Prevention of cardiovascular disease in rheumatoid arthritis

Highlights

• Although the CV morbidity in RA resembles that in diabetes, CVP strategies in RA have not been sufficiently implemented into the clinical practice. To improve CVP in RA, there is a need for RCTs elucidating the optimal preventive strategies and for feasible clinical guidelines.

• It is important to incorporate CV outcomes in clinical trials on anti-rheumatic drugs, to search both for protective and deteriorating effects.

• While awaiting results from high-quality RCTs, the CVP in RA be based on available evidence and on careful clinical judgment of the individual patient.

• RA should be considered an independent risk factor of CVD.

• In CVP in RA, there should be a focus on reducing modifiable traditional and non-traditional CV risk factors, including disease activity.

• As CVD in RA is often unrecognized and related to a poor prognosis, a pro-active screening for CVD in RA is warranted.

Abstract

The increased risk of cardiovascular disease (CVD) in rheumatoid arthritis (RA) has been recognized for many years. However, although the characteristics of CVD and its burden resemble those in diabetes, the focus on cardiovascular (CV) prevention in RA has lagged behind, both in the clinical and research settings. Similar to diabetes, the clinical picture of CVD in RA may be atypical, even asymptomatic. Therefore, a proactive screening for subclinical CVD in RA is warranted. Because of the lack of clinical trials, the ideal CVD prevention (CVP) in RA has not yet been defined. In this article, we focus on challenges and controversies in the CVP in RA (such as thresholds for statin therapy), and propose recommendations based on the current evidence. Due to the significant contribution of non-traditional, RA-related CV risk factors, the CV risk calculators developed for the general population underestimate the true risk in RA. Thus, there is an enormous need to develop adequate CV risk stratification tools and to identify the optimal CVP strategies in RA. While awaiting results from randomized controlled trials in RA, clinicians are largely dependent on the use of common sense, and extrapolation of data from studies on other patient populations. The CVP in RA should be based on an individualized evaluation of a broad spectrum of risk factors, and include: 1) reduction of inflammation, preferably with drugs decreasing CV risk, 2) management of factors associated with increased CV risk (e.g., smoking, hypertension, hyperglycemia, dyslipidemia, kidney disease, depression, periodontitis, hypothyroidism, vitamin D deficiency and sleep apnea), and promotion of healthy life style (smoking cessation, healthy diet, adjusted physical activity, stress management, weight control), 3) aspirin and influenza and pneumococcus vaccines according to current guidelines, and 4) limiting use of drugs that increase CV risk. Rheumatologists should take responsibility for the education of health care providers and RA patients regarding CVP in RA. It is immensely important to incorporate CV outcomes in testing of anti-rheumatic drugs.

Introduction

RA is associated with a wide range of CV manifestations, including atherosclerosis, thrombosis, heart failure, valvular heart disease, arrhythmia, aortic aneurysms, myo-, peri- and endocarditis, vasculitis, rheumatoid cardiac nodules, and cardiac amyloidosis. Atherothrombosis, especially in the form of coronary artery disease (CAD), has a particular clinical impact as it substantially contributes to the mortality excess in RA [1].

Rheumatoid vasculitis may present with similar symptoms as atherosclerosis. The differential diagnosis of these conditions is challenging, which may lead to under-treatment of vasculitis. The diagnostics may be facilitated by novel imaging techniques and biopsy examination of patients undergoing cardiovascular surgery [2].

The cause of the accelerated atherosclerosis in RA has not been fully elucidated. Traditional CV risk factors and medication side effects do not explain the excess in the CV morbidity, and other factors, such as immune mechanisms and inflammation, appear to play an important role [3], [4], [5], [6], [7].

Atherosclerotic disease in RA is commonly under-diagnosed, and therefore untreated [8], [9]. This fact is likely to be caused by several factors: first, chest pain may be erroneously attributed by the patient and/or physician to non-cardiac, especially musculoskeletal, pathology, especially if a low index of suspicion for the CVD is present (particularly in younger patients, especially women). Second, due to a sedentary lifestyle, the patients are less likely to develop exertional angina. Furthermore, similar to diabetes, both stable CAD and acute coronary syndromes in RA may occur with an atypical clinical picture, including absence of symptoms. This may be partly explained by a large proportion of women among RA patients because CVD in women in general is known to frequently present with an atypical picture. In theory, the atypical CVD symptoms in RA may also be related to a medication or disease-related reduction of pain perception.

Moreover, there may be differences in the pathogenesis of acute coronary events in RA compared to non-RA individuals. RA patients may have a higher susceptibility to plaque instability and thrombogenesis causing acute coronary syndromes without prodromal symptoms from obstructive atherosclerotic lesions [10], [11], [12], [13]. Also other mechanisms, including vascular inflammation and vasoconstriction, might play a role. These mechanisms may partially explain the observed discordance between a relatively high occurrence of acute coronary syndromes and a relatively low occurrence of stable angina in RA, and also highlight limitations of traditional imaging techniques for detection of subclinical CVD to predict CV events. There is a huge need to clarify the contribution of these mechanisms in the pathogenesis of CVD in RA in order to determine optimal prevention.

Besides the increased occurrence of CV events, RA patients appear to have a more severe course of acute coronary events than non-RA patients, with higher recurrence and mortality rates, though data on this point are conflicting [14], [15], [16], [17]. One might speculate that late and inappropriate diagnostics and/or treatment, or different disease mechanisms (e.g., increased thrombophilia and plaque vulnerability), may play a role.

While the scope and clinical features of CVD in RA resemble those in diabetes, the approach to these two conditions has been dramatically different, with far less emphasis on CV prevention (CVP) in RA [18].

Although the increased CV risk in RA was first recognized in the 1980s, CVP is still poorly integrated into the management of RA patients [19]. For example, dyslipidemia and hypertension in RA are frequently undertreated [20], [21], [22], [23], [24].

One of the reasons is a limited awareness of the problem among physicians and patients. Moreover, due to a lack of appropriate clinical studies, optimal strategies for CV risk calculation and reduction in RA have not been determined yet. The paucity of randomized control trials (RCTs) focusing on CVP has been attributed to a relatively low prevalence of RA. However, such RCTs have been successfully accomplished in patients with other relatively rare diseases, such as in kidney transplant recipients, resulting in implementation of adequate CVP strategies in these groups [25], [26].

Of note, there are data available from subgroup analyses of RA patients participating in studies conducted on the general population, suggesting a protective effect of statins in RA [27].

The CVP in RA is also complicated by the fact that there exist different recommendations for CVP, e.g., according to co-morbidity (such as diabetes) and place of residence [28], [29], [30], [31], [32]. Although use of the local recommendations is usually suggested, this rule does not apply universally. For example, updated international recommendations may be more appropriate than older local recommendations. Thus, there may be a doubt which of the available recommendations should be used as a base for estimating of CV risk in RA.

