Negative pressure wound therapy for surgical wounds healing by primary closure

Gill Norman; Chunhu Shi; En Lin Goh; Elizabeth MA Murphy; Adam Reid; Laura Chiverton; Monica Stankiewicz; Jo C Dumville

doi:10.1002/14651858.CD009261.pub7

. 2022 Apr 26;2022(4):CD009261. doi: 10.1002/14651858.CD009261.pub7

Negative pressure wound therapy for surgical wounds healing by primary closure

Gill Norman ^1,^✉, Chunhu Shi ¹, En Lin Goh ², Elizabeth MA Murphy ³, Adam Reid ⁴, Laura Chiverton ⁵, Monica Stankiewicz ⁶, Jo C Dumville ¹

Editor: Cochrane Wounds Group

PMCID: PMC9040710 PMID: 35471497

Abstract

Background

Indications for the use of negative pressure wound therapy (NPWT) are broad and include prophylaxis for surgical site infections (SSIs). Existing evidence for the effectiveness of NPWT on postoperative wounds healing by primary closure remains uncertain.

Objectives

To assess the effects of NPWT for preventing SSI in wounds healing through primary closure, and to assess the cost‐effectiveness of NPWT in wounds healing through primary closure.

Search methods

In January 2021, we searched the Cochrane Wounds Specialised Register; the Cochrane Central Register of Controlled Trials (CENTRAL); Ovid MEDLINE (including In‐Process & Other Non‐Indexed Citations); Ovid Embase and EBSCO CINAHL Plus. We also searched clinical trials registries and references of included studies, systematic reviews and health technology reports. There were no restrictions on language, publication date or study setting.

Selection criteria

We included trials if they allocated participants to treatment randomly and compared NPWT with any other type of wound dressing, or compared one type of NPWT with another.

Data collection and analysis

At least two review authors independently assessed trials using predetermined inclusion criteria. We carried out data extraction, assessment using the Cochrane risk of bias tool, and quality assessment according to Grading of Recommendations, Assessment, Development and Evaluations methodology. Our primary outcomes were SSI, mortality, and wound dehiscence.

Main results

In this fourth update, we added 18 new randomised controlled trials (RCTs) and one new economic study, resulting in a total of 62 RCTs (13,340 included participants) and six economic studies. Studies evaluated NPWT in a wide range of surgeries, including orthopaedic, obstetric, vascular and general procedures. All studies compared NPWT with standard dressings. Most studies had unclear or high risk of bias for at least one key domain.

Primary outcomes

Eleven studies (6384 participants) which reported mortality were pooled. There is low‐certainty evidence showing there may be a reduced risk of death after surgery for people treated with NPWT (0.84%) compared with standard dressings (1.17%) but there is uncertainty around this as confidence intervals include risk of benefits and harm; risk ratio (RR) 0.78 (95% CI 0.47 to 1.30; I² = 0%). Fifty‐four studies reported SSI; 44 studies (11,403 participants) were pooled. There is moderate‐certainty evidence that NPWT probably results in fewer SSIs (8.7% of participants) than treatment with standard dressings (11.75%) after surgery; RR 0.73 (95% CI 0.63 to 0.85; I² = 29%). Thirty studies reported wound dehiscence; 23 studies (8724 participants) were pooled. There is moderate‐certainty evidence that there is probably little or no difference in dehiscence between people treated with NPWT (6.62%) and those treated with standard dressing (6.97%), although there is imprecision around the estimate that includes risk of benefit and harms; RR 0.97 (95% CI 0.82 to 1.16; I² = 4%). Evidence was downgraded for imprecision, risk of bias, or a combination of these.

Secondary outcomes

There is low‐certainty evidence for the outcomes of reoperation and seroma; in each case, confidence intervals included both benefit and harm. There may be a reduced risk of reoperation favouring the standard dressing arm, but this was imprecise: RR 1.13 (95% CI 0.91 to 1.41; I² = 2%; 18 trials; 6272 participants). There may be a reduced risk of seroma for people treated with NPWT but this is imprecise: the RR was 0.82 (95% CI 0.65 to 1.05; I² = 0%; 15 trials; 5436 participants). For skin blisters, there is low‐certainty evidence that people treated with NPWT may be more likely to develop skin blisters compared with those treated with standard dressing (RR 3.55; 95% CI 1.43 to 8.77; I² = 74%; 11 trials; 5015 participants). The effect of NPWT on haematoma is uncertain (RR 0.79; 95 % CI 0.48 to 1.30; I² = 0%; 17 trials; 5909 participants; very low‐certainty evidence). There is low‐certainty evidence of little to no difference in reported pain between groups. Pain was measured in different ways and most studies could not be pooled; this GRADE assessment is based on all fourteen trials reporting pain; the pooled RR for the proportion of participants who experienced pain was 1.52 (95% CI 0.20, 11.31; I² = 34%; two studies; 632 participants).

Cost‐effectiveness

Six economic studies, based wholly or partially on trials in our review, assessed the cost‐effectiveness of NPWT compared with standard care. They considered NPWT in five indications: caesarean sections in obese women; surgery for lower limb fracture; knee/hip arthroplasty; coronary artery bypass grafts; and vascular surgery with inguinal incisions. They calculated quality‐adjusted life‐years or an equivalent, and produced estimates of the treatments' relative cost‐effectiveness. The reporting quality was good but the evidence certainty varied from moderate to very low. There is moderate‐certainty evidence that NPWT in surgery for lower limb fracture was not cost‐effective at any threshold of willingness‐to‐pay and that NPWT is probably cost‐effective in obese women undergoing caesarean section. Other studies found low or very low‐certainty evidence indicating that NPWT may be cost‐effective for the indications assessed.

Authors' conclusions

People with primary closure of their surgical wound and treated prophylactically with NPWT following surgery probably experience fewer SSIs than people treated with standard dressings but there is probably no difference in wound dehiscence (moderate‐certainty evidence). There may be a reduced risk of death after surgery for people treated with NPWT compared with standard dressings but there is uncertainty around this as confidence intervals include risk of benefit and harm (low‐certainty evidence). People treated with NPWT may experience more instances of skin blistering compared with standard dressing treatment (low‐certainty evidence). There are no clear differences in other secondary outcomes where most evidence is low or very low‐certainty. Assessments of cost‐effectiveness of NPWT produced differing results in different indications. There is a large number of ongoing studies, the results of which may change the findings of this review. Decisions about use of NPWT should take into account surgical indication and setting and consider evidence for all outcomes.

Plain language summary

Dressings that use negative pressure for closed surgical wounds

Key messages

Negative pressure wound therapy (NPWT) probably results in fewer surgical site infections (SSIs) than standard dressings in people with closed wounds after surgery.

NPWT probably makes no difference to the proportion of people with wound reopening (dehiscence) after surgery and may make little or no difference to the number of people who die.

NPWT may increase the number of people with skin blistering after surgery but may make little or no difference to other outcomes.

The cost‐effectiveness of NPWT and how certain we are about this depends on the type of surgery.

What are surgical wounds healing by primary closure?

Surgical wounds healing by primary closure are incisions created by surgery where the edges have been brought together, usually by using stitches or staples. Most surgical wounds heal in this way. A potential complication of surgery is SSI, an infection at the site of a surgical wound. The proportion of people who develop an SSI after surgery can be as high as 40%. An SSI can cause pain and discomfort, as well as increasing a person’s length of hospital stay and cost of treatment.

What did we want to find out?

NPWT is a sealed wound dressing attached to a vacuum pump which sucks fluid away from the wound. This may assist with wound healing and reduce risk of infection. We wanted to find out whether NPWT was better compared with standard wound dressings (usually gauze and tape) for treating people who had had surgery and had wounds which had been closed. We were interested in complications including SSI; wound reopening (dehiscence) and death for any reason. We also looked at several other outcomes including the need for another operation, the need to be admitted to hospital again, pain, quality of life, as well as some specific types of complications (haematoma (an accumulation of blood under the skin), seroma (an accumulation of clear fluid under the skin), skin blisters).

We also wanted to find out whether NPWT was cost‐effective for treating people who had closed surgical wounds.

What did we do?

We searched for randomised controlled trials (clinical studies where the treatment people receive is chosen at random). This type of study design provides the most reliable evidence about the effects of a treatment. We searched for studies that compared any type of NPWT with standard dressings in people who had had surgery and had a wound which had been closed. We compared and summarised their results, and rated our confidence in the evidence.

What did we find?

We found 62 studies which compared NPWT with standard dressings and looked at surgical site complications. A variety of NPWT systems was used. A total of 13,340 people have been included in this review. A wide variety of surgeries was included such as knee and hip operations, caesarean sections, operations for broken bones and abdominal surgeries. There were more women than men included in the review because several large trials included only women having caesarean sections. Most of the people included in the review live in North America, Europe or Australasia.

Eleven studies (6384 people) reported on risk of death and found that there may be a lower risk with NPWT compared with standard dressings but this is not clear. Forty‐four studies (11,403 people) looking at SSI were combined, and found that NPWT probably reduced the risk of SSI compared with standard dressings. Twenty‐three studies (8724 people) found that there is probably little or no difference in wound reopening between NPWT and standard dressings. For most other outcomes, the evidence showed that there may not be clear differences between the treatments, or that we are uncertain about the true effect of the treatments. The exception was skin blistering where NPWT may increase the proportion of people who experience this after surgery.

Six cost‐effectiveness studies were included in the review. These studies looked at women who had had caesarean sections, people with lower limb fractures, knee and hip surgeries, vascular surgery and heart surgery. All these studies used clinical information from trials included in this review. NPWT is probably cost‐effective for caesarean section wounds in obese women and probably not cost‐effective for fracture surgery wounds but we are less sure about its cost‐effectiveness in the other types of surgery.

What limited our confidence in the evidence?

Our confidence in the evidence was limited by different reasons for different outcomes. Given the small number of people who died, the results for death are likely to change with more evidence. For SSI, approximately half the people were in studies using methods likely to introduce errors. For wound reopening and most other outcomes, our confidence was reduced by a combination of these reasons. For skin blistering, our confidence was reduced by differences between the studies as well as study methods.

How up to date is this review?

This review is up to date to January 2021.

Summary of findings

Summary of findings 1. Negative pressure wound therapy compared with standard dressing for surgical wounds healing by primary closure.

Negative pressure wound therapy compared with standard dressing for surgical wounds healing by primary closure
Patient or population: adult patients with surgical wounds healing by primary closure Setting: general surgical, orthopaedic or obstetric wards in acute‐care hospitals Intervention: negative pressure wound therapy (NPWT) Comparison: standard dressing
Outcomes	*Anticipated absolute effects (95% CI)**		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with standard dressing	Risk with NPWT
Mortality (proportion of participants dying in each group at follow‐up of between 30 days and six months)	Study population		RR 0.78 (0.47 to 1.30)	6384 (11 studies)	⊕⊕⊝⊝ Low¹	There may be a reduced risk of death after surgery for people treated with NPWT compared with standard dressings but there is uncertainty around this as confidence intervals include risk of benefit and harm.
	12 per 1000	3 fewer deaths per 1000 people (6 fewer to 4 more)
Surgical site infection (proportion of participants in each group with SSI; follow‐up of 30 days except where other time points specified as primary outcome measure in study)	Study population		RR 0.73 (0.63 to 0.85)	11403 (44 studies)	⊕⊕⊕⊝ Moderate²	Treatment with NPWT after surgery probably decreases the incidence of surgical site infection compared with a standard dressing.
	117 per 1000	31 fewer SSI per 1000 people (10 fewer to 43 fewer)
Dehiscence (proportion of participants in each group with wound dehiscence; follow‐up of 30 days except where other time points specified as primary outcome measure in study)	Study population		RR 0.97 (0.82 to 1.16)	8724 (23 studies)	⊕⊕⊕⊝ Moderate³	There is probably little or no difference in wound dehiscence between people treated with NPWT and standard dressings after surgery.
	70 per 1000	2 fewer dehiscence per 1000 people (13 fewer to 11 more)
Reoperation (proportion of participants in each group requiring reoperation for reasons related to wound; follow‐up of 30 days except where other time points specified as primary outcome measure in study)	Study population		1.13 (0.91 to 1.41)	6272 (18 studies)	⊕⊕⊝⊝ Low³	There may be a higher incidence of reoperation in people treated with NPWT compared with standard dressings but there is uncertainty around this as confidence intervals include risk of benefit and harm.
	50 per 1000	6 more reoperations per 1000 people (4 fewer to 20 more)
Seroma (proportion of participants in each group with seroma; follow‐up of 30 days except where other time points specified as primary outcome measure in study)	Study population		RR 0.82 (0.65 to 1.05)	5436 (15 studies)	⊕⊕⊝⊝ Low⁴	There may be a lower incidence of seroma in people treated with NPWT compared with standard dressings but there is uncertainty around this as confidence intervals include risk of benefit and harm.
	43 per 1000	8 fewer seroma per 1000 people (15 fewer to 2 more)
Haematoma (proportion of participants in each group with haematoma; follow‐up of 30 days except where other time points specified as primary outcome measure in study)	Study population		RR 0.79 (0.48 to 1.30)	5909 (17 studies)	⊕⊝⊝⊝ Very low⁵	It is uncertain what the effect of NPWT compared with standard dressing is for incidence of haematoma.
	14 per 1000	3 fewer haematoma per 1000 people (7 fewer to 4 more)
Skin blisters (proportion of participants in each group with at least one skin blister; follow‐up of 30 days except where other time points specified as primary outcome measure in study)	Study population		RR 3.55 (1.43 to 8.77)	5015 (11 studies)	⊕⊕⊝⊝ Low⁶	NPWT may increase the risk of developing skin blisters compared with a standard dressing.
	19 per 1000	48 more blistering cases per 1000 people (8 more to 146 more)
*The risk in the intervention group (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; NPWT: negative pressure wound therapy; RR: risk ratio
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate; the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited; the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate; the true effect is likely to be substantially different from the estimate of effect.

