COVID Transmissions for 2-19-2021

A single dose of the Pfizer-BioNTech vaccine has effects, but let's not get ahead of ourselves

Feb 19, 2021

Good morning and welcome to COVID Transmissions.

It has been 459 days since the first documented human case of COVID-19. 459, on an alphanumeric keypad, will get you “ILY,” or “I love you.” I know, it’s kind of a stretch. But I wanted to get somewhere nice as we close out this week. Have a good weekend!

Today I am bringing together two reports of real world evidence and one letter in The New England Journal of Medicine to explore the effects of a single dose of the Pfizer-BioNTech vaccine. The results are not definitive, but I think they’re interesting. We’re in a messy period of time right now where it’s hard to say anything for sure, except this: the vaccines are working, and we need to give them the time and support to keep doing what they’re designed to do.

As usual, bolded terms are linked to the running newsletter glossary.

Keep COVID Transmissions growing by sharing it! Share the newsletter, not the virus. I love talking about science and explaining important concepts in human health, but I rely on all of you to grow the audience for this, which you can do by using this button here:

Share COVID Transmissions

Now, let’s talk COVID.

A letter in NEJM tries to make the case for one dose of the Pfizer vaccine

Several people have showed me this letter to the editor from The New England Journal of Medicine, which attempts to use existing clinical trial data to make the case for single-dose protection from the Pfizer-BioNTech vaccine: https://www.nejm.org/doi/full/10.1056/NEJMc2036242

What the authors of this letter did was the following:

They know that there is a 3-week period between doses in the published trial
They know that immunity takes about 2 weeks following most vaccinations to be present in notable amounts
They figured they could look at COVID-19 cases after 2 weeks had passed but before the second dose was administered, and compare the vaccinated group to placebo to get an assessment of efficacy

Their assessment based on this was a vaccine efficacy of 92.6% for a single dose of the Pfizer-BioNTech vaccine. They’re not the first to make this point, but they’re the first to make it so boldly in such a high-profile journal. They seem pretty confident. Sounds reasonable, right?

It’s not, and there are a lot of reasons it’s not.

Clinical trials are designed prospectively and the statistical plans are decided in advance to analyze the trial data as planned according to the protocol. One of the reasons for this is that, if you cut the data as you see fit after you have collected them, you can often make them say whatever you want. Doing this kind of retrospective analysis is like challenging someone to a coin-flipping contest and then saying “Best of 5?” once you’ve one three tosses. It’s bad science and I don’t support it.

I have other specific criticisms. One key point comes from the lawyer-commercial cliché, “past performance is no guarantee of future results.” Looking at the one week right after immune responses should be at their strongest, and cutting it off after one week only, strikes me as simply not enough follow up. I’m willing to believe that one dose of the Pfizer-BioNTech vaccine gives you a 92.6% efficacy for the third week after vaccination, sure. I can’t tell you what happens at week 4 or week 20 from this dataset. No one can, because the experiment didn’t look at this. This is a dangerous game to play. Yes, the protection might stay strong. Or it might get markedly worse. This clinical trial cannot tell you and it’s dishonest to pretend it can.

This is not how we do science. I don’t endorse this kind of thing. Unfortunately, I’m pretty sure that the ultimate message is correct—but only by accident. There are better data out there now that happen to confirm the general message about one dose of the Pfizer vaccine. These aren’t it and this is bad practice. Being right by chance is no excuse for doing bad science.

That said, I think there’s something we can learn from this analysis. Specifically, I think it’s pretty clear that if you are going to receive both doses of the vaccine on schedule, there is some amount of protection in the week before you get that second dose. That’s good news, because it tells me you’re not likely to catch COVID-19 while waiting in line to get your second dose. That’s the takeaway that I have from this.

As for the efficacy of one dose? That really needs its own trial, but I do think there’s a country we can watch for better information on this. Specifically, the UK.

A single dose of the Pfizer-BioNTech vaccine seems to have an impact in the UK

This article in The Financial Times shows us some effects of a single dose of the Pfizer-BioNTech vaccine: https://www.ft.com/content/eb311423-dd2f-48f8-af57-8d8f0103c313

Earlier in the pandemic, the UK decided that it would experiment on its population by delaying the second dose of this vaccine in favor of giving more people a first dose. I was against this type of open-air medical experiment then and I’m against it now. However, I am not going to ignore the data they obtain just because I don’t think the decision was ethical. It happened, I can’t stop it, so we might as well learn something from it.

What we see here is that once 50% of the population in the UK over the age of 80 had received one dose of the Pfizer-BioNTech vaccine, cases and hospitalizations start to drop. A national lockdown probably also contributed to this, but I think we’re probably seeing a genuine vaccine effect here.

That said, the followup here is also short. It looks to be about a month after they reached that 50% mark, but that means that a number of patients in that cohort have more than a month of followup. We haven’t hit the magic number of three months after that 50% mark, which I think is really very important, but the signs are pointing to at least some protective effect from a single dose of this vaccine, and that it can last for some weeks.

This isn’t very definitive either, but I feel more confident looking at it because it’s data from hundreds of thousands of people. I think we need to continue to watch the UK experience to understand how the dynamics of protection look over time. They’re planning to give second doses of the Pfizer-BioNTech vaccine 3 months after the first dose. That means we’ll have a solid 10 weeks of single-dose followup, and I think that’s going to be pretty meaningful on this question.

Instead of retrospectively recutting clinical trial data, this is what we should be looking at to learn what a single dose of this vaccine can do.

The Israeli experience, continued

I also wanted to share a reprint from Israel, to continue my exploration of vaccine effects there. In this reprint, we are again looking at the efficacy of a single dose in the 2 weeks before the second dose is administered. Instead of looking at protection, though, the authors looked at the amount of viral RNA present in patients who tested positive in Israel in that 2-week period. This is similar to what we discussed yesterday, but looking at 1 dose now instead of 2.

