Demographic characteristics and immune response in the participants.
We obtained blood samples for antibody testing from a total of 225, 220, 211, and 210 vaccine recipients on days 7, 28, 60, and 90 post-1st shot, respectively (Table 1). Demographic characteristics of the participants are shown in Table 1. As of the time of enrollment, the average age of the participants was 41.8 years (range: 21 to 72 years), and 69.8% of the participants were female serving as a physician, nurse, paramedical staff, or administrative staff. None of the participants was in the immunodeficient state or was receiving immunosuppressants or steroids.
Table 1: Study protocol and demographic characteristics of the participants.
|
|
|
Day7
(Post 1st dose)
|
(%)
|
Day28
|
Day60
|
Day90
|
All
|
|
225
|
|
220
|
211
|
210
|
Age
|
20-39
|
97
|
43.1
|
92
|
84
|
84
|
|
40-59
|
110
|
48.9
|
110
|
109
|
108
|
|
≥60
|
18
|
8.0
|
18
|
18
|
18
|
|
(Average)
|
(41.8 y.o.)
|
|
|
|
Gender
|
Men
|
68
|
30.2
|
68
|
63
|
61
|
|
Women
|
157
|
69.8
|
52
|
148
|
149
|
Job
|
Physicians
|
36
|
16.0
|
|
|
|
|
Nurses
|
125
|
55.6
|
|
|
|
|
Others
|
64
|
28.4
|
|
|
|
We first determined the neutralizing activity against SARS-CoV-2 in serum samples taken on day 7 post-1st shot from 225 participants; however, none of the samples showed detectable neutralization activity (NT50 < 20-fold). We then determined the neutralizing activity in samples taken on day 28 post-1st shot from 220 vaccinated participants. As shown in Fig. 1A, NT50 levels were substantially diverse among the participants: the geometric mean of NT50 values was 375.2 (range 25.6–2680.0). Very low or no correlation of the NT50 values with ages was identified (Fig. 1A: Spearman’s ρ=-0.22; 95% CI -0.34 to -0.09). We also examined the levels of S1-binding-IgG and -IgM levels using the HISCL system that enables quantitative and highly sensitive determination of S1-binding-IgG and -IgM levels 21. The geometric mean of S1-binding-IgG values was 527.0 (range 44.6-3212.2), while that of S1-binding-IgM was 85.1 (range 10.3-1406.5). There was a high positive correlation of the NT50 values with S1-binding-IgG levels (Spearman’s ρ = 0.71; 95% CI 0.63 to 0.77) as examined on day 28 post-1st shot, while there was only a low positive correlation of the NT50 values with S1-binding-IgM levels (Spearman’s ρ = 0.43; 95% CI 0.31 to 0.53), suggesting that the neutralizing activity largely resides in IgG fraction of the serum of vaccinated participants around on day 28 post-1st shot (Figs. 1B and 1C). However, when examined on day 60 post-1st shot, the correlations of NT50 levels with both IgG and IgM levels became moderate or low (Spearman’s ρ = 0.56 and 0.32, respectively)(Supplementary Fig. 1).
The occurrence of adverse effects has no association with the BNT162b2-elicited neutralizing activity levels.
Commonly observed adverse events reported following BNT162b2 vaccination include injection-site pain, systemic fever, headache, and fatigue 10. In the present study, the events were observed largely more often following the 2nd shot (Supplementary Fig. 2) as previously reported by Polack et al. 10 Pains in the inject-site were reported by 67.6 and 61.6% of the participants and systemic fever (≥ 37.1℃) was reported by 3.6 and 46.4% of the participants following the 1st and 2nd shots, respectively. Since the severity of pains can be relatively more quantitatively rated than that of other adverse effects such as headache and fatigue, the possible correlate of the NT50 values with the severity of pains rated with the short form McGill pain questionnaire 22 was first examined. No correlation was seen between the NT50 values and the pain grades assessed following the 2nd shot (Spearman’s ρ = 0.14; 95% CI 0.00 to 0.26). The correlation was also negligible between the NT50 values and the incidence of systemic fever (Spearman’s ρ = 0.26; 95% CI 0.13 to 0.38)(Figs. 2A and 2B).
The average half-life of neutralizing activity in the vaccinees is approximately 67.8 days and the average time-length for their serums to lose the detectable activity is 198.3 days.
