Introduction

Being overweight, obese, and underweight may have a major risk factor for multiple disorders in the later stage of life. Earlier studies have shown a relationship between BMI and mortality among diabetic patients (Chang et al. 2016), mental behavioral, neurological (Bhaskaran et al. 2018), and Parkinson’s disease (Chen et al. 2014).

Influence of infection on obesity

A U-shaped (non-linear relationship) elevated infection rate was seen among overweight/obese and underweight old age persons (Dobner and Kaser 2018), increased CVD risk with greater malignancy rates (Gluvic et al. 2016; Tune et al. 2017; Vucenik and Stains 2012), and obesity with impaired immune responses was seen (Carbone et al. 2016). Overweight/obesity was reported as a risk factor for Clostridium difficile colitis, pneumonia, bacteremia infections (Falagas and Kompoti 2006; Huttunen et al. 2013), and increased surgical site infection (Crabtree et al. 2004; Lillenfeld et al. 1988; Olsen et al. 2003; Thelwall et al. 2015; Tjeertes et al. 2015; Vilar-Compte et al. 2000; Zahr et al. 2011). Similarly, waist circumference was reported as a better predictor of septicemia than BMI (Wang et al. 2013). In a study, adipose tissue serves as a source for human influenza A virus, Trypanosoma gondi, HIV, adenovirus Ad-36, cytomegalovirus, and mycobacterium tuberculosis (Kassir 2020). The elderly population is reported to be more susceptible to infections (Dorner et al. 2010a). Higher mortality risks of influenza were reported among the underweight and obese old population (L. Yang et al. 2013).

Obesity and coronavirus infection

In many countries, the obese population is more vulnerable to most of the non-communicable diseases. Further, the COVID-19 pandemic has put this population at higher risk (Dayal 2020). The measures introduced in some countries, e.g., restriction on leaving home for several weeks even for daily walk, will have an impact on mobility, and these infection control measures resulted in physical inactivity, and even short periods of restriction can increase the risk of metabolic disease in future. Overweight and obesity among severe COVID-19 patients were reported as an independent risk factor (Jose and Manuel 2020). Increased adiposity undermines the pulmonary function contributing the viral pathogenesis as a secondary cause of infection (Dorner et al. 2010) has been reported in obese among older individuals (> 60 years of age) (Andersen et al. 2016). Following the 2009 influenza A virus H1N1 pandemic (Sun et al. 2016) and adult respiratory distress syndrome (ARDS) (Ni et al. 2017), obesity was reported as a risk factor. Decreased mortality due to ventilator-induced lung injury resulting in chronic pro-inflammatory status in obese patients has been reported (Repine and Fernandez-Bustamante 2013). Infection of influenza A showed that overweight/obesity risk elevates with virus shedding (42% longer) duration among symptomatic patients with obesity (Milner et al. 2015). In H1N1 disease, obesity was reported as a risk factor with increased hospitalization and death rates (Maier et al. 2018). Obesity has been shown to increase vulnerability to infections, serving as a risk factor among COVID-19 patient’s mortality rate (Misumi et al. 2019). Potential confounders for obesity and infection risk: smoking (Baik et al. 2000; Nuorti et al. 2000), well-being (Lasselin et al. 2016), physical activity (Krüger et al. 2016), weight loss (Dobner and Kaser 2018), co-morbidities (Pearson-Stuttard et al. 2016), and nutrition (Urwin et al. 2014) are reported independent of age and sex.

Methods

Search strategy

PRISMA guideline was used for this meta-analysis (Moher et al. 2009). Boolean operators: (BMI) AND (COVID-19), (Obesity) AND (COVID-19), ((Underweight) AND (COVID-19)) OR (CASRS-Cov-2), (BMI) OR (Body Mass Index) AND (COVID-19) were used for PubMed database and for Google Scholar, MEDLINE, EMBASE, Scopus, Medrixv and BioRixv using appropriate keywords (e.g., “Obesity”, “BMI”, “Body Mass Index”, “2019-nCov”, “COVID-19, “novel coronavirus”, “coronavirus disease”). The study included published literature without language restriction until 20th April 2020.