Moreover, some of the recommendations may be considered complicated and lengthy, which may lead to their superficial reading and even avoidance. This may be one of the reasons to that the current CV risk calculators, e.g. SCORE (suggested by the European Guidelines on CVD prevention in clinical practice) and NORRISK (suggested by the Norwegian guidelines), are often used and interpreted in an oversimplified and incorrect manner [30], [33]. Thus, sometimes only variables listed directly in the risk estimation chart are taken into account (e.g., age, blood pressure, smoking and cholesterol levels) although the explanatory text clearly emphasizes that the risk estimation should be done individually, in the context of other relevant factors, such as psychosocial factors, co-morbidity and its severity and results from imaging and laboratory methods (Table 1) [33].

The Norwegian guidelines have introduced several coefficients to adjust the CV risk estimated by NORRISK in accordance with factors not listed in the chart (e.g., coefficients for different levels of HbA1C and number of relatives with premature CVD), which illustrates the complexity of the current strategies.

An individual evaluation has to be performed for individuals with age below or above the age span included in the risk estimation tools.

It is important to keep in mind that clinical guidelines are meant to facilitate, but not replace, clinical judgment.

There are indications that CV risk in RA is relatively less dependent on traditional CV risk factors than the CV risk in the general population, probably due to a strong influence of systemic inflammation and other RA related factors [34], [35]. Therefore, CV risk in RA appears to be underestimated by calculators for the general population (e.g., SCORE, Framingham score) if the RA related factors are not taken into account [36], [37], [38], [39], [40], [41], [42]. It has been demonstrated that even Reynolds score, which is partly adjusted for inflammation (in terms of high sensitivity CRP), underestimates the CV risk in RA [37]. Nevertheless, the traditional CV risk factors are likely to significantly influence the CV risk, implicating a need for their favorable modification [41], [43], [44].

Besides underestimating, the current CV risk algorithms may also overestimate the real CV risk in some RA patients, and, therefore, lead to their overtreatment [34]. For example, although adding RA along with ethnicity and deprivation to the other traditional CV risk factors has been observed to significantly improve the accuracy to identify high-risk individuals by QRISK2 (CV risk algorithm used in the United Kingdom), Arts et al. reported that QRISK2 overestimates the CV risk in RA [45], [46].

For the time being, the recommendations for CVP in RA are mainly based on expert opinion. However, the current strategies have some drawbacks (Table 2), and the expert opinions vary substantially [47], [48], [49]. For instance, while UpToDate recommends to follow the European League against Rheumatism (EULAR) recommendations for CV risk management in RA, others have not accepted these recommendations, and have preferred other alternatives [47], [48], [50]. Some suggest adding 5–15 years to the age of RA patients when using the general CV risk estimating tools. The Norwegian NOKAR study estimates the CV risk in RA using the SCORE risk chart for countries with a high CV risk, although Norway belongs to countries with a low CV risk [33]. However, it is still unclear how these adjustments depict the real CV risk in RA. For example, the risk in young patients, especially women, might remain underestimated.

The EULAR recommendations, based on expert consensus opinion, propose estimating risk in RA using the same tools as for the corresponding general population, i.e. the national guidelines or SCORE, and multiplying it by 1.5 in patients fulfilling two of the following criteria: 1. positive rheumatoid factor and/or anti-CCP; 2. disease duration > 10 years; 3. extra-articular manifestations. Unfortunately, the relatively strict qualifiers identify only a small part of the RA population, while it is likely that also other RA subgroups are at increased CV risk and would benefit from an intensified CVP [34], [46], [48], [51]. Furthermore, the suggested risk stratification seems to underestimate the real CV risk even in RA patients fulfilling the three qualifiers [41], [46], although it might also overestimate the CV risk in other patients [34].

Corrales A. et al. demonstrated that only 2% of RA patients categorized as having moderate CV risk by SCORE were re-classified as having high/very high risk using adjustments according to mSCORE [42]. Similarly, Karpouzas et al. observed modest risk reclassification ranging from 4 to 12%, depending on calculator used (classic Framingham, D'Agostino Framingham, SCORE, Reynolds), by applying the EULAR modification coefficient, and most reassignments occurred from low to moderate stratum [52]. Importantly, using carotid ultrasound, 63% of RA patients with a moderate CV risk according to mSCORE were re-categorized as having a very high CV risk due to the presence of plaques in their carotid arteries (which are known to strongly predict the occurrence of CV events both in RA and non-RA population) [42]. Moreover, carotid plaques have been reported in 24% RA women with a very low CV risk (SCORE = 0), with a highest occurrence (38%) among women of age > 49.5 years and with TC > 5.4 mmol/L [41]. As SCORE is build up on the fact that men are generally more predisposed to CVD than women, its predictive value may be substantially attenuated in RA women (especially at a younger age), who represent the majority of the RA population, and who may in reality be at a high/very high CV risk due to their RA.

The plaques occurred frequently both in patients with and without the three qualifiers suggested by EULAR. Thus, these results support the notion that even patients with seronegative RA without extraarticular manifestations and with relatively short disease duration might be at a very high CV risk, possibly due to the importance of inflammation already at the early phase of the disease [35], [42], [53].

Hence, it seems sensible to integrate carotid examination, a relatively feasible non-invasive method that may be performed by musculoskeletal ultrasound scans, more broadly into the assessment of RA patients as it may possibly aid in their protection from premature death. However, further studies are needed to validate the efficacy of CVP strategies guided by carotid ultrasound assessment on the occurrence of subsequent CV events in RA.

It is important to keep in mind that, in contrast to the presence of carotid plaques, increased carotid intima media thickness is not considered an equivalent for clinical CVD, but only a factor associated with increased CV risk [33].

Due to the limitations of the current EULAR recommendations, their revision is underway. The proposed revised recommendations suggest applying the multiplying coefficient of 1.5 in all RA patients, i.e. independently of the presence of the 3 qualifiers. Furthermore, they suggest considering carotid ultrasound assessment in the CV risk evaluation [54]. However, the full version of these recommendations has not been published yet.

In the future, the CVP in RA will hopefully be facilitated by a new CV risk chart that is under development by the ATACC-RA (= A TransAtlantic Cardiovascular Risk Calculator for Rheumatoid Arthritis) consortium.

Preliminary results from a current study revealed that, in Caucasians, using Framingham Score with a lower cut off level than that used in the general population might be a useful surrogate marker of atherosclerosis in RA [55].

The CV risk score may change considerably with fluctuations of inflammatory activity and with pharmacological treatment. Therefore, the CVP should be adjusted during the course of the disease. While UpToDate recommends to screen CV risk profile in RA in patients from the age of 50, others recommend to initiate the screening earlier. It is noteworthy that the Canadian Cardiovascular Society recommends monitoring of lipid profile in all RA patients independently of age [56]. The EULAR recommendations proposed a yearly CV follow-up of RA patients, but more seldom (e.g., every 2–3 years) in those with inactive RA with a low CV risk [48]. Furthermore, they recommend reconsidering the CV risk assessment over the course of the disease.

