Meta-Analysis Comparing Outcomes of Two Different Closed Incision Negative Pressure Systems in Breast Surgery and Implications to Cost of Care

© 2024 HMP Global. All Rights Reserved.

Any views and opinions expressed are those of the author(s) and/or participants and do not necessarily reflect the views, policy, or position of ePlasty or HMP Global, their employees, and affiliates.

All surgical procedures carry the risk of surgical site complication (SSC) development. Common SSCs associated with surgery include surgical site infection (SSI), dehiscence, cellulitis, necrosis, and delayed wound healing. For breast surgery, SSC rates have been reported between 2.25% and 53%.^1-5 However, the development of SSCs may not be fully attributed to failure of therapy as both surgical and patient factors contribute to an individual’s risk of SSC development. Surgery type and procedure duration along with patient age, body mass index, tobacco use, and type and number of patient comorbidities can affect the risk of SSC development.^6-9

Advancements in patient care have created a number of options for incision management to help mitigate the risk of SSCs. Closed incision negative pressure therapy (ciNPT) offers clinicians the ability to manage incisions using the continuous delivery of negative pressure. ciNPT has previously been reported to help improve patient outcomes after orthopedic, abdominal, vascular, and thoracic procedures.^10-13 Additionally, multiple ciNPT devices are commercially available; however, not all devices are the same. Device differences can include amount of negative pressure provided by the device, type of dressings used, device dysfunction alarms, power source, and presence of a canister for exudate collection. This systematic review and meta-analysis compared the potential to mitigate SSCs, SSIs, and dehiscence in breast surgery between two different ciNPT systems: ciNPT-MLA (PICO Single Use Negative Pressure Wound Therapy System, Smith + Nephew) and ciNPT-F (3M Prevena Incision Management System, Solventum Corporation) against standard of care (SOC) dressings.

This work conformed to the statement and reporting checklist of the preferred reporting items for systematic reviews and meta-analysis (PRISMA).¹⁴ The systematic review and meta-analysis were conducted to compare the potential to help mitigate SSCs in breast surgery between 2 ciNPT systems (Prevena Incision Management System, Solventum Corporation; and PICO Single Use Negative Pressure Wound Therapy System, Smith + Nephew) and SOC dressings. Outcomes assessed included SSC (defined as all surgical site complications including SSI, dehiscence, seroma, hematoma, necrosis, etc), SSI, and dehiscence in breast surgery.

Literature Search

A systematic literature search was conducted using MEDLINE, EMBASE, PUBMED, QUOSA, and a search of references. Literature published between January 1, 2005 and June 31, 2022 were examined. Two separate literature searches were conducted, one for ciNPT-F and one for ciNPT-MLA. The ciNPT-F literature search utilized the following search terms (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). The ciNPT-MLA literature search utilized the following search terms Negative pressure wound therapy, NPWT, Negative pressure, Negative therapy, Topical negative pressure. The ciNPT-MLA literature and search utilized the following search terms (Pico, ciNPT, Prophylactic npwt, Prophylactic negative pressure, Preventative npwt, Preventative negative pressure, Incisional npwt, Incisional negative pressure, Surgical, Surgery).

Meta-Analysis Inclusion/Exclusion Criteria

Study inclusion criteria consisted of English language peer-reviewed manuscripts or published abstracts reporting on SSC, SSI, or dehiscence in patients (age >18 years) with plastic surgery. The search results for ciNPT-MLA identified articles reporting on closed incisions following breast surgery. As such, the study inclusion was limited to articles reporting on closed surgical incisions after breast surgery with ciNPT-MLA or ciNPT-F applied postoperatively on a closed incision in comparison with SOC or any non-negative pressure wound therapy dressing. Study types considered for analysis were randomized controlled trials (RCTs), prospective comparative studies, prospective cohort compared with a historical cohort, and retrospective comparative studies.

Study exclusion criteria consisted of unpublished, non-peer-reviewed studies in a non-English language, reporting on patients with open surgical incisions, nonsurgical wounds, or patient population <18 years of age. Studies with non-ciNPT negative pressure therapy, no mention of ciNPT product type, or mixed use of negative pressure wound therapy devices in 1 treatment arm were also excluded. Studies that did not report on SSC, SSI, or dehiscence outcomes were excluded. Preclinical studies, case studies, case series, reviews, or meta-analyses were also excluded from the analysis.

