Download Article (PDF)

Home , Diatrizoate, Iobitridol

In-vitro evaluation of contrast media for assessment of regional perfusion distribution by Electrical Impedance Tomography (EIT)

N Ch Hellige1, B Meyer1, T Rodt1, J VogelClaussen1, G Hahn2, G Hellige2 1 Hannover Medical School, Institute for Diagnostic and Interventional Radiology, Hannover, Germany, (email: hellige.niels@mhhannover.de); 2 University Hospital Goettingen, Centre of Anaesthesiology, Intensive and Critical Care Medicine, Goettingen, Ger many

Abstract Electrical impedance tomography (EIT) has become an accepted tool for monitoring regional ventilation. Interest is growing to derive additional information on pulmonary perfusion and ventilation/perfusion distribution. Favoured ap proach is the impedance indicator dilution technique. Until now hyperosmolar sodium chloride solutions (5.8% and 20%NaCl) have been used as indicators in experiments which, however, may be problematic in view of side effects and are not approved for application in patients. Different radiographic contrast media (CM) approved for use in patients and five reference solutions were tested in vitro. Resistivities were measured in blood at 37°C by a conductivity meter (LF39, Meinsberg GmbH, Germany) in the indica tor content range from 0 to 28 Vol%. Three nonionic CM: Iodixanol, Iomeprol, Iobitridol and two ionic CM: Ioxaglate and Diatrizoate were examined. As references 0.9%, 5.8% and 20%NaCl as well as isotonic (4.5%) and hypertonic (18.2%) glucose solutions were included. Proportionality Proportionality between test solution concentration and deter mined resistivity changes was proven by regression/correlation analysis. Nonionic CM induced linear increases of blood resistivity by 80% (Iodixanol), 75% (Iomeprol) and 66% (Iobitridol) at the highest test concentration of 28Vol%. Ionic CM changed resistivity by 13% (Ioxaglate) and 24% (Diatrizoate). The determined r2values were > 0.988. Both hyperosmolar NaCl solutions caused marked but nonlinear resistivity lowering by 82% resp. 95% with poor r2values < 0.6. Glucose solutions increased resistivity by 17% resp. 29%, r2value 0.98. Hyperosmolar NaClsolutions effect impressive changes of resistivity but may imply errors in the evaluation of indicator dilution curves by nonlinearity. The resistivity changes by the approved nonionic radiographic CM are not complicated by linearity problems. A comparison with the simple blood dilution effect by isotonic nonionic reference solutions indi cates additional mechanisms of resistivity increment by CM. In conclusion CM should be evaluated further on in view of EITbased perfusion assessment in patients.

1 Introduction tion related impedance changes and 2) use of an impe dance indicator dilution technique. Ventilation distribution in the lungs varies physiologically The separation of ventilation and heart action related sig dependant on body position due to changes of gravity in nal components by filtering is questionable due to signal fluence. The distribution normally is in congruence with component interference of ventilation and vascular pulsa the variations of intrapulmonary perfusion distribution. tion related impedance change amplitudes [1]. This leads to a regional matching of both ventilation and Consequently the indicator dilution approach which in perfusion as a prerequisite of optimisation of gas exchange principle is well established for perfusion assessment is in the lungs. Pathologically vascular occlusion e.g. by favoured. Up to now sodium chloride solutions have been pulmonary embolism, obstruction of airways or severe used. In 1992 Brown et al. [2] applied an isotonic 0.9% lung injury often cause a mismatch of local ventilation and NaCl solution. About 10 years later the first usage of a hy perfusion (V˙/Q˙ mismatch). Subsequently such conditions pertonic 1molar (5.8%) NaCl solution in animals was pub may result in increased dead space ventilation or shunt per lished [3]. This approach was later on used in animal ex fusion in the lungs leading to critical hypoxemia by failure periments in view of estimation of local ventilation and of gas exchange. perfusion distribution (V˙/Q˙ ratio) [4]. In an actual publi Electrical Impedance Tomography (EIT) provides the faci cation a 20% solution was applied [5]. lity to acquire noninvasively regional ventilation as well Extremely hypertonic sodium solutions may cause major as perfusion related signals and by this opens the opportu water shift and changes of resistance in pulmonary circula nity to detect V˙/Q˙ mismatch. Whilst the assessment of tion. The appearance of the high sodium bolus might be ventilation by EIT is well validated the estimation of per dangerous in patients in view of cardiac contractility and fusion is still in discussion. To evaluate perfusion distribu electrophysiology. Additionally increasing sodium chlo tion two approaches are possible 1) evaluation of heart ac ride concentration causes fairly proportional increase of

