Skip to main content

A Real-World Look at COVID-19 Vaccines Versus New Variants

Posted on by Dr. Francis Collins

A woman receiving a vaccine from a doctor
Credit: Getty Images/Andrey Popov

Clinical trials have shown the COVID-19 vaccines now being administered around the country are highly effective in protecting fully vaccinated individuals from the coronavirus SARS-CoV-2. But will they continue to offer sufficient protection as the frequency of more transmissible and, in some cases, deadly emerging variants rise?

More study and time is needed to fully answer this question. But new data from Israel offers an early look at how the Pfizer/BioNTech vaccine is holding up in the real world against coronavirus “variants of concern,” including the B.1.1.7 “U.K. variant” and the B.1.351 “South African variant.” And, while there is some evidence of breakthrough infections, the findings overall are encouraging.

Israel was an obvious place to look for answers to breakthrough infections. By last March, more than 80 percent of the country’s vaccine-eligible population had received at least one dose of the Pfizer/BioNTech vaccine. An earlier study in Israel showed that the vaccine offered 94 percent to 96 percent protection against infection across age groups, comparable to the results of clinical trials. But it didn’t dig into any important differences in infection rates with newly emerging variants, post-vaccination.

To dig a little deeper into this possibility, a team led by Adi Stern, Tel Aviv University, and Shay Ben-Shachar, Clalit Research Institute, Tel Aviv, looked for evidence of breakthrough infections in several hundred people who’d had at least one dose of the Pfizer/BioNTech vaccine [1]. The idea was, if this vaccine were less effective in protecting against new variants of concern, the proportion of infections caused by them should be higher in vaccinated compared to unvaccinated individuals.

During the study, reported as a pre-print in MedRxiv, it became clear that B.1.1.7 was the predominant SARS-CoV-2 variant in Israel, with its frequency increasing over time. By comparison, the B.1.351 “South African” variant was rare, accounting for less than 1 percent of cases sampled in the study. No other variants of concern, as defined by the World Health Organization, were detected.

Graph showing percentages of virus variants. B.1.1.7 is nearly 100% by March
Caption: Changing variant frequencies during the study. Credit: Adapted from Kustin T, medRxiv, 2021

In total, the researchers sequenced SARS-CoV-2 from more than 800 samples, including vaccinated individuals and matched unvaccinated individuals with similar characteristics including age, sex, and geographic location. They identified nearly 250 instances in which an individual became infected with SARS-CoV-2 after receiving their first vaccine dose, meaning that they were only partially protected. Almost 150 got infected sometime after receiving the second dose.

Interestingly, the evidence showed that these breakthrough infections with the B.1.1.7 variant occurred slightly more often in people after the first vaccine dose compared to unvaccinated people. No evidence was found for increased breakthrough rates of B.1.1.7 a week or more after the second dose. In contrast, after the second vaccine dose, infection with the B.1.351 became slightly more frequent. The findings show that people remain susceptible to B.1.1.7 following a single dose of vaccine. They also suggest that the two-dose vaccine may be slightly less effective against B.1.351 compared to the original or B.1.1.7 variants.

It’s important to note, however, that the researchers only observed 11 infections with the B.1.351 variant—eight of them in individuals vaccinated with two doses. Interestingly, all eight tested positive seven to 13 days after receiving their second dose. No one in the study tested positive for this variant two weeks or more after the second dose.

Many questions remain, including whether the vaccines reduced the duration and/or severity of infections. Nevertheless, the findings are a reminder that—while these vaccines offer remarkable protection—they are not foolproof. Breakthrough infections can and do occur.

In fact, in a recent report in the New England Journal of Medicine, NIH-supported researchers detailed the experiences of two fully vaccinated individuals in New York who tested positive for COVID-19 [2]. Though both recovered quickly at home, genomic data in those cases revealed multiple mutations in both viral samples, including a variant first identified in South Africa and Brazil, and another, which has been spreading in New York since November.

