A Five-Year Follow up of Coronary Microvascular Dysfunction and Coronary Artery Disease in SLE: Results from a Community-Based Lupus Cohort

Abstract

Objective:

We investigated prospective change in the prevalence of coronary microvascular dysfunction (CMD) and obstructive coronary artery disease (CAD) in a cohort of systemic lupus erythematosus (SLE) subjects initially evaluated for anginal chest pain (CP).

Background:

Prior work documented a relatively high prevalence of CMD in the absence of obstructive CAD in subjects with SLE.

Methods:

Twenty female SLE subjects with CP who underwent stress cardiac magnetic resonance imaging (CMRI) and coronary computed tomography angiography (CCTA) at baseline were re-evaluated at 5 years.

Results:

Seventeen subjects (85%) were available and re-enrolled, of which eleven (65%) had persistent chest pain on follow up. Fourteen subjects had complete follow up CMRI, of which 36% (5/14) demonstrated CMD on follow up. Further, 25% (¼) of the originally abnormal MPRI at baseline were lower at follow up, while 2 additionally abnormal MPRI values on follow up were noted in previously normal MPRIs. Both prevalence of CMD and non-obstructive/obstructive CAD were unchanged between baseline and follow up, respectively (both p=ns). During follow up, 33% (5/15) had adverse cardiac outcomes including pericarditis, unstable angina or intracranial aneurysm clipping procedure.

Conclusions:

At 5-year follow up of SLE subjects with CP evaluated at baseline and follow up, a majority had persistent CP and nearly half had similar or worse myocardial perfusion consistent with CMD without obstructive CAD. These findings propose an alternative explanation for CP in SLE subjects compared to the more common SLE-related accelerated obstructive CAD accounting for CP and adverse outcomes. These findings support further studies of CMD as an etiology for cardiac morbidity and mortality in SLE.

Significance:

Cardiovascular disease (CVD) is the major contributor to morbidity and mortality in SLE. Our research suggests that coronary microvascular dysfunction (CMD) measured by cardiac magnetic resonance imaging (CMRI) plays a role in SLE.

Innovation:

Our results support that obstructive CAD is not related to persistent chest pain, while persistent/worsened evidence of CMD is associated with persistent chest pain. Advanced imaging allows for a non-invasive evaluation of CVD including CMD, which appears to be prevalent in SLE.

Introduction:

Systemic lupus erythematosus (SLE) is a female predominant, systemic autoimmune disease with cardiovascular disease (CVD) as a leading cause of morbidity and mortality. CVD in SLE appears to have a complex pathogenesis that is incompletely understood and has traditionally been attributed to accelerated atherosclerosis and obstructive coronary artery disease (CAD).(1,2) While obstructive CAD has historically been thought to be a disease of older males, individuals with SLE (female-predominant and often occurring in the child-bearing years) bear a CVD incidence and adverse cardiac event rate of more than seven times the general population.(1) Our group has previously described a high prevalence of coronary microvascular dysfunction (CMD) in women with evidence of ischemia in the absence of obstructive coronary artery disease that is associated with adverse CVD outcomes.(3) Multiple additional contemporary studies report that CMD and associated myocardial ischemia due to endothelial and non-endothelial pathways have an associated increased risk of major adverse cardiac events. (3–9). Furthermore, ischemic heart disease related to CMD more commonly occurs in women, frequently is characterized as atypical and has an annual new myocardial scar rate of almost 2%, often not detected clinically, as we and others have published.(10)

Traditional CAD diagnostics can fail to detect non-obstructive CAD and CMD. Advanced noninvasive assessment by cardiac magnetic resonance imaging (CMRI) has the capacity to identify myocardial ischemia indicated by subendocardial perfusion abnormalities and abnormal myocardial perfusion reserve index (MPRI) in women in the absence of obstructive CAD. (11) We have previously demonstrated a high (44%) prevalence of CMD in in a cohort of 20 female subjects with SLE and chest pain in the absence of obstructive CAD.(12) To further understand the pathophysiology of CVD in SLE and the clinical significance of the findings of CMD in a lupus cohort, we performed a follow up study to evaluate serial changes in the chest pain, CMD and obstructive CAD in a 5-year follow up study of the original cohort.

