Current EULAR guidelines encourage rheumatologists to assess and coordinate CVD risk management in RA patients [
9]. Yet, despite the increasing knowledge of the high CV risk in RA, the optimal means of minimizing it remain unclear due to scarceness of comparative studies and limited understanding of the precise physiologic effects of these drugs on CV risk. With aims to address this gap in knowledge, The Treatments Against RA and Effect on FDG PET-CT (TARGET trial, NCT02374021) is an ongoing clinical trial that directly evaluates the degree to which reductions in inflammation and disease activity with different therapeutic agents reduce CV risk in RA [
62]. Based on data suggesting a close relationship between lower disease activity and reduced CV risk, current EULAR guidelines recommend aggressive control of RA disease activity in order to mitigate both joint damage and CV risk with effective DMARD use [
9,
23]. Current guidelines prioritize disease control over the particular choice of therapy. While data remain limited, available data suggest a differential impact of nonsteroidal anti-inflammatory drugs (NSAIDs), glucocorticoids, non-biologic DMARDs, biologics, and small-molecule-based therapy, on CV risk [
63‐
67] (Table
1). Larger studies with longer observation periods are required.
Table 1
Select studies that illustrate the relationship between particular therapeutic agents and CV risk in RA
Nonbiologic DMARDs (HCQ, SSZ, MTX) | | 23,994 | Prospective cohort | 20% reduction in CV events (stroke, MI, or congestive heart failure) in the setting of recent continuous MTX, either in combination or as monotherapy |
TNF inhibitors | | 6864 | Prospective cohort | 47 ACS events occurred in 1 year. A 50% lower ACS risk was seen in TNF responders compared with non-responders |
Abatacept (vs. TNF inhibitors) | | 13,036 | Retrospective cohort | Abatacept was associated with an approximately 20% greater reduction in CV risk compared with TNF inhibitors |
Tocilizumab (vs. TNF inhibitor) | | 3080 | RCT | No significant difference in the risk of MACE between treatment groups |
Sarilumab | | 3358 | Pooled cohort | Exposure-adjusted incidence of MACE with sarilumab combination and monotherapy was no greater than that seen in the general RA population |
Anakinra | | 23 | RCT | Improved vascular and left ventricular function in RA patients treated with anakinra, particularly those with prior documented CAD |
Rituximab | Van Vollenhoven et al. [ 83] | 2578 | Pooled cohort | Similar rates of MI (0.41 per 100 person-years) in RA patients treat with rituximab compared to those treated with methotrexate and placebo |
JAK inhibitors | | 3492 | Prospective cohort | No association between baricitinib treatment and the incidence of MACE, arterial thrombotic events, or congestive heart failure |
NSAIDs and Glucocorticoids
Glucocorticoids and non-steroidal anti-inflammatory drugs (NSAIDs) are frequently utilized for pain control during episodes of acute flares. Despite the beneficial anti-inflammatory effects, the myriad of potential side effects due to these two medication classes are well known to providers. The precise CV risk imparted by NSAIDs and glucocorticoids, however, is a more nuanced question. In a prospective cohort study by Rincon et al. [
68], 779 patients RA patients with a total of 7203 person years, exposure to glucocorticoids was found to be associated with a dose-dependent increase in death from all causes with a HR of 1.07 per mg of prednisone per day (95% CI 1.05–1.08). Of the 237 patients who died during follow-up, 120 deaths were due to CV causes, yielding a CV mortality rate of 1.8 (95% CI 1.5–2.1). In a systemic review and meta-analysis by Roubille et al. [
69] including 28 studies specifically in patients with RA, corticosteroids were found to increase the risk of all CV events (RR 1.47, 95% CI 1.34–1.60;
p < 0.001). Similarly, NSAIDs increased the risk of all CV events (RR 1.47; 95% CI 1.01–1.38,
p = 0.04), though this effect size may be overestimated due to inclusion of studies specifically pertaining to relecoxib, which is now removed from the market due to increased risk of CV events [
69].
