Objective: There were two primary outcome measures of the study: 1. To demonstrate the effect of high frequency oscillation on gas exchange in children with severe respiratory failure (ARDS) in whom conventional ventilation failed. 2. To try to identify patients in whom high frequency oscillation is highly suspicious to fail, bearing a high risk of subsequent death. Study design: Retrospective analysis of the set of patients. Setting: Department of Anesthesiology and Intensive Care; ECMO Center University Children Hospital Brno. Patients: 26 patients > 1 month of age with severe hypoxemic respiratory failure and ARDS. Mean age was 3.66 years (three adult patients O 17, 19 and 24 years), maximum weight 70 kilograms, 17 boys, 9 girls. 15 patients died. 84.6% per cent of patients suffered from basic medical disease, mean admission PRISM score was 22. Nine patients met ECMO criteria. In two of them ECMO was not necessary, in 5 ECMO was contraindicated, two patients were treated with ECMO. Four patients ventilated in HFO mode were also treated with NO, in two of them partial liquid ventilation was also implemented. Methods: Patients were ventilated in PCV or PRVC modes with limited peak inspiratory pressures and permissive hypercapnia. Tracheal gas insufflation was instituted if paCO2 > 10,0 kPa and/or pH < 7.20. HFO was started if there was a need for FiO2 > 0.6 and Paw > 15 cmH2O to maintain peripheral blood saturation > 90% or due to persistent hypercapnia and/or acidosis with CMV with TGI. HFO was designed as —High Volume Strategyii to recruit the alveoli and keep the optimal lung volume. Before switching back to CMV, following criteria had to be met: Paw 15O20 cmH2O, FiO2 < 0.6, no air-leak and/or better chest X-ray, no drop in saturation during airway toilette. P atient was considered to be disconnected when following criteria were met: saturation > 90%, FiO2 < 0.4 Paw < 15 cmH2O, normal pH with respiratory rate (RR) < 30/min and PIP < 35 cmH2O. Except demographic data (sex, age, weight, admission PRISM score) and duration of ventilation (CMV before HFO, HFO, CMV after HFO and total duration of ventilation) following data were observed: pH, PAO2,, paCO2, AaDO2, oxygenation index and hypoxemia score (paO2/FiO2). Results: Within several hours, there could be observed an improvement in oxygenation, CO2 elimination and ventilation/perfusion ratio in all patients. This improvement was persistent. There were statistically significant differences in followed parameters between survivors and non-survivors. There was also significant difference between survivors and non-survivors in the duration of CMV before HFO was instituted; 42% of patients survived. Survival of a patient could be predicted if paO2/FiO2 increased by 55% in 6 th hour of HFO (sensitivity 83%, specificity 92%). If oxygenation index decreased by 30% in 6 th hour of HFO, survival could be predicted with 86% sensitivity and with 82% specificity. Conclusion: HFO was effective in all patients in whom CMV failed. HFO significantly improved CO2 elimination and oxygenation; 42 per cent of patients survived. It seems that early institution of HFO may decrease mortality of acute hypoxemic respiratory failure in pediatric patients.

        Key words: HFO O respiratory failure O children O conventional mechanical ventilation