As we await the results of placebo-controlled COVID-19 vaccine studies, what are we clinicians to do when our patients ask us whether they should get a booster this fall? What once was a no-brainer in the early days of limited immunity to the virus — and the spectacular results of the first placebo-controlled trials — is now a much more nuanced clinical question. Essentially 100% of the population has had either COVID or been vaccinated or both, the disease is on average less severe, the vaccines do have side effects — and who wants another shot anyway?
Now, along comes one of the more clever (cleverer?) observational studies to date addressing this question.
Published here in the New England Journal of Medicine, researchers from the U.S. Department of Veterans Affairs evaluated the 2024–25 COVID-19 vaccine among more than 160,000 veterans who received it on the same day as their flu shot, compared with 130,000 who received the flu shot alone. Participants were followed for six months.
This inclusion of the flu shot only group was absolutely critical, as it diminished the inevitable bias of many observational vaccine studies — which is that those who choose to be vaccinated may intrinsically have more healthy behaviors overall, and that these differences lead to better outcomes, not the vaccines. It’s a form of confounding that no amount of statistical pyrotechnics can completely remove.
Indeed, the “Table 1” of the study, showing baseline characteristics, showed that the group getting the COVID-19 vaccine was quite similar to those who did not, even before statistical adjustment. Hooray. The study included a very typical VA population: age in the early 70s, over 90% were men, and about 18% smoked.
The results?
- 29% lower risk of COVID-related emergency department visits
- 39% lower risk of hospitalization
- 64% lower risk of death
That’s impressive, and I found the mortality results particularly surprising. However, here’s where we need to read the fine print. The relative risk was reduced, but the absolute risk was really small — even in this older patient population. If you’d like a much more detailed review of the paper, I highly recommend Dr. Adam Cifu’s excellent summary, in which he calculates a “number needed to vaccinate” to reduce these endpoints:
Effectiveness is 1-RR, so the relative risks ranged from 0.36 to 0.72 … We can calculate numbers-needed-to-vaccinate (NNV). These come out to 546 for a COVID–associated emergency department visits; 1338 for COVID–associated hospitalization; 4545 for COVID–associated death; and 549 for a composite of these outcomes.
These are pretty big numbers, and would be even higher in a young, healthy cohort. And indeed, if we want a “take-away” message from the study, it’s that we should continue to recommend the COVID vaccine for our older patients, especially those with medical comorbidities.
For a healthy 50-year-old? If they want it and didn’t have side effects from previous vaccinations, I see no contraindications — but it would be inappropriate to use these data to make a broad recommendation that literally everyone should receive an annual COVID vaccine. For that message, the upcoming clinical trials will be informative about other important endpoints, such as the incidence of COVID infection, sick-days averted, and adverse effects.
Separately, you may have seen media coverage of another recent paper — this one in Nature — reporting that mRNA COVID vaccines might somehow “sensitize” tumors to immune checkpoint inhibitors. The idea is intriguing: that vaccination could trigger innate immune pathways, leading to greater tumor responsiveness to immunotherapy. That’s a result that earns splashy headlines and a few hopeful social media posts about mRNA vaccines “helping treat cancer.”
But let’s be clear — this is an early, mechanistic, and hypothesis-generating study, not a clinical trial showing that vaccination improves cancer outcomes. It’s tough-going for non PhD’s to get through the methods and results, which means that most of us clinicians will not be able to critique it appropriately. But at this stage, we can say that the proposed biologic mechanism is speculative, the human data are retrospective, and the observed survival benefit could be confounded by which patients were well enough to receive a vaccine.
So interesting idea, yes! But not a reason to recommend COVID vaccination for cancer therapy or prevention. For now, we should stick to the data that are clinically relevant, as came from the VA study: vaccines continue to reduce serious COVID outcomes in older and high-risk individuals, even if the absolute benefit is smaller than it once was.
Note to readers: My dog Louie’s death (sniff) prompted me to write some long-form pieces for a newsletter, called PaulSaxMD. (I’m not so good with clever titles, alas.) It will include way less ID esoterica and much more personal stuff. Hope you subscribe!
Cost to you? Same as this NEJM Group blog, what a bargain.
You can follow any responses to this entry through the RSS 2.0 feed.
You can skip to the end and leave a response. Pinging is currently not allowed.
The other big takeaway from that VA study is the apparent (relatively) rapid drop off in vaccine efficacy.
If you want to maximize benefit, you need to time booster vaccination of your at risk population relative to the next COVID-19 wave. Doable for self/family; very hard to accomplish at pop health scale with complicated current media/comms environment.