Kidney International, Vol. 68 (2005), pp. 14–22

PERSPECTIVES IN BASIC MEDICINE

Pathophysiology of contrast medium–induced nephropathy

PONTUS B. PERSSON, PETER HANSELL, and PER LISS

Institute of Physiology, Humboldt University, Berlin, Germany; Section of Integrative Physiology, Department of Medical CellBiology, University of Uppsala, Uppsala, Sweden; and Section of Diagnostic Radiology, Department of Oncology, Radiology andClinical Immunology, Academic Hospital, University of Uppsala, Uppsala, Sweden

Pathophysiology of contrast medium–induced nephropathy.Background. Contrast medium–induced nephropathy (CIN)

is a well-known cause of acute renal failure, but the develop-ment of CIN remains poorly understood. A number of studieshave been performed with the one aim, to shed some light ontothe pathophysiology of CIN. These have led to manifold inter-pretations and sometimes contradicting conclusions.

Methods. This review critically surveys mechanisms believedto mediate CIN by highlighting the complex pathophysiologicentity, including altered rheologic properties, perturbation ofrenal hemodynamics, regional hypoxia, auto- and paracrinefactors [adenosine, endothelin, and reactive oxygen species(ROS)], and direct cytotoxic effects. Moreover, the importanceof physicochemical properties of contrast media are made clear.

Results. The more recently developed iso-osmolar contrastmedia are dimers, not monomers as the widely used nonioniclow osmolar contrast media. The dimers have physicochemicalfeatures different from other contrast media which may be ofclinical importance, not only with respect to osmolality. Theviscosity of the commercially available dimers is considerablyhigher than blood.

Conclusion. Many experimental studies provide evidence fora greater perturbation in renal functions by dimeric contrastmedia in comparison to nonionic monomeric contrast media.Clinical trials have yielded conflicting results.

Interventional techniques, fast multislice computer to-mographies (CTs) and new three-dimensional recon-struction techniques have increased the use of iodinatedintravascular contrast media over the last decades. Today,approximately 60 million doses are applied per year. Al-though new techniques such as magnetic resonance imag-ing (MRI) and ultrasound have been introduced, and thehazard of x-ray radiation is evident, the majority of ex-aminations require iodinated contrast media for accurateand safe diagnosis and interventional procedures.

Key words: contrast media, nephropathy, viscosity, osmolality.

Received for publication June 4, 2004and in revised form August 10, 2004, and August 24, 2004Updated September 13, 2004Accepted February 21, 2005

C© 2005 by the International Society of Nephrology

This article is aimed at reviewing the mechanisms un-derlying contrast medium–induced nephropathy (CIN).The term CIN implies impairment in renal functionoccurring within 3 days following the intravascular ad-ministration of contrast medium and the absence of analternative etiology [1, 2]. An increase in serum creatinineby more than 25% or 44 lmol/L−1 (0.5 mg/100 mL) within48 to 72 hours of contrast administration is often takenas a marker for the occurrence of CIN [3–6]. The serumcreatinine concentration typically peaks on the second orthird day after exposure to contrast medium and usuallyreturns to the baseline value within 2 weeks [7, 8].

Since the 1950s, the various available contrast me-dia have been based on triiodobenzene. They arecommonly grouped according to their osmolality andionicity. In the era when high-osmolar contrast media(having osmolalities approximately six times higher thanthe plasma) were widely used, the differentiation withregard to osmolality made sense, although it has becomeclear that many of the side effects were caused by theelectric charge. Today, only the low-osmolar contrast me-dia (which still have considerably higher osmolality thanplasma) and iso-osmolar contrast media are widespread.It appears that the subdivision of contrast media accord-ing to their osmolality may require reconsideration, sinceiso-osmolar contrast media are dimers. The currentlyavailable dimeric contrast media reveal greater viscositiesthan the monomeric low-osmolar contrast media (Fig. 1).This can have important implications for renal and sys-temic hemodynamics as outlined below.

PATHOPHYSIOLOGY OF CIN

The underlying mechanism to CIN is not clear, thoughseveral suggestions have been put forward. Most likely, acombination of various mechanisms are responsible forthe development of CIN [9]. A reduction in renal perfu-sion caused by a direct effect of contrast media on thekidney and toxic effects on the tubular cells are generallyaccepted as the main factors in the pathophysiology ofCIN. However, the pathophysiologic relevance of directeffects of contrast media on tubular cells is contentious[9], as are the other proposed etiologies.

14

Persson et al: Contrast-induced nephropathy 15

4

7

10

Vis

cosi

ty, m

Pa×

s, 3

7˚C

200 300 400 500 600 700 800Osmolality, mOsm/kg H2O

Iodixanol

Iotrolan

Iomeprol

Iopamidol

Iopromide

Iohexol

Fig. 1. Osmolality and viscosity for I-concentration of 300 mg/mL.Blood viscosity refers to prearteriolar values. Viscosity and osmolalityvalues taken from information provided by the respective distributor.∗In case of Iodixanol the viscosity value for 300 mg I/mL was extrapo-lated from the values of the 150, 270, and 320 mg I/mL solutions. Theosmolality of Iodixanol is constant for all three concentrations.

Among the discussed mechanisms behind CIN thatwill be outlined here are rheologic alterations, activationof the tubuloglomerular feedback response, regional hy-poxia, cytotoxic effects on the renal epithelial cells, gen-eration of reactive oxygen species (ROS), and, finally,increased adenosine or endothelin production.

Rheology

One of the particularities of the renal vascular bed isthe length of the vessels that supply the renal medulla withblood. These vasa recta have the same diameter as usualcapillaries, but they are severalfold longer than averagecapillary vessels, meaning that vascular resistance is high.To offset the increased resistance caused by the vessellength, the viscosity of the blood flowing through the vasarecta is maintained very low. This is due to two effects.First, as with all capillaries, the blood flowing throughthese vessels has a low hematocrit since the erythrocyteshave higher flow velocities (Fahraeus-Lindqvist effect).Thus, blood viscosity in the capillary is not much higherthan plasma viscosity. The second effect is that of plasmaskimming. The afferent arterioles branch off from theinterlobular arteries at an almost right angle. Since theerythrocytes are concentrated in the center of the inter-lobular arteries (laminar flow), the plasma rich blood nearthe endothelium is skimmed off into the juxtamedullaryafferent arterioles.

