Cardiovascular complications in inflammatory bowel disease

Abstract

Over the past years, a growing number of studies have indicated that patients suffering from inflammatory bowel disease (IBD) have an increased risk of developing cardiovascular disease. Both are chronic inflammatory diseases and share certain pathophysiological mechanisms that may influence each other. High levels of cytokines, C-reactive protein (CRP), and homocysteine in IBD patients may lead to endothelial dysfunction, an early sign of atherosclerosis. IBD patients, in general, do not show the typical risk factors for cardiovascular disease but changes in lipid profiles similar to the ones seen in cardiovascular events have been reported recently. Higher levels of coagulation factors frequently occur in IBD which may predispose to arterial thromboembolic events. Finally, the gut itself may have an impact on atherogenesis during IBD through its microbiota. Microbial products are released from the inflamed mucosa into the circulation through a leaky barrier. The induced rise in proinflammatory cytokines could contribute to endothelial damage, artherosclerosis and cardiovascular events. Although large retrospective studies favor a link between IBD and cardiovascular diseases the mechanisms behind still remain to be determined.

Introduction

Inflammatory bowel diseases (IBDs), i.e. Crohn’s disease (CD) and ulcerative colitis (UC), are chronic inflammatory diseases of the intestines that bear many venous vascular comorbidities, e.g. deep venous thrombosis or portal vein thrombosis. While the risk of developing venous disease in IBD is well established [1] the question whether IBD patients are also at risk of developing ischemic vascular diseases still remains elusive. A recent meta-review suggested that patients with IBD may indeed be at an increased risk of ischemic vascular diseases [2]. Studies have shown that chronic systemic inflammation can lead to endothelial dysfunction and platelet aggregation, precursors in the development of atherosclerosis and coronary artery disease (CAD) [3]. The weight of the evidence, mechanistically, biologically, and epidemiologically, support the notion that systemic inflammation, either induced experimentally or from what has been observed in chronic inflammatory diseases like rheumatoid arthritis or chronic kidney disease, is associated with a heightened state of cardiovascular risk. High levels of circulating cytokines and C-reactive protein (CRP) are characteristic of IBD and it is therefore expected that they contribute to endothelial dysfunction and atherogenesis. During the dangerous situation of plaque rupture when the thrombogenic core of the plaque is exposed to the blood stream, thrombus formation may occur and cause acute coronary syndrome, a risk for IBD patients who are known to have a deregulated coagulation system [4]. The intestinal microbiom could play an important role in promoting arterial disease. IBD patients have a disrupted mucosal barrier and therefore bacterial products that have translocated through the gut lining may enter the circulation and directly promote inflammation by activating immune cells and endothelial cells, known triggers in the onset or progression of cardiovascular disease. This review will highlight some likely factors and mechanisms that underlie the increased risk of cardiovascular disease in IBD patients.

Clinical studies pro and contra linking IBD with CAD

The occurrence of venous thrombotic events in IBD has been extensively investigated and described in numerous studies [5-8]. All in all, the data indicate that patients with IBD have a 1.7-5.9 fold higher risk of developing venous thromboembolism, such as deep venous thrombosis, pulmonary emboli, portal vein thrombosis and cerebral venous sinus thrombosis [1]. It has been also suggested that venous thromboembolism was pathogenetically more specific to IBD than chronic inflammation itself because chronic inflammatory diseases like rheumatoid arthritis or coeliac disease neither had an increased risk of thromboembolism [9]. For a comprehensive review on IBD and venous disease the reader is referred to Tan et al. [1].

The connection between IBD and arterial thromboembolic events has not yet been clarified but the majority of the larger studies point to a modest increase in a risk for IBD patients to develop CAD. Almost all of these studies are retrospective and heterogeneous in their methodology and subjects were mostly drawn from insurance registers and national patient databases. The results of these studies are briefly summarized here.

