Eating during the biological night is associated with nausea

Abstract

Objectives:

This study assessed whether there was a time-of-day effect on nausea reports in participants during studies employing circadian protocols.

Methods:

Visual-analog-scales of nausea ratings were recorded from 34 participants (18-70 years; 18 women) during forced desynchrony (FD) studies, where meals were scheduled at different circadian phases. Subjective nausea reports from a further 81 participants (18-35 years; 36 women) were recorded during constant routine (CR) studies, where they ate identical isocaloric hourly snacks for 36-40 hours.

Results:

Feelings of nausea varied by circadian phase in the FD studies, peaking during the biological night. Nausea during the CR was reported by 27% of participants, commencing 2.9±5.2 hours after the midpoint of usual sleep timing, but was never reported to start in the evening (4-9 PM).

Conclusions:

Nausea occurred more often during the biological night and early morning hours. This timing is relevant to overnight and early morning shift workers and suggests that a strategy to counteract that is to pay careful attention to meal timing.

Introduction

Eating when the body is physiologically unpreparedto process food can result in nausea, vomiting, abdominal pain, diarrhea, constipation, bloating, heartburn, indigestion, and other gastrointestinal problems. In shift workers and transmeridian travelers, sleep deprivation, food choices, and dehydration are thought to contribute to these symptoms^1–7. However, little is known about the timing of the symptoms. Moreover, because shiftwork is associated with changes in sleep and meal patterns, higher consumption of unhealthy foods, and meal skipping^8,9, it is not clear whether the more frequent complaints of nausea and gastrointestinal symptoms experienced by shift workers, compared to dayworkers, are due to such external factors or due to endogenous circadian variations in appetite¹⁰ and metabolism.^11,12 We investigated reports of nausea in healthy young and older adults in studies using the Forced Desynchrony (FD)¹³ and Constant Routine (CR)¹⁴ protocols. These gold standard protocols allow assessment of the effects of endogenous circadian rhythms on outcome measures under conditions in which the environmental and behavioral factors that confound such measurements are controlled or minimized¹³. We hypothesized that self-reported nausea would occur more often during the biological night, independent of food intake and other behaviors, and that sleep restriction would further exacerbate nausea.

Methods

We included data from two FD and four CR studies (Table 1), that were approved by the Partners Health Care (now Mass General Brigham) Institutional Review Board (2005-P-002292, 2005-P-002382, 2010-P-000346, 2013-P-002686, 2014-P-000343, 2019-P-001247) and that adhered to the ethical principles in the Declaration of Helsinki. Participants provided written informed consent. Aspects of the FD studies^{11,12,15–17} and two of the four CR studies^18,19 have been published.

Table 1.

Demographics table. Study 2005-P-002292 had participants from two age groups; younger and older adults.

Study	2005-P-002292	2005-P-002292	2014-P-000343	2005-P-002382	2010-P-000346	2013-P-002686	2019-P-001247
Duration	3-week FD	3-week FD	3-week FD	40-hour CR	36-hour CR	36-hour CR	39-hour CR
Number of subjects	12	12	10*	36	23	10	12
Age (years)	22.4±2.6	59.6±4.5	58.2±8.3	23.0±3.1	23.7±2.6	21.6±2.3	25.6±4.3
Age range (years)	18-27	55-70	38-69	18-30	21-30	18-25	21-35
Women	6	6	6	13	12	3	8
Men	6	6	4	23	11	7	4
American Indian or Alaskan	0	0	0	1	0	0	0
Asian	0	0	0	2	1	0	3
Black	0	0	0	2	2	2	0
White	9	12	10	27	18	7	8
More than one	3	0	0	4	2	0	0
Hispanic	2	2	2	3	1	0	2
Non-Hispanic	10	10	8	33	22	9	10
BMI	24.0±3.2	23.9±2.5	27.1±4.5	n/a	23.5±2.7	23.2±2.1	22.3±1.8
O/L	55.2±6.9	64.2±7.9	67.2±7.0	48.9±18.4	54.1±8.8	52.8±5.8	58.2±8.3
PSQI	1.8±1.4	2.2±1.1	2.6±1.4	n/a	2.4±1.6	2.2±1.3	2.3±1.5
ESS	n/a	n/a	2.5±1.8	n/a	3.9±2.0	4.4±2.7	4.4±2.7
Subjects reporting nausea	7	7	4	4	14	3	1

Abbreviations: FD, Forced Desynchrony; CR, Constant Routine; BMI, Body Mass Index; O/L, Owl and Lark morningness-eveningness questionnaire; PSQI, Pittsburgh Sleep Quality Index; ESS, Epworth Sleepiness Score.

