Sjögren's syndrome associated with systemic lupus erythematosus: Clinical and laboratory profiles and comparison with primary Sjögren's syndrome

Authors


Abstract

Objective.

To address the clinical, serologic, pathologic, and immunogenetic features of sicca syndrome that occurs in systemic lupus erythematosus (SLE), as well as its similarities to, and differences from, sicca syndrome that occurs in primary Sjögren's syndrome (SS).

Methods.

A cohort of 283 consecutive unselected SLE patients was evaluated for the presence of associated SS using the American–European classification criteria. Clinical and laboratory parameters in SLE patients with SS (SLE–SS) were compared with those in SLE patients without SS (SLE–no SS) and with a group of 86 unselected patients with primary SS.

Results.

SS was identified in 26 SLE patients (9.2%); the SS preceded the development of lupus in 18 of them (69.2%). Compared with the SLE–no SS group, patients with SLE–SS were significantly older, had a higher frequency of Raynaud's phenomenon, anti-Ro/SSA, anti-La/SSB, and rheumatoid factor, but had a significantly lower frequency of renal involvement, lymphadenopathy, and thrombocytopenia. Compared with the primary SS group, SLE–SS patients displayed a clinically similar sicca syndrome, but were significantly younger and had an increased frequency of perivascular infiltrates in the salivary glands associated with anticardiolipin antibodies in the serum. SLE–SS patients had a high frequency of the DRB1*0301 allele. This HLA profile distinguished the SLE–SS group from the SLE–no SS group, who had an increased frequency of DRB1*1501 and DQB1*0602 alleles, but was similar to the HLA profile of the primary SS group, who had an increased frequency of DRB1*0301.

Conclusion.

SLE–SS appears to constitute a subgroup of patients with distinct clinical, serologic, pathologic, and immunogenetic features, in whom SS is expressed as an overlapping entity and is largely similar to primary SS.

Sjögren's syndrome (SS) is a chronic autoimmune disorder of the exocrine glands with associated lymphocytic infiltration of the affected glands. Dryness of the mouth and/or the eyes, resulting from involvement of the salivary and lacrimal glands, is most often present (1). The exocrinopathy may be encountered alone or in the presence of another autoimmune disorder.

Features of SS may be found in almost every autoimmune rheumatic disease, including rheumatoid arthritis (RA), systemic lupus erythematosus (SLE), scleroderma, and others (1). With regard to the pathogenesis, it is unclear whether this form of SS associated with another autoimmune rheumatic disease represents a distinct overlapping entity or a manifestation in the clinical spectrum of the accompanying rheumatic disorder (2). In fact, comparisons of the clinical, serologic, and immunogenetic features of patients with SS who have no evidence of another autoimmune disease and those of patients with SS associated with RA have revealed distinct profiles (3, 4). Consequently, the terms “primary SS” and “secondary SS” were proposed for the former and the latter groups of patients, respectively (3). Such a clear distinction has not been established between primary SS and the sicca syndrome that coexists in autoimmune diseases other than RA, mainly because of a lack of comprehensive studies. Nevertheless, during the recent decade, the term “secondary SS” has generally been applied to connote sicca disorder that occurs together with any autoimmune disorder, without regard for the presence of true overlap cases.

The clinical coexistence of SLE and SS (SLE–SS) was recognized almost half a century ago (5). The prevalence of SS among patients with SLE varies considerably among the published studies (from 8% to 30%), most likely due to the application of different classification and patient selection criteria (6–9). Although controversy exists (10, 11), studies have indicated that in patients with SLE–SS, the associated lupus appears to be relatively more benign, and that these patients exhibit a relatively increased frequency of autoantibodies to Ro/SSA and La/SSB RNPs and rheumatoid factor (7, 8, 12). However, these studies are limited, and for most of them, the application of SS criteria that are not well established and the low number of patients studied precludes meaningful evaluation and conclusions. Therefore, the expression of SS that coexists with SLE needs to be further addressed. In addition, several issues pertaining to SLE–SS need to be clarified, including the severity of sicca manifestations in these patients, the temporal relationship between the development of sicca disorders and lupus itself, as well as the possible association with the immunogenetic background.

