A Comprehensive Intervention Associated With ... - Oxford Journals

Original Article A Comprehensive Intervention Associated With Reduced Surgical Site Infections Among Pediatric Cardiovascular Surgery Patients, Including Those With Delayed Closure Amanda L. Adler,1 Emily T. Martin,1 Gordon Cohen,1,3 Howard Jeffries,1,2 Michael Gilbert,1,2 Julie Smith,1 and Danielle M. Zerr1,2 1

Seattle Children’s Hospital, and 2Departments of Pediatrics, and 3Surgery, University of Washington, Seattle

Corresponding Author: Danielle M. Zerr, MD, MPH, 4800 Sandpoint Way NE, R-5441, Seattle, WA 98105. E-mail: [email protected]. Received November 15, 2011; accepted January 12, 2012. Background. Surgical site infections (SSIs) cause significant morbidity and mortality in patients undergoing cardiovascular (CV) surgery. Following an increase in SSIs in this population, driven by a high rate in those with delayed closure, we implemented an intervention to reduce these infections and assessed the intervention using both population- and patient-level analyses. Methods. An intervention drawing from existing guidelines and targeting preoperative preparation of the patient, prophylactic antibiotics, and postoperative incision care was implemented. Special attention was paid to standardizing the care of the incision of patients with delayed closure. National Healthcare Safety Network criteria were used to prospectively identify SSIs. Population-level intervention effect was assessed using interrupted time series. To assess intervention adherence and effect in our patient population, retrospective chart review was performed on a cohort of patients undergoing cardiac procedures pre- and postintervention. Multivariate analysis was used to assess risk of SSI at the patient level. Results. Timely preoperative prophylactic antibiotic dosing increased from 60% preintervention to 92% postintervention, and redosing during prolonged surgeries increased from 5% to 79% (both, P < .001). At the population-level, a decrease of 6.7 infections per 100 surgeries per 6 months was observed directly following the intervention (P = .002). The SSI rate decreased from 40% to 0.8% (P < .001) in patients with delayed closure and from 4.3% to 1.8% (P = .02) in patients with immediate closure. In multivariate analyses, surgery prior to the intervention was the strongest predictor for SSI (incidence rate ratio, 3.98; 95% confidence interval, 1.59 to 9.97). Conclusions. Our intervention decreased SSIs in pediatric CV surgery patients, particularly those with delayed closures.

(See the Editorial Commentary by Coffin and Huskins, on pages 44–6.) Surgical site infections (SSIs) increase patient morbidity and mortality as well as healthcare costs. Surgical site infections are the second most common hospital-associated infection in the United States [1], with the Centers for Disease Control and Prevention (CDC) estimating that approximately 500 000 SSIs occur annually [2]. Surgical site infection rates of 0.5%–6% have been reported following pediatric

cardiovascular (CV) surgery [3–6]. There are few publications that report SSI rates in patients with delayed closure of the sternal incision, and procedures with delayed closure are not included in the CDC’s National Healthcare Safety Network (NHSN) surveillance of SSI. Limited evidence suggests SSI rates may be higher in patients with delayed sternal closure [7–10]. Several guidelines aimed at preventing SSI exist [11–13]. These guidelines address timely administration of antibiotic prophylaxis as well as other

Journal of the Pediatric Infectious Diseases Society, Vol. 1, No. 1, pp. 35–43, 2012 DOI:10.1093/JPIDS/PIS008 © The Author 2012. Published by Oxford University Press on behalf of the Pediatric Infectious Diseases Society. All rights reserved. For Permissions, please e-mail: [email protected].

36

Adler et al

evidence-based perioperative practices. Specific guidelines are not available for patients with delayed closure of the sternal incision. Following increased SSI rates in pediatric CV surgery patients, including a large increase in patients with delayed sternal closure, we implemented a comprehensive intervention to reduce SSIs in all CV surgery patients and assessed the effect of the intervention with both population- and patient-level analyses.

