SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

Objective

To compare the incidence and risk factors for thromboembolic events in systemic lupus erythematosus (SLE) patients of different ethnic backgrounds.

Methods

SLE patients who were newly diagnosed or were referred within 6 months of diagnosis between 1996 and 2002 were prospectively followed up for the occurrence of thromboembolic events. Cumulative hazard and risk factors for thromboembolism were evaluated and compared among patients of different ethnic origins.

Results

We studied 625 patients who fulfilled the American College of Rheumatology criteria for SLE (89% women): 258 Chinese, 140 African Americans, and 227 Caucasians. The mean ± SD age at SLE diagnosis was 35.7 ± 14 years. After a followup of 3,094 patient-years, 48 arterial events and 40 venous events occurred in 83 patients. The overall incidence of arterial and venous thromboembolism was 16/1,000 patient-years and 13/1,000 patient-years, respectively. The cumulative hazard of arterial events at 60 months after the diagnosis of SLE was 8.5%, 8.1%, and 5.1% for the Chinese, African Americans, and Caucasians, respectively. The corresponding cumulative risk of venous events was 3.7%, 6.6%, and 10.3%, respectively (P = 0.008 for Chinese versus Caucasians, by log rank test). Smoking, obesity, antiphospholipid antibodies, and use of antimalarial agents and exogenous estrogens were less frequent in the Chinese patients. In Cox regression models, low levels of high-density lipoprotein (HDL) cholesterol, Chinese ethnicity, oral ulcers, and serositis predicted arterial events, whereas male sex, low levels of HDL cholesterol, antiphospholipid antibodies, non-Chinese ethnicity, obesity, renal disease, and hemolytic anemia predicted venous events.

Conclusion

There are ethnic differences in the incidence of arterial and venous thromboembolism in patients with SLE that cannot be fully explained by the clinical factors studied. Further evaluation of other genetic and immunologic factors is warranted.

Premature atherosclerosis is a major cause of mortality and late morbidity in patients with systemic lupus erythematosus (SLE) (1–3). In several cohort studies, it was found that atherosclerotic cardiovascular and cerebrovascular diseases are more common causes of late deaths than active SLE itself (1, 4–6). More recent studies have demonstrated that subclinical coronary heart disease and carotid plaque were present in a significantly higher proportion of SLE patients than in age- and sex-matched control subjects with similar risk factors (7, 8). The etiology of accelerated atherosclerosis in SLE is multifactorial and cannot be fully explained by traditional risk factors (9).

Compared with individuals without SLE, the risk of myocardial infarction in SLE patients is 2–50 times higher, and the risk of stroke is 2–10 times higher (2, 9, 10). The prevalence of symptomatic coronary heart disease in SLE patients has been reported to be 6–20%, depending on the characteristics of the cohort, disease duration, study design, prevalence of antiphospholipid antibodies (aPL), and ethnic composition (11). The frequency of stroke is less well studied, but it has been reported that 3–15% of SLE patients have a nonfatal stroke (2, 12, 13).

A number of risk factors for symptomatic coronary heart disease and stroke in SLE have been identified in case–control studies. These risk factors include age at SLE diagnosis (10, 14, 15), disease duration (10, 14), hypertension (15–17), dyslipidemia (10, 15, 17, 18), hyperhomocysteinemia (18, 19), oxidized low-density lipoprotein (LDL) (18), smoking (14, 16), use of corticosteroids and their duration (10, 15, 18), aPL (14, 18, 20, 21), and valvular abnormalities of the heart (22). Other factors that may be relevant to arterial or venous thromboembolism in SLE are genetic mutations (23), pregnancy, and the use of exogenous estrogens (24) for contraception, climacteric symptoms, and control of cyclic SLE activity. The use of hydroxychloroquine in SLE, on the other hand, may be protective (25).

It is well established that aPL are associated with both arterial and venous thrombosis. Recent studies have also shown that the antiphospholipid syndrome with thrombotic complications in SLE is a major adverse factor for irreversible organ damage and survival (26). In large cohort studies, the lupus anticoagulant has been shown to be a significant risk factor for myocardial infarction (20) and stroke (27) in both Caucasian and non-Caucasian patients with SLE.

