Overwhelming evidence now that previously infected have robust immune protection against reinfection

Key points of this article:

• An article was published in Israel demonstrating those previously infected with PCR confirmed SARS-CoV-2 infections have similar immune protection as those vaccinated.

• This evidence follows a flurry of papers that have suggested that reinfections are rare and immune protection after recovery from SARS-CoV-2 infection lasts quite long.

• In the USA, this may be surprising to many given the persistent narrative that emerged in early fall 2020 that natural immunity may quickly wane based on anecdotal reports of reinfection and early papers suggesting antibody levels quickly declined, and perhaps fed by the political climate at the time.

• These results provide support for the Israeli strategy to prioritize vaccination on those not previously infected, and suggest we should also include those previously infected in any calculations estimating our progress towards herd immunity, and any places choosing to require "immunity passports" should consider documented previous infection, not just vaccination, to be sufficient evidence for immune protection.

• Recent papers have found that previously infected individuals still receive immune benefit from the first shot that effectively serves as a booster that may prolong protection, but that the second dose does not appear to confer any further benefit. This suggests previously infected people to still be vaccinated, but may impact future guidelines about whether both doses should be given or not.

A research group from Israel has posted a paper on medRXiv looking at population-level SARS-CoV-2 infection rates across the entire country, using their excellent centralized electronic medical records, for a 3 month period between December 20, 2020 and March 20, 2021, with two primary results:

First, comparing SARS-CoV-2 infection rates among those fully vaccinated (assumed to start 1 week after 2nd dose of Pfizer vaccine) with those who were not vaccinated at all nor previously infected, they found the vaccines reduced risk of PCR-documented SARS-Cov-2 infection, symptomatic or asymptomatic, by 92.8%, of hospitalization for COVID-19 by 94.2%, of severe COVID-19 by 94.4% and death by COVID-19 by 93.7%, validating that the reported vaccine efficacy from the phase 3 clinical trial is realized in the population-based setting.

Second, among those not yet vaccinated, they found a similar level of protection in those who were previously infected (verified by a SARS-CoV-2 PCR test) relative to those not previously infected, with risk of infection reduced by 94.8%, of hospitalization for COVID-19 by 94.1%, and of severe COVID,-19 by 96.4%. This demonstrates robust immune protection for previously infected on par with full vaccination.

These results are consistent with insights that have steadily emerged from the scientific literature over the past 6 month: that the vast majority of people recovered from confirmed SARS-CoV-2 infection retain substantial robust immune protection, this protection does not quickly wane, and reinfections are exceedingly rare. This may be surprising to some, as many in the USA seem to have the impression that reinfections were relatively common and natural immune protection might rapidly wane on the basis of reports early in the fall of 2020. In this post, I will first briefly summarize this study and explain why I think it is trustworthy and based on rigorous scientific procedures, and then I will look back and describe what I believe are major political and logistical reasons why the USA has been slow to acknowledge the protective immunity conferred by prior infection in spite of emerging evidence from a string of literature that have suggested this fact even before the present paper. Next, I will discuss implications of these results for vaccination and herd immunity, and highlight why immunologists still believe it is beneficial for previously infected individuals to get at least one dose of vaccine, and why for those not yet immune protected it is much safer to obtain such protection from vaccination rather than infection.

Summary of Study

As I mentioned in previous blog posts here, here, and here, the centralized Israeli health care system with rigorous electronic medical records (EMR) containing all medical information for all Israeli residents affords outstanding opportunities for population-level analyses to investigate vaccine efficacy and safety. In this study, the researchers assembled population-level data from the EMRs for every adult (age >=16) in Israel, a total of 6,351,903 entries. They stratified these individuals at each point in time into 4 groups based on vaccination and previous infection status.

1. Unvaccinated individuals without previous evidence of SARS-CoV-2 infection

2. Partially vaccinated individuals (from first dose to 1 week after second dose)

3. Fully vaccinated individuals (from 1 week after second dose)

4. Unvaccinated individuals with previous SARS-CoV-2 infection (confirmed by PCR)

Unlike the USA, Israel did not vaccinate previously or currently infected individuals.

