Thread by @michaelzlin

Another study just posted comparing Abs & T cells in #JnJers after J&J, Moderna100, or Pfizer boost

Confirms NIH and Singapore studies: Both Abs and T cells are boosted by Moderna100 > Pfizer > J&J

#JnJers should boost accordingly

h/t @ChildrenNeedUs_

medrxiv.org/content/10.110…

Booster happened at 3mo after J&J vaccination.

Fig 2A: Antibody levels after boost go as Moderna100 > Pfizer > J&J

Moderna50 would be in between Moderna100 and Pfizer, most likely

Fig 2B: p values were not done for differences in post-boost antibody levels by booster group, but for diffs in fold change within individuals by booster group. Subtle distinction, but the latter is more variable (boxes and whiskers are taller). Anyway all diffs significant.

Why would fold change in antibody levels after boosting be a more variable measurement than antibody levels in the study population? Probably because if you have lower antibodies preboost you can have a big change, and if you have higher levels preboost you may get less change.

Fig 2C: neutralizing antibody levels. ULoD = upper limits of detection; Moderna and Pfizer effects are literally off the charts

Fig 3A: Here they use IFN as a measure of T cell activation by spike protein. The Moderna100- and Pfizer-boosted T cell activity levels are higher average than the J&J-boosted level. Also a larger fraction of the J&J-boosted failed to mount a response above the positive threshold

Note the p values (numbers under the gray boxes) are all small, so all diffs are "stat sig". Doesn't mean the results are equally good in clinical or biological significance. P values provide a measure of confidence a difference exists, not the usefulness of the difference

Fig 3B: p values calculated not for diffs between boosters in post-boost T cell levels, but for fold change within individuals by booster (which again looks more variable). J&J vs Moderna100 booster is significant at the 5% level; J&Js v Pfizer booster is not.

Conclusions: 1) If you're J&J, boosting with Moderna100 or Pfizer gives higher Ab levels than boosting with another J&J. Moderna50 should be in between Moderna100 and Pfizer. 2) Pretty clear that boosting with Moderna100 or Pfizer gives more T cells than boosting with J&J too.*

*Some might try to argue that "there's no evidence that Pfizer boost is better than J&J boost for T cell function" because of Fig 3B. But this high-variance metric still trends Pfizer > J&J. Also in Fig 3A, J&J + J&J has a higher rate of no detectable T cell responses

Real-life implication: If you're J&J, you're best off boosting with Moderna50 (by interpolation), followed by Pfizer, then J&J for both antibodies and T cell function. This confirms our earlier conclusion from the the NIH and Singapore studies

https://twitter.com/michaelzlin/status/1450988678386372609

P.S. noting again that lower effectiveness of J&J booster vs Pfizer booster here may be partly due to anti-adenovirus antibodies from the initial J&J vaccination clearing some particles from the J&J booster before they can infect cells