Complete Epstein-Barr virus seropositivity in a large cohort of patients with early multiple sclerosis.

OBJECTIVE: To determine the prevalence of antibodies to Epstein-Barr virus (EBV) in a large cohort of patients with early multiple sclerosis (MS).
METHODS: Serum samples were collected from 901 patients with a clinically isolated syndrome (CIS) or early relapsing-remitting multiple sclerosis (RRMS) participating in the German National MS cohort, a prospective cohort of patients with early MS with stringent inclusion criteria. Epstein-Barr nuclear antigen (EBNA)-1 and viral capsid antigen (VCA) antibodies were measured in diluted sera by chemiluminescence immunoassays (CLIAs). Sera of EBNA-1 and VCA antibody-negative patients were retested undiluted by an EBV IgG immunoblot. For comparison, we retrospectively analysed the EBV seroprevalence across different age cohorts, ranging from 0 to >80 years, in a large hospital population (N=16 163) from Berlin/Northern Germany.
RESULTS: EBNA-1 antibodies were detected by CLIA in 839 of 901 patients with CIS/RRMS. Of the 62 patients without EBNA-1 antibodies, 45 had antibodies to VCA as detected by CLIA. In all of the remaining 17 patients, antibodies to EBV were detected by immunoblot. Altogether, 901 of 901 (100%) patients with CIS/RRMS were EBV-seropositive. EBV seropositivity increased with age in the hospital population but did not reach 100% in any of the investigated age cohorts.
CONCLUSION: The complete EBV seropositivity in this large cohort of patients with early MS strengthens the evidence for a role of EBV in MS. It also suggests that a negative EBV serology in patients with suspected inflammatory central nervous system disease should alert clinicians to consider diagnoses other than MS. © Author(s) (or their employer(s)) 2020. Re-use permitted under CC BY-NC. No commercial re-use. See rights and permissions. Published by BMJ.

Introduction

Strong and consistent evidence indicates an association of multiple sclerosis (MS) and infection with the Epstein-Barr virus (EBV).1–3 This led to the proposal that, from an epidemiological perspective, MS could be regarded as a late complication of EBV infection.4 If this was true, one would expect that there should be practically no EBV-seronegative patients with MS.5 Previous seroepidemiological studies and meta-analyses thereof indeed observed very high EBV seropositivity rates (~98% to 100%) in patients with MS or a clinically isolated syndrome (CIS).1 6–11 Nevertheless, the detection of few EBV-seronegative persons with a diagnosis of MS in some of those studies suggests that EBV-seronegative MS may occur. However, as inclusion criteria of previous studies on the seroprevalence of EBV in patients with MS were heterogeneous, it cannot be excluded that EBV-seronegative persons with a diagnosis of MS reported in the literature may occasionally have been misclassified and could in fact have diagnoses other than MS.8 Furthermore, it was shown that the EBV seroprevalence in patients with MS may depend on the sensitivity and specificity of the applied antibody assays and that in the likely most robust studies, that is, those that used two independent methods for detection of EBV antibodies, EBV seropositivity in patients with a diagnosis of MS may reach 100%.7

To systematically search for EBV-seronegative patients with MS, we analysed the EBV seroprevalence in 901 patients of the German National MS cohort, a prospective longitudinal observational cohort of patients with early MS with stringent inclusion criteria. For comparison, we retrospectively determined EBV seroprevalence rates across different age cohorts in a large hospital population (N=16 163) from Berlin/Northern Germany.

Patients and methods

Patients with early MS

The German National MS cohort is a multicentre prospective longitudinal observational cohort which recruited a total of 1212 patients between August 2010 and December 2014.12 Inclusion criteria have previously been reported in detail and comprise female and male patients aged ≥18 years and

A diagnosis of a CIS (defined as a first clinical event suggestive of inflammatory demyelination) within 6 months before inclusion and fulfilment of three of four Barkhof criteria.

A diagnosis of a CIS within 6 months before inclusion and fulfilment of two of four Barkhof criteria and intrathecal IgG production in cerebrospinal fluid (CSF) or abnormal visually evoked potentials.

