key: cord-0828038-d5uty5mb
authors: Varona, Jose Felipe; Madurga, Rodrigo; Peñalver, Francisco; Abarca, Elena; Almirall, Cristina; Cruz, Marta; Ramos, Enrique; Castellano, Jose María
title: kinetics of anti-SARS-CoV-2 antibodies over time. Results of 10 month follow up in over 300 seropositive Health Care Workers
date: 2021-05-25
journal: Eur J Intern Med
DOI: 10.1016/j.ejim.2021.05.028
sha: bee9e35d55b76718ce7a2c519bb12d1f83be38f3
doc_id: 828038
cord_uid: d5uty5mb

Background: The kinetics of the antibodies against severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) needs to be evaluated since long-term duration of antibody remains largely unknown, particularly in infected healthcare workers (HCW). Methods: Prospective study, evaluating the longitudinal profile of anti-SARS-CoV-2 antibody titers in a random sample of 331 seropositive healthcare workers (HCW) of Spanish Hospitals Group. Serial measurements of serum IgG-anti-SARS-CoV-2 were obtained at baseline (April-May,2020), and in 2 follow-up visits. Linear mixed models were used to investigate antibody kinetics and associated factors. Results: A total of 306 seropositive subjects (median age: 44.7years;69.9 % female) were included in the final analysis. After a median follow-up of 274 days between baseline and final measurement, 235(76.8%) maintained seropositivity. Antibody titers decreased in 82.0%, while remained stable in 13.1%. Factors associated with stability of antibodies over time included age≥45 years, higher baseline titers, severe/moderate infection and high-grade exposure to COVID-19 patients. In declining profile, estimated mean antibody half-life was 146.3 days(95%CI:138.6-154.9) from baseline. Multivariate models show independent longer durability of antibodies in HCW with high-risk exposure to COVID-19 patients (+14.1 days;95%CI:0.6-40.2) and with symptomatic COVID-19 (+14.1 days;95%CI:0.9-43.0). The estimated mean time to loss antibodies was 375(95% CI:342-408) days from baseline. Conclusions: We present the first study measuring the kinetics of antibody response against SARS-CoV-2 in HCW beyond 6 months. Most participants remained seropositive after 9 months but presented a significant decline in antibody-titers. Two distinct antibody dynamic profiles were observed (declining vs. stable). Independent factors associated with longer durability of antibodies were symptomatic infection and higher exposure to COVID-19 patients.

The coronavirus disease 2019 (COVID-19) pandemic, caused by severe acute respiratory syndrome 12 coronavirus 2 (SARS-CoV-2), has led to an unprecedented health emergency, causing over 125 13 million confirmed cases and over 2.5 million deaths worldwide as of March 2021. 1 Understanding 14 the long-term immunological response in infected individuals will be central in defeating the 15 COVID-19 pandemic, and understanding humoral kinetics, timing, and persistence of SARS-CoV-2 16 antibodies after natural infection, will be essential to ensure an appropriate vaccination strategy. 17

Reports studying the immunological response in previous coronaviruses outbursts have shown 18 that after infection, a specific IgG response is elicited and sustained for 1-2 years declining 19 thereafter, 2, 3 but available data for SARS-CoV-2 is very limited beyond 4 months 4 , especially in 20 health care workers (HCW), a population particularly exposed to SARS-CoV-2 with potential high 21 risk for reinfection. Indeed, the scarce published longitudinal data on seroprevalence of HCW 22 shows a wide variability on the percentage of subjects who remain seropositive after 4 months of 1 infection, ranging from 42 to 96% in predominantly small follow up cohorts. 5-8 . Furthermore, there 2 is almost no data about asymptomatic infected HCW. 3

To the best of our knowledge, this is the first study to evaluate the durability and to characterize 4 the longitudinal profile of anti-SARS-CoV-2 antibody levels beyond 6 months in a large cohort of 5 seropositive health personnel. In this context, we conducted a 10-month follow-up serological 6 quantitative study to determine the kinetics of the humoral response against SARS-CoV-2 infection 7 in a representative sample of over 300 seropositive HCW of Grupo HM Hospitales (GHM), a chain 8 of 17 Hospitals across 4 regions of Spain. Furthermore, we analyzed different variables to identify 9 clinical and demographic factors associated with stability/declining of antibody levels and 10 seroreversion. 11

