Article Text

Concise report
Factors involved in the progress of preclinical atherosclerosis associated with systemic lupus erythematosus: a 2-year longitudinal study
  1. I Rua-Figueroa1,
  2. O Arencibia-Mireles2,
  3. M Elvira3,
  4. C Erausquin1,
  5. S Ojeda1,
  6. F Francisco1,
  7. A Naranjo1,
  8. C Rodríguez-Gallego4,
  9. I Garcia-Laorden4,
  10. J Rodríguez-Perez5,
  11. C Rodríguez-Lozano1
  1. 1Rheumatology Division, Hospital Universitario de Gran Canaria ‘Doctor Negrin’, Las Palmas, Spain
  2. 2Statistics Department, Canarias Government, Canary Islands, Spain
  3. 3Radiology Service, Hospital Universitario de Gran Canaria ‘Doctor Negrin’, Las Palmas, Spain
  4. 4Immunology Service, Hospital Universitario de Gran Canaria ‘Doctor Negrin’, Las Palmas, Spain
  5. 5Nephrology Division, Hospital Universitario de Gran Canaria ‘Doctor Negrin’, Las Palmas, Spain
  1. Correspondence to Dr I Rúa-Figueroa, Rheumatology Division, Hospital Universitario de Gran Canaria ‘Doctor Negrin’, Las Palmas, Spain; iruafer{at}


Objectives To assess the changes in carotid intima-media thickness (IMT) and the associated risks factors in patients with low severity systemic lupus erythematosus (SLE).

Methods Common carotid IMT measurements were obtained by ultrasound from 101 patients with SLE at an interval of 2 years. Cardiovascular risk factors, disease activity, accumulated damage, severity (Katz index) and biochemical parameters (including high sensitivity C-reactive protein, interleukin 6, C3a, C4a, C5a and homocysteine) were also assessed. Multiple linear regression was used to assess the effect of these variables on the end IMT measurement (eIMT) adjusted to the baseline measurement (bIMT).

Results The cohort comprised 94.1% women, with a mean age at entry of 41.5 years and a mean disease duration of 12.1 years. An increase of 0.078 mm in IMT was detected over 2 years, from a mean bIMT of 0.37 mm to a mean eIMT of 0.44 mm (p<0.001). When adjusted for the bIMT, multiple linear regression identified bIMT, age at diagnosis, homocysteine, C3 and C5a as risk factors for IMT progression.

Conclusions IMT significantly increases over 2 years in patients with SLE. Age, baseline IMT, C3, C5a anaphylatoxin and homocysteine are all associated risk factors, supporting a role for complement and homocysteine in the early stages of premature SLE-associated atherosclerosis.

Statistics from

Accelerated atherosclerosis (AT) is recognised as a major cause of morbidity and mortality in patients with systemic lupus erythematosus (SLE), and subclinical AT is evident in up to 40% of patients with SLE.1 As only a few longitudinal studies have been performed, the factors that influence the progress of AT associated with SLE are poorly understood.2 3 In this longitudinal study of patients with low-severity SLE we have therefore investigated the risk factors associated with the changes in intima-media thickness (IMT) measured by carotid ultrasound, a well-accepted and non-invasive method to assess and follow the progression of subclinical AT.4 We paid special attention to complement breakdown products and to sensitive biomarkers of inflammation.



The patient population comprised 106 patients with unselected SLE (1982 ACR criteria). Patients who were pregnant or who had significant renal insufficiency were excluded. During the follow-up period one patient died and four others failed to complete the protocol.

Disease activity was assessed every 6 months using the systemic lupus activity measure (SLAM), accumulated damage was measured with the SLICC/ACR/DI (SDI; Systemic Lupus International Collaborating Clinics/American College of Rheumatology Damage Index) and severity was measured according to the Katz index.5

We considered corticoid use both at baseline (current use, any dose) and at any time after disease diagnosis (any dose), categorising the magnitude of corticoid exposure (table 1).

