Since introduced for the first time in 1975 by Riahi et al, delayed sternal closure (DSC) remains a commonly utilized strategy in congenital heart surgery (CHS) for patients with anticipated hemodynamic instability based upon complexity of the performed procedure or in patients who exhibit instability during surgery [1]. The most common surgery utilizing DSC in pediatrics is the Norwood procedure. DSC is utilized in about 10% of all CHS while in less than 2% of adult cardiac surgeries [1,2]. DSC aids hemodynamic stability by improving cardiovascular and pulmonary compliance and positively impacting cardiorespiratory mechanics. At the same time, DSC is associated with increasing incidence of complications and the most common is surgical site infections (SSI). Multiple recent studies from different institutions have reported SSI rate ranging from 3% to 18% for cases with DSC [3-6]. Overall mortality in patients with DSC ranges from 12% to 36% [7-9]. DSC is also associated with longer hospital length of stay and longer ICU length of stay which translates into substantially increased healthcare burden with suboptimal outcomes in terms of prolonged hospitalization, increased readmissions and increased morbidity and mortality [10]. A study of costs and length of stay (LOS) for children with surgical site infection after pediatric cardiothoracic surgery report an average increase of $136,950 and increased LOS of 9.5 days per case [11]. There are efforts going on at multiple CHS programs to decrease the incidence of SSI. Our research study “A Quality Improvement Initiative to Reduce Surgical Site Infections in Patients Undergoing Delayed Sternal Closure After Pediatric Cardiac Surgery” [12] was an attempt to mitigate risk factors contributing to an increased infection in DSC population.

Purpose of delayed sternal closure

Low cardiac output state is common after complex cardiac surgeries due to cardiac arrest and cardiopulmonary bypass utilized during surgery which is further impacted by sternal closure. Studies by Maggio et al. and Shalabi et al. showed that reopening the sternum significantly increases cardiac index immediately which shows further improvement over subsequent 3-11 days [1,13]. The following conditions remain the common indication for DSC: hemodynamic instability, myocardial edema, cardiac dilatation, intractable bleeding, coagulopathies, dysrhythmias, respiratory compromise, and placement of a circulatory assist device [14].

Care of the open sternum

Intraoperative care; and postoperative surgical and nursing care are the mainstay for prevention of SSI. We initiated the project [12] to examine the impact of care modifications in reducing the incidences of SSI in our DSC population. We made modifications according to recent studies in our existing protocol-based bundled approach (Table 1). A protocol-based bundled approach has been shown to decrease the incidence of SSI [4,15]. An increasing number of congenital heart programs are adopting this approach to combat the problem of SSI [9]. However, no uniform practice or national guideline exists currently for a preventative approach to decreasing SSI in patients with CHS. A previous prospective quality improvement project completed at our pediatric cardiac critical care program as well as other studies have shown decreasing SSI with adherence to a protocol-based approach in patients with CHS [15,16]. That study showed higher incidence (18%) of SSI in cases with DSC which decreased by the second year of the project [17].

Care and planning for an open sternum begins intraoperatively. Antibiotic prophylaxis is continued till 24 or 48 hours after the sternum is closed. This practice appears to be uniform across different programs but the choice of antibiotics is not and is discussed below. When the sternum is left open, the choice of dressings to cover the mediastinal viscera varies between different CHS programs. Commonly utilized approaches are closing the skin over open sternum, using polytetrafluoroethylene (PTFE) or using silastic membrane cut in the shape of incision and covered with an occlusive sterile dressing. When the sternum is provided with an appropriate coverage, reopening and irrigation on daily basis is not necessary. Care of the sterile dressing varies between programs in terms of either daily dressing changes or to be changed only if soaked. The policy about personnel doing dressing changes also variable which could be either the bedside nurse or the cardiothoracic surgery team. There is also variation in coverage provided to the open sternum among centers as detailed in introduction. Comparison of these different techniques has not been studied but a single center retrospective review by Yabrodi et al. showed SSI rate in DSC to be 0.6%. They attributed their lower rate to practice of closing the skin over open sternum [2]. There are studies available looking at outcome difference between sternum closed in ICU versus OR. STS database review study by Nelson-McMillan et al and other single center retrospective chart review studies by Bowman et al. and Shalabi et al. showed that there was not a significant difference between sternal closure at bedside in PICU versus in the operating room (OR) [1,6,18]. About the closure technique, some studies have highlighted the effectiveness of the VAC system on microcirculation and the promotion of granulation tissue proliferation [19-22]. 15 years review of 23,499 sternotomies by Baillot et al. showed lower mortality in patients with wound healing by VAC system [23].

