Tuesday, December 29, 2020

One dose of mRNA vaccines

This is Joseph

I am starting to see the hot take of "why don't we experiment with giving only one dose of an mRNA vaccine". For example, see this

So let us count the reasons that this isn't a great idea.

  1. The vaccines were tested from phase 1 to phase 3 to find an optimal formulation. People looked at the single dose antibody titers, as compared to two doses, and it was a lot worse. This casts doubt on the durability of the immunity.
  2. Vaccine manufacturers look at these subgroups. the main Moderna trial had 30,351 participants and an efficacy of 94.5% (95% CI 86.5-97.8%). In the one dose subgroup there were 996 participants in the vaccinated group, and 1,079 in the placebo group with 7 cases vaccinated group and 39 cases placebo (80.2%; 95% CI 55.2-92.5%). So we know even short term efficacy is less.
  3. The Oxford vaccine had more participants in the low-dose/high-dose subgroup, with 3 cases in the vaccinated arm and 30 in the placebo arm (efficacy 90·0%, 95% CI: 67·4 to 97·0)
  4. In US dollars and sold in the US, the Oxford vaccine is $4 per shot versus $15/$19.50 for the mRNA vaccines. So it is a lot cheaper.
  5. The Oxford vaccine looks able to make 3 billion doses in 2021, whereas the two mRNA vaccines (combined) look to be able to produce 2 billion doses. So there is more expansion in capacity by adding Oxford than splitting the Moderna or Pfizer doses up. 
  6. This ignores Sputnik V and Sinovac, which are approved in some countries already/ These vaccines claim > 90% efficacy *(Sputnik V and Sinovac). These are both based on low case counts (bad) but no worse than the one dose sub-group for Moderna. 
  7. Oxford can "be stored and transported at normal refrigerated temps of 2 degrees to 8 degrees Celsius (36 degrees to 46 degrees Fahrenheit) for at least six months" whereas Pfizer must be kept at -70 degrees Celsius until six hours before use. That is a huge logistical advantage. Moderna's vaccine is better for logistics, but it is also the vaccine with the lowest production levels. 
  8. Even if the Oxford subgroup is due to the Texas Sharpshooter fallacy, the efficacy of the pooled vaccine estimate is about 70% (either you use the sib-groups or you don't). This is likely not significantly different than one dose Moderna and will likely be more durable
So if we want to argue for an ethical imperative to move quickly, the easy way forward is to approve the Oxford vaccine. There is almost no case where the one dose of mRNA vaccine approach is going to be better than just adding in the next tier of vaccines -- which have trials. Now it is possible that the FDA might decide that they are inferior and should not be given an emergency use authorization. But the one dose approach would require trials (timed to measure durability) and would take years to complete (versus actionable data now -- seriously, the Oxford vaccine has a Lancet publication so the data can be scrutinized). 

In a similar vein, the best thing we could do with Modern is . . . make more of it. Solves the one versus two dose problem if there is enough for everyone. If we really wanted to make a difference -- why not pay Moderna to drop the patent and make it universally free? 

Or push out the Oxford vaccine faster? Hilda Bastian has some concerns with the Oxford vaccine, but at least we have the data to make an informed risk-benefit tradeoff with it (and more data is coming soon when the US trial concludes). 

1 comment:

  1. Thank you for this interesting post. I have a question that is very basic and not quite to the point of your post, but appreciate your consideration.

    I thought that confidence intervals were supposed to be symmetric about the mean (1.96 times the std err for 95%). In the discussion above, the confidence intervals do not appear to be symmetric. For instance, in 2), discussing the single dose group, 80.2%; 95% CI 55.2-92.5%. 80.2 - 55.2 = 25, but 92.5-80.2 = 12.3. Shouldn't they be equal?

    Thanks again.