Updated: Feb 8
Do the SARS-CoV-2 vaccines protect against infection and prevent transmission of the virus, or do they simply reduce symptoms of COVID-19?
This is a key question that is being widely discussed, and rightfully so as it determines what benefits would be provided by the current population-level vaccination efforts ongoing around the world.
While reduction of symptoms would provide a clear direct benefit for any vaccinated individuals and reduce population hospitalizations and deaths, the indirect benefit of vaccines on the population is potentially much greater, protecting not just the vaccinated but all those with whom they come in contact. Reducing the proportion of the population susceptible to infection strongly reduces the rate of community transmission, preventing many infections, and helping get the pandemic under control and return society more quickly to its pre-pandemic activity levels. If great enough, this could lead to herd immunity, whereby there are not enough susceptible individuals in the population for epidemic spread of the virus to be possible, effectively ending the pandemic and converting SARS-CoV-2 to an endemic, manageable virus. But this indirect benefit is only realized if the vaccines protect people against infection and transmission.
So, are the SARS-CoV-2 vaccines protecting against infection and preventing transmission? The short answer to this question is "we don't know for sure" because of limitations in the phase 3 study designs that prevent us from being able to definitively prove efficacy against asymptomatic infections or for reduction of transmission.
This is the reason for the disclaimers we see by the vaccine companies and scientific leaders that "we do not know whether the vaccines reduce infection or transmission," including the World Health Organization. But I am afraid this disclaimer is causing confusion in many about what the vaccine does or how it is designed to work.
I see numerous people boldly state "the companies admit the vaccines don't prevent infection, they only reduce symptoms" and even claim that the vaccines are "not really vaccines, they are biological treatments." As we statisticians and other scientists well know, the statements "we do not know whether the vaccines reduce infection or transmission" and "we know the vaccines do not reduce infection or transmission" are two very different things, but those who boldly claim that the vaccines are not preventing infections but only reducing symptoms are clearly confusing the two.
While it is true that limitations in the phase 3 study designs keep us from having definitive evidence that the vaccines are preventing infections or transmission, in this blog post I will explain why I think it is all but certain that the SARS-CoV-2 vaccines must be protecting against infection, and exceptionally likely that they are reducing transmission, at least to some degree.
While they may not be providing full sterilizing immunity that prevents all viral replication and provides 100% protection against infection or transmission, they are CERTAINLY doing more than "just reducing symptoms," and are likely to be protecting against infection and transmission enough to yield the intended effect once near population-wide vaccination is accomplished. I support this assertion based on a basic explanation of how the vaccines are designed, how our immune system works, and interpretation of the existing data from the phase 3 trials.
How are the vaccines designed to work?
So far, the USA has approved two vaccines for emergency use, the Moderna and Pfizer/BioNTech mRNA vaccines, and on February 18 will consider a third one, Johnson & Johnson/Janssen, which is a "viral vector" vaccine. The key component of all of these vaccines is the genetic code for making SARS-CoV-2's spike protein, the key protein of the virus that is necessary for its entry into cells and replication. The key difference between the mRNA vaccines and viral vector vaccines is the vehicle of delivery. While the Johnson & Johnson vaccine splices the spike protein genetic code into a deactivated human adenovirus, a common cold, and delivers it into the cells like a Trojan horse, the mRNA vaccines do not use a viral vector but instead encase the mRNA in lipid nanoparticles for protection and entry into the cell. In both cases, the spike protein mRNA in the cytoplasm generates spike proteins in the cell that stimulate the individual to produce an immune response against it.
There are claims circulating around social media suggesting the "SARS-CoV-2 vaccines are not really vaccines" because they are not made from deactivated or weakened versions of the target virus or because "they only reduce symptoms, don't stop infection". While it is true that the oldest vaccines were constructed from deactivated or weakened versions of the target virus, there are many other modern strategies for vaccination, including the viral vector technique underlying the Johnson & Johnson vaccine and the mRNA technique underlying Moderna and Pfizer's. Although only containing a part of the genetic code for the target virus, these are designed to do the same thing that all vaccines are designed to do: stimulating immune response for protection against future exposure to the virus. By design, they are not meant to simply be "treatments" that "only reduce symptoms", but are meant to train the body's immune system to protect against future infection and transmission.
