Radiographic contrast media and thrombosis:
the more we know, the more we need to know

Warren K Laskey
Division of Cardiology, Uniformed Services University of
the Health Sciences, Bethesda, MD, USA
Address for correspondence:
Dr WK Laskey, MD
Division of Cardiology, A-3060
Uniformed Services University of the Health Sciences
4301 Jones Bridge Road, Bethesda, MD 20814
Tel: +1-(301)-295-3623
Fax: +1-(410)-360-7306
Email: warrenlaskey@earthlink.net

Abstract
The diverse effects of radiographic contrast media (RCM) on virtually all organ systems extend to the haemostatic system. Before the advent of non-ionic RCM, little controversy surrounded these effects, i.e. inhibition of thrombin generation, disruptive effects on arterial endothelium and inhibition of platelet aggregation. Thus, ionic high osmolar contrast media (HOCM) were viewed as having ‘anti-coagulant’ and ‘anti-thrombotic’ properties. The fact that these inhibitory properties were chiefly in vitro phenomena, that significant dilution occurred in the circulation resulting in any local (inhibitory) effects being transient and that differentiation of effects caused by ionicity or osmolality were limited by the availability of RCM other than HOCM all contributed to the origin of the ‘clotting controversy’ when non-ionic RCM was introduced.

The following review takes its motivation from the clinical perspective. Thrombosis is viewed as the end result of the pathological interaction of vessel wall, blood flow and formed elements of the blood. The interaction of ionic and non-ionic RCM with each of these elements is discussed.While consensus exists with respect to the in vitro and in vivo anti-coagulant properties of RCM (inhibition of thrombin formation), less consensus focuses on the in vitro and in vivo effects of RCM on platelet function. Furthermore, extension of the in vitro data to the clinical setting is problematic given the lack of clinical data suggesting a pro-thrombotic effect of non-ionic RCM. As many of the potentially deleterious effects on endothelial and/or platelet function are the result of either RCM ionicity or osmolality, the results of studies with non-ionic, iso-osmolar RCM are discussed.

Introduction
Radiographic contrast media (RCM) are among the most widely used pharmaceutical agents in cardiovascular medicine. However, RCM have only one desirable feature - the enhancement of radiographic contrast - and a plethora of undesirable effects.

Fortunately, the majority of the latter are transient in nature and do not pose a clinically significant hazard to the patient. Serious adverse effects of RCM may be categorised by organ specificity, e.g. RCM-related nephropathy, or more generally by system specificity, e.g. cardiovascular effects. In this review, the focus will be on the interaction of RCM with those elements of the haemostatic system in man that have clinical relevance. A basic distinction between clotting and thrombosis is made at the outset: clotting is physiological and best assessed ex vivo; thrombosis is distinctly pathological and very much an in vivo process.

Virchow's triad - the hypothesis that thrombosis is the end result of the interaction of the blood vessel wall, blood flow and formed elements of the blood - remains central to our thinking, notwithstanding the exponential increase in the knowledge base in all three areas. In fact, our increasing understanding of the inter-relationship between thrombosis and inflammation can be seen as an extension of Virchow's seminal observations. Therefore, the following discussion of the issue(s) surrounding RCM and thrombosis will focus on the three elements of Virchow's triad.

RCM and the vessel wall
The least intensively studied of the properties of RCM - their interaction with vessel wall components - has perhaps the most important clinical implications for thrombosis. Ionic RCM, irrespective of osmolality, exhibit the most disruptive effects on normal endothelium in vitro1 and ex vivo [2] while non-ionic iso-osmolar dimers appear least disruptive [1]. Not surprisingly, in the setting of abnormal endothelium, the effects of RCM should be even more pronounced. Specifically, these effects extend from abnormal histological appearance, including destruction [1,3], to the inhibition of nitric oxide production [4]. The specific effects of RCM on endothelial cell release of inflammatory cytokines and, conversely, the effects of cytokines on endothelial functions are only beginning to be appreciated. Given the increasingly evident inter-dependence between inflammation and thrombosis [5], the potential for RCM to contribute to both local thrombosis and inflammation superimposed on abnormal endothelium is of immediate relevance. Activation of von Willebrand factor and release of intercellular adhesion molecules and pro-inflammatory cytokines have all been detected in clinical studies [6-10] and confirm in vitro observations.

