Cardiovascular co-morbidity in patients with rheumatic diseases

During recent years atherosclerosis, the major cause of cardiovascular disease (CVD), has been recognised as a chronic inflammatory condition in which rupture of atherosclerotic lesions appears to play a major role. The risk of CVD is raised in many rheumatic diseases. This risk is high in systemic lupus erythematosus - as much as a 50-times increase among middle-aged women has been reported. Studies on CVD and atherosclerosis in rheumatic disease could thus provide interesting information about CVD and atherosclerosis in addition to being an important clinical problem. A combination of traditional and nontraditional risk factors accounts for the increased risk of CVD and atherosclerosis in rheumatic disease. One interesting possibility is that atherosclerotic lesions in rheumatic disease are more prone to rupture than normal atherosclerotic lesions. It is also likely that increased risk of thrombosis may play an important role, not least in systemic lupus erythematosus. Further, it is not clear whether an increased risk of CVD is a general feature of rheumatic disease, or whether this only occurs among subgroups of patients. It should be emphasised that there is an apparent lack of treatment studies where CVD in rheumatic disease is the end point. Control of disease activity and of traditional risk factors, however, appears to be well founded in relation to CVD in rheumatic disease. Further studies are needed to determine the exact role of lipid-lowering drugs as statins. Hopefully novel therapies can be developed that target the causes of the inflammation in atherosclerotic lesions both in rheumatic patients and in the general population.


Introduction
Th e history of ideas and hypotheses about atherosclerosis is interesting. Th e infl ammatory nature of atherosclerosis and the involvement of immune competent cells was described by the Austrian pathologist Karl Rokitansky in the 1840s and by the pathologist and social medicine pioneer Rudolf Virchow somewhat later in the 1850s [1]. As discussed in a previous editorial [2], these two important persons in the history of medicine had an interesting argument: Rokitansky believed that the infl ammation in atherosclerosis was secondary to other disease processes, but Virchow instead suggested that atherosclerosis is a primary infl ammatory condition. Th e relevance of this discussion to cardiovascular disease (CVD) and athero sclerosis in rheumatic disease is obvious, and in fact both arguments were most probably right. Virchow clearly had a point verifi ed in an interesting paper based on studies of Rokitansky's own arterial pathological specimens. Here, activated T cells and other infl ammatory and immune competent cells are present already at a very early stage of disease, which in principle adds support to Virchow's opinions [1]. A recent meta-analysis indicates that rheu matic diseases raise the risk of premature atherosclerosis, implying that infl ammatory conditions such as in rheumatic diseases could have secondary atherosclerosis as a side eff ect [3]. Both Rokitansky and Virchow were right, in a nonmutually exclusive way.
It was not until the early 1980s that the infl ammation/ immune hypothesis in atherosclerosis surfaced [4,5], although Russell Ross came close 1977 with his response to injury hypothesis [6]. Before that, the fi eld was dominated by the lipid hypothesis -attention was paid especially to cholesterol in the blood as a risk factor. Initially, it appeared that these two ideas about the nature of the disease contradicted each other, but now there appears to be consensus that both are relevant and nonmutually exclusive, and each probably plays a diff erent role depending on patient groups.
Interestingly, statins can illustrate this dual nature of atherosclerosis and CVD. Statins are, from a commercial point of view, among the most successful medicines in history. In fact, they may be benefi cial not only due to the Abstract During recent years atherosclerosis, the major cause of cardiovascular disease (CVD), has been recognised as a chronic infl ammatory condition in which rupture of atherosclerotic lesions appears to play a major role. The risk of CVD is raised in many rheumatic diseases. This risk is high in systemic lupus erythematosus -as much as a 50-times increase among middle-aged women has been reported. Studies on CVD and atherosclerosis in rheumatic disease could thus provide interesting information about CVD and atherosclerosis in addition to being an important clinical problem. A combination of traditional and nontraditional risk factors accounts for the increased risk of CVD and atherosclerosis in rheumatic disease. One interesting possibility is that atherosclerotic lesions in rheumatic disease are more prone to rupture than normal atherosclerotic lesions. It is also likely that increased risk of thrombosis may play an important role, not least in systemic lupus erythematosus. Further, it is not clear whether an increased risk of CVD is a general feature of rheumatic disease, or whether this only occurs among subgroups of patients. It should be emphasised that there is an apparent lack of treatment studies where CVD in rheumatic disease is the end point. Control of disease activity and of traditional risk factors, however, appears to be well founded in relation to CVD in rheumatic disease. Further studies are needed to determine the exact role of lipid-lowering drugs as statins. Hopefully novel therapies can be developed that target the causes of the infl ammation in atherosclerotic lesions both in rheumatic patients and in the general population.
mechanism for which they were developed but, in addition, for pleiotropic eff ects including anti-infl ammatory eff ects (caused by infl uencing prenylation among other processes), antioxidant eff ects, decreasing lowdensity lipoprotein (LDL) oxidation, and even immune modulatory eff ects, decreasing MHC class 2 interaction with antigen [7]. Th e Jupiter study recently demonstrated that statin treatment may be benefi cial for individuals with raised high-sensitivity C-reactive protein but normal LDL [8].

