Eating habits and lifestyle behaviors during COVID-19 lockdown: The Lebanese experience

Objectives This study aimed to assess dietary intake and lifestyle habits followed during the COVID-19 pandemic and subsequent lockdowns, as well as the level of adherence to the Mediterranean diet (MD), among a sample of the Lebanese population. Methods A cross-sectional study was conducted during the government-enforced lockdown. A validated, online, questionnaire was used to collect information about dietary and lifestyle habits. The Mediterranean diet adherence screener (MEDAS) was used to assess adherence to the MD. Results A total of 1684 participants responded to the survey. Their mean age was 23.92 ± 7.62 years, and 70.4% were females. Approximately a third of the participants reported that their dietary habits did not change and 42.3% acknowledged that their eating habits deteriorated during the lockdown. Participants smoked less and slept more during the lockdown compared to before it. Approximately 19.2% of the sample reported low adherence to the MD whereas 63.9% and 16.9% described moderate and high adherence respectively. Only age was significantly associated with higher MD adherence. Conclusion Dietary intake and MD adherence were suboptimal during the COVID-19 lockdown among the sample of the Lebanese population. It is critical that the Lebanese government implements public health programs in order to promote awareness about the importance of adhering to a healthy lifestyle and making appropriate dietary and lifestyle choices.

implements public health programs in order to promote awareness about the importance of adhering to a healthy lifestyle and making appropriate dietary and lifestyle choices.

Background
In 2019, the SARS-CoV-2 virus emerged and quickly reached pandemic proportions. It manifested as a severe acute respiratory syndrome and the condition became known as COVID-19 [1]. Because of this highly contagious virus, countries all over the world resorted to lockdown and quarantine practices with the hope of controlling infection rates and preventing healthcare systems from collapsing.
Consequently, people had to stay home, work remotely when possible, and students were educated through virtual platforms. However, such preventive measures were not without consequences. It is well known that lockdowns, enforced through physical distancing and self-isolation, have an impact on dietary intake and lifestyle habits, which have an impact on health such as smoking, alcohol consumption and sleep. Such effects are not always positive [2,3]. Additionally, social distancing and isolation have a negative effect on mental health, which can be manifested through increased anxiety, depression, poor sleep quality and decreased physical activity [4].
Accordingly, alterations in macronutrient intake can result. It has been observed that stressed and bored individuals [5] may have an augmented preference for energy-dense foods, particularly those high in sugar and fat [6] since the intake of simple carbohydrates can have a positive effect on mood as it increases serotonin production [7]. Furthermore, lockdowns often result in decreased physical activity [8], which further increases stress levels and exacerbates health problems [9].
In Lebanon, the government adopted strict lockdown practices, with the aim of controlling COVID-19 infection numbers and preventing the virus from spreading. The country was simultaneously going through a severe economic crisis which impacted citizens at every level [10]. Even before such circumstances, the Lebanese people had already shifted towards consuming an atherogenic diet coupled with the alarming increase in nutrition-related cardiovascular risk factors [11].
Thus, it is of no doubt that following a healthy dietary pattern, during such turbulent times, is necessary to ward off disease and decrease inflammation and oxidative stress [12]. There is an abundance of evidence that proves that following the Mediterranean diet (MD) can be key in fighting immune-mediated inflammatory responses [13]. The MD has proven to be one of the healthiest and most sustainable diets for both the prevention and treatment of a myriad of diseases, notably cardiovascular disease and inflammation. It is a diet that is low in saturated fat and simple sugars and simultaneously high in plant protein, whole grains, and monounsaturated fats [14].
Numerous studies have assessed adherence to MD before and during the pandemic [15][16][17][18]. The Mediterranean Diet Adherence Screener (MEDAS) is one of the most commonly used tools to assess MD adherence [19][20][21][22]. It was originally created for use in the PREDIMED study and has proven to be a valid and reliable tool to address MD adherence among different populations. The screener consists of 14 questions, 12 of which inquire about food consumption and the other two about food habits as related to MD [23]. Each of the 14 items is scored with either 1 point to indicate adherence or 0 point to indicate non-adherence. Higher scores indicate higher adherence [24]. During the lockdown, a study conducted in Cyprus proved moderate adherence to MD [16], whereas, the highest scores during the pandemic lockdown were found among the Spanish population [17]. Additionally, an Italian study revealed a moderate adherence to the MD (scores ranging between 6 and 9) in the majority of respondents [3]. Similarly, Czarniecka-Skubina and colleagues estimated that among a sample of Polish adults, half changed their dietary habits significantly secondary to the pandemic [25]. These score differences may be due to a variety of geographical factors as well as sociocultural characteristics of each population influencing the implementation or adherence to the MD in each country [26]. To our knowledge, no previous study assessed MD adherence among the Lebanese population, especially during an atypical period such as domiciliary confinement worsened by an economic burden in the country.
Therefore, our study aimed to identify the dietary intakes among a Lebanese sample as related to the MD diet during the COVID-19 pandemic and subsequent lockdowns. It also intended to assess other health habits such as smoking and sleep. Such information would be vital to stakeholders, especially public health professionals, so they can create evidence-based public health programs, specifically directed towards the Lebanese population should similar circumstances arise in the future.

