Closed-incision negative pressure therapy as a strategy to reduce sternal wound infection in identified high-risk patients: a multicentre propensity matched study

Abstract OBJECTIVES The premise of this retrospective study was to evaluate the intraoperative use of closed-incision negative pressure therapy (ciNPT) to help reduce the incidence of postoperative sternal wound infections in multimorbid patients with an elevated risk of developing a sternal wound infection post cardiac surgery versus a cohort that received standard-of-care dressings. METHODS Data for all adult patients were collected from each cardiothoracic surgery unit across 3 hospitals in the United Kingdom. High-risk patients had 2 or more recognized risk factors. Fisher’s exact test (two-tailed) and unpaired t-test were used to help analyse categorical and continuous data. Propensity matching was performed to compare the 2 groups. RESULTS A total of 5,288 patients who had cardiac surgery were included. Propensity matching led to 766 matched cases. There were significantly fewer sternal wound infections in the ciNPT group [43 (5.6%) vs 119 (15.5%) cases; P = 0.0001], as well as fewer deep sternal wound infections [14 (1.8%) vs 31 (4.0%) cases; P = 0.0149] and superficial sternal wound infections [29 (3.8%) vs 88 (11.4%) cases; P = 0.0001]. A higher mean length of stay in the ciNPT group was statistically significant (11.23 ± 13 vs 9.66 ± 10 days; P = 0.0083) as was a significantly higher mean logistic European System for Cardiac Operative Risk Evaluation (EuroSCORE) (11.143 ± 13 vs 8.094 ± 11; P = 0.0001). A statistically significant higher readmission to the intensive care unit due to sternal wound infection was noted for the controls [16 (2.08%) vs 3 (0.39%) readmissions; P = 0.0042]. CONCLUSIONS The ciNPT appears to be an effective intervention to help reduce the incidence of sternal wound infection in high-risk individuals undergoing cardiac surgery.


INTRODUCTION
The development of a sternal wound infection (SWI) represents a major complication following cardiac surgery [1].Aside from potential life-threatening consequences, SWI is associated with a prolonged length of stay (LOS) in the hospital and, according to the European Health Service, may add an estimated 19 billion Euros of additional health care costs [2].It has been previously demonstrated that avoiding virulent bacterial contamination that can lead to mediastinitis and the degradation of sutured skin, especially in high-risk patients, is integral to helping prevent postoperative complications associated with SWI [3,4].This postoperative complication is most prevalent in patients with a medical history of obesity, diabetes, chronic obstructive pulmonary disease (COPD) and advanced age [5].
According to Friberg et al., reported cases of SWI immediately postoperatively are approximately 1%-2%; however, the rate increases to 5%-8% and a further 6%-10% after 2 and 3 months, respectively [6].The National Institute of Health and Care Excellence (NICE) has noted that a lack of a postoperative follow-up at 1 month may also be characterized as a risk factor and has recommended closed-incision negative pressure therapy (ciNPT) use as an intervention for cardiothoracic surgery patients determined to be at high risk of developing an SWI [7].The ciNPT is applied as an incision-management dressing that can administer -125 mmHg for up to 7 days.The mechanism of action of ciNPT resides in its ability to help hold incisional edges together and to help bolster and stabilize closed incisions [8].The ciNPT is also enlisted to remove fluid and infectious materials and to help present a barrier against external contaminants [9].
Preventing SWI is important in reducing patients' mortality and morbidity and the cost of care to putatively provide substantial savings in any surgical service by reducing the length of hospital stay [10].Our goal was to evaluate the impact of administering ciNPT immediately postoperatively to cardiothoracic patients identified as high-risk as part of an incision management strategy at 3 cardiac surgery units in the United Kingdom (UK) to help reduce the incidence of SWI.

Ethics statement
The study was approved by the Hull University Teaching Hospitals NHS Trust Clinical Audit and Governance Board as a multicentre clinical audit (approval number NA.2021.091)granted on 30November 2021.It was also approved by the clinical audit departments of all participating hospitals.Patient informed consent was not obtained due to the anonymity of all data.

