Both studies were conducted as pilot studies in a pre-post observational design. Details are already reported elsewhere [13]. The primary outcome for both studies was the proportion of days with reaching the fasting goal of ≥15 h out of the total number of days recorded per participant in the diary. According to the study protocol, secondary outcomes were, among others, changes in sleep quality and duration, and HRQoL between baseline and follow-up.
Recruiting
Participants at the Ulm University were recruited with the support of the occupational health management and by flyers. Exclusion criteria were pre-existing metabolic conditions. Patients at the GP’s office were informed about the study by flyers in the waiting room or were invited by the doctor during a consultation. Exclusion criteria were insulin dependent diabetes or any other disease for which fasting is contraindicated [3]. Finally, 63 participants at the Ulm University and 40 participants at the GP’s office were included in the studies.
Intervention
Participants in both studies were asked to limit their daily food intake to 8–9 h and subsequently extend their nightly fasting period to 15–16 h. The duration of the intervention was 3 months. At baseline, participants had an introductory conversation with the principal investigator or the physician to clarify possible questions and problems in advance, and were given an information brochure. In addition, all participants were offered to contact the respective study centre at any time if they had questions or problems.
Data assessment
Baseline assessment comprised a questionnaire to collect data on lifestyle, health behaviour and HRQoL (EQ-5D VAS) [14], and anthropometric measurements of waist, height, and weight (for details see [13]). All participants were given a diary to record the times of their first and their last meal, and the quality and duration of their sleep. The latter was assessed on a visual analogue scale ranging from 0 (worst possible sleep quality) to 100 (best possible sleep quality). The waist-to-height ratio (WHtR) was calculated by the division of waist by height in centimetre, abdominal obesity was then defined as WHtR ≥0.5, as recommended by the literature [15]. Body weight in kilogram was divided by height in meters squared to determine body mass index (BMI), and subsequently categorized into overweight (≥ 25) and obesity (≥ 30).
After 3 months, follow-up measurements were performed in the same manner, with some additional items in the questionnaire regarding the individual experience and attitudes towards TRE.
Statistical analysis
Baseline characteristics are reported descriptively for each study group and for both groups combined. Differences between groups were tested by applying t-test, Welch’s t-test or Mann Whitney U test according to distribution and heterogeneity in variance for continuous data, and Fisher’s exact test for categorical data.
Follow-up data, and computational differences between baseline and follow-up data, presented as the respective Δ, were treated the same way. Pre and post comparisons for both groups taken together were determined by the Wilcoxon signed-rank test for related samples.
For each participant, mean values and standard deviations were calculated for the data from the diaries. Time of first meal and time of last meal were utilized to determine the duration of the fasting and the eating phase. For all days recorded, the percentage of days with fasting target reached was calculated. Differences between groups were tested as described above.
To assess differences between sleep duration and quality at the beginning and at the end of the TRE intervention period, mean values were calculated for the first 10% and the last 10% of data (or days), respectively. Subsequently the differences between the first and the last 10% of the data were calculated as the respective Δ. They are reported together with the average number of days recorded per group and for the whole group.
Pearson’s correlation coefficient was applied to test bivariate correlations between continuous variables.
Linear regression analyses were conducted for the pre-post differences in HRQoL and the differences in sleep quality between the first 10% and the last 10% of days recorded. Potential factors that might correlate with the HRQoL or sleep quality were identified and, together with variables that differed at baseline between both groups, tested in a stepwise backward elimination. Sex, age, baseline values of HRQoL, the sleep quality and sleep duration on the first 10% of reported days, mean duration of fasting, percentage of fasting target reached, and finally group membership were considered as potential associated factors. Anthropometric measures represented both, potential associated factors and differences between groups at baseline. Therefore, weight, waist circumference, BMI, WHtR, overweight, obesity, abdominal obesity as well as the respective Δ between pre and post measures of the continuous variables were considered in the regression analysis. All assumptions of linear regression (linear relationship, multivariate normality, multicollinearity, auto-correlation, homoscedasticity) were examined.
The significance level for two-sided tests was set at α = 0.05. All statistical analyses were carried out by using the statistical software packages IBM SPSS Statistics for Windows, Version 25.0. (IBM Corp., Armonk, NY, USA).