The current work expands on a transversal study of seroprevalence of the entire population of 12 workers of Grupo HM-Hospitales, over 6000 HCW, to identify both symptomatic and 13 asymptomatic individuals who presented anti-SARS-CoV-2 antibodies, the results of which have 14 been recently published. 9 15

Prospective study, measuring the evolution of serum IgG anti-SARS-CoV-2 titers among employees 18 of the GHM who presented a positive result for SARS-CoV-2 antibodies after the first COVID-19 19 pandemic wave in Spain. The baseline study was carried out between April-June,2020, including 20 6038 employees (mean age:43.8;71%female) and 662(11.0%) presented IgG anti-SARS-CoV-2 at 21 baseline (39% asymptomatic), the results of which have been published. 9 22

A sample corresponding to 50% of all seropositive HCW was randomly selected by simple random 1 sampling from the baseline study, and its longitudinal profile of anti-SARS-CoV-2 antibodies was 2 evaluated (n=331). Serial blood measurements were collected at baseline (April13 th -May28 th 3 ,2020) and in 2 follow-up visits (last measurement: January12 th -February9 th ,2021). Those HCW 4 with a significant increase in the antibody titer during the follow-up were specifically studied to 5 rule out reinfection by evaluating clinical data and performing PCR test. 6

In January 2021, the COVID-19 vaccination of all healthcare personnel was implemented in all 7

Grupo HM-Hospitales workers and, as a result, 185 (55.4%) workers of the study received the first 8 dose before the last serological extraction, with a median (IQR) time between the vaccination (first 9 dose) and the third serological sample of 11 (8-13) days, without no impact in antibody titers 10 waning (slopes between the last two measurements were -0.49 [-0.71 Figure S1 in the Supplement). 13

Quantification of antibodies against SARS-CoV-2. 14 We used the Enzyme Immunoassay (ELISA) developed by DIA.PRO to measure serum anti-15 nucleocapsid (anti-N) and anti-spike (anti-S) protein IgG antibody titers against SARS-CoV-2 in a 16 venous blood sample for all the measurements of the study. This serum test has a clinical 17 sensitivity estimated at 98%-100% and a specificity >98% (Enzyme immunoassay for the 18 determination of IgG antibodies to COVID-19 in human serum and plasma DIA.PRO IFU). Serum 19

IgG titers were considered negative (non-reactive) with a result less than 0.90 arbitrary units 20 (AU)/mL (<0.90 AU/mL), positive (reactive) with a result greater than 1.10 AU/mL (>1.10AU/mL) 21 and indeterminate with a result in the interval between 0.900 and 1.10(0.90≤x≤1.10AU/mL). 22

The ELISA test used detects IgG antibodies against specific subunits of spike and nucleocapsid 1 proteins with a stablished proportion in the solid phase. This specificity would explain in part why 2 the antibody titers are not affected in vaccinated subjects as compared with other available kits. 3

Individuals were classified according to their antibody profile over time as: stable (n=40) when the 5 variation in antibody titers were lower than 25%, declining (n=251) when the loss of antibody 6 titers was higher than 25% overtime and others (n=15) for those profiles that show an increase of 7 more than 25% in antibody titer at some time point. 8

Summary statistics were performed as absolute and relative frequencies (%). Chi-squared tests (or 9

Fisher tests when needed) were used to study the dependence between stable and declining 10 profiles and age (categorized by median), sex, symptoms, infection category, grade of exposure to 11 COVID-19, receiving SARS-CoV-2 vaccine or not and baseline antibody titer (categorized by 12 terciles). Those variables with a p value <0.2 were considered for univariable and multivariable 13 logistic regression with the antibody profile (stable or declining) as dependent variable. 14 Linear mixed models were used to investigate antibody waning for those individuals with declining 15 profile. Assuming antibody levels fell exponentially, the natural log transformed titers were 16 modelled over time. Correlated random intercept and slope were allowed. The univariable effect 17 of covariates, including age, sex, grade of exposure to COVID-19, infection category and 18 symptoms, on the fixed effect intercept and slope was analysed. A multivariable model including 19 all the analysed covariates was performed to study independent effects on antibody waning. 20

