Abstract
Chronic diseases, worse socioeconomic conditions and old age can increase infection and hospitalization rate due to Coronavirus disease (COVID-19). We assessed the association between the burden of NCDs and the occurrence of infections and hospitalizations of COVID-19 in Brazil in a large COVID-19 national survey data. We analyzed only data collected between July and November 2020 (n = 1,071,782). The frequencies of positive COVID-19 diagnosis and NCD burden were estimated according to age, sex, socioeconomic strata and skin color categories. We estimated hazard ratios and 95% confidence intervals using Cox regression models. There is a non-linear dose-response inverse association between per capita income and the rates of infection and hospitalization due to COVID-19. The presence of NCDs was associated with a higher incidence of COVID-19 infection (HR1NCD = 1.34; 95% CI: 1.26; 1.43; HR2 or more NCD= 1.54 95% CI: 1.39; 1.71) and incidence of hospitalization (HR1NCD = 3.08 95% CI: 2.26; 4.19; HR 2or more NCD= 6.81 95% CI: 4.88; 9.49).The difference between the risks of infection or hospitalization of COVID-19 attributable to the burden of NCDs is non-linearly associated with the income.
Text
Worldwide, over 93 million cases and over 2 million deaths have been reported since the start of the Coronavirus disease (COVID-19) pandemic by the World Health Organization (WHO). Brazil has reported the highest number of new cases, as of 16 March 2021 (1).
According to the WHO, 15% of COVID-19 cases are likely to develop into hospitalizations and 5% require treatment in intensive care units (ICU) (2). Factors such as comorbidities, the presence of chronic diseases, worse socioeconomic conditions and old age can increase the risk of hospitalization (3). According to data from the Pesquisa Nacional de Saúde (National Health Survey) (PNS 2019), approximately 52% of Brazilian adults have at least one chronic non-communicable disease (NCD). Respiratory and cardiac diseases, diabetes and hypertension are the comorbidities leading towards the worst prognosis for COVID-19 (4,5).
In Brazil, considering the prevalence of NCDs, nationwide estimates for adults at risk of severe COVID-19 range from 34.0% (53 million) to 54.5% (86 million) (6). To date we have no evidence that the presence of comorbidities increases the risk of infection or if there is a gradient between the number of comorbidities and the risk of hospitalizations due to COVID-19 (7). Therefore, the aim of this study was to analyze the association between the burden of NCDs and the occurrence of infections and hospitalizations of COVID-19.
METHODS
Data
We used data from PNAD COVID-19 (Portuguese acronym for National Household Sample Survey COVID-19), conducted by the Instituto Brasileiro de Geografia e Estatística (IBGE) and the Brazilian Ministry of Health. PNAD COVID-19 survey followed, through monthly telephone calls, approximately 193,000 households between May and November 2020. The sample was selected from an elaborated registry of probabilistic matching between addresses of households contained in the Brazilian National Household Sample Survey (PNAD 2019) and addresses of six databases with telephone information. The procedures, databases used, questionnaires and sample plan of the research are described in technical documents from IBGE (8). For the present study, we analyzed only data from individuals between 20 and 59 years of age, collected between July and November 2020 (n=2.650.459).
PNAD COVID-19 collected information about exams, results of COVID-19, income and sociodemographic characteristics of residents. It was also asked about previous diagnosis of diabetes mellitus, high blood pressure, chronic respiratory diseases (asthma, bronchitis, emphysema and others), heart diseases (infarction, angina, heart failure, arrhythmia), depression and cancer. Three questions related to the tested specimen were also used to diagnose COVID-19 infection, without discriminating the test performed (molecular biology, antigen or antibody). The oronasal swab test uses the RT-PCR technique to identify viral RNA, usually up to the 8th day of infection (9,10). In the venous blood test it is possible to detect antigen (current infection) or antibody (current or past infection), and it is usually applied from the 8th day after the beginning of the symptoms . The capillary blood test, also available for diagnosis, allows verification of results after 15 minutes of the exam (11). We used three diagnostic outcomes: positive diagnosis in oronasal swab, positive diagnosis in venous blood, and positive diagnosis in capillary blood.
Sociodemographic information was also analyzed, such as sex (male and female), age (in years), skin color (“white”/“yellow” and “black”/“brown”/“indigenous”), information on per capita household income (PCI), which was calculated by the monthly average of individual incomes divided by the total number of individuals in the household and then divided into fifths of the distribution; schooling, complete or not, which was grouped into the following levels: no schooling, elementary school, high school, higher education and post graduation; geographic region and area of residence (urban or rural). Predicted values adjusted for NCD, age, sex, education, income and skin color were created for two probabilistic outcomes, diagnosis and hospitalization by COVID-19.
