Systemic Lupus Erythematosus
An Independent Risk Factor for Endothelial Dysfunction in Women
Background— Systemic lupus erythematosus (SLE) patients have a significantly increased risk of coronary heart disease (CHD) that is not fully explained by classic risk factors. Endothelial dysfunction is an early stage in the process of atherogenesis. Our aim was to determine whether endothelial dysfunction occurs in SLE and whether it is associated with the occurrence of classic Framingham risk factors.
Methods and Results— We studied 62 women with SLE (1997 revised criteria) and 38 healthy women. Demographic and risk factor data were collected. In patients, disease activity and treatment-related parameters were also assessed. Endothelial function was assessed by flow-mediated dilation (FMD) in the brachial artery in response to reactive hyperemia. Carotid intima-media thickness (IMT) and the presence of carotid plaques were also assessed in SLE patients. FMD was impaired in SLE patients (median, 3.6%; range, −6.3% to 13.7%; versus median, 6.9%; range, −6.6% to 17.8%, P<0.01). Using multiple regression analysis that included all subjects in which we retained all the classic CHD risk factors, we found that systolic blood pressure (P=0.019) and SLE (P=0.017) were significantly associated with impaired FMD. Within SLE patients, IMT showed a negative correlation with percent FMD (r=−0.37, P<0.01). In stepwise multiple regression of SLE patients only that also included SLE factors and IMT, IMT alone was independently associated with FMD (P=0.037).
Conclusions— Patients with SLE have endothelial dysfunction that remained significant even after adjustment for other classic CHD risk factors. Within SLE patients, endothelial dysfunction correlates negatively with IMT, another marker of early atherosclerosis. Understanding the mechanism(s) of endothelial dysfunction in SLE may suggest novel strategies for CHD prevention in this context.
Received February 9, 2004; de novo received April 13, 2004; accepted May 15, 2004.
Systemic lupus erythematosus (SLE) is associated with an increased risk of coronary heart disease (CHD). The rate ratio for myocardial infarction in women with SLE 35 to 44 years of age was found to be 52 times that of a comparative population.1 In several large clinical series, the cumulative prevalence of CHD-related events was 6% to 10%, and the annual incidence rate of CHD in SLE populations was ≈1.3% to 1.5%.2 In addition, several studies of subclinical atherosclerosis have identified that 30% to 40% of women with SLE have carotid plaques or myocardial perfusion abnormalities.3,4 Framingham risk factors do contribute to CHD risk, but even after adjustment for these factors, the risk remains increased by 8- to 10-fold.5
Endothelial dysfunction represents a widespread phenomenon, and in the context of CHD, abnormalities of the brachial and femoral arteries have been noted. Endothelial dysfunction is also present in patients with CHD risk factors such as smoking, diabetes mellitus, and hypercholesterolemia.6
SLE is a complex disease, and in addition to atherosclerosis, it is characterized by several vascular processes, namely inflammation, Raynaud’s phenomenon, and a propensity to vascular thrombosis associated with anti-phospholipid antibodies. Lima et al7 have found that patients with SLE have evidence of endothelial dysfunction, but patients with significant CHD risk factors were excluded from this study and the presence of atheroma was not assessed. Recently, endothelial dysfunction has also been found in patients with systemic vasculitis and has been reversed by administration of immunosuppressive therapy.8 Because endothelial dysfunction may represent an early stage in atherogenesis, it is important to understand the mechanisms of its development in a condition such as SLE. It is also important to determine whether it is associated with other CHD risk factors or early atheroma. Therefore, our aims were to determine whether endothelial dysfunction occurred in women with SLE, whether it was explained by the presence of classic CHD risk factors, and whether any of the vascular processes associated with SLE influence endothelial dysfunction.
Patients were recruited from the Lupus Clinic at the University of Manchester Rheumatism Research Centre. Patients fulfilled at least 4 classification criteria (1997 revised criteria9) for SLE or had 3 criteria with no other alternative diagnosis. Healthy control subjects were recruited from the secretarial and support staff at the University of Manchester Rheumatism Research Centre and School of Medicine, as well as from friends of patients. Subjects were excluded if they had had infection within the past 4 weeks or had been pregnant or lactating in the previous 6 months. All subjects gave written informed consent to take part in this study, which was approved by the Central Manchester Local Research Ethics Committee.