Injured or denuded endothelium predisposes to thrombosis via a number of pathways, most notably the activation of Factor VII by tissue factor. The critical role of platelet activation and aggregation will be discussed below but is mentioned here given the fundamental role of platelet activation by damaged endothelium. Furthermore, the ionicity and/or osmolality of RCM may be critical in the dynamic balance between local thrombosis and lysis. Controversy exists with respect to effects of RCM on the fibrinolytic system with some studies suggesting activation [11] and others indicating inhibition(Ref:12] of the lytic system. Differences between in vitro and in vivo data further confound this issue.

RCM and blood flow
In regions of high flow velocity and shear stress, platelet activation is likely to occur [13]. These abnormal haemodynamics are also seen in association with endothelial cell injury and dysfunction. Conversely, in regions of low velocity flow and low shear, stasis and activation of the clotting cascade are likely to occur. In fact, observation of the clotting cascade activation led to many of the initial concerns surrounding the facilitation of clot formation with non-ionic RCM [14]. However, the clinical implications of clot formation in static blood admixed with RCM [14] are unclear. Additional studies have shown that the anticoagulant effects of RCM were preserved in both glass and plastic syringes, irrespective of ionicity [15]. It should be emphasised that, at both extremes of shear stress, the sine qua non for thrombosis in man is the presence of an 'abnormal surface' - thus highlighting, again, the importance of endothelial integrity.

In one of the few areas of concordance between in vivo and in vitro data, both ionic and non-ionic RCM inhibit the 'clotting cascade' and, ultimately, thrombin formation.16 Prolongation of clotting times correlate with the extent of inhibition of thrombin formation in vitro. Ionic RCM, irrespective of osmolality, consistently possess greater anticoagulant properties in vitro compared with non-ionic RCM [16,17]. The relevance of this observation to the patient receiving systemic thrombin inhibitors is less clear. From a theoretical standpoint, the viscosity of RCM might contribute to altered blood flow patterns and, therefore, shear stress. However, given the extensive dilution undergone by RCM during a first pass, with resultant local concentrations generally under 1%, any significant alteration in local viscosity is unlikely [18].

RCM and formed elements of the blood
Given the central role of the platelet in physiological haemostasis and pathological thrombosis, the extensive literature on the effects of RCM on platelet function is appropriate. However, no greater area of controversy surrounding the effects of RCM exists than in the interaction of platelets with RCM. Here again, differences among in vitro studies as well as between in vitro and in vivo studies have led to considerable uncertainty about the implications of these interactions (or lack thereof) as well as their clinical relevance.

Studies in vitro were spurred by the inception of non-ionic RCM and concerns surrounding their 'pro-thrombotic' potential [14,19,20]. Unfortunately, inter-study differences in methodology, definitions and experimental conditions resulted in conflicting data with respect to the effects of RCM on 'platelet function'. This term, however, encompasses platelet adhesion, activation, degranulation and, ultimately, aggregation. It is also important to note that, in addition to these diverse measures of platelet function, clinically relevant differences exist in the extent of activation by various agonists (e.g. collagen, thrombin or thromboxane). Thus, although ionic RCM inhibited thrombin-mediated platelet 'activation' to a greater extent than non-ionic RCM, neither class of RCM activated platelets by themselves [21]. There is evidence of a 'platelet-activating effect' associated with non-ionic low osmolal RCM (degranulation and surface antigen expression) but not with ionic low osmolal RCM [22]. However subsequent studies using more physiological means of assessing platelet function (e.g. flowing blood platelet aggregometry) demonstrated the independence of degranulation and ionicity and the over-riding importance of osmolality [23]. In addition, again using a platelet function assay simulating more physiological conditions, Sakariassen et al. demonstrated a lack of concordance between degranulation and platelet thrombus formation [24,25].

