Introduction
Although there is a large body of knowledge on the increased cardiovascular disease (CVD) risk in rheumatoid arthritis (RA), there is a lack of clinical evidence on management of the increased risk.
Current treatment recommendations
Several treatment guidelines exist regarding prevention of CVD for the general population, including both national guidelines and guidelines for larger geographical areas such as Europe and US. In the latest European Society of Cardiology (ESC) guidelines for CVD prevention, the presence of RA was mentioned as a high CVD risk factor for the first time [1]. However, an RA diagnosis itself should not be considered as an indication for initiation of cardio-protective treatment, but is recommended to be taken into consideration when evaluating the total CVD risk for each patient.
European League Against Rheumatism recommendations
A task force of the European League Against Rheumatism (EULAR) published recommendations in 2010 regarding management of CVD risk in patients with RA and other inflammatory joint diseases [2]. These recommendations suggested that in RA patients fulfilling at least two out of three of specified criteria (disease duration >10 years, rheumatoid factor/anticitrullinated protein antibody positivity, or extra-articular manifestations), the calculated CVD risk should be multiplied by 1.5. This multiplication factor was derived mainly from relevant standardized mortality ratios because information from large-scale prospective cohort studies was, and still is, lacking. However, Corrales et al have shown that the discussed multiplication factor only reclassifies 3% of the patients into more appropriate CVD risk classes, but even then patients at high CVD risk and those with asymptomatic carotid atherosclerosis were not adequately identified. Hence, the 1.5 multiplication factor does not fully account for the increased risk of CVD in RA [3].
The EULAR recommendations were updated in 2015, and include several new recommendations [4], which were presented at the annual EULAR conference in 2015, and will be published in 2016. A CVD risk assessment is currently recommended to be performed at least every five years, as no evidence supports that yearly screening reduces CVD morbidity and mortality in RA patients. National guidelines developed for the general population should be followed with respect to CVD risk estimation. If no national guidelines are available, use of the Systematic COronary Risk Evaluation (SCORE) CVD risk calculator is recommended. The total cholesterol (TC) and high-density lipoprotein cholesterol (HDL-c) is preferred over use of only TC levels to estimate future CVD risk in RA patients. It was decided that the 1.5 multiplication factor should still be recommended to be used during CVD risk evaluation, but now for all RA patients. A new recommendation includes that screening for asymptomatic atherosclerotic plaques by use of carotid ultrasound may be considered as part of the CVD risk evaluation in patients with RA. The importance of a healthy diet, regular exercise, and smoking cessation was emphasized. Furthermore, statins and antihypertensive treatment may be used as in the general population. Regarding anti-rheumatic treatment, optimal control of disease activity is advocated in order to reduce systemic inflammation. However, prescription of non-steroidal anti-inflammatory drugs (NSAIDs) and corticosteroids should be done with caution.
Pharmacological cardio-protective treatment in patients with rheumatoid arthritis
Anti-hypertensive treatment
Hypertension (HT) is a modifiable risk factor contributing to the increased CVD risk in patients with RA [5]. Several mechanisms may lead to the development of HT including use of certain anti-rheumatic drugs, such as corticosteroids [6], NSAIDs [7], cyclosporine [8], and leflunomide [9]. HT is associated with the presence of subclinical atherosclerosis [10], and higher BP levels have also been shown to be the most influential risk factor for arterial stiffening in RA patients [11]. For the management of HT, there is no evidence that preferred drugs or treatment thresholds should differ in patients with RA compared to the general population. Regarding the choice of antihypertensive drugs, evidence supports that the benefit of the treatment seems to be related to the lowering of the BP, and not to the specific drug used to reach this target. The 2013 ESH/ESC Guidelines for the management of arterial hypertension recommend diuretics, beta-blockers, calcium antagonists, angiotensin-converting enzyme inhibitors, and angiotensin receptor blockers as suitable options for first-line treatment of HT [12]. These medications may be given as monotherapy or in combination. However, the different antihypertensive drug classes vary in the adverse effect profile, and the drug of choice and BP treatment targets will depend on the specific condition (Table 5.1 and 5.2).
