Graduate Institute of Clinical Medicine Science and School of Medicine, College of Medicine, China Medical University, Taichung, Taiwan
Department of Nuclear Medicine and PET Center, China Medical University Hospital, Taichung, Taiwan
Correspondence: Chia-Hung Kao, Graduate Institute of Clinical Medicine Science and School of Medicine, College of Medicine, China Medical University, Taichung, Taiwan, No. 2, Yuh-Der Road, Taichung 404, Taiwan.
Studies on the risks of deep vein thrombosis (DVT) and pulmonary embolism (PE) in patients with systemic lupus erythematosus (SLE) are limited. We evaluated the effects of SLE on the risks of developing DVT and PE in a nationwide, population-based cohort study in Taiwan.
We randomly selected patients without SLE from the National Health Insurance database (N =23.74 million), and frequency-matched four of them, on the basis of age, sex, and index year, to each SLE patient in the catastrophic illness registry of the NHI who was diagnosed with SLE between 1998 and 2008. Using a follow-up period ending in 2010, we analyzed the risks of DVT and PE with a Cox proportional-hazards regression analysis.
The 13 084 SLE patients (87.9% women; mean age of 35.6 years) and 52 336 controls were followed for 90 237 and 379 185 person-years, respectively. After adjustment for age, sex, and comorbidities, the SLE patients' risks of developing DVT and PE were 12.8-fold and 19.7-fold higher, respectively, than those of the comparison cohort. The risks of DVT and PE increased in both study groups when the data were stratified on the basis of sex, age, and comorbidities. The SLE patients aged ≤ 35 years had the highest risks of developing DVT and PE. The multiplicative increased risks of DVT and PE were also significant in SLE patients with any comorbidity.
The risks of DVT and PE are significantly higher in SLE patients than in the general population.
Systemic lupus erythematosus (SLE) is a chronic autoimmune, multiorgan inflammatory disorder that primarily affects young women, but as many as 20% of SLE patients are ≥ 50 years. Affecting nearly every organ system in the body, SLE symptoms vary widely in severity, from mild to progressively more serious symptoms that can result in fatal complications. The average annual incidence and average prevalence of SLE in Taiwan were 4.87 and 97.5 per 100 000 persons, respectively, between 2003 and 2008 .
Deep vein thrombosis (DVT) and pulmonary embolism (PE) are manifestations of potentially lethal venous thromboembolism (VTE). Patients affected by DVT most often present with swelling, discoloration and discomfort of the affected leg. Clinicians can diagnose DVT by using non-invasive venous ultrasound, venography, magnetic resonance imaging (MRI), or pathology of the thrombus following its removal during surgery. With concomitant PE, DVT can be a dangerous condition . Patients with PE may present with hypoxemia and/or dyspnea, chest pain, and anxiety or nervousness. The diagnosis of PE can be made on the basis of the results of the D-dimer blood test, lung scan, spiral computed tomography scan, pulmonary angiography, MRI, or pathology of the thrombus following its removal during surgery .
According to the Virchow triad, VTE results from altered blood coagulation, stasis, or abnormalities in the vessel wall [4, 5]. Hypercoagulable states and endothelial dysfunction may result from multiple interactions between inherited and acquired risk factors. Several studies have identified age, immobilization following cerebrovascular accidents (CVAs), heart failure, lower leg fracture, surgery, diabetes and cancer as acquired risk factors for VTE [6-13]. Some studies have indicated that autoimmune disorders, such as inflammatory bowel disease , rheumatoid arthritis , and Wegener granulomatosis , are associated with an increased risk of VTE. Inflammation is a common feature of the autoimmune diseases, and may drive arterial, venous and microvascular thrombosis .
Mok et al.  found an 11.9-fold higher risk of VTE in SLE patients than in the general population. Weng et al.  reported a 0.8% incidence of pulmonary infarction in patients with SLE. However, the generalizability of these findings of Mok et al. and Weng et al. is limited, because of the relatively small samples used in their studies. Aviña-Zubieta et al.  demonstrated increased risks of DVT and PE in a population-based SLE cohort in Canada. Therefore, we investigated whether the risks of DVT and PE are higher among SLE patients in a nationwide, population-based cohort in Taiwan.
