Introduction

Patients who have rheumatoid arthritis (RA) show a high prevalence of cardiovascular (CV) events, especially at younger ages [1].

Atherosclerosis is a multifactorial and multifocal pathological process whose progression leads to adverse CV clinical events. It is widely present in patients with RA due to chronic systemic inflammation, which seems to induce a more rapid progression in comparison to other clinical situations. In RA, higher instability of the plaques also occurs due to infiltration of macrophages and type 1T helper (Th1) cells, collagen degradation, and neovascularization.

Atherosclerosis can be easily and accurately detected and monitored by means of carotid ultrasound, which allows the detection of carotid plaques (expression of overt disease) and measurement of the intima-medial thickness (cIMT, considered a subclinical atherosclerosis index), both of which are predictors of future adverse CV events in several types of patients at higher risk, such as those with arterial hypertension and/or type 2 diabetes mellitus [2, 3]. Recent studies in RA subjects have shown a high accuracy of ultrasound-derived indices of atherosclerosis for assessing vascular damage and predicting the risk of CV events in these patients [4].

In a recent study [5], we showed that arthritic patients, compared with controls, do not have large differences in US markers of atherosclerosis when they are well matched for age and number of traditional risk factors (TRF). In our cross-sectional survey, only age was consistently associated with the same atherosclerotic indices over TRF, activity, and duration of the arthritic disease. Conversely, in a recent study by Wah-Suarez et al., RA patients had a higher prevalence of bilateral and unilateral carotid plaques and increased carotid intima-media thickness (cIMT) than matched healthy controls had [6]. The discrepancy between our work and Wah-Suarez’s investigation testifies to the high degree of uncertainty existing in this field.

Thus, the aim of the present study was to assess factors associated with the progression of subclinical atherosclerosis in RA patients with TRF who are under specific treatment. A better knowledge of these factors might improve the CV risk stratification of arthritic patients as well as tailor preventive strategies for the control of the atherosclerotic process in this setting.

Methods

One hundred thirty-seven consecutive patients aged between 18 and 75 years who had been previously diagnosed with RA and stable sinus rhythm, and who underwent a follow-up visit at the rheumatology outpatient clinic of the “Azienda Ospedaliera Universitaria Integrata” of Verona were enrolled in this prospective study from January to December 2016. Patients previously diagnosed with myocardial infarction [7, 8], ischemic heart disease [9], transient ischemic attack [10], or stroke [10] according to clinical practice guidelines were excluded from the study. Other exclusion criteria were pregnancy, active malignancy, cirrhosis, and chronic kidney disease (defined as a glomerular filtration rate lower than 60 ml/min/1.73 m2 obtained through the Crockroft–Gault formula).

The study was approved by the Ethical Committee of Verona (CESC n°1707). All participants signed a written consent form.

Clinical and biochemical evaluations of patients were performed. Blood samples were drawn from each participant within 1 month of the visit day. The Disease Activity Score-28 with C-Reactive Protein (DAS28[CRP]) was used to assess RA activity. DAS28 [CRP] refers to the ‘disease activity score’, and the number 28 refers to the 28 joints that are examined during this assessment. This formula yields the overall disease activity score. A DAS28[CRP] > 5.1 implies highly active disease, while a DAS28[CRP] < 3.2 implies low disease activity; a DAS28[CRP] lower than 2.6 corresponds to disease remission [11]. In each participant, the clinical and medical history was recorded; in particular, pharmacological anti-rheumatic treatment was assessed (nonsteroidal anti-inflammatory drugs, NSAIDs; disease-modifying anti-rheumatic drugs, DMARDs; corticosteroids, defined as a high dose >10 mg of prednisolone or equivalent; biological drugs). Rheumatoid factor (RF) and anti-cyclic citrullinated peptide (anti-CCP) status was evaluated (defined as positivity of the test at baseline; the test was performed in the Laboratory of the “Azienda Ospedaliera Universitaria Integrata” of Verona). Finally, the 10-year risk of fatal CV disease was calculated for each participant according to the SCORE system considering gender, age, systolic blood pressure, total cholesterol, and smoke status in a cohort of low-risk populations (Italian ethnicity) [12].

Of the 137 arthritis patients enrolled at baseline, 105 patients with RA agreed to undergo a second measurement of the indices for subclinical and established atherosclerosis after 12 months. Thirty-two patients did not attend the follow-up and were excluded from the analysis. None of the patients died or had CV events or hospitalizations.

In particular, cIMT, cCD (common Carotid Distensibility) and the presence of carotid plaques were considered target indices of atherosclerosis activity and progression.

To further investigate among factors contributing to atherosclerosis progression, we divided our study population into subgroups according to age (above or below 60 years) and DAS28[CRP] score (above or below 2.6).

The a priori statistical power, set for our sample size, for a two-sided error probability <0.05 and for the mean difference/standard deviation of continuous variables considered between baseline and follow-up, was calculated as 0.861 for cIMT and 0.931 for cCD.

Ultrasound measurements

All measurements, performed by ultrasound (LOGIQ P5 pro, GE, Indianapolis, USA), were performed by the same expert operator at baseline and after 12 months processed with the aid of dedicated hardware (Multimedia Video Engine II (MVE2) DSP Lab., Pisa CNR, Italy). In particular, carotid intima-media thickness (cIMT) measurements (right and left IMT mean) were performed during evaluation on the far wall of the distal common carotid artery 1 cm from the bifurcation in a segment of carotid of ~1 cm using dedicated hardware [13, 14].

The right and left common carotid artery distensibility coefficients (cCD) were determined in conjunction with brachial BP measurements and were calculated using the following formula: DC = (ΔA/A)/PPa, where ΔA is the stroke change (i.e., distension) in carotid artery cross-sectional area, which is normalized to the total diastolic carotid artery cross-sectional luminal area (A), and PPa is the pressure differential, assuming that the cross section of the artery is circular [15]. cCD is presented as the average between the right and left common carotid arteries.

Plaques were defined as focal structures encroaching into the arterial lumen of at least 0.5 mm or 50% of the surrounding IMT value or demonstrating a thickness >1.5 mm as measured from the intima-lumen interface to the media-adventitia interface [16].

Statistics

All data were analyzed using the SPSS statistical computer package (version 21.0; IBM Corporation, Armonk, NY, USA). Continuous variables are presented as the mean ± standard deviation (SD) unless otherwise specified and as the median and 25°/75° percentiles if non-parametric. Comparison between continuous variables measured longitudinally in each participant was performed using Student’s T test for paired data and/or the Wilcoxon rank test, depending on the normal distribution of the data. To compare dichotomous variables, the Fisher’s exact test was used. Multiple linear regression was used in the multivariate model, which included the variation of cIMT (ΔcIMT) and the variation of cCD (ΔcCD) as dependent variables while sex, age, body mass index (BMI), mean arterial pressure (MAP), heart rate, diabetes, antihypertensive drugs, smoking status, dyslipidemia, high-dose corticosteroid, traditional Disease Modifying Anti-Rheumatic Drugs (DMARDSs) and/or biological drug use, RF and anti-CCP status, RA duration, and DAS28[CRP] score as covariates (data presented in Table 3: beta coefficient and standard error of the mean). Multiple logistic regression was used in the multivariate model and included the development of plaques at follow-up as the dependent variable, while sex, age, BMI, MAP, heart rate, diabetes, antihypertensive drugs, smoking status, dyslipidemia, high-dose corticosteroid, traditional DMARDSs and/or biological drug use, RF and anti-CCP status, RA duration, and DAS28[CRP] score were used as covariates (the data are presented in Table 3: beta coefficient, 95% confidence interval). All tests were two-sided, and p values < 0.05 were considered statistically significant.

