How herd immunity mitigated a deadly second wave of COVID-19 in Manaus

Abstract

In Manaus, Brazil, the COVID-19 pandemic struck hard with two significant waves of infections and deaths. The first wave began in March 2020, escalating into a severe healthcare crisis by April, as hospitals became overwhelmed and mass graves had to be dug. By October 2020, a study suggested that around 76% of Manaus's population had been exposed to the virus, leading to the expectation that the community had developed some level of immunity. However, despite this high level of infection, a new variant of the virus called Gamma P.1 appeared. It triggered a second wave in November 2020, which was even more deadly than the first. This raised important questions: Why was the second wave so much deadlier, even though many people had already been infected? And what role did immunity from the first wave play during the second?

To address these questions, we used mathematical models to study how COVID-19 spread in Manaus during both waves.Our model tracks both first infections and reinfections. It distinguishes between a longer-lasting immunity and an immunity that can weaken over time. The model also estimates how well people were protected from getting reinfected or dying if they were reinfected. We used real data from Manaus, including population demographics, hospital deaths, and antibody tests. The data was divided into different age groups for our model, which also takes into account how often contacts occur between these age groups.

Our model matched the observed data, helping us estimate how many people were infected during the first wave and in turn how many were reinfected during the second wave. Since younger people had more frequent contact with others, most of them were infected during the first wave, while many older individuals remained uninfected. The longer-lasting immunity gained during the first wave offered strong protection against death in case of reinfection, but did not reduce the chance to get reinfected enough to prevent it. When the deadlier Gamma P.1 variant emerged, reinfections in younger people played a key role in spreading the virus during the second wave, allowing it to reach the remaining individuals who hadn’t been infected in the first wave.

To better understand the impact of immunity gained during the first wave, we ran “what if” scenarios. We found that if Manaus had not experienced such a widespread first wave, the total number of deaths at the end of the second wave could have been much higher. In fact, it could have possibly doubled. The strong longer-lasting protection against death was crucial in reducing fatalities during the second wave. By limiting the number of people that had no protection against the deadlier Gamma P.1 variant, the prior infection helped lessen the overall impact of the second wave.

These results highlight how different factors, such as age-related differences in contact behavior, immunity from previous infections, and the emergence of new virus strains, can influence the severity of later waves. It underscores the importance of considering these dynamics when planning public health responses to future outbreaks of infectious diseases. While immunity from past outbreaks can reduce the severity of new ones, it doesn’t necessarily prevent them, and it is in fact possible that suppression of a first wave can result in an overall increase in severity if deadlier variants emerge.