It is often thought that iso-osmolar contrast media aresuperior to low-osmolar agents, since they would not in-crease resistance (R) to a similar extent. This is not true,as indicated by Poiseuille’s law (equation 1). Osmolalityplays no role for blood flow while the viscous properties(g) are decisive:

0

20

40

60

PO

2, m

m H

g

−20 −10 0 10 20 30Time, minutes

Injection

Fig. 2. Medullary hypoxia induced by contrast media [ioxaglate (�),iopromide (�), and iotrolan (�)] in comparison to Ringer’s solution(�). Reduction in pO2 is greatest for iotrolan (iso-osmolal nonionicdimer) followed by ioxaglate (low-osmolal ionic dimer). Iopromide(low-osmolal monomer) had the least effect of the contrast media (fromLiss et al, 1998, with permission).

R = g ∗ 8 ∗ l/p ∗ r4 (Equation 1)

Thus blood flow (Q) through the vasa recta is

Q = �P ∗ p ∗ r4/g ∗ 8 ∗ l (Equation 2)

where �P is the pressure gradient, g is viscosity, l refersto the length of the vessel and r is the radius.

Among the monomeric contrast media, there is a rela-tionship between osmolarity and viscosity (Fig. 1). How-ever, the commercially available iso-osmolar contrastmedia exhibit considerably higher viscosity. Thus, iso-osmolar contrast media should impair renal medullaryblood flow to a greater extent than low-osmolar agents.Indeed, this seems true, as indicated by the particularlyreduced pO2 levels caused by iso-osmolar contrast mediain rats [10] (Fig. 2).

However, not only the intrinsic viscosity of the fluidis important, but also their interaction with blood con-stituents. For instance, high-osmolar ionic agents dimin-ish erythrocyte deformability, thereby increasing stiffnessand making it more difficult for the red blood cells toflow through the capillaries [11, 12]. Thus, trapping oc-curs, meaning that the erythrocytes are densely packedin the renal capillaries (e.g., vasa recta), and the bloodflow through these vessels may cease, as shown for therat [13, 14].

The adverse effects of augmented fluid viscosity by theuse of dimeric contrast media may be more pronouncedin the renal tubules than in the capillaries. Normally, thetubular fluid has a lower viscosity than plasma, since theultrafiltrate contains very few plasma proteins. As seen

16 Persson et al: Contrast-induced nephropathy

Plasma viscosityTubular fluidreabsorption

Vas rectaresistance

Renal tubularviscosity

Renal tubularobstruction

Renal interstitialpressure

GFR

Tubular damage

Renal medullaryhypoxia

Vas rectaperfusion

Fig. 3. Flow chart of mechanisms linking fluid osmolality to renal dam-age. GFR is glomerular filtration rate.

in the rat, use of dimeric contrast media will increasetubular fluid viscosity thereby increasing the resistanceto flow in renal tubules [15]. Tubular viscosity will in-crease markedly toward the distal sections of the kidneydue to fluid reabsorption. There is an exponential rela-tionship between concentration and viscosity; thus, whenurine becomes very concentrated, tubular fluid viscositywill increase dramatically and tubular plugging will occur.Hydration attenuates fluid reabsorption in the collectingducts and is therefore very beneficial. Ueda et al [16]measured tubular pressures after giving various contrastmedia. The dimeric contrast media increased tubu-lar pressures over the entire observation period (50minutes) to over 40 mm Hg. Accordingly, intrarenalinterstitial pressure may significantly rise as well, therebyimpairing renal blood flow through the medulla. More-over, such high tubular pressures as seen after the applica-tion of dimeric contrast media will considerably diminishglomerular filtration rate (GFR).

A scheme of the adverse effects of pronounced in-creases in viscosity on the kidney is presented inFigure 3. Taking these considerations into account, it isimportant to note that contrast media with high viscos-ity (i.e., dimeric iso-osmolar contrast media) should be

prewarmed before infusion, since this markedly reducesviscosity.

The tubuloglomerular feedback

The tubuloglomerular feedback is a powerful mech-anism in the control of renal vascular resistance andglomerular filtration. A popular explanation for thedevelopment of CIN is that hyperosmotic contrastmedia cause diuresis, which activates the tubuloglomeru-lar feedback and subsequently compromises renal bloodflow and glomerular filtration. However, this osmotic di-uresis theory is not a likely explanation for CIN. The mac-ula densa cells of the thick ascending limb sense Na+, K+,and Cl− concentrations in the tubular fluid via the Na+-K+-2 Cl− cotransporter. This transporter is effectivelyblocked by furosemide. The affinity for Cl− is very low,so in a physiologic setting there will always be enoughNa+ and K+ to keep the system running, Cl− is the limit-ing factor [17, 18]. As shown already by pioneer experi-ments with retrograde perfusions of the tubule, osmolal-ity has no effect on the tubuloglomerular feedback. Withorthograde perfusion, quite a lot of transport occurs be-tween tubular fluid and interstitium, and even nonionicfluids may occasionally be able to elicit tubuloglomeru-lar feedback response. This would leave some room for apossible contrast media effect. In this case, however, theCIN potential of a certain contrast media would not relyon its osmolality, but rather on other structural features.

The ruling out of the osmotic diuresis theory is furthersupported by experiments using mannitol, an osmotic di-uretic. Increases in osmolality, such as after mannitol in-fusion or after contrast media application, decrease NaClconcentration at the macula densa, however, simultane-ously increasing tubular flow. Therefore, the resulting netchange in the amount of NaCl passing the macula densais negligible [19].

Finally, blocking the tubuloglomerular feedback byfurosemide does not decrease serum creatinine after ap-plication of contrast media, which is usually the param-eter taken to indicate CIN [2]. Thus, taken together, thetheory that the osmolality of a contrast media causes CINvia the tubuloglomerular feedback does not appear likely.

