Abstract

Objectives. Elevated levels of serological markers of EBV infection in patients with SLE and observations that infectious mononucleosis (IM) may precede some cases of SLE suggest a possible role of EBV in the aetiology of SLE. We evaluated the relationship between EBV-associated IM and subsequent risk of SLE in a population-based cohort study.

Methods. We followed cohorts of Danes tested serologically for IM using the Paul–Bunnell (PB) heterophile antibody test between 1939 and 1989, and patients hospitalized with IM between 1977 and 2007 for subsequent first hospitalizations with SLE in the period 1977–2008. Standardized incidence ratios (SIRs) with 95% CI served as measures of relative risk.

Results. Risk of SLE was not increased either in individuals with a positive PB test (SIR = 1.1; 95% CI 0.8, 1.6; n = 27) or in individuals hospitalized with IM (SIR = 1.3; 95% CI 0.7, 2.2; n = 12). However, SLE risk in PB-negative individuals was significantly increased (SIR = 2.6; 95% CI 2.1, 3.2; n = 82), a risk that was particularly high 1–4 years after the PB test (SIR = 6.6; 95% CI 3.3, 13.2) and remained significantly elevated for >25 years.

Conclusions. EBV-associated IM does not seem to be a risk factor for SLE. The temporal pattern of increased SLE risk in individuals with a negative PB test suggests that some patients who go on to develop SLE may present with unspecific symptoms, for which they may be tested for IM, long in advance of their SLE diagnosis.

Introduction

SLE is an autoimmune disease characterized by a wide array of clinical manifestations, among which non-specific symptoms such as fatigue, malaise, lymphadenopathy, photosensitivity and arthralgia are more common [1]. More serious disease may manifest itself as widespread organ involvement, such as lupus nephritis or CNS disease.

Observations that infectious mononucleosis (IM) may precede some cases of SLE [2–4] have sparked the hypothesis that EBV might be involved in the aetiology of SLE, since primary EBV infection is the most common cause of IM [5]. That EBV might play a role in the aetiology of SLE has also been supported by findings in several case–control studies, where SLE patients have been found more often to be infected with EBV [6–9] and to have abnormally elevated EBV viral loads in their blood [7, 10, 11] as compared with healthy controls.

SLE is characterized by the production of a variety of autoantibodies, which may appear many years before SLE diagnosis [12]. Experiments have shown that a specific peptide sequence in the Smith (Sm) autoantigen is antigenic in SLE patients [13]. This peptide sequence has been found to be highly homologous to a peptide in EBNA-1 [13]. This might suggest a role for EBV in the aetiology of SLE through cross-reactivity between anti-EBNA-1 antibodies and epitopes of self-antigens that are not normally recognized as foreign by the immune system, mounting an autoimmune response. Immunizations with EBNA-1 fragments in rabbits and mice have shown that it is possible to induce the production of Sm autoantibodies [14, 15]. Furthermore, recently diagnosed IM patients have been found to transiently recognize the same cross-reactive epitope of EBNA-1 [16].

Whether EBV-associated IM plays a role in the aetiology of SLE has only been sparsely investigated. Two case–control studies with altogether 460 cases and 498 controls [17, 18] did not find any evidence of an association. However, both studies were based upon self-reported diagnoses of IM and could therefore have been subject to both underreporting and differential recall between cases and controls. To explore these matters further, we studied the possible association between IM and subsequent risk of SLE in cohorts of individuals tested for IM and patients hospitalized with IM.

Methods

Civil Registration System

The Danish Civil Registration System (CRS) is a nationwide demographic database where information on date of birth, sex, name, address, family relations, vital status and emigration status is continually updated [19]. Since 1 April 1968, all Danish residents have been assigned a unique 10-digit personal identification number, allowing identity-secure linkage of information between registers.

Serological testing for IM

We identified all persons tested serologically for IM at Statens Serum Institut in Copenhagen, Denmark, between 1939 and 1989. The diagnostic method applied was a modification of the Paul–Bunnell (PB) reaction for heterophile antibodies, including agglutination after absorption with Guinea pig kidney cells and ox erythrocytes [20]. The PB test was considered positive if agglutination occurred at titres ≥1 : 32 and negative at titres <1 : 32. Test results coupled with name, date of birth and date of testing were kept on record and computerized manually, as described elsewhere [21–24]. We established two cohorts consisting of persons testing PB-positive or PB-negative. The PB-positive cohort comprised all individuals with recorded PB-positive tests between 1939 and 1989 for whom personal identification numbers could be established [21–24]. The PB-negative cohort comprised a random subset of individuals with records of PB-negative tests in the periods 1946–67 [24] and 1979–89 [23], combined with all individuals who tested negative in the years 1968–70 and 1978 [21], for whom personal identification numbers could be established. To obtain or verify personal identification numbers of the individuals in the cohorts, PB-positive and PB-negative individuals were linked to the CRS using name, sex and date of birth and, after 1968, the CRS identification number written on the test record [21–24]. The first test was cohort defining (PB-positive or PB-negative). If the first PB test was negative, any subsequent occurrence of a positive PB test implied change of cohort. No other cohort switches were allowed, e.g. a negative PB test following a positive PB test would be ignored. Thus, only persons whose first PB test was negative and followed by a subsequent positive PB test would contribute to both PB cohorts.

