Paradox of simultaneous intestinal ischaemia and hyperaemia in inflammatory bowel disease


  • Division of Cardiovascular Medicine (O. A. Hatoum, D. D. Gutterman). Division of Gastroenterology and Hepatology (D. G. Binion), Department of Medicine (O. A. Hatoum, D. G. Binion, D. D. Gutterman), Cardiovascular Research Center (O. A. Hatoum, D. D. Gutterman), Digestive Disease Center (D. G. Binion), Froedtert Memorial Lutheran (D. G. Binion) Hospital, Medical College of Wisconsin (O. A. Hatoum, D. G. Binion, D. D. Gutterman), Milwaukee, WI.

Ossama A. Hatoum, MD, MS, Medical College of Wisconsin, 8701 Watertown Plank Road, Milwaukee, WI 53226. E-mail:


This review has focused on evidence regarding intestinal perfusion of inflammatory bowel disease (IBD). Basic investigation has defined an altered microvascular anatomy in the affected IBD bowel, which corresponds with diminished mucosal perfusion in the setting of chronic, long-standing inflammation. Diminished perfusion is linked to impaired wound healing, and may contribute to the continued refractory mucosal damage, which characterizes IBD. Alterations in vascular anatomy and physiology in IBD suggests additional possible mechanisms by which micro-vessels may contribute to the initiation and perpetuation of IBD. This begs the following questions: will angiogenesis within the gut lead to sustained inflammation, does the growing vasculature generate factors that transform the surrounding tissue and does angiogenesis generate vascular anastomosis within the gut, with shunting of blood away from the mucosal surface, impairment of metabolism and potentiation of gut damage? Further studies are required to define the mechanisms that underlie the vascular dysfunction and its role in pathophysiology of IBD.


Inflammation is a response of vascularized tissues to sublethal injury where immunological, defensive and reparative mechanisms are activated in an orchestrated response to restore tissue homeostasis [1,2]. It is designed to destroy or inactivate invading pathogens, remove waste and cellular debris and permit restoration of normal function, either through resolution or repair. This reaction may trigger a systemic response such as fever, leukocytosis, protein catabolism and altered hepatic synthesis of plasma proteins, such as C-reactive protein [3]. Inflammation can be classified according to duration as either acute or chronic, where the vasculature plays integral roles in both processes [3]. Acute inflammation is characterized by a classic vascular response involving tissue hyperaemia resulting from dilation of small blood vessels followed by fluid and granulocyte accumulation at the injured site. In contrast, chronic inflammation is marked by tissue infiltration with macrophages and lymphocytic cells. The role of the micro-vasculature in altered leukocyte recruitment is perhaps the best characterized aspect of micro-vascular physiology in chronic inflammation, which may lead to a functional but morphologically altered organ over time.

A highly controversial concept regarding the micro-vascular contribution to chronic inflammation involves the pathophysiology of tissue perfusion and blood flow. As opposed to acute inflammation, where tissue hyperaemia and increased vascular perfusion play a critical role in essentially all aspects of the response, there is compelling data suggesting that both increased tissue blood flow (i.e. hyperaemia) as well as impaired, diminished blood flow (i.e. ischaemia) can be identified within chronically inflamed tissue.

Inflammatory bowel diseases (IBD: Crohn's disease, ulcerative colitis) are lifelong disorders characterized by chronic inflammatory damage in the gastrointestinal tract. Although the etiopathogenesis of IBD has not been defined and the majority of research has focused on an inappropriate innate and acquired immune response to enteric bacterial and potentially other antigens, the vasculature is also considered to play a key role in the disease process. Data demonstrating increased perfusion of the mesenteric vasculature in active, fulminant colitis has been described. In contrast, diminished vascular perfusion of the mucosal surfaces and the arterioles in and below the mucosa has also been shown in patients with longstanding disease.

This article will review data demonstrating the vascular perfusion changes and IBD and attempt to reconcile the apparent conflict between simultaneous hyperaemia and ischaemia with a unifying hypothesis linking vascular perfusion abnormalities to the pathophysiology of IBD intestine.

