Coronavirus - Antibody tests

[Bletchleyite]

So one thing that seems clear is that so far we don't have reliable antibody tests nor any certainty about immunity, so releasing antibody tests into the community might do more harm than good, as people who think they're immune but aren't take big risks and end up in intensive care.

But what I don't understand is why a test with a known level of error, even if it's quite high (I think we are generally talking 70-80% reliable, so 20-30% error), can't be used without telling people their result to do a sample-based analysis of how many cases are in the field, which is still useful to know? Error can, after all, be brought into the calculation of an overall figure, and a figure for the population of % previously infected could still be useful even with that large level of error?

Any thoughts on why this is seemingly not being done? Or am I missing something?

 

[Domh245]

My reading of the news stories is that the concern is more that it doesn't guarantee immunity than the unreliability of the test. Sure there's a 70% chance that this person had it, but if they don't have immunity then it's a bit academic and of no use other than seeing how widespread it had been.

Of course, if (reliable!) antibody tests don't prove immunity, there's not much hope for a vaccine either, but my suspicion is that it's being said because we haven't done the trials to prove it does give immunity rather than any sort of suspicion that our understanding of immunity doesn't apply to SARS-CoV-2

 

[Bletchleyite]

My reading of the news stories is that the concern is more that it doesn't guarantee immunity than the unreliability of the test. Sure there's a 70% chance that this person had it, but if they don't have immunity then it's a bit academic and of no use other than seeing how widespread it had been.

Though that in itself is still a potentially useful/interesting figure. The reason it sprung to mind is that I read that some other countries were still going ahead with it for that reason, though I forget where I read it.

It might also be useful to be able to establish just how big a problem reinfection actually is, for instance whether it's primarily happening to people who are immunosuppressed, or whether it's more random than that.

 

[Mogster]

We need an accurate measure of antibody prevalence, mass screening with a test with 70% sensitivity would pretty much just confirm what we already know, that people are being exposed and antibody is present in the population.

There are two basic parts to test accuracy, sensitivity and specificity. Specificity means producing positive results for the right thing you want to measure. You really don’t want to be generating positives when people have been infected with one of the other common circulating coronavirus for instance. Then there’s sensitivity, how low a level of the thing you want to measure can you detect?

70% sensitivity means potentially missing 3 in 10 genuine positives (or not, you don’t know if they are there or not...) and introducing massive potential error or “confidence interval” into your estimate. Of course you should include that potential error in the findings from your work, but huge confidence intervals make the results much less meaningful.

 

[Cardiff123]

Sky News are running this story today, which more or less renders useless the government's antibody test strategy to get out of lockdown, as well as any notion that 'herd immunity' by letting healthy people get the virus, would work.

Coronavirus: Currently 'no evidence' that COVID-19 survivors have immunity, WHO warns

Epidemiologists warn there is no proof that antibody tests can show if someone who has been infected cannot be infected again.

news.sky.com

Coronavirus: Currently 'no evidence' that COVID-19 survivors have immunity, WHO warns

Epidemiologists warn there is no proof that antibody tests can show if someone who has been infected cannot be infected again.

There is no evidence that people who have recovered from coronavirus have immunity to the disease, the World Health Organisation (WHO) has said.

The UK government has bought 3.5 million serology tests - which measure levels of antibodies in blood plasma.

But senior WHO epidemiologists have warned that there is no proof that such antibody tests can show if someone who has been infected with COVID-19 cannot be infected again.

BBC News are also carrying the story:

Coronavirus: Double warning over antibody tests

Experts say screening may not prove immunity from coronavirus, and Britons are warned over unproven tests.

www.bbc.co.uk

 

[Jayden99]

Sky News are running this story today, which more or less renders useless the government's antibody test strategy to get out of lockdown, as well as any notion that 'herd immunity' by letting healthy people get the virus, would work.

Coronavirus: Currently 'no evidence' that COVID-19 survivors have immunity, WHO warns

Epidemiologists warn there is no proof that antibody tests can show if someone who has been infected cannot be infected again.

news.sky.com

So if it turns out to be the case that immunity isn't a thing, where do we go? I assume it would be a case of accepting that COVID continues to circulate in much the same way as the usual winter diseases

 

[Domh245]

It doesn't render it useless - the WHO is (quite rightly) pointing out that antibodies aren't a guarantee of immunity - but this is because no trials have been carried out. Based on our understanding of viruses and immunity it isn't a bad presumption but there will be a lot of egg on face if trials show that antibodies and catching the virus previously don't give immunity.