The National Health Insurance (NHI) program was initiated in 1995, and covers ~ 99% of the population of Taiwan (23.74 million people) . The NHI reimburses patients for the costs of outpatient, inpatient, emergency and traditional Chinese medical services, including the costs of prescription medications. We used the longitudinal health insurance database (LHID) of the NHI, which is managed and released for research purposes by the National Health Research Institute. The hospitalization data obtained from the LHID contained information regarding sex, date of birth, encrypted patient identification numbers, dates of admission and discharge, discharge diagnoses (up to five), surgical procedures (up to five), and discharge status. The diagnoses and procedures recorded in the LHID are coded according to the criteria of the International Classification of Disease, Ninth Revision, Clinical Modification (ICD-9-CM). We ensured that all data were deidentified and analyzed anonymously. In addition, this study was also approved by the Ethics Review Board of China Medical University (CMU-REC-101-012).
Patients newly diagnosed with SLE (ICD-9-CM 710.0) between 1998 and 2008 were eligible for enrollment in our study. To ensure the reliability of SLE cases, we selected only newly diagnosed SLE patients (n = 13 319) from the Registry for Catastrophic Illness Patient Database (RCIPD). In the NHI system, SLE patients receive a catastrophic illness certificate and are exempt from copayment for their SLE-related medical care. The SLE patients with missing data regarding date of birth or sex (n = 36) and those with a history of DVT (ICD-9-CM 453.8) or PE (ICD-9-CM 415.1) (n = 199) before the index date were excluded from our study. For each SLE patient, four controls without SLE during the study period were frequency-matched to each of the SLE patients on the basis of age, sex, and the index year, to serve as the comparison cohort. The same exclusion criteria were applied to non-SLE controls.
Diagnostic criteria for SLE
The currently accepted diagnostic criteria for SLE were established in 1971, revised in 1982, and updated to the classification of SLE by the American College of Rheumatology in 1997 , and include the following 11 criteria: malar rash; discoid rash; photosensitivity; oral ulcers; non-erosive arthritis; pleuritis or pericarditis; renal disorder, persistent proteinuria, or cellular casts in the urine; neurologic disorder, seizures, or psychosis; hematologic disorder, including hemolytic anemia, leukopenia, lymphopenia, or thrombocytopenia; the production of anti-nuclear antibody (Ab); and other immunologic disorders, including the production of anti-DNA, anti-Smith or anti-phospholipid Abs (Table S1). A person can receive a diagnosis of SLE if four or more of the 11 criteria are observed in series or simultaneously during any period of observation. In Taiwan, patients who receive a diagnosis of SLE on the basis of these criteria can receive a catastrophic illness certificate from the Bureau of National Health Insurance.
Outcome measurement and comorbidities
The primary outcomes were newly diagnosed DVT or PE from hospitalization records. All of the patients were followed from the index date to the date of a primary outcome, withdrawal from the NHI program, or the end of 2010, whichever came first. Nearly all patients with DVT or PE underwent a comprehensive examination before receiving intensive care. In Taiwan, DVT and PE patients’ medical reimbursements and discharge notes are scrutinized in a peer-review process. The pre-existing comorbidities included atrial fibrillation (ICD-9-CM 427.31), hypertension (ICD-9-CM 401–405), diabetes (ICD-9-CM 250), hyperlipidemia (ICD-9-CM 272), CVA (ICD-9-CM 430–438), heart failure (ICD-9-CM 428), lower leg fracture or surgery (ICD-9-CM 820–823 and procedure codes 81.51, 81.52, 81.53, and 81.54), and cancer (ICD-9-CM 140–208).
