A Systematic Review of the Impact of the First Year of COVID-19 on Obesity Risk Factors: A Pandemic Fueling a Pandemic?

Abstract Obesity is increasingly prevalent worldwide. Associated risk factors, including depression, socioeconomic stress, poor diet, and lack of physical activity, have all been impacted by the coronavirus disease 2019 (COVID-19) pandemic. This systematic review aims to explore the indirect effects of the first year of COVID-19 on obesity and its risk factors. A literature search of PubMed and EMBASE was performed from 1 January 2020 to 31 December 2020 to identify relevant studies pertaining to the first year of the COVID-19 pandemic (PROSPERO; CRD42020219433). All English-language studies on weight change and key obesity risk factors (psychosocial and socioeconomic health) during the COVID-19 pandemic were considered for inclusion. Of 805 full-text articles that were reviewed, 87 were included for analysis. The included studies observed increased food and alcohol consumption, increased sedentary time, worsening depressive symptoms, and increased financial stress. Overall, these results suggest that COVID-19 has exacerbated the current risk factors for obesity and is likely to worsen obesity rates in the near future. Future studies, and policy makers, will need to carefully consider their interdependency to develop effective interventions able to mitigate the obesity pandemic.


Introduction
With over 268 million infections and 5.2 million deaths worldwide (1), coronavirus disease 2019 (COVID-19) is one of the most serious infectious disease outbreaks in recent history. Even before the declaration of pandemic status by the WHO on 11 March 2020, many countries had begun to impose social-distancing measures (SDMs) in an attempt to reduce disease incidence. Understandably, the attention of scientists has focused on how to limit the short-term consequences of COVID-19, which were mitigated by SDMs until vaccines were released. As a result, the scientific community has prioritized the research on the determinants of mortality and morbidity of COVID-19 over the long-term implication of the virus and the necessary countermeasures, such as SDMs.
Obesity is defined by the WHO as abnormal or excessive fat accumulation that presents a risk to health, marked by a BMI (in kg/m 2 ) >30, and has reached epidemic proportions (2). Statistics suggest that the prevalence continues to follow an increasing trajectory, with over 650 million adults having obesity in 2016 (3). Various models are attempting to predict the future burden of obesity, with projections rang-ing from 44% to >50% of the population (4,5), although all agree that it is likely to encompass a significant proportion of the population. Many chronic illnesses are adversely affected by carrying excess body fat, with obesity being linked to cancers, cardiovascular disease, hypertension, and osteoarthritis, as well as a strong association with metabolic syndrome (6).
Among the factors that can increase the risk of obesity, some seem to play a more prominent role than others. For example, depression has repeatedly been shown to have bidirectional associations with obesity and overweight (7). The effect of depression on obesity is likely multifactorial, involving neuroendocrine disruption with a chronic state of elevated cortisol (8); lifestyle changes with reduced desire to exercise and increase in emotional eating (9); and, in some cases, the use of antidepressants (10). Socioeconomic status has long been linked inversely to body weight (11) and again is multifactorial with effects mediated through fewer opportunities for physical activity and healthy food and education and poorer mental health. Not only is low physical activity a risk factor for obesity but it is also an important modulator of risk conferred by excess weight (12), and so the potential effect of lockdowns on sedentary behavior may act as a multiplier for poor outcomes.              (24,36,(45)(46)(47)(48)(49)(50)(51)(52)(37)(38)(39)(40)(41)(42)(43)(44). All of the 18 studies were longitudinal and used self-reported measurements, except for Wang et al. (35), who used an accelerometer sensor to record daily step counts. A total of 16 studies reported a reduction in physical activity during COVID-19, with 1 study showing an increase in activity (46) and 1 showing no change at all (40). A study in German schoolchildren aged between 4 and 17 y found an increase in active days per week, with an 11.1% increase in adherence to WHO physical activity guidelines (46). A study of high school students found no significant increment in physical activity during COVID-19 compared with the pre-restriction baseline; however, highly active students increased their activity levels relative to baseline (47).

Relation between COVID-19 and diet.
Twenty-seven studies were included that investigated the impact of COVID-19 on dietary patterns, as summarized in Table 4.

