SARS-CoV-2 responsive T cell numbers are associated with protection from COVID-19: A prospective cohort study in keyworkers =========================================================================================================================== * David Wyllie * Ranya Mulchandani * Hayley E Jones * Sian Taylor-Phillips * Tim Brooks * Andre Charlett * Anthony E Ades * EDSAB-HOME investigators * Andrew Makin * Isabel Oliver * Philippa Moore * John Boyes * Anil Hormis * Neil Todd * Ian Reckless ## Abstract **Background** Immune correlates of protection from COVID-19 are important, but incompletely understood. **Methods** We conducted a prospective cohort study in 2,826 participants working in hospitals and Fire and Police services in England, UK during the pandemic([ISRCTN5660922](http://medrxiv.org/external-ref?link_type=ISRCTN&access_num=ISRCTN5660922)). Of these, 2,672 were unselected volunteers recruited irrespective of previous SARS-CoV-2 RT-PCR test results, and 154 others were recruited separately specifically because they previously tested positive. At recruitment in June 2020, we measured numbers of interferon-γ secreting, SARS-CoV-2 responsive T cells using T-SPOT® *Discovery* SARS-CoV-2 kits (Oxford Immunotec Ltd), and antibodies to SARS-CoV-2 proteins using commercial immunoassays. We then described time to microbiologically confirmed SARS-CoV-2 infection, stratified by immunological parameters. **Results** T cells responsive to the spike (S), nuclear (N) and membrane proteins (M) dominated the responses measured. Using the sum of the spots (responsive cells within each well of 250,000 peripheral blood mononuclear cells) for S, N and M antigens minus the control, the 2,672 unselected participants were divided into those with higher responses (n=669, 25.4%; median 30 spots (IQR 18,54)) and those with low responses (n=2016, 76.7%, median 3 (IQR 1,6)), the cutoff we derived being 12 spots. Of the participants with higher T cell responses, 367 (53%) had detectable antibodies against the N or S proteins. During a median of 118 days follow-up, 20 participants with lower T cell responses developed COVID-19, compared with none in the population with high T cell responses (log-rank test, *p*=6×10−3). **Conclusions** Peripheral blood SARS-CoV-2 responsive T cell numbers are associated with risk of developing COVID-19. Key Words * SARS-CoV-2 * Immunity * cohort study * Survival analysis * T cell * serology ## Introduction The SARS-CoV-2 virus has caused a global pandemic that has killed over a million individuals, disrupted economies, and continues to spread widely; to date, over 590,000 confirmed cases have occurred in the UK alone. Disease manifestations vary from asymptomatic or minimally symptomatic infection through to fatal pneumonia (1-4). Why some individuals develop severe disease when others are asymptomatically infected remains unclear, with immune protection being one explanation. Animal data indicate that immune protection to SARS-CoV-2 diseases can be elicited, as does limited human epidemiological data (5-11), but such immunity may not prevent infection (12) and high attack rates of clinical illness in closed-community outbreaks suggest such protection is neither absolute nor widespread (5, 13). One potential protective mechanism involves antibody generation. Antibodies to SARS-CoV-2 nucleoprotein and spike protein are generated in >90% of cases of symptomatic infection (Table S1), but in asymptomatically infected individuals generation of SARS-CoV-2 responsive T cells without antibodies has been reported (14, 15). Additionally, SARS-CoV-2 responsive T cells have been described in a proportion of the SARS-CoV-2 naive population, likely primed by infection with the endemic common cold *Coronaviridae* (CCCs) (16-19). It has been proposed, but not proven, that these T cells may provide some protection from SARS-CoV-2 infection (20). To date, T cell immune responses to SARS-CoV-2 have been studied in smaller studies in research settings (e.g. (14-19)) using advanced but bespoke flow cytometric approaches. By contrast, reproducible, standardised high throughput serological assays which have minimal cross reactivity with CCCs have been developed (21, 22), and deployed on a large scale, even before a complete picture of their utility emerges (23). The objective of this study was to describe T cell and antibody responses to the SARS-COV-2 virus in 2,847 UK keyworkers recruited to the EDSAB-HOME (Evaluating Detection of SARS-CoV-2 AntiBodies at HOME) cohort study (25), measuring the association between COVID-19 development and T-cell and serological responses to SARS-CoV-2 at recruitment. To do so, we used T-SPOT® *Discovery* SARS-CoV-2 kits (T-SPOT hereafter), which use ELISpot technology to detect IFN-γ release from immune cells after exposure to SARS-CoV-2 peptides. The test is similar to the widely deployed T-SPOT®.