key: cord-1052212-iozkxjkh authors: Boelig, R. C.; Chaudhury, S.; Aghai, Z.; Oliver, E.; Mancuso, F.; Berghella, V.; Bergmann-Leitner, E. C. title: Comprehensive Serological Profile and Specificity of Maternal and Neonatal Cord Blood SARS CoV-2 Antibodies date: 2021-12-07 journal: medRxiv : the preprint server for health sciences DOI: 10.1101/2021.12.06.21267328 sha: 2c378f8f082fde7894ad94bae10def95dc49710e doc_id: 1052212 cord_uid: iozkxjkh Objective: To describe the profile and specificity of maternal and neonatal cord-blood antibody profile in response SARS-CoV-2 virus exposure Methods: This is a Prospective cohort study of delivering patients at Thomas Jefferson University Hospital from April 2020-February 2021. Primary objective was to describe unique maternal and fetal antibody epitope titers and specificity in those patients with COVID-19 history. Serologic profile assessed with a multiplex platform. Antigens used were: HA-trimer Influenza A (Hong Kong H3), spike trimers for SARS-CoV-2, SARS-CoV-1, MERS-CoV, and betacoronaviruses HKU-1 and OC43, as well as the spike N-terminal domain (NTD), spike receptor binding domain (RBD), and nucleocapsid protein (N; full length) for SARS-CoV-2. Results: 112 maternal samples and 101 maternal and cord blood pairs were analyzed. Thirty-seven had a known history of COVID-19 (positive PCR test) in the pregnancy and of those, 17 (47%) were diagnosed with COVID-19 within 30 days of delivery. Fifteen of remaining seventy-six (20%) without a known diagnosis had positive maternal serology. For those with history of COVID-19 we identified robust IgG response in maternal blood to CoV2 nucleocapsid (N), spike (S) full-length and S (RBD) antigens with more modest responses to the S (NTD) antigen. By contrast, the maternal blood IgM response appeared more specific to S (full-length), than N, S (RBD) or S (NTD) epitopes. There were significantly higher maternal and cord blood IgG response not just to CoV2 spike (p < 10-18), but also CoV1 spike (p < 10-9) and MERS spike (p < 10-8). By contrast, maternal IgM responses were more specific to CoV2 (p < 10-19), but to a lesser degree for CoV1 (p < 10-5), and no significant differences for MERS. Maternal and cord-blood IgG were highly correlated for both S and N (R2 = 0.96 and 0.94). Conclusions: Placental transfer is efficient, with robust N and S responses. Both nucleocapsid and spike antibody responses should be studied for a better understanding of COVID-19 immunity. IgG antibodies are cross reactive with related CoV-1 and MERS spike epitopes while IgM, which cannot cross placenta to provide neonatal passive immunity, is more SARS CoV-2 specific. Neonatal cord blood may have significantly different fine-specificity than maternal blood, despite the high efficiency of IgG transfer. to provide neonatal passive immunity, is more SARS CoV-2 specific. Neonatal cord blood may have 79 significantly different fine-specificity than maternal blood, despite the high efficiency of IgG transfer. 80 81 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 December 7, 2021. Pregnant women were found to be particularly vulnerable to respiratory pathogens and are 83 more likely to need intensive care and have higher mortality due to SARS-CoV-2 infection 1 3 . SARS-CoV-2 84 infection is associated with higher rates of pre-term birth 4 , pre-eclampsia 5,6 , and placental 85 pathology 7 . 86 Pregnancy significantly alters many elements of adaptive immunity in a gestational age-specific 87 manner 8 10 . These adaptations have been studied in viral infection, e.g., in influenza A infection 11 . The 88 effect of immunological adaptions in SARS-CoV2 infection during pregnancy has not been delineated. 89 Initial studies on COVID-19 patients highlight several findings: CoV2-specific antibody (ab) responses are 90 measurable 2-3 weeks after onset of symptoms 12, 13 , and almost everyone who recovers from SARS-CoV2 91 infection develops ab responses 14 that show a wide range of neutralization activity 15 . However, many 92 questions remain such as the durability of these responses, the isotype profile at different stages of 93 infection, and determinants of neonatal passive immunity. The presence/timing of neonatal immunity 94 following maternal exposure can be a key consideration in repeat vaccination in pregnancy. As such, 95 research into placental transfer of maternal antibodies and production of fetus-derived abs have aided 96 in developing guidelines for timing pertussis and flu vaccinations for maximal maternal and neonatal 97 benefit 16, 17 . A study of serologic positive pregnant women found that IgG responses were higher in 98 symptomatic vs. asymptomatic women, maternal IgM and IgG peaked 15-30 days post symptom onset, 99 and passive IgG-immunity was found in about three-quarters of neonates and associated with maternal 100 IgG levels 18 . Another study identified a high placental transfer ratio of antibodies, increased with latency 101 between diagnosis and delivery 19 . These studies did not evaluate the range of SARS-CoV-2 antibody 102 epitopes or cross reactivity with related viruses, limiting our understanding of the breadth of SARS CoV-103 2 antibody response as well as our ability to select optimal epitopes to follow in future research studies. 