Antibiotic prophylaxis and duration

Choice of prophylactic antibiotic is variable among different programs. There are no guidelines, either adult or pediatrics, for antibiotic usage for patients whose sternums are left open after surgery. Standard surgical prophylaxis for patients whose sternums are closed is cefazolin as a preoperative dose and continuation for 24 hours postoperatively. For patients with beta lactam allergy, clindamycin or vancomycin are recommended. For patients with MRSA (Methicillin resistant Staphylococcus aureus) infection or history of MRSA infection, Vancomycin is recommended. As an adaptation for the open sternum, this prophylaxis is continued postoperatively till 24-48 hours after the sternum is closed. For the discussion of DSC in this paper, we will consider this approach as “standard” and any variation from this would be “non-standard”. Some other antibiotic practice variations are as follows: Use of Vancomycin and Gentamycin combination for all DSC cases till 48 hours after sternal closure and switching to Cefazolin until chest tubes or mediastinal drains are in place 3; Use of Cefazolin and Vancomycin for all the cases as well as changing it to Piperacillin-Tazobactam with Vancomycin if the open sternum is expected beyond 72 hours postoperatively [24-27]. Among the postoperative variables, the duration and choice of antibiotic prophylaxis as well as the role of routine mediastinal cultures in the postoperative period remains highly debated. Most common organisms identified for SSI in DSC patients have been Staphylococcal species and gram-negative bacteria. A systemic review and meta-analysis from January 2000 to July 2019 by Silvetti et al. [28] looking at antibiotic management of DSC after CHS yielded that a total of 2,203 patients requiring an open sternum after cardiac surgery, 350 (15.9%) of whom developed infections and 182 (8.3%) developed a surgical site infection. The SSI rate in patients with “non-standard” strategy was higher than in patients with “standard” strategy: 8.8% (140 reported infections/1,582 patients) versus 6.8% (42 reported infections/621 patients), p=0.001. There are no pediatric studies which investigated the most effective duration of antibiotic after sternal closure, but a recent adult study by Eckardt et al. showed that continuing antibiotic beyond 24 hours after sternal closure did not result in reduction in infections [29]. The study reporting the lowest SSI rate (2.3%) was published in 2018. This retrospective study described treatment in a 10-year period (from January 2003 to December 2012), included 259 patients with DSC, all patients received only Cefazolin as antimicrobial drug and while the duration of therapy was not reported, only 6 patients developed mediastinitis [30]. In addition to no clear evidence of benefit, using medications which require frequent drug level monitoring adds to the cost of care provided. Also, Vancomycin and Gentamycin carries an inherent risk for renal toxicity, putting patients at additional risk during an already critical physiological state. These evidence underlines the importance of need for the consensus guidelines to minimize multidrug antibiotic prophylaxis approach.

Timing of delayed sternal closure

Timing of sternal closure depends upon hemodynamic stability of the patient and reaching a physiological recovery phase which is usually achieved within 24-72 hours of the initial surgery [31-33]. Practice of sternal closure location varies between programs. It could be either at bedside in intensive care unit (ICU) or in the operating room (OR). The chest cavity and mediastinum are irrigated with normal saline, an antibiotic solution or mixer of antibiotic and antiseptic solution before chest closure. Sternum is closed in primary intention with wires and overlying skin is closed either with primary intention or a vacuum assisted closure (VAC) system is applied. The VAC system is based on the application of negative pressure by controlled suction to the wound surface. For some programs a wound culture is always taken at the time of closure [5] versus for others, it is taken only if there is a visible concern for infection.

Complications and risk factors

Surgical site infections after a congenital heart surgery are a significant problem as it leads to increased patient morbidity. DSC has been shown to be associated with a higher number of infectious complications including sepsis with blood stream infection, nosocomial infections, mediastinitis and sternal wound infection (SWI) [5,24,34]. As discussed above, SSI in DSC is also shown to be associated with increased length of mechanical ventilation, increased length of ICU stays, increased length of hospital stays and substantial increased cost to the healthcare system [11]. A major STS database review study by Nelson-McMillan et al looking at 6127 operations showed that among the infectious complications; sepsis (8.2%), SSI (6.3%) and mediastinitis (1.8%) were common and; were positively correlated with duration of open sternum [18]. In the same study, multivariate analysis found significant association of infectious complications with preoperative renal dysfunction, preoperative mechanical ventilation, Z score for the weight in neonate and Norwood procedure compared to any other index procedure. Study by Harder et al. [5] concluded that multiple periods of DSC, longer duration of DSC, greater dependence on parenteral nutrition, and extracorporeal membrane oxygenation were significantly associated with SSI in DSC cases. Among other modifiable risk factors, there are studies indicating that infection risks are higher in patients uncontrolled postoperative bleeding [10,35]. Close intraoperative and postoperative attentions to correcting the coagulopathy could have positive effects on decreasing SSI rate.

Future direction

As discussed above, protocol based bundled approach has been successful in multiple studies to decrease the SSI in DSC population. There still exists substantial variation in components of these bundles across different programs. As per currently available evidence, antibiotic usage needs to be narrowed down to follow standard surgical prophylaxis recommendations as the multidrug approach has failed to show any benefit in decreasing SSI rate. There needs to be larger multicenter study trial looking at different methods of surgical mediastinal coverage provided and difference in rate of SSI. Surgery and intensivist team need to be diligent about performing the sternal closure at the earliest permissible clinical status as each increasing day increases the SSI and mediastinitis incidence. Patients requiring ECMO should be assessed for readiness to come off the ECMO as number of days on ECMO are shown to have positive correlation with SSI. The practice of transferring the patient to OR for the sternal closure can be avoided as majority of centers do perform sternal closure at the bedside and it has not been shown to be a factor related to SSI. Also, the transfer of critical patient has its own safety risks which can be avoided. Individual studies on VAC system for sternal closure have encouraging results. A large multicenter trial comparing VAC system with the traditional approach could provide a valuable information regarding wound healing hastiness and decreasing SSI. Again, there is little in the literature about the role of preoperative renal or other organ dysfunction, preoperative mechanical ventilation and preoperative nutritional status in increasing SSI. These could be major influencing factors for patient’s general health and may limit their ability to mount enough immune response to fight infection. We could gain a greater understanding about the role of these factors by conducting a large multicenter trial to look at different patient profiles and alternative practices.

DSC approach needs to be taken judiciously as it has vast implications in terms of patients short- and long-term morbidity; and mortality. It is crucial for the cardiac surgeons and cardiac intensivists to have a clear understanding of risk factors and outcomes which influence clinical decisions. As apparent from the discussion, there is an extensive diversity among different programs about the care for DSC cases. It is imperative to unify these diversities to find out the best approach for the betterment of outcomes in DSC care and decreasing infectious complications.