Results

Table 1 shows the main clinical metabolic and inflammatory characteristics at baseline and after 1 year of RA patients who participated in the study. Regarding smoking status, none quit smoking during the study period. We did not observe significant changes between baseline and follow-up in these parameters, including the indices of disease activity, apart from a slight increase in the prevalence of hypertension. In particular, the assumption of different types of medications did not change during follow-up. The median 10-year CV risk was calculated as 2.6% for female nonsmokers, 4.7% for female smokers, 6.5% for male nonsmokers, and 12.9% for male smokers.

Table 1 Clinical, metabolic, and inflammatory variables characterizing the cardiovascular risk profile in the whole sample at baseline and after 1 year
Full size table

Predictors of progression of subclinical atherosclerosis

The cIMT and the percentage of detectable plaques were significantly increased, while cCD was significantly decreased after 12 months of follow-up in the whole sample (Table 2). No relationship was found between 10-year CV risk and atherosclerosis progression.

Table 2 Indices of atherosclerosis in the whole sample of RA patients
Full size table

In the multivariable analysis, ΔcIMT (difference between follow-up and baseline) was found to be independently associated only with the presence of diabetes mellitus at baseline (Table 3), whereas ΔcCD (difference between follow-up and baseline) was seen to be associated with the baseline MAP, diabetes, and the assumption of corticosteroids (Table 3). The increase in the number of carotid plaques (Δ plaques) after 1 year of follow-up was related to age, MAP, and male sex (Table 3) in the logistic regression analysis.

Table 3 Determinants of cIMT, cCD, and plaques modifications after 1 year of follow-up in the whole sample of RA patients in multivariable analysis
Full size table

Subgroup analysis (age ≥60 years and DAS28 score ≥2.6)

When dividing the RA patients into two groups according to the median age of the study population (<60 years old vs. more than 60 years old), cIMT and the presence of plaques were significantly increased and cCD significantly decreased in the oldest group after 1 year, whereas no significant differences were found in younger AR patients, as shown in Table 4.

Table 4 Subclinical atherosclerosis parameters at baseline and after 1 years of follow-up in patients stratified for median age of study population (60 years)
Full size table

If patients with RA were stratified according to the values of DAS28 [CRP], only patients with active disease (DAS28 [CRP] ≥2.6) showed a significant increase in cIMT and a decrease in cCD during the study period (Table 5), whereas patients with disease remission (DAS28 [CRP] <2.6) did not show any difference. Regarding the role of CV risk factors in patients with active disease, only age was strictly linked to the detection of new plaques at follow-up (beta coefficient 1.160, 95% CI [1.079–1.247], p value < 0.001) but not with ΔcIMT and ΔcCD.

Table 5 Markers of subclinical and overt atherosclerosis in RA patients stratified for activity index (DAS28[PCR] >2.6)
Full size table

Discussion

It is well established by numerous case-control and longitudinal studies [17,18,19] that a process of accelerated atherosclerosis exists in patients with RA. In particular, two longitudinal studies clearly demonstrated this behavior: Nagata-Sakurai et al. [20] reported an annual rate of cIMT progression of 0.027 mm/year, while González-Juanatey et al. [21] reached a value of 0.036 mm/year. Our study confirms these results; in particular, we showed a worsening of all indices of subclinical atherosclerosis (cIMT, cCD, and presence of plaques) at the 1-year follow-up.

Analyzing the factors associated with the progression of atherosclerosis in our cohort, older age was found to be the strongest independent condition, both in the entire sample and when the analysis was restricted by subgroup according to age, suggesting, as in our previous study [5], that age is probably the major determinant of atherosclerosis in arthritic patients. However, all CV risk factors, particularly those included in metabolic syndrome, are involved, as demonstrated in a recent study [22].

It should be considered that our sample of RA patients had a high prevalence of hypertension, and most patients were taking antihypertensive drugs, keeping their BP under control. However, BP was shown to have a considerable impact on atherosclerosis development and progression, even independently from other risk factors, such as dyslipidemia, overweight, and diabetes.

