Bloodstream Infections in Infants and Children With Congenital Heart Disease Undergoing Cardiac Surgery

Jessica Ward, PhD, MPH, MSN, CPNP, RN; Josseline Herrera-Eguizabal, MS; Keirsten Andersen; Kelsey Ryan; Melanie Guerrero, DNP, RN, CPN, CLSSGB; Marisa Glucoft, MPH, CIC, HACP; Paula Murray, PhD

Disclosures

Am J Crit Care. 2023;32(3):157-165.

Abstract and Introduction

Abstract

Background: Children with congenital heart disease undergoing cardiac surgery are at risk for laboratory-confirmed bloodstream infections (LCBIs). These infections can lead to morbidity, mortality, and increased health care costs. The role of mucosal barrier injury in causing LCBIs is unknown.

Objectives: To describe characteristics of LCBIs in patients admitted to cardiac intensive care and step-down units and to assess frequencies of National Healthcare Safety Network infection types and associations with organism classification, patient clinical factors, and infection outcomes.

Methods: A retrospective cohort analysis using manual electronic medical record data abstraction included children with congenital heart disease who developed an LCBI while receiving inpatient cardiac care between August 2011 and November 2018 at one institution. Demographic, clinical, laboratory, and outcome variables were collected and analyzed with descriptive and inferential statistics.

Results: Eighty-seven patients with congenital heart disease developed 103 LCBIs during the study time frame. The most common causative microorganisms were gram-positive bacteria, including Enterococcus faecalis and Staphylococcus epidermidis. Sixty-three percent of causative organisms were characterized as originating from mucosal barrier injury, although no infections met National Healthcare Safety Network criteria for mucosal barrier injury LCBIs.

Conclusions: Translocation of bacteria through injured gut mucosa may cause bloodstream infections in children with congenital heart disease. Further investigation is warranted to understand microbiome changes that adversely select pathogenic gut organisms. Preventive care to maintain intact gut function and a healthy microbiome should be explored for this patient population.

Introduction

Approximately 1.35 million newborns were born with congenital heart disease (CHD) every year from 1995 to 2010, with a global birth prevalence of 9.1 per 1000 live births.^[1] The prevalence of CHD in the United States was approximately 10.8 per 1000 live births in 2008.^[2] Developments in cardiac surgery have improved the morbidity and mortality rates of patients with CHD.

Despite these advances, empiric evidence indicates that hospital-acquired infections are frequent perioperative complications experienced by neonates, infants, and children with CHD. These infections are associated with longer stays, decreased quality of life, and increased hospital expenditures.^[3–5] Bloodstream infections (BSIs) are the most common hospital-acquired infections.^[5–7] Risk factors for BSI include younger age, underlying disease, exposure to medical devices, inappropriate timing of prophylactic antibiotics, hyperglycemia, and amount of blood products transfused.^[6–10]

Conditions such as immune suppression, impaired gut function, or injured gut mucosa may predispose patients to laboratory-confirmed bloodstream infections (LCBIs) that do not arise from central venous catheter contamination.

Coagulase-negative staphylococci, closely followed by Staphylococcus aureus, were identified as frequent causes of BSI in children undergoing cardiac surgery.^[4,5] The propensity for bacteria to translocate through injured gastrointestinal mucosa to the bloodstream in this patient population is not completely known. Further, evidence is limited regarding the associations between leukopenia, neutropenia, and lymphopenia and the incidence and causes of BSI in children with CHD undergoing cardiac surgery.

Definitions of Bloodstream Infection

National Healthcare Safety Network (NHSN) classifications of BSI include laboratory-confirmed BSI (LCBI) and mucosal barrier injury LCBI (MBILCBI).^[11] According to NHSN criteria, LCBIs occur when organisms are isolated from blood and not a secondary source such as a wound, urine, the respiratory tract, or a surgical site. The diagnosis of LCBIs caused by bacterial or fungal pathogens that are not considered common commensal organisms is based on the results of 1 or more blood cultures or non-culture microbiologic tests.

Common commensal bacteria typically exist on body surfaces and beneficially coexist with the human host in a nonharmful manner. However, under conditions such as altered skin or mucosal integrity and immunosuppression, these endogenous microbes may become harmful and cause infection.^[12] Laboratory-confirmed BSIs often occur when a central venous catheter (CVC) is in place, but the presence of a CVC is not required for a BSI to be designated an LCBI.^[11] Certain conditions may predispose patients to LCBIs that do not arise from CVC contamination. These conditions include immunosuppression (due to acute illness or medications to prevent organ rejection), impaired gut function, and injured gut mucosa.

