Rapid review on protective immunity post infection with SARS-CoV-2: update 3

Public Health England (2021)^{Footnote 6}

Prospective cohort

UK

Jun 2020-Jul 2021

Asymptomatic and symptomatic staff (n=25,661) working at hospital sites participating in the SARS-CoV-2 Immunity and Reinfection Evaluation (SIREN) Study were enrolled between June 18 and December 31, 2020 and were followed to estimate the relative incidence of PCR-positive test results according to baseline antibody and/or PCR results. Earlier results from this study are also reported in Hall 2021^{Footnote 4}.

A possible reinfection was defined as a participant with two positive PCR samples 90 or more days apart or an antibody positive participant with a new positive PCR test at least 4 weeks after the first antibody-positive result. A probable case additionally required supportive quantitative serological data or supportive viral genomic data from samples available.

As of July 2021 95% of the cohort was vaccinated and thus the number contributing to the reinfection rate was getting smaller.

• Reinfections increased from 7 cases (April/May) to 44 cases (June/July). This increase in reinfections coincides with the Delta variant becoming the dominant variant in the UK.
• In 2 cases of reinfection, virus neutralizing antibodies levels were below the estimated threshold for predicting immune protection (neutralization level for 50% protection being 54 U/mL, which equates to a titre of 1:10 or 1:30 in most virus neutralization assays at 5 month post infection samples and reinfection occurred at 7 months^{Footnote 21}.

Hall (2021)^{Footnote 4}

Prospective cohort

UK

Jun 2020-Jan 2021

Asymptomatic and symptomatic staff (n=25,661) working at hospital sites participating in the SARS-CoV-2 Immunity and Reinfection Evaluation (SIREN) Study were enrolled between June 18 and December 31, 2020 and followed for 7 months to estimate the relative incidence of PCR-positive test results according to baseline antibody and/or PCR results. A possible reinfection was defined as a participant with two positive PCR samples 90 or more days apart or an antibody positive participant with a new positive PCR test at least 4 weeks after the first antibody-positive result. A probable case additionally required supportive quantitative serological data or supportive viral genomic data from samples available. The effect of the B.1.1.7 variant was included in the analysis by creating a binary variable of when the S-Gene Target Failure (SGTF) PCR accounted for 50% or more of the positive results for each region.

Population:
8,278 (32%) participants had evidence of prior infection and were assigned to the positive cohort, 17,383 (68%) participants that had no evidence of prior infection were assigned to the negative cohort.

13,401 (52.2%) participants of the cohort were vaccinated during the follow-up period (between Dec 2020 and Jan 11, 2021). This included 9,468 in the negative cohort and 3,933 in the positive cohort. Once vaccinated participants no longer contributed to the reinfection data.

• On follow-up, 1,704 new infections in the negative cohort and 155 reinfections were identified. Of these, 19.7% of new infections and 49.7% of reinfections were asymptomatic.
• The incidence density was 7.6 per 100,000 in the positive cohort and 57.3 per 100,000 in the negative cohort.
• Participants in the positive cohort had 99.8% lower risk of new infection than did participants in the negative cohort, adjusted IRR (aIRR) 0.002 (95% CI 0.00–0.01).
• Restricting infections to only those who had COVID-19 symptoms, participants in the positive cohort had a 93% lower incidence of new infection than did participants in the negative cohort, aIRR 0.074 (95% CI 0.06–0.10).
• Using the most sensitive definition for reinfection, which included possible or probable cases, participants in the positive cohort had an 84% lower incidence of new infection than did the participants in the negative cohort, aIRR 0.159 (95%CI 0.13–0.19).
• Those that had an asymptomatic infection had a higher risk of reinfection compared to those with symptomatic infection (aIRR 0.48 95% CI 0.37–0.63).
• The authors did not find any evidence that increased prevalence of the B.1.1.7 variant adversely affected reinfection rates during follow-up. The models suggested that the protective effect of previous infection was similar when Alpha was present (IRR 0.18, 95% CI 0.15–0.23) compared when the original variant was dominant (IRR 0.13 (0.10–0.17), p=0.05).The median interval between the first PCR positive date or date of symptom onset of first infection and the reinfection PCR positive date was 201 days (95-297) or 241 days (90-345), respectively.

Lumley (2021)^{Footnote 5}
Preprint

Prospective cohort

UK

Sep 2020-Feb 2021

Healthcare workers (HCWs) were followed to investigate and compare the protection from SARS-CoV-2 infection conferred by vaccination (results in Table 2) and prior infection (determined using anti-spike antibody status). Individuals were followed from >60 days after their first positive antibody test to either a positive PCR test or first vaccination. To assess the impact of the B.1.1.7 variant on (re)infection risk, they analysed PCR-positive results with and without S-gene target failure (SGTF), and those confirmed as B.1.1.7 on genome sequencing.

Nasal and oropharyngeal swab

• 294/10,513 (2.7%) seronegative HCWs had symptomatic infection during follow-up vs. 1/1273 seropositive HCW (0.08%). Thus, incidence was 98% lower after 60 days in seropositive HCWs (adjusted IRR: 0.02 (95%CI: <0.01-0.18; p<0.001).
• Rates of any PCR-positive result, irrespective of symptoms, were highest in unvaccinated seronegative HCWs (635 cases), with 85% lower incidence in unvaccinated seropositive HCWs (12 cases, aIRR=0.15 (95%CI: 0.08-0.26, p<0.001).
• There was no evidence that SGTF changed the extent of protection against any PCR-positive infection in seropositive HCWs (aIRR vs. non-SGTF, 0.43, (95%CI 0.12-1.52; p=0.19). There was also no evidence that B.1.1.7 changed the extent of protection from any PCR-positive infection in those who were seropositive (aIRR vs non-B.1.1.7=0.40 (95%CI 0.10-1.64; p=0.20).
• Seronegative HCWs had the highest viral loads (average Ct: 18.3) while unvaccinated seropositive HCWs had the lowest viral loads (average Ct: 27.2).

Shinde (2021)^{Footnote 48}

Randomized controlled trial

South Africa

Aug-Nov 2020

Phase 2b trial of a NVX-CoV2373 nanoparticle vaccine. A total of 4387 participants were randomized and dosed at least once, 2199 with NVX-CoV2373 and 2188 with placebo. Serology at baseline was determined and follow-up RT-PCR testing was conducted. Whole virus genome sequencing was conducted on nasal samples.

Nasal swabs

• Of the primary endpoint cases with available whole genome sequencing, 38 (92.7%) of 41 cases were the Beta variant.
• Among placebo recipients, the incidence of symptomatic COVID-19 was similar in baseline seronegative vs. baseline seropositive participants during the first 2 months of follow-up: 5.3% (95%CI 4.3-6.6) vs. 5.2% (95%CI 3.6-7.2). This indicates that prior infection was not protective against reinfection with the B.1.351 variant.

Banerjee (2021)^{Footnote 51}
Preprint
new

Prospective cohort

India

Oct 2020-Jun 2021

Individuals who tested positive for IgG antibodies in a population-based seroprevalence study (n=1,081) in Oct 2020 were followed up by telephone after 8 months to ascertain reinfections. Reinfection was identified by individuals self-reporting a history of fever, cough, and body ache, or after having coming in contact with a positive patient and RT-PCR was done to confirm reinfection.

• Reinfection was self-reported in 13/1081(1.2%) participants.
• All self-reported reinfections were mild with 9 individuals recovering at home and 4 opting for hospital admission.
• Of the 13 self-reported reinfection cases, 3 had reported receiving the COVID-19 vaccine prior to developing symptoms.

Massimo (2021)^{Footnote 22}
Preprint

Prospective cohort

Italy

Apr 2020-May 2021

COVID-19 recovered patients (n=2723) were recruited as potential convalescent plasma donors and tested for SARS-CoV-2 antibodies and followed for possible reinfection. Reinfection was defined as any confirmed positive RT-PCR test >90 days from first episode, regardless of symptoms, with at least one, negative RT-PCR tests on specimens collected between the first and second episode. Subjects were followed for >4 months. The LIAISON® chemiluminescence immunoassay (CLIA) SARS-CoV-2 S1/S2 IgG (DiaSorin) system was used for antibody screening. The CLIA values are expressed as AU (Arbitrary Units) (≥80 AU correlate with neutralizing antibody titer ≥1:160)-

• Two cases of reinfection were reported with 1790 cumulative person-years follow-up. The estimated risk of reinfection was 1.1 x 1000 person-years in this cohort.
• The second infection was diagnosed 201 and 347 days after the first one.
• Both second infections were asymptomatic and detected during contact tracing or routine occupational screening.
• One case was unvaccinated and the other was taken 2 days after the second dose of Comirnaty (BNT162b2) vaccine.
• In the unvaccinated case, after the first infection the anti-SARS-CoV-2 CLIA value was 51 AU, after reinfection it rose to 129 AU. For the second case of reinfection only the NA titer was available after first infection and it was 1:40, the CLIA value after vaccination and reinfection was >400 AU.