In the present study, we sought to determine the clinical, serologic, pathologic, and immunogenetic features of SLE–SS using well-defined SS classification criteria (13, 14). For this purpose, we studied a cohort of unselected SLE patients and compared the expression of sicca manifestations, as well as systemic disease features in patient subgroups with and without SS, and in patients with primary SS.

PATIENTS AND METHODS

Patients.

Patients studied included 283 consecutive unselected SLE patients (15) who were followed up in our department (Department of Pathophysiology, Medical School, National University of Athens) from 1994 to 1999. All patients were Caucasians of Greek origin and were evaluated for evidence of secondary SS according to the established European classification criteria (13). These classification criteria for secondary SS do not include serologic items, and they remained unchanged in the recently published revised form of criteria proposed by the American–European Consensus group (14).

All patients were initially evaluated by questionnaire for the presence of subjective symptoms of oral or ocular sicca (13). None of the patients with subjective sicca problems had a history or clinical evidence of primary lymphoma, infection with hepatitis B virus, hepatitis C virus, or human immunodeficiency virus, head and neck radiation therapy, sarcoidosis, graft-versus-host disease, or were receiving treatment known to cause mucosal dryness. Patients with affirmative responses to the questionnaire(s) underwent objective assessment of lacrimal and salivary gland involvement, which included slit lamp examination of the eye with rose bengal staining of conjunctivae (abnormal if ≥4 of 9, according to the scoring system of Van Bijsterveld), Schirmer I test (abnormal if ≤5 mm in 5 minutes), and measurement of unstimulated salivary flow (whole saliva; abnormal if ≤1.5 ml in 15 minutes). Patients with subjective sicca problems but without objective findings were excluded from further study.

Patients with subjective complaints and objective findings of sicca consented to undergo a minor labial salivary gland biopsy under local anesthesia. Lip biopsy was not performed in 6 patients (4 declined biopsy; 2 others declined for medical reasons [anticoagulation treatment]). Tissue specimens were scored according to the method described by Chisholm and Mason (16).

SLE patients who met the classification criteria for secondary SS (14) were designated as the SLE–SS group. Data from these patients were compared with data obtained from a group of 76 consecutive, randomly selected SLE patients (1:3 randomized selection, according to the hospital registry) who had no subjective or objective findings of sicca (SLE–no SS group), as well as with data obtained from a group of 86 consecutive, unselected patients with primary SS (14).

The patients' medical records were retrospectively analyzed for demographic, clinical, and laboratory data. For all patients, the onset and duration of either SLE (for the SLE–no SS and the SLE–SS groups) or SS (for the SLE–SS and the primary SS groups) were defined from the time when the diagnosis of SLE or SS could be ascertained by the established classification criteria (14, 15). For patients with SS, the duration of sicca manifestations was determined from the date when the first sicca symptom (either xerostomia or xerophthalmia) had appeared.

Clinical features.

Various clinical manifestations were recorded and analyzed separately. Renal involvement was based on the presence of at least 1 of the following 4 criteria: 1) persistently elevated urinary pH value (≥6.0), 2) persistently elevated serum creatinine level (>1.5 mg/dl) and/or impaired creatinine clearance (<50 ml/minute), 3) persistent proteinuria (≥500 mg/day) for more than 3 months, or 4) a pathologic urine sediment (consisting of >10 red blood cells per high-power field or red blood cell casts), which was confirmed by kidney biopsy, according to standard criteria (17). Pulmonary involvement was identified by the presence of at least 1 of the following 4 criteria: 1) chronic interstitial infiltrates on chest roentgenography, 2) restrictive or obstructive disease on pulmonary function testing, 3) decreased diffusion capacity for carbon monoxide (DLCO<80%), or 4) pulmonary hypertension.

The occurrence of peripheral neuropathy was established based on symptoms associated with the clinical presence of motor or sensory deficits and compatible electrophysiologic findings. Central nervous system (CNS) involvement was defined according to the American College of Rheumatology nomenclature and case definitions for neuropsychiatric lupus syndromes (18). Myositis was documented by the presence of proximal muscle weakness associated with increased aldolase or creatine phosphokinase levels and compatible findings on electromyography and muscle biopsy. The diagnosis of antiphospholipid syndrome was based on established criteria (19). Serositis was defined as pleurisy documented by clinical examination and pleural effusion on chest radiography or pericarditis documented by pericardial effusion on echocardiography. The development of lymphoma was documented by biopsy.