Analyses were performed using R 4.0.4 and the lme4 library. 10 21 Seroreversion was defined as the loss of antibody titers below the diagnostic threshold of 1.1A.U. 1

The association of the different variables with seroreversion was analysed using Chi-squared test 2 or Fisher test if required. 3

Three hundred and six HCW were included in the final analysis (25 subjects were excluded due to 5 incomplete data). Serological testing was performance between 13-April 2020 and 9-February 6 2021, with a median (IQR) [ 

We identified two distinct antibodies dynamic profiles according to the quantitative antibody 11 concentration kinetics, which we have categorized as: declining, which included those subjects 12 who presented a significant decrease in antibody titers (defined as a decline > 25% between the 13 first and the last serological testing) (n=251; 82.0%); and stable, for those who presented a 14 variation in antibody titers less than 25% throughout the study period (n= 40; 13.1%). Fifteen subjects (4.9%) showed a significant increase (>25%) in antibody titer ( Figure S2 in the 19 Supplement shows their dynamic profile). We believe this increase probably represents either 20 baseline measurement being carried out during the increasing phase for IgG or possible re-21 infection (these HCW were specifically evaluated to detect reinfection), and these subjects were 1 therefore excluded from the analysis. 2

Factors associated with persistence/declining anti-SARS-CoV-2 IgG antibody levels are shown in 3 with stable/non-declining antibody titers. The multivariate logistic regression model showed that 10 age, baseline antibody titer and exposure risk are independently and positively associated with the 11 stable antibody profile. 12

We considered the antibody titer of the first measurement as the baseline antibody levels. It is 14 worth mentioning that 75 individuals reported a positive PCR with median (IQR) of 28 (22 -37) 15 days prior to the first measurement. 16

The mean trajectory of IgG antibody levels in each individual with declining profile was modelled 17 using a linear mixed model (Figure 2A , Table 3, Table S1 ). Estimated mean antibody half-life was 18 146.3 (95%CI: 138.6-154.9) days from baseline. The estimated mean baseline antibody titer was 19 6.55 (95% CI: 6.20 -6.93) arbitrary units (A.U.). Initial antibodies titer were correlated with longer 20 estimated antibody half-lives ( Figure 2B , Spearman's rank R 2 = 0.55, p < 0.0001). 21

The effect of demographic and other covariates on the antibody's trajectories was marginal for 1 longer half-lives for age, high-risk exposure and previous COVID-19 symptoms ( Table 3; Table S1 Table 4 . 16 In order to have a projection of the time to loss of antibodies, we used a linear mixed baseline 17 model in order to extrapolate the mean time to cross the threshold of 1.1A.U. (mean time to loss 18 antibodies), finding that, on average, subjects will have detectable concentrations of IgG anti 19 SARS-COV-2 up to 375 (95%CI:342-408) days from baseline (Figure 2A) . 20

Fifteen subjects did not present a declining or stable longitudinal antibody profile. Evaluating their 1 profile and available clinical data, we found one probable case of PCR-documented reinfection 2 ( The aim of the current study was to investigate the immunological response in antibody levels 5 over time (mean follow up over nine months) in a large and representative cohort of more than 6 300 HCW who tested positive for IgG anti-SARS-CoV-2 antibodies after the first wave of COVID-19 7 outbreak in Spain. To the best of our knowledge, this is the first report evaluating the kinetics of 8 SARS-CoV-2 antibodies in HCW beyond 6 months. These results provide further insight into the 9 nature of post-infection immunity. 10

