One of the fears of many people in relation to covid has been that the immunity that develops after infection is so short lived that the infection will just keep going around and around and re-infecting everyone (until everyone is dead, I assume).
Two pieces of evidence have been presented to support this belief. The first concerns a few cases of “re-infection” that have been broadcast widely in media, even though virtually all of these cases have been either completely asymptomatic or only very mildly symptomatic the second time around – a sure sign the the immune system still remembers covid and is doing its thing to stop it.
The second concerns the fact that antibodies fade after infection. This builds on a fundamental lack of understanding of how the immune system works. Although the actively antibody producing cells diminish after an infection, these cells (so called “plasma cells”) are not responsible for immune memory. That role is filled by special “memory B-cells”, that lie dormant in the body, waiting for the infection to reappear. When it does, they quickly spring in to action and produce massive numbers of new antibody producing clones.
Now, however, covid has been around for a while, and we’re starting to get some pretty good data on how long immunity lasts after infection. There is a pre-print up on MedRxiv about a study that sought to gain a deeper understanding of what sort of immune memory is produced after a covid infection.
Before we get in to the details of the article, let’s talk a little bit about immune memory, so everyone is on the same page. Immune memory is the ability of the immune system to remember a pathogen after a first infection (or vaccination), and thereby respond much more quickly and effectively upon re-infection. It is mediated by three main types of cell. The first is the already mentioned memory B-cell, which is basically a dormant version of the antibody producing plasma cells. The second is the “memory killer T-cell”, which is a dormant version of the regular killer T cell (a.k.a CD8+ T-cell). Killer T-cells specialize in finding virus infected cells and getting them to commit suicide in a way that prevents the virus from spreading further.
The third is the “memory helper T-cell”, which among many other functions regulates the function of the other types of immune cell. Both killer T-cells and B-cells cannot become fully activated until helper T-cells have become activated. The central function of T-helper cells is shown by AIDS (Aquired Immune Deficiency Syndrome), a disease caused by the destruction of the T-helper cells by the Human Immunodeficiency Virus (HIV) – without the T-helper cells, other parts of the immune system cannot become fully activated, and the immune system is not able to function effectively.
In case you’re curious, the reason B-cells are called B-cells is because they mature in the bone marrow, so the B is for Bone marrow. T-cells mature in the thymus, so the T is for Thymus.
Ok, now you know enough to understand the results of the study. 185 people with confirmed covid-19 were recruited and had blood samples drawn. 92% had not required hospitalization, so only a minority had had severe disease. The ages of the participants varied from 19 to 81. The blood samples were collected from several different sites across the United States.
The results of the study were based on analysis of the participants blood. 79% of participants only provided blood at a single time point, which varied from six days post-infection to more than six months post-infection, while the remainder (21%) provided blood at multiple time points. In other words, this was not really a longitudinal study, since most participants only had their blood analyzed at a single point in time, although there was some longitudinal data. 41 participants provided blood samples at six months or longer after infection, and this is really the group we’re most interested in, since this is the group that can tell us if there is still a good level of immune memory six months after infection.
Let’s look at the results.
Among the 54 individuals measured at one month post infection, 98% had antibodies. Among the 41 individuals measured at six to eight months post infection, 90% had antibodies. As mentioned before, antibodies are produced by plasma cells, and although antibodies in the blood stream decline with time as the plasma cells start to disappear, there should still be memory B-cells present for much longer, which can quickly be activated upon re-infection. That’s why it’s actually more important to look at what’s happening with memory B-cells than with antibodies, if you want to know how long your body maintains the ability to mount an antibody response to an infection. So, what did happen with the memory B-cells?
The prevalence of memory B-cells increased at each time point measured up to five months post infection, at which point they reached a stable level. There was no sign of a decline in memory B-cells after the five month mark.
Next we have the killer T-cells. At one month post infection, 61% had detectable memory killer T-cells. At six to eight months, 50% had measurable killer T-cells. it was however only possible to test for these cells in 18 individuals at the six month mark, so the confidence interval is wide, and thus it’s really impossible to say exactly what the trajectory was between the one month and six month marks. What can be said though is that a large proportion of participants still had measurable killer T-cells at six months.
Finally we have the memory helper T-cells. 94% of those measured at one month had measurable helper T-cells. Among those measured at six to eight months, that number was 89% (again, this data is based on only 18 individuals).
So, what can we conclude?
First, it’s important to note that this study had some weaknesses. The first is that, with the exception of a minority of participants, the study was cross-sectional, not longitudinal. This means that we’re not comparing people with themselves over time, we’re comparing them with other people who happen to be at a different point in the time line. It would have been better to have longitudinal data for all participants. The second is that some of the groups studied were pretty small, which creates wide margins of error. Some of the data was based on less than twenty individuals, which is really a tiny number.
