Demographics of and reported breakthrough infections in the study population
This study continued the follow-up of humoral antibody responses of healthcare workers (HCWs) who received two BNT162b2 vaccine doses with short three-week interval (n=230, short group) or two doses of BNT162b2, mRNA-1273, or ChAdOx1 plus BNT162b2/ mRNA-1273 with a long 12-week (n=202, long group) interval before the third vaccine dose. We have previously analyzed this cohort up to 3 months after the third vaccine dose7,8,10. The novel data obtained in this study is the analysis and follow-up of serum samples up to 9 months after the 3rd vaccine dose. In Finland, COVID-19 vaccinations were first carried out with a short three-week interval, and later in the spring of 2021 the interval was lengthened to 12 weeks. Thus, the vaccinees in the short interval group received their third dose before those in the long interval group. Our analysis included serum samples collected before the first vaccination and before the third vaccination up to nine months after the third vaccine dose (short and long interval groups) (Figs.1a and 2a). The age range of the participants at the time of the first vaccine dose ranged from 20 to 65 years (mean 42.7 and median 41.2 years) in the short interval group and 2267 years (mean 45.7 and median 46.3 years) in the long interval group (Table1). Eighty-nine percent of the study participants were females.
Six of the vaccinees in the short interval group and fifteen in the long interval group had had a SARS-CoV-2 infection (confirmed by PCR) before the first vaccination. The number of PCR/antigen test-confirmed SARS-CoV-2 infections increased in the study population (and in Finland) after the emergence of the Omicron variants. The majority of the vaccinees had received their third dose by the end of 2021, and thus, the breakthrough infections occurred mostly after the third vaccine dose: HCWs in the short interval group reported 98 infections after the third dose and HCWs in the long interval group reported 10 infections after two doses and 80 after three vaccine doses (Table1).
To estimate the changes in humoral immunity after vaccination with three doses of COVID-19 mRNA vaccines and potential exposure to circulating SARS-CoV-2 variants, we analyzed the changes in SARS-CoV-2 spike protein subunit 1 (S1) and nucleoprotein (N) specific IgG antibody levels in the sera of HCWs in the short interval group (Fig.1, the timeline of the vaccinations, serum collections and circulating variants is shown in Fig.1a). The third vaccination induced strong S1-specific IgG responses, which peaked at 3 weeks post the vaccination and decreased thereafter (Geometric mean of antibody levels 128.9 EIA units at 3D3wk, 99.96 EIA units at 3D3mo, 93.19 EIA units at 3D6mo, and 99.73 EIA units at 3D9mo; Fig.1b). The reported (PCR or antigen test confirmed) or serologically observable (increase in S1-specific or N-specific antibodies greater than the cut-off, 4.8 or 8.8 EIA units, respectively) SARS-CoV-2 infections increased the S1-specific and N-specific antibody levels (Fig.1b, e). Follow-up of the vaccinees showed that the majority of infected vaccinees had an increase in anti-S1 and anti-N antibodies after the time period of a reported infection (red lines in Fig.1c, f).
Five vaccinees in the short interval group had SARS-CoV-2 N-specific antibodies prior to vaccination8 (Fig.1e): Two of these had had a PCR test confirmed SARS-CoV-2 infection before the vaccination while the other three had elevated but stable anti-N antibody levels without a sign of antibody decline. Grouping of the vaccinees based on their infection status (Fig.1d, g) showed that the anti-S1 and anti-N levels were significantly different between HCWs with or without a breakthrough infection at 6 months after the third vaccine dose (non-infected 72 EIA units vs. infected 112 EIA units for S1, p<0.0001 and 1.4 EIA units vs. 22 EIA units for N, p<0.0001) after the time point of three months after the third vaccine dose. In some vaccinees, the infection-induced antibody response was delayed and, therefore, was detectable only in the later time points, while 13 infected vaccinees showed no or little changes in S1-/or N-specific antibodies (Fig.1c, f).
