Effect of lockdown on sleep-wake rhythm and alcohol use

ABSTRACT The global covid-19 pandemic has imposed radical changes in daily lives. This study reflects upon sociodemographic and clinical characteristics (sleep-wake rhythm, psychiatric symptoms, and alcohol use behavior) during the full lockdown, comparing individuals who increased their alcohol use (iAU), those who maintained a stable use (sAU), and those who did not consume alcohol (AnoU). Participants were recruited via e-mails and they were required to complete an online survey that included questionnaires, during the last week of the full lockdown. The iAU group, compared to the sAU group, presented more disturbed sleep (PSQI; p < .001), more severe insomnia (ISI; p < .001), shorter sleep duration (p < .001), longer sleep latency (p < .001), and less regular sleep-wake schedules (p = .005). They also reported more anxiety (HAD-A; p = .009), more depressive symptoms (HAD-D: p = .006) and more psychotraumatic symptoms (PCL-5: p = .018). Moreover, the sAU group, compared to AnoU, showed better quality of sleep (PSQI; p = .002) and less severe anxiety symptoms (HAD-A; p = .014). Maintaining a stable use was also related to a better quality of life associated with bigger homes with more frequent outdoors living spaces and higher monthly incomes. Individuals who increased their alcohol consumption during the Covid-19 lockdown exhibited more sleep and circadian rhythm disturbances, as well as more (severe) psychiatric symptoms.


Introduction
Emerging in December 2019 in Wuhan, the coronavirus disease 19 (Covid-19) rapidly spread throughout the world (Wang et al. 2020). Faced with this global pandemic, governments have had to develop strategies to limit the spread of infection, including worldwide use of lockdowns to limit contaminations. In France, the government has imposed a first national lockdown, from March 16th, to May 11th, 2020. Travelling was then limited to assisting vulnerable people, essential care, and work activity when telecommuting was not possible.
Alcohol remains the second leading cause of premature death in the world (Lim et al. 2012) with approximately 3 million deaths in 2018 (World Health Organization, Management of Substance Abuse Team, Global Status Report on Alcohol andHealth 2018., 2018). In France, daily alcohol users represent about 10% of the general population (Beck et al., 2015). Indeed, alcohol consumption has increased in the general population since the emergence of Covid-19 (Pollard et al. 2020) especially for individuals who received psychiatric treatments , for adolescents and young adults (Ahmed et al. 2020;Vanderbruggen et al. 2020), and for the most stressed students (Flaudias et al. 2021). European studies have shown that the increase in alcohol use seems related to changes in consumption habits such as consumption during videoconference or virtual meetings (Marsden et al. 2020). While limiting the transmission of the virus, the lockdown also had deleterious effects in terms of psychological suffering (Brooks et al. 2020) with an increase in anxiety (Flaudias et al. 2021) and depressive symptoms (Ahmed et al. 2020), as well as a progressive worsening of global psychiatric symptoms (Pandey et al. 2020). In addition, some psychiatric patients were infected with Covid-19 during lockdown, which has been associated with an increased risk of decompensation which may have been an additional risk factor for an increased consumption of alcohol (Taquet et al. 2021).
During the lockdown, French citizens had most of their life habits changed. They were faced with a modification of social rhythms (work and social times), exposure to light, and sleep-wake rhythms (Geoffroy et al. 2020a;Russell n.d.). They also were changes in diet, physical activity and substance consumption such as alcohol. The interaction between alcohol use and changes in sleep-wake rhythms has already been demonstrated. For instance, night workers or adolescents presenting shifted sleep-wake rhythms, tend to present an enhanced consumption. (Hasler et al. 2014;Swanson et al. 2016). These interactions seem to be bi-directional. Indeed, alcohol can initially be used for hypnotic effects, but it progressively worsens circadian and biological rhythms, as well as sleep architecture and quality (Koob and Colrain 2020;Meyrel et al. 2020) .To our best knowledge no studies have yet evaluated the relationship between changes in rhythms and changes in alcohol use during this peculiar period of a full lockdown.
In this context, we decided to conduct a survey in the general population to study the impact of changes in living conditions (work, social contacts and housing) during the first Covid-19 lockdown on chronobiological rhythms, emotions and substance use (Bertrand et al. 2021). The main objective of this study was to assess the relationship between changes in alcohol use patterns and changes in sleep and psychiatric symptoms during the first lockdown.