A third weakness is that this study isn’t looking at how many people get reinfected with covid after a certain amount of time, it is looking at biomarkers – in other words, it is using proxy data, which is clearly a less reliable type of information than seeing what is actually happening to people in the real world. It’s kind of like doing a statin study and looking at what happens to cholesterol levels instead of looking at how many people have died after certain time point.
Having said all that, it is clear from this study that there is significant immune memory at the six to eight month time point after infection. At six to eight months after infection, 90% of measured samples still had antibodies and T-helper cells specific for covid-19, and 50% still had measurable T-killer cells. If the decline continues linearly over time from what was seen in this study, then it is reasonable to assume that most people continue to be immune to covid after infection for at least a couple of years.
You might also be interested in my article about the number of years of life lost, on average, when someone dies of covid, or my article about whether face masks are effective against covid.
38 thoughts on “How long does immunity to covid last?”
Hi, Dr. Rushworth. Thanks for the article. Does the fact that T cells mature in the thymus gland explain why SARS-CoV-2 is so dangerous for the elderly? I’ve read that the thymus gland disappears (turns into fat) as we age. Is that true? Are there ways to preserve the thymus? Thanks!
It’s true that the thymus shrinks and largely disappears as we age. It’s possible that that is one of the reasons older people are more susceptible to infection. I’m not aware of any methods to prevent this happening.
Should herd immunity be engineered? Giving small doses of the virus to non-vulnerable sections of the population on a consent only basis would achieve herd immunity quite rapidly. Particularly as the population in question are probably the most active socially and hence most likely to transmit the virus. They could all be asked to isolate post infection obviously…
Well, now we have vaccines that seem to be pretty effective (although the trial data haven’t been published yet), so in that case herd immunity can be achieved with vaccinations.
Hi! Very nice article.
What happens if I get a reinfection after a year? Does this study mean I can get sick after a year but it will be not as bad as the first time? Because I still have the B-Cells but maybe the T-Cells are gone or are at a to low level. Or can it be as bad as the first time?
I found this study, I think from the Netherlands. https://www.medrxiv.org/content/10.1101/2020.05.11.20086439v1.full.pdf
They investigated for several years a very small amount of people the infection with the other four Corona Viruses. They found frequent reinfections after 12 months. Of course the other Corona Viruses are not as bad as Covid19 so, there is no evidence of change in mortality and so on.
Kind regards Jens
Extrapolating forwards, this study suggests that most people who get reinfected at the twelve month mark will have only mild symptoms or no symptoms. It’s hard to say though at exactly what time point after infection half the people have lost immunity, it could be after just a few years or after ten or more years. I think it’ll be years before we know to that level of detail how long immunity lasts on average.
I read don’t remind where that there was made an experiment with frozen blood samples. In presence of Sars-cov2 the T-cells react as expected. The interesting thing is that those blood samples were colected in 2019. Can you say something about? Thanks.
I’m not sure exactly which study you’re referring to, but there’s been a lot of talk about cross-reactivity, i.e. that T-cells against other coronaviruses can also react to sars-cov-2.
Really appreciate this and all of your informative posts, especially Covid related. As others have noted, the U.S. media are following the lead of a recent Wall Street Journal article headlined “Long a Holdout from Covid-19 Restrictions, Sweden Ends Its Pandemic Experiment.” The WSJ article does not quite support the tendentious headline, but it is difficult to find sources in English that give a straightforward account of current Swedish Covid policy. Any suggestions? Thank-you.
Great explanatioin! Precise and also critical.
Is it correct to say that immune people still get infected but the immune system will clean out the virus before the infection causes any damage? And if so, may these immune people still be potentially infectious, e.g. through saliva?
That is correct. The virus can still infect cells in the respiratory tract. The difference when you have immunity is that the immune system detects the infection much more quickly and therefore reacts before the virus can do as much damage. It’s hard to say whether people with immunity can be infectious, I don’t think there is any good data on that yet, at least not that I’ve seen. That is one of the big questions with the new vaccines, whether they just make you immune or whether they also prevent you spreading the virus to others.
Thank you for all the articles and informations you provide to us. As you said, Sweden is in the eye of the beholder and your expertise and your experience as a doctor at KI in Stockholm help us to understand better the Swedish answer about covi19. I think your next article should be about mask, and why Mr Tegnell is not convinced about their efficacy’s (correct me if I am wrong). Would be interesting to have your opinion about this peer reviewed article about the effectiveness of mask: https://www.pnas.org/content/pnas/early/2020/12/02/2015954117.full.pdf
And is this article a good argument for the use of mask for the people?
Thank you. I’ve already written two articles about face masks. The most recent is here: https://sebastianrushworth.com/2020/11/19/covid-19-new-evidence-on-face-masks/
The PNAS study is observational. In other words it is pretty low quality data, and it cannot say anything about cause and effect.