We compared the distributions of the baseline characteristics between the SLE and comparison cohorts by using a chi-square test. The incidence rate ratios (IRRs) and 95% confidence intervals (CIs) for DVT and PE were estimated for each study group. The IRR was determined on the basis of the Poisson assumption. To estimate the cumulative incidence of DVT and PE risks, we performed a survival analysis of both cohorts by using the Kaplan–Meier method, and the significance of the results was assessed with the log-rank test. Multivariate Cox proportional-hazards regression models were used to calculate the hazard ratios (HRs), with stratification based on age, sex, the presence of comorbidity, and follow-up duration. All statistical analyses were performed with sas Version 9.2 (SAS Institute, Cary, NC, USA). A P-value of < 0.05 was considered to indicate a statistically significant result.
Demographic characteristics and comorbidities
A total of 13 084 SLE patients and 52 336 controls without SLE were enrolled in our study. The mean ages of the comparison and SLE cohorts were 35.4 ± 16.4 years and 35.6 ± 16.1 years, respectively, and 54.1% of the patients were ≤ 35 years of age (Table 1). Females accounted for 87.9% of the patients. Table 1 presents a summary of the baseline comorbidities of the SLE and comparison cohorts. The SLE cohort included more cases of atrial fibrillation (0.52 vs. 0.15%, P <0.0001), hypertension (7.59 vs. 2.31%, P <0.0001), diabetes (2.15 vs. 1.49%, P <0.0001), hyperlipidemia (2.13 vs. 0.70%, P <0.0001), CVA (2.45 vs. 0.90%, P <0.0001), heart failure (1.88 vs. 0.30%, P <0.0001) and cancer (1.00 vs. 0.65%, P <0.0001) than did the comparison cohort.
Table 1. Demographic characteristics and comorbidity between systemic lupus erythematosus (SLE) patients and controls
N = 52 336
N = 13 084
CVA, cerebrovascular accident; SD, standard deviation. Chi-square test. *Two-sample t-test.
Sex, n (%)
46 012 (87.9)
11 503 (87.9)
Age (years), mean (SD)*
Stratification by age (years), n (%)
28 336 (54.1)
14 604 (27.9)
Comorbidity, n (%)
Lower leg fracture or surgery
Comparison of the risks of DVT and PE stratified by sex, age, and comorbidity
For DVT, the median follow-up for the SLE cohort was 6.92 years, and that for comparison cohort was 7.29 years. In total, 136 SLE patients were diagnosed with DVT, yielding an incidence of 15.1 per 10 000 person-years, whereas 43 cases of DVT occurred in the comparison cohort, yielding an incidence of 1.13 per 10 000 person-years (Table 2). The IRR for DVT in the SLE cohort was 13.3 (95% CI12.5–14.2), and the adjusted HR for DVT was 12.8 (95% CI9.06–32.8), indicating that the SLE patients were at increased risk of DVT. The highest HRs were observed for women with SLE (adjusted HR14.9, 95% CI10.0–22.3) and the SLE patients aged ≤ 35 years (adjusted HR52.8, 95% CI23.0–121.3). The SLE patients without comorbidity had a 14.9-fold higher risk of DVT than the patients in the comparison cohort (adjusted HR14.9, 95% CI10.2–21.8).
Table 2. Comparison of deep vein thrombosis (DVT) and pulmonary embolism (PE) development stratified by sex, age and comorbidity between systemic lupus erythematosus (SLE) patients and controls
CI, confidence interval; HR, hazard ratio; IRR, incidence rate ratio; PY, person-years. *Rate: incidence rate per 10 000 person-years. †Adjusted HR: multivariable analysis including age, sex, and comorbidity. ‡Comorbidity: one of the comorbidities used to classify the comorbidity group.
