Tissue Adhesives for Closure of Surgical Incisions (Review)

Tissue adhesives for closure of surgical incisions (Review)

Dumville JC, Coulthard P, Worthington HV, Riley P, Patel N, Darcey J, Esposito M, van der Elst M, van Waes OJF

This is a reprint of a Cochrane review, prepared and maintained by The Cochrane Collaboration and published in The Cochrane Library 2014, Issue 11 http://www.thecochranelibrary.com

HEADER...... 1 ABSTRACT ...... 1 PLAINLANGUAGESUMMARY ...... 2 SUMMARY OF FINDINGS FOR THE MAIN COMPARISON ...... 3 BACKGROUND ...... 5 OBJECTIVES ...... 5 METHODS ...... 5 RESULTS...... 8 Figure1...... 13 Figure2...... 14 ADDITIONALSUMMARYOFFINDINGS ...... 20 DISCUSSION ...... 31 AUTHORS’CONCLUSIONS ...... 32 ACKNOWLEDGEMENTS ...... 32 REFERENCES ...... 33 CHARACTERISTICSOFSTUDIES ...... 36 DATAANDANALYSES...... 95 Analysis 1.1. Comparison 1 Adhesive versus suture, Outcome 1 Dehiscence: all studies...... 97 Analysis 1.2. Comparison 1 Adhesive versus suture, Outcome 2 Dehiscence: sensitivity analysis...... 99 Analysis 1.3. Comparison 1 Adhesive versus suture, Outcome 3 Infection: all studies...... 100 Analysis 1.4. Comparison 1 Adhesive versus suture, Outcome 4 Infection: sensitivity analysis...... 101 Analysis 1.5. Comparison 1 Adhesive versus suture, Outcome 5 Cosmetic appearance rated by patient...... 102 Analysis 1.6. Comparison 1 Adhesive versus suture, Outcome 6 Cosmetic appearance rated by surgeon...... 102 Analysis 1.7. Comparison 1 Adhesive versus suture, Outcome 7 Patient/parent satisfaction (% satisfied)...... 103 Analysis 1.8. Comparison 1 Adhesive versus suture, Outcome 8 Patient/parent satisfaction (VAS Scale 0 to 100). . . 103 Analysis 1.9. Comparison 1 Adhesive versus suture, Outcome 9 Surgeon satisfaction (% satisfied)...... 104 Analysis 1.10. Comparison 1 Adhesive versus suture, Outcome 10 Time taken for wound closure...... 105 Analysis 2.1. Comparison 2 Adhesive versus adhesive tape, Outcome 1 Dehiscence...... 106 Analysis 2.2. Comparison 2 Adhesive versus adhesive tape, Outcome 2 Infection...... 106 Analysis 2.3. Comparison 2 Adhesive versus adhesive tape, Outcome 3 Cosmetic appearance rated by patient (VAS). 107 Analysis 2.4. Comparison 2 Adhesive versus adhesive tape, Outcome 4 Cosmetic appearance rated by patient (% satisfied)...... 108 Analysis 2.5. Comparison 2 Adhesive versus adhesive tape, Outcome 5 Cosmetic appearance rated by surgeon (VAS). 108 Analysis 2.6. Comparison 2 Adhesive versus adhesive tape, Outcome 6 Patient satisfaction...... 109 Analysis 2.7. Comparison 2 Adhesive versus adhesive tape, Outcome 7 Surgeon satisfaction...... 109 Analysis 2.8. Comparison 2 Adhesive versus adhesive tape, Outcome 8 Time taken for wound closure...... 110 Analysis 3.1. Comparison 3 Adhesive versus staples, Outcome 1Dehiscence...... 110 Analysis 3.2. Comparison 3 Adhesive versus staples, Outcome 2Infection...... 111 Analysis 3.3. Comparison 3 Adhesive versus staples, Outcome 3 Cosmetic appearance rated by patient (scar scale). . 111 Analysis 3.4. Comparison 3 Adhesive versus staples, Outcome 4 Cosmetic appearance by plastic surgeons (VAS). . . 112 Analysis 3.5. Comparison 3 Adhesive versus staples, Outcome 5 Patient satisfaction...... 113 Analysis 3.6. Comparison 3 Adhesive versus staples, Outcome 6 Time taken for wound closure...... 113 Analysis 4.1. Comparison 4 Adhesive versus other method, Outcome 1 Dehiscence...... 114 Analysis 4.2. Comparison 4 Adhesive versus other method, Outcome 2 Infection...... 114 Analysis 4.3. Comparison 4 Adhesive versus other method, Outcome 3 Patient satisfaction...... 115 Analysis 4.4. Comparison 4 Adhesive versus other method, Outcome 4 Clinician satisfaction...... 116 Analysis 4.5. Comparison 4 Adhesive versus other method, Outcome 5 Time taken for wound closure...... 116 Analysis 5.1. Comparison 5 Adhesive versus adhesive: High viscosity versus low viscosity, Outcome 1 Dehiscence. . 117 Analysis 5.2. Comparison 5 Adhesive versus adhesive: High viscosity versus low viscosity, Outcome 2 Infection. . . 117 Analysis 5.3. Comparison 5 Adhesive versus adhesive: High viscosity versus low viscosity, Outcome 3 Patient satisfaction. 118

Tissue adhesives for closure of surgical incisions (Review) i Copyright © 2014 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Analysis 5.4. Comparison 5 Adhesive versus adhesive: High viscosity versus low viscosity, Outcome 4 Clinician satisfaction...... 118 Analysis 5.5. Comparison 5 Adhesive versus adhesive: High viscosity versus low viscosity, Outcome 5 Time taken for wound closure...... 119 Analysis 6.1. Comparison 6 Adhesive versus adhesive: octylcyanoacrylate versus butylcyanoacrylate, Outcome 1 Dehiscence...... 119 Analysis 6.2. Comparison 6 Adhesive versus adhesive: octylcyanoacrylate versus butylcyanoacrylate, Outcome 2 Infection...... 120 Analysis 6.3. Comparison 6 Adhesive versus adhesive: octylcyanoacrylate versus butylcyanoacrylate, Outcome 3 Cosmetic assessmentratedbypatient(VAS)...... 121 Analysis 6.4. Comparison 6 Adhesive versus adhesive: octylcyanoacrylate versus butylcyanoacrylate, Outcome 4 Cosmetic assessmentratedbysurgeon(VAS)...... 121 Analysis 6.5. Comparison 6 Adhesive versus adhesive: octylcyanoacrylate versus butylcyanoacrylate, Outcome 5 Surgeon satisfaction(withdevice)...... 122 Analysis 6.6. Comparison 6 Adhesive versus adhesive: octylcyanoacrylate versus butylcyanoacrylate, Outcome 6 Surgeon satisfaction(withclosure)...... 122 Analysis 6.7. Comparison 6 Adhesive versus adhesive: octylcyanoacrylate versus butylcyanoacrylate, Outcome 7 Time taken forwoundclosure...... 123 ADDITIONALTABLES...... 123 APPENDICES ...... 126 WHAT’SNEW...... 132 HISTORY...... 133 CONTRIBUTIONSOFAUTHORS ...... 133 DECLARATIONSOFINTEREST ...... 134 SOURCESOFSUPPORT ...... 135 DIFFERENCES BETWEEN PROTOCOL AND REVIEW ...... 135 INDEXTERMS ...... 135

Tissue adhesives for closure of surgical incisions (Review) ii Copyright © 2014 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. [Intervention Review] Tissue adhesives for closure of surgical incisions

Jo C Dumville1, Paul Coulthard2, Helen V Worthington3, Philip Riley3, Neil Patel4, James Darcey2, Marco Esposito3, Maarten van der Elst5, Oscar J F van Waes5

1School of Nursing, Midwifery and Social Work, University of Manchester, Manchester, UK. 2Department of Oral and Maxillofacial Surgery, School of Dentistry, The University of Manchester, Manchester, UK. 3Cochrane Oral Health Group, School of Dentistry, The University of Manchester, Manchester, UK. 4Oral Surgery, University Dental Hospital of Manchester, Manchester, UK. 5Department of Surgery, Reinier de Graaf Groep, Delft, Netherlands

Contact address: Jo C Dumville, School of Nursing, Midwifery and Social Work, University of Manchester, Manchester, M13 9PL, UK. [email protected].

Editorial group: Cochrane Wounds Group. Publication status and date: New search for studies and content updated (no change to conclusions), published in Issue 11, 2014. Review content assessed as up-to-date: 12 March 2014.

Citation: Dumville JC, Coulthard P, Worthington HV, Riley P, Patel N, Darcey J, Esposito M, van der Elst M, van Waes OJF. Tissue adhesives for closure of surgical incisions. Cochrane Database of Systematic Reviews 2014, Issue 11. Art. No.: CD004287. DOI: 10.1002/14651858.CD004287.pub4.

ABSTRACT Background Sutures (stitches), staples and adhesive tapes have been used for many years as methods of wound closure, but tissue adhesives have entered clinical practice more recently. Closure of wounds with sutures enables the closure to be meticulous, but the sutures may show tissue reactivity and can require removal. Tissue adhesives offer the advantages of an absence of risk of needlestick injury and no requirement to remove sutures later. Initially, tissue adhesives were used primarily in emergency room settings, but this review looks at the use of tissue adhesives in the operating room/theatre where surgeons are using them increasingly for the closure of surgical skin incisions. Objectives To determine the effects of various tissue adhesives compared with conventional skin closure techniques for the closure of surgical wounds. Search methods In March 2014 for this second update we searched the Cochrane Wounds Group Specialised Register; The Cochrane Central Register of Controlled Trials (CENTRAL) (The Cochrane Library); Ovid MEDLINE; Ovid MEDLINE (In-Process & Other Non-Indexed Citations); Ovid EMBASE and EBSCO CINAHL. We did not restrict the search and study selection with respect to language, date of publication or study setting. Selection criteria Only randomised controlled trials were eligible for inclusion. Data collection and analysis We conducted screening of eligible studies, data extraction and risk of bias assessment independently and in duplicate. We expressed results as random-effects models using mean difference for continuous outcomes and risk ratios (RR) with 95% conﬁdence intervals (CI) for dichotomous outcomes. We investigated heterogeneity, including both clinical and methodological factors.

Tissue adhesives for closure of surgical incisions (Review) 1 Copyright © 2014 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Main results This second update of the review identified 19 additional eligible trials resulting in a total of 33 studies (2793 participants) that met the inclusion criteria. There was low quality evidence that sutures were significantly better than tissue adhesives for reducing the risk of wound breakdown (dehiscence; RR 3.35; 95% CI 1.53 to 7.33; 10 trials, 736 participants that contributed data to the meta-analysis). The number needed to treat for an additional harmful outcome was calculated as 43. For all other outcomes - infection, patient and operator satisfaction and cost - there was no evidence of a difference for either sutures or tissue adhesives. No evidence of differences was found between tissue adhesives and tapes for minimising dehiscence, infection, patients’ assessment of cosmetic appearance, patient satisfaction or surgeon satisfaction. However there was evidence in favour of using tape for surgeons’ assessment of cosmetic appearance (mean difference (VAS 0 to 100) 9.56 (95% CI 4.74 to 14.37; 2 trials, 139 participants). One trial compared tissue adhesives with a variety of methods of wound closure and found both patients and clinicians were significantly more satisfied with the alternative closure methods than the adhesives. There appeared to be little difference in outcome for different types of tissue adhesives. One study that compared high viscosity with low viscosity adhesives found that high viscosity adhesives were less time-consuming to use than low viscosity tissue adhesives, but the time difference was small. Authors’ conclusions Sutures are significantly better than tissue adhesives for minimising dehiscence. In some cases tissue adhesives may be quicker to apply than sutures. Although surgeons may consider the use of tissue adhesives as an alternative to other methods of surgical site closure in the operating theatre, they need to be aware that sutures minimise dehiscence. There is a need for more well designed randomised controlled trials comparing tissue adhesives with alternative methods of closure. These trials should include people whose health may interfere with wound healing and surgical sites of high tension.

PLAIN LANGUAGE SUMMARY Tissue adhesives for closure of surgical skin incisions Tissue adhesives or glues are increasingly used in place of stitches (sutures) or staples to close wounds. It has been suggested that tissue adhesives may be quicker and easier to use than sutures for closing surgical wounds. Tissue adhesives carry no risk of sharps injury - unlike needles that are used for sutures - and are thought to provide a barrier to infection. This may mean that they also promote healing, and the need for removal of sutures is avoided. The researchers searched the medical literature up to March 2014, and identiﬁed 33 medical studies that investigated the use of tissue adhesives for closure of wounds. They compared tissue adhesive with another method of closure such as sutures, staples, tape, or another type of tissue adhesive. The main outcomes of interest were whether wounds stayed closed - and did not break down - and whether they became infected. The results of the review showed clearly that fewer wounds broke down when sutures were used. Studies also reported that some types of tissue adhesives might be slightly quicker to use than other types. There was no clear difference between tissue adhesives and the alternative closure methods for cosmetic results or costs. Results regarding surgeons’ and patients’ preferred skin closure method were mixed.

Tissue adhesives for closure of surgical incisions (Review) 2 Copyright © 2014 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. oyih 04TeCcrn olbrto.Pbihdb J by Published Collaboration. Cochrane The (Review 2014 incisions © surgical Copyright of closure for adhesives Tissue SUMMARYOFFINDINGSFORTHEMAINCOMPARISON [Explanation]

Tissue adhesive compared to sutures for surgical incisions

Patient or population: People with surgical incisions Settings: Intervention: Tissue adhesive Comparison: Sutures

Outcomes Illustrative comparative risks*4 (95% CI) Relative effect No of participants Quality of the evidence Comments (95% CI) (studies) (GRADE)

Assumed risk Corresponding risk ) Sutures Tissue adhesive h ie os Ltd. Sons, & Wiley ohn Wound dehiscence Study population RR 3.35 736 ⊕⊕ (1.53 to 7.32) (10 studies) low1,2 13 per 1000 45 per 1000 (21 to 99)

Moderate

Wound infection Study population RR 1.72 744 ⊕ (0.92 to 3.16) 10 studies very low2,3 38 per 1000 76 per 1000 (14 to 397)

Moderate

*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI) CI: Confidence interval; RR: Risk ratio 3 oyih 04TeCcrn olbrto.Pbihdb J by Published Collaboration. Cochrane The (Review 2014 incisions © surgical Copyright of closure for adhesives Tissue

GRADE Working Group grades of evidence High quality: further research is very unlikely to change our confidence in the estimate of effect Moderate quality: further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate Low quality: further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate Very low quality: we are very uncertain about the estimate

1 Possible unit of analyses issues. A sensitivity analysis changes a statistically significant difference to a non-statistically significant difference 2 Study 95% CIs are wide 3 Possible unit of analysis issues 4 Median control (suture) group risk across studies xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx ) h ie os Ltd. Sons, & Wiley ohn 4 BACKGROUND Why it is important to do this review It is important that clinical decision-makers are able to make evidence-informed decisions regarding the use of tissue adhesives to Description of the condition close surgical incisions. This review was ﬁrst published in 2004. As tissue adhesives become more widely used, more randomised Millions of surgical procedures are conducted around the world controlled trials are conducted, and this update is required to in- each year. The majority of procedures result in surgical wounds corporate this new evidence. that will heal by primary intention - this is where wound edges are brought together (re-approximated) and held together, e.g. with sutures, to facilitate tissue healing.

OBJECTIVES Description of the intervention To determine the effects of various tissue adhesives compared with In the past the options for wound closure have been limited largely conventional skin closure techniques for the closure of surgical to sutures (needle and thread) with other alternatives such as sta- wounds. ples, adhesive tapes and tissue adhesives entering clinical practice more recently. Closure of wounds with sutures enables meticulous closure, but skin may react to sutures and they usually require METHODS removal. Tissue adhesives (glues) offer the advantages that suture removal is not required at a later date and there is no risk of needlestick injury to the surgeon or assistant. Criteria for considering studies for this review Tissue adhesives have been used in various forms for many years since the first cyanoacrylate adhesives were synthesised (Coover 1959). The early adhesives were appropriate for small superficial Types of studies lacerations and incisions, but their limited physical properties pre- Randomised controlled trials (RCTs). vented use in the management of other wounds. There were also reports of acute and chronic inflammatory reactions (Houston 1969). Further development led to the introduction of the n-2- Types of participants butylcyanoacrlyates that were purer and stronger, but did not re- People of any age and in any setting requiring closure of a surgical ceive widespread acceptance because their clinical performance skin incision of any length. was limited by their low tensile strength and brittleness (Bruns 1996; Quinn 1993). More recently stronger tissue adhesives have been developed by combining plasticisers and stabilisers to increase Types of interventions flexibility and reduce toxicity (Quinn 1997). Surgical skin incision closure with tissue adhesive compared with Tissue adhesives have been used primarily in emergency rooms and another tissue adhesive or any alternative conventional closure there is increasing support in the literature for their effectiveness in device such as sutures, staples or adhesive tapes (e.g. Steri-Strips/ the closure of various traumatic lacerations (Farion 2001; Osmond butterfly stitches). 1999; Perron 2000; Quinn 1997). Surgeons now also use tissue adhesives in the operating room for the closure of surgical skin incisions. Types of outcome measures

Primary outcomes How the intervention might work Proportion of wounds that break down (wound • The introduction of tissue adhesives was received enthusiastically dehiscence). as they may produce equivalent tensile strength, improved cosmetic appearance of the scar and a lower infection rate when compared with sutures, staples and adhesive tapes, while avoiding many Secondary outcomes of the risks and disadvantages of alternative methods (Osmond Proportion of infected wounds (using the study • 1999). As with standard adhesives, tissue adhesives are applied to investigator’s diagnosis of infection). the wound in a liquid form - following application they undergo Cosmetic appearance at or after three months where the • polymerisation and bonding and setting occurs. investigator has used a validated measure.

Tissue adhesives for closure of surgical incisions (Review) 5 Copyright © 2014 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Patient (or in the case of studies performed with #13 MeSH descriptor: [Acrylates] explode all trees 1757 • participants under the age of 16 years, parents’) general #14 acrylate* 254 satisfaction with skin incision closure technique (this is more #15 MeSH descriptor: [Bucrylate] explode all trees 0 than cosmetic appearance as other factors such as suture removal #16 bucrylate* 6 experience may have an input). #17 #4 or #5 or #6 or #7 or #8 or #9 or #10 or #11 or #12 or # Surgeon satisfaction with skin incision closure technique 13 or #14 or #15 or #16 2397 • (this may take into account the time for surgeon to close skin #18 #3 and #17 225 incision amongst other factors). The search strategies for Ovid MEDLINE, Ovid EMBASE and Relative cost of materials required for the skin incision Ovid CINAHL can be found in Appendix 3, Appendix 4 and • closure techniques being compared (this will be reported in a Appendix 5 respectively. The Ovid MEDLINE search was com- narrative form). bined with the Cochrane Highly Sensitive Search Strategy for Time taken to wound closure (at end of surgery) has been identifying randomised trials in MEDLINE: sensitivity- and pre- • added to the review as a post hoc outcome measure. The review cision-maximizing version (2008 revision); (Lefebvre 2011). We authors believe this to be a contributory factor towards both combined the EMBASE search with the Ovid EMBASE filter de- cost-effectiveness and satisfaction. veloped by the UK Cochrane Centre (Lefebvre 2011). We combined the CINAHL searches with the trial filters developed by the Scottish Intercollegiate Guidelines Network (SIGN 2011). Search methods for identification of studies For an outline of the search methods used in the original version Searching other resources of this review see Appendix 1 The reviewers checked the bibliographies of new studies included For an outline of the search methods used in the first update of in this updated review for potentially eligible references that had this review see Appendix 2. not been identified in the electronic searches outlined above.

Electronic searches For this second update in March 2014 we searched the following Data collection and analysis electronic databases: Cochrane Wounds Group Specialised Register (searched 13 • Selection of studies March 2014); The Cochrane Central Register of Controlled Trials Two review authors independently examined the titles and ab- • (CENTRAL; The Cochrane Library 2014 Issue 1); stracts of all the articles identified by the search to identify poten- Ovid MEDLINE (1946 to March Week 1 2014); tially relevant trials and then assessed the full text of these arti- • Ovid MEDLINE (In-Process & Other Non-Indexed cles independently using a standardised form to check for eligibil- • Citations, 12 March 2014); ity in the review. Disagreements about inclusion were resolved by Ovid EMBASE (1974 to 12 March 2014); consensus or a further review author where necessary. Two review • EBSCO CINAHL (1982 to 13 March 2014). authors performed validity assessment and data extraction on all • studies meeting the inclusion criteria. Studies rejected at this stage The following strategy was used to search the Cochrane Central were recorded in a table of excluded studies and reasons for exclu- Register of Controlled Trials (CENTRAL): sion recorded. #1 MeSH descriptor: [Wounds and Injuries] explode all trees 15958 #2 surgical next wound* 3679 Data extraction and management #3 #1 or #2 19357 Data were extracted by at least two review authors independently #4 MeSH descriptor: [Tissue Adhesives] explode all trees 387 using specially designed data extraction forms. The data extraction #5 MeSH descriptor: [Fibrin Tissue Adhesive] explode all trees forms were piloted on several papers and modified as required 329 before use. Any disagreements were resolved by discussion and #6 tissue next adhesive* 655 third review author consulted where necessary. All study authors #7 MeSH descriptor: [Cyanoacrylates] explode all trees 154 were contacted for clarification, or to request missing information #8 octylcyanoacrylate* 52 where necessary. Data were excluded if further clarification could #9 Dermabond 44 not be obtained. #10 MeSH descriptor: [Enbucrilate] explode all trees 42 For each trial the following data were recorded: #11 enbucrilate 62 year of publication, country of origin and source of study • #12 butylcyanoacrylate* 7 funding;

Tissue adhesives for closure of surgical incisions (Review) 6 Copyright © 2014 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. details of the participants including demographic The reviewers intended to meta-analyse the results of trials with • characteristics and criteria for inclusion; a split wound (different parts of the same wound randomised to details of the type of intervention (adhesive, suture, staples alternate treatments), and split body designs (different wounds on • or adhesive tape) and type of adhesive (butylcyanoacrylate or the same participant randomised to alternative treatments) using octylcyanoacrylate); the inverse variance method for paired data if appropriate data details of the outcomes reported, including method of were available. • assessment and time intervals. Cosmetic appearance was included if the authors stated that was measured on a validated scale. ’Summary of findings’ tables Assessment of risk of bias in included studies In this second update we also present the main results of the review Two review authors independently assessed each eligible study for in ’Summary of findings’ tables. These tables present key informa- risk of bias using the Cochrane ‘Risk of bias assessment tool’. The tion concerning the quality of the evidence, the magnitude of the tool addresses six specific domains (see Appendix 6), namely se- effects of the interventions examined, and the sum of the available quence generation, allocation concealment, blinding, incomplete data for the main outcomes (Schunemann 2011a). The ’Summary outcome data, selective outcome reporting and other issues that of findings’ tables also include an overall grading of the evidence may potentially bias the study (Higgins 2011). They completed a related to each of the main outcomes using the GRADE (Grades ‘Risk of bias’ table for each eligible study, and a separate assessment of Recommendation, Assessment, Development and Evaluation) of blinding and completeness of outcome data for each outcome. approach. The GRADE approach defines the quality of a body of Discrepancies between review authors were resolved through dis- evidence as the extent to which one can be confident that an esti- cussion. Findings are presented using the ‘Risk of bias’ summary mate of effect or association is close to the true quantity of specific figure, which presents all of the judgements in a cross-tabulation interest. The quality of a body of evidence involves consideration of study by risk of bias domain. of within-trial risk of bias (methodological quality), directness of evidence, heterogeneity, precision of effect estimates and risk of publication bias (Schunemann 2011b). We present the following Assessment of heterogeneity outcomes in the ’Summary of findings’ tables: wound dehiscence; This assessment of clinical and methodological heterogeneity will • wound infection. be supplemented by information regarding statistical heterogene- • ity - assessed using the Chi² test (a significance level of P less than 0.10 will be considered to indicate statistically significant heterogeneity in conjunction with I² measure; Higgins 2003). I² exam- Subgroup analysis and investigation of heterogeneity ines the percentage of total variation across RCTs that is due to The reviewers intended to undertake a subgroup analysis of age heterogeneity rather than chance. In general I² values of 25%, or (under 18 years and 18 years or older), location of incision on less, may mean a low level of heterogeneity, and values of 75%, or body (face and body), length of incision (less than 4 cm and 4 cm more, indicate very high heterogeneity (Deeks 2011). or greater), and patient characteristics such as diabetes and source of trial funding (commercially or independently funded), however Data synthesis there were insufficient studies reporting these data to undertake these analyses. For dichotomous outcomes, the reviewers used risk ratios (RR) to express estimates of effect of an intervention, together with 95% confidence intervals (CI). For continuous outcomes, the reviewers Sensitivity analysis used mean differences and standard deviations to summarize the data for each group. Where there were studies of similar compar- Where possible we planned to undertake a sensitivity analysis to isons reporting the same outcome measures, we attempted a meta- examine the effect of randomisation, allocation concealment and analysis. Risk ratios were combined for dichotomous data, and blinded outcome assessment on the overall estimates of effect. In weighted mean differences (WMD) for continuous data, using a addition, we also planned a sensitivity analysis to examine the random-effects model (indicated as RE in the results section) as effect of including unpublished literature on the review’s findings. some heterogeneity between studies was anticipated. We planned In this second update of the review we conducted a post-hoc sen- to analyse time taken to close wound (after surgery) as survival sitivity analysis where we removed studies from key analyses where (time-to-event) data, using the appropriate analytical method (as we had concerns about possible unit of analysis issues (i.e. where per the Cochrane Reviewers’ Handbook guidance Deeks 2011), dehiscence and infection had been recorded over multiple time or as a continuous outcome, if data from all participants were periods and it was not clear from the analyses whether some par- available. ticipants had more than one event).

Tissue adhesives for closure of surgical incisions (Review) 7 Copyright © 2014 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. RESULTS UK (Dowson 2006; Jallali 2004; Keng 1989; Kent 2014; Krishnamoorthy 2009; Livesey 2009; Ridgway 2007; Sinha 2001), USA (Brown 2009; Eggers 2011; Greene 1999; Kouba Description of studies 2011; Maloney 2013; Sebesta 2004; Sniezek 2007; Switzer 2003; Tierny 2009; Toriumi 1998), and one was multicentre and inter- See Characteristics of included studies; Characteristics of excluded national (Blondeel 2004). All studies included adults, except for studies and Characteristics of studies awaiting classification . six that included children (Brown 2009; Cheng 1997; Ong 2002; Romero 2011; Toriumi 1998; van den Ende 2004). All trials were Characteristics of the trial setting and investigators of parallel group design except for two that had a split body design (different wounds on same participant randomised; Greene 1999; In the original review, of the 22 potentially eligible studies of Kouba 2011), and two that had a split wound design (different which eight were excluded (Alhopuro 1976; Gorozpe-Calvillo portions of the same wound randomised; Sniezek 2007; Tierny 1999; Jaibaji 2000; Kuo 2006; Orozco-Razon 2002; Singer 2002; 2009). Kent 2014 randomised participants with multiple wounds Silvestri 2006; Steiner 2000) and 14 were included (Blondeel (port sites) to treatments - with all wounds on the participant re- 2004; Cheng 1997; Dowson 2006; Greene 1999; Keng 1989; ceiving the same intervention. Some outcome data were presented Maartense 2002; Ong 2002; Ozturan 2001; Ridgway 2007; by wound rather than at the participant level, thus potentially fail- Shamiyeh 2001; Sinha 2001; Sniezek 2007; Switzer 2003; Toriumi ing to account for clustering. 1998). We excluded one study because no data were presented and Five included trials had more than two arms: Eggers 2011 had four none were received after we sent a written request to the authors arms and compared two tissue adhesives, staples and sutures for (Alhopuro 1976). Another study was excluded as we could not skin closure; Khan 2006 had three arms comparing tissue adhe- use the data for incision closure because they were combined with sive, sutures and staples; Maartense 2002 and Shamiyeh 2001 had data for laceration closure (Singer 2002); we wrote to the authors three arms comparing tissue adhesive, sutures and adhesive tape, to request data separated by group, but received no reply. The and Blondeel 2004 compared two types of tissue adhesive with third study was excluded because after full translation we discov- each other and also other non-tissue adhesive closure techniques ered not to be a randomised trial (Gorozpe-Calvillo 1999). We (a mixed comparison group). These studies are thus included in excluded two papers because their methodology was flawed. The multiple comparisons. first of these was excluded because the intervention group had a subcuticular suture in place that was thought to bias the outcome in favour of the intervention (Jaibaji 2000). The second was ex- Characteristics of interventions cluded because all participants had a subcuticular suture placed In total 24 of the 33 included studies compared tissue adhe- and the review authors thought this adjunct method of closure sive with sutures for incision closure (Avsar 2009; Brown 2009; would invalidate an independent assessment of the primary inter- Cheng 1997; Dowson 2006; Eggers 2011; Greene 1999; Jallali ventions (Kuo 2006). Three further studies were excluded, as two 2004; Keng 1989; Khan 2006; Kouba 2011; Krishnamoorthy were not randomised (Silvestri 2006; Steiner 2000), and the third 2009; Maartense 2002; Millan 2011; Mota 2009; Ong 2002; did not appear to be randomised, and although we sought clarifi- Ozturan 2001; Sebesta 2004; Shamiyeh 2001; Sinha 2001; cation from the trial authors, no reply was received (Orozco-Razon Sniezek 2007, Switzer 2003; Tierny 2009; Toriumi 1998; van 2002). In the current update of this review we obtained full text den Ende 2004). Three trials compared tissue adhesive with ad- for a total of 36 potentially eligible new studies. Twelve of these hesive tape (Maartense 2002; Romero 2011; Shamiyeh 2001). were excluded: three because they did not assess a relevant wound One of these studies also compared tissue adhesive with su- type - e.g. lacerations (Ak 2012; Quinn 1998; Wong 2011); three tures (Maartense 2002). Six trials compared adhesives with sta- because the studies did not report relevant outcomes (Chen 2010; ples (Amin 2009; Eggers 2011; Khan 2006; Livesey 2009; Pronio Chow 2010; Sun 2005); five were not considered to be RCTs (Giri 2011; Ridgway 2007). Three trials compared one type of tissue 2004; Matin 2003; Maw 1997; Spencker 2011; Sajid 2009), and adhesive with another type (Blondeel 2004; Kent 2014; Maloney one study because the closure approach was not the only system- 2013). Chibbaro 2009 compared a tissue adhesive with a compar- atic difference between groups (Ong 2010). We included 19 new ison arm where staples or sutures were used and Blondeel 2004 studies in this update, which led to a total of 33 included studies. compared tissue adhesives with any other skin closure techniques. Five additional studies are awaiting assessment. These comparisons are summarised in Table 1. The 33 included studies were conducted in Austria (Shamiyeh 2001), Hong Kong (Cheng 1997), the Netherlands (Maartense 2002; van den Ende 2004), Turkey (Avsar 2009; Ozturan Comparison 1: Tissue adhesive compared with sutures 2001), Ireland (Amin 2009), Italy (Chibbaro 2009; Pronio 2011), Austalia (Khan 2006), Mexico (Millan 2011), Portugal Butylcyanoacrylate versus sutures (Mota 2009), Germany (Romero 2011), China (Ong 2002);

Tissue adhesives for closure of surgical incisions (Review) 8 Copyright © 2014 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Seven studies investigated the use of butylcyanoacrylate compared Jallali 2004 compared octylcyanoacrylate tissue adhesive • with sutures (Cheng 1997; Dowson 2006; Eggers 2011; Keng with absorbable sutures in people undergoing laparoscopic 1989; Ozturan 2001; Sinha 2001; van den Ende 2004). cholecystectomy. Cheng 1997 compared butylcyanoacrylate tissue adhesive Khan 2006 compared octylcyanoacrylate tissue adhesive • • with 4.0 catgut suture in boys under 12 years of age requiring with Monocryl sutures in people undergoing either a total knee elective circumcision. arthroplasty or a total hip arthroplasty. Dowson 2006 compared butylcyanoacrylate tissue adhesive Krishnamoorthy 2009 compared octylcyanoacrylate tissue • • with interruptible, non-absorbable suture after laparoscopic adhesive with absorbable sutures in people undergoing procedures. saphenous vein harvesting. Eggers 2011 was a four-arm trial in participants undergoing Maartense 2002 compared octylcyanoacrylate tissue • • total knee arthroplasty, with one arm receiving adhesive with intracutaneous poliglecaprone interrupted sutures butylcyanoacrylate tissue adhesive and one receiving Monocryl in people requiring elective laparoscopic procedures. sutures. Millan 2011 compared cyanoacrylate tissue adhesive (no • Keng 1989 investigated butylcyanoacrylate in patients further details available) compared with monofilament sutures in • requiring groin incisions. Skin incisions were closed with either people undergoing skin biopsies. butylcyanoacrylate tissue adhesive or Dexon subcuticular Mota 2009 compared octylcyanoacrylate tissue adhesive • sutures. Three of the 43 patients had bilateral operations when with rapidly absorbably polyglactin sutures in women the left side was closed with adhesive (butylcyanoacrylate) and undergoing mediolateral episiotomy after a vaginal delivery. the right with sutures. Sebesta 2004 compared octylcyanoacrylate with • Ozturan 2001) compared butylcyanoacrylate tissue subcuticular suture in people who had undergone laparoscopic • adhesive with 6.0 polypropylene sutures for columellar skin surgery. closure after the majority of the tension had been taken up using Shamiyeh 2001 compared octylcyanoacrylate tissue • 5.0 chromic catgut. adhesive with 5.0 monofilament sutures in patients requiring Sinha 2001 compared butylcyanoacrylate tissue adhesive mini-phlebectomy with the Muller technique. This study also • with 4.0 monofilament suture in adult patients requiring hand compared tissue adhesive with adhesive tape. or wrist surgery. Switzer 2003 compared octylcyanoacrylate tissue adhesive • van den Ende 2004 compared butylcyanoacrylate tissue with a subcuticular Monocryl suture for the elective repair of • adhesive with polyglactin 5-0 (Vicryl) suture in children inguinal hernias. Ong 2002 also compared octylcyanoacrylate undergoing inguinal hernia repair. with a Monocryl subcuticular suture, but in children requiring herniotomies. Sniezek 2007 followed a split wound design comparing Octylcyanoacrylate versus sutures • octylcyanoacrylate with a cuticular polypropylene suture on head Eighteen studies investigated the use of octylcyanoacrylate tis- and neck surgical sites following the removal of carcinomas using sue adhesives versus sutures (Avsar 2009; Brown 2009; Eggers the Mohs technique. 2011; Greene 1999; Jallali 2004; Khan 2006; Kouba 2011; Tierny 2009 compared octylcyanoacrylate tissue adhesive Krishnamoorthy 2009; Maartense 2002; Millan 2011; Mota • with rapid absorbing gut sutures in people undergoing surgery 2009; Ong 2002; Sebesta 2004; Shamiyeh 2001; Sniezek 2007; for non-melanomas skin cancer. Switzer 2003; Tierny 2009; Toriumi 1998). Toriumi 1998 investigated incisions with and without Avsar 2009 compared high viscosity octylcyanoacrylate • • subcutaneous sutures and then randomised for closure with tissue adhesive with polypropylene sutures in women undergoing octylcyanoacrylate tissue adhesive or 5.0 or 6.0 nylon suture in the Pfannenstiel incision in the abdomen. people over one year of age and over requiring elective surgery Brown 2009 compared octylcyanoacrylate tissue adhesive • for benign skin lesions predominantly in the face and neck. with Monocryl sutures in children undergoing inguinal herniorrhaphy. Eggers 2011 compared octylcyanoacrylate tissue adhesive • Comparison 2: Tissue adhesives compared with adhesive with monocryl sutures in people undergoing total knee tape arthroplasty. Greene 1999 and Kouba 2011 randomised participants • requiring bilateral blepharoplasty. This model with two identical skin sites on the same patient allowed a split-body study design Octylcyanoacrylate versus adhesive tape and each participant to act as his or her own control. The left or Three studies compared octylcyanoacrylate tissue adhesive with right upper eye lid was closed with octylcyanoacrylate and the adhesive tapes. other eyelid closed with 6.0 suture.

