443 - Characteristics Associated with Death or Tracheostomy in Infants Developing BPD Following Predominant Non-Invasive Support

Poster Number: 443
Publication Number: 443.431

Heidi Morris, Children's Hospital of Philadelphia, Philadelphia, PA, United States; Nicolas A. Bamat, Childrens Hospital of Philadelphia, Philadelphia, PA, United States; Sara B. DeMauro, Perelman School of Medicine at the University of Pennsylvania, Philadelphia, PA, United States; Kathleen Gibbs, Children's Hospital of Philadelphia, Philadelphia, PA, United States; Huayan Zhang, Perelman School of Medicine at the University of Pennsylvania, Philadelphia, PA, United States; Catherine Avitabile, Perelman School of Medicine at the University of Pennsylvania, Philadelphia, PA, United States; Megan E. Reilly, Childrens Hospital of Philadelphia, Philadelphia, PA, United States; Xiaoyue Dong, Nanjing Maternity and Child Health Care Hospital, Nanjing, Jiangsu, China (People's Republic)

Background: Bronchopulmonary dysplasia (BPD) is a common and consequential complication of preterm birth. Preferential non-invasive respiratory support over mechanical ventilation has been shown to reduce the incidence of BPD and is increasingly used in clinical practice. However, inadequate respiratory support can exacerbate lung disease in infants receiving non-invasive support.

Objective: To identify characteristics associated with death or tracheostomy in preterm infants with BPD predominantly managed with non-invasive respiratory support prior to 36 weeks postmenstrual age (PMA).

Design/Methods: We performed a retrospective cohort study at Children’s Hospital of Philadelphia (CHOP). Eligible infants were of birth gestational age < 32 weeks, managed by the CHOP Newborn and Infant Chronic Lung Program between 2010-17, had BPD per the 2001 NIH consensus definition, and predominantly received non-invasive support, defined as any extubated respiratory support for > 50% of days between birth and 36 weeks PMA. We considered various clinical characteristics (Table 1) as candidate predictor variables of the primary outcome, tracheostomy or death prior to discharge. Echocardiographic evidence of pulmonary hypertension (PH) was defined as in Table 2. Characteristics associated with the outcome at a level of p < 0.10 in bivariable logistic regression were evaluated in multivariable models.

Results: Characteristics for the overall cohort (N = 134) and by outcome are displayed in Table 1.  The cohort had a median (IQR) gestational age of 26 (25-28) wks and birth weight of 765 (610-993) gms. Twenty-one (15.6%) infants died or had a tracheostomy. Only the use of pulmonary vasodilators and the severity of PH (Table 3) on echocardiogram at 36 wks PMA were associated with death or tracheostomy in bivariable analyses, while pulmonary vasodilator use remained statistically significant in adjusted multivariable models (aOR (95% CI): 7.61 (2.49, 23.26); p < 0.001) (Table 3).Conclusion(s):
We identified a strong association between PH and poor outcome among infants who develop BPD after predominant non-invasive respiratory support. Our findings emphasize the importance of screening for PH in preterm infants with BPD irrespective of respiratory support type. Early identification and management of these infants may help prevent disease progression.

Table 1. Cohort characteristics by primary outcomea Includes “Other” as denoted in electronic health record, and values with small sample sizes (Asian, n = 1; refused, n = 1)
b N = 72, represents greatest degree of missingness, excluded from consideration for multivariable modeling
c N = 74, excluded from consideration for multivariable modeling
dInclusive of inhaled nitric oxide, endothelin receptor antagonists, and enteral, IV and subcutaneous vasodilators
Table 2. Categorical assessment of pulmonary hypertension (PH) by severity on ECHORV: Right Ventricle
BP: Blood Pressure
IVS: Interventricular Septum
PDA: Patent Ductus Arteriosus