key: cord-1012152-fvdl23tu
authors: Desmecht, S.; Tashkeev, A.; Marechal, N.; Peree, H.; Tokunaga, Y.; Fombellida Lopez, C.; Polese, B.; Legrand, C.; Wery, M.; Mni, M.; Fouillien, N.; Toussaint, F.; Gillet, L.; Bureau, F.; Lutteri, L.; Hayette, M.-P.; Moutschen, M.; Meuris, C.; Desmecht, D.; Rahmouni, S.; Gilles, D.
title: Kinetics and persistence of the cellular and humoral immune responses to BNT162b2 mRNA vaccine in SARS-CoV-2-naive and -experienced subjects
date: 2022-01-27
journal: nan
DOI: 10.1101/2022.01.17.22269278
sha: 53420eab35178b349d4f3002b1c403800cbc250b
doc_id: 1012152
cord_uid: fvdl23tu

Background: Understanding and measuring the individual level of immune protection and its persistence at both humoral and cellular levels after SARS-CoV-2 vaccination is mandatory for the management of the vaccination booster campaign. Our prospective study was designed to assess the immunogenicity of the BNT162b2 mRNA vaccine in triggering the humoral and the cellular immune response in healthcare workers up to 6 months after two doses vaccination. Methods: This prospective study enrolled 208 healthcare workers from the Liege University Hospital (CHU) of Liege in Belgium. All participants received two doses of BioNTech/Pfizer COVID-19 vaccine (BNT162b2). Fifty participants were SARS-CoV-2 experienced (self-reported SARS-CoV-2 infection) and 158 were naive (no reported SARS-CoV-2 infection) before the vaccination. Blood sampling was performed at the day of the first (T0) and second (T1) vaccine doses administration, then at 2 weeks (T2), 4 weeks (T3) and 6 months (T4) after the 1st vaccine dose administration. A total of 1024 blood samples were collected. All samples were tested for the presence of anti-Spike antibodies using DiaSorin LIAISON SARS-CoV-2 TrimericS IgG assay. Neutralizing antibodies against the SARS-CoV-2 Wuhan-like variant strain were quantified in all samples using a Vero E6 cell-based neutralization-based assay. Cell-mediated immune response was evaluated at T4 on 80 participants by measuring the secretion of IFN-gamma on peripheral blood lymphocytes using the QuantiFERON Human IFN-gamma SARS-CoV-2, Qiagen. All participants were monitored on weekly-basis for the novo SARS-COV-2 infection for 4 weeks after the 1st vaccine dose administration. We analyzed separately the naive and experienced participants. Findings: We found that anti-spike antibodies and neutralization capacity levels were significantly higher in SARS-CoV-2 experienced healthcare workers (HCWs) compared to naive HCWs at all time points analyzed. Cellular immune response was similar in the two groups six months following 2nd dose of the vaccine. Reassuringly, most participants had a detectable cellular immune response to SARS-CoV-2 six months after vaccination. Besides the impact of SARS-CoV-2 infection history on immune response to BNT162b2 mRNA vaccine, we observed a significant negative correlation between age and persistence of humoral response. Cellular immune response was, however, not significantly correlated to age, although a trend towards a negative impact of age was observed. Conclusions: Our data strengthen previous findings demonstrating that immunization through vaccination combined with natural infection is better than 2 vaccine doses immunization or natural infection alone. It may have implications for personalizing mRNA vaccination regimens used to prevent severe COVID-19 and reduce the impact of the pandemic on the healthcare system. More specifically, it may help prioritizing vaccination, including for the deployment of booster doses.

Introduction age (18). However, information on how pre-existing immunity to SARS-CoV-2 would 109 be boosted by mRNA vaccination remains poorly understood. In particular, the 110 helpfulness and the timing of booster vaccine doses remain to be determined as well 111 as the impact of recovery after SARS-CoV-2 infection. Here, we analyzed the kinetic 112 of humoral response up to 6 months after BNT162b2. We also analyzed cellular 113 response to BNT162b2 6 months after the vaccination. In particular, we studied the 114 impact of previous SARS-CoV-2 infection on immune persistence. All participants 115

were monitored on weekly-based for the novo SARS-CoV-2 infection that may impact 116 immune persistence. CoV-2 infection. All enrolled participants were over 18 years of age. They all 126 acknowledged that they had understood the study protocol and signed the informed 127 consent. Participants were classified into two groups : "experienced group", that 128 includes participants with reported SARS-CoV-2 infection prior to vaccination, and 129 "naïve group", consisting of individuals without previous SARS-CoV -2 infection. 130 Blood was collected at the day of the 1st vaccine dose (T0), the day of second 131 vaccine dose (T1), then and at 14 days (T2), one month (T3) and six months (T4) 132 following the second vaccine dose. A total of about 40ml of blood was collected from 133 each subject at each timepoint. 134

