Fluid Overload and Cumulative Thoracostomy Output Are Associated With Surgical Site Infection After Pediatric Cardiothoracic Surgery

Objectives: To determine the impact of cumulative, postoperative thoracostomy output, amount of bolus IV fluids and peak fluid overload on the incidence and odds of developing a deep surgical site infection following pediatric cardiothoracic surgery. Design: A single-center, nested, retrospective, matched case-control study. Setting: A 26-bed cardiac ICU in a 303-bed tertiary care pediatric hospital. Patients: Cases with deep surgical site infection following cardiothoracic surgery were identified retrospectively from January 2010 through December 2013 and individually matched to controls at a ratio of 1: 2 by age, gender, Risk Adjustment for Congenital Heart Surgery score, Society of Thoracic Surgeons-European Association for Cardiothoracic Surgery category, primary cardiac diagnosis, and procedure. Interventions: None. Measurements and Main Results: Twelve cases with deep surgical site infection were identified and matched to 24 controls without detectable differences in perioperative clinical characteristics. Deep surgical site infection cases had larger thoracostomy output and bolus IV fluid volumes at 6, 24, and 48 hours postoperatively compared with controls. For every 1 mL/kg of thoracostomy output, the odds of developing a deep surgical site infection increase by 13%. By receiver operative characteristic curve analysis, a cutoff of 49 mL/kg of thoracostomy output at 48 hours best discriminates the development of deep surgical site infection (sensitivity 83%, specificity 83%). Peak fluid overload was greater in cases than matched controls (12.5% vs 6%; p < 0.01). On receiver operative characteristic curve analysis, a threshold value of 10% peak fluid overload was observed to identify deep surgical site infection (sensitivity 67%, specificity 79%). Conditional logistic regression of peak fluid overload greater than 10% on the development of deep surgical site infection yielded an odds ratio of 9.4 (95% CI, 2-46.2). Conclusions: Increased postoperative peak fluid overload and cumulative thoracostomy output were associated with deep surgical site infection after pediatric cardiothoracic surgery. We suspect the observed increased thoracostomy output, fluid overload, and IV fluid boluses may have altered antimicrobial prophylaxis. Although analysis of additional pharmacokinetic data is warranted, providers may consider modification of antimicrobial prophylaxis dosing or alterations in fluid management and diuresis in response to assessment of peak fluid overload and fluid volume shifts in the immediate postoperative period.