Subjects and Methods

This study was approved by the institutional review board at Cedars-Sinai Medical Center (CSMC) and all participants gave informed consent prior to baseline and follow up study participation.

Inclusion and Exclusion Criteria

Inclusion criteria consisted of the original 20 female SLE subjects (aged 23 to 59 years at baseline) with chest pain that have previously been described.(12) In brief, eligible subjects were recruited from the rheumatology clinic if they had experienced anginal chest pain and no obstructive CAD by coronary computer tomography angiography (CCTA) at baseline. Subjects were excluded from the follow up imaging study if they were pregnant or had new contraindications to CCTA or CMRI as previously detailed (13) but were still included in the non-imaging follow up.

Assessment of Cardiac Risk Factors and Events

Clinical information was obtained by subject interview and review of the medical record. Chest pain evaluation entailed completion of standardized chest pain symptom questionnaires and quality of life measures relative to chest pain (13,14) in addition to a demographic questionnaire.(15)

Traditional cardiovascular risk factor data (diabetes, smoking, hypertension, hyperlipidemia), and body mass index were collected in addition to SLE disease activity (SLEDAI) calculation (16) and use of hormones, anti-malarials or other SLE medication, non-steroidal anti-inflammatory drugs (NSAIDs), cytotoxic drugs, and steroids (including dose and duration).

Traditional cardiac risk factors and the Framingham, Reynolds and ACC/AHA risk scores were calculated. Additional medication history was obtained, focusing on anti-hypertensive agents, nitrates, and lipid-lowering agents.

Cardiac Magnetic Resonance Imaging (CMRI)

Using our published protocol, enrolled subjects underwent baseline and follow up adenosine stress CMRI at 1.5-Tesla (Siemens Sonata, Erlangen, Germany) with gadolinium first-pass perfusion followed by rest, first pass perfusion, function, myocardial characterization (T2), and delayed enhancement imaging (17)) and analyzed in the CSMC CMRI Core Laboratory. Left ventricular function, volumes, cardiac output, and mass were analyzed using quantitative CMRI analysis software (Pie Medical Imaging B.V., Maastricht, Netherlands). (18,19) Myocardial perfusion reserve index (MPRI) was derived from the total myocardium and subendocardial and subepicardial layers as previously described. (12,20) Of the 15 subjects that underwent follow up imaging, 1 declined the CMRI because of gadolinium allergy concern.

Coronary Computer Tomography Angiography (CCTA)

Subjects underwent baseline and follow up CCTA using a 64-slice dual source scanner at CSMC (Definition, Siemens Medical Systems, Fordheim, Germany). Images were analyzed by consensus of 2 experienced imaging cardiologists blinded to the clinical data, who reported the coronary artery calcium (CAC) score, type (calcified, noncalcified, mixed) and location of each coronary plaque, and the degree of coronary luminal narrowing in the CSMC CCTA Core Laboratory.

Statistical Analysis

Measurements are expressed as mean ± standard deviation. Abnormal CMRI was defined as visual evidence of subendocardial perfusion abnormality (≥5%), MPRI ≤ 1.84 or the presence of late gadolinium enhancement. Obstructive CAD was defined as any epicardial stenosis ≥50% luminal stenosis in ≥1 coronary artery, while non-obstructive CAD was defined as evidence of atherosclerotic plaque <50% in all coronary arteries, with or without calcification. With the goal of minimizing intra-observer variability, interval change in MPRI was determined by one observer using the same original software. For comparison of the MPRI, CCTA and chest pain symptoms between baseline and follow up, Wilcoxon signed-rank test was applied.

Results

From the 20 female SLE subjects that were enrolled in the baseline study, 19 were available for follow up. Two subjects subsequently declined to participate, resulting in re-enrollment of 17 (85%) subjects. Two subjects were then excluded from advanced imaging due to breastfeeding in one and development of comorbidities in the other that limited physical access to imaging follow up (wheel-chair bound).

