As clinical trials go, the measure of success for the massive coronavirus vaccine studies now underway is relatively simple. Thirty-thousand volunteers in each will receive either an experimental shot or saline solution, and the results will be judged positive if the COVID-19 rate among those receiving the vaccine is half or less than those who get placebo.
These will be the most consequential trial readouts in memory, with countries across the globe eager for signs the leading vaccines could offer a return to normalcy. The comparisons drugmakers will be making between vaccine and placebo, however, are complicated by the speed at which researchers have enrolled participants, and the short time frame in which they'll be looking for answers.
The Food and Drug Administration has said the minimum "vaccine efficacy" threshold for approval of a COVID-19 vaccine should be at least 50%, while Centers for Disease Control and Prevention Director Robert Redfield has indicated a decision could be made on as few as 150 cases in a trial. Does that mean a vaccine could be judged effective if more than 100 study participants given placebo get COVID-19, and fewer than 50 receiving vaccines?
That's possible, said Kaleab Abebe, associate professor of medicine and biostatistics at the University of Pittsburgh, but making that judgment too soon could result in authorizing a vaccine that in the long term turns out to be less effective. Cases can continue to accrue as the trial progresses, until a final analysis two years after participants are vaccinated.
Under Moderna's roadmap for the Phase 3 trial of its experimental shot, the primary comparison will be made after 151 participants get COVID-19. But an independent study committee will also look at the data at 53 and 106 cases. Declaring success early, at one of two interim data checks, could risk longer follow-up showing more cases than initially expected in the vaccine arm.
Furthermore, the committee, and later regulators, will need to feel certainty that the trend through the primary analysis will continue over the coming months or years of the study, in which more participants could fall ill.
"What's going to come into play is how many were infected, when they were infected and how many were at risk," Abebe said. "Somebody who got infected at day 31 is very different than somebody who got infected at day 758 in terms of what they've contributed [statistically]."
These judgments may need to be made very soon. The FDA has scheduled an advisory committee meeting on Oct. 22 to evaluate vaccine progress, and it's an open question how much data that group of outside medical advisers will have from trials which began at the end of July.
Pfizer and BioNTech's trial looked on track to reach the 30,000-patient enrollment mark this week, and CEO Albert Bourla said on Tuesday that 12,000 patients have received the second of two doses, which are spaced three weeks apart. Its trial begins measuring COVID-19 cases seven days after the second dose.
Moderna's longer dosing interval of four weeks, and a case analysis beginning two weeks after a booster dose, means it may have less data to share with the FDA advisory committee. In total enrollment Moderna is behind Pfizer, with just over 25,000 patients enrolled as of Wednesday, although it also reports that around 10,000 have received a second shot.
Bourla has said repeatedly he expects Pfizer will be able to deliver results by the end of October. Moderna CEO Stéphane Bancel, meanwhile, told BioPharma Dive that his company's base case is November. Looking more deeply at the trial's protocol, however, shows that estimate is not a sure thing.
Based on an assumed six-month case rate of 0.75% in the placebo arm and 0.3% in the vaccine arm, Moderna estimates it could take five months to have the 53 cases necessary to trigger the trial's first interim analysis. And at that point, Moderna's vaccine would need to be 74% effective for researchers to have confidence the trial would meet its criteria for success.
Pfizer's compressed schedule gives it a better chance. But assuming Pfizer's trial completes enrollment this week, the final study volunteers receive their second shot around two weeks before the FDA advisory committee convenes. Measurement of cases in the entire study population will begin around a week before that meeting.
Late Thursday, Pfizer released its protocol. The trial's data monitoring board will conduct its first interim analysis after 32 cases and a second after 62 cases. In order to trigger a potential positive readout, the vaccine would need to be at least 77% effective at the first check and at least 68% at the second.
The data panel will do a final analysis after 164 cases, in which the minimum threshold for efficacy would be 52%, or 111 cases in the placebo arm and 53 in the vaccine arm.
If either company makes the call for success, "what I expect to see is transparency," Abebe said. "I'd want to see published protocol that gives me a hint as to what are the original [research] questions and what were the explicit rules around how they decided to stop early."
But he acknowledged there is a hunger for results given the urgency of the pandemic. "You're competing with people dying," Abebe said.
Editor's note: This story has been updated to reflect Pfizer's release of its study protocol.