The National Patient Register

The National Patient Register was established in 1977 and contains information on admissions of patients to hospitals in Denmark [25]. Included are the dates of admission and discharge, diagnosis, department of admittance and hospital. Outpatients were included in 1995. A cohort of hospitalized IM patients (including all inpatients with IM in the period 1977–2007 and all outpatients with IM in the period 1995–2007) was established using International Classification of Diseases, eighth revision (ICD-8) code 075 (from 1977 to 1993) and ICD-10 group B27 (from 1994 to 2007). As in previous large-scale studies on the epidemiology of SLE and other autoimmune diseases [26–28], we also used data in the National Patient Register to identify all hospitalized SLE patients (including both inpatients and patients treated in ambulatory settings) in Denmark between 1977 and 2008 (ICD-8 code 73 419 and ICD-10 group M32).

Data analysis

The PB-positive and PB-negative cohorts and the cohort of hospitalized IM patients were followed from 1 year after the cohort-defining event (PB test or IM hospitalization) or 1 January 1977, whichever came later, until first hospitalization with SLE, death, emigration or 31 December 2008, whichever came first. In all analyses, we excluded the first year after the cohort-defining event to reduce the possible impact of confounding by indication, a type of selection bias that might result from increased testing for IM in patients with pre-clinical SLE. For individuals who first contributed to the PB-negative cohort and subsequently to the PB-positive cohort, follow-up started in the PB-negative cohort 1 year after the negative test and continued in this cohort until the date of the subsequent PB-positive test. One year after the positive PB test, the person entered follow-up in the PB-positive cohort. The standardized incidence ratio (SIR) was used as a measure of relative risk and was calculated as the ratio between observed and expected numbers of first hospitalizations with SLE in the cohort. The expected numbers were estimated as the sum of sex-, age- and period-specific person-years at risk in the cohorts multiplied by corresponding sex-, age- and period-specific first hospitalization rates of SLE derived from the SLE data in the National Patient Register. SIRs were estimated overall and stratified according to sex, age at PB test or IM hospitalization, time since PB test or IM hospitalization and attained age. The 95% CIs were calculated assuming a Poisson distribution of the observed cases of SLE using Wald’s test.

Ethics

The present register-based study required no patient contact. The study was approved by the Danish Data Protection Agency (approval no. 2008-41-2374).

Results

The PB-positive cohort comprised 24 880 persons (46% women) with a positive PB test between 1939 and 1989, while the PB-negative cohort comprised 33 025 persons with a negative PB test between 1946 and 1989 (50% women). The cohort of patients hospitalized with IM between 1977 and 2007 comprised 23 631 patients (45% women; Table 1).

Table 1

Characteristics of study cohorts of persons examined between 1939 and 1989 for IM by means of the PB test and of patients hospitalized with IM between 1977 and 2007

 PB-positive
 
PB-negative
 
IM
 
n Person-years n Person-years n Person-years 
Overall 24 880 680 450 33 025 840 213 23 631 347 224 
Sex       
 Women 11 433 316 923 16 515 422 733 10 714 155 070 
 Men 13 447 363 527 16 510 417 480 12 917 192 154 
Agea, years       
 <10 1271 36 415 4733 138 157 3630 51 560 
 10–19 14 919 412 369 10 299 285 981 13 684 209 993 
 20–29 7207 197 033 8444 224 853 4460 64 001 
 ≥30 1483 34 633 9549 191 222 1857 21 670 
Calendar yearb       
 <1970 11 511 330 290 14 843 396 320 – – 
 1970–79 6356 187 644 11 189 294 372 2131 60 437 
 1980–89 7013 162 516 6993 149 521 6944 156 116 
 ≥1990 – – – – 14 556 130 671 
 PB-positive
 
PB-negative
 
IM
 
n Person-years n Person-years n Person-years 
Overall 24 880 680 450 33 025 840 213 23 631 347 224 
Sex       
 Women 11 433 316 923 16 515 422 733 10 714 155 070 
 Men 13 447 363 527 16 510 417 480 12 917 192 154 
Agea, years       
 <10 1271 36 415 4733 138 157 3630 51 560 
 10–19 14 919 412 369 10 299 285 981 13 684 209 993 
 20–29 7207 197 033 8444 224 853 4460 64 001 
 ≥30 1483 34 633 9549 191 222 1857 21 670 
Calendar yearb       
 <1970 11 511 330 290 14 843 396 320 – – 
 1970–79 6356 187 644 11 189 294 372 2131 60 437 
 1980–89 7013 162 516 6993 149 521 6944 156 116 
 ≥1990 – – – – 14 556 130 671 

aAge at PB test or hospitalization with IM. bCalendar year at PB test or hospitalization with IM.