Evidence for hyperaemia in the chronically inflamed IBD intestine

Surgical appearance of the intestine in IBD

Although considerable variation exists in the presentation of the disease in surgical specimens, Crohn's disease (CD) is often recognized solely by appearance of macroscopic lesions including mucosal ulceration and thickening of the bowel wall caused by inflammatory cell infiltration and oedema. Surgical assessment of the intestine in CD reveals that the mesentery is often thickened and stiff with hypertrophied white adipose tissue [4,5]. This fat wrapping correlates with the presence of underlying acute and chronic inflammation, in addition to transmural inflammation in the form of lymphoid aggregates [6]. It may extend from the mesenteric attachment and partially cover the small and/or large intestinal circumference, and is considered a reliable indicator for CD [7]. Furthermore, peroxisome proliferation activator receptor γ (PPARγ), a pivotal mediator in the regulation of adipose tissue homeostasis and lipogenesis, is found in increased amount in bowel from subjects with CD [8]. Adipocytes may participate in the inflammatory process of CD by producing tumour necrosis factor and other inflammatory and pro-angiogenic factors [8]. Neovascularization and micro-vascular hypertrophy of the serosal surfaces of the intestine may also be present in the absence of fat wrapping [6,9]. This angiogenic process will sometimes remain as the only indication of an area previously affected by CD, once the chronic inflammatory infiltrate has subsided. Although these abnormalities of mesenteric fat in CD have long been recognized their pathophysiologic relevance is still unknown [8]. Because the origin of the confined mesenteric fat hypertrophy in CD is not defined, it is tempting to speculate that a mesenteric over-expression of PPARγ is the origin of a localized fat wrapping hypertrophy and in turn increases tissue hyperaemia and augmented blood flow in IBD.

In contrast, the remodelled ulcerative colitis (UC) intestine mostly demonstrates mucosal and muscle layer changes that may lead to shortening and narrowing of the large bowel. However, fibrosis is uncommon and smooth strictures are rarely observed in long-standing chronic disease [10,11].

Early IBD-increased mucosal perfusion – intraoperative and endoscopic studies

The importance of the vascular system in intestinal physiology and homeostasis has prompted studies into potential vascular alterations in IBD. The characterization of intestinal blood flow in IBD has been completed using both direct and indirect methods of perfusion measurement. These methods include micro-sphere techniques, laser Doppler velocimetry, hydrogen clearance, and various spectroscopic procedures. Spectroscopic measurements have been more extensively used to measure perfusion in the abdomen and duodenum and suggest an increase in mucosal blood flow early on in IBD patients [12,13]. These findings have been confirmed by other studies utilising intra-operative isotope washout [14] as well as in vivo abdominal angiography. Hulten et al. demonstrated that in severe, acute UC and CD (i.e. initial presentation) there is tissue hyperaemia, intense hyper-vascularity, with significantly augmented blood flow compared with normal controls [14,15]. These early observations have been confirmed with subsequent endoscopic Doppler ultrasonography exhibiting an augmented blood flow only in acute disease and early phases of animal models of colitis [16].

Transabdominal ultrasound studies of gut perfusion in IBD patients

Several noninvasive techniques have been employed to assess gut blood flow in human subjects. Doppler ultrasonography (US) is a suitable non-invasive method for investigating splanchnic haemodynamics [17]. Previous studies have shown US to be accurate and reproducible in various pathophysiological conditions such as intestinal angina [18], portal hypertension [16] as well as in the diagnosis and follow ups of patients with IBD [17,19–24]. An advantage of US is its ability to enable visualization and location of transmural bowel inflammation in CD, discrimination of CD from UC, determination of the extent of the lesions and detection of complications such as strictures, fistulas and abscesses [19,20,25,26]. Bolondi et al. and others have demonstrated that patients with UC or CD have significantly higher portal and mesenteric venous outflow compared with healthy controls [19–22,27]. These findings are consistent with a hyperdynamic splanchnic circulation in active IBD.