Clinical and Laboratory Assessment
All subjects had a complete history and physical examination. Classic CHD risk factors were recorded. In patients, disease activity and cumulative damage were assessed by the SLE Disease Activity Index (SLEDAI)10 and the Systemic Lupus International Collaborating Clinics/American College of Rheumatology Damage Index (SDI),11 respectively. In addition, any history of Raynaud’s phenomenon or previous thrombosis was sought and recorded. Blood was drawn for measurement of plasma glucose and lipid profile. Antibodies to ds-DNA and cardiolipin were measured with commercially available ELISA kits. The lupus anticoagulant was measured with the dilute Russell Viper Venom Test; C3 and C4 complement levels were also measured.
Measurement of Lipids and Lipoproteins
An ultracentrifugation method was used to remove VLDL cholesterol from the plasma.12 HDL cholesterol was determined after precipitation of LDL from the resulting supernatant by heparin/Mn2+ sulfate.12 Total serum cholesterol, HDL cholesterol, and infranatant cholesterol were determined by the CHOD-PAP method (ABX Diagnostics). LDL cholesterol was calculated as the difference between infranatant cholesterol and HDL cholesterol. Serum triglycerides were determined by the GPO-PAP method (ABX Diagnostics).
Assessment of Endothelial Function
Endothelial function was assessed with high-resolution B-mode Doppler (ATL HDI 5000 with a 7.4-MHz linear-array transducer) examination of the brachial artery using the protocol described by Celermajer et al.13 We measured flow-mediated dilation (FMD) in response to reactive hyperemia (endothelium dependent), and endothelium-independent dilation was measured in response to glyceryl trinitrate (GTN). All subjects were studied between 8 and 11 am after a12-hour overnight fast. They were asked not to smoke on the morning of study and to avoid alcohol for 48 hours. Antihypertensive medications were also omitted for 24 hours before the study. The brachial artery was scanned longitudinally 5 to 15 cm above the antecubital fossa. The scans were recorded on a super VHS videotape for later measurement of resting diameter and blood velocity. A blood pressure cuff was then inflated around the forearm to 300 mm Hg for 4.5 minutes. A further scan recorded from 30 seconds before to 1 minute after cuff release. Measurement of the maximal diameter was taken 45 to 60 seconds after cuff release. After 15 minutes, further measurement was taken at rest and at 3 minutes after sublingual spray of 400 μg GTN. All measurements were taken at the end of diastole coincident with the R-wave on an ECG monitor. Distance measured was from anterior to posterior M lines (media-adventitia interface), and every measurement was taken as an average of 5 consecutive cardiac cycles. The vascular technologist performing the scans was given only the subject’s name; no diagnostic information was available. The scans were read in batches every 3 to 4 weeks, and again the technologist was unaware of the diagnostic classification of each subject at the time that the scans were read.
FMD is calculated as follows: 100%×[(postdeflation diameter−resting diameter)/resting diameter]. To assess reproducibility of our technique, we looked at the reliability of reading scans on 2 separate occasions by a single blinded observer. For this, 15 scans from patients or control subjects were chosen at random. The intraclass correlation coefficients for resting diameter and FMD were 0.93 (95% CI, 0.56 to 0.95) and 0.82 (95% CI, 0.30 to 0.97), respectively.
Carotid Artery Intima-Media Thickness and Plaque
Patients but not control subjects also had the intima-media thickness (IMT) of their carotid artery measured. The common carotid artery was scanned longitudinally, and the IMT measurement was taken in the proximal part of the common carotid artery, 1 cm proximal to the carotid bulb as the maximum distance between the intima-lumen and adventitia-media interfaces in areas without carotid plaque.14 IMT was determined as the average of 6 measurements, 3 each from the left and right common carotid arteries. We also noted the presence or absence of carotid plaques, with plaque being defined using the criteria described by Li et al.15 The intraclass correlation coefficient for IMT measurements, assessed in 15 subjects on 2 separate occasions 2 weeks apart, was 0.92 (95% CI, 0.84 to 1.00).
We used version 10.1 of the SPSS statistical package. Values are quoted as median (range). Differences between numeric variables were tested with the Mann-Whitney U test. Correlation was tested with Spearman’s rank-order or Pearson’s correlation coefficient. A significance level was set at P<0.05. For comparison of categorical variables or percentages, we used Fisher’s exact and χ2 tests when appropriate. Multiple linear regression analysis was used to test for independent associations between FMD and various factors.