Tissue adhesives for closure of surgical incisions (Review) 9 Copyright © 2014 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Shamiyeh 2001 compared octylcyanoacrylate tissue Butylcyanoacrylate versus mixture of other closure approaches • adhesive with tape in adults requiring mini-phlebectomy with Chibbaro 2009 compared a group allocated to • the Muller technique. butylcyanoacrylate tissue adhesive with a group allocated to Maartense 2002 compared octylcyanoacrylate tissue receive either sutures or staples for skin closure - based on • adhesive to 76 mm x 6 mm adhesive paper (Steri-Strips) in clinician choice - in people undergoing elective cranial people undergoing elective laparoscopic surgery. supratentorial surgery. Romero 2011 compared octylcyanoacrylate tissue adhesive • with standard adhesive strips (strips applied in star-shaped manner) in children undergoing laparoscopic appendectomy. Comparison 5: Tissue adhesive compared with tissue adhesive

Comparison 3: Tissue adhesives compared with staples

High versus low viscosity adhesives Butylcyanoacrylate versus staples Blondeel 2004 compared high and low viscosity adhesive in • Two studies compared butylcyanoacrylate tissue adhesive with sta- the trial described above. ples. Livesey 2009 compared butylcyanoacrylate tissue adhesive • with staples in people undergoing a total hip replacement. Octylcyanoacrylate versus butylcyanoacrylate Eggers 2011 compared butylcyanoacrylate tissue adhesive • Three studies compared octylcyanoacrylate tissue adhesive with with staples in people undergoing total knee arthroplasty. butylcyanoacrylate tissue adhesive for skin closure. Eggers 2011 compared these tissue adhesives in people • Octylcyanoacrylate versus staples undergoing total knee arthroplasty. Kent 2014 compared these tissue adhesives in people Five studies compared octylcyanoacrylate tissue adhesives with sta- • ples. undergoing a range of laparoscopic procedures. Ridgway 2007 compared octylcyanoacrylate tissue adhesives Maloney 2013 compared these tissue adhesives in people • • with staples in people requiring thyroid and parathyroid surgery. undergoing a skin incision to remove skin cancer. Eggers 2011 compared octylcyanoacrylate tissue adhesives • with staples in participants undergoing total knee arthroplasty. Amin 2009 compared octylcyanoacrylate tissue adhesives Characteristics of outcome measures • with staples in people undergoing minimally invasive thyroidectomy. Khan 2006 compared octylcyanoacrylate tissue adhesives Wound dehiscence • with staples in people undergoing either a total knee arthroplasty Twenty-three studies reported wound dehiscence. The time of or a total hip arthroplasty. post-operative wound examination for dehiscence varied between Pronio 2011 compared octylcyanoacrylate tissue adhesives • studies. Cheng 1997 reported dehiscence at days 1, 2, 3, 7 and 30; with staples in people undergoing thyroid surgery. Toriumi 1998 reported dehiscence at 5 to 7 days; Chibbaro 2009, Ozturan 2001, Pronio 2011, Sniezek 2007 and Tierny 2009 at 7 Comparison 4: Tissue adhesives compared with other days; Shamiyeh 2001 and van den Ende 2004 at 10 days; Greene techniques 1999 at weeks 1, 2 and 4; and Sinha 2001 at 10 days, 2 weeks and 6 weeks. Ong 2002 assessed dehiscence at 2 to 3 weeks; Keng 1989 and Sebesta 2004 at 2 weeks; Dowson 2006 at day 1, 2 and Octylcyanoacrylate versus mixture of other closure at 4 to 6 weeks; Switzer 2003 at 2 and 4 weeks; Blondeel 2004 at approaches day 10; Brown 2009 at 6 weeks; Eggers 2011 at 24 hours, 3 weeks Blondeel 2004 was the only study that investigated the use and 6 weeks; Maloney 2013 at 2 weeks and 3 months; Mota 2009 • of tissue adhesives in incisions of 4 cm or longer. High viscosity from 42 hours to 68 hours; Romero 2011 at day 10 and day 90; octylcyanoacrylate tissue adhesive was compared with any other and Millan 2011 on days 5, 7, 10 and 14. commercially available device such as sutures, staples or tapes We felt there were potential unit of analysis issues in four studies and also compared with low viscosity octylcyanoacrylate. The where it was not clear whether one person had more than one high viscosity formulation is six-times more viscous than the episode of dehiscence over multiple time periods (Cheng 1997; normal adhesive and designed to reduce run-off of pre- Eggers 2011; Millan 2011; Sinha 2001). We acknowledged this in polymerised adhesive from the application site. the ’Risk of bias’ assessment and sensitivity analysis we conducted.

Tissue adhesives for closure of surgical incisions (Review) 10 Copyright © 2014 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Wound infection tionnaire. Surgical residents who were blinded to the wound clo- In total 25 studies measured wound infection. One study noted sure method also scored the cosmetic appearance with a VAS and infection, defined as wound discharge with positive bacterial cul- the Hollander Wound Evaluation Score (Hollander 1995). These ture, at days 1, 2, 3, 7 and 30 (Cheng 1997). A second study de- assessments were undertaken at two weeks and three months (with fined infection as the presence of pus (Keng 1989), and examined three month data presented). Ozturan 2001 reported cosmetic ap- wounds at 1 and 4 weeks postoperatively. A third study defined pearance at 3 months by blinded assessment of photographs using infection as an abscess that required drainage (Ozturan 2001), and VAS, with best possible scar rated as 100 and worst possible scar examined wounds at 1 week. A fourth study defined infection as as 0 (Quinn 1995), and the Hollander Scale. A third study also a spontaneous drainage of purulent fluid (Maartense 2002), and reported surgeon-assessed cosmetic appearance at three months noted this at 2 weeks and 3 months. Blondeel 2004 defined in- using a modified Hollander Wound Evaluation scale - this was fection as redness more than 3 mm to 5 mm from the wound a blinded assessment (Kent 2014). Livesey 2009, Maloney 2013 margin, swelling, purulent discharge, pain, increased skin tem- and Romero 2011 reported surgeon-assessed cosmetic appearance perature, fever or other signs of infection and assessed this at day at three months using a 100 mm VAS, where 0 represented the 10. Switzer 2003 described an infection as a draining sinus. Avsar worst outcome and 100 the best outcome - in all cases the sur- 2009 assessed infection at 2, 7 and 40 days postoperatively using geon-assessed outcome was blinded. Livesey 2009 and Maloney a definition of infection that was thought to be purulence and ac- 2013 also used the same VAS scale to assess participant data on companying erythema, but the translation we used was not com- cosmetic appearance, but it was unclear whether these assessments pletely clear about this definition. Livesey 2009 defined wound were blinded. infection as a participant requiring antibiotics specifically for sus- Pronio 2011 asked participants to assess cosmetic appearance us- pected wound infection, this was assessed at 24 hours, 3 weeks ing the Stony Brook scar evaluation scale composed of five di- and 6 weeks postoperatively. Khan 2006 classed follow-up as early, chotomous, evenly weighted categories. Mean data presented here which seemed to be the first 3 days after surgery, and then late, were calculated by the review authors from data included in the which was between 8 weeks and 12 weeks postoperatively; where study report. Kouba 2011 reported surgeon-assessed cosmetic ap- cultures were positive or there was clinical evidence of cellulitis, the pearance at one month and three months (we report three month participants were treated with a course of antibiotics and recorded data). The cosmetic presence of the wound was scored on a scale as having an ‘infection’. Romero 2011 assessed wounds at the 10th of 1 (excellent wound healing, scar matches surrounding skin) to and 90th days postoperatively and defined infection as an abscess 5 (poor scar wound healing, does not match surrounding skin); or redness more than 3 mm perpendicular to the wound. One this was a blinded assessment. study simply referred to wound complications (Dowson 2006). A A number of studies reported data that could not be used further. further 14 studies reported upon infection, but did not describe Toriumi 1998 assessed cosmetic appearance at three months and how infection would be defined for diagnosis and reporting in the one year using a modified Hollander scale at three months and study (Chibbaro 2009; Eggers 2011; Greene 1999; Kent 2014; a VAS of photographs at one year using two surgeons blinded to Ong 2002; Maloney 2013; Pronio 2011; Sebesta 2004; Shamiyeh treatment group. However, we could not use the data from this 2001; Sinha 2001; Sniezek 2007; Tierny 2009; Toriumi 1998; van study at three months as the standard deviation was not reported den Ende 2004). and one participant dropped out from an unspecified group. Data We felt there were potential unit of analysis issues in two stud- at one year could not be used, as the group(s) from which 11 ies where it was not clear whether one person had more than participants had dropped out was not clear. Ong 2002 evaluated one episode of infection over multiple time periods (Eggers 2011; cosmesis using a VAS and the Hollander Scale at three weeks and Khan 2006). While we have reported the total number of infec- three months, but the first data could not be used as it was too tions reported by group it was not clear whether some participants early, and the data at three months could not be used due to large reported more than one infection, as the number of infections was and unspecified loss to follow-up. Dowson 2006 assessed cosmetic reported rather than number of people having at least one infec- outcome using the Hollander scale at six weeks and three months. tion. We have acknowledged this in the ’Risk of bias’ assessment The six-week data were not included as it was too early and the and sensitivity analysis we conducted. three-month data did not report mean and standard deviation statistics; we contacted the authors but they did not respond, therefore the data could not be included. Sniezek 2007 also assessed Cosmetic appearance cosmesis at three months using a VAS, but there were insufficient data for inclusion. Amin 2009, Chibbaro 2009 and Tierny 2009 Eleven studies reported cosmetic appearance with data that could also reported cosmetic appearance at three months, but the data be used for further analyses. One study asked participants to score were not clear and were also not used. their own cosmetic result using a validated visual analogue scale Several studies evaluated cosmetic appearance less than three (VAS; Maartense 2002). Participants in the Dunker 1998 trial were months after surgery (Avsar 2009; Blondeel 2004; Brown 2009; also asked to complete the cosmetic scale of the Body Image Ques-

Tissue adhesives for closure of surgical incisions (Review) 11 Copyright © 2014 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Cheng 1997;Eggers 2011; Greene 1999; Keng 1989; Khan 2006; reported surgeon satisfaction immediately after closure, when the Krishnamoorthy 2009; Ong 2002; Ridgway 2007; Sebesta 2004; surgeon was asked to give his or her opinion on the time needed Sinha 2001; Switzer 2003). for wound closure and the practicality of the materials used by an- swering multiple choice questions (potential responses included: ’far too long’, ’a little too long’, ’not too long’, ’not practical’, Patient/parent satisfaction ’not very practical’ and ’very practical’; Maartense 2002). Due to Thirteen studies reported satisfaction outcomes. Patient satisfac- differences in methods used to measure this outcome, these stud- tion interviews in one study included questions about comfort, ies were not included in the meta-analysis. One study did report presence of a pulling sensation and appreciation of lack of suture surgeon satisfaction via a questionnaire that included ratings for removal at weeks one, two and four (Greene 1999). Wound com- wound cosmetic appearance, safety, effectiveness, applicability for fort at one and four weeks was assessed in another study (Keng a wide range of incisions, perceived patient satisfaction and overall 1989). A third study assessed participants at one year using a scale satisfaction (Blondeel 2004). Kent 2014 assessed surgeons’ satis- of satisfaction with scar (Shamiyeh 2001). One study assessed pa- faction with wound (considering expression, application, delivery tient satisfaction by using a questionnaire that included ratings for and ease of use for product); the options were ’satisfied’ or ’dis- cosmetic appearance, overall comfort, ability to shower, dressing satisfied’. Maloney 2013 assessed surgeon satisfaction (at time of changes, tension at the wound, hygienic problems or allergic reac- wound closure) using 100 mm VAS scales for: ease of use (0 ’im- tion and overall satisfaction (Blondeel 2004). Amin 2009 assessed possible’ to 10 ’very easy to use’) and satisfaction with device and patient satisfaction at three months using a self-completed 10 cm the closure achieved (0 ’completely dissatisfied’ to 10 ’completely VAS line where 0 was poor and 10 was excellent. Avsar 2009 as- satisfied’). sessed patient satisfaction at day 40 when participants were asked how satisfied they were with their skin closure and could choose from the following responses: very bad, poor, average, good, very Relative cost of materials good. Ong 2002 and Khan 2006 assessed participant/parent satis- Costs of closure devices were reported in five studies (Brown 2009; faction using a VAS scale that ran from 0 to 100, where 100 repre- Eggers 2011; Maartense 2002, Sebesta 2004; Shamiyeh 2001). sented maximal satisfaction. Data for Khan 2006 were presented as median values, and separately for different types of arthroplasty, and are not reported further in this review. Dowson 2006 and Time for wound closure following surgery Romero 2011 reported patient satisfaction as a simple ’satisfied’ Time for closure was reported in 24 studies (Avsar 2009; Brown or ’not satisfied’, but gave little information about the criteria 2009; Blondeel 2004; Cheng 1997; Chibbaro 2009; Dowson used. Likewise, Kent 2014 report participants’ satisfaction with 2006; Eggers 2011; Greene 1999; Jallali 2004; Keng 1989; incisional wound closure - options were ’satisfied’ or ’dissatisfied’. Kent 2014; Khan 2006; Krishnamoorthy 2009; Livesey 2009; Pronio 2011 assessed participants’ satisfaction with wound man- Maartense 2002; Maloney 2013; Ong 2002; Ozturan 2001; agement at a seven-day follow-up when participants were asked Ridgway 2007; Sebesta 2004; Shamiyeh 2001; Switzer 2003; to rate their level of satisfaction with the early postoperative man- Toriumi 1998; van den Ende 2004). The data from 15 of these agement of the wound (regarding the requirement of a return visit studies were excluded as they were insufficient (Avsar 2009; Cheng for medications, the possibility of washing oneself, and suture re- 1997; Chibbaro 2009; Dowson 2006; Eggers 2011; Greene 1999; moval) using a numerical scale ranging from 0 to 10. Data were Jallali 2004; Keng 1989; Khan 2006; Krishnamoorthy 2009; reported in merged categories (e.g. percentage of participants re- Livesey 2009; Maartense 2002; Switzer 2003; Toriumi 1998; van porting a score of 10 to 9, 8 to 7 etc.), which limited further use of den Ende 2004). Data from the remaining nine studies were el- these data. One paper reported upon participant satisfaction as an igible for inclusion (Blondeel 2004; Brown 2009; Kent 2014; outcome measure in the text (Sniezek 2007), but no results were Maloney 2013; Ong 2002; Ozturan 2001; Ridgway 2007; Sebesta reported in the analyses. We attempted to contact the authors, but 2004; Shamiyeh 2001). The time was recorded as a continuous received no reply. variable in seconds or minutes elapsed from beginning of wound closure to completion of wound closure. Mean time to closure times were only used in the review if this was measured on all Surgeon satisfaction participants, thereby avoiding concerns about time to event data In one study (Greene 1999), surgeon satisfaction included quality being analysed incorrectly. of wound closure by assessment of eversion, regularity, approximation and any difficulties in attaining approximation. This was reported at weeks one, two and four. Another study reported sur- Risk of bias in included studies geon satisfaction at 10 days using a scale of satisfaction with the scar on a five-point score (where 1 was a ’perfect cosmetic re- The results of the quality assessment for included trials is shown sult’ to 5 ’unsatisfactory result’; Shamiyeh 2001). A third study in Figure 1 and Figure 2.

Tissue adhesives for closure of surgical incisions (Review) 12 Copyright © 2014 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Figure 1. Risk of bias summary: review authors’ judgements about each risk of bias item for each included study

Tissue adhesives for closure of surgical incisions (Review) 13 Copyright © 2014 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Figure 2. Risk of bias graph: review authors’ judgements about each risk of bias item presented as percentages across all included studies

Random sequence generation were opened shortly before the intervention was given. Livesey Random sequence generation was performed adequately in twelve 2009 did not indicate that numbered envelopes were used, but studies (Amin 2009; Blondeel 2004; Eggers 2011; Khan 2006; stated that envelopes were opened by independent personnel. All Krishnamoorthy 2009; Maartense 2002; Mota 2009; Ong 2002; remaining studies were deemed to be at unclear risk of bias for Romero 2011; Shamiyeh 2001; Sniezek 2007; Switzer 2003). All allocation concealment. Some studies gave evidence that they used studies used a random sequences generated by computerised ran- envelopes that were sealed and opaque, but lacked sufficient ev- domisation programmes except for Sniezek 2007, which used coin idence that they were sequentially numbered or that the process tossing. was undertaken by an independent person. Many other studies In the remaining 20 studies, the risk of bias in the random sequence provided no information on allocation of the randomisation se- generation was unclear. This was largely because, although there quence at all. was mention of the participants being randomised into different intervention arms, there was no mention of the methods used to achieve this (Brown 2009; Cheng 1997; Chibbaro 2009; Dowson Blinding of participants and personnel 2006; Greene 1999; Jallali 2004; Keng 1989; Kent 2014; Kouba The risk of bias for blinding of participants and personnel was 2011; Livesey 2009; Maloney 2013; Millan 2011; Ozturan 2001; judged to be unclear in all studies. It was difficult to see how the Pronio 2011; Ridgway 2007; Sebesta 2004; Sinha 2001; Tierny operating surgeon could be completely blinded to an intervention 2009; Toriumi 1998). The partial translation of Avsar 2009 was he/she was supposed to undertake. Only Dowson 2006 and Mota such that no conclusion could be derived about whether there was 2009 suggest that participants remained blinded to the interven- adequate evidence of random sequence generation, therefore the tion they received. Given the difficulties faced in blinding partic- judgement remained unclear. ipants (in most cases) and personnel to the intervention, we felt it reasonable to judge all of the included studies as being at unclear risk of bias, as it is uncertain how this influenced the performance Allocation concealment of the personnel. Four studies were judged to be at low risk of bias for the domain of allocation concealment: Amin 2009, Dowson 2006, and Ong Blinding of outcome assessment 2002 reported that they used sealed, sequential envelopes, that

Tissue adhesives for closure of surgical incisions (Review) 14 Copyright © 2014 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Risk of bias for blinding of outcome assessment was undertaken whether dehiscence or other data, or both, were reported per per- for each relevant outcome, and thus was often characterised as son, or if the same wounds had more than one infection during the being at both unclear and high risk of bias within the same study. period of the trial (Cheng 1997; Eggers 2011; Jallali 2004; Kent This was often because numbers of wounds breaking down and 2014; van den Ende 2004). Avsar 2009 and Millan 2011 were wound infection were assessed in a way that was not obviously classed as being at unclear risk of bias for this domain, as we could blinded, but cosmesis was deemed to be blinded. Such examples of not make a judgement using the translations that we had. The studies where blinding of dehiscence was unclear and blinding of remaining 25 studies were deemed not to offer any other sources cosmesis was adequate were: Blondeel 2004, Dowson 2006, Kent of bias. 2014, Livesey 2009, Maloney 2013, Ong 2002, Ozturan 2001, Romero 2011, Shamiyeh 2001, and Toriumi 1998. Effects of interventions There were studies that did give enough evidence for us to in- fer a low risk of detection bias. These were Amin 2009, Brown See: Summary of findings for the main comparison Tissue 2009, and Maartense 2002, all of which revealed adequate blind- adhesive compared to sutures for surgical incisions; Summary ing of outcome assessment for all measures extracted for this re- of findings 2 Tissue adhesive compared to adhesive tape for view. In one study, there was evidence to suggest a high risk of bias surgical incisions; Summary of findings 3 Tissue adhesive (Mota 2009), as the study report indicated that those assessing the compared to staples for surgical incisions; Summary of findings 4 wounds were the authors of the paper. Tissue adhesive compared to other methods for surgical incisions; The remaining studies did not provide evidence of adequate blind- Summary of findings 5 High viscosity tissue adhesive compared ing of outcome assessments at any stage and were classed as being to low viscosity tissue adhesive for surgical incisions; Summary at unclear risk of bias for this domain. of findings 6 Octylcyanoacrylate compared to butylcyanoacrylate for surgical incisions

Incomplete outcome data Comparison 1. Tissue adhesives compared with Four studies were classed as having a high risk of attrition bias sutures (Dowson 2006; Kent 2014; Ong 2002; Sinha 2001). This was often due to high proportional rates of drop out, or unclear reasons for large numbers of drop outs, or both. Primary outcome Eleven studies were judged as being at unclear risk of bias for this domain, as reasons for drop out were not clear, or the extent to Data from 17 trials compared the use of tissue adhesive with su- which the drop-out rates affected the results were unclear (Amin tures for dehiscence, however as seven trials had no cases of dehis- 2009; Avsar 2009; Eggers 2011; Kouba 2011; Krishnamoorthy cence, only data from the remaining ten trials contributed to the 2009; Livesey 2009; Millan 2011; Mota 2009; Ridgway 2007; meta-analysis (Cheng 1997; Dowson 2006; Eggers 2011; Millan Romero 2011; Toriumi 1998). 2011; Mota 2009; Sebesta 2004; Shamiyeh 2001; Sinha 2001; The remaining 18 studies suffered fewer losses to follow-up and, Switzer 2003; van den Ende 2004). Overall a significant difference where these did occur they were fully explained in the literature, so was detected between the proportion of wounds with dehiscence these studies were judged to be at low risk of bias for this domain. (RR 3.35, 95% CI 1.53 to 7.33; Analysis 1.1), that favoured closure by suture with no evidence of heterogeneity (I2= 0). Taking an assumed control risk of 0.01 (10 per 1000) this returns a num- Selective outcome data ber needed to harm of 43. Only one study was deemed to be at Amin 2009, Kouba 2011, Livesey 2009 and Ozturan 2001 were low risk of bias for blinded outcome assessment for this outcome reported to be at high risk of bias due to possible selective reporting (Sinha 2001; Figure 1; Figure 2). Dowson 2006, Mota 2009 and of outcomes.Twenty-seven studies were deemed to have a low risk Sinha 2001 had one domain at high risk of bias. The remaining of reporting bias as they reported on all the outcomes that were studies had domains classed at low or unclear risk of bias. outlined in the methodology, We also conducted a sensitivity analysis in which we removed Two studies were judged as unclear, either due to an unclear En- studies that were at unclear risk of bias for this outcome due to glish translation (Avsar 2009), or because outcomes that were out- possible unit of analysis issues: these studies were Cheng 1997, lined as being assessed were not clearly reported (Sebesta 2004). Millan 2011, and Sinha 2001 (these authors were contacted where possible to request clarification regarding their data) and also van den Ende 2004, as we were not clear whether this trial reported Other sources of bias data at the participant level. When this analysis was undertaken a Eight studies were classed as being at unclear risk of bias for this similar effect size was found, however, there was reduced precision domain, including six studies where it was it was not possible to and the findings were no longer statistically significant (RR 2.70, exclude unit of analysis issues, for example, we could not establish 95% CI 0.95 to 7.68; P value 0.06; Analysis 1.2).

Tissue adhesives for closure of surgical incisions (Review) 15 Copyright © 2014 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Secondary outcomes package of poliglecaprone suture (used together with a dressing at EUR 0.42 each) and EUR 1.15 per package of adhesive paper tape (Maartense 2002). Eggers 2011 reported that the total cost Infection associated with surgery for each of the closure groups (including all Eighteen trials that compared the use of tissue adhesives with su- aspects of surgery associated with materials, labour, and operating tures reported wound infection data, however, as eight of these room expenses) was USD 993.20 for octylcyanoacrylate; USD had no cases of infection, only data from the remaining ten studies 878.27 for butylcyanoacrylate and USD 1056.32 for staples. No contributed to the meta-analysis (Avsar 2009; Cheng 1997; Eggers measures of variation were presented. Sebesta 2004 reported that 2011; Khan 2006; Maartense 2002; Ozturan 2001; Sebesta 2004; the mean tissue adhesive cost per patient was USD 65.10 (range Shamiyeh 2001; Switzer 2003; van den Ende 2004). There was USD 40.60 to USD 101.5; standard deviation (SD) USD 13.70) no evidence of a difference in the proportion of participants de- whilst the total cost for surgery in the octylcyanoacrylate group veloping infection in the individual trials, or when data from the (i.e. cost for operating room time, cost and cost of suture material) trials were pooled (RR 1.72, 95% CI 0.94 to 3.16; Analysis 1.3); was USD 193.32 (range USD 130 to USD 365; SD USD 49.40). again there was no evidence of heterogeneity (I2= 0). For the suture group the mean suture cost per patient was USD Three studies were considered to have the potential for unit of 7.74 (range USD 3.60 to USD 10.80; SD USD 2.05) and the analysis issues for this outcome (Eggers 2011; Khan 2006; van den total mean cost was USD 497 (range USD 295 to USD 835; SD Ende 2004). We conducted a sensitivity analysis removing these USD 139.70). studies from the meta-analysis. Again there was no evidence of a difference in the proportion of participants developing infection Time for wound closure following surgery between groups (RR 2.03, 95% CI 0.80 to 5.12; Analysis 1.4). Five studies reported the time taken for closure (Brown 2009; Ong 2002; Ozturan 2001; Sebesta 2004; Shamiyeh2001). Pooling 2 Cosmetic appearance these trials was not appropriate due to massive heterogeneity (I = 100%; Analysis 1.10). Two of the trials suggested that adhesive There was no evidence of a difference in the participants’ assess- took longer to apply than sutures, and in one of these trials the ment of cosmetic appearance (MD -2.12, 95% CI -7.20 to 2.95; difference was statistically significant (Shamiyeh 2001), although Analysis 1.5). Likewise there was no evidence of a difference for the the units for this were unclear - if measured in seconds then the surgeons’ assessment of cosmetic appearance based on a 0 to 100 difference, although significant, could be small in terms of total VAS (MD 3.00, 95% CI -3.30 to 9.30), or a scar assessment scale time saved with one method compared to the other. In remaining (MD -0.26, 95% CI -0.58 to 0.06; Analysis 1.6; Kouba 2011). three trials, suturing took significantly more time than adhesive ( Brown 2009; Ozturan 2001; Sebesta 2004). Sebesta 2004 reported that application of adhesive took on average 3.7 minutes compared Patient/surgeon satisfaction with 14 minutes on average in the suture group. Ong 2002 and There was no evidence of a difference in patient, surgeon or pa- Ozturan 2001 were both deemed to be at high risk of bias for one tient/parent satisfaction with treatment (Analysis 1.7; Analysis domain (Figure 1; Figure 2). 1.8; Analysis 1.9; Dowson 2006; Maartense 2002; Ong 2002; Shamiyeh 2001). In one further study 13/20 participants stated that they preferred the tissue adhesive while the remaining seven Summary: tissue adhesives compared with sutures preferred the sutures (Greene 1999). Avsar 2009 reported that 6/ There was an overall difference favouring sutures over tissue ad- 20 participants in the adhesive group reported average satisfac- hesives in that sutures led to a reduced risk of dehiscence, though tion; 5/20 good and 9/20 very good. This was compared to 13/ several studies that contributed data to this analysis had at least one 20 recording good and 7/20 very good in the suture arm. risk of bias domain that was classed as being at high risk. Removal of studies with a possible unit of analysis issue reduced the precision and the results then lay just outside the standard definition of Costs statistical significance.There was no evidence of a difference in the Four studies reported costs. Shamiyeh 2001 reported that one tube risk of developing a wound infection in trial groups - although the of tissue adhesive at USD 11.00 could be used to close 3.5 incisions comparison is underpowered and confidence intervals are wide. of 5 mm mean length, while one suture at USD 1.10 could be used One study was at both low and unclear risk of bias across domains: to close five incisions, and one package of tape (six pieces) at USD its results were statistically significant and suggested that using 0.24 could be used to close three incisions - this calculation assumes sutures for closure was slightly faster than using tissue adhesives. multiple incisions on the same patient. A second study reported However, a second study had opposite findings, which were also that for closure of laparoscopic trocar wounds the costs were EUR statistically significant, suggesting that closure with tissue adhesive 13.90 for one ampoule of octylcyanoacrylate, EUR 2.47 for one was significantly faster than sutures by over 10 minutes on average.

Tissue adhesives for closure of surgical incisions (Review) 16 Copyright © 2014 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Comparison 2. Tissue adhesives compared with Summary: tissue adhesive compared with adhesive tape adhesive tape There were limited data for wound dehiscence and wound infec- Three trials provided data comparing tissue adhesives with adhe- tion for the comparison of tissue adhesive against adhesive tape. sive tape (Maartense 2002; Romero 2011; Shamiyeh 2001). These Only one trial reported dehiscence as an outcome, and this was trials were rated as being at low or unclear risk of bias for domains underpowered. Three trials reported wound infection as an out- (Figure 1; Figure 2). come, and these were also underpowered. Based on these small studies there is currently no evidence of a difference in the incidence of dehiscence or wound infection when wounds are closed Primary outcome with tissue adhesive or tape. Surgeons’ assessment (blinded) of the cosmetic outcome was better in the tape group. One study re- There was no signiﬁcant difference between closure techniques in ported a signiﬁcant difference in time taken for closure and this the single trial that contributed data to this comparison for dehis- favoured adhesive tape (time units not clear). cence (RR 0.96, 95% CI 0.06 to 14.55; Analysis 2.1; Shamiyeh 2001): this comparison was underpowered, as only one participant experienced dehiscence in each group. Comparison 3. Tissue adhesives compared with staples Six studies compared a tissue adhesive with staples for wound Secondary outcomes closure (Amin 2009; Eggers 2011; Khan 2006; Livesey 2009; Pronio 2011; Ridgway 2007).

Infection Primary outcome All three trials reported wound infection, but there was no evidence of a difference between the groups in the proportion of participants Two trials that compared the use of tissue adhesives with staples with infection (RR 1.37 95% CI 0.39 to 4.81; Analysis 2.2). Again presented data for dehiscence (Eggers 2011; Pronio 2011). As there this comparison was underpowered, with only 10 infection events were no cases of dehiscence in one trial, only one trial contributed in total. data to the comparison (Eggers 2011). There was no statistically signiﬁcant difference in the proportion of wounds that dehisced in the tissue adhesive group compared to the staples group (RR Cosmetic appearance 0.53, 95% CI 0.05 to 5.33; Analysis 3.1). Maartense 2002 and Romero 2011 found no statistically significant difference for participants’ assessment of cosmetic appear- Secondary outcomes ance (Analysis 2.3; Analysis 2.4), however both studies reported evidence of a difference for the surgeons’ blinded assessment using a VAS favouring closure by adhesive tape (pooled estimate MD Infection 9.56, 95% CI: 4.74 to 14.37; I2 = 14%; Analysis 2.5). Four trials that compared the use of tissue adhesive with staples presented data on wound infection (Eggers 2011; Khan 2006; Patient/surgeon satisfaction Livesey 2009; Pronio 2011), however as there were no cases of wound infection in one trial, data from three trials contribute Shamiyeh 2001 found no difference between the groups for pa- data to the comparison (Eggers 2011; Khan 2006; Livesey 2009). tient satisfaction (Analysis 2.6). Similarly, Maartense 2002 and There was no evidence of a difference in the proportion of wounds Shamiyeh 2001 found no evidence of a difference between the that became infected in the tissue adhesive group compared to the groups with respect to surgeons’ satisfaction (Analysis 2.7). staples group (RR 1.39, 95% CI 0.30 to 6.54; Analysis 3.2).

Time to wound closure following surgery Cosmetic appearance Shamiyeh 2001 presented data for time to closure and demon- Pronio 2011 found no evidence of a difference between the groups strated that tapes were signiﬁcantly faster to use than adhesives for patient-assessed cosmetic outcome (Analysis 3.3). Similarly (MD 0.56, 95% CI 0.42 to 0.70; Analysis 2.8). (Review authors’ Livesey 2009 found there was no evidence for a difference between note: units not reported in trial, we assumed that units are min- the groups with respect to the cosmetic appearance by surgeons utes). (Analysis 3.4).

Tissue adhesives for closure of surgical incisions (Review) 17 Copyright © 2014 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Patient/surgeon satisfaction Secondary outcomes Amin 2009 found a statistically significant difference between the groups for patient satisfaction in favour of tissue adhesive (MD 1.10; 95% CI 0.41 to 1.79; Analysis 3.5). Infection Blondeel 2004 found no significant difference between the groups Costs for infection (RR 0.41, 95% CI 0.11 to 1.60; Analysis 4.2). Eggers 2011 reported that the total cost associated with surgery Chibbaro 2009 reported wound infection, but there were no events for each of the closure groups (including all aspects of surgery in either group. associated with materials, labour, and operating room expenses) was USD 993.20 for octylcyanoacrylate; USD 878.27 for butylcyanoacrylate and USD 802.79 for staples. Patient/surgeon satisfaction Blondeel 2004 reported statistically significant differences in favour of the group treated with other devices such as sutures, Time to wound closure following surgery staples and tapes for both patient satisfaction (MD 0.40, 95% CI Ridgway 2007 reported the time to closure and found a statisti- 0.10 to 0.70; Analysis 4.3) and clinician satisfaction (MD 0.53, cally significant difference favouring the use of staples (MD 67.0 95% CI 0.29 to 0.77; Analysis 4.4). seconds, 95% CI 39.3 to 94.7; Analysis 3.6). The study had risk of bias domains classified as both low and unclear. Time for closure Summary: tissue adhesive compared with staples Blondeel 2004 reported that closure took significantly less time There was no evidence of a difference in wound dehiscence, wound with tissue adhesive (MD -1.05 minutes, 95% CI -1.79 to -0.31; infection, or cosmetic appearance between use of tissue adhesives Analysis 4.5). Data reported for this outcome by Chibbaro 2009 or staples for wound closure. There was a significant difference in were unclear. patient satisfaction in favour of tissue adhesives. There was also a significant difference in time taken for closure that favoured staples, however, the difference between groups was only just over Summary: tissue adhesive compared with other techniques 60 seconds, so whilst the result was statistically significant, it was When tissue adhesives were compared with other techniques there not significant from a clinical point of view. was no evidence of a difference in wound dehiscence or wound infection between groups. When assessing surgeon and patient Comparison 4. Tissue adhesives compared with other satisfaction there was a statistical difference favouring the control techniques group (other techniques) over tissue adhesives. In this same analysis there was a statistically significant difference favouring the tissue Blondeel 2004 compared high viscosity adhesive with any other adhesive for time taken to closure - the mean time saving was one commercially available devices such as sutures, staples, tapes or minute. low viscosity adhesive. We requested further information from the authors, specifically data regarding high and low viscosity adhesives compared with the remaining commercially available devices Comparison 5. Adhesives compared with adhesives and data for the comparison of high viscosity with low viscosity adhesives for the outcomes of interest to this review. This comparison of data was not reported within the published paper, but the author made the data available to the review team. This paper was Comparison 5.1. High versus low viscosity adhesives designated as being at low risk of bias. Chibbaro 2009 compared The data for this comparison were provided by the authors as an tissue adhesive with a comparison group that received sutures or additional analysis to that found in the paper. staples.