Data obtained in vivo tend to support the results of studies of platelet function under 'physiological' conditions, i.e. no evidence of pro-thrombotic potential for non-ionic RCM [18,26]. The clinical literature provides strong support for the safety of non-ionic RCM during cardiac angiography [27] as well as a lack of evidence of enhanced thrombogenicity [28]. It is noteworthy that the act of vascular invasion, per se, is associated with systemic evidence of activation of coagulation and inflammation; [29-31] factors rarely controlled for in any study. Presently, the controversy centres on the role of RCM during coronary interventional procedures. These clinical circumstances represent a most extreme pro-thrombotic milieu and should facilitate the detection of the pro-thrombotic potential of RCM.

A meta-analysis of studies reported before 1999 found a reduced rate of abrupt vessel closure during interventional procedures with use of a low osmolar ionic agent (ioxaglate) compared with non-ionic low osmolar agents [32]. However, in this study no difference was observed in rates of abrupt closure between ioxaglate and the iso-osmolar, non-ionic dimer iodixanol. Furthermore, despite the large sample size, this study failed to identify a difference in an overall composite rate of serious adverse events between ioxaglate and non-ionic comparators [32]. Therefore, it is worth noting that several recent large-scale, multicentre, randomised, controlled trials have demonstrated the safety of non-ionic RCM during coronary interventional procedures [33,34].

Expanding on the importance of osmolality as well as ionicity, the Contrast Media Utilization in High Risk PTCA (COURT) trial demonstrated enhanced safety of the non-ionic, iso-osmolal dimer iodixanol compared with the ionic, low osmolal dimer ioxaglate during coronary interventional procedures [35]. These intriguing findings were also noted in the Visipaque™ vs. Isovue® in Cardiac Catheterization (VICC) trial [36] - a more contemporary version of the COURT trial. In the VICC trial, in-hospital adverse event rates during coronary intervention were lower with iodixanol compared with a non-ionic, low osmolar RCM, iopamidol. From a practical standpoint, it becomes increasingly difficult to demonstrate RCM-specific platelet dysfunction given the universal use of potent 'anti-platelet' agents during coronary intervention. Nevertheless, it remains intuitive that RCM, with the least likelihood of interfering with platelet and endothelial function, would be the most desirable, particularly in 'high-risk' settings.

The interaction of RCM with white blood cells is another area of clinical relevance. As white blood cells are rich sources of chemokines and cytokines that are both pro-inflammatory and pro-thrombotic, the effects of RCM osmolality and ionicity may have relevance at the local level. Osmolality, as a major stimulator of MAP kinase and NFkB [37,38], may exacerbate local inflammation and thrombosis. Such a result might be reflected in acute and/or short-term adverse sequelae, despite systemic anti-coagulation and anti-platelet treatment.

Conclusions
In summary, a considerable database of preclinical information indicates that all RCM possess anti-coagulant activity, i.e. they inhibit thrombin formation, and that ionic RCM exhibit greater inhibition of thrombin formation than non-ionic RCM. Furthermore, RCM have variable effects on specific measures of platelet function and although no agents are frankly pro-thrombotic, ionic RCM exhibit greater morphological and functional deterioration in endothelial cells compared with non-ionic agents, and the osmolality of RCM may be an important (negative) determinant of their pro-thrombotic potential.

The clinical literature supports the overall safety of non-ionic RCM during cardiac angiography. Recent randomised controlled clinical trials indicate that the use of the non-ionic, iso-osmolar dimer iodixanol during high-risk coronary interventional procedures is associated with a measurable benefit compared with either ionic or non-ionic RCM. Given the complexity of the relationship between vascular endothelial function, inflammation and thrombosis, further study of the effects of RCM on these determinants of procedural outcomes is needed.