While most reported studies have involved Caucasian patients and have focused on arterial thromboembolism, data regarding the risk of venous thromboembolism, particularly in non-Caucasian SLE patients, are scant. The present prospective study was undertaken to determine the incidence and risk factors for arterial and venous thromboembolic events in Chinese and African American SLE patients as compared with those in Caucasian SLE patients.

PATIENTS AND METHODS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

Patient population.

We studied SLE patients who were newly diagnosed or were referred within 6 months of diagnosis to the departments of medicine of the Tuen Mun and Princess Margaret hospitals (Hong Kong) and the John Hopkins University School of Medicine (Baltimore) between 1996 and 2002. Patients had to fulfill at least 4 of the American College of Rheumatology (ACR) criteria for the classification of SLE (28) at the time of diagnosis. Patients were identified and retrieved from the Hospital Information Retrieval System and the Hopkins Lupus Cohort databases, which are maintained by the respective units. Patients were followed up prospectively for the occurrence of thromboembolic events. Patients of different ethnic origins were compared.

Data collection and methods.

Data on age, sex, ethnicity, date of SLE diagnosis, cumulative clinical features at the last clinic visit, and autoantibodies were collected. Risk factors for thromboembolism at any time since the onset of SLE were recorded, including body mass index (BMI), history of smoking, lipid profile, diabetes mellitus, hypertension requiring therapy, nephrotic syndrome, impaired renal function (serum creatinine level ≥1.5 mg/dl), presence of aPL (anticardiolipin antibody or the lupus anticoagulant), ever use of medications since the diagnosis of SLE, such as exogenous estrogens (oral contraceptives and hormone replacement therapy), high-dose corticosteroids (defined as prednisolone dosage of ≥0.8 mg/kg/day for ≥8 weeks), azathioprine (AZA), cyclophosphamide (CYC), and hydroxychloroquine (HCQ). Patients were followed up prospectively by physicians of the respective units at 6–12-week intervals. More frequent followup was arranged for those with active disease or complications.

Definitions and ascertainment of thromboembolic events.

Arterial thromboembolic events consisted of signs and symptoms caused by partial or complete stenosis of major arteries. These signs and symptoms included cerebrovascular accident (stroke), transient ischemic attack (TIA), angina pectoris, acute myocardial infarction (MI), gangrene of the fingers or toes, intermittent claudication, and retinal artery occlusion. Venous thromboembolic events were defined as syndromes caused by complete occlusion of major venous vessels. These included calf and femoral vein thrombosis, pulmonary embolism, retinal vein occlusion, hepatic vein thrombosis, and superficial thrombophlebitis documented by clinical findings as well as by Doppler ultrasound, ventilation/perfusion lung scan, spiral computed tomography scan, or angiography.

Thromboembolic events were diagnosed at the participating centers. Events were validated by research fellows who have worked in both centers using the same criteria throughout.

Statistical analysis.

Comparison of categorical data among the 3 groups of patients was made by the chi-square test. Comparison of continuous data was performed by one-way analysis of variance. Post hoc correction for multiple comparisons was performed with the Tukey test for unequal samples.

The cumulative probability over time of arterial and venous thromboembolic events in the study patients was calculated by the Kaplan-Meier method. The dates of the SLE diagnosis (time zero) and the first occurrence of arterial or venous thromboembolism were noted. Thromboembolic events that occurred before the diagnosis of SLE were excluded from analyses. Data for patients who died or were lost to followup were censored at their last clinic visits.

The following covariates were considered to be predictors of thromboembolic events: age at SLE onset, sex, history of smoking, hypertension, diabetes mellitus, BMI ≥27 kg/m2, LDL cholesterol level ≥4.1 mmoles/liter, total cholesterol level ≥6.2 mmoles/liter, high-density lipoprotein (HDL) cholesterol level ≤1.0 mmoles/liter, triglyceride level ≥5.6 mmoles/liter, persistent nephrosis or serum creatinine level ≥1.5 mg/dl for >6 months, presence of any aPL, and ever use of high-dose corticosteroids, exogenous estrogens (oral contraceptives or hormone replacement therapy), azathioprine, or cyclophosphamide. Ever use of hydroxychloroquine was considered to be a potential protective factor.

