Abstract
Objective To evaluate the progression of coronary artery calcification (CAC) and associated risk factors in a systemic lupus erythematosus (SLE) cohort.
Methods We reassessed the presence of CAC in patients with SLE who were screened 9 years before, using multidetector computed tomography. Clinical variables (cumulated disease activity and damage accrual), antiphospholipid syndrome and SLE serology, and cardiovascular (CV) risk factors (hypertension, BMI [kg/m2], modified Framingham risk score, lipid profile, menopausal status) were assessed longitudinally.
Results We included 104 patients from the parent study. Most of them were women (94.2%), with a mean age of 41.0 (SD 8.3) years and mean disease duration of 14.8 (SD 2.9) years. We documented CAC in 17 patients (16.3%). Seven cases were from the parent study and 10 were incident cases. The cumulative incidence of CAC was 9% and the incidence density was 1 per 100 person-years. CAC occurred more frequently in the age groups 30-39 years and 40-44 years. All patients with previous CAC had worsening of their calcium indexes, and none developed clinical CV events. When comparing prevalent CAC cases (n = 17) vs patients without calcification (n = 87), both groups were similar in traditional CV risk factors, disease duration, Systemic Lupus Erythematosus Disease Activity Index 2000 (SLEDAI-2K) area under the curve (AUC), and Systemic Lupus International Collaborating Clinics (SLICC) score, but were more likely to be postmenopausal and have higher apolipoprotein B (apoB) levels. Patients with previous CAC had higher apoB levels, SLEDAI-2K AUC scores, and anticardiolipin IgG antibodies than incident cases.
Conclusion CAC in patients with SLE progressed over time but was not associated with adverse CV events during the first 9 years of follow-up. ApoB levels and postmenopausal status might be associated with this progression.
Systemic lupus erythematosus (SLE) is an autoimmune disease of unknown etiology characterized by diverse clinical manifestations and disease courses. Approximately 30% to 40% of patients with SLE may develop atherosclerosis, particularly early in the course of the disease.1 The occurrence of premature atherosclerosis in SLE has a multifactorial etiology involving traditional factors (hypertension [HTN], dyslipidemia), immune-related factors, nephrotic syndrome, primary and secondary disorders of lipoprotein metabolism, steroid use, endothelial damage by immune complexes, and microthrombosis secondary to antiphospholipid syndrome (APS).2
Coronary artery calcification (CAC) scanning by cardiac computed tomography (CT) is used for the identification of subclinical atherosclerosis as well as prognostic stratification of asymptomatic individuals. Previously, our research group investigated the presence of CAC using CT in a cohort of patients with recent-onset SLE. We found a prevalence of 7.2% compared with 1% of the subjects matched by sex and age (odds ratio [OR] 7.7), and the associated factors were disease duration, age at admission, and cumulative disease activity. In fact, CAC was detected in patients from age 23 years and 3 years after diagnosis.3
In this study, we reexamined the same cohort of patients 9 years later to determine the progression of coronary disease and to evaluate the associated risk factors.
METHODS
This was a cohort study in a tertiary referral center. Briefly, the cohort was established in 1999 and included patients with SLE according to the 1997 American College of Rheumatology criteria and disease duration of up to 12 months at inclusion.4 A standardized history, physical examination, and laboratory tests were performed in all patients. Every 3-6 months, disease activity was assessed using the Systemic Lupus Erythematosus Disease Activity Index 2000 (SLEDAI-2K),5 and damage accrual was assessed annually using the Systemic Lupus International Collaborating Clinics (SLICC) damage index.6 We also registered treatment, comorbidities, and traditional cardiovascular (CV) risk factors (HTN, diabetes mellitus [DM], menopausal status, smoking, cholesterol levels, apolipoprotein B [apoB] levels, BMI [calculated as weight in kilograms divided by height in meters squared], high sensitivity C-reactive protein [hsCRP], and the Framingham 10-year risk factor profile).3
Patients. In 2008, some of the patients were invited to participate in the CAC study (parent CAC study).3 In that study, we excluded patients with pregnancy as well as those with previous CV events (myocardial infarction [MI], angina, stroke, transient ischemic attack, and artery disease). The parent calcification study included 139 patients, all with a similar disease duration. These patients underwent baseline coronary artery CT and then they were followed clinically through the years. In the present study, we reperformed a coronary artery CT of all the patients from the parent study who were still under follow-up. We excluded patients with pregnancy or calcium metabolism disorders.
