To determine the rate of atherosclerosis progression as well as the relationship of traditional risk factors, systemic lupus erythematosus (SLE)–related factors, and treatment to atherosis progression in SLE patients.
To determine the rate of atherosclerosis progression as well as the relationship of traditional risk factors, systemic lupus erythematosus (SLE)–related factors, and treatment to atherosis progression in SLE patients.
Outpatients in the Hospital for Special Surgery SLE Registry underwent serial carotid ultrasound and clinical assessment in a longitudinal study.
Among 158 patients, 77 (49%) had persistent absence of atherosclerosis (carotid plaque), 36 (23%) had unchanged atherosclerosis, and 45 (28%) had progressive atherosclerosis, defined as a higher plaque score (new plaque in 25 patients and more extensive plaque in 20 patients) after a mean ± SD interval of 34 ± 9 months. Multivariate determinants of atherosclerosis progression were age at diagnosis (odds ratio [OR] 2.75, 95% confidence interval [95% CI] 1.67–4.54 per 10 years, P < 0.001), duration of SLE (OR 3.16, 95% CI 1.64–6.07 per 10 years, P < 0.001), and baseline homocysteine concentration (OR 1.24, 95% CI 1.06–1.44 per μmoles/liter, P = 0.006). SLE patients with stable plaque and progressive plaque differed only in baseline homocysteine concentration. Atherosclerosis progression was increased across tertiles of homocysteine concentration (16.2%, 36.4%, and 56.1%; P = 0.001), and homocysteine tertile was independently related to progression of atherosclerosis (OR 3.14, 95% CI 1.65–5.95 per tertile, P < 0.001). Less aggressive immunosuppressive therapy and lower average prednisone dose were associated with progression of atherosclerosis in univariate, but not multivariate, analyses. Inflammatory markers and lipids were not related to atherosclerosis progression.
Atherosclerosis develops or progresses in a substantial minority of SLE patients during short-term followup (10% per year on average). Older age at diagnosis, longer duration of SLE, and higher homocysteine concentration are independently related to progression of atherosclerosis. These findings show that aggressive control of SLE and lowering of homocysteine concentrations are potential means to retard the development and progression of atherosclerosis in SLE.
Systemic lupus erythematosus (SLE) is associated with premature myocardial infarction (1–3). Noninvasive imaging studies suggest that myocardial infarction is primarily due to accelerated atherosclerosis rather than in situ thrombosis associated with antiphospholipid syndrome (4–7). Although atherosclerosis in patients with SLE has been attributed to traditional risk factors seen in the general population (4, 8, 9), recent studies have suggested that the presence of the disease itself may be primarily responsible (5, 6, 10). In view of the importance of inflammation in the genesis and progression of atherosclerosis (11, 12), investigators have sought to determine whether specific aspects of SLE, a chronic inflammatory disease, are closely tied to the occurrence of myocardial infarction or the presence of subclinical atherosclerosis. Features of SLE that have been associated with clinical manifestations of coronary artery disease include older age at diagnosis (3, 8), longer duration of SLE (3, 8, 13), higher damage index score (13), longer duration of steroid therapy (1, 3, 8, 14), and higher concentrations of oxidized low-density lipoprotein (LDL) cholesterol and homocysteine (14). Features of SLE associated with the presence of subclinical atherosclerosis include older age at diagnosis (5), longer duration of SLE (5), longer duration of steroid therapy (4), less aggressive immunosuppressive therapy (5), and a higher damage index score (5).
The rate at which atherosclerosis progresses in patients with SLE is unknown. Likewise, it is uncertain whether, once atherosclerosis is established, specific aspects of the disease or its treatment foster more rapid progression of atherosclerosis. While myocardial infarction occurs prematurely in SLE, it is an uncommon event and is not suitable as an end point to measure the progression of atherosclerosis. Although serial coronary angiography can be used for documenting disease progression (ideally if coupled with intravascular ultrasound quantification of plaque volume), its use cannot be justified in large numbers of SLE patients who do not have indications for invasive procedures. Noninvasive carotid ultrasonography is risk-free, well tolerated, and well suited for documenting the presence, extent, and severity of atherosclerosis, particularly in large populations (15). Carotid atherosclerosis is a manifestation of systemic atherosclerosis (16) and is an independent predictor of cardiovascular events, primarily myocardial infarction (17). We used carotid ultrasonography in a longitudinal study of the prevalence and determinants of atherosclerosis in SLE to quantify the rate of progression of carotid atherosclerosis and to examine the relationship of traditional risk factors, SLE-related factors, and treatment to atherosclerosis progression.