Study	Wounds characteristics	Comparison	Time points	Mortality	SSI	Dehiscence	Note
Andrianello 2020	Pancreatic resection	Group A: PICO dressing Group B: gauze/standard surgical dressing	7, 30 and 90 days	Stated that 90‐day mortality recorded but not reported	Group A: 26/46 (4 superficial, 1 deep, 21 organ space) Group B: 27/49 (3 superficial, 3 deep, 21 organ space	‐	Study analysed number of non‐organ space SSIs.
Bertges 2021	Vascular groin surgery	Group A: ciNPT Group B: standard sterile gauze dressing	30 days	Mortality within 30 days: Group A: 1/115 Group B: 1/119	Group A: 17/115 patients(14/115 superficial; 3 deep) Group B: 16/119 patients (15/119 superficial; 1/119 deep)	Group A: 17/115 Group B: 17/119	A small number of participants in each group had bilateral wounds but data analyses per patient.
Bobkiewicz 2018 abstract	Stoma reversal surgery	Group A: ciNPWT Group B: standard dressing	Not reported	‐	Group A: 2/15 Group B: 4/15	"In the standard dressing group the incidence of wound dehiscence was higher".	Superficial SSI defined according to CDC
Bueno‐Lledo 2021	Hernia repair	Group A: PICO dressing Group B: Mepore dressing	30 days	‐	Group A: 0/72 Group B: 6/74	Group A: 2/72 Group B: 4/74
Chaboyer 2014	Caesarean section in obese women	Group A: PICO dressing Group B: Comfeel dressing	1, 2, 3, and 4 weeks post‐surgery	‐	Group A: 10/44 Group B: 12/43	‐	‐
Crist 2014	Open reduction and internal fixation of hip, pelvis, and acetabular fracture surgery	Group A: NPWT Group B: standard gauze dressing	12 months	‐	Group A: 5/49 Group B: 2/42	‐	‐
Crist 2017	Open reduction internal fixation (ORIF) for acetabular fractures	Group A: NPWT Group B: standard gauze dressing	10 to 21 days, 6 weeks, 12 weeks, and every 6 to 8 weeks thereafter until bony union occurred	‐	Group A: 5/33 Group B: 2/33 completed‐case analysis ‐ 5 lost after randomisation but group allocation not known	‐	Infection defined as "deep infection"
Darwisch 2020	Cardiac surgery ‐ median sternotomy	Group A: Group B:	Not reported	‐	Reported no sig difference in number of SSI in people with BMI < 35 (P = 0.622) or >/= 35 (P = 0.2926), n/N not reported	‐
DiMuzio 2017 Abstract	Groin wounds	Group A (59, high risk): NPWT dressing Group B (60, high risk): standard gauze dressing Group C (21, low risk): standard gauze dressing	30 days	‐	Group A: 6/59 Group B: 15/60 Group C: 1/21	Group A: 8.5% Group B: 18.3% Group C: 4.8%	Contacted authors for full text Group C not included in data analysis due to baseline heterogeneity
Engelhardt 2016	Groin wound	Group A: NPWT Group B: conventional dressing	5 and 42 days	‐	Group A: 9/64 Group B: 19/68	‐	‐
Flynn 2020	Laparotomy	Group A: PICO Group B: conventional dressing	7 days plus further follow up	‐	Group A: 13/96 Group B:14/92	Group A: 7/96 (1 fascial) Group B: 9/92 (1 fascial)
Fogacci 2019	Breast surgery	Group A: PICO Group B: standard dressing	"long‐term follow up"	‐	Group A: 0/50 Group B: 5/50	‐
Galiano 2018	Breast surgery	Group A: NPWT, 199 wounds Group B: standard dressings, 199 wounds	21 days	‐	Group A: 4/199 Group B: 6/199	Group A: 32/199 Group B: 52/199
Giannini 2018	Hip and knee prosthetic revision	Group A single use NPWT (PICO) Group B povidone‐iodine gauze and patch wound dressing	7 days	‐	The severity of wound infection measured by the ASEPSIS score (higher score = worse wound healing; a score > 10 = the increasing probability and severity of infection) mean (SD) of the score: 3.0 (1.89) in Group A; 5.1 (3.89) in Group B	‐	‐
Gillespie 2015	Primary hip arthroplasty	Group A: PICO dressing Group B: Comfeel dressing	30 days and 6 weeks post‐surgery	‐	Group A: 2/35 Group B: 3/35	Group A: 1/35 Group B: 1/35	QoL reported in Heard 2017.
Gillespie 2021	Caesarean	Group A: PICO dressing Group B: standard dressing	30 days	Group A: 0/1017 Group B: 0/1018	Group A: 75/1017 (70 superficial, 4 deep, 1 organ space) Group B: 99/1018 (93 superficial, 6 deep, 0 organ space)	Group A: 108/1017 Group B: 103/1018
Gok 2019	Abdominal surgery	Group A: NPWT, KCI Prevena incision management system Group B: standard dressing Group C: aspiration drainage	not stated; the study was terminated at day 5 postoperatively	One death was reported but the group was not clearly reported.	Stated that:"surgical site infection was detected five times less in the negative‐pressure group, compared to the standard dressing group. It is also 3.5‐fold less compared to the aspiration drainage group".	Group 1: 1/20 Group B: 7/20 Group C: 6/20
Gombert 2018	Vascular surgery (groin) for PAD	Group A: NPWT (Prevena) Group B: Cosmopore dressing	30 days	‐	Group A: 13/98 Group B: 30/90	‐	‐
Gunatilake 2017	Caesarean	Group A: NPWT Group B: standard care dressing	42 ± 10 days postoperatively (days 1, 2, 6, 14, and 42)	‐	Group A: 1/39 Group B: 4/43	Group A: 1/39 Group B: 5/43	ITT: n = 92; 82 completed the study.
Hasselmann 2019a	Inguinal vascular surgery	Group A: NPWT Group B: standard dressing	90 days	Group A: 1/75 Group B: 1/79	SSI among patients with unilateral incisions CDC criteria Group A: 7/59 (6 superficial; 1 deep) Group B: 17/61(13 superficial; 2 deep; 2 organ space) ASEPSIS score Group A: 7/59 Group B: 18/61	Group A: 12/59 Group B: 7/61	Data in Hasselmann 2019a and Hasselmann 2019b were reported in the same publication. These data of this row are for those with unilateral incisions randomised individually. Mortality mentioned only in the study flow figure, rather than specifying as an outcome
Hasselmann 2019b	Inguinal vascular surgery	Group A: NPWT Group B: standard dressing	90 days	Group A: 1/24 participants Group B: 1/24 participants	SSI among patients with bilateral incisions Group A: 1/19 (5.3%) incisions Group B: 5/19 (26.3%) incisions all superficial ASEPSIS criteria Group A: 1/19 Group B: 4/19	For bilateral incisions: Group A: 2/19 Group B: 2/19	Data in Hasselmann 2019a and Hasselmann 2019b were reported in the same publication. These data of this row are for those with bilateral incisions randomised with a split‐body design.  Mortality (mentioned only in the study flow figure, rather than specifying as an outcome) Given the split‐body design, the mortality case occurred only in one person (with bilateral incisions).
Howell 2011	Knee arthroplasty	Group A: NPWT Group B: gauze dressing	Followed up for 12 months post‐surgery	‐	Group A: 1/24 Group B: 1/36	‐	‐
Hussamy 2017	Caesarean	Group A (222): NPWT Group B (219): standard dressing	30 days postdelivery	‐	Superficial SSI Group A: 20/222 Group B: 25/219 Organ SSI Group A: 1/222 Group B: 0/219 Deep Infections: Group A: 0/222 Group B: 0/219	Group A: 4/222 Group B: 1/219	Unable to contact authors
Hyldig 2019b	Caesarean	Group A (432): NPWT Group B (444): standard dressing	30 days after operation	‐	SSI Group A: 20/432 Group B: 41/444 Deep SSI Group A: 8/410 Group B: 9/417	Group A: 62/410 Group B: 69/417
Javed 2018	Open pancreaticoduodenectomy	Group A: NPWT Group B: standard closure	30 days after operation	‐	SSI Group A: 6/62 Group B: 19/61 Superficial SSI Group A: 4/62 Group B: 17/61 Deep SSI Group A: 2/62 Group B: 2/61	‐
Karlakki 2016	Total hip or knee arthroplasties	Group A: PICO dressing Group B: Comfeel dressing	1, 2, and 6 weeks post‐surgery	‐	Group A: 2/102 Group B: 6/107	‐	‐
Keeney 2019	Hip and knee total joint arthroplasty	Group A: iNPWT Group B: conventional wound dressing	7, 14 and 35 days after operations; 2 years	‐	Superficial and late wound infection rates Group A: 7/185 Group B: 8/213	‐	Additional data on return to operating rooms; and infection outcome were presented in Characteristics of included studies.
Kuncewitch 2017 Abstract	Pancreatectomy	Group A: NPWT Group B: standard surgical dressing	30 days post‐surgery follow‐up	‐	Superficial SSI Group A: 5/36 Group B: 6/37 Deep SSI Group A: 3/36 Group B: 2/37	Group A: 1/36 Group B: 2/37	Unable to contact authors
Kwon 2018	Vascular groin incisions (high risk)	Group A: NPWT Group B: standard gauze	30 days	‐	Any Group A: 6/59 Group B: 12/60 Major Group A: 5/59 Group B: 12/60	Any Group A: 1/59 Group B: 1/60 Major Group A: 0/59 Group B: 0/60
Lee 2017a	Great saphenous vein harvest	Group A: NPWT Group B: standard surgical dressing	Initial assessment: not specified; endpoint assessment: 6 weeks	‐	Group A: 0/31 Group B: 1/25	‐	2 participants died (sepsis; stroke). 2 participants were delirious and unable to complete QoL; all other objective evaluations were done (all 4 in NPWT)
Lee 2017b	High‐risk groin wounds	Group A: NPWT Group B: standard care	30 days and 90 days	Mortality within 90 days: Group A: 1/53 Group B: 2/49	In‐hospital SSI Group A: 1/53 Group B: 1/49 30‐day SSI Group A: 6/53 Group B: 9/49 90‐day SSI Group A: 7/53 Group B: 11/49	‐	Latest time point of SSI data used for analysis
Leitao 2020 Abstract	Laparotomy for gynaecologic malignancy	Group A: NPWT, KCI Prevena™ Customizable™ Incision Management System Group B: standard gauze	Postoperative 30 days	‐	Superficial SSI Group A: 44/223 Broup B: 41/221	Superficial dehiscence Group A: 30/223 Broup B: 25/221	The development of a wound complication was reported as a composite outcome including multiple types of events; but data were not reported for each complication separately (overall, 41 patients (18%) in Group A (90% CI 14.1% to 22.7%); and 38 (17%) in Group B (90% CI 13.0% to 21.4%). Data in main table obtained from author contact
Leon 2016 Abstract	Open colorectal surgery	Group A: NPWT Group B: usual dressing	15‐day and 30‐day evaluation	‐	Group A: 5/47 Group B: 10/34	‐	Unable to contact authors
Lozano‐Balderas 2017	Laparotomised patients with class III or IV (contaminated/dirty‐infected) surgical wounds	Group A: vacuum‐assisted closure Group B: primary closure Group C: delayed primary closure	Daily when in hospital or in a 30‐day period after surgery	‐	Group A: 0/25 Group B: 10/27 Group C: 5/29	‐	Group C (delayed primary closure) not included in data analysis due to irrelevant wounds
Manoharan 2016	Primary arthroplasty	Group A: NPWT Group B: conventional dry dressing	10 to 12 days post‐surgery	‐	‐	‐	‐
Martin 2019	Hepatectomy or pancreatectomy	Group A: NPWT (PICO) Group B: sterile island dressing	30 days then biweekly for 3 months	‐	Overall Group A: 3/20 Group B: 6/20 Deep space Group A: 2/20 Group B: 4/20 Superficial Group A: 1/20 Group B: 2/20	Group A: 1/20 Group B: 2/20 all mild to moderate ‐ all associated with SSI	‐
Masden 2012	Radial forearm free flap	Group A: NPWT Group B: dry dressing	Not clear	‐	Group A: 3/44 Group B: 5/37	Group A: 16/44 Group B: 11/37	‐
Murphy 2019	Colorectal resections	Group A: NPWT Group B: standard gauze dressing	30 days	Group A: 3/144 Group B: 2/140	Group A: 46/144 Group B: 48/140	‐
NCT00654641	Caesarean sections	Group A: Negative pressure wound closure Group B: Standard wound closure	6 weeks	no serious adverse events reported	‐	‐	The development of a wound complication was reported as a composite outcome including multiple types of events; data were not reported for each complication separately (overall, 15 patients in Group A (53.57%) and 10 in Group B (38.46%)
NCT01759381	Spinal surgery	Group A: NPWT Group B: standard wound closure	3 months	Group A: 0/11 Group B: 1/8	Group A: 2/11 Group B: 1/7	‐
NCT02309944	Laparotomy for suspected gynaecologic malignancy	Group A: NPWT Group B: standard wound closure	4 weeks	Within 12 weeks after surgery: Group A: 0/43 Group B: 0/38	SSI: Group A: 3/43 Group B: 2/38	Wound complication (Wound dehiscence or infection): Group A: 9/42 Group B: 12/37	SSI was reported by the original investigators as two separate types of adverse events: superficial wound infection requiring readmission, and superficial wound infection; and the investigators stated adverse event data were collected by non‐systematic assessment. We combined data of these two types of adverse events together for SSI in this table. The investigators also considered both wound dehiscence and infection into the composite outcome of wound complication (Group A 9/42; Group B 12/37). However, data for either part of the composite outcome were not available.
NCT02461433	Obese patients who underwent any elective open surgery	Group A: Prevena negative pressure wound therapy Group B: standard dressing	5 weeks	Group A: 0/1 Group B: 0/1	Group A: 0/1 Group B: 0/1	‐	Dehiscence, seroma and hematoma were reported as number of aggregate events.
NCT01759381	Spinal surgery	Group A: NPWT Group B: control without NPWT	3 months	Group A: 0/11 Group B: 1/8	Group A: 2/11 Group B: 1/7	‐	‐
Newman 2019	Total hip or knee replacements	Group A: ciNPWT Group B: standard silver dressing	12 weeks	‐	Group A: 0/79 Group B: 1/80	Dehiscence Group A: 1/79 Group B: 4/80
Nordmeyer 2016	Spinal fractures treated with internal fixation	Group A: PICO dressing Group B: standard dressing	Day 5 and day 10 after surgery	‐	‐	‐	‐
O'Leary 2017	Open abdominal surgery	Group A: PICO dressing Group B: transparent waterproof dressing	Day 4 and day 30 post‐surgery	‐	Group A: 2/24 Group B: 8/25	‐	‐
Pachowsky 2012	Hip arthroplasty	Group A: NPWT Group B: standard dressing	Day 5 and day 10 in postoperative period	‐	‐	‐	Very small sample size
Pauser 2016	Fractures of the femoral neck treated by hemiarthroplasty	Group A: NPWT Group B: standard dressing	Day 5 and day 10 after surgery	‐	‐	‐	Very small sample size
Pleger 2018	Groin wound	Group A: NPWT (n = 58 incisions) Group B: control dressing (n = 71 incisions)	Days 5 to 7 and 30 after surgery	‐	Group A: 1/58 Group B: 10/71	Superficial wound dehiscence Group A: 3/58 Group B: 4/71 Deep wound dehiscence with fat necrosis Group A: 1/58 Group B: 4/71	Unit of analysis error: 100 participants with 129 groin incisions
Ruhstaller 2017	Unplanned caesarean section	Group A: NPWT Group B: standard care	4 weeks post‐surgery	‐	Group A: 2/61 Group B: 4/58	‐	‐
Sabat 2016 Abstract	Groin wounds in vascular surgery	Group A: NPWT Group B: conventional dressing (gauze and Tegaderm)	4 months post‐surgery	‐	Group A: 2/30 Group B: 7/33	Group A: 3/30 Group B: 8/33	‐
Schmid 2018	Inguinal lymph node dissection	Group A: NPWT (Prevena) Group B: conventional compression bandages	14 days after surgery	Within‐person study; 1 death reported	–	–	Reported wound complication as a composite (unspecified outcome): complications of any grade in 23/31 on each side; grade 3 complications in 3/31 conventional and 6/31 NPWT wounds
Shen 2017	Open resection of intra‐abdominal neoplasms	Group A: PICO dressing Group B: Comfeel dressing	30 days after surgery	Group A: 3/132 Group B: 5/133	Group A: 26/132 Group B: 28/133	Group A: 3/132 Group B: 3/133	‐
Shim 2018	Reconstructive surgery for acute hand injuries	Group A: NPWT Group B: conventional dressing	1 month and 1 year	‐	Group A: 0/30 Group B: 1/21	Wound disruption Group A: 2/30 Group B: 0/21	‐
Stannard 2012	Tibial plateau, pilon, or calcaneus fracture	Group A: NPWT Group B: standard dressing	Not stated	‐	Group A: 14/144 Group B: 23/122	Group A: 12/139 Group B: 20/122	Unit of analysis error
Tanaydin 2018	Bilateral breast reduction mammoplasty	Group A: NPWT Group B: standard care (fixation strips)	21 days	‐	‐	Group A: 5/32 Group B: 10/32	32 participants served as their own control.
Tuuli 2017	Caesarean delivery	Group A: NPWT Group B: standard dressing	30 days	‐	Group A: 3/60 Group B: 2/60	Group A: 2/60 Group B: 0/60	‐
Tuuli 2020	Caesarean delivery	Group A: NPWT Group B: standard dressing	30 days	Group A: 0/806 Group B: 0/802	Group A: 31/806. Superficial/deep/organ space: 18/11/2 Group B: 29/802 Superficial/deep/organ space: 16/11/2	Group A: 11/806 Group B: 9/802	Trial stopped due to adverse skin reactions and futility
Uchino 2016	Ileostomy closure in patients who underwent surgery for ulcerative colitis	Group A: NPWT Group B: standard dressing	6 weeks	‐	Group A: 3/28 Group B: 1/31	‐
WHIST 2019a	Lower limb fracture	Group A: NPWT Group B: standard dressing	30 days 90 days	Planned analysis could not be conducted 3 months Group A: 12/745 Group B: 15/711 6 months Group A: 18/745 Group B: 19/711	Deep infection 30 days Group A: 45/770 Group B: 50/749 90 days (available case data) Group A: 72/629 Group B: 78/590	Dehisced but not deep SSI 30 days Group A: 2/714 Group B: 7/687 90 days Group A: 2/563 Group B: 2/525 14 of those with deep infection dehisced or deliberately opened.
Wierdak 2021	Ileostomy closure in patients who underwent surgery for colorectal cancer	Group A: NPWT, NANOVA negative‐pressure dressing Group B: customary care using sterile wound dressings	7 days and 14 days postoperatively	‐	Incidence of surgical site infections: Group A: 2/35 (5.7%) patients Group B: 8/36 (22.2%) patients	‐	Incidence of wound healing complications were reported: 3/35 (8.6%) patients Group A and 11/36 (30.6%) in Group B (P = 0.020)
Wihbey 2018	Caesarean delivery	Group A: NPWT Group B: standard dressing	1 week and 30 days follow‐up	–	Superficial Group A: 12/80 Group B: 8/81 Deep Group A: 0/80 Group B: 0/81 Organ Group A: 1/80 Group B: 4/81	Group A: 14/80 Group B: 13/81	‐
Witt‐Majchrzac 2015	Coronary artery bypass surgery	Group A: NPWT Group B: conventional dressing	6 weeks follow‐up	‐	Group A: 1/40 Group B: 7/40	Group A: 1/40 Group B: 1/40	‐

