Where’s our herd immunity?

Where’s our herd immunity?

In the early stages of the COVID-19 pandemic, there was a lot of discussion about herd immunity. Once a certain number of people had become immune through being infected, the epidemic would grind to a halt. This level is known as the Herd Immunity Threshold (HIT). There was (and still is) uncertainty as to exactly where this would kick in. Let’s call it 70%.

How many people have been infected? If we just consider those diagnosed, that ignores the asymptomatic cases. If we consider those who have tested positive, that understates the position because the tests weren’t available to start with, and there may still be asymptomatic cases not detected because they have no reason to take  a test. We can start with the number of deaths (as recorded on death certificates – currently about 150k). If the case fatality rate (CFR) is 1%, that translates to 15 million cases. (The CFR is higher in elderly and otherwise highly susceptible, and lower in younger people, but let’s focus on the concept rather than the accuracy of the data.) That’s about 20% of the population.

Of course, those earlier calculations were just based on becoming immune through being infected – but we now have the vaccines to consider. The headline figures that are generally quoted (e.g., BBC TV News) are that 90% of the adult population have had one dose, and about 66% have had both doses. You might think that we could add that to the above 20% who have been infected, but it’s not that simple – the two overlap (i.e., some of the vaccinated people will have had COVID already). Nevertheless, it is clear that the percentages are at or above the HIT. That is supported by surveys carried out by ONS which show that about 90% of UK adults have antibodies to SARS-CoV-2.

The snag here is the word ‘adults’ in the previous paragraph (i.e., those 18 or over). The ONS antibody survey shows that it is effectively 100% in those 50 or over, 95% for 35-49. dropping to 80% for 25-34 years, and only 60% for those in the 16-24 group. They don’t give figures for anyone under 16. (These figures are for England only; the details are different for the other nations, but the trends are the same).

Does this matter? It seems likely that if we consider the impact of vaccination and natural immunity combined, we should be near the HIT for the whole population. To show that it does matter, we have to consider the assumptions behind the concept of herd immunity. The most basic forms of epidemic modelling make two important assumptions: i) that the population is homogeneous, and ii) that it is randomly mixed. In more simple terms, this means that whoever you are, and wherever you are, if you have COVID, the number of people you are likely to infect is the same. This is obviously not true. If you live in a remote part of Scotland, you are less likely to infect others than if you live in central London. (I should add that modern epidemic modelling is much more sophisticated than the simple model, but the basic concept of herd immunity is still affected by these factors).

The most relevant, but somewhat less obvious, limitation here is the assumption of random mixing. Is the population in a specific area randomly mixed? If you have an infected 16 year old, are they more likely to pass it on to another 16 year old or to someone at 76? (I can’t remember the last time I came into close contact with a 16 year old!). So with the younger age groups not fully protected, it is quite possible to have rampant spread of infection there, while the older age groups are well protected.

To some extent, this is borne out by looking at the age group distribution of test positivity rates. Over the last month, these have risen from a very low level to 2-3% (daily) in school year 7 to age 24, and up to 1.4% in age 2 to school year 6, and those between 25 and 34. While there were increases in older age groups, these were much less. (A pinch of salt here – if more tests are done, you might get an apparent increase in the number of positive tests without there being an actual increase in the number of cases – you are just detecting them better).

So the current increase in the number of cases is being driven, to a large extent, by infections in younger age groups, who have not been vaccinated. This highlights the decision that was made in the vaccination campaign – to start with the older age groups and work downwards. Why was this decision taken? If the focus had been on simply controlling the number of cases, it might have been better to go all out for vaccinating everyone – especially as the younger age groups tend to mix more than the elderly do. The reason for starting with the most vulnerable comes down to the effect on the NHS. Older people (and others who are more vulnerable) are more likely to require hospital treatment, and especially intensive care, while those who are younger will often have only relatively mild  symptoms.

If we are to finally get on top of this disease, it is essential that vaccine uptake is increased in the younger age groups (and also in other groups that are at present under-vaccinated: e.g., some towns, some minority ethnic groups). It is disappointing therefore that the rate of vaccination has slowed markedly. In April/May, the total number of vaccinations (first and second doses) was as high as 600,000, Now it is down to not much over 200,000. Of these only 71,000 are first doses – so when they become eligible for a second dose, the number will be correspondingly low.

However, I am optimistic. The number of deaths is still very low compared to previous waves, either because the vaccination is reducing the effects of infection, or the proportion of cases amongst younger people is higher (or both). And if younger people are being infected without becoming seriously ill, that is doing the same job as vaccinating them – increasing the level of immunity. Let’s keep hoping!