HCWs with a long vaccine interval received two doses of BNT162b2, mRNA-1273, or a combination of ChAdOx1 and BNT162b2/mRNA-1273 as the first two doses (Table1). Sequential serum samples were collected before the vaccination and at regular intervals after each vaccine dose and analyzed for S1- and N-specific IgG antibody levels (Fig.2, the timeline of the vaccinations, serum collections and circulating variants is shown in Fig.2a). The first vaccine dose induced a wide range of S1-specific antibody levels, and substantially higher levels in previously infected HCWs (Fig.2b, c, black dots). While the second and the third doses increased S1-specific antibody levels higher, follow-up of the antibody levels of each vaccinee showed increases also before the third dose and after the first time point post the third vaccine dose (Fig.2c). A vast majority of these increases coincided with a reported SARS-CoV-2 infection (PCR test or antigen test confirmed) preceding the sampling and most of these HCWs also had an increase in N-specific antibodies at the same time (Fig.2e, f).
A separate analysis of antibody responses in individuals that had or had not contracted an infection (PCR or antigen test confirmed, or serologically observable) showed a decrease (Geometric mean 115.4 EIA units at 3D3wk vs 32.94 EIA units at 3D9mo) in anti-S1 antibody levels prior to the next vaccine dose in case there was no infection (p<0.0001; Fig.2d). A SARS-CoV-2 breakthrough infection increased the S1-specific antibody levels when responses of infected and uninfected were compared three and six months, respectively, after the third dose (Geometric mean 82.30 EIA units at 3D3mo vs 130.6 EIA units at 3D3mo infected, 55.02 EIA units at 3D6mo vs 130.6 EIA units at 3D6mo infected) p<0.0001; Fig.2d). Also, the N-specific antibody levels were induced by an infection, although the N-specific response was absent or low in 25 of the infected vaccinees (Fig.2eg). A decline in N-specific antibodies was observed in the sequential samples from participants infected pre-vaccination or post-vaccination while a few vaccinees had relatively high basal N-antibody levels without an indication of a decline in antibody (blue dots above cutoff value, Fig.2f).
Of the 412 HCWs who had serum samples available after the third COVID-19 vaccine dose, 44% (182/412) reported a positive COVID-19 test (PCR or antigen) post third vaccine dose. Furthermore, 95% (173/182) of the COVID-19 test-positive HCWs showed a serological indication of infection, defined as a diagnostic increase in either S1- or N-specific antibodies (increase greater than cut-off, 4.8 EIA units for S1 after the time point of 3D3wk, and 8.8 EIA units for N after the third dose; Table2). Of the COVID-19 test negative (or untested) HCWs, 13% (30/246) had a serological indication of infection by S1- or N-specific antibody levels, making the proportion of HCWs with a breakthrough infection, confirmed either by COVID-19 test or serology, after three vaccine doses 51% (212/412). Of note, measuring only S1 or N-specific antibodies was less sensitive at detecting infected individuals, since only 64% (117/182) of the COVID-19 test-positive HCWs had an increase in both anti-S1 and anti-N antibodies. Thus, serological detection of past infection was most accurate when IgG antibodies for both S1 and N antigens were positive.
We also examined the antibody levels of HCWs and compared those to their COVID-19 test and serological status. We found that infection soon after a COVID-19 vaccination was more likely to lead into lack of serological indication of infection. Seventeen per cent (10/60) of the HCWs who had a short interval between the third COVID-19 vaccine dose and a positive COVID-19 test (vaccination less than three months before a positive COVID-19 test, Supplementary Fig.1) had no serological indication of an infection. Only 5% of the HCWs with a longer vaccine interval between COVID-19 vaccination and infection lacked serological indication of an infection (vaccination three to six months before a positive COVID-19 test, Supplementary Fig.2).
