Comparison of the effects of radiographic contrast media on dehydration and filterability of red blood cells from donors homozygous for hemoglobin A or hemoglobin S.код для вставкиСкачать
American Journal of Hematology 68:149±158 (2001) Comparison of the Effects of Radiographic Contrast Media on Dehydration and Filterability of Red Blood Cells From Donors Homozygous for Hemoglobin A or Hemoglobin S Patricia Losco,1* Gerard Nash,2 Phil Stone,2 and John Ventre3 1 Schering-Plough Research Institute, Lafayette, New Jersey Department of Physiology, University of Birmingham Medical School, Birmingham, United Kingdom 3 Nycomed Amersham Imaging, Princeton, New Jersey 2 Iodinated radiographic contrast media have traditionally been contraindicated in patients with sickle cell disease because their high osmolality may induce osmotic shrinkage of red blood cells, impair blood ¯ow through the microcirculation, and precipitate or exacerbate a sickle cell crisis. This study investigated that concept by comparing the hematological and rheological effects in vitro of four X-ray contrast media of differing osmolalities: Visipaque (290 mOsm/kg), Hexabrix (600 mOsm/kg), Omnipaque (844 mOsm/kg), and RenoCal-76 (1940 mOsm/kg). Blood was tested from 10 normal and 10 sickle cell donors at drug concentrations of 0, 1, 10, and 30% w/v in an attempt to approximate the relative concentrations of contrast medium to blood that might occur during the bolus-injection and circulation-diluted phases of drug administration. Parameters evaluated included hematology, red cell morphology, and red cell ¯ow resistance through a micropore ®lter to approximate the microcirculatory effects. Signi®cant hematological effects for both normal and sickle cell donors included a concentration dependent decrease in hematocrit and MCV, and increase in MCHC, all of which varied directly with the osmolality of the contrast media in the order of RenoCal-76 > Omnipaque > Hexabrix > Visipaque. The contrast media had minor effects on red blood cell morphology except for RenoCal-76, 10 30% in which marked echinocytosis was observed. There was no signi®cant increase in the number of irreversibly sickled cells in donors with hemoglobin S. Filterability of red cell suspensions through capillary size pores was impaired in both normal and sickle cell samples in direct proportion to the osmolality of the contrast media, as listed above. Filterability effects were greater for sickle cells than for normal red cells. Visipaque, which was closest to isotonicity, had little effect on red cell volume and had no signi®cant effect on ®lterability of normal or sickle cells. These results suggest that microcirculatory impairment following infusion of contrast media may occur in sickle patients because of the unusual rheological sensitivity of HbSS red cells, and may be avoided by choice of an isotonic medium. Am. J. Hematol. 68:149±158, 2001. ã 2001 Wiley-Liss, Inc. Key words: radiographic contrast media; red blood cells; sickle cell disease; microcirculation; blood osmolality INTRODUCTION Previous studies of iodinated water-soluble radiographic contrast media (RCM) have demonstrated adverse effects on blood viscosity, erythrocyte morphology and aggregation behavior [1 3], intracellular pH of red cells , and blood rheology [5,6]. Most effects have been attributed to either the viscosity of the media and/or to their high osmolality, up to eight times greater than blood. ã 2001 Wiley-Liss, Inc. Contract grant sponsor: Nycomed Amersham Imaging, Princeton, New Jersey. *Correspondence to: Patricia Losco, V.M.D., Schering-Plough Research Institute, P.O. Box 32, Lafayette, New Jersey 07848. E-mail: email@example.com Received for publication 28 September 2000; Accepted 15 June 2001 150 Losco et al. The latter is due to their high iodine content necessary for radiographic contrast . Hypertonic solutions induce dehydration and crenation of red blood cells and can impede their flow through the microcirculation [8,9] Studies of blood from sickle cell (HbSS) donors show similar effects of crenation of red cells and increased viscosity, which impair rheology in hypertonic media [10 13]. Clinically, hypertonic RCM have been implicated in severe adverse reactions in sickle cell patients including precipitation of vascular occlusive crisis [14 19]. Although hypertonicity alone does not induce morphological sickling of HbS red cells , the resultant dehydration and increase in intracellular Hb concentration does promote polymer formation during physiologic fluctuations in pO2. In fact, some polymer may form in cells with high HbS concentration even when fully oxygenated with consequent reduced filterability . Dehydrated sickle cells do not necessarily take on classical elongated forms in response to HbS polymerization, but may have irregular outline and increased rigidity due to formation of shorter HbS filaments . Such cells are, nevertheless, highly rheologically