So what’s the side effect that caused the J&J vaccine pause?
Blood clotting – but not the usual kind. This appears to be the same (or very similar) to the problem seem with the Oxford/AstraZeneca vaccine in Europe, and both are very similar to a known syndrome called heparin-induced thrombocytopenia. That involves unusual binding to a blood protein, platelet factor 4, and it occurs in rare patients when the blood thinner heparin is administered (leading to blood clots instead, the opposite of what it’s supposed to do). Something about the vaccine (again, in rare patients) is leading to the same kind of PF4 effect.
Is there treatment for this once it happens?
Definitely. There are a lot of blood-thinning compounds out there that work through several different mechanisms. Most of the people who have shown this syndrome recover, but the key is recognizing it and treating it the right way. When a patient presents with symptom of a blood clot, one of the first-line therapies is to give them heparin, but if they have this syndrome going on, heparin will do nothing but make the clotting worse. A signature of the condition is that blood clots are forming when the patient has low platelet counts (thrombocytopenia). Most of the time when a patient presents with that in their blood work, they’re at increased risk of throwing blood clots, because their platelet count is low for some other reason. But in these cases of heparin-induced (or vaccine-induced) thrombocytopenia, it’s different – the platelets are getting tangled up in this odd protein binding event involving PF4, making their count artificially low.
So the key for people getting these vaccines is to be aware of the symptoms of this blood clotting problem, and for physicians to know how to treat it appropriately (not heparin!) It seems to take several days to come on, so people need to keep an eye out, but on the other hand, if someone has had the J&J or AstraZeneca/Oxford vaccine weeks ago and shown no problems, they should be fine.
How many people are getting this? What are the odds?
A key question, but one that’s still coming into focus. It’s complicated by the fact that we have surely missed some incidences of the syndrome, and by whether or not some people are more at risk than others. It’s possible that younger women have a higher incidence rate, for example, but that could change as we learn more. Likewise, older patients don’t seem to be at higher risk, from what we’ve seen. Roughly, the incidence might be around the case-per-hundred-thousand-patients level, but that’s a real back-of-the-envelope number, and it will change. Obviously, this incidence rate is going to be a major factor in figuring out how to manage the whole problem and its effect on the vaccination plans. Reports from various European countries on the AZ/Oxford side effects have not always lined up, so we don’t know yet how the J&J ones fit into the picture.
Why didn’t we see this during the clinical trials? Doesn’t it mean that things were rushed too much?
We didn’t see this because no one has ever run a clinical trial large enough to get any good statistical read on such rare events. Remember, the Phase III trials (the largest ones) involved about 20,000 patients getting vaccinated versus another cohort of thousands as controls. Now, for a clinical trial that’s a huge number, but it’s still not going to be powered to catch the per-hundred-thousand (or per-million) levels of side effects. As I wrote here the other day, that goes for all drugs, not just vaccines (and there are a lot of drugs that never hit those 20,000-patient clinical trial levels, either). This is why there’s a whole field called “pharmacovigilance” (PV), which involves watching closely to see what happens after a drug is approved and goes out to a far larger number of people than it has ever seen before.
As for the idea of rushed trials, they were indeed fast, but some of that speed was because we immediately ramped up to the high levels of patients right after Phase II. A development program over several years might have picked up hints of this clotting problem, but on the other hand it might well not have – and if you’d seen one case of blood clotting it would likely not have been enough to stop things. Recall that there was a case of transverse myelitis in the AstraZeneca vaccine trial that stopped it for a while in several countries, but there has been no increase in that since the vaccine was rolled out. Single rare instances of things like this are a real headache in a trial, because many of them are indeed “just one of those things” and not necessarily related to your drug candidate. All kinds of odd stuff just happens to people out of the blue at a very low level, which is something that I don’t think enough people outside of the clinical sciences realize.
It’s worth remembering, though, that all drugs can have very low-incidence side effects in people and that these can be difficult-to-impossible to predict. Most people, for example, can take penicillin antibiotics without incidence, but some people have strong anaphylactic reactions (all the way up to death). Similarly, a few people are people are allergic to aspirin, and this holds for many other drugs. You’ll notice a common thread of immunology here, which also ties in with rare events like Reye syndrome and Guillain-Barré. In that context, it’s worth remembering that the seasonal flu vaccine itself probably leads to about one case of G-B syndrome per million people vaccinated, but that’s on top of several thousand cases that just happen anyway.
Is there any way to tell up front who might be at higher risk?
You’d hope so, but we don’t know yet. No one’s ever been able to pick out who might be at higher risk of the heparin-induced thrombocytopenia that this resembles, for example. Knowing which patient groups might be at higher risk in general would be even more helpful – you could just recommend a different type of vaccine for them without having to give every single person a blood test or something.
Did the CDC and other authorities make the right call here?
