I work with Bandim Health Project in Guinea-Bissau and wrote the protocol for this trial.

I wrote the below response to Boghuma, who commented on my LinkedIn post about the trial, after which she did not respond. I leave the post here so that others can read more about the sound ethical basis of our work.

Thanks, Boghuma, for your input and for keeping the discussion scientific, and for raising the ethical dimension so clearly. I agree with you on the specific benefit of timely HBV birth-dose for preventing perinatal transmission, especially in high-prevalence settings.

Where we differ is regarding whether there is equipoise for the question this trial is actually asking.

1) The trial is not “HBV0 vs no newborn vaccines”

Both arms receive the currently recommended neonatal vaccines in Guinea-Bissau:

BCG + OPV.

The randomization is:

Intervention: BCG + OPV + HBV0

Control: BCG + OPV (no HBV0)

This is explicit in the protocol.

2) The key uncertainty is: what are the overall health effects in the first weeks of life, including the possibility of an interaction (positive or negative) with BCG and OPV. Show me the data/studies that has investigated this question please? I can tell you that they don't exist.

The data that does exist (including RCT evidence from our group in West Africa and from Uganda and India) is that BCG and OPV (both live vaccines) is evidence suggesting substantial beneficial non-specific effects on early infant survival. In Guinea-Bissau, BCG was associated with a 42% reduction in neonatal mortality, and in India, the reduction was 17%.

What has not been established is whether co-administering an aluminum-adjuvanted, non-live vaccine (HBV0) at birth with BCG and OPV has:

- no effect

- beneficial effects,

- negative non-specific effects,

on severe morbidity/mortality in the early window before the next routine vaccines at 6 weeks, and

whether there is an interaction with the co-administered BCG+OPV that could attenuate (or modify) those benefits.

It has never been tested whether BCG+OPV is better than BCG+OPV+HBV0 (or vice versa). We have no clue, and I can tell you that we have been surprised many times about the significance of vaccine interactions.

That is why the primary endpoint is a composite of mortality and severe infectious morbidity within 42 days.

In addition:

3) The ACIP/CDC review itself supports that this question is unresolved.

The ACIP rapid systematic review (Aug 2025) concluded that the evidence regarding effects of HBV0 on mortality is inconclusive and graded very low confidence, with differences across settings and signals in some LMIC data that cannot simply be dismissed.

That is, by definition, equipoise about overall health effects in the populations where infant infectious mortality is highest - the very setting where we should be least comfortable relying on extrapolation.

4) This is exactly why Africa needs RCTs - not permanent extrapolation

And Africa is deprived of RCTs, there are almost none, which has previously been highlighted as a major concern.

I do share your concern about mistrust - but I worry more about the alternative: that policies affecting millions of African newborns are implemented largely by extrapolating from very different contexts, without randomized evidence on all-cause outcomes and programmatic co-administration.

In global health, “we already know it works elsewhere” has too often replaced “we know what it does here”. We all need to be humble about what we know and what we don't know. I don't know what the results of this trial will be - I hope it will exonerate HBV0 showing no NSEs or even positive NSEs, after which I would be the most staunch advocate for its use everywhere. But until then, we simply don't have the data to establish that mass vaccination also of all those born to HBsAg-negative mothers is the best way forward.

5) Participant protection is built into the design

A few concrete safeguards from the protocol:

The trial is conducted under Declaration of Helsinki, with informed consent in local language, DSMB oversight, and sponsor insurance coverage.

Participation (and simply being born while we conduct the trial) will increase timely access to BCG+OPV, since we will also vaccinate in the weekend and on public holidays. Previous data on non-specific effects has documented that this results in benefit for all neonates regardless of allocation, in addition to the known beneficial specific effects. Half of the participants will receive HBV0, and thus be protected from the negative effects of hepatitis B, while the other half receives what they would probably have received, BCG+OPV, if not for the trial.

6) On the “current policy” point

Your Substack post argues that “current policy” is an ethical sleight of hand, which I agree can’t be a blanket justification.

But here, the ethical question is whether it is acceptable to randomize during a defined pre-implementation period to answer a major unresolved safety/overall-health question that:

A: is directly relevant to the imminent policy roll-out plus to newborns all over Africa and elsewhere where BCG+OPV+HBV0 is co-administered, and

B: has been highlighted as uncertain in the best available evidence synthesis.

That is precisely what equipoise means in practice.

If you’d be open to it, I’d genuinely welcome continuing the discussion on what the right evidentiary threshold should be for nationwide HBV0 roll-out in settings where BCG/OPV non-specific effects may be a major determinant of survival — and where the ACIP review shows the non-specific-effects evidence base for HepB is still low certainty.

I also recommend this paper regarding the implications of non-specific effects for testing, approving and regulating vaccines: pubmed.ncbi.nlm.nih.gov…