Effects of Contrast Media and Mannitol on Renal Medullary Blood Flow and Red Cell Aggregation in the Rat Kidney

Kidney International, Vol. 49 (1996), pp. 1268—1275

Effects of contrast media and mannitol on renal medullary blood flow and red cell aggregation in the rat kidney

PER Liss, ANDERS NYGREN, ULF OLSSON, Hs R. ULFENDAHL, and UNO ERIKSON

Department of Diagnostic Radiology, University Hospital; Department of Physiology and Medical Biophysics, University of Uppsala; and Department of Statistics, Swedish University of Agricultural Sciences, Uppsala, Sweden

Effects of contrast media and mannitol on renal medullary blood flowred blood cell aggregation in the medullary vessels after intrave- and red cell aggregation in the rat kidney. Hemodynamic factors may play nous injection of CM. a role in the development of acute renal failure following administration of contrast media (CM). In this study the effect of intravenous injection of The CM investigated were: diatrizoate, ionic ratio 1.5; ioxa- contrast media and mannitol on red blood cell velocity (VRBc) and red glate, ionic ratio 3.0; iopromide, non-ionic ratio 3.0; iohexol, blood cell aggregation in renal medullary vessels was studied in 58 rats. non-ionic ratio 3.0; and iotrolan, non-ionic ratio 6.0 (Table 1). Renal medullary blood flow was investigated by a cross-correlation technique and by a visual aggregation score. The CM, namely diatrizoate, Methods iopromide, iohexol, ioxaglate, iotrolan, were given in iodine equivalent doses (1600 mg/kg body wt). Mannitol (950 mOsm/liter) and Ringer's Animals solution were used as controls. The same vessels were studied 30 minutes Studies were performed on 58 male Lewis-DA rats with an before and 30 minutes after injections. VRBC decreased significantly after injection of diatrizoate, iopromide, iohexol, iotrolan and mannitol. Ring- average weight of 152 g (SEM 2,5 g). The animals had free access er's solution and ioxaglate did not significantly alter medullary blood flow, to tap water and standard chow (R3, Ewos, Södertälje, Sweden; while iotrolan and mannitol caused the largest decreases in All CM containing 3 g sodium/kg, 8 g potassium/kg, 21% protein, 13 X 106 and mannitol caused both red cell aggregation and cessation of blood flow. Joule/kg) until the day of the experiment. They were anaesthe- The decrease in blood flow and increase in red blood cell aggregation after injection of CM and mannitol may partly explain the occurrence oftized with mactin® (Byk-Gulden, Konstanz, Germany), given contrast medium-induced acute renal failure. intraperitoneally in a dosage of 100 mg/kg body weight (body wt). Tracheostomy was performed and the rat was placed on a servo-controlled heating pad to maintain the body temperature at 37.5°C. The mechanisms underlying contrast medium-induced acute renal failure (CM-ARF) are not clear. Circulatory events in the Surgical procedures rat renal medulla have recently attracted the attention of several A polyethylene catheter was placed in the left femoral artery for groups in the search for an explanation of CM-ARF. The specialblood sampling and monitoring of blood pressure. A fall in vulnerability of the renal medulla to hypoxia and the documentedarterial blood pressure below 80mm Hg resulted in termination of ability of CM to reduce renal medullary blood flow justifies thisthe experiment. The left femoral vein was catheterized for infu- interest. Reduced blood flow in the renal medulla after CMsion of Ringer's solution (0.5 mI/hr per 100 g body wt) composed injection was demonstrated in rats by Nygren et al in 1988 [1] andof 120 mtvi NaC1, 2.5 mivi KCI, 25 mM NaHCO3, 0.75 mat CaCI2. trapping of red blood cells in the rat renal medullary vesselsThe right femoral vein was catheterized for infusion of CM. The following injection of CM and mannitol was shown by Nygren inurinary bladder was catheterized for collection of urine from the 1989 [2]. More recently Heyman et al [3] observed outer medul-right kidney. The left kidney was exposed by a left subcostal flank lary red blood cell trapping and hypoxia following injection ofincision and immobilized in a plastic cup. The kidney was embed- iothalamate and mannitol. In vivo aggregation of red blood cellsded in pieces of cotton wool soaked in Ringer's solution. The caused by CM and hypertonic solutions has been reported byrenal hilar fat was carefully removed and the renal pelvis was then several investigators [4—12]. However, the clinical significance ofopened by an incision to the margin of the renal cortex. This this phenomenon has been the subject of great controversyprocedure exposed approximately 1.5 mm of the papilla. [13—15]. The aim of this investigation was to further examine the influence of CM on the renal medullary circulation by in vivo Microscopy measurement of the red blood cell velocity and the incidence of The renal papillary vessels were studied with a Leitz Ortolux microscope (Leitz, Wetzlar, Germany) and the papilla was trans- illuminated with an Intralux® 5000 light source (Volpi AG, Zurich, Switzerland). In addition to direct visual observations with Received for publication June 8, 1995 a microscope, the image was recorded and projected onto a and in revised form October 5, 1995 television screen (Sony PVM-9OCE) by means of a video camera Accepted for publication November 27, 1995 (Ikegami ITC-410). The final magnification on the television © 1996 by the International Society of Nephrology screen was x640. The information was recorded on videotapes

