Health care-associated infections in children after cardiac surgery.

Few recent studies have assessed the epidemiology of health care-associated infections (HAIs) in the pediatric population after cardiac surgery. A retrospective cohort study was performed to assess the epidemiology of several types of HAIs in children 18 years of age or younger undergoing cardiac surgery from July 2010 to June 2012. Potential pre-, intra-, and postoperative risk factors, including adherence to the perioperative antibiotic prophylaxis regimen at the authors' hospital, were assessed by multivariable analysis using Poisson regression models. Microorganisms associated with HAIs and their susceptibility patterns were described. Overall, 634 surgeries were performed, 38 (6 %) of which were complicated by an HAI occurring within 90 days after surgery. The HAIs included 7 central line-associated bloodstream infections (CLABSIs), 12 non-CLABSI bacteremias, 6 episodes of early postoperative infective endocarditis (IE), 9 surgical-site infections (SSIs), and 4 ventilator-associated pneumonias (VAPs). Mechanical ventilation (rate ratio [RR] 1.07 per day; 95 % confidence interval [CI] 1.03-1.11; p = 0.0002), postoperative transfusion of blood products (RR 3.12; 95 %, CI 1.38-7.06; p = 0.0062), postoperative steroid use (RR 3.32; 95 % CI 1.56-7.02; p = 0.0018), and continuation of antibiotic prophylaxis longer than 48 h after surgery (RR 2.56; 95 % CI 1.31-5.03; p = 0.0062) were associated with HAIs. Overall, 66.7 % of the pathogens associated with SSIs were susceptible to cefazolin, the perioperative antibiotic prophylaxis used by the authors' hospital. In conclusion, HAIs occurred after 6 % of cardiac surgeries. Bacteremia and CLABSI were the most common. This study identified several potentially modifiable risk factors that suggest interventions. Further studies should assess the role of improving adherence to perioperative antibiotic prophylaxis, the age of transfused red blood cells, and evidence-based guidelines for postoperative steroids.