Coronavirus UK: Antibody data shows 4.2m may have been infected

At least 4.2million people in the UK have been infected with the killer coronavirus, according to one of the best estimates of the disease’s death rate.

And London, which has been at the heart of Britain’s epidemic, is home to at least 935,000 people who have had the disease already.

The estimate is based on numbers from a public health surveillance scheme in New York, which found that a quarter of the city’s population of 8million people had been exposed to the virus – meaning that the approximately 16,000 deaths recorded there represent a death rate of 0.79 per cent.

Other estimates from cities in Europe and a study by the University of California, Berkeley, put the virus’s death rate at between 0.19 and 0.5 per cent. In comparison, the death rate of flu is thought to be around the 0.1 per cent mark.

Extrapolating from these figures suggest that between 4.2 and 33.5million Britons have been infected with COVID-19, the vast majority of them untested, and between 935,000 and 3.8million people in London may have been ill already.

Officials have no idea how big the outbreak truly is because of a lack of widespread testing for people outside of NHS hospitals.

As of yesterday, 161,145 people had officially tested positive in the UK but the vast majority of these have been NHS hospital patients or staff. Countless members of the public are believed to have been infected and recovered without medical help.

Early surveys in major Western cities are now giving insights into the true death rate of the infection – that is, the proportion of people who die out of everyone who catches it, not just those who end up in hospital.

Applying these figures to the numbers of people known to have died in London and the UK as a whole can give a rough picture of how many people in the country might have been infected.

Office for National Statistics data today suggested that the true death rate of the UK’s coronavirus outbreak will ultimately be 55 per cent higher than the Department of Health’s daily updates show, meaning 33,596 people may already have died across the country along with 7,391 people in the capital city.

Professor Chris Whitty, the country’s chief medical officer, last week said he thought more than 10 per cent of people in London had caught the virus already – some 900,000 people – but that exposure was likely lower in other regions.

Death rate models from the UK Government, New York and cities in Europe suggest that anywhere between 900,000 and 7.3million people in London may have caught the coronavirus already

Preliminary death rates emerging in studies around the world suggest that anywhere between 4.2million and 33million people in the UK may have been infected with the coronavirus and recovered already

Governor for New York City, Andrew Cuomo, said this week that surveys showed a quarter (24.7 per cent) of the city’s population had recovered from COVID-19.

Officials found this out as a result of their antibody testing programme which can pick up people whose immune systems have fought off the virus.

At least 7,500 randomly-selected people have been tested in five boroughs in the city, which has the worst outbreak in the world.

The test results suggest some 2.1million people have been infected. 16,673 deaths had been recorded in NYC when Governor Cuomo made the announcement yesterday, putting the death rate at 0.79 per cent.

If this death rate persists, and turns out to be true for the UK, the number of people who have died already could mean millions have been infected.

A total death toll of 33,596 – a 54.6 per cent increase on the Department of Health’s tally (21,731) which ONS data suggests is an accurate assumption – would mean 4,252,658 people have been infected across the UK already.

The same condition applied to London would equal 935,569 cases there.

This lines up with Professor Chris Whitty’s estimate of how many people have been infected in the city.

He said in a Parliamentary committee meeting on Friday: ‘My view at the moment – I would love to be proved wrong – is that it is unlikely that any part of the UK, maybe outside London, will have a [prevalence] much above 10 per cent.

‘And I would expect, for example in the south west, for it to be lower… than for example London. I think it’s quite a small proportion of the population.’

Similar studies to the one going on in New York, carried out in Helsinki, Finland; Stockholm, Sweden; and Gangelt, Germany, provide alternative estimates.

Those cities found the death rate to be lower, at 0.19 per cent (Helsinki), 0.37 per cent (Gangelt) or 0.4 per cent (Stockholm).

A study by the University of California, Berkeley, looked at the outbreak in Italy and estimated the death rate could be 0.5 per cent.

And at the start of the outbreak, government scientists also suggested that the death rate could be as low as 0.1 per cent, which would indicate half the British population has been infected already.

Applying the various estimated death rates of the virus, the true number of people infected with the coronavirus could be one of the following:

0.1% death rate (Early Government estimate) – 33,696,000 cases in the UK
0.19% death rate (Helsinki, Finland) – 17,682,105 cases in the UK
0.37% death rate (Gangelt, Germany) – 9,072,702 cases in the UK
0.4% death rate (Stockholm, Sweden) – 8,399,000 cases in the UK
0.5% death rate (University of California, Berkeley) – 6,719,200 cases in the UK
0.79% death rate (New York state) – 4,252,658 cases in the UK

The death figures used for the estimates are based on Office for National Statistics (ONS) data released today which suggests the true number of COVID-19 victims is 54.6 per cent higher than the Department of Health’s figures show.

ONS data shows that, when non-hospital fatalities were included, there had been 22,351 coronavirus deaths in England and Wales by April 17 – a significant rise on the 14,451 counted by health chiefs.

If the same increase – 54.6 per cent – were applied to the total UK death toll confirmed today (21,731) it could mean the real number of victims was 33,596 – the number used in MailOnline’s calculations.