Observations that support an association of IBD with cardiovascular diseases

CAD and myocardial infarction

First evidence that IBD may be associated with CAD came from a study by Bernstein et al. who reported an increased risk of CAD in males and females with CD and UC (1.26 incidence rate ratio [IRR]; 95% confidence interval [CI], 1.11-1.44) [10]. Later, in a longitudinal cohort study by Yarur et al., in which patients were followed over 50 months, an increased hazard ratio (HR) (2.85 HR unadjusted; 95 % CI,1.82-4.46) for developing CAD was found in IBD patients [11]. Interestingly, the IBD patients in this study had significantly lower rates of traditional risk factors such as hypertension, diabetes, dyslipidemia, and obesity but a higher number of white blood cells, a non-traditional risk factor for CAD (1.23 HR; 95 % CI, 1.15-1.33) [11]. Using 2,831 IBD patients recruited from the Finnish National Health insurance register, a different group also reported that CAD was more frequent in IBD patients than in control subjects (p=0.004), especially in women [12]. In a nationwide Danish cohort study, Rungoe et al. observed an increased risk of CAD within the first year after diagnosis with IBD (2.13 IRR; 95% CI, 1.91 to 2.38) [13]. The authors noticed that the risk of CAD was lower among IBD patients that were on 5-ASA treatment (1.16 IRR; 95% CI, 1.06-1.26) as compared to non-users (1.36 IRR; 95% CI, 1.22-1.51). A good indication that IBD may be a risk factor for developing CAD comes from a meta-analysis that pooled maximally adjusted odd ratios (OR) and demonstrated that IBD was associated with an 18% increased risk of CAD (1.18 OR adjusted, 95% CI, 1.08-1.31) [2]. The risk increase was observed in both CD and UC and seemed to be gender-specific, predominantly occurring in females. Similarly, women who suffered from IBD and who were older than 40 had a higher risk of myocardial infarction (1.6 HR, P=0.003) whereas IBD patients as a whole did not [14].

Cerebrovascular events

Only CD, but not UC, was associated with an increased risk of cerebrovascular disease in the study by Bernstein et al. (1.32 IRR; 95% CI, 1.05-1.66) [10] while in another report, female CD and also UC patients under the age of 40 had a significant risk for stroke (2.1 HR, P=0.04) [14]. This finding corroborated results by Andersohn et al. who described that CD patients below the age of 50 had a significantly higher risk of stroke (2.93 OR; 95% CI, 1.44-5.98) than older CD patients (0.99 OR; 95% CI, 0.75-1.30), however, the results were not specific for genders [15]. After adjusting for potential confounders, a meta-analysis of five studies revealed a 18% higher risk of cerebrovascular events in both CD and UC with no significant differences between the two forms but a higher risk for females than men (1.28 OR adjusted, 95% CI,1.17-1.41) [2].

Peripheral arterial events

Only two studies so far addressed the presence of peripheral arterial events in IBD [10, 14]. Meta-analysis did not reveal a significant increase in the risk for peripheral arterial events [2].

The highest arterial thromboembolic events in people with IBD were reported for mesenteric ischemia [14]. In the study which included 17,487 IBD patients and 69,948 controls, IBD patients showed a largely increased risk of acute mesenteric ischemia (11.2 HR, P<0.001) [14]. A similar result was observed in hospitalized IBD patients in a cross sectional study using a nationwide inpatient sample database showing a strong association between IBD and mesenteric ischemia (3.4 OR adjusted; 95% CI, 2.9-4.0) and venous thrombotic diseases (1.38 OR adjusted; 95% CI, 1.25-1.53) [8].

Observations not supporting an association of IBD with cardiovascular diseases

The latter study also investigated other cardiovascular diseases, e.g. CAD, cerebrovascular occlusion and peripheral vascular disease, however, no increased risk due to IBD was observed [8]. They even found an inverse relationship between these diseases and IBD. In a different report using unadjusted analysis, patients with UC had significant higher risk of first time myocardial infarction as compared to control subjects but not after adjusting to traditional cardiovascular risk factors while CD patients did not show an increased risk either unadjusted or adjusted [16]. It should be noted that the IBD group in this study was slightly but significantly older than the control group.

