Association of non-invasive hemodynamics with arterial stiffness in rheumatoid arthritis

Abstract

Objectives. Arterial stiffness has emerged as a surrogate marker of cardiovascular disease. We investigated the role of myocardial performance and hemodynamic parameters in arterial stiffness in patients with rheumatoid arthritis (RA), which is accompanied by excess cardiovascular risk. Design. Arterial stiffness was evaluated with pulse wave velocity (PWV) in RA patients and controls. Cardiac and hemodynamic characterization was based on impedance cardiography. Cardiovascular risk factors, inflammatory markers and disease-related parameters were assessed. Results. PWV (8.2 ± 2.1 vs 7.4 ± 1.4 m/s, p = .016) was higher among RA patients (n = 104) compared to controls (n = 52). In the RA group, PWV correlated with markers of cardiac contractibility (acceleration and velocity index), myocardial blood flow (cardiac output and stroke volume), preload (thoracic fluid content) and afterload (systemic vascular resistance) (p < .05 for all). PWV tended to increase with decreasing oxygen delivery to the myocardium (r = 0.055), as well as with shortening of the ejection duration of the left ventricle (p = .058). However, these associations no longer remained significant after adjustment for classical cardiovascular risk factors, inflammation and corticosteroid use, which were independently associated with PWV. Conclusions. Among patients with RA, arterial stiffness appears as the composite of cardiovascular risk factors and inflammation, while corticosteroid use emerges as an additional adverse factor.

Keywords:

• Arterial stiffness
• myocardial function
• hemodynamic parameters
• impedance cardiography
• rheumatoid arthritis

Introduction

Cardiovascular disease is a leading cause of morbidity and mortality in patients with rheumatoid arthritis (RA), which cannot be solely attributed to the increased prevalence of classical cardiovascular risk factors [1]. Since patients with RA are less likely to exhibit typical signs and symptoms of cardiac involvement [2], surrogate markers of cardiovascular disease are urgently needed in this specific group of patients.

Cardiovascular risk can be further stratified by assessment of aortic pulse wave velocity (PWV) [3], which has been accepted as the “gold standard” measure of arterial stiffness. PWV emerges as a useful cardiovascular clinical marker and an early indicator of cardiovascular morbidity and mortality in both hypertension and normotension [4]. Baseline values of PWV predict future cardiovascular disease events in individuals free of overt cardiovascular disease [5], while in hypertensive individuals, this effect seems to extend even beyond Framingham score [6]. According to several reports, patients with RA exhibit higher values of central [7] and peripheral [8] arterial stiffness compared to non-RA individuals. However, PWV may be influenced by a wide range of factors, including anthropometric characteristics, conventional cardiovascular risk factors, and central hemodynamics [9]. At the same time, inflammation-driven structural and functional alterations of the arterial wall are considered to contribute vigorously to PWV elevation in patients with RA [7,10].

Central hemodynamics can be obtained non-invasively in resting conditions by use of impedance cardiography, which has emerged as a safe and reproducible alternative to invasive hemodynamic measurement [11]. This technique has been applied to several high-cardiovascular risk populations such as patients with hypertension and heart failure, in whom it may improve prognostic accuracy [12,13]. By contrast and rather surprisingly, hemodynamic profile of RA patients has been hardly evaluated with impedance cardiography. We have recently shown that central hemodynamics, obtained with impedance cardiography, are associated with coronary microvascular dysfunction in RA [14], yet it remains unknown whether such parameters correlate with macrovascular dysfunction in this specific group of patients. Therefore, the purpose of the present study was to further characterize the relationship of PWV with central hemodynamics in RA individuals, taking into account classical cardiovascular risk factors and systemic inflammation.

Methods

Consecutive patients who met the 1987 revised criteria of the American College of Rheumatology for the presence of RA [15] were recruited from the Rheumatology Outpatient Unit. The control group consisted of volunteers free from any known health problems, who did not receive any medication. Controls were recruited from the local community and from individuals attending their regular checkup appointments in the Internal Medicine Outpatient Unit. Written informed consent was obtained before inclusion in the study. The study protocol was approved from the institutional ethics committee and the study was conducted in accordance with the Helsinki declaration. Participants were instructed to abstain from smoking, coffee, tea or alcohol consumption, and intense physical activity before the procedures.

