We studied 107 patients with sleep-disordered breathing to confirm the effectiveness of continuous positive airway pressure (CPAP) titration using a split-night protocol. Patients spent two consecutive nights in our laboratory with complete polysomnography. On the first night, we applied a split-night protocol; the first half of the night was used as a baseline (B), and after a diagnosis was made, CPAP was applied during the second half of the night (SN). On the second night (2N), patients spent the entire night on CPAP to confirm the effectiveness of CPAP treatment. The SN and 2N both revealed a significant reduction in arousal index (37.8 +/- 27.9 on B, 13.2 +/- 12.1 on SN, 11.4 +/- 8.0 on 2N, values are mean +/- SD, p < 0.001), apnea hypopnea index (AHI) (23.6 +/- 26.3/h on B, 3.0 +/- 3.7/h on SN, 2.4 +/- 2.6/h on 2N, p < 0.001), percent total sleep time below 90% SaO2 (21.0 +/- 27.2% on B, 8.2 +/- 13.8% on SN, 4.9 +/- 10.2% on 2N, p < 0.001), and percent total sleep time below 80% SaO2 (1.1 +/- 3.8% on B, 0.0 +/- 0.1% on SN, 0.1 +/- 0.5% on 2N, p < 0.001). There were no significant differences between the SN and the 2N for these measurements. Final CPAP pressure was significantly lower at the end of the SN when compared with the 2N (8.8 +/- 2.7 cm H2O on SN, vs 10.3 +/- 2.8 cm H2O on 2N, p < 0.001). When patients were divided into three groups (AHI < 20, n = 69; 20 < AHI < 40, n = 18; AHI > 40, n = 20), the final CPAP pressure was different only in the group with AHI < 20 (8.1 +/- 2.3 cm H2O on SN, 9.6 +/- 2.3 cm H2O on 2N, p < 0.001). We conclude that a split-night protocol may be sufficient to determine the effective CPAP pressure, especially in patients with an AHI > 20.