Accelerated atherosclerosis: An extraarticular feature of rheumatoid arthritis?

Authors


Introduction

It is well recognized that rheumatoid arthritis (RA) causes significant morbidity as a result of synovial inflammation, joint destruction, and associated disability. In addition to these articular manifestations of RA, there is growing recognition of an excess mortality, which is predominantly due to increased coronary artery atherosclerosis. Approximately 50% of atherosclerotic coronary artery disease in the community occurs in the absence of traditional risk factors, such as smoking, hypertension, diabetes mellitus, and hypercholesterolemia (1). Histologic studies and animal models have demonstrated the importance of inflammatory mediators (including activated leukocytes, cytokines, and C-reactive protein [CRP]) within atherosclerotic plaque; further, in large, prospective, epidemiologic studies, elevated serum levels of markers of inflammation (e.g., CRP and serum amyloid A [SAA]) were predictive of future cardiovascular events (1–4). Chronic inflammation may act independently or synergistically with traditional atherosclerotic risk factors in the pathogenesis of atherosclerosis and may also be associated with a hypercoagulable state. RA and other chronic systemic inflammatory diseases may provide insight into these interactions.

Increased mortality in RA

The mortality of patients with RA has been evaluated by many groups, although it is difficult to directly compare these studies due to differences in RA definition, study design, patient population, disease duration, and length of followup. However, investigators in all but 4 studies (5–8) have reported increased mortality rates in RA. The standardized mortality ratios (SMRs) in the major studies performed in the past 50 years (6–26) range from 0.87 to 3.0 and are summarized in Table 1. The mean SMR on pooled analysis of survival in RA is 1.70 (27). Similarly, estimates of survival suggest an average shortening of lifespan of 3–18 years in RA patients (especially those seen in “referral centers”), and the prognosis for RA patients with severely impaired functional status has been compared with that for patients with triple-vessel coronary artery disease or stage 4 Hodgkin's disease (28–30). Increased mortality occurs in both men and women with RA (18) and does not appear to be declining over time, in contrast to the progressive improvement in survival of the general population (27, 31).

Table 1. Standardized mortality ratios in RA studies*
Study, year (reference)nSMR95% CI or PCountryDuration, yearsSetting
  • *

    RA = rheumatoid arthritis; SMR = standardized mortality ratio; 95% CI = 95% confidence interval; NA = not available.

  • P value for observed versus expected mortality.

  • 95% CI and/or P value were not given, but the report states that statistical significance was achieved.

Prior et al, 1984 (13)4483.0<0.001UK11.2Clinical
Symmons et al, 1998 (14)4482.72.4–3.1UK16–22Clinical
Krause et al, 2000 (15)2562.62.05–3.15Germany7.5–15.3Clinical
Allebeck, 1982 (11)1,1652.482.27–2.71Sweden7.5Clinical
Wolfe et al, 1994 (16)3,5012.262.16–2.36US, Canada9–35Clinical
Monson and Hall, 1976 (17)1,0351.85NAUS12–42Clinical
Mutru et al, 1985 (18)1,0001.73<0.0001Finland10Clinical
Wallberg-Jonsson et al, 1997 (19)6061.57<0.001Sweden15Clinical
Mitchell et al, 1986 (20)8051.51Canada12Clinical
Kvalvik et al, 2000 (21)1471.491.15–1.88Norway15Clinical
Reilly et al, 1990 (22)1001.4NAUK25Clinical
Myllykangas-Luosujarvi et al, 1995 (23)1,1861.371.35–1.4FinlandCommunity
Cobb et al, 1953 (9)5831.32NAUS9.6Clinical
Allebeck et al, 1981 (10)2931.32NASweden11Community
Pincus et al, 1984 (24)751.31NAUS9Clinical
Uddin et al, 1970 (25)4751.29<0.01Canada10Clinical
Sokka et al, 1999 (8)1351.280.83–1.89Finland8–14Clinical
Jacobsson et al, 1993 (12)2,9791.281.01–1.62US2–25Community
Lewis et al, 1980 (26)3111.28NAUK11Clinical
Linos et al, 1980 (7)5211.16>0.05US24Community
Lindqvist and Eberhardt, 1999 (6)1830.870.53–1.36Sweden9.8Clinical

