Introduction

Rheumatoid arthritis (RA) is a chronic systemic inflammatory disease, associated with increased cardiovascular mortality [12]. Even after adjustment for traditional cardiovascular risk factors, there is a higher rate of cardiovascular events in RA patients and this has been attributed in part to chronic inflammation [12]. Inflammation plays a significant role in the pathogenesis of atherosclerosis, and so far, there have been certain shared features between atherosclerosis and RA [12].

Tumor necrosis factor-a (TNF-a) is a pleiotropic, pro-inflammatory cytokine that plays a pivotal role in inflammation, both in RA and atherogenesis [15]. Treatment with TNF antagonists (anti-TNFs) has revolutionarized the treatment of RA by reducing joint and systemic inflammation. A number of studies have shown that anti-TNFs decrease cardiovascular events in RA patients [9], but thus far, it is unclear if this effect is due to its antiinflammatory activity, change in cardiovascular risk factors [6] or through a direct effect in the arterial wall [12].

Arterial stiffness and enhanced wave reflections are markers of cardiovascular disease and independent predictors of cardiovascular risk [21]. Increased arterial stiffness has been reported in RA, [3,10,13,19] although its role in predicting cardiovascular risk in this population has not been studied so far [12]. A number of studies have examined the effect of anti-TNFs on arterial stiffness with conflicting results [7], with limited data on the effect of synthetic DMARDs.

The aim of our study was to investigate the effect of adalimumab (ADA) on arterial stiffness in RA patients, according to its effect on disease activity.

Methods

Thirty six consecutive patients with RA (according to the 1987 American Rheumatism Association criteria) [2] were recruited from the Outpatient Rheumatology Clinic of a referral tertiary care center (2nd Department of Medicine, Athens University School of Medicine, Hippokration General Hospital, Athens, Greece). Patients with cancer, pregnancy, heart failure (NYHA III–IV), active infection, tuberculosis or age <18 years were excluded.

Two patient groups were studied: the first with resistant to DMARDs patients, received ADA (40 mg subcutaneously every other week) either as monotherapy (n = 7) or in combination with DMARDs (n = 11; methotrexate/MTX n = 6, leflunomide n = 4, hydroxychloroquine n = 1); eleven patients (61 %) were receiving also prednisolone (mean dose = 6.5 mg/day per os). The second group received standard DMARDs (methotrexate, n = 18) and prednisolone (n = 17, mean dose = 5 mg/day per os). Both groups were followed for 3 months without changes in these medications. At the end of follow-up, patients were categorized as responders (Disease Activity Score-DAS28 reduction > 1.2) or nonresponders (DAS28 reduction < 1.2). Approval was obtained from the Local Research Ethics Committee, and written informed consent was obtained from each participant. The study protocol conforms to the ethical guidelines of the 1975 Declaration of Helsinki.

Arterial stiffness and wave-reflection measurements

Carotid-femoral pulse wave velocity (PWV) was calculated from measurements of pulse transit time and the distance traveled between two recording sites (PWV equals distance in meters divided by transit time in seconds) with a validated noninvasive device (Complior, Artech Medical) that allows online automatic calculation of PWV. Augmentation index (AIx) of the central (aortic) pressure waveform was measured by a validated, commercially available system (SphygmoCor, AtCor Medical) as previously described [23].

Disease activity and functional impairment

Fasting venous blood was drawn before the recordings of PWV and AIx for measurement of lipids, renal and liver function, glucose, CRP, and erythrocyte sedimentation rate (ESR). Disease activity was measured using the DAS28 while functional impairment was assessed with the Health Assessment Questionnaire (HAQ).

Statistical methods

Data were analyzed with SPSS software (version 16). We performed unpaired two-tailed Student t tests to compare group differences. The paired samples t test was used to investigate the effect of treatment on arterial stiffness, disease activity and, functional impairment. We assessed whether the change in PWV was related to changes in disease activity (DAS28) or other disease-related or cardiovascular parameters. We used the Multiple Linear Regression Analysis with the method of least of squares, as well as the Analysis of Variance (ANOVA). We defined as dependent variable for the Multiple Linear Regression analysis the change of PWV from the 1st until the 12th week, and as independents the corresponding changes of all other parameters [AIx, ESR, CRP, DAS28, HAQ), age, total cholesterol, HDL, LDL, glucose, creatinine, systolic blood pressure (SBP), diastolic blood pressure (DBP), heart rate (HR)]. Factors for the stepwise model were considered either because they were established or putative determinants of stiffness or were based on initial regression analysis with an enter model, whereby factors were included if they achieved a significance value of <0.05. Neither AIx, ESR, CRP, DAS28, HAQ, steroid dose or use, disease duration, HDL, LDL, glucose, SBP, DBP, HR, or body mass index (BMI) entered the final model. A probability of <0.05 was considered significant. Results are expressed as mean ± SD, unless otherwise specified.