^*Seven unique subjects, three subjects (one male, two females) completed study twice under different diet conditions, four unique subjects reported nausea.

Briefly, inclusion criteria were similar for the studies: good health, no medical, psychological, or psychiatric conditions, no regular use of prescription or over-the-counter medications other than oral contraceptives, no shiftwork in the previous year, and no travel across more than 1 time zone in the previous 3 months. Participants maintained a fixed sleep/wake schedule with 8-10 hours in bed at their habitual sleep time for at least one week prior to the inpatient protocol, during which time all behavioral and environmental factors including sleep, meals, posture, physical activity, room temperature, and light were controlled. The inpatient studies were conducted at the Center for Clinical Investigation (Brigham and Women’s Hospital); all participants were habituated to the laboratory conditions for 1-7 nights while maintaining a regular 8-10-hour sleep schedule prior to the start of their FD or CR protocol.

Thirty-four participants (Table 1) underwent three weeks of FD with eighteen 28-hour “days”. They were scheduled to sleep for 6.5 hours (insufficient sleep) or 11.67 hours (sufficient sleep) per 28 hours, equivalent to 5.6 hours or 10 hours per 24 hours. Calorie- and macronutrient-controlled meals (breakfast, lunch, dinner, and snack; an additional snack was served for 5 participants) were scheduled according to time after wake, meaning that, as the protocol progressed, the meals were served at different circadian phases. They completed non-numeric visual-analog-scales for nausea before and after each meal. Participants were instructed to consume all their food. Core body temperature was measured continuously via a rectal thermistor to estimate circadian period and phase (0-359°) using non-orthogonal spectral analysis.^13,20

Eighty-one participants in the CR studies (Table 1) were awake, in a semi recumbent position, in bed for 36-40 hours, consuming identical hourly snacks (calorie- and macronutrient-controlled) consisting of a small sandwich, juice, and water. If the participant started to feel nauseated the snack was modified (e.g., saltine crackers instead of sandwich, water instead of juice) until they could resume eating their regular snacks. The time the hourly snack was given, and any complaints of nausea were documented. These complaints included e.g., “didn’t think they could eat much more without feeling sick”, “reports nausea with snacks which was resolved after switching to saltines”, and “the snacks were going to make them vomit”. In protocol 2010-P-000346, the 40-hour CR was scheduled twice with 60 hours between the two CRs. Either the time the participant first reported feeling nauseated or when their snack required modification was selected as the time of the first nausea event, whichever occurred first. The participant was counted as having nausea if any instance of nausea occurred during a CR. The time of the first report of nausea was calculated separately for each CR. Thus, as some participants underwent two CRs, each participant could contribute up to two nausea events.

In statistical analyses (SAS 9.4, SAS Institute, Cary, NC), visual-analog-scale data were dichotomized; values ≤10 were converted to zeros and values >10 were converted to 1 due to many participants reporting low scores for nausea. For the FD studies, values were binned into three ~1 week FD cycles, six ~4-hour circadian phase bins, and 5-6 ~4-hour time awake bins. Generalized linear and non-linear mixed cosine models were used due to the binary distribution. Study (sufficient vs. insufficient sleep), FD cycle, circadian phase, and their interactions were treated as fixed effects (categorical), participant was included as a random effect, and the model was adjusted for time awake. For the CR studies, descriptive statistics were reported as mean±standard deviation (SD) or percentage. Statistical significance was set at p<0.05.

