Why experts changed the vaccine dosage
The debate appears to be over: Ontario and other provinces have adopted recommendations from the National Advisory Committee on Immunization (NACI) allowing COVID-19 booster injections to be given within four months of the first jab, under most circumstances. That's a substantial lag compared to the recommended three-to-fourweek spacing for the mRNA vaccines.
We believe this is a sound decision that will get scarce vaccine into Canadian arms more quickly, but the decision was contentious, owing in part to the way that clinical data are produced and evaluated.
Soon after the first vaccine was approved, the head of Ontario's vaccine task force wondered out loud about relaxing the requirement for a four-week boost. The suggestion was shouted down, with commentators chiding him to “stick with the science.” More recently, B.C. Health Officer Bonnie Henry made headlines with a similar suggestion, pointing out that the proportion of people developing symptomatic COVID-19 dropped by 92 per cent starting 10 days after their first Pfizer dose. Her statement was based on publicly available information from Pfizer — conclusions borne out in later studies.
Still, many were understandably confused: other official statements had cited a less impressive 52-percent decline in cases following one dose of the same vaccine. These two values, 52 per cent and 92 per cent, have drastically different implications. If you're half protected by one jab, you might reasonably insist on getting a booster quickly, to finish the job. If you're almost fully protected (at least temporarily), it makes sense to share, providing others the same substantial benefit.
Where do these two numbers come from? The 52-per-cent reduction is from a “prespecified” analysis — one that researchers decided in advance to report. It covers the whole period after the first dose, including the first week, when no protection would be expected. The 92-per-cent figure, which looks only at the time after the vaccine starts working, is more relevant to understanding the benefits of that first jab. Unfortunately, it's based on a “post hoc” (not prespecified) analysis, calculated after everyone had looked at the data.
Regulators distrust post hoc analyses, for good reasons. Clinical trials produce rich data sets, ripe for cherry-picking. If statisticians could decide what `“success” means after seeing the data, researchers could report whatever findings appeared most positive. Insisting on prespecified analyses helps keep everybody honest. Hence, efficacy claims are always based on tests selected in advance. Similarly, the approved dosing for a vaccine or drug is based on procedures tested in the study.
But that's the standard practice. Facing the greatest health crisis of our time, don't the regulators get to look at the full data and decide which outcomes matter most? They can, but cautiously. In this case, they had to weigh benefit to the individual who gets the vaccine, along with the indirect benefits to society that come from providing the virus less opportunity to spread. In deciding to allow delayed boosters, Health Canada looked not just at the original study data, but also beyond it. Flexible thinking led to the current policy of flexible dosing.
Further challenges to regulators' flexibility are on the horizon. Soon, they will have to decide if people can be offered different vaccines for their first dose and their booster. In principle, this mix-and-match approach might work well, because the various vaccines all use the same viral spike protein to activate immunity against the virus. Some two-dose vaccines (like the Russian COVID-19 vaccine, not available here) were even designed with two different formulations.
It would simplify vaccine rollout to allow this approach. Ongoing studies are testing how well mixand-match works and whether it leads to unexpected ill effects. If the findings are reassuring, will authorities be willing to look beyond the original trial data once again, to permit a more flexible approach to dosing?
The mix-and-match question will likely become urgent because of concerns about blood clots in people receiving adenovirus-based vaccines, such as AstraZeneca's Vaxzevria and Johnson & Johnson's single shot. This serious side-effect is seen mostly in younger women — but so rarely that it is difficult to know who is actually at risk.
While waiting for further data, the NACI recommends Vaxzevria not be used in people under 55, whereas Health Canada has approved it and the Johnson & Johnson vaccine for all adults. One province has suspended Vaxzevria entirely, and the confusion may fuel vaccine hesitancy — highly unfortunate timing, as faster-spreading COVID-19 variants fill ICUs beyond capacity.
Meanwhile, Vaxzevria recipients still require boosters within four months. NACI's position is not to mix-and-match this vaccine with the others.
With new circumstances and new data, they may reconsider.
John Ashkenas, PhD, is a writer and scientist living in Toronto. Jim R. Woodgett, PhD, FRSC, served for 15 years as director of the Lunenfeld-Tanenbaum Research Institute and is a professor in the Department of Medical Biophysics, University of Toronto.