Suuuuure looks like us Johnson & Johnson recipients will have to wait longer for a blessing from the FDA for boosters we’re all but certain to need at some point. The New York Times reports that the preliminary review of J&J’s application for an emergency-use authorization on boosters got a less than enthusiastic analysis from the FDA committee reviewing the clinical data. The report calls into question the methodology employed, mainly in the sensitivity of the measures used to calculate the immune response.
Is this enough to torpedo the EUA?
In a new analysis, the Food and Drug Administration questioned the strength of evidence Johnson & Johnson provided in its application for booster shots. A key test used by the company was likely not sensitive enough, the agency suggested, adding that it didn’t have enough time to independently review much of the raw data from the trials.
The document, released in advance of a Friday meeting of the agency’s vaccine advisers, could have significant influence on whether the 15 million Americans who have received the one-dose vaccine will be allowed to get a second shot, or if they will instead be urged to get a different brand of vaccine for added protection.
The agency’s analysis follows a report released on Tuesday, in which Johnson & Johnson argued in favor of a booster, presenting data from a number of trials.
“A booster dose is recommended at 6 months or later, based on the strength of the immune responses,” the company wrote.
But on Wednesday, the F.D.A. said that the test used by the company to measure the immune response of a six-month-boost — known as a psVNA assay — is not sensitive enough for the task. The agency also questioned whether the increase in immune response was as big as the data suggested.
The report actually found that the data better supported a booster shot at “2-3 months after dose 1” than it did at the six-month mark:
Immunogenicity of dose 2 (5×1010 vp) administered approximately 2-3 months after dose 1
• In studies COV1001, COV1002, and COV2001, the geometric mean increase in neutralizing antibody titers against the reference strain (Victoria/1/2020), as measured by a validated wild-type virus neutralization assay, at 14-28 days post-dose 2 was approximately 1.5- to 4- fold above the pre-dose 2 baseline. These analyses, which were limited by small sample sizes (N=~25-50 samples per time point), suggest that a 3-month interval may result in a more robust neutralizing antibody booster response than a 2-month interval.
Immunogenicity of a 5×1010 vp booster dose administered 6 months after primary vaccination
• In study COV1001, the geometric mean increase in neutralizing antibody titers against the reference strain (WA1/2020 with D614G mutation), as measured by an exploratory (nonvalidated) pseudovirus neutralization assay (psVNA) at 7 and 28 days post-booster, was 6.8-fold and 10.5-fold above the pre-booster baseline, respectively. The geometric mean increase in neutralizing antibody titers against the Delta variant, also as measured by an exploratory (non-validated) psVNA at 7 and 28 days post booster, was 2.2-fold and 3-fold above the pre-booster baseline, respectively. These analyses were limited by small sample sizes (N=15-17 samples per time point), and titers from 28 days post-primary vaccination and pre-boost timepoints suggest that the assay may have low sensitivity.
Only 17 people got the booster? Why the small samples? If one reads further into the report, the studies J&J cited for evidence of their booster shot are almost all Phase 1 or 2 studies. One, however, is a Phase 3 study (COV3009), but that only studied a two-month interval for the booster. That one did have 30,000 participants, and certainly seems to be impressive, although not enough of the sample came into contact with Delta to measure the response to that variant:
Efficacy of two doses (5×1010 vp each dose) administered 56 days apart
• In study COV3009, an efficacy analysis of blinded, placebo-controlled follow-up (median of 36 days post-dose 2) among approximately 30,000 adult participants suggests a VE of 75.2% (95% CI: 54.6%, 87.3%) against molecularly confirmed moderate and severe/critical COVID-19 from 14 days post-dose 2 and a VE of 100% (95% CI: 32.6%, 100%) against molecularly confirmed severe/critical COVID-19 from 14 days post-dose 2 (estimated from only 8 cases in the placebo group and none in the vaccine group). In subgroup analyses of efficacy by country, the VE estimate against moderate and severe/critical COVID-19 in the U.S. was 93.7% (95% CI: 58.4%, 99.9%). The confidence interval for this U.S. VE estimate overlapped with that for the overall VE estimate post-dose 2 in study COV3009 and with the confidence interval for the U.S. VE estimate following a single dose from the final efficacy analyses in study COV3001 (VE estimate of 72.9%; 95% CI: 65.7%, 78.7%). Insufficient cases of COVID-19 caused by the Delta variant were identified to estimate VE post-dose 2 against this variant.
That’s not to say it won’t be effective against Delta, but that it’s simply too thin to reach any conclusions on that question. However, the FDA report notes that the real-world study included here on single-dose regimens show no decline in effectiveness against Delta:
Estimates of vaccine effectiveness from the real-world observational study, COV4002, are generally consistent with VE estimates from study COV3001 and do not suggest decreased protection over time or coincident with circulation of the Delta variant in the U.S.
So here too, this looks like a process issue rather than an effectiveness issue. The COV3009 study also allowed the committee to conclude that there are no further safety issues presented by a second shot. They do caution that follow-up has only been in place for 36 days, so there may be more issues, but that seems unlikely for people who tolerated the first dose. If significant side effects were to present themselves, onset would likely have been immediate.
What to make of this? It seems as though J&J may have been premature on the six-month booster application, but … not on the two-month booster. And if that’s the case, one would wonder why the FDA would balk at a six-month booster but approve a two-month booster, if in fact they do at all. Other than not having the specific Phase 3-scale study in hand, why not authorize both? If it’s effective and safe at two months, it seems very unlikely that a booster at six months would be ineffective or unsafe.
Given all this, it seems more unlikely that the FDA will issue an EUA on Friday (or more accurately, sometime after the Friday meeting) for a J&J booster. Once again, though, we should ask why an already approved vaccine should need another blessing from the FDA to access a booster shot, especially when the studies all show no ill effect or safety issue in the data from a second shot. That seems to be a question for doctors and patients rather than the FDA, especially in a pandemic environment and for a vaccine that looks like it probably should have been a two-shot regimen all along. Stay tuned.