Results and discussion

We found an endogenous circadian rhythm in self-reported nausea during the FD studies (p<0.0001), with a peak at ~0° and a trough at ~180°, equivalent to approximately 5 AM and 5 PM, respectively. This is similar to results from FD studies with 20-hour days and 8/24 hours of scheduled sleep in healthy younger adults where self-ratings of nausea varied by circadian phase^10,21, peaking at 20° (equivalent to 5:50 AM)¹⁰. Self-reported nausea ratings were highest, and the circadian rhythm in nausea was strongest, during the first FD week, which could be due to adaptation to study conditions/diet or because foods that made participants feel nauseated were swapped out of their diet. Nausea levels did not appear to be related to the number of calories consumed as pre- and post-meal visual analog scales indicated that nausea levels peaked at circadian phase 0°.

Moreover, while the overall pattern was similar between the two FD studies (sufficient vs. insufficient sleep) for the first week of FD, the self-rated nausea levels decreased from the first week to the second and third weeks of FD in the sufficient sleep condition, whereas in the insufficient sleep condition these ratings remained at the same level throughout the FD, but without robust circadian variation. More specifically, there was a robust circadian rhythm (p=0.002 and p=0.046 for amplitude) during the first two FD weeks in the sufficient sleep condition, while only a trend was observed in the insufficient sleep condition and only in the first week of FD (p=0.077 for amplitude).

Out of the 81 participants, 27% reported at least one instance of nausea during the CR, although this percentage varied between studies from 8% to 61% (Table 1). The 61% of participants who reported nausea in protocol 2010-P-000346 may be due to undergoing two CRs (43% reported nausea in CR1). The earliest reported time of nausea during the CR occurred on average 2.9 (±5.2 SD) hours after the midpoint of the usual sleep episode (Figure 1). Interestingly, while there were first reports of nausea at almost all other times, there were no first reports of nausea 6-11 hours before the midpoint of sleep (equivalent to ~4-9 PM), which coincides with the time of day when the circadian rhythms in alertness¹⁷, hunger¹⁰ and energy expenditure and macronutrient utilization¹² are highest. Thus, in keeping with the anticipatory nature of the circadian system²², there appears to be an optimum window for eating food when people are most likely to be hungry and least likely to experience nausea, an “appetite maintenance zone” that coincides with the wake-maintenance zone²³.This window occurs in the early evening just before the nighttime peak in melatonin, which at higher levels may inhibit hunger.^24–27 In fact, this effect of melatonin on appetite may partially explain why nausea is one of the most common side effects of melatonin supplements^28,29. Thus, night shift workers who can phase-shift their circadian clock and melatonin rhythm to be more aligned with night work and day sleep may be less likely to experience nausea.

Figure 1.

Time and frequency of first reports of nausea (in blue color) relative to the midpoint of habitual sleep (MOS) in the four Constant Routine studies. These reports occurred most frequently 2-3 hours (+2h to +3h) after MOS, which corresponds to 5-6 AM for a person who habitually sleeps from 11 PM to 7 AM and whose MOS thus occurs around 3 AM.

Finally, chronic sleep loss, commonly associated with shift work, can induce alterations in the gut microbiota and gastrointestinal function³⁰ and in the immune system,³¹ and thus exacerbate nausea and other gastrointestinal symptoms. Future studies are needed to investigate the effects of shift work and irregular schedules on nausea, and to develop countermeasures for nausea in shift workers.

Dr. Czeisler’s contributions to the work

Dr. Czeisler was the Principal Investigator for most of these studies. He established the laboratory where the studies were conducted and played a key role in developing and refining the FD and CR protocols for assessing human circadian rhythms.^13,32

Public health relevance

Over 15 million Americans work on evening, night, or rotating shifts or other irregular schedules.³³ Shift work is associated with a high prevalence of gastrointestinal problems; 47-82% of those working rotating or night shifts report nausea and other gastrointestinal symptoms.^1–3,5,34 While chronic diseases common among shift workers such as diabetes may contribute to the gastrointestinal problems, our results suggest that nausea is common even in healthy individuals who consume food during the biological night. Strategies such as scheduling the timing of meals and snacks to avoid the late biological night/early biological morning can be suggested for shift workers to counteract these problems.