The study population consisted of patients participating in a longitudinal study of cardiovascular disease in SLE (5). All patients met American College of Rheumatology (ACR) diagnostic criteria for SLE (18) and were enrolled in the Autoimmune Disease Registry at the Hospital for Special Surgery. Patients were consecutively recruited at the time of outpatient rheumatology visits. Patients were excluded from study participation if they were younger than 18 years of age, pregnant, or presented with renal failure (serum creatinine level ≥3.0 mg/dl or creatinine clearance ≤30 ml/minute). All patients were interviewed and examined using a standardized data collection instrument. Type and duration of pharmacologic therapy were ascertained by patient interview and chart review; drug usage was categorized as never, former, or current usage. Use of prednisone was further quantified based on average daily dose over the previous 5 years. Disease activity at the time of study visit was assessed using the SLE Disease Activity Index (19), whereas cumulative damage related to the disease was quantified using the Systemic Lupus International Collaborating Clinics/ACR Damage Index (20).
The presence of traditional risk factors for atherosclerosis was ascertained based on hypertension (sustained elevation of ≥140 mm Hg in the systolic pressure or ≥90 mm Hg in the diastolic pressure or the use of antihypertensive medications), presence of diabetes, smoking history, family history of premature myocardial infarction (<55 and <65 years in first-degree male and first-degree female relatives, respectively), and fasting lipid profile. A total of 203 participants underwent the complete study protocol at baseline, and 159 returned for a repeat study. Failure to undergo repeat study was due to death (n = 3), refusal (n = 10), loss of contact (n = 20), or scheduling difficulties (n = 11). The institutional review boards of our organizations approved the study, and all participants provided written informed consent.
All participants underwent carotid ultrasonography and echocardiography on the day of the study visit. Carotid ultrasonography was performed using a standardized protocol. Briefly, the extracranial segments of the left and right carotid arteries were extensively scanned for the presence of discrete atherosclerotic plaque, defined as the presence of focal protrusion (intima-media thickening) with a thickness exceeding that of the surrounding wall by at least 50%. A plaque score was calculated according to the number of left and right segments (common carotid, bulb, internal carotid, external carotid) containing plaque; thus, the plaque score ranged from 0 to 8. Intima-media thickness (IMT) of the far wall of the distal common carotid artery was measured on end-diastolic M-mode images of multiple cycles (15). IMT was never measured at the level of a plaque; IMT is presented as the average of the values of the left and right segments. Echocardiographic evaluation included detection of the presence and severity of pulmonary hypertension, which is defined as an estimated pulmonary artery systolic pressure >35 mm Hg (21). All ultrasound studies were performed by a research sonographer and were interpreted by a single highly experienced cardiologist (MJR) who was blinded to the clinical characteristics of the participants.
Blood specimens were obtained following an overnight fast. Analyses included routine chemistries, as well as determination of concentrations of high-sensitivity C-reactive protein, serum complement (C3 and C4), lipids, homocysteine, folate, and vitamin B12. Hyperhomocysteinemia was defined as a concentration exceeding 12 μmoles/liter (22). Antiphospholipid antibodies were considered positive in the presence of lupus anticoagulant or anticardiolipin (IgG or IgM) at ≥40 units/ml (23).
The study population was divided into 3 groups: patients with persistent absence of atherosclerosis (plaque score of 0 in both studies), patients with stable atherosclerosis (unchanged plaque score), and patients with progressive atherosclerosis, defined as an increase in plaque score due to the development of atherosclerosis (plaque score of 0 at the baseline study and ≥1 at the second study) or worsening of atherosclerosis (plaque score ≥1 at baseline and higher at the second study). Differences in continuous variables between the 3 groups were compared using analysis of variance, with post hoc testing for multiple comparisons; differences in categorical variables were compared using the chi-square test. Analyses were repeated using an adjustment for differences in age. Results are presented as the mean ± SD. Multinomial logistic regression analysis was used to assess the independence of association with progression of atherosclerosis compared with stable plaque or with absence of plaque using variables that were significant (P < 0.10) in univariate analyses. Results are presented as the odds ratio (OR) (95% confidence interval [95% CI]). Statistical analyses were performed using SPSS, version 11.0 (SPSS, Chicago, IL).