Patients were asked to refrain from drinking alcohol or caffeine for 24 hours before the study. All were interviewed by a single rheumatologist to record demographics, disease duration, somatometry, and CV events. At the baseline and follow-up coronary artery CT, we registered the presence of comorbidities (DM, dyslipidemia, HTN), menopausal status, Framingham risk score, modified Framingham risk score, cumulative activity (mean SLEDAI-2K), damage accrual, and lipid profile. We also assessed the presence of the following CV outcomes during the follow-up period: CV death, MI, angina, transient ischemic attack, arterial insufficiency, and stroke.
Coronary imaging. Coronary artery images were acquired in a 64-slice multidetector CT system (Somatom Sensation Cardiac 64, Siemens) using the same acquisition protocol as in the parent study, in compliance with recommended variables.7 The best diastolic cardiac phase was selected for reconstruction of images, analyzed on a dedicated workstation using calcium scoring software (Calcium Scoring, Siemens). The software uses an attenuation threshold of 130 Hounsfield units (HU) and a minimum of 3 contiguous pixels for the identification of a calcified lesion. Each focus that exceeded the minimum criteria was scored using the algorithm originally developed by Agatston et al.8 The total calcium score is extended by adding the individual lesion losses from each of the arteries, which includes the left main coronary artery, left anterior descending artery, and left and right circumflex arteries. The coronary CT reading was performed by 1 of the 2 radiologists who participated in the CAC parent study. We considered progression of CAC as an increment > 25% among cases with a previous value ≥ 1, or a value > 0 in the nonzero patients, as a variation ≤ 25% could be attributed to the method itself.
Statistical analysis. We used descriptive statistics according to the distribution of the variables. Univariate analysis was performed using the t test, Mann-Whitney U test, chi-square test, or Fisher exact test according to the variable. We calculated the area under the curve (AUC) of the cumulated disease activity (SLEDAI-2K) to include the whole activity of the disease throughout the follow-up period.
We performed backward logistic regression and reported the OR with a 95% CI. We introduced the variables with a P value < 0.05 to the model for the univariate analysis. We also reported the cumulative and density incidence of CAC. We considered a value of P ≤ 0.05 as statistically significant.
The study was approved by our local review board (IRE-508) and all patients signed an informed consent form according to the Declaration of Helsinki.
RESULTS
We included 104 of the original 139 patients who participated in the parent study. The other 35 patients were not included for the following reasons: 16 had died, 13 had been lost to follow-up at our institution, and 6 patients had been invited but never attended. None of the excluded patients experienced a CV event. When we compared the patients included in this study vs the nonparticipating group, they were similar in age at inclusion (mean [SD] 31.7 [8.4] vs 32.1 [10.1]; P = 0.87), sex (women, 94.2% vs 88.5%; P = 0.23), postmenopausal status (7.7% vs 8.5%; P > 0.99), positive CAC at first CT (6.7% vs 5.7%; P > 0.99), SLE duration at first CT (mean [SD] 5.9 [2.9] yrs vs 4.8 [3.0] yrs; P = 0.06), and cumulative SLEDAI-2K AUC (mean [SD] 4.9 [2.9] vs 5.3 [3.6]; P = 0.11).
Of the 16 patients who died, none had a positive CAC, and the causes of death were infection, neoplasia, and end-stage renal disease. However, they had a higher cumulative SLEDAI-2K AUC score (29.4 vs 23.7 points) and a higher cumulative prednisone dose (29.4 g vs 23.7 g).
Of the included patients, 94.2% were women, with a mean age of 41.0 (SD 8.3) years and had a mean disease duration of 14.8 (SD 2.9) years at last visit. Table 1 depicts the general features of these patients.
Characteristics of patients with and without CAC.
Prevalence and incidence of CAC. We documented CAC in 17 patients; thus, the prevalence was 16.3% (95% CI 10.41-24.63). Seven cases were from patients previously identified in the parent study, and 10 were new cases (incident cases). The cumulative incidence of CAC was 9% and the incidence density was 1 per 100 person-years.
