SARS-CoV-2 antibody seroprevalence in a large neuroimmunological patient cohort

The safety of patients with neuroimmunological conditions such as multiple sclerosis (MS) during the current SARSCoV-2 pandemic has been a subject of major debate [1] and an accurate estimation of the burden of SARS-CoV-2 infection is pivotal. However, given the high prevalence of asymptomatic patients, the imperfect sensitivity of polymerase chain reaction (PCR) assays performed at a single time point, and limited testing capacity, the true number of SARS-CoV-2 infections likely exceeds the reported one. Serosurveys for SARS-CoV-2 antibodies (Abs) have so far only been reported for few neuroimmunological patient cohorts and often with only suboptimal assays [2, 3]. Here, we investigate the seroprevalence of SARS-CoV-2 Abs among neuroimmunology outpatient clinic patients from May 14th to September 30th, 2020, by two different assays and compare it to the seroprevalence of the general population. Further, precautionary health behavior was assessed to understand how possible overor under-cautious demeanor may affect seroprevalence. Of all patients who were admitted to the neuroimmunology outpatient clinic of LMU Hospital in Munich, Germany, during mentioned period, 509 gave written informed consent to participate in the study. A serum sample for SARS-CoV-2 antibody serology was taken from all included patients. Clinical data were retrospectively obtained using structured documentation of routine clinical data. Precautionary health behavior data was acquired from a digital analyzable, paper-based questionnaire (Online Resource 1) which was answered by 474 out of 509 included patients. In the questionnaire, adherence to official recommendations concerning curfew, hygiene recommendations, and reduction of social contacts during and after the lockdown in Germany in spring 2020 (calendar weeks 12–19) was enquired by symmetrical, balanced Likert scales ranging from 1 (“not at all”) to 5 (“very much”), respectively. Each serum sample was tested for SARS-CoV-2 Ab by two assays at the Institute of Laboratory Medicine (LMU Hospital): Elecsys® Anti-SARS-CoV-2 electrochemiluminescence immunoassay system measuring pan-Immunoglobulin (ECLIA, Roche-Diagnostics, Basel, Switzerland; hereafter Ro-pan-Ig) and Anti-SARS-CoV-2 enzyme-linked immunosorbent assay measuring IgG only (ELISA, EUROIMMUN, Lübeck, Germany; hereafter Eu-IgG) [4]. While Ro-pan-Ig only detects antibodies against the SARS-CoV-2 nucleocapsid antigen, Eu-IgG exclusively binds Abs to the receptor binding domain (RBD). According to the manufacturer, Ro-pan-Ig has 99.5% sensitivity and 99.8% specificity ≥ 14 days after a positive polymerase chain reaction (PCR) assay, respectively; the Eu-IgG sensitivity was 94.4% and specificity 99.6% [4]. Joachim Havla and Tania Kümpfel contributed equally.


Dear Sirs,
The safety of patients with neuroimmunological conditions such as multiple sclerosis (MS) during the current SARS-CoV-2 pandemic has been a subject of major debate [1] and an accurate estimation of the burden of SARS-CoV-2 infection is pivotal. However, given the high prevalence of asymptomatic patients, the imperfect sensitivity of polymerase chain reaction (PCR) assays performed at a single time point, and limited testing capacity, the true number of SARS-CoV-2 infections likely exceeds the reported one. Serosurveys for SARS-CoV-2 antibodies (Abs) have so far only been reported for few neuroimmunological patient cohorts and often with only suboptimal assays [2,3].
Here, we investigate the seroprevalence of SARS-CoV-2 Abs among neuroimmunology outpatient clinic patients from May 14th to September 30th, 2020, by two different assays and compare it to the seroprevalence of the general population. Further, precautionary health behavior was assessed to understand how possible over-or under-cautious demeanor may affect seroprevalence.
Of all patients who were admitted to the neuroimmunology outpatient clinic of LMU Hospital in Munich, Germany, during mentioned period, 509 gave written informed consent to participate in the study. A serum sample for SARS-CoV-2 antibody serology was taken from all included patients. Clinical data were retrospectively obtained using structured documentation of routine clinical data. Precautionary health behavior data was acquired from a digital analyzable, paper-based questionnaire (Online Resource 1) which was answered by 474 out of 509 included patients. In the questionnaire, adherence to official recommendations concerning curfew, hygiene recommendations, and reduction of social contacts during and after the lockdown in Germany in spring 2020 (calendar weeks 12-19) was enquired by symmetrical, balanced Likert scales ranging from 1 ("not at all") to 5 ("very much"), respectively.
