The FDA's approval of Pfizer's maternal RSV vaccine has omitted key data and is problematic

US FDA's lax approval standards jeopardize trust in medicine

In this guest post, Dave Allely, medical student, takes a deep dive into the data around Pfizer’s maternal RSV vaccine. There is no question that if kids are better off overall with the vaccine, we ought to give it. Perhaps the decision varies between women who want to breastfeed vs those who don’t, but net benefit is the key question. The problem is that the FDA and Pfizer have not provided data to assess net benefit. For instance, we want to know if the vaccine lowers all cause hospitalization or at least all hospitalizations for lower respiratory tract disease. And yet, these data are not given. Dave is a thoughtful student of evidence based medicine and fleshes out these concerns.

Vinay Prasad MD MPH

Professor in residence

The FDA formally approved Abrysvo, Pfizer’s maternal RSV vaccine product, on August 21st, 2023. The data supporting approval are problematic, specifically:

• With respect to the efficacy of this product, Pfizer has not been required to provide: 

  • All-cause data regarding severe lower respiratory tract disease (LRTD)

  • All-cause data for LRTD hospitalizations

  • All-cause data for hospitalizations

• With respect to effectiveness, Pfizer has claimed that breastfeeding had no effect on vaccine efficacy (VE), but to my knowledge has not been required to provide this data.

• With respect to safety, there is a serious concern regarding pre-term births, which the FDA has attempted to mitigate by tightening the window of administration to 32-36 weeks gestational age (more on this later). 

Note: For any who missed my earlier posts and are interested in the details, here I critically appraise the MATISSE trial, on which this FDA approval is based. Here I lament the FDA advisory committee’s decision with respect to this approval. As a side note, LRTI and LRTD have been used interchangeably in this space, one meaning “disease”, the other “infections”, LRTD is a better term, given patients had to present with symptoms in order to be tested.

I should acknowledge that my faith in the rigor behind this kind of approval process was misplaced. When I originally appraised MATISSE, I assumed Pfizer would be required to provide all-cause data regarding each of the endpoints in the trial. I don’t understand the thought process that led to unanimous “yes” votes with respect to efficacy in the absence of these data. 

FDA attempt to mitigate safety risk

In the MATISSE trial, Abrysvo was given to pregnant women from 24-36 weeks gestational age. Pfizer sought approval for this product to be given within that same range of dates. The FDA approval restricted the window to 32-36 weeks gestational age. While I appreciate the attempt to limit the downside of this vaccine, it is not enough. 

In MATISSE, the median gestational age at vaccination was 31 weeks and 2 days. That means less than half of the women in the study received this product in the timeframe that it is now approved. The package insert shows preterm births remain elevated in this subset of the population:

While this 0.5% absolute risk increase is better than the 1% increase in the overall population, we are now relying on an increasingly small number of individuals to assess the safety of this product. As a side note, Tdap is recommended in pregnancy between 27-36 weeks gestational age. I have not evaluated the evidence behind that decision but it is worth acknowledging, so I will also just leave this here: 

What is an obstetrician to make of this? 

How does the 32-36 week time-range impact VE? 

Changing the window of vaccination could of course impact vaccine efficacy, and we have some data regarding this:

Vaccine efficacy against hospitalization due to RSV for 32-36 wks gestational age was not provided in the package insert. Zero all-cause data was provided. The above data were not adjusted for multiplicity, but for the sake of argument I will take them at face value. The burden of severe LRTD caused by RSV in the first 6 months of life was 1.6%. With Abrysvo, that risk was reduced to 0.4%. That corresponds to an absolute risk reduction ~1.2%. 

Final Thoughts and Looking Forward

The FDA has approved this product for all pregnant women. This approval is based off of a post-hoc subgroup analysis of a single phase 3 trial. High risk pregnancies were not included, and a major risk factor for severe RSV infection is pre-term birth. The risk of an increase in pre-term birth remains, and the data regarding all-cause LRTD burden have not been provided. This approval risks too much and stands to provide questionable benefit. 

Now we move forward into the land of observational trials, where we can pat ourselves on the back as the healthy user bias confirms our suspicion that these interventions are beneficial. Maybe one day we will get to see the all-cause data from MATISSE. 

---