Gallais (2021)^{Footnote 10}

Prospective cohort

France

Apr 2020-May 2021

Healthcare workers (n=1,309) with a COVID-19 history, proven either by serology (IgM and IgG against RBD or nucleocapsid proteins) at screening or by a previous RT-PCR (targeting two regions of the RdRp gene), were recruited and followed for up to 13 months. Participants that were seronegative without a history of positive RT-PCR were also recruited to evaluate incidence of infection during the follow-up period. Because the main objective of this study was to study serology over time, assessment of reinfection was based on participant reports during visits, as no RT-PCR surveillance was planned in the study. Therefore, it cannot be excluded that the COVID-19 positive participants had unnoticed asymptomatic reinfection during follow-up (although none had a significant increase of both anti-S and anti-N levels during follow-up).

• This study included 393 convalescent COVID-19 (COVID-19 positive group) and 916 COVID-19 negative HCWs. The COVID-19 positive HCWs included 345 with a history of positive SARS-CoV-2 RT-PCR and 48 with positive serology only.
• Overall, 69 SARS-CoV-2 infections developed in the COVID-19 negative group for an incidence of 12.22 per 100 person-years. In contrast, there was one reinfection (asymptomatic) in the COVID-19 positive group for an incidence of 0.40 per 100 person-years, indicating a relative reduction in the incidence of SARS-CoV-2 reinfection of 96.7% (p<0.0001).
• The one reinfection case experienced mild symptomatic COVID-19 during her first infection with a high viral load (Ct=17) and eventual anti-S and anti-N IgG seroconversion. The second episode occurred 9 months later, was asymptomatic, and was revealed by a low viral load (Ct=34), detected 6 days after exposure. The reinfection was associated with positive anti-S IgM and a rebound of anti-S and anti-N IgG titer 22 days after a second positive RT-PCR.

Cohen (2021)^{Footnote 11}
Preprint

Prospective cohort

South Africa

Jul 2020-Mar 2021

Estimated the burden and transmission of SARS-CoV-2 over the two waves un one rural and one urban community. Mid-turbinate nasal swabs were collected twice-weekly from consenting household members irrespective of symptoms and tested for SARS-CoV-2 by real-time RT-PCR (targeting E, N and RdRp genes). Serum was collected every two months and tested for anti-SARS-CoV-2 antibodies. Defined possible reinfection as >28 to 90 days between rRT-PCR positive specimens (no sequence data available) or between first seropositive specimen and rRT-PCR positive specimen; probable reinfection as >90 days between positive specimens; and confirmed reinfection as distinct Nextstrain clades on sequencing or variant PCR between rRT-PCR positive specimens meeting the temporal criteria for possible or probable.

• Among 71,759 samples from 1,189 participants, 834 (1%) were SARS-CoV-2-positive.
• By PCR detection and serology combined, 34% (406/1189) of individuals experienced ≥1 SARS-CoV-2 infection episode.
• Of 12 reinfections, 6 (50%) were classified as possible and 5 (42%) as probable and 1 (8%) confirmed. Thus 1.5% of the post infection cohort experienced a probably or confirmed reinfection.
• Documented infection on rRT-PCR or serology prior to the start of the second wave was associated with 84% protection against infection in the second wave (relative risk (RR) 0.16, 95% CI 0.07-0.35).

Kohler (2021)^{Footnote 20}
new

Prospective cohort

Switzerland

Jun 2020 -Mar 2021

Across 17 healthcare institutions in Northern and Eastern Switzerland, 4812 HCWs were tested for SARS-CoV-2 antibodies (Jun-Sep 2020) and followed for possible reinfection. Subsequently, participants were tested through nasopharyngeal swabs if they experienced any COVID-19 symptoms such as fever and/or the presence of any respiratory symptom (i.e., shortness of breath, cough, or sore throat). Participants were then asked to fill out a weekly survey to record COVID-19 symptoms and the date/result of any PCR or rapid antigen test. The median follow-up time was 7.9 months [IQR 6.7–8.2].

• At baseline, 144 (3%) participants were seropositive and 4668 (97%) were seronegative.

Nasopharyngeal swabs

• A positive test result was found in 3/67 (4.5%) participants who were seropositive at baseline and 547/2645 (20.7%) who were seronegative. This translates into a RR of 0.22 (95% CI: 0.07-0.66, P = 0.002) for a positive nasopharyngeal swab after positive baseline serology.
• The 3 cases with suspected reinfection were all diagnosed in Jan 2021 after a follow-up (time from baseline serology to 2nd positive test) of 198, 200, and 220 days. 1/3 cases were asymptomatic at the time of reinfection.

Full cohort

• Including those who did not undergo nasopharyngeal testing, the corresponding RR was 0.18 (95% CI: 0.06-0.55, P < 0.001) for protection against reinfection.
• 10/15 symptoms were reported less frequently by participants who were seropositive at baseline. The difference was statistically significant only for impaired olfaction/taste (RR 0.33, 95% CI: 0.15–0.73, P = 0.004), chills (RR 0.59, 95% CI: 0.39–0.90, P = 0.01), and limb/muscle pain (RR 0.68 95% CI 0.49-0.95, P = 0.02) which were all less likely among reinfections.

Maier (2021)^{Footnote 19}
new

Prospective cohort

Nicaragua

Mar 2020-Mar 2021

In this study, 2,338 individuals were followed to assess the incidence of SARS-CoV-2 infection and examine the degree of protection from repeat SARS-CoV-2 infection among seropositive individuals. Blood samples were collected in Mar 2020 or at enrollment, and mid-year samples were collected during Oct-Nov 2020. SARS-CoV-2 infections confirmed by RT-PCR were reported for the entire study period, and seropositive infections were reported for the period between blood samples. To examine the protection from symptomatic reinfection provided by anti–SARS-CoV-2 antibodies, the number of symptomatic RT-PCR–confirmed infections was compared by serostatus.

• Over the one year study period, 129 people tested positive by RT-PCR, an overall incidence of 5.3 infections per 100 person-years (95%CI: 4.4-6.3).
• At mid-year, the overall seroprevalence was 56.7% (95%CI: 53.5%–60.1%).
• Between the mid-year sample and end of Mar 2021, there were 12 symptomatic cases among 863 seronegative individuals (1.4%) and 1 symptomatic case among 1132 seropositive individuals (0.1%). Therefore being seropositive at mid-year was associated with 93.6% protection from symptomatic reinfection (95%CI: 51.1%–99.2%).

Finch (2021)^{Footnote 12}
Preprint

Prospective cohort

US

Apr 2020-Feb 2021

Analysed longitudinal PCR and IgG receptor-binding domain (RBD) serological testing data from a cohort of US SpaceX employees (n=4411) in four states. Reinfection was defined as a new positive PCR test more than 30 days after initial seropositive result. A multivariable logistic regression (Adjusted for race, ethnicity, state, job category and BMI) was conducted to investigate the association between baseline serological status and subsequent PCR test result. This required the authors to choose a cut-off week in order to define baseline seroprevalence and the subsequent observation period for PCR testing. A sensitivity analysis was conducted to identify the optimum cut-off date to define baseline seroprevalence. Follow-up was 6-10 months.

• Of 4411 individuals enrolled, 309 individuals tested seropositive during the study period for an overall adjusted seroprevalence of 8.2% (95% CI: 7.3-9.1).
• Both serology and follow-up PCR testing data were available for 1800 individuals.
• 14 possible reinfections were identified with a median time of 66.5 days between initial seropositive test and PCR positive test.
• Estimated an adjusted odds ratio of 0.09 (95%CI 0.005-0.48) for reinfection, with the week of 26th July 2020 as the optimal baseline time point (this week fell after the first wave). Odds ratio estimates using cut-off weeks in between the two waves of infection (between mid-July and mid-September 2020) ranged from 0.09 (95%CI 0.005-0.48) to 0.25 (95%CI 0.037-1.01).
• Overall, findings suggest that the presence of SARS-CoV-2 antibodies at baseline is associated with around 91% reduced odds of a subsequent PCR positive test, at least over a 6 month time period.

