BACKGROUND Nasally applied continuous positive airways pressure (nCPAP) ventilation is an effective treatment of obstructive sleep apnoea syndrome (OSAS). As the lowest effective nCPAP-level (nCPAPmin) inter-individually is quite variable individual pressure titration under polysomnographic monitoring is necessary. OBJECTIVE To determine whether nCPAPmin can reliably be predicted on the basis of clinical and polysomnographic variables. METHODS A set of 77 unselected OSAS-patients was investigated in the sleep laboratory. We used a standard sleep-related questionnaire, examined anthropometric variables, performed lung function tests and blood gas analyses, and measured nasal cross-sectional area as well as nasal flow resistance. In the night prior to the nCPAP titration a cardiorespiratory polygraphy was performed. Stepwise multiple regression analysis using the titrated nCPAPmin as dependent variable and the clinical and polygraphic data as independent variables revealed gender (GES), neck circumference (NCF), body mass index (BMI), and apnoea-/hypopnea index (AHI) as the optimal set of predictors of nCPAPmin in this model. RESULTS Multiple regression analysis with these factors yielded the following equation: predicted nCPAPmin = 1.95 + 0.80 x GES + 0.09 x BMI + 0.01 x NCF + 0.03 x AHI (woman: GES = 1; man: GES = 2). The reliability of the equation was tested with a second set of 180 prospective OSAS-patients. In these patients the mean nCPAP-level as titrated in the sleep laboratory was 9.1 +/- 2.0 mbar, whereas the mean predicted nCPAP-level was 8.4 +/- 3.6 mbar (p = n.s.). In 51% of the patients the difference between nCPAPmin measured and predicted was greater than +/- 1 mbar. CONCLUSIONS This equation is neither sufficient to reliably predict nCPAPmin nor to prescribe nCPAP without individual pressure titration. The calculated nCPAP-level might however contribute to choosing the most suitable flow generator prior to the titration night.