Regional hypoxia

Kidney perfusion is very high for the cortex, but themedullary portions are maintained at the verge of hy-poxia where pO2 levels can be as low as 20 mm Hg [20].This is the price paid for upholding the countercurrentmechanism for controlling urine excretion. A particu-larly vulnerable kidney region is the deeper portion ofthe outer medulla, an area remote from the vasa rectasupplying the renal medulla with blood. It is here thatthe thick ascending limbs of the loop of Henle exhibithypoxic damage (e.g., when the kidney is perfused with

Persson et al: Contrast-induced nephropathy 17

erythrocyte-free medium) [21]. The reason for the vul-nerability of the outer medullary portion of the nephronis the relative high oxygen requirements due to saltreabsorption.

Adding contrast media to the medium aggravates hy-poxic injury to this region (Fig. 3), probably by increas-ing renal vascular resistance, as in the rat [22]. It hasbeen shown also in the rat that the iso-osmolar contrastmedium, iodixanol (a dimer with high viscosity), reducesblood flow to all regions of the kidney to a greater ex-tent than low-osmolar, and even high-osmolar contrastmedia. However, this decrease in perfusion was probablydue to profound systemic effects of iodixanol, since bloodpressure dropped considerably [23].

Iothalamate, a high-osmolar agent, markedly reducesmedullary pO2 to about a third of control levels [24]. Re-markably, in the rat, the iso-osmolar contrast mediumiotrolan impairs local pO2 to a greater extent than thelow-osmolar contrast medium iopromide [10]. The de-crease in pO2 by the latter failed to reach statistical sig-nificance (Fig. 2). This underscores the shortcoming ofclassifying contrast medium simply by their osmolality.

A second factor that has been thought to mediate CINis an increased oxygen demand due to an augmentedworkload for the tubular cells. This hypothesis is not read-ily understood, since contrast media are not reabsorbedand bind to water, which is excreted with the contrast me-dia. Thus, the net NaCl load remains the same. Indirectly,however, one may be able to explain an enhanced work-load to the tubular cells. First, there is a transient increasein GFR after giving contrast media [25], and second, os-motic diuresis may reduce the paracellular reabsorptionof the proximal tubule, leading to larger amounts of NaClhaving to be taken up in the more distal segments. Agmonet al [26] have shown that contrast media can actually in-crease medullary blood flow to the kidney, even thoughpO2 decreases, which is supported also by a study ofHeyman et al [27]. These two studies indeed suggest thatan increased oxygen demand has taken place after con-trast media application.

Local renal hypoxia can be aggravated by the systemiceffects of some contrast media, such as transiently re-duced cardiac output [11], and suboptimal pulmonaryperfusion-ventilation relationship [28]. Moreover, oxy-gen delivery to the peripheral tissues may be impaired,since contrast media can increase oxygen affinity ofhemoglobin [29].

If renal outer medullary hypoxia causes CIN, block-ing the transporters in this nephron segment should havebeneficial effects on its prevention. The bulk of trans-port taking place in the medullary thick ascending limbis the Na+-K+-2 Cl− transporter, which, as mentionedabove, is blocked by furosemide. Blocking the transportwould dramatically lower local oxygen consumption andalleviate the reduced oxygen supply. In fact, this has been

demonstrated to occur in experiments in rats showing thatouter medullary pO2 is elevated after furosemide [30].However, contrast medium injection after furosemidestill reduces outer medullary pO2 although occurringat higher absolute pO2 values. Also in the rat model,Heyman et al [31] were able to demonstrate an attenua-tion of thick ascending limb damage induced by contrastmedia.

Nevertheless, furosemide given to patients just beforeangiography fails to limit increases in serum creatinine af-ter contrast media application, indicating that yet othermechanisms are involved in CIN [2]. However, atten-tion must be paid to replenish the fluid losses induced byfurosemide, otherwise the dehydration may simply over-ride a potential beneficial effect of furosemide.

Cytotoxic effects on renal tubular (epithelial) cells

In vitro investigations on cell lines are commonly usedfor assessing renal tubular cell function or damage. Aporcine proximal tubular cell line, LLC-PK1, was usedby Hardiek et al [32] to investigate CIN. An effect onapoptosis was not found, though, proliferation was im-paired. Reduced proliferation will affect renal functionwith a delay of hours to days, which may help explain theclinical course of CIN. Independent of the contrast me-dia used, tubular cell damage can occur. Vacuolizationas described by Andersen, Christensen, and Vik [33] forisolated cells is a morphologic hallmark rather than anindicator of damage. A more specific alteration of proxi-mal tubular function seems to be a perturbation of mito-chondrial enzyme activity and mitochondrial membranepotential [32] (Fig. 4). Indeed, attenuated mitochondrialenzyme activity is supported by the observed increase inadenosine following application of contrast media (seebelow). The extent of mitochondrial enzyme activity im-pairment relies primarily on two features of the contrastmedia: ionicity and the molecular structure. Remarkably,low-osmolar (monomeric) contrast media had the leasteffect, followed by the iso-osmolar (dimeric, nonionic)agents. Ionic compounds revealed the most profound ef-fects [32].

In the distal tubule, contrast media may induce apop-tosis, as indicated in the Madin-Darby canine kidney(MDCK) cell line model [34]. In part, this seems to rely onhypoxic damage [35]; however, there is also a direct influ-ence on these cells [34]. Contrast media can also open theintercellular junctions and affect the polarity of the ep-ithelial cell surface [36]. These features are important fornormal fluid and electrolyte reabsorption and may add tothe potential deleterious effects of contrast media.

ROS

Even under normal conditions, oxygen radicals are pro-duced endogenously, but the levels increase during oxida-tive stress. Among the most common oxygen radicals are

18 Persson et al: Contrast-induced nephropathy

0

25

50

75

100

125

MT

T r

educ

tion,

per

cent

of c

ontr

ol

0 25 50 75 100Contrast media, mg I/mL

IopamidolIomeprolIodixanol

IoxaglateDiatrizoate

Fig. 4. Altered mitochondrial function in a proximal tubular cell lineas determined by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazoliumbromide (MTT) reduction (24-hour treatment). A comparison of theeffects of various contrast media on MTT reduction reveal significantdifferences from one another. The least influence was found by the low-osmolar agents, followed by the iso-osmolar contrast media (Iodixanol).The ionic substances showed the greatest effect (from Hardieck et al,2001, with permission).

superoxide (O2−), hydrogen peroxide (H2O2), and hy-

droxyl radical (OH−) [37]. Superoxide and hydroxyl rad-ical are more reactive than H2O2, which is not a radical,but exhibits a greater membrane permeability.