Study Selection

Titles and abstracts of publications identified in the literature search were logged. After removing all duplicates, titles and abstracts of each paper were read to assess eligibility. The remaining publications were read to ensure all inclusion criteria and no exclusion criteria were met. References from identified publications were also reviewed for relevant studies that fit the inclusion criteria.

Data Collection

One reviewer completed data extraction from all eligible studies and a second reviewer validated the findings. Disagreements were resolved by discussion between the 2 reviewers or by the addition of a third reviewer. The following data were extracted from each study: funding source, bias assessments, study design, publication status, study date range, number of study sites, study location, surgical procedures, high-risk enrollment criteria (if applicable), study objectives, control type, number of treatment days, follow-up period, number of patients/incisions, number of patients/incisions analyzed, and definition of SSC. Endpoints evaluated were SSC, SSI, and dehiscence, which were reported as composite study end points.

Risk of Bias

Risk of study bias was assessed for RCTs only. RCTs were assessed for selection bias, performance bias, attrition bias, and reporting bias using the Cochrane collaboration tool for assessing risk of bias.¹⁵

Statistical Analysis

Meta-analyses were performed using risk ratios and random effects models. Weighted risk ratios and 95% confidence intervals (CI) were calculated to pool study and control groups in each publication for analysis. To assess data heterogeneity, I² was utilized. Treatment effects were combined, and a random effects model was used for each analysis performed, regardless of the heterogeneity assessment. When the analyses of ciNPT-F and ciNPT-MLA had similar statistical outcomes, prediction intervals were used to assess the levels of difference in the true effects. If it was assumed that the true effects were normally distributed, Tau-squared was an estimate of the between-study variance in random effects models and used to determine the prediction interval. If tau-squared was estimated as 0, the assumption was that all studies shared a common effect size, and no separate prediction interval was reported. All analyses were performed using Comprehensive Meta-Analysis Version 3 (Borenstein M, Hedges L, Higgins J, & Rothstein H. Biostat).

Health Economic Model

A health economic model was created to examine the potential impact of ciNPT usage on cost of care. This model used SSC rates to encompass all wound complication types. The baseline SSC rate was calculated as the number of SSC divided by total number of incisions from the SOC group across all included studies multiplied by 100. The SSC reduction rates for ciNPT-F from the meta-analysis were utilized to determine the relative SSC risk reduction rate. SSC rate with ciNPT-F was calculated by multiplying the base SSC rate with the reduction rate of SSC from the meta-analysis and subtracting the results from the base SSC rate. The average cost per SSC was obtained from the combined average cost of SSCs using a retrospective analysis using data from the Premier Healthcare Database of deidentified clinical and cost data.¹⁶ Total cost for SSC per patient was calculated as the SSC rate with ciNPT multiplied by the average mean cost of SSCs. ciNPT device cost was obtained from the list price of the device. Total cost per patient was calculated as total cost for SSC per patient plus the device cost. Mean per patient cost saving was calculated as the total cost per patient for SOC minus total cost per patient for ciNPT-F.

Literature Search

Two separate literature searches were performed, one for each of the ciNPT devices being assessed. For ciNPT-MLA, a total of 1093 articles were identified (Figure 1). After removal of duplicates and excluded studies, a total of 8 articles (5 RCTs, 1 prospective study, and 2 retrospective comparative studies) were included in the meta-analysis (Tables 1 and 2).^17-24 The included studies assessed patients receiving breast reconstruction, lumpectomy, mammoplasty, and mastectomy. One study did not report the type of breast surgery performed.²⁰ A total of 569 patients received ciNPT-MLA and 610 patients received SOC.

Figure 1. PRISMA flow diagram of ciNPT-MLA article selection for including into meta-analysis. ciNPT, closed incision negative pressure therapy; ciNPT-MLA, closed incision negative pressure therapy with multilayer absorbent dressing; HE, health economics.

The literature search for ciNPT-F identified 972 studies (Figure 2). After duplicates and excluded studies were removed, 4 articles (1 prospective study and 3 retrospective comparative studies) were included in the meta-analysis (Tables 3 and 4).^8,25-27 These included studies assessed patients receiving breast reconstruction, breast revision surgery, mammoplasty, mastectomy, and oncoplastic surgery. A total of 423 patients received ciNPT-F and 499 patients received SOC.

Figure 2. PRISMA flow diagram of ciNPT-F article selection for including in meta-analysis. ciNPT, closed incision negative pressure therapy; ciNPT-F, closed incision negative pressure therapy with foam dressing.