They work differently than monoclonal antibodies or convalescent plasma that can be injected into an infected COVID-19 patient to boost their immune system and help treat the disease. These introduce externally produced antibodies into a person to supplement their own immune system in fighting off a current infection, so are designed to be treatments meant to reduce symptoms in an already-infected patient. This is in contrast to the vaccines, that are designed to expose a non-infected person's immune system to a part of the virus to stimulate their body to produce its own immune response, with hope of building immune memory to help the immune system respond promptly if exposed to the virus sometime in the future.
All of these SARS-CoV-2 vaccines are designed to work in a similar fashion -- to expose the immune system to a part of the SARS-CoV-2 virus to generate a vigorous immune response against it, in order to build immune memory, such that when later exposed to the SARS-CoV-2 virus, the immune system is able to react quickly and neutralize the virus.
How does our immune system fight the virus and what does "infection" really mean?
Once exposed to an invading virus, a race between the virus and immune system commences that will determine the eventual fate of the infected person.
Our immune systems are designed to recognize an invading virus and begin constructing an immune response consisting of T-cells and antibodies to neutralize it. Unfortunately, this takes time, especially for a novel virus to which the system has not been previously exposed, and in the time the immune system is trying to construct the immune response the virus is rapidly replicating and spreading to other cells, which can be measured by increased viral load.
Viral load influences symptoms and transmissibility. For SARS-CoV-2, higher viral loads have been shown to increase transmissibility of the virus, leading to the exhalation of sufficient viral particles such that others nearby might be exposed to enough virus to infect them. Higher viral levels also tend to worsen symptoms and increased viral replication can lead to more severe disease. If the SARS-CoV-2 virus replicates sufficiently so that it spreads down the respiratory tract into the lungs, it finds fertile ground for enhanced replication, and that is when it tends to produce the most severe and life-threatening conditions that make the virus so dangerous. In those situations, the spread gets out of control like a forest fire and the immune system struggles to keep up. In those cases, the virus wins the race.
If the immune system works quickly enough, the virus can be neutralized before it can get to that point. Using the analogy of a forest fire, the hope is that the immune system recognizes and extinguishes the localized brush fire before it can become an inferno that burns the forest down. If the virus is neutralized before it replicates sufficiently to produce a positive viral test, then you could argue the person was never effectively infected even if technically the virus replicated to some degree, since their infection would be completely undetectable. And even if it produces a positive test, if it is neutralized before it can produce any sickness, symptoms, or transmit to others, then it still provides equivalent direct and indirect protection to society even if not strictly preventing replication or detection of the virus. In that case, the person's virus is as indolent as if they had not been infected at all. Thus, in practice the concept of preventing infection is not as straightforward as one might think.
By design, an effective vaccine will produce immune memory so that when the vaccinated individual is later exposed to the virus, their immune system has a head start in the race and can produce a rapid immune response against it. If sufficiently rapid, it can prevent sickness, symptoms and transmission, and possibly even prevent a positive viral test. This is the goal of vaccination, since if the exposed individual is protected against sickness the vaccine has produced a direct benefit for them, and if they cannot effectively transmit to others then it is providing the intended indirect effect for the rest of the population.
