Chronic non-communicable diseases and COVID-19: EPICOVID-19 Brazil results

ABSTRACT OBJECTIVE Describing the prevalence of chronic diseases and associated socioeconomic and demographic factors, evaluating the patterns of social distancing and the antibodies prevalence against SARS-CoV-2 and COVID-19 symptoms in carriers and non-carriers of chronic diseases. METHODS Data from 77,075 individuals aged 20 to 59 from three steps of the EPICOVID-19 Brazil (a nationwide serological survey conducted between May and June, 2021) were assessed. The presence of antibodies against SARS-CoV-2 was examined by rapid tests. Self-reported prevalence of hypertension, diabetes, asthma, cancer, chronic kidney disease and heart disease were investigated. The prevalence of mask use, adherence to isolation measures and antibodies were evaluated separately amid carriers and non-carriers of chronic diseases. The prevalence of symptoms was analyzed among carriers and non-carriers of chronic diseases with antibodies. RESULTS The prevalence of at least one chronic disease was 43%, higher in the Southeast region, among white and indigenous individuals, women, less schooled and in lower socioeconomic position. The use of masks when leaving home was similar among carriers and non-carriers of chronic diseases (98%). The proportion of participants who reported adherence to isolation measures was higher amid carriers (15.9%) than non-carriers (24.9%) of chronic diseases. The prevalence of antibodies to SARS-CoV-2 was similar amongst carriers and non-carriers (2.4% and 2.3%). The prevalence of cough, dyspnea, palpitations and myalgia was significantly higher among carriers, but the proportion of symptomatic patients was similar between groups. CONCLUSION The prevalence of chronic diseases in Brazil is high and the COVID-19 pandemic affects carriers and non-carriers of chronic diseases similarly. Carriers present more severe forms of COVID-19 and higher prevalence of symptoms. Greater adherence to social distancing measures among chronic patients is disassociated from a lower incidence of COVID-19 in this group.


INTRODUCTION
By early April, 2021, about 132 million COVID-19 confirmed cases and 3 million deaths were accounted worldwide. The Americas and Europe comprise most of the confirmed cases. Even though Brazil accounts for approximately 3% of the global population, it comprises 10% of all cases and deaths worldwide 1 . COVID-19 mortality depends not only on the number of incident cases, but also on population factors such as age structure and comorbidity burden connected to the development of severe forms of the disease 1 .
Due to the demographic and epidemiological transition in progress, Brazil displays a high load of NCDs 7,8 . The prevalence of obesity and physical inactivity, crucial risk factors for these morbidities, is also high 8,9 . This scenario is alarming in the Brazilian COVID-19 pandemic context, as it increases the risk of severe cases in a country with an elevated incidence. Therefore, the assessment of the panorama of NCDs associated with COVID-19 infection in Brazil is fundamental. This study aims at describing the prevalence of NCDs and associated socioeconomic and demographic factors, in addition to evaluating the patterns of social distancing and the prevalence of antibodies against SARS-CoV-2 among carriers and non-carriers of NCDs, as well as the symptom occurrence among individuals with antibodies against SARS-CoV-2 carrying and not carrying NCDs.

METHODS
We analyzed data from EPICOVID-19 Brazil, a series of national serological surveys carried out aiming at evaluating the magnitude, evolution and characteristics of the pandemic. The most populous city of each of the 133 geographical intermediate regions, according to the Brazilian Institute of Geography and Statistics (IBGE), were selected to take part in the study. Multiple stage sampling was carried out in each city, having the 2010 IBGE's urban census sectors as primary sampling units. We selected 25 sectors with probability proportional to size in each city. In each sector, 10 households were systematically selected. In each household, a resident was drawn and invited to participate in the study. In case of refusal, a second resident was drawn. If the second resident had also refused, the interviewers would have approached the next household, immediately on the right. The same replacement procedure was adopted if the selected household was empty or did not open the door to the interviewer. All refusals and substitutions were recorded by the field team.