Bias Assessment

To determine the impact and quality of the RCTs included in the analysis, a bias assessment was completed. For selection bias, 1 study was found to have unclear risk for both randomization methods and allocation masking.¹⁷ A single study was found to have unclear risk for allocation masking.¹⁸ The remaining 3 RCTs were found to be low risk for both randomization method and allocation masking.^22-24 Four RCTs had high risk of performance bias due to the difficulty in blinding the assessment of outcomes with ciNPT device use.^18,22-24 The remaining RCT performance bias was unclear.¹⁷ Attrition bias was unclear in 2 RCTs^17,22 and low risk in 3 RCTs.^18,23,24 Reporting bias was high risk in the Timmermans et al study;²³ the remaining RCTs had a low risk.^17,18,22,24

Postsurgical Risks in Breast Surgery

Surgical site complications. Five ciNPT-MLA studies were compared with SOC and resulted in a weighted risk ratio of 0.808 (95% CI: 0.470, 1.391), P = .442, indicating no difference in rates of SSCs between ciNPT-MLA and SOC (Figure 3A, Table 5). Four ciNPT-F studies were assessed for effect on SSC rates resulting in a weighted risk ratio of 0.498 (95% CI: 0.271, 0.917), P = .025, indicating a significant relative risk reduction of 50.2% with ciNPT-F compared with SOC (Figure 3B, Table 5).

Figure 3. Forest plot of surgical site complications. (A) ciNPT-MLA forest plot; (B) ciNPT-F forest plot; risk ratio, lower limit, upper limit, surgical site complication (SSC) total, and relative weight are shown.

Surgical site infection. Five ciNPT-MLA studies were assessed for SSI resulting in a weighted risk ratio of 0.633 (95% CI: 0.192, 2.09) and no difference in SSI rates between ciNPT-MLA and SOC (P = .453, Table 5, Figure 4A). Tau-squared, the variance of true effect sizes, was 0.812 in log units, resulting in a prediction interval of 0.020 to 20.178. The true effect size would fall in this interval for 95% of all comparable populations.

Three ciNPT-F studies were assessed for SSI, resulting in a weighted risk ratio of 0.478 (95% CI: 0.226, 1.010) indicating a trend toward reduced SSI rates with ciNPT-F use (P = .053, Table 5, Figure 4B). For this meta-analysis, the tau-squared was estimated as 0; thus, there was no dispersion of true effects.

Figure 4. Forest plot of surgical site infections. (A) ciNPT-MLA forest plot; (B) ciNPT-F forest plot; risk ratio, lower limit, upper limit, surgical site complication (SSC) total, relative weight, and prediction interval, if applicable, are shown.

In comparing the range of the true effect for reducing SSI risks compared with SOC, there was a higher chance to reduce the SSI risk (95% CI: 0.226, 1.010) with ciNPT-F compared with ciNPT-MLA with a prediction interval of 0.020 to 20.178.

Dehiscence. Four ciNPT-MLA studies were examined resulting in a weighted risk ratio of 0.499 (95% CI: 0.303, 0.822) and significant reduction in dehiscence rates with ciNPT-MLA compared to SOC (P = .006; Table 5, Figure 5A). With a tau-squared of 0.043, the prediction interval was 0.121 to 2.051. The true effect size in 95% of all comparable populations falls in this interval.

Figure 5. Forest plot of dehiscence. (A) ciNPT-MLA forest plot; (B) ciNPT-F forest plot; risk ratio, lower limit, upper limit, surgical site complication (SSC) total, relative weight, and prediction interval, if applicable, are shown.

Three ciNPT-F studies were examined yielding a weighted risk ratio of 0.349 (95% CI: 0.168, 0.725) and a significant reduction in dehiscence rates with ciNPT-F use compared with SOC (P = .005; Table 5, Figure 5B). For this meta-analysis, tau-squared was estimated as 0; as such, all studies were assumed to share a common effect size and there was no dispersion of true effects.

The prediction intervals were compared to estimate the range of the true effect for reducing the risk of dehiscence compared with SOC by ciNPT-MLA or ciNPT-F. A stronger effect to reduce dehiscence risk (prediction interval of 0.168 to 1.010) was observed for ciNPT-F compared with ciNPT-MLA (prediction interval of 0.121 to 2.051).

