Reducing Surgical Site Infections in Abdominal Surgery: Are Ring Retractors Effective? A Systematic Review and Meta-Analysis

Abstract

Background:

Surgical site infection (SSI) is one of the main causes of morbidity and death after surgical intervention. The use of physical barriers, including gloves, drapes, and gowns to reduce SSI after abdominal surgery is long-standing practice. The aim of this systematic review and meta-analysis was to determine the efficacy of ring incision retractors in reducing the risk of SSI in abdominal surgery.

Methods:

PubMed, CINAHL, the Cochrane randomized controlled trials (RCTs) Central Register, and the ISRCTN registry were searched for RCTs in which ring retractors were utilized to reduce SSI in abdominal surgery. The PRISMA guidelines and RevMan 5.3 were used for study selection and analysis. Additional subgroup analyses were performed, including trials using incision class (clean, clean-contaminated contaminated, and dirty) and trials that used the U.S. Centers for Disease Control and Prevention's SSI definition in their protocol.

Results:

A total of 19 RCTs inclusive of 4,229 patients were included. The utility of ring retractors in reducing SSI was suggested by an overall risk ratio of 0.62 (95% confidence interval 0.48–0.81). However, study heterogeneity caused by differences in effect size between individual RCTs, the non-standardized utilization of concomitant measures to reduce SSI, and an overall lack of high-quality trials was found.

Conclusion:

A reduction in SSI incidence with the use of ring retractors is suggested by the findings. However, this result must be treated with caution because in addition to some old trials poor quality and the large number of factors affecting SSI, there were substantial differences between trials in effect sizes in statistical heterogeneity. Further RCTs are needed to confirm this provisional finding.

Surgical site infections (SSIs) are common, complicating as many as 18% of abdominal operations [1]. Sequelae include prolonged hospital stays, a greater risk of morbidity, heightened cost to the healthcare systems, and an increased mortality rate [2,3]. Physical incision barriers may reduce SSIs by preventing bacterial colonization of the incision edges. Laboratory-based study of these devices has been successful; however, their efficacy in clinical practice remains unclear.

Two main types of ring retractors are available [4] (Table 1). Double-ring retractors; e.g. Alexis® (Applied Medical, Rancho Santa Margarita, CA), are cylindrical, clear plastic devices that are fitted through the incision and held in place by inner and outer rings. For laparoscopic procedures, the plastic sheath can be clipped or plugged using a GelPOINT advanced-access platform to maintain pneumoperitoneum [5] and, in some cases, to aid in specimen retrieval [6]. In the single-ring design, the outer ring is replaced by a plastic sheath that is extended to roll over the incision edges.

Table 1.

Main Ring Wound Retractor Designs

Design	Internal ring	External ring	Fixation mechanism
Single ring	Yes	No	Adhesive, clips
Double ring	Yes	Yes	Drape stretched between the two rings

Two previous systematic reviews of this topic were discovered during a literature search for this meta-analysis. Eleven of the 19 trials included in the present review also were included in these previous systematic reviews in which non-randomized clinical trials also were analyzed [4,7]. However, since their publication, several further trials have been reported, and their conclusions reached no consensus regarding the utility of retractor usage, with the conclusions ranging from support for retractor use to requests for further studies to answer the question. The aim of this review was to aid in the further evaluation of the effectiveness of ring retractors in halting SSIs, which may reduce morbidity and death.

Patients and Methods

The systematic review and meta-analysis were performed according to the PRISMA guidelines [8]. The protocol was not registered in an online database (Supplement 1).

Eligibility criteria

Databases were searched for randomized controlled trials (RCTs) that compared ring retractors with conventional retractors or surgical drapes in elective and emergency abdominal surgery. We defined abdominal surgery as laparotomy or any operation involving the abdominal and pelvic cavities via trans-abdominal wall incisions. An SSI was defined as the presence of ≥1 of the U.S. Centers for Disease Control and Prevention criteria listed in Table 2 [9]. Trials involving either single- or double-ring retractors were included. Non-randomized trials, systematic reviews, case reports, retrospective studies, case series, and studies involving abdominal surgery for penetrating abdominal injuries were excluded.

Table 2.

U.S. Centers for Disease Control and Prevention Definitions of Surgical Site Infections [ 9 ]