MATERIALS AND METHODS Study Design We performed a retrospective cohort study and employed 2 strategies to measure the impact of the intervention on the risk of SSI in patients undergoing cardiac surgery: a population-level assessment of the SSI rate over time and a multivariable analysis of risk of SSI using patient-level data. The study protocol was reviewed and approved by the Seattle Children’s Hospital institutional review board (IRB). The IRB determined that informed consent was not required.

Setting This study was performed at Seattle Children’s Hospital, a 250-bed, tertiary-care, pediatric facility located in Seattle, Washington, where the cardiac surgery team performs an average of 240 surgeries per year. Prospective surveillance of SSI in CV patients is routinely performed by the Infection Prevention Department as part of their standard work. All cases of SSI are identified using NHSN definitions.

Preintervention In late 2005, concerning increases in SSIs among pediatric CV patients prompted leadership within the Infection Prevention and Cardiac Surgery departments to form a multidisciplinary team to strategize ways to reduce SSI in this population. Team members included infectious disease physicians, infection preventionists, cardiac surgeons, anesthesiologists, pharmacists, nurses, cardio-pulmonary perfusionists, and members of the echocardiogram laboratory. The team met at an all-day retreat and reviewed the care of patients undergoing cardiac surgery from entry into our system until discharge. Existing guidelines from the CDC [11], Institute of Healthcare Improvement (IHI), and the Child Health Corporation of America [12] were reviewed, and opportunities to improve practice were identified.

Intervention In early 2006, we implemented a comprehensive intervention that targeted preoperative preparation of the patient, prophylactic antibiotics, and postoperative incision care for both children with primary closure and those with delayed closure. Pre-operative measures included application of mupirocin to the nares twice daily for 5 days prior to surgery for all patients and bathing patients with chlorhexidine gluconate (CHG) from the neck down (including focused cleansing of the proposed site of the surgical incision) for 2 nights prior to and the day of surgery. Children with documented allergy or sensitivity to CHG were bathed with regular soap and water. Using mupirocin and CHG bathing to target Staphylococcus aureus colonization had not been systematically addressed in our institution prior to this intervention. Povidone-iodine was the standard product used for skin preparation prior to the intervention. A formulation of 2% CHG/ 70% isopropyl alcohol (Chloraprep, Cardinal Health) was chosen as the new standard for skin preparation. In patients with documented allergy or sensitivity to CHG, povidone-iodine was used for skin preparation. Prior to the intervention, guidelines regarding prophylactic antibiotics for surgical patients did not exist at our institution. We developed an antibiotic administration strategy for patients undergoing cardiac surgery. This strategy included selection of appropriate antibiotic agents, timely administration of preoperative antibiotics, redosing of intraoperative antibiotics for cases lasting >3 hours (6 hours for a neonate), and administration of antibiotics in the priming volume for patients receiving bypass. Definitions of timely preoperative administration and re-dosing of antibiotics depended on the specific antibiotic agent and the age of the patient. Cefazolin was chosen as the first-line agent. Clindamycin was the alternative agent of choice if a patient had documented allergy (urticaria, angioedema, bronchospasm, hypotension, or arrhythmia) or history of severe drug reaction (drug-induced hypersensitivity syndrome, drug fever, or toxic epidermal necrolysis) to β-lactam agents. Vancomycin was recommended only in cases where β-lactams and clindamycin were contraindicated or the patient was known to be colonized with methicillin-resistant S. aureus either resistant to clindamycin or with unknown susceptibility. Targeted timing of antibiotic dosing was based on individual drug pharmacokinetics with a goal of redosing every 1–2 half-lives. Timing included