104 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 December 7, 2021. ; https://doi.org/10.1101 https://doi.org/10. /2021 8 The objective of this study was to describe the profile and specificity of maternal serum and 105 neonatal cord blood antibody response to maternal SARS-CoV-2 virus exposure 106 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. This is a prospective cohort of pregnant patients consented for collection of samples at delivery, 112 including maternal blood on admission and cord blood at delivery as part of an ongoing delivery cohort 113 biorepository. Patients were consented either on admission for delivery or as an outpatient prior to 114 planned delivery at our hospital. Per the study protocol additional maternal blood was collected on 115 admission and cord blood collected at delivery. This cohort includes participants from April 2020 Feb 116 2021, however due to COVID related research restrictions, most of the participants were recruited 117 November 2020-Feb 2021. Participants were included if they were consented and had maternal and/or 118 cordblood samples available for this study. Those with prior COVID vaccination were excluded. Subjects 119 were split into two groups-1) COVID-infected which was indicated by either documented SARS CoV-2 120 PCR test at any time in pregnancy or positive maternal spike IgG or IgM at delivery based on assays 121 below or 2) COVID negative: no history of COVID-19 and negative SARS CoV 2 PCR testing on admission 122 Electronic medical records were reviewed for demographic, medical, and obstetric history, date 124 of first positive SARS-CoV-2 PCR test, date of delivery, antenatal complications, delivery outcomes. 125 Latency between COVID-19 diagnosis and delivery was categorized as a binary outcome: within 30 days 126 of delivery or >30 days from delivery. Standard of care at our institution includes universal SARS-CoV-2 127 PCR test on admission and neonatal SARS-CoV-2 PCR test if mother has a positive diagnosis on admission 128 for delivery. 129 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. diabetes, antenatal complication such as chorioamnionitis, preterm delivery, nonreassuring fetal heart 151 rate tracing. Placental findings were categorized as described in Supplemental Table 1 . 152 Outcomes 153 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 December 7, 2021. ; https://doi.org/10.1101/2021.12.06.21267328 doi: medRxiv preprint The primary outcome was to describe maternal antibody epitope profile and specificity related 154 to COVID-19 exposure and correlate with cord blood levels. All antibody responses were analyzed 155 following log transformation of the mean luminescence intensity read out of the MSD assay. 156 Seropositivity for the CoV2 spike protein antigen was defined based on a cutoffs of 8.85 for IgM and 157 8.96 for IgG, in log transformed, as previously described 20 . Additional serological outcomes included: 158 correlation between epitope levels, relation between antibody epitopes and latency to delivery 159 (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 December 7, 2021. ; https://doi.org/10.1101/2021.12.06.21267328 doi: medRxiv preprint During the study period there were 112 maternal samples including 101 maternal and cord 168 blood pairs collected. Thirty-six had a known history of COVID-19 (positive PCR test) in the pregnancy 169 and of those, 17 (47%) were diagnosed with COVID-19 within 30 days of delivery. Fifteen of the 170 remaining seventy-six without a known diagnosis had positive maternal serology (IgG or IgM to SARS-171 CoV-2 spike) positive serology; this was reflected in positive cordblood IgG as well. This represents a 20% 172 seroprevalence rate among study participants. 173 Baseline Characteristics 174 Baseline characteristics are described in Table 1 . There were 51 in the COVID infected group 175 (n=40 with maternal and cordblood paired sample available). Black and Hispanic patients were 176 disproportionately represented in the COVID infected group. Severity of COVID illness/symptoms was 177 documented for 32 of 36 known cases with the large majority (N=30) being asymptomatic or mild, and 2 178 being moderate severity. 179 Maternal SARS-CoV-2 serology 180 Of the 36 with PCR confirmed COVID-19 infection, 92% had positive S-IgM (N=33) and 92% 181 (N=33) had positive S-IgG. We found that subjects with reported COVID-19 diagnosis have high CoV2 S 182 protein-specific IgM and IgG levels in maternal blood, and high IgG but low IgM levels in cord blood, 183 consistent with the dogma that IgM does not cross the placenta (Figure 1 ). Among subjects with no prior 184 COVID-19 diagnosis, 20% (15 of 76) were seropositive for CoV2 spike protein which is consistent with 185 serosurveillance studies 21 . 