In spite of these challenges, it is important to focus on CVP in the management of RA. The strategies should be based on the current levels of scientific knowledge on RA, extrapolations from studies on other patient groups, expert opinions, and clinical judgment. It is likely that we err less by initiating a relatively proactive CVP in RA than by a passive approach with avoidance of CVP until convincing data from large RCTs on RA population are available. It is likely that it will take a long time before conclusions from reliable RCTs are published.

Simple and clear recommendations would probably accelerate the integration of CVP into the management of RA. It is also important to educate health care providers and patients in this field.

The uncertainty regarding CVP in RA concerns first of all the question to which degree the LMT in RA should be more aggressive than that in the general population. However, besides LMT, there are opportunities for other preventive interventions that are far less controversial, and that have benefits beyond their cardioprotective effects (e.g., promotion of healthy life-style and reduction of RA disease activity) [33]. Thus, the confusion about the aggressiveness of LMT should not lead to a paradoxical avoidance of the CVP as a whole [20], [21], [22]. Moreover, when we look at the topic more closely, we see that there are some clear clues for the LMT in RA, and that the controversies concern relatively small nuances in the target lipid levels.

There is a broad agreement that most individuals, including those with a low CV risk, should have LDL < 3 mmol/L, while those with a higher risk should have lower LDL levels, some as low as < 1.8 mmol/L, depending on the degree of the risk. For instance, the European Guidelines on CV disease prevention in clinical practice (version 2012) recommend LDL < 3 mmol/L in all individuals with a low and moderate CV risk, < 2.5 mmol/L in individuals with a high CV risk, and < 1.8 mmol/L in those with a very high CV risk. The group with a very high risk comprises patients with any of the following conditions: 1. clinical or subclinical CVD, 2. complicated diabetes (diabetes with target organ damage and/or at least one other CV risk factor), 3. severe chronic kidney disease and 4. SCORE ≥ 10% 10-year risk of fatal CVD. Patients with uncomplicated diabetes, moderate chronic kidney disease, a markedly pronounced CV risk factor (such as familiar dyslipidemia or severe hypertension) or SCORE of 5–9% belong to the high risk group [33].

On the other hand, the 2013 American ACC/AHA guidelines on the treatment of blood cholesterol to reduce atherosclerotic CV risk in adults do not longer advocate treatment to specific cholesterol targets. Instead, a graded intensity of statin treatment (i.e., the type and dose of the statin) is recommended to different groups of individuals, according to the level of their CV risk (Table 3) [29]. This solution is based on the fact that baseline and on-treatment LDL levels have not been proven to alter the beneficial effect of statin therapy. Moreover, other current forms of LMT appear to be less effective and/or more likely fraught with side effects than statins, in spite of their beneficial effects on lipid levels (however, new promising lipid-modifying drugs are emerging). Hence, the ACC/AHA guidelines recommend statins for all diabetics, but with a graded intensity of the statin treatment depending on the patients total CV risk estimate (Table 3) [29]. For individuals with 1) CVD, 2) LDL ≥ 4.9 mmol/L, 3) 10-year risk of atherosclerotic CVD ≥ 7.5% and age 40–75 years or 4) diabetics with LDL > 1.8 mmol/L and age 40–75 years, a moderate- to high-intensity statin regimen is recommended.

Because the CVD burden in RA approaches that in diabetics, and because the pathogenesis of premature CVD in these two conditions share some common pathways, it has been proposed that a similar CVP strategy should be followed [57], [58]. One option would therefore be prescribing statins to all RA patients, with graded intensity of the statin regimen according to the actual CV risk level, e.g. in terms of inflammation and co-morbidities) [59].

Thus, it is possible to summarize that RA patients in general should have LDL < 3 mmol/L, but most RA patients should probably have LDL < 2.5 mmol/L, and some even < 1.8 mmol/L (Table 5). There is a need to examine if some patients with especially severe RA (in particular, those with a high inflammatory burden and/or extra-articular manifestations), but without other CV risk factors, would nonetheless benefit from stricter LDL thresholds, such as LDL < 2–1.8 mmol/L [35]. Taken together, it appears that choosing LDL target from the range 1.8–3.0 mmol/L, and perhaps considering statins for all RA patients, may be indicated. However, it is very important to follow up the effects and safety of the chosen regimens, and accordingly continuously reevaluate the CVP strategy.

In the follow-up of RA, it is important to keep in mind that LDL, HDL and total cholesterol (TC) levels are inversely related to disease activity. RA patients often have low LDL and TC during a high disease activity phase, while these levels increase with the reduction of inflammation. This effect leads to an apparent “lipid paradox” as RA patients with lower LDL and TC have higher CV risk and vice versa [60]. The “lipid paradox” is likely caused by the important role of inflammation in the pathogenesis of CVD in RA, and has been observed also in other inflammatory conditions. It has been suggested that TC/HDL ratio might be a useful marker for monitoring of CV risk in RA as these lipids usually change in parallel with the inflammatory activity [61]. For a reliable guidance of LMT, it is proposed to assess lipid levels during a low disease activity phase. Nonetheless, no RCTs have confirmed the usefulness of this approach yet. Furthermore, it is important to keep in mind that high LDL and TC may increase the CV risk in RA, but that the strong influence of inflammation in the CVD pathogenesis may confound their effects [41].

The reason for the inverse correlation between cholesterol levels and inflammatory activity, and the cause–effect relationship, is still unclear. In hypothesis, there may be a decreased synthesis of some lipoprotein particles during inflammation, or their increased consumption, e.g. due to the demands of rapidly proliferating cells, or due to other roles of cholesterol in mechanisms combating the inflammatory activity and its cause [62].

Some anti-rheumatic drugs theoretically might increase CV risk through induction of hypercholesterolemia. Nonetheless, the observed increase in TC and LDL during anti-rheumatic therapy is likely to be at least partially caused by the improvement of RA activity. Moreover, it is now clear that pro-atherogenic activity of LDL and anti-atherogenic activity of HDL does not simply mirror their circulating levels, but also their structural and functional changes ([63]. For example, HDL protective function in RA is impaired both with respect to its anti-inflammatory and anti-oxidant properties [64], and to its capacity to promote cell cholesterol efflux and reverse cholesterol transport [65]. In this context, methotrexate and anti-TNF agents have been shown to improve lipoprotein and macrophage functions, and, therefore, reduce foam cell formation [66]. The anti-inflammatory effect and the ameliorating effect on cell cholesterol transport, irrespective of lipoprotein levels, may partly underlie the cardioprotective effects of some anti-rheumatic drugs. However, there are probably differences between the individual agents, and further studies determining the potential beneficial or deteriorating effects of anti-rheumatic drugs on lipid metabolism and transport are therefore warranted [66].

