SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Anti-double stranded (dsDNA) antibodies are of considerable diagnostic value and are thought to be involved in the pathogenesis of systemic lupus erythematosus (SLE). Fluctuations in anti-dsDNA antibody levels are also used as markers for disease activity and exacerbations. In this study we sought to evaluate the anti-dsDNA antibody level in serum samples collected before the onset of SLE diagnosis. A total of 130 SLE patients were identified with stored serum samples available prior to diagnosis within the US Department of Defense serum repository. All 633 sera available from these patients were screened for anti-dsDNA antibodies using an enzyme linked immunosorbant assay (ELISA). Within this cohort 55% of cases had detectable anti-dsDNA antibodies prior to SLE diagnosis. The onset of anti-dsDNA antibodies ranged from 9.3 years before to within the same month as diagnosis (with a mean onset 2.7 years before diagnosis). In order to assess for fluctuations in anti-dsDNA levels relative to diagnosis, cases were selected with at least two positive samples, one within 6 months and a second greater than 6 months prior to diagnosis (n = 26). Seven of these cases also had samples available shortly after diagnosis (≤ 6 months) for comparison. Fifty-eight percent of the 26 cases developed a significant rise in anti-dsDNA antibody levels within 6 months of diagnosis. A significant decline in anti-dsDNA levels ensued after diagnosis (and following treatment with corticosteroids) in all seven cases with samples available. Patients with a significant rise in anti-dsDNA antibodies at diagnosis were more likely to have renal disease than those who did not (66.7% compared to 27.3%, χ2=3.94, P<0.05). These data suggest that anti-dsDNA antibodies are present in SLE patient sera much earlier than previously suspected. In addition, the data are consistent with increases in anti-dsDNA levels contributing to the onset of clinical illness in some patients with SLE.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

SLE is a rheumatic disease characterized by varied clinical manifestations and the production of high titers of autoantibodies. Antibodies to dsDNA are observed in 60–70% of lupus patients [1]. Anti-dsDNA antibodies rarely occur in patients with diseases other than SLE. Thus, they contribute substantial discriminating capacity to the diagnostic effort and are included among the classification criteria for SLE [1, 2]. Several studies have demonstrated that in many patients anti-dsDNA antibodies fluctuate in parallel with clinical disease activity [3–7]. In fact, high-avidity immunoglobulin (Ig)G antibodies to dsDNA are believed to play a major role in inducing some of the disease manifestations of SLE. Anti-dsDNA antibodies have been eluted from skin [8] and disease affected kidney samples [9] from patients with SLE. Furthermore, direct evidence for the pathogenic role of anti-dsDNA antibodies in nephritis has been demonstrated through the induction of proteinuria in normal mice following transfer of monoclonal anti-DNA antibodies [10]. On the other hand, a substantial fraction of patients with lupus nephritis do not have detectable anti-dsDNA antibodies in their sera, suggesting that other mechanisms may also operate to produce lupus nephritis.

While antibodies against DNA in lupus patient sera were first described over 40 years ago [11], serologic changes occurring prior to SLE disease onset have remained uncharacterized. Because SLE patients are usually identified after autoimmunity has already commenced, the possibility of analyzing the early autoimmune response and subsequent changes in the progression to disease onset has been virtually nonexistent. Previous studies regarding the presence of anti-dsDNA autoantibodies prior to the SLE diagnosis have been limited to the follow-up of patients who do not meet full diagnostic criteria for SLE, but warrant evaluation with autoantibody testing [12].

The relatively recent development of a large prospective serum collection, such as the US Department of Defense serum repository, has provided us with a record of immunological history and has given us a unique opportunity to view changes in the immune system before clinical illness. The results demonstrate that anti-dsDNA antibodies frequently precede clinical diagnosis by years. In addition, rises in anti-dsDNA antibodies often accompany SLE diagnosis.

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Serum samples The Department of Defense serum repository contains nearly 20,000,000 serum samples from active duty personnel of the US military collected at entry and on average every other year thereafter since 1988 and stored at −80 oC. Potential SLE cases were identified through searching military inpatient and outpatient databases for records containing the ICD-9 code for SLE [710.0]. Identified records were obtained and reviewed for criteria for SLE classification [1, 2]. In this study 130 individuals were identified having fulfilled the SLE classification criteria as well as having at least one retrievable serum sample either at, or before, diagnosis. For each case, one control was randomly selected (from four controls matched for gender, ethnicity, age, length of military service and availability of stored serum samples). Six hundred and thirty-three serum samples were retrieved from 130 patients, in addition to 130 serum samples from controls matched for age, gender ethnicity, length of military service and availability of sera in the US Department of Defense serum repository.