Study design
This is a multicentre, retrospective study to identify any putative clinical benefit of ciNPT use as a perioperative strategy in helping to reduce SWI incidence in patients considered to be at an elevated risk of developing said surgical site complications.With the approval of the institutional review boards of 3 health-care institutions, de-identified clinical data were collected from the patient databases of 3 hospitals and provided to the lead investigator for analysis by the respective cardiothoracic surgery units that participated in the study.

Setting
The 3 units that utilized ciNPT (3M Prevena Incision Management System, 3M Company, St Paul, MN, USA) postoperatively for incision management were invited to participate in the multicentre audit.Upon agreement, completed audit applications were submitted by each cardiothoracic surgery unit to their respective institutional review board.Upon approval, independent data gathering was performed by each unit using the audit criteria as a guide (Table 1) as well as an adapted scheme for characterizing wound classification, depth and description (Table 2).De-identified data were sent to the lead investigator for consolidation and analysis.Consent was not required because no cardiac surgery patients were contacted for information.

Study patients/patient selection
The study included all adult patients who underwent cardiac surgery via median sternotomy between January 2013 and December 2021 from either of 3 cardiothoracic surgery centres in the UK: Castle Hill-Hull University Teaching Hospital, Oxford University Hospital, and The Royal Wolverhampton NHS Trust.Patients were identified as having received ciNPT (ciNPT group, n ¼ 1060) or a standard dressing (control group, n ¼ 4228) (Table 3).Patients were propensity matched using the following risk factors identified from our previous study (Ariyaratnam et al.): advanced age (above 80 years), diabetes mellitus, COPD and obesity [body mass index (BMI) > 30 kg/m 2 ) as independent variables (Table 4 and Fig. 1).

Variables of interest
The variables of interest for this investigation were cardiothoracic surgery type and patients with risk factors for developing an SWI who underwent cardiac surgery.The following cardiothoracic surgical procedures were performed via a median sternotomy: isolated coronary artery bypass graft (CABG), isolated heart valve repair/replacement, CABG with other surgery, heart valve repair/replacement with other surgery and other cardiac procedures.Patients with 2 or more of the following risk factors-obesity, advanced age, COPD and diabetes mellitus-were noted.Other variables such as sex, LOS and types of wound infections were also documented.To quantify the preoperative mortality risk profile, the logistic European System for Cardiac Operative Risk Evaluation (EuroSCORE) was also calculated for patients within both cohorts (Table 3).

Data sources and study size
Patients with BMI > 30 kg/m 2 , 80 years old and above, a Spirometry test for COPD and blood glucose level for diabetes were noted, and those who had 2 or more of these conditions were considered high risk.A propensity matching was performed using only the stated conditions as covariants, which led to the size of the sample.

Bias
A 1:1 matching was performed to obtain a treatment group and a control group with similar covariates.Because the study had both continuous and nominal data, a lower caliper was used to lessen the chance of bias.

Statistical methods/analysis
The propensity matching was performed using the XLSTAT 2022.4.1.1366software (Lumivero, Denver, CO, USA) [11].This study primarily applied propensity score matching to 2 treatment cohorts rather than to various groupings, given the existence of some crucial unresolved issues (e.g.assessing balance among baseline variables, matching distance and caliper width choice) [12].
In this study, BMI, age, COPD and diabetes were used as covariates.Patients were matched 1:1 with an optimal algorithm Euclidean distance using a 0.1 caliper, with a confidence interval (CI) of 95% and 0.001 tolerance.Propensity score matching resulted in 766 patients matched in both the control and the ciNPT groups.
According to the Monte Carlo simulations, using a caliper width up to a maximum of 0.2 would putatively disregard approximately 98% of the bias in the crude estimator, thereby generating CIs that would be on the order of the correct coverage rates when at least 1 of the covariates is continuous data [12].
Whereas categorical data were conveyed as percentages, all continuous data (e.g.patient age, logistic EuroSCORE, LOS and extended LOS) were stated as mean ± standard deviation (SD).An unpaired, two-tailed t-test was used to calculate the P-value for all continuous data.The Fisher exact test of independence (for a two-tailed test) was used to calculate the P-value for all categorical data such as BMI, COPD, diabetes, mortality, sex or