The median calcium score was 247.6 (IQR 0.1-1858.1) HU. Prevalent cases of CAC were more frequent in the age ranges of 30-39 years (43.2%) and 40-44 years (24%), followed by the age groups 45-54 years (20.1%) and 55-64 years (6.7%).
Progression of CAC and its associated factors. As all the patients with previous CAC had worsening of their calcium indexes (Table 2), we were unable to make a comparison with those who did or did not show progression. However, we compared the 17 patients with prevalent CAC vs the 87 patients without CAC and found no differences in demographics and traditional CV risk factors (Table 1). Nevertheless, we documented a higher prevalence of postmenopausal women and a higher level of apoB during follow-up in the CAC group. Both groups were similar in terms of disease duration, SLEDAI-2K AUC, SLICC score, SLE and APS serology, and use of steroid, antimalarials, and hyperlipidemia treatment (Table 1). For the multivariate analysis for the outcome of prevalent CAC, we included postmenopausal status and apoB levels at follow-up in the model; however, neither remained significant.
Calcium index scores of prevalent cases.
Then, we compared the 7 patients with previous CAC vs the 10 incident cases. Again, the groups were similar in terms of demographics, disease duration, treatment, and traditional CV factors. However, the former group had higher apoB concentrations during follow-up, and a tendency toward being postmenopausal. Interestingly, the 7 patients also had a higher SLEDAI-2K AUC score and a higher prevalence of anticardiolipin (aCL) IgG antibody isotype than the incident cases (Table 3). Because of the limited number of patients in each of the groups, it was not possible to perform multivariate analysis.
Characteristics among patients with CAC.
CV outcomes. During follow-up between scans, cerebrovascular events occurred in 7 patients (none of them had CAC). Moreover, none of the patients had an acute MI, angina, transient ischemic attack, or arterial insufficiency.
DISCUSSION
Premature coronary heart disease is a major cause of death in SLE.9 Moreover, CAC may be found in approximately 7.2% to 30.7% of asymptomatic patients with SLE, even in patients < 50 years of age with an early disease onset.3,10-13 In a metaanalysis of 24 studies (918 patients with SLE and 3952 controls), the pooled prevalence for CAC was 29.8% in patients with SLE and 11.8% in controls (risk ratio 2.2).14 In addition, the risk factors associated with CAC in SLE are smoking, low high-density lipoprotein (LDL) levels,10 older age,3,11,12 male sex,11 high homocysteine levels,12 low glomerular filtration rate,12 a longer disease duration,3,12 and high SLEDAI-2K AUC.3 Previously, CAC was associated with elevated ceramides of very long-chain fatty acids; alterations in metabolic pathways of purine, arginine, and proline; and microbiota-derived metabolites. Indeed, the combination of metabolic and lipidomic variables with clinical variables (age, HTN, disease duration, and SLEDAI-2K > 6) and biological variables improved the prediction of CAC (AUC 0.88). However, in that study, there was an inverse relationship between SLE activity and CAC.15 Now, in our prospective cohort, we documented a cumulative incidence of CAC of 9%.
Regarding CAC progression, a previous study found that among asymptomatic individuals who underwent coronary calcium scoring as a part of a checkup, 11.5% experienced CAC progression during a mean follow-up of 109 months, with the risk factors being age, male sex, waist circumference, DM, and LDL cholesterol level.16 On the other hand, a randomized, double-blind, placebo-controlled trial of atorvastatin vs placebo in patients with SLE showed that the trend in calcium score (change of > 10 Agatston units) from a baseline CT to a subsequent CT performed 2 years later remained the same in 76%, worsened in 3%, and increased in 20%.17 Conversely, in our study, the calcium score increased in all patients with a prior CAC > 0, using a higher cut-off definition.