Odds ratios and 95% confidence intervals (95% CIs) for thromboembolic events based on each of the factors described above were first obtained by logistic regression models, which were constructed separately for arterial and venous events. Age, sex, ethnicity, risk factors described above, and cumulative clinical features of SLE were then entered in the Cox proportional hazards models for factors predictive of either arterial or venous thromboembolism.

P values less than 0.05 (2-tailed test) were considered significant. All statistical analyses were performed using SPSS software, version 11.5 (SPSS, Chicago, IL), for Windows Xp.

RESULTS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

Demographic characteristics and clinical manifestations.

We studied 625 SLE patients (555 women and 70 men), 258 of whom were Chinese, 227 were Caucasians, and 140 were African Americans. The mean ± SD age at diagnosis of SLE was 35.7 ± 14 years. All patients fulfilled at least 4 of the ACR criteria for SLE. Table 1 shows the cumulative clinical features of the patients since the diagnosis of SLE, categorized by ethnic group. Hemolytic anemia, thrombocytopenia, anti–double-stranded DNA and anti-Ro antibodies were more prevalent in the Chinese patients, whereas discoid skin rash, leukopenia, seizure, serositis, anti-Sm antibodies, and anti–U1 RNP antibodies were more common in the African Americans. Caucasian patients were more likely to have oral ulcers, malar rash, and photosensitivity. The prevalence of renal disease was higher in the Chinese and African Americans than that in the Caucasians.

Table 1. Cumulative clinical features since the diagnosis of SLE, by ethnic group*
 Chinese (n = 258)African Americans (n = 140)Caucasians (n = 227)P
  • *

    The manifestations of systemic lupus erythematosus (SLE) are those defined by the American College of Rheumatology 1982 criteria for the classification of SLE (28). Values are the number (%) or the mean ± SD. Anti-dsDNA = anti–double-stranded DNA.

Age, years34.6 ± 1334.7 ± 1237.7 ± 140.03
Women230 (89)128 (91)197 (87)0.38
Followup, years4.56 ± 2.45.15 ± 2.25.3 ± 2.30.001
SLE manifestations    
 Arthritis198 (77)110 (79)159 (70)0.12
 Malar erythema126 (44)50 (36)109 (52)0.03
 Discoid rash18 (7)31 (22)27 (12)<0.001
 Oral ulcers43 (17)52 (37)131 (58)<0.001
 Photosensitivity84 (33)43 (31)128 (56)<0.001
 Hemolytic anemia56 (22)13 (9)15 (7)<0.001
 Leukopenia118 (46)68 (49)82 (36)0.04
 Thrombocytopenia65 (25)22 (16)38 (17)0.02
 Seizure10 (4)16 (11)11 (5)0.01
 Psychosis10 (4)4 (3)1 (0.4)0.04
 Renal127 (49)59 (42)53 (23)<0.001
 Serositis56 (22)75 (54)90 (40)<0.001
Antibody positivity    
 Anti-dsDNA204 (79)81 (58)113 (50)<0.001
 Anti-Ro148 (57)39 (28)41 (18)<0.001
 Anti-La34 (13)17 (12)17 (7)0.12
 Anti-Sm32 (12)34 (24)22 (10)<0.001
 Anti–U1 RNP53 (21)51 (36)27 (12)<0.001
 Antiphospholipid76 (29)59 (42)105 (46)<0.001

Incidence of arterial and venous thromboembolism.

