Brazil was experiencing not one pandemic but many
Em meio à segunda onda da pandemia, pesquisadores brasileiros revelaram que, até abril de 2021, apenas quinze por cento da população havia desenvolvido anticorpos contra o coronavírus — um número que, ao ser desdobrado por regiões, expõe não uma epidemia nacional, mas muitas. Conduzido antes que a vacinação em massa tivesse começado, o estudo oferece um retrato raro da penetração real do vírus no tecido social do país, independente de qualquer intervenção imunológica artificial. A ciência, aqui, não apenas mede o passado — ela ilumina as desigualdades que moldaram o presente.
- Com casos em alta e hospitais sob pressão, a urgência de entender o real alcance da infecção no Brasil tornou-se incontornável.
- Os dados revelam uma fratura profunda: enquanto o Amazonas registrou quase um terço da população com anticorpos, o Ceará mal chegou a dez por cento — sugerindo que o país viveu epidemias paralelas, não uma só.
- O caso de Araraquara ilustra o perigo oculto nos números baixos: uma prevalência de apenas quatro por cento em janeiro antecedeu uma explosão de casos tão grave que forçou lockdown semanas depois.
- Com menos de um por cento dos participantes vacinados, o estudo captura um momento-chave: a fotografia mais fiel da disseminação natural do vírus antes que a imunização alterasse o quadro.
- Pesquisadores alertam que estratégias nacionais uniformes podem ser insuficientes — cada região exige uma resposta calibrada à sua própria curva epidêmica.
No final de abril de 2021, enquanto o Brasil enfrentava uma nova escalada de casos, um vasto inquérito epidemiológico tentava responder a uma pergunta fundamental: quantos brasileiros já haviam sido infectados pelo coronavírus? A resposta — quinze por cento — parecia simples, mas escondia uma realidade muito mais fragmentada.
O estudo Epicovid-19 BR 2, liderado pelo epidemiologista Marcelo Burattini, da Universidade Federal de São Paulo, testou 120 mil pessoas em 133 municípios entre janeiro e abril de 2021. O que emergiu não foi uma curva nacional uniforme, mas um mosaico de epidemias distintas. No Amazonas, quase um terço dos testados carregava anticorpos. No Ceará, o índice mal ultrapassava dez por cento. Rio de Janeiro chegou a vinte por cento; São Paulo, o estado mais populoso do país, ficou em treze — mas mesmo dentro do estado as diferenças eram marcantes, com cidades como Marília registrando o dobro da prevalência de Sorocaba ou Bauru.
O momento da coleta de dados era decisivo. A campanha de vacinação nacional estava apenas engatinhando quando a maioria das amostras foi coletada, e menos de um por cento dos participantes havia recebido alguma dose. Isso tornava os resultados um retrato genuíno da infecção natural — uma linha de base sem interferência vacinal.
O caso de Araraquara funcionou como um alerta incorporado nos números. Em janeiro, a cidade mostrava prevalência de apenas quatro por cento. Semanas depois, uma explosão de casos levou as autoridades a decretar lockdown para evitar o colapso das UTIs. Burattini sugeriu que a baixa imunidade prévia pode ter contribuído para o surto, embora reconhecesse as limitações do tamanho amostral.
Os resultados preliminares também desafiaram previsões anteriores: a taxa de infecção era mais modesta do que modelos baseados em casos assintomáticos haviam projetado. Para Burattini, o estudo é apenas o começo. Quando a análise completa dos 133 municípios estiver concluída e cruzada com os casos oficiais, o Brasil terá ferramentas mais precisas para entender — e enfrentar — as muitas epidemias que se desenrolaram, em ritmos diferentes, dentro de suas próprias fronteiras.
By late April 2021, as Brazil's COVID-19 cases climbed again, a sprawling epidemiological survey had an unsettling answer to a simple question: how many Brazilians had already been infected? The answer was fifteen percent. But that single number concealed a far more complicated picture.
