Second wave of COVID-19 in Manaus rekindles debate on herd immunity | AGÊNCIA FAPESP

Second wave of COVID-19 in Manaus rekindles debate on herd immunity Experts tell Agência FAPESP about factors that may be associated with the rapid rise in daily case numbers in the capital of Amazonas State, where a six-month state of emergency has been declared (photo: Mario Oliveira / SEMCOM / Manaus)

Second wave of COVID-19 in Manaus rekindles debate on herd immunity

January 20, 2021

By Karina Toledo | Agência FAPESP –  In August 2020, when new daily cases of COVID-19 had been falling sharply for three months in Manaus, the capital of Amazonas State in the North of Brazil, even with schools and retail stores reopened, some Brazilian experts raised the hypothesis that the population had acquired collective (“herd”) immunity to SARS-CoV-2, even though serological data showed seroprevalence for the North region to be below 30%.

The herd immunity hypothesis appeared to be validated in September when researchers at the University of São Paulo (USP) and collaborators published a draft paper reporting 66% seroprevalence in Manaus on the basis of blood donor samples and mathematical modeling (read more at:

This level of prevalence was close to the threshold calculated at the start of the pandemic using the classical epidemiological formula. The study was led by Professor Ester Sabino. In December, when the final version of the paper was published in Science, the estimate was that 76% of the population of Manaus had immunity against the virus. If so, how can we explain the second wave that has overwhelmed the city’s hospitals, forcing the mayor to declare a six-month state of emergency on January 5?

In Sabino’s opinion, because the virus continues to circulate throughout Brazil, daily case numbers rebounded when people resumed normal activities and will continue to rise until about 95% of the population have been infected. “The concept of herd immunity is sometimes misunderstood,” she said. “When the threshold is reached, it doesn’t mean the disease will disappear, but merely that daily case numbers won’t rise as fast as they did in the first wave. They’re unlikely to shoot up as steeply as in April [2020] – unless people have since lost their immunity and reinfection has become much more frequent than we assumed it would be.”

The main problem in Manaus, according to Sabino, is that there are too few hospitals and they quickly run out of intensive care beds. “The situation is alarming indeed,” she said. “They’ll have to triple the number of intensive care beds to avoid a disaster. Right now a person with appendicitis risks dying for lack of hospital capacity.”

For infectious disease specialist Júlio Croda, 99% of the new cases notified in Manaus involve people who have never had the disease before. In other words, they are not reinfections.

“In this second wave most patients are from comparatively high-income households and were able to remain isolated in the first wave. This is evidenced by the fact that private hospitals have filled up before public ones, in contrast with what happened in April 2020. After mobility restrictions were lifted, the virus resumed circulation more intensely and infected the most susceptible part of the population,” said Croda, a researcher at Oswaldo Cruz Foundation (Fiocruz, attached to the Ministry of Health) and a professor at the Federal University of Mato Grosso do Sul (UFMS).

The 76% seroprevalence in the Science paper is an overestimate, he added, reckoning more than 50% of Manaus’s inhabitants have not yet developed immunity to SARS-CoV-2.

“It’s important to note that the herd immunity threshold isn’t a fixed percentage,” Croda explained. “It’s calculated on the basis of the effective reproduction rate [Rt, initially estimated at 2.5-3, meaning each infected person transmits SARS-CoV-2 to two or three people on average], which depends both on the genetics of the virus and on the measures taken to control its transmission. The more transmissible variant recently identified in the UK affects the calculation of Rt and the herd immunity threshold.”

Precautions taken by ordinary citizens independently of the actions of local authorities, such as mask-wearing, hand hygiene and social distancing, help lower the reproduction rate, and the herd immunity threshold is lowered as a result. “This is probably what caused the fall in the number of cases seen in mid-2020,” he said. “But when part of the population and the authorities relaxed the control measures, the threshold returned to levels close to 70%.

“Relying too much on this kind of indicator is a mistake,” said Paulo Lotufo, an epidemiologist and professor of medicine at the University of São Paulo’s Medical School (FM-USP). “The main problem is that to calculate the herd immunity threshold you need to know Rt, which is an estimate – more of a ‘guesstimate’, actually. Nevertheless, it has a very significant impact. No disease transmitted from human to human has ever been controlled by means of herd immunity. Measles and smallpox, for example, were only controlled by vaccines. In the case of COVID-19, talk of herd immunity or early treatment just hinders efforts to control the disease by underplaying the importance of social distancing to the general public.”

