

Single-center double-blind RCT conducted from October 2005June 2007 56

Participants

Patients undergoing primary THA (n = 50) or TKA (n = 45). Baseline characteristics could not be obtained despite repeated emails to the author between 27 March-7 April 2014 Exclusion criterion: < 18 years of age; presence of an active infection with S aureus at the time of randomization; known allergy to mupirocin or chlorhexidine; pregnant or breast-feeding; use of mupirocin in the preceding 4 weeks; and the presence of a nasal foreign body

Interventions

All participants were screened on admission for the presence of S aureus in their nasal passages via a swab and real-time polymerase-chain reaction (PCR) assay. Participants who tested positive for S aureus were randomized to the following treatments: Group 1: mupirosin (Bactroban, GlaxoSmithKline) nasal ointment applied twice daily and chlorhexidine gluconate soap (Hibiscrub, Mölnlycke) for a period of 5 days. Participants still in the hospital after 3 weeks and after 6 weeks received a 2nd and 3rd 5-day course of trial medication. A cephalosporin was administered at some point, but we could not determine when or what amount of cephalosporin Group 2: placebo nasal ointment applied twice daily and placebo chlorhexidine soap for a period of 5 days. Participants still in the hospital after 3 weeks and after 6 weeks received a 2nd and 3rd 5-day course of trial medication (placebo) Same issue with cephalosporin as in group 1.

Outcomes

Primary outcome: cumulative incidence of hospital-acquired S aureus infections over a 6-week period after the procedure. Quote: "Patients were monitored for hospital-acquired S. aureus infection by means of microbiologic cultures. Attending physicians were encouraged to obtain culture samples if infection was suspected. If a culture grew S. aureus, the patient's medical record was reviewed to distinguish infection from colonization and to determine whether the infection was health careassociated according to criteria established by the Centers for Disease Control and Prevention." Secondary outcomes: all-cause in-hospital mortality, duration of hospitalization, and time from admission to the onset of healthcare-associated S aureus infections. Actual hospital costs were included in the cost analysis, which included the surgery and the 6 weeks of follow-up

57

Notes

Study conducted in the Netherlands Informed consent obtained The cost analysis was not an initial trial outcome, but was undertaken on a post hoc basis with the investigators evaluating costs while blinded to the treatment arms Repeated attempts were made via email to obtain a breakdown of the information mentioned above from the author (9 emails sent between 27 March - 7 April 2014). The author however stated she did not have the time to produce this information as she was studying for her PhD.


Authors' judgement

Random sequence generation low (selection bias) Allocation concealment (selection bias)

unclear

Support for judgement Quote: "A single list of random numbers with a permuted-block design was generated by an independent statistician" Comment: allocation concealment was not reported


Comment: the trial was double-blind, which suggests that participants were not aware of their treatment arm and that the placebo ointment and soap were identical in appearance to the active treatment and that clinicians administering the treatment were not aware of the treatment arms


unclear

Comment: it was not clear from the trial whether clinicians assessing for infection were aware to which treatment arm each participant had been allocated


low

Quote from author in 4 April 2014 email: "None of the patients were excluded from the analysis."


Comment: all outcomes defined in the methods section were reported on. However, a clinical trial protocol was not obtained

Other bias

Quote: "The manufacturers of the products provided the trial medications and placebo at no cost but did not influence the study design, data collection,

low

58

analysis, writing, or decision to submit the results for publication."

Wymenga 1992 39 Methods Participants

Multicenter (27 hospitals) prospective RCT conducted from 1 July 1986-1 July 1988 Patients receiving a THA Group 1 (n = 1327): mean age = 69 years (SD 11); male:female = 287:1040 Group 2 (n = 1324): mean age = 69 years (SD 11); male:female = 266:1058 Exclusion criteria: allergy to cephalosporin; penicillin anaphylaxis; use of antibiotics less than 48 h before the operation; the use of perioperative antibiotics other than cefuroxime; malignancy; former or current sepsis in the joint; or the use of gentamicin-impregnated bone cement for prosthesis fixation