compromised . Hypertonicity, plus the low oxygen saturation and acidosis that occurs during dosing with some RCM or can preexist in ill patients, may contribute to polymerization and sickling and increase the probability of occlusion in the microcirculation [14,22 23]. In recent years, development of newer contrast media has focused on reducing the osmolality of the RCM and, thereby, some of the toxic effects [24 29]. However, there have been no comparisons of the effects of different agents on the rheology of sickle cells. Visipaque, formulated at 290 mOsmol/kg of water, is the first contrast agent developed for broad intravascular application that has the same osmolality as whole blood. It Abbreviations RCM ISC HbSS HbAA HbS RBC Hb Hct MCV MCH MCHC SEM PBS IRFR CR CP b radiographic contrast media irreversibly sickled cells hemoglobin genotype of sickle cell donors hemoglobin genotype of normal donors hemoglobin S red blood cell count hemoglobin hematocrit mean corpuscular volume mean corpuscular hemoglobin mean corpuscular hemoglobin concentration scanning electron microscopy phosphate buffered saline initial relative flow rate clogging rate clogging particles a calculated value that represents mean cellular resistance to flow through a pore TABLE I. Contrast Media Data Contrast media Iodixanol nonionic dimer (Visipaque) Ioxaglate ionic dimer (Hexabrix) Iohexol nonionic monomer (Omnipaque) Diatrizoate sodium meglumine ionic monomer (RenoCal 76) Concentration of iodine (mgI/ml) Osmolality (mOsmol/kg water) 320 290 320 600 350 844 370 1,940 might, therefore, be expected to have minimal effects on red cell parameters and should not cause the marked deterioration in the flow properties of HbS cells that is associated with red cell dehydration. This study was performed to compare the iso-osmolar contrast agent, Visipaque, with other contrast media, covering a range of increasing osmolalities, with regard to hematological parameters, morphology of normal and HbS erythrocytes, and their resistance to flow. MATERIALS AND METHODS Blood Donors Venous blood (15 ml) was collected in EDTA from ten adults with homozygous (HbSS) sickle cell disease in the steady state and from 10 adults with normal (HbAA blood). Donors gave informed consent. HbS donors had not received a blood transfusion within the four months prior to donating blood or received treatment with any drugs that would alter their concentration of HbS. Contrast Media The contrast media studied are listed in Table I. They were selected to provide a range of osmolalities with minimal differences in viscosity. Although the concentration of iodine (mgI/ml) in the different contrast media was not identical, the maximum difference in iodine concentration between test samples was approximately 15%, which was not considered significant in affecting study results. Aliquots of 500 ll of whole blood or red cell suspension (109 cells/ml) were mixed with contrast media and phosphate buffered saline (PBS) in the proportions shown in Table II. The final mixtures contained contrast media at 0, 1, 10, or 30% v/v, and red cells at half their original concentration. Thirty percent was considered a reasonable concentration to represent the bolus-injection phase of arteriography necessary to achieve successful imaging . The blood and erythrocyte suspensions Effects of X-Ray Contrast Media on Hemoglobin S Red Blood Cells TABLE II. Volumes of Blood or Red Cell Suspension Mixed With Contrast Media before Analysis Volume of blood or red cell suspension l1 500 500 500 500 Volume of contrast media l1 0 10 100 300 Volume of PBS l1 Final concentration of contrast media (%v/v) 500 490 400 200 0 1.0 10 30 were incubated at 37°C for 10 minutes with agitation before measurement of hematological and rheological parameters. Hematological Measurements Osmolality (mOsmol/kg) was measured for treated and untreated samples of blood and red cell suspensions by freezing point depression after centrifugation of a small aliquot to remove cells. Red blood cell count (RBC) was measured by Coulter counter (Coulter Electronics Ltd, Luton, UK). Hemoglobin concentration (Hb) was measured by the cyanmethemoglobin method, and hematocrit (Hct) was measured using microhematocrit tubes centrifuged at 15,000 g for three minutes. Derived parameters were mean corpuscular volume (MCV = Hct/RBC), mean corpuscular hemoglobin (MCH = Hb/RBC), and mean corpuscular hemoglobin concentration (MCHC = Hb/ Hct). An automated counter was not used to determine MCV directly because sample-diluting media would reverse effects of hypertonic contrast media. Hematocrit (and thus MCV) may be overestimated slightly for rigid cells, which pack less well than normal cells on centrifugation, but this effect is small . Red Cell Morphology Morphology was assessed using a small aliquot of blood, fixed by dilution 1:10 with 1% glutaraldehyde solution made up in 2/3 strength PBS containing