Besides the aforementioned factors, also impaired LDL quality (decreased LDL size), increased apoB/apoA1 ratio and atherogenic index (i.e., a relative deficiency of HDL to TC), increased lipoprotein (a) level and triglyceride level, and presence of antibodies to oxidized LDL may contribute to the pathogenesis of premature atherosclerosis in RA [61], [67].

Evidence suggests that statin treatment can reduce TC and LDL-C and CV risk in RA patients to the same extent as in the general population [27], [68], [69], although this finding has not been universal [70]. Nonetheless, statins have been shown to counteract surrogate parameters of CVD in RA such as endothelial dysfunction and arterial stiffness [71], [72], [73]. Rosuvastatin treatment, aiming to reduce LDL to ≤ 1.8 mmol/L, has been observed to decrease thickness of carotid plaques during an 18 months period in RA. The thickness reduction was lowest in patients with the highest burden of RA activity during the follow-up [74]. Moreover, statins have been shown to reduce mortality [68] and CV morbidity [27] in RA. A recent meta-analysis indicates that statins have a beneficial effect in primary CVP in RA, and that statin discontinuation increases the CV risk [75].

One might question why statins should be prescribed to patients with high RA activity, who usually have relatively low cholesterol levels. However, statins have pleiotropic effects that may be independent of the cholesterol lowering properties. It has been proposed that the cardioprotective effect of statins may be partly due to immunomodulation, inhibition of inflammation (including vascular inflammation), plaque stabilization, improvement of endothelial and HDL function, reduction of sympathetic activity, and anti-oxidative and antithrombotic effects [59], [76], [77]. Notably, statins may be especially beneficial in individuals with chronic inflammatory state, independently on their cholesterol levels, as they have been observed to reduce CV risk in primary prevention of individuals without hyperlipidemia, but with elevated high-sensitivity CRP [59]. It is important to realize that an early initiation of statin therapy, even in a population with a relatively low absolute CV risk, is likely to have the greatest long-term benefits [59].

Nonetheless, the European guidelines on cardiovascular disease prevention in clinical practice declare that: “as yet there is no indication for the preventive use of lipid-lowering drugs only on the basis of the presence of autoimmune diseases” (class recommendation A, level of evidence B) [33], while the position paper 2013 of the International Atherosclerosis Society does not mention autoimmune diseases at all. On the other hand, the ACC/AHA 2013 guidelines recommend considering statin treatment, according to clinical judgment, in serious comorbidities such as rheumatologic and inflammatory disease in spite of the insufficient evidence [29].

There is a need for caution regarding statin treatment. Besides well-known side effects such as myopathy, hyperglycemia and hepatotoxicity, there are indications that statins may contribute to development of autoimmune disease including RA [78], [79], [80].

In post hoc analyses of IDEAL study, RA patients (n = 87) had a higher occurrence of adverse effects than non-RA patients (n = 8801). For example, myalgia occurred in 10.4% and 7.7% of RA vs 1.1% and 2.2% of non-RA patients receiving simvastatin and atorvastatin, respectively. Although the differences in adverse effects did not reach statistical significance, the observations should be interpreted with care as the study was not powered to draw conclusions regarding equivalence [81]. Conversely, in a large cross-sectional study, statin use was independently associated with musculoskeletal pain only in non-RA, but not in RA patients [82].

A post hoc analysis of IDEAL and TNT study, comparing 280 patients with inflammatory joint disease (including 199 RA patients) to 18 609 non-RA patients, did not reveal any significant difference in the occurrence of myopathy, including rhabdomyolysis, or significant liver enzyme elevation. However, patients with inflammatory joint diseases were more likely to discontinue statin treatment due to serious adverse events. There was also a trend towards increased occurrence of myopathy among RA patients using simvastatin 40 mg compared to atorvastatin 10 and 40 mg [27]. Simvastatin is known to be prone to adverse drug–drug interactions, and to increased risk of myopathy at high-dose treatment (80 mg) [29]. It is important to evaluate if simvastatin is less tolerated in RA patients than in other populations, e.g. due to polypharmacy or other factors.

Toms and colleagues did not detect any case of myopathy in RA patients during over 1500 patient–years of follow-up, although the patients harbored multiple risk factors for statin induced myopathy. The genetic susceptibility to statin induced myopathy appeared to be similar in RA and non-RA individuals [83]. Also findings from other studies suggest that statins may represent a relatively safe treatment in RA, which may be cost-effective and even protect from RA development, activity and progression [84], [85], [86], [87], [88], [89].

Also results from large RCTs on the general population and the long-term clinical experience from a wide patient population indicate a relatively good safety profile of statin therapy [33], [90]. In fact, due to their pleiotropic effects, statins are likely to have various positive bi-effects, including anti-cancer properties [91].

Although the evidence to date is mostly promising, there is a need for further studies examining the safety and efficacy of statins in RA. It is important to identify which statin regimens are the most suitable in RA with respect to the CV protective and anti-inflammatory effects, and to side effects. Furthermore, it is important to determine which subgroups of RA patients should use LMT, and what should the treatment targets be.

Inflammation plays an important role through all stages of the atherosclerotic process, from endothelial dysfunction, plaque formation and their destabilization, to thrombogenesis. Chronic inflammatory state in general is considered to increase CV risk via various mechanisms which contribute to oxidative stress, endothelial dysfunction, prothrombotic state and pro-atherogenic metabolic effects (in particular disturbances of lipid and glucose metabolism) [4], [92]. For example, serum amyloid protein, an acute phase marker, may replace apolipoprotein A-1 in HDL molecules, and thus impair the HDL function towards a pro-inflammatory and pro-atherogenic direction.

Thus, not surprisingly, inflammation and immune dysregulation appear to play a major role in the development of premature CVD in RA [35], [93], [94]. The CV risk in RA increases with disease activity and severity (such as positivity for anti-citrullinated protein antibodies (APCA) and rheumatoid factor (RF), presence of shared epitope, joint damage, disability and extra-articular manifestations) [48]. The importance of ACPA and RF in CVD is supported by their link to CV risk also in the general population [1]. However, it is still unclear if these antibodies themselves enhance the CV risk, or if they are bystanders of other, pathogenic, factors, such as inflammatory burden.

Although some studies reported a positive relationship between CV risk and RA duration, others could not confirm this association [95], [96], [97]. It is possible that the burden of disease over time rather than RA duration per se contributes to enhanced CV risk [96]. It is notable that CV risk may be increased already during the early course of RA, even before the formal criteria for RA diagnosis are met [8], [53], [98], [99], [100], [101], [102].

