|Home | About | Journals | Submit | Contact Us | Français|
To determine whether multiple sclerosis (MS) risk increases following primary infection with the Epstein-Barr virus (EBV), we conducted a nested case-control study including 305 individuals who developed MS and 610 matched controls selected among the over 8 million active-duty military personnel with serum stored in the Department of Defense Serum Repository. Time of EBV infection was determined by measuring antibody titers in serial serum samples collected before MS onset among cases, and on matched dates among controls. Ten (3.3%) cases and 32 (5.2%) controls were initially EBV negative. . All of the 10 EBV-negative cases became EBV positive before MS onset ; in contrast, only 35.7 % (10) of the 28 controls with follow-up samples seroconverted (exact p value = 0.0008). We conclude that MS risk is extremely low among individuals not infected with EBV, but it increases sharply in the same individuals following EBV infection.
A link between the Epstein-Barr virus (EBV) and MS is supported by the increased risk of MS, in individuals with history of mononucleosis 1, 2 or with elevated serum titers of antibodies against EBV nuclear antigens (EBNA) 3-6, and by the higher prevalence of EBV infection in MS cases than controls. 7-10 There are no longitudinal studies, however, estimating MS risk in EBV negative individuals or the temporal relation between EBV infection and MS. We therefore addressed these questions prospectively in a large population of healthy young adults
Active-duty US Army, Navy, and Marines personnel who have at least one serum sample in the Department of Defense Serum Repository (DoDSR), which stores approximately 46 million serum samples originally collected from over 8 million individuals for HIV testing. 11, 12
Cases were identified by searching the electronic databases of the Physical Disability Agencies of the US Army and US Navy for the diagnostic code corresponding to MS reported between 1992 and 2004, and then reviewing hard copy medical records. Overall, 515 cases were reviewed, of which 315 had definite (n=237) or probable (n=78) MS according to previously described criteria 12 and had at least one pre-clinical serum sample, i.e. a sample collected prior to the development of neurological symptoms (date of MS onset), as attested from the medical record. For each case, we obtained up to three pre-clinical samples (the earliest and latest available, as well as a third sample collected between those two). 12 Two controls for each case were randomly selected from the DoDSR population, matched by branch of service, age, sex, race/ethnicity, and dates of blood collection, as previously reported. 12 Ten cases could not be matched, leaving 305 cases and 610 matched controls in the analyses. By design, the controls in this study provide an unbiased estimate of the distribution of exposure and covariates among the millions of individuals who comprise the source population of the cases, and the odds ratios are unbiased estimates of the corresponding rate ratios that would be obtained by testing for EBV positivity all the individuals in the source population.13
Serum samples were sent to the laboratory ordered in triplets, each triplet including a case sample and the corresponding matched control samples in random order and without identification of case-control status. Blind quality control triplets consisting of multiple aliquots of the same serum samples were randomly interspersed amongst the study samples to monitor the reproducibility of the assays. EBV serology was performed using indirect immunofluorescence (VCA) or anticomplement immunofluorescence (EBNA complex, EBNA-1, EBNA-2) for the detection of IgG antibodies. 5, 14-16 IgG antibodies against cytomegalovirus (CMV), used here as a control herpes virus, were determined using an ELISA. 17 The serological assays for a first set of 83 cases and their 166 matched controls were performed at Virolab Inc., CA, USA. 5 Due to the closure of Virolab, the remaining samples (222 cases and 444 matched controls) were assayed at the Karolinska Institute in Stockholm, Sweden, under the supervision of one of the authors (KF); to reduce costs, antibodies to EBNA-1 and EBNA-2 were determined only in a subset of 166 cases and 332 matched controls.
Individuals were considered EBV negative (i.e. not infected with EBV) if they had no detectable anti-EBV antibodies (VCA < 20; EBNA complex, EBNA-1 or EBNA-2 < 5); a primary EBV infection was deemed to have occurred in these individuals if any of their subsequent samples became positive for anti-VCA antibodies, because these antibodies appear soon after infection and remain present indefinitely. 18 Exact logistic regression, which calculates unbiased estimates when data are very sparse, was used to estimate the relative risk of MS following EBV infection. All P values are 2-tailed. The statistical software SAS v9.1 (SAS Institute Inc, Cary, NC) was used for all analyses.
The research protocol was approved by the institutional review boards of the Walter Reed Army Institute of Research and the Harvard School of Public Health, both of which determined a waiver of informed consent was appropriate for the use of already existing medical record data and biological samples
The main characteristics of cases and controls are shown in Table 1. The relatively high proportions of males (66%) and blacks (30%) among MS cases and their matched controls reflect the demographic composition of the source population.