Subject Characteristics

Baseline demographics, clinical and laboratory characteristics and cardiac risk factors in the 17 follow up subjects are included in Table 1, demonstrating a relatively young female cohort with a low prevalence of traditional cardiac risk factors and a generally mild severity of SLE disease activity. The 10-year Framingham risk score was ≤ 1% in all subjects, (21) as was the Reynold’s risk score, (28) and the ACC/AHA ASCVD risk score. (29)

Table 1.

Follow up study demographics, clinical and laboratory characteristics and cardiac risk factors (n=17) compared to baseline (n=20)

	Baseline (n=20)	Five-year Follow up (n=17)
Age, years	40.6	46.2 (29–65)
Body Mass Index, kg/m2 (range)	25.8 (18–42)	26 (18–45)
Demographics
White, Hispanic	6	6 (35)
White, non-Hispanic	11	8 (47)
Black	1	1 (6)
Asian	1	1 (6)
Mixed	1	1 (6)
SLE Disease duration, years	12.8	18.9 (7–42)
SLEDAI
0	3	5
1–5	10	8
6–10	5	3
>10	0	1
Medication use, current, n (%)
Azathioprine	2 (10)	1 (6)
Hydroxychloroquine	12 (60)	12 (71)
Methotrexate	2 (10)	1 (6)
Mycophenolate mofetil	3 (15)	2 (12)
Corticosteroid use within past one year	16 (80)	7 (41)
Hormone-replacement therapy	6 (30)	5 (29)
Aspirin	Not collected	10 (59)
Beta blocker	1 (5)	6 (35)
ACE-inhibitor/ARB	1 (5)	2 (12)
Cholesterol-lowering agent	1 (5)	7 (41)
Baseline laboratory values, n (%)
Positive ANA,	20 (100)	ND
Positive dsDNA antibody	11 (55)	ND
Smoking history, ever	1 (5)	6 (35)
Diabetes	0 (0)	2 (12)
Hyperlipidemia	1 (5)	5 (29)
Menopausal status	6 (30)	10 (59)
Chest pain in preceding 4 weeks	Not collected	11 (65)
Reynold’s risk score	≤ 1%	≤ 1%

Symptom Change and Event Follow up

Overall, 11/17 (65%) subjects had persistent anginal chest pain at 5-year follow up. While none of the subjects reported an interval MI, stroke or revascularization procedure at 5-year follow up, 1 subject reported pericarditis, 3 endorsed hospitalization for acute coronary syndrome/unstable angina, and 1 had an intracranial vascular aneurysm clipping procedure. The 3/17 (18%) hospitalizations over five years resulted in a 0.9% annual rate, comparable to prior published rates in women with signs and symptoms of ischemia but no obstructive CAD. (30)

Change in CAD measured by CCTA

Three subjects demonstrated progression to non-obstructive or obstructive CAD. On their baseline evaluation, two of these subjects had non-obstructive CAD demonstrated by non-calcified plaque with < 50% stenosis and one was normal at baseline. This subgroup was older (average age 59 vs 48 years) and two of the three subjects reported current smoking on follow up. One subject with progressive plaque, aged 57 years, also had new subepicardial myocardial scar on CMRI. Two of the 15 (13.3%) subjects undergoing imaging demonstrated non-obstructive CAD with CAC scores of 14.9 (baseline CAC 0) and 28.4 (baseline CAC 5.9) and one (6.7%) obstructive CAD with CAC score of 312.1. All subjects with CAD endorsed chest pain in the preceding 4 weeks.

Change in CMD measured by CMRI

All subjects had normal left ventricular ejection fraction on CMRI at baseline and follow up. In total, 36% (5/14) of subjects had CMD on follow up, 60% (3/5) of whom demonstrated persistently abnormal MPRI compared to baseline, while 40% (2/5) presented newly abnormal findings from a previously normal MPRI. Additionally, 75% (¾) of subjects with persistent CMD demonstrated improved MRPI from baseline. Of note, all subjects with improved CMRI findings were on aspirin, beta blocker therapy, and a cholesterol-lowering agent at follow up (Table 1).