SLE risk in the PB-positive cohort

Starting 1 year after the test, the PB-positive cohort was followed for a total of 680 450 person-years between 1977 and 2008 (mean follow-up 27.3 years). The median age at positive PB test was 18 years (interquartile range 6 years). A total of 27 cases of SLE were identified during follow-up at a median age of 38 years (range 16–76 years). A positive PB test result was not significantly associated with risk of SLE overall (SIR = 1.1; 95% CI 0.8, 1.6) or for women or men separately (Table 2). For this reason, the remaining analyses for women and men were combined. Neither age at PB test nor time since PB test was significantly related to risk of SLE. An increased risk of SLE with onset before age of 20 years was observed (SIR = 4.9; 95% CI 1.2, 19.8) but this observation was based on only two SLE cases.

Table 2

SIRs of SLE among persons with a positive or negative PB test between 1939 and 1989

 PB-positive
 
PB-negative
 
SLE cases
 
 SLE cases
 
 
Observed Expected SIR (95% CI) Observed Expected SIR (95% CI) 
Overall 27 24.3 1.1 (0.8, 1.6) 82 31.5 2.6 (2.1, 3.2) 
Sex       
 Women 24 20.1 1.2 (0.8, 1.8) 68 26.8 2.5 (2.0, 3.2) 
 Men 4.1 0.7 (0.2, 2.3) 14 4.7 3.0 (1.8, 5.1) 
Age at PB test, years       
 <10 1.2 1.7 (0.4, 6.7) 3.6 2.5 (1.3, 4.8) 
 10–19 18 16.1 1.1 (0.7, 1.8) 23 10.6 2.2 (1.5, 3.3) 
 20–29 5.6 1.1 (0.5, 2.4) 28 8.7 3.2 (2.2, 4.6) 
 ≥30 1.3 0.8 (0.1, 5.4) 22 8.6 2.6 (1.7, 3.9) 
Time since PB test, years       
 1–4 0.7 2.8 (0.7, 11.5) 1.2 6.6 (3.3, 13.2) 
 5–9 1.6 1.9 (0.6, 5.8) 17 2.8 6.1 (3.8, 9.9) 
 10–14 2.6 1.5 (0.6, 4.1) 14 4.2 3.4 (2.0, 5.7) 
 15–19 3.2 0.9 (0.3, 2.9) 12 5.0 2.4 (1.4, 4.3) 
 20–24 3.4 0.6 (0.2, 2.4) 10 4.7 2.1 (1.1, 4.0) 
 ≥25 13 12.7 1.0 (0.6, 1.8) 21 13.6 1.5 (1.00, 2.4) 
Attained age, years       
 <20 0.4 4.9 (1.2, 19.8) 0.4 13.7 (6.1, 30.4) 
 20–29 3.0 1.7 (0.7, 4.0) 2.8 3.2 (1.7, 6.1) 
 30–39 5.8 1.2 (0.6, 2.5) 23 6.7 3.4 (2.3, 5.2) 
 40–49 6.3 0.6 (0.2, 1.7) 17 8.3 2.1 (1.3, 3.3) 
 50–59 5.3 1.0 (0.4, 2.3) 15 7.2 2.1 (1.3, 3.5) 
 60–69 2.5 1.2 (0.4, 3.7) 10 3.8 2.6 (1.4, 4.9) 
 ≥70 1.0 1.0 (0.1, 7.0) 2.2 0.9 (0.2, 3.6) 
 PB-positive
 
PB-negative
 
SLE cases
 
 SLE cases
 
 
Observed Expected SIR (95% CI) Observed Expected SIR (95% CI) 
Overall 27 24.3 1.1 (0.8, 1.6) 82 31.5 2.6 (2.1, 3.2) 
Sex       
 Women 24 20.1 1.2 (0.8, 1.8) 68 26.8 2.5 (2.0, 3.2) 
 Men 4.1 0.7 (0.2, 2.3) 14 4.7 3.0 (1.8, 5.1) 
Age at PB test, years       
 <10 1.2 1.7 (0.4, 6.7) 3.6 2.5 (1.3, 4.8) 
 10–19 18 16.1 1.1 (0.7, 1.8) 23 10.6 2.2 (1.5, 3.3) 
 20–29 5.6 1.1 (0.5, 2.4) 28 8.7 3.2 (2.2, 4.6) 
 ≥30 1.3 0.8 (0.1, 5.4) 22 8.6 2.6 (1.7, 3.9) 
Time since PB test, years       
 1–4 0.7 2.8 (0.7, 11.5) 1.2 6.6 (3.3, 13.2) 
 5–9 1.6 1.9 (0.6, 5.8) 17 2.8 6.1 (3.8, 9.9) 
 10–14 2.6 1.5 (0.6, 4.1) 14 4.2 3.4 (2.0, 5.7) 
 15–19 3.2 0.9 (0.3, 2.9) 12 5.0 2.4 (1.4, 4.3) 
 20–24 3.4 0.6 (0.2, 2.4) 10 4.7 2.1 (1.1, 4.0) 
 ≥25 13 12.7 1.0 (0.6, 1.8) 21 13.6 1.5 (1.00, 2.4) 
Attained age, years       
 <20 0.4 4.9 (1.2, 19.8) 0.4 13.7 (6.1, 30.4) 
 20–29 3.0 1.7 (0.7, 4.0) 2.8 3.2 (1.7, 6.1) 
 30–39 5.8 1.2 (0.6, 2.5) 23 6.7 3.4 (2.3, 5.2) 
 40–49 6.3 0.6 (0.2, 1.7) 17 8.3 2.1 (1.3, 3.3) 
 50–59 5.3 1.0 (0.4, 2.3) 15 7.2 2.1 (1.3, 3.5) 
 60–69 2.5 1.2 (0.4, 3.7) 10 3.8 2.6 (1.4, 4.9) 
 ≥70 1.0 1.0 (0.1, 7.0) 2.2 0.9 (0.2, 3.6) 

