Negative pressure wound therapy for surgical wounds healing by primary closure

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 June 2019, 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 type of NPWT. 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. Main results In this third update, we added 15 new randomised controlled trials (RCTs) and three new economic studies, resulting in a total of 44 RCTs (7447 included participants) and five economic studies. Studies evaluated NPWT in the context of a wide range of surgeries including orthopaedic, obstetric, vascular and general procedures. Economic studies assessed NPWT in orthopaedic, obstetric and general surgical settings. All studies compared NPWT with standard dressings. Most studies had unclear or high risk of bias for at least one key domain. Primary outcomes Four studies (2107 participants) reported mortality. There is low‐certainty evidence (downgraded twice for imprecision) showing no clear difference in the risk of death after surgery for people treated with NPWT (2.3%) compared with standard dressings (2.7%) (risk ratio (RR) 0.86; 95% confidence interval (CI) 0.50 to 1.47; I2 = 0%). Thirty‐nine studies reported SSI; 31 of these (6204 participants), were included in meta‐analysis. There is moderate‐certainty evidence (downgraded once for risk of bias) that NPWT probably results in fewer SSI (8.8% of participants) than treatment with standard dressings (13.0% of participants) after surgery; RR 0.66 (95% CI 0.55 to 0.80 ; I2 = 23%). Eighteen studies reported dehiscence; 14 of these (3809 participants) were included in meta‐analysis. There is low‐certainty evidence (downgraded once for risk of bias and once for imprecision) showing no clear difference in the risk of dehiscence after surgery for NPWT (5.3% of participants) compared with standard dressings (6.2% of participants) (RR 0.88, 95% CI 0.69 to 1.13; I2 = 0%). Secondary outcomes There is low‐certainty evidence showing no clear difference between NPWT and standard treatment for the outcomes of reoperation and incidence of seroma. For reoperation, the RR was 1.04 (95% CI 0.78 to 1.41; I2 = 13%; 12 trials; 3523 participants); for seroma, the RR was 0.72 (95% CI 0.50 to 1.05; I2 = 0%; seven trials; 729 participants). The effect of NPWT on occurrence of haematoma or skin blisters is uncertain (very low‐certainty evidence); for haematoma, the RR was 0.67 (95% CI 0.28 to 1.59; I2 = 0%; nine trials; 1202 participants) and for blisters the RR was 2.64 (95% CI 0.65 to 10.68; I2 = 69%; seven trials; 796 participants). The overall effect of NPWT on pain is uncertain (very low‐certainty evidence from seven trials (2218 participants) which reported disparate measures of pain); but moderate‐certainty evidence suggests there is probably little difference between the groups in pain after three or six months following surgery for lower limb fracture (one trial, 1549 participants). There is also moderate‐certainty evidence for women undergoing caesarean sections (one trial, 876 participants) and people having surgery for lower limb fractures (one trial, 1549 participants) that there is probably little difference in quality of life scores at 30 days or 3 or 6 months, respectively. Cost‐effectiveness Five economic studies, based wholly or partially on trials included in our review, assessed the cost‐effectiveness of NPWT compared with standard care. They considered NPWT in four indications: caesarean sections in obese women; surgery for lower limb fracture; knee/hip arthroplasty and coronary artery bypass graft surgery. They calculated quality‐adjusted life‐years for treatment groups and produced estimates of the treatments' relative cost‐effectiveness. The reporting quality was good but the grade of the evidence 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 experiencing primary wound closure of their surgical wound and treated prophylactically with NPWT following surgery probably experience fewer SSI than people treated with standard dressings (moderate‐certainty evidence). There is no clear difference in number of deaths or wound dehiscence between people treated with NPWT and standard dressings (low‐certainty evidence). There are also no clear differences in secondary outcomes where all evidence was low or very low‐certainty. In caesarean section in obese women and surgery for lower limb fracture, there is probably little difference in quality of life scores (moderate‐certainty evidence). Most evidence on pain is very low‐certainty, but there is probably no difference in pain between NPWT and standard dressings after surgery for lower limb fracture (moderate‐certainty evidence). 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.