Krutikov (2021)^{Footnote 9}

Prospective cohort

UK

Jun 2020-Feb 2021

Residents (n=682) and staff (n=1429) of 100 long term care facilities were tested for SARS-CoV-2 by RT-PCR monthly and weekly, respectively. Individuals who tested positive were not tested again for 90 days. Blood sampling was offered to all participants at three time points separated by 6-8 week intervals in June (baseline), August and October 2020 to determine antibody titers. All positive PCR tests after October 2020 were considered to indicate infection or reinfection. For reinfection cases, most participants had at least 90 days and all had two or more negative PCR tests between their baseline antibody test and PCR-positive test.

• Baseline antibodies were detected in 226 residents (33%) and 408 staff (29%).
• In residents, a total of 93/456 antibody negative individuals had a PCR positive test at follow-up (0.054 per month at risk) compared to 4/226 antibody positive individuals (0.007 per month at risk).
• In staff, a total of 111/1021 antibody negative individuals had a PCR positive test at follow up (0.042 per month at risk) compared to 10/408 antibody positive individuals (0.009 per month at risk).
• In Cox regression analysis, the relative adjusted hazard ratios for PCR positive infection was 0.15 (95%CI 0.05-0.44, p=0.0006) for seropositive residents vs. seronegative residents and 0.39 (0.19-0.82, p=0.012) for seropositive staff vs. seronegative staff.
• These results suggest that previous infection reduced the risk of reinfection by approximately 85% in residents and 60% in staff members, based on up to 10 months of follow-up.
• Of the 12 reinfection for which symptom data was available, 11 were symptomatic. None were admitted to hospital or died as a result of their infection.
• The median Ct value for reinfection cases was 36 (30.1-37.0).
• There was no difference in quantitative antibody titres against spike or nucleocapsid proteins in reinfected individuals compared with uninfected individuals with baseline antibodies.

Wilkins (2021)^{Footnote 13}

Prospective cohort

US

May 2020-Jan 2021

HCWs were invited to participate in a cohort study of SARS-CoV-2 serology and COVID-19 risk. Participants were invited to undergo serology testing between May 26th and July 10th (baseline) and then between November 9th and January 8th, 2021 (6-month follow-up). Participants who were seropositive at baseline were considered to be at risk for possible reinfection 90 days after their antibody test until end of follow-up or to first positive PCR plus one or more of the following characteristics: in-home exposure to someone infected with SARS-CoV-2, consistent symptoms, or a physician diagnosis of active infection. IRR analyses were adjusted for age, sex, race, and occupation.

• In the 6194 participants who were seronegative at baseline, 519 (8.4%) had a positive PCR after baseline serology testing (rate = 4.25/10,000 person days).
• In the 316 participants who were seropositive at baseline, 20 participants had positive PCR results during follow-up. Among those, 8 (2.5%) met the study criteria for possible reinfection, representing a possible reinfection rate of 1.27/10,000 days at risk (95% CI: 0.55 – 2.51).
• Five of these eight cases of possible reinfection during follow-up were asymptomatic and no cases were severe.
• The unadjusted and adjusted incidence rate ratios were 0.30 (95%CI 0.15 – 0.60) and 0.26 (95%CI: 0.13 – 0.53) for participants who were seropositive at baseline compared to those who were seronegative at baseline, respectively.
• In a sensitivity analyses, in which seronegative participants were not eligible for inclusion in the infection analysis until 90 days or more following their serology result, the possible reinfection rate was 6.7%. The rate of infection per 10,000 days at risk was 3.72 (95%CI 3.39 – 4.08).

Ronchini (2021)^{Footnote 61}
Preprint
new

Prospective cohort

Italy

Apr 2020-Jan 2021

Health-care, support staff, administrative and research personnel (n=1,493) at a cancer center in Milan were tested at baseline for SARS-CoV-2 infection by qPCR using the Allplex SARS-CoV-2 Assay and IgGs using an in-house ELISA assay. Participants were then followed up for up to 6-months to determine possible reinfection. Reinfection was defined as a participant with 2 positive PCR samples with a negative PCR in between and considering a positive PCR after 60 or more days.

• At baseline 266/1,493 (17.8%) infections were identified.
• 8/266 (3%) possible reinfections were reported, of which 7/8 were IgG+ at enrollment.
• 5 participants had reinfection at >60 days.
• When considering only individuals who tested positive for more than one SARS-CoV-2 gene in the qPCR assay, the frequency of reinfections dropped to 2/266 (0.75%).

Rivelli (2021)^{Footnote 23}
Preprint
new

Prospective cohort study

US

Mar 2020-Jan 2021

HCWs (n=2,625) from Illinois and Wisconsin with a COVID-19 history, proven by the presence of SARS-CoV-2 antibodies and a previous RT-PCR, were recruited and followed for up to 10 months. COVID-19 reinfection was defined by current CDC guidelines (subsequent COVID-19 infection ≥ 90 days from prior infection). For those with more than two positive PCR results, the second documented infection that was closest to 90 or more days from the prior infection was included.

• Over the 10-month study period, 156/2,625 (5.94%) experienced reinfection.
• The median days to reinfection were 126.5 (105.5-171.0), with the majority of reinfection occurring between 90-119 days (42.95%).
• Incidence rate of COVID-19 reinfection was 0.35 cases per 1,000 person-days. Participants working in COVID-clinical and clinical units were 3.77 and 3.57 times at greater risk of reinfection compared to those working in non-clinical units, respectively.

Abo-Leyah (2021)^{Footnote 14}

Prospective cohort

Scotland

May-Dec 2020

Health and social care workers (n=2063) were followed in this study. The Siemens SARS-CoV-2 total antibody assay was used to establish seroprevalence in this cohort. New infections post antibody testing were recorded to determine whether the presence of SARS-CoV-2 antibodies protects against reinfection.

• At enrolment, 300 HCWs (14.5%) had a positive antibody test.
• There was one RT-PCR-positive reinfection among the HCWs (1/300, 0.33%). The presence of antibodies was associated with an 85% reduced risk of reinfection with SARS-CoV-2 (hazard ratio 0.15, 95% CI 0.06–0.35; p=0.026), over a follow-up period of up to 6 months.
• This reinfection was in a symptomatic HCW who tested positive by RT-PCR 76 days after having detectable antibodies in their serum.

Dimeglio (2021)^{Footnote 8}
LTE

Prospective cohort

France

Jun-Dec 2020

Healthcare workers (n=8758) were screened for serum SARS-CoV-2 anti-spike antibodies and neutralizing antibody titers after the first wave of epidemic (June/July). Serology was investigated over time and new infections were identified during follow-up in Nov/Dec.

• The median follow-up was 167 days (IQR: 156-172).
• Among the seropositive group, 1.8% (5/276) were positive at follow-up compared to 12.1% (1028/8482) of the seronegative group (p<0.01).
• The five individuals who tested seropositive at baseline and then experienced infection during follow-up included two with low/undetectable neutralizing antibody titers after the first infection, and three with above-median titers.
• The data indicate that previous infection provided protective immunity for at least 167 days.

Lumley (2020)^{Footnote 7}

Prospective cohort

UK

Apr-Nov 2020

Followed asymptomatic and symptomatic staff (n=12,541) at Oxford University Hospitals for up to 31 weeks to estimate the relative incidence of PCR-positive test results and new symptomatic infection according to antibody status.

• Health care workers with positive SARS-CoV-2 anti spike IgG assays at baseline have lowers rates of PCR-positive tests at follow up than workers with negative baselines results (0.13 vs 1.09 per 10,000 days at risk).
• The incidence of positive PCR tests was inversely associated with anti-spike antibody titers, suggesting previous infection resulting in antibodies to SARS-CoV-2 is associated with protection from reinfection for at least 6 months.
• Of the three seropositive health care workers that had subsequent PCR-positive tests for SARS-CoV-2 infection, only one had previously tested positive for SARS-CoV-2, 190 days prior. This case was asymptomatic upon possible reinfection, with negative RT-PCR tests 2 and 4 days later and no subsequent rise in antibody titers.
• Reinfection results could be consistent with a re-exposure to SARS-CoV-2 that did not lead to symptoms but could also plausibly have been a false positive. Caution should be used when interpreting the results of this study.

Letizia (2021)^{Footnote 15}

Prospective cohort

US

May-Nov 2020

This analysis was performed as part of the prospective COVID-19 Health Action Response for Marines study (CHARM) which includes predominately male US Marine recruits, young healthy adults.

Baseline SARS-CoV-2 IgG seropositivity for RBD and spike proteins was assess during a 2 week quarantine period. PCR positivity was also assessed at 0, 1 and 2 weeks of the quarantine period and individuals were excluded at this stage if they had a positive PCR test.