The overall study population consisted of 159 SLE patients who underwent the study protocol twice, with a mean ± SD interval of 34 ± 9 months between visits (range 22–58 months). The demographic features, cardiovascular disease risk factors, and SLE disease characteristics of the population are shown in Table 1. At the time of the second study visit, 77 of the 159 patients (48.4%) had persistent absence of carotid atherosclerosis and 36 patients (22.6%) had persistent but stable atherosclerosis. Progressive atherosclerosis was detected in 45 patients (28.3%); this was characterized by the development of plaque in 25 patients and increasing plaque in 20 patients (Figure 1). One patient (0.6%) had a single plaque at the baseline study that was not detected on blinded reading of her followup study, and this patient was excluded from further analyses.
|No. of patients||159|
|Female sex, %||96|
|White race, %||48|
|Diabetes mellitus, %||2|
|Current smoker, %||15|
|Total cholesterol, mean ± SD mg/dl||206 ± 51|
|LDL cholesterol, mean ± SD mg/dl||116 ± 46|
|Homocysteine, mean ± SD μmoles/liter||7.7 ± 3.4|
|CRP, mean ± SD mg/dl||5.1 ± 8.0|
|ESR, mean ± SD mm/hour||28 ± 22|
|Age at diagnosis, mean ± SD years||32 ± 13|
|Disease duration, mean ± SD months||138 ± 107|
|SLEDAI score, mean ± SD||4.2 ± 5.2|
|SDI score, mean ± SD||1.3 ± 1.9|
|Renal disorder, %||32|
|Neurologic disorder, %||13|
|Hematologic disorder, %||54|
|Malar rash, %||43|
|Discoid rash, %||8|
|Oral ulcers, %||23|
|Antiphospholipid antibodies, %||21|
|Anti-DNA antibodies, %||45|
Baseline findings in the 3 groups of SLE patients with persistent absence of atherosclerosis, stable atherosclerosis, and progressive atherosclerosis are compared in Table 2. Other than older age in patients in the 2 groups with plaque, the 3 groups did not differ in traditional risk factors for atherosclerosis. Patients with stable plaque and progressive plaque were older at the time of diagnosis of SLE than patients with no plaque. Compared with patients with no plaque, those with progressive plaque had longer disease duration. Use of prednisone at any time was very common in the entire population; however, the average daily dose of prednisone was significantly lower among those with progressive atherosclerosis as compared with the group with no plaque. Use of cyclophosphamide and azathioprine was lower in the 2 groups with plaque. The prevalence of pulmonary hypertension and carotid IMT was greater in the 2 groups with plaque, as was the prevalence of white patients in the stable plaque group. However, the significance of these findings was eliminated after adjustment for age. The 3 groups did not differ in other SLE disease characteristics listed in Table 1. Lipid-lowering medication had been taken by 11% of patients and was unrelated to the progression of atherosclerosis.