Unfortunately, the technical and systematic problems of US are manifold. Major sources of error include incorrect positioning of the US probe, difficulty in obtaining an optimal angle of insonation and imprecise measurements of the lumen diameter, which may lead to the incorrect estimation of flow by greater than 25%[17]. Other non-technical limitations must also be considered. First, medications used in the treatment of IBD may exert changes in vascular physiology. Corticosteroids may increase not only systemic but also mesenteric vascular tone, thereby decreasing blood velocity and flow in the abdomen [28]. In contrast, posture-induced blood volume expansion has been shown to increase mesenteric blood flow in healthy controls [29]. Therefore, the net effect of steroids on mesenteric blood may be of minor importance, which is supported by observations that intra-arterial infusion of cortisol has no vaso-constricting or dilating effects in healthy subjects [30]. Second, splanchnic flow might be altered in patients with previous intestinal resection. After extended small bowel resection in animals, increased blood flow was observed in the ileum but not colon [31]. In contrast, human studies demonstrated decreased blood flow in the neo-terminal ileum after ileocolonic resection for CD compared with controls [32]. Third, active IBD bowel segments are thickened, oedematous, and therefore less mobile. These segments show reduced motility and compressibility and are painful on graded compression [33–36]. Finally, suboptimal distension of the colon, air in the colon or colonic contents can cause artifacts, thus hampering an accurate assessment of the bowel wall in ultrasound examinations.

Microbubble contrast enhancement of abdominal Doppler US in IBD

Limitations in the ability of traditional US to assess the micro-circulation has led to the development of new techniques in trans-abdominal US using intravenous microbubble contrast reagents (Levovist; galactose-based sonographic contrast agent; Schering AG, Berlin, Germany) [37]. Microbubbles interact with an ultrasound beam and resonate, producing harmonic enhanced signals. This agent is useful in detection of small vessel flow, especially in areas where the signal to noise ratio limits traditional US sensitivity and performance, including intracranial and deep abdominal examination [38]. Transabdominal Doppler sonography with Levovist increased the sonographic sensitivity from 70·9% to 96·7% and allowed for the differentiation of specific subsets of CD patients (i.e. active, fibrotic and relapsing diseases). Sabatino et al. conducted the first study using microbubble contrast enhanced US as a diagnostic modality for CD [39] and identified high signal intensity in basal Doppler, which was further increased after Levovist injection. Most of these CD patients had abnormal bowel wall thickness. This finding was reinforced by after achieving remission a significant reduction in bowel wall thickness was observed. In addition they reported on a particular subset of patients who did not show any basal Doppler signal, but enhanced signal intensity with additional microbubble contrast. Overall, these findings have demonstrated increased intestinal perfusion and hyperaemia in the remodelled and thickened IBD intestine.

However, the description of increased Doppler US in CD patients with Levovist-contrast has not uniformly demonstrated all hyperaemia. First, patients with previous intestinal resection demonstrated a distinct low blood flow pattern to the neo-terminal ileum, compared with non-operated CD patients. Second, Doppler US studies specifically evaluating mucosal surfaces using an intraluminal approach, as opposed to the trans-abdominal approach, have demonstrated diminished perfusion in CD [32,40]. Third, the ultrasound observations which suggest enhanced blood flow in CD patients may actually reflect the thickened CD intestine and its associated serosal fat wrapping with enhanced serosal vascularity [4,5,7,41]. These studies suggested that a simple correlation with enhanced Doppler US signal using Levovist and CD does not exist, but must take into account the location of the probe (mucosal vs. transabdominal) as well as the stage of the disease.

Ultrasound correlation with disease activity in IBD

In IBD, assessment of disease activity and prediction of relapse are of major clinical importance. There are few studies correlating endoscopic findings with US findings. Futagami et al. found a significant correlation between an US index and endoscopic/radiological score in patients with CD, whilst a weak correlation was found between disease activity index and biological indices of inflammation [42]. These data are of great value in the ongoing assessment and treatment of IBD patients. Furthermore, differentiation between UC and CD with US is controversial. Most investigators have not observed any specific US pattern, but others have reported that the US pattern can differentiate between CD and UC [26,35,43]. Some authors report lower accuracy for US in UC than CD, whereas some other studies have demonstrated better results in UC than in CD [44–46]. In general, ultrasound is more accurate in evaluation of UC than CD, which may be explained by different segmental distribution of involvement in UC vs. CD.