To determine the comparative efficiency of BNT162b2 and mRNA-1273 as the third dose, S1-specific antibody responses were analyzed after the third vaccine dose in uninfected HCWs. In both short and long-interval groups, the vaccine-induced antibody responses had similar kinetics, with the S1 antibody level peaking at 3 weeks after vaccination, followed on an average by a 25 EIA unit decline in subsequent samples (Fig.3a, b). The mRNA-1273 vaccine as the third dose provided significantly higher S1-specific antibody responses in both interval groups (p<0.0423 in the short interval group and p<0.0300 in the long interval group). The difference between the groups receiving BNT162b2 or mRNA-1273 as the third dose in the long interval group was observable already prior to the third dose (a result of different vaccine combinations before the third dose, different vaccine combination groups marked with different colors in Fig.3b), and the difference in antibody levels remained observable also after the third dose (Fig.3b).
SARS-CoV-2 S1-specific IgG antibody responses induced by Bnt162b2 or mRNA-1273 (triangles) as the third vaccine dose were compared in the sera of uninfected vaccinees who received a two doses of Bnt162b2 with a short, three-week dose interval or b two doses of Bnt162b2 (orange), mRNA-1273 (violet), or ChAdOx1 + Bnt162b2/mRNA-1273 (green) with a long dose interval before the third vaccine dose. SARS-CoV-2 S1-specific IgG antibody responses were analyzed by EIA in serum samples collected before the third dose (2D6mo or 2D8mo) and after the third vaccine dose (3D3wk, 3D3mo, 3D6mo, and 3D9mo). Geometric means geometric standard deviations of antibody levels, and number of samples in each time point are shown. Dashed lines indicate the cut-off values for seropositivity.
Our study included HCWs aged 19 to 67 years at the time of the first vaccine dose. To examine whether age affects humoral immune responses induced by COVID-19 vaccines, we analyzed anti-S1 IgG antibody levels in relation to age from uninfected HCWs (Supplementary Fig.3). Some decrease in antibody levels was observed by increasing age, but a higher age did not prevent the induction of high antibody levels. Although the oldest age group, 5567-year-olds, had, on average, the lowest antibody levels at 3 and 6 months after the third vaccine dose the antibody levels were relatively equal between all age groups (Supplementary Fig.4).
Neutralizing capacity of the sera against SARS-CoV-2 variants was first examined in the short interval group by randomly selecting a subset of 41 HCWs (no prior PCR-confirmed SARS-CoV-2 infection). MNT was used to analyze in vitro neutralizing antibody titers against the ancestral D614G variant and the five recent SARS-CoV-2 Omicron variants, BA.1, BA.2, BA.5, BQ.1.1, and XBB.1.5 in serum samples collected at six months after the second vaccine dose (2D) and 3 weeks (3wk), three months (3mo), and six months (6mo) after the third booster dose (3D). (Fig.4al).
HCWs received two doses of BNT162b2 with a three-week interval and a third dose of BNT162b2 (circle) or mRNA-1273 (triangle) eight months later. af Serum samples were collected 6 months after the second vaccine dose (2D6mo, n=41), 3 weeks (3D3wk, n=40), 3 months (3D3mo, n=41), and 6 months (3D6mo, n=39) after the third dose and analyzed with MNT for neutralizing antibodies against SARS-CoV-2 D614G and Omicron BA.1, BA.2, BA.5, BQ.1.1 (BA.5 subvariant), and XBB.1.5 (BA.2 subvariant) variants. HCWs with confirmed SARS-CoV-2 infection between three and six months after third dose (n=12) were separated (red dots and triangles; infected 3D6mo). Half-maximal inhibitory dilutions (ID50) were calculated, and titers <10 were marked as 5. Geometric mean titers for vaccine groups are indicated above each time point and shown as lines with geometric SDs. gl Sequential serum samples of each individual are connected with lines. Red lines indicate where a vaccinee has had a PCR or antigen test confirmed SARS-CoV-2 infection. mq Top and side views of trimeric SARS-CoV-2 spike protein structure (PDB: 7WK2) show amino acid differences of Omicron BA.1, BA.2, BA.5, BQ.1.1, and XBB.1.5 compared to Wuhan Hu-1 sequence as amino-acid substitutions (orange) and deletions (red).
