Abstract
Study Objectives: To assess the test-retest reliability of the polysomnography-multiple sleep latency test (PSG-MSLT) diagnostic classification and measures and to study the determinants of its variability in patients with narcolepsy type 1 (NT1) or with noncataplectic central disorders of hypersomnolence (NCHS): type 2 (NT2), idiopathic hypersomnia (IH), and unspecified hypersomnolence (unspecified excessive daytime sleepiness [UnsEDS]). Methods: PSG-MSLT in drug-free conditions was administered twice (median interval of 1.9 years) in 22 patients with NT1 (10 males, median age 31.2 years) and 75 patients with NCHS (32 males, median age 25.7 years). Results: At the first PSG-MSLT, patients with NCHS were classified as having NT2 (22.7%), IH (26.7%), or UnsEDS (50.6%). A positive PSG-MSLT was confirmed in 72.7% of NT1. The classification consistency at retesting was significantly lower for the NT2 (47.1%), IH (25.0%), and UnsEDS (42.1%) categories than NT1 (81.3%). The between-test mean sleep latency (MSL) variability was significantly different in NT1 and NCHS, with higher changes in NT2 and lower in NT1. A longer test-retest interval was associated with improved MSL and MSLT normalization. Between-test variations in SOREMP number were associated with changes in nocturnal REM sleep parameters and MSL. No association was found with the clinical decision to repeat the evaluation or the disease clinical course. Conclusion: The PSG-MSLT measures and classification are not stable in patients with NCHS, with frequent diagnostic changes, particularly for NT2 and IH, compared with NT1. MSLT needs to be repeated at regular intervals to confirm a stable hypersomnia and provide an accurate diagnosis of NT2 and IH.