Study	Wound characteristics	Comparison	Time‐points	Reoperation	Readmission	Seroma	Haematoma	Skin blisters	Pain	Quality of Life	Notes
Andrianello 2020	Pancreatic resection	Group A: PICO dressing Group B: gauze/standard surgical dressing	7, 30 and 90 days	Group A: 10/46 Group B: 3/49* *calculated using flow chart. Reop = relaparotomy		Group A: 0/46 Group B: 6/49	Group A: 2/46 Group B: 1/49	‐	‐	‐
Bertges 2021	Vascular groin surgery	Group A: ciNPT Group B: standard sterile gauze dressing	30 days	Group A: 16/115 Group B: 16/119	Group A: 8/115 Group B: 11/119 (readmissions for wound infection)	Reported composite data for seroma and haematoma: Group A: 3/115 Group B: 1/119	Reported composite data for seroma and haematoma: Group A: 3/115 Group B: 1/119	‐	EQ‐5D: Group A (n = 115) mean pain score baseline = 1.8 (+/‐0.6) 14 days = 2.2 (+/‐ 0.6) 30 days = 2.3 (+/‐ 0.6) Group B (n = 119) mean pain score baseline = 1.7 (+/‐ 0.5) 14 days = 2.2 (+/‐ 0.5) 30 days 2.4 (+/‐ 0.6)	EQ‐5D: Group A (n = 115) mean QOL score baseline = 11.7 (+/‐ 1) 30 days = 12.1 (+/‐ 1) Group B (n = 119) mean QOL score baseline = 11.8 (+/‐ 1) 30 days = 12.2 (+/‐ 1)	EQ‐5D for QoL incorporated pain score.
Bobkiewicz 2018	Stoma reversal surgery	Group A: ciNPWT Group B: standard dressing	Not reported	‐	‐	‐	"In the standard dressing group the incidence of haematoma was higher".	‐	‐	‐	Superficial SSI defined according to CDC
Bueno‐Lledo 2021	Hernia repair	Group A: PICO dressing Group B: Mepore dressing	30 days	Group A: 1/72 Group B: 2/74 revision after discharge	Group A: 2/72 Group B: 6/74	Group A: 9/72 Group B: 10/74	Group A: 1/72 Group B: 2/74
Chaboyer 2014	Caesarean section in obese women	Group A: PICO dressing Group B: Comfeel dressing	1, 2, 3, and 4 weeks post‐surgery	Group A: 1/44 Group B: 1/43	‐	‐	Group A: 1/44 Group B: 4/43	Group A: 4/44 Group B: 0/43	‐	‐
Crist 2014	Open reduction and internal fixation of hip, pelvis, and acetabular fracture surgery	Group A: NPWT Group B: standard gauze dressing	12 months	‐	‐	‐	‐	‐	‐	‐
Crist 2017	Open reduction internal fixation for acetabular fractures	Group A: NPWT Group B: standard gauze dressing	10 to 21 days, 6 weeks, 12 weeks, every 6 to 8 weeks thereafter until bony union occurred	‐	‐	‐	‐	‐	‐	‐	Infection defined as "deep infection"
Darwisch 2020	Cardiac surgery ‐ median sternotomy	Group A: PICO dressing Group B: standard surgical dressing	Not reported	‐	‐	‐	‐	‐	‐	‐
DiMuzio 2017	Groin wounds	Group A (59, high risk): NPWT dressing Group B (60, high risk): standard gauze dressing Group C (21, low risk): standard gauze dressing	30 days	‐	Group A: 6.8% Group B: 16.7% Group C: 4.8%	‐	‐	‐	‐	‐	Contacted authors for full text Group C not included in data analysis due to baseline heterogeneity
Engelhardt 2016	Groin wound	Group A: NPWT Group B: conventional dressing	5 and 42 days	‐	‐	‐	‐	‐	‐	‐
Flynn 2020	Laparotomy	Group A: PICO Group B: conventional dressing	7 days plus further follow‐up	‐	‐	Group A: 0/96 Group B: 1/92	Group A: 0/96 Group B: 2/92	Group A: 2/96 Group B: 0/92	Group A: 0/96 Group B: 1/92
Fogacci 2019	Breast surgery	Group A: PICO Group B: standard dressing	"long term follow‐up"	‐	Group A: mean 3.78 (range 2‐8) readmission as outpatient Group B: mean 4.18 (range 2‐14)	‐	‐	‐	‐
Galiano 2018	Breast surgery	Group A: NPWT, 199 wounds Group B: standard dressings, 199 wounds	21 days	‐	‐	Group A: 0/199 Group B: 1/199	Group A: 2/199 Group B: 3/199	‐	‐	‐
Giannini 2018	Hip and knee prosthetic revision	Group A: single‐use NPWT (PICO) Group B: povidone‐iodine gauze and patch wound dressing	7 days	‐	‐	‐	‐	Group A: 6/50 Group B: 15/50	Pain at dressing change Group A: mean 2.6 (median 2, range 1‐6) Group B: mean 4.8 (median 5, range 2‐7)	‐
Gillespie 2015	Primary hip arthroplasty	Group A: PICO dressing Group B: Comfeel dressing	30 days and 6 weeks post‐surgery	‐	Group A: 4/35 Group B: 0/35	Group A: 3/35 Group B: 0/35	Group A: 3/35 Group B: 1/35	‐	‐	‐	QoL reported in Heard 2017
Gillespie 2021	Caesarean	Group A: PICO dressing Group B: standard dressing	30 days	Group A: 4/1017 Group B: 5/1018	Group A: 23/1017 Group B: 13/1018	Group A: 27/1017 Group B: 26/1018	Group A: 11/1017 Group B: 6/1018	Group A: 40/996 Group B: 23/983	Group A: 21/1017 Group B: 11/1018	‐	QoL presumed to be in the forthcoming economics paper
Gombert 2018	Vascular surgery (groin) for PAD	Group A: NPWT (Prevena) Group B: Cosmopore dressing	30 days	Group A: 5/98 Group B: 6/90	‐	‐	‐	‐	Assessed but not reported	‐
Gunatilake 2017	Caesarean	Group A: NPWT Group B: standard care dressing	42 ± 10 days postoperatively (days 1, 2, 6, 14, and 42)	Group A: 1/39 Group B: 6/43	‐	‐	‐	‐	Pain reductions at rest Group A: 39/46 Group B: 20/46 Pain reductions with incisional pressure Group A: 42/46 Group B: 25/46	‐	ITT: n = 92; 82 completed the study
Hasselmann 2019a	Inguinal vascular surgery	Group A: NPWT Group B: standard dressing	90 days	For unilateral incisions: Group A: 2/59 (3.4%) Group B: 4/61 (6.6%)	Data at 30 days for unilateral incisions: Group A: 10/59 (16.9%); Group B: 5/61 (8.2%)	For unilateral incisions: Group A: 13/59 (22.0%) Group B: 14/61 (23.0%)	For unilateral incisions: Group A: 1/59 (1.7%) Group B: 4/61 (6.6%)	‐	‐	Vascuquol‐6 preoperative Group A: 9.1 (n = 41) Group B: 10.6 (n = 43) 30 days Group A: 14.9 (n = 39) Group B: 15.3 (n = 42)	QoL data taken from Svensson‐Bjork 2020
Hasselmann 2019b	Inguinal vascular surgery	Group A: NPWT Group B: standard dressing	90 days	For bilateral incisions: Group A: 1/19 (5.3%) Group B: 1/19 (5.3%)	Data at 30 days for bilateral incisions: Group A: 3/19 (15.8%) Group B: 2/19 (10.5%)	For bilateral incisions: Group A: 3/19 (15.8%) Group B: 4/19 (21.1%)	For bilateral incisions: Group A: 0/19 (0%) Group B: 0/19 (0%)	‐	‐	‐
Howell 2011	Knee arthroplasty	Group A: NPWT Group B: gauze dressing	12 months post‐surgery	‐	‐	‐	‐	Group A: 15/24 Group B: 3/36	‐	‐
Hussamy 2017	Caesarean section	Group A: NPWT Group B: standard dressing	30 days post‐delivery	Group A: 14/222 Group B: 10/219	Group A: 12/222 Group B: 9/219	‐	‐	‐	‐	‐	Abstract reported 13 participants in Group A were readmitted. Paper reported 12.
Hyldig 2019b	Caesarean section	Group A: NPWT Group B: standard dressing	30 days after operation	‐	Mentioned in ClinicalTrials.gov but outcome data not reported	‐	‐	‐	‐	EQ index value Group A: mean = 0.86, 95% CI 0.85 to 0.87 Group B: mean = 0.86, 95% CI 0.84 to 0.87 Overall self‐rated health status (EQ VAS) Group A: mean = 83, 95% CI 82 to 85; Group B: mean = 82, 95% CI 80 to 83
Javed 2018	Open pancreaticoduodenectomy	Group A: NPWT Group B: standard closure	30 days after operation	Need for reoperation (RR 0.25; 95% CI 0.03 to 2.32; P = 0.21)	Rate of 30‐day readmission for SSI (RR 0.49; 95% CI 0.13 to 1.88; P = 0.32) Rate of 30‐day readmission (RR 0.41; 95% CI 0.15 to 1.09; P = 0.07)	‐	‐	‐	‐	‐
Karlakki 2016	Total hip or knee arthroplasties	Group A: PICO dressing Group B: Comfeel dressing	1, 2, 6 weeks post‐surgery	‐	Group A: 0/107 Group B: 1/108	‐	Group A: 0/102 Group B: 1/107	Group A: 11/102 Group B: 1/107	‐	‐
Keeney 2019	Hip and knee total joint arthroplasty	Group A: iNPWT Group B: conventional wound dressing	7, 14, 35 days after operations; 2 years	Return to the operating room to manage a wound‐related concern within the first 3 months Group A: 1/185 Group B: 4/213	‐	‐	‐	‐	‐	‐	Additional data on return to operating rooms; and infection outcome were presented in Characteristics of included studies.
Kuncewitch 2017	Pancreatectomy	Group A: NPWT Group B: standard surgical dressing	30 days post‐surgery follow‐up	‐	‐	Group A: 4/36 Group B: 6/37	‐	‐	‐	‐	Unable to contact authors
Kwon 2018	Vascular groin incisions (high risk)	Group A: NPWT Group B: standard gauze	30 days	Group A: 5/59 Group B: 11/60	Group A: 4/59 Group B: 10/60	Reported composite outcome including seroma only	Any Group A: 0/59 Group B: 1/60 Major Group A: 0/59 Group B: 1/60	‐	‐	‐
Lee 2017a	Great saphenous vein harvest	Group A: NPWT Group B: standard surgical dressing	Initial assessment: not specified; endpoint assessment: 6 weeks	‐	‐	‐	‐	‐	‐	EQ‐5D‐3L: Group A (n = 26): 78 Group B (n = 17): 63 P = 0.172	2 participants died (sepsis; stroke). 2 participants were delirious and unable to complete QoL; all other objective evaluations were done (all 4 in NPWT).
Lee 2017b	High‐risk groin wounds	Group A: NPWT Group B: standard care	30 days and 90 days	Group A: 2/53 Group B: 1/49 for SSI	Group A: 2/53 Group B: 2/49 for SSI	‐	‐	‐	‐	‐	Latest time point of SSI data used for analysis
Leitao 2020	Laparotomy	Group A: NPWT Group B: standard care	30 days	‐	‐	Group A: 11/223 Group B: 14/221	Group A: 2/223 Group B: 1/221	Group A: 33/223 Group B: 3/221	Group A: 6/223 Group B: 2/221	‐	Data obtained from author contact (composite outcome in publication)
Leon 2016	Open colorectal surgery	Group A: NPWT Group B: usual dressing	15‐day and 30‐day evaluation	‐	‐	‐	‐	‐	‐	‐	Unable to contact authors
Lozano‐Balderas 2017	Laparotomised patients with class III or IV (contaminated/dirty‐infected) surgical wounds	Group A: vacuum‐assisted closure Group B: primary closure Group C: delayed primary closure	Daily when in hospital or in a 30‐day period after surgery	‐	‐	‐	‐	‐	‐	‐	Group C (delayed primary closure) not included in data analysis due to irrelevant wounds
Manoharan 2016	Primary arthroplasty	Group A: NPWT Group B: conventional dry dressing	10 to 12 days post‐surgery	‐	‐	‐	‐	Group A: 1/21 Group B: 0/21	‐	‐
Martin 2019	Hepatectomy or pancreatectomy	Group A: NPWT (PICO) Group B: sterile island dressing	30 days then biweekly for 3 months	‐	Stated that readmission within 6 months assessed but no data	‐	‐	‐	‐	‐
Masden 2012	Radial forearm free flap	Group A: NPWT Group B: dry dressing	Not clear	Group A: 9/44 Group B: 8/37	‐	‐	‐	‐	‐	‐
Murphy 2019	Colorectal resections	Group A: NPWT Group B: standard gauze dressing	30 days	Group A: 6/144 Group B: 6/140	‐	‐	‐	‐	‐	‐
NCT00654641	Caesarean sections	Group A: Negative pressure wound closure Group B: Standard wound closure	6 weeks	‐	‐	‐	‐	‐	‐	‐	The development of a wound complication was reported as a composite outcome including multiple types of events; data were not reported for each complication separately (overall, 15 patients in Group A (53.57%) and 10 in Group B (38.46%)
NCT01759381	Spinal surgery	Group A: NPWT Group B: standard wound closure	3 months	‐	‐	‐	‐	‐	‐	‐
NCT02309944	Laparotomy for suspected gynaecologic malignancy	Group A: NPWT Group B: Standard wound closure	4 weeks		Readmission due to superficial wound infection Up to 12 weeks Group A: 2/43 Group B: 0/38						Only readmission due to wound infection data were available and this outcome was considered as a type of adverse events. The original investigators stated adverse events were collected by non‐systematic assessment.
NCT02461433	Obese patients who undergo any elective open surgery	Group A: Prevena negative pressure wound therapy Group B: standard dressing	5 weeks	‐	Group A: 0/1 Group B: 0/1	‐	‐	‐	SF‐36 Group A: mean 100 (range 100 to 100), n = 1 Group B: mean 100 (range 100 to 100), n = 1	General health (SF‐36) Group A: mean 100 (range 100 to 100), n = 1 Group B: mean 65 (range 65 to 65), n = 1
Newman 2019	Total hip or knee replacements	Group A: ciNPWT Group B: standard silver dressing	12 weeks	Reoperation 2 weeks Group A: 0/79 Group B: 1/80 4 weeks Group A: 1/79 Group B: 3/80 12 weeks Group A: 2/79 Group B: 10/80	Readmission 2 weeks Group A: 5/79 Group B: 6/80 4 weeks Group A: 9/79 Group B: 9/80 12 weeks Group A: 16/79 Group B: 19/80	‐	Haematoma Group A: 1/79 Group B: 1/80	Skin blisters Group A: 0/79 Group B: 1/80	‐	‐
Nordmeyer 2016	Spinal fractures treated with internal fixation	Group A: PICO dressing Group B: standard dressing	Day 5 and day 10 after surgery	‐	‐	‐	‐	‐	‐	‐
O'Leary 2017	Open abdominal surgery	Group A: PICO dressing Group B: transparent waterproof dressing	Day 4 and day‐30 post‐surgery	Group A: 0/25 Group B: 1/25	‐	‐	‐	‐	Reported "no difference"	‐
Pachowsky 2012	Hip arthroplasty	Group A: NPWT Group B: standard dressing	Day 5 and day 10 in postoperative period	‐	‐	Group A: 4/9 Group B: 9/10	‐	‐	‐	‐	Very small sample size
Pauser 2016	Fractures of the femoral neck treated by hemiarthroplasty	Group A: NPWT Group B: standard dressing	Day 5 and day 10 after surgery	‐	‐	Group A: 6/11 Group B: 8/10	‐	‐	‐	‐	Very small sample size
Pleger 2018	Groin wound	Group A: NPWT (n = 58 incisions) Group B: control dressing (n = 71 incisions)	Days 5 to 7 and 30 after surgery	‐	‐	Group A: 0/58 Group B: 1/71	Group A: 0/58 Group B: 8/71	‐	‐	‐	Unit of analysis error: 100 participants with 129 groin incisions
Ruhstaller 2017	Unplanned caesarean section	Group A: NPWT Group B: standard care	4 weeks post‐surgery	Wound requiring opening/packing data: Group A: 3/61 Group B: 2/58	Data from Conference Abstract: Group A: 6/61 Group B: 9/58	‐	‐	Group A: 8/61 Group B: 2/58	Sharp pain at day 2 (0–10): Group A: median 6 (IQR 4 to 8), n = 61 Group B: median 5.5 (IQR 3 to 8), n = 58 Tingling pain at day 2 (0–10): Group A: median 2 (IQR 0 to 6), n = 61 Group B: median 1.5 (IQR 0 to 6), n = 58	Outcome mentioned but data not reported
Sabat 2016	Groin wounds in vascular surgery	Group A: NPWT Group B: conventional dressing (gauze and Tegaderm)	4 months post‐surgery	‐	‐	‐	‐	‐	‐	‐
Schmid 2018	Inguinal lymph node dissection	Group A: NPWT (Prevena) Group B: conventional compression bandages	14 days after surgery	Group A: 13/31 Group B: 14/31	‐	‐	‐	‐	Median pain score Group A: 2 (0‐2) Group B: 1 (0‐3)	‐	Reported wound complication as a composite (unspecified outcome): complications of any grade in 23/31 on each side; grade 3 complications in 3/31 conventional and 6/31 NPWT wounds
Shen 2017	Open resection of intra‐abdominal neoplasms	Group A: PICO dressing Group B: Comfeel dressing	30 days after surgery	Group A: 19/132 Group B: 16/133	Group A: 3/118 Group B: 6/119	Group A: 7/132 Group B: 8/133	Group A: 1/132 Group B: 0/133	‐	‐	‐
Shim 2018	Reconstructive surgery for acute hand injuries	Group A: NPWT Group B: conventional dressing	1 month and 1 year	‐	‐	‐	Group A: 0/30 Group B: 2/21	‐	‐	‐
Stannard 2012	Tibial plateau, pilon, or calcaneus fracture	Group A: NPWT Group B: standard dressing	Not stated	‐	‐	‐	‐	‐	‐	‐	Unit of analysis error
Tanaydin 2018	Bilateral breast reduction mammoplasty	Group A: NPWT Group B: standard care (fixation strips)	21 days	‐	‐	‐	‐	‐	‐	‐	32 participants served as their own control.
Tuuli 2017	Caesarean delivery	Group A: NPWT Group B: standard dressing	30 days	‐	‐	Group A: 0/60 Group B: 1/60	Group A: 0/60 Group B: 0/60	‐	Pain score (on 0‐to‐10 scale) was significantly lower with prophylactic NPWT (median (IQR): 0 (0, 1) vs 1 (0, 3), P = 0.02)	‐
Tuuli 2020	Caesarean delivery	Group A: NPWT Group B: standard dressing	30 days	‐	Group A: 2/806 Group B: 0/802	Group A: 5/806 Group B: 6/802	Group A: 4/806 Group B: 8/802	Group A: 27/806 Group B: 2/802	Pain Score 0‐10. Group A: 3 (0 to 5) at discharge. 0 (0 to 2) at day 30 Group B: 3 (0 to 5) at discharge. 0 (0 to 2) at day 30	‐
Uchino 2016	Ileostomoy reversal	Group A: NPWT Group B: standard dressing	6 weeks	‐	‐	‐	‐	‐	‐	‐
Wierdak 2021	Ileostomoy reversal	Group A: NPWT Group B: standard dressing	7/14 days	‐	‐	Group A: 1/35 Group B: 1/36	Group A: 0/35 Group B: 3/36	‐
WHIST 2019a	Lower limb fracture	Group A: NPWT Group B: standard dressing	30 days 90 days	Deliberate surgical reopening or surgical treatment of wound complications Group A: 2/715 1/573 Group B: 2/688 2/575 Further surgery Group A: 83/578: Group B 56/534	‐	‐	‐	‐	VAS (median IQR) 3 months Group A: 3.0 (1.0, 6.0) 365 Group B: 4.0 (2.0, 5.0) 339 6 months Group A: 3.0 (1.0, 5.0) 419 Group B: 3.0 (1.0, 5.0) 368 Proportion with neuropathic pain (DN4 >/= 3) also reported	DRI 3 months Group A: 51.6 (23.46) (507) Group B: 51.1 (23.92) (456) 6 months Group A: 40.6 (24.98) (469) Group B: 40.2 (26.73) 432 EQ‐5D (utility) 3 months Group A: 0.5 (0.29) 528 Group B: 0.5 (0.30) 470 6 months Group A: 0.6 (0.28) 486 Group B: 0.6 (0.29) 446 EQ‐5D (VAS) 3 months Group A: 64.1 (22.24) 531 Group B: 64.7 (22.78) 478 6 months Group A: 69.7 (21.15) 489 Group B: 69.4 (21.76) 449
Wihbey 2018	Caesarean delivery	Group A: NPWT Group B: standard dressing	1 week and 30 days follow‐up	Group A: 1/80 Group B: 1/81	Group A: 3/80 Group B: 5/81	Group A: 7/80 Group B: 6/81	Group A: 2/80 Group B: 4/81	‐	‐	‐
Witt‐Majchrzac 2015	Coronary artery bypass surgery	Group A: NPWT Group B: conventional dressing	6 weeks follow‐up	‐	‐	‐	Group A: 1/40 Group B: 1/40	Group A: 5/40 Group B: 0/40	‐	‐