So if it turns out to be the case that immunity isn't a thing, where do we go? I assume it would be a case of accepting that COVID continues to circulate in much the same way as the usual winter diseases

Increase hospital capacity, implement some level of contact tracing, and then live with it

 

[Mogster]

Long lasting immunity has been demonstrated against SARS though. People are still producing useful neutralising antibody 15 years later. Convalescent plasma from patients who have recovered from CoV2 seems to be producing good results when transfused to very ill patients although the reports are only anecdotal at this early stage.

We need people who’ve recovered from coronavirus (COVID-19) to donate blood plasma, as part of a potential clinical trial to help with the national effort against the virus.

Plasma programme

We have now completed two major trials for convalescent plasma for people in hospital with coronavirus and no longer need convalescent plasma donors to come forward to donate.

www.nhsbt.nhs.uk

 

[Bletchleyite]

I think it's quite possible some people will develop immunity and others not. Which is still useful to know.

 

[yorksrob]

So if it turns out to be the case that immunity isn't a thing, where do we go? I assume it would be a case of accepting that COVID continues to circulate in much the same way as the usual winter diseases

Try and look for drugs that can mitigate the worst effects. This work is also taking place at the same time as the search for a vaccine.

 

[Bantamzen]

To be honest I find it slightly perplexing that the media is making such a fuss about this. The presence of anti-bodies show that the virus has been in the person's body, and that the immune system has reacted to kill the infection off. There is no guarantee that the antibodies will remain active & capable of keeping the person immune from future infections of Covid-19, or indeed any other infection. None of this is anything new under the sun. What the anti-body tests will show us is a more realistic data on how far it may have spread, which will help us make decisions to allow people to get back to their lives whilst looking after those that will need help.

 

[Mogster]

Neutralising antibodies are antibodies that are proved to be active against the virus, they destroy the virus or at least prevent it working. Not all antibodies have neutralising effects against the virus they are directed against.

The paper below describes neutralising antibodies being produced against CoV2 in Chinese patients.

https://www.medrxiv.org/content/10.1101/2020.03.30.20047365v1.full.pdf

We found that most COVID-19 patients developed SARS-CoV-2-specific NAbs (Neutralising Antobodies) at the convalescent phase of infection. The titers of NAbs reached their peak at 10 to 15 days
after disease onset and remained stable thereafter in patients.

Some review responses from scientists

expert reaction to preprint on antibody responses to the COVID-19 virus (SARS-CoV-2) in a group of recovered people in Shanghai | Science Media Centre

www.sciencemediacentre.org

”In summary, the results of this study are pretty much as one might expect for any respiratory virus, suggesting that there is nothing terribly unusual about the antibody response to the Covid-19 virus. This is reassuring.”

 

[Greybeard33]

What the anti-body tests will show us is a more realistic data on how far it may have spread, which will help us make decisions to allow people to get back to their lives whilst looking after those that will need help.

Unless/until it is known that antibodies confer immunity, I do not see that there is any practical value in mass testing to determine what proportion of the population have antibodies. Everyone, whether or not they have already been exposed to the virus, must be regarded as potentially vulnerable to infection or reinfection, and so the R transmission rate must be kept below 1 indefinitely to prevent exponential growth in the number of cases. This requires some combination of, on the one hand, on-going social distancing, and/or on the other hand, mass PCR testing plus contact tracing/quarantining.

Neutralising antibodies are antibodies that are proved to be active against the virus, they destroy the virus or at least prevent it working. Not all antibodies have neutralising effects against the virus they are directed against.

The paper below describes neutralising antibodies being produced against CoV2 in Chinese patients.

https://www.medrxiv.org/content/10.1101/2020.03.30.20047365v1.full.pdf

Some review responses from scientists

expert reaction to preprint on antibody responses to the COVID-19 virus (SARS-CoV-2) in a group of recovered people in Shanghai | Science Media Centre

www.sciencemediacentre.org

I see that this study was carried out on Covid-19 hospital patients, albeit only those with mild symptoms. Presumably an "unknown unknown" is whether or not asymptomatic spreaders of the virus also develop sufficient neutralising antibodies to be detected in a test carried out weeks later?

 

[js1000]

So one thing that seems clear is that so far we don't have reliable antibody tests nor any certainty about immunity, so releasing antibody tests into the community might do more harm than good, as people who think they're immune but aren't take big risks and end up in intensive care.

But what I don't understand is why a test with a known level of error, even if it's quite high (I think we are generally talking 70-80% reliable, so 20-30% error), can't be used without telling people their result to do a sample-based analysis of how many cases are in the field, which is still useful to know? Error can, after all, be brought into the calculation of an overall figure, and a figure for the population of % previously infected could still be useful even with that large level of error?