SLE risk in the PB-negative cohort

The PB-negative cohort was followed for a total of 840 213 person-years between 1977 and 2008 (mean follow-up 25.4 years). The median age at first negative PB test was 21 years (interquartile range 17 years). A total of 82 cases of SLE were identified during follow-up at a median age of 43 years (range 13–70 years). The 82 observed SLE cases were significantly in excess of the expected (SIR = 2.6; 95% CI 2.1, 3.2; Table 2). The risk of SLE was equally elevated in women and men and was consistently elevated in all test age groups. The risk was markedly elevated in the period 1–4 years after the test (SIR = 6.6; 95% CI 3.3, 13.2), followed by a steady decline in subsequent years, but the risk remained 50% elevated even ≥25 years after the negative PB test (SIR = 1.5; 95% CI 1.00, 2.4). The risk of SLE was elevated in all age groups <70 years and was particularly high for SLE with onset before the age of 20 years (SIR = 13.7; 95% CI 6.1, 30.4).

SLE risk in the cohort of hospitalized patients with IM

The cohort of patients hospitalized with IM between 1977 and 2007 was followed for 347 224 person-years between 1978 and 2008 (mean follow-up 14.7 years). The median age at IM hospitalization was 17 years (interquartile range 6 years). A total of 12 patients were subsequently identified with SLE at a median age of 31 years (range 16–53 years), representing no unusual risk of SLE overall (SIR = 1.3; 95% CI 0.7, 2.2), or in strata of sex, age at IM, time since IM or attained age (Table 3).

Table 3

SIRs of SLE among patients hospitalized with IM between 1977 and 2007

 IM
 
SLE cases
 
 
Observed Expected SIR (95% CI) 
Overall 12 9.6 1.3 (0.7, 2.2) 
Sex    
 Women 8.3 1.1 (0.6, 2.1) 
 Men 1.3 2.4 (0.8, 7.4) 
Age at IM diagnosis, years    
 <20 7.4 0.9 (0.5, 2.0) 
 ≥20 2.1 2.3 (0.97, 5.6) 
Time since IM diagnosis, years    
 1–4 1.6 1.8 (0.6, 5.7) 
 5–9 2.3 0.4 (0.1, 3.2) 
 10–14 2.3 1.8 (0.7, 4.7) 
 ≥15 3.5 1.2 (0.4, 3.1) 
Attained age, years    
 <20 1.1 0.9 (0.1, 6.6) 
 20–29 3.9 1.0 (0.4, 2.7) 
 30–39 3.1 1.6 (0.7, 3.9) 
 40–49 1.1 0.9 (0.1, 6.2) 
 ≥50 0.4 2.7 (0.4, 18.8) 
 IM
 
SLE cases
 
 
Observed Expected SIR (95% CI) 
Overall 12 9.6 1.3 (0.7, 2.2) 
Sex    
 Women 8.3 1.1 (0.6, 2.1) 
 Men 1.3 2.4 (0.8, 7.4) 
Age at IM diagnosis, years    
 <20 7.4 0.9 (0.5, 2.0) 
 ≥20 2.1 2.3 (0.97, 5.6) 
Time since IM diagnosis, years    
 1–4 1.6 1.8 (0.6, 5.7) 
 5–9 2.3 0.4 (0.1, 3.2) 
 10–14 2.3 1.8 (0.7, 4.7) 
 ≥15 3.5 1.2 (0.4, 3.1) 
Attained age, years    
 <20 1.1 0.9 (0.1, 6.6) 
 20–29 3.9 1.0 (0.4, 2.7) 
 30–39 3.1 1.6 (0.7, 3.9) 
 40–49 1.1 0.9 (0.1, 6.2) 
 ≥50 0.4 2.7 (0.4, 18.8) 