Jeremy Dale

16 July 2021

Covid19 – another update

This is a good time to re-examine how well, or how badly, the UK has done, and some of the factors involved.

1. Number of deaths.

Data from https://www.worldometers.info/coronavirus/#countries [accessed 28-3-21]

We are frequently told that USA and Brazil have suffered greatly – and in terms of total deaths (562k, 310k respectively) that is true. But their population is much larger than that of the UK, and that has to be taken into account, by looking at the deaths per 100k population. If we do this for a selection of countries, we see the following table

CountryDeaths per 100k
UK186
Italy178
USA169
Spain160
Brazil145
France145
Germany91

I should add the caveat that the method of counting deaths varies from one country to another (even with the UK figures we get a different number from different sources), so the comparison is not absolutely reliable. It is also worth noting that some of the other countries have been catching up with us – the same comparison a month or two ago would have shown us to be further ‘ahead’. But it is quite clear that we have nothing to be proud of in the way we have dealt with the pandemic.

2. What could we have done better?

i) Lockdown and other controls. Obviously we were very slow to respond at the start; the Government delayed its response inexcusably. Would it have made any difference?

If we cast our minds back to January/February 2020, nobody knew when vaccines would become available, nor even whether effective vaccines would be possible. In the absence of a vaccine, all that can be achieved by lockdown (and other control measures such as border controls), would be a delay in the spread of the virus. It would not reduce the number of cases or deaths ultimately. The only factor that would eventually control the infection (in the absence of a vaccine) would be the much maligned concept of ‘herd immunity’ – eventually the number of people who had been infected (and hence became immune) would reach a level where further spread of the disease stopped. I should make it clear that this does not constitute advocating that as a policy; it is simply stating a fact.

That doesn’t mean that those control measures were pointless. Delaying the spread, or ‘smoothing the curve’, has a value. Spreading out the number of cases over time reduces the risk that hospitals would become overwhelmed. It also buys time for the development of a vaccine – and as it turns out that was critical. If our earlier control measures had been more effective, some of the people who died would now be available for vaccination.

ii) Border controls. Some countries have achieved a remarkable level of success by using border controls, coupled with rigorous lockdown measures. We are often told about New Zealand, Iceland, and Taiwan. for example. NZ and Iceland are much smaller countries, and much more isolated than the UK. Even Taiwan (which was cited as an example in a recent Guardian article) is only 1/3 of the size of the UK, and although it has frequent flights from China, it is more isolated than the UK. Although Britain is also an island, would it really be feasible to cut ourselves off from everywhere else in the way those other countries have done? All those ferries – and the Channel Tunnel – and including all the crews of the ferries, and the lorry drivers. To say nothing of the likelihood of informal passage from the continent. You can easily sail a small boat across the Channel, but I don’t fancy your chances of doing that to Iceland, or Taiwan; you would certainly need to know what you were doing. The experience of the Isle of Man is instructive here – they thought they had it sorted, and re-opened pubs etc. But it couldn’t be maintained, and now they have problems.

In the absence of a vaccine, such a policy would have to be continued until the disease was eliminated throughout the world.

3. We could have been better prepared.

i) General health. Many of the Covid-related deaths have been in ‘high-risk’ groups of people. Of these, I want to single out two factors: obesity and pre-existing respiratory conditions (for which air pollution is a major contributing factor). The Government has prevaricated for years over taking effective action to counter either of these problems. This is a major reason why we have done so badly in the current pandemic.

ii) The health service. The Government has treated the health service as a business, like selling soap. It’s not efficient to keep large reserve stocks of soap. You can predict how much soap people will buy, and if more is needed, you can get more supplies – ‘just in time’. That doesn’t work with the health service. Up to a point, you can predict how much spare capacity will be needed in an average winter, but that falls down when faced with an unexpected demand such a  pandemic. As we found out, it is not easy to suddenly buy in large supplies of PPE or respirators, when other countries are trying to do the same. And worse than that, you cannot suddenly increase the numbers of trained doctors and nurses. So the Government built all those Nightingale hospitals, forgetting that there were not enough staff available to run them – so they have been largely unused.

Several previous exercises had demonstrated that the health service was ill-equipped to deal with a pandemic – but these were ignored.

iii) Test and trace. We used to have a system where most hospitals had their own diagnostic laboratory, backed up by a network of laboratories run by the Public Health Laboratory Service. So samples could be taken locally (either in hospitals or by GPs), and tested locally. Hence people ddn’t have to travel long distances to get a test, and the results were available very quickly (often the same day). If contact tracing was required, that was also done locally, under the auspices of the local authority. That system could have been used to run test and trace, but much of it had been gradually eroded, by merging laboratories (so they are less local), abolishing the PHLS (the remnants being formed into the Health Protection Agency, HPA), and slashing local authority budgets, so the contact tracing teams are now a shadow of their former selves. The Government seemed to decide that was left of the system was not capable of doing the job, so they outsourced it, ending up with a system that was not fit for purpose, and was a gross waste of public funds.