Published data correlating ultrasound findings and clinical disease activity indices or other biologic parameters of disease activity in IBD have also demonstrated conflicting results. Haber et al. demonstrated that bowel wall thickness, as measured by US, is related to disease activity [45] and suggests that US is an accurate and easy, non-invasive method for monitoring the intestinal inflammatory process of the CD patient after the diagnosis is established and treatment initiated [47]. Kettretz et al. found a significant correlation between bowel wall thickness measured by US and abdominal pain score, or stool score, in children with UC but not in children with CD [48]. D’Arienzo et al. reported that the bowel wall thickness was the only parameter that could differentiate the activity of UC, which correlated significantly with a modified clinical endoscopic activity index of these patients [49]. In contrast, Mayer et al. and others demonstrated there was no correlation between the volume of inflamed bowel and clinical activity of IBD [42,50]. The above observations suggested that US measurements may evolve as a simple non-invasive method for monitoring the intestinal inflammatory process following IBD diagnosis and after treatment has been initiated, but further work is needed at this time.

Angiogenesis in IBD

Angiogenesis is defined as a growth of new capillary blood vessels from pre-existing vasculature [51]. This is of key importance in numerous physiologic and pathologic processes, including embryogenesis, tissue growth, wound healing and the female reproductive cycle. Angiogenesis may also contribute to the pathology of diseases such as cancer, psoriasis, tissue damage during reperfusion following ischaemia or cardiac failure, diabetic retinopathy and chronic inflammatory diseases in joints or gut [52–55]. Inflammation may promote angiogenesis in a number of ways [53,56–60]. Inflammatory tissue is often hypoxic and hypoxia can induce angiogenesis through up-regulation of factors such as vascular endothelial growth factor, fibroblast growth factor-1, tumour necrosis factor α, hypoxia-inducible factor-1 and other factors [61–65]. Extravasated plasma fibrinogen can stimulate neovascularization [66–68]. Inflammatory cells such as macrophages, lymphocytes, mast cells and fibroblasts can stimulate vessel growth by producing angiogenic factors [69–73]. Increased blood flow itself may stimulate angiogenesis through shear stress on the endothelium [74–76].

Currently, evidence for the presence of angiogenesis in IBD is extremely limited. Indirect evidence clues that increased vascularization actually occur in the inflamed mucosa. Up-regulation of pro-angiogenic factors [77,78] is consistent with Doppler US findings of increased vessel density of acute inflamed bowel loops. However, as cautioned earlier, this may represent an increased density of the overlapping fat micro-vessels, not mucosal vessel proliferation. Recent studies by Danese and colleagues have demonstrated angiogenic activity in an animal model of IBD, with additional demonstration of angiogenic cytokines being released from tissue biopsies from IBD patient materials [79]. Further studies have suggested that the IL-10 knock-out mouse will demonstrate clinical improvement when treated with novel angiogenesis inhibitors, which have also demonstrated benefit in models of rheumatoid arthritis [79].

Direct evidence of angiogenesis in acute and chronic inflammation, particularly human inflammatory disease, remains limited, with a majority of data from rheumatoid arthritis where angiogenesis is felt to be a central component of pannus formation in the inflamed synovium [80]. To date, the available data suggests that angiogenesis and inflammation frequently occur together, but the pathophysiologic relevance of angiogenesis in the perpetuation of chronic inflammation has been limited to animal studies [80,81,81,82]. Further studies should be designed to specifically investigate the role of angiogenesis in the chronically inflamed tissues of human IBD.

Evidence for ischaemia in the chronically inflamed IBD intestine

Diminished mucosal perfusion – intraoperative and endoscopic studies

In IBD, intestinal vascular damage has been demonstrated as an early pathologic finding that precedes mucosal ulceration. Damage is often greatest in areas where vessels penetrate through the muscularis propria and in regions of dense angiogenesis observed in distal areas of the mucosal circulation. Funayama et al.[83] investigated remodelling in the intestinal micro-circulation from CD resected bowel using tissue histometry and suggested that circulatory disturbance occurred in the early stage in CD, leading to increased vascular resistance in the area between the deep submucosal and the distal mesenteric arteries, possibly contributing to ischaemia. Furthermore, Wakefield et al.[84] identified occlusive fibrinoid lesions in the small arteries and arterioles supplying CD-affected areas of intestine that were not demonstrated in uninvolved areas of bowel using scanning electron micrographs of corrosion microcasts. Morphologically, the chronically inflamed micro-vessels were tapered and stenosed compared with vessels from areas of uninvolved CD and control bowel. Taken together, these studies suggest that the microvascular anatomy undergoes vascular remodelling and plays a role in the pathogenesis of chronic inflammatory lesions, with the extent of vascular damage correlating with the severity of intestinal injury.