*TB* test, which identifies patients infected with *M. tuberculosis* (24). ## Methods ### Study participants We studied participants in the EDSAB-HOME study (ISRCTN56609224) (25), which recruited and characterised three keyworker “streams”. Two streams (Streams A, B) recruited Fire & Rescue or Police service keyworkers, or Health care keyworkers (n=1,139, n=1,533 respectively), independent of any history of COVID-19 disease, asymptomatic SARS-CoV-2 infection, or RT-PCR test results; in Stream C, healthcare workers were purposefully recruited on the basis of a history of prior RT-PCR positive testing (n=154) (Table S2). The cohorts did not include acutely infected individuals; among the 268 (9.4%) cases who had had a prior positive RT-PCR result, the test had occurred a median of 63 days prior to recruitment. On attending a study clinic, participants provided 6ml blood anticoagulated using EDTA used for immunoassays, and 6ml or 10ml lithium heparin anticoagulated blood used for T-SPOT tests. Recruitment occurred in June 2020. ### Study endpoints In the period after recruitment, large scale SARS-CoV-2 testing capability was in place in England. For participants, symptom driven nasal/throat RT-PCR testing was available through state and employer routes for those with cough, fever, or disordered taste/smell; all such tests, irrespective of result, are recorded in a national database. Asymptomatic testing was not available, except as part of national surveillance schemes. We followed up participants, defining an endpoint as having a SARS-CoV-2 positive RT-PCR test. Follow-up was possible for all participants, with endpoints ascertained by nationaL database searches. Because the incubation period of SARS-CoV-2 is normally less than 14 days, we started follow-up 14 days after the recruitment clinic visit, to avoid inclusion of individuals who may have developed immune responses prior to symptom recognition. Additionally, results obtained from any individuals tested as part of national surveillance studies of randomly selected asymptomatic individuals were excluded, since the positive predictive value of results is much lower in the absence of symptoms(26). More detail is in Supplementary Materials. ### T-SPOT®*Discovery* SARS-CoV-2 kits Peripheral blood mononuclear cells (PBMCs) were isolated from a whole blood sample using the T-Cell *Select*™ reagent (Oxford Immunotec). After quantification and dilution of recovered cells, 250,000 PBMCs were plated into each well of a T-SPOT® *Discovery* SARS-CoV-2 (Oxford Immunotec) kit. The kit is designed to measure responses to six different but overlapping peptides pools to cover protein sequences of six different SARS-CoV-2 antigens, without HLA restriction, and includes negative and positive controls. Peptide sequences that showed high homology to endemic coronaviruses were removed from the sequences, but sequences that may have homology to SARS-CoV-1 were retained. Cells were incubated and interferon-γ secreting T cells detected. ### Laboratory Immunoassays We characterised serological responses against SARS-CoV-2 using two commercial immunoassays: Roche Elecsys® anti-SARS-CoV-2 (21), and EUROIMMUN anti-S IgG immunoassays (25). We considered assays positive if the Roche Elecsys® immunoassay signal was over 1.0, or the EUROIMMUN immunoassay index was over the negative cut-off of 0.8 (25). #### Masking None of the individuals who ran the laboratory immunoassays, either serological or T-SPOT, had access to any information about the samples. Participants were made aware of their EUROIMMUN serological results approximately one month after their clinic visit, with a warning this was not indicative of protection from