Individuals with obesity and COVID‐19: A global perspective on the epidemiology and biological relationships

3.1 Literature retrieval

We examined PubMed, Google Scholar, MedRxiv, BioRxiv, Wanfang (for Chinese literature) and other literature search engines (e.g., China National Knowledge Infrastructure Data and ICNARC) to systematically review all publications in Chinese or English that include data on COVID‐19 and BMI or individuals with obesity. We briefly reviewed the abstracts and results and located 75 publications available by 15 July 2020 that presented data on the BMIs or BMI categories of diagnosed COVID‐19 patients. We excluded literature in other languages, as we read Chinese and English only. All of our authors performed the literature searches and reviews. Table S1 presents the search terms.

3.2 Study characteristics

We found 1733 studies, 75 of which provided data we could use in this review (Figure 2). All were conducted between January and June 2020, including five case–control studies, 33 retrospective or prospective cohort studies and 37 observational cross‐sectional studies. Sample sizes varied from 24 to 109 367 diagnosed patients in more than 10 countries in Asia, Europe and North and South America. In total, we included 399 461 diagnosed patients in this study, about 55% of whom were male. Table S2 presents detailed demographic data from the studies we used, including a few studies that had inadequate data for use in the meta‐analysis. We used STATA (version 16, College Station, TX) to perform all random‐effects meta‐analysis and used residual maximum likelihood to fit all models.24

3.3 Being an individual with obesity and the risk of COVID‐19

We identified 20 studies that assessed the association between individuals with obesity and COVID‐19, all but two of which showed that individuals with obesity significantly increase the risk of COVID‐19 (Table S3).25–43 One study in Denmark showed that the prevalence of overweight and individuals with obesity was lower in SARS‐CoV‐2 positive cases than SARS‐CoV‐2 test‐negative individuals (8.6% vs. 9.9%).44 The results may be biased because body weight status was determined at hospital discharge. A study used U.K. Biobank data (n = 285 817) to show that overweight increased the risk of COVID‐19 by 44.0% (relative risk [RR] = 1.44; 95% CI, 1.08–1.92; p = 0.0100) and individuals with obesity almost doubled the risk (RR = 1.97; 95% CI, 1.46–2.65; p < 0.0001), adjusted for age, sex, ethnicity and socio‐economic deprivation as measured by unemployment, assets and household density.32 The authors tested only a small portion of individuals (0.5%) for COVID‐19, a key limitation of this study. A better way to calculate OR for this study is to compare the odds between subjects who tested positive and those who tested negative. Our pooled data analysis showed that the odds of individuals with obesity being COVID‐19 positive were 46.0% (OR = 1.46; 95% CI, 1.30–1.65; p < 0.0001) higher than those of individuals who were not obese (Figure 3).

3.4 Being an individual with obesity and COVID‐19 illness severity

Being an individual with obesity increases the odds of COVID‐19 patients being hospitalized. Among diagnosed COVID‐19 patients, the prevalence of individuals with obesity in hospitalized patients was much higher than that in nonhospitalized patients. For example, a report that included 5700 patients with obesity in New York City45 showed that 41.7% of COVID‐19 hospitalized patients were individuals with obesity, whereas the average prevalence of individuals with obesity in New York City was 22.0%.46 Many studies reported COVID‐19 hospitalizations, but only a few reported the relationship between individuals with obesity and hospitalization. We identified 19 studies that examined the relationship and included them in this analysis.1, 28, 38, 40, 44, 47–58 Table S4 presents the results1, 45, 47, 48, 59, 60; all showed a significantly higher prevalence of individuals with obesity among hospitalized patients than among patients not hospitalized or the general population. The pooled OR was 2.13 (95% CI, 1.74–2.60; p < 0.0001) (Figure 4).

Among patients with symptoms, those with severe or critical conditions had much higher BMIs and individuals with obesity prevalence than the normal population or patients who were COVID‐19 negative.32, 61–70 Two studies showed that the odds of having COVID‐19 increased by 30% (OR = 1.30; 95% CI, 1.09–1.54; p = 0.0030)61 and by 38% (OR = 1.38; p < 0.0001),32 respectively, among the individuals with obesity (Table 1).