Selection criteria (inclusion/exclusion)

Studies with the following conditions had included for the meta-analysis: (1) case report, case series, and cohort studies design; (2) BMI assessment > 25 kg/m2 and < 25 kg/m2 reported; (3) indicating the risk ratio and/or odds ratio for the obesity risk; (4) studies reporting cross-sectional were excluded; (5) age and gender were not kept as a bar for inclusion; (6) for meta-analysis, we included BMI with reported COVID-19 infections (clinical, laboratory, or both confirmed); and (7) definition of severe COVID-19 was taken (Xu et al. 2020).

Data extraction

Details of authors, total sample size, and numbers reported for obesity and other conditions, e.g., comorbidity (obesity) and clinical condition (critical/severe), were extracted and recorded independently. Data extraction was accomplished by two reviewers (Dr. Khaiwal Ravindra and Dr. Neha Chanana). Any disagreement was resolved by joint discussion. To minimize the risk of duplication, data were carefully handled. Continuous variable (BMI) was expressed as mean ± SD/median and IQR (interquartile range). The systematic review include case reports, case series, and observational and any other study types of study design which reported obesity and or underweight or its complication of COVID-19 infection. Studies for meta-analysis were pooled only if the outcome measured in the same way by all studies.

Quality assessment

The Newcastle-Ottawa scale for cohort studies was used for quality evaluation of the selected studies (Wells 2014). This scale was applied because most of the studies in this study fall under cohort studies (retrospective), analyzing the other studies by covering broad spectra of study design. However, another study used the modified Newcastle-Ottawa scale for cohort and case-control study design (Codipilly et al. 2018).

Statistical analysis and data synthesis

After extracting the results of the studies, the effects of BMI on COVID-19 patients were pooled and were examined using the random effects method. For continuous outcome, standard error (SE) with 95%CI was calculated. The heterogeneity (I2 statistic) was assessed between studies. Higgins and colleagues suggested I2 values 25% (low), 50% (moderate), and 75% (high), indicating the existence of heterogeneity (Ioannidis et al. 2007; Higgins 2003; Higgins and Thompson 2002). Data analysis was undertaken using Microsoft Excel (Neyeloff et al. 2012).

Results

Literature search

Literature searching and screening was done according to PRISMA chart, as shown in Fig. 1. Initially, 3405 published research articles were identified through a database search. After the removal of 3368 publications due to duplicates and not relevant to study criteria, only 37 research papers were taken for the full-text paper. Finally, 14 articles that met the inclusion criteria were included in the quantitative synthesis of the current systematic review.

Fig. 1
figure 1

PRISMA chart

Full size image

Study characteristics

The main characteristics are summarized in Table 1. All published research papers fall under observational (cohort) study design. Most of the studies are from China, the USA, and France. The study included articles published/available online till 20th April 2020. About the obesity criteria, BMI > 25 kg/m2 was considered as also described in selected studies.

Table 1 Patient/study characteristics
Full size table

Quality assessment

The Newcastle-Ottawa Scale (for cohort studies) was used for the qualitative evaluation of the included studies (Wells 2014). The risk of bias was assessed based on three domains, i.e., selection, comparability, and outcome, as highlighted in Table 2.

Table 2 Quality assessment: cohort study quality according to Newcastle-Ottawa Scale
Full size table

Publication bias

This study included published studies, as well as unpublished literature on MedRxiv and Bioxiv, as long as it meets study inclusion criteria. Possible publication bias was not calculated due to the limited power among studies, and the outcome was < 10 for funnel plot (Lau et al. 2006).