Causes of death in RA patients

In the majority of RA mortality studies, investigators have relied on death certificate or medical record data, which may be inaccurate or incomplete (32). In these studies, deaths due to infection (especially respiratory), lymphoproliferative disorders (but not other cancers), and gastrointestinal diseases were more prevalent in RA, but contributed little to the overall increase in SMR. In contrast, excess cardiovascular deaths predominate in most RA mortality studies. In a cohort of 448 RA patients followed for up to 22 years, circulatory disease was responsible for 34% of the excess deaths observed (14). In a study by Wolfe et al, 361 of the 898 deaths in RA (40%) were attributable to cardiovascular disease, with an expected number of 161 (a risk ratio of 2.24) (16). Cerebrovascular disease was the second greatest cause of excess mortality. The majority of the cardiovascular disease observed in RA may therefore result from premature atherosclerosis. This contention is supported by postmortem studies, in which specific cardiac involvement (e.g., pericarditis, coronary artery vasculitis, rheumatoid valvular disease) secondary to RA is rarely seen (33).

Determinants of mortality in RA

Several investigators have attempted to identify factors predictive of increased mortality in RA. In addition to expected risk factors such as age, male sex, education level, smoking, and hypertension, various markers of severity and disease activity in RA are predictive of mortality, including decreased function, higher joint count, use of glucocorticoids, presence of rheumatoid nodules, extraarticular disease, and higher erythrocyte sedimentation rate (ESR) (10, 12, 16, 20, 24, 34–38). In one recent study (38), virtually all of the excess mortality occurred in those patients who had extraarticular disease, while patients with only articular involvement had a survival expectancy similar to that of the general population.

Morbidity due to cardiovascular disease in RA

Excess cardiovascular morbidity is also apparent in RA patients. An increased prevalence of myocardial infarction, congestive heart failure, and stroke was present in RA patients compared with osteoarthritis (OA) patients of similar age, sex, ethnicity, smoking status, blood pressure, weight, and socioeconomic status (39). Banks et al (40) compared 67 RA subjects with 37 OA controls who were matched for all traditional vascular risk factors and for nonsteroidal antiinflammatory drug (NSAID) use. Using a combination of clinical evaluation, electrocardiography, adenosine-stressed myocardial perfusion studies, and coronary angiography, these investigators found that ischemic heart disease was almost twice as prevalent in RA subjects as in OA controls (49% versus 27%; P = 0.03) and was frequently clinically silent (52% versus 20%; P = 0.003). In a prospective study performed over 30 years, Gabriel et al (41) reported a significantly greater incidence of congestive heart failure in 450 RA subjects compared with age- and sex-matched community controls. Incidences of myocardial infarction and peripheral vascular disease were also increased.

Carotid artery intima medial thickness (IMT), measured by ultrasound, is being increasingly employed as a marker of early atherosclerosis and vascular risk. In a prospective study of 4,476 older adults free of cardiovascular disease, over a median duration of 6.2 years, increased carotid artery IMT was strongly predictive of myocardial infarction and stroke (42). In this study, an increase of 1 SD in IMT was associated with a relative risk of 1.36 for the combined end point of myocardial infarction and stroke after adjustment for age, sex, and other vascular risk factors. Several groups have documented increased IMT in RA patients compared with controls (43, 44), suggesting a greater prevalence of subclinical atherosclerosis.

Traditional cardiovascular risk factors in RA

Given that the excess mortality in RA appears to be largely due to coronary artery disease, the role of traditional vascular risk factors is of interest. These risk factors include smoking, lipids, and homocysteine.