Endothelial Function in SLE and Healthy Control Subjects
We studied 62 SLE patients and 38 healthy control subjects. In the patient group, age and disease duration were 48 years (range, 21 to 73 years) and 11 years (range, 1 to 23 years), respectively. Fifty-five patients (88%) were white. The SLEDAI and SDI scores were 2 (range, 0 to 12) and 0 (range, 0 to 4), respectively. Nine patients (14.5%) had an SLEDAI ≥6 at the time of study. Twenty-six patients (42%) had ≥1 item of damage on the SDI. With regard to clinical features, 10 patients (16%) had a history of renal involvement, and 34 (56%) had a history of Raynaud’s phenomenon. On autoantibody profiles, 29 patients (46.8%) currently had antibodies to native ds-DNA, and 23 (37.7%) had anti-cardiolipin antibodies (IgG or IgM anticardiolipin and/or lupus anticoagulant) on the day of study. Eight patients (13%) had a previous stroke or transient ischemic attack, and 2 patients (3%) had a previous myocardial infarction. Current therapy included steroids in 31 (50%), antimalarial drugs in 31 (50%), and immunosuppressants in 17 (27%), of whom 13 (21%) were on azathioprine and 4 (6%) were taking methotrexate. In addition, 11 patients were on aspirin, 9 were on nonsteroidal anti-inflammatory drugs (8 on conventional, 1 on COX-2–specific inhibitor), and 3 were on both aspirin and a nonsteroidal anti-inflammatory drug. One patient was on a statin, 10 patients were on hormone replacement therapy, and 3 were on contraceptive pills at the time of study.
As can be seen in Table 1, the SLE group tended to be older and had a higher proportion of postmenopausal women (47% versus 29%; P=0.09). Patients were more likely to be hypertensive (P<0.01); they also had lower HDL cholesterol and higher triglycerides than control subjects. The 10-year risk of CHD was also higher in patients (P<0.01). There was no difference in resting diameter between the groups (Table 2). Patients had significantly reduced FMD compared with control subjects (3.6%; range, −6.2% to 13.7%; versus 6.9%; range, −6.6% to 17.8%; P<0.01; Figure 1). A higher proportion of patients had an FMD ≤4.5% (54.8% versus 26.3%; P=0.04). GTN responses did not differ between groups. A subanalysis of white premenopausal patients (n=20) and age-matched controls (n=25) found similar differences in FMD (4.1%; range, 0% to 13.7%; versus 9.1; range, −3.1% to 17.8%; P<0.01).
Preliminary analysis showed that the strength of association between individual demographic and classic risk factors with FMD was similar in patients and control subjects; therefore, we included all 62 patients and 38 control subjects in a single group for linear regression analysis. In univariate analysis, factors associated with reduced FMD were resting diameter [unstandardized regression coefficient (B)=−0.083, P=0.002], systolic blood pressure (SBP; B=−22.9; P=0.027), and SLE (B=−3.35; P=0.001). In a multivariate model in which we retained classic risk factors, menopause, ethnicity, resting diameter, and SLE, independent factors associated with FMD were SBP (B=−0.108; P=0.019) and SLE (B=−2.75; P=0.017) (R2=0.30, P=0.01) (Table 3). After adjustment for other factors, SLE was associated with a 2.7% lower FMD compared with control subjects, and each increase of 10 mm Hg in SBP was associated with a 1.0% reduction in FMD. A stepwise multiple regression model confirmed these associations.
Endothelial Function in SLE Patients
For regression analysis of only SLE patients, we included demographic details and classic CHD risk factors. We also looked at specific factors associated with the different vascular manifestations of SLE. These included Raynaud’s phenomenon, anti-phospholipid antibodies, disease activity (SLEDAI), and early atherosclerosis (carotid IMT). We also included corticosteroid treatment and serum creatinine as additional variables. Neither Raynaud’s or anticardiolipin/lupus anticoagulant status was associated with endothelial dysfunction (Figures 2 and 3⇓). Similarly, there was no difference in FMD or percent GTN dilation between patients taking and not taking steroid therapy. The SLEDAI correlated positively with FMD (r=0.37, P<0.01) and negatively with resting diameter (r=−0.33, P=0.01). FMD also correlated negatively with IMT (r=−0.37, P<0.01) (Figure 4). In stepwise multiple regression analysis, we examined those variables that showed a degree of association in the univariate analysis (P<0.2). IMT only was independently associated with FMD (P=0.026) (Table 4). The association of SLEDAI with FMD was no longer significant in this model. After adjustment for other factors, each 0.01-cm increase in IMT is associated with a 0.92% decrease in FMD. Twelve SLE patients (19%) also had carotid plaque. These patients had higher resting diameter (0.36 cm; range, 0.28 to 0.43 cm; versus 0.31 cm; range, 0.24 to 0.39 cm; P=0.014) and a trend toward lower FMD (2.7%; range, −6.3% to 7.1%; versus 4.7%; range, −3.3% to 13.7%; P=0.1).