There are marked differences between the nephropa-thy induced by streptozotocin in rats and human diabeticnephropathy [38]. This should be kept in mind whencomparing investigations on diabetes mellitus, which isone of the most prominent risk factors for CIN. Indiabetic nephropathy, endothelial dysfunction in renalvessels is a common sequela. It seems that the tonic in-fluence of nitric oxide in the renal microvasculature issuppressed and contributes to the endothelial dysfunc-tion in the early stages of insulin-dependent diabetes[37]. Superoxide rapidly scavenges nitric oxide and couldtherefore explain the attenuated nitric oxide activity inthe diabetic renal microvasculature. In support of thishypothesis, superoxide production was found to be in-creased in renal cortical tissue from diabetic rats [39].Moreover, Ohishi and Carmines [40] demonstrated thatfor juxtamedullary nephrons of streptozotocin-diabeticrats, the afferent and efferent arteriolar vasoconstrictorresponse to nitric oxide synthase (NOS) inhibition is im-paired. Furthermore, in a recent study by Palm et al [41],scavenger treatment (vitamin E) normalizes the reducedpO2 found in the renal medulla of streptozotocin-diabeticrats. Since nitric oxide inhibits oxygen consumption, it istempting to speculate that reduced (scavenged) nitric ox-ide during diabetes elevates oxygen consumption therebyleading to reduced pO2 with consequences for

endothelial-epithelial structure and function. Diabeticnephropathy may be of importance with regard to theconclusion of the NEPHRIC study [1] that the use ofiso-osmolar contrast media, as opposed to low-osmolarcontrast media, results in reduced incidence of CIN. Theconclusion of that study is not in line with our currentunderstanding of CIN and may rely on the statisticallysignificant difference in the duration of diabetes betweenthe groups.

ROS may play a role in the effects of various vaso-constrictors that have been considered important for thedevelopment of CIN. Since ROS are extracellular sig-naling molecules, they may be significant in mediatingthe actions of vasoconstrictors, such as angiotensin II,thromboxane A2 (TXA2), endothelin-1, adenosine, andnorepinephrine. Moreover, various models of renal in-flammation and ischemia have shown a role of ROS inglomerular injury. The adverse effects of contrast mediaon renal function may therefore involve the generation ofROS (e.g., via adenosine formation). This notion is sup-ported by experiments in which the generation of ROSwas inhibited by allopurinol, or the amount of ROS wasreduced by superoxide dismutase. In these models, con-trast media–induced reductions in GFR are attenuated[42]. Later studies performed in humans further under-score a role of ROS in CIN [43].

Taking the evidence for a role of ROS in CIN into ac-count, it is not surprising that clinical trials have beenperformed with the aim to ameliorate CIN by scavengingROS [44–48]. In these trials, N-acetylcysteine was given inaddition to the general hydration protocols and showed apositive outcome in four of the studies [44, 45, 47, 49] andhas therefore been recommended for the prevention ofCIN in patients with mild-to-moderate renal insufficiency[50–52]. However, this recommendation is not unequiv-ocal since other trials fail to confirm a positive effect [48,53] or indicate that rigorously controlled larger trials arestill required [54].

Adenosine

Direct actions of adenosine on the renal vasculaturehave been discussed with regard to CIN. In the kidney,adenosine exerts a vasoconstrictor response of the af-ferent arteriole, due to the predominance of A1 recep-tors [55]. Early studies indicated the existence of bothA1 and A2A receptors in the kidney, which were foundto be widely distributed throughout the renal vascula-ture, juxtaglomerular apparatus, glomeruli, tubules, andcollecting ducts [56, 57]. Besides its vasoconstrictor ef-fect, A1 receptor stimulation contracts mesangial cellsin the glomerulus [58]. In 1982, Osswald, Hermes, andNabakowski [59] proposed that kidney hemodynamicsis under metabolic control and suggested adenosine asthe mediator of the tubuloglomerular feedback due to

Persson et al: Contrast-induced nephropathy 19

its particular vasoconstrictor response in renal circula-tion. This hypothesis has been considerably substantiatedby experiments demonstrating lacking tubuloglomerularfeedback responses in mice devoid of adenosine A1 re-ceptors [60–62].

Due to the prominent role of adenosine in the renal vas-cular bed, several studies have been performed targetingat the role of adenosine in CIN [25, 63, 64]. In diabetesmellitus, an even higher sensitivity of the renal vascula-ture to adenosine is found, thus, it has been suggestedthat adenosine is an important contributor to CIN inpatients suffering of this metabolic disorder [65]. In spiteof the strategic role of adenosine on renal function, therole it plays in CIN appears to be overestimated. Regard-ing the depression in outer medullary blood flow andoxygen tension caused by injection contrast media, theadenosine A1 receptor is not involved, as shown in a re-cent study in normal rat [66]. In that study, a specific A1 re-ceptor antagonist was given together with contrast media.Although a pronounced basal influence of A1 receptorson renal medullary hemodynamics was confirmed, block-ing these receptors failed to alleviate medullary hypoper-fusion and hypoxia in response to the contrast media. Infurther support of the limited role of adenosine in CIN, itwas found also that the general reduction in renal plasmaflow and GFR by contrast media is not attributable toenhanced adenosine action [63], but rather may involvemesangial cell contraction. However, it should be keptin mind that the A2 receptor enhances medullary bloodflow [67] and may therefore be a potential target for pre-venting CIN.

Endothelin

The effects of endothelin on vascular beds is verydependent upon the receptor subtype activation. En-dothelin (ET)-A receptor stimulation elicits pronouncedvasoconstriction, whereas the ET-B receptor has theopposite effect. The latter likely involves endothelin-dependent nitric oxide release. However, recently, bothsubtypes of receptors were found to mediate the vasocon-strictor action of endothelins in human blood vessels [68].The net vasoactive response to endothelin is believed tovary depending on the vascular bed in question.

