That's not what this says -- you're extrapolating that the vaccine somehow changed the immune response, when this letter is only providing data that people who already had Covid had a clear, ~immediate antibody response to vaccination. That's exactly what you expect to happen when immune people see an antigen for a second time.

In other words, the previously infected recipients already had a well-primed immune response. The vaccination isn't necessarily doing anything to increase the response -- it's just the expected reaction of an already-primed immune system, when it a bolus of antigens it already knows about is injected into the body:

> The antibody titers of vaccinees with preexisting immunity were 10 to 45 times as high as those of vaccinees without preexisting immunity at the same time points after the first vaccine dose (e.g., 25 times as high at 13 to 16 days) and also exceeded the median antibody titers measured in participants without preexisting immunity after the second vaccine dose by more than a factor of 6.

If anything, this letter (not a paper) provides further evidence that vaccination for the previously infected is not necessary.

I admit I'm not medically trained, but the graph [1] seems pretty clear to me, and the authors said:

> In contrast, participants with SARS-CoV-2 antibodies at baseline before the first vaccine injection rapidly developed uniform, high antibody titers within days after vaccination (median AUC before vaccination, 90 [43 participants];

>at 0 to 4 days, 133 [7 participants];

>at 5 to 8 days, 14,208 [15 participants];

>at 9 to 12 days, 20,783 [8 participants];

>at 13 to 16 days, 25,927 [20 participants];

>at 17 to 20 days, 11,755 [4 participants];

>at 21 to 27 days, 19,534 [14 participants]; and

>after the second dose, 22,509 [19 participants])

[1] https://www.nejm.org/na101/home/literatum/publisher/mms/jour...

Most notably, even those previously infected people with ~0 titers at baseline rapidly developed high levels of antibodies at 5-8 days after vaccination. It's a small group, but the rapid nature of the response is telling.

This data is clear that the infected people had a rapid, robust immune response to vaccination. It cannot tell you what you're trying to infer -- that the vaccination itself increased the immune response. To determine that, you'd need to do some other experiment.

I don't know of any study of the Delta variant (or other recent variants) comparing immunity acquired from infection versus immunity conferred by vaccination -- AFAIK, it's still unknown how these compare and whether a vaccination might provide additional immunity against variants. IIRC, the mRNA vaccines were developed to target key parts of the virus that are thought to be unlikely to mutate without decreasing virus viability.

We agree on this.

> I don't know of any study of the Delta variant (or other recent variants) comparing immunity acquired from infection versus immunity conferred by vaccination -- AFAIK, it's still unknown how these compare and whether a vaccination might provide additional immunity against variants.

I'm not aware of any study of this, either. So far, everyone is focusing on the vaccines.

On the contrary, from all that we know today, wouldn't we expect people with high IgA titres to even develop neutralizing immunity, at least in the weeks after titres peak?

Antibody titres aren't constant. They increase when your immune system sees an antigen, and decline gradually afterward. In individuals with previous exposure to the antigen, they increase more quickly due to immune memory. This is true for any virus. Decline in circulating antibodies does not mean that your immunity has gone away.

Vaccination is, to a first approximation, injecting a large amount of antigen into someone to induce an immune response. Therefore, data that shows that previously exposed people have faster and stronger immune responses to vaccination are simply confirming what we already know about immunity. It does not mean that the vaccination made the response in exposed people stronger than it would have been otherwise.

Maybe it did, but you can't conclude it from this kind of data.