Introduction

The reported overall incidence of postoperative infection in patients undergoing major heart surgery (MHS) is 5.9% to 16.5%1,2,3,4,5.

Surgical site infections (SSI) are among the most common infectious complications of MHS with incidences ranging from 1 to 8% and are generally associated with considerable morbidity and mortality6,7,8. These wound infections can be subdivided into superficial (sSSI) and deep-seated postsurgical mediastinitis (PSM).

Efforts to reduce the incidence of PSM consist of several measures9,10,11,12,13, and guidelines for this purpose have been recently issued by Bouza et al.14. However no clinical study has properly and prospectively compared the capacity of different types of wound dressing used after MHS to prevent both sSSI and PSM. While various types differing in terms of format, composition and cost are available, the newer systems such as absorbent foam dressings or NPWT have not yet proven superior in terms of either cost or efficacy15.

Our study prospectively compares the impact that the use of 3 different dressings has on the incidence of superficial and deep infection in patients undergoing major heart surgery.

Materials and methods

Hospital setting and patients

Our center is a general hospital with approximately 1,300 beds and around 50,000 admissions/year. In the Department of Cardiovascular Surgery, a large referral unit, more than 500 MHS procedures are performed on adults annually.

Ethical issues

The Ethics Committee of our institution (Hospital Gregorio Marañon) approved the study (code MICRO.HGUGM.2018–008) and all patients gave their written informed consent for inclusion on admission to hospital prior to surgery. All methods were performed in accordance with the relevant guidelines and regulations.

Study design

This was a randomized prospective clinical study of adult patients undergoing MHS (patients undergoing major cardiac surgery performed under median sternotomy, with or without the need for an extracorporeal pump), over the period October 10, 2019 to February 22, 2022, inclusively.

The study was registered on clinicaltrials.org on 2019–04-05 as NCT03905213.

All patients were subjected to the following measures:

Those scheduled for an MHS procedure were routinely assessed for nasal Staphylococcus aureus carrier status by PCR testing and/or culture. Patients testing positive were decontaminated with nasal mupirocin.

Before surgery, patients were instructed to shower using chlorhexidine soap. The surgical field was also decontaminated in the operating room using an alcoholic chlorhexidine (ChloraPrep™ BD) applicator with 2% chlorhexidine digluconate (20 mg/ml) and 70% isopropyl alcohol (0.70 ml/ml). A chest harness was prescribed from 24 h after surgery until hospital discharge16. All patients were subjected to strict glycemic control following the STS recommendation regarding glycemic management in cardiac surgery, intraoperatively, continuous infusion of IV regular insulin (1 I.U/ml) was started in every patient once glycemia ≥ 180 mg/dl was detected in routine point of care analysis. IV infusion was titrated to maintain glycemia < 180 mg/dl. This infusion was maintained at the critical care setting until oral intake could be restored and stable glycemic control was achieved. Capillary glycemic control was performed every hour until the glycemic target was achieved and every two hours during intravenous infusion treatment.

Routine antimicrobial surgical prophylaxis consisted of 2 g of cefazolin administered at the time of anesthesia induction and every 8 h thereafter to complete a total of three doses (patients allergic to cefazolin received 1 g of vancomycin before surgery and every 12 h thereafter, up to two doses). All the patients were closed in a standard fashion. Simple sternal wires were used to close the sternum. Fascial closure was performed using standard multifilament sutures and the skin was closed with monofilament non absorbable sutures.

All personnel involved in the study were previously trained before the start of the study. In addition, a staff nurse and an intensive care nurse were responsible for ensuring that all staff were trained in applying the different dressings.

Inclusion criteria: Patients over 18 years of age, unselected patients undergoing major heart surgery (MHS) who have signed the consent form. Exclusion criteria included: patients with surgical wound infection.

Randomization

Before surgical wound closure, a 1:1:1 randomization scheme was computer-generated to allocate each patient to a wound dressing group. This information was placed in sealed envelopes to be opened just before sternal wound closure. The three groups established were:

A: Conventional gauze dressing (Mepore®, Mölnlycke Health Care AB, Gothenburg, Sweden).

B: Absorbent polyurethane foam dressing (Mepilex® Border Post-Op, Mölnlycke Health Care).