Robustness analyses

To ascertain the validity of our findings we performed three robustness analyses. First, to reduce the impact of prevalent cases of SLE that might have been diagnosed before the establishment of the National Patient Register in 1977, we repeated all analyses restricting the follow-up period to 1980–2008. All reported associations in Tables 2 and 3 were similar in these restricted analyses: overall SIR = 1.1; 95% CI 0.8, 1.6 (n = 25 SLE cases) following a positive PB test, overall SIR = 2.3; 95% CI 1.8, 2.9 (n = 68 SLE cases) following a negative PB test and overall SIR = 1.3; 95% CI 0.7, 2.2 (n = 12 SLE cases) following hospitalization with IM. Secondly, to further evaluate the possible impact of prevalent SLE cases, we restricted the PB-negative cohort to include only those persons who were tested in the period 1978–89. We did this specifically to examine whether the observed long-term increase in SLE risk in this cohort was explained by SLE cases in individuals who were tested in the 1940s and 1950s and diagnosed with SLE before the establishment of the National Patient Register in 1977, but whose SLE diagnoses were identified only decades later as a result of hospital contacts in or after 1977. This analysis confirmed the results for the PB-negative cohort presented in Table 2 (overall SIR = 2.8; 95% CI 2.0, 3.8; n = 38 SLE cases) with a markedly elevated risk of SLE 1–4 years after a negative PB test, and a SIR that remained elevated even after ≥25 years of follow-up (SIR = 1.6; 95% CI 0.5, 5.0; n = 3 SLE cases), although SIR estimates were no longer statistically significant beyond the first 15 years of follow-up due to smaller numbers of observations. Thirdly, to increase the diagnostic accuracy of SLE cases in the National Patient Register, we repeated all analyses, this time restricting the outcome of interest to SLE diagnoses in patients with at least two separate hospitalization episodes for SLE during follow-up, using the date of the second SLE hospitalization as the date defining the outcome. In this third robustness analysis, we confirmed the increased overall risk of SLE following a negative PB test (SIR = 3.1; 95% CI 2.4, 4.0, n = 57 SLE cases), whereas SIRs were lower and not statistically significant in individuals with a positive PB test (SIR = 1.4; 95% CI 0.9, 2.2, n = 20 SLE cases) or those who had been hospitalized with IM (SIR = 1.7; 95% CI 0.96, 3.1, n = 11 SLE cases).

Discussion

It has been suggested that EBV infection may play an aetiological role in SLE [29–31]. In the present population-based study, we examined whether EBV-associated IM was associated with an increased risk of subsequent SLE, finding no evidence of such a relationship. We found that IM was unrelated to the risk of SLE both when measured with a positive PB test and when looking at IM diagnosis using hospital registers. The only noteworthy finding in these cohorts was that a positive PB test was associated with an increased risk of SLE with onset before the age of 20 years. However, this finding was based on only two cases of SLE, and there was no unusual risk of SLE with early onset in patients hospitalized with IM. Taken together, it seems unlikely that IM would be a risk factor restricted to the subset of SLE with early onset. We specifically investigated the association between clinical IM diagnosis and risk of SLE using hospital registers as a means to validate any findings in the PB-positive cohort. We assumed that the hospital register cohort would represent more severe cases of IM with a certain diagnosis. It is therefore reassuring that the lack of associations seen in the PB-positive cohort were also observed in the hospital register cohort. To our knowledge, no cohort study apart from ours has investigated the relationship between IM and risk of SLE, but two previous case–control studies have reported a similar lack of association with a history of IM [17, 18]. The consistency of negative findings in these prior studies and the present cohort study strongly suggests that EBV-associated IM is not a risk factor for SLE.

It has been argued that EBV may trigger the development of SLE based on findings that SLE patients are more commonly infected with EBV [6–9], they have increased EBV viral loads in their blood [7, 10, 11] and they have higher EBV antibody titres [6, 9] as compared with healthy controls. However, one previous study found no link between EBV viral load and SLE with onset in adolescence. Based on 13 cases of SLE, the researchers were unable to detect EBV DNA in the patients, even though all patients exhibited a primary or reactivated EBV serological response [32]. The authors suggested that the serological markers of EBV infection reflected immune dysregulation underlying the patient’s SLE or immunosuppressive treatment and thus that EBV may not be an aetiological factor for SLE.

In another study, adolescent SLE patients were more often anti-EBV antibody-positive than controls, but the absence of IgM antibodies against EBV suggested that the SLE patients had not recently been infected with the virus [8]. Our investigation suggests that SLE risk is not associated with primary EBV infection presenting as IM. However, our findings do not exclude the possibility that latent or reactivated EBV infection may drive the autoimmune response of SLE.

