key: cord-0282647-nrm40ni1 authors: Suntronwong, N.; Yorsaeng, R.; Auphimai, C.; Thongmee, T.; Vichaiwattana, P.; Kanokudom, S.; Assawakosri, S.; Nilyanimit, P.; Srimuan, D.; Thatsanatorn, T.; Sudhinaraset, N.; Wanlapakorn, N.; Poovorawan, Y. title: Immune responses to inactivated and vector-based vaccines in individuals previously infected with SARS-CoV-2 date: 2022-01-05 journal: nan DOI: 10.1101/2022.01.03.22268704 sha: bff6ca30b4ae879d899404261f5acf547e147f46 doc_id: 282647 cord_uid: nrm40ni1 Immunity wanes in individuals previously infected with SARS-CoV-2, and vaccinating those individuals may help reduce reinfection. Herein, reactogenicity and immunogenicity following vaccination with inactivated (CoronaVac) and vector-based (ChAdOx1-S, AZD1222) vaccines were examined in previously infected individuals. Immune response was also compared between short and long intervals between first date of detection and vaccination. Adverse events were mild but were higher with AZD1222 than with CoronaVac. Baseline IgG-specific antibodies and neutralizing activity were significantly higher with shorter than longer intervals. With a single-dose vaccine, IgG and IgA-specific binding antibodies, neutralizing activity, and total interferon-gamma response peaked at 14 days. Immune response was significantly higher in recovered individuals than in infection-naive individuals. Antibody response was greater with longer than shorter intervals. AZD1222 induced higher antibody and T cell responses than those of CoronaVac. Thus, to achieve immunity, individuals with prior SARS-CoV-2 exposure may require only a single dose of AZD1222 or two doses of CoronaVac to achieve the immune response. These findings supported vaccine strategies in previously infected individuals. Coronavirus disease 2019 (COVID-19) caused by the severe acute respiratory syndrome 20 coronavirus-2 (SARS-CoV-2) continues to circulate globally and has been difficult to control 21 even though COVID-19 vaccines are available 1 . Ongoing evolution and emergence of SARS-22 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted January 5, 2022. ; https://doi.org/10.1101/2022.01.03.22268704 doi: medRxiv preprint CoV-2 variants with significant mutations can make the virus more easily transmitted and less 23 susceptible to neutralizing antibodies, slowing control of the pandemic 2,3 . In addition, immunity 24 against SARS-CoV-2 wanes, even when induced through infection and vaccination, which has 25 raised questions about long-term immunity 4,5 . A recent study indicated lower antibody titers 26 against the spike protein of SARS-CoV-2 were associated with breakthrough infections 6 . Thus, 27 vaccine strategies that promote robust protective immunity and maintain antibodies over time 28 should be considered. 29 Natural infection with SARS-CoV-2 can elicit both humoral and T cell-mediated immune 30 responses; however, length and severity of illness after SARS-CoV-2 infection were generally 31 correlated with magnitude of immune response 7 . Infected individuals typically induce higher 32 titers of neutralizing antibodies that target the viral spike glycoprotein (S) 8 . The SARS-CoV-2 33 spike protein is homo-trimeric and is cleaved into two subunits, S1 and S2 9 . The S1 subunit 34 contains the N-terminal domain and a receptor-binding domain (RBD), which typically targets 35 angiotensin-converting enzyme 2 (ACE2); both regions can potentially induce neutralizing 36 activities. Simultaneously, the S2 subunit mediates viral membrane fusion during viral entry and 37 is typically conserved across coronaviruses. 38 Although SARS-CoV-2 infection can produce relatively high levels of neutralizing 39 antibodies that correlate with disease protection 10,11 , antibodies against SARS-CoV-2 appear to 40 decline over time 12, 13 . Antibody titers twelve months postinfection indicate that specific IgG 41 against receptor binding sites decrease by 68.1% compared with that in the first month 14 . Despite 42 a decline in antibody response, T cell-mediated immune response likely persists for several 43 months and may persist longer than detectable antibodies in recovered patients 15, 16 . 44 Immunological memory to SARS-CoV-2 remains detectable for up to eight months following 45 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted January 5, 2022. ; https://doi.org/10.1101/2022.01.03.22268704 doi: medRxiv preprint 7 exposure to SARS-CoV-2. As expected, although significantly higher IgG-specific SARS-CoV-2 111 antibodies were observed in short-interval participants than in their counterparts, IgG antibodies 112 in participants with long intervals were detected above the cutoff level at baseline. This result 113 indicated that IgG-specific antibodies, especially anti-RBD and anti-S IgG, persisted more than 114 one year following natural infection. However, there was no difference in anti-S1 IgA between 115 short and long intervals ( Supplementary Fig. 2a-d) . 