A participant in the South African trial of the AstraZeneca-Oxford COVID-19 vaccine has blood drawn before receiving her second dose.

AP Photo/Jerome Delay

Science‘s COVID-19 reporting is supported by the Heising-Simons Foundation.

Another COVID-19 vaccine has run into trouble in South Africa, showing less protection there than elsewhere because a SARS-CoV-2 variant that can apparently dodge key antibodies has become widespread. In the wake of the new finding, the country, which had pinned its pandemic hopes on this particular vaccine, halted an immunization campaign launched only last week.

The stakes are high for the vaccine globally because its makers, AstraZeneca and the University of Oxford, hope it will be widely used in developing countries; they project they can produce 3 billion doses this year for around $3 each, far more product at a far less lower price than any other vaccine shown to offer protection against COVID-19.

Yet the South African trial of the vaccine, conducted in about 2000 people, found such a low efficacy against mild and moderate disease, under 25%, that it would not meet minimal international standards for emergency use. But scientists are hopeful it might still prevent severe disease and death—arguably the most important job for any COVID-19 vaccine. That was impossible to tell from this placebo-controlled trial because it was small and recruited relatively healthy, young people–their average age was only 31. None of the subjects in either arm of the study developed severe disease or required hospitalization.

The new results are a “reality check,” Shabir Mahdi of the University of Witwatersrand, the trial’s principal investigator, said today at a press conference. “It is time unfortunately for us to recalibrate our expectations of COVID-19 vaccines, as well as how we go about deciding how to respond to the COVID-19 pandemic in South Africa as well as globally.”

COVID-19 vaccines made by Johnson & Johnson (J&J) and Novavax have also been shown to offer weaker protection against B.1.351 (also known as 501.V2), the SARS-CoV-2 variant that now causes the vast majority of all infections in South Africa, than against older variants. The vaccines’ efficacy against mild disease in South Africa was 57% for J&J and 49% for Novavax—lower than in any other country they were tested.

But the J&J vaccine, which was put to the test in the largest of the studies, convincingly protected against severe disease and death, even against the B.1.351 variant, and Mahdi remains “somewhat optimistic” that the AstraZeneca-Oxford vaccine will too; the results are not “all doom and gloom,” he said.

SARS-CoV-2-targeting antibodies triggered by the J&J vaccine were “very similar”, he said, to those elicited by the AstraZeneca-Oxford candidate, and the two vaccines are based on a similar technology: Both induce the body to make the spike surface protein of SARS-CoV-2 by delivering its genes in a harmless adenovirus. In a 44,000-person trial, the J&J vaccine   prevented 85% of severe cases and completely protected people from hospitalization and death in several countries, including the 15% of the participants who were from South Africa.

Mahdi and the research team had planned to report the results tomorrow, but the Financial Times on Saturday ran a story based on a leaked copy of the findings. Mahdi and his co-workers   have submitted a paper describing their data to a preprint server and expect it will post tomorrow.

The AstraZeneca-Oxford vaccine has produced confusing results from the start. Earlier preliminary results from trials in different countries showed a wide range of success rates against mild and moderate disease, but researchers have had difficulty interpreting the data because of differences in dose, intervals between doses, and variants in circulation.  Just on Friday, a study suggested the vaccine offered strong protection against a more transmissible variant, B.1.1.7, that exploded in the United Kingdom and is now spreading fast throughout Europe.

In South Africa, the vaccine was given in two doses spaced 21 to 35 days apart. Antibodies made by vaccine recipients can typically “neutralize” SARS-CoV-2, meaning they can prevent it from infecting cells in culture experiments. But lab studies show that they have far less power against B.1.351.

Mahdi stressed that the vaccine may still trigger a powerful T-cell response, which can target and eliminate cells the variant manages to infect. He presented a test-tube study showing how the mutations in the spike protein that allow B.1.351. to dodge neutralizing antibodies have little impact on T cell responses. “We believe that those T cell responses will still remain intact despite the mutations that exists in a B.1.351 variant,” said Mahdi.

The AstraZeneca-Oxford  vaccine trial, which ran from June to November, found that starting two weeks after the second dose—when participants presumably were fully immunized–19 cases of mild or moderate disease developed among the vaccinated, versus 23 in the placebo group, resulting in an efficacy of 21.9%. That is far below the 50% minimum required for emergency use authorization in many countries.

Researchers sequenced the viruses that infected trial participants and found a strong link between vaccine failure and B.1.351’s explosion in South Africa. In people who received one dose of the vaccine before the variant began to spread widely, efficacy against mild and moderate disease was still a respectable 75%.

South Africa last week received 1 million doses of the AstraZeneca-Oxford vaccine and began to offer them to healthcare workers, making it the first COVID-19 vaccine available in the country outside of clinical trials. Epidemiologist Salim Abdool Karim, who co-chairs the South African Ministerial Advisory Committee on COVID-19, said at the press conference that the rollout of the vaccine in South Africa “needs to be put on temporary hold” in light of the disappointing results. Barry Schoub, who leads a government advisory subcommittee on COVID-19 vaccines, says “we may need to look at combinations of the [AstraZeneca-Oxford] vaccine with other vaccines, which may in fact synergistically give a very good response.”

The University of Oxford team that originally designed the vaccine says it already has begun working on a second-generation candidate that targets the mutated spike protein of the B.1.351 variant. Oxford’s Sarah Gilbert, who leads that effort, suggested in a press statement that a reformulated vaccine might be given as a booster shot to the existing one.  “This is the same issue faced by all of the vaccine developers, and we will continue to monitor the emergence of new variants that arise in readiness for a future strain change,” noted Gilbert.