Coronavirus disease 2019 outbreak : does con � nement have any impact on weight variation and weight change perception ?

Chadia Haddad (  chadia_9@hotmail.com ) Research Department, Psychiatric Hospital of the Cross, Jal Eddib, Lebanon. 2 INSERM, Univ. Limoges, CH Esquirol, IRD, U1094 Tropical Neuroepidemiology, Institute of Epidemiology and Tropical Neurology, GEIST, Limoges, France Maha Zakhour Faculty of Science, Lebanese University, Fanar, Lebanon Ghina Siddik Faculty of Science, Lebanese University, Fanar, Lebanon Rima Haddad Department of Linguistics and Philosophy, Uppsala University, Uppsala, Sweden Hala Sacre INSPECT-LB: Institut National de Santé Publique, Epidemiologie Clinique et Toxicologie – Liban. Beirut, Lebanon. Drug Information Center, Order of Pharmacists of Lebanon, Beirut, Lebanon Pascale Salameh INSPECT-LB: Institut National de Santé Publique, Epidemiologie Clinique et Toxicologie – Liban. Beirut, Lebanon. Faculty of Pharmacy, Lebanese University, Hadat, Lebanon. Faculty of Medicine, Lebanese University, Hadat, Lebanon


Introduction
With the new pandemic of coronavirus disease of 2019 (COVID- 19), people are restructuring their daily lives to adjust to the new normal of social distancing and isolation [1]. Quarantine is often disagreeable for those who are experiencing it [2]. Separation from loved ones, restriction of movement, doubt about the state of the disease, and boredom may often have drastic consequences [2]. Anxiety is rising among people as cities are self-quarantining, shops sell merely essential goods, and it becomes increasingly unclear when things might return back to normal [3].
Also, people in boredom situations are motivated to restore a sense of purpose in their engagement and turn to unhealthy eating behaviors such as craving certain food which help them cope with boredom [4,5]. People who are con ned to their homes will often try to adjust to their new life away from the accustomed daily schedule; they will, as a result, often turn to food as a necessity to reframe their new lifestyle in order to encounter the stressors of con nement and to feel a false sense of control [8]. Food may be an easy source of comfort for many individuals particularly since the usual comfort options have become less accessible [9]. Irregular eating and seeking unhealthy food is often considered as a coping mechanism for other mental health issues such as stress and anxiety [2]. Infection fears tend to increase the feeling of not being in control, sometimes managed by an extreme eating control behaviors [2]. Many will concentrate on their diet and their weight as a way to control their insecurity fears [6]. We have to note that people with eating disorder express more concern during this period and have a feeling of not being in control. These concerns are often managed with dietary restrictions or other extreme weight management habits or with episodes of binge-eating [2]. Moreover, during the con nement conditions, reduction of physical activity may lead to a sedentary behavior [10]. The lack of direct physical interaction with a supportive environment ampli es people's obsession with fear of weight gain [10,11]. Therefore, frequent proximity to food, unhealthy diet, limited physical activity, sedentary behavior, and increase of sitting time will contribute to weight change during con nement [10,11].
There is no doubt that a change in daily life routine to a state of self-isolation exacerbates many personal problems with weight, food, and an overall relationship with one's body [7]. During these stressful times, individuals are emotionally vulnerable, and there is greater challenge and concern about body image and weight obsession leading to an overwhelming feeling that goes along other imposed daily changes [12]. In fact, people with anxiety are more affected by weight change and weight change perception [13]. Consequently in response to stress, an individual might fail to accurately identify their actual weight status leading to an ineffective weight management [14]. Many people with a normal weight misperceive their weight status as overweight and individuals with overweight or obesity underestimate their body weight status [15]. Several studies demonstrate that people fail to identify their actual weight [16][17][18].
Weight misperception may have a negative in uence on levels of disordered eating and could modulate eating attitudes and behaviors depending on the perceived acceptability of the weight [19].
While some people manage to make the best of their con nement stay by using their time at home to reorganize their daily life, eat nutritious food or take extra time to exercise, other people, particularly those with or recovering from eating disorders, need frequent reminders to trigger them away from unhealthy eating behavior. To the best of our knowledge, there is no research that measures the impact of con nement on weight variation. Therefore, the rst objective of the current study is to evaluate the impact of con nement during the COVID-19 outbreak on weight variation. The second objective is to evaluate factors associated with weight perception among a sample of the Lebanese population.