Two meta-analyses covering studies between 1965 and 2006 [17] and between 1941 and 2011 [18] that addressed IBD as a risk factor of cardiovascular mortality did not reveal an association. Rates of death were increased in patients with IBD but the increase was seen in deaths of all causes such as in colorectal-cancer, pulmonary disease, and nonalcoholic liver disease, rather than in cardiovascular diseases [18]. However, this does not necessarily exclude an association between cardiovascular diseases and IBD since the end point “cardiovascular mortality” may only indirectly represent cardiovascular disease. IBD primarily affects young people who would probably be under medical surveillance early in life and would therefore receive anti-ischemic medication promptly which may significantly prolong their survival. Thus, instead of cardiovascular death, ischemic events during life may represent more reliable surrogates. Typical cardiovascular risk factors other than hypertension have not been confirmed in IBD [19], an observation which probably suggests successful anti-ischemic treatment of IBD patients.

Possible links between IBD and cardiovascular diseases

Inflammatory mediators

IBD is associated with deregulation and increase in various cytokines [20]. Higher levels of these inflammatory mediators may play a role in atherogenesis at any stage. Mediators could flood the circulation and impair endothelium-dependent dilation thus favoring atherosclerosis [21]. Among specific downstream markers of inflammation is CRP which is highly elevated in serum of IBD patients in comparison to healthy controls [22]. CRP predicts cardiovascular events as previously shown by Ridker et al. [23] and may contribute to atherogenesis [24]. Although CRP is regarded as a strong biomarker for cardiovascular events convincing data on the role of CRP in atherogenesis are still missing [25].

Other typical proinflammatory mediators involved in IBD are tumor necrosis factor alpha (TNF-α), interleukin-6 (IL-6) and vascular endothelial growth factor (VEGF). In particular, TNF-α is a proatherogenic cytokine [26] because blockade of TNF-α with infliximab ameliorated severity of CD [27] and improved endothelial dysfunction in CD patients [22]. VEGF is induced in microvascular cells and may contribute to IBD by increasing angiogenesis and inflammation [28]. VEGF worsens atherosclerosis although beneficial effects through collateral vessel formation have been proposed as well [29].

Endothelial dysfunction and carotid intimal media thickness (IMT)

Endothelial dysfunction is an important event in the development of atherosclerosis [30]. In a recent study, Roifman et al. assessed shear stress (in reactive hyperemia), pulse arterial tonometry and brachial artery flow-mediated dilation to determine micro- and macrovascular functions in IBD patients [31]. They found evidence for microvascular dysfunction in both CD and UC patients but for macrovascular dysfunction, no differences were seen between IBD patients and controls. Other groups, however, reported significantly decreased brachial artery flow-mediated vasodilatation in IBD [32, 33]. In UC, endothelial dysfunction was also shown to be dependent on the severity of the disease [32].

Another surrogate marker for atherosclerosis is intimal media thickening (IMT) of the carotid artery. In two studies the carotid IMT was measured in IBD cohorts vs. controls and both of them revealed a significant increase in carotid IMT [34, 35]. Van Leuven et al. reported increased carotid IMT in a CD cohort and a mixed IBD group (18 UC and 34 CD) [36]. Also Papa et al. demonstrated that the mean carotid IMT value was significantly higher in IBD patients (0.63 ± 0.15 mm) than in controls (0.53 ± 0.08 mm) (P = 0.008) [37]. In line with these results, higher carotid-femoral pulse wave velocity values (parameter indicating arterial stiffness; [38]) were recorded from IBD patients than from controls [39]. The values were independent of traditional cardiovascular risk factors. Although these results favor an association of atherosclerosis and IBD, two studies failed to observe an association between IMT and IBD in a mixed IBD (25 CD and 23 UC) [33] and single CD cohort [40]. Also Maharshak et al. failed to identify IBD as a risk factor for IMT [41]. Thus, a first meta-review on carotid IMT in IBD patients provided only inconclusive results [42].

Endotoxins and the gut microbiom

The idea that atherosclerosis is caused by infection dates back many years [43]. The hypothesis holds that lipoproteins are part of the innate immune system that capture microorganisms and endotoxins by building aggregates that block vasa vasorum which then leads to plaque formation [44]. With the description of the gut microbiom and the awareness of its importance in the development of IBD [45], the idea has now gained more interest [46]. Inflammation of the colon, as in IBD, disrupts the intestinal barrier and allows microbial products such as lipopolysaccharide (LPS) and other endotoxins to translocate the mucosa and enter the circulation. As a fact, IBD patients do have elevated levels of LPS which correlate with disease activity as demonstrated in patients with UC [47]. High levels of LPS may induce expression of proinflammatory cytokines that contribute to endothelial damage and foam cell formation [48]. Thus, endotoximia has been suggested as a strong risk factor for early atherosclerosis [49]. Among its actions, LPS affects oxidation of LDL [50] by stimulating its oxidation, making it toxic for human endothelial cells [51]. It also activates macrophages thereby accelerating atherosclerosis [52].