Interventions

Group 1: 1 x 1500 mg dose IV cefuroxime administered 30 min prior to procedure Group 2: 1500 mg doses IV cefuroxime 30 minutes prior to procedure, plus 750 mg cefuroxime provided at 8 h and 16 h post procedure

Outcomes

Joint sepsis/infection defined as: confirmed sepsis (positive bacteriologic culture); strong evidence sepsis (4 or more possible signs of infection) or wound infection (erythema more than 1 cm from incision). Confirmed sepsis plus strong evidence of sepsis was defined as a Category 1 Reoperation Death Mean follow-up was 13 months

Notes

The reviewers were unable to break TKA recipients down into randomized groups - therefore we only reported on hip implants A power calculation was performed

59

Study conducted in the Netherlands


Authors' judgement



Quote: "Randomization of prophylaxis regimen performed using a computerized list in blocks of 10 numbers."


high

Comment: randomization was performed in advance of the procedure and the anesthesiologist (performing antibiotic prophylaxis) knew the group to which participants were allocated prior to treatment


Quote: "Label [specifying treatment group] placed on the patient's records and the anesthesia list prior to treatment"


unclear



high

Comment: 145 participants withdrew (reasons for withdrawal: other type of antibiotic used: Group 1 = 17, Group 2 = 16; wrong dose or administration: Group 1 = 42, Group 2 = 35; died: Group 1 = 4, Group 2 = 8). These participants were reported on and were free from infection. 15 participants were lost to follow-up


Comment: joint sepsis was defined in the methods section and was reported on. However, a trial protocol was not obtained

Other bias


low

60

Characteristics of excluded studies:

Bryan 1988 40 Reason for exclusion

RCT of cefazolin versus cefamandole in total joint arthroplasty primary hip, primary knee, and revision procedures. However, unable to separate the infections reported by procedure type. An attempt was made to contact the lead author, Charles S Bryan MD, to obtain this information. An email was sent to [email protected] on 14 February 2014, but bounced back as undeliverable

Chiu 2001 41 Reason for exclusion

QRCT with randomization performed on an odd-even basis according to medical record number. An RCT examining this comparison (systemic IV antibiotic prophylaxis versus systemic IV antibiotic prophylaxis plus antibiotic cement) was available (Hinarejos 2013), so this QRCT was excluded

Chiu 2002 42 Reason for exclusion

QRCT with randomization performed on an odd-even basis according to medical record number. An RCT examining this comparison (systemic IV antibiotic prophylaxis versus systemic IV antibiotic prophylaxis plus antibiotic cement) was available (Hinarejos 2013), so this QRCT was excluded

Gilliam 1990 43 Reason for exclusion

RCT comparing preoperative skin preparation using a 5 minute aqueous iodophor scrub followed by application of aqueous iodophor solution as a paint versus skin prepared with a 1-step application of water-insoluble iodophor in alcohol solution applied as a paint. 60 participants randomized. Not possible to determine if the total joint surgeries were primary or revision in nature and whether they were total knee or hip

61

Jones 1987 44 Reason for exclusion

RCT of various types of surgery comparing timentin to cefoxatime. 25 total joint patients randomized. Excluded because impossible to separate the infections and costs by procedure type (also unclear whether primary total joint implants were knee or hip)

Jones 1988 45,46 Reason for exclusion

RCT investigating cefoxatime versus other cephalosporins and ticarcillin/clavulanic acid in various surgical procedures, including total joint arthroplasty (320 participants). Excluded because impossible to separate out infections by procedure type

Liebergall 1995 47 Reason for exclusion

RCT excluded because impossible to separate out infections by procedure type

Mollan 1992 48 Reason for exclusion

RCT of teicoplanin versus cefamandole. Excluded because impossible to separate out infections by procedure type

Nelson 1983 49 Reason for exclusion

Study stated it was randomized, but a randomization scheme was not used, and participants were divided into 2 groups, according to the last digit (even or odd) of their hospital number

Periti 1994 50 Reason for exclusion

RCT; unclear how many people received a hip or knee implant. Excluded because impossible to separate infections by procedure type