After a total followup of 3,094 patient-years, 48 arterial (65% stroke/TIA and 19% acute MI/angina pectoris) and 40 venous (80% deep vein thrombosis and/or pulmonary embolism) thromboembolic events occurred in 83 patients. The overall incidence of arterial and venous thromboembolic events was 16/1,000 patient-years (19 in Chinese, 17 in African Americans, and 12 in Caucasians) and 13/1,000 patient-years (7 in Chinese, 15 in African Americans, and 17 in Caucasians), respectively.

The cumulative hazard for arterial thromboembolic events at 60 months of SLE diagnosis was 8.5%, 8.1%, and 5.1% for the Chinese, African Americans, and Caucasians, respectively. The differences were not significant among the ethnic groups. The corresponding cumulative hazard for venous thromboembolic events was 3.7%, 6.6%, and 10.3%, respectively, which was highly significant for the Chinese versus Caucasians (P = 0.008 by log rank test) and was borderline significant for the Chinese versus African Americans (P = 0.051 by log rank test). None of the study patients, including the Chinese, were known to consume natural products or herbs that might have anticoagulant effects. Figure 1 shows the cumulative risks of venous thromboembolism in the 3 ethnic groups studied.

thumbnail image

Figure 1. Cumulative risk of venous thromboembolism in 3 ethnic groups of patients with systemic lupus erythematosus: Chinese, African Americans, and Caucasians. P = 0.008 for the Chinese group versus the Caucasian group, by log rank test.

Download figure to PowerPoint

Risk factors for arterial and venous thromboembolism.

Table 2 shows the occurrence of arterial events in all patients as a function of the risk factors studied. Older age, longer disease duration, and a serum HDL cholesterol level ≤1.0 mmoles/liter were significantly more common in patients with arterial thromboembolism. A BMI ≥27 kg/m2 was also more frequent in patients with arterial events, but statistical significance was not achieved.

Table 2. Occurrence of arterial thromboembolism as a function of the risk factors studied*
 Arterial eventsOdds ratio95% CIP
Yes (n = 48)No (n = 577)
  • *

    Values are the number (%) or the mean ± SD. 95% CI = 95% confidence interval; BMI = body mass index; LDL = low-density lipoprotein; HDL = high-density lipoprotein; aPL = antiphospholipid antibodies; HRT = hormone replacement therapy; OCs = oral contraceptives.

  • Dosage ≥0.8 mg/kg/day for ≥8 weeks.

Age, years40.0 ± 16.435.4 ± 13.31.021.00–1.050.03
Followup, years5.62 ± 2.744.90 ± 2.301.011.00–1.020.04
Women41 (85)514 (89)0.720.31–1.670.44
Race     
 Chinese22 (46)236 (41)1.220.68–2.210.51
 African American12 (25)128 (22)1.170.59–2.310.65
 Caucasian14 (29)213 (37)0.700.37–1.340.29
Smoking, ever13 (27)151 (26)1.040.54–2.020.90
Hypertension19 (40)186 (32)1.380.75–2.520.30
Diabetes mellitus2 (4)34 (6)0.690.16–2.980.62
BMI ≥27 kg/m225 (52)224 (39)1.710.95–3.090.07
LDL cholesterol ≥4.1 mmoles/liter6 (13)57 (10)1.300.53–3.200.57
HDL cholesterol ≤1.0 mmoles/liter12 (25)55 (10)3.191.57–6.510.001
Total cholesterol ≥6.2 mmoles/liter6 (13)77 (13)0.930.38–2.270.88
Triglycerides ≥5.6 mmoles/liter1 (2)4 (1)3.060.34–27.90.32
Serum creatinine ≥1.5 mg/dl8 (17)58 (10)1.790.80–4.010.16
Persistent nephrosis3 (6)25 (4)1.470.43–5.060.54
Any aPL22 (46)218 (38)1.350.75–2.440.32
Medication use, ever     
 High-dose prednisolone5 (10)80 (14)0.720.28–1.880.51
 Azathioprine19 (40)215 (37)1.100.60–2.020.75
 Cyclophosphamide10 (21)132 (23)0.890.43–1.830.75
 Hydroxychloroquine32 (67)386 (67)0.990.53–1.850.97
 HRT/OCs17/41 (41)208/514 (40)0.970.53–1.800.93