Study design and sampling
A cross-sectional web-based online survey carried out between April 3 and 18, 2020, enrolled 407 participants selected from the general population in Lebanon. A questionnaire survey was distributed via social media (WhatsApp, Facebook, Instagram…), using a snowball technique. This online survey was posted on social media groups and sent by e-mail to potential participants. All people over the age of 18 with an Internet access were eligible. Out of the 1000 persons contacted, 407 responded, resulting in a response rate of 40.7%. The questionnaire required approximately 20 minutes to complete. The anonymity of the participants was guaranteed during the data collection process.

Procedure
The online survey consisted of a link to an internet-based questionnaire on Google forms with closedended questions in English and Arabic. Data from completed forms were imported into a Microsoft Excel spreadsheet and analyzed using the SPSS software, version 25.

Questionnaire
Two parts constitute the questionnaire, the rst one assessed the socio-demographic details of the participants (age, gender, marital status, educational level, employment status, region, and current value of monthly income (no income, low income <1,000 USD, intermediate income 1,000-2,000 USD, and high income >2,000 USD)), and their Body Mass Index (BMI).
The BMI was calculated by dividing the person's weight (in Kg) by the height in meters squared (m 2 ). Due to the con nement situation, the height and weight were estimated by the participants themselves. The variation of BMI was calculated by taking away the weight estimated during the con nement from the weight estimated before the con nement. The perception of weight change was assessed using one dichotomized question (positive/negative).
The second part of the questionnaire consisted of various scales used in the current study. These scales are described below.

Fear of COVID-19
Ten questions selected from previous studies were used to assess an individual's current fear of COVID-19 [20][21][22][23]. Examples of the asked questions include: "Thinking about COVID-19 makes me feel anxious", "I feel tense when I think about the threat of COVID-19", and "I feel quite anxious about the possibility of another outbreak of COVID-19". All items were measured on a 5-point Likert scale, from 1 (not at all) to 5 (extremely). The total score ranged from 10 to 50. High scores indicated a greater fear of COVID-19 infection. In this study, the Cronbach's alpha value was 0.917. By the time our data collection was completed, a study validating a fear of the COVID-19 scale was published [24], and thus could not be used in this paper.

Short Boredom Proneness Scale (SBPS)
The SBPS is a self-report questionnaire consisting of eight items rated on a 7-point Likert scale ranging from 1 (strongly disagree) to 7 (strongly agree) [25]. The total score ranged from 8 to 56. Higher scores indicated a greater tendency to boredom [25]. Permission to use the scale for the current article was obtained from the author of the questionnaire (Pr.James Danckert). In this study, the Cronbach's alpha value was 0.912.

Lebanese Anxiety Scale (LAS)
This 10-item self-report scale, recently developed and validated in Lebanon, was created to screen for anxiety [26]. Seven of the items are graded on a 5-point Likert scale (0 = not present to 4 = very severe) and the remaining three, on 4-point Likert scale (1 = almost never to 4 = almost always) [26]. The total score was obtained by summing all the responses, with higher scores indicating higher anxiety [26]. In this study, the Cronbach's alpha value was 0.884.

Anger subscale of the Buss-Perry Scale
The Buss-Perry Scale is a 29-item questionnaire composed of four factors that measure physical and verbal aggression, anger, and hostility [27]. In this study, the anger subscale (8 items) was used and was graded on a 5-point Likert scale from 1 (extremely uncharacteristic of me) to 5 (extremely characteristic of me) [27]. The total score was calculated by summing all the responses, with higher scores indicating a higher anger score. In this study, the Cronbach's alpha value was 0.865.

Eating Disorder Examination Questionnaire (EDE-Q)
The Eating Disorder Examination-Questionnaire (EDE-Q) is a 28-item self-reported tool measuring the range and severity of behavioral features of eating disorders [28,29]. It is rated using four subscales and a global score that re ect the severity of eating disorders. The four subscales are restraint, eating concern, shape concern, and weight concern.All items are scored on a 7-point rating scale (0-6), higher scores indicating greater levels of symptomatology [29]. In this study, the Cronbach's alpha values of the four subscales were as follows: restraint subscale (Cronbach's alpha = 0.835), eating concern (Cronbach's alpha = 0.745), shape concern (Cronbach's alpha = 0.902), and weight concern (Cronbach's alpha = 0.824).

