Participants’ characteristics
Of the 555 participants initially included in the study, 463 (83.4%) completed their first follow-up visit at month-6, and 377 (67.9%) completed their second follow-up visit at month-12. Of these, 367 participants had the complete set of three timepoints and were included in the final dataset (Fig. 1). Among these, 280 vaccinated participants meet the groups criteria and 190 (67.9%) of these completed the fourth questionnaire.
Demographic and clinical characteristics of the final dataset of 367 participants who completed all three follow up visits are shown in Table 1. The median age was 51 years at the end of the study, 80 (21.8%) were men and 287 women (78.2%), and 162 (44.1%) participants had at least one comorbidity. Ninety-three of 367 participants (25.3%) were HCWs and 274 (74.7%) worked in other areas of the public sector.
Table 1
Characterisation of the 367 participants at each time point
| 0-months | 6-months | 12-months |
n | 367 | 367 | 367 |
Female (n, %) | 287 (78.2%) | 287 (78.2%) | 287 (78.2%) |
Male (n, %) | 80 (21.8%) | 80 (21.8%) | 80 (21.8%) |
Age (years), median [IQR] | 50 [40; 56] | 50 [40; 57] | 51 [41; 57] |
BMI (ratio), median [IQR] | 23.7 [21.5; 27.3] | 24.0 [22.0; 27.0] | 23.7 [21.8; 27.0] |
Comorbidities (n, %) | 162 (44.1%) | 162 (44.1%) | 162 (44.1%) |
HCWs (n, %) | 93 (25.3%) | 93 (25.3%) | 93 (25.3%) |
New Infections (n, %) | 34 (9.3%) | 40 (10.9%) | 18 (4.9%) |
Total number of infected (n, %) | 34 (9.3%) | 74 (20.2%) | 92 (25.0%) |
Time between infection and sampling (days), median [IQR] | 68.5 [33.3; 164.5] | 258.5 [226.5; 374.3] | 387.5 [263.8; 479.8] |
Vaccinated (n, %) | 0 | 245 (66.8%) | 341 (92.9%) |
Unvaccinated (n, %) | 367 (100%) | 122 (33.2%) | 26 (7.1%) |
Number of vaccine doses (n, %) | 0 (367, 100%) | 0 (122, 33.2%) | 0 (26, 7.1%) |
1 (100, 27.2%) | 1 (60, 16.3%) |
2 (145, 39.5%) | 2 (244, 66.5%) |
3 (0) | 3 (37, 10.1%) |
Vaccinated and naïve to infection (n, %) | - | 208 (56.7%) | 259 (70.7%) |
Vaccinated and infected (n, %) | - | 37 (10.0%) | 82 (22.3%) |
Unvaccinated and infected (n, %) | 34 (9.3%) | 38 (10.4%) | 10 (2.7%) |
Unvaccinated and naïve to infection (n, %) | 333 (90.7%) | 84 (22.9%) | 16 (4.4%) |
Time between vaccination and sampling (days), median [IQR] | - | 39 [21; 75] | 188 [140; 230] |
Anti-SP IgG (BAU/mL), median [IQR] | 59.5 [33.75; 165.5] | 278.5 [94; 918.5] | 248 [108.8; 734] |
Anti-N IgG positivity (n, %) | - | 30 (40.5% of infected) | 26 (28.3% of infected) |
Neutralisation of Wuhan variant, median [IQR] | 2.6 [1.4; 4.3] | 7.4 [2.4; 16.6] | 6.4 [2.5; 14.0] |
Neutralisation of Alpha variant, median [IQR] | 2.2 [1.2; 3.5] | 6.0 [1.9; 14.9] | 5.1 [2.1; 12.8] |
Neutralisation of Delta variant, median [IQR] | 2.6 [1.1; 3.9] | 5.9 [1.9; 14.5] | 5.5 [2.2; 13.5] |
SARS-CoV-2 specific T cell response (IU/mL), median [IQR] | - | - | 0.41 [0.22; 0.9] |
N Number; IQR – Interquartile range; HCW – Health care workers; Anti-SP – Anti-Spike; IgG – Immunoglobulin G; Anti-N – Anti-Nucleocapsid. |
At baseline visit between July 2020 and January 2021, when no vaccination was available, 34 of 367 (9.3%) participants had a history of infection with SARS-CoV-2 as demonstrated by a positive PCR or antigenic test and/or positive serological test for anti-SP (Table 1 and Fig. 2). At the first follow-up visit at month-6 between February 2021 and July 2021, 74 of 367 (20.2%) participants had a history of SARS-CoV-2 infection, 245 (66.8%) were vaccinated with at least one dose of the SARS-CoV-2 vaccine, and 84 (22.9%) were naive to both infection and vaccination. At the second follow-up visit at month-12 between September 2021 and December 2021, 92 of 367 (25.0%) participants had a history of SARS-CoV-2 infection, 341 (92.9%) were vaccinated with at least one dose of the SARS-CoV-2 vaccine, and 16 (4.4%) were naive to both infection and vaccination. The rate of new infections decreased at the third time point with 34 (9.3%), 40 (10.0%) and 18 (4.9%) new infection at month-0, month-6 and month-12, respectively. HCWs had higher rates of infection in comparison to the participants working in other areas of the public sector (16.1% vs 6.9% at month-0; 23.9% vs 17.9% at month-6; and 27.2% vs 22.3% at month-12 for HCWs vs other public sector, respectively), likely due to the increased contact with COVID-19 patients. Similarly, their rate of vaccination was higher than in the public sector due to COVID-19 vaccine obligation in French hospitals (81.5% vs 62.0% at month-6; and 98.9% vs 90.9% at month-12 for HCWs vs another public sector, respectively). In order to compare the infection and vaccination rates of study participants to the general French population, we extracted infection and vaccination rates from online public data repositories data.gouv.fr and ourworldindata.org. In comparison to the general population, the study participants had a higher rate of infection and of vaccination than the general population at any given time period (month-0, month-6 and month-12; Fig. 2).
Considering the vaccination and infection status, participants were classified in four groups: Vaccinated and naive to infection (n = 259, 70.6%), Vaccinated and previously infected (n = 82, 22.3%), unvaccinated and naive to infection (n = 16, 4.4%) and Unvaccinated and infected (n = 10, 2.7%). Due to number of participants in each group, only the two groups of vaccinated individuals were analysed. Among Vaccinated and naive to infection, 47 individuals were excluded since they had received a third injection before month-12 or their samples collected were between dose one and two of vaccination. Only individuals with at least one sample before and after two doses were included in the analysis. In the group of Vaccinated and previously infected individuals, 14 were excluded due to uncertainty of infection occurrence before or after vaccination.
Humoral response
In the group of individuals who were Vaccinated and naive to infection, antibody levels before vaccination were below the threshold of positivity. Within three months following vaccination, only two of 116 individuals (1.7%) with a sample in this time range had no detectable SARS-CoV-2 anti-SP IgG antibodies, while 31 of 116 individuals (26.7%) remained below the protective threshold of 264 BAU/mL determined previously for the Alpha variant (Feng et al., 2021). There were significant differences in age and comorbidities between individuals above and below the protective threshold. Individuals with SARS-CoV-2 anti-SP IgG antibodies above 264 BAU/mL were younger (47 [IQR, 39; 56] vs 50 [42; 57] years, p = 0.045) and presented less comorbidities (50 (36.5%) individuals with at least one comorbidity vs 82 (49.7%), p = 0.027). The sex ratio and HCWs’s proportion were the same between the groups (Table 2). Between month-3 and month-6 post-vaccination, anti-SP IgG antibody levels started to significantly decrease. Sixty-two of 96 individuals (64.6%) were below the protective threshold of 264 BAU/mL. Over six months after vaccination, antibody levels rapidly decreased as 87 of 107 individuals (81.3%) were below the protective threshold of 264 BAU/mL. We found a significant inverse correlation between time post-vaccination and anti-SP IgG antibody levels (p < 0.0001, Spearman r=-0.62) (Fig. 3A left), i.e., post-vaccination time was longer for participants with sub-threshold anti-SP antibody levels compared to those with protective levels (176 and 73 days, respectively) (p < 0.0001) (Table 2).