Although I am cross about the waste of money, I’m not sure that it made any difference to the pandemic. Contact tracing is crucial in the control of some infectious diseases – tuberculosis and sexually transmitted diseases being two prime examples. In both cases, these are long-term infections, where the person concerned will go on infecting others for ages, so you not only have an opportunity to prevent further transmission, but you can also do contact tracing backwards – find out where they might have caught the disease, and deal with that as well. These diseases also highlight the value of having a team of trained, experienced contact tracers. It requires a good deal of tact and professionalism to ask someone where they might have contracted a sexually transmitted disease. Plus local knowledge as well. Someone with an hour or two of training, working from a call centre, is not going to get far.

The main problem with contact tracing for Covid comes down to timing. If you consider someone who is infected on day 0, then suppose they become infectious on day 4 (but not yet symptomatic). So they have started spreading it to others. They develop symptoms on, say, day 6. On day 7, they think “I’m ill, I must do something”. So they book a test, which is done on day 8. On day 9 they get the results and are told to isolate. On day 10, the contact tracers have talked to the identified contacts. But those who were infected on day 4 have now been infected for 6 days, and have started to spread it to others. I admit that the timings I have used are guesswork, but the general message holds – it is very difficult to get through the process quickly enough to interrupt transmission effectively.

4. Vaccine

Here, at last, we come to an area where the Government has done some things right. First of all, they seem to have got ahead of the curve in placing advance orders for large quantities of several potential vaccines, at a time when it was still uncertain whether any of them would work. We should also give credit to the UK’s regulator, MHRA, in giving speedy assent (on an emergency basis) to the vaccines. It is rather cheeky of the Government to claim credit for the development of the Oxford/AstraZeneca vaccine; the work of the Oxford Vaccine Group was largely based on blue-skies research in developing their vaccine delivery system.

The Government also made a good decision over vaccine delivery. They decided to keep their hands off! – and especially in not out-sourcing it. They just let the NHS get on with it, with the result that groups of GPs, community organisations, and many others moved very quickly to set up an impressive network of vaccination centres, including taking the vaccination out to the people who needed it.

But what about the variants? Are they going to compromise the success of the vaccines? The first thing to make clear is that variation is nothing new. Almost certainly, the flu pandemics that we have had from time to time will have been influenced by the occurrence of variants of the flu virus – indeed there is some retrospective evidence that this happened. And flu viruses seem to be more variable than coronaviruses. This is largely responsible for the repeated outbreaks of seasonal flu (the flu pandemics are caused by a more radical reshuffling of genes between human and animal strains). What is new about Covid-19 is that for the first time we have been able to follow the emergence of these variants as they happen.

Secondly, there are severe limitations on the extent to which the spike protein can vary. It has to be able to recognise a specific receptor on human cells. If it varies too much, it won’t interact effectively with that receptor and will lose infectivity. So the virus has to tread a narrow line between escaping immunity due to the vaccine (or prior infection) while still retaining infectivity. Furthermore, we cannot assume that a variant that becomes predominant, displacing a previous strain, is inherently more infectious. A variant that causes less severe disease can have an advantage, as those infected will be less ill, and therefore less likely to self-isolate. It has been suggested that the emergence of such a partially attenuated variant was responsible for the ending of the 1918-19 flu pandemic. In addition, our immune system is very complex. Each of us will produce a unique mixture of antibodies in response to the vaccine, or to infection; some of these antibodies will be very specific, and will fail to recognise the new variant, while other antibodies are less specific and will recognise a wide variety of related strains. So some people will be fully protected; others, perhaps less so. That’s one of the reasons why, when we get a bad year for seasonal flu, some get infected while others do not. So, there’s no need to panic – but it is sensible for the vaccine companies to make new versions of their vaccines – which they are doing, as it is relatively straightforward to do (and is already done for flu vaccines as new variants emerge).

5. What of the future?

It is certain that there will be future pandemics. We are better prepared for flu than we were for the pandemics in 1918, or even 1957 or 1968 – not only in the ability to produce new vaccines quickly but also because of the availability of antiviral agents that can be used for treatment of influenza. SARS-CoV2 is likely to remain with us after the current pandemic is over, perhaps producing seasonal outbreaks. We don’t yet have antiviral agents that can be used against coronaviruses, but they will come – there are many candidates being investigated.