Ethics approval and consent to participate
The research was carried out in strict accordance with national and international standards, as well as the Declaration of Helsinki (2000). Before participating in the study, all participants were fully informed about the study's objectives and were required to approve the consent form. Agreeing with the statement and completion of the entire questionnaire was regarded as informed consent to participate. Participants filled out a survey that was directly linked to the Google platform. To protect and preserve confidentiality, participants' personal information was anonymized, which prevents sensitive personal data from being traced in any manner. The study was approved by the Research and Ethics Committee of the School of Pharmacy at the Lebanese International University.

Design, setting, and participants
A cross-sectional study was conducted between the months of January and February 2021, when the Lebanese government started enforcing a nationwide lockdown and round-theclock curfew, in an attempt to curb the spread of coronavirus infections that got out of control following the holiday season. An anonymous online, validated questionnaire was used to collect information about people's eating and lifestyle habits during the lockdown in Lebanon's eight governorates (Mohafazat). The inclusion criteria were that participants were at least 16 years old whereas the study excluded those who were below 16 years of age.

Minimal sample size calculation
According to the G-power software (R 2 deviation from zero), based on an effect size f 2 = 0.02, a 95% confidence level, a 95% power, and a total of 10 factors to be entered in the multivariable analysis, a minimum sample size of 1229 people was required to provide adequate statistical power.

Data collection
Google forms was used to create an online survey, which was conducted in Arabic. The link to the questionnaire was forwarded to the investigators' personal connections and was spread through institutional and private social networks. Unlike probabilistic sampling, this form of administration produces a statistical collective whose population parameters cannot be controlled. It was, however, entirely beneficial for the research objectives since it allowed for the widespread distribution of the survey questionnaire during a time when there were significant territorial limits due to the pandemic.

Questionnaire
The questionnaire was created using Google Forms (https://docs.google.com/forms/d/ 16Gw17zFbRklQDQi7na6FDv7QmlmSt6Uh1Z2v0gUDNz4/prefill) and was adapted from a study conducted by Di Renzo on eating habits and lifestyle changes in Italy [15] (questionnaire publicly available in the before-mentioned reference). The questionnaire was divided into three parts and included 46 questions divided as follows: • • Part C: Lifestyle habits including grocery shopping, smoking, sleeping hours, and physical activity The MEDAS questionnaire was used to determine the level of adherence to the MD. Participants were divided into three groups based on their MEDAS scores: (i) low adherence (score 0-5), (ii) medium adherence (score 6-9), and (iii) high adherence (score �10) to the MD, with differences in compliance rates determined for each food.

