One Dose of Coronavirus Vaccine May Be Enough for Most People
A new analysis of Pfizer data suggests good protection starting 14 days after the first shot.
Data sleuths trolling through the FDA filings for the Pfizer and Moderna vaccines have uncovered, what they say, is a critical fact — these vaccines may be highly effective after just a single dose. Given that vaccine demand vastly outpaces supply, as well as the concern that more coronavirus variants will emerge if we don’t rapidly squelch the infection rate, calls for a one-shot-only approach abound.
Is that the right move? Well, like all things public health, it’s complicated.
But first let’s reconstruct what the researchers did.
If you dig a bit into Pfizer’s FDA filing, you’ll find this table, which breaks down the number of cases in the vaccine and placebo group from the very first dose.
You see 50 cases total in the vaccine group, and 275 in the placebo group, for 82% efficacy. Of course, we wouldn’t expect the vaccine to be effective the instant it is injected. That’s why the primary outcome looked at infection rate starting 7 days AFTER the second dose — that’s on the bottom there.
But they do report there were 39 cases in the vaccine group, and 82 in the placebo group between dose 1 and dose 2. And you can see the calculated efficacy in that window is only 52% — not amazing.
But it’s not fair to start the clock at the moment dose 1 goes in. Can we reconstruct how many cases happened between day +7 or day +14 and that second dose?
Turns out we can — the data is hidden in this graph.
At the bottom, the cumulative cases each week are shown.
At 7 days, we can see 21 cases in the vaccine group, at 14 days 37. So we know that there were 14 cases between day 7 and day 14. We know from the FDA document that there were a total of 39 cases in the vaccine group before the second dose was given. So a little bit of arithmetic can generate a table like this:
And that’s essentially what the authors of the New England Journal letter do, calculating a vaccine efficacy rate of 70% after 7 days from dose 1 and before dose 2 and 93% after 14 days from dose 1 but before dose 2.
But let’s remember that the time period of 14 days after dose 1 to dose 2 is like 7 days. That’s why you have so few cases.
It’s also why the confidence intervals around that efficacy calculation are so wide, ranging anywhere from 69 to 98%.
How do we interpret those wide confidence intervals? Very carefully. The data looks good — just 2 cases versus 27 cases in that time period. But if you had one more case in the vaccine group, that 93% efficacy drops to 89%. One more additional case and you’re at 85%. When there are so few cases overall, the estimates just aren’t very precise.
But are they good enough? Is that efficacy signal strong enough to conclude that one dose provides decent protection? I actually think it does, especially given the corroborating results from the Moderna vaccine.
What we don’t know is how long that protection will last, but given the robust immunogenicity of these vaccines I suspect it isn’t particularly short-lived.
The primary problem people have with this analysis is that this isn’t what the study was designed to do. It’s post-hoc math, which does certainly increase the risk that this is a spurious signal. But sometimes the perfect is the enemy of the good. Remember, it’s not like overall they found the vaccine didn’t work and we’re doing a sub-analysis that suggests it works in some esoteric time period. Overall, the vaccine was a smashing success — giving us more faith in these post hoc results. And similar data from Moderna, which has a very similar vaccine, really strengthens the case overall.
Now, this doesn’t mean that necessarily we should abandon the two-dose program, but I do think it is reasonable, while vaccine supply is limited, to provide one dose to lower risk, high-exposure individuals — like teachers and essential workers. Reserve the two-dose regimen for older adults and other high-risk populations. And, once more supplies come in, give everyone the booster.
With rollout still lagging demand, it’s worth a shot.
A version of this commentary first appeared on medscape.com.