Superficial Incisional SSI	Infection occurs within 30 days after the operation and infection involves only skin or subcutaneous tissue of the incision and at least one of the following: • Purulent drainage, with or without laboratory confirmation, from the superficial incision • Organisms isolated from an aseptically obtained culture of fluid or tissue from the superficial incision. • At least one of the following signs or symptoms of infection: Pain or tenderness, localized swelling, redness, or heat and superficial incision is deliberately opened by surgeon, unless incision is culture negative. • Diagnosis of superficial incisional SSI by the surgeon or attending physician.
Deep Incisional SSI	Infection occurs within 30 days after the operation if no implant is left in place or within 1 year if implant is in place and the infection appears to be related to the operation and infection involves deep soft tissues (e.g., fascial and muscle layers) of the incision and at least one of the following: • Purulent drainage from the deep incision but not from the organ/space component of the surgical site. • A deep incision spontaneously dehisces or is deliberately opened by a surgeon when the patient has at least one of the following signs or symptoms: Fever (>38°C), localized pain, or tenderness, unless site is culture negative. • An abscess or other evidence of infection involving the deep incision is found on direct examination, during reoperation, or by histopathologic or radiologic examination. • Diagnosis of a deep incisional SSI by a surgeon or attending physician.
Organ/Space SSI	Infection occurs within 30 days after the operation if no implant is left in place or within 1 year if implant is in place and the infection appears to be related to the operation and infection involves any part of the anatomy (e.g., organs or spaces), other than the incision, which was opened or manipulated during an operation and at least one of the following: • Purulent drainage from a drain that is placed through a stab wound into the organ/space. • Organisms isolated from an aseptically obtained culture of fluid or tissue in the organ/space. • An abscess or other evidence of infection involving the organ/space that is found on direct examination, during reoperation, or by histopathologic or radiologic examination. • Diagnosis of an organ/space SSI by a surgeon or attending physician.

SSI = surgical site infection.

Database search

PubMed, Cochrane Library (central register of controlled trials), CINAHL Plus database, and the ISRCTN registry for unpublished registered RCTs were searched between December 19, 2014 and March 3, 2015. The search terms used for the different databases are listed in Table 3.

Table 3.

Databases and Search Strategies

Database	Search strategy
PubMed	“Surgical wound infection” or “surgical site infection” or “retractor” or “protector”
Cochrane RCTs Central Register	“Surgical wound infection” or “retractor” or “protector”
CINAHL Plus	“Surgical wound infection” or “surgical site infection” or “retractor” or “protector”
ISRCTN Registry	“Outcomes of surgical site infection” or “surgical wound infection”

Study selection

Article titles were scanned to identify studies that compared surgical ring retractors with conventional retractors or surgical drapes vs. no intervention for the outcome of SSI in patients undergoing abdominal surgery. Full abstract review was then carried out for potential studies. Eligible full articles were examined according to the inclusion and exclusion criteria. Figure 1 shows the flow diagram for the search and selection process.

FIG. 1.

Flow diagram for search and selections processes.

Data extraction

Data were extracted for all studies by the first author. Another author (K. Bashar) ran an independent risk of bias analysis and data accuracy verifications. In addition to the primary outcome, the flowing study design data were recorded: Participants, number and demographics, exclusion and inclusion criteria, secondary outcomes, intervention(s), trial arm, and control groups. Factors that may affect the primary outcome also were recorded. These include: SSI definition used, skin preparation before incision, details of pre-operative and post-operative antibiotic usage, type of operation, and additional preparation used to decrease bacterial load pre-, intra-, or post-operatively. In trials with more than two arms, we included the retractor and control or conventional arms only. Two authors (D.T. and S.A.) were contacted to obtain more information on their studies. Each responded with the requested information on length of followup and incision class.

The Cochrane Collaboration's tool for assessing the risk of bias, RevMan 5.3, was used to evaluate individual studies and to generate a color-coded risk-of-bias table. Where the author did not clearly specify how he or she addressed any of the risk domains, the denotation of “unclear risk” was given. We acknowledge that performance bias is always unclear because of the technical inability to blind the intervention group during an operative procedure.

Publication bias

Funnel plots were created for the visual assessment of any asymmetry in the distribution of studies by size and their effect measures around the average, which may highlight differences between studies and the possibility of publication bias.

Statistical approach

RevMan 5.3 also was used to analyze the pooled data from all studies. The risk ratio (RR) was calculated using a random-effect model with the 95% confidence interval (CI). The CI alignment in the forest plot visually aids in the assessment of consistency within studies. The χ² test was used to assess statistical heterogeneity. As the χ² test is lower powered when small trials are included, the I² test was used to quantify inconsistency across studies, and the results were evaluated against p values from the χ² test [10].

Additional analyses

Additional subgroup analyses were performed for:

1. Studies classified by incision type (clean, clean-contaminated, contaminated, and dirty) to evaluate the effect of the degree of contamination on the incidence of SSI.

2. Studies that used the CDC definition for SSI as a standardized method.

Results

Study selection

After the exclusion of two trials lacking randomization [11,12], 19 studies were included in our systematic review [11 –33]. Three of the fully reviewed studies required translation (Silva et al. in Spanish [32], Batz et al. in German [15], and Brunet et al. in French [11]). We first included all studies that used any of the CDC SSI criteria to perform pooled data analysis and then identified trials that used the CDC criteria for a subgroup analysis. In RCTs with more than two arms, we included only the retractor arm and the control or conventional arm.

Characteristics of the studies

The characteristics of the RCTs included are provided in Table 4. Factors that may influence the SSI rate in the studies are listed in Table 5. Four trials were multi-center, the remainder being single center. Differences in age and other inclusion/exclusion criteria were evident between the studies (see Table 4).