The results showcase a high percentage of healthcare personnel maintaining detectable antibody 11 levels beyond 9 months after infection (median follow-up 274 days) with more than 75% of 12 subjects remaining seropositive, but with a significant decline in antibody levels in most of them. Our results are inconsistent with previous works in HCW which included significantly shorter 20 follow up periods. Some studies reported a rapid loss of SARS-CoV-2 specific antibodies within the 21 first three months after infection, alerting about a short-lasting humoral protection. 6,7 These 22 differences may reflect, among other issues, differences in baseline characteristics of the study 23 population (symptomatic category of SARS-CoV2 infection, age, baseline antibody titers…) and, 1 more importantly, different sensitivity of the immunoassays used to detect specific antibodies 2 against SARS-CoV-2. One of the strengths of the current work is that the same immunoassay (from 3 the same company, with same badge of production) was used in the complete population 4 throughout the study testing for both serum anti-nucleocapsid (anti-N) and anti-spike (anti-S) 5 protein IgG antibodies. 6

Here, we report a novel finding about the dynamic profile of the humoral response against SARS-7

CoV-2. While most subjects present a declining profile, a subgroup shows a stable (non-declining) 8 antibody titer profile during at least the first 10 months after infection. Further characterization of 9 this stable profile shows a significant correlation with high baseline antibody levels, but other 10 factors as well. Interestingly, a higher degree of exposure to COVID-19 patients is associated to 11 stability in antibody levels, with stable profile being observed more predominantly in first line 12 HCW. This thought-provoking observation may be signaling a potential mechanism of viral re- 

Higher baseline antibody titers have been associated with an increased probability to sustain 1 detectable antibody levels over time compared to lower baseline antibody levels. 6-8 In the current 2 study, baseline antibody levels were positively correlated with half-lives of antibody and HCW who 3 presented baseline antibody levels in the third tercile showed a significant longer half-life of 4 antibodies ( Figure 2B ). This finding has been associated with the severity of SARS-CoV-2 infection. 5

Previous reports have described that levels of antibodies against SARS-CoV-2 in the acute phase 6 correlate to clinical disease severity and then decline after 3 months since onset of illness. 17-19 7

While a recent study reported that the antibodies are persistent for at least 6 months in severe 8 COVID-19 patients, 20 reports in asymptomatic infection and pauci-symptomatic disease have 9

showed a weak and non-sustained humoral immunity response against SARS-CoV-2. 21,22 10 Our findings, expanding previous reports, suggest that moderate/severe COVID-19 patients 11 present significantly higher levels of baseline antibodies, but no significant difference in the 12 decreasing rate over time, compared to subjects with mild COVID-19 (half-life of antibodies was 13 significantly longer in symptomatic compared with asymptomatic COVID-19, but not in 14 moderate/severe compared to asymptomatic/ mild infection). In contrast, HCW with high-grade 15 risk exposure to COVID-19 patients showed not only significantly more frequency of stable 16 dynamic antibody response over time, but also longer antibody half-life and lower estimated 17 monthly decrease in antibody titers compared to HCW with low-grade exposure (Table S2 ). This 18 finding adds validity to the previously mentioned mechanism that viral re-exposure could be 19 playing a role in this more stable dynamic antibody profile in first line HCW. 20

Finally, in our study, age above 45 was also independently associated with stable dynamic 21 antibody profile and longer antibody half-life. This finding is consistent with a previous study 22

reporting that increasing age was associated with longer durability of anti-SARS-CoV-2 antibodies. 8 23 1

Whether the rate of decline in anti-SARS-CoV-2 antibodies is associated with an increased risk of 2 reinfection and disease remains to be determined. This issue has special implications in first line 3 HCW, since their inherent particular risk due to close contact with COVID-19 patients. In the 4 current work we detected a single case of possible reinfection. This is in the line with the findings 5 of Lumley et al who reported that the presence of anti-SARS-CoV-2 antibodies is associated with a 6 significant reduction in risk of SARS-CoV-2 reinfection in the ensuing 6 months, with only 2 7 asymptomatic cases of re-infection and no symptomatic infection detected in seropositive 8 subjects. 23 9