Attempts to characterize alterations in intestinal blood flow in IBD have been utilized both direct and indirect methods. In 1977, Hulten et al. demonstrated in chronically inflamed tissues (both chronically active and quiescent disease) diminished gut perfusion compared with controls, which correlated with histological evidence of persistent ulceration and intestinal fibrosis [14]. Angerson et al. demonstrated similar findings using endoscopic Doppler laser flowmetry, where chronically inflamed CD bowel demonstrated significantly decreased perfusion [32]. Similar findings were demonstrated by Tateishi and colleagues using laser Doppler flowmetry intra-operatively in CD patients undergoing resection [40]. These observations have consistently demonstrated that the most severe decrease in vascular perfusion is found in association with fibrotic strictures in patients with longstanding CD [32].

This potential contribution of mucosal ischaemia to chronic inflammation in IBD was evaluated by Wakefield et al. using scanning electron micrographs of corrosion microcasts following bowel resection [84]. They identified occlusive fibrinoid lesions in the arterioles supplying areas of intestine affected by CD, which were not found in uninvolved areas of bowel. This was described as a ‘granulomatous vasculitis’, with a histological appearance characteristic of prolonged disruption in the local vascular supply leading to microinfarction. Importantly, microvascular damage was an early pathologic finding that preceded the development of mucosal ulceration.

Microvascular dysfunction in IBD arterioles – microvascular physiologic studies

Refractory, poorly healing wounds such as foot and leg ulcers in patients with peripheral vascular disease and longstanding diabetes mellitus are frequently ischaemic in origin and have been linked to microvascular dysfunction [85–90]. In IBD intestine, similar poor healing and refractory inflammatory ulcerations occur in the gut mucosa, consistent with tissue hypoperfusion of an ischaemic etiology. This raises the question of whether impaired vasodilation and chronic hypoperfusion are responsible for the progression of disease in IBD. Based on this hypothesis, Hatoum et al.[91] directly examined the vasodilator responses in human intestinal microvessels by measuring in vitro vasodilator capacity in response to acetylcholine (Ach; endothelial-dependent vasodilator) from suspended and perfused 100–150-micron diameter intestinal arterioles isolated from IBD and control (non-IBD) specimens following intestinal resection. Non-IBD intestinal arterioles dilated dose-dependently to Ach, whilst chronically inflamed IBD arterioles (both CD and UC) demonstrated a diminished vasodilatory capacity to Ach [non-IBD microvessels maximal dilation (MD): 82 ± 2%, n = 34 compared with IBD microvessels MD: 15 ± 2%, n = 33, respectively; P ≤ 0·05]. This decreased vasodilator capacity was directly related to a loss of nitric oxide (NO) dependent function. The same vessels were found to be heavily dependent on cyclo-oxygenase (COX)-derived vasodilator compounds to maintain vascular tone and dilator capacity to Ach. The use of indomethacin, a nonselective COX isoforms inhibitor, decreased vasodilation of non-IBD vessels to Ach by approximately 40%, whilst in IBD arterioles it resulted in frank vasoconstriction [91]. The mechanism of the impaired NO-mediated dilation may have involved excess production of oxidative stress, as these compounds were measured by intravital dyes in IBD but not control arterioles [91].

Similar results were published recently using animal models of dextran-sodium sulphate (DSS) induced colonic inflammation. Mori et al. reported the time-course of changes in colonic blood perfusion and the mechanisms that may underlie these changes in blood flow. They reported a significant reduction (18–30%) in blood flow in small arterioles (< 40 µm diameter) on days four through six of DSS colitis in response to bradykinin. In addition, they found that NAD(P)H oxidase-derived superoxide plays a major role in the induction of the inflammation-induced endothelium-dependent arteriolar dysfunction [92].