Following quarantine, a closed cohort of 3076 recruits went on to basic training where three PCR tests were done at weeks 2, 4, and 6 in both seropositive and seronegative groups. SARS-CoV-2 was in circulation at the training site despite quarantines.

Time from initial infection, prior to training is not reported. Only IgG titers at enrollment are available as an indication of the potential protection against reinfection for each participant.

• 3076 participants were followed-up during the study period after quarantine for 6 weeks.
• 189 participants were seropositive at baseline, of which 19 (10%) had a subsequent positive PCR test during the 6-week follow-up (1.1 cases per person year). In contrast, 1079 (48%) of 2247 seronegative participants tested positive (6.2 cases per person-year). The incidence rate ratio for SARS-CoV-2 infection in the seropositive group was 0.18 (95% CI 0.11–0.28; p<0·001). Thus, presence of antibodies to SARS-CoV-2 conferred an 82% reduced incidence rate of SARS-CoV-2 infection.
• After adjusting for the effects of race, age, and sex on the SARS-CoV-2 infections, the hazard ratio (HR) comparing seropositive participants and seronegative participants was 0·16 (95%CI 0.10–0.25, p<0.001).
• Among seropositive recruits, infection was more likely in those that had lower baseline IgG titres than in those with higher baseline titres (hazard ratio 0.45, 95%CI 0.32-0.65, p<0.001). Seropositive cases who became infected were also more likely to lack detectable baseline neutralizing antibody activity compared to those that were uninfected (p<0·0001).
• Infected seropositive participants had viral loads that were ~10-times lower than those of infected seronegative participants (p=0.004).
• Symptomatic infection occurred in three (16%) of the baseline seropositive participants versus 347 (32%) of the baseline seronegative participants (p=0.13).

Papasavas (2021)^{Footnote 18}

Prospective cohort

US

May-Nov 2020

Healthcare workers (n=6863) were tested for SARS-CoV-2 antibodies 3 times (baseline, after 2-4 weeks, and after 3-6 months). Abbott Architect i2000 platform was used for the qualitative detection of IgG antibodies to the nucleocapsid protein of SARS-CoV-2.

• At baseline, the prevalence of SARS-CoV-2 antibody among 6863 HCWs was 6.3% (95%CI 5.7–6.9%).
• The incidence of reinfection was 0% in the seropositive group and the incidence of a positive PCR test was 1.2% in the seronegative group after a median follow-up of 5.5 months.

Cohen (2021)^{Footnote 49}

Prospective cohort

US

Jul-Oct 2020

Adults with end-stage kidney disease (ESKD) treated with in-center hemodialysis (ICHD) (n=2337) were assessed for the presence or absence of IgG against SARS-CoV-2 spike and nucleocapsid proteins at baseline and then assessed ~90 days later, and 3 more times monthly, for SARS-CoV-2 infection detected by RT-PCR. Two outcomes were considered. First, any SARS-CoV-2 infection, whether detected during routine clinical surveillance or via a protocolized PCR test at Visits 3, 4, or 5. The second outcome was only those SARS-CoV-2 infections detected during routine clinical surveillance (termed clinically manifest COVID-19), because these represent symptomatic infections.

• 9.5% were anti-SARS-CoV-2 IgG positive at baseline; 3.6% had a known history of COVID-19.
• Over 6679 patient-months of follow-up, 263 participants had evidence of SARS-CoV-2 infection, 141 of which were captured via clinical surveillance (symptomatic).
• Presence of SARS-CoV-2 antibodies at baseline was associated with a 45% lower risk of subsequent SARS-CoV-2 infection (incidence rate ratio 0.55, 95%CI 0.32-0.95) and a 79% lower risk of subsequent symptomatic SARS-CoV-2 infection (IRR 0.21, 95%CI 0.07-0.67).

Iversen (2021)^{Footnote 17}
new

Prospective cohort

Denmark

Apr-Oct 2020

Screening for antibodies against SARS-CoV-2 was offered 3 times during a 6 month period to HCWs in the Capital Region of Denmark. A total of 44,698 HCWs participated with 18,679 (42%) individuals participating in all 3 rounds. After each round, participants filled in an online survey and self-reported information about demographics, exposure to SARS-CoV-2, symptoms and SARS-CoV-2 PCR testing.

• The seroprevalence increased from 4.0% (1501/37,452) in round 1 to 5.8% (1722/29,862) in round 2, and 7.4% (2022/27,457) in round 3 (p<0.001).
• 7/801 (0.87%) of those who were seropositive and 193/25,144 (0.77%) who were seronegative self-reported a positive PCR test between rounds 1 and 2 (RR 1.14, 95%CI: 0.54-2.41, p=0.68).
• Between rounds 1 and 3, 5/760 (0.66%) of those who were seropositive and 389/20,894 (1.86%) who were seronegative self-reported having a positive PCR between the rounds (RR 0.35, 95%CI: 0.15-0.85, p=0.012).
• Between rounds 2 and 3, 3/796 (0.38%) of those who were seropositive and 210/19,280 (1.09%) who were seronegative self-reported a positive PCR test between the rounds (RR 0.35, 95%CI: 0.11-1.08, p=0.051).

Rovida (2021)^{Footnote 16}

Prospective cohort

Italy

Apr-Jun 2020

Healthcare workers (n=3810) were tested for previous SARS-CoV-2 infection according to serostatus determination (SARS-CoV-2 anti-S1 and anti-S2 IgG antibody). Nasopharyngeal swabs were collected and tested for SARS-CoV-2 RNA positivity in subjects with symptoms suggestive for SARS-CoV-2 infection or in case of contact with infected subjects.

• 336 subjects were seropositive and 3474 seronegative at baseline.
• During the second pandemic wave, SARS-CoV-2 infection was detected in 9 seropositive and 225 seronegative subjects. The 3-months cumulative incidence of SARS-CoV-2 infection was 2.68% in seropositive vs 6.48% in seronegative subjects (p=0.006), with a hazard ratio of 0.41 (95%CI 0.26-0.61). The protective effect of the immunity elicited by natural infection was 59% (95% CI 39-74%).
• Data on symptoms were available for 4 reinfection cases: one case developed mild symptoms and no patient required hospitalization.

Lu (2021)^{Footnote 29}

Prospective cohort

US

Mar 2020–Mar 2021

12 months post SARS-CoV-2 infection (n=29) this study investigated the function, phenotypes, and frequency of T-cells using intracellular cytokine staining and spectral flow cytometry. SARS-CoV-2 antibodies were also examined using CYTEK Aurora 5-laser spectral flow cytometer.

T-cells:

• At 12 months there was evidence of polyfunctional and cytotoxic T cells responsive to SARS-CoV-2.
• 75.9% (22/29) had CD4+ and CD8+ T-cells present identified using peptide pools of N, M and S SARS-CoV-2 proteins.
• CD4 was more frequently identified than CD8 T-cells in peripheral blood.
• Frequency of SARS-CoV-2 specific CD4 T cells and antibodies were higher among those that had severe COVID-19, but polyfunctional and cytotoxic T-cell responses were identified across a range of COVID-19 disease severities.

Antibodies:

• Antibodies against the S and N protein were present.
• In mild cases anti-N protein antibodies were undetectable at 12 months.

Wang 2021^{Footnote 38}

Prospective cohort

US

Feb 2020–Mar 2021

A cohort of 63 recovered from PCR confirmed COVID-19 were assessed at 1.3, 6.2 and 12 months post infection. (At 12 months 26 had received at least one dose of mRNA-1273 (Moderna) or BNT162b2 (Pfizer) 2-82 days before follow-up and are excluded from this summary.) 10% were hospitalized during infection, 44% and 14% reported persistent long-COVID symptoms at 6 and 12 months respectively. Serum SARS-CoV-2 RBD specific antibody levels were measured by ELISA, memory B-cells specific to SARS-CoV-2 RBD were measured by FLOW cytometry. Virus neutralization activity in serum samples, against the original variant and VOCs was measured.

Immune responses among the unvaccinated (n=37) at 12 months post infection were as follows:

Antibodies:

• Antibody positivity was maintained for the anti-RBD IgM (103%), IgG (82%), and IgA (72%).
• Anti-N antibody titers decreased significantly between 6-12 months.
• Virus neutralizing activity against the original variant among the unvaccinated remained relatively stable between 6-12 months post infection.
• VOC neutralizing activity against Alpha, Iota, Beta and Gamma was generally lower than against the Wuhan strain; the greatest loss of activity was against Beta.

Memory B-cells:

• The number of circulating RBD-specific memory B-cells remained relatively stable, and was only 1.35 times lower at 12 months post infection compared to 6 months.
• RBD memory B-cell clonality was expanded resulting in monoclonal antibodies with exceptional activity against a range of variants.