Economic Study	RCT base	Population and perspective	Comparison	Time points	Dressing‐related costs	Resource use	QALY	Relative cost‐effectiveness (e.g. ICER)	Notes
Heard 2017	Chaboyer 2014	Population: Obese women undergoing caesarean section Perspective: Australian public health care provider	Group A: PICO dressing Group B: Comfeel dressing	4 weeks	NPWT AUD 180 Standard AUD 5 Dressing change cost (nurse time) AUD 35 for each group	Group A (44): 2871.5 ± 182.1 AUD Group B (43): 2806.6 ± 260.4 AUD	Group A (44): 0.067 ± 0.01 Group B (43): 0.066 ± 0.01	Per SSI prevented: ICER AUD 1347 (95% CI dominant to 41,873) Per QALY gained: ICER AUD 42,340 (95% CI dominant to 884,019)	Data drawn from Chaboyer 2014
Hyldig 2019a	Hyldig 2019b	Population: Obese women undergoing caesarean section Perspective: Danish healthcare	NPWT Standard dressing	30 days	NPWT €151.40 Standard €0.67 (assumed included in cost of treatment)	Total healthcare costs NPWT: €5793.60 Standard: €5840.89 Cost difference: €47.29	NPWT: 0.863 Control: 0.856	ICER not reported for all participants; NPWT reported as dominant; subgroups reported	Data drawn from Hyldig 2019b
Nherera 2017	Karlakki 2016	Population: People undergoing total hip/knee arthroplasty Perspective: UK NHS	NPWT Standard dressing	6 weeks	Cost of NPWT: £144 (120 to 150). Standard dressings assumed zero	Group A (102): 5602 ± 7954 GBP Group B (107): 6713 ± 9559 GPB	Group A (102): 0.116 ± 0.01 Group B (107): 0.115 ± 0.01	ICER not reported but NPWT described as technically dominant	Data drawn from Karlakki 2016
Nherera 2018	Witt‐Majchrzac 2015	Population: People undergoing coronary artery bypass surgery Perspective: German Statutory Health Insurance payer	NPWT Standard dressing	6 weeks	NPWT: €153.00 (114.75 to 191.25) above standard cost	NPWT: €19,986 Standard: €20,572	NPWT: 0.8904 Standard: 0.8593 +0.0311 for NPWT	NPWT reported as dominant for both SSI avoided and QALY gained in base case analysis	Data drawn from Witt‐Majchrzac 2015
Svensson‐Bjork 2020	Hasselmann 2019a	Population: adult patients for elective vascular surgery with inguinal incisions Perspective: healthcare (other details were not specified; but the trial was run in Sweden)	NPWT dressing (PICO) Standard wound dressing	90 days	NPWT: EUR 208 Standard dressing: EUR 45	All healthcare costs in total: NPWT ‐ EUR 19,281 Standard dressing ‐ EUR 17,575 Vascular procedure‐related costs in total (including dressing‐related costs): NPWT ‐ EUR 16,621 Standard dressing ‐ EUR 16,285	‐	NPWT considered as cost‐effective over standard dressings: base case analysis of vascular procedure‐related costs ‐ ICER = approximately EUR 19 per cent in SSI incidence, meaning an increased cost of EUR 1853 per SSI avoided Sensitivity analysis using costs for vascular procedure‐related inpatient care only ‐ ICER = approximately EUR 2 per cent in SSI incidence, corresponding to an increased cost of EUR 204 per SSI avoided Sensitivity analysis of vascular procedure‐related costs and gain in Vascuqol‐6 score ‐ ICER = EUR 719 per unit of Vascuqol‐6 score increased	Data drawn from Hasselmann 2019a; no QALYs
WHIST 2019b	WHIST 2019a	Population: People undergoing surgery for lower limb fracture Perspective: UK NHS and PSS	NPWT Standard dressing	3 months 6 months	Cost of intervention including dressing (plus cast, initial inpatient care, antibiotics, dressing changes) Group A: £5420.66 (5559.95) Group B: £4774.15 (4633.18)	Total cost after initial intervention ‐ baseline to 6 months NHS and PSS Group A: £3100.83 (11251.93) Group B: £2330.83 (7863.51) Societal (inc NHS & PSS) Group A: £6248.64 (13074.32) Group B: £6027.23 (17737.28)	Group A: 0.40 (0.22) Group B: 0.41 (0.24)	£396,531/QALY gained (NHS & PSS perspective) £679,482 per QALY gained (societal perspective) £454,903 per QALY (complete case analysis)	Data drawn from WHIST 2019a
‐	DiMuzio 2017¹	Groin wounds	NPWT Standard dressing	30 days	‐	Group A: USD 30,492 Group B: USD 36,537 NPWT reduced cost per patient of USD 6045 (USD 30,492 + USD 500 (device) in NWPT group vs USD 36,537 in dressing group)	‐	‐	Data not linked to cost‐effectiveness
‐	Gillespie 2015¹	Total hip/knee arthroplasty	NPWT Standard dressing	30 days 6 weeks	Group A: AUS 38.4 ± AUS 13.6 Group B: AUS 3.01 ± AUS 1.2	‐	‐	‐	Data not linked to cost‐effectiveness
‐	Javed 2018¹	Open pancreaticoduodenectomy	NPWT Standard dressing	30 days	‐	Median cost of hospitalisation $43,823 (IQR, $36,820–$59,352) No between‐group data	‐	‐	Data not linked to cost‐effectiveness
‐	Kwon 2018¹	Vascular groin incisions (high risk)	NPWT Standard dressing	30 days	‐	Costs (hospital) Group A: $29,292 +/‐ 6 $29,320 (n = 51; range, $8816‐$192,658) Group B: $30,678 6 $23,338 (n = 55; range, $9071‐$131,464) Costs (post index procedure) Group A: $30,492 +/‐ $30,678 ($8816‐$192,658) Group B: $36,537 +/‐ $28,889 (range, $9071‐$131,4640*	‐	‐	Data not linked to cost‐effectiveness
‐	Manoharan 2016	Primary arthroplasty	NPWT Standard dressing	10‐12 days	Group A: AUS 285.94 ± AUS 28.54 Group B: AUS 43.51 ± AUS 64.23	‐			Data not linked to cost‐effectiveness
‐	Ruhstaller 2017	Caesarean section	NPWT Standard dressing	4 weeks	NPWT cost: USD $544 Standard dressing: NR	‐	‐	At a per‐device cost of $544, prevention of a single infection would cost approximately $15,000. This is based on a number needed to treat for an additional beneficial outcome of 28. Thus, the prevention of one SSI after a caesarean delivery would increase post‐surgical healthcare costs, an additional $10,300 beyond the average cost attributed to the infection itself.	Data not linked to cost‐effectiveness