Table 2
Characterization of individuals above and below 264 BAU/mL in the two groups: Vaccinated and naïve to infection and Vaccinated and previously infection
Characteristics | Vaccinated and naive to infection (VNI) | Vaccinated and previously infected (VPI) | p-value1 |
IgG Spike < 264 BAU/mL | IgG Spike > 264 BAU/mL | p-value1 | Total | IgG Spike < 264 BAU/mL | IgG Spike > 264 BAU/mL | p-value1 | Total | Total: VNI vs VPI | < 264 BAU/mL: VNI vs VPI | > 264 BAU/mL: VNI vs VPI |
n | 165 | 137 | | 212 | 27 | 68 | | 68 | | | |
M/F | 31/134 | 22/115 | 0.76 | 40/172 | 4/23 | 18/30 | 0.038* | 17/51 | 0.3 | 0.65 | 0.0038** |
Age (years), median [IQR] | 50 [42; 57] | 47 [39; 56] | 0.045* | 49 [41; 56] | 51 [37; 56] | 51 [39; 57] | 0.59 | 51 [39.3; 57] | 0.85 | 0.55 | 0.33 |
BMI (ratio), median [IQR] | 23.6 [21.8; 26.9] | 23.7 [22; 26.4] | 0.59 | 23.6 [21.7; 26.9] | 23.4 [21.9; 25.2] | 25 [22.3; 28.6] | 0.09 | 24.2 [22.2; 26.6] | 0.5 | 0.54 | 0.19 |
Comorbidities (n, %) | 82 (49.7%) | 50 (36.5%) | 0.027* | 90 (42.5%) | 15 (55.6%) | 30 (44.1%) | 0.37 | 32 (47.1%) | 0.57 | 0.68 | 0.36 |
HCWs (n, %) | 41 (24.8%) | 38 (27.8%) | 0.6 | 51 (24.1%) | 11 (40.8%) | 21 (30.9%) | 0.48 | 21 (30.9%) | 0.27 | 0.1 | 0.74 |
Time between infection and sampling (days), median [IQR] | - | - | | - | 387 [109; 453] | 285 [235.5; 398.3] | 0.42 | 236 [80; 243] | - | - | - |
Time between vaccination and sampling (days), median [IQR] | 176 [134; 212.5] | 73 [32.5; 142.5] | < 0.0001 | 145 [66; 194] | 188 [127; 230] | 102 [47.3; 153.8] | 0.0021** | 112 [47; 196] | 0.28 | 0.65 | 0.049* |
Number of vaccine doses, (n, %) | 2 (165, 100%) | 2 (137, 100%) | - | 2 (212, 100%) | 1 (14, 51.9%); 2 (13, 48.1%) | 1 (47, 69.1%); 2 (21, 30.9%) | 0.15 | 1 (43, 63.2%); 2 (25, 36.8%) | - | -- | |
Anti-SP IgG (BAU/mL), median [IQR] | 109 [63; 158.5] | 722 [465.5; 1331] | < 0.0001 | 208 [97; 630.5] | 138 [51; 185] | 914.5 [460.8; 1617] | < 0.0001 | 473 [185; 1396] | < 0.0001 | 0.3 | 0.19 |
Neutralisation of Wuhan variant, median [IQR] | 2.5 [1.5; 4.5] | 12.8 [8.2; 23.6] | < 0.0001 | 6.3 [2.4; 12.8] | 4.8 [2.1; 8.8] | 18.2 [11.7; 51.2] | < 0.0001 | 15 [6.6; 34.4] | < 0.0001 | 0.009** | 0.0013** |
Neutralisation of Alpha variant, median [IQR] | 2.1 [1.2; 3.5] | 11.1 [6.3; 20.6] | < 0.0001 | 4.8 [2; 10.5] | 4.2 [1.8; 9.4] | 17.2 [11.2; 47.8] | < 0.0001 | 13.5 [6.1; 34.3] | < 0.0001 | 0.0027** | < 0.0001 |
Neutralisation of Delta variant, median [IQR] | 2.3 [1.5; 4] | 11.6 [6.7; 21.2] | < 0.0001 | 5.1 [2.8; 10.8] | 4.4 [1.9; 8.5] | 19.6 [11.7; 51.8] | < 0.0001 | 13.8 [6.1; 41.4] | < 0.0001 | 0.01* | < 0.0001 |
SARS-CoV-2 specific T cell response (IU/mL), median [IQR] | 0.21 [0.09; 0.42] | 0.29 [0.16; 0.9] | 0.0023** | 0.23 [0.1; 0.54] | 0.2 [0.095; 0.68] | 0.65 [0.31; 1.9] | 0.0014** | 0.52 [0.23; 1.23] | < 0.0001 | 0.72 | 0.0018** |
N Number; IQR Interquartile range; HCW Health care workers; Anti-SP Anti-Spike; IgG Immunoglobulin G; Anti-N Anti-Nucleocapsid. |
