Since the discovery of an almost 100% association of HLA-DR2 with narcolepsy-cataplexy, many efforts have been made to demonstrate the intervention of immune factors in the pathogeny of the disease. Some epidemiological features could support this hypothesis: age of onset around 25, triggering factors, association with multiple sclerosis. Molecular studies at the DNA level have, up to now, failed to uncover an abnormal gene in the HLA system, which would imply that the DR2 antigen acts through its role in the immune response. However, results have been largely inconclusive as far as classical features of autoimmunity in blood and CSF are concerned. In canine narcolepsy, a linkage with a human immunoglobulin-related gene has recently been shown, and may constitute a counterpart of the HLA association in man. Thus, the hypothesis of a transient and discrete autoimmune aggression may be ruled out.