Considering that recent multiple clinical studies strongly suggest that the presence of high-level neutralizing antibodies is generally sufficient to confer protection against SARS-CoV-2 infection and that the protection against COVID-19 development is largely explained by robust SARS-CoV-2-neutralizing antibody response 8–10. If so, the once-established neutralizing antibody levels will decrease in time. We thus examined at what rate the levels of NT50 and S1-bindng-IgG and -IgM levels change by determining those levels from the data on day 28 (n = 220), day 60 (n = 211), and day 90 (n = 210) post-1st shot (Figs. 3A-C). The reduction of all NT50, IgG, and IgM levels from day 28 through day 90 post-1st shot was found to occur virtually linearly. By computation, the predicted average half-life of all the NT50 values turned out to be 67.8 days and those of S1-binding-IgG and IgM levels were 53.5 days and 43.6 days, respectively (Fig. 3D). The half-life of the NT50 values and that of S1-binding-IgG were reasonably comparable, corroborating that the neutralizing activity largely resides in the S1-binding IgG fraction. Based on the chronologically linear nature of the reduction identified, we attempted to extrapolate from such half-life values and tried to predict the average time-length for the serums of the participants having significant NT50 values to lose the activity down to under the undetectable level (UDL)(< 20-fold)(Fig. 3D). The predicted average time-length for the serums to lose the activity was computated to be 198.3 days, while that of the top 10% participants to lose the activity was 204.3 days. The time-length of the middle 10% (between the top 45% and 55%) participants to lose the activity was 187.6 days. For all participants, it was predicted that day 160 after the 1st shot was when the 80% lower limit of predicted NT50 levels drops under the detection level (UDL), while day 237 after the 1st shot was when the 80% upper limit of predicted NT50 levels drops below UDL. Similarly, for the top 10% participants, the estimated days dropping below UDL were 171 and 243 for lower and upper limits, respectively, while for the middle 10%, the estimated days were 155 and 224, respectively. As for IgG, the predicted time-length for the serums to get undetected was 346.8 days, and days reaching below UDL were 333 and 362 for lower and upper limits, respectively. Likewise, the estimated day reaching below UDL for IgM was 183.6, and the lower and upper limits were 170 and 198 days, respectively (Fig. 3D). We also asked whether the chronological decay rate of neutralization titers and S1-binding-IgG and -IgM differs among three age subgroups: (i) 20–39 yo, (ii) 40–59 yo, and (iii) 60’s and beyond. No significant difference was identified among the three age subgroups in the levels of neutralizing titers, IgG, or IgM levels (p = 0.60, 0.16, and 0.11, respectively: Supplementary Figs. 3A-C). The present data suggest that vaccinated individuals with good neutralization response would lose BNT162b2’s protection in 6 to 7 months without regard to age subgroups unless such people achieve robust immune boost response upon the future exposure to SARS-CoV-2. Otherwise, they should be protected by another booster vaccine shot or by other protective means.
Neutralization titers, S1-binding-IgG levels, and pain scores in the injection site were greater in women than in men.
We then asked whether there were differences between genders in neutralization activity levels, S1-bindng-IgG and -IgM levels, injection-site pain scores, and systemic fever grades. Statistically significant differences were identified in the levels of neutralization determined on 60 and 90 days post-1st shot (p = 0.002 and 0.002, respectively), S1-binding-IgG levels determined on 28, 60, and 90 days (p < 0.001, p = 0.001, and p = < 0.001, respectively) post-1st -shot, and S1-binding-IgM levels on 60 and 90 days post-1st shot (p = 0.025 and 0.044, respectively)(Supplementary Figs. 4A-C). The injection-site pain score was greater in women (p < 0.001) (Supplementary Fig. 4D). However, there was no difference in systemic fever grades between genders (Supplementary Fig. 4E). However, no difference was seen in the decline rates of neutralization activity, S1-binding-IgG and -IgM levels between men and women (Supplementary Fig. 5).
Some serums retain potent activity against various VOCs, but others showed substantially less potent or undetectable activity.
We finally asked whether the neutralizing antibodies elicited by BNT162b2 vaccination blocked the infectivity and replication of various variants of concerns (VOCs). To this end, we employed serum samples from 6 elite responders (NT50 values > 1,500-fold: the top 4% of all participants’ NT50 values as determined on 28 days post-1st dose) and serum samples from twelve randomly-selected moderate responders (NT50 values = 200 ~ 1,500-fold) and tested them for their inhibition of the infectivity and cytopathic effect of each variant in the VeroE6TMPRSS2 cell-based assay 23. As shown in Fig. 4, serums from the elite responders (n = 6) showed potent inhibition against SARS-CoV-205−2N (Wuhan strain, PANGO lineage B), while they showed less activity against SARS-CoV-2QHN001 and SARS-CoV-2QK002 (alpha), SARS-CoV-25356 (kappa), SARS-CoV-21734 (delta), and SARS-CoV-2TY8 − 612 (beta). Serums from moderate responders (n = 12) exerted less activity against the Wuhan strain than those from the elite responders. Some serums from the moderate responders also showed substantially low potency to all the VOCs tested. Notably, three serums from the moderate responders showed only marginal activity against SARS-CoV-2TY8 − 612 (beta). Two of those three samples had no detectable inhibitory activity against SARS-CoV-2TY8 − 612 (Fig. 4).