116 117 Binding antibody response after vaccination was compared in participants with short and 118 long intervals since infection and in those without prior infection. Levels of IgG antibodies 119 specific to N and RBD and IgA-specific S1 were measured at different time points (Fig. 3) . 120 Following a single dose of CoronaVac, anti-N IgG seropositivity increased slightly but not 121 significantly in participants with short intervals (Fig. 3a) . By contrast, anti-N IgG increased 122 significantly in those with long intervals after the first and second doses of the CoronaVac 123 vaccine (p < 0.001). As expected, the level of anti-N IgG did not increase after AZD1222 124 vaccination. 125 The anti-RBD IgG titer in all vaccinated previously infected participants increased 126 significantly and peaked at 14 d following a single dose of vaccine, compared with before 127 vaccination (p < 0.001; Fig. 3b ). At 28 d after single-dose vaccination with AZD1222, short and 128 long-interval groups had levels of anti-RBD IgG that were 12.87-fold (1,549 binding antibody 129 units per millilitre (BAU/mL) and 73.79-fold (2,258 BAU/mL) higher, respectively, than those 130 of baseline. By contrast, after a single dose of CoronaVac vaccine, anti-RBD IgG levels were 131 3.12-fold (235 BAU/mL) and 7.79-fold (257 BAU/mL) higher in short and long-interval groups, 132 respectively. Furthermore, anti-RBD IgG level increased significantly following the second dose 133 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted January 5, 2022. ; https://doi.org/10.1101/2022.01.03.22268704 doi: medRxiv preprint 8 of CoronaVac compared with that after the first dose (p < 0.01). By contrast, anti-RBD IgG level 134 decreased significantly after the second dose of AZD1222, compared with the first dose (p < 135 0.001). However, a single dose of vaccine in recovered individuals led to higher anti-RBD IgG 136 levels than those of infection-naïve individuals with complete two-dose vaccination (p < 0.01). A 137 similar response was observed in levels of anti-S1 IgG ( Supplementary Fig. 3) . 138 In addition to measuring IgG, anti-S1 IgA levels were also tested, and anti-S1 IgA was 140 detected in 84.61% (99/117) of participants before vaccination (Fig. 3c ). At 28 d after the first 141 dose of CoronaVac, anti-S1 IgA seropositivity in short and long-interval groups was 93.33% 142 (28/30) and 92.85% (26/28), respectively. At 28 d after a single dose of AZD1222, 100% of 143 previously infected participants were seropositive for anti-S1 IgA. Furthermore, with single-dose 144 vaccination in recovered participants, anti-S1 IgA levels increased significantly compared with 145 those in uninfected individuals postvaccination (p < 0.001). 146 147 Neutralizing activities against SARS-CoV-2 wild type and variants 148 In addition to the binding antibody response, antibody function was further determined 149 using a surrogate virus neutralization test (Fig. 4) . First, in serum samples collected at different 150 time points, presence of neutralizing antibodies against wild-type SARS-CoV-2 was determined 151 using a NeutraLISA assay based on blockage of the ACE2-RBD protein-protein interaction. 152 Previously infected individuals had a strong response following a single dose of vaccine, and 153 87.93% (51/58) of those with CoronaVac and 100% (59/59) of those with AZD1222 developed 154 neutralizing activities. Those levels of neutralizing activity were significantly higher than those 155 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. Neutralizing potential of sera collected one month after first and second doses of vaccines 163 against SARS-CoV-2 wild type and variants was measured using cPass TM ELISA assays ( Fig. 164 4c-f). A single dose of vaccine was very influential in broadly inducing antibodies to SARS-165 CoV-2 variants in almost all participants with prior infection, and antibody function was 166 significantly higher than that in infection-naïve individuals (p < 0.001). All participants who 167 experienced infection had detectable neutralizing activity against the wild type, based on the 168 cPass TM assay (Fig. 4c) . Nevertheless, 27/29 (93.1%) and 26/27 (96.3%) of individuals with 169 short and long intervals, respectively, showed positive neutralization against B.1.1.7 following a 170 single dose of CoronaVac (Fig. 4d ). All participants vaccinated with AZD1222 had strong 171 neutralizing activity, with blockage of ACE2-RBD exceeding 97%. As expected, there was a 172 decrease in SARS-CoV-2 neutralizing activity against the B.1.351 variant (Fig. 4e) . 