Clinical chart review Information regarding cumulative clinical and laboratory features for each case was obtained by chart review. Each clinical and serological feature included among the revised ACR criteria [1, 2] was recorded as present or absent for each patient. In addition, the time first observed was noted separately for each feature. The following demographic information was obtained: gender, ethnicity, date of diagnosis and age at diagnosis (having fulfilled four or more of the revised ACR criteria for the classification of SLE).

Autoantibody assays ELISA using recombinant plasmid dsDNA (Varelisa, Pharmacia and Upjohn Diagnostics, Freiburg, Germany) were used to evaluate sera for antibodies to dsDNA. Technical sensitivity and specificity of this assay for anti-dsDNA antibodies have been reported to be 100% in a recent independent study by Tan et al. [13]. All equivocal values were repeated. Samples that were equivocal upon repeat testing were classified as negative. Anti-dsDNA antibody levels are presented as IU/mL. (Anti-dsDNA antibodies were evaluated in all 633 SLE patient sera. Of the 130 control sera, four were positive in this assay.)

Statistical analysis In order to assess any increase in anti-dsDNA titer prior to diagnosis, cases were selected with at least two positive samples, one within 6 months and a second greater than 6 months prior to diagnosis. In the case where one time interval had more than one positive sample, those values closest to diagnosis were selected for analysis. Consistent with other reports [6, 14], changes in anti-dsDNA antibody levels ≥ 25% were considered significant. Wilcoxon signed rank test, chi-square, and Fisher's Exact test were used for analyzing categorical differences. Fisher's Exact test was used when the assumptions of the χ2 were violated owing to small sample size. All P values were two-tailed assuming equal variance, and P values <0.05 were considered to be statistically significant. P-values were not corrected for multiple comparisons. The odds ratio (OR) was calculated with 95% confidence intervals (CI). Results are presented as mean ± standard deviation unless indicated otherwise.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Patient population and serum samples

Medical charts were requested for 337 individuals identified with a potential diagnosis of SLE (based upon the ICD-9 code 710.0) and having at least one serum sample reported available in the US Department of Defense serum repository. Medical records were not available for 73 of these individuals. Of the remaining 264 individual cases reviewed, reasons for exclusion from the study included the presence of less than four SLE classification criteria or insufficient data available to confirm diagnosis (n = 101), available serum samples were obtained only after diagnosis (n = 32), or the reported available serum samples were not retrievable (n = 1). One hundred and thirty individuals were identified who met the inclusion criteria, having been diagnosed with SLE while serving on active duty in the military, and having at least one retrievable serum sample prior to disease diagnosis (Table 1). For each case, one control was randomly selected from among those matched for gender, ethnicity, age, length of military service and availability of stored serum samples.

Table 1.   Demographics of patient population and serum samples available
Patient population (n = 130)
Gender, Female (%) Ethnicity    64.6 % 
% African American61.5 % 
% European American26.2 % 
% Hispanic9.2 % 
% Asian3.1 % 
Age (mean ± SD)
at initial SLE Criterion    28.8 ± 6.5(range = 17.5–45.2)
at SLE Diagnosis    30.4 ± 6.8(range = 18.5–46.9)
Serum Samples (n = 633)
Average number (± SD) per Cases4.9 ± 2.5 (range = 1–12)
Time range of serum samples (relative to diagnosis)-9.4 to + 7.0 years
Average years (± SD) before diagnosis of first sample-4.4 ± 2.5 years

A total of 633 serum samples were retrieved from the 130 SLE cases identified, with an average of 4.9 ± 2.5 serum samples (ranging from 1 to 12) available per case. The earliest serum sample available for each individual case was on average 4.4 ± 2.5 years prior to diagnosis (ranging from 9.43 years prior to diagnosis to within the same month as diagnosis). While all patients were required to have at least one sample available prior to diagnosis, 59% also had serum samples available after diagnosis (up to 7 years later).

Among the 130 patients, 114 individuals also had sera available prior to the documentation of any clinical feature included among SLE classification criteria.