The protocol was approved by the ethics committee (full name: comité d'éthique 135 hospitalo-facultaire universitaire de Liège) of Liège University Hospital (approval 136 number 2021-54). 137 . CC-BY-NC-ND 4.0 International license It is made available under a 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 27, 2022. ; https://doi.org/10.1101 https://doi.org/10. /2022 by peripheral blood lymphocytes using the QuantiFERON Human SARS-CoV-2 140 (Qiagen, Cat. No. / ID: 626410). Briefly, blood was collected on four tubes: the 141 control set including one positive and negative one negative tube and the two original 142

Vacutainer tubes containing SARS-CoV-2 antigen 1 (Ag1) and SARS-CoV-2 antigen 143 2 (Ag2) formulated to activate CD4 T (by Ag1) and both CD4 T and CD8 T (by Ag2) 144 lymphocytes in heparinized whole blood. After blood collection and mixing, tubes 145 were incubated at 37°C for 16 to 24h, IFN-γ was then measured on plasma from the 146 stimulated samples using an CLIA based assay. The threshold for positivity was 0.15 147 IU/mL as given by Qiagen. 148

Participants were all invited to perform a weekly saliva self-sampling to detect the 150 presence of SARS-CoV-2 following the instruction provided in the starting pack. 151

Participants were asked to return the samples within 12 hours after sampling to the 152 core facility to perform the required follow-up PCR testing. (22) . Considering the sampling timeline in this study, we preferred to model 200 . CC-BY-NC-ND 4.0 International license It is made available under a 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 27, 2022. ; https://doi.org/10.1101/2022.01.17.22269278 doi: medRxiv preprint the dependency of the response variables on the number of days post-vaccine by 201 using single-slope rate using the data starting from T2. Namely, the model is: In all corresponding procedures we used non-parametric statistics, i.e. either tie-220 corrected Spearman correlation coefficient or Wilcoxon rank-sum test. We used log-221 . CC-BY-NC-ND 4.0 International license It is made available under a 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 27, 2022. ; https://doi.org/10. 1101 /2022 Our study included 208 consenting subjects among the CHU of Liège staff members 231 who received the Pfizer-BioNTech BNT162b2 vaccine during February 2021. Among 232 them, 158 were SARS-CoV-2 naive subjects while 50 had COVID-19 prior to 233 vaccination (experienced SARS-CoV-2). The cohort was skewed toward females 234 (~80%). The age was however balanced between naive (34-53 years old) and 235 experienced groups (37-55 years old) with median of 43 and 46 respectively. The 236 BMI was also balanced between the two groups with a median of 24,8 (21,3 to 27,7) 237 for naïve and 24,0 (21,4 to 27,3) for experienced. Comorbidities of the participants 238 included asthma (11% in naive group and 6,2% in experienced group), autoimmunity 239 autoimmunity (5,3% in naive group -3 with Hashimoto's thyroiditis, 3 with 240 hypothyroidism, 2 with Chron's disease, 1 with psoriasis, 1 with diabetes and 241 rheumatoid arthritis, and 2,1% in experienced group -2 with Lupus and 1 with 242 psoriatic arthritis), blood cancer (non-active, 2% in naive group and 0% in 243 experienced group), other cancers (non-active, 3,9% in naive group and 4,2% in 244 experienced group) and immunodeficiency (1,3% in naive group and 2,1% in 245 experienced group (Table 1) . 246

In total, 1024 samples were collected. Sampling was performed the day of 1 st and 2 nd 247 dose of the vaccine (T0 and T1), then 2 weeks, 4 weeks and 6 months after the 1 st 248 dose (T2, T3 and T4). 73% of naive participants donated blood at all time points 249 while 21% and 5,7% donated blood 4 and 3 times respectively. In the experienced 250 group, 82% of subjects donated blood at all time points while 10% and 8% donated 4 251 and 3 times respectively ( Table 1) . 252 increased rapidly after the 1 st vaccine dose in both groups but with higher titer in the 260 . CC-BY-NC-ND 4.0 International license It is made available under a 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 27, 2022. ; https://doi.org/10.1101/2022.01.17.22269278 doi: medRxiv preprint maximum two weeks after the 2 nd vaccine dose and start declining two weeks later. 263

