Airway obstruction in patients with sleep apnea-hypopnea syndrome (SAHS) is due to increased critical pressure (Pcrit) of the upper airway. The ideal nasal pressure (Pn) to maintain airway patency should consist of the constant term to account for Pcrit and a term (Rn . V) proportional to flow (V) to account for the dynamic pressure drop through nasal resistance (R n). Continuous positive airway pressure (CPAP) applied to avoid flow limitation results in a Pn greater than required over most of the breathing cycle. The aim was to assess a flow-dependent positive airway pressure (FDPAP) based on adapting Pn to the instantaneous flow: Pn = P0 + k . V. FDPAP was tested on collapsible airway models and its applicability was assessed in nine patients with SAHS during sleep. In models, FDPAP prevented flow limitation with lower mean P n and work of breathing than CPAP. In patients FDPAP allowed the patients to breathe normally with a mean Pn (6.6 +/- 1.2 cm H2O) systematically and significantly (p < 0.05, paired t test) lower than when applying CPAP (9.1 +/- 1.2 cm H2O). The results found in models and in patients suggest that adapting the applied nasal pressure to the instantaneous breathing flow may be of potential practical interest in SAHS.