Smoking. Smoking is a dose-dependent risk factor for atherosclerosis. Coronary artery disease in RA patients could be explained if smoking was more prevalent in this population, or if smoking was also a risk factor for the development of RA. Postulated mechanisms for the latter include the induction of rheumatoid factor (RF) by cigarette smoking (45) and an antiestrogenic effect of smoking (46). Table 2 summarizes studies that have evaluated the relationship between smoking and RA. Most are case–control studies or cross-sectional surveys and hence do not prove causation. The findings in one study were suggestive of a protective role of smoking (47); however, this was a limited cross-sectional survey, and a dose-dependent protective effect of smoking was not seen.

Table 2. Smoking and the risk of RA*
ReferenceType of studySubjectsResultsComments
  • *

    NS = not significant (see Table 1 for other definitions).

  • Presented as risk of RA (95% CI), where OR = odds ratio, adj. = adjusted, and RR = relative risk.

47Case–control135 female RA patients, 378 controlsOR (adj.) current smokers 0.61 (0.42–0.89), OR (adj.) current smokers of ≥21 cigarettes/day 0.82 (0.44–1.50)
48Prospective cohort116,779 nurses, 115 incident cases of RARR (adj.) current smokers 1.3 (0.9–2.1), RR (adj.) previous smokers 1.5 (0.9–2.3)Case ascertainment based on data from self-administered questionnaire
49Prospective cohort52,809 participants, 512 incident cases of RAAssociation between smoking and seropositive RA in men: RR (age adj.) ex-smokers 2.8 (1.4–5.6), RR (age adj.) current smokers of pipes/cigars 4.8 (2.0–11.8), RR (age adj.) current smokers of 1–14 cigarettes/day 4.4 (2.3–8.5), RR (age adj.) current smokers of ≥15 cigarettes/day 4.0 (2.1–7.7); no association for seropositive women, or for seronegative men or womenCase ascertainment based on enrollment in antirheumatic drug registry
50Case–control349 female RA patients, 1,457 controlsOR current and former smokers overall NS, OR smoking for ≥20 pack-years 1.5 (1.0–2.0)
51Twin79 monozygotic pairs, 71 dizygotic pairsOR ever smoked 3.9 (1.6–10.5), OR current smoker 3.7 (1.6–10.1)
52Case–control165 patients (90 with true RA), 165 controlsOR (adj.) current smoker 0.95 (0.56–1.61), OR (adj.) ever smoked 1.66 (0.95–3.06); smokers had higher risk of developing seropositive than seronegative arthritis
53Two population surveys361 RA patients, 5,851 controlsOR current smoker 1.46 (1.1–1.94), OR male current smoker 2.38 (1.45–3.92), OR seropositive male current smoker 4.77 (2.1–10.9)Exposure data consisted of single question: “Are you smoking daily?”
46Retrospective cohort377,481 subjects in cohort, 7,697 with reported RA, of whom 3,416 were seropositiveRR (adj.) past smoker 1.01 (0.95–1.08), RR (adj.) current smoker 1.22 (1.16–1.28), RR (adj.) current smokers of >25 cigarettes/day 1.32 (1.19–1.46); smoking associated with seropositivityExposure data obtained retrospectively; case ascertainment based on self-report of RA
54Case–control239 RA patients, 239 controlsOR ever smoked 1.81 (1.22–2.19), OR smoking for 41–50 pack-years 13.54 (2.89–63.38)

Given that more severe and/or active disease seems to predict greater cardiovascular mortality in RA, smoking could act by exacerbating RA. Two cross-sectional surveys have examined the relationship between smoking and various measures of disease activity and severity (55, 56). Investigators who conducted both surveys reported positive associations between smoking and seropositivity, radiographic erosions, and the presence of rheumatoid nodules. Measures of disease activity, such as ESR and joint count, were not altered by smoking. A 3-year prospective cohort study of 486 patients with early inflammatory polyarthritis (67% of whom had RA) showed that current smokers were more likely to be seropositive and to develop nodules (57). There was no relationship between smoking and the development of erosions or functional disability; however, smokers had significantly fewer swollen joints.

