Abstract Introduction Differences in neuroventilatory control can impact the type and severity of sleep disordered breathing between non-REM and REM sleep. We examined the distribution of central apneic episodes in polysomnograms from subjects with predominantly central sleep apnea, who received phrenic nerve stimulation. Methods Baseline in-lab polysomnograms from patients enrolled in the remedē System Pivotal Trial were scored by a central core laboratory (n=151). Participants with predominantly CSA were enrolled if the apnea-hypopnea index (AHI)≥20/hr, the central apnea index (CAI) exceeded the obstructive apnea index (OAI), and the OAI did not exceed 20% of the total AHI. This post-hoc analysis compared sleep apnea indices in REM and non-REM sleep in those with ≥5 minutes REM sleep (n=131). Within-patient median non-REM - REM differences were calculated and compared. Results REM sleep time was 40 [Q1=28, Q3=64] minutes, and non-REM sleep time was 301 [Q1=269,Q3=344] minutes. AHI in REM sleep was 22/hr [Q1=9,Q3=44] and was 46/hr (Q1=33,Q3=60) in non-REM sleep, yielding a within-patient difference between non-REM and REM sleep of 22/hr ([Q1=6, Q3=34], p<.001). CAI in REM was only 4/hr (Q1=0, Q3=11), but in non-REM was 25/hr [Q1=16,Q3=43] with all patients having a CAI≥5/hr during non-REM. In REM, 70% had a CAI>0/hr and 46% had a CAI ≥5/hr. The CAI difference between non-REM and REM sleep was 18/hr [Q1=10, Q3=30; p<.001]. The OAI and mixed apnea index (MAI) differed by <1/hr (p=0.235 and <.001, respectively). Of note, the hypopnea index [HI] did not differ between REM and non-REM sleep (12/hr [Q1=2,Q3=22] vs. 11/hr [Q1=4, Q3=20], respectively, p=0.273), yet hypopneas constituted a greater proportion of total AHI during REM compared to non-REM sleep. Conclusion Among subjects with predominantly CSA, the prevalence and severity of CSA was greater in non-REM than REM sleep, yet low-level CSA persisted in REM sleep. Stage-related differences in CAI but not OAI, MAI or HI can be attributed to alterations in ventilatory rather than upper airway control in this CSA cohort. These differences in the type and severity of sleep disordered breathing episodes comprise a key diagnostic signature, for which specific CSA therapeutic strategies are indicated. Support (If Any) ZOLL Respicardia, Inc. and NIH R01 HL144859