PC3095: Treatments and Outcomes of Pediatric Acute Respiratory Distress Syndrome Patients with Bronchiolitis versus Other Lower Respiratory Tract Infections

Background: Bronchiolitis commonly causes mechanical ventilation in the pediatric intensive care unit (PICU). While bronchiolitis is predominately an obstructive airways disease, patients diagnosed with bronchiolitis may also meet criteria for pediatric acute respiratory distress syndrome (PARDS) including hypoxemia and radiographic infiltrates. The epidemiology and outcomes of children with clinical diagnoses of bronchiolitis meeting PARDS criteria are poorly described. This study aims to compare patient characteristics, treatments, and outcomes of PARDS patients with bronchiolitis versus other lower respiratory tract infections (LRTI).

Methods: This is a secondary analysis of the Pediatric Acute Respiratory Distress Syndrome Incidence and Epidemiology (PARDIE) study, a prospective, international PICU point-prevalence study performed between May 2016 - June 2017. With waiver of consent, variables collected at 116 sites included demographics, illness severity scores (Pediatric Index of Mortality [PIM3] and Pediatric Risk of Mortality [PRISM] on admission; and Pediatric Logistic Organ Dysfunction [PELOD-2] on PARDS days 0-3), use of adjunctive therapies on PARDS days 0-3, and PARDS risk factor. Bronchiolitis patients were identified using PIM3 or PARDS risk factor. All other children with “pneumonia” as the PARDS risk factor were included in the “lower respiratory tract infection” (LRTI) group. Children with other risk factors (e.g. sepsis) were excluded. Patient outcomes (PICU mortality, 28-day ventilator free days [VFDs] and PICU-free days [PFDs]) were compared with Chi square tests and Wilcoxon rank-sums. Statistical significance was defined as p< 0.05, data shown as median (IQR).

Results: Out of the 708 PARDS patients, 452 were included because they had clinically diagnosed bronchiolitis or a PARDS trigger of pneumonia. Patients with bronchiolitis (n=113) were younger (0.3 [0.1-0.8] vs 5.3 [1.4-9.5] years, p< 0.01) and more commonly had no comorbidities (54.9% vs 30.4%, p< 0.01). At PARDS onset, LRTI patients had a greater severity of illness at PARDS onset by PIM3 (-3.4 [-4.6, -2.7] vs -5.5 [-6.1, -5.0], p< 0.01) and PRISM (6.0 [2.0, 10.0] vs 4.0 [1.0, 8.0], p=0.02), but there were no significant differences between groups in oxygenation index or PELOD-2 scores on days 0-3 of PARDS.

Groups did not differ significantly in use of inhaled nitric oxide, prone positioning, corticosteroids, neuromuscular blockade, high frequency oscillator ventilation, or extracorporeal membrane oxygenation on PARDS days 0-3. LRTI patients had higher PICU mortality (13.6% vs 5.9%, p=0.02), but there was no significant difference in VFDs or PFDs between group. When restricting analysis to children without co-morbidities (n=196), PICU mortality was similar (8.7% vs 3.2%, p=0.17).

Conclusion: PARDS patients with bronchiolitis have lower severity of illness at onset compared to patients with other LRTIs, but similar illness severity and use of adjunctive therapy on subsequent days. While unadjusted mortality was lower for children with bronchiolitis, this is likely explained by fewer underlying comorbidities in children with bronchiolitis and PARDS.