Overall, the data suggest that smoking may be a risk factor for the development of RA (particularly seropositive disease) and that it may be associated with more severe disease. However, it is unlikely that the excess cardiovascular disease in RA can be wholly attributed to smoking. Perhaps the presence of chronic inflammation enhances the atherogenic effect of cigarette smoking, and the two factors act in synergy to increase cardiovascular risk in RA patients. To date, no studies have specifically addressed whether the risk of cardiovascular disease for smokers with RA is greater than that for smokers in the general population.

Lipids. Dyslipidemia is an important risk factor for atherosclerotic coronary heart disease in the general population. In particular, decreased levels of high-density lipoprotein (HDL) cholesterol, elevated levels of low-density lipoprotein (LDL) cholesterol, and/or an elevated total cholesterol level or an elevated LDL:HDL cholesterol ratio predict increased coronary risk (58). Several studies have examined serum levels of lipids in RA patients compared with controls. Total and LDL cholesterol levels were elevated in some studies (59) and reduced in others (60, 61). A more consistent finding has been decreased levels of HDL cholesterol in active or untreated RA in both male and female subjects (59, 60, 62), which is an unfavorable profile with regard to cardiovascular risk (63). Lipoprotein(a) (Lp[a]) is a cholesterol-rich modified form of LDL which has recently been identified as an independent risk factor for coronary heart disease (64, 65). While Lp(a) levels are thought to be primarily genetically determined, Lp(a) shares structural similarity with plasminogen and may be an acute-phase reactant (65). Elevated levels of Lp(a) have been demonstrated in RA patients with both active and treated disease (62, 63, 66, 67).

The impact of drug treatment on lipid profiles in RA has been examined. Disease control with NSAIDs in combination with either prednisolone or gold salts, over the course of 9 months of treatment, increased previously depressed levels of total, HDL, and LDL cholesterol (60). Treatment of 33 patients with active, chronic, inflammatory arthritis (28 of whom had RA) with various agents (mainly prednisolone and either azathioprine or cyclophosphamide) was associated with elevation of the previously decreased total, HDL, and LDL cholesterol levels in parallel with improvement of disease activity (68). In contrast, in a subgroup of 6 patients treated with chloroquine, total cholesterol and triglyceride levels were decreased. Reduction of total and LDL cholesterol levels with antimalarial therapy in both RA and lupus patients (63, 69–71) and also in patients with diabetes (72) has been reported by several groups. The mechanism is not known, but may relate to decreased hepatic cholesterol synthesis or up-regulation of LDL receptors (73, 74). In 100 patients with active RA who were randomly assigned to treatment with either hydroxychloroquine or parenteral gold and followed up for 1 year, treatment with gold resulted in a 12% decrease in HDL cholesterol level and a 31% increase in triglyceride levels (74). In contrast, the hydroxychloroquine group had a 15% increase in HDL cholesterol level and no change in triglyceride levels. Total cholesterol levels remained the same in both groups.

In summary, it appears likely that active inflammation due to RA lowers HDL and possibly total cholesterol levels, and that improved disease control may increase total cholesterol levels. Measurement of serum levels of cholesterol in RA patients during periods of active inflammation may therefore not reflect basal levels. Antimalarial agents in particular may have beneficial effects on lipid profiles, with lowering of total and LDL cholesterol levels and elevation of HDL cholesterol level (a favorable combination), and this should be considered when treating RA patients with vascular risk factors.

