How reliable is a simplified MSLT nap termination protocol?

Abstract

ObjectiveAccording to current practical guidelines, naps of the Mean Sleep Latency Test (MSLT) must be terminated 15 min after sleep onset, which requires ad hoc scoring. For clinical convenience, some sleep clinics use a simplified protocol with fixed nap lengths of 20min. Its diagnostic accuracy remains unknown. MethodsA subset of MSLT naps of 56 narcolepsy type 1 (NT1), 98 Parkinson's disease (PD), 117 sleep disordered breathing (SDB), 22 insufficient sleep syndrome (ISS) patients, and 24 patients with idiopathic hypersomnia (IH), originally performed according to the simplified protocol, were retrospectively adjusted to standard protocol (nap termination 15min after sleep onset or after 20min when no sleep occurs). This was feasible in 60% of MSLT naps; in this subset, we compared sensitivity and specificity of both MSLT protocols for identification of patients with and without NT1. ResultsSensitivity of classical MSLT criteria for NT1, i.e. mean sleep latency ≤8.0min and ≥2 sleep onset rapid eye movement periods (SOREMPs), did not differ between protocols (95%). Specificity, however, was slightly lower (88.1% vs. 89.7%) in the simplified nap termination protocol, with 3 SDB patients and 1 ISS patient having false-positive MSLT findings in the simplified but not in the standard protocol. ConclusionsThe use of a simplified MSLT protocol with fixed nap duration had no impact on MSLT sensitivity for NT1, but the longer sleep periods in the simplified protocol increased the likelihood of REM sleep occurrence particularly in non-NT1 conditions, resulting in a slightly lower MSLT specificity compared to the standard protocol.