Context: Narcolepsy, a neurological disorder affecting 1 in 2000 individuals, is associated with HLA-DQB1*0602 and low cerebrospinal fluid (CSF) hypocretin (orexin) levels. Objectives: To delineate the spectrum of the hypocretin deficiency syndrome and to establish CSF hypocretin-1 measurements as a diagnostic tool for narcolepsy. Design: Diagnosis, HLA-DQ, clinical data, the multiple sleep latency test (MSLT), and CSF hypocretin-1 were studied in a case series of patients with sleep disorders from 1999 to 2002. Signal detection analysis was used to determine the CSF hypocretin-1 levels best predictive for International Classification of Sleep Disorders (ICSD)-defined narcolepsy (blinded criterion standard). Clinical and demographic features were compared in narcoleptic subjects with and without low CSF hypocretin-1 levels. Setting: Sleep disorder and neurology clinics in the United States and Europe, with biological testing performed at Stanford University, Stanford, Calif. Participants: There were 274 patients with narcolepsy; hypersomnia; obstructive sleep apnea; restless legs syndrome; insomnia; and atypical hypersomnia cases such as familial cases, narcolepsy without cataplexy or without HLA-DQB1*0602, recurrent hypersomnias, and symptomatic cases (eg, Parkinson disease, depression, Prader-Willi syndrome, Niemann-Pick disease type C). The subject group also included 296 controls (healthy and with neurological disorders). Intervention: Venopuncture for HLA typing, lumbar puncture for CSF analysis, primary diagnosis using the International Classification of Sleep Disorders, Stanford Sleep Inventory for evaluation of narcolepsy, and sleep recording studies. Main Outcome Measures: Diagnostic threshold for CSF hypocretin-1, HLA-DQB1*0602 positivity, and clinical and polysomnographic features. Results: HLA-DQB1*0602 frequency was increased in narcolepsy with typical cataplexy (93% vs 17% in controls), narcolepsy without cataplexy (56%), and in essential hypersomnia (52%). Hypocretin-1 levels below 110 pg/mL were diagnostic for narcolepsy. Values above 200 pg/mL were considered normal. Most subjects with low levels were HLA-DQB1*0602-positive narcolepsy-cataplexy patients. These patients did not always have abnormal MSLT. Rare subjects without cataplexy, DQB1*0602, and/or with secondary narcolepsy had low levels. Ten subjects with hypersomnia had intermediate levels, 7 with narcolepsy (often HLA negative, of secondary nature, and/or with atypical cataplexy or no cataplexy), and 1 with periodic hypersomnia. Healthy controls and subjects with other sleep disorders all had normal levels. Neurological subjects had generally normal levels (n=194). Intermediate (n=30) and low (n=3) levels were observed in various acute neuropathologic conditions. Conclusions: Narcolepsy-cataplexy with hypocretin deficiency is a genuine disease entity. Measuring CSF hypocretin-1 is a definitive diagnostic test, provided that it is interpreted within the clinical context. It may be most useful in cases with cataplexy and when the MSLT is difficult to interpret (ie, in subjects already treated with psychoactive drugs or with other concurrent sleep disorders).