Table 2 also shows the PE incidence densities and the adjusted HR for PE in the SLE and comparison cohorts. For PE, the median follow-up for the SLE cohort was 6.96 years, and that for the comparison cohort was 7.29 years. PE occurred more frequently in the SLE patients than in the controls (10.2 vs. 0.47 per 10 000 person-years), with an IRR of 21.4 (95% CI19.9–23.1) and an adjusted HR of 19.7 (95% CI11.9–32.8). In addition, the highest adjusted HRs were observed for women with SLE (adjusted HR21.2, 95% CI12.2–36.9) and the SLE patients aged ≤ 35 years (adjusted HR71.5, 95% CI22.3–228.9). The SLE patients without comorbidity had a 26.9-fold higher risk of PE than the controls (adjusted HR26.9, 95% CI14.6–49.5). Figure 1 shows the results of the comparison of the cumulative incidence rates of DVT and PE between the SLE and comparison cohorts. The incidence rates of DVT (Fig. 1A) and PE (Fig. 1B) were significantly higher in the SLE patients than in the comparison cohort (P <0.001 for both; log-rank test).
The effect of the interaction between comorbidity and SLE on the risks of DVT and PE development
Table 3 shows that the SLE patients with comorbidity had a significantly higher risk of DVT (adjusted HR 22.1, 95% CI13.0–37.6) than the comparison cohort with comorbidity (adjusted HR4.4, 95% CI1.98–9.78) or without comorbidity (reference). The risk of PE was also significantly higher among the SLE patients with comorbidity (adjusted HR66.3, 95% CI32.2–136.3) than in the comparison cohort with comorbidity (adjusted HR13.4, 95% CI4.81–37.3) or without comorbidity (reference).
Table 3. Cox proportional-hazards regression analysis for the risks of systemic lupus erythematosus (SLE)-associated deep vein thrombosis (DVT) and pulmonary embolism (PE) with interaction of comorbidity
CI, confidence interval; HR, hazard ratio. *P-value for interaction. †Adjusted HR: adjusted for age and sex.
Trends of DVT and PE event risk stratified by follow-up years
Table 4 shows the results of the comparison of the risks of DVT and PE stratified according to follow-up duration. During the first year of follow-up, the adjusted HRs for DVT and PE were 66.7 (95% CI16.0–278.7) and 91.5 (95% CI12.3–677.4), respectively. However, the adjusted HRs for DVT and PE decreased to 5.76 (95% CI2.68–12.3) and 6.66 (95% CI2.15–20.6), respectively, for patients who were followed up for > 8 years.
Table 4. Trends of deep vein thrombosis (DVT) and pulmonary embolism (PE) event risk stratified by duration of follow-up
The average prevalence of SLE was 97.5 per 100 000 persons in Taiwan in 2003–2008 , which is higher than the prevalence rates of 20 and 45 per 100 000 reported in the USA and Canada, respectively [23-25]. Our results show that the SLE patients had a 12.8-fold higher adjusted HR for DVT and a 19.7-fold higher adjusted HR for PE than the general population in Taiwan. Aviña-Zubieta et al.  reported that, in Canada, SLE patients had relative risks of 4.1 for DVT and 4.6 for PE, as compared with the general population, after adjustment for covariates. Although the underlying cause of the risk disparities remains unclear, it might be associated with racial differences and environmental factors [26, 27]. The fact that the patients in our study had been admitted to the hospitals might indicate more severe cases, in which SLE with acute illness and high disease activity may have a higher risk of thrombosis than that in patients with mild disease activity.
In Taiwan, the incidence and prevalence of SLE patients presenting with anti-phospholipid syndrome (APS) are high [19, 28]. The APS is an Ab-mediated prothrombotic disorder associated with blood clots in both arteries and veins . However, APS alone cannot explain the increased thrombotic risk in SLE. Chronic inflammation and other contributing thrombovascular risk factors also play vital roles in subsequent DVT and PE development. Inflammation modulates thrombotic responses by upregulating the activity of procoagulants, downregulating that of anticoagulants, and suppressing fibrinolysis . Because of periodic episodes of inflammation in multiple organ systems and a hypercoagulable status, SLE leads to arterial, venous and microvascular thromboses [30, 31]. The highest risks of DVT and PE in the SLE patients in our study occurred during the first year of follow-up. Although the risks decreased over time, the overall risk remained higher for the SLE patients than for the controls for up to 8 years of follow-up. Our findings are consistent with those of previous studies showing an association between SLE and an increased risk of venous thrombosis [20, 32].
