Concomitant Ro/SSA and La/SSB antibodies are biomarkers for the risk of venous thromboembolism and cerebral infarction in primary Sjögren's syndrome

Abstract Background To assess the risk of incident cardiovascular disease in patients with primary Sjögren's syndrome, overall and stratified by Ro/SSA and La/SSB autoantibody status. Methods A cohort of patients with primary Sjögren's syndrome in Sweden (n = 960) and matched controls from the general population (n = 9035) were included, and data extracted from the National Patient Register to identify events of myocardial infarction, cerebral infarction and venous thromboembolism. Hazard ratios were estimated using cox proportional hazard regressions. Results During a median follow‐up of 9.5 years, the overall hazard ratio (HR) was 1.6 (95% CI 1.2–2.1) for myocardial infarction, 1.2 (95% CI 0.9–1.7) for cerebral infarction and 2.1 (95% CI 1.6–2.9) for venous thromboembolism. Patients positive for both Ro/SSA and La/SSB autoantibodies had a substantially higher risk of cerebral infarction (HR 1.7, 95% CI 1.0–2.9) and venous thromboembolism (HR 3.1, 95% CI 1.9–4.8) than the general population. These risks were not significantly increased in Ro/SSA‐ and La/SSB‐negative patients. Among autoantibody‐positive patients, the highest HR of cerebral infarction was seen after ≥10 years disease duration (HR 2.8, 95% CI 1.4–5.4), while the HR for venous thromboembolism was highest 0–5 years after disease diagnosis (HR 4.7, 95% CI 2.3–9.3) and remained high throughout disease duration. Conclusions Primary Sjögren's syndrome is associated with a markedly increased risk of cardiovascular disease and the presence of Ro/SSA and La/SSB autoantibodies identify the subgroup of patients carrying the highest risk. These findings suggest that monitoring and prevention of cardiovascular disease in this patient group should be considered.


Introduction
Primary Sj€ ogren's syndrome is a systemic autoimmune disease, characterized by dysfunction of salivary and lacrimal glands leading to sicca symptoms [1,2]. The clinical presentation often includes arthralgia, myalgia and fatigue, and a subgroup of patients present with systemic manifestations such as cutaneous vasculitis, polyneuropathy and interstitial lung disease [3][4][5][6]. Serologically, autoantibodies to the Ro/SSA and La/SSB antigens can be detected [7][8][9]. These autoantibodies can induce production of type I IFN, and associate with disease severity and systemic manifestations [10][11][12].
Inflammation contributes to cardiovascular disease, and increased morbidity and premature mortality related to cardiovascular events have been reported for rheumatoid arthritis (RA) and systemic lupus erythematosus (SLE) [13][14][15]. In Sj€ ogren's syndrome, register-based studies have addressed the risk of cardiovascular disease, but with inconsistent results [16][17][18][19][20][21][22][23]. However, to the best of our knowledge, no study has previously investigated the long-term risk of cardiovascular disease in a large cohort of clinically validated patients with primary Sj€ ogren's syndrome. Further, the potential impact of Ro/SSA and La/SSB autoantibodies on the risk of cardiovascular disease has not been assessed in large cohorts. In the present study, we therefore investigated the incidence and relative risk of cardiovascular events (myocardial infarction, cerebral infarction and venous thromboembolism) in close to 1000 well-characterized patients with primary Sj€ ogren's syndrome compared with individuals from the general population with a median follow-up period approaching 10 years, overall, and stratified by Ro/SSA and/or La/SSB autoantibody status.

Study design
We performed a cohort study of patients with primary Sj€ ogren's syndrome included at diagnosis (incident cases) with matched general population comparators, based on prospectively recorded register data. The publicly funded Swedish healthcare system enables access to all healthcare services, including specialized care for chronic diseases, for all residents. All residents are assigned a unique personal identity number that can be used for linkage of different data resources, including several national health registers of high quality [24].