Reducing Cardiac Surgical Site Infections

cefazolin 20mg/kg preoperative dose initiated within 5 minutes of incision and repeated every 3 hours (6 for neonates) ± 30 minutes during surgery, clindamycin 10mg/kg preoperative dose initiated within 15 minutes of incision and repeated every 3 hours (6 for term neonates and 12 for pre-term neonates) ± 30 minutes during surgery, and vancomycin 15mg/kg preoperative dose initiated 90–120 minutes prior to incision and repeated every 6 hours (12 hours for neonates) ± 30 minutes during surgery. Patients receiving bypass received an additional standard dose of antibiotic in the bypass priming volume if they were receiving cefazolin or clindamycin. No additional dose was administered to the bypass priming volume for patients receiving vancomycin. No recommendations were made for patients who were already receiving scheduled antibiotics at the time of surgery. Postoperative antibiotic management was not addressed during this intervention. Postoperatively, patients continued to receive the prophylactic antibiotic that was chosen preoperatively, but dosing schedules were transitioned to standard intervals. Prophylactic agents were not chosen based on the status of the incision (immediate versus delayed closure). For patients with closed incisions, prophylactic antibiotics were continued at standard dosing intervals until all chest tubes were removed. In patients with delayed sternal closure, prophylactic antibiotics were continued at standard dosing intervals until sternal closure. Postoperative echocardiograms were performed with efforts made to avoid the new incision whenever possible. Probes were disinfected with a quaternary ammonium product between patients, and a clean sleeve was placed on the probe prior to contact with the patient. Multiuse containers of ultrasound gel were replaced with single use containers. Patients with delayed sternal closure were placed in contact isolation until their sternum was closed. Timely preoperative administration and redosing of antibiotics were tracked prospectively. Prompt feedback (within 3 days of the procedure) regarding late or missed antibiotic doses was provided to the responsible anesthesiologists. Reports tracking SSIs and timeliness of antibiotic administration were provided to surgeons, nursing leadership, anesthesia, and hospital administration on a quarterly basis.

Outcome Cases of SSI were identified prospectively by the Infection Prevention Department. Patients were

followed for 30 days for superficial infections and 1 year for deeper infections per NHSN guidelines. Surgical site infections were identified if a patient was readmitted to our institution with signs of infection and/or had a positive culture obtained from the surgical site at our institution. In patients with immediate sternal closures, SSIs were categorized as “superficial incisional” or “organ space” using NHSN definitions [1]. No infections were categorized as “deep incisional” because of the lack of deep soft tissue between the skin and the sternum. For patients with delayed sternal closure, a positive culture alone did not qualify for an SSI while the sternum was still open. Additional NHSN criteria such as purulence or abscess had to be evident on direct exam. Once the sternum was closed, NHSN criteria were applied in the same manner as patients with immediate closures.

Clinical Data Collection To assess adherence to the intervention and the effect of the intervention at the patient level, retrospective chart review was performed on all patients aged < 21 years who underwent an NHSN cardiac surgical procedure before and after the intervention ( January 2005–December 2007). All procedures involved a median sternotomy. Clinical data included demographic factors, surgical characteristics, timeliness of antibiotic dosing, and pre-operative bathing. The Risk Adjustment in Congenital Heart Surgery score was assigned prospectively by the cardiac surgery team to categorize the complexity of the cardiac surgical procedure [14]. Identification and classification of SSIs was determined using prospective Infection Prevention data as described above. Suspected cases of SSI were evaluated (using the same definitions listed above) by a trained member of the research staff. These SSI categorizations were compared with those prospectively identified by Infection Prevention. Comparison of these data revealed complete concordance. Additional infection data was collected, including organism isolated and whether additional surgery was necessary. To assess the use of mupirocin, medical record review was performed on a randomly selected 10% (n = 42) of the total procedures performed in the postintervention period ( January 2006–December 2007). Use of mupirocin was defined as application of the ointment by clinical staff for an inpatient or receipt of a mupirocin prescription for an outpatient.