62

Periti 1999 51 Reason for exclusion

RCT excluded because impossible to separate infections by procedure type

van den Brand 2001 52 Reason for exclusion

Single-center RCT. Evaluated postoperative bacteruria or UTI in participants receiving either an indwelling or intermittent Foley catheter. Trial performed on both TKA and THA. Excluded because impossible to separate incidence of postoperative bacteruria or UTI according to type of implant received

van Kasteren 2007 53 Reason for exclusion

Retrospective review of timing of antibiotic prophylaxis

Wall 1988 54 Reason for exclusion

RCT excluded because implant was an endoprosthetic hip or knee (i.e. only one portion of the joint was replaced) and was not a total hip or knee replacement

Winter 1987 55 Reason for exclusion

RCT, however, it was not clear whether the type of antibiotic prophylaxis administered was determined by the type of procedure (hip or knee). Also not clear whether implant was a total knee or hip implant as implants were identified as endoprostheses

Zdeblick 1986 56 Reason for exclusion

RCT of elective adult orthopedic surgical cases that excluded total joint arthroplasty patients

Appendix 3: Effects of interventions 63

Outcome

Comparison

Analysis finding - Risk ratio (RR); Confidence interval (CI); P value and I2 statistic for heterogeneity (where appropriate)

Infection Preoperative systemic IV antibiotic prophylaxis versus placebo THA - Infection at 6 months

Three RCTs (Ericson

RR 0.23; 95% CI 0.12 to

1973; Gunst 1984; Hill 1981)

0.43; P value < 0.0001; I2 0% (Figure 4)

THA - Infection at 2.5 years

Two RCTs (Ericson 1973;

RR 0.15; 95% CI 0.05 to

Schulitz 1980)

0.47; P value 0.001; I2 26% (Figure 5)

THA - Infection at 5+ years

Two RCTs Ericson 1973; Hill RR 0.19; 95% CI 0.10 to 1981

0.35; P value < 0.00001; I2 0% (Figure 6)

Timing and route of antibiotics - systemic IV antibiotic compared with intraoperative local

64

Outcome

Comparison

Analysis finding - Risk ratio (RR); Confidence interval (CI); P value and I2 statistic for heterogeneity (where appropriate)

antibiotic prophylaxis (antibiotic cement) THA – incidence of infection

Josefsson 1981 - compared

Incidence of deep infections

at 2-10 years

gentamycin impregnated

was significantly higher in

cement to beta-lactam

the systemic IV antibiotic

antibiotics (cloxacillin,

group than in the

dicloxacillin) or first

intraoperative local

generation cephalosporins

antibiotic-cement group at

(cephalexine) or penicillins

two years (P value < 0.01; 13

active against Gram-positive

versus three deep infections)

bacteria (i.e.

and at five years (P value

Appendix 1: MeSH search terms and data collect form used to collect

Appendix 1: MeSH search terms and data collect form used to collect

Suggest Documents

English version questionnaire used to collect data in Central ... - PLOS

Skout Collect

APPLICANTS TO COLLECT LETTERS

Where to collect seed?

minis to collect! - Thomas & Friends

Using R to Collect, Analyze and Visualize Graduate Attribute Data

How to Correctly Collect and Analyze Racial Profiling Data:

LO Dinghy 2000.collect

Mary Stewart Collect

The collect and collapse process

Using Google Forms to collect and analyze data

Preparing and conducting interviews to collect data

A Web-Based Platform to Collect Data from ESRD Patients ...

Using Paired Depth Interviews to Collect Qualitative Data - NSUWorks

Workshop Synthesis: Improving Methods to Collect Data ... - TU Dresden

How to Collect High-Quality Colorectal Cancer Surveillance Data

Workshop Synthesis: Improving Methods to Collect Data ... - TU Dresden

Non-interventional study to collect data for the ... - Springer Link

Backcountry Limnology: Engaging Citizen Scientists to Collect Data on ...

Top Tips To Collect Your Debts Quicker

How to Collect Your Payee Information

The FTC's Authority to Collect Transaction Data | American Antitrust ...

iSurv: Using iPhone to collect visitor data for ...

Using Vehicular Networks to Collect Common Traffic Data