Through animal experiments a peptide sequence in EBNA-1 has been found to be similar to a sequence in the Sm autoantigen [13], which is a target for Sm autoantibodies in SLE patients [33]. It has been shown that this particular peptide sequence in EBNA-1 may induce the production of autoantibodies against Sm autoantigen in experimental animals and give rise to SLE-like presentations in the form of thrombocytopenia, seizures, proteinuria, leucopenia and lymphopenia [14, 15]. The same peptide sequence also seems to be antigenic in IM patients [16]. Assuming that the specific peptide sequence induces the production of autoantibodies in EBV-infected individuals one would expect IM to be related to risk of SLE, an expectation that our findings do not support.

Whereas a positive PB test was unrelated to risk of SLE, we found that a negative PB test was significantly related to an increased risk of SLE. To our knowledge, no other study has investigated whether a negative test result for IM is related to SLE risk. It is possible that in some cohort members, notably those with a short interval between the negative PB test and the first hospitalization with SLE, the test may have been taken after the disease process leading to SLE had started, but before SLE was diagnosed. To reduce the impact of such pre-clinical SLE cases, we started follow-up 1 year after the PB test. The relative risk of SLE was high for 1–4 years after a negative PB test and declined gradually thereafter, but it remained significantly elevated for >25 years after the test. One possible explanation might be that some individuals who went on to develop SLE had unspecific symptoms with fever or lymphadenopathy [34], long in advance of their SLE diagnosis and some of these patients may have been tested for IM in search for an explanation for these symptoms. In Denmark, it has previously been estimated that initial manifestations of SLE precede SLE diagnosis by, on average, ∼6 years [35]. Our finding that the risk of SLE remained elevated for more than two decades after a negative PB test suggests that some SLE patients may have unspecific symptoms many years before development of the clinically recognized debut symptoms. In a robustness analysis, we addressed the theoretical concern that our findings for the PB-negative cohort might be due, at least in part, to delayed recording of SLE diagnoses, since the National Patient Register was established in 1977, while PB testing was begun decades before. Although numbers of SLE cases were considerably smaller in this robustness analysis of individuals testing PB-negative in the period 1978–89, the same pattern of increased long-term risk after a negative PB test was observed, although the elevated SIR estimates only retained statistical significance in the first 15 years after the negative PB test.

The risk of SLE with onset before age 20 years was >13-fold increased in individuals with a negative PB test. A likely explanation is that the incidence of IM is much higher than that of SLE in teenagers and young adults [5]. Consequently, pediatricians and family physicians may well suspect IM and perform diagnostic tests for IM in patients who present with unspecific initial symptoms of SLE, thus producing the observed pattern of increased risk of SLE with onset before age 20 years in patients suspected of, but not found to have, IM.

Our population-based cohort study relied upon the unique personal identification numbers assigned to all Danish citizens, thus enabling the ascertainment of both exposures and outcomes through administrative and health registers. Contributing to the validity of our study is that our data were collected in an unbiased manner and thus our study is free of possible selection and recall problems that might have biased findings in case–control studies. The size of our cohorts and the long periods of follow-up are further advantages and robustness analyses were performed to ascertain the validity of our findings. Nevertheless, our study may well have been subject to some level of misclassification of PB test results. The PB test has high specificity but relatively low sensitivity [36, 37] such that some false-negative results may have occurred, especially in the first week of illness [38] and in IM patients <13 years of age [39]. Depending on the extent of such misclassification, any true differences in SLE risk between patients with and without IM would appear smaller than they actually were. Consequently, our contrasting findings of increased risk of SLE in individuals testing PB negative and no unusual risk of SLE among PB-positive individuals occurred in spite of forces of misclassification that would reduce such differences. Another limitation of our study is that we were able to capture only those cases of SLE that were severe enough to require hospital contact. However, other researchers have shown that most Danish SLE patients are in contact with hospital departments, either as inpatients or as outpatients in ambulatory settings, in connection with either diagnosis or treatment [35], so our study most likely identified the majority of patients with SLE in Denmark during the period 1977–2008. Nevertheless, our findings may not necessarily apply to milder cases of SLE treated exclusively in specialist settings outside Danish hospitals.

In conclusion, EBV infection that manifests serologically or clinically as IM seems not to be a risk factor for SLE. Instead, we found that patients who were tested for but who did not have serological hallmarks of IM were at an increased risk of SLE for decades after the negative test. This suggests that some SLE patients may be unwell for prolonged periods before SLE diagnosis, for which reason they may be tested for IM before they subsequently become diagnosed with SLE.

graphic

Acknowledgements

Funding: This work was supported by unrestricted research grants from Aase and Ejnar Danielsen's Foundation, Civil Engineer Frode V. Nyegaard and Wife’s Foundation, Aage and Johanne Louis-Hansen’s Foundation, Max Fodgaard’s Foundation, Torben and Alice Frimodt’s Foundation and the Augustinus Foundation.

Disclosure statement: The authors have declared no conflicts of interest.