Table 4.

Characteristics of the Studies ^a

Trial	Methods	Participants	Interventions	Outcomes	Exclusion criteria	Number of participants	Control group
Batz et al., 1987 [15]	RCT	Colorectal cancer patients	Plastic ring drape vs. single ring	SSI	None specified	Ring group n = 25 Control group n = 25	Incision drape
Gamble and Hopton, 1984 [20]	RCT	Consecutive patients undergoing elective colonic surgery	Ring drape vs. single-ring retractor	SSI	None specified	Intervention n = 27 Control n = 29	No ring drape
Horiuchi et al., 2007 [21]^*	RCT	Non-traumatic gastrointestinal surgery between September 2003–March 2004	Alexis ring retractor vs. double-ring retractor	Primary outcome: SSI secondary outcomes: Operative time, blood loss, lowest intra-operative body temperature, intra-operative blood transfusion, highest post-operative blood glucose concentration, mean LOS	Severe adhesions (n = 5); prior laparotomy (n = 42); Long-term steroid use (n = 5); laparoscopic or minor surgery (n = 82); probable colon perforation (n = )	Intervention n = 111 Control n = 110	Without Alexis protection
Lauscher et al., 2012 [22]	RCT	Elective laparoscopic colorectal resection with mini-laparotomy aged 18 y or older	Wound protector vs. single-ring retractor	Primary outcome: SSI rates at 10 d, 1 mo and 6 mos post-operatively using modified CDC system Secondary outcomes: Post-operative bacterial colonialization of the abdominal wall, post-operative costs, LOS, cosmetic result	Emergency procedures, patients under 18 y of age	Intervention n = 46 Control n = 47	No wound protector
Cheng et al., 2012 [17]	RCT	Adult patients undergoing elective colorectal resection using standardized midline incision	Alexis vs. double-ring retractor	Primary: SSI within 30 d secondary: Post-operative pain	Laparoscopic approach, emergency re-laparotomy, contraindication to patient-controlled analgesia	Alexis n = 34 Control n = 30	Conventional wound protection
Reid et al., 2010 [23]^*	RCT (multi-center trial)	Adult patients undergoing open elective colorectal resection	Alexis vs. double-ring retractor	Primary: SSI within 30 d secondary: Wound protector performance as assessed by operating surgeons	Cognitive impairment, inability to consent, laparoscopic resection	Intervention n = 64 Control n = 66	No wound retractor
Mihaljevic et al., 2014 [13]^*	RCT- (multi-center trial)	Patients undergoing elective abdominal surgery requiring a median or transverse incision	Steri-drape wound (3M, St. Paul, MN) edge protector vs single-ring retractor	SSI on d 2, 4, 6, 8, 10, 14, 30, and 45	ASA grade ≥3, pregnant/breast feeding, laparotomy within the previous 60 d or planned re- laparotomy within the following 30 d, planned contaminated operation, abdominal surgery without a median or transverse incision, concurrent abdominal wall infection, severe pre-operative neutropenia, liver cirrhosis (Child-Pugh B or C), severe immunosuppression	Intervention n = 300 Control n = 294	Standard sterile drapes
Nyström et al., 1984 [24]	RCT (2 centers)	Adults undergoing elective colorectal surgery	Plastic wound ring drape (Opdrape, Triplus) vs. single-ring retractor	SSI within 30 d	Deferred surgery, change of operative plan, “unforeseen therapeutic situation”	Intervention n = 70 Control n = 70	No retractor
Nyström et al., 1980 [25]	RCT	Patients undergoing open appendectomy	Plastic wound drape (Vi-drape) vs. single-ring retractor	SSI	Patients ≤10 y old	Control n = 143 Intervention n = 132	No retractor
Ozer et al., 2006 [26]	RCT	Patients undergoing open appendectomy	Modified Alexis vs. double-ring retractor	SSI	None specified	Total 122	No retractor
Lee et al., 2009 [27]	RCT	All patients undergoing open appendectomy between 2006–2008 who provided informed consent	Alexis wound retractor vs. double-ring retractor	SSI within 21 d	Insulin-dependent diabetes mellitus Inability to follow-up because of geographic location	Intervention n = 61 Control n = 48	Standard retractors
Baier et al., 2012 [28]	RCT	All patients designated for laparotomy surgery	Plastic ring drape vs. single-ring retractor	SSI	Appendectomy, ostomy reduction, patents undergoing re-operation for non SSI-related reasons within 30 d	Intervention n = 99 Control n = 101	Standard cloth towels
Psaila et al., 1977 [29]	RCT	Patients undergoing abdominal surgery	Vi-Drape vs. single-ring retractor	SSI	None specified	Control n = 47 Adhesive drape n = 51 Ring drape n = 46 (not included in this review)	No ring drape
Redmond et al., 1994 [30]	RCT	All patients undergoing gastrointestinal surgery	Wound edge protector vs. single-ring retractor	SSI	None specified	Control n = 111 Intervention n = 102	No wound protector
Sookhai et al., 1999 [31]	RCT	Patients undergoing transabdominal gastrointestinal surgery	Impervious wound-edge retractor vs. single-ring retractor	SSI	None specified	Control n = 182 Intervention n = 170	No wound-edge protector
Silva et al., 2008 [32]	RCT	Patients ≥15 years old admitted to the emergency unit of the hospital San Martin de Quillota between May 2005 and October 2006 undergoing open appendectomy	Ring retractor vs. double-ring retractor	SSI	Other non-appendicitis pathology, normal biopsy results, wide laparotomy	Intervention n = 221 Control n = 212	No ring
Theodoridis et al., 2011 [18]	RCT	Pregnant women undergoing elective or emergency caesarean section	Alexis ring Retractor vs. Double Ring	SSI, ring retractor feasibility of use	Suspected chorioamnionitis	Total n = 231 Intervention n = 115 Control n = 116	Conventional Doyen Retractor
Pinkney et al., 2013 [19]	RCT ^*multi-center trial	Patients ≥18 y old undergoing elective or emergency laparotomy through any major incision between February 2010 and January 2012	Wound edge protection vs. single ring	Primary: Superficial SSI within 30 days Secondary: Quality of life (EQ5D3L tool), length of hospital stay, cost effectiveness, clinical efficiency of the device	Laparoscopic or Laparoscopic assisted procedures, previous laparotomy within three months	Total n = 749 Intervention n = 369 Control n = 366	No wound-edge protection
Williams et al., 1972 [33]	RCT	Patients undergoing midline or para-median laparotomy with opening of part of the bowel or biliary tract	Vi-drape abdominal wound protector vs. single-ring retractor	SSI	Death within 24 h of the procedure	Intervention n = 84 Control n = 83	No wound protector