Subanalyses of risk factors were performed separately for cardiovascular events (acute MI or angina pectoris) and cerebrovascular events (stroke or TIA) (data not shown). Older age (hazards ratio [HR] per year 1.08 [95% CI 1.03–1.13]; P = 0.001), hypertension (HR 4.19 [95% CI 1.04–16.9]; P = 0.04), serum HDL cholesterol ≤1.0 mmoles/liter (HR 4.23 [95% CI 1.03–17.3]; P = 0.04), and serum creatinine ≥1.5 mg/dl (HR 4.39 [95% CI 1.07–18.0]; P = 0.04) were significant predictors of cardiovascular events. However, no particular risk factors were significantly more common in patients with cerebrovascular events. Patients who were older at the time of diagnosis of SLE had a higher risk of stroke or TIA, but the difference was not significant (HR per year 1.02 [95% CI 1.00–1.05]; P = 0.07).

Table 3 shows a comparison of the risk factors in patients with and without venous events. Significant risk factors for venous thromboembolism were male sex, hypertension, BMI ≥27 kg/m2, serum HDL cholesterol ≤1.0 mmoles/liter, persistent nephrotic syndrome for >6 months, and the presence of aPL. Chinese patients had a significantly lower incidence of venous thrombosis compared with non-Chinese patients, while Caucasian patients had more frequent venous thrombosis compared with non-Caucasian patients.

Table 3. Occurrence of venous thromboembolism as a function of the risk factors studied*
 Venous eventsOdds ratio95% CIP
Yes (n = 40)No (n = 585)
  • *

    Values are the number (%) or the mean ± SD. 95% CI = 95% confidence interval; BMI = body mass index; LDL = low-density lipoprotein; HDL = high-density lipoprotein; aPL = antiphospholipid antibodies; HRT = hormone replacement therapy; OCs = oral contraceptives.

  • Dosage ≥0.8 mg/kg/day for ≥8 weeks.

Age, years35.3 ± 15.335.8 ± 13.41.000.97–1.020.82
Followup, years5.52 ± 2.154.91 ± 2.351.011.00–1.020.11
Women31 (78)524 (90)0.400.18–0.880.02
Race     
 Chinese8 (20)250 (43)0.340.15–0.740.007
 African American11 (28)129 (22)1.340.65–2.760.43
 Caucasian21 (53)206 (35)2.031.07–3.870.03
Smoking, ever10 (25)154 (26)0.930.44–1.940.84
Hypertension19 (48)186 (32)1.941.02–3.700.04
Diabetes mellitus2 (5)34 (6)0.850.20–3.690.83
BMI ≥27 kg/m226 (65)223 (38)3.021.54–5.900.001
LDL cholesterol ≥4.1 mmoles/liter5 (13)58 (10)1.260.48–3.350.64
HDL cholesterol ≤1.0 mmoles/liter10 (25)57 (10)3.021.41–6.510.005
Total cholesterol ≥6.2 mmoles/liter9 (23)74 (13)1.970.90–4.290.09
Triglycerides ≥5.6 mmoles/liter1 (3)4 (1)3.650.40–33.50.25
Serum creatinine ≥1.5 mg/dl7 (18)59 (10)1.890.80–4.460.15
Persistent nephrosis6 (15)22 (4)4.521.72–11.90.002
Any aPL27 (68)213 (36)3.521.78–6.97<0.001
Medication use, ever     
 High-dose prednisolone7 (18)78 (13)1.380.59–3.230.46
 Azathioprine16 (40)218 (37)1.120.58–2.160.73
 Cyclophosphamide14 (35)128 (22)1.920.98–3.790.06
 Hydroxychloroquine24 (60)394 (67)0.730.38–1.400.34
 HRT/OCs16/31 (52)209/524 (40)1.200.62–2.310.59

Ethnic differences in risk factors.