Population
This study is drawn from data collected by the Covid-RythmE survey (Bertrand et al. 2021). It is a crosssectional study conducted in the French general population. This is a convenience sample study, without exclusion criteria, with the main objective of identifying the effects of the two months of full lockdown (from March 17 th , 2020 -May 17 th 2020) on sleep, circadian rhythm and emotions.

Study design
Participants were asked to complete an online questionnaire (Table S1) delivered via email, Facebook or Twitter with newsletters from French sleep and biological psychiatric societies ("Société française de recherche et de médecine du sommeil (SFRMS) and "Association française de psychiatrie biologique et de neuropsychopharmacologie (AFPBN)). The online survey format made possible to remotely evaluate the impact of lockdown on socio-professional and biological rhythm changes (changes in eating, sleeping, and physical activity schedules). In the present analysis, we have focused on sleep-wake rhythms. The lockdown's impact on emotions and sleep was evaluated with standardized selfquestionnaires: The Hospital anxiety and depression scale (HAD) (Zigmond and Snaith 1983), the Posttraumatic stress disorder checklist-5 (PCL-5) (Blevins et al. 2015), the Insomnia severity index (ISI) (Bastien 2001), the Pittsburgh sleep quality index (PSQI) (Buysse et al. 1989) and the Epworth sleepiness scale (ESS) (Johns 1991). A more detailed description of the methods can be found on the Covid-RythmE primary study article (Bertrand et al. 2021). In addition, participants were asked to report their substance use. Pre-defined answers were: no substance, alcohol, tobacco, cannabis, cocaine, amphetamines, opiates, or other substances. Participants were also assessed for a possible increase in their substance consumption during the lockdown. For this study, because of our clinical experience during the Covid pandemic and the observation of changes in alcohol use in our patients, we a priori decided to divide the population into three comparison groups: participants who did not consume alcohol (AnoU), participants who maintained a stable alcohol consumption (sAU) and participants who increased their alcohol consumption (iAU).

Statistical analysis
Descriptive analysis was conducted for each of the collected parameters. Continuous variables were described as means with standard deviations (medians and quartiles if the distribution was asymmetric). Binary variables were described in absolute number and percentage. The characterization of sleep and psychiatric disorders within and between the groups of sAU, AnoU and iAU was studied using univariable and multivariable analyses. We conducted multivariable analyses, using a logistic regression, to confirm independent associations adjusting on age, gender, housing size, social contact and monthly income. For the multiparity analyses, we calculated the Odd Ratio (OR) and 95% confidence interval (CI95%) in order to assess the impact of alcohol on the different psychiatric and sleep symptoms. For continuous variables, the normality of the distribution was tested using the Shapiro-Wilk test, and equality of variances was tested using the Levene test. To compare results according to subgroups, for categorical variables we used a Chi 2 test, and for continuous variables we used a Mann-Whitney U t-test for non-normal distributions, and a student t-test for normal distributions. Statistical analyses were conducted using Jamovi 1.0.5 software. A significance level of 0.05 was used for all analyses.