The microbial effect on atherogenesis could be mediated via Toll-like receptors 2 (TLR2) and 4 (TLR4) since they are strongly increased in atherosclerotic plaques [53]. Expression of TLR2 and TLR4 are increased on circulating monocytes in acute myocardial infarction and unstable angina patients as compared to healthy control patients [54, 55]. In addition, in vitro studies showed that TLR2 agonists induced significantly higher amounts of TNF-α in monocytes of IBD patients than in control subjects while TLR2 expression was also increased in monocytes of these patients [56]. TLR4 polymorphism was associated with a decreased risk of atherosclerosis [57] suggesting that TLR2 and 4 could be in fact a link between endotoxins and cardiovascular disease. However, in a larger study, no association was found between TLR polymorphism and carotid artery IMT [58]. It is, therefore, still unclear whether TLRs are valuable markers for atherosclerotic risk. Whether a higher endotoxin level in IBD patients may be involved in CAD also remains unclear but a very recent study in rats revealed that antibiotic treatment led to smaller myocardial infarcts and improved recovery of postischemic mechanical function as compared with untreated controls [59].

Lipoproteins

Since oxidized LDL promotes atherosclerosis and cytokine release, studies were conducted to investigate lipids in IBD patients. Pediatric CD patients have highly disturbed lipid profiles together with triacylglycerol depletion and increased protein content in VLDL [60]. In a large retrospective review of medical records of patients diagnosed with IBD between 2000 and 2007 at an academic medical center, significantly lower total cholesterol and HDL-C (cholesterol associated with HDL particles) and higher LDL-C (cholesterol associated with LDL particles) were reported in IBD patients [61]. A persistently low level of HDL-C in chronic infection and inflammation may be undesirable because data from epidemiological studies have shown a greater risk of CAD in subjects with low HDL-C levels and there is mounting evidence that inflammation markedly changes HDL functionality independent of HDL-C levels [62]. Phospholipid depletion and enrichment of HDL with pro-inflammatory proteins like serum amyloid A or apolipoprotein C-III generate dysfunctional or even pro-atherogenic forms of HDL [62]. A study by Ripolles et al. showed that, in subjects with active IBD, inflammation leads to alterations in lipid, apolipoprotein, and lipoprotein profiles and reduced cholesterol efflux capability of serum [63], a metric of HDL functionality and important cardiovascular risk factor [64]. The changes seen in the lipoprotein profile and the reduced cholesterol efflux in adult IBD patients were similar to the changes seen that lead to atherosclerosis and cardiovascular events [63].

The findings of altered lipid profiles from these studies fuel the idea that IBD patients are at higher risk for atherosclerosis than healthy subjects. However, it is presently not clear what the underlying mechanism is and whether malabsorption or systemic inflammation plays a major role in it. High content of inflammatory cytokines in IBD patients may alter de novo synthesis and degradation of lipids and thus would lead to the altered lipoprotein profile [65]. It has been further suggested that lipid-lowering agents such as statins may be beneficial in IBD [66] because they have been shown to reduce CRP levels and thrombus formation [23]. In fact, experimental data on mouse colitis models suggest an anti-inflammatory role of atorvastatin [67]. In a retrospective study of 1,986 statin-exposed and 9,871 unexposed subjects, atorvastatin was found to significantly reduce steroid intake in IBD patients while statin-associated reduction of other hazards (use of anti-TNF agents, abdominal surgery, and hospitalization) failed to reach statistical significance [68].