Rank (2021)^{Footnote 26}
new

Prospective cohort

Germany

Jul 2021*

Antibodies (Spike1 IgG/IgA), neutralizing antibodies, Interferon gamma (IFN-γ), interleukin-2 (IL-2), SARS-CoV-2-specific CD4+T- cells were measured in 83 convalescent plasma donors at 6 weeks, 6 months, and 12 months. IgG and IgA were analyzed with the Euroimmum ELISA assay. The ELISPOT Interferon-γ (IFNγ kit and IL-2 CoV-iSpot kit measure T-cell reaction against antigens based on secreted cytokines. The activation-induced marker (AIM) assay measured CD4+T-helper cells (THC, CD25hi CD134hi) through the upregulation surface activation induced markers.

Antibodies:

• At 12 months anti-S1 IgA and IgG antibodies were detectable in 78% and 66% of participants, respectively. Median anti-S1 IgA levels were 1.9 (0.5–20.3) and anti-S1 IgG level were 1.7 (0.6–10.6) at 12 months, representing a ~50% within the first 6 months followed by a slow decline 6-12 months.
• 48% (37/77) of participants had neutralizing antibody titers at 12 months however titers decreased from 1:5 (1:1–1:640) at 6 weeks to 1:1 (1:1–1:40) at 12 months. 6-week antibody and Nab levels were highly correlated with 12 month levels.

T-cell:

• At 12 months there were fewer detectable T-cell IFN-γ in 65% (48/76) and IL-2 in 43% (30/70) compared to the 6-month follow up (40/51 (78%) and 24/32 (75%), respectively). Stimulation indices also declined for IFN-γ and IL-2.
• AIM assays found detectable specific T-helper cells (THC) in 80% (56/70) of participants at 12 months, of which 39% (8–64) were central specific T-cells and 58% (20–84) a effector memory T-cells.

Correlative Analysis:

• Older age and a longer duration of the acute phase of COVID-19 were associated with higher humoral and T-cell responses.
• A longer acute phase of COVID-19 (median over 10 days) was associated higher anti-S1 IgG levels (3.3 (0.6–9.5) vs. 1.4 (0.6–10.6)), and significantly more double-positive T-cells (39%) vs. (8%).

Feng (2021)^{Footnote 31}
new

Prospective cohort

China

Jan 2020–Feb 2021

204 convalescent patients admitted to hospital were followed for up to 12 months (280- 360 days, n=50 completing all four sampling points). Plasma S-IgG. RBD-IgA, RBD-IgG were measured through the Kangrun Biotech electrochemiluminescence immunoassay kits in addition to miconeutralization assays and T-cell responses through IFN-y ELISPOT assays.

Antibodies:

• RBD-IgG and S-IgG were stable and highly correlated between 6- 12 months with 12 month levels at 170 AU/ml and 290 AU/ml, respectively.
• At the 12 month mark average RBD-IgA and RBD-IgM were negative.
• Antibody level differences by severity of COVID-19 were not significant at 12 months.

NAbs:

• Microneutralization assays found stable neutralization capabilities from 6-12 months which were correlated with RBD-IgG.

T-cell:

• T-cell responses indicate IFN-γ to N, S1, S2 peptide pools decreased from baseline but remained stable between 6 and 12 months.
• Severely infected patients had more IFN-γ secreting cells towards S1 and S2 peptide pools than those with moderate or mild disease at 12 months.

Zhang (2021)^{Footnote 27}
new

Prospective cohort

China

Jul 2020–Jan 2021

Antibodies (NAb, IgG, and IgM) and T-cell responses were measured in 101 convalescent cases at 6 and 12 months after symptom onset. A total of 74 participants had results at the 12 month mark with 56 having results at both time points.

NAbs and Spike RBD IgG and IgM was measured both through microparticle chemiluminescence and ELISA.

IFN-γ, IL-2, TNFα T-cell responses were measured in PBMC (fresh and cultivated) with ELISpot assays with four peptide pools: S1, S2, M and N.

Antibodies:

• 99% (95%CI 93-100) of cases had detectable NAbs at 12 months (95% at 6 months).
• 95% (95%CI 87-99) of cases were RBD-IgG positive on both assays and this was stable between 6-12 months.
• 26% (95%CI 16-37) of the cases were RBD-IgM positive on both assays, RBD-IgG, RBD-IgM, and NAbs were all positively correlated to each other and with more severe disease.

T-cells:

• At 12 months positive T-cell responses were detected against SARS-CoV-2 (92%, 95%CI 83-97), S1 (78, 95%CI 67-87), S2 (68%, 95%CI 57-79), M (82%, 95%CI 71-90) and N (82% 95%CI 71-90).
• T-cell responses at 12 months were linearly correlated with severity of COVID-19.
• At 12 months, antibody responses were correlated with S protein directed T-cell responses.
• CD4+ and CD8+ and the proportions secreting IFN-γ, IL-2, TNFα were stable between 6-12 months. These groups were mainly composed of effector memory T cells (CD45RA−CCR7−), as well as naïve (CD45RA+CCR7+), central memory (CD45RA−CCR7+), and effector (CD45RA+CCR7−) cells.

15 months post infection

Patterson (2021)^{Footnote 40}
Preprint

Prospective cohort

US

Jul 2021 (est)

The presence of SARS-CoV-2 S1 protein was measured in 46 convalescent individuals. T-cell, B-cell, and monocytic subsets in both severe COVID-19 patients and in patients Post COVID-19 condition were included in the analysis.

Non-classical monocytes were sorted from post COVID-19 condition patients using flow cytometric sorting and the SARS-CoV-2 S1 protein was confirmed by mass spectrometry.

PBMCs were screened for SARS-CoV-2 RNA using quantitative droplet digital PCR (ddPCR).

15 months post infection:

• In patients with post COVID-19 condition the levels of both intermediate (CD14+, CD16+) and non-classical monocyte (CD14Lo, CD16+) were significantly increased compared with healthy controls (P=0.002 and P=0.01, respectively).
• In convalescent cases that had severe COVID-19, neither the intermediate nor non-classical monocytes were elevated.
• A statistically significant number of non-classical monocytes (CD14lo, CD16+) contained SARS-CoV-2 S1 protein in both severe (91%, P=0.004) and post COVID-19 condition cases (73%, P=0.02) out to 15 months post infection.
• 36% (4/11) and 4% (1/26) of convalescent severe COVID-19 cases and post COVID-19 condition cases respectively were ddPCR positive, these were confirmed to only be fragmented SARS-CoV-2 RNA.
• These results show the patients developed an immune response to retained viral antigens (S1 of spike), which continues to be presented by CD16+ monocytes, eliciting an innate immune response characterized by elevated inflammatory markers including interferon , IL-6, IL-10, and IL-2, among others. This may indicate an innate inflammatory dysregulation due to persistent viral protein presentation, further work is needed.

14-16 months post infection

Yang (2021)^{Footnote 33}
Preprint
new

Prospective cohort

China

Jan 2020–May 2021

COVID-19 diagnosed patients were recruited between January and April 2020 (n=214) and followed as long as they were unvaccinated or were not reinfected, up to a maximum of 480 days (16 months ). Viral inhibition was measured against variants (Alpha, Beta, Gamma, Delta and Lambda) in addition to the micro-neutralizations (NAbs) assay, the RBD and N protein IgG antibodies were measured by Sinobio.

• Samples collected between 11 to 16 months had detectable anti-RBD IgG (90.9% at 12, 13 and 16 months and -100.0% at 13 months positive rate), anti-N IgG (90.9% at 13 and 16 months and 100.0% at 12 months), and NAbs (72 % at 13 months and 87.5% at 12, 14-16 months).
• Those with severe disease had higher anti-RBD IgG and NAb titers than those with mild or asymptomatic disease. At 15 months 14.29% of mild and ~50% asymptomatic cases did not have detectable NAbs.
• Pseudotype virus neutralization assays found at 16 months, 8/14 had at least 50% inhibition against all variants: Alpha, Beta, Gamma, Delta, Lambda. Neutralization activities were decreased for Beta, Delta and Lambda.

Dehgani-Mobaraki (2021)^{Footnote 43}
Preprint

Prospective cohort

Italy

Mar 2020–Jun 2021

35 PCR confirmed COVID-19 cases were followed up at 14 months post infection. Anti-Spike-Receptor binding domain IgG CLIA was used for analysis. (Updated analysis to^{Footnote 36Footnote 62} below.)