Non-Biologic DMARDs
Conventional DMARDs, such as methotrexate (MTX), sulfasalazine (SSz), and hydroxychloroquine (HCQ) have been shown to improve CV risk [
70]. In a Canadian population-based inception cohort including 23,994 RA patients diagnosed after age 75, Widdifield et al. [
70] observed an approximately 20% reduction in CV events (stroke, MI, or congestive heart failure) in the setting of recent continuous MTX, either in combination or as monotherapy (hazard ratio (HR) 0.79 for continuous use vs. no use in past 12 months, 95% CI 0.70–0.88;
p < 0.0001). Similarly, a large meta-analysis of ten smaller RA cohort studies illustrated an approximately 21% decrease in CVD-related events, including MI, stroke, and death, in the setting of MTX therapy [
71]. While a precise mechanism for this effect is unclear, the beneficial effects of MTX and other conventional DMARDs are likely driven by the amelioration of chronic inflammation. Triple therapy with MTX, SSz, and HCQ has also been associated with a decrease in CV risk, by decreasing inflammation and improving lipid profiles [
70,
72]. Furthermore, in the Treatment of Early Aggressive Rheumatoid Arthritis (TEAR) trial, increases in HDL, decreases in LDL, and an improved ratio of total cholesterol to HDL were noted in those receiving triple therapy compared with patients receiving MTX monotherapy or MTX in combination with etanercept [
72].
Biologic DMARDs
TNF inhibitors have a positive impact on surrogate markers of cardiovascular disease, including improvement in cIMT, FMD, and reduction in circulating levels of CRP and IL-6 [
73,
74]. A systematic review by Barnabe et al. [
75] including 16 observational RA cohort studies showed that anti-TNF therapy was associated with a reduced risk for all CV events (pooled adjusted RR 0.46; 95% CI 0.28–0.77), MI (pooled adjusted RR 0.81; 95% CI 0.68–0.96), and cerebrovascular accidents (pooled adjusted RR 0.69; 95% CI 0.53–0.89). More recently, this association was similarly investigated by Ljung et al. [
76] in a large prospective cohort study that included 6864 RA patients initiating TNF inhibitors. With 47 acute coronary syndrome (ACS) events in the group, a 50% lower ACS risk was seen in responders (defined as a significant decrease in disease activity score [DAS] or DAS28 of ≥ 1.2, or low disease activity score: DAS ≤ 2.4 or DAS28 ≤ 3.2) compared with non-responders. Although the relatively small number of events during the 1-year study period is a limiting factor, those with a moderate response (defined as a significant change in DAS with moderate/high DAS > 3.7 or DAS28 > 5.1
or patients with a change ≤ 1.2 and > 0.6 with low/moderate disease activity), had equal risk to non-responders, implying that optimal disease control is needed to have an effect on CV events. Yet, larger studies with a longer observation period are required to adequately evaluate this clinical question.
The effects of non-TNF biologics on CVD risk, such as abatacept (a fusion protein consisting of the extracellular domain of human CTLA-4 and a modified Fc portion of human IgG1), tocilizumab and sarilumab (humanized anti-IL6 receptor monoclonal antibodies), anakinra (a recombinant IL-1 agent), and rituximab (an anti-CD20 monoclonal antibody), have also been explored, though data remains limited for most of these drugs. The cardiovascular benefits of abatacept were evaluated in comparison with TNF inhibitors in 6102 matched pairs of abatacept and TNF initiators from Medicare, as well as 6934 matched pairs from MarketScan [
77]. Among these patients, 35% and 14% of the Medicare and MarketScan subjects, respectively, had baseline CVD. After accounting for this baseline risk, abatacept was associated with an approximately 20% greater reduction in CV risk compared with TNF inhibitors. In regards to tocilizumab, despite initial concerns for worsened CV outcomes due to increases in total cholesterol levels, long-term follow-up studies show that rates of stroke and MI after tocilizumab treatment (mean duration of 2.4 years) are comparable to that of RA patients on MTX alone or in conjunction with placebo [
78,
79]. The MEASURE study by McInnes et al. [
80] showed that a possible explanation for this paradox was tocilizumab’s ability to alter HDL particles towards an anti-inflammatory composition (decreased serum amyloid A, phospholipase A2, lipoprotein A, fibrinogen, and D-dimer) that may ameliorate associated CV risk. Comparatively, limited data are available on the association between sarilumab and CV risk. Though a recent study by Fleischmann et al. [
81] showed that exposure-adjusted incidences of major adverse cardiac events with sarilumab combination (0.5 per 100 patient-years) and monotherapy (0.2 per 100 patient-years) were no greater than that seen in the general RA population (1.4 per 100 patient-years without exposure to DMARDs, 1.1 with exposure to DMARDs, and 1.2 overall). In addition, a small, double-blind, crossover, placebo-controlled study showed improved vascular and left ventricular function in RA patients treated with anakinra, particularly those with prior documented CAD [
82]. Finally, a study assessing the long-term safety of rituximab in 2578 RA patients showed similar rates of MI (0.41 per 100 person-years) to those seen in RA patients treat with methotrexate and placebo [
83]. Interestingly, in the general non-RA population, the results of the CANTOS study suggest that targeting the IL-1 pathway with canakinumab 150 mg every 3 months led to a significantly lower rate of MI compared to placebo, independent decreases in lipid levels (HR 0.85, 95% CI 0.74–0.98;
p = 0.021) [
84].