Lipid lowering medication
Statins are well established as the drug of choice to decrease cholesterol levels [1,13]. Hydroxymethylglutaryl-coenzyme A (HMG-CoA) is the precursor for cholesterol synthesis in the liver. Statins are HMG-CoA reductase inhibitors, resulting in reduced production of cholesterol, which leads to an upregulation of low-density lipoprotein cholesterol (LDL-c) receptors on the cell surface of the hepatocytes, with the consequence of increased LDL-c extraction from the blood. It is beyond all reasonable doubt that statins have a cardio-protective effect, both in primary and secondary prevention of CVD [14-16]. Furthermore, statins are generally well tolerated. The risk of myopathy is low (1/1000), but myopathy may in rare cases lead to rhabdomyolysis and renal failure. If the patients experience myalgia with no increase in creatine kinase (CK; occurs in 5–10% of treated patients), statin treatment can be continued if the muscle pain is tolerable for the patient. An elevation of CK <5 times upper limit of normal (ULN) in two blood samples is considered acceptable. Elevated liver enzymes (alanine aminotransferase and aspartate aminotransferase) occur in 0.5–2.0% of patients treated with statins, and this adverse event is dose dependent. If the increase in liver enzymes exceeds three times ULN it is recommended to discontinue the statin treatment. A small increase in incident Type 2 diabetes mellitus in statin-treated patients has been reported [17], but the CVD benefit has been shown to surpass the risk of diabetes in patients at high risk of CVD [18]. The diabetes risk for statin-treated patients were mainly in those with impaired fasting blood glucose and other major risk factors for diabetes in the Justification for the Use of statins in primary Prevention: an Intervention Trial Evaluating Rosuvastatin (JUPITER) trial [19]. The metabolism of statins is predominantly in the liver via the cytochrome P450 system (except pravastatin, rosuvastatin, and pitavastatin). Thus, statins interact with other drugs that are metabolized through the same system. Other contraindications for statin treatment are liver disease and myopathy. Combined statin-fibrate (especially gemfibrozil) therapy is shown to increase the risk for rhabdomyolysis, and is not recommended [20]. There has been some uncertainties related to cancer and statin use, but Alsheikh-Ali et al did not find a relation of statin use and development of cancer in a meta-analyses of 15 large randomized statin trials including 437,017 person-years and 5752 cases of cancer [21]. The relation of statins and cancer development in patients with RA needs further investigation.
The level of LDL-c has been used as a response indicator in almost all trials investigating the effect of lipid-lowering therapy. A 1.0 mmol/L reduction in LDL-c is associated with a 22% reduction in CVD morbidity and mortality [14]. Therefore, LDL-c is considered the major target for lipid-lowering therapy.
It has been reported that lipids are less frequently tested in patients with RA than in the general population [22]. The interaction between lipid values and inflammation [23] may indicate a need for a tighter follow-up or frequent lipid monitoring in patients with RA, although this is still an area of uncertainty. Soubrier and colleagues evaluated the prevalence of patients with RA in whom lowering LDL-c should be considered in accordance with the Adult Treatment Panel III (ATPIII) guidelines, and concluded that lipid lowering therapy was insufficiently prescribed in this patient group [24]. A significant underuse of statins for primary prevention purposes in RA patients has been reported [22,25], and the degree of under-treatment was dependent on the risk stratification method. Even after suffering a myocardial infarction, patients with RA seem to receive less CVD preventive treatment than patients without RA [26]. In addition, statin discontinuation in patients with RA has also been associated with an increased risk of death from CVD and all-cause mortality, and provides support for the importance of statin compliance in RA patients where such medication is indicated [27].
Statins are reported to have beneficial effects on the joint disease in addition to the lipid-lowering effect and reduction of CVD events. A placebo-controlled study with atorvastatin versus placebo in 116 patients with active RA revealed a significant reduction of inflammatory biomarkers, such as C-reactive protein (CRP) and erythrocyte sedimentation rate (ESR), and also a decrease in swollen joint count [28]. Reduction in RA disease activity was further confirmed in another study including 30 patients with early RA randomized to placebo or atorvastatin 40 mg daily [29]. Furthermore, statins may have favourable effects on endothelial function [29, 30], arterial stiffness [31], and on modification of HDL-c properties in patients with RA [32]. A cost-effectiveness analysis concluded that the dual effect of statins, both lipid-lowering and anti-inflammatory effects, makes this therapy cost effective in RA patients [33].
The first primary prevention trial, a randomized, placebo-controlled statin study with CVD outcome for patients with RA without diagnosed CVD, was recently presented at the American College of Rheumatology [34]. In the TRial of Atorvastatin for the primary prevention of Cardiovascular Events in patients with Rheumatoid Arthritis (TRACE-RA), 2986 RA patients from 106 centres in the United Kingdom were randomized to atorvastatin 40 mg daily or placebo. However, due to a lower event rate than anticipated, the trial was prematurely terminated. LDL-c was significantly reduced in the atorvastatin arm compared with placebo. A 34% risk reduction for a major CVD event compared with placebo was found but this difference did not reach statistical significance, possibly due to the early termination of the trial and corresponding low number of patients and CVD events. Regarding safety, there were no differences in adverse events between the atorvastatin (19.7%) and the placebo (19.5%) group (p=0.93). Comparable results have been reported from a post hoc analysis of two large statin trials (the Incremental Decrease in Endpoints Through Aggressive Lipid lowering [IDEAL] and the Treating to New Targets [TNT] studies) in patients with coronary heart disease, which also revealed that patients with and without inflammatory joint diseases had comparable lipid-lowering effect and reduction of CVD morbidity and mortality [35].