Researchers led by Marcelo Burattini, an epidemiologist at the Federal University of São Paulo, had tested 120,000 people across 133 municipalities between January and April 2021 as part of the Epicovid-19 BR 2 inquiry. What they found was not one pandemic spreading uniformly across the country, but many. In the Amazon state, nearly a third of those tested carried antibodies against the virus. In Ceará, in the northeast, the figure barely reached ten percent. The gap between the highest and lowest regions was stark enough to suggest that Brazil was experiencing not a single epidemic but a collection of them, each with its own trajectory and intensity.
Within states, the variation only deepened. Rio de Janeiro showed antibody prevalence near twenty percent, while São Paulo—the country's largest and most densely populated state—averaged just thirteen percent. But even that masked further fragmentation. In some São Paulo municipalities like Sorocaba and Bauru, fewer than one in eleven residents tested positive for antibodies. In Marília, one in five did. The pattern held across the state: some cities clustered around thirteen to fifteen percent, others fell below ten.
The timing of the survey mattered enormously. Burattini emphasized that the vast majority of those tested had not yet been vaccinated. The national vaccination campaign was barely underway when most samples were collected in January and February. Fewer than one percent of participants reported receiving any vaccine dose, and almost none had completed a full course. This meant the antibody data reflected genuine infection patterns, uncontaminated by vaccination effects—a baseline snapshot of how far the virus had actually penetrated Brazilian society by that point in the pandemic.
Araraquara, a city in the interior of São Paulo state, offered a cautionary tale embedded in the numbers. When researchers tested 258 residents there in late January and early February, they found an antibody prevalence of just four percent. Within weeks, the city erupted. Between late February and early March, cases exploded so violently that authorities imposed a lockdown to prevent hospital intensive care units from collapsing entirely. Burattini noted that the low antibody count in the earlier sample might partially explain the subsequent surge, though he stopped short of certainty—the sample size had been small, and other factors could have driven the outbreak.
The preliminary findings suggested something unexpected: the proportion of Brazilians who had been infected was significantly lower than earlier predictions about asymptomatic cases would have suggested. If so many infections went undetected, the thinking went, the true infection rate should have been much higher. Instead, the survey indicated a more modest penetration of the virus into the population, at least as of April.
Burattini framed the work as foundational. Once researchers completed their analysis of all 133 cities and compared the antibody data against officially reported COVID-19 cases, they would have crucial information for shaping public health policy. The regional variation alone suggested that one-size-fits-all national measures might miss the mark. Understanding why some areas had been hit so much harder than others, and why asymptomatic infection appeared less common than expected, could reshape how Brazil understood the disease's actual behavior within its borders. The full results were still months away, but the preliminary picture was already clear: Brazil's pandemic was not a single story but dozens of them, unfolding at different speeds in different places.
Citações Notáveis
We have multiple epidemics, not just one in Brazil. This variation extends across the entire country.— Marcelo Burattini, Federal University of São Paulo
The seroprevalence recorded is significantly lower than predictions about asymptomatic cases would suggest.— Marcelo Burattini
A Conversa do Hearth Outra perspectiva sobre a história
Why does it matter that most people tested hadn't been vaccinated yet?
Because it lets you see the real infection landscape without vaccination muddying the picture. You're measuring who actually got sick, not who got protected. That's your baseline.
But fifteen percent sounds low for a year into a pandemic. Shouldn't it be higher?
That's exactly what surprised the researchers. If asymptomatic cases were as common as people thought, the number should have been much higher. Instead, it suggests either fewer people got infected than expected, or the asymptomatic rate isn't what we assumed.
What explains why Amazonas had triple the rate of Ceará?
That's the question they're still working through. Different regions had different waves, different population densities, different behavior patterns. It wasn't one pandemic—it was many, happening at different times in different places.
The Araraquara example is striking. Four percent in February, then lockdown in March. What does that tell us?
It suggests the virus was already circulating there, just not showing up in the antibody test. Or it arrived suddenly after the survey. Either way, it shows how quickly things could change, and how a snapshot in time can miss what's coming.
So what do they do with this data now?
Compare it against official case reports and death records. If the antibody numbers don't match the reported cases, that tells you something important about testing, reporting, or how the disease actually spreads. That shapes the next round of policy.