Anti-dsDNA antibodies

Anti-dsDNA antibodies were detected in 80 cases (62%) and in four controls (3%). There was no statistical difference in the prevalence of anti-dsDNA antibodies between different ethnic groups, although the frequency was slightly higher in African-Americans (65% compared to 56% of European-Americans). Cumulative clinical criteria among patients with anti-dsDNA antibodies were compared with those among patients that did not develop anti-dsDNA antibodies (Table 2). Patients with anti-dsDNA antibodies were less likely to develop photosensitivity [26% versus 46%, χ2=5.36, P = 0.021, OR = 0.42 (0.20–0.89)] and were more likely to develop renal disease [48% versus 30%, χ2=3.9, P = 0.048, OR = 2.1 (1.00–4.44)]. In addition, patients with anti-dsDNA antibodies were more likely to have immunological features documented in their clinical record (including one or more of the following: biological false positive VDRL, lupus anticoagulant, IgG or IgM anti-cardiolipin antibodies, anti-dsDNA antibodies, and/or anti-Sm antibodies) [86% versus 58%, χ2=13.2, P < 0.001, OR = 4.5 (2.00–10.30)]. Of the 80 cases, eight had anti-dsDNA first detected in stored sera after diagnosis. Consequently, the focus of this study is on the 72 cases that had anti-dsDNA before the diagnosis was established.

Table 2.   Presence of clinical critiera among patients with and without anti-dsDNA antibodies 
 Anti-dsDNA (+)Anti-dsDNA (-)    
 n = 80(%)n = 50(%)O.R.(95% C.I.)χ2P-value
  1.  *Immunologic Disorder included biological false positive VDRL, lupus anticoagulant, IgG or IgM anti-cardiolipin antibodies, anti-dsDNA antibodies, and/or anti-Sm antibodies documented in the medial record. Results obtained from stored sera were not included.

Malar rash17(21.3)15(30.0)    
Discoid rash10(12.5)9(18.0)    
Photosensitivity21(26.3)23(46.0)0.42(0.20–0.89)5.40.021
Oral ulcers27(33.8)11(22.0)    
Arthritis58(72.5)32(64.0)    
Serositis37(46.3)17(34.0)    
Renal disease38(47.5)15(30.0)2.1(1.00–4.44)3.90.048
CNS disease5(6.3)1(2.0)    
Hematologic disorder61(76.3)35(70.0)    
Immunologic disorder*69(86.3)29(58.0)4.5(2.00–10.30)13.2<0.001
Antinuclear antibodies80(100.0)50(100.0)    

Anti-dsDNA onset relative to diagnosis

The onset of anti-dsDNA antibodies was determined relative to SLE diagnosis (meeting ≥ 4 ACR Criteria). Seventy-two cases (55% of 130) had anti-dsDNA antibodies detected prior to SLE diagnosis (Fig. 1). Among these cases, anti-dsDNA antibodies were first detected preceding diagnosis on average 2.7 years earlier (ranging 9.3 years before SLE diagnosis to within the same month). In addition 38 out of 72 cases (52.8%) were positive in the first (or only) available serum sample, indicating that onset of these antibodies is likely earlier than can be detected with available serum samples.

image

Figure 1.  Detection of anti-dsDNA antibodies prior to SLE diagnosis. Seventy-two cases (55%) had serum samples positive for anti-dsDNA antibodies prior to diagnosis. Each individual case is represented as a bar on the Y-axis. The date at which each case was first determined to be positive is shown as black bars. Thirty-eight cases (53% of positive samples) were positive in the first available serum sample. Grey bars represent the time period from which patients were last determined to be negative, to the time at which they were first determined to be positive. Although patients could have become positive any time during this time period, the times first known to be positive were used in all calculations. Earlier negative samples are not depicted in the figure. The onset of anti-dsDNA antibodies prior to diagnosis ranged from 9.3 years before to within the same month as SLE diagnosis (with a mean onset at 2.7 years).

Download figure to PowerPoint

Cases with serum samples available prior to any of the clinical features included among the SLE classification criteria (n = 114) were further analyzed. The development of anti-dsDNA antibodies occurred prior to the onset of SLE in 74.2% of anti-dsDNA positive cases. Anti-dsDNA antibodies preceded the onset of SLE classification criteria on average by 2.3 years and were found as early as seven years before presenting symptom(s). Eight patients first developed anti-dsDNA antibodies after diagnosis. They are not the focus of this report.

