In recent months variants of SARS-CoV-2 have been detected that are unusual in that they have many more genome mutations than previously found. These have been called ‘variants of concern’ (VOC) as it has been suggested that the genome mutations might impact transmission, immune control, and virulence. Below I cover each of these issues separately.
The SARS-CoV-2 lineage called B.1.1.7 arose in the United Kingdom in September 2020 and harbors 17 genomic mutations, some of which lead to amino acid changes in the spike protein (pictured). Similar but distinct variants have been detected in other locations, including South Africa (B.1.135) and Brazil, but the B.1.1.7 lineage has been best studied. A good summary of the changes can be found in this manuscript. A number of lines of evidence have led to the conclusion that viruses of the B.1.1.7 lineage may have increased transmissibility compared with previous isolates. These include the rapid displacement of previous variants in the UK within a short period of time; an apparent increase in the R index for such variants; and increased levels of viral RNA in nasopharyngeal washes as measured by PCR or RNA sequencing.
The virological definition of transmission is the movement of viruses from one host to another. In the case of SARS-CoV-2, such transmission occurs when infectious virus particles are exhaled within respiratory droplets and arrive in another host, where they initiate infection. The evidence cited above for increased transmission of the B.1.1.7 lineage are all indirect and do not prove that the variants actually transmit, in a virological sense, better between hosts. The population growth of the variant could, for example, be a consequence of changes in human behavior. The R index, a measure of transmisssibility, is influenced not only by the virus but by human behavior. The finding of increased levels of RNA in nasopharyngeal wash is also inconclusive with respect to transmission. Viral RNA is not the same as infectious virus, and no studies have been done measuring shedding of infectious virus from individuals infected with variants of the B.1.1.7 lineage compared with other variants.
There is no doubt that the B.1.1.7 lineage has rapidly displaced others in the UK. Whether this behavior is due to an increased ability of the virus to be transmitted form one host to another has not been demonstrated. The variant has also been detected in other countries and its dispersion in those locations are not consistent with increased transmission (as I have defined above). For example, we now know that the B.1.1.7 lineage was present in the US 5-6 weeks before its detection in the UK, yet as of January it comprised just 0.3% of cases nationally. After 2 months of circulation in California, the lineage is estimated to account for 0.4% of cases compared with 1.2% at a similar point in the UK. In Florida the lineage is associated with higher spread, 0.7% of cases, but this is not the situation in other US states.
These data emphasize that we cannot conclude that the B.1.1.7 lineage is biologically more transmissible. Multiple factors are likely at play, and this is why it is better to view the B.1.1.7 lineage variants and others in terms of their fitness – the reproductive success of the virus. Many factors can influence fitness, not just transmission. These could include increased physical stability of the particle, increased resistance to immune responses, longer duration of virus presence in the nasopharynx, increased infectious virus produced within the host, more efficient establishment of infection in a host, and more. A slight increase in any of these might drive a particular variant within a population but not actually affect person to person transmission. Whether such mutations are spread by founder effect – being in the right place at the right time – also must be taken into consideration.
The statistical models that have been used to approximate the transmission of SARS-CoV-2 variants cannot prove a biological property because drive through a population can be a consequence of various fitness parameters. Experiments either in animal models (in which case the relevance to humans is unknown) or measurement of infectious virus in humans is needed. So far none of the latter have been done for the current variants.
A more immediate concern is whether any of the changes in spike protein within VOC impact the ability of immune response to control infections. This question has been directly addressed for neutralizing antibodies, e.g. those which can block infection. Antibodies recognize specific protein sequences on the virus particle, and specifically the spike protein for those given the mRNA vaccine. Some of the spike changes identified in variants are in regions known to bind antibodies. Consequently an important question is whether vaccination can inhibit infection with the variant viruses.
This question has been addressed for both the Moderna and the Pfizer mRNA vaccines. Sera from persons immunized with mRNA-1273 efficiently neutralized pseudotyped viruses bearing the SARS-CoV-2 spike glycoprotein from the B.1.1.7 lineage. These sera had a reduced (6.4 fold) neutralization titer when the South African B.1.351 lineage was used. However these sera still fully neutralized B.1.351 with a titer of 1:290 which may be sufficient to prevent severe COVID-19. Nevertheless, Moderna has announced that it will advance a modified vaccine (mRNA-1273.351) encoding the B.1.351 amino acid changes.
In a separate study, sera from individuals vaccinated with the Pfizer BNT162b2 mRNA vaccine was tested in neutralization assays using SARS-CoV-2 viruses with selected spike amino acid changes from the B.1.1.7 (deletion of amino acids 69/70, N501Y, D614G) or B.1.351 (E484K + N501Y + D614G) lineages. These changes had small effects on neutralization with the sera. However, the engineered viruses do not contain the full set of changes found in the B.1.1.7 and B.1.351 viruses, which might explain the different results compared sera with antibodies induced by mRNA-1273.
These observations provide confidence that the two mRNA vaccines will provide protection against COVID-19 caused by currently circulating variants. However genomic surveillance must be increased to ensure that any new spike changes that might arise are detected quickly and their effects on neutralization determined.
A previous study did not show evidence that viruses of the B.1.1.7 lineage were associated with an increased risk of hospitalization or death. However upon examination of additional data from three separate studies NERVTAG concludes that there is a ‘realistic possibility that infection with VOC B.1.1.7 is associated with an increased risk of death compared to infection with non-VOC viruses’. This conclusion was reached by statistical analyses of reported death rates among individuals infected with VOC B.1.1.7 or non-VOC viruses. For example, in one study the relative hazard of death was 1.35 (with a 95% confidence interval of 1.08-1.68). In another study the mean ratio of case fatality ratios between cases caused by VOC or non-VOC viruses was 1.36 (95% CI 1.18-1.56). These are small differences with large confidence intervals ranging from no effect to more effect, and the authors note that the absolute risk of death remains low. The statistics are computed by analyzing a limited dataset of all COVID-19 related deaths (8%) and consequently might be in error. Furthermore, there does not appear to be an increased risk of hospitalization associated with infection by VOC viruses. My reading of this report is that it mainly serves as a warning to continue genomic surveillance of variants with respect to death risk and does not come to a conclusion on causality.
Update: Novavax just released the first results of their phase 3, spike-protein based COVID-19 vaccine. Efficacy was nearly 90% in the UK, but in a smaller trial in South Africa it was 50% against the B.1.135 variant.