1268 Liss et al: CM and renal medullaiy blood flow 1269

Table 1. Tested substances Video camera Type of contrast Iodine cone.Osmolality Substance medium mg/mi mOsm/kg Microscope Diatrizoatea ionic monomer 370 2070 (Urografin®) lohexola non-ionic monomer 350 783 Kidney (Omnipaque®) Iopromid&' non-ionic monomer 370 774 (Ultravist®) ionic dimer 320 600 Light leader under the papilla loxaglatea Ureter (1-texabrix®) lotrolana non-ionic dimer 300 320 (Isovist®) Pap ha Mannitol — 0 950 Ringer's solution — 0 310 Fig.1. Experimental set-up for studying the renal medullaiy vessels in the a Dataare from the manufacturer papilla.

using a videotape recorder (Sony AG-6200) for later off-linedetermined during the first possible 30 seconds within the first two analysis (Fig. 1). minutes of each period. With this restriction cross-correlation measurements could be performed in 91% of the total of 2,838 Choice of vessels observation periods. In 58 rats, the renal medullary blood flow was studied in a total of 258 papillary vessels with diameters of 5 to 25 jxm. In each Aggregation score experiment a good view of the papilla, with as many vessels as To determine the intravascular tendency for red blood cells to possible in focus, was sought after. Measurements were made on aggregate in the medullary vessels, a simple score method was the visible part of the papilla, as far as possible from the tip, in used. The blood flow in each vessel with a cross-correlation of 0.3 order to get as straight vessels as possible to facilitate cross- or better in the field of view was studied throughout the experi- correlation measurements. All vasa recta capillaries in the field ofment. The control period, before CM or control infusion, was view with a cross-correlation coefficient of 0.3 or better during the divided into three 10-minute observation periods. The experimen- control period were analyzed. Depending on the local anatomy tal period, after CM or control infusion, was divided into six and optical conditions, it was possible to examine one to nine 5-minute observation periods. The blood flow pattern was classed vessels in each papilla. To determine the blood vessel width, blood as (1) normal blood flow, (2) flow with appearance of red blood column diameters were measured from the television screen with cell aggregates, or (3) cessation of flow. The aggregation criterion a caliper. was fulfilled when aggregated red cells were seen continuously in Cross-correlation measurements the studied vessel for at least 30 seconds during the observation period. Analogously, the stop in flow criterion was fulfilled when Erythrocyte velocities (VRBC) in the vasa recta were deter-flow ceased in the studied vessel for at least 30 consecutive mined off-line from the video recordings using the dual-window seconds during the observation period. cross-correlation technique based on the method described by Weyland and Johnson [16] and used on the vasa recta vessels by Gussis, Jamison and Robertson [17]. Experimental protocol The Capi-flow® equipment used for the cross-correlation anal- After the surgical procedures, the animal was allowed to yses was supplied by CapiFlow AB (Kista, Sweden). The videorecover for 45 minutes. The experiment then started with a signal was passed through a video photometric analyzer generat-30-minute control period followed by a 30-minute period when ing two photometric windows that were displayed on the TVCM or control solution was injected intravenously during the first monitor. These two windows were then positioned along theeight minutes (Fig. 2). Throughout the experiment the blood flow monitored vessel. The position of the windows could be manuallyin the studied medullary vessels was recorded on videotapes. All adjusted. The velocity of the red blood cells was measured byCM were given in a dosage of 1600 mg iodine/kg body wt, which continuously computing the transit times between similar opticalis about 5 ml CM/kg body wt of a 300 mg I/mi CM in a 60 kg signals generated at the two windows by blood cells and plasmaperson. Such doses are not unusual in angiographic practice. The gaps. By dividing the distance between windows by the time delay, CM tested were: diatrizoate, 370 mg/mI (Urografin®, Schering, which resulted in maximum cross-correlation, the erythrocyteBerlin, Germany); iohexol 350 mg/ml (Omnipaque®, Nycomed velocities were determined. By replaying the videotape and repo-AJS, Oslo, Norway); iopromide, 370 mg/mi (Ultravist®, Seher- sitioning the windows, VRBC in several different vessels could being); ioxaglate, 320 mg/ml (Hexabrix®, Laboratoire Guerbet, measured during the same real time. The same vessels wereAulnay-sous-Bois, France) and iotrolan, 300 mg/mI (Isovist®, followed throughout the experiment. The measurementsSchering). Mannitol (950 mOsm/kg) or Ringer's solution was used were performed four times before the CM injection (at —30, —20, as a control (Table 1). The dose of control solution infused was 4,5 —10 and —3 minutes) and seven times after CM (at +1, +5, +10,mi/kg body wt. The amount given ranged from 0.55 to 0.95 ml. All +15, +20, +25 and +28.5 minutes). In all experiments VRBC wassolutions were at room temperature when administrated. 1270 Lisset al: CM and renal meduliaiy blood flow