Primary outcome Primary outcome Blondeel 2004 found no signiﬁcant difference between the groups Blondeel 2004 found no signiﬁcant difference between the high for dehiscence (RR 0.55, 95% CI 0.13 to 2.38; Analysis 4.1). viscosity and low viscosity adhesive intervention groups for dehis- Chibbaro 2009 reported dehiscence, but there were no events in cence (RR 3.74, 95% CI 0.21 to 67.93; Analysis 5.1). This paper either group. was designated as being at unclear risk of bias.

Tissue adhesives for closure of surgical incisions (Review) 18 Copyright © 2014 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Secondary outcomes 95% CI 0.21 to 1.88; Analysis 6.2). Maloney 2013 reported no events for either trial arm for infection.

Infection No statistically significant differences were found between inter- Cosmetic appearance vention groups for infection (Analysis 5.2). Maloney 2013 reported no evidence of a difference in participant- assessed cosmetic outcome between study groups. Likewise there was no evidence of a difference in blinded surgeon-assessed cos- Patient/surgeon satisfaction metic appearance (Analysis 6.3; Analysis 6.4). No statistically significant differences were found between intervention groups for patient or surgeon satisfaction (Analysis 5.3; Analysis 5.4). Patient/surgeon satisfaction Surgeons’ satisfaction with skin incision closure technique, in Time for closure terms of ease of use and satisfaction with closure, was assessed us- For time to closure, there was a significant difference favouring ing a VAS. There was no evidence of a difference between groups low viscosity adhesive (MD 0.54 minutes, 95% CI 0.03 to 1.05; for these measures (Analysis 6.5; Analysis 6.6.). Analysis 5.5).

Comparison 5.2. Octylcyanoacrylate vs butylcyanoacrylate Costs Three trials compared octylcyanoacrylate tissue adhesive with butylcyanoacrylate tissue adhesive (Eggers 2011; Kent 2014; Eggers 2011 reported that the total cost associated with surgery Maloney 2013). We assessed Maloney 2013 as being at low or for each of the closure groups (including all aspects of surgery as- unclear risk of bias for domains. Eggers 2011 was assessed as being sociated with materials, labour, and operating room expenses) was at unclear risk of bias for incomplete outcome data and Kent 2014 USD 993.20 for octylcyanoacrylate and USD 878.27 for butyl- for randomising at the participant level, but presenting unadjusted cyanoacrylate. data at the wound level. We were unable to re-analyse the Kent 2014 data correctly here, as we had planned to report data at the participant level only, but did not do this because of the amount of missing data. Time for wound closure following surgery Maloney 2013 reported a significant difference in time for wound Primary outcome closure favouring octylcyanoacrylate (MD -25.50, 95% CI -39.79 Eggers 2011 reported no statistically significant difference in to -11.21; Analysis 6.7). It is important to note that the units wound dehiscence between groups (RR: 0.95, 95% CI 0.06 to for this assessment were seconds - so whilst there is a statistically 14.04; Analysis 6.1). Likewise Maloney 2013 did not report a significant difference it is one of, on average, about 26 seconds statistically significant difference in wound dehiscence between between treatment arms, which probably does not translate to a groups (RR: 2.63, 95% CI 0.11 to 61.05; Analysis 6.1). These significant clinical difference. data were pooled and returned a result that showed there could be an effect in either direction, or none (RR 1.46, 95% CI 0.19 to 11.30; Analysis 6.1; very low precision due to the small sample Summary: tissue adhesive compared with other tissue size). adhesive There was no evidence of a difference in any outcome assessed Secondary outcomes when different types of tissue adhesives were compared, except for high viscosity compared with low viscosity adhesives where there was a statistical difference favouring the low viscosity adhesive for time taken for closure (mean difference approximately 1 minute). Infection Likewise, one study reported a significant difference in time taken Eggers 2011 report no statistically significant different between for wound closure favouring octylcyanoacrylate but the size of the different tissue adhesives in terms of wound infection (RR 0.63, difference was small.

Tissue adhesives for closure of surgical incisions (Review) 19 Copyright © 2014 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Overall summary of results by outcome Time to wound closure (following surgery)

One study suggested that, on average, it took 10 minutes longer Dehiscence to close a wound with sutures than with tissue adhesives, but this was not borne out in all the studies and may depend on factors When tissue adhesives were compared with sutures, pooled data such as wound type. There was a statistically significant difference from 17 studies (with 10 contributing data towards the analysis) in time taken for closure between tissue adhesives and tape-closed showed a statistically significant difference favouring sutures, as wounds, that favoured the tape group. There was also a signifi- they reduced the risk of dehiscence. It is possible this result may cant difference in time taken for closure between tissue adhesives have been influenced by unit of analysis issues, as it was not clear compared with staples that favoured the staples group, however, whether more than one dehiscence event per participant had been the difference between groups was only just over 60 seconds, so recorded in studies with multiple follow-up points. No other com- whilst it was statistically significant it was not clinically significant. parison of tissue adhesives with other closure methods showed a Another study that compared tissue adhesive with a mixed control statistically significant difference in the risk of wound dehiscence. group (i.e. staples, sutures and tape) found a statistically significant difference in time to closure that favoured tissue adhesive, Wound infection although the mean difference between groups was again approximately one minute. Finally, one study reported that there was a There was no evidence of a difference in the risk of wound infection small, but statistically significant difference in time to closure that between wounds closed with tissue adhesives and wounds closed favoured low viscosity adhesive over high viscosity adhesive, and using other methods reported here. favoured octylcyanoacrylate tissue adhesive over butylcyanoacrylate tissue adhesive. Cosmetic appearance Two studies with blinded outcome assessment reported a statistically significant difference in surgeon-assessed appearance score that favoured the tape group over tissue adhesives. ’Summary of findings’ table

We present ’Summary of findings’ tables for the dehiscence and Patient/surgeon satisfaction infection outcomes. These tables confirm our conclusion that the When tissue adhesives were compared with a mixed control group available evidence is generally underpowered and is limited in some (i.e. sutures, staples and tape), one study reported a statistically cases by a lack of reported methodological information that would significant difference in patient and surgeon satisfaction scores permit full analysis of the data. (Summary of findings for the that favoured the mixed control group, while another study that main comparison; Summary of findings 2; Summary of findings compared tissue adhesive with staples found a significant difference 3; Summary of findings 4; Summary of findings 5; Summary of in patient satisfaction that favoured the tissue adhesive. findings 6)

Tissue adhesives for closure of surgical incisions (Review) 20 Copyright © 2014 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. oyih 04TeCcrn olbrto.Pbihdb J by Published Collaboration. Cochrane The (Review 2014 incisions © surgical Copyright of closure for adhesives Tissue ADDITIONALSUMMARYOFFINDINGS [Explanation]

Tissue adhesive compared to adhesive tape for surgical incisions

Patient or population: people with surgical incisions Settings: Intervention: tissue adhesive Comparison: adhesive tape

Outcomes Illustrative comparative risks*3 (95% CI) Relative effect No of participants Quality of the evidence Comments (95% CI) (studies) (GRADE)

Assumed risk Corresponding risk ) Adhesive tape Tissue adhesive h ie os Ltd. Sons, & Wiley ohn Wound dehiscence Study population RR 0.96 50 ⊕⊕ (0.06 to 14.55) (1 study) low1 42 per 1000 40 per 1000 (2 to 606)

Moderate

Wound infection Study population RR 1.37 190 ⊕⊕ (0.39 to 4.81) (3 studies) low1,2 43 per 1000 60 per 1000 (17 to 209)

Moderate

*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI) CI: Confidence interval; RR: Risk ratio 21 oyih 04TeCcrn olbrto.Pbihdb J by Published Collaboration. Cochrane The (Review 2014 incisions © surgical Copyright of closure for adhesives Tissue

1 Study 95% CIs are very wide 2 Evidence of inconsistency in point estimates. With the point estimate from one study lying outside the 95% CIs of another 3 Control (tape) group risk in included study xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx ) h ie os Ltd. Sons, & Wiley ohn 22 oyih 04TeCcrn olbrto.Pbihdb J by Published Collaboration. Cochrane The (Review 2014 incisions © surgical Copyright of closure for adhesives Tissue

Tissue adhesive compared to staples for surgical incisions

Patient or population: people with surgical incisions Settings: Intervention: tissue adhesive Comparison: staples

Outcomes Illustrative comparative risks*3 (95% CI) Relative effect No of participants Quality of the evidence Comments (95% CI) (studies) (GRADE)

Assumed risk Corresponding risk

Staples Tissue adhesive )

h ie os Ltd. Sons, & Wiley ohn Wound dehiscence Study population RR 0.53 37 ⊕⊕ (0.05 to 5.33) (1 study) low1 105 per 1000 56 per 1000 (5 to 561)

Moderate

Wound infection Study population RR 1.39 250 ⊕ (0.3 to 6.54) (3 studies) very low1,2 71 per 1000 99 per 1000 (21 to 463)

Moderate

*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: Confidence interval; RR: Risk ratio; 23 oyih 04TeCcrn olbrto.Pbihdb J by Published Collaboration. Cochrane The (Review 2014 incisions © surgical Copyright of closure for adhesives Tissue

GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate.

1 Study 95% CIs are very wide. 2 Evidence of point estimates lying in opposite directions with the estimate for one study lying outside the 95% CI of another. 3 Control (staples ) group risk for included study. xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx ) h ie os Ltd. Sons, & Wiley ohn 24 oyih 04TeCcrn olbrto.Pbihdb J by Published Collaboration. Cochrane The (Review 2014 incisions © surgical Copyright of closure for adhesives Tissue

Tissue adhesive compared to other methods for surgical incisions

Patient or population: people with surgical incisions Settings: Intervention: tissue adhesive Comparison: other methods

Outcomes Illustrative comparative risks* (95% CI) Relative effect No of participants Quality of the evidence Comments (95% CI) (studies) (GRADE)

Assumed risk Corresponding risk

Other methods Tissue adhesive )

h ie os Ltd. Sons, & Wiley ohn Wound dehiscence Study population RR 0.55 209 ⊕⊕ (0.13 to 2.38) (1 study) low1,2 49 per 1000 27 per 1000 (6 to 117)

Moderate

Wound infection Study population RR 0.41 209 ⊕⊕ (0.11 to 1.6) (1 study) low1,2 66 per 1000 27 per 1000 (7 to 105)

Moderate

*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI) CI: Confidence interval; RR: Risk ratio 25 oyih 04TeCcrn olbrto.Pbihdb J by Published Collaboration. Cochrane The (Review 2014 incisions © surgical Copyright of closure for adhesives Tissue

1 Study 95% CIs are very wide 2 Single study with low event rate xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx ) h ie os Ltd. Sons, & Wiley ohn 26 oyih 04TeCcrn olbrto.Pbihdb J by Published Collaboration. Cochrane The (Review 2014 incisions © surgical Copyright of closure for adhesives Tissue

High viscosity tissue adhesive compared to low viscosity tissue adhesive for surgical incisions

Patient or population: people with surgical incisions Settings: Intervention: high viscosity tissue adhesive Comparison: low viscosity tissue adhesive

Outcomes Illustrative comparative risks* (95% CI) Relative effect No of participants Quality of the evidence Comments (95% CI) (studies) (GRADE)

Assumed risk Corresponding risk

Low viscosity tissue ad- High viscosity tissue ad- ) hesive hesive h ie os Ltd. Sons, & Wiley ohn Wound dehiscence Study population RR 3.74 148 ⊕ (0.21 to 67.93) (1 study) very low1,2 Could not be calculated Could not be calculated

Wound infection Study population RR 0.82 148 ⊕ (0.16 to 4.31) (1 study) very low1,2 47 per 1000 38 per 1000 (7 to 200)

Moderate

*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI) CI: Confidence interval; RR: Risk ratio 27 oyih 04TeCcrn olbrto.Pbihdb J by Published Collaboration. Cochrane The (Review 2014 incisions © surgical Copyright of closure for adhesives Tissue

1 Study 95% CIs are very wide 2 Single study with low event rate xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx ) h ie os Ltd. Sons, & Wiley ohn 28 oyih 04TeCcrn olbrto.Pbihdb J by Published Collaboration. Cochrane The (Review 2014 incisions © surgical Copyright of closure for adhesives Tissue

Octylcyanoacrylate compared to butylcyanoacrylate for surgical incisions

Patient or population: people with surgical incisions Settings: Intervention: octylcyanoacrylate Comparison: butylcyanoacrylate

Outcomes Illustrative comparative risks*2 (95% CI) Relative effect No of participants Quality of the evidence Comments (95% CI) (studies) (GRADE)

Assumed risk Corresponding risk

Butylcyanoacrylate Octylcyanoacrylate )

h ie os Ltd. Sons, & Wiley ohn Wound dehiscence 26 per 1000 38 per 1000 RR 1.46 80 ⊕⊕ (5 to 297) (0.19 to 11) (2 studies) low1

Wound infection Study population RR 0.63 37 ⊕⊕ (0.21 to 1.88) (1 study) low1 333 per 1000 210 per 1000 (70 to 627)

Moderate

1 The 95% CI estimate around the RR of 1.46 is very wide 29 oyih 04TeCcrn olbrto.Pbihdb J by Published Collaboration. Cochrane The (Review 2014 incisions © surgical Copyright of closure for adhesives Tissue 2 Median control group risk across studies xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx ) h ie os Ltd. Sons, & Wiley ohn 30 DISCUSSION metic appearance at one year (Quinn 1998), the same investigation reported that early outcome, such as at five or 10 days, is poorly predictive of the three month appearance. This can present Summary of main results a problem for trials, as attrition bias may influence the results. The only difference identified in cosmetic appearance was for the com- Traditional methods of surgical incision closure have been used parison of tissue adhesive with adhesive tape - this was a blinded for many years. However, these techniques are not without some assessment by surgeons that favoured tape. Assessments for this problems and it is therefore important to consider new develop- comparison were limited, so there was a great deal of uncertainty ments - such as tissue adhesives - which may offer some advan- regarding the comparative risk of wound dehiscence and infection tages for patients. This review included a total of 33 studies: 23 when tape was compared with tissue adhesive. studies compared tissue adhesives with sutures; three studies com- Patient satisfaction is also important when comparing alternative pared tissue adhesives with adhesive tape, which is often used for incision closure devices providing that the primary efficacy vari- smaller incisions; six studies compared adhesives with staples; two ables of dehiscence, infection and cosmetic appearance are satisfac- studies compared tissue adhesives with mixed control groups; and tory. Patient satisfaction may include ratings for cosmesis, overall two studies compared different types of tissue adhesives. A no- comfort, ability to shower, dressing changes, tension on wounds, table finding in this review was that, when data from ten studies hygiene problems, allergic reactions and overall satisfaction. An- that compared tissue adhesives with sutures were pooled, a sig- other important benefit of tissue adhesives is that they do not need nificant difference in risk of wound dehiscence became apparent, to be removed. One small study suggested that patients had a pref- with a higher risk of dehiscence associated with tissue adhesives. erence for adhesive rather than sutures (Greene 1999), although it No other comparison showed evidence for a difference in risk of was not clear what dimensions it measured. A further small study wound dehiscence for alternative treatments. It is important to also suggested that patients were had higher satisfaction with tis- note the often unclear and sometime high risk of bias of included sue adhesive than staples. Conversly, one study that compared tra- studies. Many studies were also underpowered, and the primary ditional closure methods to tissue adhesives demonstrated a sig- outcome here - dehiscence - was not used as a basis for sample size nificant patient preference for the alternatives to tissue adhesives calculation. The surgical procedures described by the studies were (Blondeel 2004): this was the largest of the studies to assess patient diverse and included blepharoplasty, circumcision, and excision of satisfaction. It would be helpful if validated questionnaires were benign skin lesions. Incision in areas of high tension and in those available to use in studies. patients whose medical health may have reduced their ability to Surgeon satisfaction assessments typically sought ratings for heal were excluded and so tissue adhesives have not been evaluated wound cosmesis or ease of use for the technique. Certainly there is in this population. no risk of needle-stick injury to the surgeon whilst using adhesive Postoperative site infection is a common problem and has a sig- rather than sutures, and this may be a factor in favour of adhesives, nificant impact on patients and healthcare systems. More than though none of the studies in this review considered this. In one 70% of surgical procedures are now performed on a outpatient study physicians completed a questionnaire that measured satis- basis, which poses major problems for surveillance of surgical site faction with the ease of use of a device, wound cosmesis, safety, infections (Emori 1993). In one study of 2345 patients under- effectiveness, applicability to a wide range of incisions, perceived going surgery, the overall incidence of surgical site infection was patient satisfaction and overall satisfaction (Blondeel 2004). This 8.5% (Malone 2002). One might anticipate that a tissue adhe- study showed that clinicians significantly preferred the alternative sive offers a barrier to micro-organisms at the site of the healing to tissue adhesive, but when comparing high viscosity adhesives incision and may therefore contribute towards reducing wound to other adhesives, operators showed no difference in preference. infection. However, the studies identified for this review did not We included time to closure as an outcome measure in the re- demonstrate any significant difference in the proportion of infec- view post hoc, as the review authors believe this to be a contributions in incisions closed with tissue adhesives compared with other tory factor towards both cost-effectiveness and satisfaction. Stud- conventional techniques. No study reported an a priori calcula- ies suggested that tissue adhesives might be significantly quicker tion for the sample size, and this may be relevant. Even the largest than suturing for some wounds, although there was heterogeneity of the studies would have been unlikely to have been adequately around this that might be related to the type and size of wound powered to show any significant difference given the relatively low being assessed, and suturing was reported to be the faster method incidence of wound infections following many types of surgery. in other studies. In other comparisons closure with tape and sta- Cosmetic outcome is an important long-term outcome of wound ples was found to be faster than with tissue adhesives - but the repair for the patient and we considered appearance at or beyond differences in time, whilst statistically significant, were generally three months after surgery to be meaningful. Some studies re- only small periods of difference - for example 60 seconds. ported cosmetic outcome earlier than this three month time point Whilst time taken is an attractive aspect for surgeons, it should be (Blondeel 2004). Whilst is has been reported that an individual noted that although there maybe a significant difference in time patient’s cosmetic outcome at three months is predictive of cos-

Tissue adhesives for closure of surgical incisions (Review) 31 Copyright © 2014 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. taken for closure, the overall time taken was still relatively small Implications for practice in both groups with even the longest closure times rarely exceed- Thirty-three studies, including 2793 participants, compared tis- ing two minutes. In clinical practice the reduction of wound clo- sue adhesives with alternative methods of surgical wound closure. sure time through using the quickest method may be insufficient These provided no evidence of a difference in rates of wound in- to increase the number of operations in a given operating room fection after surgical incision closure with tissues adhesives, su- schedule. However, a faster wound closure technique may increase tures, tapes, staples or amongst different tissue adhesives. When surgeon satisfaction and may also be attractive to patients when adhesive was compared with suture there was evidence of a benefit surgery is undertaken using only local anaesthetic. Alternatively, for using sutures for minimising dehiscence. patients may be happy with a lengthier closure time for a technique that achieves better outcomes. Although there was some evidence that surgeons and patients may Generally, tissue adhesives cost more than the alternatives. An prefer alternative methods to tissue adhesives, there was also ev- ampoule of tissue adhesive may be three times or more the cost of idence that patient satisfaction with tissue adhesives was higher the suture required to close the same length of incision. However, than for staples. Although some analyses demonstrated that alter- a dressing is not required over the tissue adhesive, but is routinely natives to tissue adhesives were less time consuming to use, this used with alternatives. Also, when taking into account the overall does not contraindicate their use, and this conclusion was refuted cost of the surgical procedure, this small difference may be of by the results of other studies. However, these trials did not con- less significance. It may be considered important to use a specific sider incisions in areas of high tension, such as the elbow and knee, closure device, even if the cost is greater, if significant superiority is which therefore have not been evaluated with tissue adhesives. demonstrated for wound healing, including cosmetic outcome and Similarly, trials excluded patients whose general health may have patient satisfaction. Thus the length of time for closure and cost had the potential to impair wound healing, so tissue adhesives may be important only after other outcomes such as dehiscence, have not been evaluated in these populations. cosmesis and infection have been considered. In summary, there is some evidence that dehiscence rates may be Implications for research higher in wounds closed with tissue adhesives than with sutures. This finding agrees with a review published in 2010 that includes Further adequately powered trials should be undertaken to investi- similar wound dehiscence data to this review (Chow 2010), al- gate the effects of tissue adhesives in clinical situations commonly though the authors did not perform a sensitivity analysis based on excluded in trials, specifically, wound closure in areas of high ten- possible unit of analysis issues, as we have here. sion and in patients with general health that could potentially im- There was no evidence of any difference between sutures and tis- pair wound healing. Trials should be sufficiently powered to de- sue adhesive for outcomes such as cosmetic appearance and satis- tect any differences in the rate of wound infection or dehiscence faction. The relative time taken using sutures for wound closure as statistically significant when comparing different incision clo- rather than tissue adhesives seems to vary, with one study showing sure devices. As any future trial would need to be large, it would use of adhesive to be significantly faster than sutures and another probably be multicentred. It would require methodological rigour study showing suturing to be significantly faster than tissue ad- to address the outstanding uncertainties in this area appropriately. hesive. Other studies reported that tissue adhesive took longer to Development and validation of patient satisfaction and surgeon apply than staples and tape. Generally, there was limited compar- satisfaction questionnaires would be helpful. ative evidence to determine whether tissue adhesives led to lower or higher levels of dehiscence, infection, satisfaction with cosmetic outcome, or patient or surgeon general satisfaction. Surgeons and patients were significantly more satisfied with the alternatives to adhesives in one analysis, but this finding was not replicated in ACKNOWLEDGEMENTS other studies. There was no evidence of a difference in outcomes between tissue adhesive types except that high viscosity adhesives We wish to thank Sally Bell-Syer and Ruth Foxlee (Cochrane took significantly more time for skin closure, adding a minute on Wounds Group) for their assistance with literature searching the average to the procedure. preparation of this review. We would also like to thank the following referees: Nicky Cullum, David Margolis, Michelle Briggs, Seokyung Hahn, James V Quinn, Ken Farion and Jac Dinnes. We also thank Matthew Fortnam who extracted data and checked AUTHORS’ CONCLUSIONS quality of data extraction for this second review update.

References to studies included in this review incisions. Journal of Laparoendoscopic and Advanced Surgical Techniques 2004 Aug;14(4):209–11. Amin 2009 {published data only} Amin M, Glynn F, Timon C. Randomised trial of tissue Keng 1989 {published data only} adhesive vs staples in thyroidectomy integrating patient Keng TM, Bucknall TE. A clinical trial of tissue adhesive satisfaction and Manchester score. Otolaryngology - Head (histoacryl) in skin closure of groin wounds. Medical Journal and Neck Surgery 2009;140(5):703–8. of Malaysia 1989;44(2):122–8. Avsar 2009 {published data only} Kent 2014 {published data only} Avsar A, Ustunnner I, Keskin L, Ozturk O, Tas E. 2- ∗ Kent A, Liversedge N, Dobbins B, McWhinnie D, octylcyanoacrylate tissue adhesive versus polypropylene Haider J. A prospective randomised controlled double- suture for skin closure of Pfannenestiel incision. Trk masked multi-center clinical trial of medical adhesives for the closure of laparoscopic incisions. Journal of Minimally Jinekoloji ve Obstetrik Derne i Dergisi 2009;6:117–22. Invasive Gynecology 2014;21:252–8. Blondeel 2004 {published data only} Khan 2006 {published data only} Blondeel P, Murphy J, Debrosse D, Nix, J, Puls LE, Khan RJK, Fick D, Yao F, Tang K, Hurworth M, Nivbrant Theodore N, et al.Closure of long surgical incisions with B, et al.A comparison of three methods of wound closure a new formulation of 2-octylcyanoacrylate tissue adhesive following arthroplasty: a prospective randomised controlled versus commercially available methods. American Journal of trial. The Journal of Bone and Joint Surgery 2006;88:238–42. Surgery 2004;188:307–13. Kouba 2011 {published data only} Brown 2009 {published data only} Kouba DJ, Tierney E, Mahmoud BH, Woo D. Optimizing Brown JK, Campbell BT, Drongowski RA, Alderman AK, closure materials for upperlid blepharoplasty: a randomized, Geiger JD, Teitelbaum DH, et al.A prospective, randomized controlled trial. Dermatologic Surgery 2011;37(1):19–30. comparison of skin adhesive and subcuticular suture for closure of pediatric hernia incisions: cost and cosmetic Krishnamoorthy 2009 {published data only} considerations. Journal of Paediatric Surgery 2009;44(7): Krishnamoorthy B, Najam O, Khan UA, Waterworth 1418–22. P, Fildes JE, Yonan N. Randomized prospective study Cheng 1997 {published data only} comparing conventional subcuticular skin closure with Cheng W, Saing H. A prospective randomised study of Dermabond skin glue after saphenous vein harvesting. The wound approximation with tissue glue in circumcision in Annals of Thoracic Surgery 2009;88(5):1445–9. children. Journal of Paediatric Child Health 1997;33:515–6. Livesey 2009 {published data only} Chibbaro 2009 {published data only} Livesey C, Wylde V, Descamps S, Estela CM, Bannister Chibbaro S, Tacconi L. Use of skin glue versus traditional GC, Learmonth ID, et al.Skin closure after total hip wound closure methods in brain surgery: a prospective, replacement: a randomised controlled trial of skin adhesive randomized, controlled study. Journal of Clinical versus surgical staples. Journal of Bone and Joint Surgery Neuroscience 2009;16(4):535–9. 2009;91(6):725–9. Dowson 2006 {published data only} Maartense 2002 {published data only} Dowson C, Gilliam A, Speake W, Lobo D, Beckingham. A Maartense S, Bemelman WA, Dunker MS, de Lint C, prospective, randomised controlled trial comparing n-butyl Pierik EGJM, Busch ORC, et al.Randomised study of the cyanoacrylate tissue adhesive (Liquiband) with sutures for effectiveness of closing laparoscopic trocar wounds with skin closure after laparoscopic general surgical procedures. octylcyanoacrylate, adhesive papertape or poliglecaprone. Surgical Laparoscopy, Endoscopy and Percutaneous Techniques British Journal of Surgery 2002;89:1370–5. 2006;16(3):146–50. Maloney 2013 {published data only} Eggers 2011 {published data only} Maloney J, Rogers GS, Kapadia M. Surgical corner: a Eggers MD, Fang L, Lionberger DR. A comparison of prospective randomized evaluation of cyanoacrylate glue wound closure techniques for total knee arthroplasty. devices in the closure of surgical wounds. The Journal of Journal of Arthroplasty 2011;26(8):1251–8. Drugs in Dermatology 2013;12(7):810–4. Greene 1999 {published data only} Millan 2011 {published data only} Greene D, Kock RJ, Goode RL. Efficacy of octyl-2- Millan P, Olivera R, Sanchez A. Efficacy and safety of 2- cyanoacrylate tissue glue in blepharoplasty. Archives of Facial octylcyanoacrylate versus simple suture in surgical wound Plastic Surgery 1999;1:292–6. closure of chronic skin inflammation. Dermatologia Revista Jallali 2004 {published data only} Mexicana 2011;55(4):185–7. ∗ Jallali N, Haji A, Watson CJ. A prospective randomized Mota 2009 {published data only} trial comparing 2-octylcyanoacrylate to conventional Mota R, Costa F, Amaral A, Oliveira F, Santos CC, Ayres- suturing in closure of laparoscopic cholecystectomy De-Campos D. Skin adhesive versus subcuticular suture

Tissue adhesives for closure of surgical incisions (Review) 33 Copyright © 2014 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. for perineal skin repair after episiotomy - a randomized Switzer 2003 {published data only} controlled trial. Acta Obstetricia et Gynecologica Scandinavica Swtizer E, Dinsmore R, North J. Subcuticular closure versus 2009;88(6):660–6. dermabond: a prospective randomised trial. The American Surgeon 2003;69:434–6. Ong 2002 {published data only} Ong C, Jacobsen A, Joseph V. Comparing wound closure Tierny 2009 {published data only} using tissue glue versus subcuticular suture for paediatric Tierney EP, Moy RL, Kouba DJ. Rapid absorbing gut surgical incisions: a prospective, randomised trial. Paediatric suture versus 2-octylethylcyanoacrylate tissue adhesive in Surgery International 2002;18:553–5. the epidermal closure of linear repairs. Journal of Drugs in Dermatology 2009;8(2):115–9. Ozturan 2001 {published data only} Ozturan O, Miman MC, Aktas D, Oncel S. Toriumi 1998 {published data only} Butylcyanoacrylate tissue adhesive for columellar incision Toriumi DM, O’Grady K, Desai D, Bagal AB. Use of octyl- closure. The Journal of Laryngology and Otology 2001;115: 2-cyanoacrylate for skin closure in facial plastic surgery. 535–40. Plastic and Reconstructive Surgery 1998;102(6):2209–19. Pronio 2011 {published data only} van den Ende 2004 {published data only} Pronio A, Di Fillippo A, Narillii P, Caporilli D, Vestri A, van den Ende ED, Vriens PWHE, Allema JH, Breslau Ciamberlaon B, et al.Closure of cutaneous incision after PJ. Adhesive bonds or percutaneous absorbable suture thyroid surgery: a comparison between metal clips and for closure of surgical wounds in children. Results of a cutaneous octyl-2-cyanoacrylate adhesive. A prospective prospective randomized trial. Journal of Pediatric Surgery randomized clinical trial. European Journal of Plastic Surgery 2004;39:1249–51. 2011;34:103–10. References to studies excluded from this review Ridgway 2007 {published data only} Ridgway D, Mahmood F, Moore L, Bramley D, Moore P. A Ak 2012 {published data only} blinded, randomised, controlled trial of stapled versus tissue Ak G, Alpk l ç Bak rt E, Kürklü E, Koray M, Tanyeri glue closure of neck surgery incisions. Annals of the Royal H, Zülfikar B. The evaluation of fibrin sealants and tissue College of Surgeons of England 2007;89(3):242–6. adhesives in oral surgery among patients with bleeding Romero 2011 {published data only} disorders. Turkish Journal of Hematology 2012;29(1):40–7. Romero P, Frongia G, Wingerter S, Holland-Cunz S. Alhopuro 1976 {published data only} Prospective, randomized,controlled trial comparing a tissue Alhopuro S, Rintala A, Salo H, Ritsila V. Tissue adhesive adhesive (Dermabond™) with adhesive strips (Steri- versus sutures in closure of incision wounds. A comparative Strips™) for the closure of laparoscopic trocar wounds in study in human skin. Annales Chirurgiae et Gynaecologiae children. European Journal of Pediatric Surgery 2011;21(3): 1976;65:308–12. 159–62. Chen 2010 {published data only} Sebesta 2004 {published data only} Chen K, Klapper AS, Voige H, Del Priore G. A randomized, Sebesta MJ, Bishoff JT. Octylcyanoacrylate skin closure controlled study comparing two standardized closure in laparoscopy. Journal of the Society of Laparoendoscopic methods of laparoscopic port sites. Journal of the Society of Surgery 2004;8(1):9–14. Laparoendoscopic Surgery 2010;14(3):391–4. Shamiyeh 2001 {published data only} Chow 2010 {published data only} Shamiyeh A, Schrenk P,Stelzer T, Wayand WU. Prospective Chow A, Marshall H, Zacharakis E, Paraskeva P,Purkayastha randomised blind controlled trial comparing sutures, tape, S. Use of tissue glue for surgical incision closure: a systematic and octylcyanoacrylate tissue adhesive for skin closure after review and meta-analysis of randomized controlled trials. phlebectomy. Dermatologic Surgery 2001;27(10):877–80. Journal of the American College of Surgery 2010;211:115–25. Sinha 2001 {published data only} Giri 2004 {published data only} Sinha S, Naik M, Wright V, Timmons J, Campbell AC. A Giri P, Das MK, Majumdar Giri A. Management of single blind, prospective, randomised trial comparing n- different types of wound by cyanoacrylate glue fixation: a butyl 2-cyanoacrylate tissue adhesive (Indermil) and sutures random study of 213 patients. Journal of the Indian Medical for skin closure in hand surgery. Journal of Hand Surgery Association 2004;102(11):624, 626. 2001;26B(3):264–5. Gorozpe-Calvillo 1999 {published data only} Sniezek 2007 {published data only} Gorozpe-Calvillo JI, Gonzalez-Villamil J, Santoya-Haro Sniezek P, Walling H, DeBloom III J, Messingham M, S. Closure of the skin with cyanoacrylate in caesarian VanBeek M, Kreiter C, et al.A randomised controlled surgery [Cierre de la piel con cianoacrilato en las cesareas]. trial of high-viscosity 2-octyl cyanoacrylate tissue adhesive Ginecologia y Obstetricia de Mexico 1999;67:491–5. versus sutures in repairing facial wounds following Mohs Jaibaji 2000 {published data only} micrographic surgery. Dermatological Surgery 2007;33(8): Jaibaji M, Liddington M, Geary P, Darcy C, Batchelor A, 966–71. Roberts A. Evaluation of the long term outcome of wounds