All studies reported that among those diagnosed, patients with obesity were more likely to be admitted to ICUs.
* However, the effect sizes in the studies with smaller sample sizes were not statistically significant.48, 72, 73 In the studies that found that being an individual with obesity did not significantly increase the odds of being admitted to the ICU, individuals with morbid obesity (defined as BMI ≥ 35) did significantly increase the odds of ICU admittance. Our pooled data (from 22 studies) showed that individuals with obesity increased the odds of being admitted to the ICU by 74% (OR = 1.68; 95% CI, 1.46–2.08; p < 0.0001) (Figure 5 and Table S5).

Reports that had smaller sample sizes from the United Kingdom and some other countries showed that patients with obesity had higher but insignificant odds of invasive mechanical ventilation (IMV) than patients without obesity.48, 72, 86 Reports from Mexico and some U.S. cities showed significantly higher odds of IMV in patients with obesity than in patients without obesity.
† The pooled data (from 14 studies) showed a 66% increase in IMV in patients with obesity (OR = 1.66; 95% CI, 1.38–1.99; p < 0.0001) (Figure 6 and Table S6).

3.5 Being an individual with obesity and COVID‐19 prognosis

The association between obesity and COVID‐19 prognosis is complex, because patients discharged from ICUs may be still hospitalized or deceased later. For example, 4.5% died after they were discharged from ICU; 11.5% remained in the hospital after leaving the ICU in one study.31 A few studies showed that individuals with obesity may decrease in‐hospital mortality.31, 75, 87, 89 Some studies showed that obesity may insignificantly decrease41, 47, 81, 90 or increase55, 67, 68, 84, 91–96 the odds of death among individuals with obesity. The majority of studies showed that obesity significantly increased the odds of death among COVID‐19 patients with obesity. The pooled data (from 35 studies) showed that patients with obesity were more likely to have unfavourable outcomes with a 48% increase in deaths (OR = 1.48; 95% CI, 1.22–1.80; p < 0.001) (Figure 7 and Table S7).
‡ We excluded two studies that had very large OR and very wide 95% CI, one study in China (OR = 32.08; 95% CI, 6.73–153)65 and one in Nevada (OR = 10.55; 95% CI, 1.07–104.45),106 from Figure 7, but included them in the meta‐analysis.

4.1 Individuals with obesity’s metabolic and physiological impairments linked to COVID‐19

Being an individual with obesity is a major risk factor for severe cases of certain infectious diseases, like influenza, hepatitis and nosocomial infections.111, 112 However, other infections, like tuberculosis, community‐acquired pneumonia and sepsis, have more favourable clinical outcomes in adults with obesity compared with lean adults.113 This supports the ‘obesity paradox’ hypothesis, where underlying characteristics of individuals with obesity influence the physiological response to infection. As with influenza infections, being an individual with obesity appears to increase COVID‐19 severity. Obesity is inherently a metabolic disease characterized by alterations in systemic metabolism, including insulin resistance, elevated serum glucose, altered adipokines (e.g., increased leptin and decreased adiponectin) and chronic low‐grade inflammation.114, 115 Strong evidence demonstrates how hormone and nutrient dysregulation in individuals with obesity can impair the response to infection.

Hyperglycaemia, a key hallmark of T2D, is highly associated with individuals with obesity. Importantly, uncontrolled serum glucose has been shown to significantly increase COVID‐19 mortality.116 During times of infection, uncontrolled serum glucose can impair immune cell function either directly or indirectly via generation of oxidants and glycation products.117 Similarly, both insulin and leptin signalling are critical in the inflammatory effector response of T cells by up‐regulating cellular glycolysis,118, 119 supporting the production of effector cytokines such as IFN‐γ and TNF‐α. These metabolic factors combine to influence immune cell metabolism,120 which dictates the functional response to pathogens, such as SARS‐CoV‐2.

