Another surge was coming. This time, Uildéia Galvão thought they were prepared.

Galvão, the lead physician in the coronavirus ward at a public hospital in the Brazilian city of Manaus, had been haunted by the wave that crashed last spring. In less than 10 days, it ruptured the city’s bewildered medical system. Sick patients were turned away. The dead were piled into mass graves.

So Galvão’s hospital organized contingency plans. Additional beds were reserved, and a detailed schedule for opening them was created.

But the new surge, when it came, was different. The virus had mutated, with a suite of alterations that probably made it more transmissible – and perhaps more lethal. Manaus was hit by what scientists call the P.1 variant. This time, it didn’t take 10 days to overwhelm Galvão’s hospital. It took 24 hours.

Even in a city as traumatized as Manaus, the horror has been unlike anything doctors have seen. The oxygen quickly ran out. Dozens of hospital patients have died of asphyxiation. Scores more, unable to get care, have died at home. Every half-hour, one doctor said, a funeral procession rumbled toward the cemetery.

“We had a plan,” Galvão said. “We increased the availability of beds. But even with that, there was strangulation.”


The humanitarian disaster unfolding in the Amazon’s largest city has shown what happens when government failures, scientific misfires and public indifference meet a new, possibly more dangerous variant of the virus that has ravaged the globe.

Believed to have been circulating in the Amazon since December, P.1 now appears to be the dominant coronavirus strain in Manaus. It’s been detected in São Paulo and as far away as Japan. A first case was identified in the United States on Monday.

More on the COVID-19 pandemic

Scientists are racing to understand the variant, one of several to have emerged in recent months. They are trying to determine whether it truly is more transmissible or has simply exploited lax behavior in a region where many people are either unable or unwilling to take precautions against the virus. The biggest unknown is whether the variant can infect people who have recovered from the more common coronavirus strain.

Doctors and front-line health workers are describing a dangerous new chapter in the struggle against the virus. The shift came suddenly: It wasn’t just the surge in patients but the severity of their cases. People started arriving at hospitals significantly sicker, lungs chewed up with disease.

“What has been said before, that this is a strain more transmissible but not more severe – that’s not what is happening in Manaus,” epidemiologist Noaldo Lucena said. “This isn’t a feeling. It’s a fact.”


The global implications could be significant. Since the beginning of the pandemic, Manaus, a city of 2 million swelling along the Amazon River, has been closely studied by scientists. Local officials shied away from lockdowns or restrictions that have been successful elsewhere. And what policies did exist, many people ignored. The virus, believed to have infected a large portion of the population, was left mostly free to spread naturally.

“Manaus represents a ‘sentinel’ population, giving us a data-based indication of what may happen if SARS-CoV-2 is allowed to spread largely unmitigated,” a team of researches wrote this month in Science.

For a time, after the wave of April and May subsided, scientists and government officials wondered whether the city had achieved herd immunity. Some scientists estimated three-fourths of the population had been infected. Many believed the worst was behind the city.

“Why Manaus will be the first Brazilian city to defeat the Covid-19 pandemic,” wrote a group of researchers from the Federal University of Amazonas.

No one is saying that now.

– – –

In late December, as the holidays were set to begin, Amazonas state Gov. Wilson Lima debated what to do. The daily counts of cases, hospitalizations and deaths had begun to pick up. Scientists were issuing increasingly urgent letters, calling on officials to institute immediate restrictions on businesses and gatherings.

“We need to save lives and not deepen the health an humanitarian disaster,” epidemiologist Jesem Orellana pleaded in one such missive. “Lives matter!”


On Christmas Eve, Lima announced the closure of all nonessential businesses. Protesters swept the city, closing roads and setting fires. Business owners and lawmakers said the economy couldn’t survive a shutdown. A third of the city’s workers are informal – street vendors, delivery men, maids. They pushed the governor to repeal the decree. And within two days, he did.

Retailers and restaurants did brisk holiday business. Massive parties – some numbering more than 4,000 revelers – gushed onto the streets. And supporters of President Jair Bolsonaro, who has made inaction the defining element of his pandemic presidency, rejoiced.

“All power emanates from the people,” tweeted Congressman Eduardo Bolsonaro, the president’s son.