An involvement of endothelin in CIN appears likelydue to the enhanced endothelin levels in plasma andurine, which is observed after radiocontrast application[69–71]. In addition, the transcription and release of en-dothelin from endothelial cells is enhanced by contrastmedia (for review see [72]). Moreover, in patients suffer-ing of impaired renal function, the increase in endothe-lin after giving radiocontrast is exaggerated [73]. How-ever, the aggregate effect of endothelin in the scenarioof CIN may not be as disadvantageous as one may as-sume from the findings mentioned above. As shown by

Wang et al [74], when both ET-A and ET-B receptors areblocked in humans receiving contrast media, the meanincrease in serum creatinine concentration is significantlygreater in patients receiving the ET-A/ET-B blocker com-pared with those who received placebo. Moreover, theCIN incidence is significantly higher in patients who re-ceive this blocker compared to placebo [74].

A potential beneficial effect of endothelin in the de-velopment of CIN may be explained by the ET-B medi-ated effects (e.g., vasodilatation). Accordingly, a selectiveET-A receptor blockade could prove to be effective in theprevention of CIN. In the study of Freed et al [75] usingthe unselective blocker, plasma endothelin-1 levels mayhave increased as shown by a study employing a simi-lar intravenous infusion of the same endothelin recep-tor antagonist, SB 209670. This increase in endothelin-1concentration is probably brought about by the ET-B re-ceptor antagonism [76], since one of the ET-B–mediatedeffects is the attenuation of further endothelin release.Thus, the potentiation of CIN induced by SB 209670 inthe study of Wang et al could be explained by ET-B re-ceptor blockade increasing plasma endothelin which thenacts on the ET-A receptor, as suggested by Haylor et al[77]. Indeed, a positive effect of ET-A selective block-ade on the renal outer medullary hypoxic response tocontrast media has been reported in the normal rat [78].Remarkably, the hypoxia to this kidney region was alle-viated without enhancing local blood flow. Hence, it wasconcluded that the oxygen requirements must have de-creased due to ET-A antagonism. This, in fact, may be animportant key for understanding the role of endothelin inCIN: It appears that BQ123, a selective ET-A antagonist,inhibits Na+/K+-ATPase activity [59, 78]. If this were tooccur in the thick ascending limbs of the loop of Henle, areduction in oxygen demand in the outer medulla wouldbe readily understood.

PREVENTION OF CIN

Unfortunately, the treatment procedures to preventCIN remain to be established. Several approaches of CINprevention have been reported, of which vigorous hy-dration may be the most important [79, 80]. Trials us-ing diuretics, dopamine, calcium channel blockers, atrialnatriuretic peptides, acetylcysteine, dopamine-1 receptoragonist fenoldopam and theophyllin have yielded con-trasting results, which have been reviewed in extent veryrecently [81].

Only periprocedural hydration is widely accepted toprevent CIN [2, 82, 83] and intravenous hydration may bebetter than oral hydration. Administration of saline leadsto an isotonic hydration, thus, the fluid remains in theextracellular space. Giving water will lead to a hypotonichydration, which mainly affects the intracellular space.Orally administered saline is excreted more rapidly. Thismay rely on the intestinal-renal endocrine axis for the

20 Persson et al: Contrast-induced nephropathy

maintenance of sodium balance by uroguanylin, a peptidehormone that regulates sodium excretion by the kidneywhen excess NaCl is consumed [84].

The reasons for the success of hydration in prevent-ing CIN are not related to an increase in renal bloodflow or GFR (as sometimes thought [85]). Unless thepatient is severely dehydrated, volume loading has littleeffect of these hemodynamic measures. It appears morelikely that medullary perfusion is increased when well hy-drated (autoregulation may not be present under theseconditions [86] and suppressed vasopressin levels aug-ment medullary blood flow [87, 88]), thus, regional pO2

is enhanced. Furthermore, the reduced concentration ofcontrast medium in the tubular system of the medulla dur-ing significant diuresis should be important and presum-ably override any smaller differences in injected contrastmedium osmolality or viscosity.

Large clinical studies and meta analyses have indicatedthat the use of low-osmolar contrast medium substan-tially reduces the risk of nephropathy in high-risk pa-tients as compared with the use of high-osmolar contrastmedium [4, 5, 74, 89, 90]. Nevertheless, the use for otherdiagnostic techniques (i.e., MRI, ultrasound) must alwaysbe considered. Using carbon dioxide has also been sug-gested for replacing iodinated contrast media [91] since itresults in lower incidence of CIN. However, special equip-ment would be required and other adverse reactions canappear. Moreover, carbon dioxide does not provide a sim-ilar attenuation as conventional contrast medium whichmay pose a problem in situations where high contrastproperties are crucial for safe and accurate diagnosis andinterventional procedures.

CONCLUSION

There seems to be no single cause for CIN. Severalpathophysiologic mechanisms may add up to impair kid-ney function. The use of newer contrast media and ex-tracellular fluid volume expansion are to be preferred inpatients with preexisting renal impairments [2, 79, 92].

With regard to the current concepts to explain CIN, therheologic properties of a fluid may not have received suf-ficient attention. Resistance depends on fluid viscosity,not osmolality (Poiseuille’s law). Moreover, osmolalitydoes not directly affect the tubuloglomerular feedback asalready shown by the pioneer work in this field [17, 18].Thus, perhaps too much attention has been directed to theosmolality of different contrast media, while neglectingthe impact of other physicochemical properties. Indeed,there is little experimental evidence that would supportthe notion that iso-osmolar contrast media are superiorto the low-osmolar agents in preventing CIN. In fact, thecontrary has been demonstrated. Furthermore, duringsimilar conditions of appropriate periprocedural hydra-tion (extracellular volume expansion) a possible advan-

tage of any iso-osmolar vs. a low-osmolar contrast mediashould be, if anything, minimal. Further studies compar-ing iso- with low-osmolar contrast media in risk patientsare required before any conclusions can be drawn as tothe possible superiority of certain contrast media. In theserequested studies, the importance of well-controlled andsufficient hydration status cannot be overestimated.