37

38

Adler et al

Statistical Analyses To assess adherence to the intervention, the proportion of patients who received timely preoperative antibiotics, appropriate antibiotic redosing, and preoperative bathing was compared pre- and postintervention using χ2 tests. The proportion of the patients receiving mupirocin postintervention was described. The population-level assessment was conducted using segmented regression analysis of interrupted time series to assess the impact of the intervention on the SSI rate over time in patients receiving cardiac surgery. The preintervention period was defined as January 2004–December 2005, and the postintervention period was defined as January 2006–December 2009. Autocorrelation was tested using the Durbin– Watson statistic. For patient-level analyses, a cohort study was performed in which the intervention period ( pre- versus post-) was the dependent variable. The preintervention period was defined as January–December 2005, and the post-intervention period was defined as January 2006–December 2007. Bivariate analyses were performed to identify potential confounders (risk factors that were unevenly distributed between intervention groups). Chi-square tests were used to evaluate dichotomous variables and if assumptions for χ2 were not met, Fisher’s exact test was used. Continuously scaled independent variables were described and tested for significance using Student’s t test or nonparametrically by the Mann–Whitney U test. Due to overdispersion of the count data in the patient level analyses, negative binomial regression was used to assess intervention affect. Variables were selected as candidates for the multivariate model based on either of the following criteria: risk factors that were unevenly distributed between intervention groups (determined by P < .15) or a priori identification as a significant risk factor for SSI (delayed closure). A step-wise multivariate model of intervention effect was constructed. Variables that were not statistically significant (P > .05) were removed from the final multivariate model. All statistical analyses were preformed using STATA version 10.0.

RESULTS Intervention Adherence Timeliness of preoperative antibiotics and antibiotic redosing increased significantly postintervention.

Timely preoperative antibiotic dosing increased from 60% in the 1 year before the intervention to 92% postintervention, and redosing when indicated increased from 8% to 79% (both, P < .001) (Figure 1). Overall, the proportion of patients that received cefazolin was not different between those with immediate versus delayed closure (96% vs 98%). One hundred sixty (93%) patients with immediately closed incisions received cefazolin preintervention versus 373 (98%) patients postintervention (P = .02), and 20 (95%) patients with delayed closure received cefazolin preintervention versus 43 (100%) patients postintervention (P = .33). Among preintervention patients, 9% received a preoperative bath per chart documentation; the remaining 91% did not have a bath documented. Postintervention, 55% of patients were documented as receiving a bath, 8% were documented as not receiving a bath, and 36% had no documentation. Mupirocin was used in 30 (71%) of the randomly selected patients. Use of mupirocin began a median 4.5 days (range, 0–16) prior to surgery.

Population-Level Analyses of SSI Rate Overall, SSIs developed in 23 (7.4%) of 310 procedures preintervention versus 16 (1.7%) of 971 procedures postintervention (P < .001). Surgical site infections developed in 22 (40%) of the 55 patients with delayed sternal closures preintervention compared with only 1 (0.8%) of 125 patients postintervention (P < .001). Similarly, 11 (4.3%) of the 255 patients with immediate sternal closures developed SSIs in the preintervention period compared with 15 (1.8%) of the 846 patients postintervention (P = .02). Sixteen (70%) of the 23 SSIs identified preintervention were classified as organ/space, whereas 14 (88%) of the 16 postintervention infections were classified as organ/space. There was a statistically significant decrease in SSIs even when restricting the analysis to organ space infections (5.2% vs 1.4%; P < .001). The pathogens associated with SSI before the intervention included S. aureus (30%), Aspergillus spp. (13%), Enterobacter spp. (13%), Candida spp. (9%), Peptostreptococcus spp. (4.5%), Pseudomonas aeruginosa (4.5%), coagulase-negative staphylococci (4.5%), and Serratia marcescens (4.5%). Seventeen percent of SSIs did not have a pathogen identified. The pathogens associated with SSI after the intervention included S. marcescens (31%), S. aureus (25%), coagulase-negative staphylococci (20%), Group A streptococcus (6%), Enterococcus spp. (6%), Candida