1Mann-Whitney test for quantitative variables; Fisher’s exact test for qualitative variables. All tests are two-sided with statistical significancy p < 0.05. *p < 0.05, **p < 0.005, ***p < 0.0005, ****p < 0.00005. |
Table 3
Characterization of individuals displaying a detectable or undetectable SARS-CoV-2 T cell response in the two groups: Vaccinated and naive to infection and Vaccinated and previously infection
Characteristics | Vaccinated and naive to infection (VNI) | Vaccinated and previously infected (VPI) | |
Undetectable SARS-CoV-2 T cell response | Detectable SARS-CoV-2 T cell response | p-value1 | Undetectable SARS-CoV-2 T cell response | Detectable SARS-CoV-2 T cell response | p-value1 | P-value: Detectable T cell response VNI vs VPI |
n | 47 | 156 | | 7 | 61 | | |
M/F | 12/35 | 26/130 | 0.2 | 0/7 | 17/42 | | 0.06 |
Age (years), median [IQR] | 43 [37; 57] | 50 (42; 56] | 0.05 | 48 [37; 57] | 51 [40; 57] | 0.6 | 0.9 |
BMI (ratio), median [IQR] | 23.6 [21.4; 26.9] | 23.7 [21.8; 26.8] | 0.6 | 23.4 [22.1; 24.2] | 25 [22.3; 27.9] | 0.1 | 0.2 |
Comorbidities (n, %) | 21 (44.7%) | 67 (42.9%) | 0.9 | 3 (42.7%) | 27 (44.3%) | > 0.9 | 0.9 |
HCWs (n, %) | 12 (25.5%) | 32 (20.5%) | 0.5 | 3 (42.7%) | 18 (29.5%) | 0.7 | 0.2 |
Time between infection and sampling (days), median [IQR] | - | - | | 387 [287.5; 404.5] | 342 [263.3; 443] | > 0.9 | - |
Time between vaccination and sampling (days), median [IQR] | 176 [133; 206] | 177 [145.3; 209.8] | 0.9 | 196 [153; 217] | 175 [102; 218] | 0.4 | 0.09 |
Number of vaccine doses, (n, %) | 2 (51, 100%) | 2 (156, 100%) | | 1 (3, 42.9%); 2 (4, 57.1%] | 1 (39, 63.9%); 2 (20, 32.8%) | 0.2 | - |
Anti-SP IgG (BAU/mL), median [IQR] | 104 [67; 193] | 164 [81.5; 349.3] | 0.014* | 182 [95; 803] | 473 [242.5; 1414] | 0.06 | < 0.0001 |
Neutralisation of Wuhan variant, median [IQR] | 2.8 [1.9; 7.4] | 4.5 [2.2; 8.3] | 0.057 | 9 [4; 18] | 13.1 [7; 34.8] | 0.2 | < 0.0001 |
Neutralisation of Alpha variant, median [IQR] | 2.2 [1.6; 5.4] | 3.5 [1.7; 6.3] | 0.09 | 9.7 [3.7; 15.5] | 13.2 [6.6; 35.6] | 0.15 | < 0.0001 |
Neutralisation of Delta variant, median [IQR] | 2.7 [1.7; 5.6] | 3.9 [1.8; 6.7] | 0.09 | 8.8 [4; 11.8] | 14 [6.7; 42.3] | 0.1 | < 0.0001 |
SARS-CoV-2 specific T cell response (IU/mL), median [IQR] | - | 0.3 [0.19; 0.71] | | - | 0.59 [0.27; 1.5] | - | 0.0002*** |
N Number; IQR Interquartile range; HCW Health care workers; Anti-SP Anti-Spike; IgG Immunoglobulin G; Anti-N Anti-Nucleocapsid. |
1Mann-Whitney test for quantitative variables; Fisher’s exact test for qualitative variables. All tests are two-sided with statistical significancy p < 0.05. *p < 0.05, **p < 0.005, ***p < 0.0005, ****p < 0.00005. |