Some vaccines, such as measles and smallpox, provide what is called sterilizing immunity, which means the immune system responds so quickly that the virus cannot replicate at all, which would prevent all sickness, symptoms, and transmission, as well as positive viral tests. In this sense, it provides 100% protection against symptoms, infection and transmission. Others, including vaccines against Hepatitis B and C, pertussis (whooping cough), inactivated polio vaccine (IPV), and inactivated flu vaccines, do not provide full sterilizing immunity, but still provide variable levels of protection against sickness, symptoms, and transmission. Some prevent sickness and suppress symptoms, but do not remove the virus from the individual's system, as it can still be detected by a viral test. In that case, if the viral levels are high enough, transmission may still be possible, although if the viral levels are sufficiently suppressed such non-sterilizing immunity might still significantly reduce transmission. One example is rotavirus vaccines that do not provide sterilizing immunity but have been shown to significantly reduce viral loads and reduce transmission.
In spite of how it is sometimes portrayed in the media (e.g. see this BBC article), the protection afforded by vaccines is not an either-or proposition of full sterilizing immunity or simple suppression of symptoms. There is a continuum of protection based on how rapidly it can produce an immune response, and what proportion of those vaccinated realize that rapid response.
For those whose response is strong and immediate, they may never experience symptoms, have sufficient virus to transmit, or even test positive for the virus if exposed.
For others, they may not experience symptoms but could still test positive for asymptomatic disease, and may or may not be able to transmit depending on their viral loads.
Still others may experience symptoms and test positive, but the vaccine might help their system neutralize the virus before it can reach the lungs or produce serious, life-threatening conditions. They might still transmit to others, but the transmissibility might be reduced if their viral loads are decreased from what they would have been sans vaccine.
For those in #2 and #3, it might be accurately said that the vaccine is "simply reducing symptoms", except perhaps they may be less prone to transmit to others, but for those in #1 it is clear that the vaccine is effectively protecting against infection and transmission. There is a middle ground whereby a vaccine might confer significant protection against infection and transmission, yet falls short of full sterilizing immunity, and that is where the evidence suggests the current SARS-CoV-2 vaccines reside (and honestly no vaccines really provide 100% sterilizing immunity)
So, what do we know about the SARS-CoV-2 vaccines and the degree of protection they are conferring?
What do the phase 3 studies tell us about the protection provided by the vaccines?
There was a national private-public partnership, ACTIV, that as part of Operation Warp Speed came up with a harmonized plan to make all USA phase 3 COVID-19 vaccine trials follow the same design to enable later meta-analysis combing results. Based on my discussion with their statistical leader, I learned that they decided to focus on assessing whether the vaccines were preventing symptomatic infections, driven by expedience, logistical considerations, lack of sufficient testing, and the primary concern of reducing the most severe cases, hospitalizations, and deaths.
Explicitly testing whether the vaccines were effective in preventing all infections, including symptomatic and asymptomatic ones, would have necessitated regular (e.g. weekly) PCR testing for infection, which would have required participants to come in for a PCR test every week, and >20 times the number of tests.
Thus, participants in the phase 3 study were only given SARS-CoV-2 PCR tests if they reported any of a number of potential COVID-19 symptoms (e.g. fever, sore throat, fatigue, coughing, loss of taste or smell). If their PCR test was positive, they were considered a symptomatic case. The primary efficacy end point of the study was then chosen to be the relative reduction in symptomatic cases in the vaccinated group relative to the placebo group.
For example, in the Pfizer/BioNTech study, they found a total of 162 symptomatic cases in the placebo group, but just 8 in the vaccinated group. The reported 95% efficacy is computed by taking the percent reduction in symptomatic cases in the vaccine group relative to the placebo group , e.g. assuming equal nunbers of placebo and vaccine subjects:
Efficacy = 100% x (162 - 8)/162 = 95.0%
If the vaccine were not effective in preventing symptomatic cases, we would expect the vaccinated group to have the same number of symptomatic cases as the placebo group and the efficacy would be 0%. The 95% efficacy means that 95% of the symptomatic cases that would have been expected to occur sans vaccine were prevented by the vaccine. In this way, it provides proof that the vaccine is effective in preventing symptomatic disease.