Resting period Controlperiod Experimental period (45 mm) (30 mm) 30 (mm) //////,,,,,Ztt//J,7z_,,,///,,,y/,z/1 ______Time t 4 Video tape recording Operation

Intravenous injection Fig. 2. Experimental protocol.

Table 2.Number of experiments, analyzed vessels, vessel diameter and VRBC before and after injection of contrast media, mannitol and Ringer's solution Number of Number of analyzed Vessel diameter Substance animals vessels rna VRBC before CM VROC after CMb Change in VRBC Diatrizoate 5 27 12.3 0.344 0.230 0.114c (Urografin) (±0.7) (±0.069) (±0.063) (±0.047) lohexol 10 39 12.4 0.447 0.320 0.127c (Omnipaque) (±0.5) (±0.052) (±0.047) (±0.035) lopromide 9 46 13.6 0.413 0.310 0.102c (Ultravist®) (±0.5) (±0.051) (±0.046) (±0.035) loxaglate 9 37 13.6 0.460 0.412 0.049 NS (Hexabrix®) (±0.7) (±0.053) (±0.048) (±0.036) lotrolan 9 40 13.2 0.441 0.216 0.225c (Isovist®) (±0.5) (±0.052) (±0.047) (±0.036) Mannitol 9 42 14.0 0.509 0.346 0.164c (±0.6) (±0.059) (±0.047) (±0.035) Ringer 7 27 13.1 0.455 0.462 —0.007 NS (±1.0) (±0.062) (±0.056) (±0.042) NS, Non-significant difference (P > 0.05) compared to the control period. Data are means SEM. a Estimatedwidth (see Methods) bAverage of period +10, +15, +20, +25, +28.5 mm after CM Significant difference (p < 0.05) compared to the control period

Blood and urine analyses displayed any type of reaction (aggregation or cessation of flow) The urine volumes were measured gravimetrically. The urinewere coded as 1 (otherwise 0). sodium and potassium concentrations (UNa, UK) were measured The urine excretion and hematocrit data were analyzed by a by flame photometry (IL 543, Instrumentation Lab., Milano,Manova model, followed by univariate tests where appropriate. Italy). Urine osmolality (U0m) was determined by the freezingAll statistical evaluation of urine analysis was performed with an point depression method (Model 3M0; Advanced Instruments,average of the two values before CM compared to the average of MA, USA). Two urine samples were collected before CM andthe value from the sampling period after CM with maximal three after CM. Blood samples were taken for hematocrit analysis,change. Statistical evaluation of the hematocrit values was per- once before and twice after injection of CM. formed by comparing the value after CM with maximal change to the value before CM. A P value < 0.05 was regarded as significant Statistical evaluation in all analyses. The data from the measurements of VRBC were analyzed using a hierarchical repeated measures multivariate analysis of variance Results (Manova) model. The GLM procedure [18] was used. The Manova was followed by univariate tests where appropriate. The Medullaiy blood flow raw data were transformed into percentage change. The analysis was based on the average percentage change in blood flow in the The average VRBC for all groups before injection was 0.435 four periods before injection of CM as compared to the average in (SEM 0.018) mm/sec (Table 2). There was no change in the the five last periods after CM injection (10, 15, 20, 25 and 28.5blood column diameter of the vessels during the experiments. mm). The average value from all studied vessels in each animalDiatrizoate caused a 32% decrease in VRBC, iopromide a 22% was used in the analysis. decrease, iohexol a 23% decrease, ioxaglate a 9% decrease, The data from the aggregation and cessation of blood flowiotrolan a 47% decrease and mannitol a 32% decrease while observations were analyzed using a repeated measures categoricalRinger's solution increased VRBC by 1%. In all groups except for data analysis model. The Catmod procedure was used [18]. Beforethe ioxaglate and Ringer's groups, VRBC in the medullaly vessels the analysis the data were converted so that all blood vessels thatwas significantly decreased (Fig. 3). Liss et al: CM and renal medullaty blood flow 1271