Six months after the second vaccine dose, 95% of the vaccinees (39/41) neutralized the D614G and 44% (18/41), 76% (31/41), 90% (37/41), 5% (2/41), and 0% (0/41) neutralized Omicron BA.1, BA.2, BA.5, BQ.1.1, and XBB.1.5 variants, respectively (Fig.4af). The third vaccine dose increased neutralizing antibodies against all variants, although the levels against BQ.1.1 and XBB.1.5 variants remained lowest. Despite the gradual decrease in the levels of neutralizing antibodies after the third dose, the geometric mean titers (GMT) were 4.47.2x higher six months after the third dose in comparison to 6 months after the second dose, (80 vs 364 for D614G, 9 vs 47 for Omicron BA.1, 24v. 173 for BA.2, and 26 vs 125 for BA.5; p<0.0001 for each pair). For Omicron BQ.1.1 and XBB.1.5 variants, the GMTs were 2.0x and 2.4x higher (5 vs. 12 for Omicron BQ.1.1, p=0.0003, and 5 vs. 10 for Omicron XBB.1.5; p=0.0001). Six months after the third dose only two samples had titers below the detection limit for Omicron BA.1, one sample for BA.2, 12 samples for BQ.1.1 and 17 samples for XBB.1.5. The results suggest that the neutralization efficiency of the induced antibodies is still reasonably high against the earlier Omicron variants, but is strongly reduced against BQ.1.1 and XBB.1.5 variants, with the number of samples below the detection limit increasing at the 3D6mo time point.
Twelve of the 41 vaccinees reported a PCR-confirmed SARS-CoV-2 breakthrough infection between the sampling of three and six months after the third dose. The breakthrough infection increased the levels of neutralizing antibodies, and the GMTs were 625x higher compared to non-infected participants (364 in non-infected vs. 2239 in infected at 3D6mo for D614G, 47 vs. 644 for Omicron BA.1, 173 vs. 997 for BA.2, and 125 vs. 910 for BA.5, 5 vs.126 for BQ.1.1, and 5 vs. 109 for XBB.1.5; p=0.0002 for D614G and Omicron BA.2, p<0.0001 for Omicron BA.1 and BA.5, and p To study differences in neutralizing antibody titers elicited by different vaccine combinations and vaccine dose intervals against D614G and Omicron BA.1. BA.2, BA.5, BQ.1.1, and XBB.1.5 sera from a representative number of HCWs (with two BNT162b2 with a short (n=41) or with a long vaccine dose interval (n=35), two mRNA-1273 (n=31), or ChAdOx1 and BNT162b2 or mRNA-1273 (n=45) before the third dose of BNT162b2 or mRNA-1273) were analyzed with MNT (Fig.5). At 6 months after the second vaccine dose, the majority of vaccinees in each vaccine combination group had neutralizing antibodies against D614G variant and Omicron BA.2 and BA.5, whereas all vaccine combination groups had lower or undetectable levels of neutralizing antibodies against Omicron BA.1, BQ.1.1, and XBB.1.5. Interestingly, 2 x mRNA-1273 induced higher neutralizing antibody titers at 2D6mo against all variants, before the administration of the third vaccine dose. The third vaccine dose increased neutralizing antibody titers in all vaccine groups resulting in neutralizing antibody titers above the detection limit against all Omicron variants, leaving only four vaccinees below the detection limit against BQ.1.1 and ten against XBB.1.5. Interestingly, in Omicron variants 3 weeks after the third dose, the short interval 2 x BNT162b2 vaccine group had the highest GMT, even 1.11.9x higher than in the 2 x mRNA-1273 group (Fig.5). Serum samples, collected 6 months after the second dose (2D6mo), 3 weeks (3D3wk), and 3 months (3D3mo) after the third vaccine dose from HCWs who received 2x BNT162b2 with a short vaccine dose interval (n=41, yellow circles and triangles) or 2x BNT162b2 (n=35, orange circles and triangles), 2x mRNA-1273 (n=31, purple circles