The results of the current study suggest an important role for serological screening and follow-up 11 and may help in our understanding of the humoral response in HCW exposed to SARS-COV-2. 12

These in turn may have potential implications in the current vaccination strategy. Results from 13 ongoing long-term studies will help better understand the kinetics of the humoral response and 14 their association to clinical manifestations of the disease. 15

The current study has several limitations. One of the main is that we have non-acute information 17 on timing of infection onset to evaluate antibody kinetics. However, our study is based on a real-18 world HCW cohort with a wide range of symptoms, including 40% of asymptomatic infected HCW, 19 and in this setting does not allow to accurately obtain this information. Another limitation that 20 needs to be pointed out is that the methodology used quantitatively measures antibodies against 21 the spike and nucleocapsid proteins of SARS-CoV-2 (anti-S and anti-N antibodies) but does not 22

identify the presence of neutralizing antibodies. The measurement of anti-S and anti-N antibodies 1 is much more accessible in a general clinical setting, and supporting this strategy, several reports 2 have shown that anti-SARS-CoV-2 antibodies to the spike protein correlates with neutralizing 3 antibodies and are responsible for neutralizing activity against SARS-CoV-2. 24,25 4

Conclusions 5

In conclusion, our results demonstrate that most of infected HCW maintain seropositivity levels of 6

IgG anti-SARS-CoV-2 antibodies after 9 months, but with a significant decline in levels. A subgroup 7 of subjects shows a stable, sustained (non-declining) dynamic profile over time, associated to high 8 baseline antibody titers, higher age and high-grade exposure to COVID-19 patients. This work has not received any funding 18

The data underlying this article will be shared on reasonable request to the corresponding author 20 

World Health Organization (WHO)

Duration of serum neutralizing antibodies for SARS-14

CoV-2: Lessons from SARS-CoV infection

Immune response following infection with SARS-CoV-17 2 and other coronaviruses: A rapid review

Antibody Response After SARS-CoV-2 19 Infection and Implications for Immunity : A Rapid Living Review

Change in antibodies to SARS-CoV-2 over 22 60 days among health care personnel in

Decline in SARS-CoV-2 antibodies after mild 24 infection among frontline health care personnel in a multistate hospital

Rapid decline of neutralizing antibodies against SARS-27

CoV-2 among infected healthcare workers

The duration, dynamics and determinants of SARS-1

CoV-2 antibody responses in individual healthcare workers

Seroprevalence of SARS-CoV-2 antibodies in over 4 6000 healthcare workers in Spain

Fitting Linear Mixed-Effects Models Using lme4

Robust neutralizing antibodies to SARS-CoV-2 8 infection persist for months

Persistence of SARS-CoV-2-specific B and T cell 10 responses in convalescent COVID-19 patients 6-8 months after the infection

Stable neutralizing antibody levels 6 months after mild 13 and severe COVID-19 episodes

Immunological memory to SARS-CoV-2 assessed for up to 8 15 months after infection

Persistence of antibodies to SARS-CoV-2 in 17 relation to symptoms in a nationwide prospective study

Seroprevalence and humoral immune durability of anti-SARS-CoV-19 2 antibodies in Wuhan, China: a longitudinal, population-level, cross-sectional study

SARS-CoV-2 neutralizing antibodies in patients 22 with varying severity of acute COVID-19 illness

A longitudinal study of SARS-CoV-2-infected patients 24 reveals a high correlation between neutralizing antibodies and COVID-19 severity

Longitudinal observation and decline of neutralizing 27 antibody responses in the three months following SARS-CoV-2 infection in humans

Effective virus-neutralizing activities in antisera from the first 30 wave of survivors of severe COVID-19

Antibody response to SARS-CoV-2 in infected 32 patients with different clinical outcome

Clinical and immunological assessment of asymptomatic 1 SARS-CoV-2 infections

Antibody status and incidence of SARS-CoV-2 3 infection in health care workers

Potently neutralizing and protective human antibodies 5 against SARS-CoV-2

Human neutralizing antibodies elicited by SARS-CoV-2 infection