However, because no regular (e.g. weekly) testing was done, there is no way to know anything about asymptomatic infections, and thus the studies cannot make any definitive statements about reducing total infections, or preventing infections. For this reason, someone could claim that it was possible that the vaccine was "just reducing symptoms," not preventing infections, and they would technically be right. This is the reason for oft quoted disclaimer that "we don't know if the vaccines protect against infection or reduce transmission".
Why the vaccines must be protecting against infection
However, for one to argue that the vaccine was simply reducing symptoms and not preventing infections or reducing transmission, they would have to assert that all 95% of those symptomatic cases that were "prevented" by the vaccines were not actually prevented symptomatic infections, but instead were simply symptomatic infections that were converted to asymptomatic infections that were equally transmissible.
This would imply that the vaccine and placebo arms in fact had the same number of total SARS-CoV-2 infections, with the placebo arm having a higher proportion of symptomatic cases and the vaccine arm having a higher proportion asymptomatic cases. Given a large (95% for Pfizer) reduction in symptomatic cases, for this to be true the unmeasured asymptomatic cases would have to be MUCH higher in the vaccine then placebo groups. And this is clearly implausible.
While regular testing for asymptomatic disease was not a primary part of the protocol, for two of the vaccines, there is partial information about asymptomatic infections that we can consider to see if indeed the asymptomatic infections are much higher in the vaccine arm.
For the Moderna study, participants were all swabbed just before the second dose and given PCR tests to detect disease. Following is the table from the FDA briefing on the results of this swab:
From this, we see that the rate of asymptomatic infection at this single time point was in fact 100% x (38-14)/38 = 63.3% lower in the vaccine arm than the placebo arm. Not only were there not much higher number of asymptomatic cases in the vaccine arm, but there were substantially and statistically significantly lower numbers of asymptomatic cases relative to placebo arm. This makes it implausible to claim that the Moderna vaccine was only reducing symptoms, not protecting against infection.
For the AstraZeneca/Oxford vaccine, a subset of 8207 subjects (4071 vaccine, 4136 placebo) in one of the UK randomized studies (COV002) did weekly at-home PCR testing to detect asymptomatic disease over a long period of time. In Table 1 of a preprint report for the Lancet presents the results for vaccine efficacy for symptomatic and asymptomatic disease:
While the overall efficacy for the AstraZeneca/Oxford vaccine for preventing symptomatic infections (66.7%) was much lower than the 95% observed for Pfizer/BioNTech and the 94.1% observed for Moderna, this still provides strong and definitive evidence that the AstraZeneca/Oxford vaccine showed efficacy for preventing symptomatic disease.
The regular at-home testing in this trial allows an investigation of whether the vaccine was indeed preventing symptomatic infections, or whether it was simply reducing COVID-19 symptoms and converting symptomatic to asymptomatic disease. From the table, we see that 73/4136 (1.8%) of the individuals in the placebo arm had asymptomatic disease while only 57/4071 (1.4%) of the individuals in the vaccine arm had asymptomatic disease. Thus, it is clear that the 66.7% of symptomatic cases that were prevented by the vaccine were not simply converted to asymptomatic disease, and in fact the vaccine demonstrated some efficacy in preventing asymptomatic disease, as well, even if at a low rate (22.2%). Looking at all infections, the trial results show a 54.1% reduction in total infections, asymptomatic and symptomatic, relative to placebo. This shows that even the AstraZeneca/Oxford vaccine, with much lower preliminary efficacy results against symptomatic disease, clearly provides some level of protection against infection.
While we lack such data for Pfizer/BioNTech, the high efficacy (94.1%) for preventing symptomatic infection means that if it were just reducing symptoms not infections, there would have to be an extreme overabundance of asymptomatic cases in the vaccine arm to cancel out this reduction in symptomatic cases, and in spite of the lack of direct evidence, it would be a real stretch for anyone to claim this is the case, and seems beyond implausible.
This means it is all but certain that the vaccines are in fact preventing infections, maybe not fully, but at least to some degree.
What about transmissions?