Statistical Analysis
For the present analysis, we carried out the following preparation: a) selection of individuals between 20 and 59 years old; b) we consider any positive result reported in any of the 3 diagnostic tests asked; c) noncommunicable Diseases (NCD) diagnoses were grouped into 0, 1, 2 or more diseases to represent the NCD burden; d) we selected all cases with complete response for variables under analysis (n = 1,071,782).
The frequencies of positive COVID-19 diagnosis and NCD burden were estimated according to age, geographic and socioeconomic strata and skin color categories. We calculated hazard ratios (HR) and 95% confidence intervals (CI) for the association between NCD and the risk of infection and hospitalization for COVID-19 using Cox regression models via maximum likelihood to account for repeated measures, as each person could contribute with one or more interviews to the study. Findings at p < 0.05 were considered statistically significant. Initially, we assessed the number of chronic noncommunicable diseases in separate models without including any covariates.. Secondly, we included a priori selected demographic variables to assess potential confounding: sex, age, race/skin color, education, household income and rural/urban residency. The adjusted prevalence of COVID-19 was predicted according to NCD burden and stratified by per capita household income. We estimated p-trends for the dose-response relationship between NCD and COVID infection and hospitalization. Finally, predicted values from the model were used to estimate COVID-19 latent frequency according to the burden of NCD as grading risk to the infection or ICU use. All analyses were performed using sampling weights in Stata® 15.1.
RESULTS
The prevalence of COVID-19 was higher among individuals aged 30 to 50 years (58%), while the prevalence of hospitalizations was higher among individuals aged 40 or over more (59%). Around 60% of the people in the age group of 40 years or over had NCDs. The prevalence of 2 or more NCDs among the richest was 1.3 times the prevalence among the poorest, however, the prevalence of hospitalizations was 1.2 times more frequent in the poorest group.
The COVID-19 infection rate as a function of the NCD burden was not statistically associated with sex. Table 2 shows the association between NCD burden and outcomes of COVID-19. The presence of NCDs was associated with a higher incidence of COVID-19 infection for 1 NCD (HR = 1.34; 95% CI: 1.26; 1.43) and for 2 or more (HR = 1.54 95% CI: 1.39; 1.71). Also, the presence of NCDs was associated with higher incidence of hospitalization for 1 NCD (HR = 3.08 95% CI: 2.26; 4.19) and for two or more NCDs (HR = 6.81 95% CI: 4.88; 9.49).
Prevalence of Chronic Noncommunicable Diseases and gross probability of positive diagnosis and hospitalization for COVID-19. Brazil PNAD COVID-19, 2020.
Cox models of risks of positive diagnosis and hospitalizations due to COVID-19 for Noncommunicable Diseases (NCD). Brazil PNAD COVID-19, 2020.
The stratification of the association between NCD and COVID-19 according to family income per capita is shown in Figure 1. The risk of infection by COVID-19 according to NCD tends to decrease as per capita income increases. This trend was not linear, since in the richest the risk of infection fluctuates slightly in a positive way. The risk of hospitalization for COVID-19 among NCD patients shows an inverse trend with per capita income, although this association is tenuous.
Adjusted prevalence of COVID-19 and hospitalization for COVID-19, according to the amount of NCDs and fifths of income, in Brazilian adults aged 20 to 59 years. Brazil PNAD COVID-19 (2020).
In Figure 1, we observed that the risk of hospitalization among patients without NCD did not vary according to income strata, even after adjusting for sex, age, education, income and skin color. Among patients with NCDs infected with COVID-19, the chance of hospitalization was higher among the richest than the poorest.
DISCUSSION
Our results suggest that: a) patients with NCDs are at high risk for infection and hospitalization with the Sars-CoV-2 virus. The risk of hospitalization is approximately twice the risk of infection; b) there is a non-linear dose-response type association between the risks of infection and hospitalization due to COVID-19 and per capita income; c) the difference between the risks of infection or hospitalization of COVID-19 due to the burden of NCDs is non-linearly associated with the per capita income.
Worldwide, approximately one in three adults suffers from multiple chronic diseases (12). In Brazil, approximately 52% of the adult population has at least one chronic disease, which is responsible for 72% of the causes of death in the country (6,13).
The COVID-19 pandemic highlights the role of NCDs as risk factors for infection, hospitalization and death. Global data show that 1.7 billion people, 22% of the world population, have at least one underlying condition that puts them at increased risk for COVID-19 infection(14) In Brazil, from one third (53 million) to half (86 million) of adults have at least one risk factor for COVID-19 (7).