disease. Participants were not informed of their Roche or T-SPOT results. All analyses used R 4.0.2 for Windows. See also Supplementary Material for more details. #### Ethics EDSAB-HOME study was approved by NHS Research Ethics Committee (Health Research Authority, IRAS 284980) on 02-Jun-2020 and PHE Research Ethics and Governance Group (REGG, NR0198) on 21-May-2020. All participants gave written informed consent. #### Statistics To explore and depict individuals’ multi-dimensional immunological results (T-SPOT counts to six pools of peptides; immunoassay results), we performed hierarchical clustering, displaying results as heatmaps. To depict associations between immunological and clinical metadata (see Table S2), we computed correlation coefficients ρ regarding ρ different from 0 if the associated *p* value was less than 0.01, adjusted for multiple comparisons using Bonferroni’s method, depicting correlations as heatmaps. To compare T-SPOT responses in groups of individuals, we used Wilcoxon Ranked-sign tests. To analyse the T-SPOT result of individuals, we computed the sum of T-SPOT responses to groups of panels, subtracting the background (see Table S3). To divide individuals based on T-SPOT results, we identified a subgroup of cases with proven SARS-CoV-2 infection before study recruitment (PCR positive, Stream C, n=154), and a subgroup at low risk of prior SARS-CoV-2 infection before recruitment (n=1,126; see Supp. Methods Fig. S4, and results). We computed a cutoff from the T-SPOT result optimally separating these two populations as described by Youden’s method (27). This cutoff was derived prior to 29-Sept-2020 when follow-up data was first examined. To determine whether T-SPOT results or serostatus was associated with protection from COVID-19, we compared time-to-event (see ‘Study Endpoints’ above), stratifying by dichotomised immunological measures (seropositive/negative; high vs low SARS-CoV-2 responsive T cell numbers). We tested for differences using log-rank tests. We also estimated hazard ratios, adjusting for serostatus, using Cox Proportional Hazards models with profile penalised likelihood confidence intervals(28) (R *coxphf* package). ## Results ### T cell responses in the general population We performed SARS-CoV-2 serology and T-SPOT testing on 2,867 key workers (Figure 1). In 14/2,867 (0.5%) of cases, T-SPOT tests failed due to raised background. Among the 2,826 (98.5%) participants in whom complete T-SPOT and immunoassay results were available for analysis (Figure 1), we observed T cells responding to a range of SARS-CoV-2 proteins in the vast majority of seropositive individuals, while similar responses were also observed in a proportion of seronegative individuals (Figure 2). ![Figure 1.](http://medrxiv.org/https://www.medrxiv.org/content/medrxiv/early/2020/11/04/2020.11.02.20222778/F1.medium.gif) [Figure 1.](http://medrxiv.org/content/early/2020/11/04/2020.11.02.20222778/F1) Figure 1. Flow chart illustrating participant flows in the EDSAB-HOME project. ![Figure 2.](http://medrxiv.org/https://www.medrxiv.org/content/medrxiv/early/2020/11/04/2020.11.02.20222778/F2.medium.gif) [Figure 2.](http://medrxiv.org/content/early/2020/11/04/2020.11.02.20222778/F2) Figure 2. Hierarchical clustering of responses to S1, S2, Nucleoprotein, Membrane protein, structural proteins and Envelope proteins, as well as anti-Nucleoprotein and S1 serological responses in 2,826 unselected keyworkers from streams A,B and C. Data were log-transformed prior to clustering; units are arbitrary. The previous PCR positivity status of the participants is shown in a guide bar on the left of the main heatmap. There were positive correlations between T-SPOT responses to Spike S1, S2 domains, Membrane and Nucleoproteins (r ∼= 0.6) and between Envelope and structural protein responses and other viral components (r ∼= 0.4) (Supp. Materials, Fig. S1). In view of this, we analysed the sum of T-SPOT responses to Spike, Membrane and Nucleoprotein (T-SPOT SNM results, see Methods) and the sum of T-SPOT Envelope and Structural protein responses (T-SPOT ES results) in further analyses. In Streams A and B, comprising keyworkers recruited without any requirement for previous RT-PCR testing, including 114 individuals with proven past infection (Figure 1), 478/2,826 (16.7%) of the population were seropositive for SARS-CoV-2. Figure 3 depicts T-SPOT SNM results *vs*. anti-S immunoassay results; depiction *vs*. anti-N immunoassay results is similar (Fig. S2). Participants who were seropositive had higher T-SPOT SNM results than individuals who were seronegative (median 28 (IQR: 15,52.5) vs. 3 (1,7) spots per 250,000 PBMC; Wilcoxon test, p < 10−12). We also observed high T-SPOT SNM results associated with seropositivity (median 47.5, (25.5, 72.5)) in a separate cohort (Stream C) comprising only cases of previously confirmed infection (Figure S3). This suggests T-SPOT SNM responses are generated, along with specific antibody, in SARS-CoV-2 infection. In contrast with T-SPOT SNM results, T-SPOT ES results were in general much lower and differed little between seronegative (median (IQR) 1, (0,3)) and seropositive individuals (median (IQR) 1, (0,3)), suggesting durable responses against these components of the T-SPOT kit were not induced by natural infection. ![Figure 3.](http://medrxiv.org/https://www.medrxiv.org/content/medrxiv/early/2020/11/04/2020.11.02.20222778/F3.medium.gif) [Figure 3.](http://medrxiv.org/content/early/2020/11/04/2020.11.02.20222778/F3) Figure 3. SNM responsive T cells responses and their relationship to anti-S1 IgG serological responses in 2,826 individuals (A) distribution of anti-Spike S1 IgG antibody responses (EUROIMMUN), and its relationship to symptoms. The vertical line is at 0.8, a manufacturer specified cutoff. (B) bivariate plots of anti-S1 IgG responses (EUROIMMUN) and the sum of Spike, Nucleoprotein and Membrane protein responsive T cell numbers. The horizontal line corresponds to 12 spots / 250,000 cells. (C) distribution of the sum of Spike, Nucleoprotein and Membrrane protein responsive T cell numbers. (D) As in B, but for Envelope and structural protein responsive T cell numbers; (E) Distribution of Envelope and structural protein responsive T cell numbers. In marginal histograms (A,C,E), the number of individuals reporting symptoms is depicted in red. ### T cell responses differ in individuals with and without prior SARS-CoV-2 infection We examined T-SPOT SNM results in individuals with prior SARS-CoV-2 infection and in those without. Because the T-SPOT test uses fresh cells, we did not have the option of analysing pre-pandemic material. We identified a cohort of 1,126 individuals from Streams A, B whom we considered to have a low risk of previous SARS-CoV-2 infection, since they reported no personal or household symptoms compatible with COVID-19 since January 2020, and were seronegative (Figure S4). We compared this low risk cohort with 154 individuals known to have been infected, based on previous PCR positivity (Stream C members) (Fig. S4, S5). T-SPOT SNM results differentiate these populations with area under curves of 0.96, and an optimal cutpoint based on Youden index(27) of 12 cells per 250,000 cells (Figure S6A). By contrast, T-SPOT ES results do not differentiate these populations (Figure S6B). We refer to the participants with T-SPOT SNM results above vs. below this exploratory cutoff as having ‘Higher T-SPOT SNM’ vs. ‘Lower T-SPOT SNM’ results going forward. ### COVID-19 symptom history is correlated with T-SPOT results only in seropositive individuals We observed seronegative individuals with higher T-SPOT SNM results: in Streams A and B, 283/2197 (12.9%) participants fell into this category. Among the 2,672 Stream A&B keyworkers, we noted T-SPOT SNM results and serological results were significantly associated with well-characterised COVID-19 symptoms, including fever, muscle aching, fatigue, and with abnormal sense of taste and smell in both the individual and in their household (Figure 4A). Importantly, this effect is driven entirely by seropositive individuals; restricting analysis to the 2,197 seronegative individuals, we observed no significant associations, either positive or negative, between T-SPOT SNM results and self-reported symptoms (Figure 4B). ![Figure 4.](http://medrxiv.org/https://www.medrxiv.org/content/medrxiv/early/2020/11/04/2020.11.02.20222778/F4.medium.gif) [Figure 4.](http://medrxiv.org/content/early/2020/11/04/2020.11.02.20222778/F4) Figure 4. Correlation matrices showing the relationship between individual or household self