Laboratory features.

The occurrence of laboratory abnormalities (verified on at least 2 successive measurements) was also recorded, including hemolytic anemia (features of hemolysis with a hemoglobin level <12 gm/dl), leukopenia (white blood cell count <4,000/mm3), lymphopenia (lymphocyte count <1,500/mm3), and thrombocytopenia (platelet count <100,000/mm3), or the presence of serum cryoglobulins, rheumatoid factor (by latex test; positive at a titer ≥1:80 or a value ≥20 IU/ml), antinuclear antibodies (ANAs) (by indirect immunofluorescence; positive at a titer ≥1:160), antibodies to Sm, U1 RNP, Ro/SSA, and La/SSB cellular antigens (by counterimmunoelectrophoresis), IgG antibodies to double-stranded DNA (dsDNA) (by enzyme-linked immunosorbent assay [ELISA]), or IgG and IgM antibodies to cardiolipin (by ELISA) (20).

Patients also underwent typing for HLA–DRB, DQA, and DQB, which was performed by polymerase chain reaction and hybridization with sequence-specific oligonucleotide probes. These results were compared with the data obtained from 246 unrelated healthy controls.

Statistical analysis.

Statistical analyses were performed using SAS software (version 6; SAS Institute, Cary, NC). The data from the 3 patient groups (SLE–no SS, SLE–SS, and primary SS) were analyzed for statistically significant associations using a multinomial logit model (21) or Fisher's exact test for binary values and the Kruskal-Wallis and the Mann-Whitney nonparametric tests for continuous responses. The multinomial logit model is a vigorous method for the simultaneous detection of significant variables that can distinguish the 3 groups. The multinomial logit model was not utilized for variables that were not applicable to all 3 groups or had unbalanced data. For such isolated comparisons, Fisher's exact test was applied, whereas a logit model was used for the evaluation of differences in sicca manifestations between the 2 SS groups (SLE–SS and primary SS). The association of phenotypic or allelic HLA frequencies in the various groups of patients was tested using the chi-square test with Yates' correction. P values less than 0.05 were considered significant.

RESULTS

Taken as a whole, the SLE and primary SS patient groups studied were similar to other larger patient cohorts previously described (22, 23), as supported by the comparison of clinical and laboratory profiles. Fifty-one of the 283 SLE patients (18.0%) had subjective complaints of oral and/or ocular sicca, whereas 35 of the 283 patients (12.4%) also had objective evidence of ocular and/or oral glandular involvement. Twenty-six of these 35 patients with SLE and sicca manifestations (9.2% of the total SLE population) fulfilled the American–European classification criteria for secondary SS (14 met 4 criteria; 12 met 3 criteria) (14) and were designated as the SLE–SS group. Full-blown SS (subjective complaints of both xerostomia and xerophthalmia corroborated by findings of objective assessments) was observed in 16 of the 26 patients with SLE–SS (61.5%; 5.7% of the total SLE population).

Demographic data.

The demographic and general characteristics of the 3 patient groups (SLE–SS, SLE–no SS, and primary SS) are summarized in Table 1. Patients with SLE–SS were significantly older than those with SLE–no SS (P < 0.001) but were younger than those with primary SS (P = 0.007). As determined by the time when patients met the clinical diagnosis of SLE, the development of lupus was relatively delayed in patients with SLE–SS compared with those with SLE–no SS (P < 0.001). Compared with the primary SS group, SS was clinically recognized relatively earlier in the SLE–SS group (P = 0.010).

Table 1. Comparison of the general features of the study patients, by diagnostic group*
FeatureSLE–no SS (n = 76)SLE–SS (n = 26)Primary SS (n = 86)P
SLE–SS vs. SLE–no SSSLE–SS vs. primary SS
  • *

    The onset (and therefore, duration) of systemic lupus erythematosus (SLE) or Sjögren's syndrome (SS) was defined from the time when the diagnosis of SLE or SS could be ascertained by the established classification criteria (14, 15). Except where indicated otherwise, values are the mean ± SEM. P values less than 0.05 are considered significant. NA = not applicable.