The same rule was applied to London, where the NHS has recorded 4,781 victims, increasing the city’s current death toll to an estimated 7,391.

Applying the death rates to London produces the following estimates:

0.1% death rate (Early Government estimate) – 7,391,000 cases in London
0.19% death rate (Helsinki, Finland) – 3,890,000 cases in London
0.37% death rate (Gangelt, Germany) – 1,997,567 cases in London
0.4% death rate (Stockholm, Sweden) – 1,847,750 cases in London
0.5% death rate (University of California, Berkeley) – 1,478,200 cases in London
0.79% death rate (New York City) – 935,569 cases in London

Understanding the true numbers of people who have had the virus and not become seriously ill is the only way scientists will be able to work out the virus’s true death rate.

While some countries in Europe are recording death rates of more than 10 per cent among hospital patients, in wider society the likelihood of death appears consistently below one in 100 patients (one per cent).

Early antibody survey results show that more than 30 per cent of people have been infected in Chelsea, Massachusetts, 14 per cent in the German district of Heinsberg, around three per cent in Oise, northern France, and 11 per cent in Stockholm.

Early results from antibody surveys reveal wild variations in infection levels in communities across the US and Europe. Most are based on small samples in localised areas, but wider data is expected in the coming month

WHAT CAN ANTIBODY TESTS SHOW US ABOUT THE VIRUS’S TRUE DEATH RATE?

The true death rate of COVID-19 is unknown because of the way the virus is being tracked around the world.

Most countries are only diagnosing people in hospitals or those with bad symptoms, meaning the number of patients they record is significantly lower than the reality.

Many patients, scientists say, develop only mild illnesses or don’t get any symptoms at all, meaning they are never tested and never counted.

As a result, the ratio of people dying to people diagnosed is artificially high.

Using antibody tests to understand the true – much larger – numbers of people who have caught the virus, and then working out what proportion of them have died, will provide more accurate death rates in the future.

Early antibody surveys have already started to shine light in this area:

LOS ANGELES, CALIFORNIA

Blood samples in Los Angeles suggest the coronavirus death rate could be around 0.18 per cent.

A study of 846 people found roughly 4.1 per cent of the county’s 3.9million population has antibodies to the virus.

It means that roughly 330,000 people have already caught the illness and built up some immunity to it.

There were officially 600 COVID-19 deaths when the research was conducted on April 20.

This suggests that around 0.18 per cent of patients fall victim to the disease.

CHELSEA, MASSACHUSETTS

US researchers in Boston found almost a third (31.5 per cent) of residents in the suburb of Chelsea had antibodies for the virus.

They collected blood samples from 200 random volunteers and said roughly 63 people had probably caught the illness.

The city of Chelsea is home to around 40,160 people. If the results were to be extrapolated to the whole city, it suggests 12,650 may have actually been infected.

When the study was published on April 17, Chelsea had suffered 39 deaths to coronavirus.

The finding suggests the true death rate it around 0.31 per cent.

GANGELT, GERMANY

Scientists studying Gangelt, dubbed the ‘German Wuhan’, found as many as 15 per cent of people may have already been infected with the virus.

Data shows around 12,500 people live in the municipality, which sits in the North-Western state of North Rhine-Westphalia.

If the results were to be extrapolated to the whole of Gangelt, it would mean that around 1,900 people have already caught the deadly virus.

It is not clear exactly how many people had died in Gangelt by the time that the University of Bonn study of 1,000 people was published.

But the team – whose work was not scrutinised and published in a journal – estimated the true death rate was in the region of 0.37 per cent.

THE NETHERLANDS

An antibody surveillance scheme in the Netherlands suggested the death rate for COVID-19 could actually be in the region of 0.63 per cent.

Dutch researchers found antibodies in three per cent of blood donors, after analysing samples from around 7,000 people aged between 18 and 69.

The head of the Netherlands’ National Institute for Health told MPs it meant that ‘several hundred thousand people’ may have already been infected.

Around 17.28million people live in the Netherlands. Three per cent of the country’s population would equate to approximately 518,400.

When results were published on April 16, official figures showed that 3,315 people had died after testing positive for COVID-19 in the Netherlands.

HELSINKI, FINLAND

Finnish researchers analysed around 150 blood samples by mid-April and found 3.4 per cent had antibodies for the coronavirus.

The samples were all taken from the region of Uusima, which is home to approximately 1.7million people – most of whom live in the capital of Helsinki.

At the time, only 2,000 cases had been confirmed by laboratory tests. But 3.4 per of the region’s population would equate to around 57,800.

Only 110 deaths have been registered in Uusima to-date – suggesting that the true fatality rate is closer to the 0.19 per cent mark.

The study was released on April 15 – but the region’s death toll has barely changed in the past week. It was not published in a journal.

The World Health Organization said the antibody surveys it has seen suggest the world is nowhere near developing herd immunity against the coronavirus.

Dr Tedros Adhanom Ghebreyesus, director-general of the WHO, predicted that between two and three per cent of the global population has had COVID-19.

If true, this suggests 190million people worldwide have caught the virus in the four months since it emerged.