This decrease continued over time between two and six months after the 1 st vaccine 264 dose with a half-life of 55 days. The kinetics of anti-Spike antibodies production was 265 similar between the naive and experienced groups. The levels of produced 266 antibodies were however significantly higher in the experienced group over the naive 267 group at all time points (Figure 1 A, B) . 268

To evaluate the quality of secreted antibodies, we measured neutralization capacities 269 at all time points and all subjects using previously reported protocols (19). More than In clinical practice, measuring antibodies neutralization capacities is complicated and 288 time consuming. Therefore, and since measuring anti-Spike IgG level is routinely 289 used, we modeled the ability to use anti-Spike antibodies levels to predict whether a 290 person still has some neutralization capacity at some reasonably far time point after 291 vaccination (6 months in the case of our study) (Figure 1E, F) . Logistic regression 292 classifier allowed to separate individuals with no detectable neutralization at the last 293 time point and the others by the anti-Spike IgG at the last time point (sensitivity: 294 . CC-BY-NC-ND 4.0 International license It is made available under a 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 27, 2022. ; https://doi.org/10.1101/2022.01.17.22269278 doi: medRxiv preprint mean=94%, std=3%; specificity: mean=70%, std=14% across five stratified folds of 295 cross-validation). The decision value estimated on the whole dataset was 326 296 BAU/ml. Using anti-Spike IgG level at earlier time points did not allow the 297 classification ( Figure S1) . (Figure 2C) , or 6 months after vaccination (T4) (Figure 2D ). Among 310 all tested clinical parameters, only age was significantly associated with the 311 neutralizing status (decreasing chances to have neutralizing capacity), and even its 312 effect size was far less (1.4-5.3x) than that of the parameter representing anti-Spike 313

IgG level. No effect of sex, smoking status, and BMI were detected at any 314 significance level (Figure 2) . 315

We next tested the association of humoral response parameters themselves with the 317 abovementioned set of clinical parameters across all time points (Figure 3) . Since 318 NT50 and anti-Spike IgG levels were strongly correlated at each time point for both 319 SARS-CoV-2-experienced and naive individuals (Figure 3A) , we tested them in 320 separate models using either multivariate (Figure 3B, D) or univariate (Figure 3C , E) 321 multiple linear regression, i.e. considering values across time points as a single 322 vector or separately. Again, only age showed a significant association with the 323 outcomes explaining 25% and 21% of conditional variance in multivariate models for 324 anti-Spike IgG level and NT50, respectively. Interestingly, effects of age in univariate 325 models were negative (i.e. decreases humoral response parameter value) at all time 326

points except the first one, at which the effect was significant but of the opposite sign. 327

One plausible hypothesis could be that older individuals are more likely to face cold 328 . CC-BY-NC-ND 4.0 International license It is made available under a 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 27, 2022. To evaluate T-cell specific immune response to BNT162b2 vaccine, heparinized 333 whole blood was stimulated with SARS-CoV-2 specific peptides contained in the 334

QuantiFERON-SARS-CoV-2 tubes designed to activate both CD4T and CD8T cells. 335 IFN-γ was then measured by CLIA. Since this assay was only commercially available 336 after we completed the collection of samples at T0 to T3, we only assayed cellular 337 immune response at the last time point (T4) (Figure 4) . Cellular response was 338 evaluated on 80 participants with 63 in the naive group and 17 in the experienced 339 group. In terms of IFN-γ secretion, we did not observe any significant difference 340 between naive and experienced groups for none of the antigen used (Ag1 and Ag2) 341

( Figure 4A, B) . However, when subjects were clustered based on their neutralizing 342 capacities at 6 months after vaccination, the difference was highly significant 343 between those with high neutralizing titers versus the one with no detectable 344 neutralization ( Figure 4B) . We next investigated if there was any association 345 between IFN-γ secretion, anti-Spike total antibodies and neutralizing antibodies titers. 346

As shown in figure 4A , there was no correlation between anti-S IgG and IFN-γ in the 347 naive group for both Ag1 and Ag2. Interestingly, IFN-γ secretion was highly 348 correlated to anti-Spike IgG in the experienced group (Figure 4C top panel) . IFN-γ 349 secretion was also correlated to neutralization titers in all groups, although the 350 correlation was most significant in the experienced group for Ag1 (p=3e-4) over the 351 naive group (p=0,0011) and vice versa for Ag2 (p= 0.0026 for experienced and 352 p=0,00029 for naive) (Figure 4C lower panel) . There was no statistically significant 353 association between cellular immune response parameters and any of the available 354 clinical parameters (Figure S3) . 355 356 COVID-19 testing of the study participants.