Homocysteine. Homocysteine is an intermediary amino acid formed during the conversion of methionine to cysteine. Severe elevations in plasma and urine levels of homocysteine occur in rare autosomal-recessive genetic defects of the enzymes cystathionine β-synthase and methylene tetrahydrofolate reductase (MTHFR). Less marked elevations of homocysteine levels are more common and can occur as a result of heterozygous mutations in metabolizing enzymes, dietary deficiency of vitamin B12 or folate, liver disease, or renal failure (75). By inhibiting dihydrofolate reductase, methotrexate reduces levels of both plasma and red cell folate, which in turn increases levels of homocysteine via reduced activity of MTHFR. Homocysteine is directly toxic to endothelial cells, potentiates oxidation of LDL, and has prothrombotic effects (76). Recently, it has been demonstrated that an elevated total level of circulating homocysteine is an important independent risk factor for atherosclerosis in the general community (76) and also in lupus patients (77). Increased serum levels of homocysteine have been demonstrated in RA patients, including those treated with methotrexate (78–82). The combination of methotrexate and sulfasalazine appears to result in even greater elevations of homocysteine levels than does methotrexate alone (81). Supplementation with folic acid (≥5 mg/week) may reduce homocysteine levels in methotrexate-treated patients (75) and also in the wider community (83).

Other traditional risk factors. There are very few data on the role of other traditional vascular risk factors in RA. The prevalence of diabetes is not increased in RA (36), although insulin resistance has been reported in patients with RA and other systemic inflammatory diseases (84, 85). In a retrospective cohort study, hypertension was an independent predictor of mortality in RA, along with male sex and greater age at disease onset (36).

Effect of treatment of RA on mortality

If disease activity and severity are predictive of mortality in RA, effective treatment might be associated with improved survival. In a prospective mortality study of 805 RA patients, treatment with gold or glucocorticoids did not appear to affect mortality (20). In contrast, a subsequent retrospective study evaluating the influence of gold therapy on mortality showed that long-term (>120 months) treatment with gold was associated with significantly lower mortality (86). Wallberg-Jonsson et al (19) followed up a cohort of 606 seropositive RA patients for 15 years, ∼80% of whom received treatment with at least 1 disease-modifying antirheumatic drug (DMARD). In univariate analysis, treatment with a DMARD significantly decreased the risk of death. In a 25-year study of 1,842 RA patients, treatment with methotrexate was associated with a 39–50% reduction in mortality risk, an effect not seen with other DMARDs (87). Without specifically assessing the role of RA treatment on mortality, it is noteworthy that the 4 studies (5–8) that did not demonstrate increased mortality in RA predominantly included patients with short disease duration and early institution of DMARD therapy.

Krause et al (15) examined the effect of DMARD response on mortality. A cohort of RA patients who had been started on methotrexate between 1980 and 1987 was reevaluated in 1995/1996. The cohort was retrospectively divided into 4 groups based on clinical status after 1 year (>50% improvement, 20–50% improvement, no improvement but continued methotrexate, and discontinued methotrexate due to side effects). At baseline, there were no differences in disease status or comorbidities among the first 3 groups, while the fourth group had a greater prevalence of hypertension, coronary heart disease, and heart failure, as well as poorer functional status. At the time of reevaluation, the SMRs (95% confidence intervals [95% CIs]) in these groups were 1.47 (0.84–2.10), 1.85 (0.97–2.73), 4.11 (2.56–5.66), and 5.56 (3.29–7.83), respectively.

Although these studies are by no means conclusive, DMARD therapy does not appear to increase mortality in RA. Conversely, it appears that effective suppression of disease activity in RA may in fact be associated with decreased mortality and that this benefit is seen with a number of different DMARDs. Data are insufficient to determine whether this apparent protective effect is specific for cardiovascular deaths or for all causes of death. However, given that most excess RA deaths are cardiovascular in nature and are not fully explained by traditional cardiovascular risk factors, it may be that atherosclerosis is potentiated by systemic inflammation in uncontrolled RA and that the antiinflammatory effects of drug therapy in RA reduce atherogenesis.