Fluctuations in Anti-dsDNA antibodies relative to diagnosis

Fluctuations in anti-dsDNA antibody levels were observed over time among anti-dsDNA positive cases. In several cases, a significant rise in anti-dsDNA levels directly preceded SLE diagnosis followed by a sharp decline. Data from three such patients are presented (Fig. 2). In order to determine if these trends were common, cases with at least two positive samples (one within 6 months of diagnosis, and one earlier than 6 months before diagnosis) were evaluated (n = 26, Table 3). The average time interval between compared samples was 1.5 years (ranging from 6.4 to 0.2 years). A significant proportion of the patients demonstrated an increase in their anti-dsDNA level within 6 months of diagnosis (73%, P = 0.003, Wilcoxon's signed rank test). For 15 of the cases (58%) the rise in anti-dsDNA levels was substantial (ranging from a 1.57-fold to 40-fold increase in antibody levels). The mean (± SEM) anti-dsDNA level within 6 months of diagnosis was 743 ± 212 units while the value beyond 6 months prior to diagnosis was 227 ± 37 units, yielding a three-fold increase in antibody levels.

image

Figure 2.  Fluctuations in anti-dsDNA antibody levels relative to the time of SLE diagnosis. (A) The patient (African-American female) presented to the clinician with arthritis, fever of unknown origin, chronic fatigue and symptoms attributed to Raynaud's phenomenon. Approximately 4 months later she was diagnosed with SLE based upon oral ulcers, arthritis, and abnormal serology. Analysis of stored sera samples revealed that anti-dsDNA antibodies had preceded symptoms by more than 1 year; however, a significant increase in the anti-dsDNA antibody level was found at the time of diagnosis. The patient was subsequently treated with nonsteroidal anti-inflammatory agents, prednisone and hydroxychloroquine. (B) The patient (African-American male) presented with a 3-month history of fatigue and weight loss and unexplained fevers. He was diagnosed with SLE based upon the presence of arthritis, diffuse proliferative lupus nephritis, leukopenia, lymphopenia, and serological abnormalities. Analysis of stored serum samples revealed anti-dsDNA antibodies preceded symptoms by more than 2 years; however, a significant increase in the anti-dsDNA antibody level was found directly at the time of diagnosis. The patient was subsequently treated with prednisone and intravenous cyclophosphamide. (C) The patient (African-American female) presented with pleurisy in 1989 at the age of 20 years. At the age of 28 years she presented again with pleuritic chest pain accompanied by arthritis and was diagnosed with SLE based upon the presence of arthritis, pleuropericarditis, lymphopenia, and abnormal serology. Analysis of stored sera samples revealed that anti-dsDNA antibodies had preceded diagnosis by 1.5 years; however, a significant increase in the level of anti-dsDNA antibodies was found at the time of diagnosis. The patient was subsequently treated with nonsteroidal anti-inflammatory agents, prednisone and hydroxychloroquine.

Download figure to PowerPoint

Table 3.   Comparison of anti-dsDNA antibody levels prior to SLE diagnosis among cases with multiple positive samples
 Sample 1Sample 2 
CaseIU/mL(Time)IU/mL(Time)Ratio
  1.  Time is presented in years prior to diagnosis. Sample 1 was collected greater than 6 months before diagnosis. Sample 2 was collected within 6 months before diagnosis. Anti-dsDNA levels for all available serum samples from cases 23, 24 and 25 are shown in Figure 2 (A–C, respectively).

1633(-1.11)236(-0.32)0.37
2143(-0.55)94(-0.26)0.65
3484(-2.34)329(-0.49)0.68
4803(-2.27)637(-0.02)0.79
5100(-0.91)84(-0.09)0.85
686(-0.52)80(-0.28)0.93
7233(-1.30)217(-0.06)0.93
8115(-6.53)121(-0.10)1.05
9101(-1.43)107(-0.43)1.06
10126(-1.54)135(0.00)1.07
11237(-2.13)281(-0.41)1.18
12209(-2.50)328(-0.27)1.57
13343(-1.75)549(-0.13)1.60
14425(-0.55)687(-0.18)1.62
15269(-1.45)442(-0.04)1.64
1667(-4.15)123(-0.08)1.82
1760(-0.86)133(-0.23)2.23
18269(-0.97)810(-0.24)3.01
19117(-0.92)410(-0.15)3.50
2066(-0.52)330(-0.06)4.98
21109(-0.94)633(-0.06)5.79
22127(-0.71)769(-0.16)6.07
23249(-1.75)2280(-0.14)9.16
24297(-0.69)3490(-0.05)11.75
25131(-1.44)1741(-0.02)13.29
26107(-2.57)4283(-0.03)40.02