Participants were all invited to perform a weekly saliva self-sampling to detect the 359 presence of SARS-CoV-2. Nevertheless, the adherence to the weekly SARS-CoV-2 360 . CC-BY-NC-ND 4.0 International license It is made available under a 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 27, 2022. ; https://doi.org/10.1101/2022.01.17.22269278 doi: medRxiv preprint positive during the study timeline. 362

Our study confirms previous evidence for an earlier, stronger and more persistent 364 humoral immune response in individuals previously infected with SARS-CoV-2 365 versus naïve individuals following vaccination with the BNT162b2 mRNA vaccine 366 (24). The anti-spike antibody and neutralization capacity levels six months after 367 vaccination protocol were significantly higher in SARS-CoV-2 experienced HCWs 368 compared to naïve HCWs. We did not observe such differences regarding cellular 369 immune response six months following 2 nd dose vaccination. Reassuringly, most 370 participants had a detectable cellular immune response to SARS-CoV-2 six months 371 after vaccination. These findings are in line with a recent observation from 372

Samanovic and colleagues who also reported more pronounced differences in 373

recently SARS-CoV-2 experienced subjects (25). 375

BNT162b2 mRNA vaccine, we observed a significant association between age and 377 persistence of humoral response. The more elderly is a participant, the less durable 378 was the humoral response. This is in line with results from other studies 379 demonstrating a similar decrease of anti-SARS-CoV-2 antibodies in all age groups 380 few months after the second vaccine dose, especially among 65 years-old or older 381 persons (26). Regarding cellular immune response, although not statistically 382 significant, we observed a trend towards a negative impact of age. 383

Our results confirm previous findings indicating that anti-spike antibodies are very 384 well correlated with neutralizing antibodies (27)(28). This may allow to set a 385 threshold of anti-spike antibodies predicting neutralization capacity with a high 386 sensitivity. Although correlates of protection from SARS-CoV-2 are not fully defined 387 yet, Khoury and colleagues showed that neutralization level is highly predictive of 388 immune protection, reinforcing the results of other reports suggesting that 389 neutralization titer is an important predictor of vaccine efficacy (21)(29)(30). Our 390 findings may thus have important implications for waning vaccine immunity and 391 vaccination strategies. Our results may also contribute to better define patients who 392 . CC-BY-NC-ND 4.0 International license It is made available under a 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 27, 2022. ; https://doi.org/10.1101 https://doi.org/10. /2022 SARS-CoV-2 mAb products (PMID: 35016195) (PMID: 34087172). Available stocks 394 of mAbs that retained activity against Omicron are extremely limited in most settings, 395 while many vulnerable patients can be considered as eligible for this treatment. 396

Predictors of remaining neutralization capacity may help clinicians with the difficult 397 selection of patients for whom this intervention would be most beneficial. 398

The humoral immune response as measured through quantification of anti-spike IgG 399 or neutralizing antibodies correlated with cellular immune responses six months 400 following vaccination. Interestingly, this correlation was much stronger in the group of 401 participants with a history of SARS-CoV-2 infection, indicating that the level of anti-402 spike antibodies in this group may not only predict the level of neutralizing antibody 403 but also the level of cellular immunity as well. 404

In conclusion, our data strongly reinforce the relevance of previous SARS-CoV -2 405 infection for understanding vaccine immune responses. It may have implications for 406 personalizing mRNA vaccination regimens used to prevent severe COVID-19 and 407 reduce the impact of the pandemic on the healthcare system. More specifically, it 408 may help prioritizing vaccination, including for the deployment of booster doses. 409 410 . CC-BY-NC-ND 4.0 International license It is made available under a 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 27, 2022. CC-BY-NC-ND 4.0 International license It is made available under a 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 27, 2022. ; https://doi.org/10.1101 https://doi.org/10. /2022 . CC-BY-NC-ND 4.0 International license It is made available under a 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 27, 2022. ; https://doi.org/10.1101 https://doi.org/10. /2022 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 27, 2022. ; https://doi.org/10.1101 https://doi.org/10. /2022 CC-BY-NC-ND 4.0 International license It is made available under a 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 27, 2022. ; https://doi.org/10.1101 https://doi.org/10. /2022 CC-BY-NC-ND 4.0 International license It is made available under a 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 27, 2022. ; https://doi.org/10.1101 https://doi.org/10. /2022 