Although the SLE patients in our study had a higher prevalence of comorbidities associated with the development of DVT and PE than did the comparison cohort, SLE remained an independent risk factor for the development of DVT and PE after adjustment for the covariates. The risks of DVT and PE increased in both men and women with SLE when age and comorbidities were controlled for. Furthermore, the SLE patients had higher risks of DVT and PE than did the comparison cohort, regardless of age, after adjustment for the covariates. However, the incidence of SLE was especially high in women aged ≤ 35 years, and the younger SLE patients had the highest risk of DVT and PE. Thus, clinicians should be especially vigilant regarding the treatment of younger women with SLE. Aspirin and antimalarial medications may have a protective thrombotic effect in these SLE patients .
Our study shows that the incidence of DVT and PE increased in patients with comorbidity in both cohorts. The multiplicative increased risks of developing DVT and PE were also significant in patients with concomitant SLE and any comorbidity. These results are robust according to Cox proportional-hazards regression analysis for the increased risks of DVT and PE in SLE with interaction of any comorbidity, as shown in Table 3.
The strength of our findings is the nationwide, population-based design of our longitudinal study on the risks of DVT and PE in Asian SLE patients. In addition, the diagnoses of all of the SLE cases identified in the NHIR database claims data were confirmed through their inclusion in the RCIPD, which establishes a high level of reliability for our data. Because each resident in Taiwan is assigned a unique personal identification number, every patient could be traced through the records of the NHI for the entire follow-up period. Thus, our findings can be generalized to the entire population of Taiwan.
However, several limitations to the interpretation of our findings should be considered. A higher proportion of SLE patients than of controls were hospitalized, which might have resulted in an overestimate of the risks of DVT and PE. Ramagopalan et al.  suggested that people admitted to hospitals for autoimmune disorders may be at increased risk of subsequent VTE. However, our comparison cohort also included inpatients, which may have reduced the potential for overestimating the risk, because VTE is a common complication during and after hospitalization for acute medical illness or surgery . Although the healthcare claims data may contain potential a misclassification bias concerning primary outcomes, the auditing mechanism of the Bureau of National Health Insurance can help to minimize the diagnostic uncertainty and misclassification . The lack of drug and laboratory data, such as those for hormone replacement therapy, the use of contraceptive drugs, glucocorticosteroid treatments, and the levels of autoantibodies, may also have influenced the primary outcomes in our study.
In conclusion, our nationwide, population-based cohort study examining 13 084 SLE patients with a follow-up period of ~ 90 400 person-years shows that SLE patients have a 12.8-fold increased risk of DVT and a 19.7-fold increased risk of PE as compared with such risks in the general population. These findings highlight the importance of a multidisciplinary approach to the management of the potential risk factors for DVT and PE among SLE patients. Future studies on the biological mechanisms of SLE and their effect on the development of DVT and PE, as well as on the autoantibodies of the SLE patients and the risk of DVT and PE development (such as renal disease, chronic inflammation, and medication), are encouraged and warranted.
W.-S. Chung and C.-H. Kao: conception and design. C.-L. Lin: administrative support. W.-S. Chung, C.-L. Lin, and C.-H. Kao: data analysis and interpretation. All authors: collection and assembly of data, manuscript writing, and final approval of manuscript.
This work was supported by the study projects (DMR-103-012) in our hospital: Taiwan Ministry of Health and Welfare Clinical Trial and Research Center for Excellence (DOH102-TD-B-111-004), Taiwan Ministry of Health and Welfare Cancer Research Center for Excellence (MOHW103-TD-B-111-03), and International Research-Intensive Centers of Excellence in Taiwan (I-RiCE) (NSC101-2911-I-002-303). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. No additional external funding was received for this study.
Disclosure of Conflict of Interests
The authors state that they have no conflict of interest.