A diagnosis of a CIS within 6 months before inclusion and fulfilment of the McDonald 2010 criteria for relapsing–remitting multiple sclerosis (RRMS).

A diagnosis of RRMS based on the McDonald 2005 criteria within 2 years before inclusion.12

Exclusion criteria comprise previous use of any disease-modifying therapy for MS (except for short-term relapse treatment), primary or secondary progressive MS, concurrent progressive neurological diseases and conditions interfering with the assessment plan, for example, contraindications to MRI. The assessment plan includes standardised collection of demographic and clinical data, assessment of the Expanded Disability Status Scale (EDSS) score and standardised sampling of biospecimens. All patients were recruited at specialised MS centres and study data were monitored with a query system.

Sera from therapy-naïve patients were collected, centrifuged and aliquoted during the baseline visit at the participating centres according to standard operating procedures and were shipped overnight to the Department of Neurology, Technische Universität München, where they were stored at −80°C. Baseline serum samples of 901 patients were available for EBV antibody testing and were sent on dry ice to the Department of Neurology, Charité – Universitätsmedizin Berlin.

Detection of EBV antibodies

Serum IgG antibodies to Epstein-Barr nuclear antigen (EBNA)-1 and to the EBV viral capsid antigen (VCA) were measured by Liaison (DiaSorin, Saluggia, Italy) automated quantitative chemiluminescence immunoassay (CLIA) at Labor Berlin GmbH, Berlin, Germany. As only limited amounts of serum (~50 µL) from participants of the German National MS cohort were available, all sera had to be measured in dilution. EBNA-1 IgG antibodies were measured in serum samples diluted either 1:20 (n=40) or 1:10 (n=861) in assay dilution buffer. With these dilutions, EBNA-1 IgG levels <3 U/mL were considered negative, and EBNA-1 IgG levels ≥3 U/mL were considered positive. All EBNA-1 IgG-negative serum samples were tested at a dilution of 1:10 in assay dilution buffer for antibodies to VCA. With this dilution, VCA IgG levels <10 U/mL were considered negative and VCA IgG levels ≥10 U/mL were considered positive. As per the manufacturer’s instructions, the assay range of the EBNA-1 IgG CLIA is 3–600 U/mL and that of the VCA IgG CLIA is 10–750 U/mL.

Since we had to determine EBNA-1 and VCA IgG by CLIA in diluted sera of participants of the German National MS cohort, we retested EBNA-1 and VCA IgG-negative sera without dilution by an EBV IgG immunoblot (recomLine EBV IgG, Mikrogen, Germany). As only a small volume of serum (20 µL) is needed for the EBV IgG immunoblot, we were able to use undiluted sera for these measurements. Serum samples that were positive in at least one of the three assays (EBNA-1 IgG CLIA, VCA IgG CLIA or EBV IgG immunoblot) were considered EBV seropositive.

Retrospective analysis of EBV seroprevalence in a large hospital population

We retrospectively analysed results of EBV serologies, which were performed for routine diagnostic purposes in 16 163 persons at Labor Berlin GmbH, Berlin, Germany, between January 2014 and December 2016. Sera were sent for EBV serological testing from persons treated as inpatients or outpatients at university hospitals (Charité – Universitätsmedizin Berlin) located in Berlin or community hospitals located in Berlin and Northern Germany. The 16 163 persons were included irrespective of diagnoses or the reason for ordering EBV serologies. For individuals tested more than once during the study period, only the results of the first EBV serology were considered. From all persons included in the analysis, results of serological tests for EBNA-1 IgG and VCA IgG and VCA IgM had to be available. Notably, EBNA-1 IgG is a marker of past EBV infection, VCA IgG can be found in primary and past EBV infections, and VCA IgM is a marker of primary EBV infection.13 Testing for EBNA-1 IgG, VCA IgG and VCA IgM was performed by Liaison automated quantitative CLIA using undiluted sera. Thus, the same method (Liaison CLIA) was used for testing of patients with CIS/RRMS and patients of the hospital population, with sera of patients with CIS/RRMS being measured in dilution and sera of patients from the hospital population being measured without dilution. According to the manufacturer’s recommendations, in undiluted sera, EBNA-1 IgG levels <5 U/mL were considered negative, levels between 5 and 20 U/mL were considered equivocal and levels ≥20 U/mL were considered positive; VCA IgG levels <20 U/mL were considered negative and VCA IgG levels ≥20 U/mL were considered positive; and VCA IgM levels <20 U/mL were considered negative, levels between 20 and 40 U/mL were considered equivocal and VCA IgM levels≥ 40 U/mL were considered positive. Persons with EBNA-1 IgG and VCA IgG and VCA IgM below the respective cut-offs were considered EBV-seronegative. Persons in whom at least one of the three antibodies, EBNA-1 IgG, VCA IgG or VCA IgM, was above the respective cut-offs were considered EBV-seropositive. From the age of 5 years onwards, persons were grouped in 5-year age cohorts. To analyse the EBV seroprevalence in early life in more detail, newborns and infants below 5 years of age were grouped in smaller age cohorts.