The comparative effect of different classes of biologic DMARDs has not been optimally studied by direct head-to-head assessment, yet there are data to suggest a differential impact of specific therapies on CV risk. In a retrospective review of 47,193 Medicare RA patients without CAD at the time of initiation of a biologic therapy, the incidence of acute MI was significantly elevated among anti-TNF initiators (adjusted HR 1.3; 95% CI 1.0–1.6) compared with those initiated on abatacept [
64]. Interestingly, tocilizumab initiators had a reduced risk of the composite outcome (acute MI and/or need for coronary revascularization) compared with those initiated on abatacept (adjusted HR 0.64, 95% CI 0.41–0.99) [
64]. Furthermore, a recent head-to-head randomized controlled trial (RCT), ENTRACTE, comparing the cardiovascular safety of tocilizumab and etanercept in 3080 RA patients followed for a mean duration of 4.9 years showed no significant difference in the risk of major adverse CV events between treatment groups (HR 1.05, 95% CI 0.77–1.43) [
84]. To reconcile these findings, Singh et al. [
66] performed a systematic review and meta-analysis of 14 cohort studies evaluating the risk of CV events in RA patients treated with conventional DMARDs, TNF inhibitors, or non-TNF biologics. Upon review, they noted tocilizumab to be associated with a lower risk of major adverse cardiac events (MACE) compared with TNF inhibitors; with no difference in such risk seen when comparing tocilizumab with abatacept. While adjustment for RA disease activity, baseline CV risk factors, and methodological differences between studies were accounted for, a key weakness was that with the exception of ENTRACTE, the analyzed studies were observational and not RCTs [
66].
Kinase Inhibitors
Data remain relatively scant regarding the effects of newer agents such as the small molecule inhibitors of Janus kinase (JAK), tofacitinib and baricitinib, on CV risk. A 10–20% increase in total and LDL cholesterol levels, similar to that seen with tocilizumab therapy, has been noted in the tofacitinib RCTs, raising concerns about worsened CV outcomes [
85,
86]. Evaluation of the effects of 24 weeks of tofacitinib treatment showed that increases in HDL cholesterol levels and decreases in the total cholesterol to HDL cholesterol ratio were associated with a diminished risk for future MACE, whereas increases in total cholesterol and LDL cholesterol levels lacked this protective effect [
87,
88]. The occurrence of venous thromboembolisms with JAK inhibition, however, has further accentuated initial concerns, and a formal comparison of the effects of tofacitinib on the risk of major adverse cardiac events is being investigated in a phase IIIb/IV prospective comparative study versus TNF inhibitors (NCT02092467) [
88,
89]. In addition, recent data from a pooled cohort of 3492 RA patients with over 7860 patient-years of exposure to baricitinib showed no association between baricitinib treatment and the incidence of MACE (incidence rates (IR) per 100 patient-years, placebo vs. 4 mg baricitinib = 0.5 vs. 0.8), arterial thrombotic events (0.5 vs. 0.5), or congestive heart failure (4.3 vs. 2.4). In regards to deep vein thrombosis (DVT) or pulmonary embolism (PE) risk, six events occurred in patients treated with 4 mg baricitinib, with no cases seen in the placebo group; though all six cases were in patients who had pre-existing risk factors for venous thromboembolism. In an extended 2 mg vs. 4 mg baricitinib analysis, IRs of DVT/PE were comparable between the doses (IR per 100 patient-years in the 2 mg vs. 4 mg baricitinib doses = 0.5 vs. 0.6) [
90].
Ultimately, given the myriad of therapeutic options and a movement towards precision medicine, it is becoming increasingly important for rheumatologists to incorporate these data in the context of an individual patient’s unique traditional CV risk factors and intrinsic RA features to generate treatment plans that optimally mitigate not only the articular manifestations of RA but also the excessive cardiovascular events and overall mortality. Clinical trials and prospective studies comparing the relative impact of different DMARDs on CVD risk in RA are currently ongoing, and will further shed light on the optimal DMARD choice in specific subsets of RA patients using a precision medicine approach.