Today’s blog post is in response to Dolly Hodes’ question about studies that have “examined how vaccination protects the unvaccinated.” Her specific question was – “does vaccination reduce community transmission” and, second,  “does any protective effect change over time”? For those that live in households with children that cannot receive the vaccine or perhaps a family member that would not respond to a vaccine (due to underlying immunosuppressive disease or medication), this series of questions is of particular interest in knowing how best to protect them.

So does vaccination reduce community transmission?  This question has been evaluated at the population level by Milman et al. article published in Nature Medicine early June 2021 entitled “Community-level evidence for SARS-CoV-2 vaccine protection of unvaccinated individuals” (link: https://www.nature.com/articles/s41591-021-01407-5). It has specific applicability to children who were not eligible to receive the vaccine during the time period studied (9 December 2020 to 9 March 2021). In this study they looked at 177 communities (out of 246 eligible based on the number of tests and people) that had a low (<10%) infection rate. While this would seem to introduce some selection bias, the authors were attempting to eliminate or minimize the protective effect of community acquired infection, which could otherwise mask the effect of vaccination. The 10% threshold is a reasonable one – in my Santa Monica based practice we have a seroprevalence of natural infection antibodies of 7.8%.

The authors studied declines in positive test fractions at fixed time intervals (3 weeks) to account for the time that it takes for the protective effect of a vaccine to take hold as well as time patterns of change in the infection rates (i.e. surges). What they found was a strong inverse relationship between population vaccination percentage and positive test fraction among unvaccinated individuals.  The effect was such that for each 20% increase in vaccinated individuals, the positive test fraction among unvaccinated individuals decreased twofold. This effect was most robustly seen after 24 days (which is consistent with the time that it takes for a vaccination to take hold).

What I also found interesting is that the largest protective effect on the unvaccinated were when those 16-50 years of age were vaccinated.  There was little influence protection among the unvaccinated with those over the age of 65 were vaccinated. This likely points to the decreased frequency of interaction between the group 16-50 years and those over 65. But it also serves as a point of warning to the gaps in the US vaccination rates.  Vaccine uptake has been significant among those over the age of 65 years but, unfortunately, this has little effect in reducing risk among the unvaccinated. So to best protect the unvaccinated, those 16-50 (or 12-50) need higher rates of vaccination.

Now the delta variant throws a wrench in what would otherwise be a neat, logical line of biologic consistency. And I think it is important we acknowledge the mess delta creates, rather than trying to tell an entirely cohesive story. A number of studies have shown that the viral load, and by extension, the infectiousness, of the delta variant is indistinguishable between the vaccinated and unvaccinated (link to my favorite study showing this: https://www.thelancet.com/action/showPdf?pii=S2589-5370%2821%2900423-5). This observation fits with what I have seen in clinic with rapid tests that produce a very dark line among those with positive tests as delta became predominant in Los Angeles, as compared with a more faint line with prior variants.

Although the initial viral load with delta is equivalent between vaccinated/unvaccinated, there is some good news associated with vaccination in the era of deta.  A recent study of 218 participants (130 mRNA unvaccinated, 71 vaccinated, 13 with only one dose, 4 with non-mRNA vaccine) in Singapore (link: https://www.medrxiv.org/content/10.1101/2021.07.28.21261295v1) all with delta variant showed a faster rate of viral load decline among the vaccinated as compared to unvaccinated. Of interest, both groups started at an equivalent viral load at the time of diagnosis (measured by Median Cycle Value on PCR: 18.8 for vaccinated and 19.2 for unvaccinated) as well as with symptom onset (21.9 vaccinated versus 19.2 unvaccinated).  However, among those that had received the vaccine a identifiable lower viral load became apparent as early as day 5 of illness, dropping sharply by day 14 (Figure 1 below). This pattern of a shorter duration of viral shedding, according to the authors, “may allow a shorter duration of isolation for vaccinated individuals” and further “it seems likely that vaccination reduces secondary transmission.” Certainly in my (limited) clinical experience with vaccinated to vaccinated delta variant spread, there is a rapid termination of transmission even within a household.