Studies that used the Centers for Disease Control and Prevention (CDC) criteria for surgical site infection (SSI) in the sub-analysis are marked ^*.

ASA = American Society of Anesthesiogists, LOS = length of stay; RCT = randomized controlled trial.

Table 5.

Factors That May Influence Surgical Site Infection (SSI) Rates

	Antibiotics
Trial	Pre-operative	Intra/Post-operative	Skin preparation	SSI definition	Other preparation	Surgical intervention
Batz et al., 1987 [15]	Prophylactic (type unspecified)	None	None specified	Pus discharge; swabs from edge of the wounds	None	Colorectal tumor resection
Gamble et al., 1984 [20]	Metronidazole 200 mg/6 h for 48 h	Metronidazole 200 mg IV/8 h for 24 h and ampicillin 500 mg/8 h for 24 h	Povidone– iodine	Any wound discharge	Bowel preparation, 1 L of neomycin solution; 1% rectal wash	Colorectal surgery
Horiuchi et al., 2007 [21]	Upper GI: 1–2 g of ampicillin + cefazollin or flomoxef twice daily for 1–4 d Lower GI: 1–2 g of cefotiam, flomoxef, or cefmetazol twice daily for 3–4 d	If infection is suspected, the administration period was extended	Wound closed after the facia was sutured, incision margins were washed with 500 mL of saline	CDC criteria	Gauze placed around anastomosis to prevent contamination, 3,000 mL of saline wash given after upper GI surgery, 5,000 mL saline wash after colorectal surgery. A 10,000 mL saline wash was given if contamination suspected; 2 L of polyethylene for bowel preparation	Gastrointestinal surgery
Lauscher et al, 2012 [22]	3 g of ampicillin/ sulbactam 30 min before skin incision	None	Iodine for subcutaneous compartment	Modified CDC classification Physical examination performed at 10 d post-operatively, telephone follow-up at 1 and 6 mos	Bowel preparation with Klean-Prep for total colectomy, enema for subtotal and sigmoid resections; two abdominal drains were placed for coloproctectomy and rectal resection; one drain for sigmoid resection or left hemicolectomy; no drains for right hemicolectomy	Laparoscopic colorectal surgery
Cheng et al., 2012 [17]	Cefoperazone 2 g + metronidazole 500 mg IV	No		CDC criteria within 30 d follow-up	Bowel preparation for ultra-low anterior resection	Elective open colorectal resection
Reid et al., 2010 [23]	Cephazolin 2 g + metronidazole 500 mg IV	Repeated dose of cephazolin 2 g after 3 h if procedure still in progress	Betadine, left to dry before incision	CDC criteria within 30 d	Incision wash with warm normal saline after mass closure	Elective colorectal surgery
Mihaljevic et al., 2014 [13]	Prophylactic	No	Isopropyl, povidone-iodine based	CDC criteria, site examined on d 2, 4, 6, 8, 10 and 30 up to 45 d		Open elective abdominal surgery
Nyström et al., 1984 [24]	Two hospitals involved used doxycycline 400/200 mg ivtinidazole 800 mg	Doxycycline 100 mg for continued for 5 d	2 g of ampicillin powder applied to contaminated incisions	Pus emptying spontaneously or on incision		Elective colorectal surgery
Nyström 1980 [26]	Operating surgeon discretion(200 mg doxycycline)	Operating surgeon discretion(100 mg doxycycline for 5 d)	0.5% chlorhexidine in 70% alcohol	Spontaneous pus drainage or infection that required incision and drainage	After removal of the appendix, a double invagination of the stump in saline-moisture velvet pad	Open appendectomy
Ozer et al., 2006 [26]	1 g cefazolin 5 min prior to skin incision	None	Not specified	Hyperemia, swelling, pain, or pus drainage	Gloves and surgical instruments changed after appendectomyRubber drain inserted for cases with perforation or spillage of intestinal content and (removed d 2 or 3)	Open appendectomy
Lee et al., 2009 [27]	3.37 g of IV piperacillin- tazobactam 1 h prior to skin incision, 400 mg IVmoxifloxacin for penicillin allergic	Simple appendicitis for 24 hComplicated appendicitis: Antibiotic continued until patient was afebrile	Not specified	Infection requiring opening or antibiotics; any patient who was prescribed antibiotics after discharge from hospital	Not specified	Open appendectomy