The nature of atherosclerosis and cardiovascular disease
Atherosclerosis is an infl ammatory process in large and middle-sized arteries, where activated monocytes/ macro phages and T cells are present in the intima [9,10]. Proinfl ammatory cytokines are produced by immune competent cells in the lesions [9][10][11]. In addition to chronic infl ammation, atherosclerosis also shares characteristics with autoimmune diseases -as indicated by studies where adoptive transfer of β 2 -glycoprotein Ireactive lymphocytes enhances atherosclerosis in animal models [12].
Of note, atherosclerosis per se is widespread in the population. Atherosclerosis has been detected in Egyptian mummies, and also in young adult humans. Further, and surprisingly, even foetal atherosclerosis in the form of early changes (fatty streaks) has been determined [13]. One could therefore debate whether atherosclerosis indeed is a disease and not a normal aspect of human aging.
Even though narrow lumens caused by noncomplicated atherosclerosis may be a problem in some cases, however, it is the development of atherosclerotic plaques into more complicated lesions -where fi ssures and even microthrombae and eventually plaque rupture occur -that leads to CVD, including stroke and acute coronary syndrome, heart failure (as a later consequence of acute coronary syndrome), and claudication. One major issue is therefore the cause of plaque rupture. Infl ammation plays a major role, although the exact mechanisms are not known. Activation of proinfl ammatory cytokines and chemokines are prominent features of plaque rupture. One interesting possibility is therefore that the proinfl ammatory state in rheumatic disease per se may promote atherosclerotic plaque rupture.
An interesting development is the possibility of immunisa tion, active or passive (administering anti bodies), against atherosclerosis and/or CVD. Not unexpectedly, LDL is a target -examples of antigens as culprits include apolipoprotein B peptides [14] (apolipoprotein B being the major carrier protein in LDL) or antigens in the phospholipid moiety such as phorphorylcholine (PC). Natural IgM antibodies against PC (anti-PC) are negatively associated with human atherosclerosis [15] and low levels of anti-PC predict increased risk of CVD independent of other risk factors [16][17][18].
In the above-mentioned meta-analysis where rheu matic disease and atherosclerosis were determined, it was demon strated that there is indeed a premature atherosclerosis in general. In this study, cases and matched controls where identifi ed through systematic analysis on PubMed and 68 comparisons from 60 diff erent studies were made. Taken together, of patients included in this meta-analysis, 37% had rheumatoid arthritis (RA), 35% had systemic lupus erythematosus (SLE), 9% had systemic sclerosis, and 19% had other rheumatic diseases [3].
In both SLE and RA, an association between CVD and extent of atherosclerosis has been established [19,20]. Th is association points to atherosclerosis as a major underlying factor in co-morbidity between rheumatic diseases (at least SLE and RA) and CVD.