Modeling every day between December 20, 2021 and March 20, 2021, they compared the risk of SARS-CoV-2 infection among these four cohorts. Since this study was observational and not randomized, it was important to adjust for potential confounding factors, since it is possible that there may be systematic differences in the vaccinated and unvaccinated groups that induce differential risk of infection. They adjusted for these potential differential risks profiles by including relevant covariates in the model including calendar time, age, sex, municipality, medical risk factors, and number of past PCR tests (0, 1, or 2+). The analysis had professional statistical support and I found their modeling to be rigorous and reasonable. Although covariate adjustment is not always sufficient to adjust for confounding factors in observational studies, it appears to have done a good job of adjusting for differential age distributions in the vaccinated/not cohorts, based on the fact that their analysis split out by age groups showed consistent results in each age group.

Validation of Vaccine Efficacy: From this analysis, they found full vaccination offered the immune protection described above, with 92.8% lower risk of infection, 94.2% lower risk of hospitalization, 94.4% lower risk of severe illness, and 93.7% lower risk of death. This protection was relatively stable across age groups, but with slightly lower efficacy (85.6%) in the oldest (80+) age group. Note that in Israel, PCR tests were given after reporting symptoms or via contact tracing, so these results include symptomatic as well as some asymptomatic infections.

These results are close to the reported 95% efficacy vs. symptomatic disease observed in the phase 3 clinical trial, and provide a strong population-level validation of the efficacy of the Pfizer vaccine. Note that the time of the study coincided with the rapid emergence of the UK variant B.1.1.7 in Israel, so it is clear that the Pfizer vaccine maintains strong efficacy for this dominant variant.

Previous infection conferred similar immune protection as vaccination: They also found that previous infection conferred similar immune protection as vaccination, with those previously infected but not vaccinated having 94.8% lower risk of infection, 94.1% lower risk of hospitalization, 96.4% lower risk of severe illness, and with only a single death in the previously infected cohort. Again, this protection was relatively stable across age groups, but with slightly lower efficacy (91.4%) in the oldest (80+) cohort. This level of protection is on par with what was provided by full vaccination.

These results may be surprising to many in the USA, given the high level of media attention given to anecdotal reports of reinfections and the persistent discussion of the uncertainty of how long after recovery immune protection lasts, with an implication that it might only last a few months. Unlike Israel, the CDC vaccination guidelines in the USA did not distinguish between those previously infected or not in its vaccination priority, in fact seemingly going out of its way to make no distinction. Also, in his discussion of vaccination proportions necessary to reach herd immunity, Dr. Fauci completely disregarded any level of societal protection provided by those recovered from previous infection who were not vaccinated yet.

There are pragmatic and logistical reasons for this, but as I will now describe, I think that based on early reports of reinfection and waning antibody levels after recover from infection a narrative emerged suggesting immune protection may not last long. Reinforced by the political climate at the time, it seems to me that this narrative has proven persistent and slow to change even as substantial evidence has emerged from numerous scientific studies over the past six months supporting the rarity of reinfection and durability of immunity.

Fear of reinfection, suggestions of rapidly declining antibody levels, and emergence of the Great Barrington Declaration: In late August and early September 2020, the first reports of SARS-CoV-2 reinfections were confirmed and widely reported in the media -- not just cases of long COVID-19 with multiple positive tests but positive tests months apart with mRNA sequencing clearly demonstrating it was indeed a new infection. The first case was reported in Hong Kong in late August, and then another in Nevada, and then over the coming weeks other anecdotal cases were reported in the popular news media.

Contemporaneously, some influential papers were published that suggested antibody levels may wane very quickly after recovery.