173 Notably, more than 96% of individuals with previous infection had a robust increase in 174 neutralizing antibodies against B.1.617.2 after a single dose of CoronaVac, and the percentage 175 reached 100% for the AZD1222 vaccine (Fig. 4f) . Overall, neutralizing activity increased 176 significantly after two doses of the CoronaVac vaccine. However, there was no significant 177 increase in neutralizing activity against SARS-CoV-2 variants following the second dose of 178 AZD1222, except against B.1.1.7. Furthermore, participants with long intervals had higher levels 179 of neutralizing activity than those with short intervals. These findings indicated that neutralizing 180 activity after vaccination was greater in participants with longer intervals than in those with 181 shorter intervals. 182 183 Total interferon-gamma response following immunization 184 Total IFN-γ responses stimulated by CD4+ epitopes derived from the S1 subunit (RBD) 185 (Ag1) and by CD4+ and CD8+ epitopes derived from S1 and S2 subunits of the spike protein 186 When total interferon-gamma response was compared between first and second doses, 194 CoronaVac recipients had a slight increase in interferon-gamma response for Ag2 after two 195 doses of the vaccine (Fig. 5e) . By contrast, level of interferon-gamma response after two doses of 196 AZD1222 decreased significantly (Fig. 5f ). These results suggest that a single dose of AZD1222 197 and two doses of CoronaVac are sufficient to induce total interferon-gamma response in 198 individuals previously infected with SARS-CoV-2. 199 200 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted January 5, 2022. individuals following a single dose of mRNA vaccine, with antibody titers gradually increasing 215 over seven days and reaching a peak at day 14 after immunization 24,31 . CoronaVac contained the 216 SARS-CoV-2 nucleocapsid protein, whereas AZD1222 did not. Therefore, only immunization 217 with the inactivated vaccine significantly induced anti-N IgG. A long-term study indicates that 218 anti-N antibodies can be detected for several months after natural infection 32 . Additionally, anti-219 N antibodies may protect against SARS-CoV-2 disease by promoting T-cell immunity 33 . 220 After boosting, anti-RBD and S1-specific IgG levels were significantly higher with 221 AZD1222 than with CoronaVac. Although there was greater decay of anti-S1 and anti-RBD IgG 222 after AZD1222 vaccination, a relatively high level of neutralizing activity against the wild type 223 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted January 5, 2022. variants 41 . Hybrid immunity suggests that those exposed to SARS-CoV-2 infection with 246 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted January 5, 2022. CoV-2-specific T-cell response is maintained for several months after natural infection and 261 vaccination 16 . Additionally, a recent study highlighted the potential for variants to escape from 262 neutralizing humoral immunity but not from cell-mediated immunity 47 . Moreover, in a previous 263 study, mutations in spike epitopes did not impair T-cell responses, and escaping neutralizing 264 antibodies indicated that T cells might play a significant role in broadly protecting against 265 SARS-CoV-2 variants 48 . 266 With both vaccines, a long interval between infection and vaccination led to a better 267 immune response than that of a short interval. These findings are consistent with those of a study 268 that found an extended gap between infection and vaccination increased peak antibody response 269 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted January 5, 2022. There were some limitations to the study. First, neutralizing titers were not examined. 281 Laboratory testing was limited, because titers of anti-S1 IgA, neutralizing activity, and total 282 interferon-gamma responses exceeded upper detection limits. Thus, exact levels of immune 283 response after boosting could not be summarized. Moreover, other SARS-CoV-2-particular 284 proteins, such as matrix and nucleocapsid proteins, which induce total interferon-gamma 285 release 52 , did not stimulate SARS-CoV-2-specific T cells. Last, small sample size was a 286 limitation. Participants were not randomly assigned to receive different types of vaccine, and 287 vaccine groups were assigned on the basis of convenience. 