Sections	Items	Item number	Heard 2017	Hyldig 2019a	Nherera 2017	Nherera 2018	Svensson‐Bjork 2020	WHIST 2019a
Title and abstract	Title	1	✓	✓	✓	✓	✓	X
Title and abstract	Abstract	2	✓	✓	✓	✓	✓	≠*
Introduction	Background and objectives	3	✓	✓	✓	✓	✓	✓
Methods	Target population and subgroups	4	✓	✓	✓	✓	✓	✓
	Setting and locations	5	✓	✓	✓	✓	✓	✓
	Study perspectives	6	✓	✓	✓	✓	≠	✓
	Comparators	7	✓	✓	✓	✓	✓	✓
	Time horizon	8	X	✓	✓	✓	✓	✓
	Discount rate	9	✓	✓	✓	✓	N/A	✓
	Choice of health outcomes	10	✓	✓	≠	✓	✓	✓
	Measurement of effectiveness	11a	≠	≠	N/A	≠	✓	≠
	Measurement of effectiveness	11b	N/A	N/A	✓	≠	N/A	N/A
	Measurement and valuation of preference‐based outcomes	12	≠	✓	≠	X	N/A	✓
	Estimating resources and costs	13a	✓	✓	N/A	N/A	✓	✓
	Estimating resources and costs	13b	N/A	N/A	✓	✓	N/A	N/A
	Currency, price date and conversion	14	✓	✓	≠	✓	✓	✓
	Choice of model	15	≠	≠	≠	✓	≠	≠
	Assumptions	16	✓	✓	✓	✓	✓	✓
	Analytical methods	17	✓	✓	✓	✓	✓	✓
Results	Study parameters	18	✓	✓	✓	✓	✓	✓
	Incremental costs and outcomes	19	✓	✓	✓	≠	✓	✓
	Characterising uncertainty	20a	≠	≠	N/A	N/A	✓	✓
	Characterising uncertainty	20b	N/A	N/A	≠	≠	N/A	N/A
	Characterising heterogeneity	21	X	✓	✓	N/A	N/A	N/A
Discussion	Study findings, limitations, generalisability and current knowledge	22	✓	✓	✓	✓	✓	✓
Others	Source of funding	23	✓	✓	X	X	✓	✓
Others	Conflicts of interest	24	✓	✓	✓	✓	✓	✓
Total			20/24 (83.3%)	22.5/24 (91.7%)	20.5/24 (85.4%)	20/23* (87.0%)	19/24 (79.2%)	20.5/23* (89.1%)

Date	Event	Description
11 March 2022	New search has been performed	Updated; conclusions changed. 18 new RCTs; one new economic evaluation; one new author joined the review team.
11 March 2022	New citation required and conclusions have changed	Updated, conclusions changed.

Date	Event	Description
9 June 2020	Amended	Republished as Open Access.
2 May 2020	New citation required and conclusions have changed	Updated. Conclusions changed.
1 March 2019	New search has been performed	Second update: new citation: conclusions not changed. New search, 25 new studies included. 'Summary of findings' table added. Four new co‐authors added, Gill Norman, Zhenmi Liu, Jo Dumville and Laura Chiverton.
27 August 2014	New citation required but conclusions have not changed	Four trials added (Crist 2014; Masden 2012; Petkar 2012; Stannard 2012), no change to conclusions.
27 August 2014	New search has been performed	First update, new search
13 November 2013	Amended	Acknowledgement added to the funders.
16 May 2012	Amended	Adjustments to text.

Term	Description
Dehiscence	Wound dehiscence is a complication of surgery in which a wound breaks open along the line of the surgical incision.
Negative pressure wound therapy (NPWT)	Negative pressure wound therapy is based on a closed, sealed system that produces negative pressure to the wound surface. The wound is covered or packed with an open‐cell foam or gauze dressing and sealed with an occlusive drape. Intermittent or continuous suction is maintained by connecting suction tubes from the wound dressing to a vacuum pump and liquid waste collector. Standard negative pressure rates range between −50 mmHg and −125 mmHg (Ubbink 2008; Vikatmaa 2008).
Risk ratio (RR)	The risk ratio, or relative risk (RR) is the probability that a member of a group who is exposed to an intervention will develop an event relative to the probability that a member of an unexposed group will develop that same event.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1.1 Mortality	11	6384	Risk Ratio (M‐H, Random, 95% CI)	0.78 [0.47, 1.30]
1.2 Surgical site infection	44	11403	Risk Ratio (M‐H, Random, 95% CI)	0.73 [0.63, 0.85]
1.2.1 Orthopaedic: Hip/knee arthroplasties	4	836	Risk Ratio (M‐H, Random, 95% CI)	0.69 [0.32, 1.49]
1.2.2 Orthopaedic: Limb fractures	3	1676	Risk Ratio (M‐H, Random, 95% CI)	1.15 [0.61, 2.20]
1.2.3 Obstetric: Caesarean	9	5529	Risk Ratio (M‐H, Random, 95% CI)	0.78 [0.65, 0.95]
1.2.4 Vascular: peripheral bypass	6	895	Risk Ratio (M‐H, Random, 95% CI)	0.54 [0.38, 0.77]
1.2.5 Vascular: cardiac surgery	2	136	Risk Ratio (M‐H, Random, 95% CI)	0.17 [0.03, 0.96]
1.2.6 General: abdominal	13	1823	Risk Ratio (M‐H, Random, 95% CI)	0.78 [0.57, 1.05]
1.2.7 General: Hepatopancreatiobiliary	3	258	Risk Ratio (M‐H, Random, 95% CI)	0.58 [0.23, 1.45]
1.2.8 Mixed	4	250	Risk Ratio (M‐H, Random, 95% CI)	0.46 [0.17, 1.27]
1.3 SSI grouped by contamination class	44	11403	Risk Ratio (M‐H, Random, 95% CI)	0.73 [0.63, 0.85]
1.3.1 Clean	17	2288	Risk Ratio (M‐H, Random, 95% CI)	0.58 [0.41, 0.81]
1.3.2 Clean‐contaminated	21	7282	Risk Ratio (M‐H, Random, 95% CI)	0.83 [0.72, 0.96]
1.3.3 Contaminated	3	1649	Risk Ratio (M‐H, Random, 95% CI)	0.78 [0.28, 2.14]
1.3.4 Dirty	3	184	Risk Ratio (M‐H, Random, 95% CI)	0.27 [0.06, 1.12]
1.4 SSI (superficial)	22	5539	Risk Ratio (M‐H, Random, 95% CI)	0.70 [0.53, 0.92]
1.5 SSI (deep)	22	8521	Risk Ratio (M‐H, Random, 95% CI)	0.95 [0.76, 1.18]
1.6 Dehiscence	23	8724	Risk Ratio (M‐H, Random, 95% CI)	0.97 [0.82, 1.16]
1.6.1 Orthopaedic: hip/knee arthroplasty	2	229	Risk Ratio (M‐H, Random, 95% CI)	0.43 [0.08, 2.35]
1.6.2 Orthopaedic: limb fracture	1	1401	Risk Ratio (M‐H, Random, 95% CI)	0.27 [0.06, 1.32]
1.6.3 Obstetric: caesarean	6	5113	Risk Ratio (M‐H, Random, 95% CI)	1.01 [0.82, 1.24]
1.6.4 Vascular: peripheral	3	473	Risk Ratio (M‐H, Random, 95% CI)	0.99 [0.51, 1.92]
1.6.5 Vascular: cardiac	1	80	Risk Ratio (M‐H, Random, 95% CI)	1.00 [0.06, 15.44]
1.6.6 General: abdominal	6	1156	Risk Ratio (M‐H, Random, 95% CI)	0.88 [0.57, 1.38]
1.6.7 General: hepatopancreatiobiliary	1	40	Risk Ratio (M‐H, Random, 95% CI)	0.50 [0.05, 5.08]
1.6.8 Mixed	3	232	Risk Ratio (M‐H, Random, 95% CI)	0.86 [0.16, 4.55]
1.7 Reoperation	18	6272	Risk Ratio (IV, Random, 95% CI)	1.13 [0.91, 1.41]
1.8 Readmission	15	5853	Risk Ratio (M‐H, Random, 95% CI)	0.98 [0.70, 1.38]
1.9 Seroma	15	5436	Risk Ratio (M‐H, Random, 95% CI)	0.82 [0.65, 1.05]
1.10 Haematoma	17	5909	Risk Ratio (M‐H, Random, 95% CI)	0.79 [0.48, 1.30]
1.11 Skin blisters	11	5015	Risk Ratio (M‐H, Random, 95% CI)	3.55 [1.43, 8.77]
1.12 Pain	2	632	Risk Ratio (M‐H, Random, 95% CI)	1.52 [0.20, 11.31]

