COVID-19 Outcomes: The Impact of Obesity Within Racial Groups

ABSTRACT

Background: The COVID-19 pandemic has thrust the obesity epidemic into the spotlight, revealing that obesity is no longer a disease that harms in the long run, but one that can have acutely devastating effects. Studies have suggested that obesity not only increases your risk of hospitalization, mechanical ventilation and death but increases your risk of becoming infected. Rapid data collection started in late 2019 as the COVID-19 spread from its site of origination, Wuhan, China. Initial observational studies noted that Hypertension, Type 2 Diabetes and Coronary Artery Disease were the most common comorbid conditions in patients with more severe complications. Although data was collected in China regarding BMI, it was not until April 2020 when the editors of the American published Obesity medical journal stated that obesity would likely prove to be independent risk factor for more severe disease. This observation was likely yet to be realized in the Chinese data given the significantly lower rates of obesity than the United States. In 2019, the overall rate of obesity among Chinese adults was 13.58%. The rate of obesity in the United States in 2017-2018, the last reported data from National Health and Nutrition Examination Survey, was 42.4%. COVID-19 also revealed conspicuous health care disparities that have been documented for decades with little national spotlight until July 2021. Federal data reported that the Black American life expectancy decreased almost 3 years from 2019 to 2020 compared to 1.2 years for White Americans. In our study we examine the association of four clinical complications from COVID-19 infection among Black/African American and White/Caucasian patients with overweight and obesity while controlling for co-morbid conditions. Our aim is to identify if different risk exists between ethnicity groups in overweight and obese patients, controlling for additional comorbidities. We hypothesize that racial disparities in COVID outcomes persist. Methods: A random sample of 3,000 records was extracted from the COVID19 DataMart through the University of Virginia Health System EMR. This included patients who tested positive by RT-PCR for SARS-CoV-2 using swab specimens between March 2020 and July 2021. We limited this study to Black/African American or White/ Caucasian patients age > 18 with a BMI >25, for a total of 1,904 patients. Measurements Four clinical outcomes from COVID-19 were examined;hospitalization, length of hospital stay (LOS), ventilator dependence and mortality. LOS was measured by an indicator constructed within the EMR calculating the difference between admission and discharge dates. Ventilator dependence and mortality were binary indicators related to COVID-19. Race was dichotomized as Black/ African American or White/Caucasian. We controlled for comorbidies with the Charlson Comorbidity Index (CCI), which accounts for 17 conditions. Patients were divided into four groups based on the CCI score. CCI score and corresponding weighted category were computed using the "comorbidity" package in R. Based on patients' BMI, we computed a variable indicating BMI category 25< and <30;>30 and <35;>35. Sex and age were controlled. Sample Size Calculation We used the "pwr" package in R to compute the minimum required sample size that would offer 80% power and significance level of 0.05 to detect a small effect size. This was attainable in our analysis. Statistical Analysis Means and standard deviations were applied to continuous variables. We ran bivariate comparisons between the four outcomes of interest and race. Unadjusted odds ratios with corresponding 95% Confidence intervals (CI) are presented for hospitalized, ventilation required and mortality. A t-test is presented for mean comparisons of LOS by race. Four regression models were constructed for each respective outcome of interest while controlling for race, CCI category, BMI category, age and sex. Logistic regression models estimated the odds of hospitalization, ventilation dependence and mortality, and a linear model to estimate effe ts of covariate on LOS. All hypothesis tests were 2-sided with a significance level of 5%. R version 4.1.1 was used for all analyses. As a sensitivity analysis, we replicated these models in a sub-sample of patients who were infected between March 2020 and February 2021 (n = 756). Results: 1,904 patients were included in the main cohort and 756 in the sensitivity cohort. The mean age was 45.7 and 46.3 in the sensitivity cohort. Females comprised 58.5% and 81% of the sensitivity cohort. Average LOS was significantly greater for Black patients than WHITE patients. Hospitalization rates for White patients were significantly lower than those for Black patients. Ventilation did not differ significantly between White and Black patients. Mortality rates were lower for White patients compared to Black patients. In the regression model White patients had significantly shorter LOS compared to Black patients. Patients with low and mild CCI risk had significantly shorter LOS compared to patients in the severe risk category. Patients with a BMI category or >35 had shorter LOS compared to those with BMI <30;males had significantly longer LOS compared to females. Race did not have a significant effect on hospitalization after controlling for CCI category, sex, BMI category and age. Patients with low and mild CCI had significantly lower odds of being hospitalized. Males had significantly greater odds of being hospitalized compared to females. Race did not have a significant effect on ventilation. Male patients had significantly higher odds of needing ventilation. Patients with low CCI risk had significantly lower odds of needing ventilation. Increase in age was associated with increased odds of needing ventilation. Race was significantly associated with mortality. White patients had significantly lower odds of death compared to Black patients. Patients in the low and mild CCI risk had lower odds of death. White patients had significantly shorter LOS compared to Black patients. Patients with low and mild CCI risk had significantly shorter LOS compared to patients in the severe risk category. Race did not have a significant effect on hospitalization. Patients with low and mild CCI risk had significantly lower odds of being hospitalized. Race did not have a significant effect on ventilation. Patients with low and moderate CCI risk had significantly lower odds of needing ventilation. Increase in age was associated with increased odds of needing ventilation. Race was associated with mortality after controlling for CCI category, sex, BMI category and age at a p value = 0.05. White patients had lower odds of death compared to Black patients. Conclusions: Our findings emphasize that racial differences persist when we look at only patients that have overweight and obesity. The racial disparities are not due to just obesity and its common co-morbidities. There are other factors contributing to these outcomes. These factors are likely multifactorial. It could be due to a combination of social and environmental factors that prevent or delay these patients from receiving care. Furthermore there may be differences in the care these patients receive once in the health care system. Further examination of these socioeconomic factors are just as important as mechanistic causes in order to decrease healthcare disparities.