One New England Journal of Medicine paper was published on September 10 that suggested antibodies decayed very rapidly after recovery, estimating a half-life (i.e. reduction of 50% towards baseline levels) of only 1-2 months. In October, a Nature Biotechnology paper was published whose primary conclusion was that antibody levels for some returned to baseline within 2 months, and a large British study tracking antibody levels in 365,000 UK residents over 3 months reported that antibody levels declined from 6.0% to 4.4% over three months, concluding from this that immunity may quickly wane and risk of reinfection was substantial. All three of these papers were very influential, with the NEJM paper viewed >260,000 times, the Nature Biotechnology paper >68,000 times, and the British paper >10,000 times as of April 25, 2021, and all highlighted in numerous popular media reports.

On the heels of media reports of reinfection, these papers fed a narrative that natural immunity after recovery from infection might not last very long, and reinfection was a substantial risk.

I also believe the political environment in the USA further fed this narrative. On October 2, 2020, a small group of scientists published The Great Barrington Declaration (GBD), which argued the current mitigation strategies were misguided and suggested an alternative strategy of protecting the older and most vulnerable members of society, but lifting all restrictions on the rest of society and allowing the infection to spread quickly through these younger, healthier people, with the hope of rapidly reaching herd immunity and bringing the pandemic to an end. Epidemiologists, immunologists, and other infectious disease experts were rightfully concerned about the danger of this strategy, whose substantial flaws I outlined in a previous blog post. The strong support of this strategy by President Trump and Scott Atlas, appointed as the lead advisory of his pandemic response team in August 2020, also contributed to the urgency and strength of opposition to these ideas.

In this political environment, the reports of reinfections and papers suggesting rapidly waning immune protection were convenient elements to use in arguments against the GBD, since if reinfections commonly happened in a matter of months, there would be no way to reach herd immunity naturally. It is possible that many people got into such a reflexive habit of citing these reports that they were slow to acknowledge emerging evidence suggesting reinfections were exceptionally rare and immune protection more durable than originally thought as it emerged.

Rarity of reinfections, and laboratory and epidemiological evidence of strong immune protection from previous infection.

Rarity of reinfections: While the anecdotal reports proved that reinfections could happen several months after recovery, even at that time a cursory quantitative analysis of the data would demonstrate that the risk of reinfection must be very low. As I first pointed out in a blog post in late September, it is straightforward to compute a basic estimate of the number of reinfections one would expect if people lost their immunity after a certain period of time. Since it is clear that the actual number of reinfections must a small fraction of this number, this implies that the vast majority of recovered individuals had substantial protection against reinfection for that period of time.

To explain, here I have updated these calculations as of 3/31/2021. For the sake of this exercise, suppose that every person lost their immunity precisely 3 months after infection, and suppose that everyone in the population is equally likely to be exposed and infected.

The table below shows the number of cases in the USA for each month and the percent of the USA population that represents. From this, we can compute the number of expected reinfections based on the infection rates from 3 months after infection until the present.

For example, for the 192,177 infected in March 2020, if they lost their immunity 3 months later, then summing the percent infection rates from June 2020 through March 2021, their cumulative probability of reinfection is 8.73%, which would suggest we should have seen 16,786 reinfections from this group if immunity was lost after 3 months. Summing over time, we see that under the assumptions we would expect a total of >650,000 reinfections in the USA by March 31, 2021, in which both tests were confirmed by positive PCR tests. While we know that not all reinfections are reported and acknowledged, it is safe to say that it is unlikely that there were anywhere near 650,000 reinfections. If one thinks the true number of reinfections was more like 32,500 (which still might be high) then the resulting 95% reduction in infections relative to those not previously infected would imply that the vast majority were protected against reinfection for longer than 3 months. The table also includes similar estimates if people lost immunity after 6 months, and the number of expected infections is >250,000, still far more than one would think has occurred. If the number is more like 25,000, then this would imply a 90% reduction in infections through 6 months relative to those not previously infected.

Now these estimates are admittedly obtained through a very simplistic model assuming all immune protection is lost at a fixed time and everyone in society is equally susceptible to exposure and infection, but it provides an idea of the order of magnitude of reinfections one would expect. Even back in September 2020, it was clear that there were not close to that number of reinfections, and this suggested the vast majority retained substantial immune protection against reinfection. but few considered this statistical argument.