150 vessels in all the CM groups and in the group injected with hyperosmolar mannitol. Ringer's solution did not cause a tendency to aggregation. The tendency of red cells to aggregate was most pronounced after iotrolan and mannitol injection. The lowest incidence of aggregation caused by CM was seen after injection of ioxaglate. 100 Most measurements of renal blood flow have concerned the total renal blood flow. Our group has previously used a laser 0 Doppler technique for studies of medullary blood flow after CM it injection [1] and produced results similar to those obtained in the > present study. As the follow-up period after injection of CM was 50 only 30 minutes in this study, it was not possible to determine whether acute renal failure would develop in any of the rats. _____ Thereduction in VRBc seen after injection of CM and hyperosmolar mannitol might be explained in several ways, for example, Injection passive compression of the vessels, rheological effects, or a vasoactive effect on the medullary vessels. 0 Blood flow may be reduced by compression of the vessels due to —30 —20 —10 0 10 20 30 increased tissue pressure in the kidney. Hypertonic CM and mannitol cause a pronounced osmotic diuresis that distends the Time, minutes tubules and collecting ducts leading to renal swelling and an Fig.3. Relative changes in red blood cell velocity (V) in the renalincrease in intrarenal venous pressure [19]. The isotonic non-ionic medullaiy vessels following intravenous injection of contrast media (CM), dimer iotrolan increases the renal interstitial pressure, presum- mannitol and Ringer's solution. Symbols are: (•) diatrizoate*; (A) iopro- ably by another mechanism, as demonstrated by Ueda et al [20, mide; (•) iohexol*; () ioxaglate; (0)iotrolan*;() mannitol*; (0)21]. This CM caused a moderate diuresis compared to that Ringer's solution. *significant difference (P < 0.05) compared to the control period. resulting from low osmolar CM. lotrolan concentrates down- stream from the tubules, and consequently the intratubular urine viscosity rises because of the high molecular weight of iotrolan, resulting in increased intratubular hydrostatic pressure. Aggregation Aggregation in the medullary vessels, which is most pronounced In all CM groups and in the mannitol group the incidence of redwith the non-ionic CM and least pronounced with ionic ioxaglate, blood cell aggregation and cessation of blood flow increasedchanges with papillary blood flow. This suggests that aggregation significantly, while no change was found in the Ringer's group.may be one of the causes of the observed blood flow decrease. The incidence of aggregation and flow cessation was significantlyHowever, it is a well known fact that aggregation is also observed lower after ioxaglate injection than after injection of all other CMin association with a low blood flow [22, 23], which means that the and mannitol. However, the change between the period beforereverse is just as likely. Aggregation caused by a decrease in shear and after CM did not differ significantly between the ioxaglate andrate in a small-caliber vessel will increase the viscosity [24], which iopromide groups (Fig. 4). in turn will reduce the blood flow according to Poiseuille's law and thereby cause further aggregation. Urineand blood data In recent years the effects of CM on blood coagulation mech- Urine flow (UV) increased 7-to17-fold after injection ofanisms and platelet activity have been thoroughly discussed. In diatrizoate, iohexol, iopromide, ioxaglate or mannitol. Urine flowstudies in vitro, ionic CM seemed to inhibit coagulation and after iotrolan injection increased threefold and after Ringer'splatelet activation to a greater extent than non-ionic CM [25—27]. 1.4-fold (Fig. 5A). The increase in urine flow was significant in allSome workers have cast doubt on the hypothesis that non-ionic but the Ringer's group. Urine sodium output (UNaV) increasedCM are prothrombotic [28]. Support for the idea of prothrom- three- to 34-fold; the increase was not significant in the iotrolan orbotic activity of non-ionic CM is provided by the findings of the Ringer's group (Fig. 5C). Urine potassium output (UKV) increased coagulation and platelet activity after infusion of non- increased four- to sixfold in the groups receiving CM and manni-ionic CM reported by Hardeman et a! [29, 30]. to!. The increase in urine potassium was significant in all groups In the present study, more red blood cell aggregates were found (Fig. SB). Uosm decreased significantly in all groups except thein the rats receiving non-ionic CM than in those in the ioxaglate Ringer's group (Fig. 5D). There was a slight decrease in hemat-group. The explanation for this finding is unclear, but it is possible ocrit in all groups; this was significant only in the Ringer's groupthat the negative electric charge of ionic CM molecules leads to (Fig. 5E). repulsion of the red blood cells and thereby inhibits aggregation. Although ioxaglate and diatrizoate are both ionic CM, more Discussion aggregates are observed after injection of diatrizoate. This may In this study, using a cross-correlation technique, we found thatindicate that factors other than the electric charge of the CM renal medullary VRBC decreased after injection of CM andmolecule are of importance. As the osmotic diuresis and thereby mannitol but did not change in animals that received Ringer'sthe interstitial hydrostatic pressure were the same in the groups solution and ioxaglate. In the same experiments it was found thatthat received ionic and non-ionic low osmolar CM, the difference red blood cells had a tendency to aggregate within the medullarybetween these types of CM as far as aggregation tendency and 1272 Liss et al: CM and renal medullaiy blood flow

