Background: Positive airway pressure is the main treatment modality in obstructive sleep apnea. The level of pressure needed for each patient is defined by a positive airway pressure titration study. Predicting those who need higher pressure or bi-level instead of continuous pressure is of a great clinical significance. This study was designed to evaluate whether voluntary breath-holding maneuver could predict Bilevel Positive Airway Pressure (BiPAP) or Continuous Positive Airway Pressure (CPAP) as the final optimal pressure in a split night study of patients with obstructive sleep apnea. Polysomnography (PSG) parameters specially oxygen saturation (SaO 2 ) during non-rapid eye movement (NREM), rapid eye movement sleep (REM), and duration of REM in diagnostic part may help determine the type of devices as soon as possible in split night study. Methods: The present research was conducted as a cross-sectional study of adults diagnosed as obstructive sleep apnea patients undergoing positive airway pressure (PAP) titration. Demographic, anthropometric, and polysomnographic data were collected. Patients were instructed to hold their breath as long as they could after five tidal breaths in the supine position. Baseline, post breath-holding phase, and recovery SaO 2 were recorded. These data were used to predict the pressure level and type of device. Results: Seventy-eight participants (56.4% male) with the mean age of 55.7 ± 13.9 years were included in the study. Mean and SD of apnea hypopnea index (AHI) and oxygen desaturation index (ODI) were 55.9 ± 34.4 and 38.3 ± 24 per hour, respectively. Most of the participants (65.3%) were categorized as CPAP group. CPAP and BiPAP groups were similar in terms of age, gender distribution, body mass index (BMI), neck circumference, and certain polysomnographic variables. Voluntary breath-holding maneuver showed a significant correlation between minimum SaO 2 and the need for changing to BiPAP during titration. Baseline and post breath-holding SaO 2 were not significantly lower in either group. NREM minimum SaO 2 and REM duration were the statistically significant correlated variables that predicted the need for BiPAP. Conclusion: Minimum oxygen following voluntary breath-holding maneuver along with higher BMI and larger neck circumference are predictors of the need to use BiPAP. Shorter REM duration and NREM minimum SaO 2 were other predictors of higher chance of BiPAP during titration of the present OSA participants.