Baier et al., 2012 [28]	Prophylactic	None	Standard disinfection	CDC criteria within 30 d	Colonic irrigation in 81 patients	Laparotomy (appendectomy or and ostomy reduction excluded)
Psaila et al., 1977 [29]	None	None	1:30 Savlon + Hibitane spirit	Erythema around sutures or incision edge with accompanying pyrexia, exudate, orpus discharge; wound breakdown	Adhesive skin drapes	Abdominal surgery
Redmond et al., 1994 [30]	Prophylactic	None	Standard preparation	Overt pus or culture-positive wound assessed at 5, 10, and 30 d	None	Gastrointestinal surgery
Sookhai et al., 1999 [31]	Prophylactic	None	Povidone–iodine	Purulent discharge, culture-positive discharge, pain/tenderness, localized swelling, erythema, or cellulitiswithin 30 d	None	Gastrointestinal surgery
Silva et al., 2008 [32]	Gentamicin 1.5 mg/kg + single dose of 1 g IV chloramphenicol	24 h or 3–5 d based on intraoperative assessment of the appendix	Antiseptic solution	CDC criteria	Electric shavers for hair remove hair + wash with soap	Open appendectomy
Theodoridis et al., 2011 [18]	Single dose of 1.5 g of cefuroxime	Repeated dose at 12 h	Povidone–iodine	CDC criteria, follow-up between 35–41 d post-op (information obtained from study author)	None	Caesarean section
Pinkney et al., 2013 [19]	Surgeon preference	Surgeon preference	Surgeon preference	CDC criteria at 30 d	None	Laparotomy
Williams et al., 1972 [33]	41 patients (20 prophylactically, 11 for LRTI, 4 for acute cholecystitis, 10 for miscellaneous reasons)	94 patients (9 continued their pre-op course)	Not specified	Erythema, exudate, or pus on d 7 and at 10 d	Wound lavage, drainage or local antibiotics at surgeon's discretion	Abdominal surgery involving alimentary tract opening

CDC = U.S. Centers for Disease Control and Prevention; IV = intravenous; LRTI = lower respiratory tract infection.

Participants

There were 4,229 patients in the 19 trials. The sample size ranged from 50 patients in the smallest, single-center trial (Batz et al. [15]) to 735 in the largest, multi-center trial (Pinkney et al. [19]).

Interventions

Twelve trials used single-ring retractors, whereas seven used double rings. One of the double-ring groups used a modified Alexis retractor (Ozer et al. [16]), whereas another made the retractor locally using two rings and surgical gloves (Silva et al. [32]). All others used standard double-ring retractors.

Outcomes

All trials designated SSI as the primary outcome. The SSI definition and the reporting and duration of post-operative follow-up were substantially different in various trials.

We performed a subgroup analysis inclusive of the nine trials that specified the CDC criteria as the definition of SSI (Fig. 6). One trial (Lauscher et al.[22]) used modified CDC criteria by physical examination at 10 d and telephone followup at 1 and 6 mos.

Control group

The designation of the control group was assigned to the cohort without intervention with ring retractors. Conventional drapes and retractors were used according to the different hospital protocols.

Reported confounders that may influence SSI rates

Antibiotic usage

There was substantial variation between trials in antibiotic usage in the pre-operative, intra-operative, and post-operative periods. There was not enough documentation for antibiotic usage comparing the control and intervention groups. Some trials did not specify their prophylactic antibiotic dose and timing (n = 7), and others used antibiotics according to surgeon preferences or intra-operative findings (n = 2). We summarize these findings in the first column of Table 5.

Preparation

The main skin preparation material was povidone-based solution. Other trials added extra skin closing techniques to decrease contamination. At least one trial used locally applied antibiotics. Additional preparations were applied to decrease the bacterial load: Bowel preparations, washout, and change of contaminated gloves in addition to other methods, especially in colorectal surgery.