Health Economic Model

A health economic model was created to estimate the potential per-patient cost savings with ciNPT-F use following breast surgery (Table 6). The overall baseline SSC rate of 24.0% across all 4 ciNPT-F studies identified was used. The relative SSC reduction rate for ciNPT-F was taken from the SSC meta-analysis (50.2%; Table 5). The cost per SSC was estimated at $9526.¹⁶ The total cost per patient was estimated to be $1793 for ciNPT-F, resulting in a potential savings of $493 per patient with ciNPT-F use compared with SOC.

Development of SSCs leads to increases in required patient care, such as infection management, reoperation, and prolonged wound care, resulting in increased length of hospital stay and health care costs. ciNPT has been previously reported to aid in the reduction of SSCs in other surgical specialties.^10-13 However, device-specific detailed examination of ciNPT use in breast surgery has been limited. A systematic literature review was performed and identified 12 articles reporting on the use of 2 different ciNPT systems compared with SOC in breast surgery.^{8,17-23,25-27} Meta-analysis reported ciNPT-MLA use was associated with reduced rates of dehiscence in breast surgery compared with SOC, while ciNPT-F use was associated with reduced SSC and dehiscence rates with a trend toward reducing SSI rates compared with SOC.

Three previously published meta-analyses reported pooled results from multiple ciNPT systems together for the ciNPT group in patients undergoing breast surgery.^28-30 Liew et al examined ciNPT use compared with SOC in breast surgery and reported reduced rates of SSC with ciNPT use.²⁸ Cagney et al also evaluated the use of ciNPT in breast surgery compared with SOC and reported reduced rates of SSC, SSI, and dehiscence with the use of ciNPT.³⁰ Similarly, Song and colleagues performed a meta-analysis comparing ciNPT with SOC in breast cancer surgery and found reduced SSC, SSI, and dehiscence rates in the ciNPT group.²⁹ Only 1 meta-analysis examined the individual ciNPT systems compared with SOC.³¹ Singh et al associated reduced rates of SSI with ciNPT-F compared with SOC and no differences in SSI rates between ciNPT-MLA use and SOC.³¹ However, Singh et al reported SSI outcomes across all surgical procedures, which did not include any results on breast-specific surgeries. Nonetheless, these findings are similar to this breast surgery-specific meta-analysis, indicating the potential for reduced SSC, SSI, and dehiscence rates following ciNPT use after breast surgery.

The risk ratio per individual studies for this meta-analysis ranged from 0.231 to 2.000 for ciNPT-MLA, resulting in a nonsignificant SSC overall risk ratio even though small single-center studies indicated a positive outcome. When all studies including multiple larger studies are assessed, overall SSC reduction was found to be nonsignificant. For ciNPT-F, the individual study risk ratios ranged from 0.291 to 0.883, resulting in an overall significant reduction of SSC by 50.2% with ciNPT-F use. While other ciNPT devices have been examined for effect on SSC, these are small, single-center studies.³² Reductions in SSCs with other ciNPT devices indicate the potential of these ciNPT devices to help reduce SSCs after breast surgery. However, it is not feasible to compare the performance of other ciNPT devices with ciNPT-F or ciNPT-MLA as further investigations are needed with increased patient population sizes to provide a more complete picture of ciNPT device use.

The meta-analysis results indicated that ciNPT-F showed a greater reduction in SSC rates compared with SOC. However, this option has a higher cost per application, raising the question if the incremental cost can be justified by the reduced cost for managing complications. To build a health economic model, the SSC rate of SOC treatment group across all ciNPT-F studies was estimated at 24.0%. This was a more conservative approach as the baseline SSC rate was 48.7% for ciNPT-MLA studies. In our health economic model, ciNPT-F was found to have a potential per-patient savings of $493 when compared with SOC. A model comparing the 2 ciNPT systems based on SSC reductions was not created as the SSC reduction outcome for ciNPT-MLA was nonsignificant. We conducted sensitivity analyses for the comparison of ciNPT-MLA with SOC for SSCs by evaluating the effect of removing 1 study at a time. This would help determine if the nonsignificant result was driven by a single study. The analyses did not alter the reported results.

