Closed Incision Negative Pressure Therapy versus Standard of Care in Reduction of Surgical Site Complications: A Systematic Review and Meta-analysis

Abstract

Introduction: 

Closed incision negative pressure therapy (ciNPT) has been utilized to help manage closed incisions across many surgical specialties. This systematic review and meta-analysis evaluated the effect of ciNPT on postsurgical and health economic outcomes.

Methods: 

A systematic literature search using PubMed, EMBASE, and QUOSA was performed for publications written in English, comparing ciNPT to standard-of-care dressings between January 2005 and August 2021. Study participant characteristics, surgical procedure, dressings used, treatment duration, postsurgical outcomes, and follow-up data were extracted. Meta-analyses were performed using random-effects models. Risk ratios summarized dichotomous outcomes. Difference in means or standardized difference in means was used to assess continuous variables reported on the same scale or outcomes reported on different scales/measurement instruments.

Results: 

The literature search identified 84 studies for analysis. Significant reductions in surgical site complication (SSC), surgical site infection (SSI), superficial SSI, deep SSI, seroma, dehiscence, skin necrosis, and prolonged incisional drainage were associated with ciNPT use (P < 0.05). Reduced readmissions and reoperations were significant in favor of ciNPT (P < 0.05). Patients receiving ciNPT had a 0.9-day shorter hospital stay (P < 0.0001). Differences in postoperative pain scores and reported amounts of opioid usage were significant in favor of ciNPT use (P < 0.05). Scar evaluations demonstrated improved scarring in favor of ciNPT (P < 0.05).

Discussion: 

For these meta-analyses, ciNPT use was associated with statistically significant reduction in SSCs, SSIs, seroma, dehiscence, and skin necrosis incidence. Reduced readmissions, reoperation, length of hospital stay, decreased pain scores and opioid use, and improved scarring were also observed in ciNPT patients.

Takeaways

Question: Does the use of a single closed incision negative pressure therapy device (ciNPT) impact patient and health economic outcomes?

Findings: Use of ciNPT was associated with reductions in surgical site complications, surgical site infection, seroma, dehiscence, and skin necrosis incidence. Reduced readmissions, reoperation, length-of-hospital stay, patient pain scores, opioid use, and improved scarring were also observed.

Meaning: Improved patient and health economic outcomes with ciNPT use suggest it is effective for surgical incision management across the adult surgical patient population.

INTRODUCTION

Surgical site complications (SSCs), such as surgical site infection (SSI), dehiscence, seroma, hematoma, and skin necrosis, are associated with increased patient morbidity and mortality.¹ Patients who develop SSCs often experience delayed wound healing, decreased quality of life, and increases in length of stay (LOS), medical procedures performed, pain, and potential reoperation and readmission.^1–3 Increased use of resources with management of SSCs directly impacts cost of care. American College of Surgeons National Surgical Quality Improvement Program (ACS NSQIP) data from 2012 to 2018 at a single institution reported a 1.5-fold mean relative increase of indirect hospital costs associated with the treatment of any postoperative complication.⁴ Direct hospital costs were increased by 1.98 times when reoperation was required, 1.3 times with SSI treatment, and 1.6 times with dehiscence management.⁴ Additionally, Merkow et al⁵ reported increases in 2016 hospital costs of $6477 for SSI, $8524 for readmission, and $15,166 for reoperation for four ACS NSQIP hospitals.