Statistical analyses

Continuous data were summarised using medians and IQRs. Categorical data are reported as absolute and relative frequencies (%). The significance of different EBV seropositivity rates between patients with early MS and persons in the hospital population was assessed by two-tailed Fisher exact tests (https://www.graphpad.com/quickcalcs/contingency2/). A p value of <0.05 was considered significant.

Results

EBV seroprevalence in early MS

The median (IQR) age of the 901 patients included in this study was 33 (27–41) years, and 630/901 (69.9%) were women. At the time of blood sampling, 380 (42.2 %) had a diagnosis of a CIS and 521 (57.8%) had a diagnosis of RRMS according to the inclusion criteria of the German National MS cohort. The median (IQR) EDSS score of patients with CIS and RRMS was 1.5 (1.0–2.0, data available from n=899 patients).

IgG antibodies to EBNA-1, as measured in diluted sera by CLIA, were positive in 839 of 901 (93.1%) patients with CIS/RRMS. Of the 62 patients without antibodies to EBNA-1, 45 (72.6%) had positive IgG antibodies to VCA, as measured in diluted sera by CLIA. Of the 17 remaining patients, 17 (100%) had IgG antibodies to EBV as detected in undiluted sera by EBV IgG immunoblot. Detailed results of EBV IgG immunoblots are provided in online supplementary table 1. In sum, 12 of 17 sera reacted with the EBNA-1 p72 antigen and 17 of 17 sera reacted with at least one of two VCA antigens (p18 and p23) included in the EBV IgG immunoblot. Altogether, 901 of 901 (100%) patients with CIS/RRMS investigated in this work were EBV seropositive.

EBV seroprevalence in a large hospital population

To compare the EBV seroprevalence in patients with early MS with the general EBV seroprevalence in a similar geographical region, we analysed the EBV seroprevalence across different age groups in a large hospital population (n=16 163) from Berlin/Northern Germany. Data on sex were available from 16 036 of these persons, of which 7714 (48%) were female and 8322 (52%) were male. Figure 1A summarises the EBV seroprevalence in the entire hospital population; the respective source data are provided in online supplementary table 2. EBV seropositivity was high in newborns (<0.25 years) and lowest in 0.75–<1 year olds. Subsequently, EBV seropositivity increased until adulthood, with steepest increases in the age cohorts of 1.5–2.0 and 15–19 years. The increases of EBV seropositivity in these age cohorts were paralleled by an increased detection of VCA IgM antibodies, indicating primary EBV infections. While EBV seropositivity further increased with increasing age and was ≥98% in all 5-year age cohorts from 45 to 79 years, it did not reach 100% in any of these age cohorts. EBV seropositivity tended to be overall slightly higher in women than in men, but the course of EBV seropositivity across the life span was similar in both sexes (figure 1B).