The role of drug therapy for RA in atherosclerosis

The net effect on atherosclerosis of the drugs used to treat RA is unclear. Corticosteroids could increase the risk of atherosclerosis via deleterious effects on lipids, glucose metabolism, and blood pressure. However, corticosteroids could also decrease the risk of atherosclerosis by controlling inflammation. Equally, use of corticosteroids could just be an indicator of more severe RA. As discussed, antimalarials appear to have a beneficial effect on lipid profiles and have also been postulated to have antithrombotic properties (73, 88). Cyclosporine frequently leads to adverse lipid changes and hypertension (69, 89). Methotrexate treatment was associated with improved survival in the study by Krause et al (15). Conversely, use of methotrexate has been shown to increase homocysteine levels and has been associated with an increased risk of mortality in patients with preexisting atherosclerosis (81, 90).

Conventional NSAIDs confer vascular protection by virtue of antiplatelet effects, and widespread use of nonselective NSAIDs in RA may have contributed to an underestimation of the true vascular risk in this population. The conventional NSAIDs are increasingly being replaced by cyclooxygenase 2 (COX-2)–specific inhibitors, which selectively inhibit production of prostaglandin E2 and prostacyclin (a powerful inhibitor of platelet aggregation) without inhibiting platelet production of thromboxane A2 (a vasoconstrictor and proaggregant) (91). The resulting imbalance could potentially increase the risk of vascular thrombotic events, revealing an even greater cardiovascular burden in RA than was previously apparent. In the Vioxx Gastrointestinal Outcomes Research (VIGOR) study, which was designed to compare the gastrointestinal tolerability of rofecoxib with that of naproxen, there was a 4-fold increase in the rate of myocardial infarction in the rofecoxib group (92). Furthermore, there is some evidence that low-dose aspirin may have a reduced antithrombotic effect in the presence of COX-2–specific inhibitors (93). Dietary supplementation with fish oil, rich in omega-3 fats, has demonstrated efficacy in the treatment of RA and may also reduce cardiovascular risk (94–96). In patients with previous myocardial infarction, dietary supplementation with omega-3 fats was associated with a significant reduction in the risk of recurrent myocardial infarction or death (96).

In summary, there is no clear evidence that the medications used to treat RA increase the risk of atherosclerosis, or that they are directly protective. Research in this area is complicated by frequent changes in medications in individual RA patients, changing practices in RA therapeutics over time (e.g., widespread use of methotrexate, frequent use of combination therapy, introduction of new agents), and the long lead time of atherosclerosis prior to the development of symptomatic disease. Selective COX-2 inhibitors may be prothrombotic, and the need for further research on the clinical importance of this has been highlighted recently (97). Overall, DMARD treatment appears to be associated with decreased mortality, and this effect has been observed with various agents. These observations support the contention that the mortality benefit of DMARDs is due to control of inflammation.

The role of inflammation in atherosclerosis

There is increasing interest in the role of inflammatory and immunologic mechanisms in the initiation and progression of atherosclerosis. This reassessment is based on a number of observations, including the abundance of monocytes, macrophages, and T lymphocytes in atherosclerotic plaques. CRP, SAA, and activated complement components are also present in plaque tissue (3, 4). In animal models of atherosclerosis induced by a high-cholesterol diet, the earliest cells to adhere to the endothelium are monocytes, which migrate to subendothelial layers, engulf oxidized cholesterol, and differentiate into macrophages (1). Activated macrophages and T lymphocytes release or induce a variety of inflammatory mediators, including cytokines (e.g., interleukin-1 [IL-1], tumor necrosis factor α [TNFα]), growth factors, adhesion molecules, and matrix metalloproteinases (1). This results in further recruitment of inflammatory cells, migration and proliferation of endothelial and smooth muscle cells, collagen breakdown, platelet aggregation, in situ thrombosis, loss of endothelial nitric oxide, and release of oxygen free radicals (1). These processes contribute to the formation of atherosclerotic plaque and share many features with the pathology of RA.