Reprint requests to Professor Dr. Pontus B. Persson, Institut fur Veg-etative Physiologie, Humboldt Universitat, Berlin Medizinische Fakultat(Charite) Tucholskystr. 2 10117 Berlin, Germany.E-mail: pontus.persson@charite.de

REFERENCES

1. ASPELIN P, AUBRY P, FRANSSON SG, et al: Nephrotoxic effects in high-risk patients undergoing angiography. N Engl J Med 348:491–499,2003

2. SOLOMON R, WERNER C, MANN D, et al: Effects of saline, mannitol,and furosemide to prevent acute decreases in renal function inducedby radiocontrast agents. N Engl J Med 331:1416–1420, 1994

3. BARRETT BJ, PARFREY PS, VAVASOUR HM, et al: Contrast nephropa-thy in patients with impaired renal function: High versus low osmo-lar media. Kidney Int 41:1274–1279, 1992

4. RUDNICK MR, GOLDFARB S, WEXLER L, et al: Nephrotoxicity of ionicand nonionic contrast media in 1196 patients: A randomized trial.The Iohexol Cooperative Study. Kidney Int 47:254–261, 1995

5. TALIERCIO CP, VLIETSTRA RE, ILSTRUP DM, et al: A randomizedcomparison of the nephrotoxicity of iopamidol and diatrizoate inhigh risk patients undergoing cardiac angiography. J Am Coll Car-diol 17:384–390, 1991

6. MANSKE CL, SPRAFKA JM, STRONY JT, WANG Y: Contrast nephropa-thy in azotemic diabetic patients undergoing coronary angiography.Am J Med 89:615–620, 1990

7. WAYBILL MM, WAYBILL PN: Contrast media-induced nephrotoxic-ity: Identification of patients at risk and algorithms for prevention.J Vasc Interv Radiol 12:3–9, 2001

8. LEVY EM, VISCOLI CM, HORWITZ RI: The effect of acute renal failureon mortality. A cohort analysis. JAMA 275:1489–1494, 1996

9. THOMSEN HS, MORCOS SK: Contrast media and the kidney: Euro-pean Society of Urogenital Radiology (ESUR) guidelines. Br J Ra-diol 76:513–518, 2003

10. LISS P, NYGREN A, ERIKSON U, ULFENDAHL HR: Injection of low andiso-osmolar contrast medium decreases oxygen tension in the renalmedulla. Kidney Int 53:698–702, 1998

11. DAWSON P: Cardiovascular effects of contrast agents. Am J Cardiol64:2E–9E, 1989

12. SCHIANTARELLI P, PERONI F, TIRONE P, ROSATI G: Effects of iodinatedcontrast media on erythrocytes. I. Effects of canine erythrocytes onmorphology. Invest Radiol 8:199–204, 1973

13. LISS P, NYGREN A, OLSSON U, et al: Effects of contrast media andmannitol on renal medullary blood flow and red cell aggregation inthe rat kidney. Kidney Int 49:1268–1275, 1996

14. NYGREN A, HELLBERG O, HANSELL P: Red-cell trapping in the ratrenal microcirculation induced by low-osmolar contrast media andmannitol. Invest Radiol 28:1033–1038, 1993

15. UEDA J, NYGREN A, HANSELL P, ERIKSON U: Influence of contrastmedia on single nephron glomerular filtration rate in rat kidney. Acomparison between diatrizoate, iohexol, ioxaglate, and iotrolan.Acta Radiol 33:596–599, 1992

16. UEDA J, NYGREN A, HANSELL P, ULFENDAHL HR: Effect of intra-venous contrast media on proximal and distal tubular hydrostaticpressure in the rat kidney. Acta Radiol 34:83–87, 1993

17. SCHNERMANN J, PLOTH DW, HERMLE M: Activation of tubulo-glomerular feedback by chloride transport. Pflugers Arch 362:229–240, 1976

18. BRIGGS JP, SCHNERMANN J, WRIGHT FS: Failure of tubule fluid os-molarity to affect feedback regulation of glomerular filtration. AmJ Physiol 239:F427–F432, 1980

19. LEYSSAC PP, HOLSTEIN-RATHLOU NH, SKOTT O: Renal bloodflow, early distal sodium, and plasma renin concentrations during

Persson et al: Contrast-induced nephropathy 21

osmotic diuresis. Am J Physiol Regul Integr Comp Physiol279:R1268–R1276, 2000

20. BREZIS M, ROSEN S: Hypoxia of the renal medulla—Its implicationsfor disease. N Engl J Med 332:647–655, 1995

21. BREZIS M, ROSEN S, SILVA P, EPSTEIN FH: Selective vulnerability ofthe medullary thick ascending limb to anoxia in the isolated perfusedrat kidney. J Clin Invest 73:182–190, 1984

22. HEYMAN SN, BREZIS M, REUBINOFF CA, et al: Acute renal failurewith selective medullary injury in the rat. J Clin Invest 82:401–412,1988

23. LANCELOT E, IDEE JM, COUTURIER V, et al: Influence of the viscosityof iodixanol on medullary and cortical blood flow in the rat kidney:A potential cause of nephrotoxicity. J Appl Toxicol 19:341–346, 1999

24. HEYMAN SN, REICHMAN J, BREZIS M: Pathophysiology of radio-contrast nephropathy: A role for medullary hypoxia. Invest Radiol34:685–691, 1999

25. ARAKAWA K, SUZUKI H, NAITOH M, et al: Role of adenosine in therenal responses to contrast medium. Kidney Int 49:1199–1206, 1996

26. AGMON Y, PELEG H, GREENFELD Z, et al: Nitric oxide and prostanoidsprotect the renal outer medulla from radiocontrast toxicity in therat. J Clin Invest 94:1069–1075, 1994

27. HEYMAN SN, GOLDFARB M, CARMELI F, et al: Effect of radiocontrastagents on intrarenal nitric oxide (NO) and NO synthase activity.Exp Nephrol 6:557–562, 1998

28. NEAGLEY SR, VOUGHT MB, WEIDNER WA, ZWILLICH CW: Transientoxygen desaturation following radiographic contrast medium ad-ministration. Arch Intern Med 146:1094–1097, 1986

29. KIM SJ, SALEM MR, JOSEPH NJ, et al: Contrast media adversely affectoxyhemoglobin dissociation. Anesth Analg 71:73–76, 1990

30. LISS P, NYGREN A, ULFENDAHL HR, ERIKSON U: Effect of furosemideor mannitol before injection of a non-ionic contrast medium onintrarenal oxygen tension. Adv Exp Med Biol 471:353–359, 1999