Table 4 shows a comparison of the risk factors for arterial and venous thromboembolism by ethnic group. Smoking, obesity, the presence of aPL, and the use of HCQ and exogenous estrogens were less frequent in Chinese patients. African Americans were more likely to be obese and have renal insufficiency and hypertension. The Caucasians were older at the time of SLE diagnosis and more likely to be smokers and have aPL. However, the use of AZA or CYC was less frequent in the Caucasian patients.

Table 4. Comparison of risk factors among the different ethnic groups*
 Chinese (n = 258)African Americans (n = 140)Caucasians (n = 227)P
  • *

    Values are the number (%) or the mean ± SD. BMI = body mass index; LDL = low-density lipoprotein; HDL = high-density lipoprotein; aPL = antiphospholipid antibodies; HRT = hormone replacement therapy; OCs = oral contraceptives.

  • Dosage ≥0.8 mg/kg/day for ≥8 weeks.

Age, years34.6 ± 1334.7 ± 1237.7 ± 140.03
Women230 (89)128 (91)197 (87)0.38
Smoking, ever32 (12)45 (32)87 (38)<0.001
Hypertension62 (24)64 (46)79 (35)<0.001
Diabetes mellitus10 (4)8 (6)18 (8)0.16
BMI ≥27 kg/m235 (14)90 (64)124 (55)<0.001
LDL cholesterol ≥4.1 mmoles/liter34 (13)14 (10)15 (7)0.06
HDL cholesterol ≤1.0 mmoles/liter23 (9)19 (14)25 (11)0.35
Total cholesterol ≥6.2 mmoles/liter38 (15)15 (11)30 (13)0.53
Triglycerides ≥5.6 mmoles/liter3 (1.2)1 (0.7)1 (0.4)0.67
Serum creatinine ≥1.5 mg/dl22 (9)26 (19)18 (8)0.002
Persistent nephrosis10 (4)9 (6)9 (4)0.45
Any aPL76 (29)59 (42)105 (46)<0.001
Medication use, ever    
 High-dose prednisolone31 (12)25 (18)29 (13)0.24
 Azathioprine157 (61)35 (25)42 (19)<0.001
 Cyclophosphamide75 (29)39 (28)28 (12)<0.001
 Hydroxychloroquine138 (53)99 (71)181 (80)<0.001
 HRT/OCs17/230 (7)82/128 (64)126/197 (64)<0.001

Multivariate analyses of risk factors for thromboembolism.

Cox proportional hazards models were constructed separately for arterial and venous thromboembolism and included age, sex, ethnicity, all the risk factors studied, and the cumulative features of SLE as covariates. Table 5 shows predictors of arterial and venous thromboembolism that were statistically significant. Age at SLE diagnosis, Chinese ethnicity (compared with non-Chinese), serum HDL cholesterol level ≤1.0 mmoles/liter, oral ulcers, serositis, and the absence of leukopenia were significant predictors of arterial thromboembolism.

Table 5. Significant predictors of arterial and venous thromboembolism in the Cox regression models*
 Hazards ratio (95% CI)P
  • *

    95% CI = 95% confidence interval; HDL = high-density lipoprotein; BMI = body mass index; aPL = antiphospholipid antibodies.

Arterial thromboembolism  
 Chinese versus non-Chinese4.06 (1.56–10.6)0.004
 HDL cholesterol ≤1.0 mmoles/liter2.96 (1.44–6.12)0.003
 Oral ulcers3.02 (1.49–6.12)0.002
 Leukopenia0.47 (0.23–0.97)0.04
 Serositis2.37 (1.20–4.68)0.01
Venous thromboembolism  
 Women0.32 (0.12–0.85)0.02
 Chinese versus non-Chinese0.26 (0.08–0.87)0.03
 Diabetes mellitus0.19 (0.04–0.96)0.05
 BMI ≥27 kg/m22.29 (1.01–5.19)0.05
 HDL cholesterol ≤1.0 mmoles/liter3.39 (1.43–8.01)0.005
 Any aPL2.66 (1.29–5.51)0.008
 Hemolytic anemia3.87 (1.62–9.24)0.002
 Leukopenia0.45 (0.20–0.98)0.04
 Renal disease2.99 (1.26–7.09)0.01

Independent risk factors for venous thromboembolism were male sex, non-Chinese ethnicity, BMI ≥27 kg/m2, aPL, serum HDL cholesterol level ≤1.0 mmoles/liter, hemolytic anemia, renal disease, and the absence of diabetes mellitus or leukopenia.