General characteristics of the population
The survey had a response rate of 58%, among the 1627 participants. 74.3% of respondents (n = 1204) were women and 80% (n = 1301) were between 18 and 45 years old ( Table 1). The difference in the number of participants can be accounted to missing data Alcohol users groups (sAU and iAU), presented a higher proportion of men (sAU vs. AnoU: p < .001 and iAU vs. AnoU:p < .001). Their income was generally higher than 2000 € (sAU vs. AnoU:p < .001 and iAU vs. AnoU:p < .001), and they more often consumed cannabis (AnoU vs. sAU: p = .024 and AnoU vs. iAU: p = .0 40).
Compared to other participants, individuals of the sAU group had more frequently changed their accommodation during the lockdown (sAU vs AnoU = 0.015), more frequently had an outdoor living space (sAU vs AnoU: p = .002) and stayed in larger housing (sAU vs iAU: p = .050). Individuals of the iAU group smoked more frequently (iAU vs. sAU: p < .001) and had less history of depression (AnoU vs. iAU: p = .019).
Compared to participants who increased their alcohol use (iAU), there were more women in the non users group (NoAU) (AnoU vs iAU: p < .001), and respondents less frequently reported a salary above 2000 euros (AnoU vs iAU: p < .001) Furthermore, participants in the NoAU group more frequently reported an absence of psychiatric disease (AnoU vs iAU: p = .026). Finally, Table 1. Socio-demographic characteristics, from" sleep and circadian rhythm characteristics in individuals from the general population during the French COVID-19 full lockdown" (Bertrand et al., 2021 they reported less tobacco (AnoU vs iAU: p < .001) and cannabis use (AnoU vs iAU: p = .040) ( Table 2).

Sleep impact
Individuals who increased their alcohol use, compared to those who maintained a stable use during the lockdown, reported a more frequent alteration of sleep quality (sAU vs iAU: p < .001), shorter sleep duration (sAU vs iAU: p < .001) and greater loss of regularity of the sleep schedules (sAU vs iAU: p = .005) ( Table 3).
The PSQI assessment was congruent with these results, showing greater impairment of global sleep quality for individuals who increased their alcohol use compared to those who did not use alcohol. Non users participants had more frequent sleep disturbances than participants who maintained stable alcohol habits (Table 4).

Emotional impact
Regarding emotional impact, participants who increased their alcohol use during lockdown reported more anxiety symptoms (iAU vs sAU: p = .009), more depressive symptoms (iAU vs. sAU: p = .006) and more traumatic symptoms (iAU vs. sAU: p = .018) compared to those who maintained stable alcohol habits during the lockdown. Participants who did not consume alcohol had more severe anxiety symptoms than those who had a stable consumption (AnoU vs sAU: p = .014) ( Table 5).