31. HEYMAN SN, BREZIS M, GREENFELD Z, ROSEN S: Protective role offurosemide and saline in radiocontrast-induced acute renal failurein the rat. Am J Kidney Dis 14:377–385, 1989

32. HARDIEK K, KATHOLI RE, RAMKUMAR V, DEITRICK C: Proximaltubule cell response to radiographic contrast media. Am J Phys-iol Renal Physiol 280:F61–F70, 2001

33. ANDERSEN KJ, CHRISTENSEN EI, VIK H: Effects of iodinated x-raycontrast media on renal epithelial cells in culture. Invest Radiol29:955–962, 1994

34. HIZOH I, STRATER J, SCHICK CS, et al: Radiocontrast-induced DNAfragmentation of renal tubular cells in vitro: Role of hypertonicity.Nephrol Dial Transplant 13:911–918, 1998

35. BEERI R, SYMON Z, BREZIS M, et al: Rapid DNA fragmentation fromhypoxia along the thick ascending limb of rat kidneys. Kidney Int47:1806–1810, 1995

36. HALLER C, SCHICK CS, ZORN M, KUBLER W: Cytotoxicity of radio-contrast agents on polarized renal epithelial cell monolayers. Car-diovasc Res 33:655–665, 1997

37. SCHNACKENBERG CG: Physiological and pathophysiological roles ofoxygen radicals in the renal microvasculature. Am J Physiol RegulIntegr Comp Physiol 282:R335–R342, 2002

38. JANLE-SWAIN E: Animal models of diabetic nephropathy, in Hand-book of Animal Models of Renal Failure, edited by Ash S, ThornhillJ, Boca Baton, CRC Press, 1985

39. ISHII N, PATEL KP, LANE PH, et al: Nitric oxide synthesis and oxida-tive stress in the renal cortex of rats with diabetes mellitus. J AmSoc Nephrol 12:1630–1639, 2001

40. OHISHI K, CARMINES PK: Superoxide dismutase restores the influ-ence of nitric oxide on renal arterioles in diabetes mellitus. J AmSoc Nephrol 5:1559–1566, 1995

41. PALM F, CEDERBERG J, HANSELL P, et al: Reactive oxygen speciescause diabetes-induced decrease in renal oxygen tension. Diabetolo-gia 46:1153–1160, 2003

42. BAKRIS GL, LASS N, GABER AO, et al: Radiocontrast medium-induced declines in renal function: A role for oxygen free radicals.Am J Physiol 258:F115–F120, 1990

43. KATHOLI RE, WOODS WT, JR., TAYLOR GJ, et al: Oxygen free radicalsand contrast nephropathy. Am J Kidney Dis 32:64–71, 1998

44. TEPEL M, VAN DER GIET M, SCHWARZFELD C, et al: Prevention ofradiographic-contrast-agent-induced reductions in renal functionby acetylcysteine. N Engl J Med 343:180–184, 2000

45. DIAZ-SANDOVAL LJ, KOSOWSKY BD, LOSORDO DW: Acetylcysteineto prevent angiography-related renal tissue injury (the APARTtrial). Am J Cardiol 89:356–358, 2002

46. BRIGUORI C, MANGANELLI F, SCARPATO P, et al: Acetylcysteine andcontrast agent-associated nephrotoxicity. J Am Coll Cardiol 40:298–303, 2002

47. SHYU KG, CHENG JJ, KUAN P: Acetylcysteine protects against acuterenal damage in patients with abnormal renal function undergoinga coronary procedure. J Am Coll Cardiol 40:1383–1388, 2002

48. ALLAQABAND S, TUMULURI R, MALIK AM, et al: Prospective ran-domized study of N-acetylcysteine, fenoldopam, and saline for pre-vention of radiocontrast-induced nephropathy. Catheter CardiovascInterv 57:279–283, 2002

49. KAY J, CHOW WH, CHAN TM, et al: Acetylcysteine for prevention ofacute deterioration of renal function following elective coronary an-giography and intervention: A randomized controlled trial. JAMA289:553–558, 2003

50. WALKER PD, BROKERING KL, THEOBALD JC: Fenoldopam and N-acetylcysteine for the prevention of radiographic contrast material-induced nephropathy: A review. Pharmacotherapy 23:1617–1626,2003

51. BIRCK R, KRZOSSOK S, MARKOWETZ F, et al: Acetylcysteine for pre-vention of contrast nephropathy: Meta-analysis. Lancet 362:598–603, 2003

52. ALONSO A, LAU J, JABER BL, et al: Prevention of radiocontrastnephropathy with N-acetylcysteine in patients with chronic kidneydisease: A meta-analysis of randomized, controlled trials. Am J Kid-ney Dis 43:1–9, 2004

53. DURHAM JD, CAPUTO C, DOKKO J, et al: A randomized controlledtrial of N-acetylcysteine to prevent contrast nephropathy in cardiacangiography. Kidney Int 62:2202–2207, 2002

54. KSHIRSAGAR AV, POOLE C, MOTTL A, et al: N-acetylcysteine for theprevention of radiocontrast induced nephropathy: A meta-analysisof prospective controlled trials. J Am Soc Nephrol 15:761–769,2004

55. WEIHPRECHT H, LORENZ JN, BRIGGS JP, SCHNERMANN J: Vasomotoreffects of purinergic agonists in isolated rabbit afferent arterioles.Am J Physiol 263:F1026–F1033, 1992

56. WEAVER DR, REPPERT SM: Adenosine receptor gene expression inrat kidney. Am J Physiol 263:F991–F995, 1992

57. SPIELMAN WS, AREND LJ: Adenosine receptors and signaling in thekidney. Hypertension 17:117–130, 1991

58. OLIVERA A, LAMAS S, RODRIGUEZ-PUYOL D, LOPEZ-NOVOA JM:Adenosine induces mesangial cell contraction by an A1-type re-ceptor. Kidney Int 35:1300–1305, 1989

59. OSSWALD H, HERMES HH, NABAKOWSKI G: Role of adenosine in sig-nal transmission of tubuloglomerular feedback. Kidney Int (Suppl12):S136–S142, 1982

60. SUN D, SAMUELSON LC, YANG T, et al: Mediation of tubuloglomerularfeedback by adenosine: Evidence from mice lacking adenosine 1receptors. Proc Natl Acad Sci USA 98:9983–9988, 2001