Statistical analysis
Statistical analysis was performed using SPSS v.22 (IBM, Chicago, IL, USA). Cronbach's alpha value was calculated for the MEDAS scale. There were no missing data since all questions were required in the online form. The normality of distribution of the MEDAS score was confirmed via a calculation of the skewness and kurtosis; values for asymmetry and kurtosis between -1 and +1 are considered acceptable in order to prove normal univariate distribution [27]. The marginal homogeneity test was used to examine the relationship between categorical variables, whereas the Student t-test and ANOVA were done to look for differences between two and three or more means respectively. Pearson's test was used to correlate two continuous variables. Finally, a linear regression was conducted taking the MEDAS score as the dependent variable; independent variables entered in the final model were those that showed a p<0.25 in the bivariate analysis [28]. P<0.05 was considered statistically significant.

Results
A total of 1684 participants enrolled in the study. Their mean age was 23.92 ± 7.62 years, with 70.4% being females. More details about the sociodemographic characteristics are reported in Table 1. The Cronbach's alpha value for the MEDAS scale in this study was 0.43.
The mean MEDAS score was 7.42 ± 2.26 (media = 7); 323 participants (19.2%) had low adherence to the MD, whereas 1076 (63.9%) and 285 (16.9%) respondents had moderate and high adherence respectively. Details about the number of portions consumed per day of certain food items during the lockdown are summarized in Table 2.
The results revealed that 556 participants (33.0%) reported no change in their dietary habits, 713 respondents (42.3%) admitted that their eating habits became less healthy, whereas 415 (24.6%) believed that their eating habits improved. Description of the food items that were consumed more often during the lockdown is summarized in Table 3. Participants reported increased consumption of fish during the lockdown most (38.4%), followed by sweets, fruits, nuts and meat products respectively.
A significantly higher percentage of participants reported smoking less than five cigarettes per day during the lockdown as compared to the period before the pandemic, whereas a significantly higher percentage of participants acknowledged sleeping for more than 9 hours per night during the lockdown (Table 4).

Bivariate analysis
Older age was significantly but weakly, associated with a higher MEDAS score (r = 0.05; p = 0.035), however, the MEDAS score was not significantly associated with any of the other sociodemographic characteristics collected (Table 5).

Multivariable analysis
A stepwise linear regression, taking the continuous MEDAS score as the dependent variable, was conducted; variables entered in the model were age and district of residency. The results showed that older age (Beta = 0.02; p = 0.035; 95% CI 0.001-0.03) was significantly associated with a higher MEDAS score.

Gender differences
Mean body mass index (BMI) score was significantly higher in males when compared to females (25.89 ± 5.42 vs 22.60 ± 3.82; p<0.001), but no difference was found between genders in terms of the MEDAS score (7.45 ± 2.41 vs 7.41 ± 2.19; p = 0.742).

