Furthermore, our sample was heterogeneous

in different aspects: based on the baseline mean of 178 min duration of PA and total score of Baecke questionnaire included participants were of normal fitness level, however, the high standard MDV3100 nmr deviation also points out that the sample covers fit and unfit persons. Based on BMI classification of the World Health Organization, participants were classified into normal weight but were close to the borderline of being overweight. Because fitness level41 and BMI42 are possible confounders or mediators in sleep we controlled for those variables in our statistical analysis. However, the fitness level and BMI should be included in future studies as independent variable to see whether exercise shows different sleep-promoting effects for fit or unfit and/or for normal or overweighed persons. The study relied on self-report data, except the pedometer data. From a methodological point of view, the

mixed results from the studies so far might be explained by the different assessment of PA and sleep, e.g., the measure of PA ranged from not validated questionnaire items to objectively measures by pedometers and from subjective sleep data (thus assessing KPT-330 concentration the psychological, but not the physiologic part of sleep) to sleep measures via actigraphy or sleep-EEG. Missing data especially for the baseline week could have been avoided by a preliminary meeting to clarify possible problems with the written informed consent about exercise log and sleep log. Further, we are aware of the missing pedometer data for the baseline week, but we decided to not hand out the pedometer at baseline because of possible motivational effects on PA which might have increased the habitual daily activity amount of the participants.43 Two further aspects are the kind of sport and the time of day in which exercise aminophylline is carried out. In our study the focus was on endurance sport (e.g., Nordic walking), however, there is also evidence for improved sleep due to resistance training.44 It would be interesting to contrast endurance and strength

training in an intervention study to see what kind of sport shows better results. Furthermore, in our study the time of day for performing exercise was monitored on the protocol, but because of underrepresentation of morning exercise no statistical analysis was assessed. Therefore from our study no conclusion can be drawn at which time of day exercise should be performed, nevertheless, Passos et al.31 showed that sleep promoting effects did not vary between morning and late-afternoon exercise. Our findings on sleep are mainly based on subjective estimates which may not correspond with objective measures.45 Thus it might be interesting to record also objective measures of sleep by polysomnography or ambulant sleep recording devices (e.g., actigraph). However, for the participants’ point of view the subjective sleep data are most important and therefore the present findings are quite important by itself.