Background
Rheumatoid arthritis (RA) is an inflammatory autoimmune disease associated with a higher prevalence of cardiovascular (CV) disease, both of which are linked to the presence of traditional risk factors for atherosclerosis as well as chronic inflammation. Accelerated atherosclerosis of the coronary and cerebrovascular arteries is the main cause of morbidity and mortality in RA [
1]. Patients with RA have a 45 % higher risk for all CV events than the general population, the risk for myocardial infarction is even higher (68 %) [
2]. In addition, 30-day mortality after the first acute coronary event is 80 % higher in patients with RA than in the general population [
3]. Early identification of high-risk patients and the initiation of preventive measures is, therefore, very important.
Classical CV Systematic Coronary Risk Evaluation (SCORE) may underestimate the actual risk in patients with RA, in whom inflammation plays a major role [
4,
5]. Thus, the European League Against Rheumatism (EULAR) recommends multiplying the risk measured by SCORE by a factor of 1.5 in selected RA patients [
6]. This applies to patients meeting 2 or more of the following criteria: duration of disease >10 years; positivity for rheumatoid factor or anti-citrulline antibodies; and the presence of certain extra-articular symptoms [
6].
Currently, several imaging methods for the detection of coronary atherosclerotic lesions are available. Non-contrast CT is frequently used for assessment of calcium burden using Agatston calcium score [
7]. Coronary CT angiography (cCTA) is, at present, used routinely in the diagnosis of coronary heart disease (CHD). This non-invasive method has high accuracy in detection of coronary stenosis and particularly predictive value in patients with negative findings [
8‐
10].
The aim of the present study was to detect the following in asymptomatic patients with RA: the prevalence of significant coronary stenosis measured by cCTA; how coronary atherosclerosis diagnosed by CT examination correlates with classical or modified cardiovascular SCORE systems; and the predictive value of biomarkers in assessing the severity of coronary stenosis.
Methods
Study design
The prospectively designed single centre study included patients with RA who were at least 1 year on biological treatment observed in specialized outpatient clinic for rheumatological diseases during March and December 2013. Exclusion criteria were: 1) no other CV disease apart from arterial hypertension; 2) negative family medical history for diseases of the CV system; 3) current smoking; 4) symptoms suggesting the possibility of CHD; 5) contraindication to iodine contrast agent administration (allergy on iodine contrast agent and severe renal insufficiency). As RA is 3 times more frequent in women than in men, in those subjects eligible for this study, were only 2 men. To achieve higher homogeneity of our study we decide not to include them into present analysis. We further not included 1 patient because of allergy to iodine, 2 other women refused to participate.
All included patients were referred to cCTA and to take blood samples for laboratory assessment. Basic medical data, including co-morbidities, medications used, resting blood pressure, heart rate, height and weight were collected. Body mass index was calculated as body weight (kg)/height2 (m2). The stage of RA and the diagnosis of arterial hypertension, diabetes mellitus and dyslipidaemia were identified from medical documentation. The dose of corticosteroids used was recalculated as the equivalent dose of prednisone.
The CV SCORE (SCORE) and modified SCORE (mSCORE) were calculated in accordance with previously published recommendations [
4,
6]. The CV SCORE takes into account patient’s sex, age, smoking status, systolic blood pressure and concentration of total cholesterol. Resulting number represents individual 10 year risk for fatal CV event. Modified SCORE multiplies the risk measured by SCORE by a factor of 1.5 in selected RA patients [
6].
The Ethical Committee at the Faculty Hospital in Pilsen validated the study protocol, and all patients provided written informed consent. The study conformed to the provisions of the World Medical Association’s Declaration of Helsinki.