. CC-BY-NC-ND 4.0 International license It is made available under a 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 27, 2022. CC-BY-NC-ND 4.0 International license It is made available under a 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 27, 2022. ; https://doi.org/10.1101/2022.01.17.22269278 doi: medRxiv preprint

. CC-BY-NC-ND 4.0 International license It is made available under a 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 27, 2022. . CC-BY-NC-ND 4.0 International license It is made available under a 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 27, 2022. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted January 27, 2022. ; https://doi.org/10.1101/2022.01.17.22269278 doi: medRxiv preprint 501 502 503 Figure S1 : Inability to separate individuals neutralizing and non-neutralizing at last 504 time point by anti-Spike IgG values at earlier time points. G-mean is a geometric 505 mean of sensitivity and specificity, a single metric used to evaluate the model. 506 507 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted January 27, 2022. ; https://doi.org/10.1101/2022.01.17.22269278 doi: medRxiv preprint 509 Figure S2 : Absence of impact of time after recovery on humoral response 510 parameters before and after 1st vaccine dose. 511 512 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted January 27, 2022. ; https://doi.org/10.1101/2022.01.17.22269278 doi: medRxiv preprint Figure S3 : Effects of clinical parameters on cellular response at the last time point. 515 516 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted January 27, 2022. ; https://doi.org/10.1101/2022.01.17.22269278 doi: medRxiv preprint 519 Figure S4 : Statistics of COVID-19 testing of the study participants. Only one 520 participant was tested as positive during the study timeline. 521 522 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted January 27, 2022. ; https://doi.org/10.1101/2022.01.17.22269278 doi: medRxiv preprint mRNA-1273 against Delta, Mu, and other emerging variants. medRxiv. 2021;

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted January 27, 2022. ; https://doi.org/10. 1101 /2022 

Safety and 524 Efficacy of the BNT162b2 mRNA Covid-19 Vaccine

Efficacy and 526 Safety of the mRNA-1273 SARS-CoV-2 Vaccine

Effectiveness of BNT162b2 Vaccine against Delta Variant in Adolescents

Evaluation 531 of mRNA-1273 SARS-CoV-2 Vaccine in Adolescents

Evaluation of the BNT162b2 Covid-19 Vaccine in Children 5 to 11 Years of Age

Transmission of SARS-CoV-2 after COVID-19 Screening and Mitigation Measures for 537

The Risk of Underestimating the 540

Contribution of Children to Severe Acute Respiratory Syndrome Coronavirus

CoV-2) Pandemic

Waning Immunity after the BNT162b2 Vaccine in Israel

Waning of BNT162b2 Vaccine Protection against SARS-CoV-2 Infection in Qatar

BNT162b2 vaccination induces durable SARS-CoV-2-specific T cells with a stem cell 555 memory phenotype

mRNA vaccines 557 induce durable immune memory to SARS-CoV-2 and variants of concern

Neutralization of the SARS-CoV-2 Mu Variant by Convalescent and Vaccine Serum

et 563 al. mRNA-based COVID-19 vaccine boosters induce neutralizing immunity against 564 SARS-CoV-2 Omicron variant

Effects of BNT162b2 Covid-19 Vaccine Booster in Long-Term Care Facilities in Israel

Effectiveness 571 of a third dose of the BNT162b2 mRNA COVID-19 vaccine for preventing severe 572 outcomes in Israel: an observational study

Biologic Methods for SARS-associated Coronavirus Infection

Evolution of 577 immune responses to SARS-CoV-2 in mild-moderate COVID-19

Fast implementation for normal mixed effects models with censored 23

Scaling regression inputs by dividing by two standard deviations

Influence 586 of past infection with SARS-CoV-2 on the response to the BNT162b2 mRNA vaccine 587 in health care workers: Kinetics and durability of the humoral immune response

Robust immune responses are observed after one dose of BNT162b2 mRNA 591 vaccine dose in SARS-CoV-2 experienced individuals

Humoral Response to BNT162b2 Covid-19 Vaccine over 6 Months

Binding and Neutralization Antibody Titers after a Single Vaccine Dose in Health Care 598

Workers Previously Infected with SARS-CoV-2

Immunological memory 600 to SARS-CoV-2 assessed for up to 8 months after infection

Neutralizing antibodies correlate with protection from SARS-CoV-2 in humans during a 603 fishery vessel outbreak with a high attack rate

Antibody Status and Incidence of SARS-CoV-2 Infection in Health Care Workers