appropriate volumes of contrast media to match osmolality to test samples. Wet films were examined by light microscopy, and 400 red cells were classified by shape as discocytes, stomatocytes, echinocytes, or irreversibly sickled cells (ISC). Changes in morphology were confirmed by examining treated and untreated samples from one normal donor and two sickle cell donors (selected at random) by scanning electron microscopy (SEM). The same glutaraldehyde-fixed blood suspensions were used. Electron microscopy specimens were prepared as follows: round (13 mm) coverslips were dipped into a so- 151 lution containing 10% gelatin dissolved in distilled water and were air dried. On each of the coverslips was placed 100 ll of the glutaraldehyde fixed RBCs. An additional 100 ll of 2% glutaraldehyde in PBS was added, and allowed to fix for 30 minutes at room temperature. The prepared coverslips were washed in PBS and were then fixed in 1% OsO4 in PBS for 30 minutes at room temperature. The samples were dehydrated in an ascending series of ethanol concentrations and dried with CO2 at 38°C and 1,200 psi. The glass coverslips were coated with gold using a Pelco SC-6 sputter coater and viewed on a Hitachi 2460N scanning electron microscope. Images from selected samples were digitally recorded using Quartz PCI software. Rheological Measurements Resistance of red cells to flow through capillarysized pores was measured using suspensions of purified red cells. Leukocytes and platelets were removed from blood by filtration through Imugard IG 500 cotton wool (Terumo Corporation, Tokyo, Japan), and red cells were washed twice with PBS. Cells were counted in a Coulter counter and adjusted to109/l with PBS. The deformability of red cells in suspension (final concentration = 5108 cells/ml) was assessed by measuring flow rate through 5 lm pore filters in a St. George's filtrometer . The device measures the time taken for three equal, sequential volumes of red cells to pass through a vertically mounted filter. The flow rate is expressed relative to the flow rate of suspending medium alone (PBS with the appropriate concentration of contrast media added). Linear regression of the relative flow rate versus volume filtered allows calculation of the initial relative flow rate (IRFR) and the clogging rate (CR), or rate of decrease of flow per milliliter. IRFR and cell concentration are used to calculate the parameter b, which represents a measure of the average resistance of individual cells to flow through the pores [33,34]. CR and the known number of filter pores allow calculation of the clogging particles (CP/ml), which gives a measure of the concentration of cells that blocked the pores. The viscosities of suspending media were measured by a capillary viscometer (Coulter Electronics, Ltd.). Statistical Methods Treatment related effects, associated with red blood cell shrinkage, were observed in the hematocrit and the calculated values of MCV and MCHC. MCV was selected as the best representative parameter to test the effect statistically. Treatment related effects on red cell suspension filterability were also observed and tested statistically. 152 Losco et al. For each agent, pairwise comparisons between 0% and each of the remaining concentrations (1, 10, and surrogate 30%) were made for MCV, and the filterability parameters of b and CP. A two-way analysis of variance (ANOVA) was used, with factors for concentration and donor (blocking factor) included in the model . Pairwise comparisons were made using linear contrasts comparing the specified levels of concentration. Each contrast was tested at the Bonferroni corrected level of a = 0.0167 (0.05 divided by the number of comparisons). Linear regression was used to determine the relationship between MCV and osmolality. The model regressed the parameter 1/osmolality onto MCV. In the case of sickle blood treated with RenoCal-76, the 30% concentration was not included in the analyses on b and CP due to stopped flow. For comparisons of the four media within each concentration, the two-way ANOVA was again used, including factors for medium and donor. All pairwise comparisons between the four mediums were performed using linear contrasts. Each contrast was tested at the a = 0.01 level. These analyses were modified if heterogeneity of variance between the concentrations or media was noted. In this case, paired t-tests were performed to compare either the media or concentrations. RESULTS Osmolality Osmolality was measured using the plasma after centrifuging blood/contrast agent mixtures to remove the red cells. However, it was not possible to reliably separate red cells for any of the contrast media at 30% concentration. The red cells either floated to the top or remained suspended throughout the liquid column with portions at the top and/or bottom. This indicated that either the density of the fluid phase was greater than that of the red cells, or the red cells