Skip to main content

Study Finds Nearly Everyone Who Recovers From COVID-19 Makes Coronavirus Antibodies

Posted on by Dr. Francis Collins

Credit: NIH

There’s been a lot of excitement about the potential of antibody-based blood tests, also known as serology tests, to help contain the coronavirus disease 2019 (COVID-19) pandemic. There’s also an awareness that more research is needed to determine when—or even if—people infected with SARS-CoV-2, the novel coronavirus that causes COVID-19, produce antibodies that may protect them from re-infection.

A recent study in Nature Medicine brings much-needed clarity, along with renewed enthusiasm, to efforts to develop and implement widescale antibody testing for SARS-CoV-2 [1]. Antibodies are blood proteins produced by the immune system to fight foreign invaders like viruses, and may help to ward off future attacks by those same invaders.

In their study of blood drawn from 285 people hospitalized with severe COVID-19, researchers in China, led by Ai-Long Huang, Chongqing Medical University, found that all had developed SARS-CoV-2 specific antibodies within two to three weeks of their first symptoms. Although more follow-up work is needed to determine just how protective these antibodies are and for how long, these findings suggest that the immune systems of people who survive COVID-19 have been be primed to recognize SARS-CoV-2 and possibly thwart a second infection.

Specifically, the researchers determined that nearly all of the 285 patients studied produced a type of antibody called IgM, which is the first antibody that the body makes when fighting an infection. Though only about 40 percent produced IgM in the first week after onset of COVID-19, that number increased steadily to almost 95 percent two weeks later. All of these patients also produced a type of antibody called IgG. While IgG often appears a little later after acute infection, it has the potential to confer sustained immunity.

To confirm their results, the researchers turned to another group of 69 people diagnosed with COVID-19. The researchers collected blood samples from each person upon admission to the hospital and every three days thereafter until discharge. The team found that, with the exception of one woman and her daughter, the patients produced specific antibodies against SARS-CoV-2 within 20 days of their first symptoms of COVID-19.

Meanwhile, innovative efforts are being made on the federal level to advance COVID-19 testing. The NIH just launched the Rapid Acceleration of Diagnostics (RADx) Initiative to support a variety of research activities aimed at improving detection of the virus. As I recently highlighted on this blog, one key component of RADx is a “shark tank”-like competition to encourage science and engineering’s most inventive minds to develop rapid, easy-to-use technologies to test for the presence of SARS-CoV-2.

On the serology testing side, the NIH’s National Cancer Institute has been checking out kits that are designed to detect antibodies to SARS-CoV-2 and have found mixed results. In response, the Food and Drug Administration just issued its updated policy on antibody tests for COVID-19. This guidance sets forth precise standards for laboratories and commercial manufacturers that will help to speed the availability of high-quality antibody tests, which in turn will expand the capacity for rapid and widespread testing in the United States.

Finally, it’s important to keep in mind that there are two different types of SARS-CoV-2 tests. Those that test for the presence of viral nucleic acid or protein are used to identify people who are acutely infected and should be immediately quarantined. Tests for IgM and/or IgG antibodies to the virus, if well-validated, indicate a person has previously been infected with COVID-19 and is now potentially immune. Two very different types of tests—two very different meanings.

There’s still a way to go with both virus and antibody testing for COVID-19. But as this study and others begin to piece together the complex puzzle of antibody-mediated immunity, it will be possible to learn more about the human body’s response to SARS-CoV-2 and home in on our goal of achieving safe, effective, and sustained protection against this devastating disease.


[1] Antibody responses to SARS-CoV-2 in patients with COVID-19. Long QX, Huang AI, et al. Nat Med. 2020 Apr 29. [Epub ahead of print]


Coronaviruses (NIH)

NIH Begins Study to Quantify Undetected Cases of Coronavirus Infection,” NIH News Release, April 10, 2020.