Favorable changes in dietary behavior.
A total of 5 studies reported an increase in home-cooked meals during the pandemic (23,61,68,74,80). Three studies reported an overall reduction in the frequency of fast food (26,74,79). Of the studies looking at alcohol consumption, only 1 study found a decrease in alcohol consumption during the pandemic in the Spanish general population (77). This decline in alcohol was correlated with higher adherence to the Mediterranean diet.
A cross-sectional study of the general population in Italy found an increase in the consumption of fruit, vegetables, nuts, and legumes and a significant decrease in junk food consumption (66). Second, a Spanish cross-sectional study focusing on patients with type 2 diabetes found a significant increase in vegetable consumption during the pandemic (69). Third, a study looking at healthy Chinese adults found an increase in vegetable, fruit, and milk consumption (70) relative to before the pandemic. The last change reported by the studies was a reduction in overall food consumption during the pandemic (26,82). A longitudinal study of adults older than 62 y in the Netherlands found that 12% of the sample were eating less than usual. However, this change in dietary habits was not reflected by a statistically significant reduction in weight (64).

Relation between COVID-19 and socioeconomic status.
Eleven studies were included in this review that investigated the impact of COVID-19 on financial status, as summarized in Table 3. Out of these studies, one reported a statistically significant worsening of financial well-being among 5550 benefits-eligible university staff (94). The remaining studies did not report a P value or 95% CI but reported a detrimental impact of COVID-19 on financial status, resulting in either reduced income (53,54,58,60,62) or job loss (56,57,(59)(60)(61)(62). Two of the papers showed that COVID-19 resulted in alarming the participant and increasing their fear of job insecurity (55,62), with Wilson et al. (55) reporting that 31.9% of participants had financial fears during the pandemic and only 19.6% of the sample had no concerns at all.
Ten studies reported a statistically significant increase in depressive symptoms during the pandemic (59,89,91,(93)(94)(95)(96)(99)(100)(101). Two of the studies looked at the general population in the United States (57) and Austria (88). Three of these studies investigated clinical staff including obstetricians and midwives (96), nurses (98), and physicians (91). Four studies looked at a younger cohort of participants including schoolchildren (85) and students (86,87,100). Finally, one of the studies looked at the impact of COVID-19 on the LGBT (lesbian, gay, bisexual, transgender) population in the United States and found a significant increase in depressive symptoms, particularly in those with a negative baseline screen (92). Although the P value was not reported in 7 studies (89,90,93,94,97,99,100), 6 of them reported a trend of increased depression scores during COVID-19 (89,90,93,97,99,100). Only 1 study found no increase in depressive symptoms during COVID-19 and looked at US physician trainees (94).

