Risk factors and outcomes of pediatric non-invasive respiratory support failure in Latin America

Abstract

Background: Noninvasive respiratory support (NRS) is standard in pediatric intensive care units (PICUs) for respiratory diseases, but its failure can lead to complications requiring invasive mechanical ventilation (IMV). This study aimed to identify risk factors for NRS failure in children with acute respiratory failure (ARF) in PICUs, and compare complications and outcomes between IMV-only and NRS failure patients.

Methods: We conducted a cohort study using data from the LARed Network prospective registry (April 2017-November 2022), in children under 18 years admitted to PICUs for ARF. Cases were divided into subgroups: those managed with IMV only, those who experienced NRS failure requiring IMV, those who received NRS successfully, and those who did not require NRS or IMV. Exclusions included patients with home respiratory support prior to admission, patients without PICU discharge at the cutoff date of the analysis and those with incomplete data. Multivariate mixed models analyzed NRS failure risk factors, and complications between the IMV-only and NRS failure groups, using centers as a random effect.

Results: A total of 7374 children met the inclusion criteria, with 6208 in the NRS group and 1166 in the IMV-only group. The NRS success rate was 85.3 %. Risk factors for NRS failure included age (median of 4.6 months, interquartile range of 2.1-14.2 months), history of prematurity (adjusted odds ratio [aOR]=1.53, 95 % confidence interval [CI]: 1.20 to 1.95) or malnutrition (aOR=1.85, 95 % CI: 1.18 to 2.91), suspected bacterial infection (aOR=5.12, 95 % CI: 4.05to 6.49), FiO₂ >30 % (aOR=1.52, 95 % CI: 1.18 to 1.97), severe hypoxemia with SpO₂/FiO₂ ≤150 (aOR=1.85, 95 % CI: 1.48 to 2.30), tachypnea (aOR=1.42, 95 % CI: 1.18 to 1.72), tachycardia (aOR=1.77, 95 % CI: 1.47 to 2.12), and lung consolidations (aOR=1.45, 95 % CI: 1.14 to 1.85) or interstitial infiltrates (aOR=1.29, 95 % CI: 1.05 to 1.58) on chest X-ray. There were no significant differences in morbidity, mortality, duration of IMV, or PICU length of stay between patients who received IMV only and those who experienced NRS failure. However, patients who experienced NRS failure were more likely to develop withdrawal symptoms related to sedative or opioid discontinuation and/or delirium (aOR=2.57, 95 % CI: 1.85 to 2.57).

Conclusion: This study identified key risk factors for predicting NRS failure in children with acute ARF in PICUs, including younger age, prematurity, malnutrition, suspected bacterial infection, FiO₂ >30 %, severe hypoxemia (SpO₂/FiO₂ ≤150), tachypnea, tachycardia, and radiological findings such as lung consolidation and interstitial infiltrates. Compared to patients managed with IMV from the start, those who experienced NRS failure were more likely to develop withdrawal symptoms and/or delirium, although clinical outcomes such as mortality, IMV duration, and PICU length of stay were similar in both groups.

Keywords: Continuous positive airway pressure; Noninvasive ventilation; Oxygen therapy; Pediatric intensive care unit; Pediatrics.

Figure 1

Flowchart of patients enrolled in the study. IMV, Invasive ventilation; NRS, Non-invasive respiratory support; PICU, Pediatric intensive care unit.

Figure 2

Patient trajectories and respiratory support interventions in the intensive care unit. This Sankey diagram illustrates the treatment pathways of 8664 patients admitted to the intensive care unit, showing the types and transitions of respiratory support provided at three critical stages. Upon arrival, the diagram indicates the type of respiratory support each patient received as they entered the ICU. The “Initial” stage represents the type of support determined to be necessary after the initial assessment in the unit. The “Final” stage shows the most intensive level of support utilized during the patient's stay. The widths of the flow paths reflect the proportions of patients transitioning between different support types, and absolute numbers at each node provide a clear count of patients in each category. The key identifies NRS, IMV, and cases where No RS was provided. IMV, Invasive mechanical ventilation; No RS, No respiratory support; NRS, Non-invasive respiratory support.

Figure 3

Kaplan–Meier graph with distribution of PICU length of stay (days) according to the four study groups. This graph illustrates the PICU length of stay distributed among four different groups: NRS success, NRS failure, IMV only, and No RS. The log-rank test revealed significant differences in length of stay between these groups (χ²(3)=309.4, P <0.0001). IMV, Invasive mechanical ventilation; No RS, No respiratory support; NRS, Non-invasive respiratory support; PICU, Pediatric intensive care unit.