Tissue adhesives for closure of surgical incisions (Review) 34 Copyright © 2014 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. closed with “Indermil” adhesive. European Journal of Plastic European Journal of Pacing, Arrhythmias and Cardiac Surgery 2000;23:330–2. Electrophysiology 2011;13(3):416–20. Kuo 2006 {published data only} Steiner 2000 {published data only} Kuo F, Lee D, Rodgers G. Prospective, randomised, blinded Steiner Z, Mogilner J. Histoacryl vs Dermabond study of a new wound closure ﬁlm versus cutaneous suture cyanoacrylate glue for closing small operative wounds. for surgical wound closure. Dermatological Surgery 2006; Harefuah 2000;139(11-12):409-11, 496. 32:676–81. Sun 2005 {published data only} Matin 2003 {published data only} Sun J, Chen Q-M, Zhang M, Shi C. Octylcyanoacrylate Matin SF. Prospective randomized trial of skin adhesive versus absorbable suture in the repair of skin wounds in versus sutures for closure of 217 laparoscopic port-site children. Zhongguo Linchuang Kangfu 2005;9(19):186–7. incisions. Journal of the America College of Surgeons 2003; Wong 2011 {published data only} 196(6):845–53. Wong EM, Rainer TH, Ng YC, Chan MS, Lopez V. Cost- Maw 1997 {published data only} effectiveness of Dermabond versus sutures for lacerated Maw JL, Quinn JV, Wells GA, Ducic Y, Odell PF, Lamothe wound closure: a randomised controlled trial. Hong Kong A, et al.A prospective comparison of octylcyanoacrylate Medical Journal 2011;17 Suppl 6:4–8. tissue adhesive and suture for the closure of head and neck incisions. Journal of Otolaryngology 1997;26(1):26–30. References to studies awaiting assessment Ong 2010 {published data only} Gennari 2004 {published data only} Ong J, Ho KS, Chew MH, Eu KW. Prospective randomised study to evaluate the use of DERMABOND ProPen (2- Handschel 2006 {published data only} octylcyanoacrylate) in the closure of abdominal wounds Jan 2013 {published data only} versus closure with skin staples in patients undergoing Nipshagen 2008 {published data only} elective colectomy. International Journal of Colorectal Disease 2010;25(7):899–905. Yoon 2006 {published data only} Orozco-Razon 2002 {published data only} Additional references Orozco-Razon LF, Millan-Guerrero RO, Vera-Rodriguez SE. Butylcyanoacrylate tissue adhesive for columellar Bruns 1996 incision closure [Cyanoacrylate compared with traditional Bruns TB, Simon HK, McLario DJ. Laceration repair using surgery in tension–free incision closure [Spanish]]. Gaceta a tissue adhesive in a children’s emergency department. Medica de Mexico 2002;138(6):505–9. Pediatrics 1996;98:673–5. Quinn 1998 {published data only} Coover 1959 Quinn J, Wells G, Sutcliffe T, Jarmuske M, Maw J, Stiell I, Coover HN, Joyner FB, Sheere NH. Chemistry and et al.Tissue adhesive versus suture wound repair at 1 year: performance of cyanoacrylate adhesive. Journal of the Society randomized clinical trial correlating early, 3-month, and 1- of Plastic Surgery of England 1959;1:5–6. year cosmetic outcome. Annals of Emergency Medicine 1998; Deeks 2011 32(6):645–9. Deeks JJ, Higgins JPT, Altman DG (editors). Chapter 9: Sajid 2009 {published data only} Analysing data and undertaking meta-analyses. In: Higgins Sajid MS, Siddiqui MR, Khan MA, Baig MK. Meta-analysis JPT, Green S (editors). Cochrane Handbook for Systematic of skin adhesives versus sutures in closure of laparoscopic Reviews of Interventions. Version 5.1.0 (updated March port-site wounds. Surgical Endoscopy 2009;23(6):1191–7. 2011). The Cochrane Collaboration, 2011. available from Silvestri 2006 {published data only} www.cochrane-handbook.org. Silvestri A, Brandi C, Grimaldi L, Nisi G, Brafa A, Calabro Dunker 1998 M, et al.Octyl-2-cyanoacrylate adhesive for skin closure and Dunker MS, Stiggelbout AM, van Hogezand RA, Ringers prevention of infection in plastic surgery. Aesthetic Plastic J, Grifﬁoen G, Bemelman WA. Cosmesis and body image Surgery 2006;30:695–9. after laparoscopic assisted and ileocolic resection for Crohn’s disease. Surgical Endoscopy 1998;12:1334–40. Singer 2002 {published data only} Singer AJ, Quinn JV, Clarke RE, Hollander JE. Closure Emori 1993 of lacerations and incisions with octylcyanoacrylate: a Emori TG, Gaynes RP. An overview of nosocomial multicentre randomised controlled trial. Surgery 2002;131: infections, including the role of the microbiological 270–6. laboratory. Clinical Microbiology 1993;6:428–42. Spencker 2011 {published data only} Farion 2001 Spencker S, Coban N, Koch L, Schirdewan A, Mueller D. Farion KJ, Russell KF, Osmond MH, Hartling L, Comparison of skin adhesive and absorbable intracutaneous Klassen TP, Durec T, et al.Tissue adhesives for traumatic suture for the implantation of cardiac rhythm devices. lacerations in children and adults. Cochrane Database

Tissue adhesives for closure of surgical incisions (Review) 35 Copyright © 2014 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. of Systematic Reviews 2001, Issue 4. [DOI: 10.1002/ of lacerations during competitive athletics. American Journal 14651858.CD003326] of Emergency Medicine 2000;18:261–3. Higgins 2003 Quinn 1993 Higgins JPT, Thompson SG, Deeks JJ, Altman DG. Quinn JV, Drzewiecki A, Li MM, Stiell IG, Sutcliffe T, Measuring inconsistency in meta-analyses. BMJ 2003;327 Elmslie TJ, et al.A randomised, controlled trial comparing a (7414):557–60. tissue adhesive with suturing in the repair of paediatric facial Higgins 2011 lacerations. Annals of Emergency Medicine 1993;22:1130–5. Higgins JPT, Altman DG, Sterne JAC (editors). Chapter 8: Assessing risk of bias in included studies. In: Higgins Quinn 1995 JPT, Green S (editors). Cochrane Handbook for Systematic Quinn JV, Drzewiecki AE, Stiell IG, Elmslie TJ. Appearance Reviews of Interventions. Version 5.1.0 (updated March scales to measure cosmetic outcomes of healed lacerations. 2011). The Cochrane Collaboration, 2011. Available from American Journal of Emergency Medicine 1995;13:229–31. www.cochrane-handbook.org. Quinn 1997 Hollander 1995 Quinn J, Wells G, Sutcliffe T. A randomized trial comparing Hollander JE, Singer AJ, Valentine S, Henry MC. Wound octylcyanoacrylate tissue adhesive and sutures in the registry: development and validation. Annals of Emergency management of lacerations. JAMA 1997;277(19):1527–30. Medicine 1995;25:675–85. Schunemann 2011a Houston 1969 Schünemann HJ, Oxman AD, Higgins JPT, Vist GE, Houston S, Hodge JW Jr, Ousterhout DK, Leonard F. The Glasziou P, Guyatt GH. Chapter 11: Presenting results effect of alpha-cyanoacrylates on wound healing. Journal of and ’Summary of findings’ tables. In: Higgins JPT, Biomedical Materials Research 1969;3(2):281–9. Green S (editors). Cochrane Handbook for Systematic Lefebvre 2011 Reviews of Interventions Version 5.1.0 (updated March Lefebvre C, Manheimer E, Glanville J. Chapter 6: Searching 2011). The Cochrane Collaboration, 2011. Available from for studies. In: Higgins JPT, Green S (editors). Cochrane www.cochrane-handbook.org. Handbook for Systematic Reviews of Interventions. Version 5.1.0 (updated March 2011). The Cochrane Collaboration, Schunemann 2011b 2011. Available from www.cochrane-handbook.org 2011. Schünemann HJ, Oxman AD, Higgins JPT, Deeks JJ, Malone 2002 Glasziou P, Guyatt GH. Chapter 12: Interpreting results Malone DL, Genuit T, Tracy JK. Surgical site infections: and drawing conclusions. In: Higgins JPT, Green S reanalysis of risk factors. Journal of Surgical Research 2002; (editors). Cochrane Handbook for Systematic Reviews 103:89–95. of Interventions Version 5.1.0 (updated March 2011). The Cochrane Collaboration, 2011. Available from Osmond 1999 www.cochrane-handbook.org. Osmond MH. Pediatric wound management: the role of tissue adhesives. Pediatric Emergency Care 1999;15(2): SIGN 2011 137–40. Scottish Intercollegiate Guidelines Network (SIGN). Search Perron 2000 filters. http://www.sign.ac.uk/methodology/filters.html# Perron AD, Garcia JA, Hays EP. The efficacy of random 2011. cyanoacrylate-derived surgical adhesive for use in the repair ∗ Indicates the major publication for the study

Characteristics of included studies [ordered by study ID]

Amin 2009

Methods RCT with 3 months follow-up. 72 participants randomised, but only 60 had 3-month outcome data collected and reported

Participants 72 participants receiving minimally invasive thyroidectomy using either video-assisted thyroidectomy or minimal incision thyroidectomy Surgery performed in 1 centre in Ireland. All cases were performed by the senior surgeon Exclusion criteria: autoimmune thyroiditis, diabetes mellitus and/or poor general health Trial conducted in Dublin, Ireland

Interventions Group 1 (n = 38; results reported for 33): 2-octylcyanoacrylate (Dermabond®) tissue adhesive (Ethicon Inc, a Johnson & Johnson company, Somerville, New Jersey, USA) Group 2 (n = 34; results reported for 27): staples

Outcomes Cosmetic appearance by participant and by surgeon (at 3 months) using the Manchester scar scale Patient satisfaction (self assessed at 3 months): collected using a 10 cm VAS line where 0 was poor and 10 was excellent. Data for overall patient satisfaction score (n = 60) was calculated by the review authors using summary data presented in the study report. The patient satisfaction assessment form also measured: cosmesis; ability to shower same day (as operation); need to visit GP for wound; pain on removing clips; pain/tightness of wound; overall wound comfort and allergic reactions. Only ability to shower data was reported in addition to overall satisfaction score - ability to shower data are not presented here

Notes Cosmetic appearance outcome data not clearly presented for participants or surgeons. Authors contacted Study also reports pain at 1 and 10 days after surgery - not reported here Cost of tissue adhesive reported as EUR22 - no corresponding data for staples

Risk of bias

Bias Authors’ judgement Support for judgement

Random sequence generation (selection Low risk Quote: “Randomization was performed prior bias) to commencement of the study as follows: Opaque envelopes were numbered sequentially from 1 to 75. A table of random numbers was generated by a computer program andused for groupassignment;if the last digit of the random number was from 0 to 4, the assignment was to Group A (tissue adhesive) , and if the last digit was from 5 to 9, the assignment was to Group B (staples).” Comment: adequate method of random se-

quence generation

Allocation concealment (selection bias) Low risk Quote: “The assignments were then placed into the opaque envelopes and the envelopes sealed. As eligible participants were entered into the trial, these envelopes were opened in sequential order to give each patient his or her random group assignment. The envelopes were opened by the operating surgeon following patient consent and just prior to the surgical procedure.” Comment: use of sealed, numbered opaque envelopes considered robust

Blinding of participants and personnel Unclear risk Quote: “Blindingwas not feasible at the time (performance bias) of skin closure, nor was it of any beneﬁt at the All outcomes two-week postoperative visit, as wound cosmesis at this stage is not considered predictive of the long-term cosmetic outcome.” Comment: understandably difﬁcult to blind operating surgeon and participants to the intervention.Staff would have been aware of allocation on wound closure and during short-term post-operative assessment - however none of these outcomes are reported here. Participants were not blinded. Unclear whether this would have led to bias in the study in terms of satisfaction in favour of 1 treatment

Blinding of outcome assessment (detection Low risk Quote: “Blinding was achieved during eval- bias) uation of the surgical wound at or after three All outcomes months postoperatively as surgeons were un- aware of the wound closure method used. A three-month appointment was arranged for each participant for wound evaluation by a surgeon who was not involved in the patient management.” Comment: blinded outcome assessment undertaken at 3 months for surgeon cosmetic outcome data although we were unable to extract these data for the review. Par- ticipant assessment not blinded (as noted above)

Incomplete outcome data (attrition bias) Unclear risk Quote: “Sixty out of the 72 patients agreed All outcomes to come back for the three-month postoperative evaluation. The remainder did not attend do to the long travelling [sic]/commuting

distance to the tertiary service” Comment: data from only 60 of the 72 participants randomised were available at 3 months. Also possible differential attrition (adhesive 13%, staples 21%). Deemed to be at unclear risk of bias

Selective reporting (reporting bias) High risk Comment: multiple dimensions of patient satisfaction assessed with ability to shower selectively reported - data not presented here. Study protocol not sought

Other bias Low risk None noted

Avsar 2009

Methods RCT with 40 participants with 40 days follow-up

Participants 40 women undergoing the Pfannenstiel incision Exclusion criteria: known allergy to trial products Conducted in Turkey

Interventions Group 1 (n = 20): high viscosity 2-octylcyanoacrylate (High Viscocity Dermabond®; Ethicon Inc, a Johnson & Johnson company, Somerville, New Jersey, USA) tissue adhesive Group 2 (n = 20): polypropylene sutures All participants had skin cleansed with povidone iodine and 1 g antibiotic prophylaxis

Outcomes Wound infection (at 2, 7 and 40 days after surgery: 7-day data used for analyses, as these data were most clear from the translation) Time for skin closure Participant satisfaction at day 40. Participants were asked how satisﬁed they were with their skin closure they could select from the following responses: very bad, poor, average, good, very good

Notes Data extraction based on English language abstract and partial translation of report text that was published in Turkish The study reports an outcome - wound disruption - as yet we have been unable to translate whether this can be interpreted as wound dehiscence Cometic appearance not used in this review as assessed at less than 3 months

Risk of bias

Bias Authors’ judgement Support for judgement

Random sequence generation (selection Unclear risk Comment: unable to gauge from transla- bias) tion

Allocation concealment (selection bias) Unclear risk Comment: unable to gauge from translation

Blinding of participants and personnel Unclear risk Comment: unable to gauge from transla- (performance bias) tion All outcomes

Blinding of outcome assessment (detection Unclear risk Comment: unable to gauge from transla- bias) tion All outcomes

Incomplete outcome data (attrition bias) Unclear risk Comment: unable to gauge from trans- All outcomes lation. We acknowledge risk of selection bias from reporting infection outcome data from the time point where statistical significance was observed

Selective reporting (reporting bias) Unclear risk Comment: unable to gauge from translation

Other bias Unclear risk Comment: unable to gauge from translation

Blondeel 2004

Methods RCT with 1-month follow-up

Participants 217 adults requiring any skin incision closure 4 cm or greater in length. Multicentre study conducted at: Department of Plastic Surgery, University Hospital Gent, Gent, Belgium; Intitute Mutualiste Montsouris, Paris, France; Crestwood Hospital, Huntsville, AL, USA; Gynecologic Oncology Research and Development, LLC, Greenville, SC, USA; Naval Medical Centre, San Diego, CA, USA; University Dental Hospital, Manchester, UK Exclusion criteria: a history of peripheral vascular disease; insulin-dependent diabetes; a blood-clotting disorder or taking anticoagulants within 7 days of surgery; concurrent use of steroids; or a personal or family history of keloid formation or hypertrophy; impaired wound healing; or a known allergy to cyanoacrylate or formaldehyde

Interventions Group 1 (n = 106): high viscosity 2-octylcyanoacrylate (High Viscocity Dermabond®; Ethicon Inc, a Johnson & Johnson company, Somerville, New Jersey, USA) tissue adhesive Group 2 (n = 103): any other commercially available device such as sutures, staples, or tapes or compared with low viscosity 2-octylcyanoacrylate (n = 42; Dermabond®, Ethicon Inc, a Johnson & Johnson company, Somerville, New Jersey, USA)

Outcomes Wound dehiscence, infection, patient satisfaction and surgeon satisfaction at 10 days

Notes Cosmetic appearance data not used as measured at less than 3 months

Risk of bias

Bias Authors’ judgement Support for judgement

Random sequence generation (selection Low risk Quote: ” . . . patients at each study site were randomised bias) by computer generated code in a 1:1 ratio to epidermal incision closure with high viscosity 2-octylcyanoacrlate or a commercially available device“ Comment: use of computer generated code constitutes low risk

Allocation concealment (selection bias) Unclear risk Quote: ”Sealed envelopes containing the concealed treatment allocations were opened in the operating room immediately before epidermal closure.“ Comment: judgement of unclear risk of bias made as not clear whether envelopes were sequentially numbered and opaque. Not clear who was responsible for allocation

Blinding of participants and personnel Unclear risk Quote: ”Blinding to type of treatment was not feasible“, (performance bias) due to the nature of the personnel-led intervention All outcomes Comment: understandably difﬁcult to fully blind participants and personnel to intervention

Blinding of outcome assessment (detection Unclear risk Quote: ”On day 10, wounds were assessed for adequate bias) progress in healing, wound-related infection, and indica- All outcomes tors of an acute inﬂammatory reaction.“ Similarly, ”.. . on day 10, physicians completed a questionnaire that measured satisfaction with each use of a device.“, also, “a counterpart questionnaire [given to patients] for patient satisfaction for cosmesis, overall comfort, ability to shower, dressing changes, tension at the wound, hygienic problem, allergic reaction, and overall satisfaction” Comment: there is no mention that the wound was assessed by a blinded assessor. Such assessment would be difﬁcult for satisfaction scores - as noted above

Incomplete outcome data (attrition bias) Low risk Quote: “The study was not completed by 4 patients as- All outcomes signed to high viscosity group 2 octylcyanoacrylate because of voluntary withdrawal (n=2), or loss to follow up (n= 2) and by 6 patients to the control devices because of voluntary withdrawal (n=2), loss to the follow up (n=3), or death (n=1) [attributed to Burkitt’s Lymphoma]”. Comment: the numbers were fairly small and were fairly evenly distributed between the treatment arms. Therefore we do not believe this represented a signiﬁ- cant risk of bias

Selective reporting (reporting bias) Low risk No direct quotations, but all the variables outlined in the methodology were accounted for in the results Comment: judged as low risk. Study protocol not sought

Other bias Low risk No other sources of bias were detected

Brown 2009

Methods Parallel RCT with 6 weeks follow-up. No withdrawals reported

Participants 134 children undergoing inguinal herniorrhaphy (ages ranged from 1 month to 12 years) All operations performed by 1 of 4 paediatric surgeons No other inclusion or exclusion data reported Undertaken in 1 centre in the USA

Interventions Group 1 (n = 64): 2-octylcyanoacrylate (Dermabond®) tissue adhesive (Ethicon Inc, a Johnson & Johnson company, Somerville, New Jersey, USA) Group 2 (n = 70): sutures (5.0 Monocryl) Group numbers calculated from data in paper as not reported directly

Outcomes Wound dehiscence (6 weeks) Time for skin closure Relative costs of materials required for skin closure

Notes Surgeon cosmetic appearance data not used as measured at < 3 months

Risk of bias

Bias Authors’ judgement Support for judgement

Random sequence generation (selection Unclear risk Quote: “Just before wound closure, a sealed bias) envelope indicating randomization to skin adhesive or suture closure was revealed.” Comment: no information about how the randomisation sequence was generated

Allocation concealment (selection bias) Unclear risk Quote: “Just before wound closure, a sealed envelope indicating randomization to skin adhesive or suture closure was revealed.” Comment: no information about whether envelopes were opaque and sequential and who prepared or opened these

Blinding of participants and personnel Unclear risk Quote: “Masking as to skin adhesive vs su- (performance bias) ture closure was not possible for the operating All outcomes surgeon because of the nature of the intervention.”

Comment: understandably difﬁcult to blind operating surgeon and participants to the intervention. Staff would have been aware of allocation on wound closure

Blinding of outcome assessment (detection Low risk Quote: “However, subsequent interviewers bias) were masked as to group during assessment of All outcomes cosmetic outcome measures, as well as those related to efﬁciency, cost, and complications of wound closure.” Comment: deemed at low risk of bias for debridement and costs outcomes

Incomplete outcome data (attrition bias) Low risk No report of any loss to follow-up All outcomes Comment: judged as low risk

Selective reporting (reporting bias) Low risk No direct quotations, but all the variables outlined in the methodology were accounted for in the results Comment: judged as low risk. Study protocol not sought

Other bias Low risk No other sources of bias were detected

Cheng 1997

Methods RCT with 1-month follow-up

Participants 86 healthy male patients under the age of 12 years requiring elective circumcision. Study conducted at Duchess of Kent Hospital, Hong Kong. No speciﬁc exclusion criteria were described

Interventions Group 1 (n =40): butylcyanoacrylate (Histoacryl®) tissue adhesive Group 2 (n = 46): 4.0 catgut sutures

Outcomes Dehiscence and infection at days 1, 2, 3, 7 and 30. Cosmetic appearance, bleeding and wound inﬂammation were also assessed at these time points. Bleeding and wound inﬂammation were not of interest to this review

Notes Cosmetic appearance data were not used as measured at < 3 months

Risk of bias

Bias Authors’ judgement Support for judgement

Random sequence generation (selection Unclear risk Quote: “ . . . [patients were] randomised into two groups” bias) Comment: no method of random sequence generation was actively discussed; therefore the risk of bias for this

domain is unclear

Allocation concealment (selection bias) Unclear risk Quote: Patients were “ . . . randomised into two groups” as previously mentioned Comment: no method of allocation concealment was made clear. Therefore, the risk of bias is again unclear

Blinding of participants and personnel Unclear risk Not reported whether the participants or personnel (performance bias) were blinded to the intervention All outcomes Comment: as the participants were under the age of 12 and under general anaesthesia, it is reasonable to as- sume that they were blinded to the intervention. How- ever, the personnel were probably not blinded as they carried out the intervention

Blinding of outcome assessment (detection Unclear risk Quote: “The wounds of all the patients were then assessed bias) on day 1, day 2 and day 3, 1 week and 1 month after the All outcomes operation. Wound inﬂammation, infection, bleeding, cosmetic result and dehiscence were assessed. A questionnaire was completed” Comment: no mention of whether the assessors of the wound and the cosmetic results were blinded to the intervention. It is not clear if the questionnaire was ﬁlled out by parents of the children, who may have been blinded to the intervention, or the assessors, who were likely to know which intervention was given. The judgement for this domain is therefore unclear

Incomplete outcome data (attrition bias) Low risk No direct quotations, but no losses to follow-up were All outcomes reported Comment: therefore judged as low risk

Selective reporting (reporting bias) Low risk No direct quotations, but the results account for the all the assessment methods identiﬁed to qualify the success of the intervention Comment: no evidence of reporting bias, therefore judged as low risk

Other bias Unclear risk Possible unit of analysis issue for dehiscence outcome

Chibbaro 2009

Methods RCT with up to 12 months follow-up, with assessment at 1, 3, 5-7 and 14 days, then 1, 3, 6 and 12 months

Participants 40 participants undergoing elective cranial supratentorial surgery Exclusion criteria: undergoing a procedure for infective pathology; head trauma; admission to the intensive care unit; a dermatologic disease; previous cranial radiotherapy; diabetes mellitus; known blood-clotting disorders; and allergy to cyanoacrylate products

Undertaken in 1 centre in Italy

Interventions Group 1 (n = 20): n-butyl-cyanoacrylate (Liquiband®; MedLogic Global Ltd, Plymouth, Devon, UK) tissue adhesive Group 2 (n = 20): sutures (Monosof 2/0, Tyco United States Surgical, Norwalk, CT, USA) or staples (Auto Suture Appose ULC 35, Tyco United States Surgical) All participants received the same antibiotic prophylaxis (1 dose of 3rd-generation cephalosporin 20 min prior to surgical incision) Wound dressings were not used for those allocated to tissue adhesive because this formed its own waterproof and antimicrobial wound dressing Participants in the tissue adhesive group (Group 1) were permitted to shower from the day after the surgical procedure, while those participants allocated to sutures were instructed to keep their wounds clean and dry until the removal of the sutures

Outcomes Wound dehiscence (collected to 7th day postoperatively) Wound infection (collected to 7th day postoperatively; infection not deﬁned) Cosmetic appearance (nurse-blinded) using modiﬁed Hollander Wound Score scale (up to 12 months) Skin closure time

Notes Cosmetic appearance (patient and surgeon) using a VAS of 1 to 10 where a score of 10 reﬂected optimal cosmetic outcome. It is not clear at what time points these data were collected. Authors were contacted Supplmentary material referenced in the main paper also checked

Risk of bias

Bias Authors’ judgement Support for judgement

Random sequence generation (selection Unclear risk Quote:”At the moment of skin closure, they bias) [the patients] were randomly allocated into one of two groups (A or B) of 20 patients each. “ Comment: no reporting on how the patients were randomly allocated or the method used

Allocation concealment (selection bias) Unclear risk Quote: ”The randomization sequence was arranged on the same day as surgery; each patient’s name was randomly assigned to one of 40 envelopes (20 marked as LiquiBand, 10 as TTS and 10 as SC).“ Comment: no indication that the allocation was concealed

Blinding of participants and personnel Unclear risk No direct quotations. Did not report that (performance bias) the participants or personnel were blinded All outcomes to the intervention Comment: understandably difﬁcult to

blind the operating surgeon to the intervention

Blinding of outcome assessment (detection Unclear risk Quote: ”All of the patients were followed up bias) post-operatively on days 1, 3, 5-7 by the same All outcomes ward nurse who initially recorded details regarding their wound aspects“ ”After discharge, a second nurse (not from theneurosurgicaldepartment),usingthesame scale, continued the follow-up, initially at 2 weeks, and then at 1, 3, 6 and 12 months post-operatively”. Noted in abstract that this second nurse was blinded Comment: although the study intimates that the nurse was not from the department and the second nurse was blinded it is not clear that the ﬁrst nurse was blinded

Incomplete outcome data (attrition bias) Low risk Quote: “Only 39 patients (19 in group All outcomes A and 20 in group B) were available at the 12 month follow up, as one patient in group A developed a post-operative intracra- nial haematoma. This required an emergency evacuation procedure and the wound was subsequently closed with stitched” Comment: given that 39/40 participants had 12-month data available, we made an overall judgment of low risk of attrition bias

Other bias Low risk No other sources of bias detected

Dowson 2006

Methods Parallel group RCT with 3-month follow-up

Participants 168 patients recruited for laparoscopic procedures at Queen’s Medical Centre, Notting- ham. 154 participants included in the ﬁnal analysis. Withdrawals were accounted for. Exclusion criteria included insulin dependant diabetes, prolonged corticosteroid use; known keloid scarring; wounds greater than 5cm in length without deep layer sutures; those undergoing emergency surgery; patients unable to give informed consent

Interventions Group 1 (n = 76): n-butyl-cyanoacrylate (Liquiband®; MedLogic Global Ltd, Plymouth, Devon, UK) tissue adhesive Group 2 (n = 78): interrupted, non-absorbable suture

Outcomes Time to closure, dehiscence, satisfaction, cosmesis and infection (we used 24 to 48 h ﬁgures to avoid unit of analysis issues for dehiscence and infection)

Notes Unclear reference to subcutaneous layers being dealt with. Authors contacted for mean and standard deviations of time to closure and cosmesis, but as we received no reply, this information was not included

Risk of bias

Bias Authors’ judgement Support for judgement

Random sequence generation (selection Unclear risk Quote: “Randomisation was performed using bias) consecutively numbered, sealed envelopes by a person not involved with the study” Comment: although allocation was concealed, there was no speciﬁc mention of how the sequence was randomly generated

Allocation concealment (selection bias) Low risk Quote: “Randomisation was performed using consecutively numbered, sealed envelopes by a person not involved with the study” Comment: allocation was concealed adequately

Blinding of participants and personnel Unclear risk Quote: “Participants in the trial were not in- (performance bias) formed of the wound closure method they had All outcomes been allocated to until they had undergone surgery. The investigators were not blinded” Comment: the risk of performance bias was unclear as it was not possible to blind the investigators to the procedure they were car- rying out

Blinding of outcome assessment (detection Unclear risk Quote: “Patients were followed up at 24 to bias) 48hours,4to6weeksand3monthspostoper- All outcomes atively by C.C.D, A.D.G., or W.J.S. At these times it was documented if wounds showed any of the following characteristics; erythema, oedema, tenderness, inﬂammation, drainage/ discharge and/or malodorous smell.” Concurrently, at the 3 month cosmetic evaluation, “a blinded assessment by a quali- ﬁed nurse or surgical registrar not involved in the study also being made at 3 months”

Comment: adequate blinding of assessment at 3 months. Potential for high risk of detection bias for outcomes reported prior to this

Incomplete outcome data (attrition bias) High risk Flow chart suggests no loss to follow-up All outcomes in ﬁrst 48 h postoperatively. At 4-6 weeks 20 participants were lost to follow-up in the suture arm and 15 in the adhesive arm. There was further loss to follow-up at 3 months: the suture arm suffered a total loss of 14 (with 7 missing both the 4-6 week and the 3-month follow-up), while the adhesive group had 8 missing the 3-month follow- up (with 6 missing both follow-ups) Comment: relative high attrition of data from study for outcomes collected at later time points

Other bias Low risk No other sources of bias identiﬁed

Eggers 2011

Methods 4-arm parallel RCT with 90 participants with 6 weeks of follow-up (assessment at 24 h, 3 and 6 weeks). Only data on 75 of participants were reported

Participants Participants undergoing total knee arthroplasty Exclusion criteria: medical conditions or personal circumstances that would prevent participation and completion of physical therapy and follow-up visits; current participation in another clinical trial; preoperative systemic infections; uncontrolled diabetes; diseases or conditions known to effect the wound healing process; known hypersensitivity to cyanoacrylate, formaldehyde, or the dye D&C Violet #2 (Aesculap, Inc, Center Valley, PA); prior knee hardware ﬁxation devices; prior knee incisions greater than 9 cm, and arthroﬁbrosis as evidence by limited ROM of 80° or higher Study performed in 1 centre in the USA

Interventions Group 1 (n = 19): subcutaneous closure method: sutures at 1.5/cm; skin closure method: 2-octylcyanoacrylate (Dermabond®) tissue adhesive (Ethicon Inc, a Johnson & Johnson company, Somerville, New Jersey, USA) Group 2 (n = 18): subcutaneous closure method: sutures at 1.5/cm; skin closure method: butylcyanoacrylate tissue adhesive (Histoacryl Blue tissue adhesive B Braun Corp) Group 3 (n = 19): subcutaneous closure method: sutures at 1.0/cm; skin closure method: staples (Visistat 35W Stapler (Teleﬂex Corp)

Group 4 (n = 19): subcutaneous closure method: sutures at 1.0/cm; skin closure method: Monocryl suture (poliglecaprone 25; Ethicon) Data on 15 participants were excluded from analysis data: it was not reported which trial groups these participants were from. We note there were slight difference to the procedures in each group for the method of closure of the sub-cutaneous layer. Details for (1) sub-cutaneous closure methods and (2) skin closure method are provided above

Outcomes Wound dehiscence (24 h, 3 weeks, and 6 weeks postoperatively) Wound infection (24 h, 3 weeks, and 6 weeks postoperatively; infection not deﬁned. We have taken the total number of infections reported over the 6-week period, however, it is not clear whether some participants reported more than 1 infection as the number of infections was reported rather than number of people having an infection Relative costs of materials required for skin closure Skin closure time

Notes Surgeon cosmetic appearance data not used as measured at < 3 months

Risk of bias

Bias Authors’ judgement Support for judgement

Random sequence generation (selection Low risk Quote: “ . . . the eligible subjects were bias) randomly categorized (via a pseudo-random number generator algorithm) into 1 of 4 co- horts” Comment: adequate random sequence generation

Allocation concealment (selection bias) Unclear risk Quote: “ . . . the eligible subjects were randomly categorized (via a pseudo-random number generator algorithm) into 1 of 4 co- horts”. The study also states, “The surgeon was blinded to the closure technique before and during the operation until subcutaneous closure” Comment: although sequence generation was adequate and there is evidence to suggest that allocation was concealed, there is not enough evidence to suggest how this was done and how allocation was concealed to the personnel

Blinding of participants and personnel Unclear risk Quote: “The surgeon was blinded to the clo- (performance bias) sure technique and during the operation until All outcomes subcutaneous closure” Comment: although attempts were made to conceal the allocated treatment, it would be difﬁcult to blind the operating surgeon

to the intended intervention, therefore our judgement is unclear. No information was provided about the extent to which participants were blinded

Blinding of outcome assessment (detection Unclear risk Quote: “At each visit, a physical exami- bias) nation was conducted; and adverse events All outcomes were recorded. The physical examinations after TKA included evaluation of peripheral edema, infection, and dehiscence as well as pain level (0-10 scale) and cosmesis (100- mm visual analogue scale (VAS)) evaluation by patient. General health and wellness were further evaluated by an SF-12v2 survey” Comment: the extent to which the participants were blinded to the intervention is unclear; it is also unclear whether those conducting the examinations were aware of the treatment given. The outcome of cosmesis was evaluated before the our spec- iﬁed time-frame for the purposes of our review. The judgement remains unclear in this case

Incomplete outcome data (attrition bias) Unclear risk Quote:“Of the 90 subjects recruited, 15 were All outcomes excluded because of screen failure; 6 were di- agnosed with arthroﬁbrosis after surgery, 4 failed to follow preferred physical therapy, and 5 sustained unrelated co-morbidities prevent- ing study completion.” “Consequently, a total of 75 subjects, 19 per cohort with the exception of 18 for the n- butyl-2 adhesive cohort, completed the study; and their data are presented in the following section.” Comment: it seems that these participants were excluded post-randomisation

Other bias Unclear risk Possible issues with clustering for infection outcome data

Methods Randomised split-body design study with 1-month follow-up. No participants lost to follow-up

Participants 20 adults requiring bilateral blepharoplasty for functional or aesthetic indications (40 eyelids were treated). Procedure conducted at: Division of Otolaryngology - Head and Neck Surgery, Plastic and Reconstructive Surgery, Standford University Medical Centre and Palo Alto Veterans Healthcare System, Palo Alto, California, and Department of Otolayngology - Head and Neck Surgery, Cleveland Clinic Florida, Naples, USA. Used a blepharoplasty model with identical skin sites on the same participant and each participant acted as his or her own control. No speciﬁc exclusion criteria were described

Interventions Group 1 (n = 20): left or right upper eye lid incision closed with 2-octylcyanoacrylate (Dermabond®, Ethicon Inc, a Johnson & Johnson company, Somerville, New Jersey, USA) Group 2 (n = 20): other eyelid incision closed with 6.0 suture (10 fast-absorbing gut or 10 polypropylene, Prolene, Ethicon Inc, a Johnson & Johnson company, Somerville, New Jersey, USA) Blepharoplasties were closed on the tissue adhesive side by using Castroviejo forceps to approximate the skin edges in 15 participants and by using 3-4 sutures as handles to facilitate apposition and eversion of edges in 5 participants

Outcomes Dehiscence, infection, patient satisfaction, and surgeon satisfaction at 1, 2, and 4 weeks. Time for closure at end of procedure

Notes Cosmetic appearance data could not be used as measured at < 3 months

Risk of bias

Bias Authors’ judgement Support for judgement

Random sequence generation (selection Unclear risk Quote: ”Each patient had been randomised bias) to have either the right or left upper eyelid serve as the experimental closure [with adhesive] . . . and the opposite eyelid as the control with sutures“ Comment: despite stating that each patient was randomised to treatment group, there was no speciﬁc description of how the randomisation process was done or achieved

