The impact of humidified high nasal flow oxygen in different groups of intensive care patients with respiratory failure

Methods We retrospectively studied respiratory -, oxygen derivedand hemodynamic parameters before and one hour after start of HFNC in 116 patients during 2010 and 2011. All patients were treated in a mixed medical, surgical, neurological ICU of a teaching hospital. We compared non-invasive oxygen therapy like venturi mask, non-rebreathing mask and non-invasive positive pressure ventilation with HFNC. The HFNC used an air-oxygen blender with adjustable Fi02 (0.21-1.0), delivering a modifiable gas flow of 40 l/min (Optiflow, Fisher&Paykel, Auckland, New Zeeland) in combination with humidification. Data from MetavisionPDMS and Mediscore. We used oneway ANOVA to compare hemodynamic and oxygen related variables before and one hour after start of HFNC. Results 116 patients were included, 66 men and 50 women, mean age 66, SD 14, mean APACHE 4 on admission 64, SD 21. Mean duration of HFNC was 24, SD 31 hours. Indications for HFNC could be divided in 6 categories; 1. Hypoxia: n = 41, 2. Weaning from NPPV: n = 25, 3. Comfort: n = 18, 4. No acceptance of non-invasive positive pressure ventilation (NPPV): n = 17, 5. Respiratory distress/comfort: n = 9, 6. Other: n = 6. In 21 patients we measured arterial blood gases (ABG). The oxygenation was significant better with HFNC, Pa02 (p = 0,019) and Pa02/Fi02 (p = 0,002). The PaC02 was significant lower (p = 0,048) with HFNC. In 116 patients the peripheral oxygen saturation (Sp02)/ fraction of inspired oxygen (Fi02) ratio was significant better (p = 0,000) with HFNC, the respiratory rate was significant lower (p = 0,000) with HFNC. Also the hemodynamic variables heart rate and mean arterial pressure were significant better with HFNC. Despite the use of HFNC, in 35 patients (30%) intubation was unavoidable and 29 patients (25%) died.