Assessment of arterial stiffness

Vascular measurements were performed in patients and controls, to detect differences between groups. After a 15-minute resting period in the supine position, the carotid-femoral PWV was non-invasively assessed with applanation tonometry using the Sphygmocor device (AtCor Medical, Sydney, Australia). Briefly, surface distance between the 2 recording sites was measured (sternal notch to carotid site and sternal notch to femoral site) and sequential waveforms at the right common carotid and right femoral site were recorded. Wave transit time was calculated using a simultaneously recorded electrocardiogram as a reference. PWV was automatically calculated as the distance between the carotid and the femoral sampling site divided by time (PWV = Δd/Δt), with higher levels indicating greater arterial stiffness.

Traditional cardiovascular risk factors

Demographic and anthropometric characteristics, medical history and medication use were recorded and body mass index (BMI) was calculated in kg/m². Systolic/diastolic blood pressure was measured in the sitting position in the arm with the higher blood pressure using a validated oscillometric device (Microlife Exact BP, Microlife AG, Widnau, Switzerland), and was determined as the mean of the second and third value of three consecutive measurements. Hypertension was defined as office systolic and/or diastolic blood pressure ≥140/90 mm Hg, and/or current antihypertensive medication use [16]. Previous cardiovascular events were defined as history of stroke or transient ischemic attack, angina, and myocardial infarction based on self-report and ascertained by a medical record and medication review. After completion of physical examination and vascular and other non-invasive measurements analyzed below, fasting total cholesterol, triglycerides, LDL- and HDL-cholesterol were estimated in blood samples.

Hemodynamic characterization

Markers of cardiac function and hemodynamic profile were non-invasively assessed with impedance cardiography (Cardioscreen 1000 impedance cardiography system, Medis, Germany). Impedance cardiography offers noninvasive, continuous, beat-by-beat measurements of central hemodynamics. As low-amplitude alternating electrical current is applied via external sensors in the thorax, changes of thoracic impedance during the cardiac cycle are most dependent on blood volume changes of the thoracic aorta. Following digital process of baseline thoracic impedance (Zo), pulsatile impedance/time changes (dZ/dt), and electrocardiogram recordings, impedance cardiography is able to calculate indices of myocardial performance and hemodynamic parameters. The reproducibility of this technique has been validated in stable outpatients, and accuracy has been established in comparison with invasive methods in patients with various cardiovascular disorders [17].

A number of cardiac indices were assessed with impedance cardiography. Cardiac output and stroke volume reflect the myocardial blood flow. Pre-ejection period and left ventricular ejection time (LVET) can reflect myocardial contraction (prolonged pre-ejection period and LVET shortening correspond to compromised cardiac function).

Myocardial contractibility was evaluated with acceleration index and velocity index. Acceleration index is the maximum rate of change of blood velocity and representative of blood acceleration in the aorta. Velocity index is the maximum rate of impedance change, and is representative of blood velocity in the aorta.

Assessment of central hemodynamics additionally included thoracic fluid content, (extravascular, intravascular and chest water content), an index of preload, and systemic vascular resistance, an index of afterload. Left cardiac work, which denotes increased myocardial oxygen consumption at increased rates, and delivered oxygen were additionally calculated. These indices were normalized for body size by indexing to each patient’s body surface area to obtain: cardiac index; stroke index; thoracic fluid content index (TFCI); systemic vascular resistance index (SVRI); left cardiac work index (LCWI), and delivered oxygen index.

Disease-related parameters

Erythrocyte sedimentation rate (ESR) and C-reactive protein (CRP) were estimated from patients’ blood samples. Disease activity was assessed during the visit by the treating rheumatologist, using the Disease Activity Score with 28 joints (DAS28), which was calculated based on swollen joint count (28), tender joint count (28), visual analogue scale indicating pain/discomfort due to RA (0–100 scale), and ESR [18]. DAS28-ESR thresholds of ≤3.2, 3.2–5.1, and >5.1 are indicative of low, moderate and high disease activity, respectively.