Participants and descriptive data
The total number of initial study participants was 5,288; the ciNPT group had n ¼ 1060 and the control group had n ¼ 4228.
All adult patients with complete data shown in Table 1 were included Patient demographics and risk factors between the control and the ciNPT groups were compared.Both groups had a similar average age, incidence of comorbidities (BMI, COPD and diabetes) and type of surgery performed (Table 3).Using 4 identified risk factors (age, BMI, COPD and diabetes), the groups were propensity matched, which resulted in a total of 1,532 patients (control group, n ¼ 766; ciNPT group, n ¼ 766) included in the study (Table 4 and Fig. 1).Both groups were also examined using the following variables: cardiac surgery type, logistic EuroSCORE, SWI cases, SWI classification, length of postoperative stay from day of surgery to discharge, ICU re-admission due to SWI and subsequent hospital stay due to SWI (Tables 3, 5, and 6).A total of n ¼ 109 (ciNPT) and n ¼ 779 (control) patients were excluded in the matching due to incomplete data.
Compared to the ciNPT group, the control group contained a higher percentage of CABG þ other (27.8% vs 8.9%, P ¼ 0.0001) and valve þ other cases (14.9% vs 5.8%, P ¼ 0.0001) (Table 3).However, the ciNPT group had a higher percentage of cases for cardiac operations such as isolated CABG (48% vs 37.5%; P ¼ 0.0001), valve repair/replacement alone (18.11 vs 16.3%; P ¼ 0.0012) or other cardiac operations (8.2% vs 6.10%; 3).For the control and ciNPT cohorts, the predicted operative mortality of patients was calculated using the logistic EuroSCORE model (Table 3).The mean logistic EuroSCORE captured statistical significance between cohorts (control group ¼ 8.094 ± 11 vs ciNPT group ¼ 11.143 ± 13; P ¼ 0.0001) and helped to prevent the underestimation of individual patient risk to offer a more exact forecast where preoperative risk factors coexisted (Table 3).Contrariwise, a total of 43 (5.6%) patients in the ciNPT group developed an SWI that was managed with a regimen of antibiotics.Within these patients, there were 29 SWIs characterized as superficial and 14 that were determined to be deep SWIs.Patients with SWI within the ciNPT group also had their wounds managed via debridement (n ¼ 7), sternal rewiring (n ¼ 4), and negative pressure wound therapy (n ¼ 9); however, no patients were referred to plastic surgery for surgical revision.A lower incidence of SWI was noted among patients within the ciNPT group [n ¼ 43 patients (5.6%)] relative to the control group [n ¼ 119 patients (15.5%)], which was statistically significant (P ¼ 0.0001).Additionally, patients within the ciNPT group also demonstrated a statistically significant lower incidence of deep SWI compared to the control group [n ¼ 14 patients (1.8%) vs n ¼ 31 (4.0%);P ¼ 0.0149].

Postoperative distribution of sternal wound infection between cohorts.
Length of inpatient hospital stay.Given its higher mean logistic EuroSCORE, the ciNPT group demonstrated a higher mortality risk profile evidenced by a longer mean LOS (11.2 ± 13.12 vs 9.6 ± 10.31 days; P ¼ 0.0083), which was statistically significant (Table 6).Conversely, the control group demonstrated a higher rate of ICU readmission resultant of SWI (16 cases vs 3 cases, P ¼ 0.0042), which was statistically significant.For patients with SWI who experienced a subsequent hospital stay due to wound infection, the total number of days for the control group was sevenfold higher than the ciNPT group (256 days vs 36 days) (Table 6).However, the mean hospital stay due to SWI was not statistically significant (16 ± 14 days vs 12 ± 5.35 days; P ¼ 0.6358) (Table 6).