The diagnosis of LCBIs caused by common commensals is based on the results of 2 or more cultures obtained on separate occasions. Patients with common commensal LCBIs must experience signs or symptoms of infection such as fever, hypotension, or chills during the infection window period. The infection window period is defined as the 7-day window during which the infection occurred, including the date the positive blood culture (or other diagnostic test) result was identified, the 3 calendar days before this date, and the 3 calendar days after this date.

Mucosal barrier injury LCBIs caused by intestinal organisms are subsets of LCBIs and must meet LCBI criteria and a set of criteria specific to the MBI-LCBI category. The causative organism must be intestinal in origin, viridans streptococci, and/or Rothia species for the LCBI to meet MBI-LCBI criteria. Patients must have concurrent neutropenia, as defined by an absolute neutrophil count and/or white blood cell count of less than 500/μL on at least 2 separate days during the infection window period. Patients who received an allogeneic hematopoietic stem cell transplant (HSCT) and have active gastrointestinal graft-vs-host disease or high-volume diarrhea also meet MBI-LCBI criteria, even in the absence of neutropenia or leukopenia.

Past allogeneic HSCT, neutropenia, and leukopenia during the infection window period are the only clinical criteria included in the MBI-LCBI definition. These NHSN MBI-LCBI criteria do not account for increased intestinal permeability^[13] and gut barrier dysfunction^[14] during the perioperative period in children with CHD. This pathophysiology enables the translocation of gut bacteria into the bloodstream. We hypothesized that children with CHD undergoing cardiac surgery are susceptible to MBI-LCBIs caused by translocation of pathogenic gut organisms to systemic circulation.

Objectives

To our knowledge, there are no published studies discerning the incidence of MBI-LCBIs in children with CHD in the perioperative inpatient setting. Evidence is needed to evaluate the possibility of widening the NHSN MBI-LCBI definition to include disease states that may predispose patients to translocation of gut organisms through injured and permeable gut. The purpose of this study was to address this gap in the literature by evaluating the characteristics of documented LCBIs that occurred during an 8-year period in a cohort of children admitted to the inpatient cardiac units at a single institution. The study objectives were to (1) describe characteristics of LCBIs that occurred in patients admitted to the inpatient cardiac intensive care and step-down units from August 2011 through November 2018 and (2) determine the frequencies of NHSN LCBI types and associations with organism classification, patient clinical factors, and LCBI outcomes such as infection resolution, CVC removal, shock, and death.

1 2 3 4 5

Next Section

Table 1. Characteristics of 87 patients included in the study
Characteristic	No. (%)^a
Age at time of bloodstream infection, mean (SD), y	1.2 (3.3)
Sex
Female	43 (49)
Male	44 (51)
Race
White	28 (32)
Black or African American	13 (15)
American Indian/Alaskan native	1 (1)
Asian or Pacific Islander	6 (7)
Other	36 (41)
Not identified	3 (3)
Ethnicity
Hispanic or Latino	41 (47)
Not Hispanic or Latino	29 (33)
Not identified	17 (20)
Death during hospital stay with bloodstream infection
Yes	17 (20)
No	70 (80)
Length of stay, mean (SD), d	101.7 (80.7)
Diagnoses (n = 168 diagnoses)
Hypoplastic left heart	28 (16.7)
Ventricular septal defect	22 (13.7)
Pulmonary atresia/pulmonary stenosis	18 (10.7)
Patent ductus arteriosus	14 (8.3)
Double-outlet right ventricle	10 (6.0)
Diagnosis other than congenital heart disease	8 (4.8)
TGA	7 (4.2)
Atrial septal defect	6 (3.6)
TAPVR, coarctation of aorta, tetralogy of Fallot/pulmonary atresia, aortic stenosis/aortic atresia (n = 5 each)	5 (3.0)
Truncus arteriosus, valvular defect, atrioventricular canal (n = 4 each)	4 (2.4)
Single ventricle, cardiomyopathy, hypoplastic aortic arch (n = 3 each)	3 (1.8)
Patent foramen ovale, tricuspid atresia, interrupted aortic arch (n = 2 each)	2 (1.2)
Ebstein anomaly, heart block, dextrocardia, inferior vena cava thrombus, cardiac failure, hypoplastic right ventricle, double-inlet right ventricle, total anomalous left coronary artery (n = 1 each)	1 (0.6)
Surgical procedures (n = 102 procedures)
Atrial or ventricular septal defect repair	24 (23.5)
Norwood/Sano Norwood procedure	22 (21.6)
Patent ductus arteriosus ligation	7 (6.9)
Glenn, TAPVR repair, Blalock-Taussig shunt (n = 6 each)	6 (5.9)
Coarctation of the aorta repair	4 (3.9)
Atrial switch repair, tetralogy of Fallot repair, pulmonary valve repair (n = 3 each)	3 (2.9)
Atrioventricular canal defect repair, Ebstein anomaly repair, delayed sternal closure, heart transplant (n = 2 each)	2 (2)
ALCAPA repair, aortic valve repair, diaphragm plication, TGA repair, mitral valve repair, closure of patent foramen ovale, ventricular outflow tract right obstruction repair, repair of truncus arteriosus, Bentall procedure, thoracic duct ligation (n = 1 each)	1 (1)