Dietary consumption of fatty acids can also influence inflammatory responses. Prostaglandins, the derivatives of long chain fatty acids, are acute phase pyrogens that initiate the local inflammatory response during infection. Omega‐3 polyunsaturated fatty acids can induce anti‐inflammatory responses through cyclooxygenase (COX) activity, whereas omega‐6 fatty acids mediate the pro‐inflammatory COX production of prostaglandins.121–123 Current dietary intakes favour omega‐6 fatty acids over omega‐3s, with U.S. consumption currently in a 10:1 ratio due to the widespread consumption of vegetable oils.124 Fatty acid derivatives can directly influence COVID‐19 in individuals with obesity. Preclinical data suggest a role for fatty acid derived pro‐resolving lipid mediators, as they can be deficient in individuals with obesity and thus are not able to appropriately resolve inflammatory responses during infection.125

Other fatty acids, such as cholesterol, are essential in the spread of enveloped RNA viruses, like respiratory syncytial viruses and influenza. SARS‐CoV, the nearest relative to SARS‐CoV‐2, uses cholesterol to facilitate viral budding following S protein binding of cellular ACE2 receptors, allowing the spread to neighbouring cells. Depletion of cholesterol in ACE2 expressing cells results in markedly reduced viral S protein binding.126 Being an individual with obesity also increases the risk of COVID‐19 severity among patients with metabolic associated fatty liver disease, where adults with obesity had a greater than six fold higher risk for severe COVID‐19 regardless of age, sex or comorbidities, such as hypertension, diabetes and dyslipidaemia.127

Physical features of individuals with obesity also likely increase COVID‐19 severity and risk. Obstructive sleep apnoea and other respiratory dysfunctions in the individuals with obesity often increase risk of hypoventilation‐associated pneumonia, pulmonary hypertension and cardiac stress.128 Large waist circumference and greater body mass increase the difficulty of care in hospital settings for supportive therapies, such as intubation, mask ventilation and prone positioning to help reduce abdominal tension and increase diaphragm capacity.129 Thus, the prognoses of COVID‐19 patients with obesity may be complicated by the increased clinical care burden among this already vulnerable group.

4.2 Being an individual with obesity impairs the immune response to SARS‐CoV‐2

Being an individual with obesity has modulatory effects on key immune cell populations critical in the response to SARS‐CoV‐2. Specifically, increased BMI is associated with greater frequency of the anti‐inflammatory CD4 T cell subsets Th2 and T regulatory cells.130 Increased anti‐inflammatory cells may inhibit the ability to reduce the infection, as inflammatory responses are needed to control viral spread. Regulatory T cells (Tregs) primarily resolve immune cell mediated inflammation following infection. Tregs from hyperinsulinaemic mice with obesity have reduced interleukin 10 (IL‐10) production131 and, despite being in higher abundance in the lungs during influenza infections, are 40% less suppressive.132 Functional responses to RNA viruses, like SARS‐CoV‐2, rely on type 1 inflammatory responses by Th1 cells for protection with optimal anti‐inflammatory Treg responses for immune resolution following infection. Severe cases of individuals with influenza and COVID‐19 share remarkably similar reliance on type I interferon activation, with TNF/IL‐1β‐driven inflammation present in severe but not mild cases.133 Any imbalance in these T cell subsets or functions is likely to impair the immune response to SARS‐CoV‐2.