Laboratory assessment
Standard biochemical parameters were assessed in the laboratory at the Faculty Hospital in Pilsen. The estimated glomerular filtration rate (eGFR) was calculated using the Modification of Diet in Renal Disease (MDRD) formula. Albuminuria is expressed as a urinary albumin-to-creatinine ratio (ACR). High-sensitivity cardiac troponin I (hsTnI) was measured using the Architect i2000 platform with STAT High Sensitive Troponin-I assay (Abbott Diagnostics, USA) (limit of quantification 4–10 ng/l, 10 % coefficient of variation 4.7 ng/l, 99th percentile 26.2 ng/l, percentage of measurable values in healthy population >80 %). Plasma levels of osteoprotegerin were measured using multiplex immunoanalyses based on xMAP® technology, commercially available kits MILLIPLEX MAP Human Bone Panel 1A magnetic (Merck-Millipore Corporation, USA) and the MAGPIX instrument (Luminex Corporation, USA). xMAP® technology and the MAGPIX system uses color-coded magnetic microspheres coated by antibodies to perform quantitative sandwich immunoanalysis of proteins in a variety of sample matrices.
CT examination
Scanning parameters
CT examinations were performed using a second-generation dual-source CT device (SOMATOM Definition Flash, Siemens HealthCare, Forchheim, Germany). Initially, a non-contrast scan for calcium scoring quantification was performed. Cardiac CTA was monitored by retrospective electrocardiographic gating (128 × 0.6 mm; rotation 280 ms; adaptive pitch factor 0.2 − 0.4; optimized setting of tube voltage and tube current using CareDose and CarekV). Fifty millilitres of iodine contrast agent (400 mgI/l) was administered at a rate of 6 ml/s.
Image analysis
Assessment of the examinations was performed in consensus by 2 radiologists experienced in CT heart examination (20 and 10 years, respectively). Calcium score was calculated using a dedicated Syngo. CaScoring application (Siemens Healthcare, Forchheim, Germany). The total calcium score, without differentiation of individual coronary arteries, was used for the analysis.
Coronary arteries were assessed according to the degree of luminal stenosis (mild: 10 % − 40 %; moderate: 40 % − 70 % and severe: >70 %). The significance of the degree of stenosis was assessed for each patient, and the characteristics of plaques were assessed according to individual segments [
11].
Statistical analysis
For statistical analysis, SAS software version 9.3 (SAS Institute Inc, USA) was used. The results are presented as arithmetic mean ± standard deviation, median with inter-quartile range (IQR) or as a proportion (percentage). For the purpose of statistical analysis, the subjects were divided into three groups according to maximal stenotic involvement of coronary arteries: without stenosis or mild stenosis according to cCTA (coronary stenosis <40 % [n = 28]); moderate stenosis (40 % − 70 % [n = 12]); and a group with severe stenosis (>70 % [n = 4]).
Differences among the groups were assessed using the paired Student’s t test, the Kruskal-Wallis test and the Fisher’s exact test. Correlations were expressed using the Spearman correlation coefficient. For the purposes of regression analysis, non-normally distributed variables were normalized using logarithmic transformation. Factors associated with a positive finding in CTAG were analysed using logistic regression. Significant attention was devoted to the selection of potential confounders. First, potentially significant covariables (age, arterial hypertension, dyslipidaemia, stage of RA, albuminuria, hs troponin I, dose of corticoids) were sought with a stepwise logistic regression analysis, which included only factors demonstrated to be significant in the univariate analysis. The final regression model included only parameters that remained in the models permanently.
Discussion
In our study involving 44 women with rheumatoid arthritis without clinical symptoms of ischemic heart disease, we found that 9 % of them had severe coronary stenosis detected by coronary CT angiography. The classical cardiovascular risk SCORE or modified risk SCORE were not associated with coronary stenosis. On the other hand, high sensitivity Troponin I was correlated with the number and severity of coronary artery lesions.
RA is a disease associated with accelerated atherosclerosis and higher mortality after myocardial infarction in affected individuals [
2,
3]. Early detection of significant CHD, thus, plays an important role in patient prognosis. Coronary CT angiography is non-invasive and safe method with possibility of excellent depiction of coronary plaques and stenosis assessment [
12,
13].