Table 1

Baseline clinical characteristics, biochemical assessments and cardiovascular risk factors

Traditional cardiovascular risk factors (defined by standard criteria) were recorded at the onset of the study and the baseline laboratory measurements included high-sensitive C-reactive protein (hsCRP); lipoprotein (a) and the complete lipid profile; C3 and C4; interleukin 6 (IL6, assessed by chemiluminescence: cut-off point <3.4 pg/ml); antiphospholipid antibodies (anticardiolipin antibody or lupus anticoagulant); and dsDNA antibodies. C3a, C4a, C5a were assayed in blood collected in the presence of protease inhibitors and a plasma stabiliser (Futhan; BD Biosciences, San Jose, California, USA) using a Cytometric Bead Array (Becton-Dickinson, Franklin Lakes, New Jersey, USA: normal range in our laboratory 0.0–83.1 ng/ml for C3a, 85.7–1965.0 ng/ml for C4a and 162.3–1676.0 ng/ml for C5a). Plasma homocysteine was measured using a fluorescence polarisation immunoassay (Imx-Homocysteine; Abbott Laboratories, Abbott Park, Illinois, USA,), cut-off point ≤15 μmol/dl. An adapted Framingham risk score (for the Spanish population) was used to estimate the 10-year coronary risk6 (see further details in online supplement).

A longitudinal Doppler ultrasound scan (Toshiba Aplio 80) of the carotid arteries was performed on all participants using a standard protocol. Two scans were performed on each patient, one at baseline and one 2 years later. The far wall common carotid IMT measurements were made at predefined sites free of plaques, 1 cm and 1.5 cm from the carotid bulb, defining plaques according to Mannheim consensus criteria.7 The carotid plaques were counted in each territory and the IMT was determined as the average of four measurements. All measurements were performed by the same radiologist, blind to the clinical data and previous scan findings. The intraclass correlation coefficient for common carotid IMT measurements was 0.96 (95% CI 0.80 to 0.99).

Statistical methods

Continuous variables were compared by the Wilcoxon rank test (non-parametric) and the Student t test (parametric) for paired data. Categorical variables were compared in two-way tables with the McNemar exact test.

A multiple linear regression model was used to asses the association between the baseline risk factors and the end IMT (eIMT) adjusted for the baseline IMT (bIMT). These IMT variables were transformed into natural logarithms (ln). The variables entered in the stepwise linear regression included those that displayed significant bivariate relationships with the ln eIMT (p<0.10), or those that were thought to be potential explanatory variables. Analytical and graphical diagnostic methods were implemented to assess the fit of the model. All analyses were carried out using SPSS 15.0 (Chicago, Illinois, USA).


The number of patients finally included in the analysis was 101 and their clinical and biochemical characteristics are shown in table 1.

While 38.6% of patients suffered lupus nephritis, renal insufficiency was only evident in 4% at entry or during the study, and in no case were creatinine levels ≥2. The mean (SD) SLAM score at baseline was 3.8 (2.8) and it remained similar throughout the study (mean of five measurements 3.7), indicating weak lupus activity. The mean (SD) baseline SDI was 1.15 (1.60) and the mean (SD) Katz severity index was also low at 3.02 (2.14). With regard to the cardiovascular risk factors and biochemical parameters (table 1), despite the low levels of overall activity, persistently low levels of complement were detected throughout the study in 49.5% of patients (low C3 and/or C4 at least twice). Homocysteine levels were raised in 42.9% of patients (>15 μmol/dl). As an indication of the low global coronary risk, a low 10-year risk score for coronary events (<5%) was calculated for 88.4% of patients (n=95). At baseline, four patients (4%) had a history of stroke and two more suffered a stroke during the study, while no patient had a history of myocardial infarction or other arterial events.

The mean interval between ultrasound studies was 2.03 years (range 1.49–2.82). The mean (SD) bIMT was 0.37 (0.08) mm and the mean (SD) eIMT was 0.44 (0.09) mm, reflecting a mean (SD) 2-year difference in IMT of 0.078 (0.071) mm (p<0.001).

At baseline, 20 patients had at least one plaque (19.8%) and this figure did not increase significantly during the study, reaching a total of 24 by the end of the study (23.8%, p=0.388).