Any thoughts on why this is seemingly not being done? Or am I missing something?

From what I have read so far the anti-body tests to date are apparently detecting those who have had a severe form of the Covid illness (i.e. requiring hospitalisation) and where the body has had a strong immune response. But for those with a strong immune system and who may have had an asymptomatic case of Covid the anti-body tests are not detecting the presence of anti-bodies. This is a particular problem.

The anti-body tests would have been a game-changer for those who are key workers in the health and social care sector. If the population is to acquire immunity as has been discussed, and in reality this will form one of the avenues out of this, then this should naturally form in the age demographic who are least affected. In this case this would logically be the under 30/40s who spend a lot of time socialising anyway.

But if we cannot detect anti-bodies in these age groups because they are younger and more healthy and the body does not need a strong immune response then this is clearly concerning as there could be younger 'silent' carriers passing this onto to older population and those with underlying conditions. Nor do we have an accurate percentage for how many cases are asymptomatic without conclusively being able to say someone has had it.

There will be some immunity to this illness but it may take a few bouts over a number of years in the younger population to adapt to it. One of the interesting things about the Spanish flu was the older population was not badly affected - it was mainly those between 20 and 40 years old. They say that the "Russian Flu" in 1889 30 years earlier provided the over 40s with some acquired immunity and why fewer died. Interesting theory.

 

[cactustwirly]

To be honest I find it slightly perplexing that the media is making such a fuss about this. The presence of anti-bodies show that the virus has been in the person's body, and that the immune system has reacted to kill the infection off. There is no guarantee that the antibodies will remain active & capable of keeping the person immune from future infections of Covid-19, or indeed any other infection. None of this is anything new under the sun. What the anti-body tests will show us is a more realistic data on how far it may have spread, which will help us make decisions to allow people to get back to their lives whilst looking after those that will need help.

No the antibodies won't stay active, but the immune response creates memory cells, it's these that provide immunity, allowing the immune system to produce the correct antibody quickly.

 

[Bletchleyite]

No the antibodies won't stay active, but the immune response creates memory cells, it's these that provide immunity, allowing the immune system to produce the correct antibody quickly.

Which is how a vaccine works - you teach the immune system to create those antibodies by showing it something that is very similar but much less harmful. For instance, the first vaccine worked by actually giving people cowpox (they did suffer the full course of the disease) because it turned out that the exact same antibody also was effective at fighting smallpox (which killed people). If it doesn't work at all for any people who have had the actual disease, then there is no hope for a vaccine either - which means attention must turn to finding a treatment for it, and how we can design a society to live with it for many, many years or unless/until it mutates to become less harmful (which is what viruses normally do, though not always). However, that "disaster scenario" seems unlikely given that it isn't true for any other coronaviruses that circulate widely in humans.

(With regard to colds, some of which are coronaviruses, the reason you keep getting them is not that you're not immune to a specific one, but because there are loads of different ones. When a cold is going round your family each person only gets it once. If there was no immunity, it'd keep going round and never go away)

 

[Greybeard33]

(With regard to colds, some of which are coronaviruses, the reason you keep getting them is not that you're not immune to a specific one, but because there are loads of different ones. When a cold is going round your family each person only gets it once. If there was no immunity, it'd keep going round and never go away)

There are only four coronaviruses that are endemic in humans and cause upper respiratory tract infections, comprising about 15% of common colds (rhinoviruses, not coronaviruses, cause the majority of colds). These coronaviruses continually circulate, which must mean that any immunity to them is short lived - you can catch the same one more than once in your life. There is no vaccine against any of these four common cold coronaviruses, nor against SARS-CoV or MERS-CoV.

Six species of human coronaviruses are known, with one species subdivided into two different strains, making seven strains of human coronaviruses altogether. Four of these coronaviruses continually circulate in the human population and produce the generally mild symptoms of the common cold in adults and children worldwide: HCoV-OC43, HCoV-HKU1, HCoV-229E, HCoV-NL63. Coronaviruses cause about 15% of commons colds. The majority of colds are caused by rhinoviruses. The four mild coronaviruses have a seasonal incidence occurring in the winter months in temperate climates. There is no preference towards a particular season in tropical climates.

Four human coronaviruses produce symptoms that are generally mild:

Human coronavirus OC43 (HCoV-OC43), β-CoV
Human coronavirus HKU1 (HCoV-HKU1), β-CoV
Human coronavirus 229E (HCoV-229E), α-CoV
Human coronavirus NL63 (HCoV-NL63), α-CoV
Three human coronaviruses produce symptoms that are potentially severe:

Middle East respiratory syndrome-related coronavirus (MERS-CoV), β-CoV
Severe acute respiratory syndrome coronavirus (SARS-CoV), β-CoV
Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), β-CoV

https://en.wikipedia.org/wiki/Coronavirus

 

[Adam Williams]

But what I don't understand is why a test with a known level of error, even if it's quite high (I think we are generally talking 70-80% reliable, so 20-30% error), can't be used without telling people their result to do a sample-based analysis of how many cases are in the field, which is still useful to know?