In the group of previously infected individuals, only eight of 52 participants (15.4%) had anti-SP IgG antibody levels above the protective threshold of 264 BAU/mL prior to vaccination (Fig. 3A right), independently of the time elapsed since infection (data not shown). Anti-SP IgG antibody levels significantly increased post-vaccination (p < 0.0001) and were more stable over time as only 20 of 59 (33.9%) individuals were below 264 BAU/mL of anti-SP IgG antibodies three months after vaccination in comparison to 149 of 203 (73.4%) Vaccinated individuals naive to infection (p < 0.0001, Table 2). The inverse correlation between number of days since vaccination and anti-SP IgG antibody levels was significantly lower in Vaccinated and previously infected individuals than in Vaccinated and previously uninfected individuals (p = 0.018, Spearman r=-0.22) (Fig. 3A). Despite this, within three months after vaccination there was no significant difference in anti-SP IgG antibody levels between naive or previously infected individuals. No association was found in Vaccinated and previously infected individuals between anti-SP IgG levels and the number of vaccine doses (data not shown).
In the groups of both Vaccinated and naive to infection and Vaccinated and previously infected individuals, the neutralization capacity of anti-SP IgG antibodies was correlated to their levels (p < 0.0001, Spearman r = 0.9; p < 0.0001, Spearman r = 0.86, respectively) (Fig. 3B). In previously infected individuals, the neutralization capacity of anti-SP IgG antibodies significantly increased after vaccination (p < 0.0001) (Fig. 3C). Independently of the variant (Wuhan, Alpha or Delta), Vaccinated and previously infected individuals exhibited a higher median neutralization capacity (15.0, 13.5, 13.8 unit/mL, respectively) than Vaccinated individuals naive to infection (6.3, 4.8, 5.1 unit/mL, respectively) (p < 0.0001) (Table 2).
As the neutralization capacity of the Wuhan variant was higher than the neutralization of Alpha and Delta variants, a percentage of neutralization of the Alpha and Delta variants was calculated for each individual in comparison to the Wuhan variant. The neutralization of Alpha variant was significantly lower in comparison to the Wuhan variant in the group of individuals naive to infection (80.7%) and in the group of Vaccinated and previously infected individuals (91.5%, p < 0.0001). Similar results were obtained for the Delta variant (85.6% relative neutralization in the group of Vaccinated individuals naive to infection vs 94.1% in the group of Vaccinated and previously infected individuals, p < 0.0001). (Fig. 3D). The median neutralization capacity against the Alpha variant in Vaccinated and previously infected individuals was 91.5% vs 80.7% in Vaccinated individuals naive to infection (p < 0.0001). Similarly, Vaccinated and previously infected individuals had a 94.1% neutralizing capacity of the Delta variant vs 85.6% in Vaccinated individuals naive to infection (p < 0.0001) (Fig. 3D).
Cellular response
Using IGRA, we measured the SARS-CoV-2-specific T cell response in participants after infection and/or vaccination at the third and last sampling point at month-12 after inclusion in the study. As the cellular response was measured only at month-12, a selection of samples within nine months after vaccination was performed to normalize the time between vaccination and sampling.