Discussion
This cross-sectional observational study found that all participants whether they consumed alcohol or not, showed significantly altered sleep and habits during the first lockdown. We found that individuals who increased their alcohol use during the lockdown had more sleep disturbances and more altered sleep-wake rhythms than those who maintained their use during the lockdown. The individuals who have increased their use also presented more psychiatric symptoms (more severe anxiety, depressive and traumatic symptoms) than the other participants. Moreover, we found that the sociodemographic characteristics of the individuals differed between groups. Participants who maintained their use over time seemed to have a better quality of life (larger homes, more frequent outdoor living spaces, and higher monthly incomes) than participants who maintained their alcohol use and those who did not use The observation of an altered sleep quality for more than half of the population during the first lockdown is consistent with previous findings made in Italy (Marelli et al. 2021). We also found that participants who increased their alcohol consumption during lockdown had a poorer quality of sleep. This was expected since correlation between alcohol consumption and sleep alterations are well known and bidirectional (Chakravorty et al. 2016;Geoffroy et al. 2020c). In addition, we found an increase in severity of depressive and anxious symptoms during the first lockdown. These results are consistent with those found in the literature (Brooks et al. 2020). The rate of alcohol consumption found in this study during the first lockdown was higher than in other French studies: 41.3% vs 24.8% ). This previous French study assessed substance use, including alcohol, in the general population during the first lockdown and the characteristics of the participants in terms of socioeconomic level, gender, age, and psychiatric history were similar to those of the participants in our study. The main difference between those studies was the duration of assessment. Our study assessed alcohol consumption over a longer period (i.e., the two months of lockdown) whereas Rolland et al.
(2020)'s study assessed consumption between the 8th and 13th day of the lockdown. Therefore, it is possible that alcohol consumptions had continued to increase during the lockdown, as we observed an increase at the end of the period. Another hypothesis could be a switch of use between tobacco and alcohol, as the percentage of smokers in our study was rather lower (15.2%) than in Rolland et al.'s (2020) study (35.6%). Moreover, our study confirmed that the subjects most at risk of increasing their consumption were those with a higher level of education (Niedzwiedz et al. 2020;Rolland et al. 2020).
This study also emphasizes that participants who had fewer social contacts were those who did not drink. This is rather unusual, as the data from the literature usually suggests that individuals who don't use alcohol have more social contacts than individuals who do (Holt-Lunstad et al. 2010). In addition, participants who did not consume alcohol presented more severe anxiety symptoms than participants with stable use, whereas in the literature we find an association between anxiety severity symptoms and alcohol use disorder (Lai et al. 2015). This finding may be explained by the acute anxiolytic effect of alcohol, which quickly decreases with chronic use and disappears with the pharmacological tolerance leading to the risk of increase consumptions or comorbid psychiatric disorders such as depression (Carrigan and Randall 2003;Kushner et al., 1990). This may explain the worsening of anxiety and depressive symptoms in individuals who increased their alcohol consumption. Indeed, this is not a population of patients suffering from substance use disorders (as a reminder, only 1.6% of participants reported a history of substance    use disorder). In addition, the most socially isolated participants had more sleep problems. These results are consistent with another study that highlighted the worsening of sleep disorders and loneliness during lockdown (Grossman et al. 2021). Aa slight over-representation of higher-income was observed in our study (around 50% had >2000€ monthly income) compared to general French population (the median income in France is 1771 euros in 2018) (Insee -Données 2018 -© Observatoire des inégalités 2020). Thus we can hypothesize that a pleasant living environment, as well as a certain standard of living, might play a role as protective factors against the negative consequences of alcohol consumption and thus limit the loss of control over substance use (Henkel 2011). However, individuals who increased their alcohol use had a more disturbed sleep than other participants, and it is well known that alcohol has a direct and genuine impact on sleep quality and on circadian rhythms (Ebrahim et al. 2013;Geoffroy et al. 2020b;Meyrel et al. 2020). Furthermore, alcohol is an addictive substance (Volkow & Ting-Kai, 2005), and loss of control can be rapid and frequent (Beck et al. 2015). We could hypothesize that alcohol users who did not increased their consumption may be less vulnerable to stress and addictions. The sAU group of individuals may thus be closer in terms of biological vulnerability to the group of AnoU than to the individuals in the iAU group who increased their alcohol use. Also, we have shown that individuals who maintained a stable alcohol use during the lockdown more frequently had an outdoor living space in their home. We can therefore hypothesize that they were more regularly exposed to natural light, allowing them to have greater regularity in their circadian rhythm and to preserve their sleep-wake rhythm (Blume et al. 2019) This study has several limitations that should be acknowledged. First, it is a declarative study, with the information being self-reported by the participants without any medical validation. Therefore, the impact of stigmatization (Schomerus et al. 2011) and the desirability bias may have affected this study with underreported alcohol use for instance (Conigrave et al. 1995). Regarding substances in themselves, we only had access to qualitative information concerning consumption and its increase, without access to quantitative information (Kranzler and Soyka 2018) and we did not evaluate the severity of the substance use. Finally, our convenient population was not representative of the general population, and so may have misestimated some prevalence. Indeed, we had a larger proportion of women in our sample, which may impact the prevalence of alcohol for instance. Nevertheless, this cross-sectional study made during the full lockdown, allowed to gather initial information on a naturalistic cohort of participants. It found interesting relationships and associations between alcohol use and sleep and psychiatric symptoms. In addition, we used standardized questionnaires allowing a reliable and objective evaluation of psychiatric symptoms and sleep-wake cycles, with all assessments made during the same last week of the full lockdown.

Conclusion
This study highlights the major impact on lifestyle that the Covid 19 lockdown had. The increase in alcohol use during the lockdown had a negative impact on sleep, sleep-wake rhythms and anxiety and depressive symptoms. Nevertheless, beyond this increase in alcohol use, isolation and deprivation of social contacts seem to have been the most impactful factors on well-being and chronobiological rhythms.

Disclosure statement
No potential conflict of interest was reported by the authors.

Funding
The author(s) reported there is no funding associated with the work featured in this article.