Seven of these 26 cases also had samples available shortly after diagnosis (within 6 months) and were evaluated for additional changes in antibody levels following diagnosis (Table 4). There was on average a 0.4-year interval between samples (ranging from 0.2 to 0.7 years). All seven cases demonstrated a significant drop in anti-dsDNA levels following diagnosis (ranging from 47% to 98%). Each of these patients received corticosteroids during the sample interval.

Table 4.   Anti-dsDNA antibody levels compared from before and after SLE diagnosis
 Pre-diagnosisPost-diagnosisPercent
CaseIU/mL(Time)IU/mL(Time)ReductionTreatment during interval
  1.  Time is presented in years relative to diagnosis.

8121(-0.10)33(0.18)72.7%Prednisone 10mg QD
16123(-0.08)66(0.29)46.5%Hydroxychloroquine 200 mg BID Prednisone 20 mg QD
17133(-0.23)41(0.39)69.2%Prednisone 10mg QD Hydroxychloroquine 200mg BID
19410(-0.15)160(0.05)61.0%Azathioprine 75mg QD Prednisone 20 mg QD Hydroxychloroquine 200 mg QD
20330(-0.06)168(0.38)49.1%Solumedrol 100mg IV BID Prednisone 20mg BID Methotrexate NSAIDs
22769(-0.16)30(0.49)96.1%Prednisone 10mg QD
232280(-0.14)38(0.37)98.3%Prednisone 60 mg QD Hydroxychloroquine 200mg QD NSAIDs

Clinical criteria were compared between individuals that demonstrated a significant increase (≥25%) in anti-dsDNA antibodies with those anti-dsDNA positive cases who did not (Table 5). Cases with a significant rise in anti-dsDNA levels were more likely to have lupus nephritis [67% versus 27%, χ2=3.9, P = 0.047, OR = 5.3 (1.0–28.8)].

Table 5.   Presence of clinical critiera among patients demonstrating a signficant rise in anti-dsDNA levels at diagnosis compared with those anti-dsDNA (+) patients who did not
 Significant rise in anti-dsDNA levels at diagnosis    
 PresentAbsent    
 n = 15(%)n = 11(%)O.R.(95% C.I.)χ2P-value
  1.  *Immunologic disorder included biological false positive VDRL, lupus anticoagulant, IgG or IgM anticardiolipin antibodies, anti-dsDNA antibodies, and/or anti-Sm antibodies documented in the medial record. Results obtained from stored sera were not included.

Malar rash1(6.7)1(9.1)    
Discoid rash2(13.3)1(9.1)    
Photosensitivity1(6.7)3(27.3)    
Oral ulcers4(26.7)2(18.2)    
Arthritis12(80.0)9(81.8)    
Serositis8(53.3)6(54.5)    
Renal disease10(66.7)3(27.3)5.3(0.99–28.78)3.90.047
CNS disease0(0)1(9.1)    
Hematologic disorder13(86.7)8(72.7)    
Immunologic disorder*15(100)9(81.8)    
Antinuclear antibodies15(100)11(100)    

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

In this study we have combined the power of a large prospective serum collection within the US Department of Defense with retrospective chart review in order to evaluate the onset of anti-dsDNA antibodies prior to SLE onset and diagnosis. The cumulative prevalence of anti-dsDNA antibodies detected in stored serum samples in this study (62%) is similar to the prevalence reported in studies where only data after diagnosis are available [1].

As expected, patients with anti-dsDNA antibodies were more likely to have immunological features characteristic of lupus documented in the medial record [86% versus 58%, χ2=13.2, P<0.001, OR = 4.5 (2.00–10.30)]. Although all patients with detectable anti-dsDNA antibodies would have met immunological criteria as a result of information obtained from stored sera, immunological data were not obtained at the time of clinical evaluation in all cases. The various assays for anti-dsDNA used by clinical sites in the military may also explain immunological SLE criterion not being recorded in the medical record. Some suspect that the ELISA method may have a higher sensitivity than some other methods [15–17]. Assay methods noted during chart review varied by the location of the clinic and included Crithidia luciliae, ELISA, or Farr assay, or was not specified.