Age, years36.1 ± 1.648.3 ± 2.556.5 ± 1.5<0.0010.007
% female81.610093.00.0180.333
Age at SLE onset, years29.7 ± 1.544.8 ± 2.4NA<0.001NA
Age at SS onset, yearsNA44.7 ± 2.751.5 ± 1.6NA0.010
Duration of SLE, years5.9 ± 0.53.5 ± 0.5NA0.011NA
Duration of SS, yearsNA3.7 ± 0.74.9 ± 0.4NA0.068

Sicca manifestations.

Sicca manifestations preceded the onset of SLE by 1–15 years in 18 of the 26 patients with SLE–SS (69.2%). In fact, these 18 patients were being followed up and treated primarily for the sicca symptoms and mild extraglandular manifestations. Among all 26 SLE–SS patients, the median onset of sicca manifestations was −4 years relative to the onset of SLE (range −15 years to +3 years) (Figure 1). Another 4 of the 26 patients experienced the concurrent development of symptoms and findings that signified the diagnoses of both SLE and SS. In the remaining 4 patients of this group, the diagnosis of SLE preceded the diagnosis of SS by 2–3 years.

Figure 1.

Onset of sicca manifestations in relation to the onset of systemic lupus erythematosus (SLE) in the 26 patients with SLE–Sjögren's syndrome.

Twenty-three of the 26 SLE–SS patients (88.5%) had subjective complaints of both xerophthalmia and xerostomia, whereas the 3 remaining patients reported only xerophthalmia (2 patients) or only xerostomia (1 patient). Xerophthalmia was self-reported by 15 patients (57.7%) and xerostomia was self-reported by 14 patients (53.8%) of the SLE–SS group. Objective assessments revealed the occurrence of both ocular and oral sicca involvement in 19 of the 26 patients (73.1%), whereas 6 patients had only ocular involvement and 1 had only oral involvement. Comparison of sicca manifestations between the SLE–SS and primary SS groups revealed distinct similarities and differences (Table 2).

Table 2. Comparison of sicca manifestations between patients with SLE–SS and patients with primary SS*
Sicca manifestationSLE–SS (n = 26)Primary SS (n = 86)P
  • *

    P values less than 0.05 are considered significant. SLE = systemic lupus erythematosus; SS = Sjögren's syndrome; ND = not done; NA = not applicable.

Subjective sicca symptoms   
 Age at onset, mean ± SEM years40.7 ± 2.748.5 ± 1.50.015
 Duration prior to SS diagnosis, mean ± SEM years3.9 ± 0.94.1 ± 0.50.798
Xerophthalmia   
 % positive96.294.20.547
 Duration, mean ± SEM years6.3 ± 0.98.4 ± 0.60.139
Xerostomia   
 % positive92.397.70.118
 Duration, mean ± SEM years7.4 ± 0.98.2 ± 0.60.748
Dyspareunia   
 % positive46.220.90.004
Salivary gland enlargement   
 % positive46.251.20.857
 Duration, mean ± SEM years7.6 ± 1.3NDNA
Objective assessment of ocular involvement   
 % positive on Schirmer I testing76.984.90.115
 % positive on rose bengal staining65.483.70.022
Objective assessment of salivary gland involvement   
 % with positive findings75.074.30.781
 Salivary flow, mean ± SEM ml/15 minutes1.4 ± 0.41.0 ± 0.30.260
Histologic features of minor salivary gland biopsy   
 % with positive findings91.397.60.205
 No. of foci/4 mm2, mean ± SEM1.6 ± 0.31.7 ± 0.10.130
 % with perivascular infiltrates38.92.4<0.001

As in patients with primary SS, in the majority of SLE–SS patients, minor salivary gland lesions were characterized by periductal lymphocytic infiltrates, which occasionally were extensive. Nevertheless, lymphocytic infiltrative lesions that were primarily located perivascularly were recorded in a considerable proportion of SLE–SS patients (38.9%), but not in those with primary SS (P < 0.001) (Table 2 and Figure 2). Compared with patients without perivascular infiltrates, SLE–SS patients with perivascular infiltrates had longer durations of disease (mean ± SEM 16.6 ± 3.9 versus 7.4 ± 1.3 years; P = 0.030), sicca manifestations (9.0 ± 1.4 versus 4.8 ± 1.0 years; P = 0.037), and salivary gland enlargement (9.2 ± 1.9 versus 2.8 ± 0.6 years; P = 0.019).