At baseline, 10 (3.3%) of the 305 MS cases and 32 (5.2%) of the 610 controls were EBV seronegative (p=0.2). At least one follow-up blood sample was available for all the EBV-seronegative cases and for 28 of the 32 seronegative controls. During the follow-up, all of the 10 initially seronegative cases became EBV positive before the onset of MS, while only 10 of the 28 controls seroconverted (p, exact test = 0.0008); results adjusted for duration of follow-up and serum levels of 25(OH) vitamin D were similar (p, exact test = 0.005). The seroconversion rate among controls was 11.2% per year, which is similar to that in previous longitudinal studies of military recruits or university students. 19, 20 The relation between time of seroconversion and MS onset is shown in Figure 1, and demographic characteristics and antibody titers in cases and controls who were EBV negative at baseline are shown in Table 2. In all cases, the anti-VCA titers increased to 320 or higher before MS onset. The mean interval between the date of the first EBV positive serum sample and MS onset was 3.8 yrs (range: 1.7 years to 7.0 years). Based on the assumption that EBV infection occurred, on average, at the mid-point between the last seronegative sample and the first seropositive sample, the mean interval between primary EBV infection and MS onset would be 5.6 years (range 2.3 years to 9.4 years). Because none of the EBV-negative individuals developed MS, the relative risk of MS following EBV infection cannot be directly estimated; the estimator involves a division by zero and the relative risk would tend to infinity. In addition, a statistically nonsignificant relative risk of 2.0 (95% confidence interval: 0.8 to 5.0) for MS was found for individuals who seroconverted during the follow-up as compared with individuals who were already EBV positive at baseline.
Seropositivity to CMV at baseline was lower in cases (49%) than in controls (58%), but there was no significant association between CMV seroconversion and MS risk (relative risk = 0.94; p=0.9).
In this large prospective investigation, we found a total absence of incident MS among individuals without detectable serum antibodies to EBV. About one third of these individuals seroconverted during the follow-up, and after seroconversion, manifested a rate of MS similar to that of individuals of the same age and sex who were already EBV positive at the study baseline.
These findings suggest that EBV infection increases the risk of MS, but several alternative interpretations need to be considered, including laboratory errors, reverse causation, and confounding. The detection of anti-EBV antibodies by indirect immunofluorescence provides a highly sensitive and specific marker of EBV infection, as demonstrated by observations in longitudinal studies that seronegative individuals have a high risk of EBV-induced acute infectious mononucleosis, from which EBV seropositive individuals are completely protected. 21, 22 The occurrence of EBV infection is confirmed by the observation that in all cases the anti-VCA antibodies, undetectable in the initial sample, increased to titers of 320 and above before MS onset, and anti-EBNA complex or anti-EBNA-1 titers also became positive in all cases. Further, the annualized rate of EBV seroconversion that we observed among controls during the follow-up (11.2%) is remarkably similar to that in previous studies in the U.S. Military Academy (12.3%) and college students (12.0 – 15 %). 19, 20 In those studies, seroconversion was associated with a high frequency of heterophile antibody-positive clinical mononucleosis, a marker of primary EBV infection. Thus, the appearance of anti-EBV antibodies in previously seronegative individuals is a robust marker of primary EBV infection. In contrast, seroconversion to CMV occurred at a similar rate in cases and controls, and was thus unrelated to MS risk.
The pathological process leading to MS starts before the clinical symptoms. 23 The results of our study could therefore be explained if pre-clinical MS resulted in increased susceptibility to EBV infection. While we cannot exclude this possibility, it should be noted that such an increase should be selective for EBV, because the prevalence of CMV infection was not higher among MS cases than controls, and rather extreme to result in a 100% prevalence of EBV infection before MS onset, neither of which seems likely.
Finally, a positive association between EBV infection and MS could be explained if both were affected by a common factor or confounder. A genetic resistance to both EBV infection and MS is not a plausible explanation, because many EBV negative individuals eventually became EBV infected and, following infection, susceptible to MS. Environmental factors, such as latitude, vitamin D, or cigarette smoking, being less strong risk factors for MS than EBV infection itself 24, and at most only weakly associated with EBV infection, are unlikely to explain the EBV-MS association. Confounding by an infectious agent that is co-transmitted with EBV cannot be excluded, but there is no evidence supporting this.
The most likely interpretation of the main results of our study, therefore, is that EBV infection itself increases the risk of MS.
Acknowledgment: We thank Mark Rubertone, MD, MPH, Armed Forces Health Surveillance Center, Silver Spring, MD for control and sample identification and retrieval; Noel Howard, MD, Department of the Navy, Secretary of the Naval Council of Review Boards, Washington, DC for MS case identification in the US Navy and Marines; David Armitage, MD, JD, US Army Physical Disability Agency, Washington, DC, for MS case identification in the US Army; Mona Hedenskog of the Centre for Microbiological Preparedness, Swedish Institute for Infectious Disease and Control, Solna, Sweden for laboratory assistance; and Leslie Unger Department of Nutrition, Harvard School of Public Health, Boston, Mass, for technical assistance. This work was supported by grant NS46635 and NS042194 from the National Institute of Neurological Diseases and Stroke.
Role of the funding source
The sponsor of this study had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; and preparation, review, or approval of the manuscript. Nor did they participate in the decision to submit for publication.
The views expressed are those of the authors and should not be construed to represent the positions of the Department of the Army, Department of the Navy, or Department of Defense.
Conflict of Interest Statement
None of the authors have any conflicts of interest to disclose.