Although rates of SSCs have declined with improved infection prevention and sterile surgical methods, they are still a substantial burden. To help further mitigate the risk of SSC development, surgeons utilize postoperative surgical site management. Options exist for surgical site management, including the use of closed incision negative pressure therapy (ciNPT). One manufacturer’s ciNPT device (3M Prevena Incision Management System, 3M Company, St. Paul, Minn.) is a disposable, single-use system with a replaceable canister and a variety of one-piece reticulated open-cell foam dressing options for different incision lengths and suitable for a variety of anatomical locations. This system is placed in the operating room over closed incisions and provides continuous negative pressure at −125 mm Hg. The dressings can be utilized for up to 7 days without requiring a dressing change. The use of ciNPT has been shown to manage the environment of closed surgical incisions, help hold the incision edges together, remove fluid and infectious materials, and act as a barrier to external contamination.^6–8 This −125 mm Hg ciNPT device was the first surgical incision management system cleared for use by the US Food and Drug Agency to aid in reducing the incidence of seroma and superficial SSI in class 1 and 2 wounds in adult patients at high risk for postoperative infections. The use of ciNPT is not intended to treat SSIs or seroma, and clinicians should consult the manufacturer’s instructions as needed before application. Other available ciNPT devices have different negative pressure settings, multilayer absorbent dressings, and small or no canister for exudate collection, and may lack an audible alarm for negative pressure failure. These system differences may impact postsurgical outcomes. For instance, two recent meta-analyses observed differences in rates of SSIs, complications, and drainage for patients who received −125 mm Hg ciNPT versus traditional dressings compared to patients who received −80 mm Hg ciNPT versus traditional dressings.^9,10 However, clinical studies that specifically compare these two systems are lacking,¹⁰ and there may have been underlying differences in patient and wound comorbidities, surgical procedures, and study protocols that were not controlled for and may have biased observed outcomes.

The effectiveness of −125 mm Hg ciNPT in reducing the incidence of SSIs and seroma in all surgical procedures and populations has not yet been fully assessed. Published literature has reported reduction of SSCs, including SSI, dehiscence, and seroma with use of ciNPT for various surgical procedures.^11–16 A previously published meta-analysis by Singh et al¹⁷ demonstrated a statistically significant reduction in SSI with the use of a −125 mm Hg ciNPT device compared to standard-of-care (SOC) dressings. Given the emerging literature, the gap in knowledge regarding the effect of ciNPT on surgical complications and health economic outcomes across surgical procedures, and the heterogeneity of ciNPT devices and patient populations, the objective of this systematic review and meta-analysis was to expand on the previous work of Singh et al by assessing SSC rates and the health economic impact of −125 mm Hg ciNPT (Prevena Incision Management System, 3M Company) use over closed incisions versus SOC. In addition to broadening the scope of the analysis, we included literature published between 2005 and 2021, resulting in a larger, more current body of evidence. Observed postsurgical outcomes included SSI, hematoma, seroma, dehiscence, skin necrosis, surgical site drainage, antibiotic usage, opioid analgesia usage, and patient-reported pain. Health economic outcomes, such as hospital LOS, readmission, and reoperation, were also assessed.

METHODS

This work conformed to the statement and reporting checklist of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses.¹⁸ The systematic literature review and meta-analysis were conducted to assess the effect of ciNPT (Prevena Incision Management System, 3M) versus SOC use over closed incisions. Outcomes assessed included the number of SSCs (surgical site infections, hematoma, seroma, dehiscence, and skin necrosis), health economic endpoints (hospital LOS, readmissions, and reoperations), and additional endpoints of drainage, antibiotic use, patient-reported pain, opioid analgesia use, and scar quality.

Literature Search

A systematic literature search was conducted on August 5, 2021 using PubMed, EMBASE, and QUOSA. Literature between January 1, 2005 and August 5, 2021 was assessed. The following search terms were used: (“negative pressure wound therapy” OR “negative pressure” OR “negative pressure therapy” OR “NPWT”) AND (“Prevena” OR “ciNPT” OR “prophylactic NPWT” OR “preventive NPWT” OR “incision management” OR “incisional management” OR “closed incision negative pressure wound therapy” OR “closed incision negative pressure therapy”). Inclusion criteria consisted of published English language abstracts or articles comparing the use of one manufacturer’s ciNPT (Prevena Incision Management System, 3M) over closed incision to SOC. Exclusion criteria were meta-analyses, preclinical studies (animal or bench studies), veterinary studies, pediatric patient population, noncomparative studies, or use of a non-ciNPT negative pressure device.

Studies were selected for inclusion after a review of titles and abstracts to identify additional articles for further review. Abstracts and articles that met all the inclusion criteria and none of the exclusion criteria underwent a full-text assessment by two independent reviewers. When disagreement occurred, a third person reviewed the article, and eligibility was decided by consensus.