Figure 2.

Characteristic presentation of perivascular mononuclear cell infiltrates in the minor salivary gland tissue of a patient with systemic lupus erythematosus–Sjögren's syndrome.

Clinical and laboratory features.

The prevalence of distinct disease manifestations in the groups of patients with SLE–no SS, SLE–SS, and primary SS are presented in Table 3. As derived from the multinomial logit model, a particular set of clinical variables, namely, Raynaud's phenomenon, renal involvement, lymphadenopathy, and thrombocytopenia, was found to be important overall for discriminating SLE–SS patients from SLE–no SS patients. In contrast, the set of Raynaud's phenomenon, arthritis, serositis, lymphadenopathy, and CNS involvement was found to be important for discriminating SLE–SS patients from primary SS patients (Table 3).

Table 3. Comparison of various disease manifestations between the study groups*
Disease manifestation% positiveP
SLE–no SS (n = 76)SLE–SS (n = 26)Primary SS (n = 86)SLE–SS vs. SLE–no SSSLE–SS vs. primary SS
  • *

    P values less than 0.05 are considered significant. SLE = systemic lupus erythematosus; SS = Sjögren's syndrome; NA = not applicable; CNS = central nervous system.

Raynaud's phenomenon43.480.843.0<0.0010.033
Arthritis51.376.931.40.2770.002
Mucocutaneous involvement80.388.5NA0.552NA
Livedo reticularis39.515.411.60.0820.402
Purpura36.830.825.60.6730.634
Myositis2.63.83.50.6220.558
Lung involvement11.811.59.30.8910.610
Kidney involvement55.311.54.70.0050.148
Peripheral nerve involvement5.311.511.60.1290.783
CNS involvement19.711.50.00.5520.012
Antiphospholipid syndrome22.47.70.00.1440.052
Serositis36.850.04.70.779<0.001
Lymphadenopathy46.119.244.20.0040.002
Hemolytic anemia18.47.70.00.3480.055
Thrombocytopenia26.37.71.20.0300.104
Leukopenia51.338.544.20.3310.790
Lymphopenia52.650.034.90.7410.222
Lymphoma0.00.07.0NA0.113

Histologic evidence of kidney disease was obtained in 3 patients with SLE–SS and in 35 patients with SLE–no SS. Among the patients with SLE–SS, 2 had histologic lesions of focal segmental proliferative glomerulonephritis (GN) (World Health Organization [WHO] class III), and 1 had diffuse membranous GN (WHO class V). Among those with SLE–no SS, kidney disease was classified as mesangial hyperplasia (WHO class II) in 3 patients, focal segmental proliferative GN (WHO class III) in 12, diffuse proliferative GN (WHO class IV) in 7, membranoproliferative GN (WHO class V) in 11, and advanced sclerosing GN (WHO class VI) in 2. Along with GN, 2 of the SLE–no SS patients also had evidence of interstitial nephritis. Among the patients with primary SS, 4 had evidence of renal involvement (4.7%); 3 of these 4 patients had distal tubular acidosis type I and nephrocalcinosis (kidney biopsy in 2 of them revealed changes of interstitial nephritis), and 1 had lesions of membranoproliferative GN.

CNS involvement in the 15 patients with SLE–no SS included cerebrovascular disease (6 patients), seizures (5 patients), and psychosis (2 patients), as well as myelopathy and chorea (1 patient each). The neuropsychiatric manifestations in the SLE–SS group were cerebrovascular disease (2 patients), cognitive dysfunction (1 patient), and chorea (1 patient). Six of the patients with SLE–no SS and 2 with SLE–SS had features of antiphospholipid syndrome along with CNS involvement.

Serologic analyses.

The serologic profiles of the 3 groups of patients are presented in Table 4. All patients who were evaluated had positive findings on tests for ANAs. Anti-Ro/SSA autoantibodies were found in the presence or absence of anti-La/SSB, whereas anti-La/SSB autoantibodies were always detected in association with anti-Ro/SSA. Therefore, the occurrence of anti-Ro/SSA without anti-La/SSB antibodies (i.e., anti-Ro/SSA alone) was analyzed separately.