There has been most focus on the potential role of systemic inflammatory markers, such as C-reactive protein ( CRP), tumor necrosis factor (TNF) and interleukin-6 (IL-6), for CV risk in RA. Nonetheless, also inflammation in the CV system, in particular in vessels (both outside and inside of atherosclerotic lesions) and in the heart, may play a major role in the pathogenesis of premature CVD in RA, and may not be detected by measuring levels of the circulating markers [4], [13], [103], [104]. Thus, there is a need to clarify which drugs reduce inflammation in the CV system, and how this influences the CV morbidity and mortality. An interesting PET scan study from United Kingdom demonstrated that RA patients had more inflammation in their large arteries than non-RA individuals with stable CVD. Importantly, this inflammation was inhibited by 8-weeks anti-TNF therapy, and the reduction was related to the improvement in vascular stiffness [104].

An emerging circulating marker of CV risk is pentraxin 3 (PTX3), which appears to more accurately mirror vascular inflammation and CV risk than CRP. Increased PTX3 levels have been observed in inflammatory rheumatic diseases including RA [105]. Although PTX3 levels are positively related to CV risk, the role of PTX3 in CVD may be protective [4].

Reduction of time-averaged RA disease activity has been observed to be associated with fewer CV events [35]. Therefore, anti-rheumatic treatment may play an essential role not only in protection from joint damage, but also from CVD. Indeed, anti-rheumatic treatment is reported to reduce CV morbidity and mortality in RA (with a better effect in responders than non-responders) [106], [107]. The evidence is so far best for anti-TNF treatment and methotrexate [108], [109], [110], [111], [112]. However, some observations suggest that also other anti-rheumatic drugs, including sulfasalazine, anakinra, abatacept and rituximab, may reduce CV risk [113], [114], [115], [116], [117], [118]. Hydroxychloroquine (HCQ) has anti-thrombotic properties, improves lipid and glucose metabolism, and appear to protect from diabetes in RA and to reduce CV morbidity and mortality in systemic lupus erythematosus [119]. Thus, HCQ, which represents a relatively inexpensive and safe intervention, might be useful in the reduction of CV risk. Hence, there may be a reason for a more liberal prescription of HCQ in RA, similar to lupus. However, there is a need for RCTs examining the effect of these drugs on CV events and mortality before definitive conclusions can be drawn. Currently, a RCT testing the effect of HCQ on CV risk in patients with kidney disease is ongoing (clinicaltrials.gov: NCT01537315).

The cardioprotective effect of methotrexate has been observed to be potentiated by folic acid substitution and by combination with HCQ and sulfasalazine [98], [120].

Paradoxically, in spite of significant improvements in anti-rheumatic therapy, the mortality gap between RA and general population has appeared to widen [19]. The reason to this phenomenon is unclear, and might be caused, among other factors, by persistent low-grade inflammation even in apparently well-treated patients. It is possible that an early and effective long-term inhibition of inflammation results in a better CV prognosis [1]. Also, we have to keep in mind that we currently do not know how much does the modern aggressive anti-rheumatic therapy and management of modifiable CV risk factors improve CV mortality in RA, as the evaluation requires a longer observation.

This cardioprotective effect of anti-rheumatic therapy has been mostly attributed to the inhibition of systemic inflammation, with consequent improvements of the related metabolic disturbances. However, also inhibition of vascular inflammation might play a crucial role as vascular inflammation promotes plaque progression and destabilization [4]. It is notable that atherosclerosis-related vascular inflammation involves some of the immune factors known to play a key role in synovitis, such as TNF [4]. Moreover, it is possible that cardiac inflammation (involving TNF) contributes to the increased risk of heart failure in RA, which may partly explain why the risk of heart failure in RA is higher than that expected on the base of CAD. Therefore, it is possible that anti-inflammatory drugs, such as methotrexate, could protect from heart failure [103], [121], [122], [123], [124], [125]. Although anti-TNF treatment is suspected to promote heart failure in the general population, it is still unclear whether and into which extent anti-TNF really promotes heart failure in RA. In fact, some results indicate that anti-TNF even might protect from heart failure in RA [98].

Interestingly, anti-rheumatic treatment may ameliorate CV risk also via actions that are at least partially independent of their anti-inflammatory activity, e.g., via decreased Lp(a) levels and via reduced uptake and increased efflux of cholesterol by macrophages, which implies a reduced foam cell formation [66], [126].

Glucocorticoid therapy has adverse effects that negatively influence CV risk, e.g., disturbances of lipid and glucose metabolism and increase in blood pressure. Glucocorticoids also can act directly on macrophages, increasing cholesterol accumulation [127]. However, these pro-atherogenic effects can be to some extent counteracted by the anti-inflammatory properties that are likely to protect from CVD, in particular in chronic inflammatory conditions. Indeed, RA ameliorating treatment, which includes glucocorticoids, is associated with a decreased CV risk in RA [114]. Furthermore, glucocorticoids attenuate further CV risk in RA patients with a history of CVD [128], [129].

However, available data is conflicting, and a large body of evidence indicates that glucocorticoid treatment is a strong independent CV risk factor in autoimmune diseases, including RA [7], [130], [131], [132], [133]. The risk of CVD and CV and total mortality appear to be dose-dependent, increasing with higher daily and cumulative dose [134]. The CV toxicity appears to be minimal at daily doses corresponding to prednisone ≤ 5–7.5 mg [130], [132], [134], [135]. For this reason, current recommendations, including the EULAR recommendations, state the need to limit the use of glucocorticoids to the lowest dose and for the shortest possible duration [1], [48].

Also other factors, such as genetic factors (e.g., the shared epitope alleles), poor lifestyle, comorbidity and laboratory aberrations (including metabolic syndrome) and medication-side effects may contribute to CV risk excess in RA [1], [4], [136], [137]. Interestingly, some CV risk factors (such as smoking, dyslipidemia, obesity, periodontitis and genetic factors) appear also to promote RA development [1]. These facts support the hypothesis that RA and CVD may share some common pathogenetic mechanisms [138].

However, it is important to realize that the traditional CV risk factors only partially explain the CVD excess in RA. Additionally, the relative contribution of the individual traditional CV risk factors to the total CV risk may be different in RA compared to the general population [7]. Interestingly, traditional CV risk factors and inflammation appear to interact, as the impact of systemic inflammation on atherosclerosis reportedly depends on the traditional CV risk burden extent in RA. The disease related factors and traditional CV risk factors have been observed to associate most strongly with atherosclerosis in younger and older RA patients, respectively [44], [139].

Smoking has multiple adverse effects, such as increasing the risk of CVD, respiratory disease and cancer. The link between smoking and CV risk has been confirmed also in RA [43]. It is noteworthy that smoking also seems to play a pathogenic role in RA, promote RA activity and reduce response to anti-rheumatic therapy [140], [141]. Thus, facilitation of smoking cessation represents one of the most important interventions in RA.