Abbreviations: ALCAPA, anomalous left coronary artery from the pulmonary artery; TAPVR, total anomalous pulmonary venous return; TGA, transposition of the great arteries.
^aUnless otherwise indicated.

Table 2. Organisms isolated
Organism	No. (%) of isolates
Blood culture (n = 112)
Enterococcus faecalis	22 (19.6)
Staphylococcus epidermidis	18 (16.1)
Staphylococcus aureus	12 (10.7)
Klebsiella oxytoca	9 (8.0)
Klebsiella pneumoniae	8 (7.1)
Enterobacter cloacae complex	8 (7.1)
Pseudomonas aeruginosa	7 (6.3)
Candida species^a	5 (4.5)
Enterobacter aerogenes	5 (4.5)
Serratia species^b	4 (3.6)
Other^c	14 (12.5)
Urine culture^d (n = 18)
Enterococcus faecalis	4 (22)
Enterobacter cloacae	4 (22)
Escherichia coli	2 (11)
Staphylococcus epidermidis	1 (6)
Klebsiella species	1 (6)
Pseudomonas species	1 (6)
Not specified	5 (28)
Respiratory culture^d (n = 23)
Pseudomonas species	7 (30)
Klebsiella species	4 (17)
Serratia species	4 (17)
Staphylococcus aureus	3 (13)
Stenotrophomonas species	2 (9)
Escherichia coli	2 (9)
Fungi	1 (4)
Surgical site^e (n = 3)
Methicillin-resistant Staphylococcus aureus	1 (33)
Stenotrophomonas species	1 (33)
Staphylococcus aureus	1 (33)

^a Candida krusei, Candida parapsilosis, Candida tropicalis.
^b Serratia liquefaciens, Serratia marcescens.
^c Rhodotorula species, Citrobacter freundii, Escherichia coli, Acinetobacter baumannii, Pantoea species, Elizabethkingia meningoseptica, Streptococcus oralis, Clostridium tertium, Enterococcus faecium, Staphylococcus hominis, coagulase-negative Staphylococcus species (n = 2), Bacillus species (n = 2).
^dOrganisms isolated from sources other than bloodstream infections were different from organisms associated with bloodstream infections and were isolated during the infection window period.
^eIn sternal/mediastinal locations and developed before the bloodstream infections.

Table 3. Laboratory data during the infection window period ^a
Laboratory test	No. of results	Mean (SD)	Range
White blood cell count, x1000/μL	515	13.46 (8.0)	1.1–68.7
Absolute neutrophil count,/μL	497	9415.73 (6728.1)	258.5–50 340.0
Absolute lymphocyte count,/μL	497	2299.63 (1619.5)	0.0–9159.4
C-reactive protein, mg/dL	389	7.77 (7.4)	0.5–45.0
Prothrombin time, s	206	18.11 (10.8)	10.6–77.5
International normalized ratio	165	6.85 (8.3)	1.0–41.6
Blood glucose, mg/dL	571	106.84 (45.3)	36.0–472.0

SI conversion: to convert blood glucose to mmol/L, multiply by 0.0555.
^aThe infection window period is defined as the 7-day window during which the infection occurred, including the date the positive blood culture (or other diagnostic test) result was identified, the 3 calendar days before this date, and the 3 calendar days after this date.