Cost comparison of surgical incision management modalities in patients with closed-incision negative pressure therapy.
A simple cost analysis comparing the cost of treating control group patients who developed SWI and required readmission to the ICU against the cost of treating patients in the same category in the ciNPT group was performed.Sixteen patients in the control group, who were readmitted to the ICU due to SWI, had a total LOS of 256 days beyond the routine 5-day post-surgical LOS (Table 6).According to the data, the average cost of inpatient stay per day in the UK  ).However, the 3 patients who received ciNPT, who were readmitted to the ICU due to SWI, had a total LOS of 36 days beyond the routine 5-day postsurgical LOS (Table 7).The extra cost for the ciNPT group was £12,924 (36 days x £356).According to NICE, the cost associated with ciNPT use is £299 per unit, whereas the standard dressing costs £3.87 per unit [7].The costs for managing the SWIs of the 16 control group patients and the 3 ciNPT group patients were £61.92 (£3.87 x 16 standard dressings) and £897 (£299 x 3 ciNPT units), respectively.Taken together, the total cost of an extended LOS post ICU readmission with incision management was £91,965.92(£91,904 þ £61.92) for the control group vs £13,821 (£12,924 þ £897) for the ciNPT group.The £78,144.92 cost difference between cohorts would reflect a savings when ciNPT was utilized as a strategy to help aid in reducing the incidence of SWI (Table 7).

DISCUSSION
Evidence from various studies suggests that the use of ciNPT in adult cardiac surgery is a valid adjunctive therapy where the maximum benefit seems to be in helping to mitigate the   [18].In this study, the application of ciNPT to high-risk patients helped significantly to reduce the cases of SWIs (superficial and deep) and hospital LOS.A preliminary cost-of-care analysis suggests that ciNPT use may provide some putative cost saving in cardiac surgery relative to the standard-of-care dressing.Diabetes, COPD, advanced age and obesity represented the most common comorbidities in the study cohorts; however, patient demographics and comorbidities were similar between the control and the ciNPT cohorts because both groups were propensity matched.The current study also took into consideration a patient's critical preoperative state by calculating the logistic EuroSCORE to assign a preoperative mortality risk grade.We reported statistical significance between the mean logistic EuroSCORE of the control cohort versus the ciNPT cohort.Additionally, certain surgical approaches may protract wound healing, which may then facilitate the occurrence of SSIs.For example, the use of internal thoracic arteries or mammary arteries in CABG operation combined with sternotomy has been reported as a potential operation-related risk factor for the development of SWIs [19,20].Because combined procedures (CABG þ other, valve repair/replacement þ other and other cardiac operation) were the most common cardiothoracic surgical interventions in our study, this may also be indicative of an elevated surgical risk factor for the development of SWI within our study population.
From a clinical research standpoint, SWI treatment post cardiac surgery is no longer sufficient because greater emphasis presently resides with SWI prevention as well as the health economics associated with managing this hospital-acquired infection [21].Using a risk-stratification model, Fowler's Score, Atkins et al. [10] cited diabetes and obesity as SWI risk factors.Colli and Camara [22] also used the same model, citing that ciNPT use was effective in preventing SWI complications up to the 30-day follow-up appointment.Grauhan et al. [23] assessed the effectiveness of ciNPT use on a larger group (n ¼ 150) of obese patients.They reported that ciNPT use helped to significantly reduce the incidence of SWI in the obese population at 90-day follow-up appointments.A subsequent study by Grauhan et al. [24] assessed the effect of ciNPT utilization in the general population of post-sternotomy patients and confirmed their previous findings after 30 days follow-up.Our study appears to align with current evidence, because it demonstrated that high-risk patients receiving ciNPT exhibited a statistically significant lower incidence of SWI compared to the control group (5.6% vs 15.5%; P ¼ 0.0001).Furthermore, 31 (4.0%) and 14 (1.8%) cases of deep SWI were noted in the control group and ciNPT group, respectively.The difference in these rates was also statistically significant (P ¼ 0.0149) (Table 5).Our study reported that more control group patients required debridement (n ¼ 19), sternal rewiring (n ¼ 13) and intervention via plastic surgery (n ¼ 7), whereas fewer ciNPT group patients had debridement (n ¼ 7) or had sternal rewiring (n ¼ 4) performed, and no plastic surgery referrals.
A mini review on complications of ciNPT by Sartipy et al. [25] and Li and Yu [26] raised potential safety concerns over its use.However, Dohmen et al. [27] produced cardiothoracic specific consensus recommendations on the use of ciNPT.Eleven cardiac and orthopaedic randomized controlled trials, retrospective studies and case studies were analysed.Taking into account previous studies, the authors were able to propose clear preoperative risk factors that predisposed patients to major infection after cardiac surgery.Both Dohmen et al. [27] and Biancari et al.
[2] agreed that not all patients require ciNPT, but that the correct identification of those high-risk individuals was key in the provision of a cost-effective intervention.In our study, the mean LOS was higher in the ciNPT group (11.23 ± 13 days) compared to the control group (9.66 ± 10 days) (P ¼ 0.0083), which is statistically significant.This result might be due to the greater representation of acute patients enrolled in the ciNPT group as evidenced by an elevated logistic EuroSCORE.On the other hand, the control group had significantly higher rates of readmission to the ICU due to SWI (3 vs 16 cases, P ¼ 0.0042).Mean hospital stays post ICU readmission due to SWI were not statistically significant (16 ± 14 days vs 12 ± 5.35 days; P ¼ 0.6358) despite the control group noting a sevenfold higher total number of days (256 days vs 36 days) (Table 6).
In the systematic review and meta-analysis by Biancari et al.
[2], pooled rates were calculated from 10 included investigations and showed that 22.2 patients should be treated with ciNPTbased systems to help avoid a solitary SWI event.Although this result may seem to be a large number of patients, the putative benefit of ciNPT should be viewed in the context of an increased burden on resource utilization due to prolonged treatment of the complication.This is in addition to the significant clinical implication for the patient's recovery, which could increase the risk of mortality.Multiple reviews on the financial burden of treating SWI have been published [3,28].Recent data from the European Health Service noted an extra cost of 19 billion euros to treat SWIs [29].Our simple cost analysis, which was based on the most current figures published by the King's Fund [28], an organization that monitors key facts and figures about the UK National Health Service, determined that the average cost of hospitalization per day is £359.Given the extended total LOS for the control group (256 days) and the average cost of a simple surgical procedure (£2,817), the total expenditure for patients with SWI who require plastic surgery intervention is £111,623 (£91,904 þ £19,719) (Table 7).Conversely, the cost associated with ciNPT use according to NICE and the total LOS of 36 days would amount to £13,821 (£897 þ £12,924) (Table 7).
Admittedly, ciNPT may not be cost effective if used routinely when considering the cost of each device.Hawkins et al. [30] suggested that this treatment method may be cost effective only when used on patients at risk of developing deep SWI (1.3%), which could mean that fewer than half of identified high-risk cardiac patients may benefit clinically from the potential savings.However, there is no stratified scoring system available at present to measure the probability risk on the degree of SWI.

