Late, but Not Early, Night Sleep Loss Compromises Neuroendocrine Appetite Regulation and the Desire for Food

**Background:** Reduced sleep duration is known to increase hunger, appetite, and food intake, contributing to metabolic diseases such as obesity and type 2 diabetes. However, the impact of sleep timing—independent of sleep duration—on appetite regulation is less understood. This study aimed to evaluate whether sleep loss during the late versus early part of the night differentially affects neuroendocrine appetite regulation and the desire for food.

**Methods:** Fifteen healthy, normal-weight (BMI: 23.3 ± 0.4 kg/m²), young (age: 24.6 ± 0.7 years) men were studied in a randomized, balanced, crossover design with three conditions spaced at least three weeks apart: (1) 'late-night sleep loss' (4 h sleep, bedtime 22:30–03:00 h), (2) 'early-night sleep loss' (4 h sleep, bedtime 02:15–06:45 h), and (3) 'regular sleep' (8 h sleep, bedtime 22:15–06:45 h). Participants arrived at 19:15 h, received a standardized 380 kcal dinner at 20:15 h, and were allowed only water thereafter. Polysomnographic recordings were performed. Blood samples for plasma total ghrelin and leptin (measured by radio-immunoassay) were collected at multiple time points from evening through the next morning (until 11:00 h). Subjective feelings of hunger, appetite, and desire for food (general, hearty, and sweet) were assessed at 60-min intervals using semi-quantitative symptom rating scales (0–10). Statistical analyses used repeated-measures ANOVA with factors 'condition' and 'time', with Greenhouse-Geisser correction and non-parametric tests for pairwise comparisons.

**Key Results:** Total sleep time was 426 ± 43 min in the regular sleep condition, 249 ± 18 min in late-night sleep loss, and 263 ± 10 min in early-night sleep loss (p = 0.041 for condition effect; p = 0.771 between sleep loss conditions). Leptin concentrations increased with sleep onset independent of condition (p = 0.031 for time effect; p = 0.812 for condition × time interaction) and showed no differences between conditions in the morning (p = 0.213 for condition × time interaction; AUC p = 0.32). Ghrelin concentrations increased with sleep onset in the late-night sleep loss and regular sleep conditions but not in early-night sleep loss (p = 0.001 for condition × time interaction). In the morning, ghrelin was significantly elevated after late-night sleep loss compared to both regular sleep and early-night sleep loss (p = 0.004 for condition effect; p ≤ 0.008 for pairwise comparisons; AUC p = 0.006). Feelings of hunger and appetite were markedly increased in the morning after late-night sleep loss compared to both other conditions (p ≤ 0.039 for Friedman test; both p ≤ 0.008 for Wilcoxon tests). General desire for food at 07:00 h was highest after late-night sleep loss (p = 0.001 for condition effect), particularly for hearty foods (p = 0.05 for condition effect; p = 0.038 for late-night vs. early-night sleep loss). By 11:00 h, hunger, appetite, and desire for food converged and were no longer significantly different between conditions.

**Clinical Implications:** This study demonstrates that the timing of sleep restriction critically modulates its effects on appetite regulation. Late-night sleep loss (early awakening) elevates ghrelin and increases hunger, appetite, and desire for food, whereas early-night sleep loss (delayed bedtime) does not produce these effects despite equivalent sleep duration. These findings highlight that chronobiological factors—not just total sleep time—should be considered in sleep recommendations. Late-night sleep loss may represent a specific risk factor for overeating and metabolic diseases such as obesity and type 2 diabetes. Limitations include the small sample size (n=15), inclusion of only young healthy men, assessment after only a single night of sleep restriction, and measurement of total rather than acylated ghrelin. Further studies are needed in women, at-risk populations (e.g., obesity, T2D), and under longer-term or chronic circadian disruption conditions.