Two hundred and seventy of 367 participants (71.8%) had a positive SARS-CoV-2-specific cellular response. Among Vaccinated participants naive to infection, 156 of 203 (76.8%) were positive to the test, in comparison to 61 of 68 (89.7%) Vaccinated and previously infected participants (p = 0.023). There were no significant differences in age, comorbidities or gender or HCWs between groups (Table 2b). There was no difference in the levels of cellular response in Vaccinated and previously infected individuals depending on the number of vaccine doses (data not shown).
Contrary to humoral response, in the group of Vaccinated individuals naive to infection there was no correlation between the level of the cellular response and the time since vaccination (p = 0.3, Spearman r = 0.07) (Table 2b, Fig. 4A left). In the Vaccinated and previously infected group, a slight decline of the cellular response was observed corresponding to a significant negative correlation with time since vaccination (p = 0.004, Spearman r=-0.3) (Fig. 4A right). Despite this decline, median values of IFNγ in Vaccinated and previously infected individuals with a detectable SARS-CoV-2 T cell response were significantly higher than in Vaccinated individuals naive to infection (0.6 IU/mL and 0.3 IU/mL, respectively, p = 0.0002) (Table 2b).
In the group of Vaccinated participants naive to infection, individuals with anti-SP IgG antibody levels below 264 BAU/mL had a median of 0.21 IU/mL of IFNγ vs 0.29 IU/mL in individuals with more than 264 BAU/mL (p = 0.0023). In Vaccinated and previously infected individuals, IFNγ secretion was 0.20 and 0.65 IU/mL for participants with antibody levels over 264 BAU/mL and below 264 BAU/mL, respectively (p = 0.0014). Between the two groups, median IFNγ levels did not differ in individuals with anti-SP IgG levels below 264 BAU/mL (0.21 IU/mL vs 0.20 IU/mL, p = 0.7). Above this threshold, Vaccinated and previously infected individuals displayed significantly higher median IFNγ levels than Vaccinated individuals naive to infection (0.65 IU/mL vs 0.29 IU/mL, respectively, p = 0.0018) (Fig. 4B).
Infection predictive factors
In the second part of the study, information about infections occurring between December 2021 and April 2022 was collected and correlated to the prior humoral and cellular response as measured during the third sampling time point at month-12 between September 2021 and December 2021 to measure the performance of these different vaccination schedules. The last sample was used to evaluate the existing immune response level against SARS-CoV-2 just before the new wave of Delta- and Omicron-related contaminations in the beginning of 2022. Of 212 individuals who were Vaccinated and naive to infection at month-12, 169 (88.9%) answered the questionnaire and so this was the only group analysed. Thirty five of 169 (20.7%) were tested positive to SARS-CoV-2 between December 2021 and April 2022 presenting either no symptoms (n = 5, 14.3%) or mild symptoms without requiring hospitalization (n = 30, 85.8%).
During this period, there were no significant differences between infected and not infected participants in gender ratio, comorbidities nor being HCW (more exposed to the virus). Infected individuals were significantly younger than non-infected individuals (median 44 vs 52 years, respectively, p = 0.0026). Time since the last vaccine dose and the questionnaire between the vaccinated individuals who got infected (109 days) and those who did not get infected (96 days) was significantly different (p < 0.0001). However, infection happen in median 74 days after the last dose. There was also a difference in the number of vaccine doses, as only 22 of 35 infected individuals (62.9%) had received a third dose vs 126 of 134 individuals (94.0%) who did not get infected over the same period of time (p < 0.0001) (Fig. 5A). Vaccinated individuals who got infected tend to have slightly lower IFNγ release after SARS-CoV-2 T cell stimulation after two doses (0.2 IU/mL) than vaccinated individuals who did not got infected (0.3 IU/mL) (p = 0.06) (Table 4).