Table 4. Comparison of serologic features between the study groups*
Serologic feature% positiveP
SLE–no SS (n = 71)SLE–SS (n = 26)Primary SS (n = 86)SLE–SS vs. SLE–no SSSLE–SS vs. primary SS
  • *

    Anti-Ro/SSA antibodies were detected either alone or in the presence of anti-La/SSB antibodies. Anti-La/SSB antibodies were always detected in the presence of anti-Ro/SSA antibodies. Values in parentheses are the number of patients positive/number tested. P values less than 0.05 are considered significant. SLE = systemic lupus erythematosus; SS = Sjögren's syndrome; anti–U1 nRNP = anti–U1 nuclear RNP; anti-dsDNA = anti–double-stranded DNA.

Anti-Ro/SSA antibodies alone23.938.533.70.0080.291
Anti-La/SSB antibodies7.038.529.1<0.0010.192
Anti-Sm antibodies11.37.70.00.9990.056
Anti–U1 nRNP antibodies12.711.51.20.9990.013
Anti-dsDNA antibodies77.3 (58/75)69.20.00.436<0.001
Anticardiolipin antibodies52.9 (37/70)45.8 (11/24)10.3 (4/39)0.639<0.001
Rheumatoid factor28.6 (14/49)64.0 (16/25)61.7 (50/81)<0.0010.999
Cryoglobulins14.7 (5/34)15.8 (3/19)15.1 (8/53)0.9990.999

Compared with the SLE–no SS group (but not the group with primary SS), significantly more patients in the SLE–SS group exhibited serum anti-Ro/SSA antibodies alone (P = 0.008), anti-La/SSB antibodies (P < 0.001), as well as rheumatoid factor (P < 0.001). The statistical difference in the presence of anti-Ro/SSA alone between the SLE–SS group and the SLE–no SS group was revealed by the logit model (which also included the data on anti-La/SSB as a separate variable), but not by Fisher's exact test (which involved only separate analyses; P = 0.202), a fact that likely indicates an interdependence between anti-Ro/SSA and anti-La/SSB autoantibodies. The logit model also revealed significant differences in the prevalence of anti-Ro/SSA alone (P = 0.012) and anti-La/SSB (P < 0.001) between the SLE–no SS and primary SS groups. In contrast, SLE–SS patients had a significantly higher frequency of anti-dsDNA (P < 0.001), anticardiolipin (P < 0.001), and anti–U1 nuclear RNP (P = 0.013) antibodies than did the patients with primary SS, but not the patients with SLE–no SS. A high frequency of anticardiolipin antibodies was observed in the subgroup of SLE–SS patients with perivascular infiltrates on lip biopsy (100%) compared with the remaining SLE–SS patients without this feature (9.1%) (P = 0.003).

Immunogenetic analyses.

The immunogenetic analysis of HLA–DRB, DQA, and DQB alleles in the 3 groups of study patients revealed several statistically significant associations as compared with the healthy control group (Table 5). Compared with healthy controls, patients with SLE–SS had an increased frequency of the DRB1*0301 allele (P = 0.042). In addition, an increased frequency of the DQB1*0201 allele was found (43.5% versus 24.4% in controls), but this did not reach statistical significance (P = 0.081), most likely because of the relatively low number of observations. These HLA associations distinguished the SLE–SS group from the SLE–no SS group, in whom there were increased phenotype and allele frequencies for DRB1*1501 (P = 0.020 and P = 0.015, respectively) and DQB1*0602 (both P < 0.001). Also in the SLE–no SS group, the frequency of the DQA1*0102 allele was marginally higher than that in the control group (35.4% versus 25.2% in controls, P = 0.073). However, there was a similarity between the SLE–SS group and the primary SS group, in whom there were increased phenotype and allele frequencies for DRB1*0301 (P = 0.012 and P = 0.009, respectively, versus controls). A relatively increased phenotype frequency for the DRB1*1104 allele was also noted in the primary SS patients studied (47.3% versus 34.1% in controls), but the difference was not statistically significant (P = 0.094).

