Objective: There were two objectives of the submitted investigation: 1. To demonstrate the effect of high frequency oscillation on the gas exchange in children with severe respiratory failure (ARDS) where conventional mechanical ventilation had failed. 2. To attempt identification of patients where failure of high frequency oscillation is very probable and thus there is a great risk of death. Type of study: Retrospective analysis of a group. Site: ARO and ECMO Centre of the Children™s Hospital Brno. Patients: 26 patients above 1 month of age with severe hypoxaemic respiratory failure and ARDS. Mean age 3.66 years (three adults - 17, 19 and 24 years), maximum body weight 70 kg, 17 boys, 9 girls. Fifteen patients died, 84.6% patients suffered from some internal disease, mean PRISM score on admission 22. Nine patients met the ECMO criteria, in two ECMO was not necessary, in another five patients ECMO was contraindicated, two patients were connected with ECMO. To four patients during HFO iNO was administered, in two moreover partial liquid ventilation was used.Method: The patients were connected to a PCV or PRVC regime with a limited peak pressure and permissive hypercapnia. If PaCO2 > 10.0 kPa and/or pH < 7.20 was recorded, continuous tracheal gas insufflation was used. When FiO2 > 0.6 and Paw > 15 cm H2O for peripheral blood saturation to a minimum of 90% had to be used or in case of persisting hypercapnia and/or acidosis in CMV with TGI the patients were switched to HFO. The HFO strategy was —High Volume Strategyii - recruitment of alveoli and achievement of the optimal pulmonary volume. For switching back to CMV the patient had to meet the following criteria: Paw 15 - 20 cm H2O, FiO2 < 0.6 without an air leak and/or improvement of the X-ray finding without desaturation during aspiration of the airways. The patient was considered detached if he met the following criteria: saturation > 90%, FiO2 < 0.4, Paw < 15 cm H2O, normal pH at a respiration rate (RR) < 30/min. and PIP < 35 cm H2O. In addition to demographic data (sex, age, body weight, PRISM score on admission) and period of ventilation (CMV before HFO, HFO, CMV after HFO and total ventilation period) the following were assessed: pH, PaO2, PaCO2, AaDO2, oxygenation index and hypoxaemic score (PaO2/FiO2). Results: Within several hours the oxygenation improved, CO2 was eliminated and the ventilation-perfusion state returned to normal in all patients. This improvement was permanent. Between surviving and not surviving patients there are significant differences in the investigated parameters, moreover there were marked differences in the CMV before HFO. 42% patients survived. If the PaO2/FiO2 increases by 55% during the sixth hour of HFO it may be assumed with a 83% sensitivity and 92% specificity that the patient will survive. When OI declines by 30% during the 6th hour of HFO the prediction of survival is with a 86% sensitivity and 82% specificity. Conclusion: HFO was effective in all patients where CMV failed, CO2 elimination and oxygenation improved, 42% patients survived. It is possible that early HFO administration can reduce the mortality of acute hypoxaemic respiratory failure in children.

        Key words: high frequency oscillation (HFO), respiratory failure, children, conventional mechanical ventilation