Population surveys seeking to find out how widespread asymptomatic Covid-19 cases really are show that a herd immunity strategy might struggle, Marc Daalder reports
Conventional wisdom about Covid-19 – insofar as it exists – holds that it is a dangerous illness which can hospitalise one in every four or five cases, force around 5 percent of patients onto a ventilator and kill between 0.1 and 3.4 percent of people who catch it.
But what if it wasn’t that bad – what if a surprising percentage of the population already caught it, had mild or no symptoms, and then recovered? What if the bulk of some countries were already immune to the virus?
That’s the argument that opponents of global economic and social lockdowns have seized upon: The virus isn’t that bad and herd immunity – through exposing most people to Covid-19 – will get us through it. New Zealand academics who wanted to see the country move straight from Level 4 to Level 2 made this argument, pointing to the town of Gangelt in Germany where 80 percent of the population was tested and a preliminary survey found 14 percent had antibodies – indicating they had contracted the virus and then recovered.
The Gangelt study has come under fierce criticism, however, and other population surveys – where a random but significant number of people have been tested, in order to also seek out asymptomatic cases – have presented conflicting results. At the same time, experts say, none of these surveys can be directly applied to New Zealand.
“There’s a lot of conflicting estimates. Obviously, because it’s early on,” Janine Paynter, a research fellow in population health at the University of Auckland, told Newsroom.
Gangelt survey controversial
The Gangelt survey made headlines for its robust claims. If Covid-19 was as widespread as the study claimed, then the death rate would be much lower than previously thought – as low as 0.37 percent. This bolstered the anti-lockdown argument, which holds that the virus is not so deadly that we should be shutting down the bulk of economic activity.
Herd immunity advocates, who believe that allowing the virus to sweep through the population of those under 70 and with no vulnerable conditions will be faster and safer than waiting for a vaccine, have also seized on the Gangelt study. They say that official testing had indicated just 1 percent of the town’s population had Covid-19, so the disease could be as much as 14 times more prevalent in communities than previously thought. But expert commentary from the United Kingdom’s Science Media Centre criticised jumping to such conclusions.
“The finding that 14 percent of the population of Gangelt in Germany have antibodies to the virus causing Covid-19 is interesting, but does not in any way prove that 14 percent of the population are immune, as these reports suggest,” University of Reading associate professor of cellular microbiology Simon Clarke said.
“While it’s undeniably a good sign, merely having some antibodies does not necessarily prove immunity, as is often casually asserted. We don’t know what the correlates of protection, the measurable signs that someone has immunity to a disease, are for this virus and neither do we know how long any immunity would last.”
Professor David Heymann, an expert in infectious diseases at the London School of Hygiene & Tropical Medicine, agreed with Clarke. “The immune response to SARSCoV2 [the coronavirus that causes Covid-19] is not yet understood – it is known for some of the other coronaviruses that infect humans that reinfection with the same virus is possible,” he said.
“Although the authors of this work have made reference to possible herd immunity, it is far too premature to make these kinds of statements – it is necessary to understand if and how long antibody to SARSCoV2 signals protection first.”
The Diamond Princess cruise ship, where everyone on board was tested, can also serve as a population survey, although the population is heavily weighted towards older people and the dynamics of physical interaction and population density on a cruise ship are different from most other environments.
“The Diamond Princess, which was sort of early on, gave us an estimate of asymptomatic transmission of 75 percent, which was very high. We just have to mark that one off as a very unusual situation, too,” Paynter said.
Iceland and Italy show different results
Other surveys have also found far lower rates of infection, even when looking for asymptomatic cases.
The town of Vo, which saw Italy’s first Covid-19 death, was tested when Italy’s lockdown began and then once again two weeks later. In the first round of testing, 85.9 percent of the population was tested and 2.6 percent of the population at the time tested positive. The second round of testing, which covered 71.5 percent of the population, found just 1.2 percent of the population were positive for Covid-19. Notably, across the two testing rounds, 43 percent of positive cases were asymptomatic.
Vo seems to indicate that many cases are asymptomatic but doesn’t indicate that the virus is as widespread as the Ganglet survey shows. Similarly, a study of Iceland’s population where 6 percent of people were tested found low rates of transmission. Of “high-risk” tests – those of symptomatic patients, people who had been in contact with an infected person or people who had travelled to an overseas hot spot recently – 13.3 percent were positive. But when researchers allowed anyone who wanted a test to get one, just 0.8 percent tested positive. When researchers then began randomly testing the population – removing self-selection bias – just 0.6 percent of the sample tested positive.
Asymptomatic transmission in Iceland was about as common as in Vo – among high-risk cases, just 7 percent were asymptomatic, while 43 percent of people who tested positive in the two population screening scenarios reported no symptoms.
“The Iceland and Italian studies, which are based on fairly large sampling of the populations, puts asymptomatic transmission much lower. But they’re probably more reliable measures,” Paynter said.