Overall group mean MPRI was not significantly different from baseline and follow up (2.03 ± 0.39 vs 2.12 ± 0.58, respectively, p=ns). Among the 11 subjects with persistent chest pain within 4 weeks of follow up visit, 5/11 (45%) had abnormal MPRI at follow up. Table 2 lists all subjects who had any imaging abnormality, either at baseline or 5-year follow up. When categorized as to the previously described MPRI threshold ≤1.84 consistent with CMD (29), among the 4 subjects with a baseline MPRI ≤1.84, 3 again demonstrated an MPRI ≤1.84 while 1 subject’s MPRI exceeded 1.84. Two subjects with MPRI >1.84 at baseline was ≤1.84 on follow up. Only 1 (25%) subject with baseline abnormal MPRI demonstrated a lower MPRI at follow up (p=ns). Further review of the data demonstrated that 4/5 subjects with abnormal follow up MPRI (≤1.84) were obese, had an elevated serum c-reactive protein, and/or reported a current or prior history of smoking and current use of systemic steroids. On follow up, all abnormal MPRI subjects also endorsed depression, for which they were on therapy.

Table 2.

Abnormal imaging: CMRI and CCTA, Baseline and Follow up

Subject	Age	Baseline CMRI (mean MPRI)	Follow up CMRI (mean MPRI)	Baseline CCTA	Follow up CCTA
1	38	1.45 Abnormal	1.78 Abnormal	Normal	Normal
2	34	2.12 Normal	1.46 Abnormal	Normal	Normal
3	56	1.64 Abnormal	1.73 Abnormal	CAC 0, LAD ulcerated plaque	CAC 14.9, new plaque
4	52	1.84 Abnormal	1.95 Normal	Normal	Normal
5	57	2.15 Normal	2.45 New scar Abnormal	Normal	CAC 312.1 with obstructive CAD
6	59	1.62 Abnormal	1.33 Abnormal	Normal	Normal
11	65	2.05 Normal	2.64 Normal	CAC 5.9	CAC 28.4

CAC: coronary artery calcium score; CMRI: cardiac magnetic resonance imaging; CCTA: coronary computed tomography angiography

In reviewing visual perfusion defects on CMRI, 4/14 (29%) subjects demonstrated abnormal hypoperfusion without evidence of CAD on CCTA. One subject, aged 57 years, had abnormal follow up CMRI with obstructive CAD on CCTA and new scar formation that was not seen on baseline imaging (Figure 1). She lacked a family history of cardiovascular disease, denied a history of smoking, and had a BMI of 18 and SLEDAI of 0. However, she did suffer from depression, anxiety, and hypertension and endorsed use of statin, hormone replacement therapy and prednisone medication.

Figure 1. CMRI findings.

A: First-pass perfusion, normal perfusion, mid-ventricle (rest), B: Circumferential subendocardial hypoperfusion, mid-ventricle indicated by arrows (stress), C: Late gadolinium enhancement demonstrating subepicardial hyper-enhancement in the basal inferolateral wall, indicated by arrows, new compared to baseline CMRI 12/2008.

Discussion

At 5-year follow up of SLE subjects with chest pain evaluated at baseline and follow up including advanced imaging with CMRI and CCTA, a majority (11/17, 65%) had persistent anginal chest pain, of which 45% (5/11) demonstrated CMD measured by CMRI. Interestingly, 75% (¾) of subjects with baseline CMD demonstrated improvement on follow up CMRI, one of whom showed resolution of prior findings. Regarding information on interval therapy, Table 1 documents a higher use of cardiac medication on follow up compared to baseline. Relative to specific cardiac therapy and patient improvement, we note that all individuals with improved CMRI findings (¾, 75%) were concomitantly on aspirin, beta blocker therapy, and a cholesterol-lowering agent at follow up, consistent with a prior study by our group and colleagues linking cardiac therapy with improved CMD measured by MPRI.(22)