Table 4
Characterization of Vaccinated and naïve to infection individuals who respond to the post-12-months questionnaire
Characteristics | All post-12-months | Not infected post-12-months | Infected post-12-months | p-value1 |
Univariable | Multivariable |
n | 169 | 134 | 35 | | |
M/F | 32/137 | 27/107 | 5/30 | 0.6 | |
Age (years), median [IQR] | 50 [42; 57] | 52 [43; 58] | 44 [35; 52] | 0.0026 | 0.0084 |
BMI (ratio), median [IQR] | 23.5 [21.7; 27] | 23.6 [21.7; 27.1] | 23.3 [21.2; 26.2] | 0.63 | |
Comorbidities (n, %) | 67 (39.6%) | 55 (41%) | 12 (34.3%) | 0.6 | |
HCWs (n, %) | 38 (22.5%) | 26 (19.4%) | 12 (34.3%) | 0.07 | 0.25 |
Time between last dose and questionnaire (days), median [IQR] | - | 96 [75.5; 107.3] | 74 [39; 130] | 0.0001 | 0.14 |
Time between vaccination and sampling (days), median [IQR] | 181 [144; 224] | 181.5 [144; 223.3] | 180 [139; 238] | 0.95 | |
Number of vaccine doses before sampling (n, %) | 2 (169, 100%) | 2 (134, 100%) | 2 (35, 100%) | - | |
Current number of vaccine doses (n, %) | - | 3 (126, 94%) | 3 (22, 62.9%) | < 0.0001 | 0.28 |
Anti-SP IgG (BAU/mL), median [IQR] | 147 [70; 320.5] | 149.5 [70.8; 331] | 129 [62; 259] | 0.5 | |
Neutralisation of Wuhan variant, median [IQR] | 3.9 [1.9; 7.9] | 4.1 [2; 7.9] | 3.8 [1.7; 7] | 0.7 | |
Neutralisation of Alpha variant, median [IQR] | 3.3 [1.5; 6] | 3.4 [1.5; 6.2] | 3.2 [1.5; 5.4] | 0.7 | |
Neutralisation of Delta variant, median [IQR] | 3.1 [1.6; 6.4] | 3 [1.5; 6.5] | 3.5 [1.8; 5.4] | 0.9 | |
SARS-CoV-2 specific T cell response (IU/mL), median [IQR] | 0.23 [0.1; 0.53] | 0.25 [0.12; 0.58] | 0.15 [0.08; 0.32] | 0.06 | 0.04 |
N Number; IQR Interquartile range; HCW Health care workers; Anti-SP Anti-Spike; IgG Immunoglobulin G; Anti-N Anti-Nucleocapsid. |
1Mann-Whitney test for quantitative variables; Fisher’s exact test for qualitative variables. All tests are two-sided with statistical significancy p < 0.05. To identify infection and bad vaccine responder’s predictive factors, all variables with p < 0.15 in unadjusted analysis were used for multivariable logistic regression models. *p < 0.05, **p < 0.005, ***p < 0.0005, ****p < 0.00005. |
Multivariable analysis of these factors showed than when an individual increases his age of one year, his risk of infection decreases 1.5% (p = 0.0084). In the same way for each additional unit of IFNγ secreted after specific T cell activation the risk of infection decreases of 73.5% (p = 0.04, Table 4).
Characterization of bad vaccine responders
Crossing data from infections during the Delta-Omicron wage, with both T cell and B cell responses after two doses of SARS-Cov-2 vaccines, we performed a ROC curve analysis for the levels of anti-SP IgG antibodies and SARS-CoV-2-induced IFNγ secretion. A threshold of 244 BAU/mL anti-SP IgG antibodies (100.0% sensibility, 78.6% specificity) and 0.285 IU/mL of IFNγ secretion (100% sensibility, 72.1% specificity) was established (Fig. 5B) defining four quadrants (Q). This allow to distinguish an area of values where no infections were found (Q4) from an area were individuals got infected (Q1-3) (Fig. 5C). Individuals in Q4 area were considered good vaccine responders and individuals in Q1-3 were bad vaccine responders or individuals with a declining SARS-CoV-2 immunity.