Antibodies at 14 months:

• 96.8% (31/32) of cases were positive for anti-S-RBD IgG antibodies.
• There was a significant association between anti-S-RBD IgG titers at 14 months and disease severity.
• Cases that developed loss of taste and smell during acute disease had higher titers.
• Cases that were anti-N protein at 10 months, were anti-S-RBD at 12-14 months.

11-13 months post infection

Haveri (2021)^{Footnote 35}
Preprint
new

Prospective cohort

Finland

Oct 2020–May 2021

2586 confirmed COVID-19 patients were identified and invited to provide serum samples 5.9-9.9 months after infection. Among the 652 subjects that were unvaccinated at 1 year 367 were randomly selected for another sample 11.7-14.3 months post infection.

Neutralization was tested by micro-neutralization assays detected NAbs against the original variant and variants of concern (VOC).

SARS-CoV-2 fluorescent multiplex immunoassay (FMIA) has been previously described on the MAGPIX system. Microspheres conjugated with SARS-CoV-2 N and spike full length (SFL) and RBD of the spike protein were used to detect IgG antibodies.

Antibodies:

• S-IgG positivity did not appreciably decrease from 8-13 months with between 96-100% of participants positive at 13 months.
• S-IgG positivity was lowest (96%, 117/122) among men >60 years old that had mild disease.
• NAb positivity was 100% (n=47) among those with severe disease at 13 month.
• NAb positivity for people with mild disease was 84% (99/1187) for men >60 years and 88% (151/171) among women >60 years and 93% (14/15) for men <60 years and 100% for women <60 years.
• N-IgG was 67% for severe infections and 32% among those with mild disease at 13 months.
• N-IgG titers were higher among those >60 years than <60 years with mild disease at 8 and 13 months.
• RBD-IgG titers were higher among those with severe disease than in those with mild disease and among those >60 than those <60 years.

VOCs:

• Compared to the original variant at 13 months, there was a 77%, 69%, and 42% decrease in Beta, Delta, and Alpha Nabs, respectively.

Miyakawa (2021)^{Footnote 28}
Preprint
new

Prospective cohort

Japan

Jan–Mar 2021

358 patients (over 20 years old) who had a positive COVID-19 result (RT-PCR, RT-LAMP, or antigen tests) were recruited to submit serum between 5-8 months and 11.5-14.5 months.

N-IgG and RBD IgG were measured using the Tosoh immunoassay AIA-CL1200 while neutralizing activity was determined through pseudovirus and rapid qualitative neutralizing assays.

Antibodies:

• 96% (344/358) had NAbs at 12 months.
• Significant decreases for N-IgG, RBD-IgG, and NAbs.
• NAbs were correlated with RBD-IgG titers, but not with N-IgG titers.
• Significant differences were seen between those who experiences moderate or severe disease compared to mild disease in N-IgG, RBD-IgG and NAbs.
• Multiple regression analysis of NAb titers at 12 months controlling for age, sex, BMI, and smoking found a significant association between higher disease severity and older age.

VOCs:

• At 12 months significant decreases were seen in neutralizing activities against Beta, Gamma, and Kappa. Neutralizing activity was maintained for Alpha and Delta.
• RBD-IgG titers against Beta and Gamma were lower than for the original variant or Alpha strains.
• Those with severe or moderate disease maintained high levels of NAb positivity (90-100%) at 12 months.
• Nab positivity among those with mild disease varied more with original variant at 12 months (94%, 2% decrease from 6 months), Alpha (79%, 6% decrease), Beta (69%, no change), Gamma (76%, 5% decrease), Delta (75%, 4% decrease), and Kappa (69%, 6% decrease).

Dehgani-Mobaraki (2021)^{Footnote 36}

Prospective cohort

Italy

Mar 2020–May 2021

35 PCR confirmed COVID-19 cases were followed up at 12 and 13 months post infection.

Anti-Spike-Receptor binding domain IgG CLIA was used for analysis. (Updated analysis to 62 below.)

Antibodies at 12-13 months:

• 97% (34/35) cases were positive for anti-SARS-CoV-2 RBD IgG.
• Titers were higher in patients that had severe disease (19 AU/ml) compared to those with mild (6 AU/ml).

Shi (2021)^{Footnote 46}
new

Prospective cohort

China

Jan 2020–Feb 2021

102 COVID-19 recovered inpatients had blood drawn at 7, 14, 30 days post symptom onset (POS), and 1-2, 2-4, 4-7 and 7-13 months POS to measure serum antibodies (IgG, IgM, IgA) against S, N and RBD and nAbs. Antibodies were measured by quantum dot (QD)-labeled lateral flow immunochromatographic assay (LFIA), in vitro microneutralization assay, and immunofluorescence.

NAbs:

• NAbs were detected in 95.2% of samples collected at the 7 to 13 m range and were stable with no significant decrease over long term follow-up.

Antibodies at 7-13 months:

• S2 or N IgM positivity was not maintained, but 7.1% were RBD-IgM positive.
• 2.4% of participants had samples positive RBD-IgA and 9.5% of N-IgA were positive.
• IgG positivity was more robust with 19% positivity for RBD-IgG, 52.4% for N-IgG, and 85.7% for S2-IgG.
• Cumulative S2/N-IgG/IgA and RBD/S2/N-IgG/IgM/IgA positivity were both 88.1%.

Pradenas (2021)^{Footnote 58}
new

Prospective cohort

Spain

Mar 2020–Jun 2021

Patients with asymptomatic to severe COVID-19 from three waves (March- June 2020, July - December 2020, January - June 2021) were recruited to determine longitudinal neutralizing antibody responses with data on vaccination (not reported here) and VOCs. 139 unvaccinated individuals were followed up for a maximum of 458 days from symptom onset in the first wave.

Neutralization Assays, pseudoviruses expressing SARS-CoV-2 S protein and Luciferase were generated against the original variant and variants of concern (VOC).

Neutralizing antibodies:

• Long-lasting neutralizing antibodies were observed in all cases at 12 months and the decay rate slope was flat (533 day half-life). For hospitalized cases the kinetics fit a two-phase exponential decay model with a steep decline in the first phase (half-life 26 days) and a slower decrease in the second phase (half-life 533 days) whereas mild or asymptomatic convalescents had a steady slow decay slope.

VOCs:

• Samples from 60 unvaccinated individuals with follow up beyond 300 days (>10 months) were used in neutralizing activity against variants.
• 33% of participants had low neutralizing titers (neutralizing activity under 250) against the original variant and Alpha compared to 52% against Beta.
• Outpatients more frequently had low neutralizing titers than those were hospitalized.

Xiao (2021)^{Footnote 37}

Prospective cohort

China

Jan 2020–Mar 2021

51 PCR confirmed COVID-19 cases were followed for 12 months after discharge.

IgG and IgM were measured monthly by Antibody Detection Kit (magnetic particle chemiluminescence method) for Novel Coronavirus (2019-nCoV)

Antibodies:

• IgG titers gradually decreased in the first 6 months and then remained stable to 12 months.
• The more severe COVID-19 cases had higher IgG levels.
• At 6 months 8% and at 12 months 11.8% were IgG negative.
• At 6 months 50% and at 12 months 64.7% were IgM negative.

Masiá (2021)^{Footnote 30}

Prospective cohort

Spain

Mar 2020–Apr 2021

From 80 PCR confirmed COVID-19 cases, sequential samples were collected at 1,2, 6 and 12 months post discharge.

S and N IgG protein levels were measured.

Antibodies at 12 months:

• 91.2% (71/80) were positive for S-IgG.
• 43.8% (35/80) were positive for N-IgG
• Logistic regression showed that seroreversion was inversely associated with peak IgG for both S and N.

Renk (2021)^{Footnote 41}
Preprint

Prospective cohort

Germany

May 2020–Jun 2021

A group of 553 children and 726 adults from 328 households with exposure to SARS-CoV-2 were studied. Samples collected at approximately 4 months and 12 months.

Neutralization in a surrogate assay was used to evaluate neutralization potential.

Antibodies at 12 months:

• Between 4-12 months 3.78% of children and 17.11% of adults seroreverted.
• Children’s sera had higher neutralization compared to adult sera (p=0.02).

Chansaenroj (2021)^{Footnote 60}
Preprint

Prospective cohort

Thailand

Mar 2020–May 2021

A longitudinal cohort of 531 PCR confirmed COVID-19 cases was followed for 12 months with sampling points at 3, 6, 9 and 12 months. Only 229 provided multiple time point samples.

Blood samples were tested for SARS-CoV-2 anti-N IgG by chemiluminescent microparticle immunoassay using the commercially available automated ARCHITECT system.