Meta-analysis

The outcomes of the meta-analysis (Table 3) and forest plot (Figs. 2 and 3) are shown. A random-effects model (I2 = 53.45%) was used on 9 articles for those reported ≥ 25 BMI kg/m2 for COVID-19 infection involving 689 patients to analyze the risk factors of ≥ 25 BMI kg/m2 [prevalence, 95%CI: 0.34 (0.23–0.44)] for patients with COVID-19 than those of having < 25 BMI kg/m2 [prevalence, 95%CI: 0.60 (0.34–0.86)] with random-effect model (I2 = 0%). BMI < 25 kg/m2 random effect models I2 were negative, which indicates no observed heterogeneity for this group (Higgins 2003; Melsen et al. 2014).

Table 3 The meta-analysis of body mass index for COVID-19 patients
Full size table
Fig. 2
figure 2

Forest plot of risk factor BMI > 25 kg/m2 with COVID-19 patients

Full size image
Fig. 3
figure 3

Forest plot of risk factor BMI < 25 kg/m2 with COVID-19 patients

Full size image

Discussion

This study summarized available studies from retrospective about BMI and COVID-19 infection. We observed prevalence rate from 0.11 (95%CI: 0.07–0.15) to 0.86 (95%CI: 0.69–1.02) among COVID-19 patients having BMI > 25 kg/m2. Similarly, those having BMI < 25 kg/m2 show prevalence rate from 0.16 (95%CI: 0.08–0.23) to 0.80 (95%CI: 0.70–0.90) studies included in the meta-analysis. This study observed heterogeneity of 53.45% among COVID-19 patients with BMI > 25 kg/m2. This might be due to the various variables, age, and gender. The previous study suggests 30% prevalence of obesity in MERS-CoV infection (Badawi and Ryoo 2016) with pooled prevalence in obesity of 16 ± 2% (95%CI: 12–19%). Similarly, other studies observed the relationship between weight gain with pneumonia (Baik et al. 2000), infection rate among weight-reduced patients (Anderin et al. 2015), and thromboembolic risk in obese (Kassir 2020) population. COVID-19 patient with high BMI has moderate to high risk of medical complication with infection; hence, their periodic health status of co-morbidities, e.g., hypertension and diabetes, should be evaluated. Also, there is geographical variation in fatality case rates in South Korea (0.8), China (2.3), and Italy (7.2) that has been reported (Onder et al. 2020) with risk factors of smoking, pollution, and aging. In US patients, < 60 years and with BMI 30–40 (OR (95%CI) 1.8: 1.2–2.7) were more likely to be admitted to acute and critical care to individuals with a BMI < 30 (Lighter et al. 2020a) that has been reported for the first time. In a study, gender (male), age, and heart disease were the main risk factors of COVID-19-related death (Guo et al. 2020). Adolescents and young adults might play a key role in the worldwide spread of COVID-19 disease because they study overseas and frequently travel (Liao et al. 2020). Meta-analysis showed that elderly male patients with high BMI have greater chances of being into critically ill patient category (53). This study observed prevalence rate 0.60 (95%CI: 0.34–0.86) of COVID-19 infection among < 25 kg/m2 BMI population with no heterogeneity as compared to the > 25 kg/m2 BMI, 0.34, 95%CI: 0.23–0.44 with heterogeneity of 53.4% among study population.

Limitations

Population, continuous variable, clinical condition, and statistical methods have the potency to differ and may cause heterogeneity among studies included for the meta-analysis. Further, the study reviewed only the risk of BMI (> 25 kg/m2 and < 25 kg/m2) of COVID-19 subjects and their severity.

Conclusion

The BMI plays a significant role in COVID-19 infection and severity at all ages, especially elderly population. COVID-19 patients with higher BMI should be reviewed by a panel for the risk factors. Further, there should be a procedure for increased vigilance, testing priority, and therapy for patients with obesity and COVID-19 disease, whose illness has entered 7–10 days, having age > 50 years, and elevated CRP levels. The severity of COVID-19 found has a significant burden on intensive care resources in hospitals worldwide and specifically in lower-and-middle income countries due to lack of health finance and resources. Hence, the patients having higher BMI with other comorbidities should be given special attention to reduce morbidity and mortality associated with COVID-19 infection.