This study demonstrated Brown and Black individuals were more likely to die from COVID-19 than White individuals, especially Brown women. The presence of obesity was an effect modifier of the association between race/skin color and COVID-19 mortality. When comparing White vs. Brown and Black populations, the likelihood of mortality was higher in individuals without obesity, except for the group of Brown women with obesity, who had higher odds of death than Brown women without obesity. Finally, a direct association between obesity and mortality was observed only in White men and women and in Brown women.
Our findings are consistent with those of previous studies showing higher COVID-19 mortality in racial/ethnic minorities [20,21]. In Brazil, descriptive studies have indicated higher odds of mortality in Black individuals, followed by Brown individuals [6,22,23]. A study analyzing data from more than 70 million Brazilians diagnosed with mild/moderate or severe COVID-19 between February 26, 2020, and November 15, 2021, demonstrated that Brown (odds ratio [OR] 1.11, 95% CI 1.10–1.12) and Black (OR 1.34, 95% CI 1.32–1.36) individuals had the highest odds of COVID-19 mortality compared with the White population [23].
In this study, although the descriptive analyses also indicated a higher prevalence of mortality in the Black population, the analyses stratified by gender identified Brown women as the main risk group. This highlights the importance of applying theory-based analytical approaches, such as intersectionality theory. The intersectionality approach allows the analysis of risk experiences in groups that are predominantly investigated as homogeneous, such as women and Black individuals. This approach assumes that multiple social categories, such as gender, race/skin color, and socioeconomic status, interact at the individual level and reflect interrelated systems of privilege and oppression at the macro level, such as racism and sexism [24–26]. Based on this theory, the intersectionality analysis of gender and race/skin color allows the recognition of multiple social identities that are hidden in traditional approaches, which analyze these identities separately [27].
This study used hospitalization data, which may have led to underreporting of COVID-19 mortality in the Black population. As a result of racial and social inequalities, this population may have died without even accessing tertiary health facilities, such as hospitals. Also, the population self-identified as Black in Rio Grande do Sul represents 21% of the state’s population (13.9% as Brown and 5.9% as Black), and in this study it accounted only for 7.8%28. Studies have indicated an increase in the number of people who self-identify as Black and Brown in recent years, as a result of a gradual increase in racial discussions in Brazil [29], a phenomenon involving mainly young people. The mean age of our sample was approximately 60 years, that is, a population of middle-aged and older adults, who also lived in the region with the smallest number of Black individuals in Brazil. Therefore, they may tend to report a lighter skin color (Brown)[30], underestimating the association estimates for Black individuals.
The increased number of severe and fatal cases of COVID-19 in the Black population in Brazil may be explained by the racism experienced by this group, which results in worse socioeconomic and labor conditions, limited access to healthy food, higher prevalence of chronic diseases, and limited access to health services and immunization [31,32]. Such difficulties are the result of the systematic degradation of the living and health conditions of Black people in Brazil, based on historical, social, and political events that support Brazilian racism9. Racism is a structuring system that results in practices, beliefs and prejudices that underlie avoidable and unfair inequalities between social groups, based on race or ethnicity, and that, institutionally, restricts access to goods, services, and opportunities [33].
Although the presence of obesity has been identified as an important risk factor for COVID-19 mortality since the beginning of the pandemic, meta-analyses have produced conflicting results [34–36]. A review of meta-analyses found an association between obesity and increased risk of mortality in patients with COVID-19, but a critical appraisal of the quality of the evidence indicated a high risk of bias and low certainty of evidence [37]. In this study, obesity was associated with higher odds of COVID-19 mortality only in White people and in Brown women. Studies that analyzed the association between obesity and COVID mortality did not consider characteristics such as race/skin color, which could be related to this association. Furthermore, most of the studies were conducted in the Global North. A possible explanation for the present results is that the Black population, because it is the most exposed to poverty and situations of malnutrition and food insecurity in Brazil [28], is more exposed to the obesity paradox, which is the inverse association between body mass index and mortality in critically ill patients, which has already been observed in some studies in the setting of COVID-19 [38].
Finally, Brown women with obesity were the most likely to die from COVID-19 in this study. Data from a recent technical report produced in the state of Rio Grande do Sul show worse socioeconomic and living conditions (education, income, work, internet access, quality of life, political participation, and racial violence) for the Black population than for the White population, and when people self-identifying as Brown and Black are compared, all indicators are worse for the Black population [28]. These data refer to the concepts of colorism and pigmentocracy, both of which are based on the idea that people are treated differently based on their skin tone and assume that a Black person with a lighter skin tone will suffer less prejudice, as it is understood to approximate whiteness39. This difference in position in the social hierarchy would still have gender and class implications, in an intersectionality perspective, where Black women would be in a position of lower social prestige [40]. In Brazil, this becomes clear when looking at income statistics: Black women, for example, have the lowest income [41]. In this respect, it can be hypothesized that Black women accumulate vulnerabilities, such as poverty, food insecurity, and malnutrition, so that the presence of obesity does not play a decisive role in increasing the risk of COVID-19 mortality, or may even be a protective factor (obesity paradox) [42]. In Brown women, however, who are in a social position less affected by vulnerabilities, comparatively, obesity is a factor that increases the risk of mortality. The prevalence of obesity was the same in the two groups in this study.
Limitations and strengths
This study has some limitations. Both obesity and race/skin color were self-reported or reported by the accompanying person, and in patients with more severe disease, they were probably more often reported by the accompanying person, which may have led to differential classification error in these cases. Regarding obesity, overweight people may have reported having obesity and vice versa, which may have introduced classification bias. Also, it was not possible to classify individuals according to different degrees of obesity. Some variables for which mandatory reporting is not required had a considerable number of missing data. The variable ‘level of education’ was not included in the adjustments because of the high percentage of missing data (57.2%), although we performed an analysis adjusted for education, and even with the reduced sample size, the results remained consistent. The variables ‘race/skin color’ and ‘mortality’ had, respectively, 0.4% and 4.4% of missing data, so a random pattern of missing data was assumed. Finally, logistic regression may have overestimated the association measures found in this study.
Strengths of this study include the longitudinal design, with a large sample of persons hospitalized for COVID-19 in the first two years of the pandemic in Brazil. The analyses followed a previous conceptual model and were stratified by gender, which allowed an intersectional analysis of the findings. Finally, to our knowledge, this is the first study to explore the role of race/skin color in the association between obesity and COVID-19 mortality, contributing to a better understanding of the impact of racial inequalities on the COVID-19 pandemic.