But what about further novel pathogens? We have to recognise that in a sense we have been lucky over the last century or two. Influenza and coronaviruses are both only moderately transmissible and moderately pathogenic. After the experience of the last 12 months, that may sound a surprising statement. But an Ro value of 2-4 (for Covid) looks small compared to measles (Ro 15-20). And the case-fatality rate CFR (the chance of dying if you have been infected) of perhaps 1% for Covid is low compared to that for plague or smallpox (about 50%), or HIV/AIDS (where, before the advent of anti-retroviral therapy, it was effectively 100%). If you want a sleepless night, try imagining a pandemic caused by a pathogen with the infectivity of measles and the CFR of HIV.

Jeremy Dale

29/3/21

COVID-19 – a bit of perspective: update

An update

According to ONS, in the week up to Sept 11th, there were 9,215 deaths from all causes in England. Of these, COVID-19 accounted for 97 deaths – that is about 1% of the total. Of course, we would expect that figure to rise as the number of cases increases. For the whole period of the epidemic (up to Sept 11), COVID deaths (52,482) account for about 12% of the total number of deaths (434,618) in England and Wales..

If we assume a case fatality rate (the number of deaths as a percentage of the number of cases) of 1%, this means that just under 9% of the population have been infected.  But we don’t know the case fatality rate for sure; I have seen estimates that put it as low as 0.3%. If that is true, then the number who have been infected (England and Wales) would be as high as 17.5 million, or 29% of the population.  That would mean, given the highly uneven spread of cases, that there would be some areas that would, at some point during the ‘second wave’, approach the level required for herd immunity to kick in (usually taken as about 60%).

Jeremy Dale 25/9/20

COVID-19 – a bit of perspective

Every death is a tragedy for those affected. Especially so for COVID, where someone may die in isolation, unable to see, or be seen by,  their loved ones. In no sense do I want to diminish that – and if you have been affected by the virus you may not want to read on. But in the midst of all the publicity about COVID, I feel it is necessary to try to put it into perspective.

First of all, let’s look at the data for COVID deaths. This is very confusing. The Government dashboard (1 August) puts it at 46,193 in the UK. This is the number who have died having had a positive test result. This has been much criticised, as it would seem to mean that even if you get run over by the proverbial bus, having been tested positive months ago, you are still counted as a COVID-related death. On the other hand, if you die from COVID but have never been tested (or the test didn’t work), you wouldn’t be counted.

If you look at the data from the Office of National Statistics (ONS), you get a different figure,  50.800, in England and Wales only. This comes from death certificates, and is the number of times COVID was mentioned (even if other causes such as pneumonia were also mentioned). I’ll stick with the ONS figure, mainly because I want to compare it with other data from ONS.

So, 50,000 deaths. That’s a lot of tragedies. But death is a part of life. During the period of the epidemic, 245,000 people have died from all causes – so that’s getting on for 200,000 people have died from something other than COVID. And during each of the last five weeks, more people have died from what is recorded as Influenza/pneumonia than from COVID.

Another way of looking at the impact of COVID is to consider the excess deaths – that is the number of people who have died from any cause, compared to the average number who died in the same period over the last five years. This shows that since the epidemic started there have been over 53,000 excess deaths. That measure includes the possible indirect effects of COVID, e.g., people who didn’t get appropriate treatment in time. If we look at the weekly breakdown of excess deaths, we see that during the last five weeks it has been negative – that is, fewer people are dying than expected. The likely reason for this is that one effect of COVID has been to cause the death of some people who would otherwise have died soon anyway.

Historical comparisons

The current crisis has highlighted the fact that we are no longer used to people dying in large numbers from infective diseases. Medical advances, including antibiotics and vaccines, coupled with improvements in nutrition, housing, public health and other environmental issues, have in general made such diseases of historic interest only, at least in countries like the UK. (This is not of course true for most of the world, where diseases such as malaria and tuberculosis are causing death and suffering on a large scale – in low income countries, communicable diseases represent 5 of the top 10 causes of death).

A look at the death statistics (from ONS) for 100 years ago (1915) illustrates the point. In that year, there were 66k deaths from pneumonia/bronchitis and 39k deaths from tuberculosis. We can add others – 13k deaths from measles. 5k each from diphtheria and flu (not an epidemic year), 4k from whooping cough and nearly 2k from scarlet fever.

In the more distant past, there are numerous examples of devastating infections. The Black Death (1381) is thought to have killed a third of the population. In the nineteenth century, there were repeated epidemics of cholera, with tens of thousands of deaths, and tuberculosis was rampant (at its height, causing a third of all deaths).