Table 5. Statistically significant immunogenetic associations for HLA–DRB, DQA, and DQB alleles in the patient groups compared with the healthy control group*
Allele% frequency
SLE–no SS (n = 41)SLE–SS (n = 23)Primary SS (n = 55)Healthy controls (n = 246)
Phenotype (n = 41)Allele (n = 82)Phenotype (n = 23)Allele (n = 46)Phenotype (n = 55)Allele (n = 110)Phenotype (n = 246)Allele (n = 492)
  • *

    Only statistically significant differences (P < 0.05) between the patient groups and the healthy controls are shown. SLE = systemic lupus erythematosus; SS = Sjögren's syndrome.

  • P = 0.042 versus healthy controls.

  • P = 0.012 versus healthy controls.

  • §

    P = 0.009 versus healthy controls.

  • P = 0.020 versus healthy controls.

  • #

    P = 0.015 versus healthy controls.

  • **

    P < 0.001 versus healthy controls.

  • ††

    P < 0.001 versus healthy controls.

DRB1*030119.59.830.415.227.314.5§12.66.5
DRB1*150126.814.6#8.74.312.77.311.86.3
DQB1*060229.3**15.9††8.74.312.77.36.93.5

DISCUSSION

With the exception of SS associated with RA (3, 4), comparative studies of well-defined groups of patients manifesting SS in association with another autoimmune rheumatic disease and patients with primary SS are lacking. Despite this fact, the term “secondary SS” (denoting SS identifiable in the context of another disorder) has been arbitrarily granted to SS that occurs in patients with any autoimmune disorder. More specifically, in regard to sicca syndrome that occurs in the context of SLE, studies have not satisfactorily addressed whether such a form of SS represents one more feature in the spectrum of lupus (i.e., SLE with “secondary SS”) or whether it represents a separate coexisting entity (i.e., a lupus–SS overlap disease). In the present study, in an attempt to better define the nature and expression of SS that accompanies SLE, we conducted a more in-depth study of the clinical, serologic, pathologic, and immunogenetic features of these patients. For this purpose, a fairly large number of unselected patients with well-defined SLE were evaluated for evidence of SS, and their data were compared with the data obtained from patients with well-defined primary SS. The analysis of our population of 283 unselected SLE patients revealed the occurrence of sicca manifestations in 12.4%, whereas SS was found in 9.2% according to the established criteria for secondary SS (13, 14).

Overall, our data indicate that SS that occurs together with SLE resembles, in several aspects, primary SS and that such SLE–SS patients constitute a subgroup of patients with SLE characterized by milder lupus and distinctive clinical, serologic, pathologic, and immunogenetic features. These findings corroborate and extend certain lines of evidence presented previously (7, 8, 24–26). Patients with SLE–SS were found to be older with a later onset of disease compared with patients with SLE–no SS, but younger with an earlier onset of disease compared with patients with primary SS. In terms of clinical manifestations, there was a significantly higher frequency of Raynaud's phenomenon, but a lower frequency of renal involvement, thrombocytopenia, and lymphadenopathy, in patients with SLE–SS than in patients with SLE–no SS.

Our data indicate that in most respects, the clinical presentation of sicca syndrome that occurs in SLE–SS patients is very similar to that in primary SS patients, including the presence of salivary gland enlargement, which has been shown to be infrequent in RA patients with secondary SS (3). Dyspareunia was reported significantly more frequently by patients with SLE–SS than by patients with primary SS, possibly related to vasculitic processes that are common in lupus. In addition, our results have established that SLE–SS patients possess a classic primary SS–associated autoantibody profile, namely, antibodies to the Ro/SSA and La/SSB RNP complex and rheumatoid factor, which further distinguishes them from patients with SLE–no SS. However, the SLE–SS patients displayed 2 features that differentiated them from patients with primary SS and that apparently result from the influence of background factors of lupus. The first is the serologic autoantibody indices of lupus (i.e., anti-dsDNA and anticardiolipin antibodies), which are retained by patients with SLE–SS and do not distinguish them from patients with SLE–no SS. The second is the presence of perivascular infiltrates in the minor salivary gland lesions of SLE–SS patients, which are clearly different from the classic periductal lymphoepithelial infiltrates of primary SS, as first described by investigators from our group (27).

