A population-level analysis of changes in diel rhythms and sleep and their association with negative emotions during the outbreak of COVID-19 in China


 Introduction: From December 2019, COVID-19 (novel coronavirus pneumonia) began spreading in China and has significantly affected the industrial economy and peoples’ daily lifestyle. Beginning on January 23, the public was asked to constantly stay at home for quarantine and community containment.Methods: To assess the effects of the changes in diel rhythms and sleep and their association with negative emotions during the COVID-19 outbreak, a questionnaire was administered to 451 responders for analysis between January 20, 2020, and January 31, 2020, in China.Results: We found that 34.6% of the participants reported diel rhythm disturbance. Moreover, 67.2% of the participants presented negative emotions regarding the pandemic situation, including worry, fear, downheartedness, anxiety, depression, and stupefaction; among them, worry was the most prevalent. Gender and age were significant factors for changes in the diel phases and emotions. There was a correlation between diel rhythm alterations and negative emotions. Three factors, i.e., the Spring Festival holiday, quarantines and concern regarding the pandemic situation, were associated with changes in diel rhythms, sleep, and negative emotions during the pandemic period. Holiday jet lag, quarantine (or community containment), and concerns regarding the pandemic situation had significant effects on diel rhythms, sleep and negative emotion in a substantial part of the population. Conclusions: Our findings suggest that diel rhythms and sleep and their association with negative emotions in COVID-19 patients and the normal population need to be considered. Moreover, the adjustment of diel rhythms could help relieve negative effects and improve the global health during the pandemic period.

Introduction COVID-19 (novel coronavirus pneumonia) is a newly emerging severe acute respiratory infectious disease that is caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) [1]. To suppress COVID-19 spread in China, 31 provinces, autonomous regions and municipalities declared a level 1 health emergency; furthermore, the public was required to stay at home for quarantine and community containment [2,3]. This resulted in extensive lifestyle changes with the public continuously staying at home for long periods since January 23, 2020, and only leaving discreetly to shop for food and necessities was allowed.
The circadian clock controls the rhythmicity of physiology and behavior, which endows the adaptability to cycling environmental cues with a period of 24 h [4,5]. Circadian rhythms display over a period of 24 h, i.e., entrained by natural daily environmental cycles, which are called diel, diurnal or nyctohemeral rhythms [6]. Moreover, circadian rhythms are affected by social cues, including working schedule, shift work, and jet lag [7]. Circadian rhythm disturbances cause disorders, metabolic diseases, and decreased immunity [8]. Circadian desynchronization is related to several mood disorders, including depression, seasonal affective disorder, and bipolar syndrome [9][10][11][12][13]. Con nement is also a social factor that has The Kolmogorov-Smirnov test was conducted to assess the normality of samples [18]. The Wilcoxon signed-rank test was performed to assess between-period differences in diel rhythms and sleep [19]. Between-group differences were analyzed using the Mann-Whitney U test with statistical signi cance set at P ≤ 0.05 [20]. Bivariate (Spearman) correlation analysis was used to assess the between-variable correlation for each comparison [21]. SPSS 26 (IBM, USA), GraphPad Prism 8 (Graphpad Software Inc., USA), and RStudio 3.6.1 (RStudio Inc., USA) were used for statistical analysis and picture drawing. P ≤0.05 was considered signi cant with the signi cant values being * P <0.05 and ** P <0.01. Additional details regarding the method and the results of correlation coe cient analysis can be found online in the suppl. protocol.

Results
Changes in the diel rhythm, sleep-wake cycle, dining and exercise We found that 85.6% (386) of participants self-reported changes in at least one variable related to diel rhythmicity. The sleep schedule became irregular and regular in 31.0% (140) and 14.6% (66) of the participants. To assess the changes in the diel phases, we classi ed the participants into three chronotypes based on their sleep onset time during the survey period as follows: early-type, slept before 22:00; intermediate type, slept between 22:00 and 24:00; late-type, slept after 24:00. In addition, irregular type denotes people who showed no xed sleep time. Both the time of falling asleep and waking up are diel phase references. In this study, we used the former parameter to assess the changes in the diel phase since there was a signi cant positive correlation between both parameters (suppl. le 3).