In more recent times, the best comparison is with pandemics of influenza. (Technical note: Influenza viruses are classified by their H and A antigens, the most common type being H1N1. Various H1N1 strains are similar but not identical in both, so you get a degree of cross-immunity, while another type say H2N2 differs in both and there is no cross-immunity between them. Major pandemics usually occur with a virus that has ‘shifted’ to different H and N types)

In 1957-58 there was a pandemic of so-called ‘Asian flu’ (H2N2), which caused some 20-30k deaths in the UK. Then in 1968-69, we had an H3N2 strain (labelled ‘Hong Kong’ flu) for which the estimates of the number of UK deaths go up to 80k. In neither case was the official response anything like what we are currently seeing with COVID-19. And the media managed to find plenty of other news to cover.

More recently, there was some concern about ‘swine flu’ (2009). The incidence rose to about 110k cases per week in July, before dropping off, and then re-emerging in the autumn to about 84k cases per week in October. However, mortality was low (<1,000 deaths in UK), probably because this was an H1N1 strain, and older people had already encountered H1N1 strains and so had significant immunity to it.

Why all the fuss?

Why is it that fifty years ago we could face a disease that caused up to 80k deaths, not exactly with equanimity but at least without the massive sacrifices that we are currently making for a disease of (apparently) similar magnitude? Of course, we have to recognise that without the control measures it might have been much worse. Based on what was known about the disease, the initial assessment was that, if left unchecked, the disease would spread until Herd Immunity was achieved, and that would happen when about 60% of the population had been infected. Assuming a case fatality rate of 1%, that implied something like 350,000 deaths, which was deemed unacceptable. Of course we will never know if that would have happened, but a comparison with other countries is interesting. We hear a lot about the numbers of deaths in the USA and Brazil, but if we look at the numbers of deaths per million population, we are still some way ahead of either of them (UK 680, USA 477, Brazil 440) – although I am well aware of the dangers of reading too much into such comparisons, given the different methods and reliability of reporting deaths. But superficially, it could mean that our lockdown didn’t have much effect, and the original estimate of 350,000 deaths was over the top.

I’m not saying that we should all ignore the advice, and go out and party. But let’s keep a sense of perspective. At the individual level, unless you are in an extremely vulnerable category, there are plenty of other ways of dying that we don’t bother too much about. But collectively, we still have a duty to try to limit transmission so as to protect those who are more vulnerable. Above all, don’t panic!

Jeremy Dale

2 August 2020

Moonshine Project

Is it a good idea to regularly test the whole population for infection with SARS-CoV-2?

Superficially, you might well say the answer is ‘Yes, obviously’. But I want to look at it in more depth.

First of all, obviously, there are the practical considerations. The current system, fragmented as it is (some would say deliberately fragmented between the private and public sector), is proving to be quite unable to cope with testing those who really need it. And the tracing aspect is even worse. Remember when you hear the figures quoted, e.g., 70% of infections are detected, and 70% of contacts are traced – that is 70% of 70%, or 49%. So even with the Government’s claims, only about half of potential contacts are traced. And it is far too slow – several days to get the test result, and several more days (often longer) to reach the contacts, who thus may have been spreading the disease for days before being told to isolate,

Secondly, it depends on technology that doesn’t yet exist. There is much talk about rapid tests that are being introduced. These are antibody tests, which only tell you whether someone has been infected at some time in the past, not whether they are currently infected, Antibody tests don’t give a positive result until the infection is, in most cases, over.

Leaving the practicality aside, let’s assume that somehow we do have a test that can be easily and rapidly done for vast numbers of people. Would it be a good idea then, to screen the whole population?

To answer that question, we need to look at some concepts related to the test. I should clarify that what I mean by ‘test’ is what happens in the laboratory with the sample you have provided. The other parts of the process – taking the sample, transmitting it to the lab, and reporting the results – all have their problems, but I’m not considering those.

No laboratory diagnostic test is perfect. Biological systems are very complex, and all sorts of things can lead to a ‘wrong’ result. This can mean that someone who is infected gives a negative result (a ‘false negative’). For example, there might be something in the sample that interferes with the test process. The ability of the test to detect real positives is referred to as the sensitivity, which is defined as the number of positives detected divided by the number of true positives.

The other side of the coin is the specificity, which roughly means the ability of the test to correctly identify someone who is not infected. This is defined as the number correctly identified as negative divided by the number who are really not infected. This gives us a measure of how often we would get a false positive – i.e., a positive test result for someone who does not have the disease.