To date, other health economic assessments have also identified potential cost savings with ciNPT use.^19,33-35 Two studies have published cost savings with ciNPT-MLA versus SOC in breast reconstructions.^19,35 Irwin and colleagues assessed resource use in 7 patients following prepectoral breast reconstruction failure.¹⁹ As reconstruction failures occurred only in the SOC group, a cost savings of ₤426 was associated with ciNPT-MLA use.¹⁹ A separate study by Murphy et al reported an expected cost savings of ₤1706 per patient with ciNPT-MLA use in the UK compared to SOC.³⁵ The Murphy study utilized a hypothetical model with complication rates obtained from the Irwin et al study for ciNPT-MLA use and baseline complication rates from a separate study assessing mastectomy and immediate implant-based breast reconstruction in 2108 patients across the UK.³⁵ Similarly, Nherera et al reported that ciNPT-MLA use was cost effective compared with SOC in both the UK and US.³³ However, this study used SSI rates for all surgery types. Gabriel et al developed a hypothetical cost model assessing ciNPT-F use after breast reconstruction compared with SOC.³⁴ In this model, ciNPT-F was estimated to have a $218 per-patient cost savings over SOC.³⁴ This lower cost savings was based on a single study that had lower baseline SSC rates and slightly lower ciNPT-F reduction rates when compared with the baseline and SSC rates of our cost model. No other cost analysis for ciNPT-F use in breast surgery has been published. However, a meta-analysis examining ciNPT-F use in plastic surgery reported a $904 cost savings in patient care over SOC.³⁶ Similarly, estimated cost savings have been reported with ciNPT-F use compared with SOC in orthopedic, abdominal, and vascular surgeries, and in patients with high risk of developing complications in cardiothoracic surgery.^37-40 Our health economic model supports the previously published literature that ciNPT-F may provide a cost benefit compared with SOC.

Limitations for this meta-analysis include analysis of a mix of observational studies and RCTs, differences in data reporting, and assessment of only 2 ciNPT systems. Currently, limited high-level published evidence exists comparing use of ciNPT with SOC in breast surgery. This lack of high-level clinical evidence led to the inclusion of studies with lower levels of evidence, such as nonrandomized prospective studies and retrospective comparison studies, along with a handful of RCTs in the meta-analysis. Differences in study data reporting also exist as some included studies did not report on specific outcomes, leading to a smaller number of studies assessed for the individual clinical outcome. Other factors such as variations in SSC definitions and patient populations may partially explain the differences in SSC, SSI, and dehiscence reduction outcomes between ciNPT-MLA and SOC and ciNPT-F and SOC. The study quality and limited evidence reporting may have reduced the strength of the analyses, though all meta-analyses were conducted using the random-effects model to help mitigate any potential data heterogeneity. While significant differences were observed between ciNPT groups and SOC, the small number of studies assessed in the SSI and dehiscence sub-analyses do limit the strength of the results. Additionally, the small number of studies included in the meta-analysis hindered the ability to perform breast surgery-specific subgroup analyses. This limited the analyses to breast surgery overall despite known variations in patient populations, surgical procedure, and use of chemotherapy or radiation therapy across thevarious breast surgery procedures. Furthermore, only 2 ciNPT systems were evaluated. Individual ciNPT systems may exert different impacts on patient outcomes. System differences such as type of dressing utilized and amount of negative pressure provided may result in differences of clinical outcomes. As only 2 ciNPT systems were considered, the results of this meta-analysis may not be applicable to other available systems. Similarly, this meta-analysis utilized an indirect comparison between 2 ciNPT systems and SOC, as no evidence exists directly comparing these two systems, also limiting the strength of the analysis.

Surgical procedures carry a risk of SSC development that is affected by patient comorbidities and type of surgical procedures. In breast surgery, SSC rates between 2.25% and 53% have been reported.^1-5 While ciNPT systems have been associated with SSC rate reductions in other surgical procedures, limited evidence exists for ciNPT use in breast surgery. This systematic review and meta-analysis compared use of ciNPT from 2 systems (ciNPT-MLA and ciNPT-F) against SOC to assess SSC, SSI, and dehiscence outcomes in breast surgery. Compared with SOC, ciNPT-MLA significantly reduced rates of dehiscence. ciNPT-F use resulted in significantly reduced rates of SSC and dehiscence with a trend toward reducing SSI. Caution should be used when interpreting these outcomes as only 2 ciNPT devices were assessed and results may differ for other ciNPT systems. To identify ciNPT system-specific effects on clinical and health economic outcomes, a ciNPT system-specific assessment is recommended as there is high potential for differences in clinical effectiveness between all ciNPT systems.