*Study characteristics*
Methods	Study design: randomised controlled trial Study grouping: parallel Ethics and informed consent: Ethics Committee of the Provinces of Verona and Rovigo; written informed consent at hospital admission Follow‐up period: 30 days Sample size estimate: yes; rate of SSI in patients considered at high risk accounted for 33% of cases. Hypothesised that NPWT could reduce this to 10% of cases (based on literature). With the significance threshold set at 0.05 and power set at 80%, the sample size calculation suggested a recruitment target of 94 patients (47 in each group). Considering a possible dropout rate of 5% to 10% of cases because of re‐laparotomy for bleeding or septic complications, the final recruitment target was extended to 100 patients (50 in each group). ITT analysis: yesnumber randomised: 100, number analysed: 95 Funding: this work was supported by Associazione Italiana per la Ricerca sul Cancro (AIRC n.12182 and n.17132); the Italian Ministry of Health (FIMPCUP_J33G13000210001); and the FP7 European Community Grant Cam‐Pac (n. 602783). Smith & Nephew Healthcare (Hull, UK) supplied the devices used for the study. Preregistration: NCT03700086
Participants	Location: Italy Intervention group: 50,control group: 50 Mean age: intervention group median (IQR) 69 (12), control group median (IQR) 64 (17) Inclusion criteria: patients scheduled for a major clean‐contaminated surgical procedure for periampullary neoplasms ‐ namely, PD total pancreatectomy or gastro‐jejunal and biliary bypass with the presence of at least one of the following indicators of high risk for SSI (assessed after wound closure): body mass index 30 kg/m² , diabetes mellitus, chronic use of steroids, neoadjuvant therapy, American Society of Anesthesiologists score 3, Charlson comorbidity index 1, time of surgery 360 min, and estimated blood loss 1 L. Exclusion criteria: previous abdominal surgery, no indicator of high risk for SSI
Interventions	Aim/s: to assess whether a disposable device for NPWT could reduce the incidence of SSI in patients at high risk for SSI after surgery for periampullary neoplasms, given the standardised surgical technique and clinical pathway of a high‐volume center for pancreatic surgery. Group 1 (NPWT) intervention: PICO portable NPWT device (Smith & Nephew Healthcare, Hull, UK); dressing changed after 3 days, NPWT maintained for 7 days. Group 2 (control) intervention: sterile gauze dressing until day 3 then standard sterile dressing (OPSITE, Smith & Nephew) Study date/s: July 2018 to October 2019
Outcomes	SSI Mortality Seroma Haematoma Reoperation Validity of measure/s: CDC definition for SSI Time points: 7 days, 30 days, 90 days for mortality
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "At this point [post wound closure], randomization was provided in a 1:1 ratio, using a computer‐generated randomization list kept by independent data managers and concealed to the investigators". Comment: computer‐generated randomisation sequence
Allocation concealment (selection bias)	Low risk	Quote: "At this point [post wound closure], randomization was provided in a 1:1 ratio, using a computer‐generated randomization list kept by independent data managers and concealed to the investigators" Comment: sequence was kept by independent data managers and concealed from the investigators.
Blinding of participants and personnel (performance bias) All outcomes	High risk	The nature of the intervention means that blinding of participants and personnel was unlikely.
Blinding of outcome assessment (detection bias) All outcomes	High risk	Quote: "All objective evaluations were performed by different clinicians involved in patient care during their postoperative course. All subjective evaluations, including the VAS and SBSES, were performed by a single physician (L.I.) who was blinded to the type of wound closure". Comment: blinded outcome evaluation reported but not for all outcomes ‐ and seemed probable not for outcomes considered in review.
Incomplete outcome data (attrition bias) All outcomes	Low risk	An ITT analysis included 95% of 100 randomised participants; reasons for 5 participants not being included were provided and related to people not receiving the intervention.
Selective reporting (reporting bias)	High risk	There were disparities between the planned and the reported outcomes. In particular, mortality was not reported although it was planned to be assessed.
Other bias	Low risk	No evidence of other bias

*Study characteristics*
Methods	Study design: RCT Study grouping: parallel Ethics and informed consent: approved by the Ethics Committee of the La Fe University Hospital; written informed consent. Follow‐up period: 30 days Sample size estimate: yes; assuming that the percentage of patients developing SSOs at 30 days post‐surgery following elective hernia repair is at most 30% and can be reduced to 10% in the treatment group, a sample size of 150 patients was required to achieve 80% power with an at‐risk of 5%, including an anticipated 10% loss to follow‐up. Thus, 75 patients were required in each group. ITT analysis: no; analysis performed on per‐protocol basis (150 randomised, 146 analysed). Funding: study sponsored by La Fe University Hospital. Authors stated the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Preregistration: NCT03576222
Participants	Location: Spain Intervention group: 75,control group: 75 Mean age: intervention group 51.6 (23.2), control group 51.3 (19.4) Inclusion criteria: male and female patients > 18 years’ old with IH type W2 (transverse hernia defect with 4–10 cm) or W3 (transverse hernia defect over 10 cms) according to European Hernia Society (EHS) classification. Exclusion criteria: patients under the age of 18 years, patients unable to give written consent, patients who had abdominal surgery reintervention within 30 days before the hernia repair, patients who had undergone emergency hernia surgery, pregnant patients, and patients with hepatic cirrhosis and IH not involving the midline.
Interventions	Aim/s: to evaluate whether the prophylactic application of a specific single‐use negative pressure (sNPWT) dressing on closed surgical incisions after incisional hernia (IH) repair decreases the risk of surgical site occurrences (SSOs) and the length of stay. Group 1 (NPWT) intervention: sNPWT (PICO; Smith & Nephew, London, UK). Dressing left in situ for 6 days/during hospital stay. Group 2 (control) intervention: conventional dressing (MEPORE pro; Molnlycke, Goteborg, Sweden) Study date/s: May 2017 to January 2020
Outcomes	SSI Dehiscence Seroma Haematoma Validity of measure/s: SSI was defined as an infection that occurred at the site of a surgical incision or in an organ space within 30 days of the surgery. Wound dehiscence was defined as the splitting apart or rupturing of the margins of a previously closed wound along some or all of its length. Time points: 12 days; 30 days
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Randomization was performed on a 1:1 basis to either the treatment group or the control group. The sequence was generated on www.randomization.com and allocation was concealed using closed envelopes." Comment: computer‐generated randomisation
Allocation concealment (selection bias)	Unclear risk	Quote: "Randomization was performed on a 1:1 basis to either the treatment group or the control group. The sequence was generated on www.randomization.com and allocation was concealed using closed envelopes." Comment: closed envelopes but unclear if opaque or sequentially numbered.
Blinding of participants and personnel (performance bias) All outcomes	High risk	Quote: "The operating surgeon was not blinded to the dressing being applied to the wound." Comment: personnel not blinded; participant blinding unlikely due to nature of the intervention.
Blinding of outcome assessment (detection bias) All outcomes	High risk	Quote: "Patients were assessed again at 12 (wound clip removal) and 30 days in the outpatient clinic and the wound examined for evidence of SSOs by the same study assessor, who was a member of the operating surgical team and who was not blinded to the treatment group." Comment: outcome assessor not blinded.
Incomplete outcome data (attrition bias) All outcomes	Low risk	146/150 participants included in analysis; no difference between groups and reasons for loss reported.
Selective reporting (reporting bias)	Unclear risk	Some outcomes which were not prespecified were reported (readmission; reoperation). All prespecified outcomes were reported.
Other bias	Low risk	No evidence of other sources of bias.

*Study characteristics*
Methods	Study design: randomised control trial Study grouping: parallel design Ethics and informed consent: informed consent obtained; approved by the ethics committees of the Royal Brisbane and Women’s Hospital and Griffith University Follow‐up period: 30 days Sample size estimate: calculated the sample size based on the proportion of women who developed a SSI within 30 days of CS. Conservatively estimated that 15% of women in the control group were likely to develop an SSI; determined that an absolute reduction in rate of SSI of 5 percentage points would be clinically important. The sample size required to detect a reduction in the cumulative incidence of SSI at 30 days from 15% to 10% was 950 per group (90% power and 5% significance level; inflated the sample size by 10% to allow for loss to follow up (n = 1045/group; total sample size 2090). ITT analysis: yes; number randomised: 2035, number analysed: 2035 Funding: Australian National Health and Medical Research Council Preregistration: ANZCTR identifier 12615000286549
Participants	Location: Australia (4 tertiary hospitals) Intervention group: 1017,control group: 1018 Mean age: 31 (5.5) vs 31 (5.4) Inclusion criteria: women booked for elective (category 4) or for semi‐urgent (categories 2‐3) caesarean section; who recorded a pre‐pregnancy BMI of ≥ 30 kg/m² and were able to give informed consent. Exclusion criteria: women who needed an urgent CS (category 1), had an infection in hospital including during labour or immediately prior to CS, had participated in the trial in a previous pregnancy, or were unable to speak or understand English with no interpreter present.
Interventions	Aims: to determine the effectiveness of closed incision negative pressure wound therapy (NPWT) compared with standard dressings in preventing surgical site infection (SSI) in obese women undergoing caesarean section. Group 1 (NPWT) intervention: women assigned to the NPWT group received a PICO™ dressing (Smith & Nephew, Hull, UK), which was left intact for approximately 5‐7 days as recommended by the manufacturer. Group 2 (control) intervention: women received the standard hospital dressing. The choice of standard dressings was based on the treating obstetrician’s usual choice of dressing, (e.g. hydrocolloid or transparent) applied according to the manufacturer’s recommendations. The standard dressing was left intact for 5‐7 days.
Outcomes	Primary SSI (including superficial, deep or organ/body space) Dehiscence Mortality (included in the serious adverse events outcome) Secondary Haematoma Seroma Blistering Readmission Reoperation Pain Validity of measure/s: CDC definition for SSI Time points: 30 days
Notes	Full cost‐effectiveness analysis planned to be reported separately (protocol).
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "We used a web based central randomisation service to randomly assigned eligible, consenting women (1:1) just before the caesarean procedure to receive either a closed incision NPWT dressing or the standard hospital dressing. To ensure that equal numbers of participants were assigned to each group, we used random block sizes of four, six, and eight, stratified by hospital". Comment: appropriate methods used to generate randomisation sequence
Allocation concealment (selection bias)	Low risk	Quote: "Allocation was concealed until after skin closure". Comment: central randomisation service (above) and concealed allocation reported
Blinding of participants and personnel (performance bias) All outcomes	High risk	Quote: "The nature of the intervention meant that women, clinical staff, and research staff were not blinded to treatment after allocation". Comment: unblinded participants and personnel
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Quote: "Data were reviewed by two independent, blinded outcome assessors to determine primary and secondary wound endpoints, and discrepancies were adjudicated by a third blinded assessor. Principal investigators, including the trial statistician, were also blinded to group allocation". Comment: blinded outcome assessment
Incomplete outcome data (attrition bias) All outcomes	Low risk	All randomised women were included in the ITT analysis; only for AE was a per‐protocol analysis adopted (this impacts the outcome of blistering) and was prespecified; attrition to PP analysis was low and balanced.
Selective reporting (reporting bias)	Low risk	All planned outcomes were fully reported.
Other bias	Low risk	No evidence of any other source of bias; well reported.