Loss of endothelial generation of NO in IBD endothelial cells

Early studies into the potential role of the endothelium in IBD pathogenesis focused on histological evaluation, characterizing the morphology of the microvasculature in chronically inflamed intestine. Dvorak et al. evaluated CD intestinal specimens using transmission electron microscopy, and demonstrated endothelial cell abnormalities, characterized by loss of monolayer integrity with tissue oedema and extravasation of red blood cells, focal venular endothelial necrosis adjacent to areas of undamaged endothelial cells as well as endothelial cell hypertrophy [3]. To more fully define the contribution of microvascular endothelial cells in chronic intestinal inflammation, Binion et al. developed protocols for the routine isolation and long-term culture of pure populations of human intestinal microvascular endothelial cells (HIMECs) from small and large intestinal resections. Endothelial cultures were generated from areas of chronically inflamed and uninvolved CD and UC intestine. Human intestinal microvascular endothelial cells isolated from both chronically inflamed CD and UC demonstrated a significantly enhanced capacity to adhere leukocytes, compared with control HIMEC cultures. The phenomenon of leukocyte hyperadhesion was only present in chronically inflamed IBD HIMECs, as cultures derived from uninvolved areas in close proximity failed to demonstrate increased leukocyte binding [93]. The mechanisms underlying leukocyte hyperadhesion in the chronically inflamed IBD HIMECs did not appear to involve altered patterns of cell adhesion molecule expression, but instead were linked to reduced NO production [94,95]. The NO exerts a potent anti-inflammatory effect within the vasculature, down-regulating the activation of vascular endothelial cells as well as their capacity to bind circulating leukocytes (normally an early, rate-limiting step in the inflammatory process). Control HIMECs displayed distinct patterns of NO generation through both constitutive endothelial nitric oxide synthase (eNOS; NOS3) as well as inducible NOS (iNOS; NOS2). Inhibition of iNOS function in control HIMECs produced increased patterns of leukocyte binding, which were similar to IBD HIMECs. Investigation of the mechanisms of reduced NO generation in the IBD HIMECs confirmed that a loss of iNOS gene transcription and expression resulted from elevations in cytokines and lipopolysaccharides.

Enhanced vasoconstrictor products in IBD intestine

Additional evidence for mucosal hypoperfusion in CD and UC comes from studies investigating specific vasoconstrictor peptides. The endothelin peptides (ET-1, ET-2 and ET-3) are potent and long-acting vasoconstrictors that have been implicated as causative agents in a number of pathological states [96,97]. Because IBD micro-vessels often display heightened vasomotor tone, Murch et al. performed initial studies demonstrating increased levels of endothelins in IBD. Both ET-1 and ET-2 were found to be increased in the colonic mucosa from patients with CD and UC [98]. This observation provided an important link between the hypotheses that immunological hypersensitivity [99] and vascular abnormalities [100,101] contributed to the pathophysiology of IBD. Interestingly, several of the pro-inflammatory cytokines [102,103] can stimulate production of endothelin's [104,105]. What Klemm et al. have shown indicates that administration of tumour necrosis factor-alpha and interleukin-2 to anaesthetized rats caused a marked elevation in the circulating plasma level of ET-1 and a very pronounced ET-1-dependent coronary vasoconstriction. In addition, using rats suffering from adjuvant polyarthritis, in which there is marked joint inflammation and associated cytokine production, there was dramatic increases in coronary perfusion pressure which were absent when rats were treated with an endothelin receptor antagonist.

Despite these earlier observations, other investigators have not confirmed an increase in tissue ET-1 and ET-2 in IBD, with Rachmilewtz et al. actually reporting reduced levels of both mediators in UC [106]. These observations suggest that cytokines could induce the production of endothelin and promote vasoconstriction which may lead to tissue damage. Furthermore, despite the presence of endothelin binding sites to ET-1 in the colonic neural plexus and mucosa [107,108], the authors believe that further studies are needed to test the relative expression and roles of the individual endothelin isopeptides within the human gastrointestinal tract.