LIMITATIONS
Our present study is limited by several considerations.First, despite the duration of the market availability of ciNPT, large, adequately powered clinical studies in the cardiac surgery sphere investigating ciNPT, which may help inform the surgeon or health care institution on the wider acceptance of this active wound care modality, are still wanting.Additionally, the high price point of ciNPT as an incision management approach relative to standard-of-care dressings, is still considered to be a cost prohibitive intervention in the absence of sufficient health economic evidence to justify its routine intraoperative administration during cardiac surgery to help reduce the incidence of SWI.Another possible limitation is that consideration was not afforded to microbiological characterization, postoperative low output syndrome, sternal stability and other perioperative concerns (cardiothoracic surgery type, procedure duration, urgency of the procedure, number of vessels reclaimed for total arterial revascularization) beyond multiple comorbid conditions that might have an influence upon the variance in SWI incidence between the cohorts.Our study is also limited by the rudimentary cost-comparison presented, where we cite a cost difference of £78,144.92 between members of the cohorts who developed SWI.Development of improved economic models designed to sensitively capture the quality-adjusted life years, risk stratification models, as well as studies incorporating health economic analysis, may assist in better defining cost offsets and economic end points to help mitigate the uncertainty of the clinical efficacy and cost-effectiveness of ciNPT in managing poststernotomy incisions.

CONCLUSION
The evidence presented in this multicentre retrospective study demonstrates that intraoperative ciNPT use should be viewed as an effective adjunctive intervention to help reduce SWI incidence in cardiac surgery patients with multiple comorbidities considered to have a high-risk of SWI development.Our study supports these findings, as demonstrated by the reduction of SWI cases in patients with an elevated risk for developing SWI.Although ciNPT use did not translate to a shorter hospital stay, it did reveal a lower incidence of readmission to the ICU and repeat surgical procedures that may render potential financial savings.