Each participant answered a questionnaire on socioeconomic and demographic characteristics, social distancing, mask use, presence of COVID-19 symptoms, previous diagnosis of NCDs, among others. Participants underwent the WONDFO SARS-CoV-2 rapid Antibody Test (Wondfo Biotech Co., Guangzhou, China), which detects the presence of antibodies against SARS-CoV-2 (IgG and IgM). The test was previously validated by the responsible research team 10,11 . The interviewers received standardized training on the questionnaire application, rapid test administration, and results analysis. All positive rapid tests and 20% of the negative ones were evaluated by a second observer. EPICOVID-19 Brazil's detailed methodology was previously published 12 .
Data were collected in three EPICOVID-19 Brazil stages, conducted from May 14 to 21 (n = 25,025), from June 4 to 7 (n = 31,165) and from June 21 to 24 (n = 33,207), 2020. The analyses here were restricted to adults aged 20 or older.
To assess the existence of chronic diseases, patients were asked if they had ever been diagnosed with hypertension, diabetes, asthma, cancer, chronic kidney disease and heart disease ("yes" or "no"). Individuals who reported a diagnosis of one or more NCDs were considered carriers of NCDs and will be mentioned in this way throughout the text.
The presence of antibodies against SARS-CoV-2 was evaluated by the rapid test results: when the result was positive, the individual was considered to have antibodies against the virus. Mask use was assessed by the question "Do you wear a mask when you leave your house?" ("yes", "no" or "I do not leave my house"). Adherence to social distancing was assessed by the following questions: • "How much of the social distancing guidelines recommended by health authorities (i. e., staying at home and avoiding contact with other people) do you follow?" ("very little", "little", "more or less", "quite a lot" or "mostly isolated from everyone"); • "How has your daily routine been?" ("I stay at home all the time", "I leave my house only for essential things, such as buying food", "I go out from time to time to shop and stretch my legs", "I leave my house every day for some activity", "I go out every day, all day, to work or to do other regular activities"); • "Thinking about the household routine, who has been attending the house?" ("only family members who live together, if any, and no one else", "some close relatives visit once to twice a week", "some close relatives visit almost every day", "friends, relatives or others visit once to twice a week", "friends, relatives or others visit almost every day").
To assess the presence of COVID-19 symptoms, participants were asked if they had the following symptoms: fever, cough, sore throat, changes in smell or taste, myalgia, twitching, headache, breathing difficulties, palpitation, diarrhea and vomiting. Participants who reported the occurrence of at least one symptom were considered symptomatic. The recall period of questions about symptoms differed between the stages of EPICOVID-19 Brazil. The first two stages evaluated the occurrence of symptoms in the 15 days prior to the interview and the third stage evaluated the period from March 2020 to the interview date.
Since antibodies against SARS-Cov-2 become detectable in the rapid test about 15 days after infection, the analysis on the occurrence of symptoms was restricted to the participants of the third stage.
The evaluated socioeconomic and demographic variables were geographic region (North, Northeast, Southeast, South or Midwest), self-reported skin color (white, black, brown, yellow or indigenous), gender (female or male), age (in categories) and schooling levels (elementary school, high school or higher education). To assess the socioeconomic position of households, the National Economic Indicator (IEN) was used. This index was created by the analysis of main components based on a set of household characteristics and goods (computer, internet access, color television, air conditioning, motor vehicles, cable television, number of rooms used for sleeping and number of bathrooms). The first component was extracted and divided into quintiles. The first quintile represents the poorest 20% and the fifth quintile, the richest 20% 13 .
The analyses included the sample description and the prevalence calculation of each NCD and the prevalence of carriers of NCDs. The prevalence of carriers of NCDs was assessed according to region, skin color, gender, age, IEN and schooling levels. The prevalence This study respected all the ethical precepts and legislation governing research with human beings and was approved by the National Commission for Research Ethics (CAAE 30721520.7.1001.5313). All participants signed the informed consent form.