Analysis of changes in sleep duration and quality
The questionnaire was used to assess the quantity and quality in terms of daily total sleep duration, daytime nap duration, nocturnal wake-up times, and KSS-rated sleepiness. During the quarantine period, 59.7% (269), 3.1% (14), and 37.3% (168) of the participants reported an increase, decrease, and no change in the daily total sleep duration during the quarantine period (Fig. 2F). Nocturnal wake-up episodes were associated with sleep quality and caused sleep fragmentation and sleep cycle disturbance. The female gender was positively associated with KSS-rated sleepiness (r=0.133, P=0.005). Education level was positively associated with prolonged daytime napping (r=0.154, P=0.001). Age was positively associated with increased waking episodes (r=0.122, P=0.011) and prolonged daytime napping (r=0.236, P< 0.001) and negatively associated with KSS-rated sleepiness (r=-0.115, P=0.015) ( Fig. 2G and suppl. le 3).
Compared to that in the control period, there was an increase in the incidence of nocturnal waking episodes (53, 11.8%) during the pandemic period. An assessment of KSS-rated sleepiness showed that most participants had a score of 5 (33.9%, 153). In addition, 123 (27.3%) participants had a score of 1-3, 319 (70.7%) had a score of 4-7, and 9 (2.0%) had a score of 8-10 ( Fig. 2E). Furthermore, 12.6% (57) reported a subjective decrease in sleep quality (Fig. 2F). Before January 20, 2020, 27.9% (126) of the participants did not have a daytime napping habit; among them, 80.2% (101) maintained this habit after the emergency launch. Some of the participants with previous daytime napping did not report them during quarantine. The proportion of participants without daytime napping habits during the quarantine period was 49.0% (221) (Fig. 2F). Exercise was negatively associated with prolonged total sleep duration (r=-0.157, P=0.001) ( Fig. 2G and suppl. le 3).

Subjective assessments of the effect of quarantine on diel rhythms and sleep
We assessed three potential factors: the Spring Festival holiday, concerns regarding the epidemic situation and constant quarantine and community containment. In this subjective assessment, the effects on diel rhythms and sleep were considered to be one entity. The responders reported the subjective assessments of these factors on diel rhythms and sleep using the corresponding scales 1-5. Among the 451 participants, 89.6% (404), 76.7% (346), and 88.0% (397) of the participants considered constant quarantine, holiday, and community containment, as the causative factors, respectively (rating ≥ 2) for diel rhythms and sleep disturbance (Fig. 4C,D and suppl. le 8). Moreover, 80 participants (17.7%) listed that other factors could affect diel rhythms and sleep including family (38.8%, 31/80), work or schoolwork (30%, 24/80), environmental comfort (17.5%, 14/80), and economic pressure (2.5%, 2/80). Regarding environmental comfort we found that noise (3.8%, 3/80) and ambient temperature (2.5%, 2/80) could contribute to effects on diel rhythms and sleep (suppl. le 8).