Each serum sample was tested for SARS-CoV-2 Ab by two assays at the Institute of Laboratory Medicine (LMU Hospital): Elecsys ® Anti-SARS-CoV-2 electrochemiluminescence immunoassay system measuring pan-Immunoglobulin (ECLIA, Roche-Diagnostics, Basel, Switzerland; hereafter Ro-pan-Ig) and Anti-SARS-CoV-2 enzyme-linked immunosorbent assay measuring IgG only (ELISA, EURO-IMMUN, Lübeck, Germany; hereafter Eu-IgG) [4]. While Ro-pan-Ig only detects antibodies against the SARS-CoV-2 nucleocapsid antigen, Eu-IgG exclusively binds Abs to the receptor binding domain (RBD). According to the manufacturer, Ro-pan-Ig has 99.5% sensitivity and 99.8% specificity ≥ 14 days after a positive polymerase chain reaction (PCR) assay, respectively; the Eu-IgG sensitivity was 94.4% and specificity 99.6% [4]. Statistical analyses were performed using R-studio version 4.0.2 and WinBUGS (V 1.4.3, Imperial College and MRC, UK). To estimate the seroprevalence two different approaches were applied: (i) For comparison with the seroprevalence of the general population in Munich, which was reported earlier using the Ro-pan-Ig assay only [5], Ro-pan-Ig seroprevalence in our cohort was adjusted for the manufacturer's sensitivity and specificity. 95% confidence intervals (CI) were calculated by Poisson distribution.
(ii) Additionally, for a more robust estimation of SARS-CoV-2-Ab seroprevalence in our study population, a Bayesian approach was applied which considers both Ro-pan-Ig and Eu-IgG Ab assay results [6].
To compare two seroprevalences, a Chi 2 test was applied. Questionnaire data were automatically retrieved by EvaSys software (ElectricPaper Evaluationssysteme, Lüneburg, Germany) and group comparisons were conducted by clustered Wilcoxon rank sum test.
Cohort details are provided in Table 1. In total, 11 out of 509 patients were identified as seropositive for SARS-CoV-2 Ab by either Ro-pan-Ig (n = 4) or Eu-IgG (n = 2) or both assays (n = 5), yielding an estimated seroprevalence of 1.27% (CI 0.46-2.44; approach (ii)) in our cumulated patient cohort (cumulated seroprevalence; cSP). The sensitivity-and specificity-adjusted cSP regarding Ro-pan-Ig only test results was slightly higher (1.58%; CI 0.49-2.67; approach (i)). The latter was compared to the SARS-CoV-2 Ab seroprevalence of the general population in Munich (1.82%; CI 1.28-2.37; Ro-pan-Ig), as obtained from the Representative COVID-19 Cohort Munich [5]. Even though a tendency can be observed with a slightly lower cSP in our patient cohort, this difference was not significant (Chi 2 test, p = 0.96).
Six of all 11 seropositive patients received immunotherapy at the time of serum sampling, among whom two were B-cell depleted due to anti-CD20 therapy. The sensitivity-and specificity-adjusted seroprevalence for Ro-pan-Ig of all patients receiving immunotherapy at the time of serum sampling (iSP) was 1.39% (CI 0.09-2.69; approach (i)). Compared to the SARS-CoV-2 Ab seroprevalence of the general population in Munich [5], iSP was lower, but again not significantly reduced (Chi 2 test, p = 0.77).
Our cohort largely adhered to official regulations both during and after the first lockdown in Germany (Fig. 1) without significant differences between participants with and without immunotherapy (clustered Wilcoxon rank sum test, p = 0.63). Of note, also all 11 SARS-CoV-2 Ab seropositive patients strictly adhered to recommendations as assessed in our study. Overall, the seroprevalence of SARS-CoV-2 Ab of neuroimmunological patients did not differ significantly from the seroprevalence of the general population. The high rate of incidental seropositives (72.7%) including two patients on anti-CD20 treatment further highlights that a serology study may be useful for estimating the true burden of SARS-CoV-2 infection also in patients undergoing immunotherapy. Nevertheless, false-positive Ab results, particularly due to cross-reaction with other endemic coronaviruses, should be considered [8]. Inter-assay discrepancies may be due to different test reactivities, as reported previously [4].
In accordance with previous data [9], precautionary health behavior recommendations were strictly adhered to in our whole cohort which may have contributed to the low number of infections in our vulnerable patient population. Also, self-reported behavior was independent of immunotherapy use, suggesting a general tendency toward cautious health behaviors during the current pandemic in our patient cohort.
This study has limitations. First, the Representative Covid-19 Cohort Munich Study used for reference was conducted from April to June 2020 and, thus, does not correspond exactly with the time period of our study. Comparison of seroprevalence is further limited due to different cutoffs for the Ro-pan-Ig assay (0.4 in the Representative Covid-19 Cohort Munich Study instead of 1.0 as intended by the manufacturer and as applied in this study). Also, seroprevalence in mentioned study may have been underestimated since they exclusively relied on Ro-pan-Ig assay results in the final analysis. Furthermore, in our study, precautionary health behavior was self-reported and appropriate controls are lacking, restricting any direct causal link between behavior and seroprevalence. Generally, the presented data are limited to a low-prevalence setting. Nonetheless, this study emphasizes the importance of serosurveys as a public health practice to avoid underestimation of SARS-CoV-2 infection burden in patients with neuroimmunological disease.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.