A further imbalance in immune cell subsets occurs with accumulation of pro‐inflammatory cells, including macrophages, dendritic cells, cytotoxic T cells and Th1 cells, in the adipose tissue of obese individuals. This influx of immune cells contributes to the development of insulin resistance and chronic inflammation.134 These pro‐inflammatory immune cells along with hypertrophic adipocytes are responsible for increased serum inflammatory cytokines, such as IL‐6, C‐reactive protein and type I and type III interferons.135, 136 This immune phenotype can be further distinguished between nondiabetic and people with diabetes and obesity through increased Th17 inflammation driven by impaired immune cell oxidation of fatty acid metabolites.137, 138

These changes in systemic immune cell populations and their accumulation in adipose tissue have been proposed as key mediators of COVID‐19 severity in individuals with obesity.139 Recently, mice with obesity infected with lymphocytic choriomeningitis virus (LCMV) were shown to have increased LCMV viral titres and LCMV‐specific immune cells in white adipose tissue, which upon secondary infection resulted in greater inflammation and mortality in mice with obesity compared with mice that are lean.140 Accumulation of adipocytes and adipocyte‐like cells can increase immune activation and cytokine production during coronavirus infection.141 In addition to being nutrient‐rich storage pools, lipid accumulation and adipocyte hypertrophy might be an immune reservoir that in individuals with obesity becomes saturated with pro‐inflammatory immune cell subsets.

Alterations in immune cell frequencies in individuals with obesity have been proposed for SARS‐CoV‐2 severity, which uses the angiotensin‐converting enzyme 2 (ACE2) for viral entry and is highly expressed in vascular tissues like the lungs and adipose tissue.141 Viral entry via ACE2 cleavage by the serine protease TMRPSS2 spike protein allows viral replication not only in the respiratory tract but also in other tissues expressing ACE2, including the intestinal enterocytes, liver, heart and kidneys.109, 142 This mechanism is thought to drive increased incidence of ischemic and coagulopathy conditions in COVID‐19 patients.

4.3 Inflammatory considerations of COVID‐19 in individuals with obesity

ARDS and acute lung injury (ALI) are two of the primary causes of morbidity and mortality among adults infected with SARS‐CoV‐2.143 Presentation of ARDS and ALI is characterized by respiratory failure due to excessive pro‐inflammatory cytokine production. This inflammatory state leads to extensive lung damage, hypoxemic respiratory failure regardless of oxygen administration and pulmonary oedema not caused by congestive heart failure.144 Patients who develop ARDS are typically administered mechanical ventilation with positive end‐expiratory pressure and high FiO₂. Currently, adults with obesity infected with SARS‐CoV‐2 have higher burdens of mechanical respiratory therapy support and ARDS development.66

Gong et al. previously demonstrated that, compared with lean adults (BMI 18.5–24.9), adults with obesity are more likely to develop ARDS.9 A 2016 meta‐analysis investigating how BMI influences ARDS/ALI outcomes demonstrated significantly lower ARDS‐related mortality in adults with obesity compared with lean adults despite confirming greater odds for developing ARDS.145 However, a retrospective multicentre study in Wuhan, China, found higher ARDS‐related mortality among COVID‐19 patients, which was predicted by elevated serum IL‐6.143 Similarly, elevated IL‐6 is a hallmark of severe SARS‐CoV,146 MERS‐CoV147 and pandemic H1N1 influenza A viral infections.148 Additionally, severe COVID‐19 cases have been associated with lymphopenia149 and lower expression of IFN‐γ by CD4 T cells.150 IFN‐γ is an important antiviral protein, and reduced production of this cytokine in response to influenza has been documented previously in both models of mice with obesity and human populations with obesity.132, 151

Men also experience a higher burden of COVID‐19 than women.59 Being a man with obesity increases aromatase activity, which can convert testosterone to estradiol.152 Oestrogen receptor signalling can subsequently down‐regulate IL‐6 expression through inhibition of NF‐κB,153 which has been shown to confer protective effects against influenza A virus in women through stimulation of neutrophil and virus‐specific CD8 T cell responses.154 Interestingly, however, men with obesity have impaired oestrogen receptor signalling, which leads to increased androgenic hormones and elevated oestrogen production from adipose tissue.155 Recently, androgen depletion therapy has been shown to protect against COVID‐19 in male prostate cancer patients.156 However, more information is needed to understand the mechanism of action of androgens and androgen depletion therapy. Nonetheless, adequate control of pro‐ and anti‐inflammatory responses during SARS‐CoV‐2 infections is critical to limit nonspecific tissue damage and subsequent development of ARDS, which has a higher burden among COVID‐19 cases with obesity.