Diatrizoate* Iohexol* 70 70 0 60 60 (I) 50 o0 50 C (0 0 40 Injection 40 Injection (0 -0 0) 30 30 ci) 0) 0) +1 0) 20 20 10 ri 10 1flMM 0 !iii —30 —20 —10 0 51015202530 —30 —20 —10 0 51015202530 Time,minutes Time,minutes

Iotrolan* Ringer (NS) 70 70 60 60 0 50 o0 50 (0 C,) 40 ? 40 0C 0 30 Injection 30 0)(0 0) Injection ci) 20 20 0) 0) 0) 10 0) 10 0 .+.i I —30 —20 —10 0 51015202530 —30 —20 —10 0 51015202530 Time,minutes Time,minutes Iopromide* Ioxaglate* 70 70 o0 60 0 60 50 o 50 0) Cl) 40 . 40 -0 0 30 Injection 30 0) 0) Injection 6) 20 20 0) 0) + I )10 io — __InI —30 —20 —10 0 51015202530 —30 —20 —10 0 51015202530 Time,minutes Time,minutes Mannitol*

70 60 o0. 50 Ci, 40 0 Injection 30 Fig. 4. Incidence of aggregates of red blood cells 0) and stop in blood flow in the renal medulla,y 20 +11 0) vessels before and after intravenous contrast 10 - — media (CM), mannitol and Ringer's solution. 0 Il1]III Symbols are: (El) vessels with aggregates; (•) —30 —20 —10 0 51015202530 vesselswith a stop in blood flow. "Significant difference (P < 0.05) compared to the control Time,minutes period.

VRBC are concerned cannot be explained by intrarenal pressurethis mechanism in the present study cannot be ruled out. How- differences. ever, we found a smaller decrease in papillary blood flow in the One drawback of the opened pelvis approach to papillary blooddiatrizoate, injected rats, which had the most pronounced diuresis flow measurements is a possible influence on papillary blood flow(Fig. 5A), than in the iotrolan group (which was the least diuretic as studied by Schmidt-Nielsen et a! [31,32]. Theyhave describedcompound tested). These findings indicate that the milking effect pelvic milking of the urine out of the papilla, a mechanismwas of minor importance for the present results. This is also postulated to be most important during high urine flows. Thissupported by the findings by Lemley et a! [33], who found no enhancement of urine transport from the collecting ducts maydifference in VRBC, measured through an opened pelvis in have secondary effects on blood flow. The possible influence ofanti-diuretic compared to water-diuresis induced rats. Liss et a!: CM and renal medullaty blood flow 1273

A B 40 3 Diatrizoate* 2.5 Diatrizoate* 30 lopromide* iopromide* 2 Iohexol* I ohexol" 0 a.20 E 1.5 —0—— > —0-Ioxaglate* Ioxaglate* __o__ IotroIan' Iotrolan* 1 10 Mannitol* —6-— Mannitol* 0.5 Ringer _.Q_ Ringer' 0 0- —30 —15 0 10 20 30 —30 —15 010 20 30