These findings in Israel and the United States also highlight the importance of tracking coronavirus variants and making sure that all eligible individuals get fully vaccinated as soon as they have the opportunity. They show that COVID-19 testing will continue to play an important role, even in those who’ve already been vaccinated. This is even more important now as new variants continue to rise in frequency.

Just over 100 million Americans aged 18 and older—about 40 percent of adults—are now fully vaccinated [3]. However, we need to get that number much higher. If you or a loved one haven’t yet been vaccinated, please consider doing so. It will help to save lives and bring this pandemic to an end.


[1] Evidence for increased breakthrough rates of SARS-CoV-2 variants of concern in BNT162b2 mRNA vaccinated individuals. Kustin T et al. medRxiv. April 16, 2021.

[2] Vaccine breakthrough infections with SARS-CoV-2 variants. Hacisuleyman E, Hale C, Saito Y, Blachere NE, Bergh M, Conlon EG, Schaefer-Babajew DJ, DaSilva J, Muecksch F, Gaebler C, Lifton R, Nussenzweig MC, Hatziioannou T, Bieniasz PD, Darnell RB. N Engl J Med. 2021 Apr 21.

[3] COVID-19 vaccinations in the United States. Centers for Disease Control and Prevention.


COVID-19 Research (NIH)

Stern Lab (Tel Aviv University, Israel)

Ben-Shachar Lab (Clalit Research Institute, Tel Aviv, Israel)

NIH Support: National Institute of Allergy and Infectious Diseases


  • Steve White says:

    In the US, the South African B1351 and Brazilian P1 variants were reported at about the same very low levels in January 2021, .1%, but now, both have grown in prevalence – B1351 to .9%, but P1 to 3.5%.
    If these big differences are not due to sampling errors, P1 is the variant, (well,ONE of the variants) we have to worry about.

    Reported vaccination rates by state do not yet vary much – and yet, some states have very tiny numbers of cases – California for example – while nearby states have growing numbers. There has been a huge outbreak of P1 in British Columbia, and growing numbers of Covid cases Washington and Oregon. I can not find data on what variants the WA and OR cases are caused by, but it is starting to look reasonable to fear they are P1 and it is evading vaccines. Or, hopefully, there were not enough vaccines in arms until recently.and they will work. But right now, it looks like no one should get very complacent.

    If I could suggest something to NIH – about 30% of US population has been infected – many know it – many believe they do not need a vaccine because of it – it might be good, after we confirm which vaccines work best against the variants we end up with, to message those folks that their previous infection will not protect them?

    • Mike H says:

      I think Dr. Collins did a good job of covering this concern in last weeks post:

      • Steve White says:

        Thanks for the link. I am not trying to be a pain, and a layman may worry about the wrong things, but the basis for claiming our US mRNA vaccines work against the P1 variant, as best as I can tell, seems to be test tube experiments showing binding of antibodies produced after vaccination with Sars CoV2 pseudoparticles.

        The experiments were comforting as they showed a lot of reaction, though hugely reduced from the variants the vaccines were designed for, but what if the antibodies are NOT actually neutralizing? I do my best to understand but end up unsure.

        The US Federal Government (CDC, White House? I do not know) decided to apportion the vaccines by state population, as they became available. This was criticized as being not a good way to fight an epidemic – perhaps it was better to vaccinate heavily at the sites of outbreaks. It seems to me it also lead to getting less data back. If the US had concentrated vaccinations in variant outbreak areas, we might have data back on how well those vaccines worked against the variants.

        Certainly, with a huge reported outbreak of P1 in British Columbia, we should have sent our vaccines up there to see how they did.

        I just do not feel confident we know for sure.

        • matt says:

          Thanks a lot Steve for the information. I got a trip going to Sao Paulo, Brazil next weekend, and I am really struggling if I should cancel this trip. It is a pretty important trip for me, but I’d cancel it if it will risk my life. I am fully vaccinated (2 doses of pfizer a couple of months ago). I have been searching and researching, and lots of information contradicts to each other. Basically, I want to know if pfizer can prevent serious illness/death. I know it still has some chances to get infected after fully vaccinated with 2 shots of pfizer as long as I don’t get serious disease/hospitalized. Lots of people say it’s 100% protection from serious illness/death. Pfizer’s CEO claimed that no variants found so far can escape pfizer two days ago.
          What’s your opinion? will someone still be possible to get covid, become seriously ill or even die after two shots’ pfizer?

