INTRODUCTION
Acute viral bronchiolitis (AVB) is a leading cause of lower respiratory tract infection and hospitalization in young infants, notably those aged less than 1 year [1]. Management of moderate-to-severe AVB is based on respiratory support, with as the first step administration of a heated and humidified mixture of air and oxygen with high‐flow nasal cannula (HFNC) [2]. While several studies have confirmed HFNC effectiveness to reduce the work of breathing in AVB [3-5], failure occurs in 30-40% of these patients [6-8]. In these infants, evolving respiratory failure requires escalation in therapeutic measures, including transition to continuous positive airway pressure (CPAP), then if necessary to noninvasive ventilation or intubation [9].
Early identification of patients who are most likely to fail with HFNC is critical for care organization in the pediatric emergency department [10]. Indeed, these infants will need to be referred to a pediatric intensive care unit, necessary for the monitoring of any non-invasive ventilation technique including CPAP and bilevel positive airway pressure, while those who will improve with HFNC can potentially be transferred to a general pediatric ward [11]. Currently, some patient characteristics have been individualized as predictors of respiratory deterioration, notably younger age or initial severity [12, 13]. However, neither isolated physiologic parameters, such as respiratory rate (RR), fraction of inspired oxygen (FiO2), or venous/capillary partial pressure of carbon dioxide (pvCO2), nor clinical scales that incorporate different vital signs have demonstrated a consistent association with the risk of HFNC failure and are discriminating enough to be used as triage tools [14-18].
Recently, Roca et al. developed a tool to assess the risk for HFNC failure in adult patients with hypoxemic acute respiratory failure [19]. ROX (Respiratory rate-OXygenation) index corresponds to the ratio of patient oxygenation, which has been associated to HFNC success, over RR, which has been associated to HFNC failure. Subsequent studies confirmed it was a good predictor of HFNC failure in lower respiratory tract infections, including those caused by virus [20, 21]. In pediatric patients with acute respiratory failure, ROX index application 24 and 48 hours after hospital admission also appeared a good marker for predicting the risk of HFNC failure [22]. In the specific context of AVB in <2 years infants managed in a pediatric emergency department, patients with ROX index in the lowest quartile at HFNC initiation were three times more likely to require CPAP compared to those in the highest quartile [23]. The single-center nature of this study, with lack of standardization for failure criteria, incited to test the relevance of ROX index in an homogeneous population of <6 months infants requiring HFNC for severe AVB, recruited in the framework of a multicenter study with predefined HFNC failure criteria [24].
The primary objective of this study was to assess the relationship between ROX index collected early (i.e. before HFNC initiation and 1 hour after), and HFNC failure occurring in the following 48 hours in patients admitted for severe AVB.