Patients completed the Health Assessment Questionnaire Disability Index (HAQ-DI), which is one a widely implemented and well-validated measure of physical function in patients with rheumatic disorders. Obtained scores range from 0 (no disability) to 3 (completely disabled) [19].

The HAQ visual analogue scale for pain was used to assess arthritis-related pain on a horizontal linear scale, which is labeled from zero (no pain) at the left anchor point to 100 (severe pain) at the right anchor point.

Statistical analysis

Data analysis was performed using SPSS (Statistical Package for Social Sciences, SPSS Inc., Chicago, IL, USA) software, version 22. Results were expressed as frequencies for qualitative variables, and as mean ± standard deviation (m ± SD) or median (interquartile range) for continuous variables. Comparison of frequencies was performed by Pearson chi-square test. Student’s t-test or Mann–Whitney test was used to estimate differences between mean values. Correlations between PWV and the main continuous variables were assessed using the parametric Pearson or the nonparamentric Spearman’s Rho correlation coefficient. Multiple linear regression analysis with the “Enter” method was used to identify statistically significant associations with PWV after adjustment for other factors. Due to multicollinearity, hemodynamic parameters were not entered in a single model altogether. Where needed to transform a non-to-normal distribution, we used the logarithmic mean of the parameter. A probability value of p ≤ .05 was considered statistically significant.

Results

A total of 156 individuals, 104 RA patients and 52 controls, were studied. Baseline characteristics of the study population, including information on disease characteristics and medical treatment, cardiovascular risk factors and hemodynamic parameters, are presented in Table 1. None of the patients presented symptoms or signs of heart failure upon clinical examination, nor was heart failure identified in their medical records. The vast majority of patients (90%) presented low- to moderate disease activity. Patients did not significantly differ with controls in terms of age, sex, BMI, and blood pressure. Higher levels of PWV were observed among patients with RA, compared to the control group (8.2 ± 2.1 vs 7.4 ± 1.4 m/s, p = .016).

Table 1. Baseline characteristics of study population.

	Rheumatoid Arthritis Group (n = 104)	Control Group (n = 52)
	Mean ± SD/ Median (IQ range)	Mean ± SD/ Median (IQ range)	p value
Cardiovascular risk factors
Age (years)	60.0 ± 11.4	57.2 ± 8.2	.068
Male sex, n (%)	22 (21.2)	11 (21.2)	.396
Systolic blood pressure (mmHg)	125 ± 15	122 ± 10	.171
Diastolic blood pressure (mmHg)	75 ± 8	77 ± 8	.140
Heart rate (beats/min)	70 ± 11	67 ± 8	.169
BMI (kg/m^2)	26.5 ± 4.8	26.9 ± 4.1	.433
Hypertension (%)	49.0	0	<.001
Diabetes mellitus (%)	6.7	0	.053
Cardiovascular events (%)	7.7	0	.039
Smoking (%)	26.7	17.4	.218
Total cholesterol (mg/dl)	206 ± 40	218 ± 33	.017
Triglycerides (mg/dl)	111 (82–142)	94 (72–120)	.055
HDL-C (mg/dl)	54 (45–63)	51 (45–60)	.641
LDL-C (mg/dl)	127 ± 36	144 ± 28	.001
Disease characteristics
Disease duration (years)	9 (5–15)	NA
ESR (mm/hr)	19 (10–29)	NA
CRP (mg/l)	3.19 (3.19–4.20)	NA
Rheumatoid factor (U/ml)	14 (11–83)	NA
DAS28-ESR	3.5 ± 1.2	NA
HAQ-DI	0.81 (0.06–1.50)	NA
VAS-Pain	32 (9–64)	NA
Medication
sDMARDs (%)	75.8	ΝΑ
bDMARDs (%)	43.0	ΝΑ
NSAIDs (%)	3.8	ΝΑ
Glucocorticoids (%)	27.9	ΝΑ
Statins (%)	22.8	0
Anti-hypertensive drugs (%)	44.1	0
Hemodynamic parameters
Systemic vascular resistance index (dyn·s·cm^−5/m^2)	2001 (1581–2319)	1918 (1626–2156)	.543
Stroke index (ml/m^2)	49 ± 13	53 ± 11	.090
Cardiac index (l/min/m^2)	3.4 ± 0.8	3.6 ± 0.7	.373
Thoracic fluid content index (l/kOhm/m^2)	18.1 ± 3.3	16.4 ± 2.5	.004
Acceleration index (1/100/s^2)	88 ± 37	86 ± 35	.864
Velocity index (1/1000/s)	55 ± 16	56 ± 18	.572
Left cardiac work index (kg·m/m^2)	3.9 ± 1.0	4.0 ± 1.0	.346
Delivered oxygen index (ml/min^−1/m^−2)	574 ± 136	603 ± 120	.240
Pre-ejection period (ms)	95 ± 19	98 ± 16	.326
Left ventricular ejection time (ms)	322 ± 54	336 ± 44	.145