Any thoughts on why this is seemingly not being done? Or am I missing something?

I believe this is definitely being done, albeit on a small scale to begin with - https://oxford.onlinesurveys.ac.uk/coronavirus-disease-covid-19-in-the-uk-community

Note that this test looks for antibodies (via the finger-prick blood test) as well as identifying current infections by looking for genetic material from the virus collected using the provided swabs - these are the polymerase chain reaction (PCR) and nanopore sequencing tests referred to in the study.

You can sign up as a volunteer, but you won't receive results. They're starting with folks who live in Oxford and expanding afterwards.

I have signed up, but I really don't like blood so I'm sure it'll be a "fun" experience if they send me a kit.

 

[TheEdge]

One of the interesting things about the Spanish flu was the older population was not badly affected - it was mainly those between 20 and 40 years old. They say that the "Russian Flu" in 1889 30 years earlier provided the over 40s with some acquired immunity and why fewer died. Interesting theory.

No. Most studies into the Spanish Flu seem to point the oddly weighted death rates towards the virus managing to trigger cytokine storms. These are massive overreactions of the immune system and trigger a positive feedback loop of inflammation which can cause death. They are normally worse in the healthier and stronger immune system of the young adult, hence the odd death rate.

Luckily Covid doesn't appear to regularly trigger these.

 

[Bletchleyite]

No. Most studies into the Spanish Flu seem to point the oddly weighted death rates towards the virus managing to trigger cytokine storms. These are massive overreactions of the immune system and trigger a positive feedback loop of inflammation which can cause death. They are normally worse in the healthier and stronger immune system of the young adult, hence the odd death rate.

Luckily Covid doesn't appear to regularly trigger these.

Indeed, though it does appear to be the cause of death when people die of it. The key bit of research at the moment would be (a) why those individuals, and (b) is there any way to prevent it. If (b) could be established, the death rate would go right down and it'd just be a nuisance. Even if only (a) could be established the shielding policies could be amended to protect those people.

 

[Mogster]

There are only four coronaviruses that are endemic in humans and cause upper respiratory tract infections, comprising about 15% of common colds (rhinoviruses, not coronaviruses, cause the majority of colds). These coronaviruses continually circulate, which must mean that any immunity to them is short lived - you can catch the same one more than once in your life. There is no vaccine against any of these four common cold coronaviruses, nor against SARS-CoV or MERS-CoV.

https://en.wikipedia.org/wiki/Coronavirus

This editorial was published in the Lancet a week or so ago. It appears there is some evidence that we produce useful long immunity to SARS and as they are closely related probably to CoV2. This is just opinion though, no one really knows yet. The article suggests CoV2 is much more closely related to SARS than the other 4 commonly circulating coranovirus.

https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(20)30788-1/fulltext

Hibberd argues that once people produce antibodies against a particular coronavirus, they probably have immunity for life. Indeed, Wang's laboratory has investigated how long immunity against SARS-CoV and Middle East respiratory syndrome coronavirus lasts. “17 years later, a SARS survivor still has neutralising antibodies against SARS—we found that not only were the antibodies there, but they could still neutralise the SARS virus.”

I think the bottom line is that it’s surprising how little we know about very common virus’s that circulate routinely through human populations. Until a more virulent form develops they really aren’t interesting to healthcare scientists.

More on CoV2 antibody from JAMA today. The discussion below the article is interesting also.

The Promise and Peril of Antibody Testing for COVID-19

This Medical News article discusses how antibody tests should—and should not—be used during the SARS-CoV-2 pandemic.

jamanetwork.com

Scientists around the world will be working to understand what sort of protection infection bestows, both in the laboratory and by following up recovered patients to see if reinfections occur. So far, the novel coronavirus doesn’t appear to mutate quickly. This, coupled with experience with other viral infections, suggests that people with SARS-CoV-2 antibodies may be protected at least for some time, Wetzler said.