RESULTS
We analyzed EPICOVID-19 Brazil's data from 77,075 participants aged 20 or older. Most were female (59%) and non-white (62%). About a quarter were older adults (aged 60 or older). Around 40% of the participants reported having attended elementary school (39%) and lived in the region. Considering the three stages of the study, the crude proportion of participants with antibodies against SARS-CoV-2 was 2.3% (95%CI 2.3-2.5). Circa 45% of those reported carrying one or more NCDs. The most prevalent NCDs were arterial hypertension (30%), followed by diabetes (13%), asthma (9%) and heart disease (8%) ( Table 1). Table 2 shows the prevalence of one or more NCDs according to socioeconomic and demographic characteristics. The prevalence was higher in the Southeast and South (48%), among indigenous people (47%), women (47%) and individuals with lower schooling levels (57%). The prevalence of one or more NCDs presented an inverse relationship with age and quintiles of the IEN. All differences were statistically significant (p-value < 0.001).
The results on social distancing are presented in Table 3. Around 25% of individuals carrying one or more NCDs reported being isolated from virtually everyone. Among non-carriers, this percentage was 15%. Regarding the daily activities, 13% of those with one or more NCDs reported leaving their houses every day to work, while among those without NCDs this proportion was 25%. Both among carriers and non-carriers, the proportion that reported that only the residents themselves attended the household was approximately 50%. A higher proportion of carriers with one or more NCDs reported receiving close relatives (44%) when compared to non-carriers (39%).
Except for individuals who reported a diagnosis of cancer, the differences in the prevalence of antibodies against SARS-CoV-2 between carriers and non-carriers of NCDs were minimal (Table 4). Table 5 shows the prevalence of COVID-19 symptoms among individuals with antibodies against SARS-CoV-2, both carriers and non-carriers of NCDs. Carriers of one or more NCDs reported 50% more twitching occurrences than non-carriers, in addition a to a 26% and 29% higher occurrence of coughing and dyspnea, respectively. Myalgia was 16% more prevalent among carriers of NCDs. The prevalence of the other evaluated symptoms and the proportion of asymptomatic individuals were statistically similar.

DISCUSSION
Our results showed that Brazil has a high burden of chronic diseases. Despite greater adherence to social distancing measures among carriers of NCDs, the prevalence of antibodies against SARS-CoV-2 was similar between groups.
Each morbidity assessed here depicted a higher prevalence when compared to the National Health Survey results, in which 21% of participants reported a diagnosis of hypertension 14 (versus 30% in the present study), 6.2% of diabetes 15 (versus 13%), 4.4% of asma 16 (versus 9%), 1.8% of cancer 17 (versus 3%), 1.4% of chronic kidney disease 18 (versus 4%) and 4.2% heart disease 19 (versus 8%). The results were also higher than the 2018 VIGITEL ones, which estimated self-reported prevalence diagnosis of hypertension and diabetes of 25% and 8%, respectively 8 . The higher prevalence of NCDs evidenced here is possibly related to the  high percentage of people aged 60 or older in the sample (26%), higher than the percentage of this group among Brazilians in general (14%). As shown in Table 2 and corroborated by other studies, the prevalence of NCDs increases with age 7 .
There are two probable reasons for the higher proportion of older adults in this study compared to the rest of the population. Possibly because of the pandemic, most of the older adults living in the households included in the sample were present at the time of the study, unlike younger people. A population-based study on social distancing patterns conducted in Rio Grande do Sul showed that older adults are more likely to stay home or leave their houses only for essential activities (80%) than individuals aged between 20 and 59 (60%) 20 .
Another relevant aspect is that we assessed 133 sentinel cities, which are more populous, have a higher Human Development Index and greater access to health services than the cities excluded from the study. Because of those characteristics, residents of these cities tend to live longer, so the proportion of older adults tends to be higher.
The mask use and physical distancing drastically reduce the transmission of COVID-19 21 .
The percentage of participants who reported wearing a mask when leaving home was almost universal in both groups (98%). However, we must consider an information bias when interpreting the results. Knowing the legal requirements of mask use during the pandemic, participants might have reported that they always use masks because this is the correct behavior in the current context.
