Study Estimates 76 Percent of Brazilian City Exposed to SARS-CoV-2
Study Estimates 76 Percent of Brazilian City Exposed to SARS-CoV-2

Study Estimates 76 Percent of Brazilian City Exposed to SARS-CoV-2

The number, extrapolated from antibodies present in blood donors in Manaus, should be treated with caution, experts warn.

Ignacio Amigo
Dec 14, 2020

ABOVE: © ISTOCK.COM, DANIEL FIGUERIEDO

In a study published in Science December 8, a group of researchers estimate that 76 percent of the population of Manaus, the first Brazilian city to be hard hit by the coronavirus pandemic, could have been infected with the SARS-CoV-2 virus so far, based on antibodies present in samples from blood banks. The report also estimates that in São Paulo, Brazil’s largest city, that percentage is much lower, accounting for 29 percent of the population. According to the paper, these results highlight the failure of Manaus to control the spread of the virus. 

The findings are based on the analysis of samples from donor blood centers. Because Brazilian law establishes that blood banks should keep samples for six months, the researchers were able to go back in time and trace the unravelling of the pandemic between February and October. In the case of Manaus, they observed that the proportion of blood donors with antibodies against SARS-CoV-2 was around 5 percent in April, but increased steeply to almost 40 percent in May. After adjusting the results for age and sex, for the sensitivity of the test, and for the decay in antibodies in the samples, the researchers extrapolate that by June, 66.2 percent of the population had been infected, a figure that rose to 76 percent by October. An average of around 900 to 1,000 samples were analyzed for each month and city.

“The sixty-six percent figure sets the theoretical threshold for herd immunity, but cases are still increasing in Manaus,” says Ester Sabino, an immunologist at the University of São Paulo and lead author of the study.

An early version of the paper, uploaded to a preprint repository in September, stated that the high presence of antibodies against SARS-CoV-2 in the population meant that Manaus would have achieved herd immunity. The manuscript was met with fierce backlash by part of the medical and scientific community, who considered that such a claim could make the population take fewer precautions and could be used by politicians to lift restrictions. Just a few days after the preprint came out, the city experienced a second wave of cases after a period of stability. While both events happened too close in time to be causally connected, the paper may have contributed to a delay in the response and a failure to acknowledge the severity of the situation, according to news outlet Amazonia Real.

With COVID-19 cases continuing to be very high in Manaus, and nearly all intensive care unit beds currently occupied, there is clearly a substantial amount of the population still vulnerable.

To address this criticism, the peer-reviewed version of the study has downplayed the statements referring to herd immunity, noting that “prior infection might not confer long-lasting immunity.”

“It looked like we had said that by reaching herd immunity the epidemic was over, but that was not what we meant, so we rewrote the text to try to improve that,” says Sabino.

Located in the heart of the Amazon forest, hundreds of miles away from the country’s economic and political centers, Manaus is an industrial city with a population of more than 2 million. 

Since the pandemic started, the city has had more than 73,000 confirmed positive cases of COVID-19 and some 3,000 deaths, according to official figures. However, experts agree that the real figures are probably much higher, as few tests are performed. The shocking photographs of mass graves in the city’s cemetery back in April were some of the earliest graphic testimonies of the severity of the pandemic. According to a study published in a Brazilian journal, the number of total deaths in Manaus in April was more than 4.5 times higher than in previous years.

Extrapolation from blood donors

All the experts who spoke with The Scientist agree that the main limitation of this new study is that blood bank donors are not representative of the total population. The researchers adjusted for some of the demographic skew in the sample by comparing it to the population census, but the approach is far from perfect.

Jesem Orellana, an epidemiologist from Manaus who works at the Fundação Oswaldo Cruz, a public research institution, notes the prospect of a free COVID-19 test upon donating blood may have prompted specific groups of people to contribute.

“Many people may have donated blood to know their serological status, especially those who suspected they had been infected and with less financial resources,” Orellana says. He adds that the blood center even partnered with Uber to offer a discount to donors heading to the facility. All of this might have influenced the outcome of the study.

See “Researchers in Brazil Struggle to Get Solid COVID-19 Death Counts”

Another challenge for the researchers was dealing with antibody waning. The kit used in the study was able to detect IgG antibodies against SARS-CoV-2 fairly well if the patient had been infected recently. But as weeks passed, the signal became weaker and more difficult to identify, according to the study.

To overcome this, the authors developed a mathematical model to correct for cases that were likely positive based on a weak antibody signal, but overlooked by the assay. This proved especially important for the samples from Manaus, as most infections appear to have happened in the early days of the pandemic, around April. People who were infected in April or May, but donated in September or October, had by then experienced a decay in the number of antibodies, which made it difficult for the researchers to identify these cases as positive.

In São Paulo, where the infection curve was more contained, this correction wasn’t so relevant for the calculation, according to the authors, as the antibody waning from old infections was compensated by new infections.

The pandemic in Manaus

In Manaus, the study has been received with skepticism by the scientific and medical communities. The extensive media coverage that the early September version of the study received, filled with claims that the city had reached herd immunity, made it increasingly difficult for health officials to convey the need for keeping restrictions and social distance measures. Political decisions in the region, such as reopening schools, were influenced by the study, according to the magazine Época. So while the new version of the study puts less weight on herd immunity, many still worry that the paper may do more harm than good.

Cristiano Fernandes, the technical director of the Fundação Vigilância e Saúde do Amazonas, a department of the State Health Secretariat that has been at the frontline of the pandemic response, criticizes the paper’s repeated use of the term “unmitigated” to refer to how the pandemic unfolded in the city.

“Saying that Manaus did not have a mitigation plan to restrain the disease is just not true. To make such a claim was inelegant and unnecessary, to say the least,” says Fernandes, who notes that there might have been other factors that contributed to the severity of the pandemic in Manaus.

One of the factors pointed out by Fernandes is the existence of at least eight different strains of the new coronavirus circulating in the state where Manaus is located. This suggests that the virus arrived through multiple paths, which could have amplified the development of the pandemic in the city.

According to several researchers who spoke with The Scientist, the city’s precarious health system may have also played a role in the pandemic response. Despite its size and wealth, Manaus has few intensive care unit beds, which serve not only the city, but the whole state of Amazonas, an area which is approximately 3.5 times bigger than California.

“Manaus absorbed a high load [of patients] from the municipalities in the interior [of the state], where there is no appropriate health infrastructure,” says Felipe Naveca, an epidemiologist at the Fundação Oswaldo Cruz who was critical of the new study. “There are still many questions and few reliable answers.”

A previous study using serological household surveys estimated that in June, 14 percent of the population had been infected—much lower than the 66.2 percent figure found in this new study for that month. While the household survey also has some limitations, the disparity between the studies calls for caution in interpreting the results.

Microbiologist Natalia Pasternak, the director of the Instituto Questão de Ciencia, a private organization that promotes scientific integrity and has been monitoring the pandemic in Brazil since its early days, describes the study in a positive light. However, she warns that the 76 percent figure should be seen as an estimate. With COVID-19 cases continuing to be very high in Manaus, and nearly all intensive care unit beds currently occupied, there is clearly a substantial amount of the population still vulnerable.

“I think we need to stress that herd immunity will only be achieved through vaccines,” says Pasternak.

L.F. Buss et al, “Three-quarters attack rate of SARS-CoV-2 in the Brazilian Amazon during a largely unmitigated epidemic,” Science, doi:10.1126/science.abe9728, 2020.