To the Editor

Malek-Ahmadi et al. [1] evaluated the association between self-reported sleep duration and cognitive impairment in 189 cognitively normal older adults. Adjusted odds ratio (95 % confidence interval) of short sleep duration (≤7 h) against long sleep duration (≥ 9h) for cognitive impairment was 0.86 (0.76–0.98). The authors concluded that old subjects with long sleep duration had a risk of decreased cognitive performance. I have two comments on their study.

First, Chen et al. [2] reported the association between self-reported sleep duration and cognitive decline or mild cognitive impairment (MCI)/dementia in older women. Although they concluded that both short and long sleep duration had a risk for cognitive decline and MCI/dementia, adjusted hazard ratio of short sleep duration against normal sleep duration for MCI/dementia only showed statistical significance. Although Chen et al. presented some biological mechanisms on the association between short sleep duration and MCI/dementia in older women, further study is needed to confirm the association.

Second, the authors used self-reported sleep duration, and sleep evaluation methods should be handled with caution for the risk assessment of cognitive decline or MCI/dementia [3, 4]. Deterioration of sleep was predominantly progressed by psychological distress, and causal association between sleep loss and subsequent cognitive decline or MCI/dementia would be mediated not only by depression but also anxiety [5]. In addition, classification of self-reported short (long) sleep duration needs validation study. Furthermore, Malek-Ahmadi et al. handled cognitively normal older adults. Taken together, the association between sleep duration and cognitive impairment should be evaluated by a cohort or intervention study to confirm the causal association .