Data Extraction

Data extraction from all eligible studies was completed by one reviewer and validated by a second reviewer. Disagreements were resolved by discussion between the two reviewers or by the addition of a third reviewer. Extracted data included surgical procedures, patient demographics, SOC type, the number of treatment days, follow-up period, the number of patients/incisions, and the number of patients/incisions analyzed. Patient outcomes data such as the number of SSCs (a reported composite endpoint including SSCs such as SSI, dehiscence, seroma, hematoma, and skin necrosis), SSIs, type of SSI, dehiscence, seroma, hematoma, skin necrosis, readmission, reoperations, deaths, amputations, drainage amount, the number of days with surgical drain, patients with postoperative pain, patients with postoperative antibiotic treatment, and quality of life/patient satisfaction measures were extracted. Additionally, health economic measures, including hospital LOS, readmissions, reoperations, and reported costs, were recorded.

Statistical Analysis

The effect of ciNPT use compared to SOC was assessed for SSC, SSI, seroma, hematoma, dehiscence, skin necrosis, readmission, reoperation, length of stay, drainage, antibiotic usage, patient-reported pain, opioid analgesia use (morphine equivalents), and surgical site scarring. Subgroup analyses include RCT only and observational studies only for SSC and SSI.

To assess the effect of dichotomous variables, weighted risk ratios (RRs) were calculated to pool study and control groups in each publication for analysis. Difference in means as ciNPT minus SOC was used to assess differences between continuous variables reported on the same scale. Standardized difference in means was used if outcomes were reported with different scales or measurement instruments. A sensitivity analysis using Duval and Tweedie’s Trim and Fill method was performed for the meta-analyses.¹⁹ I² was utilized to assess data heterogeneity. Treatment effects were combined, and the random-effects model was used for each analysis performed, regardless of the heterogeneity assessment. The more conservative random-effects model, which often has higher estimated variances and, therefore, wider confidence intervals, was chosen over the fixed-effects model, which assumes low data heterogeneity. All analyses were performed using Comprehensive Meta-Analysis Version 3 (Borenstein M, Hedges L, Higgins J, and Rothstein H. Biostat Englewood, NJ 2013).

RESULTS

Literature Search

A total of 972 publications were identified. After removal of 362 duplicate publications, 610 abstracts and titles were screened against the inclusion and exclusion criteria. Reasons for exclusion are listed in Figure 1. Eighty-four articles were included for analysis (Fig. 1; see Table, Supplemental Digital Content 1, which summarizes the SSC meta-analyses, https://links.lww.com/PRSGO/C323). Twenty-four articles were randomized controlled trials (RCTs), 17 were prospective studies, and 43 were retrospective studies. The 84 studies identified encompassed cardiac surgery (n = 8), general abdominal surgery (n = 21), obstetric and gynecologic surgery (n = 8), orthopedic surgery (n = 16), plastic surgery (n = 16), and vascular surgery (n = 15).^{11–16,20–97}

Fig. 1.: Flowchart showing the identification of studies for the meta-analysis.

Study Characteristics

The included studies represent a global surgical patient population with 7451 patients receiving ciNPT and 16,085 patients receiving SOC across 16 countries.^{11–16,20–97} The countries represented included the USA (n = 42), Germany (n = 11), Australia (n = 6), Canada (n = 5), Italy (n = 4), Spain (n = 3), the UK (n = 3), Denmark (n = 2), and one each from Austria, Brazil, France, Ireland, the Netherlands, South Africa, Taiwan, and Turkey. Dressings used in the SOC population varied and included gauze dressings, antimicrobial hydrofiber dressings, absorbent dressings, hydrocolloid dressings, occlusive dressings, or foam dressings.^{11–16,20–97}

Surgical Site Complication Outcomes

Results from meta-analyses of SSCs are reported in Table, Supplemental Digital Content 1 (https://links.lww.com/PRSGO/C323). Forty-six studies reported on a composite end point of SSCs. (See Table, Supplemental Digital Content 1,https://links.lww.com/PRSGO/C323.) Significant differences in SSC rates were found in favor of ciNPT use {RR, 0.543 [95% confidence interval (CI), 0.448–0.657]; P < 0.001}, indicating that the probability of an SSC with ciNPT was approximately cut in half compared to the probability of an SSC with traditional dressings with a relative risk reduction of 46%. (See Table, Supplemental Digital Content 1,https://links.lww.com/PRSGO/C323; Fig. 2.) The subgroup analyses by study design demonstrated similar results for the 13 RCTs [RR, 0.653 (95% CI, 0.481–0.885); P = 0.006] and the 33 observational studies [RR, 0.501 (95% CI, 0.395–0.636); P < 0.001]. (See Table, Supplemental Digital Content 1,https://links.lww.com/PRSGO/C323; Fig. 3.)