and triangles), or ChAdOx1+BNT162b2/mRNA-1273 (n=45, green circles and triangles) with a long vaccine dose interval, and a third dose of BNT162b2 (circle) or mRNA-1273 (triangle) were compared for neutralizing antibody responses against D614G and Omicron variants BA.1, BA.2, BA.5, BQ.1.1, and XBB.1.5. HCWs with confirmed SARS-CoV-2 infection between the samplings at 3 weeks and three months after the third dose (3D3wk and 3D3mo) were separated (3 doses + infection, n=24, red circles and triangles). Half-maximal inhibitory dilutions (ID50) were calculated, and titers <10 were marked as 5. Geometric mean titers (GMTs) for each vaccine group are shown as lines with geometric SDs. Three months after the third vaccine dose, the decrease in neutralizing antibody titers against D614G, and the Omicron variants was similar in all vaccine combination groups, and the neutralizing titers remained slightly higher in the 2 x BNT162b2 and 2 x mRNA-1273 vaccine groups than in the other groups. The difference in neutralization of Omicron BA.1 and XBB.1.5 was significant when the titers in short 2 x BNT162b2 group were compared to titers in ChAsOx1+BNT162b2/mRNA-1273 group (GMT 40 vs. 99 against BA.1, p=0.0014; GMT 8 vs. 16 against XBB.1.5, p=0.0024) and 2x mRNA-1273 group (GMT 8 vs. 16 against XBB.1.5, p=0.040). Between the sampling of 3 weeks and three months after the third dose, 24 of the HCWs with the long vaccine interval reported a COVID-19 test-positive SARS-CoV-2 breakthrough infection (red dots in Fig.5). The neutralizing antibody titers against the four variants were significantly higher in infected than in non-infected HCWs. Only one HCW with three vaccine doses and an infection had neutralizing antibody titers against Omicron XBB.1.5 below the detection limit. Altogether, these results indicate that the studied vaccine combinations elicit high titers of SARS-CoV-2 neutralizing antibodies against Omicron BA.1, BA.2, and BA.5 variants, while the titers against the Omicron BQ.1.1 and XBB.1.5 were yet relatively low. An infection within three months after three vaccine doses elicits high neutralizing antibody titers against all Omicron variants. The estimation of neutralization capacity of COVID-19-vaccinated individuals against different variants is central in deciding the need for further vaccine doses. Here, we compared the neutralizing antibody titers of HCWs without breakthrough infection against D614G and Omicron variants BA.1, BA.2, BA.5, BQ.1.1, and XBB.1.5 in the order these variants emerged (Fig.6). Regardless of the vaccine combination, the neutralizing capacity of the antibodies was 4.811.5x reduced 3 weeks and 3 months after the third vaccine dose when moving from D614G variant to Omicron BA.1, 2.14.6x increased from Omicron BA.1 to BA.2, and again1.52.3x reduced from Omicron BA.2 to BA.5, and 4.99.9x further reduced to BQ.1.1 and to XBB.1.5. The only exception was the 2 x mRNA-1273+BNT162b2/mRNA-1273 group, as in this group there was no significant difference between the neutralization capacity against Omicron BA.2 and Omicron BA.5 (GMT 503 vs 359) at 3 weeks after the third dose. Comparison of neutralizing antibody titers against D614G and Omicron variants BA.1, BA.2, BA.5, BQ.1.1, and XBB.1.5 at 3 weeks (3D3wk) and three months (3D3mo) after the third vaccine dose of HCWs without SARS-CoV-2 infection within each vaccine combination group (2x BNT162b2 with a short, n=40 at 3D3wk and at 3D3mo or a long vaccine dose interval, n=34 at 3D3wk and n=27 at 3D3mo; 2x mRNA-1273 n=31 at 3D3wk and n=24 at 3D3mo; or ChAdOx1+BNT162b2/mRNA-1273, n=44 at 3D3wk and n=35 at 3D3mo). Half-maximal inhibitory