Limits of self-report in assessing sleep terrors in a population survey.

Sleep terrors are less frequent compared to other parasomnias, and there are no prevalence studies on adults. We performed a questionnaire study in a well-defined population-based sample, the Finnish Twin Cohort. The study population consisted of 11,220 subjects aged 33-60 years, responding to questions on the frequency of sleep terrors in childhood and as adults. In the first questionnaire about 9% reported sleep terrors often or sometimes in childhood, and 3.5% at least once monthly as adults. However, in a second more-detailed questionnaire, only 1% of those with at-least-monthly attacks in adulthood presented with features compatible with the minimal diagnostic criteria for sleep terrors of the International Classification of Sleep Disorders. There was also a strong correlation between current occurrence of nightmares and the report of sleep terrors. Although a clinically definable entity, sleep terrors seem to be unknown to lay people, at least in Finland. Therefore, the use of single items or brief question series on sleep terrors may give inaccurate results in questionnaires. An interview of a person who has witnessed the nocturnal attack suspected to be sleep terror is essential because of the patient's impaired recall of the episode. Our results also support the general view that sleep terrors are rare in adults.