Figure 1

(A) EBV seropositivity rates by age cohorts in a large hospital population (N=16 163). The percentage positivity of IgG antibodies to EBNA-1 and of IgG and IgM antibodies to VCA in different age cohorts is shown. Note that the x-axis is not proportional. EBV seropositivity was defined as seropositivity to at least one of the three antibodies, EBNA-1 IgG, VCA IgG or VCA IgM. The number of persons analysed in each age cohort is indicated above the graph. (B) EBV seropositivity rates by age cohorts in a large hospital population are shown separately for female (n=7714) and male (n=8322) persons. EBV seropositivity was defined as seropositivity to at least one of the three antibodies, EBNA-1 IgG, VCA IgG or VCA IgM. The number of female/male persons analysed in each age cohort is indicated above the graph. EBNA-1, Epstein-Barr nuclear antigen-1; EBV, Epstein-Barr virus; VCA, viral capsid antigen.

When comparing the EBV seropositivity rates in patients with early MS and in participants of the hospital population in the age range of 20–40 years, that is, the typical age of clinical onset of MS,14 EBV seropositivity among patients with MS (610/610, 100%) was higher than among participants of the hospital population (4134/4343, 95.2%; p<0.0001). In a comparison of EBNA-1 IgG seropositivity by age between patients with early MS and patients from the hospital population, the difference in EBNA-1 IgG seropositivity was strongest in the youngest analysed age cohort (20–24 years) and subsequently declined with increasing age (online supplementary figure 1).

Discussion

The key result of this study is a complete EBV seropositivity in a large cohort (n=901) of patients with early MS. While this finding is consistent with the known high EBV seroprevalence in MS,1 6–11 the absence of any EBV-seronegative patients with early MS in our cohort appears remarkable and further strengthens the evidence for an association of EBV infection and MS. The already 100% EBV seropositivity in the 380 patients with a CIS, that is, the earliest clinically detectable manifestation of MS, complies with the concept that EBV infection precedes the clinical onset of MS and suggests that EBV exerts its role early in the development of MS.15

The German National MS cohort is a prospective longitudinal cohort with stringent inclusion criteria, requiring MRI, CSF or electrophysiological evidence supportive of MS in patients with a CIS.12 Thus, the likelihood of inclusion of patients not having a true CIS, that is, a CIS as a first clinical manifestation of MS, or not having true MS into this well-characterised cohort of patients with CIS/early RRMS appears very low. Assuming that there is a genuine association of EBV and MS, the high degree of diagnostic certainty in patients participating in the German National MS cohort may therefore explain the 100% EBV seropositivity observed in this cohort. This conclusion is supported by findings of a previous meta-analysis, which in a post hoc analysis found higher ORs for EBNA-1 and VCA IgG seropositivity in serological studies of EBV prevalence in MS that included confirmed cases of MS as compared with studies that included confirmed and probable cases of MS.8 The most plausible explanation for this observation is that probable cases of MS are more likely to comprise misclassified patients who actually do not have MS. The EBV seroprevalence of those misclassified patients would be expected to correspond to that of the general population and to thus be lower than the EBV seroprevalence of patients with true MS, which could explain the occasional detection of EBV-seronegative persons in some former studies on EBV seroprevalence in MS.8

Our study corroborates results obtained in 1047 retrospectively collected patients with a CIS, only one of whom was found to be EBV seronegative.10 16 Altogether, the present evidence suggests that EBV-seronegative patients with MS, if they should exist at all, occur extremely rarely. An implication of these findings of relevance for clinical practise is that a negative EBV serology in a patient with suspected inflammatory central nervous system disease should alert clinicians to consider diagnoses other than MS.2 5 7 Given that the difference in the EBV seropositivity rates between patients with MS and controls declines with age (see also online supplementary figure 1), the younger the age of the patient, the more informative testing for EBV should be. Future studies on this issue may therefore focus in particular on children with suspected inflammatory central nervous system disease, as the difference between the likewise high EBV seroprevalence in children with MS and the EBV seroprevalence in paediatric controls is rather pronounced.17

The high EBV seropositivity in newborns and the rapid decline until the age of 1 year observed in the hospital population is explained by placental transmission and subsequent disappearance of maternal EBV antibodies. The subsequent steep increases of EBV seropositivity in early infancy and in the age cohort of 15–19 years correspond to previous data on the natural course of EBV infection in industrialised countries of the northern hemisphere.18–21 The somewhat higher seroprevalence of VCA IgG as compared with EBNA-1 IgG in patients of the hospital population is consistent with findings of previous large seroepidemiological studies and likely related to the known phenomenon that a certain proportion of persons infected with EBV does not develop antibodies to EBNA-1.8 13 22 Importantly, we found that about 30% of 10–14 year olds in our hospital population from Northern Germany were EBV seronegative. This is of relevance as these individuals are particularly prone to develop symptomatic primary EBV infection in the form of infectious mononucleosis, which is associated with an about twofold increased risk of MS.23 24