Time,minutes Time,minutes

C D 3 3000

2.5 Diatrizoate* Diatrizoate*

A loprom ide* ci> 2 2000 Iopromide* Iohexol* —0-Iohexol* E1.5 a. Injection --Ioxaglate* -0-Ioxaglate* a. >. -a--lotro Ian E Iotrolan* (0 1000 -6--Mannitol* 0 —6--Mannitol* 0.5 -0-Ringer Injection -0-Ringer 0 0 —30 —15 0 10 20 30 —30 —15 0 10 20 30

Time,minutes Time,minutes E 60 • Diatrizoate 50 A lopromide • lohexol 0 040 --0—loxaglate cci E -0-lotrolan ci) I30 t -&-Mannitol Injection -0-Ringer* 20 Fig.5. Changes in (A) urine flow, excretion of (B) potassium and (C) —30 —15 0 10 30 sodium, (D) urine osmolality and (E) hematocrit following injection of contrast media, mannitol and Ringer's solution. *signiflcant difference (P < Time,minutes 0.05)comparedto the control period.

The possible role of vasoactive substances in CM-inducedreduction in medullary blood flow and aggregation of red blood changes in renal blood flow cannot be ruled out by this study, butcells as observed in the present study may be due to the vascular previous work has failed to identify any definite relationship toarchitecture of the renal medulla. The counter-current system, renin, angiotensin or prostaglandin systems [34—36]. consisting of the ascending and descending vasa recta, builds up Only 10 to 15% of the total renal blood flow that passes throughan oxygen gradient in the medulla, with a high P°2 in the outer the kidney perfuses the medullary regions [37]. Even majorpart of the medulla and a low P°2 in the inner part. The high 02 reductions in renal medullary blood flow are thus not necessarilyrequirement for active transport of sodium in the thick ascending reflected in total renal blood flow and the changes in total renalpart of the loop of Henle and the collecting ducts makes this blood flow are in fact moderate after injection of CM [38—40]. Aregion especially vulnerable to hypoxia [3, 41, 42]. 1274 Lisset al: CM and renal medullaiy blood flow

