Abstract

Objectives: To determine factors associated with health-related quality of life (HRQL) decline among pediatric acute respiratory distress syndrome (PARDS) survivors. Design: Retrospective cohort study. Setting: Academic children’s hospital. Patients: Three hundred fifteen children 1 month to 18 years old with an unplanned PICU admission from December 2011 to February 2017 enrolled in the hospital’s Outcomes Assessment Program. Interventions: None. Measurements and Main Results: Pre-admission baseline and median 6-week post-discharge HRQL were assessed using the Pediatric Quality of Life Inventory or the Functional Status II-R. Patients meeting retrospectively applied Second Pediatric Acute Lung Injury Consensus Conference criteria for PARDS were identified, and PARDS severity was classified using binary (mild/moderate, severe) and trichotomous (mild, moderate, severe) categorization for noninvasive ventilation and invasive mechanical ventilation (IMV). PARDS occurred in 41 of 315 children (13.0%). Clinically important HRQL decline (≥ 4.5 points) occurred in 17 of 41 patients (41.5%) with PARDS and 64 of 274 without PARDS (23.4%). On multivariable generalized linear regression adjusted for age, baseline Pediatric Overall Performance Category, maximum nonrespiratory Pediatric Logistic Organ Dysfunction score, diagnosis, length of stay, and time to follow-up, PARDS was associated with HRQL decline (adjusted relative risk [aRR], 1.70; 95% CI, 1.03–2.77). Four-hour and maximum PARDS severity were the only factors associated with HRQL decline. HRQL decline occurred in five of 18 patients with mild PARDS at 4 hours, five of 13 with moderate PARDS (aRR 2.35 vs. no PARDS [95% CI, 1.01–5.50]), and seven of ten with severe PARDS (aRR 2.56 vs. no PARDS [95% CI, 1.45–4.53]). The area under the receiver operating characteristic curve for discrimination of HRQL decline for IMV patients was 0.79 (95% CI, 0.66–0.91) for binary and 0.80 (95% CI, 0.69–0.93) for trichotomous severity categorization. Conclusions: HRQL decline is common among children surviving PARDS, and risk of decline is associated with PARDS severity. HRQL decline from baseline may be an efficient and clinically meaningful endpoint to incorporate into PARDS clinical trials.