Carriers of NCDs reported greater adherence to social isolation and universal mask use when leaving home. Nevertheless, the prevalence of antibodies against SARS-CoV-2 in this group was similar to that evidenced among non-carriers of NCDs. This may suggest the occurrence of intra-household transmission of COVID-19, since the proportion of participants who are visited by close relatives and friends was about 50% in both groups. COVID-19 transmission in clusters, especially family ones, plays an important role in the rapid spread of the disease 22,23 . The almost universal mask use outside the house is probably rejected at home. Thus, family members who adhere less to isolation and, therefore, present similar infection levels to those evidenced among non-carriers of NCDs would contaminate the carrier ones. However, the adoption of preventive measures such as wearing a mask and physical distancing at home, as well as in-house social distancing, was not evaluated in this study, which prevents us from formally testing this hypothesis. Besides, as previously mentioned, mask use might be overestimated or masks may be worn incorrectly. We must also consider reverse causation. Participants with antibodies against SARS-CoV-2 may have switched to wearing a mask after experiencing COVID-19.
The analysis on the symptom prevalence was restricted to individuals with antibodies against SARS-CoV-2. Among the evaluated COVID-19 symptoms, the prevalence of cough, dyspnea, myalgia and twitching was higher among carriers of NCDs than among non-carriers. Despite the higher prevalence of these symptoms among carriers of NCDs, we must consider an eventual non-responder bias. COVID-19 tends to evolve into severe forms in carriers of NCDs 2-4 , and severe cases tend to leave after-effects even months after the disease 24 . Therefore, carriers of NCDs possibly developed severe forms of the disease with prolonged after-effects and presented a higher refusal rate, which may have decreased the difference in the symptom prevalence amid groups. In addition, COVID-19 mortality among carriers of NCDs is higher when compared to non-carriers. Hence, participants with a positive result on the rapid test here may represent, to some extent, surviving cases.
This study has other limitations. We assessed data from 133 sentinel cities, which are more developed, have a better health services structure and higher Human Development Index than the cities excluded from the sample. We excluded residents of rural areas, which correspond to 15% of the country's population. We ignored the adoption of protective behaviors, such as mask use and physical distancing during the interaction with close acquaintances, as well as adherence to distancing recommendations during daily in-house activities. The evaluation of self-reported symptoms may have introduced classification bias. However, we revealed the rapid test results only at the end of the interview, so the participants were blind to them while answering the questionnaire. This possibly decreased the bias likelihood.
The strengths of this study include: its national scope, being one of the world's largest COVID-19 serological surveys in sample size and territorial coverage carried out so far; the blinding of participants about the test results; and the multidimensional character of the research, which evaluates both seroprevalence and several individual aspects related to COVID-19, such as the presence of comorbidities, adoption of preventive measures against COVID-19 and symptom occurrence.
The results here improve the understanding of the pandemic dynamics in Brazil, where the NCDs burden, as well as incidence and mortality due to COVID-19 are high. Since 2016 Brazil has been suffering from fiscal austerity policies, which constricted the Unified Health System (SUS). Consequently, the supply and quality of actions and services decreased, compromising the health of a large portion of Brazilians who depend on it 25 . Combined with this context, the absence of a national policy to deal with COVID-19 based on scientific evidence 26,27 contributes substantially to Brazil being placed among the top countries at the global ranking of new infections and deaths from the disease 1 .
The pandemic control in Brazil requires large-scale immunization of the population. Immunization against COVID-19 in Brazil began in mid-January 2021, firstly applied to health professionals working on the front line against the virus. Carriers of NCDs are part of the third priority group at the Brazilian Ministry of Health's National Immunization Plan 28 . Despite having one of the largest and most efficient immunization networks in the world, vaccination in Brazil is still incipient and the country currently ranks 78th in the world ranking of countries with the highest number of citizens vaccinated against the disease (10.3 /100 inhabitants) 29 . The immunization of NCDs carriers lacks a prediction.
In conclusion, the prevalence of chronic diseases in Brazil is high and the COVID-19 pandemic affects carriers and non-carriers of chronic diseases similarly. Carriers of NCDs present more severe forms of COVID-19 and higher symptom prevalence. Greater adherence to social distancing measures among carriers of NCDs is disassociated from a lower incidence of COVID-19 in this group. More than ever, we must stimulate protective measures such as physical distancing and wearing a mask at home in this group.