reported symptoms, T cell numbers and immunoassay results in (A) 2,672 individuals from Streams A,B (B) 2,197 seronegative individuals. Clustering was performed independently in the two datasets. The order of the columns minimises differences between individuals. Only statistically significant correlations are shown. Colour scale reflects Spearman’s correlation coefficient, ρ In (A), the box denotes a correlation between various immune parameters and symptoms. Comparable correlations are not seen if symptomatic individuals are excluded (B). Serological data is not included in (B) as in (B) all individuals are seronegative. From an alternative analysis, we reached similar conclusions. Analysing dichotomised T-SPOT SNM results, we observed higher T-SPOT SNM results in 665/2672 (24.8%) of keyworkers, including 377/475 (79.4%) of seropositive individuals and 283/2197 of (12.9%) seronegative individuals. We identified 24 clinical features significantly associated with seropositivity in univariate analyses (Table S6). However, in the 2,197 person seronegative subset, none of the risk factors for seropositivity were positively correlated with higher T-SPOT SNM results (Table S6). Not only were COVID-19 compatible symptoms not significantly associated with higher T-SPOT SNM results in seronegative participants, neither were other SARS-CoV-2 exposure-associated risk factors including health care worker occupation and household COVID-19 disease (Table S6). There was one shared risk factor, age, but the direction of the effect differed. While older seropositive individuals have higher T-SPOT SNM results than younger subjects (p < 10−4, Wilcoxon test), the opposite is true in seronegative populations: in under 30s who were seronegative, 21.7% had higher T-SPOT SNM results; in 60+ adults, this had reduced to just 5.9% (association with age, p <10−4) (Fig. S7). ### Higher T-SPOT SNM results are associated with decreased COVID-19 risk We followed up all participants until 18-10-2020, a median of 118 days. 20 (0.71%) participants developed the endpoint (SARS-COV-2 test positive COVID-19 disease), all of whom were seronegative and had lower T-SPOT SNM results (Figure 5, Fig. S8). COVID-19 development is more common in individuals with lower T-SPOT SNM results (hazard ratio (HR) 30.0 (2.53, 178000), p=0.001; log-rank p=0.007). This effect persisted after adjustment for serostatus (adjusted HR 13.5 (1.05,83000), p=0.04). Restricting to seronegative individuals, a similar estimate was obtained (HR 12.2 (1.03, 72300), p=0.046); log-rank p=0.08. The association of low T cell SARS-CoV-2 numbers with having positive RT-PCR tests is specific to positive results; similar associations were not observed when having a test, as opposed to testing positive, was considered (Fig. S9). ![Figure 5.](http://medrxiv.org/https://www.medrxiv.org/content/medrxiv/early/2020/11/04/2020.11.02.20222778/F5.medium.gif) [Figure 5.](http://medrxiv.org/content/early/2020/11/04/2020.11.02.20222778/F5) Figure 5. Kaplan-Meier survival curves in all subjects (A), or seronegative subjects (B), showing time to testing positive for SARS-CoV-2. ## Discussion About 25% (669/2,672) unselected UK keyworker volunteers had elevated levels of SARS-CoV-2 S, N, and M protein responsive T cells, when using a cutoff of 12 spots per 250,000 PBMC. Of these 25%, only 55% (367/669) were seropositive using either one of two sensitive SARS-CoV-2 specific immunoassays. The other 45% lack both risk factors for SARS-CoV-2 infection, and detectable antibodies against the N or S antigens; they most likely harbour SARS-CoV-2 responsive T cells primed by CCCs, rather than SARS-CoV-2 (16-19). The absence of epidemiological risk factors for SARS-CoV-2 in this population argues against, but does not completely exclude, an alternative scenario involving T cell priming by minimally symptomatic SARS-CoV-2 infection without antibody generation/ persistence (14, 15). While our findings are compatible with flow cytometric studies identifying interferon-γ secreting SARS-CoV-2 specific T cells (16-19), the study’s much larger scale allows us to demonstrate that high levels of SARS-CoV-2 responsive T cells are associated with protection from symptomatic SARS-CoV-2 infection, at least over the time period studied. A potential bias