Figure 1 :
Figure 1: Standardized coefficients of the logistic EuroSCORE and 4 risk factors for sternal wound infection development.ciNPT: closed-incision negative pressure therapy; COPD: chronic obstructive pulmonary disease.

Table 1 :
Audit criteria Sternal wounds severity are classified using Table2.

Table 2 .
Wound classification based on anatomical site plus (adapted from Suelo-Calanao et al. 2020) � number of infections.To calculate the P-value for all categorical data (e.g.BMI, COPD, diabetes, mortality, sex or number of infections), the Fisher exact test of independence (for a twotailed test) was used.A P-value of < 0.05 was recognized as statistically significant between the groups.
Table 5 catalogues the distribution of SWI among patients receiving conventional dressings and ciNPT dressings.Of the 766 patients in the control group, a total of 119 (15.5%) patients had SWI.Of these, 88 patients had an SWI classified as a superficial infection and 31 were classified as deep sternal.A course of antibiotics treatment was administered to the 119 control group patients with SWI.Additionally, control group patients with SWI received the following interventions: debridement (n ¼ 19), sternal rewiring (n ¼ 13), application of conventional negative pressure wound therapy) (n ¼ 28), and surgical revision by a plastic surgeon (n ¼ 7).

Table 3 :
Patient demographics, perioperative risk factors and type of surgery BMI: body mass index; CABG: coronary artery bypass grafting; ciNPT: closed-incision negative pressure therapy; COPD: chronic obstructive pulmonary disease; EuroSCORE: European System for Cardiac Operative Risk Evaluation; SD: standard deviation.

Table 4A -
F. Propensity matching data †A.Summary statistics: closed-incision negative pressure therapy/non-closed-incision negative pressure therapy D. Correspondence between categories of the response variable and the probabilities: closed-incision negative pressure therapy/non-closed-incision negative pressure therapy †Interactions/level: 2; matching method: optimal algorithm Euclidean distance; number of matches: 1:1; caliper: 0.10.The measure and parameters used by the software to identify a 1:1 matched ratio between the 2 groups.� Sigma; confidence interval (%): 95; tolerance: 0.001; number of removed observations: 888.80 years old and above.as of 2022 was £359 [3].The extra expenditure for the control group totaled £91,904 (256 days x £359

Table 5 :
Distribution of sternal wound infections among the matched controls and the closed-incision negative pressure therapy group [17]: negative pressure wound dressing ; SWI: sternal wound infection .�SWIclassification is described in Table 1.incidence of deep SWI cases post sternotomy [1, 13, 14].SWI is a grave postoperative complication associated with cardiac surgery, and recent data have indicated a 30% mortality rate among high-risk populations, with recent data reporting 30% associated mortality among high-risk populations [2].Willy et al. noted obesity (BMI � 30 kg/m 2 ), diabetes mellitus, respiratory insufficiency and tobacco use as risk factors for the development of SWI[15].A 2-centre randomized controlled trial previously reported that 1 independent predictive factor of an increased SWI rate is an elevated preoperative BMI in patients considered to be obese (BMI > 30 kg/m 2 )[16].However, per a multivariate exact logistic regression model in Allen et al., BMI was not a significant predictor for SWI or any sternal complication[17].Our previous research in Ariyaratnam et al. identified 4 comorbidities (BMI, age, diabetes and COPD) that increase the risk of developing SWI

Table 6 .
Postoperative length of stay, readmission to the intensive care unit, extended length of stay and deaths

Table 7 .
Cost comparison between patients in the control group who developed sternal wound infections and received standard dressings and patients who received closed-incision negative pressure therapy -incision negative pressure therapy; ICU: intensive care unit; LOS: length of stay; NHS: National Health Service; SWI: sternal wound infection.#The Author(s) 2024.Published by Oxford University Press on behalf of the European Association for Cardio-Thoracic Surgery.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/),which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited.Interdisciplinary CardioVascular and Thoracic Surgery, 2024, 38, 1-9 https://doi.org/10.1093/icvts/ivae056Original article