Anti-N protein antibodies:

• 87.5% (328/375) at 3 months were seropositive
• 38.6% (93/241) at 6 months
• 23.7% (49/207) at 9 months
• 26.6% (38/143) at 12 months

Zeng (2021)^{Footnote 47}

Cross-sectional

China

Mar 2021

538 PCR confirmed COVID-19 cases were enrolled during their 1 year post COVID-19 follow-up in March 2021 in Wuhan.

Blood samples were analysed using a CLIA for IgM and IgG antibodies to SARS-CoV-2.

Antibodies at 12 months:

• 12.8% (69/538) were IgM seropositive
• 82.9% (446/538) were IgG seropositive
• No difference by sex
• Analyzing IgG antibody levels by age showed younger cases were associated with lower IgG levels.

Gallais (2021)^{Footnote 10}

Prospective cohort

France

Apr 2020–May 2021

SARS-CoV-2 S protein and N protein antibodies were longitudinally measured in healthcare workers, including COVID-19 negative (n=916) and previously infected (n=393) individuals, using lateral flow assay and CLIA. Infected individuals were sampled at 1, 7-9, 11-13 month intervals.

• Seropositivity rates differed widely depending on antibody isotypes (IgM or IgG), antibody specificity (N or S protein), assays and serum collection time points.

Lateral flow assay results:

• 11-13 months post infection: significant reduction in sample seropositivity was observed ( p<0.0001) for the following antibodies:
  • 51.8% S protein IgM vs. 1 month = 91.3%
  • 56.8% S protein IgG vs. 1 month = 83.7%
  • 20.1% N protein IgG vs. 1 month= 85%

CLIA results:

• Serum S protein IgG antibody seropositivity in the sample remained approximately 97.1%, from one to eleven months post infection.

Petersen (2021)^{Footnote 45}

Prospective cohort

Faroe Islands

Spring 2020, Fall 2020

Serum samples were collected longitudinally from Faroe Island residents with PCR confirmed COVID-19, at various time points between 1-12 months post infection, during the first and second infection waves in the region. Serum RBD specific IgG levels were measured by two ELISA assays.

IgG RBD antibodies:

• 95% of the sample remained seropositive at all sample collection time points, via both assays.
• IgG titers declined over time in both waves (p < 0.001). Pairwise comparison of samples found IgG levels rapidly declined significantly over time until 7 months post infection (p < 0.001) and remained fairly stable from 7-12 months after infection.
• IgG titers followed a U-shaped curve by participant age, with higher antibody levels among the oldest (67+) and the youngest (0– 17) age groups compared to intermediate groups (p < 0.001).

Li (2021)^{Footnote 44}

Prospective cohort

China

Feb 2020–Jan 2021

869 donors for convalescent plasma transfusion were recruited and sampled up to 12 months, all had confirmed COVID-19.

CE-marked coronavirus IgG antibody detection kit was used to test the titer of RBD specific IgG.

RBD IgG positivity rates post diagnosis (titer cutoff <1:80):

• 89.4% at 6-7 months
• 81.4% at 11-12 months
• 5.4% of convalescent plasma donors did not have detectable titers at any point.
• After 9 months, the RBD-IgG titers began to stabilize at a GMT of ~200. The RBD-IgG titer 12 months was 70% lower than month 1 following diagnosis.

Dobaño (2021)^{Footnote 32}

Prospective cohort

Spain

Mar 2020–Apr 2021

Antibody levels and seropositivity was evaluated in a sample of primary health care workers (n=173), 149-270 days after symptom onset; serum samples were collected at 3 time points. A subset of unvaccinated HCWs were also tested at 322-379 days. Infections were confirmed by PCR.

The majority of the sample was mild to moderate cases, and 14% were hospitalized. Levels of S protein and RBD IgM, IgA and IgG antibodies were measured by assay (not specified). Factors associated with higher levels of antibodies were identified by stepwise multivariable regression analyses.

Antibodies at 11.5–12.5 months:

• Seropositivity was 96.9% (95.3% IgG, 82.8% IgA.)

Antibodies at 5-9 months after symptom onset:

• 92.5% (90.2 IgG, 76.3% IgA, 60.7% IgM) of participants were positive for at least one immunoglobulin isotype, indicative of highly stable and persistent immunity.
• Factors associated with higher levels of antibodies at follow-up were hospital admission, fever, anosmia and/or hypogeusia, and previous allergies.

Violán (2021)^{Footnote 34}
Preprint
new

Prospective cohort

Spain

Mar 2020–May 2021

Healthcare professionals (303 healthy, 72 asymptomatic, 367 mild-moderate, and 39 severe-critical) were recruited March 2020 with follow-up to May 15 2021. Repeat serological testing was carried out at 15, 30, 60, 90,180, 270, and 360 days after baseline visit. At the 360 day mark results were available for 109 individuals.

Commercially available antiSARS-CoV-2 IgG and IgM anti-N ELISA kits were used but kit name is not reported. Anti-spike (S) IgG ELISA using DECOV190 allowed for the quantitative determination of IgG class antibodies.

Antibodies at 12 months:

• Only anti-N IgG and IgM were tested and 67% and 43% were positive respectively.
• Anti-N IgG and IgM at month 12 were higher among those who experienced severe-critical disease compared to those with mild-moderate or asymptomatic disease and there was no difference by sex.

Antibodies at 9 months:

• Anti-N IgG and IgM were tested and 67% and 46% were positive respectively. Anti-S IgG was 80%.
• Those who were asymptomatic or had mild to moderate symptoms were below the threshold for positivity for IgM after 9 months.
• Anti-S IgG was positive in most participants and those with more severe disease had higher levels of values.

Peng (2021)^{Footnote 42}
LTE
new

Prospective cohort

China

Aug 2021(est)

85 recovered patients were recruited from the Yongchuan Hospital, China to measure long term humoral immunity with measurements at 1, 3, 8, and 12 months. Anti- S IgM, IgG, and IgA and NAbs were measured using Bioscience Magnetic Chemiluminescence Enzyme Immunoassay Kits.

Antibodies at 12 months:

• 95.5% anti-S IgG and 93.2% NAbs (n=44)
• IgM and IgA was below detection.
• NAbs, IgG, IgM, and IgA decreased over the study period.
• S-IgG and NAb was significantly and positively correlated across the study period.

IgG decreased 82.8% from month 1 to month 12, IgM decreased 96.4% from month 1 to month 12, and IgA decreased 89.4% month 1 to month 12.

Zhan (2021)^{Footnote 29}
new

Prospective cohort

China

Jan 2020–Jan 2021

121 hospitalized COVID-19 patients were recruited to analyze antibody level one year after infection (10 – 12 months). Data on antibodies were also collected from prior clinical trials within 1 month, 1- 2, and 3-months.

Antibody tests included Livzon Diagnostics immunochromatographic assays, InnoDx chemiluminescence microparticle immunoassay, and ELISA measuring RBD-IgG. Pseudovirus neutralization assays were conducted for neutralization activities.

Antibodies:

• IgM positivity was 4% (95% CI, 2–10%).
• IgG positivity was 62% (95% CI, 54–71%). Total RBD IgG was stable over the whole year period.
• Total RBD had a log linear relationship with RBD-IgG and to a lesser degree N-IgG.
• NAb titers were correlated to total RBD antibodies at 12 months.
• It was predicted that 50% protection against the original variant infection at 12 months was 17% (95% CI, 11–24%) and 87% (95% CI, 80–92%) against severe infection for mild and severe cases respectively.
• Generalized linear regression finds a positive association between total anti-RBD antibodies and age, having a severe disease, time from discharge and length of stay, and negatively associated with persistent symptoms and male sex.

VOCs:

• 42 recovery phase serum and 23 convalescent sera with the highest titers were used to determine neutralization against VOCs.
• At 10-12 months decreases in neutralizing activity was seen for Alpha and Beta (10-fold lower) variants compared to the original variant).

est= date study conducted is approximated using publication date, LTE= letter to the editor. AIM = activation induced marker, ELISA = enzyme linked immunoassay, E protein = envelope, HCW = healthcare worker, Ig= Immunoglobulin antibodies, ICS = intracellular cytokine staining, IFN=interferon LTE= letter, LIPS = luciferase immunoprecipitation system assay, M protein = membrane, MN – microneutralization assay, NAb = neutralizing antibodies, N protein = nucleocapsid, ORF = open reading frame, PBMC = peripheral blood mononuclear cells, S = spike protein, RBD= receptor binding domain (d.a.o)=days post symptoms onset THC= specific T-helper cells Tcm = central memory T-cell (Tem) = effector memory T-cell

Chivese (2021)^{Footnote 63}
Preprint

Systematic review

NA

Apr 2021 (est)

A systematic review of 6 databases was conducted with a search date of April 1, 2021.