Hypercoagulability and the CD40/CD40L system

One important factor linking IBD with cardiovascular disease is hypercoagulability. Many abnormalities of the coagulation system and its cellular elements have been described in IBD [4]. Fig. 1 summarizes coagulation abnormalities found in IBD patients. Because of the formation of thrombi at sites of atherosclerotic plaque ruptures, abnormalities in the coagulation cascade and their cellular components are not only important for the development of venous but also arterial thromboemboli [3]. Prothrombin fragments, thrombin-antithrombin complexes, fibrinogen, activated factor XI, IX, VIII and V were all significantly elevated in patients with active UC as compared to patients in quiescent phase [69]. A decrease in the thrombin inhibitor antithrombin III is also prevalent in IBD patients [70] and occurs together with a diminished ability of fibrinolysis. The question of fibrinolysis has been addressed by DeJong et al. and others who showed that levels of tissue plasminogen activator (tPA) were reduced but those of plasminogen activator inhibitor (PAI) were increased in IBD [71, 72]. From platelets it is known that they cooperate with leukocytes during leukocyte recruitment into inflamed tissue. Platelet-leukocyte aggregates that are sensitive to thiopurine treatment form during IBD [73]. These aggregates are thought to provide a surface for both coagulation and inflammation. Activated platelets express adhesion proteins such as P-selectin that may support leukocyte rolling on surface-adherent platelets [74]. IBD patients exhibit an increased amount of activated circulating platelets and platelet aggregates that could contribute to an increased risk of thromboembolism [75]. This is of considerable interest regarding the CD40 ligand which is expressed on activated platelets and immune cells [76]. CD40 ligand induces inflammation of the endothelium [77] and blockade of the CD40/CD40L system has been shown to reduce atherogenesis in mice [78]. IBD patients have increased plasma levels of soluble CD40 ligand which reflects increased release of CD40 ligand from platelets [79]. Taken together, platelet-leukocyte interaction and an increased CD40/CD40L system may be important links in the generation of atherosclerosis and IBD. For an in depth review on coagulability and IBD see Scaldaferri et al. [4].

Fig. 1. Abnormalities in the coagulation cascade in inflammatory bowel diseases(IBD).

Aberrant levels of coagulation factors are found in IBD patients that may lead to hypercoagulability. The coagulation cascade can be devided in an intrinsic, extrinsic and common pathway. Fibrinolysis (Fibrinolytic system) and thrombin generation (Thrombin-generating system) takes also part in the coagulation system [4]. Within the extrinsic pathway, higher levels of factor VIIa has been been found in IBD patients vs. quiescent patients [69] while in the intrinsic pathway, Factors XIa and IXa were found elevated in ulcerative colitis (UC) patients vs. patients with inactive UC [69]. Additionally, Factors Va and Xa of the common pathway showed higher levels in active vs. inactive UC [69]. Within the thrombin generating system, prothrombin fragments 1+2 (F1+2) and the thrombin-antithrombin complex (TAT) were also significantly elevated in IBD patients vs. patients in quiescent phase [69] whereas thrombin inhibitor antithrombin III was decreased [70]. Within the fibrinolytic system, inhibitor of plasminogen activator-1 (PAI-1) and thrombin-activatable fibrinolysis inhibitor (TAFI) were increased while tissue plasminogen activator (tPA) was decreased in IBD patients vs. healthy controls [71, 72, 94] suggesting promotion of prothrombotic mechanisms in IBD.

Homocysteine

A deranged homocysteine metabolism has been implicated in atherogenesis [80]. In a population-based prospective cohort study, plasma homocysteine was found to be higher in subjects with preexisting cardiovascular disease than in subjects without it [81]. Homocysteine is produced from methionine and a study by Kanani et al. shows that experimentally-increased homocysteinemia by oral methionine overload leads to endothelial dysfunction as measured by forearm blood flow [82]. Homocysteine may damage the endothelium through generation of oxidative stress, by inhibition of vasodilator properties of nitric oxide synthase and by inciting the release of cytokines [83]. IBD patients typically have high levels of homocysteine in plasma [84] and colon mucosa [85] where it may contribute to colon inflammation [86]. Plasma homocysteine levels were also found to be higher in IBD patients with active disease than in those in remission [87]. These observations suggest a role for homocysteine in atherogenesis of IBD patients. In support of this idea, homocysteine levels in patients with a history of arterial and venous thromboembolic events were investigated before and after methionine loading [88]. The group found a significant increase in homocysteine after methionine loading in the IBD arterial thrombosis group as compared to IBD controls. The results have been challenged by a meta-review that did not indicate a higher risk of hyperhomocysteinemia among IBD patients who had a history of thromboembolic complications [84]. Because of the low number of studies (only four) and IBD patients included in the meta-review, the authors rightfully pointed out that, regarding the risk of hyperhomocysteinemia-related thrombosis in IBD, firm evidence-based conclusions could not have been drawn from the available data [84].