Incidence of thromboembolism by aPL positivity.

Table 6 shows the cumulative risks of arterial and venous thromboembolism in the 3 different ethnic groups with regard to the presence and absence of aPL. The difference in the incidence of arterial thrombosis between Chinese and Caucasian patients became greater when aPL was negative, but the difference still failed to reach statistical significance. For venous thromboembolic risk, significant interethnic differences could not be demonstrated in either aPL-positive or aPL-negative patients, but the trend toward a lower risk in the Chinese patients persisted.

Table 6. Cumulative incidence of thromboembolism at 3 years and 5 years after the diagnosis of SLE, by aPL status and ethnic group*
aPL status, ethnic groupArterial thromboembolismVenous thromboembolism
3 years since SLE diagnosis5 years since SLE diagnosis3 years since SLE diagnosis5 years since SLE diagnosis
  • *

    Values are percentages. SLE = systemic lupus erythematosus; aPL = antiphospholipid antibodies.

Positive    
 Chinese8.010.36.76.7
 African Americans9.29.212.612.6
 Caucasians5.77.313.615.6
Negative    
 Chinese5.77.90.62.6
 African Americans4.07.42.62.6
 Caucasians3.33.31.75.8

DISCUSSION

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

This was a prospective study on the incidence and risk factors of arterial and venous thromboembolism in a large number of patients with new-onset SLE. Toloza et al (14) reported that 7.1% of Caucasian SLE patients experienced arterial vascular events over a mean observation period of 73.8 months, whereas in the study by Manzi et al (10), 6.6% of Caucasian SLE patients developed MI or angina pectoris over a mean followup of ∼7 years. Esdaile et al (9) studied 263 Caucasian patients with SLE who were followed up for a mean of 8.6 years. The incidence of MI or death related to MI was 13%, and the incidence of stroke or death related to stroke was 6%. A study of 397 SLE patients by Rahman et al (29) showed that 9% of patients had coronary heart disease. The incidence of arterial events in the present study (cumulative risk at 5 years 5.1% in Caucasians) is similar to that in these previous studies.

However, few studies have reported the incidence of arterial thrombosis in African American patients with SLE. In the LUMINA (LUpus in MInorities, NAture versus nurture) study by Toloza et al (14), 7.5% of African American SLE patients developed arterial vascular events. This incidence is very similar to the 8.1% found in our study. In fact, in the previous study of the Hopkins Lupus Cohort (the participating institute of the current study) that was reported more than 10 years ago (15), the incidence of MI, angina pectoris, and cardiac death was 8.3%. At that time, data from Caucasian and African American patients were not separated, and the analyses did not include other arterial events, such as stroke and peripheral vascular disease. This precludes direct comparison with the current results.

The risk of thromboembolism in Chinese SLE patients is virtually unknown. In a recent survival analysis of Chinese SLE patients followed up at Tuen Mun Hospital, a participating institute of the present study, cardiovascular and cerebrovascular events accounted for 31% of the deaths in a 10-year period (30). It is intriguing to note that Chinese patients had a higher incidence of arterial vascular events and that Chinese race (as compared with non-Chinese) emerged as an independent predictor of these events in the multivariate model. From the comparative analysis of risk factors, it appeared that the only unfavorable factor for arterial events in the Chinese patients was the lower frequency of the use of HCQ, which might have a beneficial effect on the lipid profile (25, 31). However, HCQ was not a significant protective factor in the current study. The greater interethnic variation in the incidence of arterial events in aPL-negative versus aPL-positive patients suggests that risk factors other than aPL might make a greater contribution to the observed difference. While differences in the health care delivery systems of the two centers, as well as the health-seeking and compliance behaviors of the patients, might contribute to the observed interethnic difference in the risk of arterial thrombosis in the current study, factors that were not addressed, such as serum homocysteine levels, genetic factors, including mannose-binding lectin polymorphisms (23), and overall SLE activity throughout the entire observation period, warrant further evaluation.