186 In terms of the magnitude and epitope specificity of the CoV2 antibody response, we found that 187 subjects with prior COVID-19 infection had robust IgG response in maternal blood to CoV2 N, S (full-188 length) and S (RBD) antigens that were approximately 20-fold, 150-fold, and 10-fold higher, respectively, 189 than what was found in subjects without prior infection, with more modest responses to the S (NTD) 190 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 December 7, 2021. ; https://doi.org/10.1101/2021.12.06.21267328 doi: medRxiv preprint 13 antigen ( Figure 1B ). By contrast, the maternal blood IgM response appeared more specific to S (full-191 length) , than N, S (RBD) or S (NTD) epitopes. For example, maternal blood IgM response to S (full-length) 192 was approximately 20-fold higher in COVID-19 subjects than in subjects with no prior history, but only 2-to 3-fold higher for N or S (RBD) ( Figure 1B) . 194 SARS CoV-2 positive and relation cross reactivity to related viruses 195 We evaluated maternal IgG and IgM response to different coronavirus spike proteins to assess 196 the cross-reactivity of these antibody responses across coronaviruses. For IgG responses on maternal 197 blood and cordblood, we found that participants with prior COVID-19 history had significantly higher 198 responses not just to CoV2 spike (p < 10 -18 ), but also CoV1 spike (p < 10 -9 ) and MERS spike (p < 10 -8 ), 199 suggesting a largely cross-reactive IgG response. By contrast, for IgM responses in maternal blood, we 200 found the response more specific to CoV2, with participants with prior COVID-19 history showing 201 significantly higher response to CoV2 (p < 10 -19 ), but to a lesser degree for CoV1 (p < 10 -5 ), and no 202 significant differences for MERS ( Figure 1A) . and that maternal samples are clearly distinguishable from cordblood samples (Figure 2) . 210 We found a high correlation of IgG (R 2 = 0.96 and R 2 = 0.94), but not IgM responses (R 2 = 0.13 211 and R 2 = 0.01) between paired maternal and cord blood samples for CoV2 spike and N antigens, 212 respectively ( Figure 3) . A linear fit between maternal and cordblood IgG responses shows a slope of 1.01 213 for both CoV2 spike and CoV2 N antigens respectively. 214 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 December 7, 2021. Placental histopathology was available for N=52 participants (N=27 COVID negative and N=25 COVID 224 exposed) and described in Table 2 (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. There is highly efficient transfer in maternal to cordblood IgG, with IgG response to nucleocapsid 237 and spike and IgM to spike (full length) showing the highest specificity. IgG antibodies are cross reactive 238 with related CoV-1 and MERS spike epitopes while IgM, which largely does not cross the placenta, is 239 highly SARS Co-V-2 specific. Our results suggest: 1) Both nucleocapsid and full length spike IgG and IgM 240 antibody epitopes should be included in evaluating serologic response in future studies of disease or 241 vaccine development 2) Serological profile functions as a proxy for latency from disease exposure and in 242 the absence of known disease (i.e. lack of access to testing) and 3) Cord blood, which is effectively 243 depleted of the highly specific maternal IgM response, may have significantly different fine-specificity 244 than maternal blood, despite the high efficiency of IgG transfer. 245 This study adds to existing studies through a comprehensive evaluation of maternal and 247 cordblood serological response. One study including 83 mother/baby dyads examined only spike-RBD 248 and IgG and IgM and similarly identified strong correlation between mother and cordblood as well as 249 latency to delivery 19 . Another study of 88 mother/baby dyads identified that IgM peaked at 15 days 250 while IgG peaked at 30 days, which is consistent with our findings regarding maternal and cordblood N-251 IgG titers and latency >30 days from delivery, although that testing platform was semi-quantitative using 252 a combination of S-thus did 253 not provide granular detail on antibody epitopes and specificity 18 . A third study 22 looked at 63 254 mother/baby pairs and examined anti-RBD and anti-N IgG. The limited positive serology reported (65-255 70% of PCR+ mothers) demonstrates the limitations of examining isolated antibody epitopes, and 256 specifically of RBD only epitope. 257 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 December 7, 2021. with other studies in non-pregnant adults, highlights the potential import of N-antibodies in SARS-CoV-2 268 immunity and target for therapy. Finally, similar to the reports cited above 18,19,22 , we found a high degree 269 of correlation and a linear fit with a slope of 1.0, between maternal and cordblood IgG with cordblood 270 IgG concentrations, indicating highly efficient transfer of CoV-2 IgG antibodies. Finally, in examining 271 serology across latency we found that antibody titers rise rapidly after 7 days and persist >100 days is 272 associated with increased antibody responses, specifically against nucleocapsid and spike epitopes 273 demonstrating these antibody responses appear durable rather than short lived. 