Fig. 2.: SSC forest plot. Each study is displayed with the risk ratio, upper and lower limit, total number of events, and relative weight. The total represents the overall weighted risk ratio and limits using a random-effects model.

Fig. 3.: SSC in randomized controlled trials forest plot. Each study is displayed with the risk ratio, upper and lower limit, total number of events, and relative weight. The total represents the overall weighted risk ratio and limits using a random-effects model.

Individual SSCs, including SSI, seroma, hematoma, dehiscence, and skin necrosis, were also assessed. A total of 65 studies were included in the analysis of SSIs. (See Table, Supplemental Digital Content 1,https://links.lww.com/PRSGO/C323.) For this end point, there was a significant difference in favor of ciNPT use with a risk ratio of 0.530 (P < 0.001; Fig. 4). (See Table, Supplemental Digital Content 1, https://links.lww.com/PRSGO/C323.) Additional analyses evaluating the effect of ciNPT on SSIs were performed by study design. Reductions in SSIs were found in favor of ciNPT for the 19 RCTs [RR, 0.678 (95% CI, 0.546–0.843); P < 0.001; Fig. 5]. (See Table, Supplemental Digital Content 1, https://links.lww.com/PRSGO/C323.) Similar findings were seen for the 46 observational studies [RR, 0.470 (95% CI, 0.367–0.602); P < 0.001]. (See Table, Supplemental Digital Content 1, https://links.lww.com/PRSGO/C323.) Significant differences were also found in favor of ciNPT use for SSSI (RR, 0.505; P < 0.001), DSSI (RR, 0.469; P = 0.002), seroma (RR, 0.677; P = 0.004), dehiscence (RR, 0.644; P = 0.022), and skin necrosis (RR, 0.466; P = 0.001; Figs. 6–10). (See Table, Supplemental Digital Content 1, https://links.lww.com/PRSGO/C323). Hematoma rates were similar between the two groups (P = 0.183) (see Table, Supplemental Digital Content 1, https://links.lww.com/PRSGO/C323; Fig. 11).

Fig. 4.: SSI forest plot. Each study is displayed with the risk ratio, upper and lower limit, total number of events, and relative weight. The total represents the overall weighted risk ratio and limits using a random-effects model.

Fig. 5.: SSI in randomized controlled trials forest plot. Each study is displayed with the risk ratio, upper and lower limit, total number of events, and relative weight. The total represents the overall weighted risk ratio and limits using a random-effects model.

Fig. 6.: Superficial SSI forest plot. Each study is displayed with the risk ratio, upper and lower limit, total number of events, and relative weight. The total represents the overall weighted risk ratio and limits using a random-effects model.

Fig. 7.: Deep SSI forest plot. Each study is displayed with the risk ratio, upper and lower limit, total number of events, and relative weight. The total represents the overall weighted risk ratio and limits using a random-effects model.

Fig. 8.: Seroma forest plot. Each study is displayed with the risk ratio, upper and lower limit, total number of events, and relative weight. The total represents the overall weighted risk ratio and limits using a random-effects model.

Fig. 9.: Hematoma forest plot. Each study is displayed with the risk ratio, upper and lower limit, total number of events, and relative weight. The total represents the overall weighted risk ratio and limits using a random-effects model.

Fig. 10.: Dehiscence forest plot. Each study is displayed with the risk ratio, upper and lower limit, total number of events, and relative weight. The total represents the overall weighted risk ratio and limits using a random-effects model.

Fig. 11.: Skin necrosis forest plot. Each study is displayed with the risk ratio, upper and lower limit, total number of events, and relative weight. The total represents the overall weighted risk ratio and limits using a random-effects model.