The present study was able to reconfirm these observations in a completely different patient population. Perivascular infiltrates were observed among vessels that were located in close proximity to ducts in a considerable proportion of SLE–SS patients. The occurrence of such perivascular infiltrates in SLE–SS patients was found to correlate significantly with a longer duration of lupus-related and of SS-related symptoms, possibly indicating a time-dependent complication. In addition, in a manner reminiscent to an antiphospholipid-associated vasculopathy (28), the presence of perivascular infiltrates on lip biopsy was found to correlate highly significantly with the presence of serum anticardiolipin antibodies.

The statistically significant association of anti-Ro/SSA and anti-La/SSB antibodies with SLE–SS deserves special attention. These particular autoantibodies appear to represent a meaningful link between sicca syndrome that occurs in primary SS as well as in SS that coexists with lupus. Previous assessments of anti-Ro/SSA– and/or anti-La/SSB–positive SLE patients and mothers of offspring with neonatal lupus indicated the immunogenetic similarity between these patients and patients with primary SS (29, 30), a fact that strengthens the association between these autoantibodies and SS. In fact, patients with primary SS and patients with SLE–SS have previously been reported to manifest a similar peptide and protein specificity of autoantibody responses to Ro/SSA and La/SSB (31–33). These findings are especially important not only for classification purposes (2), but also because they strongly suggest a relationship between the biologic processes that generate these specific immunologic responses and the autoimmune sicca disorder itself.

The distinctive clinical and serologic presentation of SLE–SS patients among the total SLE population may be the result of immunogenetic variability, which in turn would be associated with the expression of particular disease phenotypes. In fact, the existence of certain HLA–DR and DQ alleles has been shown to correlate highly significantly with the development of SLE as well as primary SS (30, 34, 35). Previous studies of primary SS patients revealed associations with various HLA–DR alleles (DRB1 genes), including DR3 (DRB1*0301) and DR5 (DRB1*1104). In addition, significant associations with alleles at the HLA–DQ locus, such as the DQA1*0501, DQB1*0201, and DQB1*0301 alleles, have been also demonstrated in patients with primary SS with autoantibody responses to Ro/SSA and La/SSB RNPs (34, 35).

To our knowledge, however, no previous studies have addressed the immunogenetic profile in well-defined SLE–SS patients in comparison with that in SLE patients without SS and in patients with primary SS. Despite the relatively low number of SLE–SS patients tested, this is the first study to demonstrate that this particular subgroup of SLE patients bears an immunogenetic background that strongly resembles that of primary SS and differs from that of SLE–no SS. Similar to patients with primary SS, those with SLE–SS were found to display an increased prevalence of DRB1*0301 (DR3). In contrast, our findings clearly distinguished SLE–SS patients from SLE–no SS patients, as illustrated by the high frequency of the DRB1*0301 allele, coupled with the significantly lower frequency of the lupus-related alleles DRB1*1501 and DQB1*0602.

Taken together, our findings strongly imply that SS occurring in association with SLE represents a coexisting entity that is distinct from lupus (i.e., a lupus–SS overlap disease). In direct contrast, the occurrence of SS in patients with RA has not been shown to be associated with the primary SS–related immunogenetic background, but with that of RA, which suggests that this form of SS corresponds to a feature in the spectrum of RA (i.e., “secondary SS” in the context of RA) (3, 4). Consistent with this primary SS–associated immunogenetic profile, we found that in the majority of SLE–SS patients, sicca manifestations had begun well before the outbreak of signs indicative of lupus. A few cases of patients with longstanding primary SS who eventually developed a systemic disorder that fulfilled the criteria for SLE have also been described previously (36, 37). According to our data these SLE manifestations are mild and usually include arthritis, serositis, and mucocutaneous manifestations, whereas in the majority of such patients, the development of lupus is denoted by high titers of anti-dsDNA antibodies.

In conclusion, SS that occurs in association with SLE appears to be a form of the primary sicca disorder that progresses to, or is followed by, lupus that results in a form of “lupus–SS overlap” disease. From this point of view, disease expression in patients with SLE–SS may represent a continuous spectrum, ranging from the classic presentation of primary SS with minimal lupoid manifestations to a more recognizable lupus process with sicca features. Thus, SLE–SS may be considered a specific subgroup of primary SS. SLE–SS patients have a immunogenetic profile compatible with that of primary SS, but they develop sicca manifestations at a younger age, often have serum anti-dsDNA antibodies and perivascular infiltrates in the salivary glands, and usually proceed to develop mild features of SLE.

Ancillary