dilutions (ID50) were calculated, and titers <10 were marked as 5. Geometric mean titers (GMTs) for each vaccine group are indicated above bars and shown as lines with geometric SDs. The follow-up of neutralizing antibodies in HCWs who had a breakthrough infection showed that the infection boosted the titers of neutralizing antibodies 1.77.9x against the tested variants (Fig.7). The fold difference between the Omicron variants was relatively similar at all time points (1.52.1x between BA.1 and BA.2, 1.41.5 between BA.2 and BA.5 etc.), while saturation of titers for D614G affected the measured GMT values (Supplementary Fig.5). The comparable fold difference between the variants in both pre-infection and post-infection samples from vaccinees suggests that an Omicron variant infection equally enhances the existing neutralizing antibody response against various variants. Neutralizing antibodies against D614G and Omicron variants BA.1, BA.2, BA.5, BQ.1.1, and XBB.1.5 of a 12 HCWs with short vaccination interval and b 23 HCWs with long vaccination interval with a breakthrough infection were compared. Half-maximal inhibitory dilutions (ID50) were calculated before third COVID-19 vaccine dose (2D6mo), and 3 weeks (3D3wk), 3 months (3D3mo), and 6 months (3D6mo, only for short vaccine interval group) after receiving the third COVID-19 vaccine dose. Uninfected individuals are marked with blue (long interval) or yellow (short interval) diamonds, infected individuals with red diamonds. Titers <10 were marked as 5. Titers of each vaccinee in different time points are connected with lines. Red line indicates the period of a PCR- or serology-confirmed infection. Geometric mean antibody titers with standard deviations are shown as bars and lines at each time point. To analyze the correlation of the antibody responses induced by the vaccinations against the Omicron variants BA.1, BA.2, BA.5, BQ.1.1, and XBB.1.5 and the ancestral D614G variant, the neutralization efficiency of 499 serum samples from 155 vaccinees was pairwise compared between the six SARS-CoV-2 variants (Supplementary Fig.6). Neutralizing antibody titers against the D614G variant were higher than against any of the Omicron variants. Even though the neutralization efficiency between different variants varied, they correlated well with each other (r=0.77710.8766, p<0.0001 for all pairwise comparisons). Parallel examination of neutralizing antibody titers (for BA.5 as an example) and IgG responses for S1 and N showed similar kinetics following the third vaccine dose and breakthrough infections in both short and long vaccine interval groups (Supplementary Fig.7). In the long vaccine interval group the breakthrough infections (n=23) occurred closer to the third vaccine dose (between 3D3wk and 3D3mo) than in the short vaccine interval group (n=12, between 3D3mo and 3D6mo). Of the 12 short vaccine group HCWs with a confirmed breakthrough infection 92% (11/12) had an increase in anti-N and 83% (10/12) in anti-S1 antibodies, and 83% (10/12) had at least a 4x-increase in Omicron BA.5-specific neutralizing titers. In the long-interval vaccine group serological evidence for a breakthrough infection was detectable in anti-N antibodies in 65% (15/23), in anti-S1 antibodies in 78% (18/23), and for Omicron BA.5 neutralizing titers in 48% (11/23) of infected HCWs. Three HCWs with a COVID-19 test-positive infection showed no increase in S1- or N-specific IgG antibodies, however, two of these had their serum sampling close (810 days) to the positive COVID-19 test date, and this may have been too early after the infection to detect newly formed antibodies. See the article here:
