Introduction
Rheumatoid arthritis (RA) is an independent risk factor for cardiovascular disease (CVD) that is associated with at least a 1.5-fold increased risk for a fatal coronary event compared to the general population [
1]. This is based predominantly on data derived from epidemiological studies conducted during the past two or three decades, which included many RA patients diagnosed and managed before the era of aggressive, target-driven therapy with disease-modifying antirheumatic drugs (DMARDs), including biologic agents [
2]. Complementary data deriving from cross-sectional studies have shown that RA patients have evidence of advanced preclinical carotid atherosclerosis compared to healthy controls [
3] to a magnitude similar to that observed in patients with diabetes mellitus (DM) [
4]. High-resolution B-mode ultrasonography of the carotid artery provides a noninvasive, valid and reproducible method for identifying atherosclerotic plaques, which reflect prevalent, clinical or preclinical CVD and may represent predictors of future CVD events [
5]. Notably, in RA patients without traditional CVD risk factors or events, an increased intima-media thickness (IMT) of the common carotid artery and evidence of focal plaques were each predictive of incident CVD events [
6,
7].
There is an increased burden of traditional CVD risk factors in RA patients [
8], including hypertension [
9], dyslipidaemia [
10], smoking [
11], insulin resistance [
12], obesity and/or altered body composition [
13,
14] and physical inactivity [
15], but this accounts only partly for the excess CVD mortality [
16]. Because inflammation plays a major role in the atherosclerotic process [
17], interest has focused on the impact of the systemic inflammatory burden and chronic activation of the immune system [
18] present in RA [
19,
20]. Given the paucity of prospective data, the extent to which traditional CVD risk factors and RA-related parameters (inflammatory burden, activity and/or remission and treatment modalities) interact and/or contribute to atherosclerosis acceleration remains inconclusive [
21]. According to a recent study, both traditional CVD risk factors and markers of RA severity at baseline contribute to models predicting cardiovascular (CV) events in the subsequent 22 months [
22].
In the present study, we assessed (1) the factors associated with the formation of at least one new carotid plaque per subject during the follow-up period in nondiabetic RA patients and (2) whether the number of newly formed plaques in these patients differs from the respective number observed in non-RA subjects carefully matched for traditional CVD risk factors.
Discussion
In this prospective, longitudinal study, we examined factors leading to new carotid plaque formation in nondiabetic patients with RA without CVD at baseline. As recently reported, carotid atherosclerosis predicts future coronary events in RA, whereas those patients with carotid plaques, multiple CV risk factors (particularly DM or hypertension), many swollen joints and a high cumulative dose of glucocorticoids, as well as male RA patients, are at even higher risk [
7]. We focused on new carotid plaque formation and not on IMT progression, although both biomarkers have been associated with CV events in RA and non-RA populations and may have additive predictive value [
6,
32]. Our decision to exclude IMT from the analysis was based on the fact that, though focal plaque is the direct manifestation of atherosclerosis, IMT is considered a surrogate not only of diffuse atherosclerosis but also of medial hypertrophy, unrelated to atherosclerosis, that is driven primarily by hypertension [
32].
Our study produced three main novel findings which were consistently verified by different statistical approaches. First, in well-treated RA patients in clinical remission for more than two-thirds of the follow-up period, new carotid plaque formation was predominantly associated with traditional CVD risk factors. Second, of all RA-related parameters evaluated, only chronic corticosteroid use was independently associated with new carotid plaque formation. Third, over an average 3.5 years follow-up period, the number of new carotid plaques in the subgroup of RA patients and the tightly matched for traditional CVD risk factors non-RA subjects, was similar.
In the present study, a relatively small sample size was available, probably resulting in insufficient statistical power to detect small differences, especially at baseline, between patients with or without new plaque formation during follow-up. Another limitation is the fact that we focused solely on new plaque formation and did not assess the progression of preexisting plaques. Such assessments ideally require three-dimensional models that were not available to us, whereas the potential error that would have been introduced by the available two-dimensional models might have been too large.
Cross-sectional studies have shown that the severity of inflammatory burden in long-standing RA measured by CRP, as well as disease duration, are associated with increased preclinical atherosclerosis in patients without traditional CVD risk factors [
33]. In the present prospective study, however, we did not observe any association between new carotid plaque formation and RA disease-related parameters, including CRP and ESR levels. We think that, in a prospective study, the influence of patients' characteristics at baseline is less important than those present during the follow-up period. The lack of such an association may be explained by the low residual inflammation present in our optimally treated patient cohort during the follow-up period as well as by the relatively high burden of traditional CV risk factors in the present cohort, which constituted a typical RA population. All patients were managed according to the latest approaches of treating to target (that is, disease remission) with very frequent follow-up, rapid escalation of disease-modifying therapy and unrestricted access to biologic agents (given in 52% of patients), thus resulting in clinical remission for more than two-thirds of the follow-up time. As most recently reported in another prospective study assessing longitudinal predictors of progression of carotid atherosclerosis in RA, inflammation was indeed a contributor. However, the probability of atherosclerosis progression was minimal up to an average CRP threshold of 12 mg/L, whereas, more importantly, the average CRP (median of cumulative average during study period of 3.1 mg/L) in nondiabetic RA patients was not associated with carotid atherosclerosis progression [
34].
The finding that the formation of new carotid plaques in patients with well-controlled RA depends mainly on traditional CVD risk factors was reproduced in the comparative cohort study. In this comparative study, patients had higher BP than controls by almost 5 mmHg, yet they developed a similar number of new plaques over a 3.5-year period. Because the younger age of patients in the RA matched population versus the remaining patients could constitute a selection bias, part of the results may not be representative and thus cannot be extrapolated to other age groups.
The findings presented herein add credence to the current European Union League Against Rheumatism guidelines on CVD risk management in RA [
1]. Because our present study was conducted before the publication of these guidelines, our patients did not receive aggressive CVD risk reduction management. Smoking, which has a higher prevalence in Greece than in other European Union countries [
35], and age were the main factors defining the new carotid plaque formation in the present population, as expected in non-RA populations [
36,
37]. This finding clearly shows the need to optimally control classical CVD risk factors in RA patients, factors which are often neglected in everyday clinical practice. Of note, no interaction between inflammation and classical CVD risk factors and no synergy between them were found. Also, we did not find any interaction between age- and RA-related parameters (inflammation or drugs) regarding their effect on new plaque formation. However, since the age at onset of RA was greater in patients who developed new plaques than in those who did not (61.2 ± 7.8 versus 47.3 ± 12.7 years of age respectively,
P < 0.001), it is possible that older arteries may also respond less well to inflammatory burden and/or steroids and hence the process of atherosclerosis may be accelerated.
In contrast to the known detrimental role of high-dose corticosteroids in CVD risk, until recently there was no clear evidence for the role of a chronic low daily dose of corticosteroids in patients with inflammatory arthritis [
1,
9,
10]. In general, the net effect of corticosteroids on vascular function in chronic inflammatory disease is obscure [
38]. As reported by Evans
et al., higher cumulative glucocorticoid dose does confer a higher risk of acute coronary syndromes in RA [
7]. Even more recently, Giles
et al. reported prospective data suggesting that subclinical carotid atherosclerosis progression in RA is potentially modified detrimentally by cumulative prednisone exposure [
34]. Our results also suggest that low doses of corticosteroids influence significant new plaque formation in RA. It should be noted, however, that these are not randomised study-derived data and may thus represent a low level of proof. Moreover, a detrimental effect of corticosteroids is a confounding factor by indication, because it could reflect higher disease activity and may be associated with RA severity [
39].
Although biologic therapies appear to decrease the risk of CVD events in RA patients [
40], inconsistent data have been reported in several papers on the effects of biologics, primarily TNF blockers, on carotid atherosclerosis, depending on the treatment duration and clinical efficacy of treatment [
41,
42]. Whether the positive effect of these drugs is related to their anti-inflammatory actions, leading to a decreased risk of unstable plaques, rather than to an effect on the size of carotid IMT deserves further studies. In the initial multivariate analysis, we found a marginally significant beneficial effect of biologic DMARDs on the atherosclerotic process, but the significance was lost in the final model (Table
2). Other prospective observations suggest that subclinical carotid atherosclerosis progression in RA is potentially modified favourably by TNF inhibitors [
34]. Taken together, these data imply that particular treatment modalities that induce disease remission, such as corticosteroids and TNF inhibitors, may be very important or even more important than disease remission
per se as far as the subclinical atherosclerosis in RA is concerned.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
EZ acquired ultrasound data, reevaluated the baseline images offline, analysed the data and drafted the manuscript. AP analysed the data and helped with statistical analysis and drafting the manuscript. KS reevaluated the baseline images offline and analysed the data. KF and CGK were responsible for patient follow-up. KK acquired ultrasound data and reevaluated the baseline images offline. CMP acquired ultrasound data. MM helped in drafting the manuscript. PN was responsible for the statistical analysis. GDK helped with data analysis and interpretation and the drafting of the manuscript. PPS conceived and supervised the study and drafted the manuscript. All authors read and approved the final version of the manuscript.