|No plaque (n = 77)||Stable plaque (n = 36)||Progressive plaque (n = 45)||P|
|Age, years||36 ± 9||51 ± 12†||50 ± 14†||<0.001|
|Study interval, months||34 ± 9||32 ± 9||34 ± 9||0.655|
|Female sex, %||97||97||91||0.268|
|White race, %||40||72†||56||0.006‡|
|Diabetes mellitus, %||0||3||5||0.219|
|Current smoker, %||14||14||16||0.962|
|BMI, kg/m2||26.5 ± 6.7||26.9 ± 6.6||25.7 ± 5.6||0.695|
|Systolic BP, mm Hg||107 ± 16||114 ± 21||111 ± 16||0.076‡|
|Diastolic BP, mm Hg||70 ± 9||72 ± 10||71 ± 9||0.502|
|Total cholesterol, mg/dl||201 ± 60||210 ± 45||212 ± 37||0.469|
|LDL cholesterol, mg/dl||112 ± 54||118 ± 41||123 ± 32||0.475|
|HDL cholesterol, mg/dl||58 ± 16||62 ± 17||62 ± 18||0.399|
|Triglycerides, mg/dl||155 ± 70||155 ± 91||141 ± 58||0.552|
|Serum creatinine, mg/dl||0.91 ± 0.57||1.06 ± 0.70||0.96 ± 0.37||0.420|
|Homocysteine, μmoles/liter||7.0 ± 3.02||7.4 ± 2.9||9.1 ± 4.1§||0.005|
|ESR, mm/hour||29 ± 23||26 ± 24||27 ± 19||0.884|
|CRP, mg/dl||4.7 ± 7.2||6.0 ± 11.5||5.2 ± 5.9||0.723|
|Folate, ng/ml||15.8 ± 9.8||18.4 ± 14.5||22.6 ± 36.5||0.259|
|B12, pg/ml||588 ± 303||786 ± 495||707 ± 427||0.419|
|Age at diagnosis, years||27 ± 9||39 ± 14†||36 ± 15†||<0.001|
|Disease duration, years||116 ± 87||146 ± 128||168 ± 114¶||0.027|
|SLEDAI score||5.0 ± 5.7||3.4 ± 4.5||3.4 ± 4.6||0.165|
|SDI score||1.1 ± 1.7||1.8 ± 2.5||1.1 ± 1.6||0.172|
|Pulmonary hypertension, %||8||25¶||27§||0.008‡|
|Antiphospholipid antibodies, % positive||22||25||18||0.721|
|Prednisone use, %#||95||81||87||0.140|
|5-year average daily prednisone dose, mg||12.5 ± 8.9||7.1 ± 8.5||7.2 ± 6.2¶||0.010‡|
|Cyclophosphamide use, %#||29||6¶||13¶||0.005‡|
|Azathioprine use, %#||40||21||21||0.057‡|
|Carotid IMT, mm||0.56 ± 0.09||0.66 ± 0.17†||0.68 ± 0.18†||<0.001‡|
Homocysteine concentration was significantly higher among patients with progressive plaque compared with the group with no plaque, with a trend toward higher concentration than in the stable plaque group (P = 0.079), although vitamin B12 and folate concentrations were similar in the 3 groups. Additionally, homocysteine and folate levels were unrelated (r = −0.078, P = 0.339) in the entire population. Homocysteine concentrations were elevated (≥12 μmoles/liter) in only 17 patients, 9.1% with persistent absence of plaque, 8.3% of those with stable plaque, and 15.5% with progressive plaque. Because of the small number of patients with elevated homocysteine concentrations, the population was divided according to tertile of homocysteine concentration: ≤5.8 μmoles/liter, 5.9–7.8 μmoles/liter, and ≥7.9 μmoles/liter. Progression of atherosclerosis significantly increased across tertiles of homocysteine concentration (Figure 2) despite the comparability of age, disease duration, and concentrations of folate and vitamin B12 in the 3 groups.
In logistic regression analysis including baseline variables that were significant in univariate analyses (Table 3), only age at diagnosis, duration of SLE, and homocysteine concentration (or tertile) were independent predictors of progression of atherosclerosis as compared with persistent absence of plaque, whereas only the homocysteine concentration distinguished the groups with stable plaque and progressive plaque (lower homocysteine concentrations in the group with stable plaque). The groups with no plaque and stable plaque differed in disease duration and age at diagnosis.
|Comparison, variable||OR (95% CI)||P|
|Progressive plaque versus no plaque|
|Age at diagnosis (per 10 years)||2.75 (1.67–4.54)||<0.001|
|SLE duration (per 10 years)||3.16 (1.64–6.07)||<0.001|
|Homocysteine (per μmole/liter)||1.24 (1.06–1.44)||0.006|
|Homocysteine tertile†||3.14 (1.65–5.95)||<0.001|
|Stable plaque versus no plaque|
|Age at diagnosis (per 10 years)||2.99 (1.76–5.09)||<0.001|
|SLE duration (per 10 years)||2.68 (1.33–5.38)||0.006|
|Stable plaque versus progressive plaque|
|Homocysteine (per μmole/liter)||0.80 (0.67–0.97)||0.021|
|Homocysteine tertile†||0.42 (0.21–0.81)||0.010|
Our study documents the progression of atherosclerosis in 28% of SLE patients, or 10% per year over a mean period of 34 months. Other than older age, none of the risk factors traditionally associated with the presence of atherosclerosis was related to the progression of atherosclerosis. Disease-related factors associated with progressive atherosclerosis included duration of SLE, presence of pulmonary hypertension, and trends toward less aggressive use of immunosuppressive agents, suggesting a cumulative impact of duration and severity of inflammation and less aggressive immunosuppressive therapy on atherosclerosis progression. These findings are consistent with our earlier observations regarding correlates of the presence of atherosclerosis (5).