References

1. Nyman U, Almen T. Effects of contrast media on aortic endothelium. Experiments in the rat with non-ionic monomeric and monoacidic dimer contrast media. Acta Radiol Suppl 1980;362:65-71.
2. Barstad RM, Buchmann MS, Hamers MJAG, et al. Effects of ionic and nonionic contrast media on endothelium and on arterial thrombus formation. Acta Radiologica 1996;37:954-61.
3. Riemann CD, Massey CV, McCarron DL, et al. Ionic contrast agent-mediated endothelial injury causes increased platelet deposition to vascular surfaces. Am Heart J 1993;125:71-8.
4. Hutcheson IR, Griffith TM, Pitman MR, et al. Iodinated radiographic contrast media inhibit shear stress-and agonist-evoked release of NO by endothelium. Br J Pharmacol 1999;128:451-7.
5. Esmon CT. Role of coagulation inhibitors in inflammation. Thromb Haemost 2001;86:51-6.
6. Blann AD, Adams R, Ashleigh R, et al. Changes in endothelial, leucocyte and platelet markers following contrast medium injection during angiography in patients with peripheral artery disease. Br J Radiol 2001;74:811-7.
7. Goldberg A, Zinder O, Zdorovyak A, et al. Diagnostic coronary angiography induces a systemic inflammatory response in patients with stable angina. Am Heart J 2003;146:819-23.
8. Laskey WK, Gellman J. Inflammatory markers increase following exposure to radiographic contrast media. Acta Radiol 2003;44:498-503.
9. Dehmer GJ, Nichols TC, Li S, et al. Effects of an ionic and nonionic contrast agent on von Willebrand factor assessed during coronary angiography. Am J Cardiol 1999;84:223-5.
10. Vermeiren GL,Willems R, Claeys MJ, et al. Influence of ionic and nonionic radiographic contrast media on leukocyte adhesion molecules. Mediators Inflamm 2003;12:269-75.
11. Chandra R, Abraham J. Preliminary studies on in vitro and in vivo effects of 50% hypaque on coagulation in man. Angiology 1973;24:199-204.
12. Farrehi PM, Zhu Y, Fay WP. An analysis of mechanisms underlying the antifibrinolytic properties of radiographic contrast agents. J Thromb Thrombolysis 2001;12:273-81.
13. Einav S, Blustein D. Dynamics of blood flow and platelet transport in pathological vessels. Ann NY Acad Sci 2004;1015:351-66.
14. Robertson HJF. Blood clot formation in angiographic syringes containing nonionic contrast media. Radiology 1987;163:621-2.
15. Grabowski EF, Kaplan KL, Halpern EF. Anticoagulant effects of nonionic versus ionic contrast media in angiography syringes. Invest Radiol 1991;26:417-21.
16. Stormorken H. Effects of contrast media on the hemostatic and thrombotic mechanisms. Invest Radiol 1988;23:S318-25.
17. Fay WP, Parker WC. Effects of radiographic contrast agents on thrombin formation and activity. Thromb Haemost 1998;80:266-72.
18. Zhu Y, Tomasiak R, Fay WP. Do clinically relevant circulating concentrations of radiographic contrast agents inhibit platelet-dependent arterial thrombosis? Thromb Res 2002;105:413-8.
19. Grollman JH Jr, Liu CK, Astone RA. Thromboembolic complications in coronary angiography associated with the use of nonionic contrast media. Cathet Cardiovasc Diagn 1988;14:159-64.
20. Idee J-M, Corot C. Thrombotic risk associated with the use of iodinated contrast media in interventional cardiology: pathophysiology and clinical aspects. Fundam Clin Pharmacol 1999;13:613-23.
21. Li X, Gabriel DA. Differences between contrast media in the inhibition of platelet activation by specific platelet agonists. Acad Radiol 1997;4:108-14.
22. Chronos NAF, Goodall AH,Wilson DJ, et al. Profound platelet degranulation is an important side effect of some types of contrast media used in interventional cardiology. Circulation 1993;88:2035-44.