had a density spanning a range including the fluid density. If cells floated to the top, osmolality could be measured for the fluid column, but in many cases this was not possible. In no case could the Hct be measured reliably. To obtain representative values for the 30% concentration, a surrogate sample was generated by adding concentrated saline with the same osmolality as the contrast media to the blood. This sample was then used to measure the hematological parameters for the 30% concentration. It should be noted that red cell morphology could be measured in all samples, as centrifugation was not required. Filtration measurements were also unaffected, as they used samples with known cell count rather than hematocrit. Table III shows the data for osmolality. Values for 30% mixtures are shown where possible (generally n < 10), along with values for the surrogate samples (n = 10). The undiluted media had osmolality in the order TABLE III. Mean Values for Osmolality of Blood With Contrast Media Added Agent 0%a A. Normal blood Visipaque 295 3 Hexabrix 295 3 Omnipaque 295 3 RenoCal-76 295 3 B. Sickle blood Visipaque 291 2 1% 10% Osmolality (mOsmol/Kg) 296 1 311 3 331 7 n 10 302 3 342 6 429 6 n 2 302 2 357 6 491 3 n 7 314 5 439 18 801 15 n 6 294 3 309 3 291 2 298 4 337 3 Omnipaque 291 2 300 1 354 5 RenoCal-76 291 2 311 4 428 8 Hexabrix 30%b 330 n 1 n 0 484 n 1 778 n 2 Surrogate 30% 339 2 480 18 543 6 813 10 337 3 474 4 531 4 793 9 *Data are mean SD from 10 measurements except where stated. a The same control sample was used for the comparisons of all contrast media. b Data are from n samples where red cells separated from medium upon centrifugation (other samples were not measurable). Visipaque < Hexabrix < Omnipaque < RenoCal-76 (Table I), and the samples diluted in blood retained this order. At 30% suspension, all media were hypertonic, including Visipaque. Suspensions containing RenoCal76 had by far the greatest osmolality. Hematological Parameters The sickle donors had RBC and Hb approximately half those of the normal donors. However, Hb, RBC, and MCH were not significantly altered by exposure to any of the contrast media (data are not shown). Values for Hct (Table IV) and MCV (Table V) were decreased, and values for MCHC (data not shown) were increased in direct proportion to the osmolality of the contrast media added. Because all three parameters equally reflect the degree of cell shrinkage, MCV was chosen for discussion here. For normal blood, all media induced significant reduction in MCV at 30% v/v (P < 0.0167). Omnipaque, Hexabrix, and RenoCal-76 also caused significant shrinkage at 10%. For sickle blood, Omnipaque, Hexabrix, and RenoCal-76 induced significant changes at 10 and 30% v/v (P < 0.0167), but Visipaque caused no significant changes at any concentration. No agent caused significant shrinkage at 1% for either normal or sickle blood. For both types of red cells, the degree of shrinkage was consistently ordered among the four media, regardless of concentration, as RenoCal-76 > Omnipaque > Hexabrix > Visipaque. Significant differences Effects of X-Ray Contrast Media on Hemoglobin S Red Blood Cells 153 TABLE IV. Mean Values for Hct of Blood With Contrast Agents Added* 0%a 01% 10% Surrogate 30% A. Normal blood Visipaque Hexabrix Omnipaque RenoCal-76 0:220 0:013 0:220 0:013 0:220 0:013 0:220 0:013 0:219 0:013 0:218 0:012 0:219 0:012 0:211 0:012 0:215 0:014 0:208 0:013 0:203 0:010 0:182 0:014 0:204 0:012 0:173 0:011 0:166 0:005 0:144 0:009 B. Sickle blood Visipaque Hexabrix Omnipaque RenoCal-76 0:128 0:018 0:128 0:018 0:128 0:018 0:128 0:018 0:130 0:016 0:127 0:017 0:128 0:016 0:125 0:015 0:129 0:016 0:121 0:015 0:118 0:016 0:110 0:013 0:121 0:014 0:103 0:012 0:099 0:011 0:088 0:010 Agent *Note that each blood sample has been diluted by 50% (500 ll blood/ml) and, therefore, values are half the actual patientÕs value for Hct. a The same control sample was used for the comparisons of all agents. TABLE V. Mean Values for MCV of Red Cells in Blood With Contrast Agents Added 0%a TABLE VI. Summary of Signi®cant Pairwise Comparisons of MCV Among the Four Tested Media 1% 10% Surrogate 30% A. Normal blood Visipaque 87 4 Hexabrix 87 4 Omnipaque 87 4 RenoCal-76 87 4 MCV (fl) 89 4 87 2 86 4 86 3 Concentration of agent % v/v Significant differences in levels of MCV 86 4 84 4b 80 4b 72 4b 83 4b 69 3b 65 3b 59 3b A. Normal blooda 1% 10% Surrogate 30% No differences V>R, V>O, H>R, O>R V>H, V>O, V>R, H>O, H>R, O>R B. Sickle blood Visipaque Hexabrix Omnipaque RenoCal-76 91 13 89 12 88 12 88 13 91 12 84 14b 81 13b 77 10b 87 12 72 10b 68 10b 62 9b B. Sickle blooda 1% 10% Surrogate 30% V>O V>O, V>H, V>R, H>R V>H, V>O, V>R, H>O, H>R, O>R Agent 90 15 90 15 90 15 90 15 a The same control sample was used for the comparisons of all agents. b P < 0.0167. (P < 0.01) between the mediums depended on the level of concentration and generally supported the order of osmolalities of the contrast