Figure 1. Percentage of timely antibiotic dosing ( preoperative and redosing) and percentage of patients who developed a surgical site infection (SSI) per 6-month period before and during the intervention. Data from 2005–2007 reflects the patient-level cohort study period; 2008–2009 reflects data from ongoing Infection Prevention surveillance to demonstrate persistence of the intervention and its effect.

spp. (6%) and Viridans streptococci (6%). Staphylococcus aureus accounted for 7 (30%) of the SSIs preintervention and 16 (25%) postintervention. Surgical site infections attributed to S. aureus also significantly decreased during the postintervention period (2.3% vs 0.41%; P = .002). The percentage of SSIs attributed to S. marcescens increased postintervention (4.5% vs 31%). However, when examining the rate of SSIs attributed to S. marcescens, there was not a significant difference between intervention groups (0.3% vs 0.5%; P = .65). Using segmented regression, the preintervention SSI rate was 1.49 SSIs per 100 surgeries per 6 months and was increasing at a rate of 1.7 SSIs per 100 surgeries per 6 months (P = .02) (Figure 2). The postintervention level change was a decrease of 6.7 SSIs per 100 surgeries (P = .002), and the trend change was −1.7 SSIs per 100 surgeries (P = .02). The Durbin–Watson statistic showed no evidence of autocorrelation for these data.

preintervention period were more likely to receive autologous or allogeneic transplanted material during surgery (P < .001) and have a surgery lasting >6 hours (P = .02), whereas those in the postintervention group were more likely to have noncardiac comorbidities (P = .03) (Table 1). Other measured variables did not differ between groups. Single predictors for SSI that met criteria for inclusion in the multivariate model were preintervention status, delayed sternal closure, surgery lasting >6 hours, use of transplanted material during surgery, and presence of a noncardiac comorbid condition. In multivariate analyses, preintervention status and delayed closure were significantly associated with increased rates of SSI. Surgery prior to the intervention was the strongest predictor for SSI, even after controlling for delayed closure and long surgeries (incidence rate ratio, 3.98; 95% CI, 1.59–9.97) (Table 2).

DISCUSSION Patient-Level Analyses Chart review was performed on 618 procedures. Of these, 194 were from the preintervention period and 424 were postintervention. Patients in the

We describe the impact of a comprehensive intervention aimed at reducing SSIs in children undergoing CV surgery. The intervention was followed by increases in preoperative CHG bathing, timely administration of

39

40

Adler et al

Figure 2. Percentage of patients undergoing cardiovascular surgery who developed a surgical site infection (SSI) per 6-month period before and during the intervention.

preoperative and intraoperative antibiotics, and a significant decrease in SSI rates in both patients with immediate and delayed closures. Procedures with delayed sternal closures are not included in NHSN surveillance. However, we felt it was important to include this population in surveillance efforts and the intervention. Reports suggest patients with delayed sternal closure are at increased risk of SSI, other hospital-acquired infections, and mortality. [7–10, 15–17] Johnson et al reported that centers that used delayed sternal closure in >25% of their cases of stage 1 palliation for hypoplastic left heart had a significantly higher rate of SSI in this population [8]. Additionally, Long and associates reported that delayed sternal closure was an independent risk factor for Gram-negative mediastinitis [7]. Conversely, other reports found that delayed sternal closure was a significant predictor for SSI in univariate analyses but did not remain significant in multivariate analyses [9, 10]. Although the baseline rate of SSI was much higher in delayed versus immediately closed incisions (40% vs 3%), significant reductions in the rate of SSI were achieved after the intervention in both patient populations (40% to 0.8% and 4.3% to 1.8%, respectively).