Interestingly, the assumption of corticosteroids was independently associated with the variation of cCD, whereas in previous studies involving patients with RA, it was associated with an increase in cIMT [23, 24]. In addition, in a study by Avina-Zubieta et al. [25], steroids showed a dose-dependent effect on CV mortality, and the association persisted even after correction for traditional CV risk factors and markers of inflammation. This is a well-known unfavorable condition characterizing patients with chronic rheumatic disorders, which is typically related to a lack of physical activity, which promotes obesity, systemic inflammation, and use of drugs, such as steroids [26,27,28,29,30,31].

When RA patients were stratified according to the degree of disease activity, we observed that indices of subclinical atherosclerosis significantly worsened exclusively in the group of patients with a high arthritic activity (DAS28 [CRP] ≥2.6). Carotid distensibility and carotid IMT, in fact, were significantly lower in patients with active disease than in those with remission of the inflammatory state. This result is in line with evidence provided by other investigations showing an association between the worsening of subclinical atherosclerosis and DAS28 [19, 23], even if the clinical characteristics of the populations of these studies were quite different from ours. Moreover, atherosclerosis progression in RA patients with disease control was shown to be comparable to non-AR and nondiabetic subjects with similar basal CV risk profiles [32], underlining the role of the chronic inflammatory background in this context. The inflammatory mechanisms promoting pro-atherogenic activation and endothelial dysfunction are individuated in C reactive protein (CRP) [33], tumor necrosis factor (TNF) [34], fibrinogen cytokines and interleukins (IL-s), such as IL-6, IL-18, and IL-33, which are abnormally high in RA. All these mediators are closely associated with alterations in lipid levels, insulin resistance, and oxidative stress [35, 36], which are clinically expressed as hypertension and diabetes, which are the traditional CV risk factors we found to be associated with the progression of subclinical atherosclerosis in our RA patients. In this context, the use of DMARDs and biologic drugs blocks the pathogenetic mechanisms of vascular aging and damage that characterize atherosclerosis [37]. Furthermore, recent evidence links CV disease to gut microbiota, mainly due to chronic inflammation related to abnormal crosstalk between vascular, immune, and microbiological systems [38]; in this context, a close relationship between RA-related chronic inflammation, enteric dysfunction, and CV disease cannot be excluded.

Traditional CV risk factors, and in particular hypertension, have been shown to be predictive for a worse vascular status (higher cIMT and lower cCD). In this context, both the inflammatory process and the classical agents of CV damage should be treated to prevent rapid atherosclerotic degeneration.

Conclusion

Our study confirms that TRF, especially age, are the major risk factors for carotid atherosclerosis progression in RA patients and demonstrates that arthritic disease activity also contributes to the worsening of subclinical atherosclerosis. Furthermore, the use of steroid drugs, to some extent, seems to act independently in terms of increasing this risk, at least for CD.

Thus, our findings contribute to underlining the importance of a careful and regular CV risk assessment of arthritic patients and prompt us to consider all necessary measures (i.e., lifestyle changes, drug therapies, and hypertension control) to correct CV risk factors, also according to the recommendations of the EULAR 2016 [39]. Our data also suggest that the remission of the inflammatory activity of arthritis may play a role in slowing the progression of atherosclerotic disease.

Strength and limitations

This study evaluated a real-life population of rheumatic patients with the aim of assessing the contribution of chronic inflammation together with CV risk factors in atherosclerosis progression. The strength consists of the selection of a group of relatively homogeneous patients, excluding those with previous CV events, and the accurate measurement of vascular properties.

The study has some limitations, mainly, the absence of a control group to compare vascular parameters. Pack years of smoking were not available; however, none of the smokers at enrollment stopped during the study period.

In our sample, almost all RA participants had remitted or low-grade disease activity; in this setting, a specific analysis of patients with highly active disease could not be performed. Furthermore, measured levels of blood pro-inflammatory cytokines (TNF-α, IL-6) might have been useful for correlation with atherosclerosis progression.

Regarding technical study limitations, although vascular ultrasound was performed by a single expert operator, data on intraobserver variability are not available because of the lack of measurement repetition at different study points.