Agmon et al have reported an increase in outer medullary 9. RING J, ENDRICH B, INTAGLIETrA M: Histamine release, complement blood flow following injection of iothalamate [43]. Our results do consumption, and microvascular changes after radiographic contrast not have to be in conflict with these, as papillary blood flow does media infusion in rabbits. J Lab Clin Med 92:584—594, 1978 10. READ RC, MEYER M: The role of red cell agglutionation in arterio- not necessarily reflect flow changes in the whole medulla, as graphic complications. S Forum 10:472—475, 1960 shown by Agmon et al [43] and Hellberg, Kallskog and Wolgast11. WIEDEMAN MP: Vascular and intravascular responses to various [441.Onthe contrary, many studies have demonstratied outer contrast media. Angiology 14:107—109, 1963 medullary trapping of red blood cells in acute renal failure settings 12. JOHNSON JH, KNIESLY MH: Intravascular agglutionation of the flow- and following administration of contrast media [3, 44, 45]. A ing blood following the injection of radiopaque contrast media. decrease in blood flow can be expected after this trapping. This is Neurology 12:560—570, 1962 13. ASPELIN P, SCHMID-SCHONBEIN H, MALOTTA H: Do nonionic contrast what Hellberg et al [44] found in postischemic kidneys, where media increase red cell aggregation and clot formation? Invest Radiol trapping in outer medulla was observed together with a decrease 23(Suppl 2):S326—S333, 1988 in the outer medullary blood flow (measured by laser-Doppler). 14. ASPELIN P, SCHMID-SCHONBEIN H: Effect of ionic and non-ionic Different control systems for the outer and inner medulla may contrast media on red cell aggregation in vitro. Acta Radiol 19:766— also exist. 784, 1978 The red blood cell aggregation after injection of CM demon- 15. DAWSON P, COUSINS C, BRADSHAW A: The clotting issue: Etiologic factors in thromboembolism. II. Clinical considerations. Invest Radiol strated in vivo in other capillary vessels (mesenteric [4, 8, 46], pia 28(Suppl 5):S31—S36, 1993 mater, muscle [10], bat wing [11], conjuctival [7, 12]) has been 16. WEYLAND H, JOHNSON PC: Eiythrocyte velocity measurement in judged to have no clinical significance [13—15]. Opposing this view microvessels by a two-slit photometric method. JAppl Physiol 22:333— are recent findings by Heyman et al [3] of both trapping of red 337, 1967 blood cells in renal medullary vessels and medullary hypoxia after17. GUSSIS GL, JAMISON RL, ROBERTSON CR: Determination of eiythrocyte velocities in the mammalian inner renal medulla by a video CM. The present study demonstrated aggregation of red blood velocity-tracking system. Microvasc Res 18:370—383, 1979 cells and a decrease in medullary blood flow following injection of18. SAS/Stat User's Guide (version 6, 4th ed). Cary, NC, 1989 CM and hyperosmolar mannitol, supporting the theory that a 19. KATZBERG RW, SCHULMAN G, MEGGS LG, CALDIC0'vr WJH, DAMI- disturbance of renal medullary blood flow is a possible mechanism ANO MM, HOLLENBERG NK: Mechanism of the renal response to underlying CM-ARF. contrast medium in dogs: Decrease in renal function due to hyperto- nicity. Invest Radiol 18:74—80, 1983 Conclusions 20. UEDA J, NYGREN A, HANSELL P, ERIKSON U: The influence of contrast media on singel nephron glomerular filtration rate in the rat kidney. Contrast media and hypertonic mannitol decrease red blood Acta Radiol 33:596—599, 1992 cell flow in the rat renal medullary vessels and cause erythrocyte21. UEDA J, NYGREN A, HANSELL P, ULFENDAHL HR: Effect of intrave- aggregation, which may contribute to the onset of acute renal nous contrast media on proximal and distal tubular hydrostatic failure following infusions of CM. Non-ionic CM cause more flow pressure in the rat kidney. Acta Radiol 34:83—87, 1993 22. COPLEY AL, KING RG, CHIEN KS, USAMI 5, SKALAK R, HUANG CR: disturbances than ionic ioxaglate. Microscopic observations of viscoelasticity of human blood in steady and oscillatory shear. Biorheology 12:257—263, 1975 Acknowledgments 23. SCHMID-SCHONBEIN H, GAEHTGENS P, HIRSCH H: On the shear rate dependence of red cell aggregation in vitro. J Clin Invest 47:1447— This study was supported by the Gunvor and Josef Anér Foundation 1454, 1968 and the Marcus and Amalia Wallenberg Foundation. Thanks are due to 24. CHIEN S, USAMI S, SKALAK R: Blood flow in small tubes (Section 2, Lena HoIm for lending us the cross-correlation equipment. part 2), in Handbook of Physiology, The Cardiovascular System, edited by RENKIN EM, MICHEL CC, Bethesda, American Physiology Society, Reprint requests to Per Liss, MD., Department of Diagnostic Radiology, 1984, pp 217—249 University Hospital, S-751 85 Uppsala, Sweden. e-mail: [email protected] 25. COROT C, PERRIN JM, BELLEVILLE J, AMIEL M, ELOY R: Effect of iodinated contrast media on blood clotting. Invest Radiol 24:390—393, References 1989 26. BELLEVILLE J, BAGUET J, PAUL J, ELOY R: In vitro study of the 1. NYGREN A, ULFENDAHL HR, HANSELL P, ERIKSON U: Effects of inhibition of coagulation induced by different radio contrast mole- intravenous contrast media on cortical and medullary blood flow in cules. Thromb Res 38:149—162, 1985 the rat kidney. Invest Radiol 23:753—761, 1988 2. NYGREN A: Contrast media and regional renal blood flow. 1989, 27. DAWSON P: Non-ionic contrast agents and coagulation. Invest Radiol 23(Suppl):S310—S317, 1988 Thesis, Uppsala University 3. HEYMAN SN, BREZIS M, EPSTEIN FE, SPOKES K, SILVA F, ROSEN 5: 28. Thromboembolism during angiography. (Editorial) Lancet 339:1576— Early renal medullary hypoxic injury from radiocontrast and indo- 1578, 1992 methacin. Kidney mt 40:632—642, 1991 29. HARDEMAN MR, GOEDHART P, KOEN IY: The effect of low-osmolar 4. BROWN R, YOUMANS CR, LIVANEC G, DERRICK JR, BOND TP, GUEST ionic and nonionic contrast media on human blood viscosity, erytro- MM: Cinemicrographic observations of the effects of contrast media cyte morphology, and aggregation behavior. Invest Radiol 26:810—819, on the microcirculation. Vasc Surg 2:109—115, 1968 1991 5. KLOPP R, KRAUSE W: Prevention of contrast media induced distur- 30. HARDEMAN M, KONIJNENBERG A, STURK A, REEKERS JA: Activation bance of microcirculation by ileoprost. 18th International Congress of of platelets by low-osmolar contrast media: Differential effects of ionic Radiology, Singapore, 1994 and nonionic agents. Radiology 192:563—566, 1994 6. KLoPP R, NIEMER W, SCHIPPEL W, MONSTER W: Tierexperimentelle 31. SCHMIDT-NIELSEN B, GRAVES B: Changes in fluid compartments in Untersuchungen zur Mikrozirkulation ionischer und nichtionischer hamster renal papilla due to peristalsis in the pelvic wall. Kidney mt ROntgenkontrastmittel. Radiol Diagn 30:324—329, 1989 22:613—625, 1982 7. SOBIN SS, FRASI-IER WG, JACOBSON G, VAN EECKENHOVEN FA: 32. REINKING LN, SCHMIDT-NIELSEN B: Peristaltic flow of urine in the Nature of adverse reactions to radiopaque agents. Preliminary report. renal papillary collecting ducts of hamsters. Kidney mt 20:55—60, 1981 JAMA 170:1546—1547, 1959 33. LEMLEY Ky, SCHMITT SL, HOLLIGER C, DUNN MJ, ROBERTSON CR, 8. MEYER MW, REA RC: Red cell aggregation from concentrated saline JAMISON RL: Frostaglandin synthesis inhibitors and vasa recta eryth- and angiographic media. Radiology 82:630—635, 1964 rocyte velocities in the rat. Am J Physiol 247:F562—F567, 1984 Liss et a!: CM and renal medullay blood flow 1275