288 Further investigation is warranted to assess the durability of antibody and T-cell 289 responses after vaccination of people with previous SARS-CoV-2 infection, as well as the 290 interplay between natural and vaccine-induced immunity. Notably, some vaccinated participants 291 in this study failed to produce a booster response, particularly a T-cell response; thus, further 292 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted January 5, 2022. IgA should be further investigated. 296 In conclusion, previously infected SARS-CoV-2 individuals developed a more robust 297 immune response after AZD1222 or CoronaVac vaccination than that in naïve individuals. 298 Notably, vaccination with AZD1222 produced higher antibody levels, neutralizing activity, and 299 T-cell responses than those with the CoronaVac vaccine. In addition, a longer interval between 300 infection and first vaccination improves the immune response more than that of a shorter one. 301 The results suggest that a single dose of AZD1222 or two doses CoronaVac can boost antibody 302 and T-cell responses in individuals with prior infection, which is a conclusion that can be 303 paramount in facilitating vaccine policies. 304 305 One hundred and seventeen recovered COVID-19 patients who were healthy adults aged 308 ≥18 years and previously infected with SARS-CoV-2 (defined as anti-nucleocapsid positivity 309 (IgG) or a history of positive SARS-CoV-2 detection) were enrolled with written consent. 310 According to the duration between the first date of positive SARS-CoV-2 detection and 311 vaccination, there were two groups of participants: the short-interval group with 2 to 5 months 312 between dates (infection between January 2 and April 13, 2021), which included 60 participants, 313 and the long-interval group with 13 to 15 months between dates (infection between March 3 and 314 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted January 5, 2022. ; https://doi.org/10.1101/2022.01.03.22268704 doi: medRxiv preprint April 4, 2020), which included 57 participants (Fig. 1) (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted January 5, 2022. The IgG antibodies against SARS-CoV-2 S1/RBD were quantitatively measured using CE 353 marked SARS-CoV-2 Quant IgG II (Abbott Diagnostics), and antibody level is expressed as 354 BAU/mL, with values ≥7.1 BAU/mL defined as positive. In addition, anti-S1 IgG was measured 355 using LIASON SARS-CoV-2 Trimeric S IgG (DiaSorin, S.p.A, Saluggia, Italy), with results 356 reported in BAU/mL and with a cutoff ≥33.8 BAU/mL defined as positive. 357 358 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted January 5, 2022. Quantification of interferon-gamma response 381 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted January 5, 2022. ; https://doi.org/10.1101/2022.01.03.22268704 doi: medRxiv preprint April 13, 2021, and long-interval participants were infected from March 3 to April 4, 2020. 567 Those two groups each included one group that received CoronaVac and one group that received 568 AZD1222. The short-interval group had a single-dose vaccination, and the long-interval group 569 had a two-dose vaccination. Blood samples were collected at four different time points (0, 14, 570 and 28 d after the first dose and one month after the second dose). 571 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted January 5, 2022. Swelling was graded as mild (<5 cm), moderate (5 cm to <10 cm), and severe (≥10 cm). 579 Symptoms were graded as follows: mild, no limitation on normal activity; moderate, some 580 limitation of daily activity; and severe, unable to perform or daily activity prevented. 581 582 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted January 5, 2022. using Kruskal-Wallis tests with Dunns' post hoc correction (*p < 0.05, **p < 0.01, and ***p < 596 0.001). 597 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted January 5, 2022. ; https://doi.org/10.1101/2022.01.03.22268704 doi: medRxiv preprint and n=27 for 13-15 months) vaccine on days 0, 14, and 28 after the first dose were stimulated by 616 a Ag1, which is a CD4+ epitope derived from RBD, minus negative control (Nil), and b Ag2, 617 which is CD4+ and CD8+ epitopes derived from S1 and S2 subunits, minus negative control 618 (Nil). Total interferon-gamma response of those with the AZD1222 (n=30 for 2-5 months and 619 n=29 for 13-15 months) vaccine to SARS-CoV-2 c Ag1 and d Ag2 after the first dose at 620 different time points. Interferon-gamma responses above cutoff values (0.15 international units 621 per millilitre (IU/mL)) were detected in those with e CoronaVac (n = 12) and f AZD1222 (n = 622 27) vaccines after the second dose. Horizontal bars indicate the median. Two-tailed pair-matched 623 comparisons were performed using Wilcoxon signed-rank tests. Statistical significance is 624 indicated as follows: *p < 0.05, **p < 0.01, and ***p < 0.001. 