Study finds 76% of Manaus residents have COVID-19 antibodies, far exceeding São Paulo's 29%

Brazilian researchers published findings this week in the journal Science that painted a stark picture of how differently the pandemic had moved through two major Brazilian cities. In Manaus, the capital of Amazonas state, three-quarters of the population—76 percent—had developed antibodies to the coronavirus by October. In São Paulo, a city of more than 12 million people, the figure was 29 percent. The gap was not a matter of chance or timing. It reflected the speed and severity with which the virus had swept through Manaus, overwhelming hospitals and funeral services alike.

The study, led by epidemiologist Ester Sabino from the University of São Paulo's Institute of Tropical Medicine, analyzed blood samples collected between March and June, then extended the data through October. Researchers tested more than 6,000 samples from blood donors at Manaus's hemotherapy foundation and compared them against 1,000 samples from São Paulo's blood bank. The work was methodical: they measured IgG antibodies, the type that persists after infection, and accounted for false negatives by building mathematical models around the data. By June, when the pandemic peaked in Manaus, about 52 percent of residents had already been infected. In São Paulo during the same month, the figure was 13.6 percent.

Manaus had been hit with exceptional force. The first confirmed case arrived on March 13, and by early December, the city had recorded 72,876 cases and 3,152 deaths. The healthcare system fractured under the weight. Funeral services could not keep pace. The antibody prevalence of 76 percent exceeded the theoretical threshold for herd immunity—the point at which enough of a population has immunity that transmission slows dramatically—which epidemiologists estimate at 60 to 70 percent. Yet this milestone came at an enormous human cost, not as a sign of triumph.

The researchers had initially presented preliminary findings in September showing 66 percent prevalence, before peer review. The October data pushed the number higher. They had also tested blood from patients with moderate COVID-19 symptoms to account for the possibility that blood donors—who must test negative on RT-PCR screening—might have lower antibody rates than the general population. Many people carry asymptomatic infections without knowing it, and these lighter cases still trigger antibody production, though at lower levels than severe illness. When the team modeled all possible false negatives, they arrived at a maximum estimate of 76 percent.

What the antibodies revealed was not immunity from further spread, but rather a population that had already paid the price of exposure. After the June peak, the rate of new infections continued but at a slower pace than it would have without any precautions. Sabino noted that if no preventive measures had been taken—no distancing, no masks—the prevalence could have reached 94 percent. The measures that were adopted did not stop transmission; they merely slowed it. By August, the prevalence had climbed to 66 percent in Manaus while São Paulo remained at 30 percent, with 20,000 deaths recorded in the larger city.

Why Manaus burned through the population so much faster remains partly mysterious. Population density was one hypothesis, but cell phone mobility data showed that residents of Manaus had reduced their movement starting in March, just as São Paulo residents had. Sabino suggested that the nature of movement itself—much of it on foot, some by boat—may have accelerated transmission in ways the data could not fully capture. The pandemic's trajectory in Manaus was simply more compressed and more severe.

Questions about immunity and protection remain open. Vaccines in late-stage trials appeared to generate stronger immunity than natural infection, but the durability of antibodies from infection was unknown. Whether people who had recovered from COVID-19 should still be vaccinated was not yet settled. Sabino was careful not to overstate what the antibodies meant: they indicated some protection, but the extent of that protection against reinfection was unclear. The study had mapped the pandemic's path through one city with precision, but it had not answered all the questions about what came next.

The research team plans to extend the work to six other Brazilian capitals: Rio de Janeiro, Belo Horizonte, Salvador, Recife, Fortaleza, and Curitiba. Each city would tell its own story about how the virus moved, how populations responded, and what immunity looked like when it emerged not from vaccination but from the hard passage of disease through a community.