We have several considerations about our results. First, we found a low prevalence and incidence of CAC compared to other SLE studies. This finding might be influenced by the races and ethnicities of our patients. In the Multi-Ethnic Study of Atherosclerosis (MESA) cohort, White men had the highest rate of nonzero calcium scores, whereas Black women and Hispanic women had the lowest rate.18 Second, the calcium scores of patients with SLE in our study are strikingly high when compared to both the subgroup of the Hispanic population of the MESA cohort18 and the Genetics of Atherosclerotic Disease (GEA) cohort, which included Mexican patients free from CV disease.19 For instance, in the GEA cohort, the expected normal values of the calcium score (tertile 75) for women aged 45 to 54 years, 56 to 64 years, and 65 to 74 years were 0, 0, and 7.2, respectively. Third, interestingly, we did not observe any CV events during follow-up. In the setting of rheumatoid arthritis, a calcium score > 100 implies a higher rate of major CV events20; however, only half of our patients scored above that range.
Regarding risk factors associated with progression of CAC, Kiani et al found no association between the Safety of Estrogens in Lupus National Assessment (SELENA)-SLEDAI, anti-DNA, and progression of CAC during a follow-up period of 2 years. However, age, current smoking, and low hsCRP were associated with progression of CAC.17
In our parent study, we found that age, disease duration, and moderate/severe cumulative disease activity were associated with CAC.3 These results were not replicated in the present work. However, we now found that menopausal status was associated with CAC in the univariate analysis. Menopausal status may be an indirect variable of age and may favor the development of atherosclerosis, pericardial fat, and calcium index scores.21 In addition, we also found an association between apoB and CAC. In the overall population, the prevalence of CAC increased according to the quartile values of apoB levels.22 Moreover, a randomized, double-blind, placebo-controlled trial investigating the effect of atorvastatin on the incidence of CAC in SLE found that after 1 year of follow-up, there was a statistically signficant increase in the calcium index score, from a mean of 32.1 (SD 39.1) to 59.5 (SD 64.4) in the placebo group but not in the statin group (44.8 [SD 50.4] to 54.9 [SD 62.56]).23
Regarding disease activity, in our study, the patients who had CAC at the first screening had a higher SLEDAI-2K AUC score than the incident cases. This finding is in agreement with Urbain et al, who considered that the deleterious effect of disease activity occurs primarily in younger and less disease-exposed patients with SLE rather than older patients with longer disease duration.15 Moreover, our incident cases had a lower prevalence of acL IgG. Conversely, Urbain et al did not find any association between the antiphospholipid antibody profile and CAC.15
Finally, according to the American College of Cardiology/American Heart Association guideline on primary prevention of CV disease,24 the CAC measurement candidates are men aged 55-80 years and women aged 60-80 years with a low risk factor burden and who question whether they should benefit from statin therapy, as well as middle-aged adults (aged 40-55 yrs) with a pooled cohort equations–calculated 10-year risk for atherosclerotic CV disease (ASCVD) < 7.5% who have risk factors. As SLE is a recognized risk-enhancing factor, CV-asymptomatic patients with SLE aged > 40 years with an ASCVD risk < 7.5% should be screened for CAC. However, we should stress that younger patients with SLE might present a positive calcium score even within a few years of disease onset. Moreover, in the non-SLE population, the timing to reperform a calcium score with a previous value of 0 is every 5 to 10 years if no other risk factor was added, whereas the timing in patients with SLE is unknown.24
Our study has some potential limitations. First, we must consider the risk of attrition bias in our cohort, as we lost 25% of our patients. Nevertheless, we did not find differences when we compared the patients who were lost to follow-up to those who continued. Second, our negative results regarding progression risk factors in the multivariate analysis are likely because of the low incidence of CAC; as previously mentioned, these results might be driven by the races and ethnicities of our patients. Thus, a large multicohort study is needed to fully address this point. However, our study has the following strength: it was conducted in a well-established cohort followed prospectively at regular intervals. To our knowledge, our study has the longest follow-up assessing calcification progression in the literature.
In conclusion, CAC in patients with SLE progressed over time but was not associated with adverse CV events in the first 9 years of follow-up. ApoB levels and postmenopausal status might be associated with this progression.
Footnotes
This manuscript was funded by grants INCMNSZ-AI-024 and CONACYT SALUD-2014-214395 (to JRD).
The authors declare no conflicts of interest relevant to this article.
- Accepted for publication June 14, 2024.
- Copyright © 2024 by the Journal of Rheumatology