A major focus of our intervention was the development of a standard antibiotic prophylaxis procedure based upon recommended guidelines. Our strategy focused on 2 concepts: (1) that the initial dose of antibiotic is timed such that the level of drug reaches therapeutic concentration prior to incision, and (2) that therapeutic levels of antimicrobial agent in serum and tissues are maintained throughout the procedure. Recommendations call for antimicrobial prophylaxis to be initiated within 60 minutes prior to incision (120 minutes for vancomycin). Prior to our intervention, only 60% of patients we reviewed received preoperative prophylaxis following this standard. This is consistent with adult studies that report preintervention compliance of 5%–72% [18, 19]. To maintain therapeutic levels during surgery, recommendations call for redosing antibiotics every 1–2 half-lives until the incision is closed [11, 12]. Our baseline estimate of intraoperative redosing was 8%, similar to the 12% reported by Bratzler et al [18]. Ultimately, appropriate antibiotic redosing increased to >95% postintervention. Timely preoperative antibiotic dosing was achieved faster than timely redosing. The real-time feedback and quarterly reports on antibiotic dosing timeliness and SSI rates served to keep staff members


Table 1. Demographic and Clinical Characteristics of Patients Included in the Patient-Level Cohort Study Undergoing Cardiovascular Surgery by Intervention Period Preinterventiona (n = 194)

Postinterventiona (n = 424)

10 (0–235) 3–65

9 (0–254) 3–47

P Value

Demographics Age, median months (range) IQR

.33

Neonatal age

24 (12%)

71 (17%)

.16

Prematurity (gestational age 6 hours

.02

6 (1.5%)

2 (1%) 228 (58–840)

2 (0.5%) 212 (28–598)

.07

114 (59%)

246 (58%)

.86

18 (16%) 96 (84%)

51 (21%) 197 (79%)

26 (13%)

32 (8%)

Neonates Nonneonates

3 (12%) 23 (88%)

7 (22%) 25 (78%)

Use of bypass

166 (86%)

368 (87%)

Duration of bypass, mean minutes (range) Ischemic time, mean minutes (range) Delayed sternal closure Use of implantsc Xenografts

91 (0–504) 45 (0–378)

83 (0–528) 44 (0–212)

21 (11%)

43 (10%)

.80

285 (67%) 77 (27%)

.33

45 (33%)

112 (39%)

Nonporous (Gore Tex) Mechanical

99 (72%) 8 (6%)

228 (80%) 24 (8%)

Autologous Allogeneic

51 (26%)

51 (12%)

47 (92%) 4 (8%)

50 (98%) 1 (2%)

The preintervention period represents January–December 2005 and the postintervention period represents January 2006–December 2007. Abbreviations: IQR, interquartile range; RACHS, risk adjustment in congenital heart surgery. a

Continuous variables reported as described; dichotomous variables are reported as no. (%).

b c

.68 .20 .73

138 (71%) 68 (49%)

Porous (Dacron)

Use of transplanted material

.02

P value obtained using Fisher’s exact test. Values will add up to more than 100% as patients could have received >1 implant type during a single surgery.

6 hours

4.47 (1.74–11.52) 4.19 (1.56–11.27)

.002 .004

3.26 (1.23–8.66) 2.14 (.76–6.0)

.02 .15

.75

a

a

.001

3.98 (1.59–9.97)

.003

Transplanted material Intervention status

1.2 (.4–3.64) 4.63 (1.84–11.68)

a

Abbreviations: CI, confidence interval; IRR, incidence rate ratio. a

Variable did not meet criteria for inclusion in the final model.