*Study characteristics*
Methods	Study design: cost‐effectiveness analysis. Data drawn from the Chaboyer 2014 RCT Analytical approach: trial‐based evaluation Effectiveness data: data from pilot RCT (N = 87) (Chaboyer 2014). Key effectiveness inputs were SSI and quality of life (SF‐12) at up to 4 weeks post‐discharge in trial. Perspective: Australian public health care provider Utility valuations: QALYs were calculated from SF‐12 data. QoL indices (utility weights) were calculated using the method of Brazier and Roberts. QALYs were estimated from the utility weights using the standard area under the curve method. Adjustment: QALYs were adjusted for differences in baseline SF‐12 indices using the regression‐based adjustment of Manca, Hawkins and Sculpher. Measure of benefit: surgical site infection avoided; QALY Cost data: measured in AUSD; in hospital resource use data were collected by direct observation or chart audit during the trial. Included cost of intervention, nursing time for dressing changes, hospital (inpatient) care. No discount rate was applied due to the short time horizon. Analysis of uncertainty: a nonparametric bootstrap with 1000 replications was used to construct 95% percentile method confidence intervals (CIs) for the estimates. A sensitivity analysis used only post‐discharge QALYs, ignoring the period of hospitalisation (the base case analysis calculating QALYs from utility weights assumed that the change in QoL over the hospital stay was linear). Funding: Office of Health and Medical Research, Queensland Health, the National Health and Medical Research Council Centre of Research Excellence in Nursing and a Gold Coast University Hospital Private Practice grant.
Participants	Location: Obstetric unit, Australia Intervention group: n = 46,control group: n = 46 (obese women (> 30 BMI) undergoing elective CS) Mean age: intervention group = 30.6 years (SD 5.5),control group = 30.7 years (SD 5.0) Inclusion criteria: booked for elective CS; pre‐pregnancy BMI > 30; able to provide consent. Exclusion criteria: women whose condition changed to require urgent CS; previous participation in the trial; existing infection.
Interventions	Aim/s: to evaluate whether NPWT is cost‐effective compared with standard care in preventing surgical site infection among obese women undergoing caesarean section.  Group 1 (NPWT) intervention: NPWT: PICO (Smith and Nephew) dressing applied over the primarily closed incision by the surgeon in the operating room. Dressing was left on for 4 days, or longer if drainage continued, unless soiled or dislodged (n = 44) in Heard 2017 trial). Group 2 (Comparator) intervention: Comfeel dressing applied over the primarily closed incision by the surgeon in the operating room. Dressing was left on for 4 days, or longer if drainage continued, unless soiled or dislodged (n = 43 in a trial). Study date/s: July 2012 to April 2014
Outcomes	For data see Heard 2017 and for clinical data see Chaboyer 2014 in additional table 1 Surgical site infection Costs (AUSD) QALY (measure of benefit). ICER with 95% CI (AUSD per unit outcome) to inform probability of intervention being cost‐effective
Notes	Authors' conclusions: NPWT may be cost‐effective in the prophylactic treatment of surgical wounds following elective caesarean section in obese women. Larger trials could clarify the cost‐effectiveness of NPWT as a prophylactic treatment for SSI. Sensitive capture of QALYs and cost offsets will be important given the high level of uncertainty around the point estimate cost‐effectiveness ratio which was close to conventional thresholds. Quality rating according to the CHEERS checklist was 83.3%.

*Study characteristics*
Methods	Study design: pragmatic randomised controlled trial Study grouping: parallel Ethics and informed consent: yes Follow‐up period: 30 days Sample size estimate: yes; a sample size of 870 for a reduction in surgical site infection of 50% in the intervention group compared with an expected baseline event rate of 10% in the control group, with a two‐sided 5% significance level and a power of 80%. ITT analysis: yes (for surgical site infection only),number randomised: 876, number analysed: 876 for surgical site infection and 827 for other outcomes Funding: grants from the University of Southern Denmark, Odense University Hospital, the Region of Southern Denmark, Lundbeckfonden, and an unrestricted grant from the iNPWT device manufacturer Smith & Nephew (devices and operating funding) Preregistration: yes; ClinicalTrials.gov (NCT 01890720)
Participants	Location: Denmark (2 tertiary referral centres and 3 Danish teaching hospitals) Intervention group: n = 432, control group: n = 444 (6 received iNWPT dressing) Mean age: a range from 18 to 46 years across groups; intervention group: 32 (SD 5), control group 32 (SD 5) Inclusion criteria: pregnant women undergoing elective or emergency caesarean section, aged >= 18 years; who had a prepregnancy body mass index >= 30 kg/m², and could read and understand Danish. Exclusion criteria: women who had given informed consent but subsequently delivered vaginally
Interventions	Aim/s: to investigate the effectiveness of prophylactic iNPWT after caesarean section in obese women; hypothesis: iNPWT would be associated with fewer surgical site infection and other wound complications (i.e. wound exudate and dehiscence) compared with standard postoperative dressing.  Group 1 (NPWT) intervention: incisional negative pressure wound therapy (iNPWT; PICO, SIZE 10 X 30 cm or 10 X 40 cm, Smith & Nephew, Hull, UK) in which dressing was left in situ for approximately 5 days Group 2 (control) intervention: standard postoperative dressing in which dressing was left in situ for at least 24 hours Study date/s: September 2013 to October 2016
Outcomes	Surgical site infection, those infections requiring antibiotic treatment within the first 30 days after caesarean section Deep surgical site infection, those infections requiring surgery Minor dehiscence, defined as a gap between the sides of the wound Health‐related quality of life (EQ‐5D‐5L) Readmissions to hospital/contact to the general practitioner on suspicion of infection following caesarean section (listed in ClinicalTrials.gov) Validity of measure/s: Time points: within the first 30 days after surgery
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Participants were randomised in the operating theatre during surgery using a web‐based randomisation programme with a 1:1 allocation ratio and random block sizes of 4–6, stratified by centre and type of caesarean section (emergency versus elective)." Comment: low risk of bias due to valid random sequence generation
Allocation concealment (selection bias)	Low risk	Quote: "The random allocation sequence was generated by an external data manager with no clinical involvement in the study". Comment: low risk of bias due to likely appropriate approach taken to conceal randomisation process
Blinding of participants and personnel (performance bias) All outcomes	High risk	Quote: "Blinding was not possible due to the nature of the intervention". Comment: high risk of bias because it was clearly stated no blinding of participants and personnel
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not reported
Incomplete outcome data (attrition bias) All outcomes	Low risk	ITT analysis was conducted for surgical site infection and for other outcomes; only 22 of 432 in Group 1 and 27 of 444 in Group 2 were excluded from analyses. Low risk of attrition bias.
Selective reporting (reporting bias)	High risk	Readmission to hospital/contact to the GP was listed on ClinicalTrials.gov but not presented in the full text. High risk of reporting bias.
Other bias	Low risk	None detected

*Study characteristics*
Methods	Study design: prospective, randomised, multicentre study Study grouping: parallel Ethics and informed consent: not reported Follow‐up period: not reported Sample size estimate: not reported ITT analysis: yes, number randomised: 81, number analysed: 81 Funding: not reported Preregistration: not reported
Participants	Location: Spain Intervention group: n = 47,control group: n = 34 Mean age (SD): not reported Inclusion criteria: patients undergoing open and programmed colorectal surgery Exclusion criteria: not stated
Interventions	Aim/s: to evaluate the benefits of negative pressure therapy to reduce surgical site infection rate in open colorectal surgery.  Group 1 (NPWT) intervention: NPWT Group 2 (control) intervention: usual dressing group Study date/s: not reported
Outcomes	SSI rate Validity of measure/s: not described Time points: a daily evaluation through hospitalisation and a 15‐ and 30‐day evaluation
Notes	Only the abstract was available.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Not stated
Allocation concealment (selection bias)	Unclear risk	Not stated
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not stated
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not stated
Incomplete outcome data (attrition bias) All outcomes	Low risk	All enrolled participants were accounted for in the analyses.
Selective reporting (reporting bias)	Unclear risk	Not stated
Other bias	Unclear risk	Abstract only

*Study characteristics*
Methods	Study design: 1:1 parallel‐group randomised controlled trial Study grouping: parallel Ethics and informed consent: yes Sample size estimate: no Follow‐up period: 4 months ITT analysis: no Funding: not stated Preregistration: not stated
Participants	Location: Philadelphia, USA Intervention group: n = 33 wounds,control group: n = 30 wounds (total 49 participants) Mean age: not reported Inclusion criteria: people undergoing open vascular surgery involving a groin incision Exclusion criteria: not stated
Interventions	Aim/s: to compare the effect of postoperative negative pressure therapy to conventional dressings on wound occurrences.  Group 1 (NPWT) intervention: NPWT device Group 2 control: conventional dressing (gauze and Tegaderm) Study date/s: not stated
Outcomes	SSI Wound dehiscence
Notes	Abstract only; unit analysis
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Not stated
Allocation concealment (selection bias)	Unclear risk	Not stated
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not stated
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not stated
Incomplete outcome data (attrition bias) All outcomes	Low risk	All those recruited appeared to have been included in the analysis.
Selective reporting (reporting bias)	Unclear risk	Results for outcomes identified in the Methods section were reported. We did not see the original protocol.
Other bias	Unclear risk	Unit of analysis issue ‐ unclear if accounted for.

*Study characteristics*
Methods	Study design: multicentre randomised controlled trial (four centres, each a level 1 trauma centre) Ethics and informed consent: ethics approved and consent obtained Sample size calculation: no Follow‐up period: not reported ITT analysis: wounds, not people were assessed. Funding: "funds from corporate/industry were received from Kinetic Concepts, Inc to support this work".
Participants	Location: Columbus, Ohio, USA Intervention group: n = 130, participants; 141 fractures,control group: n = 119 participants; 122 fractures Mean age: not stated Inclusion criteria: people > 18 years of age who had sustained a high‐energy tibial plateau, pilon, or calcaneus fracture and were able to comply with research protocol and willing to give informed consent. Exclusion criteria: non‐operative calcaneus, tibia plateau, or pilon fractures; patients with open calcaneus fractures; tibial plateau or calcaneus fractures receiving definitive surgery more than 16 days after injury; pilon fractures receiving definitive surgery more than 21 days after injury; prisoners; pregnant women; patients with one of these fractures as a result of a low‐energy mechanism of injury; patients or family members unable or unwilling to sign study informed consent; and patients unable to comply with the protocol.
Interventions	Aim/s: "to investigate the use of NPWT to prevent wound dehiscence and infection after high‐risk lower extremity trauma" Intervention/s in both groups: dressings or NPWT were applied in the operating room and then changed on postoperative day 2 and every 1 to 2 days thereafter. Group 1 (NPWT) intervention: NPWT over the surgical incision after open reduction and internal fixation of the fracture. Group 2 (control) intervention: standard postoperative dressing (dressing not described). Study date/s: not stated
Outcomes	Wound infection and dehiscence Time to discharge from hospital Validity of measure/s: "all infections were confirmed with cultures". Time points: not stated ‐ unclear for how long participants were followed up.
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Evidence: quote: "patients were enrolled and then randomised to receive either standard postoperative dressings (control) or NPWT (study)". Comment: additional author information: "the randomization was done via a computer generated randomization program".
Allocation concealment (selection bias)	Unclear risk	Comment: method not clarified
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Evidence for participants: not possible Comment: unlikely to affect outcomes Evidence for personnel: not possible Comment: unlikely to affect outcomes
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Evidence: quote: "a patient was diagnosed as having an infection when a combination of clinical signs and symptoms (purulent drainage, erythema, fever, chills, etc) and laboratory data documented the infection. All infections were confirmed with cultures. Wound dehiscence was defined as any separation of the surgical incision that required either local wound care or surgical treatment". Comment: not clear whether those assessing outcomes were aware of group assignment.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Comment: a total of 249 participants were recruited. The same number of participants were reported for both acute and long‐term follow‐up (follow‐up period not defined). Given that 4 hospitals were involved in the study, it seemed unusual that complete follow‐up would have occurred, suggesting that an available‐case analysis may have been performed.
Selective reporting (reporting bias)	Low risk	Comment: registered in CTR (NCT00582998) 9 months after final data collection date, so it was unclear whether reported outcomes matched the original protocol. However, infection and dehiscence were the expected outcomes.
Other bias	High risk	Comment: unequal number of participants in each group; appeared from the protocol that data collection was over many years, but no dates or explanation in manuscript; results reported per fracture, so there was a potential unit of analysis issue.

*Study characteristics*
Methods	Study design: randomized controlled trial Study grouping: Parallel Ethics and informed consent: institutional review board approval was obtained from the Dartmouth Committee for the Protection of Human Subjects on April 21, 2015 (#00005211) and from the Southern New Hampshire Medical Center Clinical Trials Office (#2015‐01). Women were recruited and consented to participate in this study before the onset of active labor during any routine prenatal visit or inpatient admission. Follow‐up period: 30 days Sample size estimate: Yes. 400 women (200 prophylactic negative pressure wound therapy, 200 standard dressing) would need to be recruited to have an 80% power to detect a 50% decrease in superficial surgical site infection (assuming P < 0.05). ITT analysis: Yes, number randomised: 166, number analysed: 166 Funding: the devices used in this study were provided by an unrestricted research grant from KCI Medical (San Antonio, Texas). Preregistration: Yes. This trial was registered with clinical‐trials.gov (Clinical Trial Registration: NCT02390401).
Participants	Location: 2 centres (USA) Intervention group: n = 80, control group: n = 86 Mean age: intervention group 31 ± 6, control group 30.2 ± 5 Inclusion criteria: women undergoing caesarean delivery for a viable neonate and their BMI on admission to the labour and delivery floor was 35 or higher. Exclusion criteria: women who were younger than 18 years old, did not speak English, had an allergy to silver or adhesives products, or who had a skin incision that would not fit the device or standard dressing (e.g. “T” skin incision).
Interventions	Aim/s: to compare the occurrence of superficial surgical site infections in women with class II or III obesity as defined by the Centers for Disease Control and Prevention using prophylactic negative pressure wound therapy compared with standard dressings after caesarean delivery.  Group 1 (NPWT) intervention: Prophylactic NPWT supplied by KCI Medical (San Antonio, Texas) was applied at the time of primary skin closure at caesarean delivery and was placed over the closed surgical incision under sterile conditions and removed between postoperative day 5 and 7 at the time of incision check. Group 2 (control) intervention: standard dressing after caesarean delivery was applied using a sterile technique. If subcuticular closure was used, sterile slim adhesive strips (also known as Steri‐Strips) were applied. For both sub‐cuticular and staple closure, the dressing consisted of a sterile nonadherent wound dressing (also known as Telfa), a sterile gauze, and a waterproof transparent adhesive dressing (also known as Tegaderm). The standard dressing was removed on postoperative day 2. Study date/s: January 2015 to January 2017
Outcomes	Primary outcome: occurrence of surgical site infection defined according to Centers for Disease Control and Prevention criteria (superficial SSI) Composite wound complication, including superficial, deep, or organ‐space surgical site infection Wound dehiscence Seroma within 30 days of surgery Haematoma within 30 days of surgery 30‐day readmission, 30‐day reoperation Validity of measure/s: Centers for Disease Control and prevention criteria were used. Time points: 1 week and 30 days postoperatively
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "A randomization program (www.randomization. com, Alberta, Canada) was used to generate sealed opaque envelopes with study assignment. Women were randomised at the conclusion of the cesarean delivery, during skin closure, when the envelopes were opened by a circulating operating room nurse. Two randomization strata were created using permuted blocks with varying block sizes for women with BMIs from 35 to less than 40 and women with BMIs of 40 or higher and for each site to ensure equal distribution of study allocation across these two separate BMI categories and sites." Comment: appropriate method of sequence generation.
Allocation concealment (selection bias)	Low risk	Quote: "A randomization program (www.randomization. com, Alberta, Canada) was used to generate sealed opaque envelopes with study assignment. Women were randomised at the conclusion of the cesarean delivery, during skin closure, when the envelopes were opened by a circulating operating room nurse. Two randomization strata were created using permuted blocks with varying block sizes for women with BMIs from 35 to less than 40 and women with BMIs of 40 or higher and for each site to ensure equal distribution of study allocation across these two separate BMI categories and sites." Comment: centrally‐generated sequence of sealed opaque envelopes. Sequential numbering of envelopes may be inferred.
Blinding of participants and personnel (performance bias) All outcomes	High risk	Quote: "We conducted a randomised controlled, nonblinded, multicenter study". Comment: not blinded
Blinding of outcome assessment (detection bias) All outcomes	High risk	Quote: "We conducted a randomised controlled, nonblinded, multicenter study". Comment: not blinded
Incomplete outcome data (attrition bias) All outcomes	Low risk	Small attrition rate. Worst case scenario analysis performed for patients lost to follow‐up.
Selective reporting (reporting bias)	Low risk	All prespecified outcomes fully reported.
Other bias	Low risk	No evidence of other sources of bias; adequate reporting.

