OBJECTIVE:Continuous positive airway pressure (CPAP) is an effective treatment for obstructive sleep apnea (OSA). After performing an initial titration study, most physicians do not have the CPAP equipment retitrated unless the patient complains about the CPAP use. Several automated CPAP devices are used clinically that can detect upper airway obstructive events and provide information about residual events while patients are on CPAP. The aim of this study was to compare the apnea-hypopnea index (AHI) determined by automated CPAP devices to that obtained from polysomnography. METHODS:Patients with OSA underwent polysomnography for CPAP titration using the REMstar Auto M-series. The initial two hours of CPAP titration were spent at a subtherapeutic pressure of 4 cmH(2)O so that more breathing events could be observed. The correlations between the simultaneous determination of the AHI with polysomnography (AHI-PSG) and the automated device (AHI-RAM) during the subtherapeutic, therapeutic and overall phases were evaluated. In addition, the apnea index (AI) and the hypopnea index (HI) were each evaluated separately. RESULTS:Sixty patients were enrolled. The mean AHI on diagnostic PSG was 35.2±2.6 events/hour. Strong correlations were observed between the AHI-PSG and the AHI-RAM (subtherapeutic: r=0.958, p<0.001; therapeutic: r=0.824, p<0.001; overall: r=0.927, p<0.001). A slightly stronger correlation was observed between the AI values, whereas a weaker correlation was observed between the HI values in all three phases. CONCLUSION:Strong correlations between the AHI-PSG and the AHI-RAM were observed. The correlations were weakened when the analysis was limited to the HI and the therapeutic phase.