Endothelial dysfunction is believed to represent a widespread phenomenon that occurs at an early stage in the atherogenic process. There is a correlation between endothelial function measured in the brachial and coronary circulations.16 We have found evidence of endothelial dysfunction in SLE patients that is not fully explained by classic CHD risk factors. Also, within SLE patients, endothelial dysfunction is associated with carotid IMT, an early marker of atherosclerosis. Two other studies have reported similar findings. In a study from Sao Paulo, Lima et al7 noted that the mean±SD FMD in SLE was 5.0±5.0% compared with 12.0±6.0% in healthy control subjects. In this study, postmenopausal women and subjects with known CHD risk factors were excluded. Piper et al17 found in a UK cohort that SLE women had a median FMD of 5.6 (interquartile range [IQR], 3.1% to 7.2%) compared with 8.0% (IQR, 6.3% to 9.3%) in control subjects. The differences in subject selection and interlaboratory variations in technique will clearly influence the absolute values between studies. The median percent FMD in our control subjects appears low. Potential reasons for this include the wide age range of control subjects studied and the lack of exclusions based on prevalent CHD risk factors. In a subgroup of premenopausal control subjects, the median percent FMD was 9.1% (IQR, −3.1% to 17.8%), which accords with other studies.7,17 Nevertheless, endothelial function is significantly impaired in SLE. The role of classic CHD factors in this process is less clear from these studies. Piper et al17 found a significant negative correlation between total cholesterol and FMD (r=−0.442). In contrast, Lima et al7 found no association with cholesterol or blood pressure, although it should be noted that this study excluded patients with established hypertension or hyperlipidemia. We chose not to exclude patients with known risk factors to examine the full range of association with FMD in SLE. A recent large cohort control study found that SLE patients have differences in several key cardiovascular risk factors. Specifically, SLE patients are more likely to have hypertension and to be postmenopausal at a particular age; they also have more risk factors per patient.18 In our population, patients also had lower HDL cholesterol. To adjust for these differences, we performed a multivariate analysis on all patients and control subjects. Of the classic risk factors, SBP was most strongly associated with reduced FMD. Each 10–mm Hg increase in SBP was associated with a 0.89% reduction in FMD. Importantly, however, even after adjustment for classic risk factors, SLE remains independently associated with endothelial dysfunction. This finding is important because SLE is associated with an increased risk of CHD compared with the general population. It also accords with the clinical observation by Esdaile et al5 that even after adjustment for Framingham risk factors, SLE patients still have increased CHD risk. It may also suggest that interventions that improve endothelial function in the context of SLE will have the potential to improve long-term CHD outcomes.
Multivariate analysis of SLE patients showed that FMD was associated with carotid IMT. Twelve patients had carotid plaque, and in this group, there was also a trend toward lower FMD. We did not measure IMT in our control subjects, but this finding in SLE confirms that, as in the general population, endothelial dysfunction in SLE correlates with other markers of atherosclerosis development. Piper et al17 found an association of FMD with total cholesterol. We could not confirm this finding, but IMT as an early marker of atherosclerosis will represent the final common pathway of several other risk factors and will make it difficult for any single risk factor to remain in a multivariate analysis. Endothelial function may act in a similar fashion as an integrated index of all atherogenic and atheroprotective factors present in an individual.19 Interventions that modify atherosclerosis risk factors may therefore be of benefit in SLE.