625 626 13 (11.1 %) All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted January 5, 2022. Data presents as n (%), mean (range). 630 DM: Diabetes Mellitus, HT: Hypertension, CHD: Coronary heart disease 631 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted January 5, 2022. ; https://doi.org/10.1101/2022.01.03.22268704 doi: medRxiv preprint Twelve-month specific IgG response to SARS-CoV-2 receptor-binding 467 domain among COVID-19 convalescent plasma donors in Wuhan Cellular immunity in COVID-19 convalescents with PCR-470 confirmed infection but with undetectable SARS-CoV-2-specific IgG Robust SARS-CoV-2-specific T cell immunity is maintained at 6 months 473 following primary infection Longitudinal analysis shows durable and broad immune memory after 475 SARS-CoV-2 infection with persisting antibody responses and memory B and T cells Immunological memory to SARS-CoV-2 assessed for up to 8 months 478 after infection Assessment of SARS-CoV-2 Reinfection 1 Year After Primary Infection 480 in a Population in Lombardy, Italy SARS-CoV-2 vaccines in development Antibody responses after a single dose of SARS-CoV-2 mRNA vaccine Antibody responses to the BNT162b2 mRNA vaccine in individuals 485 previously infected with SARS-CoV-2 Effect of previous SARS-CoV-2 infection on humoral and T-cell 487 responses to single-dose BNT162b2 vaccine The SARS-CoV-2-specific T-cell response was evaluated by conducting a whole-blood 382Interferon-Gamma Release Assay according to the manufacturer's instructions (QuantiFERON, 383 Qiagen, Hilden, Germany). In this assay, whole heparinized blood was collected and then 384 transferred to blood collection tubes coated with SARS-CoV-2-specific antigen peptides to 385 stimulate cell-mediated immunity. The blood collection tubes consisted of two antigen tubes and 386 positive and negative controls. Antigen tubes were coated with CD4+ epitopes derived from the 387 S1 subunit (RBD) (Ag1) and CD4+ and CD8+ epitopes derived from S1 and S2 subunits of the 388 spike protein (Ag2). After 1 mL of whole blood was added, QFN collection tubes were incubated 389for 24 h at 37 °C and then centrifuged to collect the plasma. The interferon-gamma release from 390 stimulated samples was detected using ELISA according to the manufacturer's guidelines 391 (Qiagen). Following ELISA, interferon-gamma concentration was quantified based on the eight-392 point standard (0.125 to 8 IU/mL) and calculated as IU/mL by QuantiFERON RD (v5.03) 393 software. The lower detection limit was 0.065 IU/mL, and IFN-g values ≥10 IU/mL were 394 defined as 10 IU/mL. Results of IFN-g from Ag1 and Ag2 were subtracted from unstimulated 395 controls (Nil) to represent the interferon-gamma response related to the SARS-CoV-2-specific T-396 cell stimulation. After Outcomes are reported as geometric mean titer (GMT) with 95% confidence intervals (95% CI) 404 All rights reserved. No reuse allowed without permission.(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint this version posted January 5, 2022. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint this version posted January 5, 2022. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint this version posted January 5, 2022. ; https://doi.org/10.1101/2022.01.03.22268704 doi: medRxiv preprint All rights reserved. No reuse allowed without permission.(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint this version posted January 5, 2022. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint this version posted January 5, 2022. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint this version posted January 5, 2022. ; https://doi.org/10.1101/2022.01.03.22268704 doi: medRxiv preprint All rights reserved. No reuse allowed without permission.(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint this version posted January 5, 2022. and ***p < 0.001). 612 All rights reserved. No reuse allowed without permission.(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint this version posted January 5, 2022. ; https://doi.org/10.1101/2022.01.03.22268704 doi: medRxiv preprint 33 613 antigens. Serum samples from participants that received the CoronaVac (n=30 for 2-5 months 615 All rights reserved. No reuse allowed without permission.(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint this version posted January 5, 2022. ; https://doi.org/10.1101/2022.01.03.22268704 doi: medRxiv preprint