The reprint is here: https://www.medrxiv.org/content/10.1101/2021.02.06.21251283v1.full.pdf

I think this is really the key figure, though:

Image shows two line graphs, separated into panels (a) and (b). The top graph, (a), shows the proportion of patients in days 1-11 after the first dose who had RT-PCR values of certain numbers. Two lines are shown, and they overlap each other: vaccinated patients in blue and unvaccinated patients in yellow. In (b), the same kind of graph is shown, but covering days 12-28 after the first dose is administered. This time there is a gap between the two lines.

The bottom line is that this figure shows that, in the period 12-28 days after a single dose of the vaccine, there is a reduction in viral load (genome count) in patients who got that vaccine.

For those of you who want more details, let’s walk through how to interpret the data. These graphs show “cumulative proportion” of patients with a “Ct” value from RT-PCR. When we talked about RT-PCR before, I mentioned that RT-PCR gives us these “Ct” values. A “Ct” is the number of RT-PCR cycles it takes to detect the presence of the virus genome. The higher the Ct, the more times you needed to duplicate the original RNA before you could detect it. So a higher Ct means less RNA is present.

Cumulative proportion means the graph is showing us what percentage of patients had a Ct value of X or less; when graphed this way, when you get to the extreme right of the graph, both groups should approach 1.0 (100%), because all of these patients being measured tested positive with at least some Ct value.

So what these graphs show us is the general level of RNA that was present in vaccinated and unvaccinated patients who tested positive for SARS-CoV-2 infection in the period 1-11 days after the first vaccine dose vs 12-28 days after the first vaccine dose. If the vaccine has an effect, you would expect more patients in the vaccinated group to have higher Ct values (less RNA) than unvaccinated patients, in that 12-28 day period. This “shift” of the two curves is present, and it tells us an important fact: once there is an immune response to the first dose of the Pfizer vaccine, that immune response appears to be capable of reducing the overall viral load in vaccinated patients who become infected.

This builds on what we looked at yesterday to illustrate a couple of things: (1) a single dose of the Pfizer vaccine does have some effect during this particular period between 12-28 days and (2) that effect manifests with a reduction in viral load.

This doesn’t necessarily tell me that one dose is “enough” or that it will make for durable immunity. But it does tell me that if you give a lot of people one dose of the Pfizer vaccine, you’ll start to see real population-level differences in the viral load in vaccinees. If we compare that to the UK data, where we saw impacts on disease, we start to get a really clear picture: the Pfizer vaccine can impact the viral load in patients who receive it, which is associated with a reduction in cases in the population, and is associated with a reduction in hospitalizations.

These effects can be seen, at least for a short time, with only one dose. They are likely reinforced by two doses.

All in all, we need more data, but this is all looking good. The vaccines are out in the world, doing their job. We might really beat this pandemic in a matter of months. Stay strong.

What am I doing to cope with the pandemic? This:

Cooking: a simpler, cooler stir fry

Image is a plate of chicken and snow peas on rice.

Now that I’ve gotten used to the wok, I’ve been trying to nail the fundamentals. There’s a temptation with cooking to overcomplicate things because it feels fancier, but I prefer to avoid that. “Elegant” is a synonym for “simple,” and I try to take that to heart. It’s better to cook a few ingredients well, particularly when you’re learning a new technique. At least, I think so.

Last night, I wanted to work on lower-temperature stir-frying, which takes better to sauces. When I say lower temperature, though, I mean “simply extremely hot” instead of “positively screaming hot.” The main difference is that the lower temperature is moderated by more liquid in the wok, which cools all the ingredients and allows development of a sauce.

I used 6 major ingredients in this, with some seasonings. The first was sliced chicken breast, which I marinated in soy sauce, honey, cornstarch, and a little vinegar. Truth be told, I should have used less vinegar.

I let this marinate about 30 minutes at room temperature. In the meantime, I chopped some garlic, scallions, ginger, and shiitake mushrooms (that’s 5 things now).

I started with the shiitakes and garlic in the wok, then added in the scallions once I started to see some nice color on the mushrooms and garlic. With the scallions broken down a little, I started to add the chicken. Most of the sauce developed from marinade absorbed by the chicken as it cooked out. And honestly, that was enough. I didn’t have to add anything else. Keep in mind that I didn’t add the leftover marinade itself—that’s not good to do, as you’ll get off flavors from the chicken that you wanted to leave behind. Instead, my sauce came from absorbed marinade that cooked out of my chicken.

For the chicken, I stir-fried it, adding it progressively, until there was a little color on each back. A bit of browning is what I want to see on chicken, but only just—let it go too far and you have slices of rubber.

Once the chicken was done, I added snow pea pods (ingredient 6) and let them get just to the edge of cooked. What I’m looking for there is the green color of the pods to brighten. When you cook vegetables, they get brighter green as the heat breaks down darker pigments in the vegetable. If you want a nice, crisp veggie, you take them to their brightest green and then get them off the heat. So, right as the green color started to go up on these snow pea pods, I tossed in the ginger I’d chopped and took everything off the heat right when that had become aromatic and the green color on the snow pea pods had started to peak. Here’s how that turned out, on a bed of white rice:

I have my critiques of this—namely the chicken needs more work—but if I’d ordered this from a restaurant I wouldn’t have been angry about it.

You might have some questions or comments! Send them in. As several folks have figured out, you can also email me if you have a comment that you don’t want to share with the whole group.

Join the conversation, and what you say will impact what I talk about in the next issue.

Also, let me know any other thoughts you might have about the newsletter. I’d like to make sure you’re getting what you want out of this.

Viral Transmissions

Discussion about this post