Health Economic Outcomes

Readmission rates were evaluated for 24 studies, and a significant reduction was seen in patients treated with ciNPT compared to traditional dressings with a risk ratio of 0.773 (P = 0.039; Table 1). Similarly, there was a statistically significant reduction in reoperation rates for ciNPT patients (RR, 0.64; P < 0.001; Table 1). A total of 25 studies reported hospital LOS statistics that could be included in the meta-analysis (Table 1). LOS was 0.9 days shorter for ciNPT patients (P < 0.001).

Table 1. - Summary of Health Economic Outcomes Meta-analyses

Outcome	Studies (n)	Study References	Statistic (95% CI)		I2	P -value
Readmission	24	13,22,25,26,30,32,36,46,53–57–,59,62,65,67,69,74,87,90–92–,95	RR	0.773 (0.606–0.987)	24.453	0.039
Reoperation	40	11–13–,16,20–22–,25,26,30,32–35–,37–39–,42,47,49,53–57–,59,62,64,66,73–75–,78,81,85,88,91–93–,95	RR	0.64 (0.519–0.789)	12.271	<0.001
LOS	25	13,16,25,32,33,36,40,46,48,59,62,65,66,68,74,78,80,81,85–89–,91,92	DM	−0.901 (−1.257 to −0.544)	80.458	<0.001

DM, difference in means.P-values in bold represent statistically significant values.

Other Patient Outcomes

Six studies reported the fluid volume removed from the surgical site via drains (Table 2). Patients receiving ciNPT had approximately 122 mL less fluid volume (95% CI, −253 mL to 7.7 mL; P = 0.065), although this was not statistically significant. The rate of postoperative antibiotic use was reported in nine studies and was lower in ciNPT patients, though not statistically significant (P = 0.053; Table 2). In the three studies that reported patient pain levels, patients in the ciNPT group reported significantly lower levels of postoperative pain (P < 0.001; Table 2). Opioid analgesia use (morphine equivalents) was reported in two studies (Table 2). Patients who underwent ciNPT received 20 mg (95% CI, 5 mg–35 mg) less than patients who received traditional dressings (P = 0.009). Three studies reported that scar evaluations were analyzed (Table 2). At 90-day follow-up, scars were evaluated using the Vancouver Scar Scale (VSS) (n = 2), resulting in significant improvement in incisional scar quality in favor of ciNPT (P < 0.001). Scar evaluation at 12 months (n = 2) also showed improvement in favor of ciNPT (P = 0.048). The overall scar evaluation (n = 3) demonstrated a significant scar improvement for patients receiving ciNPT (P = 0.014; Table 2).

Table 2. - Summary of Other Patient Outcomes Meta-analyses

Outcome	Studies (n)	Study References	Statistic (95% CI)		I2	P -value
Drainage (mL)	6	21,43,58,72,77,79	DM	−122.8 (−253.2 to 7.7)	83.4	0.065
Antibiotics	9	16,30,37,49,54,55,58,90,95	RR	0.697 (0.484–1.005)	68.811	0.053
Pain score	3	43,46,81	SDM	−0.499 (−0.668 to −0.331)	0	<0.001
Opioid use (mg)	2	52,86	DM	−20.098 (−35.198 to −4.999)	0	0.009
Scarring at 90 d (VSS)	2	20,43	DM	−5.111 (−5.935 to −4.287)	45.172	<0.001
Scarring at 12 mo	2	15,43	SDM	−1.728 (−3.440 to −0.017)	72.100	0.048
Scarring overall	3	15,20,43	SDM	−2.543 (−4.564 to −0.521)	82.075	0.014

DM, difference in means; SDM, standardized difference in means; VSS, Vancouver Scar Scale.P-values in bold represent statistically significant values.

DISCUSSION

This work expands on Singh et al¹⁷ and addresses limitations identified in the published article. This systematic review and meta-analysis examined the effect of a single ciNPT device on multiple SSC types and patient and health economic outcomes. By expanding to include other outcomes (such as dehiscence, skin necrosis, readmission rates, scar quality, and antibiotic usage) and increasing the number of studies assessed (84 versus 30), a more complete picture of the potential benefits of ciNPT use across a broad surgical patient population is created. Additionally, the included articles were published from 2013 to 2021, resulting in the assessment of ciNPT use among current trends in surgical site management. In this meta-analysis, ciNPT use was associated with statistically significant reductions in SSCs, SSIs, seroma, dehiscence, skin necrosis incidence, readmissions, reoperation, LOS, pain score, opioid use, and significant improvement in scarring.