Conclusion and future perspectives

In summary, there is convincing evidence that IBD is associated with an increased risk of cardiovascular events. Though the number of publications on this topic is limited and despite the lack of prospective studies, the evidence is multifold and strong. The available evidence supports that IBD is associated with CAD, myocardial infarction, cerebrovascular ischemic events and mesenteric ischemia. Stronger epidemiological data and timely prospective studies are needed to substantiate the risk, since depending on the risk increase, special attention or even surveillance may be necessary. The mechanisms behind the increased cardiovascular risk associated with IBD are still unknown. Most of the presently available data suggests that inflammation or consequences of the systemic inflammation may contribute. Pivotal studies demonstrate endothelial dysfunction and increased IMT are involved. The role of systemic inflammatory mediators seems self-evident but clarifying studies are necessary to reveal the involved mechanisms. Further attention is also required to detail the mechanisms behind the dyslipidemia in IBD since dyslipidemia is a well-established cardiovascular risk factor and IBD patients may benefit from medications altering the lipoprotein profile. Additional changes in IBD like hypercoagulability and hyperhomocysteinemia may contribute to the increased cardiovascular risk but these associations are less well established. Fig. 2 summarizes potential links between IBD and cardiovascular disease.

Fig. 2. Potential factors linking inflammatory bowel diseases (IBD) with cardiovascular disease.

Cytokines, C-reactive protein (CRP) and LDL-cholesterol are increased in patients with IBD and may promote atherogenesis and cardiovascular disease. Decreased levels of HDL-cholesterol observed in IBD patients and impaired HDL functionality may contribute as well. A further characteristic of IBD is a leaky mucosal barrier. Microbial products and endotoxins could translocate through the colon mucosa and enter the circulation to directly activate immune cells and endothelial cells thus contributing to atherosclerosis. IBD patients also exhibit high levels of homocysteine and coagulation factors known to be implicated in atherosclerosis and thrombus formation.

With regard to cardiovascular diseases, commonalities with other chronic inflammatory diseases like rheumatoid arthritis (RA), for which a strong link to cardiovascular diseases has been established [89], exist. Similar to IBD, RA patients display elevated levels of inflammatory markers, such as TNF-α, IL-6 and, in particular, CRP [90]. RA patients also show increased arterial stiffness [91] and elevated levels of homocysteine [92]. A recent literature analysis has revealed findings of coagulation abnormalities and increased platelet masses in RA patients suggesting similar mechanisms in the generation of cardiovascular disease between IBD and RA [93].

Finally it has to be considered that also short and long term medications used in the treatment of IBD may contribute to the increased cardiovascular risk. Future studies on the impact of IBD on cardiovascular disease and the underlying mechanisms are imperative since with a better understanding we may be able to proactively avoid cardiovascular events in this patient population.

Abbreviations

5-ASA

5-aminosalicylic acid

CAD

coronary artery disease

CD

Crohn’s disease

CI

confidence interval

CRP

C-reactive protein

HDL

high-density lipoprotein

HDL-C

HDL-cholesterol

HR

hazard risk

IBD

inflammatory bowel disease

IRR

incidence rate ratio

IL-6

interleukin 6

IL-1β

interleukin 1 beta

IMT

intimal media thickness

LDL

low-density lipoprotein

LDL-C

LDL-cholesterol

LPS

lipopolysaccharide

OR

odds ratio

P

probability

PAI-1

inhibitor of plasminogen activator-1

TAFI

thrombin-activatable fibrinolysis inhibitor

TAT

thrombin-antithrombin complex

TNF-α

tumor necrosis factor alpha

TLR2

Toll-like receptor 2

TLR4

Toll-like receptor 4

tPA

tissue plasminogen activator

UC

ulcerative colitis

VEGF

vascular endothelial growth factor

VLDL

very low density lipoprotein