[What are the most reliable predictive factors of non-invasive ventilation failure in paediatric intensive care units?].

INTRODUCTION: Despite there being limited evidence, non-invasive ventilation (NIV) has become a common treatment for acute respiratory failure (ARF). The aim of this study was to identify the predictive factors of NIV failure, in order to enable early detection of patients failing the treatment.

PATIENTS AND METHODS: Prospective cohort study was conducted that included all ARF patients that received NIV as the initial treatment between 2005 and 2009 in a fourteen-bed Paediatric Intensive Care Unit (PICU) of a tertiary university hospital. Information was collected about the NIV, as well as clinical data prior to NIV, at 2, 8, 12, and 24hrs. The haemoglobin saturation (SpO2)/fraction of inspired oxygen (FiO2) ratio (S/F) was retrospectively calculated. NIV failure was defined as the need for intubation or requiring rescue with bi-level pressure (BLPAP). Univariate and multivariate statistical analyses were performed.

RESULTS: A total of 282 patients received non-invasive support, with 71 receiving Continuous Pressure (CPAP), and 211 with BLPAP treatment. The overall success rate was 71%. Patients receiving BLPAP vs. CPAP, patients with higher S/F ratios at 2hours (odds ratio [OR] 0.991, 95% CI 0.986-0.996, P=.001], and patients older than 6 months (Hazard ratio [HZ] 0.375, 95% CI 0.171-0.820, P=.014], were also more likely to fail. Patients with higher heart rates (HR) at 2hours (OR 1.021, 95% CI [1.008-1.034], P=.001) and higher inspiratory positive airway pressure (IPAP) at 2hours were more prone to failure (HZ 1.214, 95% CI [1.046-1.408], P=.011).

CONCLUSIONS: Age below 6 months, S/F ratio, HR, and IPAP at 2hours are independent predictive factors for initial NIV failure in paediatric patients with ARF admitted to the PICU.