Study population
Patients with primary Sj€ ogren's syndrome (n = 960) diagnosed between 1987 and 2013 at the Departments of Rheumatology at the University Hospitals in Gothenburg, Malm€ o/Lund, Link€ oping, € Orebro and Uppsala, as well as the Karolinska University Hospital in Stockholm, Sweden were included in the study. All patients received their diagnosis by a specialist in rheumatology at each centre, and all fulfilled the American-European consensus group (AECG) criteria [25]. Clinical parameters related to diagnosis, including autoantibody status, were collected through patient chart review.
For each patient, ten controls from the general population (matched on sex, age and region of residency 10 years before the matched case's diagnosis date) were randomly selected from the Swedish Total Population Register at Statistics Sweden (www.scb.se/en). Matching prior to Sj€ ogren's syndrome diagnosis date was chosen due to patients reporting onset of disease-specific symptoms several years before diagnosis [5]. Controls were required to be alive and resident in Sweden at the date of the matched case's diagnosis, resulting in a cohort of n = 9035 controls. The study was approved by the Regional Ethical Review Board in Stockholm, Sweden.
Data sources used to detect outcomes during follow-up Using the national personal identification number, we linked the cohort of patients with primary Sj€ ogren's syndrome and the matched general population comparator cohort with data from the following registers, for which data were available through 31 December 2013: The Swedish National Patient Register (NPR), the Population Register and the Cause of Death Register. The NPR has a nationwide coverage of hospitalizations (inpatient care) since 1987 and of specialist outpatient care (excluding primary care by general practitioners) since 2001. The register lists date of admission, date of discharge and the discharge diagnosis (primary and secondary diagnoses) as set by the discharging physician and classified according to the calendar year-specific version of the International Classification of Diseases (ICD). The coverage is 99% for hospitalizations and 80% for outpatient care, with the latter lower coverage mainly due to lower reporting from private healthcare providers [26]. The Population Register includes information on deaths, emigration and immigration for the entire Swedish population. The Cause of Death Register holds information on cause of death since 1962 (primary and secondary diagnosis), coded according to ICD. Through these linkages, we identified all hospitalizations and nonprimary care outpatient visits after primary Sj€ ogren's syndrome diagnosis date, and all emigrations and deaths during follow-up.

Definition of outcomes and follow-up
Using the NPR and Cause of Death Register, we assessed the risk of three types of cardiovascular disease: Myocardial infarction (MI), cerebral infarction and venous thromboembolism (VTE).
MI was defined as hospitalization listing a primary ICD-10 diagnosis of I21 (acute myocardial infarction), or an acute myocardial infarction listed as an underlying cause of death. This definition has a positive predictive value (PPV) of 95% [27]. Cerebral infarction was defined as a primary ICD-10 diagnosis of I63 in the inpatient register, or listed as an underlying cause of death, where diagnosis validation has shown a PPV of > 95% [26]. VTE was defined as a composite measure of pulmonary embolism (ICD-10: I26) and deep vein thrombosis (DVT) (ICD-10: I80.1-I80.2, I81, I82.2-I82.9). Primary and secondary diagnoses in the inpatient register and cause of death register were included; diagnoses of DVTs in outpatient care were also included. A previous study examining the validity of VTEs in the NPR suggests that this definition holds a satisfactory validity [28]. For records of diagnoses predating 1997, corresponding ICD-9 codes were used (Table S1).
The primary Sj€ ogren's syndrome and comparison cohorts were followed from the Sj€ ogren's syndrome diagnosis date until the first cardiovascular event, death, emigration, or 31 December 2013, whichever came first.