The results of this study demonstrate that anti-dsDNA antibodies frequently develop early in the clinical course of SLE (with anti-dsDNA antibodies preceding diagnosis in 90% of anti-dsDNA positive cases). In fact, anti-dsDNA antibodies were first detected prior to the development of any SLE classification criteria in 74% of anti-dsDNA positive cases.

The data indicate that the onset of anti-dsDNA antibodies frequently precedes SLE diagnosis by years [as many as 9.3 years earlier in one case (Fig. 1)]. Although anti-dsDNA antibodies were first detected on average at 2.7 years before diagnosis, over half of these patients had anti-dsDNA antibodies in the first (or only) available serum sample. There is an ascertainment bias in these data. The observed mean of 2.7 years in this group is a lower bound and the true average in this population is longer, perhaps substantially longer, than these data suggest. Consider that of the nine cases with sera available 6 years or more before diagnosis, eight have anti-dsDNA in the first serum available (Fig. 1). Therefore, the time from a positive anti-dsDNA test to diagnosis is underestimated. In this study the interval from the appearance of anti-dsDNA to diagnosis is similarly underestimated in about half of the cases. Consequently, this study has not produced an accurate estimate of the average time from serologic conversion of anti-dsDNA, except to be confident that this interval is at least, and probably substantially longer than 2.7 years before diagnosis, an extraordinary and unexpected result.

In other experiments with these samples anti-nuclear antibodies were assessed in these same 72 patients who are anti-dsDNA positive at diagnosis. Anti-nuclear antibodies are present on average longer than are anti-dsDNA in these patients [median = 3.5 versus 2.1 years before diagnosis, P = 0.0012 (Wilcoxan ranked sum) relative to the anti-dsDNA antibodies (unpublished observations)].

These results reinforce the impression that the presence of autoantibody specificities in the natural history of lupus is cumulative. Anti-dsDNA appearing after diagnosis (n = 8) may reflect this tendency of lupus patients to accumulate autoimmune specificities. That the level of anti-dsDNA increases just before diagnosis (Fig. 2) adds additional support for both a qualitative and quantitative crescendo of autoimmune phenomena that culminates in clinical illness. Indeed, diagnosis and subsequent modern management clearly has important immune consequences (Fig. 2). One suspects that if the crescendo of autoimmune abnormalities were allowed to continue and the natural history of the disease were uninterrupted, as it was for many patients observed 50 years ago, then the data presented herein also provide a perspective from which to consider the cumulative immune disregulation that produced the 50% 5 year mortality of lupus typically observed at that time [18].

It is intriguing that anti-dsDNA antibodies, considered to be highly specific for SLE, were found years before disease onset in a presumably healthy population. In this case, anti-dsDNA antibodies may represent a serological marker for patients at risk for subsequent development of SLE. A prospective study by Swaak and Smeenk [12] found that non-SLE patients with anti-dsDNA antibodies detected by Farr assay were likely to subsequently fulfil preliminary classification criteria for SLE. In fact, 69% did so within 1 year of initial anti-dsDNA detection. In contrast, the percentage of patients in our study meeting SLE classification criteria [1, 2] within 1 year of developing anti-dsDNA antibodies was much smaller (29%). Differences between these results may be owing to the fact that individuals in the study by Swaak and Smeenk were identified after seeking medial attention for complaints warranting evaluation with anti-dsDNA testing, and thus, do not truly represent ‘pre-clinical’ disease.

Because autoantibodies are generally believed to contribute to the disease pathogenesis, it seems logical that these autoantibodies must precede the development of clinical manifestations. However, it is unclear in patients with long-standing autoimmunity what may trigger the transition from asymptomatic (or subclinical) autoimmunity (with established autoantibodies) to the onset of clinical disease. In these instances the onset of disease could be related to diverging specificity of established autoantibodies (as seen with epitope spreading), increasing avidity of autoantibodies (as in affinity maturation), localization of antibodies in pathogenic sites and/or by an increase in antibody levels (as demonstrated in this study).