BMI: Body Mass Index; HDL-C: High-Density Lipoprotein Cholesterol; LDL-C: Low-Density Lipoprotein Cholesterol; ESR: Erythrocyte Sedimentation Rate; CRP: C-reactive protein; DAS28: Disease Activity Score of 28 joints; HAQ-DI: Health Assessment Questionnaire-Disability Index; VAS: Visual Analogue Scale; sDMARDs: synthetic Disease Modifying Anti-Rheumatic Drugs; bDMARDs: biologic Disease Modifying Anti-Rheumatic Drugs; NSAIDs: Non-steroid Anti-inflammatory Drugs; NA: Not Applicable.

Classical cardiovascular risk factors

In the RA group, PWV was associated with the majority of classical cardiovascular risk factors, including age (r = 0.583, p < .001), systolic (r = 0.671, p < .001) and diastolic (r = 0.400, p < .001) blood pressure, BMI (r = 0.245, p = .014), triglycerides (r = 0.227, p = .034) and HDL cholesterol (r=–0.213, p = .045). PWV was higher in hypertensive compared to normotensive patients (9.2 ± 2.1 vs 7.1 ± 1.6 m/s, p <. 001), and in patients with diabetes mellitus compared to non-diabetics (10.4 ± 2.9 vs 8.0 ± 2.0 m/s, p = .034). Male patients exhibited higher values of PWV compared to females (9.4 ± 2.0 vs 7.9 ± 2.1 m/s, p = .002).

Disease-related characteristics

PWV significantly correlated with both inflammatory markers, ESR (r = 0.214, p = .035) and CRP (r = 0.271, p = .009), as well as with disease activity, DAS28 (r = 0.233, p = .025). Increased PWV was observed with increasing levels of physical disability and pain, evaluated with HAQ-DI (r = 0.220, p = .049) and pain (r = 0.304, p = .010) scores, respectively. Disease duration and RF were not significantly associated with PWV.

Medication use

Patients receiving glucocorticoids exhibited elevated PWV, compared to those who were not prescribed such medications (9.1 ± 2.6 vs 7.9 ± 1.9 m/s, p = .032). Treatment with antihypertensives was associated with increased values of PWV (9.3 ± 2.2 vs 7.4 ± 1.7 m/s, p < .001) compared to patients who did not receive such agents. PWV did not differ in patients treated with statins, biologics, or other disease modifying antirheumatic drugs, compared to patients who did not receive the respective drug category.

Central hemodynamics

Table 2 presents a summary of the univariate correlations between PWV and central hemodynamics in patients with RA. PWV positively correlated with SVRI (p < .001) and inversely with stroke index (p < .001), cardiac index (p = .005), TFCI (p = .041), and contractibility indices (acceleration index and velocity index, p < .001 for both). A trend towards statistical significance was observed for the associations with delivered oxygen index (p = .055) and LVET (p = .058).