34. WoRKMAN RJ, SHAFF MI, JACKSON RV, DIGGs J, FRAZER MG, and iopamidol in euvolemic and dehydrated dogs. Invest Radiol BRISCOE C: Relationship of renal hemodynamic and functional 21:793—797, 1986 changes following intravascular contrast to the renin-angiotensin 40. LEEMING BWA, SPoms KC, SILVA P: Effect of meglumine system and renal prostacyclin in the dog. Invest Radiol 18:160—166, iothalamate on renal hemodynamics and function in the diabetic rat. 1983 Invest Radio! 20:971—977, 1985 35. KATZBERG RW, MORRIS TW, BURGERNER FA, KAMM DE, FISCHER 41. BREzIs M, ROSEN S, SILVA P, EPSTEIN FH: Selective vulnerability of HW: Renal renin and hemodynamic responses to selective renal the medullary thick ascending limb to anoxia in the isolated perfused artery catheterization and angiography. Invest Radio! 12:381—388, 1977 rat kidney. J Clin Invest 73:182—190, 1984 36. LUND G, EINzIG S, RYSAVY J, SALOMONOWITZ E, CASTANEDA-ZUNIGA 42. BREZIS M, ROSEN S, SILVA P, EPSTEIN FH: Renal ischemia: A new W, AMPLATZ K: Effect of prostaglandin inhibition on the renal perspective. Kidney mt 26:375—383, 1984 vascular response to ionic and non-ionic contrast media in the dog. 43. AGMON Y, PELEG H, GREENFELD Z, ROSEN S, BREZIS M: Nitric oxide Acta Radiolol 25:407—410, 1984 and prostanoids protect the renal outer medulla from radiocontrast 37. WOLGAST M: Renal medullary red cell and plasma flow as studied with toxicity in the rat. J Clin Invest 94:1069—1075, 1994 labelled indicators and internal detection. Acta Physiol Scand 88:215— 44. HELLBERG P0, KALLSKOG 0, WOLGAST M: Red cell trapping and 225, 1973 postischemic renal blood flow. Differences between cortex, outer and 38. ALDICOTr H, HOLLENBERG NK, ABRAMS HL: Characteristics of inner medulla. Kidney Int 40:625—631, 1991 response of renal vascular bed to contrast media. Invest Radio! 45. NYGREN A, HELLBERG 0, HANSELL P: Red-cell trapping in the rat 5:539—547, 1970 renal microcirculation induced by low-osmolar contrast media and 39. KATZBERG RW, MORRIS TW, LASSER EC, DIMARCO PL, MERGUE- mannitol. Invest Radiol 28:1033—1038, 1993 RIAN PA, VENTURA JA, PABICO RC, MCKENNA BA: Acute systemic 46. READ RC, JOHN JA, VICK JA, MEYER MW: Vacular effects of and renal hemodynamic effects of meglumine/sodium diatrizoate 76% hypertonic solutions. Circ Res 8:538—548, 1960