Rigorously demonstrating vaccines prevent transmission is much more difficult than providing proof of prevention of infections. One could use cluster designs where entire families are entered into a study and randomized or secondary transmission designs that regularly test contacts of participants for virus and then compare the secondary transmission from individuals in vaccine and placebo groups -- both could rigorously study transmission and provide scientific assessments of the potential benefit of vaccines to reduce transmission. However, for many of the same reasons they chose not to do broad asymptotic testing, including expediency, logistics, and testing availability, plus additional reasons of wanting to get broad representation of important subgroups like minorities, elderly, and those with pre-existing reasons, the ACTIV public-private partnership decided to forego these designs and focus on simpler designs randomizing individuals instead. This limits their ability to draw rigorous conclusions about transmission.
However, if a vaccine is efficacious in preventing infections, it stands to reason that it is reducing transmission since individuals who are not infected are also not transmitting to others. While it is technically possible that even someone with a negative PCR test who is determined to be neither symptomatically or asymptomatically infected may have some undetected level of virus that is technically transmissible, these individuals would also have very low nasal viral loads, so would be much less likely to transmit to others. Assessments of nasal viral concentrations could be used as a surrogate that might be used to assess likelihood of transmission in future trials.
There is far less evidence for reduction of transmissibility from the phase 3 results themselves. However, as vaccination is rapidly progressing in many countries in the world, careful analysis of observational data that adjusts for other confounding factors can shed light on whether the vaccines are indeed reducing transmissions. Comparisons of regions undergoing early or late vaccination, for examples, can shed some light on this question. If it is clear that many places experiencing early vaccination show more rapid decline than those vaccinating later, and if regions approaching full vaccination show near full reduction of transmission, then this will provide strong evidence for the hypothesis of reduced transmission.
As mentioned in another blog post, Israel is rapidly vaccinating their entire population, already vaccinating a vast majority of there >60 population, with >90% receiving their first dose and >75% receiving both doses, and early evidence comparing the >60 and <60 population and cities vaccinating early or later providing support for reduced infection and transmission. Time will tell if these results are due to vaccination, as if it is we should see continued sharp declines in the coming weeks and months.
While it has not been definitively shown that the SARS-CoV-2 vaccines prevent infections and transmissions, it is clear from how the vaccines work and the available data that it is not simply reducing symptoms -- it must be preventing infections to some degree and likely also reducing transmission.
It is clear that the vaccines do not produce the nearly complete sterilizing immunity that measles or smallpox vaccines provide that prevents nearly all replication, infection, and transmission. However, complete sterilizing immunity is not necessary for a vaccine to provide a strong direct protection for recipients and strong indirect protection for the rest of the population.
As long it is significantly reducing risk of infection in the vaccinated group, it should lead to much lower levels of viral infections and lower transmission rates in society, strongly reducing the number of individuals exposed to the virus. Even with less than 100% efficacy in preventing infection, if we get a high enough proportion of society vaccinated, we can reduce transmission to a trickle and perhaps even reach herd immunity whereby there are not enough truly susceptible people to sustain epidemic growth. And this proportion does not need to be near 100%, but does need to be substantial, clearly more than half.
Even if the vaccines have strong efficacy in reducing infections, many questions remain. The emergence of new variants, some of which appear to be able to resist current immune response may reduce the realized efficacy and require variant-specific boosters, raise concerns. But early indications are that vaccines should retain most of their efficacy against these variants, plus for the mRNA vaccines experts claim variant-specific boosters can be developed and validated in as little as 6 weeks. Also, durability of immunity is an unanswered question as well -- it is not clear how long a person retains effective immunity vs. infection after recovery from asymptomatic, mild, moderate, or severe disease, nor is it clear whether the vaccine's immunity will be more or less durable than natural infections. Given that the vaccines have been shown to produce stronger neutralizing antibody responses than natural infections makes it likely that they produce immunity at least as durable as natural infection, but time will tell whether this is the case or not.
Image from BusinessInsider