C: Negative-pressure wound dressing (NPWT) (PICO, Smith & Nephew S.A., Barcelona, Spain). This is a canister-free single-use negative-pressure wound therapy (NPWT) system consisting of a sterile pump and multilayered adhesive dressing that applies negative pressure across the whole dressing surface to ensure a vacuum is delivered to the wound and surrounding tissue.

Dressings were managed as follows:

Gauze dressings were replaced on the second day when drainage tubes were removed. If on Day 4 post surgery the dressing was not stained, the wound was left uncovered. The absorbent dressing ( Mepilex®) is an absorbent foam dressing and were removed if visible soiled and withdrawn at day 7 if the wound was closed and the PICO NPWT dressings were removed after 7 days.

Once the patient was randomized to one type of dressing, he could no longer change to other type of dressing even if he was reoperated.

Data collection

Clinical data were recorded according to a preestablished protocol. Presurgical information was obtained including epidemiologic data, underlying diseases, and standard scores (ie, European System for Cardiac Operative Risk Evaluation (Euro-SCORE), and chronic health evaluation APACHE II score on admission to the ICU). Surgical information included type of surgery, indication, duration, CPBP time, aortic cross-clamp time, transfusion needs, reinterventions, antimicrobial prophylaxis, and need for inotropic support.

Postsurgical outcome events included ICU and hospital length of stay and days on mechanical ventilation (MV). Infections other than sSSI and PSM were also recorded. Outcome variables also included antimicrobial use determined as daily defined doses (DDDs), ICU stay, hospital stay, mortality, associated costs of dressing and dressing-related adverse events. Analysis was performed by a blind statistician.

Follow-up

Physicians from the Departments of Anesthesia and Infectious Diseases monitored patients daily to check for the presence of infection. Our infection control team is multidisciplinary and includes physicians and nurses from the ICU, microbiologists, infectious disease specialists, and health care workers from the Department of Preventive Medicine. All data obtained from patients undergoing MHS were regularly and prospectively collected in a database. Follow-up of patients was continued until hospital discharge and up to 6 months after discharge. Patients who were discharged before the follow-up period were reviewed by a study investigating cardiac surgeon. After discharge, all patients were followed up by telephone with a monthly phone call and were seen in the day hospital when deemed necessary by the research team or by the patient. Patients could contact the research team at any time by phone or e-mail.

Our study has not intervened in the policy of prophylactic and therapeutic use of antimicrobials, which in our case is in line with the Society of Thoracic Surgeons guidelines.

Definitions

Superficial Surgical Site Infection (sSSI) was defined as follows: infections signs that involve only skin or subcutaneous tissue of the incision and at least one or more of the following: purulent drainage from the superficial incision or organisms isolated from an aseptically obtained culture of fluid or tissue from the superficial incision and at least one of the following signs or symptoms of infection: pain or tenderness, localized swelling, redness, or heat, or superficial incision is deliberately opened by a surgeon, unless the incision is culture-negative or diagnosis of a superficial incisional sSSI by a surgeon or attending physician.

The CDC describes mediastinitis as an infection that involves the mediastinal tissues and is characterized by one or more of the following criteria: purulent drainage from the mediastinal area, organisms identified from fluid or tissue in the mediastinal area obtained through a culture or non-culture-based microbiologic testing. Evidence of infection seen during reoperation or by histopathologic examination. Clinical symptoms of mediastinitis such as fever, chest pain, or sternal instability, along with a positive culture or radiographic evidence consistent with infection17,18,19.

Microbiological samples

All microorganisms were identified using standard methods. Antimicrobial susceptibility was determined according to Clinical and Laboratory Standards Institute recommendations20.

Primary endpoint

Primary endpoints of the study were overall incidences of sSSI and PSM.

Secondary endpoints

The following outcome variables were compared between the three groups: consumption of antimicrobials, hospital stay and ICU stay (days), mortality, costs of acquiringeach dressing group incurred during wound care and adverse events.

Statistical analysis and sample size

Qualitative variables are provided as frequency distributions and percentages. Quantitative variables are summarized as their mean and standard deviation (SD) or median and 25th and 75th percentiles (IQR). The three dressing groups were compared using the non-parametric Kruskal–Wallis test. Pearson’s chi-squared test or Fisher’s exact test was used to examine associations between categorical variables.