Surgical interventions

Trials included patients undergoing appendectomy (n = 4), colorectal resection (n = 6), a gastrointestinal procedure (n = 3), a mixed abdominal operation (n = 5), and cesarean section (n = 1).

Risk of bias assessment

We conducted a risk-of-bias assessment for the individual studies using the Cochrane assessment tool within RevMan 5.3 to produce a color summary and graph (Figs. 2 and 3).

FIG. 2.

Risk of bias assessment for all trials combined. Color image is available online at www.liebertpub.com/sur

FIG. 3.

Summary of risk of bias assessment for individual trials. Color image is available online at www.liebertpub.com/sur

Random sequence generation

Most of the studies, especially the older trials, demonstrated an unclear to high risk of bias. One trial used alternating patient assignment to the two arms of the trial (Ozer et al. [16]).

Allocation concealment

Whereas six of the trials addressed this issue, in the majority of the trials, there was no clear information about blinding. We agree that blinding is technically difficult or impossible, given the nature of the tested intervention. However, the risk of favoring one arm is obvious.

Blinding of outcome assessment

Eight of the studies assigned blinded assessors.

Attrition bias

Six studies succeeded in preventing attrition, whereas four studies showed a high risk of bias because of number discrepancy. The rest of the studies showed unclear risk because of the number of patients lost to follow-up or, in one case, a crossover of patients between the two study arms after the study was started [13].

Selective reporting

The risk of selective reporting was unclear in all the trials.

Other bias

The risk of other biases is unclear in all studies because of the large numbers of confounders of SSI.

Publication bias

A funnel plot was created to assess the risk of publication bias (Fig. 4). The overall asymmetry can be explained by differences in study sizes and their effect measures. The asymmetry observed in the left lower aspect of the plot accommodates three studies (Cheng et al. [17], Ozer et al. [16] and Theodoradis et al. [18]). All these are small and favored the ring retractor, suggesting publication bias. However, all three trials were within the 95% CIs, suggesting low precision of effect estimate to cause this asymmetry. The RR (95% CI) were 0.07 (0.00, 1.16), 0.10 (0.01, 1.83), and 0.14 (0.01, 2.76), respectively.

FIG. 4.

Funnel plot of comparison of ring retractor vs. control for the outcome of surgical site infections.

Pooled data analysis

The studies ranged from having an unclear to a high risk of bias. We first calculated RR with 95% CI for the pooled data from all studies using random-effect models (Fig. 5). We then performed the same analysis for the trials that specified the use of CDC criteria to define SSI (Fig. 6). The nine studies that fulfilled the CDC criteria are more recent and addressed more risk-of-bias domains compared with the older studies. The overall results were RR 0.62, 95% CI 0.48, 0.81 and RR 0.59, 95% CI 0.40, 0.87 for the CDC sub-group with no marked differences between the two, both favoring ring retractors.

FIG. 5.

Forest plot of comparisons of all studies inclusive of ring retractor vs. control groups for the outcome surgical site infections.

FIG. 6.

Forest plot of comparisons between studies using Centers for Disease Control and Prevention criteria to define SSI with ring retractor vs. control with the outcome of surgical site infection.

Additional Analysis

We conduct sub-group analysis for studies by incision classification (Fig. 7). When no information was provided or surgery could not be categorized under any class, the data were not analyzed.

FIG. 7.

Forest plot of comparisons of studies by classification of incisions with ring retractor vs. control for the outcome surgical site infections. Color image is available online at www.liebertpub.com/sur

The ring retractor seems to be more effective in the dirty-incision group, but there was no substantial evidence of this, with the pooled effects almost touching the no-effect line. In the random effects model, the clean incision group had an RR of 0.83 (95% CI 0.21, 3.26), the clean contaminated group had an RR 0.79 (95% CI 0.54, 1.16), the contaminated group had an RR of 0.64 (95% CI 0.39, 1.06), and the dirty group had an RR of 0.63 (95% CI 0.40, 1.01).

Discussion

Summary of evidence

This systematic review of 4,229 patients involved in 19 RCTs provides some evidence that SSI is reduced by using ring retractors. The overall pooled effects favored ring retractors (RR 0.62; 95% CI 0.48, 0.81), and limiting the RCTs to those that used the CDC criteria produced a similar effect size (RR 0.59; 95% CI 0.40, 0.87).

The incision classification sub-group analysis suggests that retractors may be more effective in relation to the degree of contamination. Contaminated and dirty procedures showed more benefits from using ring retractors than did the clean and clean contaminated procedures, but the numbers are small (Fig. 7).

These results should be treated with caution, as there is significant heterogeneity in the pooled results. This may be attributable to differences in RCT quality (see risk of bias in Fig. 3), differences in populations of patients in terms of incision contamination (Fig. 7), and differential use of antibiotics in terms of type, administration time, duration, and management protocol of SSI (see Table 5).

Context of previous studies

Two previous systematic reviews collectively included 11 of the 19 studies we reviewed [4,7]. Another systematic review published on March 27, 2015, after our final search, includes 16 trials [14]. The conclusion of these reviews ranges from the need for more high-quality RCTs [7] to support for the use of the retractors in non-trauma abdominal operations [4] and suggesting strong evidence of site protector effectiveness compared with standard care [14].