Looking at SARS-CoV-2 strong RBD similarities with SARS-CoV, and the longitudinal studies done on immunity markers (IgG and specific T-cells) for that particular coronavirus, one could deduce the immunity response would last for at lesat 6 years.
Again, the sample was not large – but the results were very consistent.
That is certainly very possible.
Thank you for this series.
It is fall influenza virus season, as well as the very first covid fall season. How do they co-exist? In a normal there is 20-30% extra mortality compared to summer?
Do they outcompete each other “survival of the fittest” style?
Or do they assist each other, first one creating “easy prey” for the next one?
Any chance for cross immunity, the “spanish flu” saved the elderly 100 years ago, presumably due to earlier flu?
The influenza season hasn’t really started yet, but will be interesting to see what happens. We saw in spring that when covid arrived, influenza dropped drastically, which suggests that the viruses compete for hosts. One thing that is certainly true is that you can only die of one thing, and influenza and covid both preferentially kill the very old and frail, so if a lot people that would normally have died of influenza have instead died of covid, then that suggests the influenza season this year will be very mild. We’ll know in another couple of months.
Thanks for sharing this new data that now starts to be available. In this context it might be interesting to look back at discussions about these question early in this pandemic. These podcasts are entertaining as well as educational about immunology and immune defence that may be worth sharing.
TWiV 597: A lot of immunology and some COVID-19 with Jon Yewdell
TWiV 620: Antibodies and T cells in COVID-19 patients
Thank you Doctor, I have completly missed this article about the mask!
Reversibility of the thymic involution and of age-related peripheral immune dysfunctions by zinc supplementation in old mice. https://pubmed.ncbi.nlm.nih.gov/8582782/
Effect of zinc on thymus of recently malnourished children. https://pubmed.ncbi.nlm.nih.gov/72960/
Hello Dr. Rushworth,
You state that about the competition between Influenza and Sars-cov2 in spring that “Influenza dropped drastically”. I would like to know if you can say this because you saw that the Influenza tests that were made in spring turned out negative more and more, or is this statement only a statistical observation that there was a sharp decline in the amount of positive Influenza tests reported? Because in this last case it might easily be a fact that Influenza testing dropped/stopped in favour of Sars-cov2 testing. And that does not mean that the Influenza virus disappeared or was ‘beaten’ by Sars-cov2…
It is very conceivable, since symptoms following infection by either of the viruses are very much identical, that a positive Sars-cov2 test does not exclude an Influenza (co-)infection. To assume that last year every viral syndrome with a positive Sars-cov2 PCR-test is a Covid-case is bad diagnostic practice of course, if you did not exclude other viral pathogens that can cause similar symptoms! (and might just be the real culprit!). Co-infection is not a rare thing, especially with a very sensitive PCR test you can imagine.
So I would like to know if in spring Influenza testing was still being done in every patient with a viral syndrome, and if relatively more of these Influenza tests started to turn out negative as the Sars-cov2 virus spread and so pushed Influenza away. Or if it is just clinical prejudice that with all the focus on Covid everywhere that Influenza was/is just ‘forgotten’…?
Thanks and keep up the good work!
Thank you. I haven’t seen data on the number of influenza tests done, just data showing the number of positive influenza tests, so, as you say, the virus might still have been circulating but could merely have appeared to have disappeared due to a decline in testing.
We’ve tested for covid to an extent that we’ve never tested for respiratory diseases before. Pre-covid, if a child came in with respiratory symptoms conducive with RSV (respiratory syncytial virus), for example, but wasn’t sick enough to be admitted, we wouldn’t test for it, because it didn’t change management and was therefore considered a waste of tax payer money. The same is true if an adult showed up with signs of a respiratory viral infection, but wasn’t sick enough to be admitted. With covid, we’re testing everyone, even people who are totally asymptomatic, so it’s a completely new situation.
From my own hospital, I know that when we admit people with signs of a respiratory infection, we usually only test for covid, not for any other respiratory viruses, so, as you say, the risk of a patient falsely being diagnosed with covid when they actually have influenza or some other respiratory virus is real. In previous years, we would probably have been testing more broadly when a patient was admitted.
I personally know two people who was reinfected in november efter being infected in april first time. The wife is working for a karolinska lab and was verified with pcr both times and also had all the classic covid symptoms. Her husband had a mild infection first time and was confirmed the second time with pcr. Both had similar symtoms first and second time. She had a tough time both times, 3 weeeks with fever, extreme fatigue, loss of taste and smell. Hard to breathe. The son in the family, below 10 in age, had no symptoms whatsoever.