274 Unique to our study, we also evaluated cross reactivity to related viruses by comparing baseline 275 (COVID-19 negative) antibody response against related coronavirus spike antigen to the antibody 276 response in those with COVID-19. We found that while maternal IgG antibodies were cross reactive with 277 related MERS and CoV-1 spike epitopes, IgM was highly specific for CoV-2 and had similar response to 278 MERS and CoV-1 as those without COVID-19. Given that placental antibody transfer is IgG limited, this 279 suggest that cord blood antibody responses may be more cross-reactive and lack some of the CoV2-280 specific responses found in the IgM of maternal blood. Previous work on adult patients from South 281 Korea indicate that IgM and IgG have a distinct antibody profiles 20 . There is also indication that the 282 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 December 7, 2021. ; https://doi.org/10.1101 https://doi.org/10. /2021 profile of the IgM vs IgG response is indicative of functional activities and subsequent disease severity 26 . 283 Cross-reactivity between the different CoVs has been a subject of debate. As the antibody profiles of 284 IgM and IgG differ and are associated with distinct biological functions, it is likely that the lack of IgM in 285 the cord blood may fail to transfer at least some of the protective immunity from the mother to the 286 child. This is further highlighted by recent work demonstrating the importance of IgM in SARS-CoV-2 287 neutralizing activity in adults 27,28 and another study comparing pediatric and adult responses which 288 found pediatric response was predominantly anti-spike IgG in contrast to adult IgG, IgM, and IgA against 289 both spike and nucleocapsid epitopes, and had reduced neutralizing activity compared to adults 29 . These 290 systematic differences in isotype and epitope specificity raise the possibility that cordblood, which is 291 effectively depleted of the highly specific maternal IgM response, may have significantly different fine-292 specificity and may therefore have different neutralizing activity than maternal blood, despite the high 293 efficiency of IgG transfer. 294 As in prior studies 30 , we have identified an increased rate of preterm in the setting of COVID-19 296 exposure, even in this small cohort. We also identified a positive correlation with higher SARS CoV-2 297 antibody titers and placental maternal vascular malperfusion. This placental finding with COVID-19 has 298 been previously reported 31,32 , and our finding of higher titers related to this finding rather than just 299 COVID exposure, suggest that time from illness, and potentially chronic or downstream effects from 300 COVID-19 lead to placental pathology rather than the acute infectious state (ie within 7 days from 301 infection). 302 This study has a number of strengths. Other evaluations of COVID-19 serology in the 304 mother/baby dyad focused on limited antibody epitopes. We have provided a comprehensive evaluation 305 of maternal and neonatal cord blood serological response looking across the array of SARS CoV-2 306 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 December 7, 2021. ; https://doi.org/10.1101/2021.12.06.21267328 doi: medRxiv preprint antibody epitopes and specificity of response through evaluation of cross reactivity to related viruses. 307 This is a unique examination of serological profile not only in the mother, but also in what is passively 308 acquired by the neonate, and the implications for passive immunity. Although a single institution, our 309 population is diverse, improving external validity. 310 This study has limitations as well. The number of mother/baby dyads limited our ability to provide a 311 more detailed evaluation of serological changes over time. There may be antibody epitope correlations 312 we were not powered to detect. Finally, while we identified differences in maternal and cord blood 313 antibody signature that could impact immunity, we did not evaluate functional activity. 314 Maternal COVID-19 exposure is associated with specific maternal and cord blood antibody 316 signature, with nucleocapsid and full length spike epitopes demonstrating highest specificity in 317 distinguishing exposed vs non-exposed individuals. Serologic profile relates to latency from exposure. 318 There is highly efficient transfer in maternal to cord blood IgG antibodies. IgG antibodies are cross 319 reactive with related CoV-1 and MERS spike epitopes while IgM, which cannot cross placenta to provide 320 neonatal passive immunity, is highly SARS CoV-2 specific, suggesting there may be important qualitative 321 distinctions between maternal immunity and neonatal passive immunity. 