media (Table VI). Values for MCV showed a close inverse correlation with osmolality (R2 = 0.96 for linear regression of MCV vs. 1/Osmolality for normal blood, (R2 = 0.98 for sickle blood). The regression line for normal blood was MVC/ MCVo = 0.43 + 0.58 (295/osmolality), where MCVo is the volume in plasma alone (295 mOsmol/kg). The regression line for sickle blood was MVC/MCVo = 0.49 + 0.52 (291/osmolality), where MCVo is the volume in plasma alone (291 mOsmol/kg). Red Cell Morphology Morphology results are presented in Table VII, sections A and B, for normal and sickle blood, respectively. Evaluation of red blood cells by SEM confirmed the results of the light microscopic examination of the cells. In control normal donor samples, the vast majority of red cells were discocytes. The control sickle samples displayed a variety of red cell forms, including stomato- a R = RenoCal-76, O = Omnipaque, H = Hexabrix, V = Visipaque. cytes, reticulocytes, and discocytes, as well as elongate boat shaped cells (irreversibly sickled cells), which predominated over the classical ÔÔsickleÕÕ shape. None of the contrast media produced significant changes in red cell shape at 1%, and only RenoCal-76 caused a noticeable change at 10%, with an increase in echinocytes (shrunken cells). At 30%, Visipaque produced a slight increase in stomatocytes, but showed no change in the incidence of echinocytes for either normal or sickle blood (Fig. 1). At 30%, Hexabrix and Omnipaque produced a slight increase in echinocytes (4 and 9%, respectively, for normal donors), whereas RenoCal-76 caused a marked increase in echinocytes for both normal and sickle donors (Figs. 2 and 3, respectively). The effects of RCM on morphology of discocytes were similar for normal and sickle blood, although sickle blood contained a greater variety of red cell shapes before media were added and showed a wider range of morphologic changes when exposed to contrast media. Notably, shrinkage of sickle cells at high RCM concentrations was not associated with an increase in classical sickling. Also, the ISC or sickled component of blood from HbS donors usually did not show echinocytosis. 154 Losco et al. TABLE VII. Morphology of Red Cells in Blood/Contrast Media Mixtures Agent A. Normal blooda None Visipaque Hexabrix Omnipaque RenoCal-76 B. Sickle blooda None Visipaque Hexabrix Omnipaque RenoCal-76 (%) Stomatocytes (%) Discocytes (%) Echinocytes (%) 0 1 10 30 1 10 30 1 10 30 1 10 30 65 85 85 13 9 74 53 22 64 33 11 54 11 0 93 5 92 5 91 5 86 9 92 3 92 2 94 2 93 3 94 2 91 10 93 3 67 17 0 11 11 11 11 11 32 42 11 33 9 10 21 33 17 100 0 1 10 30 1 10 30 1 10 30 1 10 30 43 53 53 76 32 32 11 43 22 21 33 11 11 65 10 66 10 66 11 63 10 63 11 61 10 61 12 66 11 63 11 57 15 63 10 44 16 33 11 6 10 7 11 9 12 9 14 8 17 9 18 8 11 7 15 9 22 13 13 8 37 18 74 13 Sickled cells (%) (ISC) 20 8 20 11 22 16 18 9 21 10 20 10 20 11 19 9 20 10 19 10 21 10 18 9 21 11 a Mean values for 10 patients each. Fig. 1. Scanning electron micrograph of red cells from a sickle donor exposed to 30% Visipaque. Red cell appearance is similar to 0% control. Fig. 2. Red cells from a normal donor exposed to 30% RenoCal-76 showing 100% echinocyte formation. Filterability of Red Cell Suspensions Filtration measurements allowed calculation of the average resistance of individual cells to flow through the pores (b) and the concentration of pore blocking particles (CP). The filterability measurements were made relative to suspending medium alone, so that changes in viscosity were taken into account in calculation of the filtration parameter b. Viscosity of the contrast media in PBS was measured independently. Each RCM caused marked and comparable increase in viscosity (Table VIII). Effects of X-Ray Contrast Media on Hemoglobin S Red Blood Cells 155 Fig. 3. Red cells from a sickle donor exposed to 30% RenoCal-76. Most nonsickled red cells are echinocytes, but sickled cells show little change. TABLE VIII. Values for the Viscosity of PBS/Contrast Media Mixtures Viscosity (mPa.s) Agent 0%a 1% 10% 30% Visipaque Hexabrix Omnipaque RenoCal-76 0.90 0.90 0.90 0.90 0.91 0.91 0.92 0.92 1.03 1.01 1.05 1.03 1.50 1.42 1.53 1.49 a The same control sample was used for the comparisons of all constrast media. Data for b are presented in Fig. 4 and for CP in Fig. 5. Values for b and CP were higher for sickle than normal red cells. There was much greater variation in filtration parameters and responses to media for sickle blood compared with normal blood. Although the trends were often the same, statistical significance was not reached as often for sickle blood, even though the magnitude of changes induced by media was generally larger. Visipaque did not significantly alter filterability of either normal or sickle cells. For normal cells, Hexabrix significantly increased b and CP, but only at 30%. Omnipaque significantly increased b at 10 and 30%, and CP at 30%. RenoCal-76 significantly increased