A new British preprint study adds to the evidence that COVID-19 vaccines are less effective at preventing infection by the especially contagious delta variant of the coronavirus, although they still sharply reduce that risk. The study also indicates that vaccinated people infected by delta tend to have lower viral loads and less severe symptoms than unvaccinated people.
The study presents data from the two latest rounds of REACT-1, a survey that questions and tests a random, nationally representative sample of people in England. Round 12, which included 108,911 participants, was conducted from May 20 to June 7. Round 13, which included 98,233 participants, was conducted from June 24 to July 12. Overall, 0.63 percent of participants tested positive for COVID-19 in Round 13, up from 0.15 percent in Round 12, which is consistent with the surge in new cases that the U.K. saw between late May and mid-July. Delta accounted for 78 percent of infections in Round 12 and 100 percent in Round 13.
Among participants who reported that they had been fully vaccinated, 0.4 percent tested positive in Round 13, up from 0.07 percent in Round 12. The corresponding figures for unvaccinated people were 1.21 percent and 0.24 percent, respectively.
When they assessed vaccine effectiveness, the researchers limited their analysis to participants ages 18 to 64, based on the assumption that vaccination among older people is associated with more cautious behavior, which would independently reduce the risk of infection. In Round 13, they estimated, self-reported vaccination reduced the risk of infection by 49 percent, down from 64 percent in Round 12. For symptomatic infections, the risk reductions were 59 percent and 83 percent, respectively.
The researchers, most of whom are affiliated with Imperial College London, did separate calculations for participants who agreed to allow verification of their vaccination status. Here the risk reductions for all infections were higher: 62 percent in Round 13, down from 75 percent in Round 12. But the authors note the possibility of "selection bias for consent to linkage": The share of participants who agreed to verification fell from 87 percent in Round 12 to 62 percent in Round 13.
The REACT-1 researchers also looked at the infection rate among Round 13 participants who reported contact with a confirmed COVID-19 carrier. It was 3.84 percent for fully vaccinated people, compared to 7.23 percent for unvaccinated people.
The effectiveness rates for Round 13 of REACT-1 are substantially lower than those seen in studies involving earlier iterations of the coronavirus, which generally found that mRNA vaccines reduced the risk of infection by 86 percent to 99 percent. They are also lower than the effectiveness rates in several earlier studies conducted after delta's emergence. "These estimates are lower than some others," the authors note, "but consistent with more recent data from Israel."
The fact that risk reductions were higher when the analysis was limited to symptomatic infections is consistent with other data indicating that vaccines remain effective at preventing severe cases, notwithstanding the delta variant. A recent study from Singapore found that patients infected by delta had fewer symptoms and shorter illnesses. In the United States, CNN recently calculated based on data from the Centers for Disease Control and Prevention (CDC), "less than 0.004% of fully vaccinated people had a breakthrough case that led to hospitalization and less than 0.001% of fully vaccinated people died from a breakthrough Covid-19 case."
The authors of the REACT-1 study note that "estimates of effectiveness against serious outcomes of greater than 90% have been reported" for people who received two doses of the Pfizer or AstraZeneca vaccines. "This is in keeping with our observation of a weakening of the association between infections and hospitalisations and deaths from mid-February to early April 2021 when Alpha variant was dominant," they say. "However, in our more recent data (since mid-April 2021), infections and hospitalisations began to re-converge, potentially reflecting the increased prevalence and severity of Delta compared with Alpha, a changing age mix of severe cases, and possible waning of protection."
That last point is worth keeping in mind, since the trends that generally have been attributed to the proliferation of delta also might be explained by waning vaccine effectiveness. Robert Wachter, chair of the Department of Medicine at the University of California at San Francisco, raised that possibility in an interview with The Washington Post last week. "If we're seeing more breakthroughs, is it just because the virus is better and the vaccines don't hold up quite as well, or is the efficacy of the vaccines beginning to wane, independent of the delta?" Wachter asked. "This is three-dimensional chess. There's a hundred things going on at the same time."
The question of whether vaccine booster shots are prudent remains controversial. Pfizer has been promoting the idea, and last week Israel began offering booster shots to vaccinated people older than 60. But so far the CDC and the Food and Drug Administration are unpersuaded. Anthony Fauci, the Biden administration's top COVID-19 adviser, told CNN last month that "we don't need" boosters "right now," but "there's a lot of work going on to examine this in real time to see if we might need a boost."
Based on cycle threshold (Ct) values from the positive PCR tests, the REACT-1 researchers found that vaccinated people had lower viral loads than unvaccinated people, which is relevant in assessing the likelihood that vaccinated carriers will transmit the virus. That finding contrasts with the results that the CDC reported in its study of a July outbreak in Provincetown, Massachusetts, where Ct values for vaccinated and unvaccinated people infected by the delta variant were similar.
Since the CDC's study was based on cases that came to its attention, the authors noted, "asymptomatic breakthrough infections might be underrepresented because of detection bias." The REACT-1 study, by contrast, was based on a random sample of the population, so its viral-load findings are probably more representative. The Singapore study found that viral loads in vaccinated and unvaccinated COVID-19 patients infected by delta were initially similar but declined faster in the vaccinated group.
The effectiveness of vaccines in the U.K.—where more than 70 percent of adults have been fully vaccinated, including more than 90 percent of those 65 or older—is reflected in daily COVID-19 deaths, which have fallen precipitously since late January. While the seven-day average of newly identified cases rose 24-fold between May 5 and July 20, the seven-day average of daily deaths—82 as of yesterday, per Worldometer, down from more than 1,200 in late January—has risen much less dramatically since May. As Reason's Elizabeth Nolan Brown notes, the current case fatality rate (deaths as a percentage of confirmed infections) in the U.K. is about one-tenth the rate seen earlier in the pandemic, when the alpha variant predominated.
The REACT-1 researchers note that "participants aged between 5 and 24 years," who are less likely to be vaccinated than older adults but also far less likely to die from COVID-19, "were over-represented among infected people in our study, contributing 50% of infections…while only representing 25% of the population of England aged 5 years or above." They add that "the highest prevalence of infection was among 12 to 24 year olds," consistent with their conclusion that "the third wave of infections in England" has been driven "primarily by the Delta variant in younger, unvaccinated people."
In the United States—where half of the population has been fully vaccinated, including the vast majority of Americans 65 or older—the recent increase in COVID-19 deaths likewise has been much smaller than the surge in cases. Newly identified COVID-19 cases in the U.K. have fallen sharply since July 20, and Fauci this week told MedPage Today he expects the U.S. will see a similar "turnaround" as the vaccination rate rises.