4.4 Implications for treatment and vaccination strategies for being an individual with obesity

Obesity may also impair therapeutic treatments during COVID‐19 infections. ACE inhibitors, which are commonly used to treat hypertension, may increase COVID‐19 severity in T2D patients, especially those with poorly controlled blood glucose.157 While discontinuing use of ACE inhibitors is not advisable at this time due to offsetting cardiovascular benefits,158 current clinical trials are investigating mitigation of the spread of SARS‐CoV‐2 through inhibition of ACE2 binding. How these treatments in patients with obesity contribute to COVID‐19 severity, however, will be a key question in their overall effectiveness. The IL‐6 receptor (IL‐6R) antagonist tocilizumab may reduce IL‐6 signalling in severe COVID‐19 cases where cytokine release syndrome is a major factor of mortality.159 As noted above, chronic inflammation is a hallmark of individuals with obesity, which includes elevated levels of IL‐6. Preliminary data suggest tocilizumab treatment can reduce fever and oxygen requirement.160 However, subjects with obesity with chronically elevated IL‐6 may not benefit from acute treatment. Dexamethasone, a corticosteroid commonly used for inflammatory treatment of arthritis, allergic reactions or other immune inflammatory disorders, has been shown in preliminary data to reduce mortality in severe COVID‐19 patients by 8–26%.161 These data from the RECOVERY trial provide evidence of reduced 28‐day mortality with treatment of 6‐mg dexamethasone over a consecutive 10‐day period during the symptomatic phase, resulting in inclusion of dexamethasone as a treatment option under the National Health Service COVID‐19 treatment protocol,162 joining remdesivir as the only approved treatment options for COVID‐19. There remains limited information on other treatments, such as statins, nonsteroidal anti‐inflammatory drugs and angiotensin receptor blockers, regarding their effectiveness against COVID‐19 in the individuals with obesity. Considering that almost all countries today have a prevalence of individuals with overweight/obesity greater than 20% and that in certain countries, such as the United States and the United Kingdom, two‐thirds of the population is overweight/obese, understanding how these therapies work in the host with obesity is critical information to determine their long‐term effectiveness.

Importantly, being an individual with obesity has also been shown to impair the development of immunological memory. Influenza vaccination in adults with and without obesity results in equivalent influenza‐specific antibody titres at 30 days post vaccination, but antibody titres wane significantly more in adults with obesity compared with adults who are lean at 1 year post vaccination.163 Compared with influenza‐vaccinated lean adults, vaccinated adults with obesity have impaired CD4 and CD8 T cell production of key inflammatory cytokines IFN‐γ and granzyme B.151 Adults with obesity also have two times greater odds of influenza or influenza‐like illness despite a robust antibody response.164 Preclinical evidence demonstrates that adjuvant vaccines confer less protection against influenza viruses in diet‐induced mice that are obese.165 Similar impairments in vaccine effectiveness have been reported individuals with obesity for tetanus,166 hepatitis A and B and rabies.167

Data from recovered COVID‐19 patients show greater than 95% of infected patients develop neutralizing antibodies against SARS‐CoV‐2. However, early evidence suggests a waning of antibody production over a period of weeks to months,168 suggesting vaccines strategies designed for antibody seroprotection may not have as long‐lasting effects. This fast decline in circulating neutralizing antibodies is more similar to common coronaviruses as opposed to SARS‐CoV, which has a longer sustained level of antibody titres of ~2 years.169 Promising data from multiple groups find cross‐reactive T cell responses in 70–100% of COVID‐19 patients.170, 171 Le Bert et al. showed 36 convalescent COVID‐19 patients all had CD4 and CD8 T cells capable of recognizing and responding to the NP protein of SARS‐CoV2. Importantly, they demonstrate presence of long‐lived memory T cells in 23 patients who recovered from SARS‐CoV.172 Several other key papers find T cell mediated immune responses to SARS‐CoV‐2 across cohorts, suggesting generation of memory T cell populations is critical for any future COVID‐19 vaccine.173 Unfortunately, as T cell responses have been shown to be impaired in individuals with obesity, this suggests that a future COVID‐19 vaccine may be less effective in an population with a high prevalence of individuals with obesity. Therefore, it is urgent that any vaccine trials and studies include BMI as a potential confounder for vaccine effectiveness and protection.