Now we can look at what this means in reality. Let’s assume we have a test that is 99% sensitive and 99% specific. That would be regarded as a really good test. And let’s assume we have a prevalence of 1 in 1,000. So, in a population of 100,000, there would be 100 cases, and the test would detect 99 of them. That’s not bad. The crunch comes with the specificity. In this situation, the test would return a positive result for 999 people who do not have the disease (false positives). In other words, there would be about 10 false positives for every correctly identified positive. If we scale that up to 60 million people, there would be nearly 600,000 people who would be told they were infected when they weren’t. Would they all have to isolate themselves? And even thinking about contact tracing on that scale gives me a headache.

These considerations apply to all occasions when you are screening random samples for some relatively rare event, whether it’s screening donated blood for an infectious agent such as HIV, or mass screening for cancer. The standard way of dealing with it is to use two independent tests. So, you take all the samples that gave a positive result in the first test and re-test them, with a different test. (It has to be a different test as the reason for the ‘wrong; result might be inherent in the sample). If you did that, with the above example, then after re-testing all the positives from the first test, you would find that fewer than 10% of the positives would now be false positives.

In effect this is what happens when you use a test to confirm a clinical diagnosis, or if you are applying a COVID-19 test to people who have symptoms. The clinical picture is the first test, giving a group of people who are much more likely to have the disease than the general population is.

Is a two-step procedure feasible? It could be argued that the first step could use the hitherto unknown and entirely conjectural rapid test, and the positive samples submitted to the existing PCR-based test – but only if this hypothetical new test was something entirely different. And it would still demand adding an additional 600,000 samples to the existing testing workload. Since that system is currently unable to meet the present demands, this seems unlikely.

If you would like to see the calculations I have used, and play around with them, send me an email and I’ll send you the spreadsheet.

Jeremy Dale

20 Sept 2020

COVID-19 – a bit of perspective

An update

According to ONS, in the week up to Sept 11th, there were 9,215 deaths from all causes in England. Of these, COVID-19 accounted for 97 deaths – that is about 1% of the total. Of course, we would expect that figure to rise as the number of cases increases. For the whole period of the epidemic (up to Sept 11), COVID deaths (52,482) account for about 12% of the total number of deaths (434,618) in England and Wales..

If we assume a case fatality rate (the number of deaths as a percentage of the number of cases) of 1%, this means that just under 9% of the population have been infected.  But we don’t know the case fatality rate for sure; I have seen estimates that put it as low as 0.3%. If that is true, then the number who have been infected (England and Wales) would be as high as 17.5 million, or 29% of the population.  That would mean, given the highly uneven spread of cases, that there would be some areas that would, at some point during the ‘second wave’, approach the level required for herd immunity to kick in (usually taken as about 60%).

Jeremy Dale 25/9/20

COVID-19 – a bit of perspective

Every death is a tragedy for those affected. Especially so for COVID, where someone may die in isolation, unable to see, or be seen by,  their loved ones.

In no sense do I want to diminish that – and if you have been affected by the virus you may not want to read on.

But in the midst of all the publicity about COVID, I feel it is necessary to try to put it into perspective.

First of all, let’s look at the data for COVID deaths. This is very confusing. The Government dashboard (1 August) puts it at 46,193 in the UK. This is the number who have died having had a positive test result. This has been much criticised, as it would seem to mean that even if you get run over by the proverbial bus, having been tested positive months ago, you are still counted as a COVID-related death. On the other hand, if you die from COVID but have never been tested (or the test didn’t work), you wouldn’t be counted.

If you look at the data from the Office of National Statistics (ONS), you get a different figure,  50.800, in England and Wales only. This comes from death certificates, and is the number of times COVID was mentioned (even if other causes such as pneumonia were also mentioned). I’ll stick with the ONS figure, mainly because I want to compare it with other data from ONS.

So, 50,000 deaths. That’s a lot of tragedies. But death is a part of life. During the period of the epidemic, 245,000 people have died from all causes – so that’s getting on for 200,000 people have died from something other than COVID. And during each of the last five weeks, more people have died from what is recorded as Influenza/pneumonia than from COVID.

Another way of looking at the impact of COVID is to consider the excess deaths – that is the number of people who have died from any cause, compared to the average number who died in the same period over the last five years. This shows that since the epidemic started there have been over 53,000 excess deaths. That measure includes the possible indirect effects of COVID, e.g., people who didn’t get appropriate treatment in time. If we look at the weekly breakdown of excess deaths, we see that during the last five weeks it has been negative – that is, fewer people are dying than expected. The likely reason for this is that one effect of COVID has been to cause the death of some people who would otherwise have died soon anyway.

Historical comparisons

The current crisis has highlighted the fact that we are no longer used to people dying in large numbers from infective diseases. Medical advances, including antibiotics and vaccines, coupled with improvements in nutrition, housing, public health and other environmental issues, have in general made such diseases of historic interest only, at least in countries like the UK. (This is not of course true for most of the world, where diseases such as malaria and tuberculosis are causing death and suffering on a large scale – in low income countries, communicable diseases represent 5 of the top 10 causes of death).