involving seropositive participants being tested less (perhaps because they believed they were immune) does not explain this effect. We did not observe COVID-19 in seropositive individuals, but note a protective association of having higher T-SPOT SNM results in seronegative populations, questioning whether seropositivity might appear protective in part because seropositivity is commonly associated with high SARS-CoV-2 responsive T cell numbers. There are several limitations. First, numbers of individuals developing illness during follow-up remain small at present, with about 0.7% of those followed up developing illness. While disease acquisition is associated with T-SPOT SNM results (p=0.007), we cannot precisely quantify the T cell number / protection relationship. The 12 spot cutoff used here was selected since it discriminates individuals at low risk of SARS-CoV-2 from those with proven past COVID-19 disease, but it is not necessarily optimally predictive of disease risk going forward. We plan on publishing updated analyses as case numbers rise which may help address this, while also permitting co-modelling of disease risk using clinical COVID-19 risk factors (e.g. age, ethnicity (1-4)). Secondly, we counted circulating SARS-CoV-2 responsive interferon-γ secreting cells, but did not phenotype them (16), and so may have missed prognostic immunophenotypes and mucosally restricted T-cell populations. Finally, we only measured symptomatic COVID-19 infection; we did not investigate asymptomatic infection (which may be important for transmission, and over which immune control is at present uncertain (20)). We considered potential confounders of the association we observed. Participants knew their symptom history and antibody status; seropositive or previously symptomatic participants may have been more likely to allow themselves to be exposed to SARS-CoV-2. Secondly, some individuals may have been infected without being tested (and acquired protection) subsequent to the clinic visits. Thirdly, individuals at the highest risk of occupational exposure on follow-up may have been at highest risk of having been infected prior to recruitment. Importantly, all these biases would be expected to dilute any immunology/protection association. Overall, this study suggests that serology may underestimate the working age population at lower risk of clinical SARS-CoV-2 infection, something which might impact outbreak kinetics (29) and which has been suspected on epidemiological grounds (30). We would speculate that the declining number of individuals with high levels of SARS-CoV-2 responsive T cells with increasing age may explain higher illness incidence and severity in older age (1-4). Intriguingly, our data indicate individual level risk stratification may be possible using T-cell assays, including the standardised assay kits used in this study which, being in the same format as the widely used T-SPOT®.*TB* tests for latent TB infection, would be readily deployable at scale (24). (2,700 words) ## Supporting information Supplementary Materials [[supplements/222778_file03.pdf]](pending:yes) ## Data Availability Data available on reasonable request to corresponding author. ## Funding statement The study was commissioned by the UK Government’s Department of Health and Social Care. It was funded and implemented by Public Health England, supported by the NIHR Clinical Research Network (CRN) Portfolio. Oxford Immunotec Ltd did T-SPOT tests at their own cost as part of the study. Department of Health and Social had no role in the study design, data collection, analysis, interpretation of results, writing of the manuscript, or the decision to publish. DW acknowledges support from the NIHR Health Protection Research Unit in Genomics and Data Enabling at the University of Warwick. HEJ, AEA, MH and IO acknowledge support from the NIHR Health Protection Research Unit in Behavioural Science and Evaluation at University of Bristol. STP is supported by an NIHR Career Development Fellowship (CDF-2016-09-018). The views expressed are those of the author(s) and not necessarily those of the NHS, NIHR or the Department of Health and Social Care. ## Conflict of Interest Oxford Immunotec Ltd have a filed a patent relevant to the T-SPOT technology described in this work