A new study points the finger at viral antigens.

It’s pretty clear at this point that myocarditis – inflammation of the heart muscle – is a complication, albeit a very rare one, of the mRNA COVID vaccines. The big question, of course, is why?

To date, it has been a mystery, like Glass Onion.

And in the spirit of all the great mysteries, to get to the bottom of this, we’ll need to round up the usual suspects.

Appearing in Circulation, a new study does a great job systematically evaluating multiple hypotheses linking vaccination to myocarditis, and eliminating them, Poirot-style, one by one until only one remains.  We’ll get there.

But first, let’s review the suspects.  Why do the mRNA vaccines cause myocarditis in a small subset of people?

There are a few leading candidates.

Number one – antibody responses. There’s two flavors here. The quantitative hypothesis suggests that some people simply generate too many antibodies to the vaccine – leading to increased inflammation and heart damage.

The qualitative hypothesis suggests that maybe it’s the nature of the antibodies generated rather than the amount – they might cross-react with some protein on the surface of heart cells for instance.

Or maybe it is driven by t-cell responses, which, of course, are independent of antibody levels.

There’s the idea that myocarditis is due to excessive cytokine release – sort of a like what we see in the multisystem inflammatory syndrome in children.

Or it could be due to the viral antigens themselves – the spike protein the mRNA codes for that is generated after vaccination.

To tease all these possibilities apart, researchers led by Lael Yonker at Mass General performed a case-control study. 16 children with post-vaccine myocarditis were matched by age to 45 control children who had been vaccinated without complications.

The matching was… ok… but as you can see here there were more boys in the myocarditis group, and the time from vaccination was a bit shorter in that group as well. We’ll keep that in mind as we go through the results.

OK let’s start eliminating suspects.

First – quantitative antibodies.  Seems unlikely. Absolute antibody titers were really no different in the myocarditis versus the control group.

What about the quality of the antibodies? Would the kids with myocarditis have more self-recognizing antibodies present?  It doesn’t appear so. Autoantibody levels were similar in the two groups.

Take antibodies off the list.

T-cell responses come next, and, again, no major differences here, save for one specific t-cell subtype that was moderately elevated in the myocarditis group. Not what I would call a smoking gun frankly.

Cytokines give us a bit more to chew on. Levels of Il-8, Il-6, TNF-alpha, and IL-10 were all substantially higher in the kids with myocarditis.

But the thing about cytokines is that they are not particularly specific. OK, kids with myocarditis have more systemic inflammation than kids without – that’s not really surprising. It still leaves us with the question of what is causing all this inflammation? Who is the arch-villian? The kingpin? The don?

It's the analyses of antigens – the protein products of vaccination – that may hold the key here.

In 12 out of 16 kids with myocarditis, the researchers were able to measure free spike protein in the blood – that is to say spike protein, not bound by anti-spike antibodies.

These free spikes were present in wait for it – zero – of the 45 control patients. That makes spike protein itself our prime suspect. J’accuse free spike protein!

Of course, all good detectives need to wrap up the case with a good story – how was it all done?

And here’s where we could use Agatha Christie’s help. How could this all work? The vaccine gets injected – mRNA is taken up into cells, where spike protein is generated and released, generating antibody and t-cell responses all the while. Those responses rapidly clear that spike protein from the system – this has been demonstrated in multiple studies – in adults at least. But in some small number of people, apparently, spike protein is not cleared. Why? It makes no damn sense. Compels me though. Some have suggested that inadvertent intravenous injection of vaccine, compared to the appropriate intramuscular route, might distribute the vaccine to sites with less immune surveillance. But that is definitely not proven yet. We are on the path for sure, but this is, as Benoit Blanc would say, a twisted web – and we are not finished untangling it. Not yet.

A version of this commentary first appeared on Medscape.com.