Physical activity index
The physical activity index is a frequently used indicator of physical activity at a population level. This index is based on responses to a series of questions about the intensity, frequency and duration of participation in leisure-time physical activity. The Total Physical Activity Index is based on the multiplication of the intensity, duration and frequency of daily activity [30].
Translation procedure A forward and backward translation was conducted for all the scales except for the LAS-10 already available in Arabic. One translator was in charge of translating the scales from English to Arabic, and a second one performed the back translation. Discrepancies between the original English version and the translated one were resolved by consensus.

Statistical analysis
Data were analyzed using Statistical Package for Social Sciences (SPSS software version 25). A descriptive analysis was done using the mean and standard deviation for continuous measures and counts and percentages for categorical variables. Student t-test and ANOVA F tests were used to assess the association of continuous variables with two or more levels and Pearson correlation analyses were used for continuous variables. Chi-square and Fisher exact tests were used to assess the association of categorical variables.
Four linear regressions were conducted taking the declared BMI variation as the dependent variable. In addition, a backward logistic regression was conducted, taking the weight change perception as the dependent variable. In order to eliminate potentially confounding factors as much as possible, all variables that showed a p<0.1 in the bivariate analysis were considered as important variables to be entered in the model. A p<0.05 was considered signi cant. The Cronbach's alpha was used to assess scales' reliability.

Sample description
The socio-demographic characteristics of the participants are summarized in Table 1. The results showed that the mean age of the participants was 30.59 ± 10.10 years, with 51.3% females. The majority (90.9%) had a university level of education and were single (75.0%). More than half of the participants (56.6%) were employed, reported having a low monthly income (51.4%), resided in Mont Lebanon region (58.6%), and lived in cities (60.1%). The mean con nement duration was 26.05 ± 10.69 days. The mean BMI of the participants before the con nement was 25.02 ± 4.69 Kg/m2 and during the con nement was 25.08 ± 4.44 Kg/m2. No signi cant variation was found for BMI before and during the con nement (p = 0.40). Bivariate analysis: correlates of declared BMI variation A signi cantly higher mean declared BMI variation was found in married participants as compared to single participants. Moreover, higher anxiety, boredom, eating concern and weight concern were signi cantly associated with higher declared BMI variation. However, higher physical activity index scale and higher restraint were signi cantly associated with lower declared BMI variation (Table 2). Bivariate analysis: correlates of weight change perception Bivariate analysis taking the perception of weight change as the dependent variable is shown in Table 3. A signi cantly higher mean con nement duration, anxiety, anger, boredom, restraint, eating concern, shape concern and weight concern was found in participants with positive as compared to negative weight change perception. No signi cant difference was found between the socio-demographic characteristics and weight change perception.

Multivariable analysis
The results of a rst linear regression, taking the variation of declared BMI as the dependent variable and the socio-demographic features as independent variables, showed that having a university education (Beta = -0.55) and low income (Beta = -0.41) were signi cantly associated with a lower declared BMI variation ( Table 4, Model 1).
A second linear regression, taking the variation of declared BMI as the dependent variable and the fear of COVID-19, psychological scales and the socio-demographic characteristics as the independent variables showed that a higher anxiety score (Beta = 0.01) and higher boredom (Beta = 0.01) were signi cantly associated with greater declared BMI variation. However, having a university education (Beta = -0.49) and low income (Beta = -0.36) were signi cantly associated with a lower declared BMI variation (Table 4, Model 2).
A third linear regression, taking the variation of declared BMI as the dependent variable and the fear of COVID-19, eating behavior, physical activity and socio-demographic characteristics as the independent variables showed that higher eating concern (Beta = 0.24) was signi cantly associated with greater declared BMI variation. However, longer duration of con nement (Beta = -0.01), higher restraint (Beta = -0.22), higher physical activity (Beta = -0.006) and low income (Beta = -0.58) were signi cantly associated with a lower declared BMI variation (Table 4, Model 3).
A fourth linear regression, taking the variation of declared BMI as the dependent variable and the Fear of COVID-19, psychological scales, eating behaviors, physical activity and sociodemographic characteristics as the independent variables showed that higher eating concern (Beta = 0.21) was signi cantly associated with greater declared BMI variation. However, longer duration of con nement (Beta = -0.01), higher restraint (Beta = -0.21), higher physical activity (Beta = -0.006) and low income (Beta = -0.59) were signi cantly associated with a lower declared BMI variation (Table 4, Model 4).
A backward logistic regression, taking the negative/positive weight change perception as the dependent variable, showed that longer con nement duration (ORa = 1.07), higher anxiety (ORa = 1.05) and higher eating concern (ORa = 1.81) were signi cantly associated with greater weight change perception. A higher fear of COVID-19 score (ORa = 0.96) was signi cantly associated with lower positive weight change perception (Table 4, Model 5). Variable entered: Length of con nement, Fear of COVID-19 scale, anxiety, physical activity index scale, boredom, anger, EDE restraint subscale, EDE eating concern subscale, EDE shape concern subscale, and EDE weight concern subscale.

