BACKGROUND: Prone position improves oxygenation in some patients with ARDS. According to some authors, prone position can also improve the deteriorated hemodynamics induced by PEEP. But these respiratory and hemodynamic effects of prone position has not yet been fully established. METHODS: Twentythree consequtive patients with ARDS(M : F = 11:12, 62.1+/-20.8yrs) were the subjects for this study. ABGA, static compliance of the respiratory system, mean arterial pressure and pulse rate were obtained in supine position and at 5mm, 0.5h and 2h of prone position. Positive respiratory response was defined as 20mmHg or more increase in PaO2/FIO2 within 2h of prone position. Early of late respiratory responses were defined if the positive response was observed within of after 3 day of ARDS onset, respectively. Positive hemodynamic response was defined as l0mmHg or more increase in mean arterial pressure at 5mm of prone position. RESULTS: Fifteen patients (65%) showed positive respiratory response. In the respiratory responders, PaO2 was 69.8+/-17.6mmHg in supine position, 83.2+/-22.6mmHg in prone position 0.5h, 96.8+/-22.7mmHg in prone position 2h(p<0.001), and PaO2/FIO2 was 108+/-41mmHg, 137+/-57mmHg, 158+/-50mmHg, respectively(p=0.001). Age, sex, cause of ARDS, supine PaO2, and PaO2/FIO2 were not different between the respiratory responders and the nonresponders. The respiratory responders, however, showed higher mean arterial pressure than the nonresponders(91.1+/-13.1mmHg vs. 76.0+/-18.7mmHg, p=0.035), and tendency of higher survival rate(9/15 vs. 2/8, p= 0.074). Static compliance of the respiratory system was decreased in prone position 0.5h(28.4+/-7.9ml/cm H2O vs. 23.8+/-7.6ml/cm H2O, p=0.007). The overall rate of early response(n=23) and late response(n=11) were similar(14/23 vs. 7/11, p>0.05). But patient without early response showed late response only in 25% (1/4), while patient with early response showed late response in 85.7% (6/7)(p=0.072). Five patients(22%) showed positive hemodynamic response, two of them being respiratory nonresponders. There were no differences in the baseline mean arterial pressure or the level of PEEP applied in supine between the hemodynamic responders and the hemodynamic nonresponders. CONCLUSIONS: Prone position either improved oxygenation or increased arterial pressure in significant proportion of patients with ARDS. And the respiratory response to prone position was thought to be determined in the early stage of ARDS.