          • Steve White says:

            I just can not find definite information on this. I am not a scientist or doctor and only know what I read online, but what I see does not prove efficacy. I will look around some more and make another post, IF I find anything definite.

  • Catherine Hofmann says:

    I have heard of a few people who have suffered from shingles within a week after receiving the J&J vaccine. Has anyone else heard of this happening?

    • Chelsie B says:

      There was a small study on shingles following covid vaccine out of Israel (though I think it wasn’t J&J) for individuals w/ rheumatic/autoimmune disease

  • Patrice C. says:

    I don’t find it surprising that people either suffer from another viral infection or a dose of COVID after a first vaccine dose. The immune system is challenged by this. I find the same thing after a flu shot or any time my immune system is challenged, such as after a long flight. I usually get a cold sore or a cold of some sort. So shingles or any other latent virus is opportunistic. Lower the immune system and expect something to happen, depending on the person’s state of health, stress levels, genetics and environment.

    • Steve White says:

      I have a different theory about that. Consider a person who never had some viral disease and has no learned immunity to it being infected with the disease very recently – and another person who never had it, has no immunity, is recently infected, but also got a vaccine recently.
      Both of them have a small number of antibodies, but in the case of the person who was not recently vaccinated, all the antibodies the person can make are being directed at the still relatively small number of virus particles in their body. They are sick but fighting the virus effectively.
      But if they had been vaccinated at the wrong time – they put out antibodies, but some very high portion of their antibodies attach to the antigens in the vaccine – in the case of mRNA, made by the body but does not matter, they are still decoys, and the body is wasting it’s antibody “ammo” on them. So the disease progresses much faster and further, and that is observed as the increase in infection rates among those recently vaccinated. I was extra cautious around the time I took the shot, and after, worried about this effect.

      • Gregory S. says:

        The principle of Occam’s Razor applies here: The simplest explanation is far more likely to be the correct one. The simplest explanation is that people get Shingles – period. The same frequency of people getting shingles should be evident in people who are unvaccinated to Covid-19 as are vaccinated to Covid-19, and therefore, there WILL BE (not may be) people getting shingles on the same day as a Covid-19 vaccination, 2 days later, a week later and two weeks later. People get shingles. End of mystery. There would have to be evidence of an INCREASED RATE of getting shingles after Covid-19 for there to be anything to explain. Absent such evidence, put your personal experience in the VAERS database and wait for enough others to do so to see a pattern, but do not talk about personal experience as evidence of a pattern. That’s irresponsible. That’s what anti-vaxers do with people who die after getting vaccinated, use it as evidence that vaccinations cause death, when in fact, the death RATE goes DOWN after vaccination, when you control for pre-existing conditions and age.

  • T Kwan says:

    While there is no information on the breakthrough rate of the P.3.5.1 variant with respect to mRNA vaccines occurring two weeks after the final dose, do researchers know if B cell and T cell involvement in neutralizing the variant is effective, especially with the Moderna vaccine since it showed less cell mediated response after vaccination? Most of the experiments as to vaccine neutralization of variants are done with blood of vaccinated persons and pseudoviruses, but would the T cells in such blood samples still be intact and viable in such experiments?

  • John Hasty BS MT(ASCP) Retired says:

    Since the Spike Protein has been found to damage blood vessels it follows that the ideal vaccine might be one that provides a broad spectrum immune response without the Spike Protein antigen. This could be accomplished using the Synthetic Attenuated Viral Engineering (SAVE) platform . . . They could simply de-optimize the Spike Protein codon pairs in addition to others if necessary to produce their live attenuated nasal vaccine against COVID-19.