Discussion
This systematic review of over 350,000 participants from across the globe attempted to describe the indirect impact that the SDMs due to the COVID-19 pandemic had on population body weight by altering the most important risk factors-namely, diet, physical activity, mental health, and financial status. Although the impact of the countermeasures used to curb the COVID-19 pandemic was evident on obesity risk factors, none of the studies included in our research explored the direct impact of the risk factors on obesity itself.
The general trend seen in included studies was a worsening in the obesity risk factors. There were, however, notable exceptions. A German study in schoolchildren found an improvement in physical activity (46) due to recreational sporting activities. This discrepancy is likely due to contextual factors, such as how stringent the SDMs were in the specific countries. For example, in China, outdoor physical activity was banned during the first wave of COVID-19 (46).
Differences were also seen in dietary changes, with some studies showing an improvement in diet. However, those studies showing improvements in diet were looking at very different subgroups of the population (66,69,70), including the elderly or those with underlying medical conditions. The age of participants appears to have an impact, with the largest sample-size studies (25,34) showing a significant weight increase in those under age 25. The same was seen in a US sample of stu- LGBT population PHQ-9 Significant increase in PHQ-9 depression score in the total population during COVID-19 (P < 0 .001) Significant decrease in PHQ-9 depression score in those with a positive baseline screen (P < 0.001) Significant increase in PHQ-9 depression score in those with a negative baseline screen   dents (35). This may reflect the widespread reduction in activity and greater sedentary time in this group of people across multiple nations (36,38,43,46,50). It may also suggest a disproportionate impact of SDMs on the younger population. However, a comparable group of undergraduate students in Italy (30) did not show an increase in weight, which suggests a potential cultural role. The proximity to COVID-19 exposure may have played a role in the likelihood to report increased stress or depressive symptoms, as was seen in several cohorts of health care workers (89,91,99). These studies did, however, tend to occur earlier in the course of SDMs, which could also have played a role as uncertainty was at its greatest early on in the pandemic.
The COVID-19 pandemic, and its related SDMs, led to a worsening of obesity risk factors in the majority of studies-albeit some beneficial effects were observed in the dieting domain, such as higher consumption of home-cooked meals and healthy food (e.g., vegetables). On the other hand, the overall food and alcohol consumption showed an increasing trend, which could have been either the result or the cause of poorer mental health (102).
An unavoidable consequence of the SDMs and, in the most extreme cases, of the national lockdowns was financial hardship and job loss. A large body of evidence suggests that financial stress is linked to mental illness, which, then, could have fueled the obesity risk factors mentioned previously (103). Another element adding an extra level of complexity is the bidirectional relation between financial hardship, mental illness, and the other obesity risk factors, which makes it problematic to draw a conclusion on which is the leading factor during stressful circumstances, such as a pandemic.
There are several notable papers in the literature that have been published during the writing of this report, which go some way to supporting our conclusions. Jia (104), Browne et al. (105), and Knebush et al. (106) all discuss similar findings with the interaction between the coronavirus pandemic and obesogenic risk factors. Jia (104) highlights the multifactorial impact of the pandemic on the obesogenic environment in adolescents, including increased sedentary time and dietary changes. Upstream factors, such as changes in food environments and interaction with the built environment, might help to explain some of our findings; however, as noted by Jia, more modern measurement techniques are needed to better quantify this. An important issue raised is the difficulty in following up cohorts during periods of lockdown and how this will affect future data trends.
Browne et al. (105) also considered the change in the obesogenic environment affecting children during the COVID-19 pandemic. Increased stress has arisen from changes to home and school environments, in concert with less engagement in physical activity and increased familial financial stress. As we have found the case to be in adults, this review suggests that COVID-19 has exacerbated the obesity pandemic in children. An additional consideration in this paper was the deleterious impact of weight stigma, which can further increase the psychological and physical sequelae of obesity.
Knebush et al. (106) again noted similar patterns of reduced physical activity, increased screen time, and dietary changes. School closures have had a marked impact on each of these risk factors at critical points in a child's development.
These papers all highlight a similar pattern of an increasingly obesogenic environment that children have been subjected to during multiple SDMs throughout the pandemic. Of interest will be the effect of this in years to come as these children become adults, perpetuating the trend for increasing weight.
A BMJ feature (107) highlights the voice of Christina Marriott, chief executive of the Royal Society of Public Health, on the topic of obesity in the COVID-19 pandemic, who states that there has not been sufficient action to address the root causes of obesity. For this to happen, the complex relation between the obesity risk factors should be explored in quantitative studies. Our review acts to emphasize the areas in which further data are required. In addition to this, there is a clear need for cost-effective policies able to mitigate the impact on obesity of stressful circumstances, such as a pandemic.
Our research is the first to attempt to summarize the multifactorial implications that the SDMs due to the COVID-19 pandemic had on obesity. A very broad search strategy was adopted to capture as thorough a picture as possible, aiming to include papers noting an association between COVID-19 SDMs, obesity, and risk factors together. None of the studies included in our research investigated the link between 1) SDMs, 2) obesity risk factors, and 3) obesity itself. The absence of studies linking (1) to (2) and, thus (3), led us to focus our review on the impact of SDMs on obesity risk factors. As a consequence, our review cannot provide a conclusion on which elements have driven the increment in BMI during the COVID-19 pandemic (15). While this is the most important weakness of our study, our broad literature review allowed us to identify the studies on the effects of the pandemic on obesity and its risk factors.
Although our contribution is not sufficient to draw a conclusion, it represents a necessary step to develop new studies able to determine the key drivers of obesity in stressful circumstances, such as a pandemic. In addition to the absence of evidence necessary to draw a conclusion, many of the included studies focused either on self-reported body weight or BMI. Although these are widely used and validated measures of identifying individuals at risk of overweight or obesity, they do not account for factors that more reliably and objectively link to health outcomes, such as total body fat percentage.
Another limitation of our review is the high proportion of crosssectional studies, which makes it problematic to establish a causal link. Likewise, the high heterogeneity in methodology, samples, and socioeconomic characteristics made comparisons difficult. Many of the studies had a significantly higher response rate in females, which may somewhat limit the application of our conclusions to the general population. Several studies also focused on specific groups, many of which used health care workers or students. Once again, this may limit the generalizability of our conclusions.
These limitations are acknowledged in our quality assessment of the included studies. However, given the circumstances in which many of these studies were carried out, amid national lockdowns, in-person data collection was often unfeasible and so the majority of studies were affected by this measurement issue.
While this review does not provide a conclusive answer on the driver of obesity during the COVID-19 pandemic, it provides useful information to direct future research aiming at strengthening the link between stressful circumstances and a rise in risk factors for obesity and weight gain. This is important as establishing a link enables us to effectively target the risk factors in preventative public health measures. There is a need for longitudinal studies to elucidate the nature of the association.