Allocation concealment (selection bias) Unclear risk Quote: ”Each patient had been randomised ...“ Comment: no description of how allocation was undertaken and whether it was concealed

Blinding of participants and personnel Unclear risk Not reported whether participants or per- (performance bias) sonnel were blinded to the intervention All outcomes Comment: understandably difﬁcult to

blind surgeon to the intervention

Blinding of outcome assessment (detection Unclear risk Quote: ”The photographs were shown to 5 bias) observers blinded to the technique of closure“ All outcomes to evaluate wound quality post operatively.” Comment: blinding of outcome assessment achieved for cosmetic appearance, which was not included in this review. Not clear whether blinded assessment was undertaken for other outcomes

Incomplete outcome data (attrition bias) Low risk Quote: “The 5 blinded observers using the All outcomes visual analogue scale to rate the 40 treated eyelids did not ﬁnd any statistically signiﬁcant difference between the wound quality . . .” Comment: suggested that all participants were accounted for, as the study had 20 participants (40 eyelids)

Other bias Low risk No other sources of bias detected

Jallali 2004

Methods RCT in which participants (with multiple port wounds) were randomised. Follow-up was for between 6 and 8 weeks. No withdrawals reported

Participants Participants undergoing laparoscopic cholecystectomy No other inclusion or exclusion information Undertaken in 1 UK centre

Interventions Group 1 (12 participants; 48 wounds): 2-octylcyanoacrylate Group 2 (13 participants; 51 wounds): absorbable sutures Participants in the suture arms had a dressing placed over the wound. Those in the tissue adhesive arm did not require a dressing

Outcomes Skin closure time (not clear if these data were collected for multiple wounds on the same person)

Notes Outcome reporting was unclear, so not sure if results were reported for a reference wound for each participant or if outcome data from multiple wounds were collected Cosmetic appearance data could not be used in the review as measured at < 3 months Wound complications reported as an outcome - but nature of the complications was not

clear

Risk of bias

Bias Authors’ judgement Support for judgement

Random sequence generation (selection Unclear risk Quote: “Randomization was performed by bias) asking the patient to select an envelope out of a hat” Comment: although it seems efforts were in place to randomise patients, it is not wholly clear that a truly randomised sequence was generated

Allocation concealment (selection bias) Unclear risk Quote:“Randomization was performed by asking the patient to select an envelope out of a hat” Comment: there was no mention of whether allocation was concealed to the investigators

Blinding of participants and personnel Unclear risk No direct quotations about whether the (performance bias) participants or personnel were blinded in All outcomes this study. It would be difﬁcult to blind personnel in a surgical procedure

Blinding of outcome assessment (detection Unclear risk No direct quotations about whether the bias) participants or personnel were blinded in All outcomes this study. It would be difﬁcult to blind personnel in a surgical procedure. The only outcome reported here relevant to the review was time to skin closure. It is not clear if this was assessed by a blinded assessor

Incomplete outcome data (attrition bias) Low risk Quote: “All patients were followed up. In the All outcomes suture group one patient refused to have a photograph taken” Comment: adequate outcome data gath- ered

Other bias Unclear risk Based on information collected it is not clear whether data were presented for an incorrectly analysed cluster trial

Methods RCT with 7-month follow-up. 3 participants lost to follow-up: 1 from the suture group and 2 from the tissue adhesive group

Participants 43 people requiring groin incisions (for: inguinal hernia, femoral hernia, sapheno liga- tions, testicular operations and lymph node biopsies). Skin incisions were closed with either butylcyanoacrylate (Histoacryl) tissue adhesive or Dexon subcuticular suture. In bilateral operations the left side was closed with Histoacryl and the left with Dexon. Conducted at Burton General Hospital, Burton-on-Trent, UK

Interventions Group 1: butylcyanoacrylate (Histoacryl) tissue adhesive Group 2: Dexon suture on straight needle using anchoring knot both ends or opposing the wound with forceps

Outcomes Infection, inﬂammation, cosmetic appearance, wound closing time, wound comfort and haematoma

Notes Exclusion criteria were not stated. Cosmetic appearance data not used in the review, as measured at < 3 months

Risk of bias

Bias Authors’ judgement Support for judgement

Random sequence generation (selection Unclear risk Quote:“The patients were randomised just bias) prior to skin closure into two groups. Even numbers were closed with Dexon subcuticular suture (Dexon group) and odd numbers were closed with Histoacryl-Blue tissue adhesive (Histoacryl)” Comment: whilst use of odd and even numbers is detailed in terms of how the randomisation was implemented there was no detail about how the sequence was generated

Allocation concealment (selection bias) Unclear risk Quote:“Even numbers were closed with Dexon subcuticular suture (Dexon group) and odd numbers were closed with Histoacryl- Blue tissue adhesive (Histoacryl)” Comment: no indication mention that this sequence was concealed from surgeons

Blinding of participants and personnel Unclear risk No direct quotation. Not reported whether (performance bias) participants or personnel were blinded to All outcomes the intervention. It is possible that participants may have been aware of 2 different methods of closure. “When the patients had bilateral operations, the left side

was closed with Histoacryl and the right side with Dexon”. Comment: understandably difﬁcult to blind operating surgeon to intervention

Blinding of outcome assessment (detection Unclear risk Quote: “Cosmesis was assessed by an indepen- bias) dent observer (the experienced clinic sister) on All outcomes a scale of one to ﬁve . . . ” “The patients and wounds were assessed at one week and one month postoperatively in the surgical outpatients clinic. A simple scheme was used to assess the wound . . .” Comment: blinding of outcome assessment achieved for wound appearance, but this was not used in this review. Blinding not clear with regard to the other outcomes

Incomplete outcome data (attrition bias) Low risk Comment: 3/43 participants were lost to All outcomes follow-up. Considered low risk of bias

Selective reporting (reporting bias) Low risk All the variables outlined in the methodology were accounted for in the results Comment: judged as low risk. Study protocol not sought

Other bias Low risk No other sources of bias detected

Kent 2014

Methods Parallel RCT in 433 participants with 3 months follow-up

Participants 433 participants undergoing a range of laparoscopic procedures Exclusion criteria: known sensitivity to cyanoacrylates; pregnancy or breastfeeding; or conditions known to interfere with wound healing (no further information provided) 4 UK centres

Interventions Group 1 (216 participants; 636 treated wounds): high viscosity 2-octylcyanoacrylate (High Viscocity Dermabond®; Ethicon Inc, a Johnson & Johnson company, Somerville, New Jersey, USA) tissue adhesive Group 2 (217 participants; 618 treated wounds): n-butyl-cyanoacrylate (Liquiband®) tissue adhesive (LiquiBand. MedLogic Global Ltd, Plymouth, Devon, UK)

Outcomes Wound dehiscence (evaluated at 14 days and 3 months postoperatively) Wound infection (evaluated at 14 days and 3 months postoperatively; infection not defined) Cosmetic appearance by surgeon (at 3 months) using a modified Hollander Wound Evaluation scale Participants’ satisfaction with incisional wound closure - options were ’satisfied’ or ’dis-

satisfied’ Sugeons’ satisfaction with wound (considering expression, application, delivery and ease of use for product): options were ’satisfied’ or ’dissatisfied’ Skin closure time

Notes Unit of randomisation was person with some data reported per wound (with multiple port incisions on each participant) Cosmetic appearance by participants and surgeon (at 3 months). Participants and surgeons were asked to assess whether they were ’satisﬁed’ or ’dissatisﬁed’ with the overall appearance of the wound. Whilst this was referred to as satisfaction in the study we deemed it to be an unvalidated measure of cosmetic appearance

Risk of bias

Bias Authors’ judgement Support for judgement

Random sequence generation (selection Unclear risk Quote: ”Subjects were randomised into 1 of bias) 2 treatment groups: LB or DB. Although the adhesive user was not masked, both the study subject and evaluators were blinded to the randomised study treatment assignment“ Comment. it is unclear how the randomisation sequence was achieved. Judged as unclear

Allocation concealment (selection bias) Unclear risk Quote:”Subjects were randomised into 1 of 2 treatment groups: LB or DB. Although the adhesive user was not masked, both the study subject and evaluators were blinded to the randomised study treatment assignment“ Comment: no method of allocation concealment described, therefore judged as unclear

Blinding of participants and personnel Unclear risk Quote: ”The adhesive user was not masked, (performance bias) both the study subject and evaluators were All outcomes blinded to the randomised study treatment assignment“ Comment: this indicates that personnel were not blinded to the procedure, whilst participants were blinded. Given that this was a study investigating two adhesives, blinding might have been possible. The judgement is unclear

Blinding of outcome assessment (detection Unclear risk Quote: ”Wound cosmesis was evaluated by a bias) masked panel at the 3-month visit. Cosme- All outcomes sis was measured using a modiﬁed 6-point Hollander Wound Evaluation Scale (HWES)

score.“ Comment: adequate blinding of outcome assessment Quote: Satisfaction measures were obtained from different study participants throughout the course of this trial. First, the surgical user was asked to indicate his/her satisfaction with the expression, application, delivery as ease of use of products. The user marked on a sheet that he/she was either ”satisfied“ or ”dissatisfied“. At the 3-month follow up visit, the masked evaluators were requested to state whether they were satisfied with the healing and the overall appearance of the incisions” Comment: satisfaction outcome blinded adequately Not clear if wound dehiscence and infection were blinded

Incomplete outcome data (attrition bias) High risk Quote:“A total of 76 subjects withdrew from All outcomes the study, primarily because they were lost to follow up, resulting in a loss to follow up rate of 17.6%. Final data analysis was performed on 373 subjects and a total of 1089 incisions” Comment: high number of missing data at the participant level

Selective reporting (reporting bias) Low risk All the variables outlined in the methodology were accounted for in the results Comment: judged as low risk. Study protocol not sought

Other bias Unclear risk Data presented and analysed as an individual trial, but this is a cluster trial (wounds clustered per person)

Khan 2006

Methods 3-arm parallel trial with 187 participants. No withdrawals reported Follow-up was classed as ’early’, which seems to have been the ﬁrst 3 days after surgery, and then ’late’ which was between 8 and 12 weeks after surgery

Participants 187 participants undergoing either a total knee arthroplasty or a total hip arthroplasty Exclusion criteria: having a revision or with a previous incision in the operative ﬁeld; a history of keloid formation; allergy to superglue; regular anticoagulation therapy; or an underlying malignancy The study was performed in 1 centre in Australia

Interventions Group 1 (n = 60): 2-octylcyanoacrylate (Dermabond®) tissue adhesive (Ethicon Inc, a Johnson & Johnson company, Somerville, New Jersey, USA) Group 2 (n = 64): continuous 3.0 subarticular absorbable poliglecaprone suture (Monocryl, Johnson and Johnson) Group 3 (n = 63): skin staples

Outcomes Wound infection: (report states that“where cultures were positive or there was clinical evidence of cellulitis, the patients were treated with a course of antibiotics and recorded as having an ‘infection”) data collected at 2 time points - early and late Patient satisfaction with the techniques of skin closure was assessed with a VAS between 0 and 100, where 100 represented maximal satisfaction Skin closure time

Notes Cosmetic appearance data (surgeon) could not be used in this review as measured at < 3 months

Risk of bias

Bias Authors’ judgement Support for judgement

Random sequence generation (selection Low risk Quote: “The patients were randomised using bias) a computer-generated method . . . ” Comment: Adequate random sequence generation method utilised

Allocation concealment (selection bias) Unclear risk Quote: “The patients were randomised using a computer-generated method stored in sealed identical opaque envelopes. Allocation took place in the operating theatre after closure of the deep layers. Enrolment, generation of the allocation sequence and assignment of the patients was performed by the lead author.” Comment: not clear if envelopes were numbered or otherwise labelled or stored so that allocation was concealed

Blinding of participants and personnel Unclear risk Quote: “The patients and assessors remained (performance bias) blinded to the treatment allocated until the All outcomes dressings were changed, prior to discharge. At follow-up, the assessors were not informed of the technique of closure.” Comment: no mention of whether the personnel were blinded to the procedure, although this would be understandably dif- ﬁcult for a surgical procedure

Blinding of outcome assessment (detection Unclear risk Quote: “The patients and assessors remained bias) blinded to the treatment allocated until the All outcomes dressings were changed, prior to discharge. At

follow-up, the assessors were not informed of the technique of closure.”

Incomplete outcome data (attrition bias) Low risk No direct quotation, but no account of a All outcomes loss to follow-up, therefore judged as low risk

Selective reporting (reporting bias) Low risk No direct quotation, but no outcomes selectively reported, therefore judged as adequate

Other bias Unclear risk Possible issues with clustering for infection outcome data

Kouba 2011

Methods Split-body design RCT with 12 weeks of follow-up. Data missing for 1 participant

Participants Adults undergoing upper lid blepharoplasty who had not previously undergone this procedure Participants were recruited from a singe site, Henry Ford Health System, USA 1 cosmetic surgeon performed whole procedure Exclusion criteria: taking salicylates and/or anticoagulants; taking oral retinoids in the last 3 months; an unexplained history of excessive bleeding; a history of acute glaucoma or Sjogren’s syndrome; having resurfacing or laser techniques for the eyelid

Interventions Group 1 (24 participants; 24 eyes): 2-octylcyanoacrylate (Dermabond®) tissue adhesive (Ethicon Inc, a Johnson & Johnson company, Somerville, New Jersey, USA) Group 2 (24 participants; 24 eyes): standard monoﬁlament suture (6-0 fast-absorbing gut or polypropylene) The study authors randomised each participant to have either (1) one eye treated with adhesive and the other with gut sutures; or (2) one eye treated with adhesive and the other with polypropylene sutures; or (3) one eye treated with gut sutures and the other treated with polypropylene sutures. For this review the adhesive data and the suture data from the two separate ’sub-studies’ involving adhesives have been pooled together and treated as one study

Outcomes Wound dehiscence: assessed at 1 week Cosmetic appearance (surgeon-rated): assessed at 1 month and 3 months (we report 3- month data). Scoring performed using a scale of 1 (excellent wound healing, scar matches surrounding skin) to 5 (poor scar wound healing, does not match surrounding skin). This was a blinded assessment. The data presented here were calculated by the review authors from raw data presented in the study report

Notes The following outcome was reported, but was not thought to be a validated measure for cosmetic appearance, “Participant cosmetic evaluation: (thickness, width, texture, color change, overall cosmetic outcome) was assessed at 3 months. A composite score was calculated as the sum of the scores for thickness, width, texture, and color change.”

No clear information was reported on wound dehiscence No comparative baseline data reported Funding source not reported

Risk of bias

Bias Authors’ judgement Support for judgement

Random sequence generation (selection Unclear risk Quote:“They [the patients] were randomised bias) for treatment in three subgroups in which each eyelid was repaired with a different material” Comment: no indication that there was randomised sequence generation

Allocation concealment (selection bias) Unclear risk Quote: “They [the patients] were randomised for treatment in three subgroups in which each eyelid was repaired with a different material” Comment: no indication allocation was concealed to the personnel

Blinding of participants and personnel Unclear risk No direct quotation, but no indication that (performance bias) there was blinding of the participant or the All outcomes personnel during the procedure. However, there are obvious difﬁculties in blinding the operating surgeon to the procedure

Blinding of outcome assessment (detection Unclear risk Quote: “Incidence of wound dehiscence and bias) side effects of itching, bleeding, and pain were All outcomes assessed at 1 week” Comment: no indication that this wound evaluation was undertaken by a blinded assessor Quote. [At 3 months] “A blinded physician assessed cosmetic outcome of wound closure technique using standardized photographs” Comment: adequate blinding of this outcome assessment

Incomplete outcome data (attrition bias) Unclear risk No direct information regarding drop out All outcomes rates, however different numbers used in total participants at 1 month to those at 3 months. Therefore the judgement is unclear

Selective reporting (reporting bias) High risk Quote: “Incidence of wound dehiscence and side effects of itching, bleeding, and pain were assessed at 1 week” Comment: incidence of wound dehiscence not clearly reported in the results

Other bias Low risk None reported

Krishnamoorthy 2009

Methods A parallel RCT with follow-up at 7 days and 6 weeks postoperatively. No missing data reported, but the number included in analysis was not clear

Participants 106 participants undergoing saphenous vein harvesting (for coronary artery bypass graft- ing). Participants were recruited from a single UK site Excluded patients with high risk of vein harvest failure (deﬁned as those with varicose veins, those with small or thin legs and those who required emergency or repeat procedures)

Interventions Group 1 (n = 53): 2-octylcyanoacrylate (Dermabond®) tissue adhesive (Ethicon Inc, a Johnson & Johnson company, Somerville, New Jersey, USA) Group 2 (n = 53): sutures (monoﬁlament synthetic absorbable Biosyn 3-0; Covidien PLC, Dublin, Ireland) In the suture group the wound was dressed with Mepore dressing (Mölnlycke Health Care, Manchester, United Kingdom),and a pressure bandage was applied for 48 h In the adhesive group a pressure bandage and Steri-Strips (3M, St Paul,MN) were applied to hold the edges together for 24 h

Outcomes Skin closure time

Notes Cosmetic appearance (surgeon-reported) data not used in this review as measured at < 3 months An outcome that study authors referred to as a ’patient satisfaction score’ was also measured, however, this seemed to focus on satisfaction of cosmetic appearance so was deemed a cosmetic evaluation; as it was collected at 6 weeks after surgery it is not reported here

Risk of bias

Bias Authors’ judgement Support for judgement

Random sequence generation (selection Low risk Quote: “A computerized randomization sys- bias) tem was used to place patients into two groups of 53 each.” Comment: judged as adequate evidence of sequence randomisation

Allocation concealment (selection bias) Unclear risk Quote: “A computerized randomization system was used to place patients into two groups of 53 each.” Comment: no indication that allocation was concealed from personnel

Blinding of participants and personnel Unclear risk No direct quotation or indication that ei- (performance bias) ther participants or personnel were blinded All outcomes from the study. However it is understandably difﬁcult for personnel to be blinded in this case as it is a surgical procedure

Blinding of outcome assessment (detection Unclear risk Quote:“Two surgeons independent of the bias) study who were blinded to the type of skin clo- All outcomes sure rated photographs at the same time points using a visual analogue scale evaluating cosmetic appearance and the previously validated Hollander wound evaluation scale.” Comment: although this would correlate to a perceived low risk of bias, the cosmesis outcome should be excluded at less than 3 months as per the systematic review protocol. No direct quotation available regarding blinding related to the time to skin closure study outcome

Incomplete outcome data (attrition bias) Unclear risk Quote “We identiﬁed and excluded as addi- All outcomes tional 12 patients . . . ” Comment: this quote is presented in the results. It is not clear if these were post- randomisation exclusions

Selective reporting (reporting bias) Low risk All the variables outlined in the methodology were accounted for in the results Comment: judged as low risk. Study protocol not sought

Other bias Low risk None detected

Livesey 2009

Methods RCT with a 3-month follow-up period. In total 13 participants did not contribute data at 3 months

Participants 90 people having a total hip replacement Exclusion criteria: revision total hip replacement; a previous incision in the operative ﬁeld, local skin conditions such as psoriasis, eczema or dermatitis; history of keloid formation; underlying malignancy; peripheral vascular disease; insulin-dependant diabetes; allergy to skin adhesive or metal staples Undertaken in 1 UK centre

Interventions Group 1 (n = 45): butylcyanoacrylate (LiquiBand®. MedLogic Global Ltd, Plymouth, Devon, UK) Group 2 (n = 45): staples (appose UCL 35W, Tyco, Norfolk, Connecticut)

Data at 3 months only available for 38 in Group 1 and 39 in Group 2 thanks to missing data

Outcomes Wound infection: deﬁned as participant requiring antibiotics speciﬁcally for suspected wound infection Cosmetic appearance (participant-rated) at 3 months: used a 100 mm VAS where 0 = worst outcome and 100 = best outcome. Median rather than mean data presented for this outcome Cosemtic appearance (surgeon-rated) at 3 months: used a 100 mm VAS where 0 = worst outcome and 100 = best outcome

Notes Data on participant satisfaction with scar and appearance of the wound in relation to expected appearance were also collected using a VAS scale (as above). Reviewers considered this to be a variation of cosmetic appearance and so it is not reported in the review Surgeons also reported cosmetic appearance using modified version of the Hollander wound evaluation score and modified Vancover scar score. The modified versions were deemed to be not validated (they only included 3 items) Time to wound closure data and ease of wound closure were collected on a sub-set of 10 participants in each trial arm. There was no information about how these sub-sets were selected, so these data could not be considered as data from an RCT per se and have not been presented here Follow-up appointments took place a mean of 14.4 weeks after surgery (no evidence of significant difference between groups)

Risk of bias

Bias Authors’ judgement Support for judgement

Random sequence generation (selection Unclear risk Quote:“Randomisation was performed, us- bias) ing the sealed envelope method, which involved identical sealed envelopes containing a card stating either ’skin adhesive’ or ’staple’ being opened by an independentresearcher on the day of surgery” Comment: no indication that adequate randomised sequence generation was performed

Allocation concealment (selection bias) Low risk Quote: “Randomisation was performed, using the sealed envelope method, which involved identical sealed envelopes containing a card stating either ’skin adhesive’ or ’staple’ being opened by an independentresearcher on the day of surgery. The operating surgeon was blinded to the skin closure method until the patient was in theatre” Comment: although it was not indicated that the envelopes were numbered sequen-

tially, there is reasonable evidence here to describe the allocation concealment as adequate, and as it was undertaken by an independent researcher it has been judged ac- cordingly

Blinding of participants and personnel Unclear risk Quote: “All patients had the same post-oper- (performance bias) ative care pathways and were blinded to the All outcomes method of skin closure until the dressings were changed post-operatively” Comment: adequate blinding of participants, however study design unable to blind personnel effectively, as a surgical intervention required

Blinding of outcome assessment (detection Unclear risk Quote: “A researcher collected information bias) on the presence of oozing and wound infec- All outcomes tion, which was deﬁned as the patient requiring antibiotics speciﬁcally for suspected wound infection. Patients whose wound was closed with staples had these removed between ten and 14 days post-operatively” Comment: indication that those evaluating the wounds were not blinded to the procedure, however this is not explicit, therefore the judgment remains unclear Quote: “An orthopaedic surgeon (AWB) also evaluated the scars using the VAS for cosmetic appearance. Both the plastic and orthopaedic surgeon were blinded to which method of skin closure had been used to each patient, and to each others’ scores” Comment: adequate blinding of cosmesis outcome

Incomplete outcome data (attrition bias) Unclear risk Quote: “In all, 12 patients (of 90) were All outcomes lost to follow-up because of non-attendance or cancellation of the three-month outpatient appointment. These patients have been seen subsequently with no reported adverse occur- rences, but were not included in the ﬁnal analysis as the time point for the last follow-up in the study was three months” Comment: impact of these missing data unclear

Selective reporting (reporting bias) High risk Quote: “Evaluation of the cosmetic appearance of the scars was completed by a plastic surgeon (CME) using a modiﬁed version of

the Hollander wound evaluation score and the Vancouver scar score, a Likert Scale and a VAS”. . Comment: The results section reports only the Vancouver scar score, the Likert Scale and the VAS . The Hollander wound evaluation score has been omitted the judgement here is of high risk of bias

Other bias Low risk None reported

Maartense 2002

Methods Randomised parallel group study with 16-month follow-up. There were no withdrawals, however 7 patients treated with paper tape and 3 with tissue adhesive were converted to the suture group

Participants 140 adults requiring elective laparoscopic surgery. Patients were excluded if they had undergone previous laparotomy or were pregnant. The study was undertaken at 2 centres, Department of Surgery, Academic Medical Centre, Amsterdam; and The Netherlands and Department of Surgery, Isala Clinics, Zwolle, the Netherlands

Interventions Group 1 (n = 48): octylcyanoacrylate (Dermabond®, Johnson & Johnson, Amersfoot, the Netherlands) tissue adhesive Group 2 (n = 42): 76 mm x 6 mm adhesive paper tape (SteriStrip® Bioplasty/Uroplasty, Geleen, the Netherlands) Group 3 (n = 50): intracutaneous poliglecaprone (Monocryl®) 4/0, Johnson & Johnson) interrupted sutures

Outcomes Infection, cosmetic appearance, and surgeon satisfaction at 2 weeks and 3 months, and costs

Notes We wrote to the authors who conﬁrmed that there were no withdrawals

Risk of bias

Bias Authors’ judgement Support for judgement

Random sequence generation (selection Low risk Quote: “Patients were allocated to one of the bias) three groups using a computer randomization” Comment: random sequence generation undertaken by computer

Allocation concealment (selection bias) Unclear risk Quote: “Patients were allocated to one of the three groups using a computer randomization” Comment: not reported whether this allocation was concealed from the operating surgeon

Blinding of outcome assessment (detection Low risk Quote: “At follow up [at 10-14 days and bias) 3 months], the incidence of wound infection All outcomes and cosmesis were scored. Patients were also asked to score their own cosmetic results. Sur- gical residents scored wound infection and cosmetic results, they were blinded to the method used for wound closure” Comment: this infers a low risk of detection bias at both intervals for outcomes assessed by surgeons

Incomplete outcome data (attrition bias) Low risk No direct relevant quotations, although All outcomes some patients who underwent closure with adhesive tape or adhesive liquid had to be converted intra-operatively to suture closure. These seem to be analysed in the groups to which they were randomised (intention-to-treat analysis) Comment: judgement of low risk of attrition bias

Selective reporting (reporting bias) Low risk No direct relevant quotations, but all variables listed in the methods section are listed in the tables in the Results section Comment: a low risk of reporting bias is inferred

Other bias Low risk No other sources of bias were detected

Maloney 2013

Methods Parallel RCT in 43 participants. Follow-up was for 3 months (assessment at 2 weeks and 3 months after surgery). No withdrawals noted

Participants 43 participants undergoing a skin incision to remove skin cancer Exclusion criteria: known sensitivity to cyanoacrylates; wounds under high tension forces; wounds < 2 cm or > 5 cm in length; or participant pregnant or breastfeeding Study undertaken in USA, number of centres involved was unclear

Interventions Group 1 (n = 23): high viscosity 2-octylcyanoacrylate (High Viscocity Dermabond®; Ethicon Inc, a Johnson & Johnson company, Somerville, New Jersey, USA) tissue adhesive

Group 2 (n = 20): n-butyl-cyanoacrylate (Liquiband®) tissue adhesive (LiquiBand. MedLogic Global Ltd, Plymouth, Devon, UK)

Outcomes Wound dehiscence (2 weeks and 3 months after surgery - no events reported at 3 months and 1 event only at 2 weeks, no unit of analysis issue) Wound infection (not defined) Cosmetic appearance at 3 months (blinded evaluators - panel of 6 doctors) using 100 mm VAS scale where 0 = worst scar and 10 = best scar Cosmetic appearance at 3 months (participant-reported) using VAS scale on worst scar (0) to best scar (10) Surgeon satisfaction (at time of wound closure) assessed using 100 mm VAS scales for: ease of use (0 = impossible; 10 = very easy to use) and satisfaction with device and the closure achieved (0 = completely dissatisfied; 10 = completely satisfied)

Notes Wound infection listed as outcome, but not clearly reported Cosmetic appearance assessed by blinded evaluator was also measured using Modifed Hollander Wound Evaluation Scale (modiﬁed to measure 5 not 6 items). This was not deemed to be validated, and so reviewers did not present the data

Risk of bias

Bias Authors’ judgement Support for judgement

Random sequence generation (selection Unclear risk Quote: “Subjects were randomly assigned to bias) receive either LB or DB for wound closure. Randomization occurred immediately following placement of the subcuticular sutures” Comment: unclear how the randomised sequence was generated

Allocation concealment (selection bias) Unclear risk Quote: “Randomization occurred immediately following placement of the subcuticular sutures. Due to the fact that the devices are packaged differently, it was not possible for the surgeon to be masked from the knowledge of the randomised treatment assignments” Comment: no evidence of adequate concealment of allocation

Blinding of participants and personnel Unclear risk Quote: “Due to the fact that devices are pack- (performance bias) ageddifferently, it was not possible for the sur- All outcomes geon to be masked from the knowledge of the randomised treatment assignments, however, the study subjects and the follow up assessments were masked to the device being used” Comment: adequate evidence that participants were blinded to the intervention, but personnel were not blinded. This is under- standable due to the nature of the interven-

tion

Blinding of outcome assessment (detection Unclear risk Quote: “ . . . subjects were asked to return at bias) two weeks and three months post surgery at All outcomes which time any applicable wound complications (erythema, pain, infection, and dehiscence) were captured along with any reported adverse events” Comment: no indication that the 2 week assessment was undertaken by a blinded assessor Quote “ . . . the photos were independently evaluatedbyamaskedpanelforcosmesisusing the same 100mm VAS as well as a modiﬁed Hollander Wound Evaluation Scale (HWES) .” “The masked panel consisted of 6 physicians: 2 general dermatologists, 1 dermatologic surgeon, 1 oculoplastic surgeon, 1 facial plastic surgeon, and 1 plastic surgeon.” Comment: adequate blinding of assessment outcome of cosmetic appearance at 3 months

Incomplete outcome data (attrition bias) Low risk Quote: “A total of 43 subjects participated in All outcomes this trial completing all study visits including 2-week and 3-month follow up” Comment: adequate evidence that no loss to follow-up occurred

Selective reporting (reporting bias) Low risk All the variables outlined in the methodology were accounted for in the results Comment: judged as low risk. Study protocol not sought

Other bias Low risk None detected

Millan 2011

Methods RCT with 60 participants and 14 days of follow-up

Participants 60 people undergoing a skin biopsy with chronically inﬂamed skin Exclusion criteria: infection at location where biopsy was planned and known hypersensitivity to cyanoacrylate or other tissue adhesive Surgery was undertaken in 1 centre in Mexico

Interventions Group 1 (n = 30): 2-octylcyanoacrylate tissue adhesive (no further details) Group 2 (n = 30): monoﬁlament suture (no further details)

Outcomes Wound dehiscence (on days 5, 7, 10 and 14 after surgery) Skin closure time

Notes Data extraction based on English language abstract and partial translation of report text which was published in Spanish

Risk of bias

Bias Authors’ judgement Support for judgement

Random sequence generation (selection Unclear risk Comment: study described as randomised. bias) No further detail

Allocation concealment (selection bias) Unclear risk Comment: no detail reported

Blinding of participants and personnel Unclear risk Comment: no detail reported (performance bias) All outcomes

Blinding of outcome assessment (detection Unclear risk Comment: no detail reported bias) All outcomes

Incomplete outcome data (attrition bias) Unclear risk Comment: no detail reported All outcomes

Selective reporting (reporting bias) Low risk Comment: wound dehiscence was the only outcome listed in the methods and reported

Other bias Unclear risk Comment: not noted from the translation available

Mota 2009

Methods Parallel RCT with 100 women and 30-day follow-up Data on the primary outcome was not available for 2 participants in the adhesive arm and 1 participant in the suture arm

Participants 100 women undergoing mediolateral episiotomy after a vaginal delivery in the absence of any other perianal or vaginal lesions Exclusion criteria: existing local infections or lesions; body mass index 35 (kg/m2); severe pulmonary disease; collagen disease; known immunodeﬁciency; diabetes mellitus; or currently receiving immunosuppressive treatment The study was conducted in 1 hospital centre in Portugal

Interventions Group 1 (n = 53): 2-octylcyanoacrylate (Dermabond®) tissue adhesive (Ethicon Inc, a Johnson & Johnson company, Somerville, New Jersey, USA) Group 2 (n = 47): continuous subcuticular suture (rapidly absorbable polyglactin 910)

Outcomes Wound dehiscence (42 h-68 h post-partum)

Notes The main outcome measure for the paper was self-assessed perineal pain during the ﬁrst 30 days after delivery. This was not deemed to be an outcome relevant to the review given the outcomes listed Skin closure time not reported - only time for whole procedure

Risk of bias

Bias Authors’ judgement Support for judgement

Random sequence generation (selection Low risk Quote: “Allocation was decided by opening a bias) previously prepared, opaque, sealed envelope containing a 1:1 computer-generated random number, assigning the patient to one of the two arms” Comment: use of computer-generated randomisation adequate

Allocation concealment (selection bias) Unclear risk Quote: “Allocation was decided by opening a previously prepared, opaque, sealed envelope containing a 1:1 computer-generated random number, assigning the patient to one of the two arms” Comment: use of opaque, sealed envelopes but not noted whether numbered in such a way as to ensure maintenance of robust allocation concealment. Not clear if person who opened the envelopes was independent

Blinding of participants and personnel Unclear risk Quote: “Women were not informed of the (performance bias) technique that was used to close the perineal All outcomes skin throughout the study period.” “As both groups required stitching of other layers, it is believed that they were not able to see or feel thedifferencebetweenthetwotypesofperineal skin closure, at the time of repair or thereafter. ” Comment: participants were blinded. Un- derstandably difﬁcult to blind the operating surgeon to the intervention

Blinding of outcome assessment (detection High risk Quote: “Other outcomes evaluated were . bias) . . duration of surgical repair and perineal All outcomes wound complications observed at hospital discharge, 42-68 hours post -partum. This ob- servation was carried out in all cases by one

of two authors (RM, FC)” Comment: as the authors were aware of the intervention undertaken, the judgement is that this potentially carries a high risk of bias

Incomplete outcome data (attrition bias) Unclear risk Quote: “Intra-operative data was available All outcomes for all participants. Data from 42-68 hours post-op was missing for 2 participants in the adhesive arm and one in the suture arm.” Quote: “Eighty-six women returned the questionnaires, 49 (92%) from the skin adhesive arm and 37 (79%) from the subcuticular suture arm. This difference was very close to achieving statistical difference p= 0.05 . . .”. The results also indicated that the reason for loss of follow up was the participants “did not return questionnaire”, rather than “discontinued intervention” Comment: unclear risk of bias

Selective reporting (reporting bias) Low risk Quote: No direct quotations, but all variables listed in methods section were listed in the tables in the results section Comment: a low risk of reporting bias is inferred

Other bias Low risk None detected

Ong 2002

Methods RCT with 3-month follow-up. 50 withdrawals at 3-month follow-up

Participants 59 patients requiring unilateral or bi-lateral herniotomies at KK Womens’ and Childrens’ Hospital, Singapore

Interventions Group 1 (n = 26): 2-octylcyanoacrylate (Dermabond®) tissue adhesive (Ethicon Inc, a Johnson & Johnson company, Somerville, New Jersey, USA) Group 2 (n = 33): subcuticular polyglecaprone (Monocryl®) suture

Outcomes Infection, dehiscence, parent satisfaction and time for closure

Notes Cosmesis scores at 3 weeks not usable as too early and those at 3-months not usable due to large loss to follow-up. Viewed as a high risk of bias with 50 participants dropping out

