A vaccine from China’s Sinovac protected health care workers in Brazil from developing severe COVID-19.

REUTERS/Thomas Peter

Sciences COVID-19 reporting is supported by the Pulitzer Center and the Heising-Simons Foundation.

Brazil, with more COVID-19 cases than any other country after the United States and India, is on the verge of having its first authorized vaccine for the pandemic disease. At a press conference today, Brazilian researchers reported that a vaccine made by a Chinese company, Sinovac, was safe and had 78% efficacy in preventing mild cases of COVID-19 in a study of more than 12,000 health care workers. It also completely prevented moderate and severe illness caused by SARS-CoV-2 infections, the team said. “The result we are seeing today is fantastic,” Rosana Richtmann, a physician from the Emilio Ribas Institute of Infectious Disease from São Paulo, said at the press conference.

Dimas Tadeau Covas, head of the Butantan Institute–a state-owned vaccine-maker in São Paulo that is co-sponsoring the trial—expects Brazil’s regulatory agency to authorize the vaccine, called CoronaVac, for emergency use in the coming days. “We have now all the documentation to make the request,” he said. Brazil also has efficacy trials underway of a COVID-19 vaccine made by AstraZeneca and Oxford University and many scientists there expect it, too, to receive authorization soon. But the good news about CoronaVac was shadowed by complaints that the announcement included little data—and that Sinovac has muzzled its Brazilian collaborators. Indeed, Brazilian researchers had previously held a press conference to announce the vaccine was a success but were not allowed then to give an exact efficacy figure.

Most of the vaccines authorized in other countries rely on high-tech approaches, such as messenger RNA coding for the  surface protein of SARS-CoV-2 or—the approach taken by the AstraZeneca-Oxford vaccine, which uses a harmless viral vector carrying the gene for that protein. But Sinovac turned to a more established approach. Its vaccine, which is also in efficacy trials in Turkey and Indonesia, relies on the whole coronavirus, chemically crippled so it can’t cause disease. The reported efficacy against mild disease falls short of the roughly 95% achieved by two mRNA vaccines. But the “main purpose” of COVID-19 vaccines, Covas said, is to prevent infected people from progressing to severe disease. A different whole, inactivated SARS-CoV-2 vaccine made by Sinopharm’s China National Biotec Group has nearly identical results, that Chinese company reported last week from a large efficacy trial in several Middle Eastern countries.

Both Sinovac and Sinopharm have kept a tight lid on what their partners can reveal about their vaccines. At the 23 December press conference held in São Paulo at the governor’s residence, researchers said that because of a contractual agreement with Sinovac they could only report CoronaVac had greater than 50% efficacy, an internationally accepted standard for emergency authorization. Today’s presentation revealed the specific efficacy rate, but researchers were coy about describing the exact number of cases who developed disease in the vaccinated and placebo arms.

In a formal presentation of the data at Butantan today,  Covas made no mention of case numbers, saying the data will be reported in a scientific publication and to Brazilian regulators. But when pressed by journalists, he allowed that there were 218 cases of mild disease. “The exact number is 160 something in the placebo group and less than 60 in the vaccinated group,” he said. (This would equal 73%, not 78%, efficacy.) CoronaVac appeared to worked as well in the elderly as in other age groups, the researchers added.

The lack of data—let alone the efficacy discrepancy—predictably led to immediate skepticism. “Until they show us these numbers, it is nothing more than an empty announcement,” tweeted epidemiologist Denise Garrett, a vice president of the nonprofit Sabin Vaccine Institute.

Details have been in short supply in the initial efficacy trial announcements by other COVID-19 vaccine makers as well. But a researcher who ran one of the 16 sites that conduct the CoronaVac trial in Brazil, Esper Kallas of the University of São Paulo, said even study investigators still don’t know all the data. Kallas, who spoke with Science after the press conference, said he is “happy” about the results. “We know the 78% is true.” But he is frustrated that “we don’t have the granularity of it.” As for the case numbers Covas mentioned, Kallas said, “I couldn’t understand those figures.”

Sinovac did not reply to Science’s request for more information. But Kallas says the root of the disagreement between Butantan and Sinovac is what constitutes a confirmed case of COVID-19, as definitions can vary. The greater the number and severity of symptoms a trial subject must have to count as a case, the higher a vaccine’s  efficacy will appear to be. The Brazil team and the vaccine maker were at odds about whether a case has to show one recognized COVID-19 symptom or two, in addition to a positive PCR test for the virus.

Researchers testing CoronaVac in Turkey apparently faced no such hurdles. On 24 December they revealed preliminary data showing 26 COVID-19 cases out of 570 participants who received the placebo and 3 cases in the 752 vaccinated volunteers. This translated to 91.25% efficacy and no serious safety issues surfaced.  The next analysis will occur at 40 cases, and Murat Akova, a research at Hacettepe University who is helping to run the study. Asked about the restrictions on  Brazilian researchers, Akova says there is “no such contract with Sinovac about releasing data.”

The CoronaVac trial has come under fire from Brazilian President Jair Bolsonaro who has a strong bias against China and the vaccine itself and is a political rival of the governor of São Paulo state. But Sinovac, which has negotiated with São Paulo’s governor, will send Brazil enough vaccine for 46 million doses. (Two doses are needed for full protection.) Butantan also started making the vaccine itself and hopes to produce about 1 million doses a day. Covas says the institute has begun negotiations with several Latin American countries that want to obtain the vaccine.

Kallas, who has conducted about 15 clinical trials of different vaccines, says he has never seen more aggressive attempts to control the release of data, and fears that they undermine a highly positive message. “If you can prevent someone being seen by a doctor by 78% and prevent hospital admissions by 100%, let’s give a toast and celebrate,” he says.