Risk of bias was conducted.

Random-effects meta-analysis of proportions was conducted.

• 54 studies were included from 18 countries and >12 million observations. Follow-up was up to 8 months.
• 6-8 months post infection 90% of people had SARS-CoV-2 specific immunological memory:
  • IgG – 90.4% (95%CI 72.2-99.9, I²=89.0%, 5 studies)
  • CD4+ - 91.7% (95%CI 78.2 – 97.1, one study),
  • memory B cells 80.6% (95%CI 65.0-90.2, one study)
  • The pooled prevalence of reinfection was 0.2% (95%CI 0.0 – 0.7, I² = 98.8, 9 studies). Individuals who recovered from COVID-19 had an 81% reduction in odds of a reinfection (OR 0.19, 95% CI 0.1 - 0.3, I² = 90.5%, 5 studies).

Chen (2021)^{Footnote 64}

Systematic review

NA

Jul 2021 (est)

A systematic review of 6 databases was conducted with a search date of July 8, 2021. PROSPERO registration no. CRD42021256932.

50% neutralization titers were extracted.

No risk of bias was conducted.

Random-effects meta-analysis of GMTs was conducted.

• Included 106 studies, 65 and 10 were on previously infected with original variant and VOC participants respectively. 15 included vaccinated participants.
• Neutralization was conducted in live virus neutralization assays (n=48 studies), lentivirus-vector pseudovirus neutralization assay (n=39) and VSV-vector pseudovirus neutralization assay (n=24).
• They provide pooled GMT for original variants and VOCs showing 4.2 and 3.3 fold reductions in neutralization of Beta and Delta respectively.
• Vaccine recipient titers are also presented with high heterogeneity across studies and reduced neutralization for Beta and Delta. Potency of immunity depended on the vaccine platform.
• For vaccinated individuals that had previously been infected, neutralization was significantly higher than for uninfected vaccinated individuals.
• Data is not analysed for changes in neutralization titers over time.

Arkhipova-Jenkins (2021)^{Footnote 3}

Living rapid review

NA

Mar 2021 (est)

A rapid review that aims to synthesize evidence on the prevalence, levels, and durability of detectable antibodies after SARS-CoV-2 infection to determine if antibodies to SARS-CoV-2 confer natural immunity. Relevant literature between Jan 1 and Dec 15, 2020 was included in the review. 444 observational studies were included in the review.

• Evidence suggests most adults develop detectable levels of antibodies (i.e. IgM, IgG, and NAb) that peak between 20-30 days post symptom onset.
• The estimated duration was IgM 115, IgG 120, and IgA 140 days.
• Most adults generated neutralizing antibodies (99% NAb, 95% IgG and 80% for IgM), which persisted for several months after infection.
• Age, disease severity, and the presence of symptoms may be associated with higher antibody levels (low level of evidence).
• Some adults did not develop antibodies after SARS-CoV-2 infection, for reasons unclear.
• The summarized evidence was assessed to be low to moderate in quality.

Poland (2020)^{Footnote 1}

Review

NA

Oct 2020 (est)

This review discusses what was known about human humoral and cellular immune responses to SARS-CoV-2 as of the search date Sept 24, 2020.

• The article reviews humoral and cellular immunity and presents some data on kinetics and durability of antibody response and correlation with T-cell response. Many inconsistencies were noted in the initial research.
• Knowledge gaps include high-quality studies on duration of protection by neutralizing antibodies and a good understanding of how the immunological measures being used correlate to protection.

Post (2020)^{Footnote 52}

Systematic review

NA

Jun 2020 (est)

A systematic review on antibody response to SARS-CoV-2 with a search date of June 26, 2020. 150 papers were included. Inclusion criteria included follow-up of greater than 28 days and measured antibody titres.

High variability across includes studies and study designs was reported by the author.

See appendix 2 for a figure on antibody kinetics over time.

• Inconsistency in antibody correlates were seen across the literature.
• IgM (seroconversion 4-14d, peak 2-5 weeks and declining to undetectable levels around 6 weeks) was consistently detected before IgG.
• IgG (seroconversion 12-15 d, peak 3-7 weeks, plateaued until at least 8 weeks with longest follow-up of 12 weeks still detecting antibodies).
• IgA infrequently studied showed seroconversion between 4-11 days, with outliers reporting 24 days.
• Neutralizing antibodies detected 7-15 days after symptom onset, peaking 14-22 days and then declining over 6 weeks. AT 39 days one study had 79% of participants with low neutralizing antibody titres, 3% with high titres. Mild cases had lower neutralizing antibodies.
• Animal studies show promising initial results for protective immunity; however studies were small and short in duration.
• There are studies that have demonstrated correlations with disease severity. An inverse relationship with viral load has been inconsistently reported and no association with re-detection was reported. Studies cannot speak to lasting immunity.

Shrotri (2021)^{Footnote 65}

Systematic review

NA

Jun 2020 (est)

A systematic review that critically evaluates and synthesises published and pre-print literature from Jan 2020-Jun 26 2020 on T-cell mediated immunity post SARS-CoV-2 infection.

61 publications included in the review.

• Symptomatic adult cases consistently show a reduction in peripheral T cell counts in the acute infection phase, which positively correlates with increased disease severity, duration of RNA positivity, and non-survival. The observed relative reductions in CD4+ and CD8+ T cell were variable.
• Asymptomatic and paediatric cases display preserved T-cell counts.
• Severe or critical COVID-19 cases developed more robust, virus-specific T-cell responses. Elevated levels of pro-inflammatory cytokines, interleukin-6 (IL-6), to lesser degree, interleukin-10 (IL-10), and tumour necrosis factor alpha (TNF-α) were identified in severe cases.
• Longitudinal follow-up (14-44 days post infection) suggested recovery of T-cell subset counts alongside clinical recovery and viral clearance.
• T-cell memory and effector function in early convalescents (up to approximately 3 months post onset) was demonstrated against viral antigens S, M and N proteins. T cell response breath and magnitude were generally enhanced among individuals recovering from sever infections. Cytokine producing activity of CD 8+ T cells specific to M and N proteins displayed wider functionality than those targeting S proteins among individuals with mild disease. CD3+ T cells were reduced in severe infections.
• Cross-reactive T-cells among unexposed or individuals previously exposed to other coronaviruses (e.g., pre-pandemic seasonal corona virus strains, SARS-CoV-1) were often identified and appear long-term; in some cases maintained up to 17 years post infection. Cross-reactive T-cells targeting viral S protein and N proteins were the identified cross-reactive immune cells. The impact of cross-reactivity on SARS-CoV-2 infections remains largely unclear, but assumed to be low due to variability in coronavirus epitopes.

Reinfection (n=3)

Shenai (2021)^{Footnote 66}

Systematic Review

NA

Aug 2021 (est)

A systematic review of studies reporting on the rate of infection among recovered and vaccinated individuals. Search was Aug 31, 2021 and included published and preprint papers.

Risk of bias was New Castle Ottawa Scale.

• 9 studies included, 3 RCTs, 4 retrospective cohorts, 1 prospective cohort and 1 case control.
• 4 studies looked at vaccinated vs. infection immunity, RR was 1.86 [95% CI 0.77-4.51, P=0.17] was not significant for superior protection of vaccination over infection immunity.
• Vaccination immunity among individuals that had COVID-19 was significant RR of 1.82 [1.21-2.73, P=.004], compared to those that recovered and did not get vaccinated.
• The analysis does not include data on time since vaccination or infection and waning immunity which may in future provide more clear benefit of cVOID-19 vaccinations to those previously infected.

Kojima (2021)^{Footnote 67}

Systematic Review

NA

Aug 2021 (est)

A systematic review of studies reporting on reinfections. Search was Aug 18, 2021 and included published and preprint papers.

Risk of bias was not conducted.

• 10 studies were included in the review with observation periods of 1-10.3 months.
• The weighted risk reduction against reinfection was 90.7% (SD 7.7%).
• This review cannot establish the longer term protection from reinfection and does not include studies on high risk groups.

Lo Muzio (2021)^{Footnote 24}

Systematic Review

NA

Jul 2021 (est)

A systematic review of studies reporting on reinfections. Search was Jul 31, 2021 and included published and preprint papers.

Risk of bias was not conducted.

A quality of reporting and risk of bias checklist was used.

• 117 articles were included with 260 confirmed cases of reinfection. (only 14 studies met all risk of bias criteria)
• 92/ 260 had severe COVID-19 and 14 died of the reinfection.
• Different clades or lineages were confirmed between infections in 52/260 cases.

(est)=Search date or publication date when search date was not available was used.