informed and engaged in the intervention. This level of engagement and cooperation among key groups such as cardiac surgery and anesthesia was critical in creating a change in culture regarding antibiotic dosing. Approximately 85% of our patients required the use of cardiopulmonary bypass (CPB). Previous data has shown that CPB reduces the serum concentration of antibiotics [20–22]. Fellinger et al showed that an additional dose of antibiotic at the initiation of CPB corrected this drop and resulted in therapeutic concentrations of cefazolin for the 2 most common Gram-positive organisms causing postoperative infection [20]. In our intervention, patients receiving cefazolin or clindamycin had a standard dose of antibiotic added to the priming volume and administered directly into the extracorporeal circuit. Staphylococcus aureus was the most common pathogen causing infection in the preintervention group (30%) and accounted for 25% of the SSIs postintervention. We observed a statistically significant decrease in SSIs when restricting the analysis to S. aureus infections (2.3% vs 0.41%). Staphylococcus aureus colonization is detectable in 37% of children and often involves extranasal sites [23]. Consequently, to target S. aureus colonization, our standardized guideline included application of mupirocin to the nares twice daily for 5 days prior to surgery and preoperative bathing and skin preparation with CHG-based products. Both mupirocin and CHG have been shown to be effective in S. aureus decolonization and reducing risk of S. aureus infection. In a randomized, blinded, placebo-controlled trial of nasal mupirocin, Perl et al demonstrated that nasal carriage of S. aureus was eliminated in 81.3% of carriers who received 3–5 doses of mupirocin and that colonized patients receiving nasal mupirocin had fewer postoperative S. aureus infections than those receiving placebo ointment [24]. More recently Bode et al demonstrated a decrease in hospital-associated

S. aureus infections in S. aureus–colonized patients randomized to nasal mupirocin and CHG baths versus placebo (3.4% vs 7.7%, respectively) [25]. Previous risk factors for development of SSI have included neonatal age, duration of surgery, and delayed sternal closure [3–7]. Patients in the preintervention group were more likely to have a surgery lasting >6 hours. In multivariate analyses, surgery lasting >6 hours and delayed closure were independently associated with SSI. However, preintervention status was the strongest predictor for SSI, with an incident rate ratio of 3.98, even after controlling for delayed closure and long surgeries. This study is limited by the before-after design and because it represents a single institution. Additionally, we were unable to audit all aspects of the intervention equally. Strengths of the study include prospective surveillance of SSIs using strict definitions and prospective surveillance of timely antibiotic dosing. This allowed us to provide timely feedback to medical providers, and we believe this had a considerable impact on the success of our intervention. We implemented a comprehensive intervention that improved timeliness of preoperative administration and intraoperative redosing of prophylactic antibiotics and dramatically improved SSI rates, particularly in patients with delayed closures. Surgical site infection reduction strategies may be beneficial in high-risk patients like those with delayed sternal closure.

References 1. Edwards JR, Peterson KD, Mu Y, et al. National Healthcare Safety Network (NHSN) report: data summary for 2006 through 2008, issued December 2009. Am J Infect Control 2009; 37: 783–805. 2. Wong ES. Surgical site infection. In: Mayhall CJ, ed. Hospital epidemiology and infection control.


3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

2nd ed. Philadelphia: Lippincott, Williams, & Wilkins, 1999:189–210. Allpress AL, Rosenthal GL, Goodrich KM, Lupinetti FM, Zerr DM. Risk factors for surgical site infections after pediatric cardiovascular surgery. Pediatr Infect Dis J 2004; 23:231–4. Nateghian A, Taylor G, Robinson JL. Risk factors for surgical site infections following open-heart surgery in a Canadian pediatric population. Am J Infect Control 2004; 32:397–401. Mehta PA, Cunningham CK, Colella CB, Alferis G, Weiner LB. Risk factors for sternal wound and other infections in pediatric cardiac surgery patients. Pediatr Infect Dis J 2000; 19:1000–4. Sohn AH, Schwartz JM, Yang KY, Jarvis WR, Guglielmo BJ, Weintrub PS. Risk factors and risk adjustment for surgical site infections in pediatric cardiothoracic surgery patients. Am J Infect Control 2010; 38:706–10. Long CB, Shah SS, Lautenbach E, et al. Postoperative mediastinitis in children: epidemiology, microbiology and risk factors for Gram-negative pathogens. Pediatr Infect Dis J 2005; 24:315–9. Johnson JN, Jaggers J, Li S, et al. Center variation and outcomes associated with delayed sternal closure after stage 1 palliation for hypoplastic left heart syndrome. J Thorac Cardiovasc Surg 2010; 139:1205–10. Costello JM, Graham DA, Morrow DF, et al. Risk factors for surgical site infection after cardiac surgery in children. Ann Thorac Surg 2010; 89: 1833–41; discussion 41–2. Holzmann-Pazgal G, Hopkins-Broyles D, Recktenwald A, et al. Case-control study of pediatric cardiothoracic surgical site infections. Infect Control Hosp Epidemiol 2008; 29:76–9. Mangram AJ, Horan TC, Pearson ML, Silver LC, Jarvis WR. Guideline for prevention of surgical site infection. Hospital Infection Control Practices Advisory Committee. Infect Control Hosp Epidemiol 1999; 20:250–78; quiz 79-80. Institute for Healthcare Improvement. Pediatric surgical site infection guideline. Available at: http:// www.ihi.org/IHI/Programs/Campaign/SSI.htm. Accessed 14 March 2010. Anderson DJ, Kaye KS, Classen D, et al. Strategies to prevent surgical site infections in acute care hospitals. Infect Control Hosp Epidemiol 2008 Oct;29 Suppl 1:S51-61.