Animal models illustrate the importance of inflammation in atherosclerosis. Apolipoprotein E gene knockout (apoE−/−) mice develop profound and accelerated atherosclerosis in response to a high-cholesterol diet. When apoE−/− mice were crossed with SCID/SCID mice (which lack T and B cells), there was a substantial reduction in aortic fatty streak lesions (98). Adoptive transfer of apoE−/− CD4+ T cells to apoE−/−/SCID/SCID mice resulted in preferential infiltration of the atherosclerotic lesion by donor T cells and an increase in lesion size. Reduction of atherosclerotic lesions in the apoE−/− mouse model was also seen with polyclonal immunoglobulin and by induction of neonatal tolerance to oxidized LDL (99, 100). Similarly, in LDL-receptor knockout mice, inhibition of CD40 signaling via administration of anti-CD40 antibody decreased atheroma size, lipid content, leukocyte numbers, and vascular cell adhesion molecule 1 (VCAM-1) expression (101). A link between inflammatory arthritis and blood vessel wall inflammation was observed in IL-1 receptor antagonist knockout mice, which developed an inflammatory polyarthritis when backcrossed with BALB/cA mice and severe arterial inflammation when backcrossed onto the C57BL/6J strain (102, 103). These experiments demonstrate the importance of proinflammatory cytokines, such as IL-1, in the development of atherosclerosis, and important genetic interactions in the two strains.

Results of epidemiologic studies also indicate an inflammatory component in atherosclerosis. Several population studies of healthy men and women have demonstrated a relationship between CRP levels (even within the population “normal” range) and risk of future myocardial infarction and ischemic stroke (104–107). A recent meta-analysis of 11 studies demonstrated a risk ratio for coronary heart disease of 2.0 (95% CI 1.6–2.5) in individuals in the top tertile compared with those in the bottom tertile of baseline CRP measurements (2). CRP levels also have prognostic significance in patients with stable and unstable angina (108). In prospective studies, higher baseline levels of SAA predicted coronary artery disease (65). Intercellular adhesion molecule 1 (ICAM-1) is an endothelial cell membrane–bound adhesion molecule which mediates leukocyte–endothelial cell adhesion and transmigration of leukocytes. IL-6 is a circulating cytokine that is central to the acute-phase response and one of the major stimuli for hepatic CRP synthesis and secretion. Increased baseline serum levels of soluble ICAM-1 (the circulating form of ICAM-1) and IL-6 have both been reported to be predictive of future myocardial infarction in apparently healthy men (109, 110).

Potential mechanisms of accelerated atherosclerosis in RA

If atherosclerosis is in some measure an inflammatory disease, and the excess coronary artery disease in RA is not entirely explained by traditional vascular risk factors, how then might RA contribute to the initiation and/or progression of the atheromatous plaque? Potential mechanisms are summarized in Figure 1.

Figure 1.

Possible biologic mechanisms contributing to atherosclerosis in rheumatoid arthritis. RF = rheumatoid factor; Anti-oxLDL = antibodies to oxidized low-density lipoprotein; CRP = C-reactive protein; SAA = serum amyloid A; ICAM-1 = intercellular adhesion molecule 1; VCAM-1 = vascular cell adhesion molecule 1; TXA2 = thromboxane A2; PGI2= prostacyclin; IL-1 = interleukin-1; TNFα = tumor necrosis factor α.

The event(s) which initiate RA may also initiate or potentiate a T cell–mediated process focally or diffusely in the vessel wall. Expanded populations of CD4+CD28− T cells have been demonstrated in the peripheral blood of RA patients, especially those with vasculitic complications (111). Clonal expansion of CD4+CD28− T cells has also been demonstrated in blood and atherosclerotic plaque of patients with unstable, but not stable, angina pectoris (112). Circulating RF or other immune complexes may cause direct injury to endothelial cells. Autoantibodies to oxidized LDL have been detected in patients with atherosclerosis and acute coronary syndromes (113–115); however, it is not known whether they are directly pathogenic or simply markers of arterial disease. These autoantibodies have been detected in patients with systemic lupus erythematosus and the antiphospholipid syndrome (116, 117), but to date have not been reported in RA patients.