Several studies have reported an association between disease activity and levels of anti-dsDNA antibodies in SLE, especially in patients with renal involvement [6, 19, 20]. These data generally support our observations at disease onset (Tables 3 and 4 and Fig. 2). Relapses are often found with a prominent rise in anti-dsDNA antibodies and followed by a dramatic decline. Ter Borg et al. found 89% of exacerbations accompanied by positive anti-dsDNA antibodies (24/27) were preceded by a significant increase in anti-dsDNA antibody levels [6]. Furthermore, this rise was seen in all exacerbations with renal involvement as the only or predominant manifestation (approximately 50% of the total relapses observed).

Following diagnosis patients in this study demonstrated a fall in anti-dsDNA levels, which correlated with corticosteroid treatment. The ability of steroids to decrease anti-dsDNA antibody levels has been previously documented [21]. Furthermore, treatment of patients when anti-dsDNA antibody levels start to rise can be effective in preventing a significant number of relapses. In a recent study by Bootsma et al. [14] patients demonstrating a significant (≥25%) rise in anti-dsDNA antibodies were randomly assigned to receive conventional treatment or 30 mg of prednisone added to the existing regimen. The frequency of exacerbation was significantly less in the groups treated with additional prednisone than in the conventional treatment group (2 versus 20, P < 0.001).

Murine models of lupus also provide parallels and insights relevant to these results. It has long been appreciated that autoantibodies precede clinical manifestations. For example, in both NZB and NZB/NZW mice the clinical manifestations of Coombs positive autoantibodies and antinuclear autoantibodies (of a number of specificities) precede evidence of, respectively, haemolytic anaemia and lupus nephritis by a few months [22, 23]. These murine findings demonstrate a much shorter time interval than observed in the lupus patients reported herein. However, when the life expectancy differences between mice and humans are considered and the interval treated as a fraction thereof, the proportion of the life expectancy from appearance of autoantibodies to clinically recognizable disease is more similar.

Some data and investigators strongly support the possibility that anti-dsDNA antibodies are pathogenic. If so, then the observation that anti-dsDNA autoantibodies precedes clinical illness by more than 2 years may contradict this position. There are two obvious explanations for these observations that would not require the rejection of the hypothesis that anti-dsDNA antibodies are pathogenic. First, perhaps the level of anti-dsDNA does not overwhelm the mechanisms in place to reduce their impact. These would include the processing of immune complexes and the generation of complement, among other possibilities. Second, perhaps the assay used in the laboratory for this study does not directly measure the pathogenic anti-DNA autoantibodies. Perhaps, other anti-dsDNA assays might be more closely associated with the more pathogenic antibodies [24], recently thought to be anti-nucleosome autoantibodies [25]. The possibility of progression in the autoantibody specificities observed also implies the obvious correlate that the autoantibodies change during the course of the illness. If so, then there must also be the possibility that environmental exposures and influences help trigger the process that culminates in sufficient quantity of pathogenic autoantibodies to cause clinical illness.

Thus, we show that anti-dsDNA antibodies precede the onset of clinical illness in SLE. Astonishingly, these antibodies are present in some patients for many years before the clinical onset of lupus. Furthermore, once present anti-dsDNA do not usually spontaneously disappear in people who subsequently develop clinical lupus. In fact, they appear to increase in titer until the diagnosis is made and modern therapy is instituted.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