A number of factors have been reported to increase the risk of clinical and subclinical atherosclerotic vascular disease in patients with SLE. These include increasing age (7, 8, 10, 14, 15, 32–35), male sex (7), smoking (14, 16), longer disease duration (8, 10, 14, 15), dyslipidemia (8, 10, 15, 16, 18, 32–34, 36), hyperhomocysteinemia (18, 19, 36), presence of aPL (14, 18, 20), higher damage scores (8), duration and cumulative doses of corticosteroids (10, 15, 16, 18, 33), systolic hypertension (15, 17, 32–34), and obesity (15, 36).

In the current study, traditional risk factors, such as age, smoking, diabetes mellitus, and hypertension, were not found to be significant predictors of arterial thrombosis. Two possible explanations can be offered. First, atherosclerosis is a slow and longstanding process. The relatively short exposure of our patients to these risk factors (<5 years) may not be adequate for the development of arterial events, especially when risk factors that were present at any time since the diagnosis of SLE were used in the analyses. Second, the sample size of the cohorts may not be large enough to allow a demonstration of the significant contribution of these risk factors to arterial thrombosis.

Dyslipidemia, as manifested by high levels of serum total and LDL cholesterol, high levels of triglycerides, and low levels of HDL cholesterol, is a well-recognized conventional risk factor for arterial thrombosis in SLE patients (10, 15, 16, 18, 36). Although hypertriglyceridemia and LDL hypercholesterolemia were not found to predict arterial events in the present study, a low serum level of HDL cholesterol was found to be a significant risk factor in both univariate and multivariate analyses. This is consistent with the findings of other studies (16, 18). Intriguingly, a low serum level of HDL cholesterol was also found to be an independent risk factor for venous thrombosis. The importance of a low HDL cholesterol level in vascular thrombosis is illustrated by a recent study that showed a high prevalence of antibodies to apolipoprotein A-I in SLE patients, which was associated with the presence of aPL (37). Sera with antibodies to apolipoprotein A-I showed a high affinity for mature HDL cholesterol molecules, which might account for the low serum HDL levels observed in SLE patients at risk of vascular thrombosis.

Most studies of venous thromboembolism in SLE have been cross-sectional or retrospective. The incidence of venous thromboembolism in SLE patients with disease duration of >10 years ranges from 3% to 16% (38–40). The cumulative risk of venous thromboembolism in Caucasian SLE patients in the current study (10.3% at 5 years of diagnosis) fell into the expected range. However, Chinese patients had a significantly lower risk of venous thrombosis as compared with Caucasian patients, even after adjustment for other factors in the full multivariate model. This is consistent with the clinical observation that Chinese and Asian patients generally have a lower incidence of deep vein thrombosis and pulmonary embolism compared with Caucasian patients (41, 42). While the lower prevalence of aPL and the less frequent use of exogenous hormones might contribute to the lower incidence of venous thrombosis observed in our Chinese patients, other factors not addressed in this study may be responsible for this ethnic difference. These include differences in the prevalence of acquired activated protein C resistance, factor V Leiden mutation, β2-glycoprotein I genetic mutation, and antiprothrombin antibodies, which have been reported in association with venous thromboembolism in SLE patients (43–47).

The procoagulant activity of aPL is related to their effects on coagulation factors, platelets, and endothelial cells. These antibodies are associated with both arterial and venous thrombosis in patients with SLE. The demonstration of aPL as one of the independent predictors of venous thrombosis in the current study is consistent with this.

In summary, interethnic differences in the incidence of arterial and venous thromboembolism in patients with SLE are important and may have therapeutic implications. These differences cannot be fully explained by differences in the prevalence of the risk factors we studied. Further study of other genetic and immunologic factors will be necessary to account for these racial differences.

REFERENCES

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES