Background: Prediction of effective CPAP level to treat obstructive sleep apnea syndrome (OSAS) before CPAP titration helps the effectiveness of titration. Several algorithms that predict the optimal continuous positive airways pressure (CPAP) level have been developed. However, a standard algorithm has not been composed.There are two aims of this study. First, to examine the factors that account for the variability in CPAP levels required to abolish apnea and hypopnea in patients with OSAS. Second, to obtain a formulation of predicting the lowest effective pressure from some anthropometric and polysomnographic variables for Turkish patients with OSAS.Methods: We retrospectively have evaluated 127 patients with mild-severe obstructive sleep apnea who were applied CPAP titration. Anthropometric (body mass index(BMI), age, gender) and polysomnographic (sleep efficacy, epworth sleepiness scale (ESS) score, multiple sleep latency test (MSLT) mean sleep latency, apnea hypopnea index, minimum O2 saturation) parameters in patients are determined. Then, we have evaluated the relationship between optimum CPAP level and those parameters.Results: We found a significant correlation between CPAP level and BMI (p<0.028) and AHI (p<0.001) and min O2sat (p<0.001). However, we did not find a correlation between CPAP level and BMI in presence of multiple variables. Also, there wasn't any correlation between CPAP level and sleep efficacy and ESS score and MSLT mean sleep latency.Conclusions: Predicting optimal CPAP level formulation: logCPAP=0.921-(0.002xminO2sat)+(0.001xAHI). According to this formulation, BMI is not a parameter that affect CPAP level. On the other hand, min O2sat and AHI are important that determine to CPAP level.