Resident Physician Boston Children's Hospital / Boston Medical Center, United States
Abstract:
Introduction: In adults, early tracheostomy is associated with reduced mortality, shorter ICU stays, and less sedation use. In pediatric cardiac patients with a tracheostomy, optimal timing and factors predicting mortality are unclear.
Objectives: To characterize epidemiology of patients requiring tracheostomy in the pediatric Cardiac Intensive Care Unit (CICU) and identify associations with mortality.
Methods: This was a retrospective single-center cohort study of Boston Children’s Hospital CICU admissions with tracheostomy placement (7.1.2011 to 7.1.2020). Patients were compared based on duration of total respiratory support (invasive and non-invasive ventilation) prior to tracheostomy using two definitions of prolonged mechanical ventilation (PMV): >3 weeks (literature definition) and >7 weeks (fitting a receiver operating characteristic curve for mortality). The primary outcome was survival at current follow up (7.1.2022). Patient variables were gathered from the medical record. Risk factors for mortality were examined using cox proportional hazard regression.
Results: 61 patients had a tracheostomy (0.5% incidence of admissions). Incidence increased in the second half of the study period (p < 0.01). Median duration of respiratory support pre-tracheostomy was 7.0 weeks (IQR 3.1, 11.4) and 18.0% had < 3 weeks of PMV pre-tracheostomy. Patients with upper airway obstruction (UAO) as the indication for tracheostomy were more likely to have < 3 weeks of PMV pre-tracheostomy. Single ventricle physiology (SV) and age < 1 year were associated with >7 weeks PMV, whereas patients with genetic diagnoses had shorter ventilation pre-tracheostomy. No statistical difference in length of stay or sedation needs were observed post tracheostomy according to PMV [Table]. Overall mortality was 59.0% (median follow up 7.8 months, IQR 2.6, 30.0) with a statistical difference in survival for PMV >7 weeks (vs < 7 weeks, Hazard Ratio [HR] 2.3; 95%CI 1.2, 4.5, logrank p=0.0133, [Figure]). Bivariate models revealed Pulmonary Vein Stenosis (PVS) to be associated with mortality whereas genetic disorders and UAO were associated with improved survival. Using forward stepwise selection of these variables, PMV >7 weeks remained independently associated with mortality (adjusted HR 4.2; 95%CI 1.9, 9.0) as were PVS (aHR 5.2; 95%CI 2.5, 10.8) and UAO (aHR 0.17; 95%CI 0.1, 0.5).
Conclusions: Patients with SV or age < 1 year were more likely to have PMV pre-tracheostomy, whereas patients with genetic diagnoses or UAO had shorter ventilation duration. Mortality in CICU patients undergoing tracheostomy is high. CICU patients with >7 weeks of ventilation pre-tracheostomy have increased mortality risk.