Thromboxane A2 is a pro-inflammatory eicosanoid derived from isomerisation of prostaglandin H2 by thromboxane synthase. Thromboxane A2 has a very short half life and is rapidly metabolized to the more stable product, thromboxane B2[109]. Mucosal production of thromboxane is increased, both in the spontaneous model of colitis [110,111] and in colitis induced by exogenous agents [112–114]. Within human IBD, thromboxane production in cultured isolated intestinal epithelial cells from patients with UC and those with CD is increased [115,116]. Additionally, in active colonic tissue homogenates from patients with UC [117,118] and in cultured biopsies from both patients with UC or CD excess thromboxane production was observed [119–121]. Moreover, rectal dialysis in patients with UC and CD showed increased rectal mucosal production of thromboxane in active disease in vivo[122–125]. In vitro, studies have shown that thromboxane has pro-inflammatory actions and induces the activation of neutrophils [126], the production of leukotriene B4 [127] and enhances neutrophil adhesion to the endothelium [128] with subsequent diapedesis [129]. Furthermore, it induces apoptosis [130], modulates T-cell function [131] and induces vasoconstriction [132] and platelet aggregation [109]. The platelet aggregates have been identified in the capillaries of inflamed rectal biopsies in IBD [133], and microvascular thrombosis has been proposed as an early pathogenic factor in CD [101]. Therefore, thromboxane can perpetuate ischaemic injury and thereby play a key role in the pathogenesis of IBD.

Clinical response to pharmacological vasoconstrictors in IBD: non-steroidal anti-inflammatory agents

It has been documented that nonsteroidal anti-inflammatory drugs (NSAIDs) can have deleterious effects on small bowel and colonic mucosa in addition to the well-known adverse effects on the upper gastrointestinal tract [134–140]. The clinical manifestations of NSAID-associated small and large bowel tract toxicity are extensive and include diarrhoea, abdominal pain, blood and protein loss, colitis and perforation. The influence of NSAIDs has been reported to cause the initial onset of IBD and to be associated with reactivation of quiescent disease. This conclusion has been based largely on case reports and small series of patients [134–140]. More recently, authors have noted a higher than expected use of NSAIDs among patients admitted to the hospital for flares of IBD [140], leading some physicians to recommend that IBD patients strictly avoid use of NSAIDs [135,141]. The mechanism by which NSAIDs may provoke inflammatory bowel disease is not clear, but may involve a mechanism of altered arachidonic acid metabolism similar to that of NSAID-induced enterocolitis that is not related to IBD [142–145]. Some prostaglandins, end products of the cyclooxygenase pathway, serve a protective role in the gut [146–149] and NSAIDs inhibit their formation. Similar findings have been seen in an animal model of experimental colitis, which demonstrated marked elevation of cyclooxygenase-2 mRNA and increased prostaglandin synthesis by the colon. Treatment with selective cyclooxygenase-2 inhibitors resulted in exacerbation of inflammation-associated colonic injury and perforation leading to early death in the treated group [150].

Specific inhibition of thromboxane synthase, but not cyclooxygenase or prostacyclin synthase, leads to reduced formation of thromboxane, overproduction of prostaglandins I2, E2 and D2 and no change in leukotriene production. Ridogrel (thromboxane synthase inhibitor and receptor antagonist) [151–153] ameliorates experimental colitis in animals and has shown promise in preliminary studies in patients with UC [113,154]. Furthermore, a pilot open-label trial with a different thromboxane synthase inhibitor and receptor antagonist, picotamide, also suggested an improvement in CD activity index after 6 weeks of treatment in nine patients with active CD [155–157]. In contrast, studies that failed to achieve sufficient inhibition of thromboxane synthesis and of platelet function demonstrated no clinical improvement [155,158]. If overproduction of mucosal thromboxane and excessive platelet activation are of major pathogenic importance in CD, confirmation of this hypothesis would depend on new comprehensive large, randomized, controlled clinical trials using different agents and doses to achieve potent inhibition of the synthesis and/or actions of this eicosanoid.

Unifying hypothesis for altered vascular physiology in IBD

The evidence presented in this review strongly suggests that the vasculature plays a central role in the pathology of human inflammatory bowel disease. In addition to the known abnormalities of leukocyte recruitment, the physiology of vascular perfusion is fundamentally altered in IBD. These data demonstrate that in chronic intestinal inflammation there are alterations in vascular perfusion, which may be dependent on location of the vascular bed (i.e. mucosal vs. serosal surface) and may be integrally linked to the phase of the disease (early disease is characterized by hyperaemia, while long-standing, remodelled bowel is associated with diminished vascular perfusion in the mucosa). Clearly, a unifying hypothesis is needed regarding the vascular pathophysiology that accounts for the mesenteric, serosal and bowel wall hyperaemia, which exists in concert with mucosal ischaemia in IBD.