Study	Reason for exclusion
Abesamis 2019	Ineligible study design ‐ not an RCT
Albert 2012	No acute wounds were included.
Al‐Inany 2002	Ineligible intervention
Anderson 2014	Feasibility study. Predefined criteria used to assess feasibility included: recruitment (> 75% participation); loss to follow‐up (< 10%); intervention fidelity (= 95%); and interrater reliability (kappa = 0.8). Assessment of clinical outcomes was not planned or conducted.
Athanasiou 2018	Commentary on an RCT; not original research
Banasiewicz 2013	Included infected wounds
Bi 2017	Ineligible intervention
Bondokji 2011	Prospective cohort study
Braakenburg 2006	Chronic and acute wounds were reported together, and further information was not available.
Cantero 2016	Ineligible study design ‐ not an RCT
Chang 2018	Discussion article
Chiang 2017	Open wounds
Chio 2010	Skin graft study
Cocjin 2019	Ineligible patient population ‐ open wounds
Costa 2018	Ineligible population ‐ wounds healing by secondary intention
De Rooij 2020	Ineligible patient population
Dorafshar 2012	The study used NPWT to treat existing non‐healing skin graft wounds.
Dragu 2011	Ineligible study design ‐ not RCT
Echeribi 2015	Ineligible study design ‐ not an RCT‐based economic evaluation
Eisenhardt 2012	Skin graft study; no inclusion of wounds healing by primary closure
Erne 2018	Ineligible intervention
Fang 2020	Ineligible study design ‐ not an RCT
Fleming 2018	Ineligible study design ‐ not an RCT
Frazee 2018	Ineligible comparison
Grauhan 2013	Quasi‐randomised study: "A total of 156 patients were enrolled and allocated to 2 study groups, alternating according to the time of operation".
Hu 2009	Acute, subacute, and chronic wounds were included. Acute wounds were defined as those that had been "open" for less than 1 week.
Johannesson 2008	The intervention dressing was not a continuous negative pressure device.
Joos 2015	Commentary on an RCT in wounds healing by secondary intention
Kim 2007	The study was not a randomised controlled trial.
Kim 2015	Ineligible study design ‐ not an RCT
Kim 2020	Ineligible study population
Krishnamoorthy 2012	Use of NPWT was not the only difference between the groups.
Li 2016	Quasi‐randomisation (by odd and even numbers)
Licari 2020	Ineligible study design ‐ not RCT‐based economic evaluation
Llanos 2006	Skin graft study
Lychagin 2021	Ineligible intervention
Moisidis 2004	Skin graft study; no inclusion of wounds healing by primary closure
Monsen 2015	Ineligible patient population
Mouës 2004	No inclusion of acute wounds
Mouës 2007	No inclusion of acute wounds
Mujahid 2020	Ineligible patient population ‐ people with skin grafts
Muller‐Sloof 2018	Ineligible population
Muoghalu 2019	Ineligible study design ‐ not an RCT
NCT00724750	Non relevant population (not closed surgical wounds)
Pellino 2014	Non‐randomised study in people with Crohn's disease
Petkar 2012	Skin graft study
Rahmanian‐Schwarz 2012	Included chronic and acute wounds, and these were not separately reported
Seidel 2020	Ineligible patient population
Sinha 2016	Ineligible population; infected wounds
Stannard 2006	Ineligible population; not closed incision wounds
Stannard 2009	Ineligible patient population
Stapleton 2015	Ineligible study design ‐ not an RCT
Svensson‐Bjork 2018	Non‐randomised subgroup of RCT participants
Trofa 2019	Ineligible comparison
Visser 2017	The vacuum therapy device was a syringe inserted subcutaneously into the dressing, which was used to create a vacuum. Consequently, it was not a standard, continuous pressure device.
Walker 2018	Ineligible intervention
Wang 2019	Ineligible patient population
Xu 2019	Ineligible intervention
Yongchao 2017	Ineligible patient population
Yu 2017	A drain was left inside the wound, so not strictly a primarily closed wound.
Zhang 2020	Ineligible intervention
Zhuang 2020	Ineligible study design ‐ not an RCT
Zotes 2015	Ineligible population; infected wounds

Study name	Do suction assisted negative pressure dressings reduce the incidence of surgical site infections after abdominal surgery: a randomized controlled trial
Methods	Randomised controlled trial
Participants	Patients undergoing laparotomy (where abdominal incision breaches peritoneum, and wound is large enough to at least fit the surgeon's hand)
Interventions	Negative pressure wound therapy versus standard dressing used with a clear film with an absorbent layer
Outcomes	Wound infection; patient satisfaction
Starting date	2015
Contact information	peeyau.tan@monashhealth.org
Notes

Study name	The effect of PICO dressings on surgical site infection following bowel resection: a randomised controlled trial
Methods	Randomised controlled trial
Participants	All adults (aged 18 and over) undergoing elective or emergency small or large bowel resection
Interventions	Negative pressure wound therapy (PICO dressing) versus standard dressing
Outcomes	SSI; Patient and Observer Scar Assessment Scale (POSAS); patient satisfaction
Starting date	2018
Contact information	Alexandra.Gordon@midcentraldhb.govt.nz
Notes

PERMALINK

Negative pressure wound therapy for surgical wounds healing by primary closure

Gill Norman

Chunhu Shi

En Lin Goh

Elizabeth MA Murphy

Adam Reid

Laura Chiverton

Monica Stankiewicz

Jo C Dumville

Abstract

Background

Objectives

Search methods

Selection criteria

Data collection and analysis

Main results

Authors' conclusions

Plain language summary

Summary of findings

Summary of findings 1. Negative pressure wound therapy compared with standard dressing for surgical wounds healing by primary closure.

Background

Description of the condition

Description of the intervention

How the intervention might work

Why it is important to do this review

Objectives

Methods

Criteria for considering studies for this review

Types of studies

Types of participants

Types of interventions

Types of outcome measures

Primary outcomes

Secondary outcomes

Search methods for identification of studies

Electronic searches

Searching other resources

Data collection and analysis

Selection of studies

Data extraction and management

Assessment of risk of bias in included studies

1.

Measures of treatment effect

Unit of analysis issues

Dealing with missing data

Assessment of heterogeneity

Assessment of reporting biases

Data synthesis

Economic analyses

Subgroup analysis and investigation of heterogeneity

Sensitivity analysis

Summary of findings and assessment of the certainty of the evidence

Results

Description of studies

1. Primary outcome data.

2. Secondary outcome data.

3. Economic outcome data.

Results of the search

2.

Interventions search

Economic analysis search

Included studies

Types of participants

Types of interventions

Types of economic assessments

Excluded studies

Ongoing studies

Studies awaiting classification

Risk of bias in included studies

3.

4. Quality assessment of economic studies using the Consolidated Health Economic Evaluation Reporting Standards (CHEERS) checklist.

Risk of bias in economic studies

Effects of interventions

Comparison 1: NPWT compared with standard dressing (62 trials, 13340 participants)

Primary outcomes

Mortality (follow‐up period 30 days to 90 days or unspecified)

1.1. Analysis.

Surgical site infection (follow‐up period 30 days to 12 months or unspecified)

1.2. Analysis.

Study name	EffiCacY of neGative pressure wound therapy in the preventioN of surgical woUnd complicationS in the cesarean section at risk population: a randomised multi‐centre trial, the CYGNUS trial
Methods	Randomised controlled trial
Participants	Pregnant women between 18‐50 years undergoing caesarean section
Interventions	Negative pressure wound therapy versus standard dressing
Outcomes	SSI; wound dehiscence
Starting date	2018
Contact information	kylie.sandy‐hodgetts@uwa.edu.au
Notes

Study name	Negative pressure wound therapy to reduce incisional wound infections ‐ a randomised control trial
Methods	Randomised controlled trial
Participants	People aged at least 18 years undergoing emergency laparotomy with primary closure
Interventions	NPWT (Prevena) Standard dressings
Outcomes	SSI
Starting date	02 March 2020 (actual) planned 10 June 2019
Contact information	neil.strugnell@nh.org.au
Notes

Study name	Negative pressure wound therapy (NPWT) on closed incisions to prevent surgical site infection in high‐risk patients in hepatopancreatobiliary surgery: the NP‐SSI trial
Methods	Randomised controlled trial
Participants	People undergoing hepatopancreatobiliary surgery who are aged at least 49 years
Interventions	NPWT (Prevena) Conventional gauze dressings
Outcomes	SSI (superficial and deep) Complications including seroma, haematoma, dehiscence) Quality of life Reoperation
Starting date	1 May 2019
Contact information	frank.brennfleck at ukr.de
Notes	DRKS00015136

Study name	Prevention surgical site infection with using negative pressure wound therapy in abdominal incision
Methods	Parallel randomised controlled trial
Participants	High‐risk patients: including abdominal surgery for malignancy, colorectal, abdominal wall reconstruction
Interventions	Negative pressure wound therapy versus routine approach
Outcomes	Rate of surgical site infection
Starting date	2015
Contact information	hpzhangly@163.com
Notes

Study name	To study the effect of negative pressure dressings in preventing surgical site infection after emergency midline abdominal surgery; to evaluate negative pressure dressings in decreasing surgical site infections after emergency lapatotomy: a randomized controlled study.
Methods	Randomised controlled trial
Participants	Patients of 18‐65 years of age undergoing emergency midline laparotomy for peritonitis
Interventions	NPWT Conventional saline dressing
Outcomes	SSI Dehiscence
Starting date	09 September 2019
Contact information	drkartiksahni@gmail.com
Notes

Study name	Usage of vaccum therapy in closed abdominal incision in reducing surgical site infection; prophylactic Negative Pressure Wound Therapy in reducing surgical site infections in closed abdominal incision: a randomized controlled trial.
Methods	Randomised controlled trial
Participants	People with intestinal disease aged 18 to 80 years
Interventions	NPWT Sterile gauze
Outcomes	SSI Seroma Haematoma Dehiscence Readmission
Starting date	02 November 2020
Contact information	maharjanmanik44@gmail.com
Notes

Study name	Prophylactic negative wound therapy in laparotomy wounds (PROPEL trial): randomized controlled trial
Methods	Randomised controlled trial
Participants	People aged over 18 undergoing emergency or elective laparotomy for benign or malignant conditions
Interventions	NPWT Prevena NPWT PICO Standard surgical dressings
Outcomes	Superfical SSI Seroma Haematoma Dehiscence
Starting date	6 November 2019
Contact information	donlonn@tcd.ie
Notes	NCT03871023

Study name	Postoperative negative pressure incision therapy following open colorectal surgery: a randomized‐controlled trial
Methods	Randomised controlled trial
Participants	Patients undergoing planned elective open colorectal surgery via median or transverse laparotomy
Interventions	Negative pressure wound therapy versus standard wound dressings
Outcomes	Rate of SSI; length of hospital stay; rate of reoperations; rate of antibiotic therapy; duration of postoperative negative pressure incision therapy (intervention arm only); wound pain assessed with VAS; rate of wound complications other than wound infections; rate of serious adverse events
Starting date	1 October 2015
Contact information	Unclear
Notes

Study name	Evaluation of negative pressure incisional therapy in urgent gastrointestinal surgery for reduction of superficial surgical site infections compared to non‐occlusive conventional plaster ‐ a prospective, randomized, controlled, multicenter clinical trial
Methods	Randomised controlled trial
Participants	Patients undergoing urgent laparotomy due to an acute gastrointestinal disorder
Interventions	Negative pressure wound therapy versus non‐occlusive conventional plaster
Outcomes	SSI; prolongation of hospitalisation due to SSI; cosmetic result; safety endpoints: AEs, SAEs
Starting date	21 September 2016
Contact information	Unclear
Notes

Study name	Single use negative pressure wound therapy system (Prevena ™) compared to standard wound care after spinal surgery
Methods	Randomised controlled trial
Participants	Patients aged at least 18 years undergoing spinal surgery in the thoracic or lumbar spine from dorsal approach
Interventions	NPWT (Prevena) Conventional OpSite dressing
Outcomes	Wound healing disorders Revision (reoperation)
Starting date	25 June 2020
Contact information	ahmed.bassem at diakovere.de
Notes