The occurrence of hypertension in RA appears to be increased, although this finding is not universal. The discrepancies between study results may be caused, among other factors, by differences in the examined patient populations and in hypertension criteria, and immunosuppressive therapy. Regardless, hypertension is one of the most important CV risk factors in RA [23], [67], [139], [142]. The adverse effect of hypertension on CV risk may be stronger in RA than non-RA patients. Hypertensive RA patients have been reported to display significantly higher numbers of coronary segments harboring plaque, greater degree of coronary stenosis and greater plaque burden than matched non-RA individuals [143].

Various factors, including co-morbidities (e.g., kidney disease and sleep apnea), obesity, physical inactivity, stress and medication side effects (e.g., due to non-steroidal anti-rheumatic drugs (NSAIDs), glucocorticoids, leflunomide) may contribute to blood pressure increase in RA. Importantly, also inflammation and immune dysregulation might play a role. Hypertension is associated with increased levels of inflammatory markers such as CRP and pro-inflammatory cytokines (TNF, Il-6 and Il-17), shifts in T cell sub-populations (such as down-regulation of regulatory T cells), and immune cell recruitment into the vascular adventitia and perivascular tissue and to the kidneys [144], [145]. The relationship between inflammation and hypertension is reciprocal, and hence may cause a vicious circle: hypertension may induce inflammation (e.g. due to vascular damage secondary to mechanic and oxidative stress and due to humoral mechanisms), while inflammation appears to play a pathogenic role in hypertension and the related target organ damage (e.g., by enhancing vascular stiffness and kidney impairment). In keeping with this notion, increased levels of hs-CRP predict the development of hypertension. It seems that both the specific and innate immune system play a pivotal role in hypertension development [144], [145]. Autoimmunity, in particular to isoketal adducts induced by oxidative stress, is likely to significantly contribute to the hypertension associated inflammation. Interestingly, premature senescence of T cells, which is likely to play a role in RA and atherosclerosis, may be important also in the pathogenesis of hypertension [144]. It is tempting to speculate that the inflammation in deep vascular layers observed in CAD patients with and without inflammatory rheumatic diseases might be a common denominator of hypertension as well as atherosclerosis and some types of vascular aneurysms, which may imply shared pathogenic mechanisms in these conditions [38], [146]. Hypertension has been observed to be associated with significantly higher burden of vulnerable (mixed) coronary plaques in RA compared to non-RA individuals [143]. It remains to be examined in further studies if hypertension itself promotes destabilization of plaques, or if their destabilization might be triggered by the vascular inflammation associated with hypertension.

Besides antihypertensive medication, also non-pharmacological interventions, including regular physical activity, should be emphasized in hypertension treatment. Furthermore, it has been suggested that inhibition of inflammation might protect from hypertension. Notwithstanding, although mycophenolate mofetil (which suppresses T-cells) has been reported to reduce blood pressure, such effects have not been observed for several other immunosuppressive regimens, including anti-TNF therapy, that may even increase blood pressure [144], [147]. This topic is though complicated as the potentially beneficial effect of the anti-inflammatory drugs might be confounded by other, toxic, effects. Hence, the role of different anti-inflammatory and immunomodulating agents in obtaining optimal blood pressure control in RA requires further study [23].

RA is associated with arterial stiffness, which is known to predict CV events and mortality independently of traditional CV risk factors including hypertension [148]. Although arterial stiffness has been traditionally considered a consequence of hypertension, it also appears to be involved in its pathogenesis. Arterial stiffness has been largely attributed to calcifications and changed characteristics of the extracellular matrix. However, it is now increasingly recognized that inflammation substantially underlies its development. Indeed, anti-inflammatory treatment has been observed to rapidly reverse arterial stiffness [4], [104].

RA is associated with increased prevalence of diabetes and increased insulin resistance. Insulin resistance, which increases the risk of diabetes type 2 development, has been found to be related to CV risk in RA, and is dependent on disease activity. It is noteworthy that drugs improving insulin resistance are under development.

RA patients are predisposed to changes in body composition, including increased fat/lean body mass ratio and abdominal adiposity [149], [150], [151], [152], [153], [154]. Thus, similar to the general population, waist circumference may be a better marker of CV risk in RA than BMI [154], [155], [156], [157]. While RA inflammation may promote accumulation of the body fat mass, it may lead to degradation of skeletal muscles [151], [158]. These facts may underlie the fact that increased fat mass occurs even in RA patients with normal or low BMI, such as in rheumatoid cachexia [159], [160].

The CV risk attributed to adiposity is likely to be due to several factors, including pro-inflammatory effects of adipokines secreted by adipose tissue [161]. It is known that adipose tissue at different locations has different impact on the CV risk: for example, visceral obesity is associated with a greater CV risk than is subcutaneous obesity. One might speculate that within the visceral obesity, the adipose tissue depots related to the heart and vessels are of a particular importance, e.g. due to its mechanic, metabolic and secretory properties, and because it may contain inflammation influencing the adjacent structures [4]. Increased volume of adipose tissue in the epicardium and perivascular tissue, which has been observed also in patients with autoimmune rheumatic diseases, is related to CV risk [4], [162], [163], [164]. Thus, there is a need to explore the clinical importance, causes and therapeutic implications of adiposity including “CV adiposity” in RA.

It is important to keep in mind that not only overweight, but also underweight increases CV risk in RA (and in non-RA population) [165]. In fact, BMI is inversely related to mortality in RA. This finding has been attributed to an underlying inflammation because high disease activity may lead to weight loss (e.g., in RA cachexia) [166]. As underweight is related to increased total and CV mortality in CAD patients, the possible adverse role of underweight in CVD deserves more focus [167], [168].

Arthritis may curtail physical activity. However, appropriate physical exercise in RA may be of a great importance as it may reduce CV risk as well as improve joint mobility [169], [170], [171]. Exercise has been shown to ameliorate both macrovascular and microvascular function and cardiorespiratory fitness in RA [172], [173]. An appropriate chronic physical activity may also exert anti-inflammatory effects, e.g. reduce expression of CRP and pro-inflammatory cytokines, and protect from development of severe RA [173], [174], [175], [176].

It is noteworthy that skeletal muscles secrete myokines regulating immune system and metabolic functions. These factors are likely to mediate some of the cardioprotective effects of exercise, e.g. by counteraction adverse reactions of pro-inflammatory adipokines. As the secretion of myokines appear to increase during muscle contraction, it is possible that exercise could reduce CV risk in RA via increased myokines secretion due to increased muscular activity and due to improved fat/lean mass ratio [177]. However, also other factors may play a role, such as a direct beneficial effect of shear stress on the vessel wall (e.g., reduced release of microparticles due to increased mitochondrial biogenesis) [178], [179].