Risk of bias

Bias Authors’ judgement Support for judgement

Random sequence generation (selection Low risk Quote: “All enrolled patients were allocated bias) to glue or suture by opening serial sealed envelopes prepared with computerised randomisation” Comment: computerised randomisation probably represents randomised sequence generation

Allocation concealment (selection bias) Low risk Quote: “All enrolled patients were allocated to glue or suture by opening serial sealed envelopes prepared with computerised randomisation” Comment: the use of sequential sealed envelopes would reduce risk of selection bias. We have taken serial to mean that envelopes were kept in a pre-arranged and transparent ﬁxed order and therefore judged this study to be at low risk of bias for this domain

Blinding of participants and personnel Unclear risk Not reported whether the patients or per- (performance bias) sonnel were blinded to the intervention All outcomes Comment: understandably difﬁcult to blind the operating surgeon to the procedure

Blinding of outcome assessment (detection Unclear risk Quote: “ . . . assessment was done by an in- bias) dependent, blinded observer (staff nurse) us- All outcomes ing a previously validated score [The Hollan- der score] . . . Parent satisfaction with wound cosmesis was recorded at the same time on a 100mm visual analogue scale (VAS)” Comment: reasonable to deduce that this represents a low risk of detection bias for nurse-assessed cosmetic outcome, but not necessarily participant-assessed outcomes. Not clear if other outcomes such as wound infection or dehiscence were collected via blinded assessment

Incomplete outcome data (attrition bias) High risk Quote: “We were unable to do a 3-month All outcomes wound assessment on most of the patients, as only 9 returned for late follow up” Comment: it is difﬁcult to draw any conclusions from the results of the long-term outcomes due to high rates of losses to follow-up. This leads to a high risk of attrition bias

Selective reporting (reporting bias) Low risk No direct quotations, but all variables listed in the methods are listed in the tables in the results section Comment: a low risk of reporting bias is inferred

Other bias Low risk No other sources of bias were detected

Ozturan 2001

Methods RCT with 3-month follow-up. There were no withdrawals

Participants 101 people requiring rhinoplasty or septorhinoplasty entered the study Exclusion criteria: a history of peripheral vascular disease; diabetes mellitus; a clotting disorder; keloid or hypertrophic scarring; or an allergy to cyanoacrylate or formaldehyde Conducted at Inonu University Hospital, Turkey

Interventions Group 1 (n = 34): butylcyanoacrylate (LiquiBand®. MedLogic Global Ltd, Plymouth, Devon, UK) tissue adhesive Group 2 (n = 67): 6.0 polypropylene sutures for columellar skin closure after the majority of the tension had been taken up using 5.0 chromic catgut

Outcomes Dehiscence and infection at 1 week Cosmesis at 3 months by blinded assessment of photographs using VAS and Hollander scale Time required for skin closure

Notes We wrote to the authors to clarify the numbers in each group randomised by coin toss and received conﬁrmation that the numbers were correct. We also received clariﬁcation that the standard deviations were presented after the means in the results section of the paper

Risk of bias

Bias Authors’ judgement Support for judgement

Random sequence generation (selection Unclear risk Quote: “Patients were randomly allocated by bias) coin toss” Comment: random sequence generated

Allocation concealment (selection bias) Unclear risk Quote: “Patients were randomly allocated by coin toss” Comment: no mention of allocation concealment from personnel

Blinding of participants and personnel Unclear risk No mention of blinding of participants or (performance bias) personnel All outcomes Comment: understandably difﬁcult to

blind the operating surgeon to the procedure

Blinding of outcome assessment (detection Unclear risk Quote: “All patients were examined once ev- bias) eryweek in theﬁrstpost-operativemonth,and All outcomes once a month for the second and third months. Columellar wounds were examined for infection, inﬂammation, dehiscence and scarring in addition to cosmetic and functional nasal evaluations”. “The three month basal view photograph of each subject were given to two otolaryngologic surgeons who were blinded to the method of repair of the columellar incision” Comment: 3-month cosmetic assessment blinded. Unclear whether assessments at prior time points were blinded

Incomplete outcome data (attrition bias) Low risk Quote: “One hundred and one patients . All outcomes . . were eligible for inclusion”; this ﬁgure corresponds with the 101 participants accounted for in the results table Comment: no evidence of loss to follow- up

Selective reporting (reporting bias) High risk No direct quotations, but all variables listed in methods are listed in the tables in the results section Comment: a low risk of reporting bias is inferred

Other bias Low risk No other sources of bias were identiﬁed

Pronio 2011

Methods RCT with follow up at 7 days, 15 days, 1 month, 3 months, 6 months and 12 months. There were no withdrawals

Participants 70 people who underwent thyroid surgery from November 2005 to May 2007 No inclusion/exclusion criteria were stated In all cases, participants underwent thyroidectomy following a cervical incision The study took place in Italy, although no further information is given regarding the exact location of the trial

Interventions Group 1 (n = 32): 2-octylcyanoacrylate (Dermabond, Ethicon Inc) Group 2 (n = 38): skin staples (Proximate, Ethicon Inc) Prophylactic antibiotics in the form of ceftriaxone were administered to both groups

Outcomes Wound dehiscence (7days) Wound infection (7 days; not deﬁned) Cosmetic appearance (assessments by participants at 3, 6 and 12 months relevant here) . Data were recorded using the Stony Brook scar evaluation scale composed of 5 dichotomous, evenly weighted categories. Scars are assigned 0-1 point for the presence or absence of a width greater than 2 mm at any point of the scar, a raised (or depressed) scar, a darker coloration than surrounding skin, any hatch or staples marks, an overall poor appearance, the total score ranging from 0 (worst) to 5 (best) Patient’s satisfaction with wound management (7 days). Participants were asked to rate their level of satisfaction with the early postoperative management of the wound (regarding the requirement of a return visit for medications, the possibility of washing oneself, the suture removal) using a numerical scale ranging from 0-10

Notes Study records how many closures were considered rapid (between 30 and 60 seconds). These data were not extracted as it was felt that they reported skin closure time in a way that could not be summarised meaningfully Participants were also asked to provide a score using a verbal rating response regarding their scar (at 7 days, 15 days and 3, 6, 9 and 12 months after surgery). Scores could range from 0 to 10: 0-4, poor; 5-6, mild; 7-8, good; and 9-10, excellent. These data were only presented categorically and are not extracted - instead the scar evaluation scale data are presented

Risk of bias

Bias Authors’ judgement Support for judgement

Random sequence generation (selection Unclear risk Quote: “Only when the surgical proce- bias) dure was completed, after the closure of the platysma, each patient was randomly assigned to the two treatment groups” Comment: no indication that there was evidence of randomised sequence allocation

Allocation concealment (selection bias) Unclear risk Quote: “Only when the surgical procedure was completed, after the closure of the platysma, each patient was randomly assigned to the two treatment groups” Comment: no indication of allocation concealment from personnel

Blinding of participants and personnel Unclear risk No direct quotations, but no indication (performance bias) that the participants or the personnel were All outcomes blinded to treatment. It is understandably difﬁcult to blind personnel to the intervention arm given during an operation

Blinding of outcome assessment (detection Unclear risk Quote: “In the follow-up over 7 and 15 bias) days, 1-month, 3, 6 and 12-months peri- All outcomes ods wound-healing process was monitored by clinical assessment and documented by digital photographs.” Comment: there was no indication that the clinical assessments were made by people who were blinded to the intervention. Therefore the risk of bias here is unclear

Incomplete outcome data (attrition bias) Low risk No direct quotations, but no losses to fol- All outcomes low-up encountered

Selective reporting (reporting bias) Low risk No direct quotations, although all methods of evaluating the interventions were accounted for in the results Comment: adequate outcome reporting

Other bias Low risk None detected

Ridgway 2007

Methods RCT with 6-week follow-up. 1 withdrawal but no explanation

Participants 30 participants recruited for cervicotomy for thyroid and para thyroid surgery at Scun- thorpe General Hospital, North Lincs, UK. 29 successfully randomised. No reference to inclusion/exclusion criteria

Interventions Group 1 (n = 14): 2-octylcyanoacrylate (Dermabond®) tissue adhesive (Ethicon Inc, a Johnson & Johnson company, Somerville, New Jersey, USA) Group 2 (n = 15): staples (Vivistat®)

Outcomes Time to closure, neck mobility, cosmesis and adverse events

Notes Cosmesis results could not be included as they were measured too early to be usable. No explanation of what characterised an adverse event or the occurrence of such outcomes was provided

Risk of bias

Bias Authors’ judgement Support for judgement

Random sequence generation (selection Unclear risk Quote: “Randomisation to glue or stapled bias) closure was performed following induction of general anaesthesia by random envelope allocation” Comment: method of random sequence allocation not stated

Allocation concealment (selection bias) Unclear risk Quote: “Randomisation to glue or stapled closure was performed following induction of general anaesthesia by random envelope allocation” Comment: It is not clear whether the envelope was sealed or ordered and marked to be sequential, therefore without further information, the judgement is of an unclear risk of selection bias

Blinding of participants and personnel Unclear risk No statement made about whether the per- (performance bias) sonnel were blinded to the procedure All outcomes Comment: understandably difﬁcult to blind the operating surgeon to the intervention

Blinding of outcome assessment (detection Unclear risk Quote: “Scar cosmesis assessment by both pa- bias) tient, surgeon and independent blinded asses- All outcomes sor at 6 weeks.” Comment: this assessment was not included in the review (as conducted at less than 3 months) no other information about blinded assessment of other outcomes provided

Incomplete outcome data (attrition bias) Unclear risk No direct quotations, although no losses All outcomes to follow-up were reported, and all participants accounted for in results Comment: judged as low risk

Selective reporting (reporting bias) Low risk No direct quotations although all methods of evaluating the interventions were accounted for in the results Comment: adequate outcome reporting

Other bias Low risk None detected

Romero 2011

Methods RCT with 90 days of follow-up (assessment at 10 and 90 days postoperatively). Assumed all participants’ wounds assessed at day 10. There were 6 withdrawals from 90 day follow- up (2 from adhesives and 4 from strips)

Participants 49 children undergoing laparoscopic appendectomy Exclusion criteria: presence of concomitant chronic diseases; immunosuppression; ma- lignancies; and conversion to laparotomy or intraoperative enlargement of incisions for intact specimen extractions Conducted in 1 centre in Germany

Interventions Group 1 (n = 23): 2-octylcyanoacrylate (Dermabond®) tissue adhesive (Dermabond, Ethicon, Sommerville, NJ) Group 2 (n = 24): standard adhesive strips ( strips applied in star-shaped manner) Postoperative antibiotic therapy with cefotaxime and metronidazole was administered intravenously

Outcomes Wound dehiscence (assessment at 10 and 90 days after surgery) Wound infection (deﬁned as abscess or redness > 3 mm perpendicular to incision) Cosmetic appearance (surgeon-rated; at day 10 and day 90 - day 90 data reported here) using a 100 mm VAS where 0 = best scar and 100 = worse scar and assessed by 2 blinded surgeons using macrophotos of each scar (3 photos per participant). Mean value of the 2 assessments has been taken Cosmetic appearance (participant-rated; at day 10 and day 90 - day 90 data reported here) assessed using a dichotomous question (wording of question not clear but results were reported as % expressing dissatisfaction with cosmetic result)

Notes

Risk of bias

Bias Authors’ judgement Support for judgement

Random sequence generation (selection Low risk Quote: “For randomization, a computer- bias) generated randomization pattern was implemented and patients were intraoperatively assigned to a procedure by means of a blind envelope system” Comment: adequate random sequence generation

Allocation concealment (selection bias) Unclear risk Quote: “For randomization, a computer- generated randomization pattern was implemented and patients were intraoperatively assigned to a procedure by means of a blind envelope system” Comment: although attempts made to conceal allocation of the intervention, it was not stated whether the envelopes were sealed sequentially, whether they were sealed, or whether they were opaque, therefore the judgement remains unclear

Blinding of participants and personnel Unclear risk No direct quotations, however it is under- (performance bias) standably difﬁcult to blind the operating All outcomes surgeon from the intervention, therefore the judgement is unclear

Blinding of outcome assessment (detection Unclear risk Quote: “Follow up examinations were car- bias) ried out on the 10th and 90th postoperative All outcomes day . . . Follow-up included clinical investigation and a questionnaire to assess satisfaction with the cosmetic results, postoperative pain at port sites, wound infections, wound dehiscence, and any other adverse events associated with the wound” Comment: no indication that the follow- up examination on the 10th postoperative day was undertaken by a blinded assessor Quote: [At 90 days] “This assessment was completed by 2 pediatric surgeons blinded to the method of wound repair” Comment: adequate blinding of cosmesis outcome at 3 months

Incomplete outcome data (attrition bias) Unclear risk Quote: “A total of 49 patients (24 Der- All outcomes mabondT M , 25 Steri-StripT M) were enrolled between August 2007 and August 2008”. “Photographs were taken from 21 patients in the Steri-StripT M Group and from 22 patients in the DermabondT M Group” Comment: there is a discrepancy concerning 6 participants that were unaccounted for the in the ﬁnal outcome assessment without clear reporting of why they were not included in the results. Therefore there is a risk of attrition bias, leaving the judgement unclear

Selective reporting (reporting bias) Low risk No direct quotations, although all methods of evaluating the interventions were accounted for in the results Comment: adequate outcome reporting

Other bias Low risk None detected

Sebesta 2004

Methods RCT with 2-week follow-up. No participants were lost to follow-up

Participants 59 participants were enrolled, in whom 228 trocar sites were closed following undergoing laparoscopic surgery performed by one surgeon in the department of Urology, Wilford Hall Medical Center, Texas, USA No inclusion/exclusion criteria were stated

Interventions Group 1 (30 participants; 118 incisions): 2-octylcyanoacrylate (Dermabond, Ethicon, Sommerville, NJ) Group 2 (29 participants; 110 incisions): subcuticular suture (4-0 absorbable sutures, either Vicryl or Monocryl) Data presented at participant, not wound, level The fascia of all sites > 1 cm were closed with absorbable suture. Wounds in both groups that did not closely approximate to 1 cm received interrupted, subcutaneous sutures Those who received subcuticular sutures had their wounds dressed with steri-strips, a 2 x 2xm gauze pad, and tape or a Tegaderm dressing No dressings were applied to the octylcyanoacrylate-closed wounds

Outcomes Wound dehiscence (2 weeks after surgery) Wound infection (2 weeks after surgery; not deﬁned) Mean closure time Cost per patient

Notes Cosmetic appearance measured at < 3 months so not included

Risk of bias

Bias Authors’ judgement Support for judgement

Random sequence generation (selection Unclear risk Quote: “All patients undergoing laparoscopic bias) surgery by one surgeon (JTB) were ran- domisedtoreceiveskinclosurewitheithersub- cuticular suture of octylcyanoacrylate” Comment: no further information given regarding how the randomised sequence was generated, and therefore the judgement is unclear

Allocation concealment (selection bias) Unclear risk Quote: “All patients undergoing laparoscopic surgery by one surgeon (JTB) were ran- domisedtoreceiveskinclosurewitheithersub- cuticular suture of octylcyanoacrylate” Comment: no further information given regarding how the allocation was concealed to the operating surgeon (if at all), therefore the judgement remains unclear

Blinding of participants and personnel Unclear risk No direct quotations, but no information (performance bias) given regarding blinding of participants or All outcomes personnel in this study. It is understandably difﬁcult to blind the operating surgeon to the intervention. Therefore the judgement here is unclear

Blinding of outcome assessment (detection Unclear risk Quote: “Patients were evaluated 2 weeks post- bias) operatively for evidence of infection, dehis- All outcomes cence, seroma, and general cosmetic appearance.” Comment: unclear whether this evaluation was undertaken by a blinded assessor

Incomplete outcome data (attrition bias) Low risk No direct quotations, but no reported loss All outcomes to follow-up, therefore judgment of low risk

Selective reporting (reporting bias) Unclear risk Quote: “Postoperative wound complications were similar for both groups. Five patients in the octylcyanoacrylate group had wound complications in 9 incisions. Two patients experienced skin separation in 5 incisions. One patient had a minor wound infection at one incision site treated with oral antibiotics. Two patients experienced small seromas at 2 incision sites, requiring incision opening and healing by secondary intention. We noted small seromas in 2 patients receiving suture closure. These healed by secondary intention after skin opening and drainage. Ta- ble 4 presents a summary of the results.” Comment: however the table shown does not compare rates of wound complication and the reporting of the outcomes is unclear, it is this author’s judgement therefore that the risk of bias here is unclear

Other bias Low risk None detected

Shamiyeh 2001

Methods RCT with 9-month follow-up. 2 participants were lost to follow-up from the suture group due to failure to attend and could not be traced by mail or phone

Participants 79 adults requiring varicose vein surgery on the leg. Trial conducted at Ludwick Boltz- mann Institure, Linz, Austria Exclusion criteria: a history of chronic venous insufﬁciency with dermatosclerosis; previous phlebectomies; or allergy to plaster or octylcyanoacrylate

Interventions Mullerian phlebectomy performed creating an average wound length of 5 mm. Used 5 min wound compression followed by skin closure with: Group 1 (n = 26): octylcyanoacrylate tissue adhesive Group 2 (n = 28): 5.0 monoﬁlament suture Group 3 (n = 25): tape

A small plaster was placed over each wound

Outcomes Wound dehiscence, infection at 10 days, and patient and surgeon satisfaction, and cosmetic appearance at 1 year and costs. Time required for incision closure was also recorded

Notes

Risk of bias

Bias Authors’ judgement Support for judgement

Random sequence generation (selection Low risk Quote:“Randomisation was done by a com- bias) puter [during] the morning of the operation day” Comment: this was judged to be an adequate method of generating the randomisation sequence i.e. ‘by computer’

Allocation concealment (selection bias) Unclear risk Quote:“Randomisation was done by a computer [during] the morning of the operation day”. Comment: there is no mention of the method used to conceal allocation

Blinding of participants and personnel Unclear risk No mention of blinding of participants or (performance bias) personnel All outcomes Comment: understandably difﬁcult to blind the operating surgeon to the procedure

Blinding of outcome assessment (detection Unclear risk Quote: “The follow up by the operating sur- bias) geon was performed[at] 10 days and 6 weeks All outcomes after the operation” and, “Eight to 14 months after the operation all patients had been investigated by one senior dermatologist who was blinded to the method of skin closure, the scars were examined for color, width, and cosmetic appearance . . .” Comment: blinded outcome assessment achieved for outcomes collected after 8 months. Not clear if blinding was implemented for earlier follow-up times when wound dehiscence and infection were assessed

Incomplete outcome data (attrition bias) Low risk Quote:“Only 2 of 79 patients were lost to fol- All outcomes low-up, resulting in 77 patients with postoperative control” Comment: no reasons reported for those

lost to follow-up, but as they only represented a small percentage of the total participants, this was judged overall to be at low risk of attrition bias

Selective reporting (reporting bias) Low risk No direct quotations, but all variables listed in methods are listed in the tables in the results section Comment: a low risk of reporting bias is inferred

Other bias Low risk No other sources of bias were identiﬁed

Sinha 2001

Methods RCT with 6-month follow-up. 6 participants were lost to follow-up: 5 in tissue adhesive and 1 in suture group

Participants 50 adults requiring hand or wrist surgery (carpal tunnel syndrome, trigger ﬁnger, De Quervain’s tenosynovitis, ganglions of wrist and hand, and cysts of ﬁngers) Exclusion criteria: requiring surgery for Dupuytren’s contracture; repeat surgery; a history of skin allergy or keloid formation; diabetes; or corticosteroid use Trial conducted at Monklands Hospital, Airdre, UK

Interventions Skin approximated with skin hooks then: Group 1 (n = 20): application of butylcyanoacrylate adhesive (Indermil), or Group 2 (n = 24): suturing with 4.0 monoﬁlament All cases had local anaesthetic inﬁltration with or without general anaesthesia

Outcomes Dehiscence, infection, cosmetic appearance at 10 days, 2 weeks and 6 weeks

Notes Cosmetic appearance data not used as assessed at < 3 months

Risk of bias

Bias Authors’ judgement Support for judgement

Random sequence generation (selection Unclear risk Quote: “Patients were randomised to wound bias) adhesive or suture on the basis of 50 previously prepared and sealed envelopes (25 of each)” Comment: whilst the trial was reported as randomised there was no information about how the sequence was generated

Allocation concealment (selection bias) Unclear risk Quote: “Patients were randomised to wound adhesive or suture on the basis of 50 previously prepared and sealed envelopes (25 of each)” Comment: judgement of unclear risk of

bias, as not clear if envelopes were sequentially numbered and opaque. Not clear who was responsible for allocation

Blinding of participants and personnel Unclear risk No information provided about blinding (performance bias) of participants or surgeons All outcomes Comment: it would be difﬁcult to blind the operating surgeon

Blinding of outcome assessment (detection Low risk Quote: “Patients were subsequently assessed bias) at 2 and 6 weeks post-surgery by a designated All outcomes tissue viability nurse who was blinded to the method of closure” Comment: outcome assessment blinded

Incomplete outcome data (attrition bias) High risk Quote: “Of the study 50 participants, six were All outcomes lost to follow up (ﬁve in the adhesive group and one in the suture group)” Comment: the sample size of the study is small, 6 participants lost represents over 20% of the total sample. No information reported on reasons for withdrawal

Other bias Low risk No other sources of bias were identiﬁed

Sniezek 2007

Methods RCT with 3-month follow-up. No reference to withdrawals

Participants 14 participants recruited for removal of basal cell carcinoma or squamous cell carcinoma of the head and neck using the Mohs technique at the Department of Dermatology at University Hospital Iowa

Interventions Split wound design, so both methods of closure were used for part of each wound Intervention 1: 2-octylcyanoacrylate (Dermabond®; Ethicon Inc, a Johnson & Johnson company, Somerville, New Jersey, USA) Intervention 2: polypropylene cuticular suture

Outcomes Dehiscence, infection and cosmesis

Notes As a split wound design the data were paired. The authors were contacted for the mean difference and the standard deviation of the cosmetic scores on a participant basis. No reply was received, therefore this information was not included in the review

Risk of bias

Bias Authors’ judgement Support for judgement

Random sequence generation (selection Low risk Quote: “Half of the surgical wound was ran- bias) domly selected (by coin toss) for epidermalap- proximation . . .” Comment: use of coin toss judged to con- stitute adequate random sequence generation

Allocation concealment (selection bias) Unclear risk Quote: “Half of the surgical wound was randomly selected (by coin toss) for epidermalap- proximation . . .” Comment: no information given about whether the personnel were concealed to the allocation

Blinding of participants and personnel Unclear risk No reporting of either participants or per- (performance bias) sonnel being blinded to the intervention All outcomes Comment: as the nature of the intervention was that the two methods of closure were placed together to close the same incision, there is potential that participants may have understood the method of closure. Similarly it is understandably difﬁcult to blind the operating surgeon to the intervention. The judgement is therefore unclear

Blinding of outcome assessment (detection Unclear risk Quote: “Patients returned in 7 days for suture bias) removal and were evaluated for early compli- All outcomes cations such as wound separation, or dehiscence, inﬂammation or infection. High res- olution photographs were obtained immediately post-operatively and 3 months postoperatively” and, “The primary outcome measure was scar cosmesis, evaluated by comparing both halves of the scar from the 3 month photographs by ﬁve blinded dermatologists” Comment: unclear whether assessment at day 7 was blinded. 3-month assessment of cosmesis was blinded

Incomplete outcome data (attrition bias) Low risk Quote: “All 14 patients completed all study All outcomes end points” Comment: judged as low risk.

Other bias Low risk No other sources of bias were identiﬁed

Switzer 2003

Methods RCT with 4-week follow-up. 2 withdrawals accounted for

Participants 45 participants recruited for elective repair of inguinal hernias at the Eisenhower Army Medical Centre, Fort Gordon, Georgia, USA Inclusion/exclusion criteria not reported

Interventions Group 1 (n = 24): 2-octylcyanoacrylate (Dermabond®) tissue adhesive (Ethicon Inc, a Johnson & Johnson company, Somerville, New Jersey, USA) Group 2 (n = 22): subcuticular polyglecaprone (Monocryl®) suture

Outcomes Dehiscence, infection, cosmesis and time taken to closure

Notes The cosmetic data were not included as the data were taken at < 3 months. The authors were contacted for the mean and standard deviations for the time to closure, but since no response has been received these data were not included

Risk of bias

Bias Authors’ judgement Support for judgement

Random sequence generation (selection Low risk Quote: “ . . . patients were then randomised bias) with the use of a computerized random number generator to receive either 2-octylcyanoacrylate tissue adhesive or subcuticular running 4-0 poliglecaprone 25 (Monocryl, Ethicon, Inc) skin closure” Comment: judgement of a low risk of selection bias due to use of computer to gen- erate random number sequence

Allocation concealment (selection bias) Unclear risk Quote: “ . . . patients were then randomised with the use of a computerized random number generator to receive either 2-octylcyanoacrylate tissue adhesive or subcuticular running 4-0 poliglecaprone 25 (Monocryl, Ethicon, Inc) skin closure” Comment: No speciﬁc mention of how sequence was allocated.

Blinding of outcome assessment (detection Unclear risk Quote: “Patients were monitored postopera- bias) tively for complications of their closures, and All outcomes all patients were scheduled for post-operative visits at 2 weeks and at 4 weeks” Comment: it is not reported whether those monitoring for post-operative complications were blinded to the intervention. Cosmesis assessment was blinded, but these data were not extracted here as taken at < 3 months after surgery

Incomplete outcome data (attrition bias) Low risk Not directly stated, but no participants All outcomes were reported to have been lost to follow- up and all were accounted for in results Comment: judged as low risk of attrition bias

Other bias Low risk None detected

Tierny 2009

Methods Split body design RCT with 3 months follow-up

Participants 8 participants undergoing surgery for non-melanomas skin cancer Exclusion criteria: on immuno-suppressive medication and/or recipients of organ trans- plants Undertaken in 1 centre in the USA

Interventions 8 participants, each having half of incisional wound randomised to each intervention Intervention 1 (8 half wounds): 2-octylcyanoacrylate (Dermabond®) tissue adhesive (Ethicon Inc, a Johnson & Johnson company, Somerville, New Jersey, USA) Intervention 2 (8 half wounds): rapid absorbing gut suture

Outcomes Wound dehiscence (1 week postoperatively) Wound infection (1 week postoperatively - no deﬁnition provided)

Notes Cosmetic appearance (surgeon) at 3 months and cosmetic appearance (participant) at 3 months were collected but the data are not clear and not presented here. The report suggests that outcome data were reported on a 1 to 4 point scale but some data are > 4 so this is unclear. Also unclear what the scale measured Patient reference data not extracted

Risk of bias

Bias Authors’ judgement Support for judgement

Random sequence generation (selection Unclear risk Quote: “Patients were randomised for epi- bias) dermal closure with one half of the wounds (chest (n=6) and upper extremities n=2)) with rapid absorbing gut structure and on the other half with 2-octylethylcyanoacrylate tissue adhesive” Comment: no information given about the method of randomisation

Allocation concealment (selection bias) Unclear risk Quote: “Patients were randomised for epidermal closure with one half of the wounds (chest (n=6) and upper extremities n=2)) with rapid absorbing gut structure and on the other half with 2-octylethylcyanoacrylate tissue adhesive” Comment: no information given about whether allocation was concealed from the participants

Blinding of participants and personnel Unclear risk Quote: “All wounds were closed by a single (performance bias) surgeon (DJK) using a linear, bilayered clo- All outcomes sure method . . .” Comment: no information given about whether the participants or the personnel were blinded to the intervention. However, it would be difﬁcult to blind personnel to a surgical intervention

Blinding of outcome assessment (detection Unclear risk Quote: “Patients were seen for evaluation bias) at postoperative visits at both 1 week and All outcomes 3 months after the procedure. Incidence of wound dehiscence and side effects of itching, bleeding, and pain were assessed at 1 week” Comment: unclear whether those evaluating the wounds at 1 week were blinded to the intervention

Incomplete outcome data (attrition bias) Low risk No direct quotations, although there was All outcomes no loss to follow-up reported within the study

Other bias Low risk None detected

Toriumi 1998

Methods RCT with 1-year follow-up. 11 participants were lost to follow-up, but the groups from which they came were not speciﬁed

Participants People over 1 year of age requiring elective surgery for benign skin lesions predominantly in face and neck Exclusion criteria: a history of signiﬁcant trauma; peripheral vascular disease; diabetes mellitus; blood clotting disorder; keloid or hypertrophy scarring; known allergy to cyanoacrylate or formaldehyde Trial conducted at University of Illinois, Chicago, USA

Interventions Incisions with and without subcutaneous sutures were randomised for closure with: Group 1: 2-octylcyanoacrylate, or Group 2: 5.0 or 6.0 nylon suture

Outcomes Dehiscence, infection, cosmesis and closure time

Notes Participants lost to follow up were not reported by group

Risk of bias

Bias Authors’ judgement Support for judgement

Random sequence generation (selection Unclear risk Quote: ”Using clinical indications as evalu- bias) ated by the surgeon, patients were assigned to one of the two treatment groups“. Later trial report mentions ”Patients were randomised ...“ Comment: no method of randomised sequence generation stated. The initial quotation suggests clinicians had some auton- omy in deciding which participants were to receive which treatment, which in turn represents a high risk of bias, but this is unclear

Allocation concealment (selection bias) Unclear risk Quote: ”Using clinical indications as evaluated by the surgeon, patients were assigned to one of the two treatment groups“ Comment: no mention of allocation concealment and possible evidence that surgeon was aware of allocation

Blinding of participants and personnel Unclear risk No mention of blinding of participants or (performance bias) personnel All outcomes Comment: due to the nature of the surgical procedure, it would be difﬁcult to blind either the operating surgeon or participants

Blinding of outcome assessment (detection Unclear risk Quote: ”At each postoperative visit, wounds bias) were examined for infection, inﬂammation, All outcomes wound dehiscence or separation, and scarring. “ and, ”At the 90 day follow up visit, the wounds were graded for cosmesis using the modiﬁed Hollander wound evaluation scale”, and, at one year, the patients were assessed by way of a standardised method of pho- tography of the wound, “The photographs were then given to two facial plastic surgeons that were unfamiliar with the study design, the purpose or the site of the surgical incision, or the type of treatment received” Comment: no indication of blinded outcome assessment for outcomes assessed prior to one year. Blinded outcome assessment of wound appearance at 12 month follow-up

Incomplete outcome data (attrition bias) Unclear risk No direct quotations given, but results All outcomes show that of the 100/111 (90%) patients enrolled in the study for their long term 1- year follow-up wound evaluation Comment: 10% loss of follow-up data at 12 months

Other bias Low risk No other sources of bias were identiﬁed

Methods 2-arm RCT with 6 week follow-up

Participants 100 children undergoing correction of hernia inguinalis No inclusion or exclusion criteria stated

Interventions Group 1 (n = 50): butylcyanoacrylate (Indermil) tissue adhesive Group 2 (n = 50): polyglactin 5-0 (Vicryl) sutures

Outcomes Wound dehiscence (10 days after surgery) Wound infection (not deﬁned; 10 days after surgery) Mean closure time

Notes It is noted that 100 participants were randomised and that some had surgery on both sides of the abdomen. It was not clear from the study results whether outcomes were presented at the participant level. In one instance where % data are presented the denominator used was 50 for each group suggesting that there is no unit of analysis issue but this is not clear Wound cosmesis assessed at 6 weeks - not presented here

Risk of bias

Bias Authors’ judgement Support for judgement

Random sequence generation (selection Unclear risk Quote: “Patients were selected randomly to receive bias) wound adhesive or suture on the basis of 100 previously prepared and sealed envelopes containing slips for either suture closure or the use of Indermil (50 of each)” Comment: use of opaque, sealed envelopes, but not noted whether numbered in such a way as to ensure robust allocation concealment maintained. Not clear if person who opened envelopes was independent

Allocation concealment (selection bias) Unclear risk Not reported

Blinding of participants and personnel Unclear risk No mention of blinding of participants or person- (performance bias) nel All outcomes Comment: due to the nature of the surgical procedure, it would be difﬁcult to blind either the operating surgeon

Blinding of outcome assessment (detection Unclear risk No mention of blinded outcome assessment bias) All outcomes

Incomplete outcome data (attrition bias) Low risk No mention of loss to follow-up All outcomes

Selective reporting (reporting bias) Low risk No direct quotations, but all variables listed in methods are listed in the tables in the results section Comment: a low risk of reporting bias is inferred

Other bias Unclear risk Comment: it is noted that 100 participants were randomised and that some had surgery on both sides of the abdomen. It is not clear from the study results if outcomes are presented at the participant level. Where % data is presented in one case the denominator used was 50 for each group suggesting that there is no unit of analysis issue but this is not clear

Abbreviations h = hour(s) min = minute(s) RCT = randomised controlled trial ROM = range of motion VAS = visual analogue scale

Characteristics of excluded studies [ordered by study ID]

Study Reason for exclusion

Ak 2012 Not relevant wound type (not surgical wound)

Alhopuro 1976 No data given in the paper. We have contacted authors but did not receive a reply

Chen 2010 Does not include any outcome relevant to this review

Chow 2010 Does not include any outcome relevant to this review

Giri 2004 Not an RCT

Gorozpe-Calvillo 1999 After full translation, found not to be an RCT

Jaibaji 2000 All participants in the intervention group had a precautionary subcuticular suture placed. The reviewers viewed this as an inherent bias to the outcome of this group and thus the study was excluded

Kuo 2006 All participants had a subcuticular suture placed and described the intervention as a method of “superﬁcial” closure. The reviewers viewed this as unnecessary and counter-intuitive as the aim of the study is to assess the adhesive as an alternative to other methods of closure of incisions, not as an adjunct

Matin 2003 Not an RCT - see quotation below - considered alternation Quote: “Randomization was performed weekly. Each odd numbered week of the study (ﬁrst week, third week, etc. was randomized to either OCA or suturing by a biostatistician, and then the subsequent even numbered week (second week, fourth week, etc) received the opposite treatment.”