The association between NCDs, diagnosis and hospitalization for COVID-19, observed in our study, is in line with results from other countries (15–19). A meta-analysis carried out with Chinese data showed that the most prevalent comorbidities among patients with COVID-19 were hypertension (21%), diabetes (10%), cardiovascular disease (8%) and respiratory system disease (2%). The combined OR for presenting a severe case among those with hypertension, respiratory system disease and cardiovascular disease was 2.36 (95% CI: 1.46; 3.83), 2.46 (95% CI: 1.76; 3.44) and 3.42 (95% CI: 1.88; 6.22) respectively (15). Analyzes based on American clinical data indicate that the majority of outpatients with COVID-19 studied were African American (72%), and had at least one comorbidity (94%), including hypertension (64%), chronic kidney disease (39 %) and diabetes (38%). Severe obesity (OR = 2.0; 95% CI: 1.4; 3.6) and having chronic kidney disease (OR = 2.0; 95% CI: 1.3; 3.3) were the strongest predictors for admission to the intensive care unit (16).
Increasing evidence suggests that NCDs are determinants of the prognosis for infectious diseases (20). The gradient between the number of NCDs and the evolution to a bad prognosis of COVID-19, as found in the present study, brings up new challenges. The policy makers need to take into account the management of actions that strengthen equity attention to the social and biologically vulnerable population when elaborating strategies for health care (19).
Our results indicate that poorer individuals are more vulnerable to contagion and hospitalization due to COVID-19 in Brazil. Previous population epidemiological data (H1N1 and SARS) demonstrates the impact of social inequality as determinants for higher contagion and severity of these diseases (21–23). Factors such as lower wages, quality of public transport, unsanitary housing conditions, inefficient health care and difficulty in maintaining social isolation without significant loss of income corroborate the increased vulnerability of this population. (24,25). With regard to inequality in access to health and the needs to cope with the pandemic of COVID-19, there is an increase in the demand for health care and intensive care beds (ICU) with a range of attention to the population with dependence on public health care (3,26).
In Brazil, 71.5% of the population exclusively uses public health services, showing a direct relationship with income ranges per capita (27). The recommendation of WHO about the adequate ratio of Intensive Care (ICU) beds is 1 to 3 beds for every 10,000 inhabitants. In Brazil, the SUS (portuguese acronym for National Public Health System) has an average of 1.4 beds for every 10,000 inhabitants against 4.9 in private health service (28).
Our findings revealed that among NCDs patients infected with COVID-19, the chance of hospitalization was higher for the richest relative to the poorest. Income determinants seem to be associated with less access to health and intensive treatment of the disease. In our interpretation, poorer individuals, even though they are more vulnerable to the contagion of the disease, are less likely to obtain access to treatment and hospitalization for COVID-19.
As a limitation of the study, the classic limitations of observational studies can be cited, such as: self-report of information and lack of specificity and limited amount of information on clinical diagnoses. Thus, our analyzes were conditioned to the volume and type of information available in the database. As strengths of the study, longitudinal monitoring and the national representativeness of the home-based sample can be pointed out, with a high response rate. In addition, the sample size allows for consistent stratifications throughout the analysis (29).
These findings and methods of estimating the impact of vulnerable groups on the progression of the pandemic may be useful for planning and managing prevention and treatment strategies in Brazil. A better management and direction of policies that aim to obey the principle of universality of rights in health care, covering the population in a state of biological and social vulnerability is necessary. It is imperative to protect groups at greatest risk, such as people with chronic diseases and comorbidities in scenarios of broad measures of social isolation, particularly while the capacity for immunization is limited.
CONCLUSION
The burden of NCDs increases the risk of infection and hospitalization by COVID-19. This is troublesome in a context of increasing social and health inequalities as inBrazil. This situation, associated with the prospects caused by COVID-19, projects a scenario of increased demand for clinical care in the coming years.
Data Availability
The data used in this work can be obtained free at the Data service (https://www.ibge.gov.br/en/statistics/experimental-investigations/experimental-statistics/27975-weekly-release-pnadcovid1.html?=&t=microdados) Further information regarding the sample design and data collection methods can be found on the survey website (https://www.ibge.gov.br/en/statistics/experimental-investigations/experimental-statistics/27975-weekly-release-pnadcovid1.html?=&t=notas-tecnicas).
Author Bio
Researcher and member of the Laboratory of Nutritional Assessment of Populations of the University of São Paulo. Currently study epidemiological and nutritional developments in populations.