First of all, a positive PCR test is not confimation of anything. Unless you’re simultaneously testing for a lot of different viruses, and all the other tests are negative, you can’t say for sure that the symptoms are due to covid, even if the person has a positive PCR test (especially if you are considering the result positive even if the cycle threshold is high). Second, say 1 in 1000 can get re-infected within one year, just as an example, that would still mean 999 out of 1,000 can’t, which would mean pretty good immunity at the one year point. So anecdotal stories really mean very little. Third, if immunity is so bad that people aren’t even immune for one year after infection, then vaccination is largely pointless, because the vaccine is unlikely to provide better immunity than real infection.
Also, the other four known corona viruses still being among us is giving a short immunity . Why would cov2 be different?
Sebastian, something I really have tried to find an aswer to is, what viruses except cov2 will put a heavy pressure over your chest as well as pain in the chest, very hard to fill your lungs and extreme fatigue for 3 weeks plus without any fever and the lungs sound normal, also without giving more then normal cold symptoms on the kids?
Thatś what Ím afraid of Sebastian, that immunity is less then a year. We know that people with antibodies can get reinfected within 6-8 months, to many reports of that.. The question is of course how common that is. I certainly hope you are right!
Hi. I just saw an interview wuith Sucharit Bhakdi in
Around minute 18 or so
So what do you think of the long term T cell Studies from SARS in 2003 that show immunity 17 years later?
Titled: SARS-CoV-2-specific T cell immunity in cases of COVID-19 and SARS, and uninfected controls”
Great article in Nature magazine showing T cell immunity for SARS-CoV-2 from the 2003 SARS-CoV-1 spread 17 years ago.
Titled: SARS-CoV-2-specific T cell immunity in cases of COVID-19 and SARS, and uninfected controls”
I remain sceptical about the vaccine and partly that is because I am also sceptical about the benefits of the flu vaccine which has been around much longer and data should be available about it. We have the flu vaccine every year so there is no long term immunity from the vaccine. I also understand that the flu vaccine given is based on the strains expected in each particular year and that could be inaccurate. We hear about mutations/strains of the Covid virus but we are constantly told that the vaccines will still be effective, but with no explanation why. The figures from the Cochrane Institute say that 2% of people not vaccinated for flu are likely to get it and this is reduced to 1% for people who have the flu vaccine. Why bother with the flu vaccine on this basis?
Most people who get Covid have minor symptoms, unlike flu, but surely, the key issue is does the vaccine prevent death. I have seen nothing to say that the flu vaccine prevents death and many thousands die each year with flu and we have no information about it. How many of those dying with flu received the vaccine? I have never seen the question asked, let alone answered.
I am inclined to think it is a money making racket for big pharma.
The effectiveness of vaccine against the flu among the group most at risk of hospitalization and death from it — sickly weak old folk — is estimated by the CDC to have been in recent years as low as 20%.
Re Claire Palmer, first comment.
Ummm, may not be able to repair the thymus gland but there’s LOTS we can do to support thymus function, is your head up with eyes and ears open? Just out of school and have never heard of Thymosin A1?
“In an astounding study the public has not heard about, the subcutaneous (under the skin) injection of just 1.6 milligrams of a thymus gland hormone (thymosin alpha-1) for seven days among patients with severe deathbed COVID-19 coronavirus disease dramatically increases T-cell counts and slashes the mortality rate by 272% (from 30% to 11%) without side effects.”
Re Claire Palmer, continued. Ever learn or did they teach you anything about Zinc? “Zinc supplementation raises thymus hormone levels among zinc-deficient individuals by 5-fold. The indirect way to boost thymosin-A1 is to supplement the diet with zinc since thymosin-A1 is a zinc-dependent string of 28 amino acids (peptide). Zinc supplementation raises T-cell counts.” Even more here: https://knowledgeofhealth.com/dietary-supplements-cured-cancer-covid-19-2/
And ever learn anything about Lactoferrin and immunity? “Lactoferrin belongs to the innate immune system. Apart from its main biological function, namely binding and transport of iron ions, lactoferrin also has antibacterial, ANTIVIRAL, antiparasitic, catalytic, anti-cancer, and anti-allergic functions and properties.” More here: https://en.wikipedia.org/wiki/Lactoferrin#Antiviral_activity
Before you start going off like everybody else about how Wikipedia “isn’t a good source” remember that it’s an ENCYCLOPEDIA, it’s function is to give you the 30 thousand foot overview of the terrain and make you aware of things you might not have known about. And it’s loaded with footnotes citing the academic reference material so if in doubt just search those sources and start reading, your REAL education starts here and now.
Good morning Doctor !
I did go for antibody test after 18 month of infection. The results show 18.86Uml SARSCov2 with ECLIA test. Is it a fair number in case of re-infection and I don’t want to get vaccinated with ARN messenger thinking that is would change my immune response to a possible new attack from covid… What do you think about that ?