These findings propose an alternate explanation for chest pain in SLE patients compared to the traditional concept of SLE-related accelerated obstructive CAD as the primary or major cause for chest pain and adverse CVD events in SLE. These preliminary findings support further studies in CMD as an etiology for cardiac morbidity and mortality in SLE. Despite the evidence of myocardial ischemia on CMRI, obstructive CAD on CCTA was most often not present in this cohort at baseline or follow up. The CVD risk scores, including the Framingham (21), Reynold’s risk (23), and ASCVD (29) were all <1% 10 year CVD risk in our subjects. These results support prior reports of underestimating the prevalence of CVD risk in SLE subjects.(21,24)

These data bring to the foreground the possibility of different CVD phenotypes in SLE. Heretofore prior studies in the SLE population have described both traditional and non-traditional CAD risk factors that appear to play a role in endothelial dysfunction and accelerated atherosclerosis (25). Reports from the WISE study attribute less than 20% of conventional atherosclerosis risk factors to CMD (26), suggesting alternative risk variables for CMD. Reis et al (27) identified age and the number of years post-menopause to correlate inversely with reduced coronary flow reserve and further work is needed to investigate this in SLE subjects. While our sample size is limited to test associations, additional studies have associated an increased risk of CVD events in SLE with the use of ≥10mg/day of prednisone, increased SLEDAI, increased dsDNA titer (28), current smoking, longer follow up time, and elevated CRP (29,30).

Limitations of this study include the exclusion of men, although it is important to note that the majority of SLE subjects are women. Additionally, our sample size is relatively small and limits analyses relating risk factors associated with CMD and obstructive CAD progression. Specifically, that the subset of 4 of 5 subjects with abnormal follow up studies were obese or had a history of steroid use or smoking may not be substantiated in larger studies. However, the 5-year interval between assessments and use of advanced imaging to observe objective interval change in our cohort provides robustness to the data interpretation. Specifically, this is the first study to evaluate interval progression of CMRI and CCTA in subjects with SLE and anginal chest pain. While CMRI was used for non-invasive evaluation of myocardial perfusion reserve to detect CMD rather than invasive coronary reactivity testing, a recent report from the WISE study compared traditional (invasive) detection of coronary microvascular disease by coronary reactivity testing and found non-invasive CMRI assessment of MPRI was predictive for presence of abnormal reactivity testing.(31)

In summary, our 5-year follow up evaluation documents a high frequency of persistent chest pain and CMD associated with minimal obstructive CAD in an outpatient SLE population with baseline chest pain. The study findings support contemplating an alternate etiology for persistent chest pain in lupus outside of the dominant paradigm of accelerated obstructive CAD as etiological for chest pain and adverse CVD events in SLE. Furthermore, we have previously published improvement in MPRI with ant-ischemic treatment, implying practical applications of our clinical findings for improved long-term outcomes.(22) Larger longitudinal studies are warranted to further define and understand the phenotypes of coronary artery dysfunction in patients with SLE and chest pain.

Significance and Innovation:

Cardiovascular disease (CVD) is a major contributor to morbidity and mortality in SLE, presumed to be due to accelerated obstructive coronary artery disease (CAD). Traditional cardiac risk factors do not adequately explain the observed CVD risk in SLE. While prior research focused on obstructive CAD, our research suggests that coronary microvascular dysfunction (CMD) measured by cardiac magnetic resonance imaging (CMRI) plays a role in SLE.

Innovation:

Our results challenge the theory of SLE-related accelerated CAD. Specifically, obstructive CAD is not related to persistent chest pain, while worsened evidence of CMD is associated with persistent chest pain. CMRI allows for a non-invasive evaluation of CVD including CMD, which appears to be prevalent in SLE.

Abbreviations List:

CAC

Coronary Artery Calcium

CAD

Coronary Artery Disease

CCTA

Coronary Computed Tomography Angiography

CMD

Coronary Microvascular Dysfunction

CMRI

Cardiac Magnetic Resonance Imaging

CSMC

Cedars-Sinai Medical Center

CVD

Cardiovascular Disease

MPRI

Myocardial Perfusion Reserve Index

NSAID

Non-steroidal Anti-inflammatory Drug

SLE

Systemic Lupus Erythematosus

SLEDAI

Systemic Lupus Erythematosus Disease Activity Index