Cyclic variation of heart rate (CVHR) associated with sleep apnea/hypopnea episodes has been suggested as a marker of sleep disordered breathing (SDB). This study examined the utility of ECG-based CVHR detection for diagnosing SDB using simultaneous polysomnography as the reference standard. We used a previously developed automated CVHR detection algorithm (autocorrelated wave detection with adaptive threshold, ACAT) that provides the number of CVHR per hour (CVHR index). The ACAT was refined using a polysomnographic database of 194 subjects with various severities of SDB and then, applied to a single channel ECG obtained during standard overnight polysomnography in 862 consecutive subjects referred for SDB diagnosis. Using multiple thresholds of CVHR index ≥ 38 and <27, positive and negative predictive values of 95.6% and 95.1%, respectively, were achieved for detecting and excluding subjects with apnea-hypopnea index (AHI) ≥ 30, leaving 58 (6.7%) unclassified subjects. Positive and negative likelihood ratios (LRs) were 97.3 and 0.23, respectively. Also, thresholds of CVHR index ≥ 29 and <7 provided 96.1% and 95.1% of positive and negative predictive values, respectively, for subjects with AHI ≥ 15 (LRs, 50.6 and 0.11), leaving 426 (49.4%) unclassified subjects. The CVHR correlated with the AHI (r = 0.86) and showed the limits of agreement with the AHI of 19.6 and -18.6. Automated detection of CVHR by the ACAT algorithm provides useful screening tool for both increasing and decreasing probability of moderate and sever SDB with adequate thresholds.