To identify factors associated with mediastinitis given the low event rate, zero-inflated negative binomial regression was used. Variables included were: age > 70 years, diabetes, COPD, bilateral internal mammary artery (IMA) harvesting, and use of NPWT dressing. These variables were included because they emerged as significant in the univariate analysis and/or because they were considered clinically relevant. In addition, a Kaplan-Meyer analysis was performed to measure the time elapsed until mediastinitis. The log-rank test was used to compare groups.

The analysis of the groups was performed blindly.

Significance was set at p < 0.05. All statistical tests were performed using the software package IBM SPSS Statistics for Windows, version 25.0. Armonk, NY: IBM Corp. and Epidat 3.1

A sample size of 900 patients was calculated based on the hypothesis of a reduction in the rate of superficial wound infection of 10% with a significance level of 5% and a power of 80%

Results

Study population

Over the study period (October 10, 2019 to February 22, 2022), 1,078 patients underwent MHS, of whom 178 were excluded (3 prior sternal infection, 41 no median sternotomy performed,12 no immediate sternal closure, 3 allergy to dressing, 6 cardiac surgery plus another procedure, 44 urgent surgery). Figure 1a.

Fig. 1
figure 1

a.Consort diagram b. Patient recruitmen.

Full size image

The remaining study population of 900 patients were randomly assigned to the different wound dressing groups (gauze, absorbent foam and NPWT; N = 300 each). The Fig. 1b represents patient recruitment.

Participant characteristics

The preoperative and surgical characteristics of the patients in the three groups are compared in Table 1.

Table 1 Characteristics of the study population.
Full size table

Although we found differences in terms of age between the three groups, there were no differences observed in surgical risk classifications. Median (IQR) EuroSCOREs for the gauze, foam and NPWT groups were respectively: 6 (4–8); 6 (4–8) and 6 (3–8); p = 0.257.

Of all the interventions, 49.2% were for valve disease, 30.6% were for CABG and the remaining were other types of surgeries or mixed surgeries (valvular and coronary). We found no significant differences between the three groups in underlying conditions or surgical procedures.

Of all the coronary procedures, 29.4% involved bilateral mammary artery harvesting. These patients were distributed in the groups as follows: gauze 27.6%, absorbent foam 30.8% and NPWT 29.6% (p = 0.889). A total of 7.9% of patients required one or more re-interventions with the following distribution: 8.7%, 8.7% and 6.3%; p = 0.407 in the 3 groups on study.

Superficial wound infection (sSSI)

Rates of superficial wound infection were 2.3%, 3% and 3% respectively for the three groups (p = 0.848). (Table 1).

Postsurgical mediastinitis (PSM)

Sixteen episodes of PSM in 14 patients were recorded among the 900 participants (1.7%). Rates were 2%, 2% and 0.7% respectively for the three groups (p = 0.313).

Of the 14 patients who developed mediastinitis, 5 patients were treated with a single mammary graft (3 conventional dressing and 2 with absorbent dressing) and the other 5 patients had double mammary graft (3 with absorbent dressing and 2 with vacuum dressing).

Median times post-surgery to mediastinitis (in days) were 35 (10.25–57.50), 29 (7.00–44.75) and 28 (15.00–28.00) respectively for the three groups (p = 0.315) (Fig. 2).

Fig. 2
figure 2

Cumulative incidence of mediastinitis.

Full size image

The surgical treatment of patients with mediastinitis varied depending on the extent, etiology and situation of the patients. The 14 patients with PSM were all treated surgically with wound reopening and debridement in all episodes. In addition to this, a pectoral crossing was performed in 3 cases and in 2 other patients a vacuum therapy was performed. Sternal refixation was performed in all surviving patients.

Microorganisms causing of sSSI and PSM

The microorganisms causing the infection of the superficial wound were as follows: Gram positive bacteria caused 53.9% of all episodes, 42.3% Gram-negative bacteria and 3.8% Fungi. There were no significant differences in the microorganisms isolated in the three groups of patients.

The etiology of the episodes of mediastinitis is reported in Fig. 3. Gram-positive bacteria caused 83.33% of all episodes. Two episodes were recorded as polymicrobial. There were no significant differences in the microorganisms isolated in the three groups of patients.