and its applications. DW is named as an inventor in the patent. AM is an employee of Oxford Immunotec Ltd. Other authors declare no conflicts of interest. ## Footnotes * David.wyllie{at}phe.gov.uk * Ranya.mulchandani{at}phe.gov.uk * Hayley.jones{at}bristol.ac.uk * s.taylor-phillips{at}warwick.ac.uk * Tim.brooks{at}phe.gov.uk * Andre.charlett{at}phe.gov.uk * t.ades{at}bristol.ac.uk * amakin{at}oxfordimmunotec.com * isabel.oliver{at}phe.gov.uk * philippa.moore1{at}nhs.net * johnboyes{at}nhs.net * Anilhormis{at}nhs.net * Neil.todd{at}york.nhs.uk * Ian.reckless{at}mkuh.nhs.uk * Received November 2, 2020. * Revision received November 2, 2020. * Accepted November 4, 2020. * © 2020, Posted by Cold Spring Harbor Laboratory The copyright holder for this pre-print is the author. All rights reserved. The material may not be redistributed, re-used or adapted without the author's permission. ## References 1. 1.Guan WJ, Ni ZY, Hu Y, Liang WH, Ou CQ, He JX, et al. Clinical Characteristics of Coronavirus Disease 2019 in China. The New England journal of medicine. 2020;382(18):1708–20. [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1056/NEJMoa2002032&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=http://www.n&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2020%2F11%2F04%2F2020.11.02.20222778.atom) 2. 2.Wölfel R, Corman VM, Guggemos W, Seilmaier M, Zange S, Müller MA, et al. Virological assessment of hospitalized patients with COVID-2019. Nature. 2020;581(7809):465–9. [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1038/s41586-020-2196-x&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=http://www.n&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2020%2F11%2F04%2F2020.11.02.20222778.atom) 3. 3.Wu Z, McGoogan JM. Characteristics of and Important Lessons From the Coronavirus Disease 2019 (COVID-19) Outbreak in China: Summary of a Report of 72lJ314 Cases From the Chinese Center for Disease Control and Prevention. Jama. 2020;323(13):1239–42. [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1001/jama.2020.2648&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=http://www.n&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2020%2F11%2F04%2F2020.11.02.20222778.atom) 4. 4.Yang R, Gui X, Xiong Y. Comparison of Clinical Characteristics of Patients with Asymptomatic vs Symptomatic Coronavirus Disease 2019 in Wuhan, China. JAMA network open. 2020;3(5):e2010182. 5. 5.Addetia A, Crawford KH, Dingens A, Zhu H, Roychoudhury P, Huang M, et al. Neutralizing antibodies correlate with protection from SARS-CoV-2 in humans during a fishery vessel outbreak with high attack rate. 2020:2020.08.13.20173161. 6. 6.Houlihan CF, Vora N, Byrne T, Lewer D, Kelly G, Heaney J, et al. Pandemic peak SARS-CoV-2 infection and seroconversion rates in London frontline health-care workers. Lancet (London, England). 2020;396(10246):e6–e7. 7. 7.Hassan AO, Case JB, Winkler ES, Thackray LB, Kafai NM, Bailey AL, et al. A SARS- CoV-2 Infection Model in Mice Demonstrates Protection by Neutralizing Antibodies. Cell. 2020;182(3):744-53.e4. 8. 8.Rogers TF, Zhao F, Huang D, Beutler N, Burns A, He WT, et al. Isolation of potent SARS-CoV-2 neutralizing antibodies and protection from disease in a small animal model. Science (New York, NY). 2020;369(6506):956–63. 9. 9.Mercado NB, Zahn R, Wegmann F, Loos C, Chandrashekar A, Yu J, et al. Single- shot Ad26 vaccine protects against SARS-CoV-2 in rhesus macaques. Nature. 2020. 10. 10.Zost SJ, Gilchuk P, Case JB, Binshtein E, Chen RE, Nkolola JP, et al. Potently neutralizing and protective human antibodies against SARS-CoV-2. Nature. 2020;584(7821):443–9. 11. 11.Chandrashekar A, Liu J, Martinot AJ, McMahan K, Mercado NB, Peter L, et al. SARS-CoV-2 infection protects against rechallenge in rhesus macaques. Science (New York, NY). 2020;369(6505):812–7. 12. 12.sermet i, temmam s, huon c, behillil s, gadjos v, bigot t, et al. Prior infection by seasonal coronaviruses does not prevent SARS-CoV-2 infection and associated Multisystem Inflammatory Syndrome in children. 2020:2020.06.29.20142596. 13. 13.Fontanet A, Cauchemez S. COVID-19 herd immunity: where are we? Nature reviews Immunology. 2020;20(10):583–4. 14. 14.Gallais F, Velay A, Wendling M-J, Nazon C, Partisani M, Sibilia J, et al. Intrafamilial Exposure to SARS-CoV-2 Induces Cellular Immune Response without Seroconversion. 