In univariate analysis, we noted a positive association between disease activity and FMD. It should be noted that SLEDAI was associated with a reduced resting diameter, which itself predicts increased FMD; after adjustment for resting diameter, the effect of SLEDAI was lost. Inflammation itself may be associated with endothelial dysfunction. This has been suggested by studies of endotoxin-induced experimental inflammation and in patients with primary systemic vasculitis.20,21 Therefore, it would be reasonable to expect that inflammatory disease activity in SLE should impair endothelial function, but this was not found in this study or in the study by Lima et al.7 Both of these studies, however, were cross-sectional and included patients mainly with low disease activity. A prospective study of patients in a clinical flare before and after therapy is therefore warranted to study the influence of SLE inflammation on endothelial function. It is still an attractive hypothesis that chronic inflammation is the key factor contributing to atherosclerosis risk in SLE, yet other mechanisms mediating endothelial dysfunction—eg, insulin resistance, hyperhomocysteinemia, or ADMA, a recently described inhibitor of nitric oxide synthase—could be more important in SLE.22 These require further investigation in the context of SLE.
This study has several limitations. First, the cross-sectional nature may fail to estimate the true magnitude of the contribution of variables such as disease activity and therapy; a prospective study might demonstrate a greater effect over time. Second, this study was underpowered to determine the association of endothelial function with clinical outcomes of relevance. Even so, we did find a significant association with carotid IMT and a trend toward lower FMD in patients with carotid plaque. Carotid IMT is itself a valid surrogate for early atherosclerosis and has value in predicting future CHD events in the general population. This association, as well as the trend to lower FMD in patients with plaque, supports the validity of endothelial dysfunction as a marker of interest, and we plan to follow up our cohort to determine whether endothelial dysfunction does indeed predict future CHD events.
In conclusion, we have found impaired endothelial function in SLE that is not fully explained by classic CHD risk factors. Within SLE patients, endothelial dysfunction is associated with carotid IMT and plaque. Interventions to improve endothelial function in SLE may therefore have an impact on the risk of future CHD. A better understanding of the factors underlying endothelial dysfunction that may be peculiar to SLE is also required to develop novel approaches to reducing the CHD burden in these patients.
This study was funded by the arc Epidemiology Unit, University of Manchester, and Research Endowment funds from Central Manchester and Manchester Children University Hospitals NHS Trust, Manchester, UK.
Manzi S, Meilahn EN, Rairie JE, et al. Age-specific incidence rates of myocardial infarction and angina in women with systemic lupus erythematosus: comparison with the Framingham study. Am J Epidemiol. 1997; 145: 408–415.
Bruce IN, Gladman DD, Ibanez D, et al. Single photon emission computed tomography dual isotope myocardial perfusion imaging in women with systemic lupus erythematosus, II: predictive factors for perfusion abnormalities. J Rheumatol. 2002; 30: 288–291.
Lima DS, Sato EI, Lima VC, et al. Brachial endothelial function is impaired in patients with systemic lupus erythematosus. J Rheumatol. 2001; 29: 292–297.
Raza K, Thambyrajah J, Townend JN, et al. Suppression of inflammation in primary systemic vasculitis restores vascular endothelial function: lessons for atherosclerotic disease? Circulation. 2000; 102: 1470–1472.
Mackness MI, Durrington PN. Lipoprotein analysis for clinical studies. In: Converse CA, Skinner ER, eds. Lipoprotein Analysis: A Practical Approach. Oxford, UK: IRL Press; 1992: 1–42.
Sidhu PS, Desai SR. A simple and reproducible method for assessing intimal-medial thickness of the common carotid artery. Br J Radiol. 1997; 70: 85–89.
Li R, Duncan BB, Metcalf PA, et al. B-mode-detected carotid artery plaque in a general population: Atherosclerosis Risk in Communities (ARIC) Study Investigators. Stroke. 1994; 25: 2377–2383.
Piper MJ, Heaton SJ, Gardner-Medwin JM, et al. A study of endothelial function in systemic lupus erythematosus (SLE). Rheumatology. 2001; 40 (suppl 1): 113.
Bonetti PO, Lerman LO, Lerman A. Endothelial dysfunction: a marker of atherosclerotic risk. Arterioscler Thromb Vasc Biol. 2003; 23: 168–175.
Hingorani AD, Cross J, Kharbanda RK, et al. Acute systemic inflammation impairs endothelium-dependent dilatation in humans. Circulation. 2000; 102: 994–999.
Filer AD, Gardner-Medwin JM, Thambyrajah J, et al. Diffuse endothelial dysfunction is common to ANCA associated systemic vasculitis and polyarteritis nodosa. Ann Rheum Dis. 2003; 62: 162–167.
Cooke JP. Does ADMA cause endothelial dysfunction? Arterioscler Thromb Vasc Biol. 2000; 20: 2032–2037.