b and CP at 10 and 30%. For sickle cells, Hexabrix and Omnipaque significantly increased b and CP only at 30%. RenoCal-76 significantly increased CP, but not b, at 10%. Neither b nor CP could be evaluated for RenoCal76 at 30%, due to stoppage of flow. However, there was clearly massive increase in flow resistance. For either type of red cell, results for b and CP were consistently ordered as RenoCal-76 > Omnipaque > Hexabrix > Visipaque. Statistical comparisons between contrast Fig. 4. Effect of increasing concentration of radiographic contrast media on mean resistance to ®ltration (b) of red cells from (A) HbAA donors and (B) HbSS donors. Data are mean the standard error of the mean from measurements on 10 donors. For 30% RenoCal-76, ¯ow rate was too slow to allow measurement (*). media for b and CP showed significant variation depending on the level of concentration (P < 0.01), but the changes were inconsistent relative to defined parameters of osmolality, and did not provide useful insight in interpreting the data (data not shown). DISCUSSION Patients with sickle cell disease characteristically suffer from periodic painful crises caused by occlusion of microvessels. HbSS red cells are abnormally resistant 156 Losco et al. Fig. 5. Effect of increasing concentration of radiographic contrast media on the number of clogging particles during ®ltration of red cells from (A) HbAA donors and (B) HbSS donors. Data are mean the standard error of the mean from measurements on 10 donors. For 30% RenoCal-76, ¯ow rate was too slow to allow measurement (*). to flow even when oxygenated, and this abnormality is exacerbated on deoxygenation when polymerization of HbS causes formation of elongated rigid cells [36 39]. The impaired flow of oxygenated HbSS red cells is largely attributable to the existence of a population of unusually dense cells, including both dense discocytes containing HbS polymer and boat-shaped irreversibly sickled cells [20,36]. In the present study, the preexisting rheological abnormality was evident as increased values for b and CP for untreated sickle cells compared with normal red cells. The likelihood of new polymer formation occurring during transit through the microcirculation in vivo is increased if the intracellular concentration of HbS is increased, because the delay time for gelation is highly dependent on the HbS concentration . Thus, infusion of hypertonic RCM may cause adverse reactions in sickle patients, both because of deterioration in the flow properties of oxygenated cells and also because decrease in MCV and associated increase in MCHC will increase the chance of sickling and further impairment. Hypertonic RCM also caused dehydration of normal red cells and impairment of flow properties, in agreement with previous reports [3 6]. The increase in flow resistance was not so dramatic as with HbSS cells, and probably arose from an increase in viscosity of the cytoplasmic HbA solution. At MCHC around 500 g/l, the cytoplasm of HbAA cells shows transformation to solidlike behavior, but the rheological effect is not so dramatic as that caused by sickling of HbS . The implications of these studies are, therefore, that dehydration of red cells by RCM is undesirable in all blood because it can increase flow resistance, and it is especially to be avoided in sickle blood where more dangerous changes in flow properties may ensue. Osmolality was the major determinant of changes in most of the parameters studied, judging from the relative magnitude of changes in osmolality, and in hematological and rheological parameters for the different contrast media. Changes in MCV were linearly, inversely related to changes in osmolality, so that shrinkage was fully explained by the increases in osmolality caused by the contrast media. Linear regression of MCV vs. l/Osmol-1 ality for normal red cells yielded an equation that agreed well with literature reports obtained using saline . The equation for sickle cells indicated that they shrank less than normal cells in hypertonic medium. This suggests that there is slightly less osmotically active or exchangeable water in sickle cells. Hemoglobin, red cell counts, and MCH were not affected by addition of contrast media in either normal or sickle blood. Hemoglobin levels of samples were not expected to vary on addition of media, because equal volumetric dilutions were used for all samples in this protocol. Constancy of red cell counts indicated that there was negligible hemolysis of red blood cells due to any RCM. Nor was there any discoloration of the plasma that might indicate free Hb. MCH, calculated from Hb/RBC, would also not be affected. Of the agents tested here, only Visipaque had no significant effect on filterability parameters for either normal or sickle red cells. The other RCM caused large changes in filterability at 30%, whereas RenoCal-76 also caused marked changes at 10%. In the circulation, RCM concentrations