Rheumatoid synovium is characterized by the production of multiple cytokines, including TNFα, IL-1, and IL-6. These cytokines recruit leukocytes via increased endothelial cell adhesion molecule expression and contribute to cartilage damage via induction of destructive enzymes, such as collagenase and stromelysin. Circulating cytokines could contribute to a similar process in the vessel wall, either initiating plaque growth or destabilizing an existing nonocclusive plaque. TNFα and IL-6 also induce hepatic synthesis of CRP, the prognostic significance of which has been mentioned. It has been recently demonstrated that CRP induces expression of adhesion molecules (ICAM-1, VCAM-1, and E-selectin) in human endothelial cells, suggesting a direct role of CRP in promoting intercellular adhesion and atherosclerosis (118). Increased levels of activated complement components have been reported in RA, and this may be directly mediated by CRP (119). Activated complement components up-regulate expression of proinflammatory cytokines and endothelial cell expression of leukocyte adhesion molecules (120).

Like CRP, SAA is an acute-phase reactant, the serum levels of which are increased up to 1,000-fold in systemic inflammation (4). Although SAA is mainly produced in the liver, SAA protein and messenger RNA expression has been detected in macrophages, endothelial cells, and smooth muscle cells within human atherosclerotic plaque (121). SAA has multiple functions which could contribute to atherogenesis, including both pro- and antiinflammatory effects on platelets and leukocytes and a role in cholesterol transport and metabolism (4).

Thrombosis, superimposed upon an unstable atherosclerotic plaque, is the triggering event in most acute coronary syndromes, such as myocardial infarction (122). In prospective studies of healthy subjects, elevated baseline levels of fibrinogen, von Willebrand factor (vWF), and plasminogen activator inhibitor 1 (PAI-1) were predictive of myocardial infarction (123, 124). Systemic inflammation may be associated with a hypercoagulable state due to thrombocytosis and elevated levels of fibrinogen, vWF, and PAI-1 (125, 126). In a recent survey of 76 RA patients and 641 community controls, the RA patients had significantly elevated levels of fibrinogen, vWF, tissue plasminogen activator antigen (tPA antigen), and fibrin D-dimer (127). In a separate cohort of 74 RA patients followed up for 8 years, increased baseline levels of vWF, PAI-1, and tPA were predictive of cardiovascular events (128). Thus, inflammation in RA may contribute to both the atherosclerotic and the thrombotic components of acute vascular events.

Conclusions

Atherosclerosis and its clinical sequelae appear to be more prevalent in RA than expected. Although traditional risk factors, such as smoking and dyslipidemia, and some newer risk factors (e.g., hyperhomocysteinemia) are clearly important, systemic inflammation associated with RA may play a significant role. Inflammation may be equally important in “typical” atherosclerosis. Determination of highly sensitive CRP levels may be useful in vascular risk stratification, and antiinflammatory therapies may be used to treat atherosclerotic disease in the future. Interestingly, the “statin” group of drugs lower CRP as well as cholesterol (129), which may contribute to primary and secondary prevention of coronary artery disease.

Rheumatologists need to be aware of the increased risk of atherosclerotic disease in RA and remain vigilant for vascular risk factors in their patients. Effective suppression of disease activity may reduce the risk of vascular disease and is yet another argument for early, aggressive, and sustained treatment of RA. The choice of DMARD may be influenced by the vascular profile of the patient: hydroxychloroquine appears to have favorable vascular effects, while cyclosporine, methotrexate, and COX-2 inhibitors may be detrimental in patients with established atherosclerosis. However, more research into the determinants of atherosclerosis in RA is required before we can make therapeutic decisions with confidence. Early intervention studies are needed to test whether disease control and vascular risk factor modification will decrease atherosclerosis prevalence and cardiovascular mortality in RA patients. Also, surrogate markers for early and clinically silent atherosclerosis in RA need to be evaluated. Meanwhile, lifestyle modifications, such as cessation of smoking, maintenance of physical activity, and dietary fat modification (including supplementation with omega-3 fatty acids), are low-risk strategies which may have added importance in RA.

Ancillary