Discussion
To the best of our knowledge, this is the rst study to assess the factors correlated with declared BMI variation and weight change perception during the COVID-19 con nement among 407 Lebanese participants. Our results showed that 52.1% had a positive perception of weight change and no signi cant variation was found for BMI before and during the con nement, indicating no major weight changes during lockdown. These results indicate a discrepancy between perceived weight status and actual weight status demonstrating a weight misperception among the participants. Other studies have reported a positive correlation between weight perception and actual weight variation [31][32][33]. The selfreporting errors of weight in the current study might have underestimated the actual body size. In fact, many individuals tend to underestimate their weight and overestimate their height [34].
Additionally, the absence of declared BMI variation in our results might be explained by the fact that people are adapting some healthy habits and are bene tting from the lockdown to take care of their eating habits and daily exercise.. Evidence suggests that variation in weight can be detected 12 weeks after starting a diet and performing exercise modi cation interventions [35]. This was not the case in our study, however, since the mean duration of the con nement was 26 days, nearby four weeks, which is considered a short time to detect any effective weight change.

Declared BMI Variation
The results of the multivariate analysis showed no signi cant association between fear of COVID-19 and weight variation. Like SARS, fear of COVID-19 is signi cantly correlated to depression and anxiety [37]. Many people are experiencing higher fear and anxiety reactions to the COVID-19 that may contribute to weight change through variation of eating habits [38]. Furthermore, long terms of social isolation increase depression and anxiety [39]. Few studies show that anxious people tend to eat less than non-anxious people even when they are hungry [40,41] where as other studies reveal that anxiety and fear are associated to overeating and so to high BMI [42]. Apparently, no studies show a correlation between fear of COVID-19 and BMI variation. Only a couple of studies show that depression, anxiety and emotional eating predicted higher BMI [43,44]. Indeed, mood disorders were positively associated to weight gain: People dealing with mental health problems are much more likely to indulge in low levels of physical activity and eat poor quality of food, which result in the rise of body weight [13]. Evidence shows that in times of stress, people are more likely to seek energy food such as (sugars, fats and carbohydrates) that act like a tranquilizer to calm the stressful situation [45]. This energy food intake could be stored as body fat and can thus contribute to excess energy intake and subsequent weight gain [46]. Also, longer duration of con nement was correlated to lower declared BMI variation. Social isolation can affect differently in individuals. Eating or living alone may be associated with higher prevalence of obesity and unhealthy eating behaviors [47]. Eating with others can protect against nutritional issues such as obesity, unhealthy eating, and disordered eating [48,49]. The parameter of living alone or with others should be studied further to understand the association between the duration of con nement on BMI variation. We have to note that the data collection of the current study was realized within one month of con nement. Future studies should check the association between fear of COVID-19 and duration of con nement with BMI variation.
Also, the results of the multivariable analysis revealed that higher restraint eating was related to lower declared BMI variation. Most of the previous studies show positive [50][51][52], negative [53] or no signi cant [54] associations between dietary restraint and body weight change or BMI variation. Dietary restraint was shown to play double roles either as a response for weight gain [51,55] or as a protector for weight gain [56]. Studies show that exible dietary restraint is negatively associated with BMI while rigid dietary restraint can be either positively or negatively associated with BMI [56][57][58]. The type of restraint eating used by the participants and the level of the dietary restraint should be de ned to clarify the correlation between restraint eating and BMI. Low Dietary Restraint was associated with increased BMI, while high restraint eating was associated with low BMI [57]. Regarding eating concern, results showed a positive correlation with BMI variation. In fact, dieting is a major factor for weight gaining and obesity [59,60].
Indeed an increase in weight was associated with more eating concern [61,62].
Higher physical activity was related to lower declared BMI variation. There is a general agreement that a sedentary lifestyle is one of the most prominent risk factors for the increase in BMI. So, any kind of physical activity is bene cial in avoiding weight gain [63,64]. Garimella et al. showed that physical inactivity is one of the factors that closely correlates with obesity [65]. Individuals that perform physical activity tend to maintain their weight comparing to non-performed physical activity individuals [66]. High physical activity people reported less weight gain than those who practice low physical activity [67]. In addition, the intensity of physical activity did not matter in lowering or maintaining BMI [68]. Other factors or habits may compensate the effect of physical activity and do not mean the more hours of exercise, the lower BMI [69].