in the range of 30% concentration exist only for brief periods during the bolus injection phase of drug administration before being diluted by the circulating blood. Clinical studies have shown, however, that achieving high concentrations of 20 65% in the target organ being radiographed is common and considered Effects of X-Ray Contrast Media on Hemoglobin S Red Blood Cells necessary for achieving good vascular imaging . Therefore, the effects of high RCM concentrations on flow resistance constitute a valid concern for adverse effects on the patient. RCM also caused modification of morphology of the red cells. However, only RenoCal-76 induced marked changes, with high levels of echinocytes formed. This may be partly an effect of high ionic strength. However, it was notable that even at comparable osmolality, 30% Hexabrix or Omnipaque did not induce the same level of echinocytosis as 10% RenoCal-76. It is uncertain how great the effect of the echinocytosis was on flow resistance in this study. In general, spheroechinocytes have high resistance to deformation, but lesser degrees of echinocytosis have little effect . Sickle blood was slightly more resistant to formation of echinocytes than normal blood. This may reflect resistance to transformation of rigid dense cells, including sickle discocytes containing HbS polymer. It may also be related to the fact that the irreversibly sickled population (about 20% of RBCs in our cohort) were already dehydrated cells. Shrinkage did not increase the percentage of ISC, consistent with the finding that dense cells (sickle discocytes) do not form classical elongated shapes when dehydrated . This study addressed only the physical properties of RCM and their interactions with red blood cells. Significant additional effects on flow in large conduit blood vessels and the microcirculation may arise in response to RCM administration secondary to the release of cytokines from white blood cells, platelets, and the vascular endothelium [42 45]. The effects of abnormalities in pH and oxygen saturation, both of which may be compromised in sickle cell patients in crisis, also were not investigated here. Nevertheless, this study supports the concept that hypertonic contrast media may increase the risk of microcirculatory occlusion in sickle cell patients. Use of RCM with nearly physiologic osmolality may not eliminate all adverse effects of contrast arteriography, but it is likely to reduce interference with red cell passage through the microcirculation, particularly for sickle cell patients. ACKNOWLEDGMENTS The authors wish to thank J. Suhan for technical assistance in electron microscopy and V. Sansone and T. Santillo for technical assistance in the planning and preparation of this manuscript. We are grateful to S. Stevens for liasing with blood donors and carrying out phlebotomy, and to Drs. D. Bareford and J. Wilde for enabling recruitment of donors with homozygous sickle cell disease. 157 REFERENCES 1. Aspelin P, Schmid-Schönbein H. Effect of ionic and non-ionic contrast media on red cell aggregation in vitro. Acta Radiolic Diag 1978;19:766 784. 2. Aspelin P, Stöhr-Liessen M, AlmØn T. Effect of Iohexol on human erythrocytes, I. Changes of red cell morphology in vitro. Acta Radiol Supplementum 1980;362:117 122. 3. Hardeman M, Goedhart P, Koen I. The effect of low-osmolar ionic and nonionic contrast media on human blood viscosity, erythrocyte morphology, and aggregation behavior. Invest Radiol 1991;26: 810 819. 4. Rosenthal A, Mesrobian A. The relationship between angiography, intraerythrocytic pH and hemoglobin oxygen equilibrium. Invest Radiol 1975;10:140 47. 5. Lerche D, Hennicke G. The effects of different ionic and non-ionic X-ray contrast media on the morphological and rheological properties of human red blood cells. Clin Hemorheol 1992;12:341 355. 6. Strickland N, Rampling M, Dawson P, Martin G. The effects of contrast media on the rheological properties of blood. Clin Hemorheol 1992;12:369 379. 7. Pugh N. Haemodynamic and rheological effects of contrast media: the role of viscosity and osmolality. Eur Radiol 1996;6:S13 S15. 8. Dawson P, Harrison M, Weisblatt E. Effect of contrast media on red cell filtrability and morphology. Brit J Radiol 1983;56:707 710. 9. Le Mignon M, Nain Dit Ducret M, Bonnemain B, Donadieu A. Effect of contrast media on whole blood filtrability. Acta Radiol 1988; 29:593 597. 10. Perillie P, Epstein F. Sickling phenomenon produced by hypertonic solutions: a possible explanation for the hyposthenuria of sicklemia. J Clin Invest 1963;42:570 580. 11. Murphy J, Wengard M, Brereton W. Rheological studies of Hb SS blood: influence of hematocrit, hypertonicity, separation of cells, deoxygenation, and mixture with normal cells. J Lab Clin Med 1976;87:475 486. 12. Clark M, Guatelli J, Mohandas N, Shohet S. Influence of red cell water content on the morphology of sickling. Blood 1980;55: 823 830. 