Experiences regarding lipid lowering treatment in patients with RA have been reported from a preventive cardio-rheuma clinic. Patients with inflammatory joint diseases referred for a CVD risk evaluation from a rheumatology outpatient clinic and general practitioners in the time period 2009–12 were assessed to reveal the proportion of patients in need of CVD preventive measures [36]. Furthermore, the effects of lipid-lowering therapy were evaluated with regard to achievement of guideline-recommended lipid goals. CVD risk stratification was performed at the first consultation, and all patients received advice about physical activity and cholesterol friendly diet. Smokers were offered the opportunity to join smoking cessation programmes and diabetics were referred to a specialist clinic for optimization of glucose control if needed. Hypertensive patients were treated with antihypertensive medication aiming at a BP goal <140/90 mmHg. For patients without documented CVD (CVD event or asymptomatic atherosclerosis) the SCORE calculator was used to distinguish between patients in need of primary prevention and those with no indication for CVD preventive measures. After initiation of statins (atorvastatin, simvastatin, rosuvastatin, or pravastatin), the patients were followed until at least two lipid targets were reached. Of the 426 patients referred to the Preventive Cardio-Rheuma clinic, 63.4 % were in need of CVD preventive treatment, and secondary prevention was indicated in 77% of these patients. Achievement of at least two lipid targets was successful in approximately 90% of all the treated patients, using on average less than three consultations to obtain these goals (Figure 5.1).
This high overall lipid goal attainment in patients with inflammatory joint diseases is encouraging, but whether it will sustain over time is uncertain, considering the relatively low goal attainment of slightly more than 40% reported in the general population [37]. Despite the presence of CVD, asymptomatic carotid atherosclerosis, or a calculated CVD risk by SCORE ≥5%, a high proportion of inflammatory joint disease patients did not receive recommended cardio-protective treatment at referral date to the Preventive Cardio-Rheuma clinic. Ultrasound of the carotid arteries revealed carotid plaques in approximately 50% of the patients, which resulted in a correct stratification to indication for statin treatment with secondary CVD prevention lipid goals. Our results highlight the importance of performing carotid ultrasound during CVD risk evaluation in patients with inflammatory joint diseases.
In the general population there exists a linear relationship between lipid levels and risk of CVD [38]. Patients with inflammatory joint diseases have lower cholesterol levels compared with persons without inflammatory joint disease [35, 39, 40]. The complex interaction between inflammation and lipid levels has been investigated thoroughly [23,41–43]. Inflammation increases the risk of CVD in RA patients [44] but the corresponding low lipid levels may camouflage the actual risk for the executive physician. Patients treated with statins at the Preventive Cardio-Rheuma clinic, who obtained the guideline recommended LDL-c goal, were evaluated to reveal if baseline systemic inflammation (measured by CRP and ESR) and lipid levels were of importance regarding statin dose needed to obtain LDL-c targets [45]. Intensive statin dose was defined as rosuvastatin ≥20 mg and atorvastatin and simvastatin at the highest dose of 80 mg, and conventional lipid-lowering treatment was defined as all lower doses in accordance with the drug efficacy across doses obtained by the various statins in the STELLAR (the Statin Therapies for Elevated Lipid Levels compared across doses to Rosuvastatin) trial [46]. Change or up-titration of statins was done in cases of adverse events or failure to obtain LDL-c targets. Systemic inflammation or lipid levels at baseline were not associated with statin dose needed to achieve lipid targets (Figure 5.2).
