REM sleep latency changes after version 2.1 of the AASM manual for scoring sleep

The classical criteria for scoring REM sleep changed in version 2.1 of the AASM manual for scoring sleep, by allowing N1 epochs with atonia precedent and contiguous to definite REM sleep to be scored as REM sleep in the absence of rapid eye movements when the EEG was compatible. This may shorten the REM latency in the Multiple Sleep Latency Test (MSLT) in naps with wake/N1 to REM transitions, characteristic of narcolepsy type 1. Since REM latency of <5 or <6 min is a biomarker of NT-1 we have assessed the impact of this change in scoring REM sleep in the MSLT.Ninety-two consecutive five-nap MSLT studies (460 naps) performed in our center between 2013 and 2019 for evaluation of hypersomnolence with ≥1 sleep onset REM (SOREM) naps were included. REM latencies were measured using both classical and new criteria.SOREMs occurred in 255 (55.9%) naps, 134 directly from wake/N1. By using the new criteria REM latency shortened in 29.1% of these naps (mean 0.2 ± 0.5, range 0-3 min, p < 0.01), predominantly in females. Twenty-eight percent of MSLTs had at least one nap with a shortened REM latency (mean 0.1 min ± 0.2, p < 0.01). Only two MSLTs changed their REM latency to <5 min and none to <6 min with the new rules.The criterion to define REM sleep onset significantly influences its latency and should be considered when comparing studies performed before or after version 2.1 modification. The clinical relevance of this scoring change is probably minimal.