Several RA associated factors may have a negative impact on the psychosocial situation, and lead to social isolation and limited opportunities to achieve high education and income. These and other factors can lead to stress, which in turn can promote CV risk via metabolic changes (e.g., hyperglycemia, insulin resistance and hypertension) and immune dysregulation characterized by a proinflammatory state. There has been observed an association between symptoms of tension as estimated by the Arthritis Impact Measurement Scale and the occurrence of carotid plaque in African RA patients [180]. Taken together, there is a need to explore the role of stress in premature CV morbidity in RA, and its reversibility by stress management strategies.

Depression is common yet often unrecognized and untreated in RA [181]. Interestingly, there is a reciprocally reinforcing association between depression and inflammation: systemic inflammation increases risk of depression, while depression seems to induce immune dysregulation, with signs of autoimmunity, and a chronic pro-inflammatory state (including elevated TNF, IL-1, IL-6 and Th17 response and increased proliferation of B-cells). Both inflammation and other factors related to depression, including hypercoagulation, poor lifestyle and reduced compliance regarding chronic therapies, may contribute to the increased CV risk. Depression is suspected to play a role both in development of atherosclerotic plaques, and in triggering of acute coronary syndromes (probably due to increased tendency to plaque destabilization and thrombogenesis). RA patients with depression appear to have a higher CV risk and mortality, and a lower response to anti-rheumatic therapy. Besides its effect on mood, anti-depressive treatment has also been observed to decrease inflammation, CV morbidity, pain and fatigue. In theory, it may also improve the patients' compliance regarding anti-rheumatic treatment [4]. An AHA Scientific Advisory Board has recommended routine screening and treatment of depression in CVD, and it is likely that such screening would be beneficial also in RA [182]. Although the beneficial effects of anti-depressive treatment on CV outcomes has not been verified in large RCTs yet, appropriate treatment of depression is generally recommended.

Kidney disease occurs relatively frequently in RA. It may be caused by RA related manifestations (glomerulonephritis, renal vasculitis and secondary amyloidosis), concurrent disease not related to RA, and side effects to anti-rheumatic, anti-inflammatory and other medications. As kidney disease, e.g. in terms of renal failure and proteinuria, has a significant effect on CV risk, there is a need to monitor RA patients for the development of renal disease, and, if necessary, provide adequate treatment.

Although obstructive and/or central sleep apnea may occur as relatively common co-morbidity in RA, this problem has received relatively little attention [183], [184], [185]. Yet it may be of a major importance because sleep apnea may significantly increase the morbidity (e.g., from fatigue and CVD) and mortality [4], [186], [187]. Interestingly, sleep apnea is suspected to promote development of autoimmune diseases including RA [188], and appears to be linked to inflammation [189].

Obstructive sleep apnea (OSA) is characterized by recurrent collapse of the upper airways during sleep, resulting in substantially reduced or complete cessation of airflow despite ongoing breathing efforts. The development of OSA is facilitated by narrowing of upper airways due to accumulation of adipose tissue, reduced muscle tone, or other factors. Central sleep apnea, which is more rare than OSA, is caused by mechanisms impairing the function of the breath center in the brainstem. The central and OSA may occur in combination, as a complex sleep apnea.

In RA, some disease specific factors, such as the arthritis of the temporomandibular joint or affection of the cervical spine, may cause OSA [185]. As indicated by observations from patients with spondyloarthritis, in whom the occurrence of OSA was reduced by anti-TNF treatment, it is possible that inflammation plays a role in the pathogenesis of OSA [190]. However, a study performed in RA patients suggested improved sleep quality after anti-TNF treatment, but no apparent improvement of sleep apnea [184]. Thus, there is a need for more research on the clinical impact, causes and treatment of sleep apnea in RA.

The link between periodontitis and CVD in the general population is well-established. Although the cause-effect relationship is not entirely clear, there is a suspicion that periodontal bacteria, such as Porphyromonas gingivalis, may play a pathogenic role in CVD [4]. Thus, the increased occurrence of periodontal disease in RA might contribute to the development of premature atherosclerosis, and may play a role in the pathogenesis of RA itself. Interestingly, there is a suspicion that periodontitis might play a pathogenic role also in RA [4]. Porphyromonas gingivalis, possessing peptidylarginine deiminase, is able to induce citrullination of arginine, and therefore formation of ACPA, i.e. autoantigens that play a crucial role in RA. Periodontal bacteria have been observed in joints of RA patients. Importantly, severity of periodontitis is related to severity of RA disease, and periodontal therapy appears to reduce RA activity [4].

The dental status in RA may be negatively influenced by secondary Sjøgren's syndrome, and by impaired oral hygiene due to arthritis of the temporomandibular joint and due to limited dexterity. It is notable that intensive dental hygiene appears to reduce CV risk [191], [192], [193], [194], [195]. The effect of dental hygiene on CV health in RA has not been tested yet. However, as good dental hygiene is a generally desirable, cheap and safe intervention, there is no significant reason not to recommend it, and the effects of dental hygiene on CV morbidity warrants further studies.

Patients with RA may have an increased susceptibility to thrombosis due to systemic inflammation and due to the presence of antiphospholipid antibodies (aPLs), which may contribute to the increased frequency of CV events. However, the role of aPL in premature CVD in RA is still unclear.

Influenza epidemics are associated with an increased rate of CV events, and influenza vaccination in the general population appears to be a cost-effective CVP modality. The effect is probably greatest in high risk groups [196]. The European guidelines recommend influenza vaccination in secondary CVP [33]. Similarly, the CV risk may also be reduced by pneumococcal vaccine [197]. Because RA patients are at increased risk of infections due to immune dysregulation and immunosuppressive treatment, appropriate vaccination should be addressed.

RA is often associated with low levels of vitamin D, which is known to regulate the immune system. Based on epidemiological studies and on observed biologic effects of vitamin D, vitamin D deficiency is suspected to contribute to development of a variety of diseases including atherosclerotic CVD, hypertension and autoimmune disease [145], [198]. Thus, vitamin D might represent a relatively safe and cheap treatment in these conditions [199], [200], [201], [202]. However, the effect of vitamin D supplementation on the reduction of RA activity and CV risk is still uncertain, and further studies are therefore needed [201]. In the meanwhile, as optimal levels of vitamin D are desirable due to several reasons, in particular due to the increased risk of osteoporosis in RA, adequate vitamin D supplementation in patients with insufficient vitamin D levels appears reasonable [200], [203].

There are observations indicating association between high serum vitamin D levels (> 100 nmol/L) and adverse effects including increased CV mortality. Due to lack of RCTs examining this topic, the cause–effect relationship is though still unclear [204].

