The effect of sleep disturbances on the incidence of dementia for varying lag times

BACKGROUND: Few studies have addressed the association of sleep disturbances with incident dementia with long lag times. We add to this literature by investigating how lag times varying from 2.2 to 23.8 years affect the relationship between sleep disturbance and incident dementia in a Dutch cohort study on aging. METHODS: Using eight waves of data from the Longitudinal Aging Study Amsterdam, we investigated the association of hours of sleep, difficulty falling asleep, interrupted sleep, and waking up early with incident dementia. For dementia an algorithm was used based on repeated measurements of cognitive tests and other data sources that provide strong indications of dementia. Sleep disturbances were assessed with a self-report questionnaire. RESULTS: Of 2,218 participants, 237 (11%) developed dementia in the period 1992/3 to 2015/6. Participants ≥70 years more often reported sleep disturbances compared to those <70. Only for a short lag time (3 years), sleeping ≥9 h was associated with incident dementia. Sleeping ≤6 h, interrupted sleep and waking up early were associated with incident dementia, particularly for lag times ≥15 years. DISCUSSION: We found that the association of sleep disturbances with incident dementia becomes stronger with longer lag times (particularly ≥15 years). Studies with lag times <15 years may suffer from reverse causation due to the changes in sleep patterns caused by the prodromal phase of neurodegenerative disease. The association of sleeping ≥9 h and the incidence of dementia in analyses with a short lag time seem to be the result of reverse causation.