For herd immunity to develop, a community’s rate is likely to need to be higher than 66 per cent, the UK’s chief scientist, Sir Patrick Vallance, has suggested.

Herd immunity protects people from a disease by virtue of so many people being immune to it that it cannot spread through the population.

Of early surveys to work out the levels of immunity in the population, the city of Chelsea in Massachusetts, USA, has shown the largest scale of infection.

A total 31.5 per cent of people there tested positive in a small sample of 200 random passers-by, carried out by scientists at the Massachusetts General Hospital.

Meanwhile, rapid testing in New York city and state revealed that some 21 per cent of people in the city had antibodies against the virus, and 14 per cent statewide.

Antibodies are substances which people’s immune systems develop to remember how to fight off the virus the next time they come into contact with it. They are only producible through real-world infection or a vaccine – and there is no vaccine yet for the virus.

The figures from New York suggest millions of people among its population of 19million have caught the virus and recovered without being recorded.

The district of Heinsberg in Germany, and a town inside it – Gangelt – showed similar rates of immunity in blood bank screening in March.

Fourteen per cent of people there tested positive for antibodies in a sample of around 1,500 people.

In Sweden’s capital, Stockholm, where the government has staunchly refused to put the country into lockdown, screening in blood donation centres showed that at least 11 per cent of residents have been exposed to the virus.

Swedish officials appear to be pursuing plans to develop herd immunity – which were met with outrage when suggested in Britain – and say that they expect between a quarter and a third of the nation will already have been infected by the end of this month.

One academic at Stockholm University, Tom Britton, predicts that half of the people in the capital city will have caught the virus by May 1.

Despite business there continuing as usual with only advice to people to try social distancing, the country has recorded few deaths, with just 2,000 recorded victims.

Exposure rates are considerably lower in other areas.

In Oise, a small region in the north of France, blood bank screening found that three per cent of the population had evidence of past COVID-19 infection.

The same study found that, in a small school community there, this rate soared to 26 per cent, showing there is scope for wide variations even within regions.

Tests in California showed similarly low levels of infection. Testing from Santa Clara County pointed to approximately 2.8 per cent of people having been infected, while Los Angeles County testing put the figure at 4.1 per cent.

These were pulled from studies done on 3,300 people and 846 people, respectively.

Finland’s capital region, Uusimaa, which is centred around Helsinki, showed around 3.4 per cent of people had evidence of past infection in a study of 442.

A World Health Organization scientist, Dr Maria Van Kerkhove, said the number of people testing positive for antibodies was lower than experts had hoped for.

Dr Kerkhove said: ‘Initially, we see a lower proportion of people with antibodies than we were expecting. A lower number of people are infected,’ The Guardian reported.

Dr Joe Grove, a virologist at University College London, told MailOnline: ‘Antibody testing is important because the better we understand the virus, the better we can respond to it.

‘The true death rate allows public health experts and epidemiologist to asses what the effects of another epidemic would be.

‘A lot of our current policy has been determined by the predictions of computer simulations. But those models are only as good as the data you put into them.

‘So there would’ve been estimates of death rates and infections, but as we get firmer numbers we can run more accurate simulations and predict with more confidence what might happen in future.

‘This is critical for working out if given epidemic will overwhelm the healthcare system again.’

Current antibody surveys are limited by the quality of the tests being used, the small numbers of people recruited in the studies and a poor understanding of immunity.

Because the virus has never been seen before, health authorities around the world are still scrambling to find reliable test to spot the antibodies created to fight it.

In some cases, patients don’t seem to produce detectable levels of antibodies at all, which has led scientists to fear that people do not become immune to COVID-19 after recovering from it. If that was the case, the virus could become unstoppable.

WHY EARLY ANTIBODY SURVEYS SHOULD BE TAKEN WITH A PINCH OF SALT

Current antibody surveys are limited by the quality of the tests being used, the small numbers of people recruited in the studies and a poor understanding of immunity.

Because the virus has never been seen before, health authorities around the world are still scrambling to find reliable test to spot the antibodies created to fight it.

In some cases, patients don’t seem to produce detectable levels of antibodies at all, which has led scientists to fear that people do not become immune to COVID-19 after recovering from it, as they would with measles or strains of cold and flu viruses.

The tests that are being used vary from country to country.

Some being used in Spain have an accuracy of just 80 per cent, while some made in Belgium claim to be 100 per cent accurate.

The one commercial test approved by the Food & Drug Administration in the US is around 95 per cent accurate.

Low accuracy in tests means false positives show antibodies where there are none, or miss them in false negatives from people who have been ill.

And the studies that have been done so far are very small by normal scientific standards – they would not be considered acceptable in drug trials or studies into the causes of diseases.

The largest of those listed above was on 7,000 people in the Netherlands – 0.04 per cent of the country’s population – and the smallest was done on just 200 people in Chelsea, Massachusetts.

Although this may still give useful insight, such small studies are at a very high risk of being confounded – thrown off course – by random chance factors or errors.

For example, if a small study like the one in Chelsea, MA, happened to have been done in a place with a COVID-19 outbreak much more ferocious than the state as a whole – because of a high population of old people, for example – its results would not be representative of other areas, even if they were nearby.