6.1 Policy implications

Vaccination remains the best protection against infectious diseases like COVID‐19. Therapeutics targeted at limiting viral replication or remediating complications of infection may help limit severe cases and moderately reduce mortality. Public health experts agree that viral spread will continue to cause tremendous health and economic problems until we reach vaccination and/or community‐acquired herd immunity. Current models project that intermittent times of social distancing and lockdown measures will be required until a viable vaccine can be widely produced,197 and these measures are likely to extend into the foreseeable future. This paper highlights another concern—that is, vaccines may not be as effective in individuals with overweight/obesity. Given the large prevalence of the world population that is composed of individuals with overweight/obesity, it is imperative that governments ensure that testing and research focus not only on the general efficacy of vaccines and therapeutics but also on how they will impact individuals with obesity.

Furthermore, we must carefully monitor and regulate the consumption of ultraprocessed foods and beverages through fiscal policies such as taxation and regulating marketing and promotion of such foods. If as expected this behaviour is increasing, it will exacerbate other health concerns, including risks of increased adiposity and major NCDs. When compounded with reduced physical activity and increased sedentary behaviour, the risk of increased adiposity is clearly an important concern. Finally, the poor around much of the globe also face increased hunger and with it the potential for elevated stunting and its consequences, including the long‐term risks of central visceral adiposity and many NCDs. Increasing hunger and stunting can have long‐term adverse impacts on health and well‐being in multiple ways, and major policies to mitigate this effect are critical when resources are available.

In addition to COVID‐19’s critical economic constraints, its impacts on diets may pose lifelong risks to populations around the globe. Food habits developed during this period, particularly the intake of ultraprocessed foods, represent a major health risk. Exact policy prescriptions will be country specific, and clearly, the concerns for higher and middle‐income countries will differ from those of low‐income countries. NCD and individuals with obesity risks are far more predominant in the former, whereas the latter face high levels of the double burden of malnutrition, in which slow declines in stunting are likely to shift to increased stunting and wasting accompanied by rapid increases in individuals with obesity.

Creative policies to reduce consumption of ultraprocessed foods and increase consumption of healthier foods, such as legumes, selected whole grains, vegetables and fresh fruits, are important for all countries. A recent World Bank report suggests that the multipurpose Chilean model effectively administers multiple regulations and laws that reinforce each other and are impactful.12, 198, 199 It is quite that likely Chile’s policies could significantly reduce the current growth in consumption of ultraprocessed foods. Moreover, a tax accompanying purchases of those foods would potentially increase fiscal space in countries suffering from the economic impacts of COVID‐19, albeit few countries have successfully allocated these resources for health or nutrition programmes. Some countries are finding ways to provide boxes of fresh vegetables and fruits to the elderly such as one programme in several Chilean cities200; however, most low‐ and middle‐income countries do not have the resources for such efforts though combined with taxation and marketing controls, such efforts would be more feasible. All countries need to consider how to enhance consumers’ selections of healthy food options while reducing incentives to purchase ultraprocessed foods and beverages. To date, no country has combined these fiscal and regulatory policies. However, Israel’s Chilean‐style warning labels and promotion of healthy eating comes closest.201

The COVID‐19 pandemic challenges all countries enormously. Our systems, institutions, health and welfare will feel the impacts for many years. The high prevalence of individuals with obesity exacerbates the threat to everyone’s health, and the economic, social distancing and stay‐at‐home components compound the impacts. We will need creative solutions quickly to prevent undesirable dietary patterns and promote healthy eating, which is so critical to our future health and for building resilience against future threats.