Statistical analyses
The aim of this study was to investigate the risk of incident cardiovascular disease in patients with primary Sj€ ogren's syndrome. Therefore, participants with a record of MI, cerebral infarction or VTE before the Sj€ ogren's syndrome diagnosis date were excluded from the respective follow-up. Differences in the proportions of individuals with prior cardiovascular disease were assessed using a conditional logistic regression conditioned on the matching factors.
Crude incidence rates are expressed as events per 1000 person-years. Confidence intervals were estimated for incidence rates by assuming that the number of events followed a Poisson distribution. To compare the risk of incident cardiovascular disease in patients with Sj€ ogren's syndrome with that in individuals in the comparison cohort, Cox proportional hazard models were used to estimate the hazard ratio (HR) and 95% confidence interval (CI), using time since Sj€ ogren's syndrome diagnosis as time scale. Proportionality of the hazards was tested using scaled Schoenfeld residuals. No evidence of departure from this assumption was observed for the outcomes considered in this study. In addition, incidence and HR were assessed separately by time since Sj€ ogren´s syndrome diagnosis (0 to <5, 5 to <10, ≥10 years), age (<50, 50 to 70, >70 years), and by the presence of Ro/SSA and La/SSB autoantibodies.
All analyses were performed using STATA/MP version 13.0 (StataCorp LP, College Station, TX, USA). Statistical significance was defined by an alpha level of 0.05.

Demographic and clinical characteristics of the cohort
Of the included 960 patients with primary Sj€ ogren's syndrome, 391 (41%) were seropositive for both Ro/ SSA and La/SSB autoantibodies (SSA/SSB doublepositive), 278 (29%) were positive for only one of either Ro/SSA or La/SSB antibodies (SSA/SSB single-positive), and 274 (29%) tested negative for both antibodies (SSA/SSB-negative) ( Table 1). Seventeen patients did not have any available record on autoantibody status. The mean age at diagnosis was 55 years; SSA/SSB double-positive patients were diagnosed on average 4.8 and 7.3 years earlier in life than SSA/SSB single-positive (P < 0.0001) and SSA/SSB-negative patients (P < 0.0001), respectively. The median time of follow-up (interquartile range) was 9.5 (4.5-14.6) years in the Sj€ ogren's syndrome patient cohort and 9.5 (4.5-15.5) years in the comparison cohort.

Cerebral infarction
The incidence rate of cerebral infarction was 3.6 (95% CI 2.6-5.0) per 1000 person-years in patients with Sj€ ogren's syndrome and 3.0 (95% CI 2.7-3.4) in the comparison cohort, associating with a hazard ratio of 1.2 (95% CI 0.9-1.7) ( Table 2). Stratifying based on disease duration, a significantly increased risk was observed in patients after ≥10 years of follow-up (HR 1.6, 95% CI 1.0-2.7) (Fig. 2a). No cerebral infarction was observed in patients younger than 50 years. The incidence rate of cerebral infarction increased with age, yet with no corresponding increase in observed relative risk in the aggregated group of patients (Fig. 2b, Table S2).
When stratifying by autoantibody profile, SSA/SSB double-positive patients displayed a significantly increased risk for cerebral infarction (HR 1.7, 95% CI 1.0-2.9), which was not observed in patients single positive or negative for these autoantibodies ( Table 2). The increased risk in the SSA/SSB     Table S2).

Venous thromboembolism
Overall, patients with Sj€ ogren's syndrome had a twofold higher risk of VTE (HR 2.1, 95% CI 1.6-2.9) compared to the general population controls, with incidence rates of 5.4 (95% CI 4.1-7.1) and 2.6 (95% CI 2.3-2.9) per 1000 person-years, respectively ( Table 2). An increased risk of VTE was observed already during the first five years after Sj€ ogren's syndrome diagnosis (HR 2.1, 95% CI 1.2-3.5), as well as between 5 to <10 years (HR 2.8, 95% CI 1.6-4.9), and ≥10 years after diagnosis (HR 1.8, 95% CI 1.1-2.9) (Fig. 3a). The incidence rate of VTE in patients with Sj€ ogren's syndrome increased with age, and was consistently higher than that in controls across all ages, with the highest relative risk observed in patients younger than 50 years (HR 4.4, 95% CI 1.5-12.6) (Fig. 3b). Moreover, patients with primary Sj€ ogren's syndrome were significantly more likely to have a history of VTE at the time of diagnosis compared to controls (Table S4).
Separate analyses of pulmonary embolism and DVT resulted in hazard ratios similar to those of VTE. Similar to VTE, the highest relative risk of pulmonary embolism and DVT was observed in SSA/SSB double-positive patients (Table S5).