“Regardless of the alarmist newscasts, Manaus has seen a large drop in deaths since June, showing collective (or herd) immunity,” tweeted Osmar Terra, a former Bolsonaro cabinet member.

But that belief – which seems to have seduced many in Manaus into a false sense of security – was quickly proved a fiction. Soon after the holidays, deaths and hospitalizations exploded. The hospital system buckled. The number of confirmed coronavirus deaths at home rose from a total of 35 from May through December to 178 so far this month, according to city health officials.

That stunned Brazilian researchers who last month published a paper in Science proclaiming that 76% of Manaus’s population had already been infected with the virus.


“How can you have 76% of people infected and, at the same time, have an epidemic that’s bigger than the first?” asked author Ester Sabino. “This was a concern from the moment cases started to rise.”

To understand what was happening – and why the city wasn’t protected from a debilitating second wave – the team started sequencing fresh samples, to see if any changes in the virus could explain it.

On Jan. 10, Japan announced the discovery of a new variant, found to have infected four travelers from Brazil’s Amazon region. Then Sabino’s team published preliminary findings showing that the strain accounted for 42% of the coronavirus cases sampled in December.

As viruses course through a population, they inevitably mutate, although most such genetic changes are functionally insignificant. The coronavirus has spawned countless variants around the world. But P.1 – along with variants found in South Africa and Britain – is provoking particular concern.

Not only does it have a spike protein mutation that could lead to a higher infection rate, it possesses what’s called an “escape mutation.” Also found in the South Africa variant, the mutation, known as E484k, could help it evade coronavirus antibodies.

Sylvain Aldighieri, a senior official with the Pan American Health Organization who has been tracking the Manaus outbreak, said there is no evidence to suggest that reinfections are driving the health crisis. “We would have many more reports,” he said. “We have to use our common sense at this point. Herd immunity in Manaus was not achieved.”


Other scientists have expressed doubt that 76% of people in Manaus were infected.

Doctors said they haven’t seen many reinfections but cautioned that it’s nearly impossible to know. The city was swept by the disease at a time when shortages in supplies meant few could get tested. That early failure has seeded today’s: Without previous testing, it’s impossible to confirm a reinfection.

One case, however, has been confirmed by scientists. Dozens more are under analysis.

Mariana Leite, 31, an engineer in Manaus, said she tested positive for antibodies in June and felt a “sense of relief.” She didn’t think it would be possible to be reinfected, but she said she was. Her polymerase chain reaction test came back positive Jan. 8.

“It’s caused so much anxiety in everyone,” she said. “We feel like it’s never going to end.”

Meanwhile, the P.1 variant appears to have widened its reach: In January, according to a sample of 48 cases, it represented 85% of the infections.


– – –

The toll has been clear. By mid-January, the hospital system hadn’t just run out of beds, as it did during the first wave, but also oxygen. Wards had been transformed, in the words of one epidemiologist, into “chambers of asphyxiation.” Hundreds of patients were shipped out of the city, some to the other side of the country.

The federal government was warned of the looming disaster, according to an investigation requested by the supreme court, but didn’t do enough to avert it.

On Jan. 3, local health officials told federal officials the health system would probably fail within 10 days. Then the company White Martins, which supplies the public health system in Manaus with oxygen, warned state and federal health officials it couldn’t keep up with demand. On Jan. 14 and 15, dozens of people suffocated to death.

The Health Ministry has defended its response, arguing it sent doctors and helped the city open more beds for patients. Federal Health Minister Eduardo Pazuello, now under investigation by the attorney general over allegations of inaction, has stationed himself in the city indefinitely.

“We had a jump in contaminations since the beginning of January, tripling the number of people infected,” Pazuello said Tuesday. “This was a completely unknown situation for the whole world; it was very fast.”

Lima, the state governor, has now closed all nonessential businesses and imposed a curfew. He has warned that the weeks ahead could be still more difficult. Demand for oxygen will be 70% higher in February than in the worst days of this month, he told the newspaper O Globo.


It’s difficult for Lucena, the epidemiologist, to imagine how it could get much worse.

“You are watching deaths every day, deaths that could have been avoided,” he said. “You choose who lives and who dies, who gets oxygen and who doesn’t.”

“It’s like we’re in a horror film.”

– – –

Data on cases is from the Amazonas Health Surveillance Foundation. The foundation did not report data on Nov. 6 and 7, 2020.