322 323 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 December 7, 2021. ; https://doi.org/10.1101 https://doi.org/10. /2021 Acknowledgements 324 We would like to acknowledge our senior research coordinator Brandy Firman as well as the Thomas 325 Jefferson University Hospital Labor and Delivery teams who assisted in sample collection. We would like 326 to thank Ms. Jessica Bolton (WRAIR) for technical assistance in the serological analysis. (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 December 7, 2021. ; https://doi.org/10.1101 https://doi.org/10. /2021 (57) 19 (31) 2 (3) 5 (8) 18 (35) 23 (45) 0 (0) 10 ( (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. and OC43. Participants with prior COVID-19 history had significantly higher responses not just to CoV2 439 spike (p < 10 -18 ), but also CoV1 spike (p < 10 -9 ) and MERS spike (p < 10 -8 ). By contrast, for IgM responses 440 in maternal blood, participants with prior COVID-19 history showing significantly higher response to 441 CoV2 (p < 10 -19 ), but to a lesser degree for CoV1 (p < 10 -5 ), and no significant differences for MERS. B) 442 Fine specificity of SARS-CoV-2 specific IgM (top) and IgG ((bottom) responses in maternal sera and 443 cordblood samples to SARS-CoV2 epitopes ,i.e., nucleoprotein (N), the full-length spike protein (spike-FL) 444 and its functional subdomains, i.e., receptor binding domain (RBD) and N-terminal domain (NTD). 445 Subjects with prior COVID-19 infection had robust IgG response in maternal blood to CoV2 N, S (full-446 length) and S (RBD) antigens that were approximately 20-fold, 150-fold, and 10-fold higher, respectively, 447 than what was found in subjects without prior infection, with more modest responses to the S (NTD) 448 antigen. By contrast, maternal blood IgM response to S (full-length) was approximately 20-fold higher in 449 COVID-19 subjects than in subjects with no prior COVID-19 infection history, but only 2-to 3-fold higher 450 for N or S (RBD). 451 452 Figure 2 : Principal Component Analysis of antibody responses. IgM and IgG responses to CoV2 antigens 453 N, spike, and RBD clearly distinguish between prior COVID-19 cases and maternal from cord blood. 454 455 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 December 7, 2021. ; https://doi.org/10.1101/2021.12.06.21267328 doi: medRxiv preprint (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 December 7, 2021. ; https://doi.org/10. 1101 /2021 Maternal death due to COVID-19 Birth and Infant Outcomes Following Laboratory-343 Confirmed SARS-CoV-2 Infection in Pregnancy SET-NET, 16 Jurisdictions Update: Characteristics of Symptomatic Women of 346 Reproductive Age with Laboratory-Confirmed SARS-CoV-2 Infection by Pregnancy Status 347 United States Clinical manifestations, risk factors, and maternal and 350 perinatal outcomes of coronavirus disease 2019 in pregnancy: Living systematic review and 351 meta-analysis Coronavirus disease 2019 (COVID-19) in pregnant women: A report 353 based on 116 cases An immune clock of human pregnancy Characterizing the pregnancy immune phenotype: Results 362 of the viral immunity and pregnancy (VIP) study Antibody Detection and Dynamic Characteristics in Patients with 373 COVID-19 Antibody responses to SARS-CoV-2 in patients with COVID-19 Convergent antibody responses to SARS-CoV-2 in 377 convalescent individuals Pertussis and influenza immunisation 379 during pregnancy: a landscape review Optimal timing of influenza vaccine during pregnancy: 381 A systematic review and meta-analysis. Influenza Other Respi Viruses Severe acute respiratory syndrome coronavirus 2 serology 384 levels in pregnant women and their neonates SARS-CoV-2 Antibodies and Placental Transfer Ratios. JAMA Pediatr Serological profiles of pan-coronavirus-specific responses 390 in COVID-19 patients using a multiplexed electro-chemiluminescence-based testing platform 391 Short title: Establishing Pan-CoV serological profiles using multiplex assay 4 5. medRxiv SARS-CoV-2 seroprevalence among parturient women 395 in Philadelphia Assessment of Maternal and Neonatal SARS-CoV-2 Viral Load Evolution of antibody immunity to SARS-CoV-2 Kinetics of SARS-CoV-2 specific IgM and IgG responses in COVID-19 402 patients Biochemical characterization of SARS-CoV-2 nucleocapsid protein SARS-CoV-2 infection induces sustained humoral immune responses 406 in convalescent patients following symptomatic COVID-19 Role of Immunoglobulin M and A Antibodies in the Neutralization 409 of Severe Acute Respiratory Syndrome Coronavirus 2 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 December 7, 2021. ; https://doi.org/10.1101/2021.12.06.21267328 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. (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 December 7, 2021. ; https://doi.org/10. 1101 /2021