Short-term changes in human metabolism following a 5-h delay of the light-dark and behavioral cycle

Abstract

Experimental inversion of circadian and behavioral rhythms by 12 h adversely affects markers of metabolic health. We investigated the effects of a more modest 5-h delay in behavioral cycles. Fourteen participants completed an 8-day in-patient laboratory protocol, with controlled sleep-wake opportunities, light-dark cycles, and diet. The 5-h delay in behavioral cycles was induced by delaying sleep opportunity. We measured melatonin to confirm central circadian phase, fasting markers and postprandial metabolism, energy expenditure, subjective sleepiness, and appetite, throughout the waking period. After the phase delay, there was slower gastric emptying at breakfast, lower fasting plasma glucose, higher postprandial plasma glucose and triglycerides, and lower thermic effect of feeding. Any changes were abolished or attenuated within 48-72 h. These data extend our previous findings, which showed no time-of-day effect in healthy adults on daytime energy expenditure or thermic effect of feeding when accounting for circadian variation in resting metabolic rate.

Keywords: biological sciences; human metabolism; physiology.

Graphical abstract

Figure 1

Study protocol and experimental measures (A) Illustration of the phase shift protocol with example of a participant on a 23:00- to 07:00-h sleep schedule (waking at 07.00 h and sleep time at 23.00 h). Light bars indicate lights on and wakefulness. Black bars indicate lights off and sleep opportunity. B, breakfast; L, lunch; D, dinner; S, snack. Meals started 1 h after waking and were separated by 5 h. On Day 2, lights-on was extended by 5 h to induce the phase shift. On Day 2, a snack was provided at 23.00 h to maintain a consistent 16-h fasting period to breakfast on Day 3. Lunch and dinner on Day 2 are aligned in clock time to breakfast and lunch after the phase shift to permit comparisons between meals occurring at the same clock time during aligned versus misaligned conditions. (B) Test days (Day 2, Day 3, Day 5, and Day 7) protocol. Lights-on in the individual sleep laboratories occurred 1 h before breakfast. Bodyweight was measured following voiding upon waking. Baseline subjective questionnaires (KSS/VAS) and the baseline 30-min resting metabolic rate (RMR) measure were completed; 5-min prior to breakfast, baseline breath samples were taken for gastric emptying in addition to baseline fasted blood samples. Breath samples for gastric emptying were taken after breakfast at 15, 30, 45, 60, 75, 90, 120, 150, 180, 240, and 300 min postprandial. Blood samples were taken 15, 30, 60, 90, and 120 min in each postprandial period for glucose, insulin, and triglycerides. Melatonin sampling began in the afternoon ∼11.5 h before habitual bedtime, and 12 samples were taken over 11.5 h. Questionnaires were completed every hour across the waking period. Meals were initiated 1 h after waking, with lunch and dinner separated by 5 h, respectively.

Figure 2

Effects of a 5-h phase delay on melatonin timing (A) Dim-light melatonin onset (DLMO) between study days and (B) time interval between waking and DLMO between study days. ∗∗∗p < 0.001 for the difference in DLMO comparing Day 2 to Day 3, Day 5, and Day 7. ###p < 0.001 for the difference in DLMO relative to wake time comparing Day 2 to Day 3 and Day 5; ##p = 0.005 for the difference in DLMO relative to wake time comparing Day 2 and Day 7. Data are presented as mean ± SEM.

Figure 3

Effects of a 5-h phase delay on daily profiles of sleepiness and alertness (A) Karolinska Sleepiness Scale (KSS) scores. The KSS is a 9-point scale of discreet numbers between 1 and 9 corresponding with a verbal anchor from: 1 “Extremely Alert” to 9 “Fighting Sleep”. (B) Alertness Scale, which contains two verbal anchors and a corresponding discrete number: 1 “Very Alert” to 9 “Very Sleepy” (i.e., higher scores indicate lower alertness). Data are presented as mean ± standard error. ∗∗∗p < 0.001 for the difference in KSS and Alertness from 14 to 17 h after waking, compared to the morning, on Days 3, 5, and 7. The truncated X axis represents the additional time points included for the 5-h extended waking period on Day 2 to induce the phase shift. Black = Day 2; Red = Day 3; Green = Day 5; Purple = Day 7. Data are presented as mean ± SEM.

Figure 4

Whole-day energy expenditure and thermic effect of food responses (A) Time course of whole-day (16.5 h) energy expenditure (kJ/min) between study days (Black = Day 2; Red = Day 3; Green = Day 5; Purple = Day 7). (B, L, and D) indicate the timing of breakfast, lunch, and dinner, respectively, relative to energy expenditure measurements. The fasting resting metabolic rate measure was conducted for 30 min upon waking, which is represented as 0 on the X axis. Data are presented as mean and standard error. (B) Thermic effect of feeding [kJ/d] between test days. Data are presented as bars with mean and standard error. ∗∗p < 0.01 for the difference in TEF comparing Day 3 to Day 2; ##p < 0.01 for the difference in TEF comparing Day 7 to Day 3. Data are presented as mean ± SEM.

Figure 5

Effects of a 5-h phase delay on gastric emptying (A) Gastric emptying T-½ between study days. (B) Gastric emptying T-Lag between study days. Data are presented as bars with mean and standard error. ∗p < 0.05 for the difference in T-½ comparing Day 3 to Day 2. Data are presented as mean ± SEM.

Figure 6

Effects of a 5-h phase delay on fasting and postprandial metabolism (A) Fasting plasma glucose between study days. ∗∗p < 0.01 for the difference in fasting glucose comparing Day 3 to Day 2; ∗p < 0.05 for the difference fasting glucose comparing Day 5 to Day 2. (B) Postprandial plasma glucose incremental area under the curve (iAUC; mmol/L · min) for the main effect of day. ##p < 0.01 for the difference in plasma glucose iAUC on Day 3 compared to Day 2. (C) Fasting plasma insulin between study days. (D) Postprandial plasma insulin iAUC (pmol/L · min) for the main effect of day. (E) Fasting plasma triglycerides (TGs) between study days. (F) Postprandial plasma TG iAUC (mmol/L · min) for the main effect of day. ⋅⋅p < 0.01 for the difference in iAUC on Day 3 compared to Day 2 and ⋅ p < 0.05 for the difference in on Day 5 compared to Day 2. Data are presented as mean ± SEM.