  • Bradley says:

    Alcohol consumption is a cultural aspects and there is much more per capita alcohol consumption in some places than others. This brings up a relevant factor in terms of vaccine ADME for covid. NEJM articles on EUA enabling trials do not go into depth whether ALT/AST functions are monitored and for how long? Once the mRNA is taken up by cells, what is the half life of expression? These are after all modified mRNAs most likely cleared through the liver.

    Given the dose differences between Pfizer and Moderna shots and listed side effects, what is the difference in molarity between the two as modifications and PEGylation will effect molecular weight. Are they equivalent? If someone were to get noticeable side effect from the first shot, does it not make scientific sense to lower the 2nd dose? NEJM would suggest multiple doses were checked and there is no reason to suspect that a 100ug dose in 200lb male is equivalent to a 50ug dose in a 100lb woman. Most of the serious adverse events reported have been in women. Doesn’t a weight dependent dosage schedule make sense? Isn’t one being tested for pediatric indications?

    In terms of off target effects, is heart rate/blood pressure being monitored? And for how long? After all ACE2, which the spike protein targets, places a critical role in that aspect. For an older person may be less of a concern given projected life span. For a 12 year getting a Pfizer jab now, these may be rather salient points for consideration. As for off-target effects of neuropilin, that an altogether different headache (no pun intended).

    • Betty Darner says:

      If the trials dosed by weight, the vaccine would have to be given to the public by weight also, right? No one asked me what my weight was for any vaccine: COVID-19, Flu, or Shingrix.

  • Steve White says:

    Just my own speculation, and I have no knowledge of virology or immunology, but as far as B1117 being more likely to infect people with one vaccine shot than people who are unvaccinatd:
    After vaccination, a person has some tremendous number of antibody targets circulating in his blood. Far more than he has antibodies, at least at first.
    After natural infection, for some period of time, a person has a very, very low number of virus particles circulating, compared to the antigens from a vaccination, so even with very small numbers of antibodies, at first, there may be enough antibodies to counter the infection, and stop it such that it never becomes detected.
    But if the person gets infected, and vaccinated, at about the same time, nearly all his antibodies may attach to the vaccine generated antigens – leaving more virus particles unimpeded to make him sicker, until the infection becomes detectable.

    Since the first dose of Pfizer induces a lot of immunity, this effect would not be seen around the time of the second dose – the immunity provided by the first dose is enough to suppress a concurrent infection, even with the added burden of more antibody targets from the second vaccination.

    This may all be wrong, but it is very interesting to see it in a study, as I was worried about this as a possible effect when I got my first shot and was extra cautious not to get infected at the time.

    • Alan says:

      Immunity and inflammation are a fine balance. Initial exposure is different from subsequent exposure when there is something called class switching of antibodies from primarily IgM to IgG (which confers longer term protection). This takes time from repeated exposure. Which is why how long an antigen is expressed is relevant. In case of Pfizer/Moderna the mRNA is expressed for a certain period of time as you correctly point as a bolus shot. For J&J vaccine, given the the history of viral vectors used, that expression may be for a longer period, which is why the J&J shot is a single shot.
      Merck, who discontinued their covid vaccine program given efficacy, has relevant data from viral vectors used for HIV vaccines. Something that a lot of non-profits in the sector should be well aware of. As for mRNA, in and of itself, it can be immunogenic. Since it’s encapsulated in a proprietary lipid formulation, that can dictate what is a safe level in terms of immunity/inflammation balance.
      Now like any complex scientific endeavor, entities that do a shoddy job in any part of the vaccine manufacturing can be putting people at risk for autoimmune or perhaps higher risk of getting infected. The part that is relevant here is whether the infection then results in life threatening or life long complications.
      I think we should all be prepared that covid is here to stay. Now if the equation is such that it keeps people out of the hospital, it becomes as common as a rhinovirus that causes the common cold.

  • Mirta H. says:

    I am a person with autoimmune disorder and treated with monoclonals. Until now there is no information regarding vaccines and their use in these cases. Can the vaccine interfere with the immune and medication status?

Leave a Comment