Table 2. Results of univariate analysis of PWV with cardiac and hemodynamic parameters.

Variable	Univariate analysis for PWV
	r	p
Systemic vascular resistance index (dyn·s·cm^−5/m^2)	0.451	<.001
Stroke index (ml/m^2)	–0.351	<.001
Cardiac index (l/min/m^2)	–0.282	.005
Thoracic fluid content index (l/kOhm/m^2)	–0.205	.041
Acceleration index (1/100/s^2)	–0.465	<.001
Velocity index (1/1000/s)	–0.473	<.001
Left cardiac work index (kg·m/m^2)	–0.060	.551
Delivered oxygen index (ml/min^−1/m^−2)	–0.195	.055
Pre-ejection period (ms)	–0.046	.652
Left ventricular ejection time (ms)	–0.190	.058

PWV: Pulse Wave Velocity.

Multivariate analysis for arterial stiffness

Table 3 presents the results of multiple regression analysis of various clinical variables, to evaluate their independent associations with PWV. All models consistently indicated that increasing age, male sex, presence of hypertension and diabetes were independently associated with increased arterial stiffness in RA. In Models 1, 2 and 3, inflammation emerged as an additional determinant, and a trend towards statistical significance was observed in Model 3. Glucocorticoid use was independently associated with PWV in Models 2 and 3, and a trend was observed in the rest models (1 and 4). On the contrary, the impact of cardiac indices and hemodynamic parameters no longer remained significant after adjustment for the above factors.

Table 3. Multiple regression models for PWV, accounting for cardiovascular risk factors, inflammation and hemodynamic parameters.

Variable	Model 1		Model 2		Model 3		Model 4
	β	p	β	p	β	p	β	p
Dependent variable: PWV
Age (years)	0.260	.012	0.286	.006	0.290	.005	0.278	.007
Sex^a	–0.220	.016	–0.244	.007	–0.245	.007	–0.229	.013
BMI (kg/m^2)	–0.011	.912	–0.021	.841	–0.033	.765	–0.040	.692
Hypertension	–0.280	.005	–0.303	.003	–0.305	.003	–0.297	.003
Diabetes	–0.210	.017	–0.205	.022	–0.207	.020	–0.204	.021
ESR (mm/hr)	0.174	.047	0.179	.044	0.179	.044	0.171	.055
Triglycerides (mg/dl)	–0.091	.320	–0.073	.429	–0.073	.429	–0.066	.471
Glucocorticoids	–0.157	.065	–0.177	.040	–0.176	.044	–0.167	.052
SVRI (dyn·s·cm^−5/m^2)	0.133	.154
Cardiac index (l/min/m^2)			–0.014	.882
TFCI (l/kOhm/m^2)					–0.016	.869
Acceleration index (1/100/s^2)							–0.073	.416

PWV: Pulse Wave Velocity; BMI: Body Mass Index; ESR: Erythrocyte Sedimentation Rate; SVRI: Systemic Vascular Resistance Index; TFCI: Thoracic Fluid Content Index.

Model 1: R²=0.552, p < .001; Models 2&3: R²=0.539, p < .001; Model 4: R²=0.543, p < .001.

a 1 = male, 2 = female.

Discussion

To our knowledge, this is the first study to examine the association between central hemodynamics, using the non-invasive method of impedance cardiography, with arterial stiffness among patients with RA. In the univariate analysis (Table 2), PWV correlated with markers of preload (TFCI), afterload (vascular resistance), myocardial blood flow (cardiac index, stroke index) and myocardial contractibility (acceleration index and velocity index). Nevertheless, these associations were no longer significant after adjustment for other factors (Table 3). On the contrary, both inflammation and classical cardiovascular risk factors (age, male sex, hypertension, diabetes) were independently associated with PWV among patients with RA. Pathophysiologically, these factors are known to induce structural and functional alterations on the arterial wall that result in large artery stiffening [20]. In addition, treatment with glucocorticoids was independently associated with PWV. Epidemiological studies have shown a dose-dependent increased risk of total and cardiovascular mortality, especially myocardial infarction, in RA patients treated with corticosteroids [21]. According to experimental data, corticosteroids promote endothelial dysfunction and increase the production of reactive oxygen species within the vessels [22]. Although endothelium only modestly modulates arterial stiffness [20], it is possible that this unfavorable effect may partly explain the observed negative impact on PWV. In any case, the negative effect of corticosteroids on arterial stiffness may contribute to the increased cardiovascular mortality associated with corticosteroids in RA, but this hypothesis warrants further examination.