A look at the death statistics (from ONS) for 100 years ago (1915) illustrates the point. In that year, there were 66k deaths from pneumonia/bronchitis and 39k deaths from tuberculosis. We can add others – 13k deaths from measles. 5k each from diphtheria and flu (not an epidemic year), 4k from whooping cough and nearly 2k from scarlet fever.

In the more distant past, there are numerous examples of devastating infections. The Black Death (1381) is thought to have killed a third of the population. In the nineteenth century, there were repeated epidemics of cholera, with tens of thousands of deaths, and tuberculosis was rampant (at its height, causing a third of all deaths).

In more recent times, the best comparison is with pandemics of influenza. (Technical note: Influenza viruses are classified by their H and A antigens, the most common type being H1N1. Various H1N1 strains are similar but not identical in both, so you get a degree of cross-immunity, while another type say H2N2 differs in both and there is no cross-immunity between them. Major pandemics usually occur with a virus that has ‘shifted’ to different H and N types)

In 1957-58 there was a pandemic of so-called ‘Asian flu’ (H2N2), which caused some 20-30k deaths in the UK. Then in 1968-69, we had an H3N2 strain (labelled ‘Hong Kong’ flu) for which the estimates of the number of UK deaths go up to 80k. In neither case was the official response anything like what we are currently seeing with COVID-19. And the media managed to find plenty of other news to cover.

More recently, there was some concern about ‘swine flu’ (2009). The incidence rose to about 110k cases per week in July, before dropping off, and then re-emerging in the autumn to about 84k cases per week in October. However, mortality was low (<1,000 deaths in UK), probably because this was an H1N1 strain, and older people had already encountered H1N1 strains and so had significant immunity to it.

Why all the fuss?

Why is it that fifty years ago we could face a disease that caused up to 80k deaths, not exactly with equanimity but at least without the massive sacrifices that we are currently making for a disease of (apparently) similar magnitude? Of course, we have to recognise that without the control measures it might have been much worse. Based on what was known about the disease, the initial assessment was that, if left unchecked, the disease would spread until Herd Immunity was achieved, and that would happen when about 60% of the population had been infected. Assuming a case fatality rate of 1%, that implied something like 350,000 deaths, which was deemed unacceptable. Of course we will never know if that would have happened, but a comparison with other countries is interesting. We hear a lot about the numbers of deaths in the USA and Brazil, but if we look at the numbers of deaths per million population, we are still some way ahead of either of them (UK 680, USA 477, Brazil 440) – although I am well aware of the dangers of reading too much into such comparisons, given the different methods and reliability of reporting deaths. But superficially, it could mean that our lockdown didn’t have much effect, and the original estimate of 350,000 deaths was over the top.

I’m not saying that we should all ignore the advice, and go out and party. But let’s keep a sense of perspective. At the individual level, unless you are in an extremely vulnerable category, there are plenty of other ways of dying that we don’t bother too much about. But collectively, we still have a duty to try to limit transmission so as to protect those who are more vulnerable. Above all, don’t panic!

Jeremy Dale

2 August 2020

 

COVID-19 tests

COVID-19 tests

There has been a lot of talk about tests for COVID-19 (or the virus that causes it,  SARS-CoV-2). The purpose of this piece is to try to clarify the nature of the tests that are usually referred to, and some other possible tests.

In general, most diagnostic tests for an infectious disease fall under the following headings:

1.Tests for the presence of the whole (‘live’) pathogen

2. Tests for the presence of some part of the pathogen. These can be divided into

a) tests for the nucleic acid of the pathogen (RNA in this case).

b) tests for an antigen of the pathogen

3. Tests for antibodies produced by the body in response to the presence of the pathogen.

1. Live pathogen tests.

For most bacterial diseases, this is the ‘gold standard’. Bacterial pathogens can (mostly) be grown easily in the lab, so this is a good way of detecting them.

It is not so easy for viruses. They don’t grow on their own – they have to get into a cell and use that cell to make more copies of the virus. (This makes it very debatable as to whether we consider a virus as ‘alive’; ‘viable’ is a better term). So you need to start with a culture of a human or other mammalian cell line that you can try to infect with the virus. There is a wide variety of such cell lines in common use in virology laboratories.

Viruses are often fussy – they will grow on some cell lines but not on others. Also, you want to be able to see the effect the virus has on the cells – a ‘cytopathic effect’. Fortunately, previous work with the related virus that caused SARS indicated which cell lines were most likely to be successful. So, quite soon after the first samples became available, several laboratories around the world had successfully cultivated SARS-CoV-2.