Discussion
The current study provides valuable insight about dietary intake, MD adherence and lifestyle habits during the lockdown that was imposed because of the COVID-19 pandemic in Lebanon. Significant changes in sleep quantity and dietary intake were reported. The only variable that was found to be associated with MD adherence was older age. More than 40% of our study sample reported that their dietary intake had changed negatively because of the pandemic and its restrictions, whereas almost a quarter of the participants described positive improvements in their diets. Our findings are in line with several other studies that have been conducted in different parts of the world. For example, Di Renzo and colleagues reported that amongst a sample of Italians, 37.2% felt that they were following inferior dietary intakes and 16.7% felt that they had improved their dietary habits, due to lockdowns that had taken place because of COVID-19 [15]. Additionally, in a study that took place in Kuwait, Husain & Ashkanani indicated significant changes in eating practices among their Kuwaiti sample [29]. Furthermore, Radwan et al. also reported that 31.8% of Emiratis increased food consumption [30]. Cheikh et al also reported that close to 40% of the Lebanese participants in their study were not consuming fruits and vegetables on a daily basis [31], while Dimassi et al reported a significant increase in consumption of fruits and vegetables [32]. Nevertheless, as a result of the pandemic, almost 40% of the participants have reported weight gain almost [33]. Unhealthy dietary patterns and weight gain are especially problematic during a pandemic since healthy dietary habits are the cornerstone to enhanced immunity, which could help ward of the severe consequences of viruses [34]. We also found that participants in this study reported smoking less during the pandemic. This is similar to the results reported by Di Renzo and colleagues who also described that smoking frequency was reduced in their study sample during the lockdown [15]. This is a positive outcome that may be attributed to the smokers' concern about COVID-19-related respiratory distress, which led to lower exposure to smoking [35].
Additionally, we found that participants reported sleeping more during the pandemic. This could be due to lockdowns and restrictions that were imposed and changes that resulted in virtual workplaces. Getting adequate sleep may be a positive finding since improved sleep improves immunity [36]. It may also be a negative finding since increased sleep hours are associated with psychiatric disease conditions and a higher BMI [37]. Unfortunately, we did not assess sleep quality and our findings are based simply on the number of hours of sleep reported during the pandemic. Likewise, Di Renzo and colleagues also reported increased sleep hours during lockdown among the study participants [15]. Future studies should focus on assessing the quality of sleep during a lockdown and its effect on dietary intake and immunity.
In terms of the MD dietary pattern, only a small number of participants reported high adherence and only age was positively correlated with a higher MEDAS score, which is a common finding in the scientific literature [17,[38][39][40]. However, increased adherence to the MD diet was not inversely correlated with BMI which was a surprising finding since many studies have reported contradicting results [38,41,42]. We hypothesize that this could be a result of increased food intake among our sample, secondary to COVID-19 restrictions, regardless of the quality of the diet as it was previously noted that lockdowns are associated with increased preparation of homemade meals more frequently [43][44][45].
Consuming more portions of fruits, legumes, olive oil, nuts and wine would give a higher MEDAS score, which suggests more adherence but could also result in increased weight because of increased caloric intake. This was an interesting finding in our study, as over a third of the individuals reported increasing their intake of fruits, vegetables, and grains. Weight gain is more likely to occur when food intake is increased while physical activity levels decline, as is often the case during lockdowns [33,43]. Although we did not assess physical activity using validated tools, similar studies have found that physical activity levels often fall during lockdown [43]. Future studies should investigate this association further [16,46].

Implications
Hence, it is easy to conclude that the lockdowns and closures imposed by governments worldwide have had some impact on the nutritional behavior in a significant number of people and the picture is no different in Lebanon. Since nutritional interventions may have a role in COVID-19 infection and mortality rates, our results should be used to guide public health policymakers to provide programs that focus on sending appropriate nutritional advice to better help in the management and prevention of the pandemic consequences [17].

Limitations and strengths
Our study is not without limitations, as our design was cross-sectional, which does not allow us to confer causation. Participants recruited had to self-report and describe lifestyle habits, and anthropometric parameters as our survey was web-based and thus respondents may not describe their dietary intake adequately. Additionally, the questionnaire was sent out in English, and although this could lead to a sampling bias, it is worth noting that a great proportion of the Lebanese population is fluent in the English language. Besides, the Cronbach's alpha value was low, therefore, results should be interpreted with caution. Furthermore, our sample had a high number of female respondents compared to males. Yet, research has consistently documented that females are more likely to participate in surveys [47]. Our results should be interpreted with caution since correlations were significant, but weak.
However, our study has several strengths that are worthy to mention. First, our study was adequately powered, as we were able to recruit a large number of participants. Additionally, the questionnaire that was used was validated, thorough, and detailed. Furthermore, we were able to recruit participants from all different governorates in Lebanon, which makes the study somewhat representative of the Lebanese population. To our knowledge, this is also the first study of its kind to be conducted in Lebanon that examines how lockdowns and countrywide closures affect people's lifestyle changes and adherence to the MD.

Conclusion
This study assessed lifestyle, dietary habits and MD adherence among a Lebanese sample during lockdown. We concluded from the findings in this study that there is room for improvement in the Lebanese dietary and lifestyle habits as MD adherence was suboptimal during the COVID-19 lockdown. It is critical that the Lebanese government implements public health programs in order to promote awareness about the importance of adhering to a healthy lifestyle and making appropriate dietary and lifestyle choices.