23. Grabowski EF, Jang IK, Gold H, et al. Variability of platelet degranulation by different contrast media. Acad Radiol 1996;3:S485-7.
24. Sakariassen KS, Barstad RM, Hamers MJ, et al. Iohexol, platelet activation and thrombosis. I. Iohexol-induced platelet secretion does not affect thrombus formation in native blood. Acta Radiol 1998;39:349-54.
25. Sakariassen KS, Barstad RM, Hamers MJ, et al. Iohexol, platelet activation and thrombosis. II. Iohexol-induced platelet secretion does not affect collagen-induced or tissue-factor induced thrombus formation in blood that is anticoagulated with heparin and aspirin. Acta Radiol 1998;39:355-61.
26. Markou CP, Chronos NA, Hanson SR. Antithrombotic effects of ionic and nonionic contrast media in nonhuman primates. Thromb Haemost 2001;85:488-93.
27. Barrett BJ, Parfrey PS, Vavasour HM, et al. A comparison of non-ionic, low osmolality radiocontrast agents with ionic, high-osmolality agents during cardiac catheterization. N Engl J Med 1992;326:431-6.
28. Davidson CJ, Mark DB, Pieper KS, et al. Thrombotic and cardiovascular complications related to nonionic contrast media during cardiac catheterization: Analysis of 8,517 patients. Am J Cardiol 1990;65:1481-4.
29. Cousins C, Dutka DP, Bradshaw A, et al. Effect of arterial cannulation and contrast agents on blood coagulation. Acad Radiol 1995;2:663-6.
30. Biancardi M, Cimminiello C, Toschi V, et al. TAT, F1+2 and D-dimer levels in patients after coronary angiography with nonionic or ionic contrast media or after cardiac catheterization. Semin Thromb Hemost 1996;22(Suppl 1):61-5.
31. Hoffmann JJ, Tielbeek AV, Krause W. Haemostatic effects of low osmolar non-ionic contrast media: a double blind comparative study. Br J Radiol 2000;73:248-55.
32. Cucherat M, Leizorovicz A. Effects of non-ionic contrast media on abrupt vessel closure and ischemic complications after angioplasty. A meta-analysis. Am J Cardiol 1999;84:98P (Abstract).
33. Lefevre T, Adjeroud N, Royer T, et al. Influence of contrast media on the results of percutaneous transluminal coronary angioplasty with provisional stenting: a comparative study. J Invasive Cardiol 1998;10:380-4.
34. Bertrand ME, Esplugas E, Piessens J, et al. Influence of a nonionic, iso-osmolar contrast medium (iodixanol) versus an ionic, low-osmolar contrast medium (ioxaglate) on major adverse cardiac events in patients undergoing percutaneous transluminal coronary angioplasty: A multicenter, randomized, double-blind study. Visipaque in Percutaneous Transluminal Coronary Angioplasty. Circulation 2000;101:131-6.
35. Davidson CJ, Laskey WK, Hermiller JB, et al. Randomized trial of contrast media utilization in high-risk PTCA: The COURT trial. Circulation 2000;101:2172-7.
36. Harrison, KJ, Hermiller JB, Vetrovec GW, et al. A randomized study of 1276 patients undergoing PCI using iodixanol (Visipaque) vs. iopamidol (Isovue): Comparison of in-hospital and 30 day major adverse cardiac events. The results of the VICC trial. Circulation 2003;108(Supplement IV):IV-354-5.
37. Shapiro L, Dinarello CA. Osmotic regulation of cytokine synthesis in vitro. Proc Natl Acad Sci U S A 1995;92:12230-4.
38. Hoffmeyer A, Grosse-Wilde A, Flory E, et al. Different mitogenactivated protein kinase signaling pathways cooperate to regulate tumor necrosis factor alpha gene expression in T lymphocytes. J Biol Chem 1999;274:4319-27.

December 2004, 1098/OS

 

Source: C2I2, Winter 2004

Read also - related articles