Simple associations

In the linear regression analysis, the variables associated with ln eIMT when adjusted for the bIMT alone were age at entry to the study (p<0.001), age at diagnosis (p<0.001), postmenopausal status (p<0.001), C3 (p=0.006) and C5a (p=0.029).

Multiple regression analysis

After bIMT and sex adjustment, the age at SLE diagnosis, SLE duration and homocysteine, C3 or C5a levels were identified as significant independent predictors of eIMT (table 2). Furthermore, the eIMT significantly increased in accordance with the increase in the tertile ranges of C5a when adjusted for the rest of the model's variables (figure 1).

Figure 1

Plot of residuals showing changes in Ln eIMT by tertiles of C5a levels when adjusted for the rest of the model's variables. *Ln eIMT, natural logarithm of end intima-media thickness.

Table 2

Simple associations and stepwise multiple linear regression model (R2=0.631) with the natural logarithm of eIMT as the dependent variable


The patients in this study experienced a significant increase in carotid IMT over 2 years which was associated with SLE duration, age at diagnosis, as well as the levels of homocysteine, C3 and C5a in the multiple regression analysis. To our knowledge, the changes in IMT in SLE have only previously been studied twice and, in one of these earlier SLE cohorts, IMT progression was associated with LDL-cholesterol, global disease activity and the accumulated prednisone dose.8 The discrepancy between these results and those presented here could reflect the fact that this was an inception cohort, suggesting that the factors associated with AT in SLE might vary during the course of the disease. Another recent longitudinal controlled study concluded that renal insufficiency was the only factor associated with IMT progression3; however, patients with significant renal insufficiency were excluded from our study.

The association between homocysteine levels and both AT and arterial thrombosis in SLE is well known,9 and our results reinforce the strong association between homocysteine levels and subclinical carotid AT in SLE found previously.2 The exclusion of patients with renal insufficiency in both studies indicates that another mechanism is probably responsible for the hyperhomocysteinaemia in SLE.

Although it was proposed that complement activation may initiate the inflammatory reaction characteristic of AT,10 11 our study is the first to associate C5a serum concentrations with early AT in SLE, suggesting that the complement system could also play a role in this process. Indeed, elevated serum C5a levels seem to be associated with increased cardiovascular risk in patients with atherosclerosis, irrespective of the acute phase response.12 Human endothelial cells (ECs) express the C5a receptor in AT plaques, and it is known that C5a can induce the expression of adhesion molecules in ECs.13 This anaphylatoxin is known to be a potent chemotactic factor and it may participate in initial monocyte recruitment, thereby influencing the artherosclerotic inflammatory milieu. The association of C5a and subclinical AT in SLE and its potential to predict the development of the disease should be confirmed in larger longitudinal studies, perhaps using different end points to measure AT such as plaque formation or coronary calcium.

We also observed a positive association between increased C3 levels and IMT progression. Increased C3 is a recognised marker of atherogenic risk, and it may be a sensitive marker for the early development and progression of atherosclerosis in both the general population and in patients with SLE.3 14 The close relationship proposed between increased C3 and insulin resistance, and consequently with metabolic syndrome,15 could explain the association between C3 and early AT.

Some limitations of this study must be acknowledged. An association between IMT and plaques is not always found in SLE. In fact, in a recent longitudinal study, the annual progression of IMT in patients with SLE was no different from that detected in control patients.3 Doubts have therefore been raised regarding the validity of IMT as a surrogate marker of AT in SLE. However, IMT progression was recently shown to predict cardiovascular events independent of plaque formation in patients with SLE,16 and prospective studies provided solid evidence that IMT measurements may indeed serve as an indicator of subclinical AT in the general population.17

In conclusion, our data support a possible role of C3, C5a anaphylatoxins and homocysteine in the early stages of premature atherosclerosis associated with SLE.


Supplementary materials


  • Funding This research was financed by a FUNCIS Grant 06/03 (Canarias Government) and by a research grant from Schering-Plough, Spain.

  • Ethics approval This study was conducted with the approval of the Hospital Doctor Negrin and patients gave written informed consent.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.