Fig. 3
figure 3

Etiology of the episodes of mediastinitis.

Full size image

Days of ICU and hospital stay

No differences in median length of ICU stay (IQR) emerged between the three groups: 4 (2–6), 4 (3–7), 4 (3–7) days, respectively (p = 0.236.) Median hospitalization days (IQR) were: 15 (10–24), 16.5 (11–28.8) and 15 (10–24) respectively (p = 0.201) (Table 1).

Mortality

Overall mortality was 4.4%. Again, no significant differences were observed in mortality between the three groups 4.7%, 5.7% and 3% respectively (p = 0.277).

Use of antimicrobial agents

Figures for the use of antimicrobials were a defined daily dose (DDD) median (IQR) of 7.3 (4–18.6), with no differences between the groups: 7.5 (3–18.3), 8 (5.1–21) and 6.6 (3.5–18.2) respectively (p = 0.377).

Other infections

Of the 900 patients who underwent MHS, 144 (16.0%) had one or more nosocomial infections, including SSI, during their hospital stay. There was no significant difference in the incidence of these infections among the three groups of patients (Table 1). Main non-SSI infections in these patients were: 66 urinary tract infection, 34 mechanical ventilation-associated pneumonia, 26 tracheobronchitis, 27 non-ventilator associated nosocomial pneumonia, 17 bacteremia, 5 catheter-associated bacteremia, 14 Clostridioides difficile-associated diarrhea, 1 cholecystitis, and 6 other infections.

Wound dressing costs

The cost derived exclusively from the acquisition of dressings in the 3 different groups of patients was: conventional dressings 646 €, absorbent dressings 4,088 € and vacuum dressings 43,335 € respectively.

Adverse events related to the use of dressings

In the absorbent foam dressing group, some patients had a macerated wound requiring more frequent dressing changes. Only 32 of the 300 patients in this group were able to maintain the dressing for one week, in the rest dressings had to be replaced more often. No adverse events were recorded in the other two groups.

Multivariable analysis

In a multivariable analysis where we included the risk variables of developing PSM: age > 70 years, diabetes, COPD, double breast mammary artery harvesting, and applying a vacuum dressing, we only found that being diabetic increased the risk of developing PSM 5.5 times more. (Table 2).

Table 2 Multivariate analysis of risk factors for mediastinitis in patients who underwent MHS.
Full size table

Considering patients who had 2 or more risk factors (age > 70 years, diabetes, COPD, double breast) for PSM, the risk of mediastinitis increase almost 4 times.

Postoperative healing with any of the three types of dressings did not make a significant difference in the risk of PSM, neither in the univariate nor in the multivariate model but not having vacuum dressing increase risk 2 times more although not significant (p = 0.378). (Table 3).

Table 3 Multivariate analysis of risk factors for mediastinitis in patients who underwent MHS.
Full size table

We performed subgroup analyses of patients and the only subgroup of patients who benefited statistically from healing with vacuum dressings were patients with coronary by-passes (301 patients) in whom one or more of the mammary arteries were used as grafts. In this particular subgroup, the treatment with vacuum dressings decreased the risk of having mediastinitis 13 times. p = 0.024; IC (1.394 -123.08). Of these subgroup 10 patients had PSM, of which 8 (80%) had diabetes mellitus.

Therefore, our recommendation after this study is to use it only in that subpopulation.

Discussion

This prospective, randomized study shows that there are no significant differences in the incidence of surgical site infection (SSI) with the use of three different types of commercially available dressings whose economic cost is however very different.

The incidence of infection reported after MHS varies greatly from 1% to 16,5% (5–8) and in our opinion, this is due to the variability and biases of many of these studies. Generally, the reports select certain types of patients (coronary or valvular,…), specific clinical syndromes (ventilator associated pneumonia, mediastinitis, …) and specific periods during which infections occur (stay in the ICU, long term…)21. A study recently performed in our institution, without selecting patients by any of those criteria or by etiology, yielded figures of 15.5% of infectious complications of all types from the time of MHS to hospital discharge5.

The incidence of superficial infections (sSSI) has been estimated at 0.5%–16.5% and 0.2%–8% for deep wounds, or PSM1,22 23,24,25,26. PSM is a particularly serious condition with a mortality that may be as high as 50%26,27,28.