Maw 1997 Not considered an RCT

Ong 2010 Skin closure method was not the only systematic difference between groups

Orozco-Razon 2002 There was no reference to randomisation. The authors have been contacted but we have received no reply

Quinn 1998 Not relevant wound type (not surgical wound)

Sajid 2009 Not an RCT

Silvestri 2006 Not an RCT

Singer 2002 Data could not be used because combined with data for lacerations. We wrote to the authors requesting data by group but did not receive a reply

Spencker 2011 Not an RCT

Steiner 2000 Not an RCT

Sun 2005 No relevant outcomes (based on translation)

Wong 2011 Not relevant wound type (not surgical wounds)

Abbreviation RCT = randomised controlled trial

Characteristics of studies awaiting assessment [ordered by study ID]

Gennari 2004

Methods

Participants

Interventions

Outcomes

Notes Citation retrieved through handsearching awaiting full text retrieval

Methods

Participants

Interventions

Outcomes

Notes Citation retrieved through handsearching awaiting full text retrieval

Jan 2013

Methods

Participants

Interventions

Outcomes

Notes Conference abstract with limited data - contacting authors to ﬁnd out if there is a pending publication or if further information available

Nipshagen 2008

Methods

Participants

Interventions

Outcomes

Notes Citation retrieved through handsearching awaiting full text retrieval

Yoon 2006

Methods Requires translation

Participants

Interventions

Outcomes

Notes

Comparison 1. Adhesive versus suture

No. of No. of Outcome or subgroup title studies participants Statistical method Effect size

1 Dehiscence: all studies 17 1225 Risk Ratio (M-H, Random, 95% CI) 3.35 [1.53, 7.33] 2 Dehiscence: sensitivity analysis 13 935 Risk Ratio (M-H, Random, 95% CI) 2.70 [0.95, 7.68] 3 Infection: all studies 18 1239 Risk Ratio (M-H, Random, 95% CI) 1.72 [0.94, 3.16] 4 Infection: sensitivity analysis 15 977 Risk Ratio (M-H, Random, 95% CI) 2.03 [0.80, 5.12] 5 Cosmetic appearance rated by 2 199 Mean Difference (IV, Random, 95% CI) -2.12 [-7.20, 2.95] patient 5.1 VAS 0 to 100 2 199 Mean Difference (IV, Random, 95% CI) -2.12 [-7.20, 2.95] 6 Cosmetic appearance rated by 2 Mean Difference (IV, Fixed, 95% CI) Totals not selected surgeon 6.1 VAS scale 0 to 100 1 Mean Difference (IV, Fixed, 95% CI) 0.0 [0.0, 0.0] 6.2 Scar assessment (0 to 5 1 Mean Difference (IV, Fixed, 95% CI) 0.0 [0.0, 0.0] scale) 7 Patient/parent satisfaction (% 2 206 Risk Ratio (M-H, Random, 95% CI) 1.01 [0.96, 1.07] satisﬁed) 8 Patient/parent satisfaction (VAS 1 Mean Difference (IV, Fixed, 95% CI) Subtotals only Scale 0 to 100) 9 Surgeon satisfaction (% satisﬁed) 2 150 Risk Ratio (M-H, Random, 95% CI) 1.12 [0.58, 2.19] 10 Time taken for wound closure 5 Mean Difference (IV, Random, 95% CI) Subtotals only

Comparison 2. Adhesive versus adhesive tape

No. of No. of Outcome or subgroup title studies participants Statistical method Effect size

1 Dehiscence 2 Risk Ratio (M-H, Random, 95% CI) Subtotals only 2 Infection 3 190 Risk Ratio (M-H, Random, 95% CI) 1.37 [0.39, 4.81] 3 Cosmetic appearance rated by 1 Mean Difference (IV, Fixed, 95% CI) Subtotals only patient (VAS) 4 Cosmetic appearance rated by 1 Risk Ratio (M-H, Fixed, 95% CI) Subtotals only patient (% satisﬁed) 5 Cosmetic appearance rated by 2 139 Mean Difference (IV, Random, 95% CI) 9.56 [4.74, 14.37] surgeon (VAS) 6 Patient satisfaction 1 Risk Ratio (M-H, Fixed, 95% CI) Subtotals only 7 Surgeon satisfaction 2 141 Risk Ratio (M-H, Random, 95% CI) 0.87 [0.63, 1.19] 8 Time taken for wound closure 1 Mean Difference (IV, Fixed, 95% CI) Subtotals only

No. of No. of Outcome or subgroup title studies participants Statistical method Effect size

1 Dehiscence 2 107 Risk Ratio (M-H, Random, 95% CI) 0.53 [0.05, 5.33] 2 Infection 4 320 Risk Ratio (M-H, Random, 95% CI) 1.39 [0.30, 6.54] 3 Cosmetic appearance rated by 1 Mean Difference (IV, Fixed, 95% CI) Subtotals only patient (scar scale) 4 Cosmetic appearance by plastic 1 Mean Difference (IV, Fixed, 95% CI) Subtotals only surgeons (VAS) 5 Patient satisfaction 1 Mean Difference (IV, Fixed, 95% CI) Subtotals only 6 Time taken for wound closure 1 Mean Difference (IV, Fixed, 95% CI) Subtotals only

Comparison 4. Adhesive versus other method

No. of No. of Outcome or subgroup title studies participants Statistical method Effect size

1 Dehiscence 2 249 Risk Ratio (M-H, Random, 95% CI) 0.55 [0.13, 2.38] 2 Infection 2 249 Risk Ratio (M-H, Random, 95% CI) 0.41 [0.11, 1.60] 3 Patient satisfaction 1 187 Mean Difference (IV, Fixed, 95% CI) 0.40 [0.10, 0.70] 4 Clinician satisfaction 1 209 Mean Difference (IV, Fixed, 95% CI) 0.53 [0.29, 0.77] 5 Time taken for wound closure 1 209 Mean Difference (IV, Fixed, 95% CI) -1.05 [-1.79, -0.31]

Comparison 5. Adhesive versus adhesive: High viscosity versus low viscosity

No. of No. of Outcome or subgroup title studies participants Statistical method Effect size

1 Dehiscence 1 Risk Ratio (M-H, Fixed, 95% CI) Subtotals only 2 Infection 1 Risk Ratio (M-H, Fixed, 95% CI) Subtotals only 3 Patient satisfaction 1 Mean Difference (IV, Fixed, 95% CI) Subtotals only 4 Clinician satisfaction 1 Mean Difference (IV, Fixed, 95% CI) Subtotals only 5 Time taken for wound closure 1 Mean Difference (IV, Fixed, 95% CI) Subtotals only

Tissue adhesives for closure of surgical incisions (Review) 96 Copyright © 2014 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Comparison 6. Adhesive versus adhesive: octylcyanoacrylate versus butylcyanoacrylate

No. of No. of Outcome or subgroup title studies participants Statistical method Effect size

1 Dehiscence 2 80 Risk Ratio (M-H, Random, 95% CI) 1.46 [0.19, 11.30] 2 Infection 2 Risk Ratio (M-H, Fixed, 95% CI) Subtotals only 3 Cosmetic assessment rated by 1 Mean Difference (IV, Fixed, 95% CI) Subtotals only patient (VAS) 4 Cosmetic assessment rated by 1 Mean Difference (IV, Fixed, 95% CI) Subtotals only surgeon (VAS) 5 Surgeon satisfaction (with 1 Mean Difference (IV, Fixed, 95% CI) Subtotals only device) 6 Surgeon satisfaction (with 1 Mean Difference (IV, Fixed, 95% CI) Subtotals only closure) 7 Time taken for wound closure 1 Mean Difference (IV, Fixed, 95% CI) Subtotals only

Analysis 1.1. Comparison 1 Adhesive versus suture, Outcome 1 Dehiscence: all studies.

Review: Tissue adhesives for closure of surgical incisions

Comparison: 1 Adhesive versus suture

Outcome: 1 Dehiscence: all studies

Studyorsubgroup Adhesive Suture RiskRatio Weight RiskRatio M- M- H,Random,95% H,Random,95% n/N n/N CI CI Brown 2009 0/64 0/70 Not estimable

Cheng 1997 2/40 0/46 6.8 % 5.73 [ 0.28, 115.97 ]

Dowson 2006 4/76 0/78 7.3 % 9.23 [ 0.51, 168.63 ]

Eggers 2011 1/19 1/19 8.5 % 1.00 [ 0.07, 14.85 ]

Greene 1999 0/20 0/20 Not estimable

Millan 2011 3/30 0/30 7.2 % 7.00 [ 0.38, 129.93 ]

Mota 2009 3/51 2/46 20.2 % 1.35 [ 0.24, 7.74 ]

Ong 2002 0/26 0/33 Not estimable

Ozturan 2001 0/34 0/67 Not estimable

Sebesta 2004 2/29 0/30 6.9 % 5.17 [ 0.26, 103.21 ]

Shamiyeh 2001 1/26 0/26 6.2 % 3.00 [ 0.13, 70.42 ]

0.002 0.1 1 10 500 Favours adhesive Favours suture (Continued ... )

Tissue adhesives for closure of surgical incisions (Review) 97 Copyright © 2014 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. (... Continued) Studyorsubgroup Adhesive Suture RiskRatio Weight RiskRatio M- M- H,Random,95% H,Random,95% n/N n/N CI CI Sinha 2001 3/20 2/24 21.6 % 1.80 [ 0.33, 9.73 ]

Sniezek 2007 0/14 0/14 Not estimable

Switzer 2003 4/24 0/22 7.5 % 8.28 [ 0.47, 145.50 ]

Tierny 2009 0/8 0/8 Not estimable

Toriumi 1998 0/54 0/57 Not estimable

van den Ende 2004 13/50 0/50 7.9 % 27.00 [ 1.65, 442.17 ] Total (95% CI) 585 640 100.0 % 3.35 [ 1.53, 7.33 ] Total events: 36 (Adhesive), 5 (Suture) Heterogeneity: Tau2 = 0.0; Chi2 = 6.52, df = 9 (P = 0.69); I2 =0.0% Test for overall effect: Z = 3.02 (P = 0.0025) Test for subgroup differences: Not applicable

0.002 0.1 1 10 500 Favours adhesive Favours suture

Tissue adhesives for closure of surgical incisions (Review) 98 Copyright © 2014 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Analysis 1.2. Comparison 1 Adhesive versus suture, Outcome 2 Dehiscence: sensitivity analysis.

Review: Tissue adhesives for closure of surgical incisions

Comparison: 1 Adhesive versus suture

Outcome: 2 Dehiscence: sensitivity analysis

Studyorsubgroup Adhesive Suture RiskRatio Weight RiskRatio M- M- H,Random,95% H,Random,95% n/N n/N CI CI Brown 2009 0/64 0/70 Not estimable

Dowson 2006 4/76 0/78 12.9 % 9.23 [ 0.51, 168.63 ]

Eggers 2011 1/19 1/19 15.0 % 1.00 [ 0.07, 14.85 ]

Greene 1999 0/20 0/20 Not estimable

Mota 2009 3/51 2/46 35.8 % 1.35 [ 0.24, 7.74 ]

Ong 2002 0/26 0/33 Not estimable

Ozturan 2001 0/34 0/67 Not estimable

Sebesta 2004 2/29 0/30 12.1 % 5.17 [ 0.26, 103.21 ]

Shamiyeh 2001 1/26 0/26 10.9 % 3.00 [ 0.13, 70.42 ]

Sniezek 2007 0/14 0/14 Not estimable

Switzer 2003 4/24 0/22 13.3 % 8.28 [ 0.47, 145.50 ]

Tierny 2009 0/8 0/8 Not estimable

Toriumi 1998 0/54 0/57 Not estimable Total (95% CI) 445 490 100.0 % 2.70 [ 0.95, 7.68 ] Total events: 15 (Adhesive), 3 (Suture) Heterogeneity: Tau2 = 0.0; Chi2 = 2.71, df = 5 (P = 0.74); I2 =0.0% Test for overall effect: Z = 1.87 (P = 0.062) Test for subgroup differences: Not applicable

0.01 0.1 1 10 100 Favours adhesive Favours sutures

Tissue adhesives for closure of surgical incisions (Review) 99 Copyright © 2014 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Analysis 1.3. Comparison 1 Adhesive versus suture, Outcome 3 Infection: all studies.

Review: Tissue adhesives for closure of surgical incisions

Comparison: 1 Adhesive versus suture

Outcome: 3 Infection: all studies

Studyorsubgroup Adhesive Suture RiskRatio Weight RiskRatio M- M- H,Random,95% H,Random,95% n/N n/N CI CI Avsar 2009 3/20 0/20 4.4 % 7.00 [ 0.38, 127.32 ]

Cheng 1997 1/40 0/46 3.7 % 3.44 [ 0.14, 82.13 ]

Dowson 2006 0/76 0/78 Not estimable

Eggers 2011 4/19 5/19 27.8 % 0.80 [ 0.25, 2.53 ]

Greene 1999 0/20 0/20 Not estimable

Keng 1989 0/21 0/22 Not estimable

Khan 2006 7/60 2/64 15.7 % 3.73 [ 0.81, 17.27 ]

Maartense 2002 5/48 3/50 19.5 % 1.74 [ 0.44, 6.87 ]

Ong 2002 0/26 0/33 Not estimable

Ozturan 2001 0/34 3/67 4.3 % 0.28 [ 0.01, 5.22 ]

Sebesta 2004 1/29 0/30 3.7 % 3.10 [ 0.13, 73.14 ]

Shamiyeh 2001 1/26 0/26 3.7 % 3.00 [ 0.13, 70.42 ]

Sinha 2001 0/20 0/24 Not estimable

Sniezek 2007 0/14 0/14 Not estimable

Switzer 2003 1/24 0/22 3.7 % 2.76 [ 0.12, 64.41 ]

Tierny 2009 0/8 0/8 Not estimable

Toriumi 1998 0/54 0/57 Not estimable

van den Ende 2004 4/50 2/50 13.5 % 2.00 [ 0.38, 10.43 ] Total (95% CI) 589 650 100.0 % 1.72 [ 0.94, 3.16 ] Total events: 27 (Adhesive), 15 (Suture) Heterogeneity: Tau2 = 0.0; Chi2 = 5.66, df = 9 (P = 0.77); I2 =0.0% Test for overall effect: Z = 1.75 (P = 0.080) Test for subgroup differences: Not applicable

0.002 0.1 1 10 500 Favours adhesive Favours suture

Tissue adhesives for closure of surgical incisions (Review) 100 Copyright © 2014 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Analysis 1.4. Comparison 1 Adhesive versus suture, Outcome 4 Infection: sensitivity analysis.

Review: Tissue adhesives for closure of surgical incisions

Comparison: 1 Adhesive versus suture

Outcome: 4 Infection: sensitivity analysis

Studyorsubgroup Adhesive Suture RiskRatio Weight RiskRatio M- M- H,Random,95% H,Random,95% n/N n/N CI CI Avsar 2009 3/20 0/20 10.2 % 7.00 [ 0.38, 127.32 ]

Cheng 1997 1/40 0/46 8.5 % 3.44 [ 0.14, 82.13 ]

Dowson 2006 0/76 0/78 Not estimable

Greene 1999 0/20 0/20 Not estimable

Keng 1989 0/21 0/22 Not estimable

Maartense 2002 5/48 3/50 45.4 % 1.74 [ 0.44, 6.87 ]

Ong 2002 0/26 0/33 Not estimable

Ozturan 2001 0/34 3/67 10.0 % 0.28 [ 0.01, 5.22 ]

Sebesta 2004 1/29 0/30 8.6 % 3.10 [ 0.13, 73.14 ]

Shamiyeh 2001 1/26 0/26 8.6 % 3.00 [ 0.13, 70.42 ]

Sinha 2001 0/20 0/24 Not estimable

Sniezek 2007 0/14 0/14 Not estimable

Switzer 2003 1/24 0/22 8.7 % 2.76 [ 0.12, 64.41 ]

Tierny 2009 0/8 0/8 Not estimable

Toriumi 1998 0/54 0/57 Not estimable Total (95% CI) 460 517 100.0 % 2.03 [ 0.80, 5.12 ] Total events: 12 (Adhesive), 6 (Suture) Heterogeneity: Tau2 = 0.0; Chi2 = 2.79, df = 6 (P = 0.83); I2 =0.0% Test for overall effect: Z = 1.49 (P = 0.14) Test for subgroup differences: Not applicable

0.01 0.1 1 10 100 Favours adhesive Favours sutures

Tissue adhesives for closure of surgical incisions (Review) 101 Copyright © 2014 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Analysis 1.5. Comparison 1 Adhesive versus suture, Outcome 5 Cosmetic appearance rated by patient.

Review: Tissue adhesives for closure of surgical incisions

Comparison: 1 Adhesive versus suture

Outcome: 5 Cosmetic appearance rated by patient

Mean Mean Studyorsubgroup Adhesive Suture Difference Weight Difference N Mean(SD) N Mean(SD) IV,Random,95% CI IV,Random,95% CI

1 VAS 0 to 100 Maartense 2002 48 76 (24) 50 78 (17) 37.7 % -2.00 [ -10.26, 6.26 ]

Ozturan 2001 34 70.3 (14.9) 67 72.5 (16.8) 62.3 % -2.20 [ -8.62, 4.22 ] Total (95% CI) 82 117 100.0 % -2.12 [ -7.20, 2.95 ] Heterogeneity: Tau2 = 0.0; Chi2 = 0.00, df = 1 (P = 0.97); I2 =0.0% Test for overall effect: Z = 0.82 (P = 0.41) Test for subgroup differences: Not applicable

-20 -10 0 10 20 Favours suture Favours adhesive

Analysis 1.6. Comparison 1 Adhesive versus suture, Outcome 6 Cosmetic appearance rated by surgeon.

Review: Tissue adhesives for closure of surgical incisions

Comparison: 1 Adhesive versus suture

Outcome: 6 Cosmetic appearance rated by surgeon

Mean Mean Studyorsubgroup Adhesive Suture Difference Difference N Mean(SD) N Mean(SD) IV,Fixed,95% CI IV,Fixed,95% CI

1 VAS scale 0 to 100 Maartense 2002 48 82 (20) 50 79 (10) 3.00 [ -3.30, 9.30 ]

2 Scar assessment (0 to 5 scale) Kouba 2011 24 1.26 (0.45) 24 1.52 (0.67) -0.26 [ -0.58, 0.06 ]

-10 -5 0 5 10 Favours suture Favours adhesive

Tissue adhesives for closure of surgical incisions (Review) 102 Copyright © 2014 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Analysis 1.7. Comparison 1 Adhesive versus suture, Outcome 7 Patient/parent satisfaction (% satisﬁed).

Review: Tissue adhesives for closure of surgical incisions

Comparison: 1 Adhesive versus suture

Outcome: 7 Patient/parent satisfaction (% satisﬁed)

Studyorsubgroup Adhesive Suture RiskRatio Weight RiskRatio M- M- H,Random,95% H,Random,95% n/N n/N CI CI Dowson 2006 76/76 76/78 76.6 % 1.03 [ 0.98, 1.07 ]

Shamiyeh 2001 25/26 26/26 23.4 % 0.96 [ 0.87, 1.07 ] Total (95% CI) 102 104 100.0 % 1.01 [ 0.96, 1.07 ] Total events: 101 (Adhesive), 102 (Suture) Heterogeneity: Tau2 = 0.00; Chi2 = 1.33, df = 1 (P = 0.25); I2 =25% Test for overall effect: Z = 0.37 (P = 0.71) Test for subgroup differences: Not applicable

0.5 0.7 1 1.5 2 Favours sutures Favours adhesive

Analysis 1.8. Comparison 1 Adhesive versus suture, Outcome 8 Patient/parent satisfaction (VAS Scale 0 to 100).

Review: Tissue adhesives for closure of surgical incisions

Comparison: 1 Adhesive versus suture

Outcome: 8 Patient/parent satisfaction (VAS Scale 0 to 100)

Mean Mean Studyorsubgroup Adhesive Suture Difference Weight Difference N Mean(SD) N Mean(SD) IV,Fixed,95% CI IV,Fixed,95% CI

Ong 2002 26 78 (19) 33 81 (15) -3.00 [ -11.92, 5.92 ] Subtotal (95% CI) 0 0 0.0 [ 0.0, 0.0 ] Heterogeneity: not applicable Test for overall effect: Z = 0.0 (P < 0.00001) Test for subgroup differences: Not applicable

-100 -50 0 50 100 Favours suture Favours adhesive

Tissue adhesives for closure of surgical incisions (Review) 103 Copyright © 2014 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Analysis 1.9. Comparison 1 Adhesive versus suture, Outcome 9 Surgeon satisfaction (% satisﬁed).

Review: Tissue adhesives for closure of surgical incisions

Comparison: 1 Adhesive versus suture

Outcome: 9 Surgeon satisfaction (% satisﬁed)

Studyorsubgroup Adhesive Suture RiskRatio Weight RiskRatio M- M- H,Random,95% H,Random,95% n/N n/N CI CI Maartense 2002 20/48 13/50 45.7 % 1.60 [ 0.90, 2.85 ]

Shamiyeh 2001 15/26 18/26 54.3 % 0.83 [ 0.55, 1.26 ] Total (95% CI) 74 76 100.0 % 1.12 [ 0.58, 2.19 ] Total events: 35 (Adhesive), 31 (Suture) Heterogeneity: Tau2 = 0.17; Chi2 = 3.58, df = 1 (P = 0.06); I2 =72% Test for overall effect: Z = 0.34 (P = 0.73) Test for subgroup differences: Not applicable

0.01 0.1 1 10 100 Favours suture Favours adhesive

Tissue adhesives for closure of surgical incisions (Review) 104 Copyright © 2014 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Analysis 1.10. Comparison 1 Adhesive versus suture, Outcome 10 Time taken for wound closure.

Review: Tissue adhesives for closure of surgical incisions

Comparison: 1 Adhesive versus suture

Outcome: 10 Time taken for wound closure

Mean Mean Studyorsubgroup Adhesive Suture Difference Weight Difference N Mean(SD) N Mean(SD) IV,Random,95% CI IV,Random,95% CI

Sebesta 2004 29 222 (67.8) 30 845 (360) -623.00 [ -754.16, -491.84 ]

Ozturan 2001 34 55 (10) 67 155 (25) -100.00 [ -106.87, -93.13 ]

Brown 2009 64 1.4 (0.8) 70 2.4 (1.1) -1.00 [ -1.32, -0.68 ]

Shamiyeh 2001 26 1.14 (0.24) 28 0.64 (0.23) 0.50 [ 0.37, 0.63 ]

Ong 2002 26 181 (61) 33 161 (45) 20.00 [ -8.03, 48.03 ] Subtotal (95% CI) 0 0 0.0 [ 0.0, 0.0 ] Heterogeneity: Tau2 = 0.0; Chi2 = 0.0, df = 0 (P<0.00001); I2 =0.0% Test for overall effect: Z = 0.0 (P < 0.00001) Test for subgroup differences: Not applicable

-100 -50 0 50 100 Favours adhesive Favours suture

Tissue adhesives for closure of surgical incisions (Review) 105 Copyright © 2014 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Analysis 2.1. Comparison 2 Adhesive versus adhesive tape, Outcome 1 Dehiscence.

Review: Tissue adhesives for closure of surgical incisions

Comparison: 2 Adhesive versus adhesive tape

Outcome: 1 Dehiscence

Studyorsubgroup Adhesive Tape RiskRatio Weight RiskRatio M- M- H,Random,95% H,Random,95% n/N n/N CI CI Romero 2011 0/24 0/25 Not estimable

Shamiyeh 2001 1/26 1/25 0.96 [ 0.06, 14.55 ] Subtotal (95% CI) 0 0 0.0 [ 0.0, 0.0 ] Total events: 1 (Adhesive), 1 (Tape) Heterogeneity: not applicable Test for overall effect: Z = 0.0 (P < 0.00001) Test for subgroup differences: Not applicable

0.001 0.01 0.1 1 10 100 1000 Favours adhesive Favours tape

Analysis 2.2. Comparison 2 Adhesive versus adhesive tape, Outcome 2 Infection.

Review: Tissue adhesives for closure of surgical incisions

Comparison: 2 Adhesive versus adhesive tape

Outcome: 2 Infection

Studyorsubgroup Adhesive Tape RiskRatio Weight RiskRatio M- M- H,Random,95% H,Random,95% n/N n/N CI CI Maartense 2002 5/48 2/42 62.7 % 2.19 [ 0.45, 10.69 ]

Romero 2011 0/24 1/25 15.9 % 0.35 [ 0.01, 8.12 ]

Shamiyeh 2001 1/26 1/25 21.4 % 0.96 [ 0.06, 14.55 ] Total (95% CI) 98 92 100.0 % 1.37 [ 0.39, 4.81 ] Total events: 6 (Adhesive), 4 (Tape) Heterogeneity: Tau2 = 0.0; Chi2 = 1.13, df = 2 (P = 0.57); I2 =0.0% Test for overall effect: Z = 0.49 (P = 0.62) Test for subgroup differences: Not applicable

0.005 0.1 1 10 200 Favours adhesive Favours tape

Tissue adhesives for closure of surgical incisions (Review) 106 Copyright © 2014 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Analysis 2.3. Comparison 2 Adhesive versus adhesive tape, Outcome 3 Cosmetic appearance rated by patient (VAS).

Review: Tissue adhesives for closure of surgical incisions

Comparison: 2 Adhesive versus adhesive tape

Outcome: 3 Cosmetic appearance rated by patient (VAS)

Mean Mean Studyorsubgroup Adhesive Tape Difference Weight Difference N Mean(SD) N Mean(SD) IV,Fixed,95% CI IV,Fixed,95% CI

Maartense 2002 48 76 (24) 42 79 (20) -3.00 [ -12.09, 6.09 ] Subtotal (95% CI) 0 0 0.0 [ 0.0, 0.0 ] Heterogeneity: not applicable Test for overall effect: Z = 0.0 (P < 0.00001) Test for subgroup differences: Not applicable

-50 -25 0 25 50 Favours tape Favours adhesive

Tissue adhesives for closure of surgical incisions (Review) 107 Copyright © 2014 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Analysis 2.4. Comparison 2 Adhesive versus adhesive tape, Outcome 4 Cosmetic appearance rated by patient (% satisﬁed).

Review: Tissue adhesives for closure of surgical incisions

Comparison: 2 Adhesive versus adhesive tape

Outcome: 4 Cosmetic appearance rated by patient (% satisﬁed)

Studyorsubgroup Adhesive Tape RiskRatio Weight RiskRatio n/N n/N M-H,Fixed,95% CI M-H,Fixed,95% CI Romero 2011 19/22 19/20 0.91 [ 0.75, 1.10 ] Subtotal (95% CI) 0 0 0.0 [ 0.0, 0.0 ] Total events: 19 (Adhesive), 19 (Tape) Heterogeneity: not applicable Test for overall effect: Z = 0.0 (P < 0.00001) Test for subgroup differences: Not applicable

0.01 0.1 1 10 100 Favours tape Favours adhesive

Analysis 2.5. Comparison 2 Adhesive versus adhesive tape, Outcome 5 Cosmetic appearance rated by surgeon (VAS).

Review: Tissue adhesives for closure of surgical incisions

Comparison: 2 Adhesive versus adhesive tape

Outcome: 5 Cosmetic appearance rated by surgeon (VAS)

Mean Mean Studyorsubgroup Adhesive Tape Difference Weight Difference N Mean(SD) N Mean(SD) IV,Random,95% CI IV,Random,95% CI

Maartense 2002 48 82 (20) 42 69 (18) 33.8 % 13.00 [ 5.15, 20.85 ]

Romero 2011 24 28.9 (10.4) 25 21.1 (8.3) 66.2 % 7.80 [ 2.52, 13.08 ] Total (95% CI) 72 67 100.0 % 9.56 [ 4.74, 14.37 ] Heterogeneity: Tau2 = 1.87; Chi2 = 1.16, df = 1 (P = 0.28); I2 =14% Test for overall effect: Z = 3.89 (P = 0.00010) Test for subgroup differences: Not applicable

-50 -25 0 25 50 Favours adhesive Favours tape

Tissue adhesives for closure of surgical incisions (Review) 108 Copyright © 2014 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Analysis 2.6. Comparison 2 Adhesive versus adhesive tape, Outcome 6 Patient satisfaction.

Review: Tissue adhesives for closure of surgical incisions

Comparison: 2 Adhesive versus adhesive tape

Outcome: 6 Patient satisfaction

Studyorsubgroup Adhesive Tape RiskRatio Weight RiskRatio n/N n/N M-H,Fixed,95% CI M-H,Fixed,95% CI Shamiyeh 2001 25/26 22/25 1.09 [ 0.93, 1.29 ] Subtotal (95% CI) 0 0 0.0 [ 0.0, 0.0 ] Total events: 25 (Adhesive), 22 (Tape) Heterogeneity: not applicable Test for overall effect: Z = 0.0 (P < 0.00001) Test for subgroup differences: Not applicable

0.5 0.7 1 1.5 2 Favours tape Favours adhesive

Analysis 2.7. Comparison 2 Adhesive versus adhesive tape, Outcome 7 Surgeon satisfaction.

Review: Tissue adhesives for closure of surgical incisions

Comparison: 2 Adhesive versus adhesive tape

Outcome: 7 Surgeon satisfaction

Studyorsubgroup Adhesive Tape RiskRatio Weight RiskRatio M- M- H,Random,95% H,Random,95% n/N n/N CI CI Maartense 2002 20/48 21/42 48.9 % 0.83 [ 0.53, 1.31 ]

Shamiyeh 2001 15/26 16/25 51.1 % 0.90 [ 0.58, 1.40 ] Total (95% CI) 74 67 100.0 % 0.87 [ 0.63, 1.19 ] Total events: 35 (Adhesive), 37 (Tape) Heterogeneity: Tau2 = 0.0; Chi2 = 0.06, df = 1 (P = 0.80); I2 =0.0% Test for overall effect: Z = 0.88 (P = 0.38) Test for subgroup differences: Not applicable

0.05 0.2 1 5 20 Favours tape Favours adhesive

Tissue adhesives for closure of surgical incisions (Review) 109 Copyright © 2014 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Analysis 2.8. Comparison 2 Adhesive versus adhesive tape, Outcome 8 Time taken for wound closure.

Review: Tissue adhesives for closure of surgical incisions

Comparison: 2 Adhesive versus adhesive tape

Outcome: 8 Time taken for wound closure

Mean Mean Studyorsubgroup Adhesive Tape Difference Weight Difference N Mean(SD) N Mean(SD) IV,Fixed,95% CI IV,Fixed,95% CI

Shamiyeh 2001 26 1.14 (0.24) 25 0.58 (0.27) 0.56 [ 0.42, 0.70 ] Subtotal (95% CI) 0 0 0.0 [ 0.0, 0.0 ] Heterogeneity: not applicable Test for overall effect: Z = 0.0 (P < 0.00001) Test for subgroup differences: Not applicable

-2 -1 0 1 2 Favours adhesive Favours tape

Analysis 3.1. Comparison 3 Adhesive versus staples, Outcome 1 Dehiscence.

Review: Tissue adhesives for closure of surgical incisions

Comparison: 3 Adhesive versus staples

Outcome: 1 Dehiscence

Studyorsubgroup Adhesive Staples RiskRatio Weight RiskRatio M- M- H,Random,95% H,Random,95% n/N n/N CI CI Eggers 2011 1/18 2/19 100.0 % 0.53 [ 0.05, 5.33 ]

Pronio 2011 0/32 0/38 Not estimable Total (95% CI) 50 57 100.0 % 0.53 [ 0.05, 5.33 ] Total events: 1 (Adhesive), 2 (Staples) Heterogeneity: not applicable Test for overall effect: Z = 0.54 (P = 0.59) Test for subgroup differences: Not applicable

0.01 0.1 1 10 100 Favours staples Favours adhesive

Tissue adhesives for closure of surgical incisions (Review) 110 Copyright © 2014 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Analysis 3.2. Comparison 3 Adhesive versus staples, Outcome 2 Infection.

Review: Tissue adhesives for closure of surgical incisions

Comparison: 3 Adhesive versus staples

Outcome: 2 Infection

Studyorsubgroup Adhesive Staples RiskRatio Weight RiskRatio M- M- H,Random,95% H,Random,95% n/N n/N CI CI Eggers 2011 6/18 1/19 31.0 % 6.33 [ 0.84, 47.57 ]

Khan 2006 4/60 7/63 47.5 % 0.60 [ 0.19, 1.95 ]

Livesey 2009 1/45 1/45 21.5 % 1.00 [ 0.06, 15.50 ]

Pronio 2011 0/32 0/38 Not estimable Total (95% CI) 155 165 100.0 % 1.39 [ 0.30, 6.54 ] Total events: 11 (Adhesive), 9 (Staples) Heterogeneity: Tau2 = 0.95; Chi2 = 4.04, df = 2 (P = 0.13); I2 =50% Test for overall effect: Z = 0.42 (P = 0.68) Test for subgroup differences: Not applicable

0.01 0.1 1 10 100 Favours adhesive Favours staples

Analysis 3.3. Comparison 3 Adhesive versus staples, Outcome 3 Cosmetic appearance rated by patient (scar scale).

Review: Tissue adhesives for closure of surgical incisions

Comparison: 3 Adhesive versus staples

Outcome: 3 Cosmetic appearance rated by patient (scar scale)

Mean Mean Studyorsubgroup Adhesive Staples Difference Weight Difference N Mean(SD) N Mean(SD) IV,Fixed,95% CI IV,Fixed,95% CI

Pronio 2011 32 4.94 (0.25) 38 4.86 (0.36) 0.08 [ -0.06, 0.22 ] Subtotal (95% CI) 0 0 0.0 [ 0.0, 0.0 ] Heterogeneity: not applicable Test for overall effect: Z = 0.0 (P < 0.00001) Test for subgroup differences: Not applicable

-1 -0.5 0 0.5 1 Favours adhesive Favours staples

Tissue adhesives for closure of surgical incisions (Review) 111 Copyright © 2014 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Analysis 3.4. Comparison 3 Adhesive versus staples, Outcome 4 Cosmetic appearance by plastic surgeons (VAS).

Review: Tissue adhesives for closure of surgical incisions

Comparison: 3 Adhesive versus staples

Outcome: 4 Cosmetic appearance by plastic surgeons (VAS)

Mean Mean Studyorsubgroup Adhesive Staples Difference Weight Difference N Mean(SD) N Mean(SD) IV,Fixed,95% CI IV,Fixed,95% CI

Livesey 2009 45 78.2 (10.9) 45 81.1 (7.4) -2.90 [ -6.75, 0.95 ] Subtotal (95% CI) 0 0 0.0 [ 0.0, 0.0 ] Heterogeneity: not applicable Test for overall effect: Z = 0.0 (P < 0.00001) Test for subgroup differences: Not applicable

-20 -10 0 10 20 Favours staples Favours adhesive

Tissue adhesives for closure of surgical incisions (Review) 112 Copyright © 2014 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Analysis 3.5. Comparison 3 Adhesive versus staples, Outcome 5 Patient satisfaction.

Review: Tissue adhesives for closure of surgical incisions

Comparison: 3 Adhesive versus staples

Outcome: 5 Patient satisfaction

Mean Mean Studyorsubgroup Adhesive Staples Difference Weight Difference N Mean(SD) N Mean(SD) IV,Fixed,95% CI IV,Fixed,95% CI

Amin 2009 33 9.1 (1.3) 27 8 (1.4) 1.10 [ 0.41, 1.79 ] Subtotal (95% CI) 0 0 0.0 [ 0.0, 0.0 ] Heterogeneity: not applicable Test for overall effect: Z = 0.0 (P < 0.00001) Test for subgroup differences: Not applicable

-4 -2 0 2 4 Favours staples Favours adhesive

Analysis 3.6. Comparison 3 Adhesive versus staples, Outcome 6 Time taken for wound closure.