Systemic lupus erythematosus
Since the infl ammatory nature of atherosclerosis was not in focus until the 1980s, it is not surprising that little attention was paid to associations between CVD and rheumatic disease. In an early report from 1976, however, a bimodal pattern of SLE was reported [21], According to this paper, in addition to early direct eff ects of SLE on various organ systems, a later complication was CVD [21].
Before immunosuppressive treatment was imple men ted, more acute SLE manifestations, such as nephritis, were often fatal. Early autopsy and angiographic studies also demonstrated that the prevalence of atherosclerotic lesions is high in SLE [22,23].
Th e strong association between SLE and CVD has been fi rmly established in many reports. Th is risk can be very high in some patient groups: according to one study, women aged 44 to 50 had a 50-times increased risk of myocardial infarction [24]; and an increased CVD risk in SLE is well documented [25]. Indeed, T-helper type 2 cytokines have been associated with SLE and at the same time inhibit atherosclerosis in experimental animals [25]. Even though it is clear that the risk of CVD is raised in SLE, this possibly only applies to a subgroup of SLE patients. Information and advice for rheumatic patients in relation to CVD risk should take this possibility into account. Even though CVD is associated with atherosclerosis in SLE [20], thrombosis per se possibly adds to the risk [20].
A combination of traditional and nontraditional risk factors typically accounts for, statistically, the increased risk of CVD in SLE, although there are variations in studies -for example, in relation to the role of smoking. Dyslipidaemia (typically the lupus pattern with high triglycerides), hypertension and renal disease are in most studies signifi cantly associated with CVD risk.
Non traditional factors such as infl ammation and antiphos pholipid antibodies (aPL) are also of importance in SLE, aPL more than in other rheumatic diseases. LDL is generally recognised as a risk factor in the general population, and LDL oxidation is believed to be of impor tance due to its proinfl ammatory, even toxic, eff ects and the uptake of oxidised LDL into the vascular wall, from which it is then not removed. It is therefore of interest that oxidised LDL in the circulation is raised in SLE [20,25].
A possibility deserving further study is that atherosclerotic plaques in SLE are more prone to rupture. We recently determined by carotid ultrasound that vulnerable atherosclerotic plaques are more prevalent in SLE, lending support to this notion [26].
Emerging risk factors that also implicate novel mechanisms may play a role in SLE-related CVD. Two examples of this are, fi rstly, anti-PC and, secondly, the binding of annexin A5 and interaction with endothelial cells, aPL and infl ammation.
Firstly, we reported recently that low levels of anti-PC independently predict CVD in general and that there is a negative association between anti-PC levels and development of human atherosclerosis, where high levels confer decreased atherosclerosis development after 5 years [27]. Further, low levels of anti-PC were associated with SLE in a nested case-control SLE study [28] -and in a new SLE case-control study we confi rmed and extended this association, low levels of anti-PC being associated with prevalence of atherosclerotic plaques [29].
Mechanisms by which anti-PC could be benefi cial include an anti-infl ammatory eff ect whereby anti-PC inhibits endothelial activation caused by infl ammatory phos pholipids [28]. In principle, by this anti-infl ammatory eff ect, low anti-PC could predispose to both atherosclerosis and rheumatic disease, suggesting one possible common underlying factor. Another mechanism could be decreased uptake of oxidised LDL in macrophages, which could lead to less atherosclerosis development [16].
Secondly, binding of annexin A5, which has antithrombotic properties, is decreased in individuals with SLE and CVD. Th is decreased binding is caused by aPL that outcompete annexin A5 binding, causing a prothrombotic state. We also demonstrated that annexin A5 is abundant in atherosclerotic plaques, at sites prone to plaque rupture, and suggested that this protein may stabilise plaques, protect endothelium and inhibit plaque rupture [30]. Further, pooled immunoglobulin (intravenous immuno globulin; IVIG) can neutralise aPL and restore binding of annexin A5 [31].