Measures to prevent PSM are well known, however, we were unable to find any specific data on the impact that the use of different types of dressing for surgical wound care may have on either sSSI or PSM14,29.

The types of dressing most frequently used after MHS are the classic gauze dressings (eg, Mepore®), those with absorbent components (eg, Mepilex® Border Post-Op) and those based on generating negative pressure over the covered area (eg, PICO NPWT system).

In a Cochrane review designed to assess the impacts of different dressings on surgical site complications following different surgery procedures, the results of 29 trials (5,718 participants) were examined, but all the studies included but one showed an unclear or high risk of bias. Conclusions were that it remains unclear whether any particular wound dressing is more effective than another at reducing the risk of SSI30. The Wynne study15, similar in intention and sample size to ours, however, evaluates different dressings to ours and does not include vacuum systems. However, in general terms, this study also does not demonstrate a superiority of one dressing over another.

In a prospective randomized controlled clinical trial comparing two wound dressings used after elective hip and knee replacement (Mepilex® Border Post-Op versus Aquacel® surgical), differences were detected between the two dressings but no specific comparison of infection risk was made31. Another prospective cohort study including 200 patients had as an end-point the incidence of blisters. Mepilex dressings were changed less frequently than with the use of conventional dressings but no significant differences were found in SSI32.

In a systematic review of the use of negative-pressure dressings for various surgical wounds, the NPWT vacuum system was associated with a lower incidence of SSI. However, the number of MHS patients was only 80 and these all underwent CABG using IMA13. Turning to the specific field of MHS, a recent comprehensive analysis was designed to examine the impact of NPWT versus traditional dressings used in patients undergoing cardiac surgery. NPWT emerged as more effective at reducing surgical site infections than conventional dressings but this was only demonstrated for PSM.33.

In our prospective, randomized, non selective, study, we were unable to detect any benefits of Mepilex® over the use of conventional gauze dressings. Mepilex were more expensive than conventional gauze dressings and their replacement was more frequently required. Regarding the NPWT dressings our study failed to show significant differences in the incidence of either sSSI or PSM, with the exception of the subgroup of patients undergoing CABG in which the graft is performed with mammary artery samples where PICO dressings reduced the incidence of PSM but not sSSI. Although the risk of infection when using a double mammary is higher, our work suggests that any sternal devascularization carries a higher risk of mediastinitis.

The vacuum system improves blood microcirculation and induces granulation and angiogenesis. The vacuum pressure promotes both exudate removal and arteriolar vasodilatation, this could explain the reduction of mediastinitis.

Although our study was not a cost-effectiveness analysis, only the cost of acquisition of NPWT dressings in our 301 patients undergoing CABG with a mammary artery graft was 40,635 €. Considering that the extra-cost of an episode of mediastinitis after CABG has been estimated between $19,000 to $56,000 per case34, we can speculate that pending the results of appropriate cost–benefit studies, the use of this type of dressings could be recommended in this subgroup of patients.

Among the limitations of our study is the fact that this was a single-center study although this allowed for direct and prospective follow-up of all patients by the same research team inside and outside the ICU. With the differences found so small, there is barely more than 20% power to detect them. Initially, it was expected to find higher SSI rates of at least twice what was observed. It was also expected to find larger differences between the 3 groups. Performing a multivariate analysis of such a rare event may have its limitations, but we consider that the significant variables in the univariate analysis and those that are risk factors for developing mediastinitis have been used.

Our series is the largest we are aware of in whom data on SSI prevention are compared for the use of different dressings in a prospective randomized trial.

In addition, although glucose levels in the first 48 h were maintained in the recommended range, these data have not been reported in this study.

Based on these data, we propose to perform a randomized study in patients undergoing coronary surgery with mammary artery grafting to determine the cost-effectiveness of NPWT and conventional dressings.

Conclusion

Our study carried out in patients after major heart surgery has failed to demonstrate that a systematic change of the usual wound dressings is associated with a reduction in surgical wound infection. The use of vacuum dressings, with a much higher cost than the conventional ones, has shown a greater efficacy in the reduction of post-surgical mediastinitis only in the subgroup of patients who underwent Coronary Artery Bypass Surgery performed with mammary arteries although cost-effectiveness studies are needed in this subgroup of patients.