13. Rao V, Rao A, Steiner R, Burka E, Grainger R, Ballas S. The effect of ionic and nonionic contrast media on the sickling phenomenon. Radiology 1982;144:291 293. 14. Richards D, Nulsen F. Angiographic media and the sickling phenomenon. Surg Forum 1971;22:403 404. 15. Stockman J, Nigro M, Mishkin M, Oski F. Occlusion of large cerebral vessels in sickle-cell anemia. New Eng J of Med 1972;287:846 849. 16. Darr M, Hamburger S, Koprivica B, Ellereck E. Hemolytic anemia associated with a radiopaque contrast agent in a patient with hemoglobin SC disease. South Med 1981;174:1552. 17. Rao A, Thompson R, Durlacher L, James F. Angiographic contrast agent-induced acute hemolysis is a patient with hemoglobin SC disease. Arch Intern Med 1985;145:759 760. 18. Banna M. Post-angiographic blindness in a patient with sickle cell disease. Invest Radiol 1992;27:179 181. 19. Zagoria R. Iodinated contrast agents in neuroradiology. Neuroimaging Clin N Am 1994;4:1 8. 20. Green M, Noguchi C, Keidan A, Marwah S, Stuart J. Polymerization of sickle cell hemoglobin at arterial oxygen saturation. J Clin Invest 1988;81:1669 1674. 21. Nash GB, Johnson CS, Meiselma HJ. Influence of oxygen tension on the viscoelastic behavior of red blood cells in sickle cell disease. Blood 1986;67:100 118. 22. Hofrichter J, Ross PD, Eaton WA. Kinetics and mechanism of deoxyhemoglobin S gelation: a new approach to understanding sickle cell disease. Proc Nat Acad Sci USA 1974;71:4864 4868. 158 Losco et al. 23. Bookchin RM, Balazs T, Landau LC. Determinants of red cell sickling. Effects of varying pH and of increasing intracellular concentration by osmotic shrinkage. J Lab Clin Med 1976;87:597 616. 24. Grainger R. Formulation and clinical introduction of low osmolality contrast media. Radiologe 1981;6:261 267. 25. Bettman M. Angiographic contrast agents: conventional and new media compared. Am J Radiol 1982;139:787 794. 26. Speck U, Mützel W, Weinmann H. Chemistry, physicochemistry and pharmacology of known and new contrast media for angiography, urography and CT enhancement. In: Taenzer V, Zeitler E, editors. Contrast media in urography, angiography and computerized tomography. New York: Thieme-Stratton; 1983. p 2 10. 27. Grynne B, Nossen J, Bolstad B, Borch K. Main results of the first comparative clinical studies on Visipaque. Acta Radiol 1995; 36: 265 270. 28. Pugh N, Griffith T, Karlsson J. Effects of iodinated contrast media on peripheral blood flow. Acta Radiol Suppl 1995; 399:155 163. 29. Morcos S, Dawson P, Pearson J, Jeremy J, Davenport A, Yates M, Tirone P, Cipolla P, de Haen C, Muschick P, et al. The haemodynamic effects of iodinated water soluble radiographic contrast media: a review. Eur J Radiol 1998;29:31 46. 30. Kutt H, Schick R, Rennie L, McDowell F. The quantities of sodium diatrizoate in cerebral vessels during arteriography. J Neurosurg 1963;20:515 519. 31. Nash GB, Meiselman HJ. Alteration of red cell membrane viscoelasticity by heat treatment: effect on red cell deformability and suspension viscosity. Biorheology 1985;22:73 84. 32. Dormandy J, Flute P, Matral A, Bogar L, Mikita J. The new St GeorgeÕs blood filtrometer. Clin Hemorheol 1985;5:975 983. 33. Skalak R, Impelluso T, Schmalzer E, Chien S. Theoretical modeling of filtration of blood cell suspensions. Biorheology 1983;20:41 56. 34. Nash G. Filterability of blood cells: Methods and clinical applications. Biorheology 1990;27:873 882. 35. Neter J, Wasserman W, Kutner M. Applied linear statistical models, 3rd ed. Homewood, IL: Irwin Publishing; 1990. p 915 924. 36. Kaul DK, Fabry ME, Windisch P, Baez S, Nagel RL. Erythrocytes in sickle cell anemia are heterogeneous in their rheological and hemodynamic characteristics. J Clin Invest 1983;72:22 31. 37. Linderkamp O, Meiselman HJ. Geometric, osmotic and membrane mechanical properties of density-separated human red cells. Blood 1982;59:1121 1127. 38. Baez S, Kaul DK, Nagel RL. Microvascular determinants of blood flow behaviour and HbSS erythrocyte plugging in microcirculation. Blood Cells 1982;8:127 137. 39. Hebbel RP. Beyond hemoglobin polymerisation: the red blood cell membrane and sickle disease pathophysiology. Blood 1991;77: 214 237. 40. Nash G, Meiselman H. Effect of dehydration on the viscoelastic behaviour of red cells. Blood Cells 1991;17:517 522. 41. Meiselman H. Morphological determinants of red cell deformability. Scand J Clin Lab Invest 1981;41:27 34. 42. Coel M, Lasser E. A pharmacologic basis for peripheral vascular resistance changes with contrast media injections. Am Radiol J 1971; 111:802 806. 43. Ring J, Arroyave C. Alteration of human blood cells and changes in plasma mediators produced by radiographic contrast media. Z ImmunForsch Immunobiol 1979;155:200 211. 44. Assem E, Bray K. The release of histamine from human basophils by radiological contrast agents. Brit J Radiol 1983;56:647 652. 45. Dawson P. Chemotoxicity of contrast media and clinical adverse effects. Invest Radiol (Suppl) 1985;20:S84 S91.