Previously published meta-analyses examining negative pressure therapy use over closed incisions exist; however, these use multiple therapy systems and reported reductions in SSI, dehiscence, seroma, skin necrosis, or LOS.^98–100 The inclusion of multiple negative pressure therapy systems into analyses assumes that all systems are the same and fails to tease out any differences in efficacy. The previously published Singh et al meta-analysis included 30 studies published up to April 2018 on the use of a single ciNPT device. The single SSI outcome was the publication’s focus with subanalyses by study design, RCT or observational, and surgery type. The results in our meta-analysis are similar to previously published literature reporting significant reductions in SSCs, SSIs, seroma, dehiscence, and skin necrosis incidence; however, this meta-analysis examined outcomes from a single −125 mm Hg ciNPT device across a larger volume of recently published evidence. As such, patient outcomes in favor of this single ciNPT device point to its efficacy across the global surgical patient population within the current surgical site management landscape.

This meta-analysis also assessed health economic outcomes associated with use of ciNPT. In a meta-analysis on a −80 mm Hg ciNPT device, LOS was significantly reduced compared to SOC; however, no significant differences were found in readmission rates or reoperation rates despite including nine studies in each analysis.¹⁰¹ This is in contrast to our meta-analysis where hospital readmission, reoperation, and LOS were significantly reduced following ciNPT use across 84 publications. These results indicate a cost-effective benefit of surgical incision management with the use of −125 mm Hg ciNPT. In a retrospective analysis using data from the Premier Healthcare Database, a large, US hospital–based, all-payer database, the incidence of 90-day SSCs on open surgical procedures is estimated at 6.4% with a cost per SSC of $17,317.¹⁰² This leads to a mean cost of SSC per patient of $1108. These meta-analyses have demonstrated a 46% risk reduction in SSCs with the use of this ciNPT device. A 46% reduction from the baseline of 6.4% would result in an SSC rate of 3.5% and mean cost of SSC per patient of $598. If the cost of the ciNPT device is factored in at $495 per patient, this would result in a total mean cost per patient of $1093. This indicates that the reduced SSC rate with use of −125 mm Hg ciNPT in an open surgical procedure population may occur without introduction of additional costs. This is a conservative estimate, given that the cost- effectiveness of ciNPT device use can be increased by selective application based on predefined SSC or SSI high-risk criteria.

Significant reductions in patient pain score, opioid analgesic use, and significant improvements in scarring were observed in the ciNPT population; however, only a small number of studies reporting these outcomes were available. While the potential benefit of the reduced outcomes was associated with ciNPT use, more evidence is needed to confirm these findings.

LIMITATIONS

Limitations of this meta-analysis include use of a mix of observational studies and RCTs and differences in data reporting. The studies included in the analysis are heterogeneous and range in quality from low to high. Additionally, data reporting was varied with some outcomes either not reported or unable to be evaluated. For example, hospital LOS was reported in 40 studies; however, only data from 25 studies could be assessed. Furthermore, the definition of SSC differed across studies. While study quality and data reporting discrepancies may have affected the strength of the analyses, use of the conservative random-effects model was employed to mitigate the effect of data heterogeneity. However, interpretation of the results should be made with caution. Results of certain endpoints, such as pain score, opioid analgesia use, and scar quality, were also limited due to reporting in a small number of studies. Although significant differences favoring the ciNPT population were observed, the small number of studies included in these subanalyses limits the strength of the results. Lastly, these results are related to one commercially available ciNPT system and may not be applicable to other available systems.

CONCLUSIONS

This systematic review and meta-analysis demonstrated that ciNPT use was associated with statistically significant reduction in SSCs, SSIs, seroma, dehiscence, and skin necrosis incidence. Use of ciNPT was also associated with reduced readmissions, reoperation, and LOS; decreased pain scores and opioid use; and improved scarring.

ACKNOWLEDGMENT

The authors thank Julie M. Robertson, PhD (3M), for assistance with manuscript preparation and editing.