Results

Patient characteristics

Thirty-six RA patients were included; 18 were treated with ADA (±DMARDs) and 18 with DMARDs alone (MTX). Their mean age was 60.1 ± 11.9 years and the mean disease duration was 6.2 ± 6.2 years. At baseline, there were no statistically significant differences between the two groups (Table 1).

Changes in arterial stiffness and disease activity indices

Patients treated with ADA demonstrated a statistically significant reduction in PWV (from 8.18 ± 2.03 at baseline to 7.01 ± 1.78) after 12 weeks (p = 0.00006, Fig. 1) while there was no significant change in AIx (from 27.2 ± 12.9 to 26.5 ± 9.6, p = 0.75, Fig. 1). A statistically significant decrease in DAS28 (from 6.65 ± 1.22 to 4.69 ± 1.46, p = 0.00007, Fig. 1) and ESR (from 40 ± 22.5 to 29.3 ± 21.7, p = 0.04) was noted. There was also a decrease in HAQ score (from 1.16 ± 0.66 to 0.69 ± 0.71, p = 0.21) and CRP (from 12.2 ± 15.9 to 8.83 ± 15.03, p = 0.35) but this did not reach statistical significance.

Fig. 1
figure 1

The change in DAS28 (a), pulse wave velocity (PWV, b) and augmentation index (AIx, c) of the group of patients treated with adalimumab for 12 weeks as a whole (all patients, n = 18) or in subgroups according to their response to therapy (NR = nonresponders, n = 6; R = responders, n = 12) is shown. In blue bars and light blue bars, the baseline and 12 week values (mean ± 1 standard deviation), respectively, are shown

Full size image

Twelve (67 %) patients demonstrated clinical response during ADA-based therapy. The baseline characteristics of the two groups of patients are shown in Table 2. A statistically significant reduction in PWV both in responders and non-responders was noted (responders: from 8.28 ± 2.13 to 7.05 ± 1.79, p = 0.001, nonresponders: from 7.99 ± 2.0 to 6.93 ± 1.92, p = 0.03). Responders had a statistically significant reduction in RA disease activity score (DAS28, Fig. 1) compared to nonresponders, while there was no statistically significant change in the AIx in both groups (responders: from 27.23 ± 11.71 to 27.64, p = 0.89, nonresponders: 27.28 ± 16.47 to 24.18 ± 10.1, p = 0.45, Fig. 1).

Table 1 The baseline characteristics of the patients treated with methotrexate or adalimumab-based regimens are shown
Full size table
Table 2 Baseline characteristics of patients treated with adalimumab according to their response to treatment
Full size table

In patients treated with MTX alone, there was no statistically significant change in PWV (from 8.87 ± 1.91 to 8.41 ± 2.17, p = 0.29) or AIx (from 30.06 ± 13.85 to 28.46 ± 12.29, p = 0.59), despite improvement of RA disease activity (DAS28: from 5.78 ± 1.61 to 3.81 ± 1.07, p = 0.00001) and function (HAQ score: from 0.95 ± 0.53 to 0.23 ± 0.34, p = 0.00002). PWV decreased in responders (from 8.58 ± 1.52 to 7.92 ± 1.63, p = 0.97) while it increased in nonresponders to therapy (from 8.80 ± 4.18 to 9.67 ± 3.03, p = 0.19).

Predictors of PWV change

For patients treated with ADA, applying the multiple linear model, we found that none of the independent variables affected our dependent variable—PWV (p values > 0.05). These results were also confirmed by the ANOVA analysis (p = 0.432 for the complete model). Applying the same analysis, for responders (n = 12) as well as for patients treated with MTX, similar results were obtained.

Discussion

In this study, we demonstrated for the first time that treatment with a specific TNF antagonist (ADA) reduces significantly aortic stiffness in RA patients, independently of its antirheumatic effect. In a separate control group of biologic-naïve RA patients, a traditional DMARD (MTX) did not significantly alter aortic stiffness.