Discussion
We here report that individuals with primary Sj€ ogren's syndrome have a significantly higher incidence rate of cardiovascular disease in the form of MI, cerebral infarction and VTE. The most striking excess risk was of VTE, for which the aggregated group of patients with Sj€ ogren's syndrome displayed a twofold increased risk compared to the general population. This estimate is in close proximity with that of a recent metaanalysis reviewing four studies, calculating a pooled risk-ratio of 2.05 [29]. Notably, however, by stratifying patients according to autoantibody status, we found that the risk of VTE was in fact substantially higher in patients with SSA and SSB autoantibodies, reaching hazard ratios approaching 5 during the years following  diagnosis, while not significant for patients with only one antibody specificity or lacking these autoantibodies. The risk of VTE in patients with primary Sj€ ogren's syndrome has been suggested to relate chronic inflammation, which increases coagulability [30], and is also observed in other systemic autoimmune diseases [21,28,29,31]. Moreover, type I IFN, the expression of which can be induced by Ro/SSA and La/SSB autoantibodies, is known to exert multiple adverse effects on the vasculature [11,12,[32][33][34][35]. The degree and character of the systemic inflammation in SSA/SSB double-positive patients may thus explain the high risk of VTE in this subgroup of patients.
We also observed that Sj€ ogren's syndrome was associated with an increased risk of arterial ischaemic diseases. Overall, the patients had a 1.6-fold higher relative risk of MI compared to the general population. This estimate is consistent with two previous studies of population-based cohorts in Sweden and Taiwan, estimating relative risks of 1.6 and 1.2, respectively [19,20]. By contrast, another population-based study in Taiwan did not note an increased risk [18]. This observation may however be explained by the short median follow-up time of 3.7 years. Indeed, we did not observe an increased risk of MI during the first 5 years following the Sj€ ogren's syndrome diagnosis in our cohort. Of note, no MI events were observed in patients under the age of 50 in our cohort, contrasting to RA and SLE for which increased risk is observed also in younger patients [27,36].
Cerebral infarction occurred significantly more frequently in patients with Sj€ ogren's syndrome than in controls after 10 or more years from diagnosis. The risk was confined to SSA/SSB double-positive patients, and in this group reached a HR of 1.6. The observed overall relative risk in our study was similar to the corresponding estimate of 1.3 reported by Z€ oller et al., which is slightly lower than the corresponding risk in RA and SLE [22].
Higher frequencies of subclinical atherosclerosis or endothelial dysfunction have been reported in patients with primary Sj€ ogren's syndrome, as determined by carotid intima-media thickness, ankle-brachial index or endothelium-dependent flow mediated or nitrate-mediated vasodilation [37][38][39][40]. Moreover, disease duration appears to be an important factor in the pathogenesis. Rachapalli et al. [41] demonstrated that the ankle-brachial index was significantly reduced only in patients with a disease duration of more than 10 years. These findings are consistent with our study, in which the relative risk of both MI and cerebral infarction increased with disease duration. In most cases of arterial ischaemic cardiovascular disease, the underlying cause is atherosclerosis, which is widely considered a chronic inflammatory disease [42]. Inflammation is increasingly recognized as playing a key role in cardiovascular disease development, and may be the common mechanism underlying the effect of many traditional risk factors [43]. Interestingly, Ro/SSA and La/SSB antibodies, which are associated with an activated inflammatory type I IFN system, have been reported to correlate positively with subclinical atherosclerosis and endothelial dysfunction [37,38,40]. These reports are thus consistent with our findings of higher risks of cerebral arterial events in Ro/SSA-and La/SSB-positive patients. However, intra-group variations with regards to reaching statistical significance were observed in the analyses. This may partially relate to the limited number of events in each substratum, and calls for caution in interpreting individual findings. Nevertheless, it does not preclude the joint conclusion that the presence of these antibodies associates with an increased risk.