 

[Greybeard33]

This editorial was published in the Lancet a week or so ago. It appears there is some evidence that we produce useful long immunity to SARS and as they are closely related probably to CoV2. This is just opinion though, no one really knows yet. The article suggests CoV2 is much more closely related to SARS than the other 4 commonly circulating coranovirus.

https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(20)30788-1/fulltext

I think that suggestion is not really supported by the text of the quoted article. The relevant section actually states:

There is a lot hanging on the uniqueness of the spike protein. In terms of the specificity of serological tests in which it is used, the more unique it is, the lower the odds of cross-reactivity with other coronaviruses—false positives resulting from immunity to other coronaviruses. The most similar of these is severe acute respiratory syndrome coronavirus (SARS-CoV), which led to the SARS outbreak of 2002. But another four coronaviruses cause the common cold, and ensuring there is no cross-reactivity to these is essential. “If you line up the amino acids of the spike proteins of SARS and the COVID-19 virus, there's a 75% identity”, says Lewis. Hibberd reckons the overall figure for common cold-causing coronaviruses is probably about 50–60%, but the potential for cross-reactivity really depends on whether the new tests select sections of the spike protein that are particularly distinct across coronaviruses.

The 60% figure presumably relates to the two common cold betacoronaviruses, -OC43 and -HKU1, which are in the same genus as the two SARS betacoronaviruses. 60% similarity versus 75% similarity is not a huge difference. It is curious that the article quotes two different sources for the similarity of the Covid-19 virus to the SARS and the common cold viruses - comparative estimates by the same scientist would be more meaningful.

Presumably the main reason that the early antibody tests have been giving too many false positives is that they fail to discriminate adequately between antibodies to SARS-CoV-2 and those to at least one of the common cold coronaviruses.

 

[Mogster]

Presumably the main reason that the early antibody tests have been giving too many false positives is that they fail to discriminate adequately between antibodies to SARS-CoV-2 and those to at least one of the common cold coronaviruses.

Yes definitely that will be it. The problem is made worse by the fact the other Coronavirus are commonly found within the UK. If they weren’t you could exclude their presence on the grounds of rarity, but they are very common so you can’t and the test needs to not cross react with them or it’s useless.

PHE have started a large scale antibody prevalence study from several UK regions. They must have an antibody screening test they feel is at least useful as a research tool.

 

[TheEdge]

Indeed, though it does appear to be the cause of death when people die of it. The key bit of research at the moment would be (a) why those individuals, and (b) is there any way to prevent it. If (b) could be established, the death rate would go right down and it'd just be a nuisance. Even if only (a) could be established the shielding policies could be amended to protect those people.

As I understand it the difference here is Spanish flu "did not pass go, did not collect £200 and went straight to cytokine storm" which helped it kill the young and healthy quickly unlike Covid that appears to follow the traditional kill the weakened of most infectious diseases. I thought from reports the eventual cause of death is pneumonia.

 

[Jozhua]

I think it's quite possible some people will develop immunity and others not. Which is still useful to know.

This seems reasonable, even the best vaccines aren't 100% effective, they just need to work on enough people to provide herd immunity.

There are only four coronaviruses that are endemic in humans and cause upper respiratory tract infections, comprising about 15% of common colds (rhinoviruses, not coronaviruses, cause the majority of colds). These coronaviruses continually circulate, which must mean that any immunity to them is short lived - you can catch the same one more than once in your life. There is no vaccine against any of these four common cold coronaviruses, nor against SARS-CoV or MERS-CoV.

https://en.wikipedia.org/wiki/Coronavirus

To be fair, SARS and MERS were contained and the common cold isn't that much of a killer. There's probably a lot more resources going into preventing SARS-CoV-2 infections.

 

[Jozhua]

Yes definitely that will be it. The problem is made worse by the fact the other Coronavirus are commonly found within the UK. If they weren’t you could exclude their presence on the grounds of rarity, but they are very common so you can’t and the test needs to not cross react with them or it’s useless.

PHE have started a large scale antibody prevalence study from several UK regions. They must have an antibody screening test they feel is at least useful as a research tool.

Glad to hear about the antibody prevalence study! Hopefully this can go towards shedding some light on the situation...

 

[greyman42]

I am not sure if this is the correct section but I have a question.
When you see the map of the UK showing how many people have been infected by the virus in each area, who is included in these figures. How do they compile the figures?

 

[Bletchleyite]

It's based on the number of tests.

 

[greyman42]

So it would only be people who have attended hospitals and surgeries?

 

[Bletchleyite]

So it would only be people who have attended hospitals and surgeries?

It is the number of people tested in any setting. It's a notifiable disease so any positive test is notified to the Government, whoever does it.

Helpfully the slides used at the daily briefings (which are also available here: Latest from Coronavirus (COVID-19): UK government response - GOV.UK ) separate the tests done in hospitals which made up the original figure (blue) from the tests done in other settings e.g. the new drive-in testing sites (orange), so the picture doesn't look as depressing as it might do otherwise.