NIH mobilizes national innovation initiative for COVID-19 diagnostics,” NIH News Release, April 29, 2020.

Policy for Coronavirus Disease-2019 Tests During the Public Health Emergency (Revised), May 2020 (Food and Drug Administration)


  • S. says:

    Thank you for continuing to highlight the science in this ‘evolving’ story. I would be interested to see more on comparing immune responses by age group. Can the skewing of fatalities towards the aged be understood in terms of the robustness of the immune response (declining in the elderly)?

    • John Fischer says:

      Does having a Cytokine Storm (mainly IL – 6) triggered by CoVid -19 produce the same or different types of antibodies, quantities thereof and immunity than CoVid – 19 patients that did not have a Cytotokine Storm?

      • Steveno Maconovichskya says:

        No as no complete studies have been done yet between CS and apparent Non CS patients on the onset seems like there are no major differences in initial IgG/IgM antibodies

  • Dr Victor M Shorrocks says:

    Antibody In Covid-19 patients

    In your study was the selenium concentration in the blood determined?
    If so what were the Se concentrations?
    If not could the selenium concentrations be measured now?

    • Karen Pentek says:

      What specific role does selenium play in the immune system and in COVID-19?

      • Victor M. Shorrocks says:

        The following details the chronology of referenced work linking selenium to virus occurrence and virulence, including immune function
        1970s Selenium supplementation shown to be effective on several thousands of people in China in controlling the often fatal Keshan disease.
        By 1980 Coxsackievirus (B4 or B3) implicated as an associated/contributory factor in causing Keshan disease

        By 1984 It was accepted that selenium deficiency predisposes people to viral attack, possibly by impairing the immune function.

        By 1987 High risk of HIV-related mortality is associated with selenium deficiency in the USA.

        By 1989 Incidence of infectious hepatitis in several thousands in intervention studies in China reduced by supplementation with selenium supplied in salt.

        By 2001 Influenza A strain (H3N2) was more virulent in mice raised on a selenium deficient diet.

        By 2001 Both Coxsackievirus B1 and an Echovirus 9 caused most heart damage in selenium deficient mice.

        By 2002 An increase in selenium intake improves immune function and clearance of poliovirus in adults with marginal selenium status in England.

        2020, April Claim that the only people that can be infected by the Covid-19 coronavirus have less than 98.7 µg/L of Selenium in plasma or serum.

        2020, April Selenium status was linked to the occurrence of Covid-19 in China; the better the selenium intake and status the less disease

    • Mary Hanson says:

      Has a study shown the immunity to last more than 30 days??? Dr Francis Collins post seems misleading. Many people have tested positive again after they recovered. I’ve had it twice. Thank God for memory T cells it was milder the second time. I feel that priority should be given to re-test known recovered patients at 30 days, 60 and 90 days. I think NIH should not give false hope to people. Transparency and honesty is what is needed more than false hope!! It is insanity that this virus has been known for 4 months and we don’t have a result of immunity studies at 30, 60, 90 days. People are opening their eyes to the lack of honesty and lack of transparency surrounding this virus.

      • O. says:

        Its only natural that the medical community did not waste resources on something that was not expected to happen. Its only in last month or so that evidence accumulated that tracking relapses is necessary.

        So call it complacency based on the experience that only a few diseases like malaria have significant relapse without many months or years for antibodies to be forgotten. Plus of course medical facilities are flooded. Worse is that hospitalized cases often take 2-6 weeks to recover or even longer.

        Unfortunately its now emerging that its probably not just false negatives and that a significant number of people even die during a relapse. Not all relapses actually being tested negative but often being thought close to release from hospital when a wild reverse occurs.

      • Observation2020 says:

        This is the one CRITICAL practical use for an Antibody test — to check the length of time specific antibodies are present in the blood. There are rare diseases which the immune system quickly forgets or does not sustain effective levels long in some people. Most diseases only need boosters every 2-10 or more years.