14. Jenkins KJ, Gauvreau K, Newburger JW, Spray TL, Moller JH, Iezzoni LI. Consensus-based method for risk adjustment for surgery for congenital heart disease. J Thorac Cardiovasc Surg 2002; 123:110–8. 15. Alexi-Meskishvili V, Weng Y, Uhlemann F, Lange PE, Hetzer R. Prolonged open sternotomy after pediatric open heart operation: experience with 113 patients. Ann Thorac Surg 1995; 59:379–83. 16. Levy I, Ovadia B, Erez E, et al. Nosocomial infections after cardiac surgery in infants and children: incidence and risk factors. J Hosp Infect 2003; 53:111–6. 17. Das S, Rubio A, Simsic JM, et al. Bloodstream infections increased after delayed sternal closure: cause or coincidence. Ann Thorac Surg 2011; 91: 793–7. 18. Bratzler DW, Houck PM, Richards C, et al. Use of antimicrobial prophylaxis for major surgery: baseline results from the National Surgical Infection Prevention Project. Arch Surg 2005; 140:174–82. 19. Forbes SS, Stephen WJ, Harper WL, et al. Implementation of evidence-based practices for surgical site infection prophylaxis: results of a preand postintervention study. J Am Coll Surg 2008; 207:336–41. 20. Fellinger EK, Leavitt BJ, Hebert JC. Serum levels of prophylactic cefazolin during cardiopulmonary bypass surgery. Ann Thorac Surg 2002; 74: 1187–90. 21. Hatzopoulos FK, Stile-Calligaro IL, Rodvold KA, Sullivan-Bolyai J, Del Nido P, Levitsky S. Pharmacokinetics of intravenous vancomycin in pediatric cardiopulmonary bypass surgery. Pediatr Infect Dis J 1993; 12:300–4. 22. Lonsky V, Dominik J, Mand’ak J, et al. Changes of the serum antibiotic levels during open heart surgery (ceftazidim, ciprofloxacin, clindamycin). Acta Medica (Hradec Kralove) 2000; 43:23–7. 23. Kuehnert MJ, Kruszon-Moran D, Hill HA, et al. Prevalence of Staphylococcus aureus nasal colonization in the United States, 2001–2002. J Infect Dis 2006; 193:172–9. 24. Perl TM, Cullen JJ, Wenzel RP, et al. Intranasal mupirocin to prevent postoperative Staphylococcus aureus infections. N Engl J Med 2002; 346: 1871–7. 25. Bode LG, Kluytmans JA, Wertheim HF, et al. Preventing surgical-site infections in nasal carriers of Staphylococcus aureus. N Engl J Med 2010; 362:9–17.

43