Furthermore, there was no significant impact of anti-rheumatic medication (biologic and synthetic disease modifying anti-rheumatic drugs [DMARDs], prednisolone, and NSAIDs) on the relation between baseline lipid levels or systemic inflammation on doses of statins needed to achieve LDL-c targets. In a population-based cohort from US, patients with RA were less likely to obtain LDL-c goals in comparison with non-RA subjects, and that was related to an increased ESR at baseline [47]. Firstly, the relation between inflammatory parameters and lack of LDL-c goal attainment may reflect an increased RA disease activity resulting in lower compliance regarding both medications (statins) and primary care physician follow-up. Secondly, the low LDL-c goal attainment may reflect that this was a population-based cohort, where the patients were probably mostly attending primary care, and not results from a cardiology specialist practice as our Preventive Cardio-Rheuma clinic is. In the last case, patients attending such a specialist clinic might be more motivated and hence compliance will be better. In addition, a tight control regime was applied, which may not be possible in primary care. Thirdly, a limitation to the population-based report is that data regarding statin dose used in RA patients and non-RA persons are lacking. Interestingly, systemic inflammation at baseline was comparable for patients who did and did not obtain LDL-c targets in inflammatory joint disease patients attending the Preventive Cardio-Rheuma clinic. Thus, the mechanisms associated with individual statin response in the general population may also be applicable for patients with inflammatory joint diseases [48]. The data discussed suggests that systemic inflammation may be of limited value when developing individual lipid lowering CVD preventive strategies for patients with RA.
Since effects of prospective randomized control trials with statins/placebo on longitudinal CVD outcome are scarce, studies with surrogate CVD endpoints are of interest. In the ROsuvastatin in Rheumatoid Arthritis, Ankylosing Spondylitis and other inflammatory joint diseases (RORA-AS) study, we examined the development of carotid atheroma during 18 months of intensive statin treatment with rosuvastatin [49]. The aims were to evaluate change in carotid plaque height, and whether laboratory values or clinical parameters were predictors of the potential change in carotid plaques after 18 months of intensive lipid lowering therapy. The RORA-AS study was a prospective, open intervention study where 86 statin-naive patients with inflammatory joint diseases (RA: n=55, ankylosing spondylitis (AS): n=21, and psoriatic arthritis (PsA): n=10) who had carotid plaques were treated with rosuvastatin to obtain LDL-c goal ≤1.8 mmol/L. Carotid ultrasound was performed at baseline and after 18 months to evaluate carotid plaque height. Joint disease activity was assessed by the Disease Activity Score using 28 joints (DAS28) [50] and the Ankylosing Spondylitis Disease Activity Score (ASDAS) [51]. In order to address bone damage due to accumulated disease activity, digital X-rays of hands and feet were conducted and scored in accordance with the Sharp/van der Heijde method [52]. Compliance of rosuvastatin in the RORA-AS study was 97.9%. There was a significant reduction in both TC and LDL-c. The levels of the inflammatory biomarkers (CRP/ESR) and the composite disease activity values did not change during the study period. Intensive lipid lowering for 18 months induced carotid plaque height regression in patients with inflammatory joint disease (Figure 5.3).
The degree of atherosclerotic regression was inversely related to disease activity but was not influenced by LDL-c goal achievement, degree of change in LDL-c, or the LDL-c level exposure (area under the curve) during the study period. There was a significant difference in carotid plaque height reduction between patients using and not using biologic DMARDS (bDMARDs), in favor of non-users of bDMARDs. To date, there exist conflicting results regarding the impact of bDMARDs on the risk of future CVD. The analyses in the RORA-AS study were not adjusted for confounding by indication, which is a major concern because patients with the most severe rheumatic joint disease are more likely treated with bDMARDs, and they may also be more likely to have atherosclerotic disease as carotid plaques. However, demographic data, presence of CVD risk factors/co morbidities, disease activity, joint damage measured by the Sharp/van der Heijde method, laboratory values, and medication use at baseline were comparable for users versus non-users of bDMARDs. The increased risk of CVD in patients with RA has been related to joint disease activity, and our results indicate that disease activity may also influence the effect of anti-atherosclerotic treatment. Prospective randomized statin studies with clinical endpoints are warranted to reveal whether height reduction of asymptomatic carotid plaques will have an impact on future CVD events in patients with RA.
For practical purposes, statin initiation should be done as in the general population. A starting dose of statins that moderately lowers the lipids such as atorvastatin 40 mg daily is reasonable, and should be adjusted until lipid goals are achieved. Lipid targets for both primary and secondary prevention are presented in Table 5.3.
Anti-thrombotic treatment
As in the general population, the use of aspirin for the primary prevention of CVD events in patients with RA is not recommended, due to lack of reduction in CVD mortality as well as an increase in bleeding events [53].
The role of anti-rheumatic medications
Raised levels of high-sensitivity CRP have been found to predict CVD events in the general population [54]. Inflammation plays a key role in the atherosclerotic process and further in development of CVD. Large Phase III trials are being conducted in the general population with anti-inflammatory agents such as canakinumab (the CANTOS trial) and methotrexate (MTX; the CIRT trial) to reveal the effect on CVD endpoints [55]. Therapies aimed to reduce disease activity in RA may therefore also have a positive impact on the risk of CVD by reducing the systemic inflammation. On the other hand, adverse effects of diverse anti-rheumatic medication on CVD risk have also been observed.