However, Vo and Iceland differ from Gangelt in that they used the commonplace PCR tests that identify active Covid-19 cases and would not have found recovered cases.
“The weaknesses of the work in Iceland and Italy were that they were the PCR sampling. So if someone was exposed, say, three or four weeks ago, potentially the PCR test would be negative and yet we know that that person had had that infection,” Paynter said.
Other antibody tests
There are other studies that use the antibody tests popularised by the Gangelt survey. Surveys of Santa Clara County and Los Angeles in California have been similarly controversial. The studies concluded that between 2.5 and 4.2 percent of Santa Clara had been infected, as had between 2.8 and 5.6 percent of Los Angeles.
However, both papers have been subjected to intense criticism from other statisticians for projecting the results of a small sample onto much larger populations.
The conclusions are “some numbers that were essentially the product of a statistical error,” Andrew Gelman, a professor of statistics and director of the Applied Statistics Center at Columbia University, told Mercury News.
“They’re the kind of screw-ups that happen if you want to leap out with an exciting finding and you don’t look too carefully at what you might have done wrong.”
Among the criticisms were worries that the study authors might have been biased. Two of the authors of the Santa Clara study, run out of Stanford University, had previously written an op-ed for the Wall Street Journal saying that lockdowns were an overreaction.
The antibody tests themselves came into question as well. Both studies used a test that reported around two false positives for every 371 true negatives. For the Santa Clara sample, which reported 50 positives out of 3330 tests, as many as 18 positive tests could be false.
However, a new antibody study from New York City found 21 percent of a sample group of 3000 there tested positive. While this is significant, it is also far from enough for herd immunity in a city that has already seen 15,000 deaths and 141,000 confirmed cases.
“That’s still not high enough exposure. That’s nowhere near herd immunity levels. And that’s in New York, where we know that it’s been rampaging through,” Paynter said.
So what?
Why does all of this matter? If, say, 50 percent of the population had already contracted the virus, then the worst could soon be over. Through immunity, the chains of transmission would be cut without even needing a lockdown – assuming immunity works the way it does with the flu, for example.
Likewise, if 50 percent of the population had contracted the virus, then the severity would be much lower than if just 5 percent had, because the death rate would be 10 times lower.
However, experts aren’t confident that the population surveys indicate this.
“My first go-to for herd immunity is, until we’ve got a vaccine, we won’t achieve herd immunity. That’s the key message that we need to understand. Because it’s such a new virus, we don’t understand – at the moment – its infectivity rate or the duration of recovery,” Daniel Exeter, an associate professor in health geography at the University of Auckland, told Newsroom.
“There’s been many people, internationally as well, who have said it’s just far too premature to be able to make those statements around herd immunity or reinfection because it is a new and emerging disease.
“The problem is we don’t know whether [asymptomatic cases] are protected. They may come across it again, down the track – say six months later, one year later – and they could get sick,” Paynter said. She pointed to a study in which 15 people were infected with the coronavirus that causes the common cold. Five people didn’t develop symptoms and one of them became symptomatic when reinfected a year later, indicating that lack of symptomatic response could mean that immunity doesn’t develop either.
Newsroom recently examined the significance of reported Covid-19 reinfections.
“We just don’t know. This is a new coronavirus. We don’t know how long the immunity lasts. It could only be two months that you’re protected, or it could be two to three years, or even longer. We just don’t know,” Paynter said.
On Friday, the World Health Organisation released a statement along these lines. “There is currently no evidence that people who have recovered from COVID-19 and have antibodies are protected from a second infection,” the statement read.
“As of 24 April 2020, no study has evaluated whether the presence of antibodies to SARS-CoV-2 confers immunity to subsequent infection by this virus in humans.”
Exeter also stressed that overseas population studies cannot be directly imposed on New Zealand. Just because 14 percent of Gangelt might have contracted Covid-19 doesn’t mean the same applies to all of New Zealand.
“I wouldn’t be taking a value from one of those studies and applying it to New Zealand, simply just as a one-off study. If we had more time and more data, maybe if there was a meta-analysis of all of the studies, then we could perhaps with a little more confidence respond in that way, but I wouldn’t do that for now,” he said.
However, the current outcomes from surveys indicate that, even if immunity was long-lasting and effective, the world is still a long way off from effectively implementing a herd immunity strategy, according to the WHO.
“Early data suggests that a relatively small percentage of the populations may have been infected. Not more than 2 percent to 3 percent,” WHO director-general Tedros Adhanom Ghebreyesus told a media briefing in Geneva.
Maria Van Kerkhove, an American infectious diseases expert and the WHO’s technical lead on Covid-19, said that while it was too early to make definitive statements, results this far have shown lower population spread than expected.
“Initially, we see a lower proportion of people with antibodies than we were expecting,” she said at the briefing. “A lower number of people are infected.”
Paynter agreed, saying that the survey results she had reviewed were not promising for herd immunity advocates.
“This puts the hammer on any herd immunity proposals.”