But this is not suitable as a diagnostic test. It is too slow (it takes several days or more) and needs high levels of containment. However, being able to grow the virus is important for several reasons, including:

– it is the only way of confirming that the virus is viable (and hence presumably infectious), which is necessary for investigating whether someone is shedding infectious virus, or for testing survival in the environment, for example.

– for lab-based testing of the effects of potential antiviral agents

– providing reasonably large amounts of the virus, especially for the development of some types of vaccines (especially those based on inactivated virus).

2. Tests for the presence of part of the virus.

This is the mainstay of current tests, using a technique known as RT-PCR (Reverse transcriptase-polymerase chain reaction). The RT part copies the viral RNA into DNA, and the PCR amplifies the DNA. It works through a series of cycles, each cycle doubling the amount of DNA. So after 10 cycles, you have 1,000 copies, after 20 cycles a million (106) copies, and after 30 cycles 109 copies. This can all be automated, so you put it into the machine which then tells you when you have a detectable product.

A key element in this process is that it requires the binding of specific ‘primers’ – short lengths of nucleic acid derived from the sequence of the DNA copy of the viral RNA. If the right DNA is present, the primers will bind and you get amplification. If it is not there, there’s no binding, and no amplification. So it is not only exquisitely sensitive, it is also highly specific.

You need to know the sequence of the viral RNA before you can do this. This sequence was made available very shortly after the virus was first identified.

It is important to remember that this test only detects the presence of viral RNA, not the whole virus. So a positive test does not necessarily mean that you are still infectious; it depends on how long the RNA remains after the virus has been inactivated by your immune response.

Some pathogens are detected or identified through the presence of other parts of their structure, mainly those that invoke an immune response – i.e., antigens. I have seen references to an ‘antigen test’ for this virus, but these are probably erroneous references to the RT-PCR test (to contrast it to the antibody tests, see below). There seems no point in an antigen test, which would be slower and less sensitive the RT-PCR.

3. Antibody tests.

These are highly controversial.

The principle is that during infection you produce antibodies (immunoglobulins) that are capable of reacting with a viral antigen (such as the proteins that form the ‘spikes’ on the surface of the virus). So you take a blood sample, mix it with a sample of the specific antigen, and the test detects the binding of the antibody to the antigen. So far, so good.

The problems are of two sorts. Firstly, sensitivity. During the early stages of infection, there may be only low levels of antibody. So you may not detect infection at that stage. Later on, after you have recovered, the antibody levels may drop off quite quickly – so it may not be a good test for immunity (see below).

Secondly, specificity. SARS-CoV-2 is closely related to the SARS virus, and (less closely) to a variety of other coronaviruses that cause a proportion of common colds. While it is easy to make the RT-PCR test specific (by choosing suitable primers), it is not so easy to get a highly specific antibody test.

These problems may be overcome in a lab setting, with experienced technicians, and larger samples of blood – and some laboratories are already conducting surveys using antibody tests. But we are not yet at the stage where you could reliably use a ‘pin-prick’ test in your own home.

This leads on to a related point. If a reliable antibody test says you have antibodies to the virus, what does that mean in practical terms? Especially, will you be resistant to further infection? I don’t want to get bogged down in the direct evidence, which is inconclusive – there are anecdotal reports of a second infection in individuals who have converted to RT-PCR negative (but there are other explanations for that), and on the other hand some animal experiments indicate that re-infection does not happen. The question I want to address is whether the detection of antibodies (assuming a reliable test) means you are now resistant to infection. A full answer, based on general concepts with other viruses, would be far too long and complex to be covered here -and I’m not an immunologist, so I’m not equipped to tackle it, but there are a few points worth making.

– With some viruses, you get ‘neutralising antibodies’ – i.e., binding of the antibodies to the virus is by itself enough to prevent the virus infecting cells (this can be detected in the lab). If that does not happen, the body has other ways of disposing of the virus-antibody complexes, including the activity of various cells such as macrophages and T cells. So in that case your resistance to the disease may depend on this cell-mediated immunity rather than, or as well as, the production of antibodies.

– As described above, antibody levels may fall off considerably after you recover from infection. But it’s worth noting that you can be effectively immune even if antibody levels are low – your immune system may ‘remember’ the previous exposure and produce antibodies very quickly in response to further exposure.

– Antibodies can be damaging as well as protective. The classic example is dengue, where a second infection, with a different strain, causes a much more serious disease – a phenomenon known as antibody-dependent enhancement. Although it is thought unlikely that this will happen with SARS-CoV-2, the prospect is enough to make the vaccine developers twitchy.

Jeremy Dale

24 April 2020