Rheumatoid arthritis
Th e risk of CVD is also increased in RA, although not as strikingly as in SLE [32][33][34][35]. Th e risk varies in diff erent studies, which could depend on the study populations chosen, age and other factors, including secular trends for RA per se. As in SLE, a combination of traditional and nontraditional risk factors, including infl ammation and also extra-articular manifestations, appears to explain this increased risk [32][33][34][36][37][38][39][40][41].
For example, in young women a 3.6-times increased risk of death in coronary artery disease was reported, and in a population-based cohort of RA patients the incidence of myocardial infarction and coronary heart disease was 50% higher in RA [32]. Similar results were obtained in other studies [42], and it has been suggested that RA is comparable with type 2 diabetes mellitus as an independent risk factor for CVD [43]. As in SLE, traditional CVD risk factors and infl ammation-associated factors appear to be of major importance to explain the increased risk of CVD in RA [25]. Th e risk of CVD in RA may be decreasing [44].
While it thus appears that CVD is increased in RA and SLE (and other rheumatic diseases), the exact role of atherosclerosis/CVD and potential underlying mechanisms in RA has been less clear [25].
Th e role of rheumatoid factor in this context is not known, although interestingly rheumatoid factor is often present in smokers. Further, it is not clear how immune complexes in general, or even complement, aff ect RArelated cardiovascular co-morbidity. Perhaps comple ment could play a diff erent role depending on the disease stage.
Another interesting development in RA is the role of citrullinated proteins and antibodies against these. Recent fi ndings imply that such antibodies, increasingly recognised as important novel risk markers for RA, could also play an independent role in RA-related atherosclerosis and CVD, including ischaemic heart disease [45,46].
In an interesting paper, functional polymorphisms relating to MHC-molecule expression were demonstrated to be associated with susceptibility to RA, multiple sclerosis and myocardial infarction [47], suggesting putative common mechanisms. While many studies including the above-mentioned meta-analysis support an increased prevalence of atherosclerosis as determined by ultrasound of carotid arteries [37,39,48], there are also studies where such an increase was not detected, either as the intima-media thickness or as prevalence of plaque [36]. In favour of arterial changes in RA (as in SLE) are studies in which endothelial dysfunction has been reported [49]. aPL and also antibodies against oxidised LDL are raised in RA but their clinical importance for CVD and atherosclerosis is not clear [50]. Lipid peroxidation may also play a role in RA, and oxidised LDL-containing foam cells have been described in RA synovia [51]. Further, oxidative stress is increased in RA and associated with atherosclerosis [52].
Another emerging factor in CVD is heat shock proteins, which are implicated in both RA and atherosclerosisalthough immune reactivity to heat shock proteins in RA appears to play a somewhat diff erent role, being protective in many cases while it appears to be unfavourable in atherosclerosis and CVD [53].
Dyslipidaemia is often present in RA with low highdensity lipoprotein and high triglycerides in a similar way as in infl ammatory and infectious diseases in general. An increased prevalence of potentially atherogenic, small, dense LDL particles were reported in RA, and LDL from RA patients also has an increased capacity to bind proteo glycans, which most probably is an important step in early atherogenesis [54].
As in SLE, the role played by treatment is of potential importance. Corticosteroids at moderate dose (7.5 mg prednisolone) did not aff ect atherosclerosis, but it is still possible they would aff ect atherosclerosis in higher doses -for example, the unfavourable metabolic eff ects may outweigh the anti-infl ammatory properties [55]. Th e role of metho trexate has been debated, but recent investigations suggest such treatment with folate substitution could infl uence CVD risk factors in a benefi cial way [56].
TNF inhibition could be expected to be antiatherogenic since TNF has proinfl ammatory and unfavourable metabolic eff ects [20]; in line with this, in a mouse model of atherosclerosis, TNF inhibition decreased atherosclerosis development [57]. Side eff ects such as heart failure are implicated [58]. In humans, however, TNF inhibition appears to have a favourable eff ect on CVD [59].
Even though statins may be implicated in RA (and potentially other rheumatic diseases) and indeed have an eff ect on RA per se [60], further studies are needed before general recommendations should be given.

Other rheumatic diseases and atherosclerosis/ cardiovascular disease
SLE and RA in relation to atherosclerosis and CVD have been studied more than other rheumatic diseases. In psoriatic arthritis, increased atherosclerosis has been reported [61,62] and an increased prevalence of CVD is well established with risk factors comparable with those in RA [63]. In ankylosing spondylitis, studies indicate that the risk of CVD is enhanced, but perhaps less so than in RA and SLE. Also in ankylosing spondylitis, dyslipidaemia and infl ammation per se may play a role, and positive eff ects of TNF inhibitors are discussed [64][65][66].
Gout has been associated with CVD and increased urate levels, but urate may also have antioxidant properties that may be benefi cial under some circumstances [72].

Summary and conclusions
Both atherosclerosis and the risk of CVD are increased in rheumatic diseases, especially in SLE, and the risk appears to be strikingly high. Th e underlying mechanisms are pro bably related to atherothrombosis and increased preva lence of atherosclerotic plaques where traditional and nontraditional risk factors act in concert. When treating patients with rheumatic disease, it is important to pay attention to the increased risk of CVD. Traditional risk factors as dyslipidaemia, hypertension, diabetes and smok ing should be closely monitored and disease symptoms including infl ammation should be treated. Hopefully novel therapeutic modalities will be developed that target the causes of the infl ammation present in atherosclerotic lesions.