The study has limitations to consider. We were unable to control for lifestyle factors and comorbid disorders potentially influencing the risk of cardiovascular disease as previous studies have suggested that, patients with Sj€ ogren's syndrome have an increased frequency of traditional risk factors associated with cardiovascular disease, including hypertension, hypercholesterolaemia and hypertriglyceridaemia [16,23,44], and higher prevalence of subclinical atherosclerosis and endothelial dysfunction [39]. However, our matching based on age, sex and geographical region may mitigate confounding effects from lifestyle factors, and adjustment for comorbid disorders has had marginal impact in previous studies [17][18][19]. The lack of data to account for the presence of antiphospholipid antibodies constitutes another limitation [45,46]. Further, we had no information on underlying treatment, which may potentially influence the risk of cardiovascular disease [47].
The strengths of our study relate to the contextually large size of the cohorts, the long follow-up and that cardiovascular events were identified using the NPR, which has been confirmed to hold high validity [26]. Most importantly though, we included only patients fulfilling AECG criteria [25] and through our clinical characterization could stratify analyses based on Ro/SSA and La/SSB autoantibodies. The presence or absence of these autoantibodies mark two genetically and clinically distinct subgroups, and therefore constitutes an important factor to take into account [48].
Investigations into the mechanisms behind the increased risk of cardiovascular disease in patients with primary Sj€ ogren's syndrome should involve the assessment of both traditional and diseasespecific risk factors. Previous studies of traditional risk factors have mostly not been able to stratify patients according to antibody positivity. A study of the impact of smoking and lifestyle factors using the same cohort of patients with primary Sj€ ogren's syndrome, and hence information on antibody status, is currently underway. Patients positive for SSA/SSB more often present with organ involvement, leukopenia, hypergammaglobulinaemia and high disease activity measured by the European League Against Rheumatism Disease Activity Index (ESSDAI), compared with antibodynegative patients [49,50]. It is plausible that high disease activity confers an increased risk for cardiovascular disease. While CRP is not a good marker for disease activity in primary Sj€ ogren's syndrome, a prospective study assessing clinical and laboratory manifestations as well as ESSDAI at diagnosis and follow-up would improve the understanding of disease-related risk factors. Investigation of the IFN signature, pro-inflammatory cytokines and biomarkers of endothelial damage as reviewed in Valim et al. [39], measured at diagnosis and during the disease course, stratified by SSA/SSB antibodies, would hopefully elucidate some of the immunological disease mechanisms behind the increased risk.

Conclusion
In this cohort study, we found that patients with primary Sj€ ogren's syndrome have a considerably increased risk of cardiovascular disease of both venous and arterial origin. We further add to the current knowledge by demonstrating that positivity for both Ro/SSA and La/SSB autoantibodies mark the subgroup of patients that carry most of the risk of cerebral infarction and VTE, highlighting the importance of taking patient subgroups into account for correctly defining comorbidity risks.
The observations suggest that monitoring and prevention of cardiovascular disease in patients with primary Sj€ ogren's syndrome should be considered, and that Ro/SSA and La/SSB autoantibodies may be used as biomarkers to identify the group of patients with the greater need.

Supporting Information
Additional Supporting Information may be found in the online version of this article: Table S1. ICD codes and registers used to identify cardiovascular events. Table S2. Risk of myocardial infarction, cerebral infarction, and venous thromboembolism in prevalent primary Sj€ ogren's syndrome, stratified by SSA and SSB autoantibodies and age.