In addition, Lotufo said, it is difficult to obtain precise indicators for Manaus and accurately analyze the factors that have driven the second wave of COVID-19 there. Besides the hypotheses already mentioned, such as reinfection or the emergence of a more contagious variant, it is possible that some of the patients hospitalized in Manaus are from the interior of Amazonas State where there are few hospitals and intensive care beds are virtually non-existent.

Projections versus actuality

Among the experts who concluded that Manaus must have reached the herd immunity threshold in mid-2020 was Gabriela Gomes, a Portuguese-born biomathematician currently at the University of Strathclyde in Scotland (UK). Gomes developed a model that took into account the different degrees of susceptibility and exposure to the virus in the population concerned (read more at:

Marcelo Urbano Ferreira and Rodrigo Corder, researchers at the University of São Paulo’s Biomedical Science Institute (ICB-USP), are also members of Gomes’s group.

“We compared the hospitalization rate for COVID-19 patients in Manaus projected in October using our heterogeneous model with the rate projected using the homogeneous model by researchers at Imperial College London [in the UK], and found that the actual rate is currently about halfway between the two projections,” Ferreira told Agência FAPESP. “The curve isn’t as high as ICL predicted or as low as we predicted. I believe this is due to the fact that all models have limitations.”

Some of the key variables used in the calculations are imprecise because the disease is novel and scientists have yet to gain a full understanding of its dynamics. “We’ve been working with data for antibodies to estimate the number of cases and that’s problematical regardless of the adjustments made,” Ferreira said. “We now know the quantity of antibodies tends to decrease over time. In addition, even after recovering from an infection, people risk reinfection. We didn’t know that at the start of the pandemic. Another limitation relates to the way we inserted parameters representing relaxation of control measures into the model. For example, São Paulo moved from what’s called the ‘red’ to the ‘yellow zone’ between Christmas and New Year. We also don’t know exactly how many people resumed normal activity or continued to comply with social distancing rules. All this makes modeling difficult, not to mention the emergence of novel variants of the virus that are even more transmissible.”

Gomes and her group are currently studying how mobility restrictions affect the risk coefficient, given the model’s assumption of variations in susceptibility and exposure to the virus due to genetics and immunity as well as lifestyle factors. “When everyone stays at home, however, this difference between individuals is far smaller,” Ferreira said.

Turning point

The findings of the fourth and most recent phase of EPICOVID, the serological survey conducted in 133 Brazilian cities by researchers affiliated with the Federal University of Pelotas (UFPel), were published in September 2020, indicating that the epidemic was decelerating in most of Brazil. According to epidemiologist Pedro Hallal, principal investigator for the study, the downtrend ended in November when the municipal elections were held. The magnitude of the impact will be measured by the fifth phase of EPICOVID, scheduled to take place in January.

“The fortnight preceding the runoffs for mayor was a decisive period,” Hallal said. “We then had Christmas and New Year, followed by summer vacations for many. Daily case numbers rose consistently until at least mid-December, according to the official statistics, but we’re not yet back to the level seen when the moving average was over 1,000 deaths per day. It’s hard to say if the uptrend will persist. The history of the pandemic is a day-by-day construct. Everything depends on how the general public behave.”

Hallal lamented the lack of an effective national policy to combat the disease. “We know what to do to make the curve fall rapidly: mass testing and contract tracing,” he said. “That’s never been done in Brazil. When people are diagnosed as having the disease, no one goes looking for their contacts, but it should be compulsory to do so.  Other countries implement lockdowns in response to a rapid rise in the contagion curve. The virus circulates less, and the economy can reopen. In Brazil we have halfhearted isolation, which is bad for public health and ensures that the economy can’t really reopen.”

For Hallal, the most important policy to be planned in 2021 is vaccination, which he considers the only way to achieve herd immunity without a tragically high mortality rate. Lotufo agreed and expressed optimism. “The Butantan Institute vaccine [developed by China’s Sinovac] and the AstraZeneca vaccine [developed by the University of Oxford in the UK] are effective and easy to handle. Both will be produced in Brazil. If all goes well, we’ll be able to vaccinate the population quickly,” he said.

Croda reckoned there would not be enough doses initially for herd immunity to be reached nationwide, which would require vaccination of 80% of the population. “However, according to the World Health Organization, it’s possible to reduce hospitalizations and deaths significantly by vaccinating the 20% most at risk. This alone would be a major advance,” he said.



Agência FAPESP licenses news reports under Creative Commons license CC-BY-NC-ND so that they can be republished free of charge and in a straightforward manner by other digital media or by print media. The name of the author or reporter (when applied) must be cited, as must the source (Agência FAPESP). Using the button HTML below ensures compliance with the rules described in Agência FAPESP’s Digital Content Republication Policy.

Topics most popular