The higher EBV seropositivity rate in patients with early MS (100%) as compared with persons in the hospital population (95.2%) in the age range of 20–40 years is consistent with previous data8 and appears compatible with the concept that EBV may be a necessary but not sufficient factor for the development of MS.

Of note, due to only limited amounts of serum available, sera of patients from the MS cohort had to be measured by CLIA in dilution. All EBNA-1 and VCA antibody-negative patients with CIS/early RRMS, as determined in diluted sera by CLIA, were thus retested by an EBV IgG immunoblot, which requires only a small volume of serum, enabling us to analyse undiluted sera. Detection of EBV antibodies by immunoblot in patients, who were EBV and VCA IgG negative by CLIA, is therefore explained by the fact that those sera were measured undiluted, that is, at a 10-fold higher concentration than in the CLIA. The majority of the 17 sera tested by the EBV immunoblot contained antibodies to EBNA-1, and all 17 sera contained antibodies to VCA (see online supplementary table 1), further supporting the conclusion that antibody responses in those 17 patients differed only quantitatively but not qualitatitvely from that of the other 884 patients.

Of further note, unlike patients with CIS/RRMS, EBV-seronegative participants of the hospital population, as determined by CLIA using undiluted sera, were not retested by an EBV immunoblot. We consider it very unlikely that this could have resulted in a higher rate of EBV seronegativity in the hospital population than in patients with MS for the following reasons: first, in routine diagnostic serology, absence of EBNA-1 IgG, VCA IgG and VCA IgM, as determined by CLIA in undiluted sera, is generally accepted to reliably indicate EBV seronegativity with no further confirmatory tests being required.13 21 25 Second, while we are not aware of published studies that directly compared the recomLine EBV IgG immunoblot with the liaison CLIA applied in our work, in a previous comparative study, the sensitivity of the recomLine EBV IgG immunoblot was not higher than that of another CLIA method (Architect; Abbott, Wiesbaden, Germany),25 which is similar to the CLIA Method used in our work.26 Third, when we re-tested 28 sera, which were EBNA-1 IgG and VCA IgG negative, as determined in undiluted sera by the Liaison CLIA, by EBV immunoblots, all of these sera were likewise EBV seronegative in the EBV immunoblot (unpublished observation). Altogether, the available evidence therefore clearly argues against a higher sensitivity of the EBV IgG immunoblot as compared with determination of EBNA-1 IgG and VCA IgG in undiluted sera by the liaison CLIA.

A limitation of this study is that we did not determine EBV seroprevalence in the general population, but, similar to a previous large investigation,21 used a hospital population as a surrogate instead. The hospital population may have included patients in whom EBV serologies were ordered for a suspected primary EBV infection, which could potentially result in higher EBV seropositivity rates than in the general population. Conversely, our hospital population may also have included immunosuppressed patients in whom EBV serologies could potentially result false negative. However, given the very high number of patients analysed, it seems conceivable that the data obtained in the hospital population are overall representative of EBV seropositivity rates across different age ranges in the general population. Finally, as is the case with every serological test, we cannot exclude the occurrence of rare false-positive results. Nevertheless, such rare false-positive results would be highly unlikely to explain the 100% EBV seropositivity observed in patients with CIS/RRMS in our study.

Conclusion

The complete EBV seropositivity in this large cohort of patients with CIS/RRMS strengthens the evidence for a role of EBV in MS. It also suggests that a negative EBV serology in patients with suspected inflammatory central nervous system disease should alert clinicians to consider diagnoses other than MS. The results of this study are compatible with the concept that MS could be a rare late complication of EBV infection. Future studies should focus on the clarification of the mechanisms underlying the role of EBV in MS.