High sensitivity cardiac troponins have demonstrated utility in estimating prognosis in general population and in patients with coronary heart disease, even in the range far below the 99th percentile (recommended cut-off value for diagnosis of acute myocardial infarction). deFillipi et al. [
14] found a HR of 2.91 for cardiovascular mortality in older adults without known heart failure (>65 years; hsTnT > 12.94 ng/l vs < 3 ng/l). Omland et al. [
15] (PEACE study) showed a HR 2.39 for cardiovascular mortality in stable CHD patients (hsTnT > 9.6 ng/l in men and > 7.4 ng/l in women) and Kavsak et al. [
16] observed a HR 2.2 for cardiovascular mortality in stable patients at risk for CHD or chronic heart failure (hsTnI > 10 ng/l). Accordingly, hs troponins may be useful for risk stratification in patients with RA. In our study, hsTnl level had a high negative predictive value (18 of 21 patients without any coronary lesions had hsTnI < 3 ng/l).
Bradham et al. demonstrated that patients with RA had 49 % higher hsTnl levels than healthy individuals [
17]. In the same study, the hsTnl value correlated with the calcium score in univariate analysis; however, after adjustment for age, race, sex and Framingham risk score, this association was no longer statistically significant. Our data suggest that female RA patients with the highest concentrations of hsTnI tend to have severe coronary stenosis (adjusted OR 5.39).
Of note, 99th percentile derived from our population of RA women without any signs of acute myocardial ischemia was 25.7 ng/l, whereas 99th percentile of the reference female population was 15 ng/l [
18]. This fact should be probably considered in diagnostics of acute coronary syndromes in RA patients.
Moreover, we highlighted a potentially interesting connection between albuminuria and severe coronary stenosis. Albuminuria is a condition for which testing is readily available. Moreover, the presence of albuminuria increases CV risk by a factor of 2 − 4 [
4]. Sihvonen et al. demonstrated that in patients with RA, the presence of albuminuria (in the range of 20–200 μg/min) increases the risk of mortality by 180 % compared with patients without simultaneously increased ACR [
19]. In our study, albuminuria was correlated (borderline significance) with the severity of coronary stenosis, but not with the calcium score or the number of coronary lesions.
Other studies have demonstrated a connection between inflammatory markers and CV risk [
1,
20,
21]. In our groups, however, we did not find any difference, either in erythrocyte sedimentation rate or C-reactive protein levels between women with positive and negative findings on cCTA.
Surprisingly, in our study classical or mSCORE systems did not correlate with the severity of coronary stenosis [
4,
6]. On the other hand, calcium score enables to diagnose the severity of calcification in the coronary arteries. Increased calcium scores were detected in patients with RA [
11] and were correlated with the occurrence of CV complications [
22]. In our cohort, the calcium score was a very good predictor of severe coronary stenosis (AUC = 0.94). Moreover, it was significantly correlated (r = 0.96) with the total number of coronary lesions. Given its ready availability and its non-invasive character, cCTA is a suitable method for the detection of coronary atherosclerosis. A negative finding has a particularly strong predictive value with regard to stenotic lesions [
23]. Coronary CTA is comparable with invasive angiography in its precision for assessing the degree of possible coronary arterial stenosis. However, its limitation lies mainly in its low specificity, which may overestimate stenosis, particularly in cases involving sclerotic plaque [
23]. However, high sensitivity of cCTA is more important factor in our study and a negative cCTA finding has a strong predictive value with regard to stenotic lesions [
24].
The present study had several limitations. First, it investigated a relatively small group of women with RA; our results, thus, cannot be generalized to include men or women without RA. However, our patients were comprehensively examined using a CT method that achieves excellent results in diagnosing CHD [
25]. Second, the present study was cross-sectional in design; therefore, we have no long-term follow-up information regarding patient outcomes, although individuals with severe coronary stenosis have been shown to have a higher risk for major adverse cardiac events [
26]. Third, the low prevalence of severe coronary stenosis in our group may lead to a skewed interpretation of the results, although we used several different statistical methods to analyse our data with confirmatory results. Four, we did not perform any stress test to assess functional severity of stenosis detected on cCTA. Moreover, we performed an invasive coronary angiography only in patients with finding of severe stenosis (<70 %). However, all patients were completely asymptomatic and we focused on non-invasive examination with minimized radiation and overall burden.