Non-invasive ventilation in bronchiolitis: Analysis according to a chronologic classification

OBJECTIVE: To study patients diagnosed with bronchiolitis receiving ventilatory support with non-invasive ventilation (NIV) according to a chronologic classification: initial support (i-NIV), rescue post-extubation (r-NIV) and elective post-extubation (e-NIV); and to identify predictive factors of failure for each group. Prospective observational study (January 2004-December 2009), including all the patients with bronchiolitis admitted to pediatric intensive care unit and receiving ventilatory support with NIV. Clinical data collected at 0 (pre-NIV), 1, 2, 8, 12, 24 h of treatment were analyzed. Need for intubation was considered as NIV failure. NIV was successful in 65.8% of 152 cases included. Success rates were as follows: i-NIV (52.2%); r-NIV (72.2%); and e-NIV (90.9%) (Anova P < 0.000). Bi-level modes had higher efficacy (73%) than continuous positive airway pressure (61.5%) (Fisher's-test P = 0.049). Predictive factors of success in i-NIV group were inspired fraction of oxygen (FiO(2)) at 2 h, P = 0.003, higher pulse oximeter saturation (SpO(2))/FiO(2) values at 2 h (P = 0.009), and SpO(2)/FiO(2) - 12 h (P = 0.05), lower heart rate (HR) at 12 h (P = 0.01), lower partial pressure of carbon dioxide (PCO2) previous to NIV (P = 0.009) and HR decrease - 12 h (P = 0.008), In e-NIV: respiratory rate (RR)-1 h (P = 0.02), RR decrease-1 h (P = 0.006) and higher SpO(2)/FiO(2) - 24 h (P = 0.01); in r-NIV: SpO(2)/FiO(2) - 12 h (P = 0.04), lower HR-2 h (P = 0.03) and HR-8 h (P = 0.01). Multivariate analysis identified the groups as an independent variable (P = 0.04) but didn't show any significant value in any of the analyzed groups. Separate evaluation is advisable for i-NIV, r-NIV and e-NIV.