RA is associated with hyperhomocysteinemia. Hyperhomocysteinemia may reflect deficiency of vitamin B12, B6 and folic acid, and may be induced by methotrexate, a folic acid antagonist. Although hyperhomocysteinemia is related to CV risk, it is not sure that it plays a pathogenic role, and that homocysteine-reducing therapies would reduce CV risk. Several RCTs examining the role of homocysteine-lowering therapy in primary CVP are underway. Nevertheless, an adequate supplementation in patients with hypovitaminoses is generally desirable, and folic acid supplementation in methotrexate users is recommended [58].

Hypothyroidism, often of autoimmune origin, has been reported to frequently occur in RA, with an increased occurrence already before the RA onset [205]. Hypothyroidism, including subclinical hypothyroidism, is related to increased CV risk in the general population, and may contribute to premature CVD in RA [33], [205], [206], [207], [208]. Consequently, proactive screening for hypothyroidism seems prudent in RA, as adequate replacement therapy may reduce CV risk as well as improve overall health.

Hyperuricemia is related to CV risk and mortality, but its role in RA has not been fully determined. While some studies indicated a relationship between uric acid levels and CV risk in RA [208], [209], [210], a current large longitudinal population-based study from USA demonstrated a relationship to all-cause mortality and peripheral arterial disease events, but not to other CV manifestations [211]. As a causal relationship between hyperuricemia and CVD has not been established, a urate-lowering therapy is not recommended for treatment of CVD [58]. On the other hand, there are recommendations for treatment of gout and hyperuricemia in the general population (that also apply for RA patients), with aim to ameliorate other types of clinical manifestations [212].

It has been suggested that decreased numbers and function of EPC might play a role in the development of premature CVD in RA, although findings to support this theory are inconsistent [213], [214], [215]. Decreased EPC numbers and function appear to play a pathogenic role in atherosclerosis as these cells contribute to endothelium repair and revascularization. Interestingly, low EPC count in RA has been observed to be related both to endothelial dysfunction and to high bone erosion score. There are theories that EPC and optimization of their activity might be used therapeutically in CVD. In fact, there are indications that some of the existing drugs, such as anti-TNF therapy, angiotensin-2 receptor and angiotensin converting enzyme blockers and fish oil supplementation, have a beneficial effect on EPC [214], [216], [217]. However, besides their protective role in atherosclerosis, EPC may also have detrimental effects in other instances, such as RA synovitis and cancer, because they promote pathologic neovascularization. In fact, it has been suggested that the low count of peripheral EPC in RA might be due to their migration to the joints [213]. It is possible that different subsets of EPC have different roles; research on this topic is ongoing.

HMGB1 is a chromatin protein with important nuclear functions such as organizing of the DNA structure and regulating transcription. However, upon cell activation or damage, HMGB1 may be released from the nuclei into the cytosol and transported out of the cells. The extranuclear HMGB1 exerts proinflammatory properties. Increased levels of circulating HMGB1 have been observed in RA, and may play a part in its pathogenesis. Extranuclear HMGB1 has also deteriorating effects on the CV system, and may contribute to the development of both atherosclerosis and heart failure [218], [219]. Thus, increased level of extranuclear HMGB1 may represent an important link between RA and CVD, and play an important role in the development of premature CVD in RA [220]. Accordingly, HMGB1 targeted drugs might protect both from CVD and RA, and novel biologic anti-HMGB1 therapies are being developed. Interestingly, inhibition of the negative effects of extracellular HMGB1 may partly underlie the antirheumatic effects of some current RA treatments, including gold injections, glucocorticoids and methotrexate [219]. There is a need to examine the effect of anti-HMGB1 therapies on premature CVD in RA.

Gut microbiota has gained increasing attention recently due to its significant impact on various bodily functions including digestion, metabolism, immune response and activity of hypothalamus–pituitary–adrenal axis (modifying the stress response) [221]. As a result, gut microbiota is likely to play a role in the development of various diseases including CVD as well as autoimmune disease [222], [223], [224]. Therefore, there is a need to examine if alterations of gut microflora in RA patients (due to the disease or treatment) contribute to the development of premature CVD. If this is the case, dietary and other interventions improving the composition of gut flora might be beneficial [225].

Epigenetic mechanisms refer to mechanisms that influence heritable gene expression. The epigenetic change may be induced by various causes including life-style factors and aging. Epigenetic alterations play a role in a variety of diseases, including CVD and autoimmune diseases, and knowledge about the gene–environment interactions paves the way for development of novel therapeutic strategies [226].

Most of the cyclooxygenase (COX)-2 selective (i.e., coxibs) and COX-2 non-selective NSAIDs are associated with an increased risk of adverse CV events, including CAD, hypertension, worsening of heart failure and arrhythmia. Although the CV risk may be increased early during the NSAIDs treatment, it appears to be low with short-time use. For most NSAIDs, the CV risk seems to increase with higher dose and frequency of use, and differs between medications. Notably, while naproxen has been observed to increase CV risk at low doses, high doses appear to be safe. Among coxibs, celecoxib seems to have a relatively low CV toxicity, similar to that of non-selective NSAIDs [58].

Although NSAIDs in general should be used cautiously to minimize risk, the real magnitude of their CV toxicity in RA is still unclear, and might be modest [58], [227]. A study of polyarthritis patients has even shown a protective effect of NSAIDs (ibuprofen, naproxen, diclofenac) on all-cause and CV mortality [228]. In a Danish registry, the occurrence of major CV events attributable to NSAIDs was significantly lower in RA than non-RA patients, and, except for diclofenac and rofecoxib, the NSAIDs did not seem to increase CV risk in RA [227]. Furthermore, an increased risk for myocardial infarction has been observed for several weeks after the cessation of NSAIDs in RA [98]. Thus, further research is needed to clarify if abrupt NSAIDs cessation in RA might have negative effects on CV risk.

In general, the anti-platelet effects of non-selective NSAIDs appear to be too weak to allow these drugs to replace aspirin or other antiplatelet drugs in CVP.

Although paracetamol and opioids have been suggested as safer alternatives for pain relief in RA, also these drugs are associated with a significant toxicity including adverse CV effects [229].

Thus, the CV risk profile of the individual anti-inflammatory and analgesic drugs in RA needs further clarification, and symptom-relieving therapies with a better safety profile should be integrated into the clinical practice.

A variety of other factors have been proposed as contributors to the elevated CV risk in RA (e.g., oxidative stress, accelerated immune aging (in particular the occurrence of senescent auto-aggressive T cells), low levels of atheroprotective anti-phosphorylcholine antibodies, increased number of circulating microparticles, exaggerated complement activation, increased endothelial cell apoptosis, impairment of the endothelial cell glycocalyx, impairment of vasa vasorum and small cardiac vessels, pathogens, increased formation of neutrophilic extracellular traps (NETs), and impairment in microRNAs expression), but their therapeutic implications are so far uncertain [4], [230], [231], [232], [233], [234].