Synthetic disease modifying anti-rheumatic drugs
The immunosuppressive drug MTX is standard medication as soon as the diagnosis of RA has been made [56]. Systematic literature reviews suggest that the use of MTX in patients with RA is associated with a decrease in the risk of CVD [57, 58]. The reduction in CVD related to MTX use may appear early in the joint disease course [59]. A window of opportunity may therefore also exist regarding prevention of atherosclerosis in addition to suppression of disease activity as a result of early initiation of MTX treatment. However, Greenberg et al did not reveal a cardio-protective effect of MTX in a large register study [60]. The impact of MTX on the lipid profile is still an unresolved question, as all studies addressing this subject [61–63] had a high risk of bias, according to a review performed by Westlake and colleagues [58]. Data on the effect of MTX on risk of CVD are all from observational studies, no randomized controlled trials have to date been performed, and it is not likely that such a study will be ethically justifiable to conduct considering the well-documented effect of MTX on the joint disease.
Biologic disease modifying anti-rheumatic drugs
The safety of tumour necrosis factor inhibitors (TNF-i), interleukin inhibitors, and other biologic agents has been investigated to some extent but controversies exist regarding the impact of bDMARDs on the risk of CVD in patients with RA. The adverse effects of TNF-i on lipid profile and fasting glucose levels are of particular concern [64, 65]. Systematic literature reviews have suggested that treatment with TNF-i is associated with a decreased risk of CVD [66, 67]. It has also been reported that TNF-i influence surrogate markers of CVD, such as aortic stiffness and carotid intima-media thickness [68]. The increase in lipid levels related to use of TNF-i may reflect a normalization of the lipids to the level the patient experienced prior to the RA disease [57]. Therefore, it is believed that these changes are likely to be due to the inflammatory-dampening effect of the drug. In a placebo-controlled study, atorvastatin was shown to significantly reduce tofacitinib-associated (a Janus kinase inhibitor) elevation of TC and LDL-c levels [69]. Responders to TNF-i seem to benefit the most with regard to risk reduction of CVD [70].
Non-steroidal anti-inflammatory drugs
In 2004, the cyclooxygenase-2 selective inhibitor rofecoxib was withdrawn from the market due to increased risk of CVD events associated with the drug [71]. Use of several NSAIDs has been reported to increase the risk of CVD in both observational and randomized trials [72-75]. However, naproxen seems to be the safest NSAID in terms of CVD risk [75]. It is currently advised that NSAIDs should be taken at the lowest effective dose for the shortest possible time considering both the CVD risk and the risk of gastrointestinal bleeding, but strict avoidance of NSAIDs among patients with RA, who also have presence of CVD risk factors, may not always be justified [76]. Indeed, Lindhardsen et al reported that the CVD risk associated with NSAID use in RA patients was significantly lower than in non-RA individuals [77].
Corticosteroids
There exist uncertainties regarding use of corticosteroids in patients with RA and the effect on CVD risk. Adverse effects of steroids have been thought to be related to elevation in lipids. However, some studies have also revealed beneficial effects of corticosteroids regarding cholesterol levels [78, 79]. Corticosteroids cause a mild increase in fasting glucose levels, but development of diabetes is rare in those with a normal glucose tolerance [80]. Studies from diverse diagnosis groups have reported adverse effects of corticosteroids on CVD outcome [81]. The adverse effects of corticosteroids seem to be dose dependent. Wei et al reported that high-dose corticosteroids (>7.5 mg daily) were associated with a threefold increased risk of CVD events [82]. For RA patients receiving a similar dose, the risk of CVD events was approximately five times increased. Davis and colleagues showed that in those exposed to glucocorticoids and being rheumatoid factor-positive, but not rheumatoid factor-negative, were at increased CVD risk [83]. A possible confounding factor may be that rheumatoid factor positivity is associated with more severe joint disease progression as well as extra-articular manifestations. Development of iatrogenic Cushing syndrome due to prolonged steroid treatment may be a marker of increased CVD risk [84]. In summary, whether the risk of CVD associated with use of corticosteroids increases or decreases remains unclear. A low dose of steroids to decrease disease activity in patients with inflammatory joint disease may be beneficial concerning the risk of CVD.
Hydroxychloroquine
The use of hydroxychloroquine has been reported from several studies to have both a favorable glucose- and lipid-lowering effect [85], but the influence on CVD morbidity and mortality remains unclear. Of note, hydroxychloroquine seems to have a cardio-toxic effect in some patients, and has been associated with an increased risk of cardiomyopathy [86].