There were no statistical differences in gender, age or the number of vaccine doses between the two areas. Time between sampling and the second vaccine dose was higher in individuals in the Q1-3 area than in individuals in the Q4 area (186.5 vs 160.0 days, respectively, p = 0.0171). In the Q1-3 area 62 of 144 (43.1%) individuals displayed at least one comorbidity vs only five of 25 (20.0%) in the Q4 area (p = 0.03). As there was a difference in comorbidities among the two areas, we measured the general capability of T cells and innate immunity cells to secrete IFNγ after anti-CD3 and TLR 7/8 agonist stimulation. There were no statistical differences of IFNγ secretion capability between individuals from Q4 and Q1-3 (100.0 vs 165.0 IU/mL, respectively, p = 0.39) (Table 5).
Table 5
Characterisation of individuals in Q4 versus Q1-3
Characteristics | Q4 | Q1-3 | p-value1: Q4 vs Q1-3 |
Univariable | Multivariable |
n | 25 | 144 | | |
M/F | 6/19 | 28/116 | 0.8 | |
Age (years), median [IQR] | 48 [40.5; 58.5] | 50.5 [42.0; 57.0] | 0.5 | |
BMI (ratio), median [IQR] | 23.1 [21.5; 27.5] | 23.6 [21.7; 26.8} | 1.0 | |
Comorbidities (n, %) | 5 (20.0%) | 62 (43.1%) | 0.03 | 0.03 |
HCWs (n, %) | 2 (8.0%) | 36 (25.0%) | 0.07 | 0.18 |
Time between last dose and questionnaire (days), median [IQR] | 90.0 [76; 108] | 101.0 [86.0; 112.5] | 0.22 | |
Time between D2 and sampling (days), median [IQR] | 160.0 [107.5; 203.5] | 186.5 [154.0; 225.0] | 0.02 | 0.1 |
Current number of vaccine doses, median [IQR] | 3 (22, 88.0%) | 3 (125, 86.8%) | 1.0 | |
Anti-SP IgG (BAU/mL), median [IQR] | 414.0 [311.0; 766.0] | 124.0 [67.0; 221.0] | < 0,0001 | |
Neutralisation of Wuhan variant, median [IQR] | 8.5 [5.5; 12.7] | 2.8 [1.5; 6.3] | < 0,0001 | |
Neutralisation of Alpha variant, median [IQR] | 6.6 [5.3; 10.2] | 2.3 [1.2; 4.8] | < 0,0001 | |
Neutralisation of Delta variant, median [IQR] | 6.9 [5.8; 11.9] | 2.6 [1.3; 5.1] | < 0,0001 | |
SARS-CoV-2 specific T cell response (IU/mL), median [IQR] | 0.79 [0.49; 1.1] | 0.19 [0.08; 0.33] | < 0,0001 | |
Unspecific T cell stimulation (IU/ml), median [IQR] | 100.0 [20.0; 245.0] | 165.0 [26.0; 378.0] | 0.39 | |
N Number; IQR Interquartile range; HCW Health care workers; Anti-SP Anti-Spike; IgG Immunoglobulin G; Anti-N Anti-Nucleocapsid. |
1Mann-Whitney test for quantitative variables; Fisher’s exact test for qualitative variables. All tests are two-sided with statistical significancy p < 0.05. To identify infection and bad vaccine responder’s predictive factors, all variables with p < 0.15 in unadjusted analysis were used for multivariable logistic regression models. *p < 0.05, **p < 0.005, ***p < 0.0005, ****p < 0.00005. |
Multivariable analysis of the factors showed that only the presence of comorbidities explained the difference among individuals in the two areas (p = 0.03). An individual presenting at least one comorbidity was 3.2 times more likely to be in the Q1-3 area (Table 5). In our cohort the most common comorbidities were active smoking and alcohol drinking (30 of 62, 48.5%, for both). Ten of 62 individuals (16.1%) with comorbidities had an autoimmune disease, 14 (22.6%) had high blood pressure, four (6.5%) had cancer, three (4,8%) had chronic bronchitis and two (3.2%) had diabetes.