Our study provides some additional points that merit further attention. Patients with RA exhibit higher PWV compared to non-RA individuals, in line with the findings of several studies summarized in the results of a recent meta-analysis [7]. In addition, in a relatively well-controlled setting of RA patients without typical signs or symptoms of heart failure, central hemodynamics were comparable between RA and non-RA individuals, with the exception of TFCI, which was significantly elevated among patients. It has been previously proposed that TFCI might best reflect diastolic dysfunction, which is the typical pattern of cardiac involvement in patients with RA [23], rather than both systolic and diastolic dysfunction [24]. However, this observation can be considered only as hypothesis-generating and further studies using more robust measures of cardiac function are needed to determine the net effect of RA on central hemodynamics.

Available data regarding the association between central hemodynamics and macrovascular impairment in RA are extremely limited. PWV has been associated with left ventricular dysfunction in 40 patients with RA, evaluated by use of tissue Doppler-derived myocardial performance index [25]. LVET, yet no other indices from impedance cardiography, was identified as a determinant of PWV in young, healthy males [26]. On the other hand, the negative impact of conventional cardiovascular risk factors on arterial stiffness has been observed also in RA patients [7]. Moreover, the interplay between systemic inflammation and autoimmune dysregulation in RA, which emerges as an appropriate substrate for the development and progression of macrovascular dysfunction, has been well-described [10,27]. In the present study, the non-significant effect of central hemodynamics on PWV in the multivariate analysis might be explained by the fact that with the exception of TFCI, all other indices were comparable between patients and healthy controls and therefore, not profoundly impaired (Table 1). In addition, rheological disorders and myocardial dysfunction most probably are the exact derivative of the effect of classical cardiovascular risk factors on the cardiovascular system. In any case, impedance cardiography can be helpful in terms of hemodynamic characterization of these patients.

The clinical relevance of measuring PWV in patients with RA lies in its ability to predict cardiovascular morbidity and mortality also in this specific group of patients [28], in accordance with findings from the general and high-risk populations [4–6]. Our study is limited by the divergent composition of the RA and the non-random selection of the control group. The groups differed in many important aspects: a significant portion presented with hypertension, diabetes, and cardiovascular events, while they were non-significantly older and smoked more; all factors that increase PWV. We included a real-life cohort of patients, and the difficulty of identifying healthy individuals of advanced age needs to be acknowledged. Even though our primary goal was not to compare PWV between patients and controls, it cannot be excluded that a precise matching for all cardiovascular risk factors would result in non-significant differences in PWV, as has been previously reported [29]. However, increased prevalence of traditional cardiovascular risk factors in RA appears as an integral part of the disease related to chronic inflammation. Hence, exhaustive matching of these patients’ cardiovascular risk factors with the control group might eliminate characteristics of RA disease. Considering the difficulty of predicting cardiovascular disease in this patient group, PWV may emerge as a potentially important biomarker for predicting cardiovascular disease in RA, just as is the case with the general population and other high-cardiovascular risk populations.

Conclusions

Among patients with RA, the observed associations of PWV with markers of myocardial blood flow, myocardial contractibility, preload and afterload, non-invasively assessed with impedance cardiography, were no longer significant after adjustment for other factors. In this specific group of patients, arterial stiffness appears to be independently associated with cardiovascular risk factors and systemic inflammation, while treatment with corticosteroids emerges as an additional adverse factor.

Disclosure statement

No potential conflict of interest was reported by the authors.