Review: Tissue adhesives for closure of surgical incisions

Comparison: 3 Adhesive versus staples

Outcome: 6 Time taken for wound closure

Mean Mean Studyorsubgroup Adhesive Staples Difference Weight Difference N Mean(SD) N Mean(SD) IV,Fixed,95% CI IV,Fixed,95% CI

Ridgway 2007 14 95 (52) 15 28 (10) 67.00 [ 39.30, 94.70 ] Subtotal (95% CI) 0 0 0.0 [ 0.0, 0.0 ] Heterogeneity: not applicable Test for overall effect: Z = 0.0 (P < 0.00001) Test for subgroup differences: Not applicable

-200 -100 0 100 200 Favours adhesive Favours staple

Tissue adhesives for closure of surgical incisions (Review) 113 Copyright © 2014 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Analysis 4.1. Comparison 4 Adhesive versus other method, Outcome 1 Dehiscence.

Review: Tissue adhesives for closure of surgical incisions

Comparison: 4 Adhesive versus other method

Outcome: 1 Dehiscence

Studyorsubgroup HVadhesive Other RiskRatio Weight RiskRatio M- M- H,Random,95% H,Random,95% n/N n/N CI CI Blondeel 2004 4/148 3/61 100.0 % 0.55 [ 0.13, 2.38 ]

Chibbaro 2009 0/20 0/20 Not estimable Total (95% CI) 168 81 100.0 % 0.55 [ 0.13, 2.38 ] Total events: 4 (HV adhesive), 3 (Other) Heterogeneity: not applicable Test for overall effect: Z = 0.80 (P = 0.42) Test for subgroup differences: Not applicable

0.002 0.1 1 10 500 Favours HV adhesive Favours other

Analysis 4.2. Comparison 4 Adhesive versus other method, Outcome 2 Infection.

Review: Tissue adhesives for closure of surgical incisions

Comparison: 4 Adhesive versus other method

Outcome: 2 Infection

Studyorsubgroup HVadhesive Other RiskRatio Weight RiskRatio M- M- H,Random,95% H,Random,95% n/N n/N CI CI Blondeel 2004 4/148 4/61 100.0 % 0.41 [ 0.11, 1.60 ]

Chibbaro 2009 0/20 0/20 Not estimable Total (95% CI) 168 81 100.0 % 0.41 [ 0.11, 1.60 ] Total events: 4 (HV adhesive), 4 (Other) Heterogeneity: not applicable Test for overall effect: Z = 1.28 (P = 0.20) Test for subgroup differences: Not applicable

0.001 0.01 0.1 1 10 100 1000 Favours HV adhesive Favours other

Tissue adhesives for closure of surgical incisions (Review) 114 Copyright © 2014 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Analysis 4.3. Comparison 4 Adhesive versus other method, Outcome 3 Patient satisfaction.

Review: Tissue adhesives for closure of surgical incisions

Comparison: 4 Adhesive versus other method

Outcome: 3 Patient satisfaction

Mean Mean Studyorsubgroup HVadhesive Other Difference Weight Difference N Mean(SD) N Mean(SD) IV,Fixed,95% CI IV,Fixed,95% CI

Blondeel 2004 133 4.51 (0.66) 54 4.11 (1.06) 100.0 % 0.40 [ 0.10, 0.70 ] Total (95% CI) 133 54 100.0 % 0.40 [ 0.10, 0.70 ] Heterogeneity: not applicable Test for overall effect: Z = 2.58 (P = 0.010) Test for subgroup differences: Not applicable

-2 -1 0 1 2 Favours HV adhesive Favours other

Tissue adhesives for closure of surgical incisions (Review) 115 Copyright © 2014 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Analysis 4.4. Comparison 4 Adhesive versus other method, Outcome 4 Clinician satisfaction.

Review: Tissue adhesives for closure of surgical incisions

Comparison: 4 Adhesive versus other method

Outcome: 4 Clinician satisfaction

Mean Mean Studyorsubgroup HVadhesive Other Difference Weight Difference N Mean(SD) N Mean(SD) IV,Fixed,95% CI IV,Fixed,95% CI

Blondeel 2004 148 4.53 (0.6) 61 4 (0.89) 100.0 % 0.53 [ 0.29, 0.77 ] Total (95% CI) 148 61 100.0 % 0.53 [ 0.29, 0.77 ] Heterogeneity: not applicable Test for overall effect: Z = 4.27 (P = 0.000020) Test for subgroup differences: Not applicable

-1 -0.5 0 0.5 1 Favours HV adhesive Favours other

Analysis 4.5. Comparison 4 Adhesive versus other method, Outcome 5 Time taken for wound closure.

Review: Tissue adhesives for closure of surgical incisions

Comparison: 4 Adhesive versus other method

Outcome: 5 Time taken for wound closure

Mean Mean Studyorsubgroup HVadhesive Other Difference Weight Difference N Mean(SD) N Mean(SD) IV,Fixed,95% CI IV,Fixed,95% CI

Blondeel 2004 148 1.96 (1.67) 61 3.01 (2.76) 100.0 % -1.05 [ -1.79, -0.31 ] Total (95% CI) 148 61 100.0 % -1.05 [ -1.79, -0.31 ] Heterogeneity: not applicable Test for overall effect: Z = 2.77 (P = 0.0056) Test for subgroup differences: Not applicable

-2 -1 0 1 2 Favours HV adhesive Favours other

Tissue adhesives for closure of surgical incisions (Review) 116 Copyright © 2014 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Analysis 5.1. Comparison 5 Adhesive versus adhesive: High viscosity versus low viscosity, Outcome 1 Dehiscence.

Review: Tissue adhesives for closure of surgical incisions

Comparison: 5 Adhesive versus adhesive: High viscosity versus low viscosity

Outcome: 1 Dehiscence

Studyorsubgroup HVadhesive LVadhesive RiskRatio Weight Risk Ratio n/N n/N M-H,Fixed,95% CI M-H,Fixed,95% CI Blondeel 2004 4/105 0/43 3.74 [ 0.21, 67.93 ] Subtotal (95% CI) 0 0 0.0 [ 0.0, 0.0 ] Total events: 4 (HV adhesive), 0 (LV adhesive) Heterogeneity: not applicable Test for overall effect: Z = 0.0 (P < 0.00001) Test for subgroup differences: Not applicable

0.01 0.1 1 10 100 Favours HV adhesive Favours LV adhesive

Analysis 5.2. Comparison 5 Adhesive versus adhesive: High viscosity versus low viscosity, Outcome 2 Infection.

Review: Tissue adhesives for closure of surgical incisions

Comparison: 5 Adhesive versus adhesive: High viscosity versus low viscosity

Outcome: 2 Infection

Studyorsubgroup HVadhesive LVadhesive RiskRatio Weight Risk Ratio n/N n/N M-H,Fixed,95% CI M-H,Fixed,95% CI Blondeel 2004 4/105 2/43 0.82 [ 0.16, 4.31 ] Subtotal (95% CI) 0 0 0.0 [ 0.0, 0.0 ] Total events: 4 (HV adhesive), 2 (LV adhesive) Heterogeneity: not applicable Test for overall effect: Z = 0.0 (P < 0.00001) Test for subgroup differences: Not applicable

0.005 0.1 1 10 200 Favours HV adhesive Favours LV adhesive

Tissue adhesives for closure of surgical incisions (Review) 117 Copyright © 2014 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Analysis 5.3. Comparison 5 Adhesive versus adhesive: High viscosity versus low viscosity, Outcome 3 Patient satisfaction.

Review: Tissue adhesives for closure of surgical incisions

Comparison: 5 Adhesive versus adhesive: High viscosity versus low viscosity

Outcome: 3 Patient satisfaction

Mean Mean Studyorsubgroup HVadhesive LVadhesive Difference Weight Difference N Mean(SD) N Mean(SD) IV,Fixed,95% CI IV,Fixed,95% CI

Blondeel 2004 95 4.5 (0.68) 38 4.55 (0.6) -0.05 [ -0.28, 0.18 ] Subtotal (95% CI) 0 0 0.0 [ 0.0, 0.0 ] Heterogeneity: not applicable Test for overall effect: Z = 0.0 (P < 0.00001) Test for subgroup differences: Not applicable

-0.5 -0.25 0 0.25 0.5 Favours HV adhesive Favours LV adhesive

Analysis 5.4. Comparison 5 Adhesive versus adhesive: High viscosity versus low viscosity, Outcome 4 Clinician satisfaction.

Review: Tissue adhesives for closure of surgical incisions

Comparison: 5 Adhesive versus adhesive: High viscosity versus low viscosity

Outcome: 4 Clinician satisfaction

Mean Mean Studyorsubgroup HVadhesive LVadhesive Difference Weight Difference N Mean(SD) N Mean(SD) IV,Fixed,95% CI IV,Fixed,95% CI

Blondeel 2004 105 4.52 (0.61) 43 4.54 (0.59) -0.02 [ -0.23, 0.19 ] Subtotal (95% CI) 0 0 0.0 [ 0.0, 0.0 ] Heterogeneity: not applicable Test for overall effect: Z = 0.0 (P < 0.00001) Test for subgroup differences: Not applicable

-1 -0.5 0 0.5 1 Favours HV adhesive Favours LV adhesive

Tissue adhesives for closure of surgical incisions (Review) 118 Copyright © 2014 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Analysis 5.5. Comparison 5 Adhesive versus adhesive: High viscosity versus low viscosity, Outcome 5 Time taken for wound closure.

Review: Tissue adhesives for closure of surgical incisions

Comparison: 5 Adhesive versus adhesive: High viscosity versus low viscosity

Outcome: 5 Time taken for wound closure

Mean Mean Studyorsubgroup HVadhesive LVadhesive Difference Weight Difference N Mean(SD) N Mean(SD) IV,Fixed,95% CI IV,Fixed,95% CI

Blondeel 2004 105 2.12 (1.79) 43 1.58 (1.26) 0.54 [ 0.03, 1.05 ] Subtotal (95% CI) 0 0 0.0 [ 0.0, 0.0 ] Heterogeneity: not applicable Test for overall effect: Z = 0.0 (P < 0.00001) Test for subgroup differences: Not applicable

-2 -1 0 1 2 Favours HV adhesive Favours LV adhesive

Analysis 6.1. Comparison 6 Adhesive versus adhesive: octylcyanoacrylate versus butylcyanoacrylate, Outcome 1 Dehiscence.

Review: Tissue adhesives for closure of surgical incisions

Comparison: 6 Adhesive versus adhesive: octylcyanoacrylate versus butylcyanoacrylate

Outcome: 1 Dehiscence

Studyorsubgroup Octylcyanoacrylate Butylcyanoacrylate RiskRatio Weight RiskRatio M- M- H,Random,95% H,Random,95% n/N n/N CI CI Eggers 2011 1/19 1/18 57.7 % 0.95 [ 0.06, 14.04 ]

Maloney 2013 1/23 0/20 42.3 % 2.63 [ 0.11, 61.05 ] Total (95% CI) 42 38 100.0 % 1.46 [ 0.19, 11.30 ] Total events: 2 (Octylcyanoacrylate), 1 (Butylcyanoacrylate) Heterogeneity: Tau2 = 0.0; Chi2 = 0.23, df = 1 (P = 0.63); I2 =0.0% Test for overall effect: Z = 0.36 (P = 0.72) Test for subgroup differences: Not applicable

0.01 0.1 1 10 100 Favours octyl’ Favours butyl’

Tissue adhesives for closure of surgical incisions (Review) 119 Copyright © 2014 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Analysis 6.2. Comparison 6 Adhesive versus adhesive: octylcyanoacrylate versus butylcyanoacrylate, Outcome 2 Infection.

Review: Tissue adhesives for closure of surgical incisions

Comparison: 6 Adhesive versus adhesive: octylcyanoacrylate versus butylcyanoacrylate

Outcome: 2 Infection

Studyorsubgroup Octylcyanoacrylate Butylcyanoacrylate RiskRatio Weight RiskRatio n/N n/N M-H,Fixed,95% CI M-H,Fixed,95% CI Eggers 2011 4/19 6/18 0.63 [ 0.21, 1.88 ]

Maloney 2013 0/23 0/20 Not estimable Subtotal (95% CI) 0 0 0.0 [ 0.0, 0.0 ] Total events: 4 (Octylcyanoacrylate), 6 (Butylcyanoacrylate) Heterogeneity: not applicable Test for overall effect: Z = 0.0 (P < 0.00001) Test for subgroup differences: Not applicable

0.01 0.1 1 10 100 Favours octyl’ Favours butyl’

Tissue adhesives for closure of surgical incisions (Review) 120 Copyright © 2014 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Analysis 6.3. Comparison 6 Adhesive versus adhesive: octylcyanoacrylate versus butylcyanoacrylate, Outcome 3 Cosmetic assessment rated by patient (VAS).

Review: Tissue adhesives for closure of surgical incisions

Comparison: 6 Adhesive versus adhesive: octylcyanoacrylate versus butylcyanoacrylate

Outcome: 3 Cosmetic assessment rated by patient (VAS)

Mean Mean Studyorsubgroup Octylcyanoacrylate Butylcyanoacrylate Difference Weight Difference N Mean(SD) N Mean(SD) IV,Fixed,95% CI IV,Fixed,95% CI

Maloney 2013 23 9 (1.2) 20 8.4 (2.1) 0.60 [ -0.44, 1.64 ] Subtotal (95% CI) 0 0 0.0 [ 0.0, 0.0 ] Heterogeneity: not applicable Test for overall effect: Z = 0.0 (P < 0.00001) Test for subgroup differences: Not applicable

-10 -5 0 5 10 Favours octyl’ Favours butyl’

Analysis 6.4. Comparison 6 Adhesive versus adhesive: octylcyanoacrylate versus butylcyanoacrylate, Outcome 4 Cosmetic assessment rated by surgeon (VAS).

Review: Tissue adhesives for closure of surgical incisions

Comparison: 6 Adhesive versus adhesive: octylcyanoacrylate versus butylcyanoacrylate

Outcome: 4 Cosmetic assessment rated by surgeon (VAS)

Mean Mean Studyorsubgroup Octylcyanoacrylate Butylcyanoacrylate Difference Weight Difference N Mean(SD) N Mean(SD) IV,Fixed,95% CI IV,Fixed,95% CI

Maloney 2013 23 6.8 (1.7) 20 7.2 (1.6) -0.40 [ -1.39, 0.59 ] Subtotal (95% CI) 0 0 0.0 [ 0.0, 0.0 ] Heterogeneity: not applicable Test for overall effect: Z = 0.0 (P < 0.00001) Test for subgroup differences: Not applicable

-10 -5 0 5 10 Favours octyl’ Favours butyl’

Tissue adhesives for closure of surgical incisions (Review) 121 Copyright © 2014 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Analysis 6.5. Comparison 6 Adhesive versus adhesive: octylcyanoacrylate versus butylcyanoacrylate, Outcome 5 Surgeon satisfaction (with device).

Review: Tissue adhesives for closure of surgical incisions

Comparison: 6 Adhesive versus adhesive: octylcyanoacrylate versus butylcyanoacrylate

Outcome: 5 Surgeon satisfaction (with device)

Mean Mean Studyorsubgroup Octylcyanoacrylate Butylcyanoacrylate Difference Weight Difference N Mean(SD) N Mean(SD) IV,Fixed,95% CI IV,Fixed,95% CI

Maloney 2013 23 8.8 (0.7) 20 9 (1) -0.20 [ -0.72, 0.32 ] Subtotal (95% CI) 0 0 0.0 [ 0.0, 0.0 ] Heterogeneity: not applicable Test for overall effect: Z = 0.0 (P < 0.00001) Test for subgroup differences: Not applicable

-4 -2 0 2 4 Favours octyl’ Favours butyl’

Analysis 6.6. Comparison 6 Adhesive versus adhesive: octylcyanoacrylate versus butylcyanoacrylate, Outcome 6 Surgeon satisfaction (with closure).

Review: Tissue adhesives for closure of surgical incisions

Comparison: 6 Adhesive versus adhesive: octylcyanoacrylate versus butylcyanoacrylate

Outcome: 6 Surgeon satisfaction (with closure)

Mean Mean Studyorsubgroup Octylcyanoacrylate Butylcyanoacrylate Difference Weight Difference N Mean(SD) N Mean(SD) IV,Fixed,95% CI IV,Fixed,95% CI

Maloney 2013 23 8.3 (1) 20 8.4 (1.2) -0.10 [ -0.77, 0.57 ] Subtotal (95% CI) 0 0 0.0 [ 0.0, 0.0 ] Heterogeneity: not applicable Test for overall effect: Z = 0.0 (P < 0.00001) Test for subgroup differences: Not applicable

-4 -2 0 2 4 Favours octyl’ Favours butyl’

Tissue adhesives for closure of surgical incisions (Review) 122 Copyright © 2014 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Analysis 6.7. Comparison 6 Adhesive versus adhesive: octylcyanoacrylate versus butylcyanoacrylate, Outcome 7 Time taken for wound closure.

Review: Tissue adhesives for closure of surgical incisions

Comparison: 6 Adhesive versus adhesive: octylcyanoacrylate versus butylcyanoacrylate

Outcome: 7 Time taken for wound closure

Mean Mean Studyorsubgroup Octylcyanoacrylate Butylcyanoacrylate Difference Weight Difference N Mean(SD) N Mean(SD) IV,Fixed,95% CI IV,Fixed,95% CI

Maloney 2013 23 54.5 (25.6) 20 80 (22.2) -25.50 [ -39.79, -11.21 ] Subtotal (95% CI) 0 0 0.0 [ 0.0, 0.0 ] Heterogeneity: not applicable Test for overall effect: Z = 0.0 (P < 0.00001) Test for subgroup differences: Not applicable

-100 -50 0 50 100 Favours octyl’ Favours butyl’

ADDITIONAL TABLES

Table 1. Summary of study comparisons

Tis- 2- 2- 2- Trial 2- Su- Sta- Ad- All sue Mixed Butyl- octyl octyl Butyl- octyl ID octyl Butyl- tures ples he- Mixed non- Other adhe- con- 2- 2- 2- sive su- tissue vis- sive trol cyanoacry-cyanoacry- cyanoacry- cyanoacry- tape/ tures adhe- cos- vs tis- cyanoacry-late late cyanoacry-late late cyanoacry- strips and sive ity 2- sue (Comp late vs vs late vs su- late sta- clo- octyl adhe- 4) vs sta- tape vs sutures ples sure sive staples tures meth- cyanoacry- ples (Comp (Comp1) ods late (Comp (Comp 2) (Comp 5) (Comp 3) 1) 3)

3 Amin √ √ 2009

1 Avsar √ √ 2009

5 4 Blon- √ √ √ deel 2004

1 √ √ Brown 2009

1 √ √ Cheng 1997

4 Chib- √ √ baro 2009

1 Dow- √ √ son 2006

5 3 3 1 1 Eg- √ √ √ √ gers 2011

1 √ √ Greene 1999

1 Jallali √ √ 2004

1 Keng √ √ 1989

5 Kent √ √ 2014

3 1 Khan √ √ √ 2006

1 √ √ Kouba 2011

1 Kr- √ √ ish- namoor- thy 2009

3 √ √ Livesey 2009

2 1 √ √ √ Maartense 2002

5 Mal- √ √ oney 2013

1 Mil- √ √ lan 2011

1 Mota √ √ 2009

1 Ong √ √ 2002

1 Oztu- √ √ ran 2001

3 Pro- √ √ nio 2011

3 Ridg- √ √ way 2007

1 √ √ Sebesta 2004

2 1 √ √ √ Shamiyeh 2001

1 Sinha √ √ 2001

1 √ √ Sniezek 2007

1 √ √ Switzer 2003

1 √ √ Tierny 2009

1 Tori- √ √ umi 1998

2 √ √ Romero 2011

1 van √ √ den Ende 2004 Abbreviation Comp = comparison

APPENDICES

Appendix 1. Search strategy used in the original version of this review - 2007 In order to identify studies to be considered for this review the following databases were searched: Cochrane Wounds Group Trials Register - November 2007 The Cochrane Central Register of Controlled Trials (CENTRAL) - The Cochrane Library Issue 2, November 2007 MEDLINE 1966 - November 2007 EMBASE 1980 - November 2007 The Cochrane Wounds Group Trials Register has been compiled through searching of the major health databases including MED- LINE, Cinahl and EMBASE and is regularly updated through searching of the Cochrane Central Register of Controlled Trials, handsearching of wound care journals and relevant conference proceedings. See: Collaborative review group search strategy (http:/ www.cochranewounds.org). The following search strategy was used for searching the Cochrane Wounds Group Trials Register and CENTRAL: 1 WOUNDS-AND-INJURIES*:ME 2 INCISION* 3 WOUND* 4 (SURGICAL and WOUND*) 5 (((#1 or #2) or #3) or #4) 6 TISSUE-ADHESIVES*:ME 7 ADHESIVES*:ME 8 ACRYLATES*:ME 9 (TISSUE* and ADHESIVE*) 10 ACRYLATE* 11 CYANOACRYLATE* 12 (GLU or GLUES) 13 (GLUE or GLUES) 14 FIBRIN-TISSUE-ADHESIVE*:ME 15 BUCRYLATE*:ME 16 BUCRYLATE* 17 SUTURES*:ME 18 SUTUR* 19 SURGICAL-STAPLING*:ME

Tissue adhesives for closure of surgical incisions (Review) 126 Copyright © 2014 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. 20 STAPLE* 21 TAPE* 22 (((((((((((((((#6 or #7) or #8) or #9) or #10) or #11) or #12) or #13) or #14) or #15) or #16) or #17) or #18) or #19) or #20) or # 21) 23 (#5 and #22) Search strategies were developed for Medline and Embase and these search strategies combined a sensitive search strategy for RCTs revised from phases 1 and 2 of the Cochrane Sensitive Search Strategy for RCTs (as published in Appendix 6c in the Cochrane Handbook). The subject search used a combination of controlled vocabulary and free text terms based on the search strategy for searching the Cochrane Wounds Group Trial Register. LANGUAGE There were no language restrictions. UNPUBLISHED STUDIES Authors of the identiﬁed RCTs were written to in order to obtain further information about the trial and to attempt to identify unpublished or ongoing studies. In addition, wound care product manufacturers were contacted. HANDSEARCHING Trials for Wounds Group Trials Register are identiﬁed by systematically handsearching specialised journals, relevant conference proceedings and abstracts. A list of journals currently being handsearched by the group may be found at http:/www.cochranewounds.org.

Appendix 2. Search strategy for the ﬁrst update of this review - 2009 In order to identify studies to be considered for this review the following databases were searched: Cochrane Wounds Group Specialised Register (November 2009;); The Cochrane Central Register of Controlled Trials (CENTRAL) (The Cochrane Library 2009, Issue 4); Ovid MEDLINE (1996 to November week 1 2009); Ovid MEDLINE (In-Process & Other Non-Indexed Citations, November 16 2009); Ovid EMBASE (1996 to 2009 week 46); EBSCO CINAHL (1982 to 11 November 2009) The Cochrane Wounds Group Trials Register has been compiled through searching of the major health databases including MED- LINE, Cinahl and EMBASE and is regularly updated through searching of the Cochrane Central Register of Controlled Trials, handsearching of wound care journals and relevant conference proceedings. See: Collaborative review group search strategy (http:/ www.cochranewounds.org). The following search strategy was used for searching the Cochrane Central Register of Controlled Trials (CENTRAL): #1 MeSH descriptor Wounds and Injuries explode all trees #2 surgical next wound* #3 (#1 OR #2) #4 MeSH descriptor Tissue Adhesives explode all trees #5 MeSH descriptor Fibrin Tissue Adhesive explode all trees #6 tissue next adhesive* #7 MeSH descriptor Cyanoacrylates explode all trees #8 octylcyanoacrylate* #9 Dermabond #10 MeSH descriptor Enbucrilate explode all trees #11 enbucrilate #12 butylcyanoacrylate* #13 MeSH descriptor Acrylates explode all trees #14 acrylate* #15 MeSH descriptor Bucrylate explode all trees #16 bucrylate* #17 (#4 OR #5 OR #6 OR #7 OR #8 OR #9 OR #10 OR #11 OR #12 OR #13 OR #14 OR #15 OR #16) #18 (#3 AND #17) with New in Record Status

Tissue adhesives for closure of surgical incisions (Review) 127 Copyright © 2014 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Appendix 3. Ovid MEDLINE search strategy Undertaken for second update in 2014 1 exp “Wounds and Injuries”/ (691113) 2 (surgical adj wound$).mp. (35309) 3 or/1-2 (721540) 4 exp Tissue Adhesives/ (12636) 5 exp Fibrin Tissue Adhesive/ (3859) 6 tissue adhesive$.mp. (8303) 7 exp Cyanoacrylates/ (3816) 8 octylcyanoacrylate$.mp. (103) 9 Dermabond.mp. (114) 10 exp Enbucrilate/ (1496) 11 enbucrilate.mp. (1236) 12 butylcyanoacrylate$.mp. (119) 13 exp Acrylates/ (41800) 14 acrylate$.mp. (8636) 15 exp Bucrylate/ (280) 16 bucrylate$.mp. (309) 17 or/4-16 (51986) 18 3 and 17 (3841) 19 randomized controlled trial.pt. (366703) 20 controlled clinical trial.pt. (87802) 21 randomi?ed.ab. (318385) 22 placebo.ab. (143748) 23 clinical trials as topic.sh. (168638) 24 randomly.ab. (189528) 25 trial.ti. (114737) 26 or/19-25 (860509) 27 exp animals/ not humans.sh. (3901060) 28 26 not 27 (791299) 29 18 and 28 (276)

Appendix 4. Ovid EMBASE search strategy Undertaken for second update in 2014 1 exp Surgical Wound/ (3914) 2 (surgical adj wound$).mp. (8021) 3 or/1-2 (8021) 4 exp Tissue Adhesive/ (13492) 5 exp Fibrin Glue/ (6712) 6 (tissue adj adhesive$).mp. (5179) 7 exp Cyanoacrylate Derivative/ (1595) 8 exp Cyanoacrylic Acid Octyl Ester/ (340) 9 octylcyanoacrylate$.mp. (130) 10 Dermabond.mp. (343) 11 exp ENBUCRILATE/ (2847) 12 enbucrilate.mp. (2852) 13 butylcyanoacrylate$.mp. (190) 14 exp Acrylic Acid/ (3718) 15 acrylate$.mp. (6120) 16 exp Bucrilate/ (623) 17 bucrylate$.mp. (138)

Tissue adhesives for closure of surgical incisions (Review) 128 Copyright © 2014 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. 18 or/4-17 (23772) 19 3 and 18 (181) 20 Randomized controlled trials/ (47799) 21 Single-Blind Method/ (17933) 22 Double-Blind Method/ (114456) 23 Crossover Procedure/ (38119) 24 (random$ or factorial$ or crossover$ or cross over$ or cross-over$ or placebo$ or assign$ or allocat$ or volunteer$).ti,ab. (1300258) 25 (doubl$ adj blind$).ti,ab. (143978) 26 (singl$ adj blind$).ti,ab. (14109) 27 or/20-26 (1365111) 28 exp animals/ or exp invertebrate/ or animal experiment/ or animal model/ or animal tissue/ or animal cell/ or nonhuman/ (19982416) 29 human/ or human cell/ (14519180) 30 and/28-29 (14472519) 31 28 not 30 (5509897) 32 27 not 31 (1177074) 33 19 and 32 (35)

Appendix 5. EBSCO CINAHL search strategy Undertaken for second update in 2014 S29 S16 AND S28 S28 S17 or S18 or S19 or S20 or S21 or S22 or S23 or S24 or S25 or S26 or S27 S27 TX allocat* random* S26 (MH “Quantitative Studies”) S25 (MH “Placebos”) S24 TX placebo* S23 TX random* allocat* S22 (MH “Random Assignment”) S21 TX randomi* control* trial* S20 TX ( (singl* n1 blind*) or (singl* n1 mask*) ) or TX ( (doubl* n1 blind*) or (doubl* n1 mask*) ) or TX ( (tripl* n1 blind*) or (tripl* n1 mask*) ) or TX ( (trebl* n1 blind*) or (trebl* n1 mask*) ) S19 TX clinic* n1 trial* S18 PT Clinical trial S17 (MH “Clinical Trials+”) S16 S5 and S15 S15 S6 or S7 or S8 or S9 or S10 or S11 or S12 or S13 or S14 S14 TI Dermabond or AB Dermabond S13 TI enbucrilate or AB enbucrilate S12 TI bucrylate* or AB bucrylate* S11 TI acrylate* or AB acrylate* S10 TI butylcyanoacrylate* or AB butylcyanoacrylate* S9 TI octylcyanoacrylate* or AB octylcyanoacrylate* S8 TI cyanoacrylate* or AB cyanoacrylate* S7 TI tissue adhesive* or AB tissue adhesive* S6 (MH “Fibrin Tissue Adhesive”) S5S1orS2orS3orS4 S4 TI surgical wound* or AB surgical wound* S3 (MH “Surgical Wound Care”) S2 (MH “Surgical Wound”) S1 (MH “Tears and Lacerations”)

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 a 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 were used without appropriate safeguards (e.g. if envelopes were unsealed or non opaque or not sequentially numbered); alternation or rotation; date of birth; case record number; any other explicitly unconcealed procedure.

Unclear Insufficient information provided to permit a 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.

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 unlikely to introduce bias.

Tissue adhesives for closure of surgical incisions (Review) 130 Copyright © 2014 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. 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 Any one of the following. Insufﬁcient information provided to permit a 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. 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 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 enough to induce clinically relevant bias in observed effect size. ‘As-treated’ analysis done with substantial departure of the intervention received from that assigned at randomisation. • Potentially inappropriate application of simple imputation. •

Unclear Either of the following. Insufﬁcient reporting of attrition/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.

Tissue adhesives for closure of surgical incisions (Review) 131 Copyright © 2014 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. 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 pre-specified 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. • 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 • 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. •

WHAT’S NEW Last assessed as up-to-date: 12 March 2014.

1 August 2014 New citation required but conclusions have not changed New search, risk of bias updated for all studies into current format, revisions to text, no change to conclusions

1 August 2014 New search has been performed Second update: 19 new studies added: Amin 2009; Avsar 2009; Brown 2009; Chibbaro 2009; Eggers 2011; Jallali 2004; Kent 2014; Khan 2006; Kouba 2011; Krishnamoorthy 2009; Livesey 2009; Maloney 2013; Millan 2011; Mota 2009; Pronio 2011; Romero 2011; Sebesta 2004; Tierny 2009; van den Ende 2004. Five studies are awaiting assessment: Gennari 2004; Handschel 2006; Jan 2013; Nipshagen 2008; Yoon 2006

HISTORY Protocol ﬁrst published: Issue 3, 2003 Review ﬁrst published: Issue 2, 2004

Date Event Description

3 August 2010 Amended Contact details updated.

23 November 2009 New search has been performed New search and an additional 6 studies included in the review (Blondeel 2004; Dowson 2006; Ong 2002; Ridgway 2007; Sniezek 2007; Switzer 2003). Three studies were excluded (Jaibaji 2000; Orozco-Razon 2002; Steiner 2000) and 7 studies are awaiting assessment.

23 November 2009 New citation required and conclusions have changed Time to closure as an outcome measure was included in the review at the time of this update (post hoc). The review authors believe this to be a contributory factor towards both cost-effectiveness and satisfaction. An additional review author has joined the review team for this ﬁrst update

16 May 2002 New citation required and conclusions have changed Substantive amendment

Tissue adhesives for closure of surgical incisions (Review) 133 Copyright © 2014 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. CONTRIBUTIONSOFAUTHORS Jo Dumville: coordinated the second review update. Extracted data and checked quality of data extraction. Undertook and checked quality assessment. Analysed and interpreted data. Performed statistical analysis and checked quality of statistical analysis. Performed part of writing and editing the review. Approved final review update prior to submission. Secured funding. Is the guarantor of the review. Paul Coulthard: conceived, designed and co-ordinated the initial review. Undertook the searches for the original review and screened search results. Appraised quality and extracted data, wrote to trial authors for additional information. Managed data for the review and entered data into RevMan. Analysed and interpreted data and wrote the review. Undertook previous work that was the foundation of the current review. Approved final review update prior to submission. Philip Riley: analysed and interpreted data for the second review update. Performed statistical analysis and checked quality of statistical analysis. Performed part of writing and editing the review. Approved final review update prior to submission. Helen Worthington: analysed and interpreted data for all versions of the review. Performed statistical analysis and checked quality of statistical analysis. Performed part of writing and editing the review. screened search output for the second update. Approved final review update prior to submission Neil Patel: undertook quality assessment and checked quality assessment for the first update. Performed part of writing and editing the review. Approved final review update prior to submission. Marco Esposito: developed the search strategy, undertook the searches for the original review and advised on the review and commented on all updates of the review. Maarten van der Elst: Provided general advice on the original version of the review and undertook previous work that was the foundation of the current review. Oscar JF van Waes: screened search results against the inclusion criteria and retrieved papers, appraised quality and extracted data, wrote to trial authors for additional information for the original version of the review. Undertook previous work that was the foundation of the current review. James Darcey: screened search results against the inclusion criteria, retrieved papers, appraised quality and extracted data, wrote to trial authors for additional information for the for the first review update, managed data for the first review update and entered data into RevMan. Analysed and interpreted data and wrote the first review update.

Contributions of editorial base Nicky Cullum: edited the review, advised on methodology, interpretation and review content. Approved the ﬁnal review and review update prior to submission. Sally Bell-Syer: co-ordinated the editorial process. Advised on methodology, interpretation and content. Edited and copy edited the review and the updated review. Amanda Briant: ran the searches and edited the search methods section for the update.

DECLARATIONSOFINTEREST Jo C Dumville: None known Paul Coulthard: was a co-author in the Blondeel 2004 study. This study was also commercially supported by Ethicon. Philip Riley: None known Helen V Worthington: None known Neil Patel: None known Marco Esposito: None known Maarten van der Elst: None known

Tissue adhesives for closure of surgical incisions (Review) 134 Copyright © 2014 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Oscar J F van Waes: None known James Darcey: None known

SOURCES OF SUPPORT

Internal sources The University of Manchester, UK. • Renier de Graaf Hospital, Netherlands. • The Sahlgrenska Academy at Goteborg University, Sweden. •

External sources Swedish Medical Research Council (9495), Sweden. • The Hjalmar Svensson Research Fund, Sweden. • The National Institute for Health Research (NIHR) is the sole funder of the Cochrane Wounds Group, UK. •

DIFFERENCESBETWEENPROTOCOLANDREVIEW Time to closure as an outcome measure was included in the review post hoc as the review authors believe this to be a contributory factor towards both cost-effectiveness and satisfaction.

INDEX TERMS

Medical Subject Headings (MeSH)

∗Surgical Procedures, Operative; ∗Sutures; ∗Tissue Adhesives; ∗Wound Healing; Bandages; Cyanoacrylates; Enbucrilate; Randomized Controlled Trials as Topic; Surgical Wound Dehiscence [∗prevention & control]; Surgical Wound Infection [diagnosis]

MeSH check words Humans