In our review, there was significant variation between trial inclusion/exclusion criteria, skin preparations, use of antibiotics, bowel preparation before colorectal surgery, and length of followup. The SSI definition was not unified among trials, and only nine trials used the CDC criteria. Other trials used a wide variety of clinical findings to define SSI (see Table 4), with one trial [15] depending mainly on overt pus and microbiologic findings.

Interestingly, the trials extend from 1972 to 2014, and we observed significant improvements through time in terms of design, identifying confounders, and addressing risks of bias. Overall quality assessment of the trials ranges from moderate to a high risk of bias (see Figs. 2 and 3). The observed funnel plot asymmetry can be attributed to the number of small trials favoring ring retractors [16 –18], with one of the large trials touching the no-effect line [19].

Clinical implications of results

Our review suggests some benefit for ring retractors in the reduction of SSI. The device has been used clinically for a long time with a good safety profile, but we need more studies to test these results. We need to standardize antibiotic usage protocols, skin preparations, other infections confounders (e.g., diabetes mellitus, immunosuppressive drugs), and type of procedure.

Strengths and weaknesses of this review

The main weakness of our systematic review is the poor quality of most of the trials. Abdominal surgery includes a range of interventions that may add more SSI confounders, depending on the intra-operative findings, duration, and how much manipulation was done during surgery. In addition, combined procedures in five trials (upper and lower gastrointestinal), unclear randomization methods in 11 trials, and non-standardized usage of antibiotics, added to the fact there was no or scattered documentation for these confounders among the studies (with the exception of more recent trials), cast a shadow on SSI risk stratifications and hence create a need for caution dealing with these results.

The strength of this review is the inclusion of randomized trials only in accordance with PRISMA guidelines, which may reduce the risks of bias usually observed more in other types of studies. In addition, there was no language restriction, as we translated reports from German, Spanish, and French. The authors were directly contacted for further information when needed, and we got a 66% response rate.

Implications for future research

We believe that this result may provoke more research in this area, and more surgeons may be interested in exploring the potential effectiveness of ring reactors in reducing SSI.

Conclusion

Ring retractors appear to reduce the risk of SSI. However, this result must be treated with caution because most of studies we examined were of poor quality. In addition, there are substantial differences between trials in effect sizes, which results in statistical heterogeneity. Further RCTs are needed to confirm this provisional finding.

Footnotes

Author Disclosure Statement,Funding,and Acknowledgment

The authors declare no conflict of interest. The first author, K. Ahmed, is supported by National University of Ireland Galway.

We acknowledge the great help from research librarians in James Hardiman Library National University of Ireland Galway.

References

Altemeier

. Principles in the management of traumatic wounds and in infection control. Bull N Y Acad Med, 1979; 55:123–138.

National Nosocomial Infections Surveillance (NNIS) System. Data summary from October 1986–April 1998, issued June 1998. Am J Infect Control, 1998; 26:522–533.

Kirkland

, Briggs

, Trivette

, et al. The impact of surgical-site infections in the 1990s: Attributable mortality, excess length of hospitalization, and extra costs. Infect Control Hosp Epidemiol, 1999; 20:725–730.

Edwards

, Ho

, Tee

, et al. Wound protectors reduce surgical site infection: A meta-analysis of randomized controlled trials. Ann Surg, 2012; 256:53–59.

Nishimura

, Kawahara

, Honda

, et al. Totally laparoscopic anterior resection with transvaginal assistance and transvaginal specimen extraction: A technique for natural orifice surgery combined with reduced-port surgery. Surg Endosc, 2013; 27:4734–4740.

Almeida

, Musi

, Mazzoleni

, et al. Intraoperative frozen pathology during robot-assisted laparoscopic radical prostatectomy: Can ALEXIS trocar make it easy and fast?. J Endourol, 2013; 27:1213–1217.

Gheorghe

, Calvert

, Pinkney

, et al. Systematic review of the clinical effectiveness of wound-edge protection devices in reducing surgical site infection in patients undergoing open abdominal surgery. Ann Surg, 2012; 255:1017–1029.

Moher

, Liberati

, Tetzlaff

, Altman

, Group

. Preferred reporting items for systematic reviews and meta-analyses: The PRISMA statement. Int J Surg, 2010; 8:336–341.

Mangram

, Horan

, Pearson

, et al. Guideline for Prevention of Surgical Site Infection, 1999. Centers for Disease Control and Prevention (CDC) Hospital Infection Control Practices Advisory Committee. Am J Infect Control, 1999; 27:97–132.

10.

Higgins

JPT

, Green

. Cochrane Handbook for Systematic Reviews of Interventions. John Wiley & Sons. Hoboken, NJ. 2011.

11.

Brunet

, Bounoua

, Bugnon

, Gautier-Benoit

. [The use of ring drapes in abdominal surgery] (Fre). Lyon Chirurg, 1994:438–441.