2020:2020.06.21.20132449. 15. 15.Sekine T, Perez-Potti A, Rivera-Ballesteros O, Strålin K, Gorin JB, Olsson A, et al. Robust T Cell Immunity in Convalescent Individuals with Asymptomatic or Mild COVID-19. Cell. 2020;183(1):158-68.e14. 16. 16.Grifoni A, Weiskopf D, Ramirez SI, Mateus J, Dan JM, Moderbacher CR, et al. Targets of T Cell Responses to SARS-CoV-2 Coronavirus in Humans with COVID-19 Disease and Unexposed Individuals. Cell. 2020;181(7):1489-501.e15. 17. 17.Braun J, Loyal L, Frentsch M, Wendisch D, Georg P, Kurth F, et al. SARS-CoV-2- reactive T cells in healthy donors and patients with COVID-19. Nature. 2020. 18. 18.Le Bert N, Tan AT, Kunasegaran K, Tham CYL, Hafezi M, Chia A, et al. SARS-CoV- 2-specific T cell immunity in cases of COVID-19 and SARS, and uninfected controls. Nature. 2020;584(7821):457–62. 19. 19.Mateus J, Grifoni A, Tarke A, Sidney J, Ramirez SI, Dan JM, et al. Selective and cross-reactive SARS-CoV-2 T cell epitopes in unexposed humans. Science (New York, NY). 2020;370(6512):89–94. 20. 20.Lipsitch M, Grad YH, Sette A, Crotty S. Cross-reactive memory T cells and herd immunity to SARS-CoV-2. Nature reviews Immunology. 2020:1–5. 21. 21.Evaluation of sensitivity and specificity of four commercially available SARS-CoV-2 antibody immunassays. Public Health England; 2020. 22. 22.Okba NMA, Müller MA, Li W, Wang C, GeurtsvanKessel CH, Corman VM, et al. Severe Acute Respiratory Syndrome Coronavirus 2-Specific Antibody Responses in Coronavirus Disease Patients. Emerging infectious diseases. 2020;26(7):1478–88. [PubMed](http://medrxiv.org/lookup/external-ref?access_num=http://www.n&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2020%2F11%2F04%2F2020.11.02.20222778.atom) 23. 23.Andersson M, Low N, French N, Greenhalgh T, Jeffery K, Brent A, et al. Rapid roll out of SARS-CoV-2 antibody testing-a concern. BMJ (Clinical research ed). 2020;369:m2420. [FREE Full Text](http://medrxiv.org/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiRlVMTCI7czoxMToiam91cm5hbENvZGUiO3M6MzoiYm1qIjtzOjU6InJlc2lkIjtzOjE4OiIzNjkvanVuMjNfMTEvbTI0MjAiO3M6NDoiYXRvbSI7czo1MDoiL21lZHJ4aXYvZWFybHkvMjAyMC8xMS8wNC8yMDIwLjExLjAyLjIwMjIyNzc4LmF0b20iO31zOjg6ImZyYWdtZW50IjtzOjA6IiI7fQ==) 24. 24.Abubakar I, Lalvani A, Southern J, Sitch A, Jackson C, Onyimadu O, et al. Two interferon gamma release assays for predicting active tuberculosis: the UK PREDICT TB prognostic test study. Health technology assessment (Winchester, England). 2018;22(56):1–96. [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.3310/hta22440&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=29298736&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2020%2F11%2F04%2F2020.11.02.20222778.atom) 25. 25.Mulchandani R, Taylor-Phillips S, Jones H, Ades T, Borrow R, Linley E, et al. Self assessment overestimates historical COVID-19 disease relative to sensitive serological assays: cross sectional study in UK key workers. 2020:2020.08.19.20178186. 26. 26.Watson J, Whiting PF, Brush JE. Interpreting a covid-19 test result. 2020;369:m1808. 27. 27.Youden WJ. Index for rating diagnostic tests. Cancer. 1950;3(1):32–5. [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1002/1097-0142(1950)3:1<32::AID-CNCR2820030106>3.0.CO;2-3&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=15405679&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2020%2F11%2F04%2F2020.11.02.20222778.atom) [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=A1950UD97200004&link_type=ISI) 28. 28.Heinze G, Dunkler D. Avoiding infinite estimates of time-dependent effects in small- sample survival studies. Statistics in medicine. 2008;27(30):6455–69. [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1002/sim.3418&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=18816502&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2020%2F11%2F04%2F2020.11.02.20222778.atom) 29. 29.Gomes MGM, Corder RM, King JG, Langwig KE, Souto-Maior C, Carneiro J, et al. Individual variation in susceptibility or exposure to SARS-CoV-2 lowers the herd immunity threshold. 2020:2020.04.27.20081893. 30. 30.Lourenco J, Pinotti F, Thompson C, Gupta S. The impact of host resistance on cumulative mortality and the threshold of herd immunity for SARS-CoV-2. 2020:2020.07.15.20154294.