The strong relationship of homocysteine concentration with the progression of atherosclerosis seen in the present study is consistent with observations linking homocysteine concentrations to cardiovascular disease in population-based studies. Large meta-analyses have shown a relationship of homocysteine to ischemic heart disease and stroke, following adjustment for traditional cardiovascular risk factors (24, 25). In the European Concerted Action Project, elevated homocysteine concentration was comparable to smoking and hyperlipidemia in strength of risk for vascular disease (26).
Homocysteine has also been implicated as a risk factor for vascular disease in patients with SLE. Hyperhomocysteinemia (>14.1 μmoles/liter), present in 15% of 337 patients, was an independent risk factor for arterial thromboses in the Hopkins Lupus Cohort Study (27). Similarly, Svenungsson and colleagues (14) reported significantly higher homocysteine concentrations in 26 SLE patients with clinical manifestations of cardiovascular disease compared with 26 SLE patients without these manifestations. Homocysteine concentrations showed a strong tendency to be higher in SLE patients with atherosclerosis in our baseline study of 197 patients (8.6 ± 4.3 versus 7.3 ± 2.6 μmoles/liter in those without atherosclerosis; P = 0.06) (5). In contrast, Afeltra and colleagues (28), in a study of 57 patients, 21 of whom had venous and/or arterial thromboses, noted elevated levels (>15 μmoles/liter) in 14% of them; however, homocysteine concentrations were not significantly higher in patients with thromboses compared with those without them. In the recent study by Von Feldt and colleagues (29), coronary artery calcification, as measured by electron beam computed tomography in patients with SLE, was independently associated with homocysteine concentrations. Our study is the first to associate homocysteine concentrations with progression of atherosclerosis.
Homocysteine concentrations were elevated in 11% of patients in the current study, and mean folate and vitamin B12 levels were within the normal range. Although low folate levels may be associated with cardiovascular disease independently of homocysteine concentrations (30), folate levels were not related to atherosclerosis progression in the present study. Methotrexate therapy elevates homocysteine concentrations by antagonism of folate metabolism; however, only 10% of the study population had received methotrexate therapy, and levels of folate (22.0 ± 23.0 versus 18.1 ± 22.0 ng/ml; P = 0.744) and homocysteine (8.0 ± 3.6 versus 7.7 ± 3.4 μmoles/liter; P = 0.549) were comparable in those who did and did not receive methotrexate therapy, respectively. If the higher threshold (>15 μmoles/liter, applied in the absence of folate fortification) was used to define elevated homocysteine, only 6% of our cohort would have high homocysteine concentrations. This lower prevalence, in comparison with that of other studies, may reflect the exclusion of patients with active renal disease.
Potential mechanisms whereby homocysteine might contribute to atherosclerosis include endothelial injury due to release of reactive oxygen species by oxidized homocysteine, increased oxidation of low-density lipoprotein, and stimulation of smooth muscle cell proliferation (31). Homocysteine enhances endothelial cell apoptosis (32). It is interesting to note that patients with SLE have increased levels of circulating apoptotic endothelial cells and reduced brachial artery flow-mediated dilation (33), and yet, despite accelerated damage, they have decreased numbers of circulating endothelial cell precursors, cells that are essential for vascular repair (34). Elevated homocysteine concentrations may aggravate the SLE-associated imbalance between endothelial damage and regeneration by increasing apoptosis. Our data raise the possibility that a given level of homocysteine confers more damage in SLE patients than in nonautoimmune patients and that traditional risk factors and lupus-related factors may act synergistically.