Evidence of long-lasting immune markers: The three studies mentioned above contributed strongly to the narrative that natural immunity was not long lasting, but not much was made of the studies' limitations or their methodological flaws. The NEJM study only looked at 34 subjects, and for most of these they only had measured antibody levels at two time points. They fit a line to these time points, and extrapolated this line to get an estimate of the half-life of 1-2 months. However, this ignores the known shape of antibody curves over time, as illustrated in this figure:

Early on in infection, the immune system produces IgM antibodies that quickly wane after infection. The production of IgG antibodies is somewhat delayed, and these antibodies are the ones that remain over time and provide enduring immune protection. As shown in the green curve above, these IgG levels spike during the peak of infection, and sharply decline for a time before flattening out after an inflection point and potentially enduring for a long time as it slowly declines. This NEJM paper only had data on two time early time points from each subject, and if these two time points were in the rapid decline phase of the curve it is easy to see how extrapolation of that line would exaggerate the degree of decline over time and result in a potentially erroneous conclusion that antibody levels quickly waned. The other two papers also had their own limitations -- the take home message in the Nature Biotechnology paper was that "some decreased near baseline antibody levels by 2 months", but looking at their data for most subjects the levels remained very high and there were only 3-4 of the 65, about 5-7%, whose neutralizing antibody levels decreased to near baseline levels, and these were individuals whose levels were never very high to begin with. This is consistent with the notion that most individuals retained strong immune protection, but a small minority had more rapidly waning protection. And the UK study also had its limitations -- it did not even measure antibody levels in the same individuals over time, it simply sampled a different subset each month and inferred that the antibody levels must be waning simply because the proportion with detectable antibodies slightly decreased over a three month period. This lacks the specificity or rigor of studies looking directly at reinfection or at multiple endpoints per subject. This paper still has not been published in a peer reviewed journal 6 months later.

So these early papers did not provide as much evidence for a rapid loss of immune protection as they seemed, and a subsequent flurry of papers all provided an opposing narrative that immune markers appeared to remain substantially durable over time.

• A paper published in NEJM on 10/29/20 by Gudbjartsson and colleagues looked at 1797 samples and showed that IgG increased at 2 months, and had plateaued at 4 months, following the curve shape above.

• A paper published in Clinical Infectious Disease on 11/5/20 by Peterson and colleagues looked at 2547 health care workers (HCW) in NYC and Detroit and found 93.7% still had detectable IgG levels at 2 months (100% of those hospitalized).

• A paper published in Immunity on 11/17/20 by Ripperger and colleagues followed a group of infected individuals in Arizona and found most had detectable levels of neutralizing antibodies (nIgG) for 5-7 months. Neutralizing antibodies are antibodies that target strategic locations of the virus that are sufficient to prevent the virus from replicating, in the case of SARS-CoV-2 specific locations on the receptor binding domain (RBD) of the spike protein in the virus.

• A paper published in Science on 12/5/20 by Wajnberg and colleagues followed 30,082 patients at Mt. Sinai, and found >90% had evidence of nIgG, and found these levels were still stable on the plateau at 5m post infection.

• A paper published in Lancet Infectious disease on 12/17/20 by Duysburgh and colleagues measured monthly antibody levels for 81 Belgian HCW and found 91% had evidence of nIG at 4m post infection.

When studying immune protection, many overemphasize antibody levels that, although often quite durable, eventually wane over time. Our bodies do not indefinitely retain high levels of antibodies against all antigens our body has been exposed to during our lifetimes -- these levels eventually wane but can be regenerated using other elements of the immune system. The immune system is multifocal, with antibodies playing a central role (especially nIgG antibodies), but with memory B-cells able to quickly regenerate a specific antibody when later exposed to the virus, helper T-cells (CD4+) playing a role in the elicitation of such antibody production by the memory B-cells, and killer T-cells (CD8+) themselves able to destroy viral cells. Thus, as immunologists know, immune protection does not disappear just because antibody levels wane. This makes the following paper especially important

• A research group led by Shane Crotty published a paper in Science on 2/5/21 that investigated all of these elements of the mult-focal immune system, including IgG, memory B cells, helper T cells, and killer T cells. He found that based on this system, >90% of recovered individuals had immune memory through the end of the study, which is at 5 months.

Together, all of these paper provide strong evidence that most recovered individuals showing enduring immune marker levels at least through 5-7 months, with studies showing stable levels that might endure much longer. Data at later time points are necessary to fit the nonlinear curve and accurately predict when immune protection tends to wane. While some individuals showed declining immune markers, typically a subset of those with very mild disease, these papers showed a vast majority had strong and lasting protection.

Epidemiological evidence of protection vs. reinfection: While the Israeli medRXiv paper that motivated this post provides strong evidence that previously infected individuals have reduced risk of reinfection, it is not the first epidemiological study demonstrating such results. Here are a few others:

• Lumley and colleagues published a paper in NEJM on 12/23/20 following N=12,541 health care workers in several USA sites, of which 1265 were previously infected as indicated by positive antibody tests, and were found to have an 88% reduced risk of infection relative to those not previously infected. Because they defined previously infected based on antibody tests, this included asymptomatic infected individuals whose immune protection is expected to be not as strong.

• In the Johnson and Johnson study, 10% of placebo patients were previously infected as indicated by positive antibody tests, and as I discussed in a previous blog post, Table 14 in the FDA report on 2/26/21 shows that these individuals had a 92.4% reduced risk of infection relative to the placebo patients not previously infected, a result I have not seen discussed elsewhere.

• Hansen and colleagues published a paper in the Lancet on 3/17/21 compared >11k PCR positive individuals vs. >525k PCR negative individuals in Denmark and showed those previously infected had a 80.5% reduced risk of subsequent infection (but this benefit fell to 47.1% for individuals >65yr).

• Leidi and colleagues published a paper in Clinical Infectious Diseases on 5/27/21 in which a representative sample of 498 Swiss seropositive (previously infected) individuals was taken and matched one-to-two with 996 seronegative controls based on age, gender, BMI, smoking, immunodeficiency, and education level, and followed through January 2021. They found that the infection rate was much higher in seronegative individuals (15.5%) vs. those previously infected (1.4%), indicating 94% reduced risk of reinfection.

• Pilz and colleagues published a paper in European Journal of Clinical Investigation in April 2021 performing a population based study in Austria estimating reinfection rates from those infected in the first wave (Feb to Apr 2020) during the second wave (Sept to Nov 2020) vs. those not previously infected. They found 40 reinfections among 14,840 COVID-19 survivors (0.27%) and 253,581 infections in 8,885,640 of the remaining population (2.85%), showing a 91% reduced risk of reinfection.

All of these studies include previously infected individuals from many months ago, some almost one year prior. Along with the immunological studies showing immune markers remaining high and stable over time, these studies provide strong evidence that most individuals retain strong and durable immune protection vs. reinfection after recovery from COVID-19, with caveats that a subset of individuals, perhaps older and/or with mild disease or perhaps immunocompromised, may not be as strongly protected, and that few of these studies included many with asymptomatic infections. It is possible that these same groups are less protected by vaccination, as well, so the relative strength of immune protection from recovery from previous infection and vaccination is an open question, and assessment of the durability of immune protection in both settings is pending further data.

Implications for vaccination priority: What are the implications of these results for vaccine prioritization? It seems that Israel has acknowledged the protection afforded by previous infection, since they have chosen not to vaccinate individuals with evidence of previous infection. In the USA, little distinction has been made and all adults, whether previously infected or not, have been strongly encouraged to be vaccinated. In a blog post in early December, I suggested that previously infected individuals not be included in the first priority groups for vaccination citing their existing immune protection. I thought that in the early months of vaccine distribution when supplies were severely limited, this would be an inefficient use of resources, especially during a time of high viral surge as we were experiencing all over the world in December and January.

However, I understood why the CDC did not want to do this. It is difficult enough to logistically arrange deployment of two shots to nearly 300 million residents, much less add a layer of PCR or antibody testing as a preliminary step. Also, given uncertainty about the durability of natural immunity in all subgroups, they may have been wary about implying these people never needed vaccination.

Also, there has been a flurry of recent articles looking at immune marker levels after vaccination, and many have found that the nIgG levels tend to be much higher after vaccination than after natural infection, plus demonstrate the vaccine effectively induces production of targeted memory B-cells, killer T-cells, and helper T-cells. Thus, it may end up being the case that immune protection is even more durable with vaccination than natural infection, especially for emerging viral variants, although longer term studies are necessary to determine whether this is the case or not. This is far from certain.

Interestingly, some of these articles look at the immune markers in previously infected individuals after vaccination. All have found that:

• Previously infected individuals tend to have high levels of IgG, memory B-cells, and T-cells at baseline before vaccination.

• SARS-CoV-2 naive individuals (i.e. not previously infected) don't reach maximum full immune protection until after the second shot of mRNA vaccines.

• Previously infected individuals have considerable immune protection at baseline and reach this maximum level after the first dose, but don't show any additional immune benefit from the second dose. This has been shown to be true for nearly all elements of the multifocal immune system, including neutralizing antibodies, memory B-cells, helper T-cells, and killer T-cells.

These results suggest that perhaps individuals known to be previously infected should only be given a single dose that effectively serves as a booster to ramp up their immune protection to maximal levels. This may be especially true in subgroups who might be at higher risk for serious adverse events from the vaccine, should any such subgroups be identified. Further studies are needed to see whether this is the best long-term strategy. For now, it seems prudent for previously infected people to get vaccinated if they get the chance given the results that the first dose boosts immune levels which might confer more durable immune protection.

Implications for "Immunity Passports": Some places are requiring some type of immunity passports for entry. These results make it clear that individuals recovered from documented SARS-CoV-2 infections should be considered to have immune protection, not just those that have been vaccinated.

Implications for Herd Immunity: The substantial protection vs. reinfection conferred by previous infection has clear implications for herd immunity. As I mentioned above, some individuals ignore this component when assessing the levels of vaccination required for herd immunity. Clearly the proportion previously infected but not vaccinated should be considered along with the proportion vaccinated. While not all infected individuals are confirmed cases by a PCR test, conservatively one could consider just those PCR-confirmed infections in the calculations or maybe double that number, since it seems likely that at least half of SARS-CoV-2 infections are unknown or unconfirmed in most every population.

I think that the term herd immunity has become a bit of a loaded term given the controversy surrounding the Great Barrington Declaration and related proposals, plus I think the term is broadly misunderstood by the public.

Herd immunity is not some threshold above which the virus suddenly and magically disappears from the face of the earth. It is best not thought of as a discrete threshold at all. The concept is simply that as a higher proportion of society is not susceptible to infection, because of immune protection either from vaccination or previous infection, the virus has trouble spreading as quickly because a higher proportion of people it encounters cannot serve as a host to further its spread. It is just a mathematical concept that the base of exponential growth is reduced by the proportion not susceptible to infection, and the threshold is simply the proportion at which the base of the exponent gets below 1, which implies exponential growth is no longer possible and at which point daily case counts enter an exponential decline. Even before this point, the growth is strongly suppressed, and after this point, there is still some viral spread, but it pretty much defines the point when we will no longer see sharp growth curves that produce the strong concern that motivates pushes for more aggressive mitigation strategies. Perhaps population immunity would be a better term to avoid the baggage that has come to be associated with the term herd immunity

Where is the threshold? It is expected to be somewhere around 70-80% for SARS-CoV-2, but this number depends on many factors, and could be higher for faster-spreading variants like the UK B.1.1.7. Given the theoretical threshold is only an approximation of reality anyway, it might be better to look to populations that have substantial vaccination and see what has happened there. Israel: The best example is Israel, who had the most rapid and thorough vaccination deployment in the world. Starting with the >60yr population in December, they have now fully vaccinated (with Pfizer mRNA vaccine) their entire adult population (not including previously infected), 91% of all adults which is 57% of the total population, plus they have another 9% whose previous infections were confirmed by positive PCR tests. Their daily case counts are now down 99.7% from their peak in mid-January, with just 38 new cases and 0 deaths in the entire country as of 4/24/21. It is reasonable to think that they have reached the effective population immunity, and that is without vaccinating any previously infected individuals or children under 16yrs.

United Kingdom: The United Kingdom has also been among the world leaders in vaccination deployment. They have currently vaccinated 60% of their adult population (48% of total population) with at least the first dose (mostly Pfizer mRNA vaccine) plus have another ~7% who have had confirmed PCR infections. Their daily cases counts are now down 98% from their peak in mid-January, and if you compare to other European nations that have vaccinated far fewer (e.g. Germany at 7%) you see this decline is in stark contrast to the surges in those slowly-vaccinated countries.

USA Nursing Home population: The USA prioritized nursing home residents in its early vaccination deployment and quickly vaccinated nearly all residents and workers. By 3/1/21, their daily case counts were down 96%, and likely have continued to decline since then (but I haven't been able to find updated data).

USA: So what can we expect for the entire USA? Currently in the USA, 33% of adults are fully vaccinated and 52% have received at least one dose (mostly Pfizer and Moderna mRNA vaccines), plus we have 10% who have had confirmed PCR infections. Since vaccination is not limited to those without prior infection, if we assume 33% of the previously infected have been vaccinated, this conservatively leaves us at 33%+7%=40% towards the population immunity threshold. At the current rate of full vaccination of 5% per week I estimate that by early June we should be close to the 65-70% non-susceptible proportion that appears to have stopped exponential growth in Israel and the UK. Of course there is lots of uncertainty so we will have to wait and see.

So how much vaccine hesitancy can we handle and still reach population immune protection? We certainly could handle a bit. If the projected 15-20% refuse vaccination, we should be fine and perhaps could handle as much as 30% of the population unvaccinated considering some groups of people who are vaccine hesitant were also the same that resisted compliance with mitigation policies, and thus might be more likely to have protection from previous infection. As long as we can reach 70-75% protected via vaccination or previous infection, I expect we will see a cessation of exponential growth of viral cases in the country.

Conclusions

In conclusion, we now have substantial and overwhelming evidence that previous infection confers strong immune protection for a vast majority, and that this protection appears to last at least 6 months for most, and maybe much longer as we have not really reached a time point yet in which we see the immune protection clearly waned for most. An immunology colleague believes that this protection could last as long as 3-5 years, but acknowledges that we will need more data at later time points to get a better idea.

This should assuage the fears that arose in the early fall 2020 that immune protection may wane too quickly to get the virus under control, and suggests that given a vigorous vaccination campaign, the immune protection from vaccination or previous infection will be sufficient to reach population immune protection and get the pandemic under control.

There are questions about current or future immune escape variants that evade the immunity conferred by natural immunity or vaccination, but given the vaccines' reasonable efficacy for even the worst immune escape variants to date (topic of a blog post to come) and the potential for rapid development of variant-specific boosters, I am optimistic it is possible to get this pandemic under control with vaccine deployment. It is important for wealthy, vaccinated countries to assist other countries in getting their populations vaccinated as well, which could prevent the continued spread of the virus around the world that could lead to the development of more problematic immune escape variants.

For individuals previously infected, it is likely they do not need to have the same urgency to get vaccinated as SARS-CoV-2 naive people, but given the extra immune boost that the vaccine provides and the potential this may extend duration of immunity, it is still recommended to get vaccinated. Perhaps with further study, recommendations will be altered to suggest just a single dose.

Also, it should go without saying but if you don't have immune protection yet it is clearly safer to get via vaccination than infection. In spite of the fact that infection confers strong natural immunity, any objective analysis of available data suggests that the substantial risk of death or long term health effects from infection is clearly far greater than for vaccination.