        It would be really bad if COVID-19 is quickly forgotten in some people. A normal vaccine would not work. Instead those people would need regular vaccine shots (every 20-30 days?) until the pandemic was over.

        Worse if almost everyone lost antibodies in less than a year. We would be giving everyone continual vaccines shots for years.

        Unfortunately this could be one of those genetic quirks that could be related to race or even families. Nature is not governed by laws of equality. Hopefully any problem is not so isolated to small/poor groups that the world forgets to help them nor so widespread that the world runs out of resources and trained people to fix it for all effected.

      • Scott Duncan says:

        There are 2 different mutations, really more, but 2 that are different enough that may allow you to be infected twice.

      • IgG IgM says:

        Mary, if you can provide evidence of having had it twice, you would become the first scientifically proven case of re-infection. Did you do -twice- a high quality antibody test you can share with the scientific community? So far more than 20 million people have gone through the virus, surely we should be able to see at least 10,000 people re-infected? (and that would still be a 0,05% reinfection rate). Hence we can say that as humans enjoy at least 2 months of immunity after developing the antibodies, and every day that passes, that period is extended with millions of data points to validate this (in comparison a vaccine is only tested on maximum tens of thousands of people).

      • Casual Observer says:

        False hope implies that the hope level is actually zero on a scale of 0-10… I think we can safely derive a small amount of hope because typical immune system behavior is that if there are antibodies, then a subsequent encounter with the pathogen should be less severe. Therefore, I do not believe the article is perpetuating “false hope” and I think the article manages expectations at the end.

  • Lisa Phillip Rimland says:

    A hope-giving article!!

  • Steve White says:

    I have been reading up, as a layman, on the concerns and known problems with getting a vaccine out as soon as possible.
    Based on pre-print papers, it appears the worries about Antibody Dependent Enhancement are overblown for many of the vaccine candidates.
    I really hope the FDA, CDC, every other government agency, and the attitudes of the science and medical community will adapt to this unique situation by allowing large scale testing of human volunteers, even including deliberately infecting them with the virus, to combine Phase 1, Phase 2, and Phase 3 testing and complete all 3 stages for the most reasonable candidates by the end of the summer.
    This seems doable, Nature’s “Pick of the Coronavirus Papers” highlighted a study where the RBD – Receptor Binding Domain of the S, spike protein of the virus was given to rats, (not the best model from what I read but a start) and they did not find Antibody Dependent Enhancement, and did find “neutralizing antibodies” – as I understand it, it would be better to prove the antibodies were neutralizing by doing the same study in an animal which gets sick from the virus naturally, but this seems a good start.

    I see about 30 vaccine studies are directed towards the RBD per media reports.

    It may be the case that one vaccine, RBD with the same adjuvant they used in this study, is already known – do exactly the same thing to humans, good chance they will get immunity and not have ADE.

    But then I read it will be 18 months before a vaccine because of all the testing.

    There seems to be no way widespread aggressive human testing will harm more people than allowing the pandemic to go forward for 18 months will.

    I hope the Director will seriously consider changing the rules, radically, to get all the good candidates tested in a few months, I mean totally tested, which will mean many thousands of volunteers taking serious risks the test vaccines will do them harm, and may even involve (after antibody testing to see if they have antibodies, and, hopefully, if any of those antiboides might cause ADE) directly infecting volunteers identified as being at low risk of severe infection.

    Please Mr. Director, take these actions. We can not stick with a cautious, “First, do no harm” paradigm, the harm can not be avoided by cautious action.

  • Jerry says:

    A well-written article which clarifies the difference between a diagnostic test (determining if the virus is present) and a serology test (determining if the virus HAS BEEN present in the body). The next big question to answer is whether immunity prevents re-infection and for how long. I’d also be interested to know if the virus remains latent in the body like chickenpox or Eppstein-Barr, and could flare up much later.