We’re not going to be able to settle that question very easily. They were put in a more or less impossible position, one that happens pretty often in drug regulatory matters. You either end up looking like you’re killing patients by letting unsafe therapies through, or looking like you’re killing patients by denying them useful ones. “Unsafe” and “useful” are both words with a lot of sliding scales hiding inside them. For now, I think I can safely scorn the two far ends of the response to this issue, that is “Disband the FDA and the regulatory state!” and the “Ban everything that causes any level of harm!” position. There’s a hell of a lot of territory between those two.
Is this blood clotting happening with the mRNA vaccines, too?
No. That seems quite clear – to the best of my knowledge, there have been no reports like this at all with either the Moderna or the Pfizer/BioNTech vaccine. That’s good news, and it tells us a lot. For one thing, this blood clotting problem is not a general feature of trying to vaccinate people against the coronavirus. It also means that it apparently has nothing to do with inducing the coronavirus Spike protein in people, since that’s what both the adenovirus vectors and the mRNA vaccines are doing, in the end.
The version of that protein brought on by the AZ/Oxford vaccine is slightly different from the others (it doesn’t have a key set of protein-stabilizing mutations), and when the clotting problems showed up in Europe some people were wondering if that had anything to do with it. But the appearance of such side effects with the J&J vaccine would seem to rule that out. Instead, what those two have in common is that they’re both adenovirus vector vaccines. Oxford used a chimpanzee adenovirus, and J&J picked a less-common human one. Which means that if this is a side effect shared by adenovirus vectors, it’s shared at a pretty basic level, isn’t it? I’m not enough of an adenovirus jock to tell you in detail about the similarities between the proteins in the ChAdOx vector versus Ad26, and at any rate it’s probably more about the antibody response to these things (and why, in a small number of people, that goes awry with the PF4 protein).
What does that mean for adenovirus vectors, then?
It complicates things, for sure. As with all side effects, it depends on what disease you’re treating and in which population. Vaccines are in an unusual category here, since they’re being given to a vast population of people who are not sick yet – pretty much the opposite of a normal drug rollout! At the other end of that scale, as mentioned here the other day, no one would think twice about this level of side effect for treating an otherwise incurable disease with a high death rate (pancreatic cancer or the like).
J&J, for one, has been hoping to leverage their Ad26 platform into a number of other vaccines and therapies, and I believe that Oxford has similar hopes for their adenovirus vector as well. There’s always been a cloud on that horizon, in that you wonder if people who are treated with such vectors for one condition will have too high an antibody response to the adenovirus itself to be easily treated again. The mRNA platform doesn’t have as much to worry about in that regard – the lipids involved in the formulation don’t seem to set off an immune response of their own (from what I’ve seen, although there could always be some rare exceptions), and they don’t have the rest of the viral-vector machinery as immunological baggage.
This potential clotting effect is another cloud. One of the features of the vaccine response to this pandemic has been that we were finally going to see large-scale human dosing of these two platform technologies, head to head, and at least at the moment, it looks like mRNA is winning out. In this case, it looks like it’s (1) faster to deploy, (2) has manufacturing that is not dependent on the vagaries of cell culture in its crucial steps, (3) may be somewhat more efficacious, at least in this case, and (4) so far has a better safety profile. The competition is just beginning, but so far the mRNA tech looks to be ahead.
What does this mean for vaccinating the world?
More complications, unfortunately. One of the weak points of the mRNA vaccines is that the current ones, anyway, can have more demanding cold chain storage requirements. The logistics have been working out well here in the US, but they’re not going to work out so well in many other regions of the world, and the adenovirus vector vaccines have been more promising for wide distribution. We definitely could have done without something that slows down their rollout and adds to the uncertainties about who should be dosed – or adds to the uncertainties that many people might be feeling about getting vaccinated at all.
There are several ways out of this. One could plow ahead and just vaccinate away, side effects be damned, but since there are cleaner alternatives that doesn’t seem like it’s going to be as popular a strategy. As I mentioned in the previous post linked above, if there were no other vaccine alternatives that’s exactly what we would be doing right now – but that’s not our situation. It doesn’t help that the regions of the world that would have the most difficulty with the mRNA cold chain distribution are generally the ones that would have the most difficulty doing the pharmacovigilance for blood clotting problems and subsequent treatment.
If we can narrow down the at-risk patient groups for the adenovirus vaccines, that would help a lot. If there are other ways to store and distribute the mRNA vaccines, that would similarly help. And we could also use more vaccines – I would particularly point to Novavax’s recombinant-protein candidate as another one that apparently has easier storage and distribution requirements and already has some clinical efficacy data to make its case. There are plenty of other interesting and potentially very useful candidates in the works (more on this in a coming vaccine roundup post), but they’re further behind and time is tight.