References

1
D'Cruz
DP
Systemic lupus erythematosus
Br Med J
 , 
2006
, vol. 
332
 (pg. 
890
-
4
)
2
Dror
Y
Blachar
Y
Cohen
P
Livni
N
Rosenmann
E
Ashkenazi
A
Systemic lupus erythematosus associated with acute Epstein-Barr virus infection
Am J Kidney Dis
 , 
1998
, vol. 
32
 (pg. 
825
-
8
)
3
Verdolini
R
Bugatti
L
Giangiacomi
M
Nicolini
M
Filosa
G
Cerio
R
Systemic lupus erythematosus induced by Epstein-Barr virus infection
Br J Dermatol
 , 
2002
, vol. 
146
 (pg. 
877
-
81
)
4
Bhimma
R
Adhikari
M
Coovadia
HM
Epstein-Barr virus-induced systemic lupus erythematosus
S Afr Med J
 , 
1995
, vol. 
85
 (pg. 
899
-
900
)
5
Evans
AS
Clinical syndromes associated with EB virus infection
Adv Intern Med
 , 
1972
, vol. 
18
 (pg. 
77
-
93
)
6
Rothfield
NF
Evans
AS
Niederman
JC
Clinical and laboratory aspects of raised virus antibody titres in systemic lupus erythematosus
Ann Rheum Dis
 , 
1973
, vol. 
32
 (pg. 
238
-
46
)
7
Lu
JJ
Chen
DY
Hsieh
CW
Lan
JL
Lin
FJ
Lin
SH
Association of Epstein-Barr virus infection with systemic lupus erythematosus in Taiwan
Lupus
 , 
2007
, vol. 
16
 (pg. 
168
-
75
)
8
James
JA
Kaufman
KM
Farris
AD
Taylor-Albert
E
Lehman
TJ
Harley
JB
An increased prevalence of Epstein-Barr virus infection in young patients suggests a possible etiology for systemic lupus erythematosus
J Clin Invest
 , 
1997
, vol. 
100
 (pg. 
3019
-
26
)
9
James
JA
Neas
BR
Moser
KL
, et al.  . 
Systemic lupus erythematosus in adults is associated with previous Epstein-Barr virus exposure
Arthritis Rheum
 , 
2001
, vol. 
44
 (pg. 
1122
-
6
)
10
Kang
I
Quan
T
Nolasco
H
, et al.  . 
Defective control of latent Epstein-Barr virus infection in systemic lupus erythematosus
J Immunol
 , 
2004
, vol. 
172
 (pg. 
1287
-
94
)
11
Moon
UY
Park
SJ
Oh
ST
, et al.  . 
Patients with systemic lupus erythematosus have abnormally elevated Epstein-Barr virus load in blood
Arthritis Res Ther
 , 
2004
, vol. 
6
 (pg. 
R295
-
302
)
12
Arbuckle
MR
McClain
MT
Rubertone
MV
, et al.  . 
Development of autoantibodies before the clinical onset of systemic lupus erythematosus
N Engl J Med
 , 
2003
, vol. 
349
 (pg. 
1526
-
33
)
13
James
JA
Harley
JB
Linear epitope mapping of an Sm B/B′ polypeptide
J Immunol
 , 
1992
, vol. 
148
 (pg. 
2074
-
9
)
14
James
JA
Gross
T
Scofield
RH
Harley
JB
Immunoglobulin epitope spreading and autoimmune disease after peptide immunization: Sm B/B′-derived PPPGMRPP and PPPGIRGP induce spliceosome autoimmunity
J Exp Med
 , 
1995
, vol. 
181
 (pg. 
453
-
61
)
15
Poole
BD
Gross
T
Maier
S
Harley
JB
James
JA
Lupus-like autoantibody development in rabbits and mice after immunization with EBNA-1 fragments
J Autoimmun
 , 
2008
, vol. 
31
 (pg. 
362
-
71
)
16
McClain
MT
Rapp
EC
Harley
JB
James
JA
Infectious mononucleosis patients temporarily recognize a unique, cross-reactive epitope of Epstein-Barr virus nuclear antigen-1
J Med Virol
 , 
2003
, vol. 
70
 (pg. 
253
-
7
)
17
Strom
BL
Reidenberg
MM
West
S
Snyder
ES
Freundlich
B
Stolley
PD
Shingles, allergies, family medical history, oral contraceptives, and other potential risk factors for systemic lupus erythematosus
Am J Epidemiol
 , 
1994
, vol. 
140
 (pg. 
632
-
42
)
18
Cooper
GS
Dooley
MA
Treadwell
EL
St Clair
EW
Gilkeson
GS
Risk factors for development of systemic lupus erythematosus: allergies, infections, and family history
J Clin Epidemiol
 , 
2002
, vol. 
55
 (pg. 
982
-
9
)
19
Pedersen
CB
Gøtzsche
H
Møller
JO
Mortensen
PB
The Danish Civil Registration System. A cohort of eight million persons
Dan Med Bull
 , 
2006
, vol. 
53
 (pg. 
441
-
9
)
20
Kristensen
M
Studies on the sero-diagnosis of mononucleosis infectiosa
Acta Pathol Microbiol Immunol Scand
 , 
1938
, vol. 
15(Suppl. 37)
 (pg. 
339
-
56
)
21
Haahr
S
Koch-Henriksen
N
Møller-Larsen
A
Eriksen
LS
Andersen
HM
Increased risk of multiple sclerosis after late Epstein-Barr virus infection: a historical prospective study
Mult Scler
 , 
1995
, vol. 
1
 (pg. 
73
-
7
)
22
Rosdahl
N
Larsen
SO
Clemmesen
J
Hodgkin's disease in patients with previous infectious mononucleosis: 30 years’ experience
Br Med J
 , 
1974
, vol. 
2
 (pg. 
253
-
6
)
23
Nielsen
TR
Rostgaard
K
Nielsen
NM
, et al.  . 
Multiple sclerosis after infectious mononucleosis
Arch Neurol
 , 
2007
, vol. 
64
 (pg. 
72
-
5
)
24
Hjalgrim
H
Askling
J
Rostgaard
K
, et al.  . 
Characteristics of Hodgkin’s lymphoma after infectious mononucleosis
N Engl J Med
 , 
2003
, vol. 
349
 (pg. 
1324
-
32
)
25
Andersen
TF
Madsen
M
Jørgensen
J
Mellemkjær
L
Olsen
JH
The Danish National Hospital Register. A valuable source of data for modern health sciences
Dan Med Bull
 , 
1999
, vol. 
46
 (pg. 
263
-
8
)
26
Ulff-Møller
CJ
Jørgensen
KT
Pedersen
BV
Nielsen
NM
Frisch
M
Reproductive factors and risk of systemic lupus erythematosus: nationwide cohort study in Denmark
J Rheumatol
 , 
2009
, vol. 
36
 (pg. 
1903
-
9
)
27
Nielsen
NM
Frisch
M
Rostgaard
K
, et al.  . 
Autoimmune diseases in patients with multiple sclerosis and their first-degree relatives
Mult Scler
 , 
2008
, vol. 
14
 (pg. 
823
-
9
)
28
Jørgensen
KT
Rostgaard
K
Bache
I
, et al.  . 
Autoimmune diseases in women with Turner syndrome
Arthritis Rheum
 , 
2010
, vol. 
62
 (pg. 
658
-
66
)
29
Niller
HH
Wolf
H
Minarovits
J
Regulation and dysregulation of Epstein-Barr virus latency: implications for the development of autoimmune diseases
Autoimmunity
 , 
2008
, vol. 
41
 (pg. 
298
-
328
)
30
Barzilai
O
Ram
M
Shoenfeld
Y
Viral infection can induce the production of autoantibodies
Curr Opin Rheumatol
 , 
2007
, vol. 
19
 (pg. 
636
-
43
)
31
Zandman-Goddard
G
Shoenfeld
Y
Infections and SLE
Autoimmunity
 , 
2005
, vol. 
38
 (pg. 
473
-
85
)
32
Katz
BZ
Salimi
B
Kim
S
Nsiah-Kumi
P
Wagner-Weiner
L
Epstein-Barr virus burden in adolescents with systemic lupus erythematosus
Pediatr Infect Dis J
 , 
2001
, vol. 
20
 (pg. 
148
-
53
)
33
Rahman
A
Isenberg
DA
Systemic lupus erythematosus
N Engl J Med
 , 
2008
, vol. 
358
 (pg. 
929
-
39
)
34
Hurt
C
Tammaro
D
Diagnostic evaluation of mononucleosis-like illnesses
Am J Med
 , 
2007
, vol. 
120
 (pg. 
911
-
8
)
35
Voss
A
Green
A
Junker
P
Systemic lupus erythematosus in Denmark: clinical and epidemiological characterization of a county-based cohort
Scand J Rheumatol
 , 
1998
, vol. 
27
 (pg. 
98
-
105
)
36
Evans
AS
Niederman
JC
Cenabre
LC
West
B
Richards
VA
A prospective evaluation of heterophile and Epstein-Barr virus-specific IgM antibody tests in clinical and subclinical infectious mononucleosis: specificity and sensitivity of the tests and persistence of antibody
J Infect Dis
 , 
1975
, vol. 
132
 (pg. 
546
-
54
)
37
Bruu
AL
Hjetland
R
Holter
E
, et al.  . 
Evaluation of 12 commercial tests for detection of Epstein-Barr virus-specific and heterophile antibodies
Clin Diagn Lab Immunol
 , 
2000
, vol. 
7
 (pg. 
451
-
6
)
38
Hoagland
RJ
Infectious mononucleosis
Prim Care
 , 
1975
, vol. 
2
 (pg. 
295
-
307
)
39
Linderholm
M
Boman
J
Juto
P
Linde
A
Comparative evaluation of nine kits for rapid diagnosis of infectious mononucleosis and Epstein-Barr virus-specific serology
J Clin Microbiol
 , 
1994
, vol. 
32
 (pg. 
259
-
61
)

Comments

0 Comments