  • Elizabeth K says:

    I tested positive for the CORONAVIRUS on the 8 April when I left my husband in the ER to go on the Vent. I remained in my home and away from everyone, I received another test on May 01, I tested positive once more. I am having another test tomorrow. If I am positive after my test tomorrow what do I do? My husband today is 5 1/2 days off the ventilator.

    • Jessica Barrett says:

      PCR tests amplify viable, as well as nonviable viral particles and cannot determine if the viral particles are viable or nonviable. Most people shed viral particles for 3 weeks.

  • Ed Krasinski says:

    1st. It’s autoimmune triggering capabilities.

    2nd. More importantly. Is its ability to trigger in vivo synthesis of chemical compounds that act like venom.
    Are the scientist studying these ?

    • Observation2020 says:

      Actually any virus or cancer has the potential to cause cells to produce toxic byproducts. In both cases the normal cell DNA is broken so something abnormal is going to be produced.

      For a virus those toxins are a consequence of hijacking DNA regulated processes for reproducing the virus. There is almost “virus industrial waste” left over. In contrast to the normal DNA processes where evolution has made sure every DNA regulated product has its on place in the body chemistry or there methods to excrete such wastes in a controlled manner.

  • Peter Sommer says:

    We know that an antibody response is detectable within 14 days since Drosten’s team described the first patient cohort from Bavaria. Nice to see a confirmation of this findings, but what about the neutralization potential of the humoral response? Once you have the serum, it’s a quick assay to do isn’t it?

  • Selena Johnson says:

    This information is highly appreciated. Precautionary, healthy immune etc., everything is a must also these type of infos should reach people at the right time.

  • William B Lawson says:

    These are important findings. The question is not just whether the antibody response is protective but whether as some studies show it may be destructive

  • Margaret Bradbury says:

    The article by Dr. Collins states “Though only about 40 percent produced IgM in the first week after onset of COVID-19, that number increased steadily to almost 95 percent two weeks later. Some of these patients also produced a type of antibody called IgG ” My reading of figure 1a is that 100% of the tested subjects were eventually IgG positive. This is not the same as the statement “some of these patients produced…. IgG Can this be clarified?

  • Observation2020 says:

    Presence of antibodies does not mean you are not still infectious and needing treatment. Worse you can be infected and infectious without antibodies being present at detectable levels. Antibodies can take days to develop and then many long
    hours more to ramp up in numbers.

    So anti-body tests are not that useful for quarantine or treatment. Really only useful as a history statistic to describe the struggle in one area and general abstract expectations of similar struggles. *** So the most optimistic practical use in actual COVID-19 health care is to regulate dosages of effective non-vaccine treatments. *** But having the time and patient numbers to develop such an in-depth knowledge over a broad enough population to become useful would say really bad things about defeating the pandemic.

    Are there diseases where you get well without anti-bodies? That would be more of a discovery.

    Its already known that only diseases that kill the host very quickly or are killed off quick/easily by existing anti-bodies don’t get a customized reaction eventually.

  • Melissa says:

    I had the virus it started middle of March. I took care of myself as I didn’t get deathly ill. I got over it mid April. Can I still be tested for antibodies? If so can I give blood to people who need it most?

  • Dr. Grace says:

    I’d like to know more about the possibility of antibody dependent enhancement in this specific corona virus. It sounds like there may be some suspicion that the antibodies may all not necessarily be neutralizing antibodies. In that case the virus makes the helper antibodies. No one Is talking about this.