There is strong evidence to suggest that both MTX [24] and anti-TNFs [4] reduce significantly cardiovascular events in RA patients [16]. Although there have not been any head to head comparative trials, a recent systematic review and meta-analysis, showed that anti-TNF agents decrease the cardiovascular risk by 15–54 % more compared to DMARDs (mostly MTX) [4]. These data indicate that anti-TNFs may have an additive cardiovascular protective effect compared to DMARDs.

Since long-term studies assessing the effect of antirheumatic agents in cardiovascular outcomes have not been performed, surrogate markers of subclinical vascular disease are used. These include intima-media thickness, flow-mediated endothelial-dependent vasodilatation, PWV, and AIx [12]. Most studies have focused on anti-TNFs with limited information on conventional synthetic DMARDs, showing that anti-TNF treatment (3–12 months) leads to a significant decrease in aortic PWV [1,13,18,25]. Only one study by Komai et al. did not show a change in aortic PWV [11]. In contrast, the effect in AIx was variable with most studies reporting no change, [1,5,13,18,20,25] decrease [8] or even increase [17] (treatment duration: 6 weeks–1 year).

In studies where both aortic PWV and AIx were measured, a decrease in aortic PWV without a change in AIx was observed [1,13,18,25]. Our study showed a pattern of decreased aortic PWV and unchanged AIx during anti-TNF treatment. It is currently unclear why there is such a discordant effect of anti-TNFs on aortic PWV and AIx. AIx measures vascular stiffness and peripheral resistance, while PWV purely reflects vascular stiffness. PWV and wave reflection indexes often change in parallel because PWV affects the timing of the merging of incident and reflected waves. Despite the increase in PWV by inflammation, however, wave reflections are decreased due to peripheral vasodilatation [22]. Thus, the predominant mechanism appears to be dilatation of medium and small peripheral muscular arteries and arterioles [22]. Anti-TNFs, by reducing inflammation, reduce PWV and thus AIx, but they also reduce peripheral vasodilatation, tending to increase AIx.

Only one of the studies where both aortic PWV and AIx were measured included a control group treated with synthetic DMARDs (MTX) alone [18]. Similarly to Tam et al., we did not find any significant change in aortic PWV and AIx during short-term MTX treatment.

Overall, ours and previously published data show a greater decrease in vascular stiffness in patients treated with anti-TNFs. We also demonstrated that this effect occurred irrespectively of its antirheumatic effect. Since there was no statistically significant change in traditional cardiovascular risk factors, other anti-TNF mediated distinct effects on vascular function may be implicated. More studies are needed in order to fully understand the exact mechanisms of this specific anti-TNF effect which does not occur with other biologic agents used in RA with a different mode of action. For example, treatment with a B cell depleting agent (rituximab) had no effect on aortic stiffness despite significant improvement in disease activity [14].

Our study may have important implications. Given the importance of increased aortic stiffness for cardiovascular function and for the development of left ventricular dysfunction and myocardial ischemia, we may hypothesize that the reduction of arterial stiffness with anti-TNF agents may, at least partially, explain the reduction of the increased cardiovascular morbidity and mortality, seen in RA patients. The improvement of arterial stiffness independently of the clinical response to treatment, may suggest a direct protective effect of some anti-TNF agents on arterial wall.

Limitations of our study include the small number of patients, the short period of treatment (3 months), and the absence of a direct control group. Patients treated with ADA had longer disease duration and higher disease activity despite DMARD therapy, whereas DMARD-treated patients had earlier disease and were biologic naïve. Nevertheless, the baseline characteristics did not differ strengthening our findings for a differential effect of DMARDs (MTX) versus anti-TNF agents (ADA) on arterial stiffness.

In conclusion, we report for the first time that arterial stiffness improves with ADA treatment in RA patients but not with treatment with MTX alone and this effect seems to occur independently of changes in cardiovascular risk factors or disease activity indices. These findings indicate that treatment with TNF antagonists may not only improve disease activity, but also vascular “damage” in RA patients and could explain the observed beneficial long-term effects on cardiovascular morbidity and mortality compared to MTX alone. This effect on arterial stiffness by anti-TNF agents, if proven, could influence future treatment decisions in RA patients in daily clinical practice.