Older age at diagnosis (3, 5, 8) and longer disease duration (3, 5, 8, 13, 29) have been repeatedly documented to correlate with manifestations of atherosclerosis in SLE. The present study extends the documentation of these characteristics as independent correlates of atherosclerosis progression. Duration of disease, indicative of longer exposure to an inflammatory milieu, may be a surrogate measure of inflammatory burden and thus may define the cumulative impact of the disease better than individual inflammatory markers or metrics determined at specific points to quantify disease activity and damage. The explanation for the repeated association of older age at diagnosis with the presence and progression of atherosclerosis is less apparent. It is possible that older age at diagnosis with longer duration of exposure of the vasculature to traditional risk factors is associated with a greater frequency of preexisting atherosclerosis, accelerated by the presence of SLE as a potent atherogenic stimulus. This speculation is consistent with the primacy of the aging process in the development of both atherosis and sclerosis of the conduit arteries and with the hypothesis that the vasculature in lupus patients has less regenerative capacity and is thus more vulnerable to damage.
Most studies that show serial changes in carotid ultrasound findings relate to the impact of pharmacologic interventions, such as lowering of blood pressure or LDL cholesterol levels. Furthermore, the end point is usually a change in IMT, which is frequently a composite measurement of multiple segments, some containing focal plaque. Likewise, population-based studies of progressive changes have primarily focused on IMT (35, 36). IMT in areas free of discrete plaque does not appear to be affected by the presence of SLE, based on large studies in relatively unselected populations (4, 5).
Using a definition of plaque progression similar to that in the current study, carotid atherosclerosis progressed over 4 years in 18% of women participating in the Aging Vascular Study conducted in France (37). This rate is lower on an annual basis (<5% per year) than that observed in the present study, but it also occurred in an older age group (59–71 years old at baseline). Of the 197 control subjects evaluated in our baseline analysis (5), 98 subjects underwent a second carotid ultrasound study almost 5 years after the initial study (almost twice the interval between studies in the SLE patients), with progression of atherosclerosis detected in 23% per year (<5%). These data suggest that the 10% per year rate of plaque progression in patients with SLE is substantially higher than in the general population.
Potential limitations of the current study include the dichotomous definition of progression of atherosclerosis, i.e., the presence or absence of an increase in plaque score. However, given the importance of studying large numbers of relatively unselected patients to address the goal of the study and hence the need to use a noninvasive imaging modality, the current inability to accurately quantify plaque burden using standard transcutaneous ultrasound, the absence of obstructive atherosclerosis that could be quantified using Doppler techniques, and the lack of utility of the more continuous measure of IMT, we chose change in plaque score as the most robust option. Since patients in the stable plaque group may actually have had an increase in plaque size within previously affected segments as opposed to developing a new plaque in other arterial segments, the finding of higher homocysteine concentrations in the progressive plaque group than in the stable plaque group is even more impressive. In addition, laboratory values, particularly inflammatory markers, were only measured twice, and the full extent of inflammatory burden (“area under the curve”) cannot be accurately calculated. Finally, drug exposure is imprecisely quantified, given the inherent difficulties in determining exact doses and duration of drug use in the setting of episodic and tapering therapy.
In conclusion, the present study documents atherosclerosis progression in a substantial minority of SLE patients over a relatively short period of time that is strongly related to duration of disease. The observed rate of progression appears to be substantially higher (∼10% versus 5% per year) than in individuals who do not have lupus. Quantification of the rate of progression may be useful in assessing efficacy of intervention in future treatment trials. In addition, homocysteine is identified as an important predictor of atherosclerosis progression. Our results suggest that this recognized risk factor has potent damaging effects on the vasculature in patients with SLE. Although secondary prevention trials in non-lupus populations have not demonstrated the efficacy of lowering homocysteine concentrations in preventing recurrent vascular events (38–40), the importance of homocysteine in atherogenesis has not been refuted (41), and primary prevention by reducing homocysteine concentrations in the setting of SLE may be especially efficacious.
Dr. Roman had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study design. Roman, Crow, Lockshin, Devereux, Paget, Sammaritano, Salmon.
Acquisition of data. Roman, Lockshin, Sammaritano, Levine, Davis, Salmon.
Analysis and interpretation of data. Roman, Crow, Lockshin, Devereux, Paget, Sammaritano, Levine, Salmon.
Manuscript preparation. Roman, Crow, Lockshin, Devereux, Paget, Sammaritano, Levine, Salmon.
Statistical analysis. Roman.