We would like to thank Suzanne Lapolla and James Longton for their technical support and Xana Kim for the assistance with statistical analysis. We would also like to thank all the co-operating military rheumatologists and rheumatology clinics for their assistance. The use of Army medical records in the preparation of this material is acknowledged, but it is not to be construed as implying official Department of the Army approval of the conclusions presented. This work has been supported by the National Institutes of Health Grants AI31584, AR42474, AR01981, AR45084, AR45231, AR42460, AI24717 and the US Department of Veteran Affairs.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  • 1
    Tan EM, Cohen AS, Fries JF et al. The. revised criteria for the classification of systemic lupus erythematosus. Arthritis Rheum 1982;25:12717.
  • 2
    Hochberg MC. Updating the American College of Rheumatology revised criteria for the classification of systemic lupus erythematosus. Arthritis Rheum 1997;40:1725.
  • 3
    Davis P, Percy JS, Russel AS. Correlation between levels of DNA antibodies and clinical disease activity in SLE. Ann Rheum Dis 1977;36:1579.
  • 4
    Lloyd W & Schur PH. Immune complexes, complement and anti-DNA in exacerbations of systemic lupus erythematosus (SLE). Medicine 1981;60:20817.
  • 5
    Swaak AJG, Groenwold J, Bronsveld W. Predictive value of complement profiles and anti-dsDNA in systemic lupus erythematosus. Ann Rheum Dis 1986;45:35966.
  • 6
    Ter Borg EJ, Horst G, Hummel EJ, Limburg PC, Kallenberg CGM. Measurement of increases in anti-double-stranded DNA antibody levels as a predictor of disease exacerbation in systemic lupus erythematosus. A long-term, prospective study. Arthritis Rheum 1990;33:63443.
  • 7
    Spronk PE, Limburg PC, Kallenberg CGM. Review: Serological markers of disease activity in systemic lupus erythematosus. Lupus 1995;4:8694.
  • 8
    Tan EM. Immunopathology and pathogenesis of cutaneous involvement in systemic lupus erythematosus. J Invest Dermatol 1976;67:3605.
  • 9
    Winfield JB, Faiferman I, Koffler D. Avidity of anti-dsDNA antibodies in serum and IgG glomerular eluates from patients with systemic lupus erythematosus. J Clin Invest 1977;59:906.
  • 10
    Madaio MP, Carlson J, Cataldo J, Ucci A, Migliorini P, Pankewycz O. Murine monoclonal anti-DNA antibodies bind directly to glomerular antigens and form immune deposits. J Immunol 1987;138:28839.
  • 11
    Ceppelini R, Polli E, Celada F. A DNA-reacting factor in serum of patients with lupus erythematosus diffuses. Proc Soc Exp Biol Med 1957;96:5724.
  • 12
    Swaak T & Smeenk R. Detection of anti-dsDNA as a diagnostic tool: a prospective study in 441 non-systemic lupus erythematosus patients with anti-dsDNA antibody (anti-dsDNA). Ann Rheum Dis 1985;44:24551.
  • 13
    Tan EM, Smolen JS, McDougal JS, Butcher BT, Conn D et al. A critical evaluation of enzyme immunoassays for detection of antinuclear antibodies of defined specificities. I. Precision, sensitivity and specificity. Arthritis Rheum 1999;42:45564.
  • 14
    Bootsma H, Spronk P, Derksen R et al. Prevention of relapses in systemic lupus erythematosus. Lancet 1995;345:15959.
  • 15
    Smeenk R & Hylkema M. Detection of antibodies to DNA. a technical assessment. Mol Biol Reports 1992;17:719.
  • 16
    Smeenk RJT, Berden JHM, Swaak AJG. ‘dsDNA autoantibodies’. In: Peter JB, Shoenfeld Y, eds. Autoantibodies. Amsterdam: Elsevier, 1996:22736.
  • 17
    Smeenk RJT, Van Den Brink HG, Brinkman K et al. Choice of assay in relation to clinical value. Rheumatol Int 1991;11:1017.
  • 18
    Dubois EL. ‘Results of steroid therapy in systemic lupus erythematosus, Chapter 15’. In: DuboisEL, ed. Lupus Erythematosus, 5th edn. Los Angeles: University of Southern California Press, 1974:61332.
  • 19
    Cohen AS, Reynolds WF, Franklin EG et al. Preliminary criteria for the classification of systemic lupus erythematosus. Bull Rheum Dis 1971;21:6438.
  • 20
    Swaak AJG, Aarden LA, Statius van Eps LW, Feltkamp TEW. Anti-dsDNA and complement profiles as prognostic guides in systemic lupus erythematosus. Arthritis Rheum 1979;22:22635.
  • 21
    Weisbart RH & Colburn K. Effect of corticosteroids on serum antinuclear antibodies in man. Immunopharmacology 1984;8:97101.
  • 22
    Howie JB & Heyler BJ. The immunology and pathology of NZB mice. Adv Immunol 1968;9:21566.
  • 23
    Howie JB, Heyler BJ, Casey TP, Aarons I. Renal Disease in Autoimmune Strains of Mice. Proc Third International Congress Nephrol. Basel: Karger 1967;2:15063.
  • 24
    Hahn BH & Tsao BP. ‘Antibodies to DNA, Chapter 23’. In: WallaceDJ, HahnBH, eds. Dubois' Lupus Erythematosus, 5th edn. Baltimore: Williams & Wilkens, 1997:40722.
  • 25
    Amoura Z, Piette J-C, Bach J-F, Koutouzov S. The key role of nucleosomes in lupus. Arthritis Rheum 1999;42:26.