Weight perception
The results of the multivariable analysis revealed that greater fear of COVID-19 was associated with lower weight change perception whereas higher anxiety was associated with greater weight change perception. In a time of uncertainty caused by the COVID-19 outbreak, many people are experiencing a higher fear and anxiety reactions that may contribute to weight change through variation of eating habits [70,71]. People experiencing negative emotions such as fear and anxiety would have a pessimistic perspective and perceive everything differently than the normal emotional state [72]. Studies relating the association between psychological distress and weight perception are very limited. The directional association found in several studies had related weight perception to psychological distress and showed a stronger association between weight misperceptions and anxious/depressive symptoms [73][74][75]. The mechanisms that linked weight perception to psychological distress suggest that people engaging in selfevaluation had a poor body image and may often feel isolated or discriminated in social situations that in turn contribute to the feelings of distress and depression. The reasons for the association between the fear of COVID-19, anxiety and weight perception are still unclear and more cohort studies are needed to con rm causality.
By examining the eating concern among participants, we were able to see that eating concern is signi cantly associated with perceived weight change. Studies show that eating concerns are associated with the accuracy of the body weight perception [14,76,77]. During this period of uncertainty people might have more concerns about eating and they are seeking unhealthy food as a way to cope with the stressful situation [12]. People concerned about their body and eating behavior might be very conscious about their body weight [78]. Consequently, an increased tendency to perform unhealthy weight control behaviors such as dieting and excessive exercise are performed to prevent weight gain [12].
Longer duration of con nement was signi cantly associated with perceived weight change. Prolonging con ned might cause mental health issues as showed in previous studies [79][80][81]. However, people's reaction to self-quarantine may vary, some of them may be better able to cope with stress and others may be overwhelmed with negative feelings. Some people might start adapting new habits and create eating habits that might result in effective weight perception and maintenance. However, the association between con nement duration and weight perception is still unclear and more studies are needed.

Limitation
The main limitation in the current study is the inability to justify the real variation in weight of the participants as it was self-estimated. The perception of weight was assessed by using a single question not a validated scale. The sample may not be representative to the entire group of con ned persons in Lebanon as the actual number of respondents is low compared to the total number of con ned persons.
An information bias could exist since the study questionnaire was based on an online survey and answers were self-reported. Self-reported replies may be exaggerated; respondents may choose to omit or not reveal private details. A selection bias might have occurred since the sample was not randomly selected but rather gathered by using the snowball sampling technique. Individuals selected to be studied had recruited new participants from within their network of friends thus it is likely that all participants might have the same characteristics or traits. The majority of the participants were well educated with computer literacy and Internet access and the less educated people and those unable to access the Internet were not assessed. Residual confounding bias is also possible, since there could be factors related to weight perception and BMI that were not measured in this study.

Conclusion
Our main results indicate that no signi cant change in weight was detected before and during con nement. The fear of COVID-19 was not related to weight change, however, a negative association was found with weight perception. Also, longer duration of con nement were associated with a lower declared BMI variation and greater weight change perception. These results are a step forward to do more studies concerning, not only weight change in con nement, but also the long-term effects of con nement on physical and psychological health. Also, our results suggest that implementing con nement during a virus outbreak should go alongside political and public health decisions that need to promote information Declarations