Definition
A diagnosis of narcolepsy can be carried out using polysomnography (PSG) to document sleep patterns and exclude other sleep disorders, and the Multiple Sleep Latency Test (MSLT) to record sleepiness with a short sleep latency—less than five minutes—and sleep-onset rapid eye movement (REM) episodes (SOREMPs). Recently, the diagnostic criteria have been updated to define narcolepsy with cataplexy, narcolepsy without cataplexy, and narcolepsy due to another underlying medical condition (see Table 1).

The high prevalance (>97%) of specific HLA typing has led to suggestions that it should be included in the criteria, but it is not specific for narcolepsy as HLA DQB1*0602 is present in approximately 20% of the background population. It is likely that low cerebrospinal fluid–hypocretin (CSF-Hct) may be included in the diagnostic criteria as >90% of patients with narcolepsy with cataplexy show abnormal CSF-Hct levels, whereas low CSF-Hct is found in a smaller proportion of patients with narcolepsy without cataplexy.^1–3

However, it should be noted that other neurological diseases may present intermediate or low CSF-Hct values, and a number of patients with classic narcolepsy may present normal values.³ It is likely that patients with CSF-Hct may represent a subpopulation of narcoleptic patients, which may have future diagnostic, pathophysiological, and treatment implications. A potential limitation of the definition of narcolepsy is the primary inclusion of patients with excessive daytime sleepiness (EDS). The role of isolated cataplexy or hallucinations without significant EDS has not been established.

Epidemiology
Studies of the prevalence of narcolepsy with cataplexy have been found to be between 25 and 50 per 100,000 people in European countries, Japan, and the US. Thus, narcolepsy is a relatively frequent disorder. Apart from one study in Israel that presented a low prevalence rate, a relatively uniform distribution has been found in most countries. The information regarding occurrence is limited, with one study finding the incidence of narcolepsy with cataplexy to be 0.74 per 100,000 person-years.⁴ In Olmsted County, Minnesota, the incidence ciphers were 1.72 for men and 1.05 for women.⁵ These findings highlight the fact that the disease has an early onset and a chronic disease pattern: this frequency is similar to other neurological diseases, such as multiple sclerosis or parkinsonism.

The onset of the disease manifests as bimodal distribution with disease symptoms. Studies from Montpellier, France and Montreal, Canada presented evidence for a mean age at onset of 15 years and around 35 years, with a wide age distribution. Positive family history presented an increased risk of early onset, which suggests a strong genetic component.6 There is a slight gender preference, with a male predominance of 1.4–1.6:1.⁵

Applying classic epidemiology methods to narcolepsy have not yet identified the specific risk factors for disease development, apart from HLA typing and family history. As with other diseases characterized by selective cell loss—such as Parkinson’s disease or type 1 diabetes mellitus—narcolepsy is likely caused by environmental exposure before the age of onset in genetically susceptible individuals. The main difficulty in identifying such factors is the problem of the significant delay between disease onset and diagnosis, with a delay of at least 14 years. The most thoroughly examined factors include body mass index and stressful life events; however, such associations may reflect a disease consequence rather than a cause of disease. For example, it is likely that metabolic factors may be influenced by abnormal Hct function, and stress and depressive symptoms are as likely to be consequences as causes of disease.

In some cases there is a strong familial association of narcolepsy, although this explains only a minority of disease cases.^7–9 Immune mechanisms are likely to be associated with narcolepsy. During the past few years antigens have been identified as part of the development of the disease,^10–13 and as a consequence of these findings immune globulin treatment has been suggested.^14,15 The main problem in interpreting these findings is that not all subjects present antigens, and the antigens are found in subjects with a long disease duration, which may mask a potential relation. Antigens that react against hypothalamic neurons are also found in control subjects, and other immune systems such as cellular mechanisms may be involved. Further prospective studies are needed.