61. BROWN R, OLLERSTAM A, JOHANSSON B, et al: Abolished tubu-loglomerular feedback and increased plasma renin in adenosine A1receptor-deficient mice. Am J Physiol Regul Integr Comp Physiol281:R1362–R1367, 2001

62. PERSSON PB: Tubuloglomerular feedback in adenosine A1 receptor-deficient mice. Am J Physiol Regul Integr Comp Physiol 281:R1361,2001

63. OLDROYD SD, FANG L, HAYLOR JL, et al: Effects of adenosine recep-tor antagonists on the responses to contrast media in the isolatedrat kidney. Clin Sci (Lond) 98:303–311, 2000

64. KATHOLI RE, TAYLOR GJ, MCCANN WP, et al: Nephrotoxicity fromcontrast media: Attenuation with theophylline. Radiology 195:17–22, 1995

65. PFLUEGER A, LARSON TS, NATH KA, et al: Role of adenosine in con-trast media-induced acute renal failure in diabetes mellitus. MayoClin Proc 75:1275–1283, 2000

66. LISS P, CARLSSON PO, PALM F, HANSELL P: Adenosine A(1) receptorsin contrast media-induced renal dysfunction in the normal rat. EurRadiol 14:1297–1302, 2004

67. DINOUR D, AGMON Y, BREZIS M: Adenosine: An emerging role inthe control of renal medullary oxygenation? Exp Nephrol 1:152–157, 1993

22 Persson et al: Contrast-induced nephropathy

68. SEO B, OEMAR BS, SIEBENMANN R, et al: Both ETA and ETB recep-tors mediate contraction to endothelin-1 in human blood vessels.Circulation 89:1203–1208, 1994

69. BAGNIS C, IDEE JM, DUBOIS M, et al: Role of endothelium-derivednitric oxide-endothelin balance in contrast medium-induced acuterenal vasoconstriction in dogs. Acad Radiol 4:343–348, 1997

70. CLARK BA, KIM D, EPSTEIN FH: Endothelin and atrial natriureticpeptide levels following radiocontrast exposure in humans. Am JKidney Dis 30:82–86, 1997

71. HEYMAN SN, CLARK BA, KAISER N, et al: Radiocontrast agents in-duce endothelin release in vivo and in vitro. J Am Soc Nephrol3:58–65, 1992

72. OLDROYD SD, MORCOS SK: Endothelin: What does the radiologistneed to know? Br J Radiol 73:1246–1251, 2000

73. FUJISAKI K, KUBO M, MASUDA K, et al: Infusion of radiocontrastagents induces exaggerated release of urinary endothelin in pa-tients with impaired renal function. Clin Exp Nephrol 7:279–283,2003

74. WANG A, HOLCSLAW T, BASHORE TM, et al: Exacerbation of radio-contrast nephrotoxicity by endothelin receptor antagonism. KidneyInt 57:1675–1680, 2000

75. FREED MI, WILSON DE, THOMPSON KA, et al: Pharmacokinetics andpharmacodynamics of SB 209670, an endothelin receptor antago-nist: Effects on the regulation of renal vascular tone. Clin PharmacolTher 65:473–482, 1999

76. WEBER C, SCHMITT R, BIRNBOECK H, et al: Pharmacokinetics andpharmacodynamics of the endothelin-receptor antagonist bosentanin healthy human subjects. Clin Pharmacol Ther 60:124–137, 1996

77. HAYLOR JL, MORCOS SK: An oral ET(A)-selective endothelin recep-tor antagonist for contrast nephropathy? Nephrol Dial Transplant16:1336–1337, 2001

78. LISS P, CARLSSON PO, NYGREN A, et al: Et-A receptor antagonistBQ123 prevents radiocontrast media-induced renal medullary hy-poxia. Acta Radiol 44:111–117, 2003

79. MORCOS SK, THOMSEN HS, WEBB JA: Contrast-media-inducednephrotoxicity: A consensus report. Contrast Media Safety Com-

mittee, European Society of Urogenital Radiology (ESUR). EurRadiol 9:1602–1613, 1999

80. KATZBERG RW: Urography into the 21st century: New contrast me-dia, renal handling, imaging characteristics, and nephrotoxicity. Ra-diology 204:297–312, 1997

81. MORCOS SK: Prevention of contrast media nephrotoxicity—Thestory so far. Clin Radiol 59:381–389, 2004

82. MURPHY SW, BARRETT BJ, PARFREY PS: Contrast nephropathy. JAm Soc Nephrol 11:177–182, 2000

83. TRIVEDI HS, MOORE H, NASR S, et al: A randomized prospective trialto assess the role of saline hydration on the development of contrastnephrotoxicity. Nephron Clin Pract 93:C29–C34, 2003

84. LORENZ JN, NIEMAN M, SABO J, et al: Uroguanylin knockout micehave increased blood pressure and impaired natriuretic response toenteral NaCl load. J Clin Invest 112:1244–1254, 2003

85. GAMI AS, GAROVIC VD: Contrast nephropathy after coronary an-giography. Mayo Clin Proc 79:211–219, 2004

86. NAFZ B, BERGER K, ROSLER C, PERSSON PB: Kinins modulate thesodium-dependent autoregulation of renal medullary blood flow.Cardiovasc Res 40:573–579, 1998

87. FRANCHINI KG, COWLEY AW, JR.: Sensitivity of the renal medullarycirculation to plasma vasopressin. Am J Physiol 271:R647–R653,1996

88. FRANCHINI KG, COWLEY AW, JR.: Renal cortical and medullaryblood flow responses during water restriction: Role of vasopressin.Am J Physiol 270:R1257–R1264, 1996

89. BARRETT BJ: Contrast nephrotoxicity. J Am Soc Nephrol 5:125–137,1994

90. BARRETT BJ, CARLISLE EJ: Metaanalysis of the relative nephrotoxic-ity of high- and low-osmolality iodinated contrast media. Radiology188:171–178, 1993

91. HAWKINS IF, CARIDI JG: Carbon dioxide (CO2) digital subtractionangiography: 26-year experience at the University of Florida. EurRadiol 8:391–402, 1998

92. THOMSEN HS, MORCOS SK: Radiographic contrast media. BJU Int86 (Suppl 1):1–10, 2000