  • Osman (Ozzie) Ahmed, MD says:

    The national response to the Coronavirus pandemic has been fragmented and inadequate. As a family physician and epidemiologist, I find it mystifying that our efforts are haphazard and uncoordinated. We live in America, a continent of 48 contiguous states (=countries) and 2 non-adjacent (HI, AK) states. Any attempt to contain the virus must include instituting rigorous interstate unnecessary travel and movement controls, especially in and out of hard-hit states/ communities. We live in a free mobile society and unless we curtail the spread by limiting the importation/ exportation of cases, we will experience successive outbreaks in various localities at different times making the virus endemic in our country with significant loss of lives and resources. Our states (=countries) have no borders, the virus has no borders and we have to have the best measures in place or we will fail even with social distancing, contact tracing, case identification, testing, isolation and treatment. Leadership matters. We must develop and implement a coordinated interstate and intrastate “antiviral” plan. We ought to draw to the international community a road map for success and recovery while we wait to develop natural or acquired immunity. This is the only way to save precious lives.

  • Cris says:

    Hi, i’m exposed to Covid19. My one roommate has it and pass it on my other roommate and all in our house. They all infected but not me. My test came negative. I did not feel any sickness or symptoms compare to them. Am I immune to it?

  • Paul Han says:

    These tests have being validated on HOSPITALIZED patients, but the major application would be in patients who are in the COMMUNITY. Non-hospitalized patients would may not have the same immune response as sicker, hospitalized patients. These tests should be validated in the community. The sensitivity may not be as high as advertised. When accuracy is reported, the demographics of the test group should be emphasized. “Nearly Everyone [Who Was Hospitalized For COVID-19]” (in China).

  • Greg Miller says:

    Given that we are now coming off of “flu season” in North America, I have a question relating to the the accuracy of current IgG and IgM antibody test kits in determining whether someone has has been infected with SARS CoV-2, vs any number of alternate infections. I was reading the sales brochure for the Abbott IgG/IgM test kit and it stated the following disclaimer: “Positive results may be due to past or present infection with non-SARS-CoV-2 coronavirus strains, such as coronavirus HKU1, NL63, OC43, or 229E. It is not known for certain whether individuals infected with SARS–CoV-2 who subsequently recover will be protected, either fully or partially, from future infection with SARS–CoV-2 or how long protective immunity may last”. From what I can tell, coronavirus OC43 and 229E are nothing more than the common cold, and other than having an outer envelope and single RNA strand, share little else in common with SARS CoV-2. So my question to the NIH is as follows: Can the IgG/IgM test kits currently on the market identify antibodies that are related specifically to SARS CoV-2, and if so, what are these antibodies and their sub types? If not, how is this type of testing useful in determining whether someone has developed immunity specific to SARS CoV-2? Please advise.

  • Robert G says:

    I tested positive for COVID-19 almost 30 days ago, went through classic symptoms, and was tested today for antibodies. The results are negative for antibodies. How can this be?

    • Kwaaikat says:

      Anti body tests are calibrated to never give false positives, so it errs on the side of sometimes giving false negatives.

      Another possibility is cell mediated immunity, which for almost every disease at least some people will develop as an alternative strategy to antibodies. Such immunity does not produce antibodies.

    • Pete says:

      I read the other day that anecdotal evidence suggests that although severe (hospitalized) cases develop antibodies, only about 10% of people with mild symptoms do. If this turns out to be true, it would not bode well for those insisting we go ahead and get it to develop herd immunity. Perhaps yours was a mild case?

      • kwaaikat says:

        You are right it would not, if that is true.

        But the fact that infections in all territories where with cases in Feb, has slowed and peaked after no change in restrictions, suggests that some level of immunity in the wider (non serious case) population is definitely growing.

        It would be interesting to read the report if you still have the link.

        One way to settle that question is to test all passengers who were on board the Diamond Princess, which tested nearly everybody on board, so practically if not all Covid-19 cases (mild and asymptomatic) are known. If all these known cases (half of who were asymptomatic) have antibody tests done now, we will have an idea.

  • GNB says:

    Coronavirus has affected all the world. The higher class people are not much affected by the middle and lower class are badly affected. The treatment of COVID19 is just to stay home and stay safe.
    But the small business will deal with this situation and for their livelihood as well?

  • 1 2 3

Leave a Reply to giorgiCancel reply