12.

Maxwell

, Ford

, Peterson

, Richards

. Abdominal wound infections and plastic drape protectors. Am J Surg, 1969; 118:844–848.

13.

Mihaljevic

, Schirren

, Ozer

, et al. Multicenter double-blinded randomized controlled trial of standard abdominal wound edge protection with surgical dressings versus coverage with a sterile circular polyethylene drape for prevention of surgical site infections: A CHIR-Net trial (BaFO; NCT01181206). Ann Surg, 2014; 260:730–737.

14.

Mihaljevic

AéL

, Fau-Mihaljevic

, Müller

, et al. Wound edge protectors in open abdominal surgery to reduce surgical site infections: A systematic review and meta-analysis. (1932-6203 (electronic)).

15.

Batz

, Marcus

, Rothmund

. [Value of ring drape and incision drape to prevent wound infection in colorectal surgery: A controlled randomized study] (Ger). Aktuel-Chir [Internet], 1987; 22:149–152. Available at www.onlinelibrary.wiley.com/o/cochrane/clcentral/articles/271/CN-00200271/frame.html.

16.

Ozer

, Yigit

, Uzar

, et al. Can incision barrier decrease the risk of surgical site infection after appendectomy?. Saudi Med J, 2006; 27:1259–1261.

17.

Cheng

, Roslani

, Sehha

, et al. ALEXIS O-ring wound retractor vs conventional wound protection for the prevention of surgical site infections in colorectal resections (1). Colorectal Dis, 2012; 14:e346–e351.

18.

Theodoridis

, Chatzigeorgiou

, Zepiridis

, et al. A prospective randomized study for evaluation of wound retractors in the prevention of incision site infections after cesarean section. Clin Exp Obstet Gynecol, 2011; 38:57–59.

19.

Pinkney

, Calvert

, Bartlett

, et al. Impact of wound edge protection devices on surgical site infection after laparotomy: Multicentre randomised controlled trial (ROSSINI Trial). BMJ, 2013; 347:f4305.

20.

Gamble

, Hopton

. Plastic ring wound drapes in elective colorectal surgery. J R Coll Surg Edinb, 1984; 29:232–233.

21.

Horiuchi

, Tanishima

, Tamagawa

, et al. Randomized, controlled investigation of the anti-infective properties of the Alexis retractor/protector of incision sites. J Trauma, 2007; 62:212–215.

22.

Lauscher

, Grittner

, Stroux

, et al. Reduction of wound infections in laparoscopic-assisted colorectal resections by plastic wound ring drapes (REDWIL)? A randomized controlled trial. Langenbeck's Arch Surg, 2012; 397:1079–1085.

23.

Reid

, Pockney

, Draganic

, Smith

. Barrier wound protection decreases surgical site infection in open elective colorectal surgery: A randomized clinical trial. Dis Colon Rectum, 2010; 53:1374–1380.

24.

Nystrom

, Broome

, Hojer

, Ling

. A controlled trial of a plastic wound ring drape to prevent contamination and infection in colorectal surgery. Dis Colon Rectu. 1984; 27:451–453.

25.

Nystrom

, Brote

. Effects of a plastic wound drape on contamination with enterobacteria and on infection after appendicectomy. Acta Chir Scand, 1980; 146:65–70.

26.

Ozer

, Yigit

, Uzar

, et al. Can incision barrier decrease the risk of surgical site infection after appendectomy?. Saudi Med J, 2006; 27:1259–1261.

27.

Lee

, Waxman

, Taylor

, Yim

. Use of wound-protection system and postoperative wound-infection rates in open appendectomy: A randomized prospective trial. Arch Surg, 2009; 144:872–875.

28.

Baier

, Kiesel

, Kayser

, et al. Ring drapes do not protect against surgical site infections in colorectal surgery: A randomised controlled study. Int J Colorectal Dis, 2012; 27:1223–1228.

29.

Psaila

, Wheeler

, Crosby

. The role of plastic wound drapes in the prevention of wound infection following abdominal surgery. Br J Surg, 1977; 64:729–732.

30.

Redmond

, Meagher

, Kelly

. Use of an impervious wound-edge protector to reduce the postoperative wound-infection rate. Br J Surg, 1994; 81:[1811 p.]. Available at http://onlinelibrary.wiley.com/o/cochrane/clcentral/articles/523/CN-00830523/frame.html

31.

Sookhai

, Redmond

, Deasy

. Impervious wound-edge protector to reduce postoperative wound infection: A randomised, controlled trial. Lancet, 1999; 353:1585.

32.

Silva

, Vargas

, Moreno

, Becerra

. Utilidad del retractor elástico abdominal para disminuir el riesgo de infección de herida operatoria en apendicitis aguda. Rev Chil Cirurg, 2008; 60:527–533.

33.

Williams

, Oates

, Brown

, et al. Abdominal wound infections and plastic wound guards. Br J Surg, 1972; 59:142–146.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.03 MB

0.00 MB