Discussion
In this study, we surveyed the changes in diel rhythms and sleep of 451 Chinese people during the traditional Spring Festival who had experienced quarantine and community containment for approximately two weeks during the outbreak of COVID-19. We show that from January 20, 2020, to January 31, 2020, the diel rhythms, sleep and negative emotions of a substantial part of the population were subject to change, owing to at least three factors: holiday, quarantine and concerns regarding the pandemic situation. In this study, we performed correlation coe cient analysis to address the possible associations among demographic characteristics, environmental factors, variables related to diel rhythms and sleep, and negative emotions.
Over the study period, there were changes in the diel rhythms of 386 individuals (85.6%) with 36.3% of these individuals developing irregular diel rhythms and sleep disturbance. Approximately 49.7% (224/451) of the responders showed a delayed wake-up time (Fig. 1D). There were sleep changes during the quarantine period, which was indicated by several parameters. First, 59.6% (269/451) of the participants showed increased total sleep duration (Fig. 2F). Nocturnal wake-up episodes are indicative of sleep disturbance; in our study, 47 (10.4%) participants reported increased nocturnal wake-up episodes, which was negatively associated with subjectively self-assessed sleep quality (suppl. le 3). There are close correlations among circadian rhythms, sleep, and emotions [9][10][11][12][13]. Our ndings revealed extensive diel rhythm and sleep alterations during the pandemic period, which were negatively correlated with the six negative emotions (Fig. 4B).
Cain et al. analyzed diel changes in serum melatonin and core body temperature and reported that compared with men, women displayed more advanced phases of these variables [22]. Women have a shorter free-running period than men, which indicates gender-based differences in circadian patterns [23].
Consistently, in this survey female participants reported an earlier time of waking up and falling asleep than men did before January 20, 2020 (suppl. le 9). In contrast, after January 20, 2020, female participants reported a later time of waking up and falling asleep; moreover, they had higher ratios of delayed wake-up time and di culty falling asleep than men did (Fig. 2C and suppl. le 9). There was a positive correlation between female gender and irregular sleep-wake cycles (Fig. 2D). These ndings suggest that women could be more susceptible to the effects of the holiday, quarantine or concerns regarding the pandemic situation on the assessed parameters.
The majority of the responders (67.2%, 303/451) reported at least one negative emotion type, with worry being the most prevalent (57.0%; 257/451). Across all negative emotions, 7.1%-16.4% of participants reported very high scores (scales ≥ 4) (Fig. 3A,B and suppl. le 7), which suggests that a substantial part of the population had negative emotions. Circadian misalignment causes emotional deterioration [24][25][26]. We observed a negative relationship between fear and delayed waking up (Fig. 4B). The negative emotions could be more attributed to the pandemic situation than to the Spring Festival holiday.
Older individuals have been shown to have less negative emotions, which could be attributed to increased ability to regulate emotions [27,28] Consistent with these previous studies, we observed a negative correlation between age and worry, stupefaction, downheartedness, anxiety, and fear ( Fig. 3D and suppl. le 3). Exercise was negatively related to the degree of diel rhythms and sleep components (Figs. 2D, G and suppl. le 3), which suggests that appropriate exercise might improve sleep and diel rhythms. However, exercise was not associated with the six negative emotions (Fig. 3D and suppl. le 3).
Social jet lag caused by night work, shift work and transmeridian travel could cause circadian misalignment [29]. The traditional Spring Festival holiday lasted between January 24, 2020, and February 2, 2020, while the level 1 emergency was launched on January 23, 2020. Therefore, the periods for the three factors (Spring Festival holiday, quarantine, and concerns regarding the pandemic situation) might overlap.
To discriminate the effects of these three factors on diel rhythms and sleep, we analyzed their relationships with subjective self-assessed variables. We found that 63.2% (251/397) of the participants reported the holiday as a causative factor for diel rhythm and sleep disturbance. Moreover, a coe cient analysis revealed positive correlations between the Spring Festival and total sleep duration, sleep quality, KSS-rated sleepiness, delayed wake-up time, and self-reported changes in diel rhythms. However, among the six negative emotions, only worry was positively correlated with Spring Festival holiday ( Fig. 4E and suppl. le 3), which suggests that holiday was not the primary cause of negative emotions.
Compared with the holiday, quarantine was correlated with more negative emotions, including worry, stupefaction, downheartedness, anxiety, and fear. Furthermore, 89.6% of the participants (404/451) reported that quarantine affected a number of diel rhythm components, including the growing irregularity of diel rhythms, di culty falling asleep, delayed wake-up time and dining times. In addition, 346 (76.7%) of the participants reported that concern about the pandemic situation was signi cantly related to several sleep and diel rhythm variables, as well as all six negative emotions ( Fig. 4E and suppl. le 3).

Limitations
This study was conducted in the very early stage of pandemic outbreak hurriedly, bias may exist due to the limited quantity of the responders. This is the main limitation of this study implying that some of our ndings should be interpreted with caution. For instance, though the overall negative association between age and negative emotions has been revealed, the quantity of participants who were < 20 years or > 60 years were very few (suppl. le 6). Therefore, the situation of these two parts of population remains unclear.

Conclusions
Our ndings show that quarantine and concerns regarding the pandemic situation contributed to the changes in sleep, diel rhythms, and negative emotions. Moreover, some factors, e.g., gender and age, might be causes for of the difference in some of the changes. These ndings suggest that people who are not COVID-19 patients may also be subject to extensive in uence. More than 4.0 million cases of COVID-19 have been reported worldwide as of May 14, 2020, and pandemic surveillance may last for additional months or even years [30]. Our ndings could contribute to improvements in diel rhythms, sleep, and emotion during the pandemic period.