conductivity which, however, is inverse to resistivity 2.3 Reference solutions change (conductivity = 1/ resistivity) [6]. The significance As references isotonic and hypertonic NaCl and glucose of these properties of high concentrated NaCl solutions solutions were included. The substance listing presents needs to be clarified. substance concentration (g per 100 ml), electrolytic pro These potential limitations provide the basis to investigate perties, osmolality and measured resistivity: also clinically approved xray contrast media used for an 0.9% NaCl (ionic, isotonic, 0.3 osm/kg, 64 Ωcm) giography in regard to their influence on resistivity of 5.8% NaCl (ionic, hypertonic, 1.9 osm/kg, 12 Ωcm) blood and consequently in respect of an usability as indica 20% NaCl (ionic, hypertonic, 6.6 osm/kg, 0.5 Ω cm) tors for electrical impedance tomography indicator dilution 4,5% glucose (nonionic, isotonic,0.28osm/kg, 54 kΩcm) technique. 18.2% glucose (nonionic, hypertonic, 1osm/kg, 59 kΩcm) 2.4 Data evaluation 2 Methods The data are presented by relative changes of resistivity Different radiographic contrast media (CM) approved for referred to the respective native blood sample (control use in patients were tested by an in vitro experimental condition). Proportionality between test solution concen setup with the aim to evaluate the resulting changes of tration and determined resistivity changes was proven by blood resistivity in dependence of the respective test solu linear regression/correlation analysis. tion concentrations. Additionally five reference solutions were examined simulating specific physicochemical pro 3 Results perties of contrast media. 2.1 Experimental setup 3.1 Contrast media Fresh human venous blood samples (10 ml, Hct 4648%, The resistivity changes obtained by addition of contrast resistivity 186211 Ω cm) were incubated at 37°C in a media to blood are presented in figure 1. thermostat water bath. Test solutions were added by micro pipette in 0.2 ml steps up to 1ml and afterwards in 1ml steps up to 4ml resulting in test solution contents up to Contrast Media in Blood at 37°C 28,6 volume percent (Vol%). 200 Resistivities were measured by a conductivity meter (LF39, Meinsberg GmbH, Germany) using a twoelectrode 180 conductivity cell with graphite electrodes and builtin tem perature probe cell without activated temperature compen sation. Data are presented as resistivities in Ωcm. 160

2.2 Contrast media 140 Five CM were examined representing four different types Iodixanol in view of molecular structure (monomeric resp. dimeric) 120 and electrolytic properties (ionic resp. nonionic). Iomeprol Iobitridol The substance listing presents generic name, trade name, 100 Ioxaglate electrolytic properties, molecular structure, osmolality, Diatrizoate measured resistivity and pharmaceutical company: 80

® ) control of % ( resistivity Iodixanol, (Visipaque 320 , nonionic and dimeric, 60 0.29 osm/kg, 1395 Ωcm, GE Healthcare Buchler GmbH & Co.KG, München, Germany) 40 Iomeprol (Imeron 400®, nonionic and monomeric,

0.72 osm/kg, 10580 Ωcm, Bracco Imaging Deutschland 20 GmbH, Konstanz, Germany), Iobitridol (Xenetix 350®, nonionic and monomeric, 0 0.915 osm/kg, 3720 Ωcm, Guerbet, Roissy CdG Cedex, 0 10 20 30 France) Ioxaglate (Hexabrix 320®, ionic and dimeric, contrast media amount ( Vol% ) 0.58 osm/kg, 216 Ωcm, Guerbet, Roissy CdG Cedex, France) and Fig. 1 Resistivity changes in blood by addition of contrast Diatrizoate (Urografin76®, ionic and monomeric, media (CM) up to 28.6 volume percent. Nonionic CM 2.1 osm/kg, 265 Ωcm, Bayer HealthCare, Leverkusen, cause marked increases of resistivity. The low osmolality Germany) ionic CM Ioxaglate increases slightly, the hyperosmolar ionic Diatrizoate decreases resistivity exhibiting a non linear shape of the concentration vs. resistivity plot.

Nonionic CM induced fairly linear increases of blood re 4 Discussion sistivity by 80% (Iodixanol), 75% (Iomeprol), 66% (Iobi tridol) at the highest test concentration of 28Vol%. Ionic 4.1 Mechanisms of resistivity changes CM changed resistivity by 13% (Ioxaglate) and 24% Solutions applied as indicators for regional perfusion (Diatrizoate). monitoring by Electrical Impedance Tomography (EIT) The r2values for all nonionic CM were > 0.988 indica must act by changing the resistivity of blood. This may be ting linearity between CMcontent and resistivity change. achieved by different approaches: The r2value for Diatrizoate was 0.91 as a hint on deviation Addition of high concentrated ionic solutions like hyper from linearity. tonic sodium chloride solutions increases the plasmatic ion 3.1 Reference solutions concentration. Resistivity is lowered markedly but non The corresponding data for the reference solutions are pre linear. This is due to the fact that an increase in ion content sented in figure 2. The hypertonic NaCl solutions (5,8% increases conductivity linearly which, however, is inverse and 20%) caused a marked but nonlinear resistivity lower to resistivity. ing by 82% resp. 95% with poor r2values < 0.6. Isotonic Isotonic (0.9%) NaCl solution causes a less marked but 0.9% NaCl solution showed a linear decrease of resistivity linear resistivity reduction by increasing the plasma frac up to 44% of control. Glucose solutions (4.5% and 20%) tion of blood and consecutively by lowering hematocrit. increased resistivity by 17% resp. 29%, r2values >0.98. The hematocrit of blood is one of the major parameters de termining the resistivity of whole blood. Addition of nonionic solutions reduces the ion content of Reference Solutions in Blood at 37"C plasma. In parallel hematocrit is reduced by dilution. Since the data show resistivity increases the decrease in plas 200 matic ion content prevails the hematocrit effect. Hypertonic nonionic solutions in general cause osmotic 180 water shift from erythrocytes and extravascular tissue into the plasma of intravascular blood. By this the plasmatic 160 ion concentration will be further reduced and consecu tively the resistivity increases more markedly. 140 The described mechanisms help to explain the resistance changes obtained by the tested solutions. The results of the 120 Gluc. 18,2% ionic and nonionic reference solutions can clearly be ex Gluc. 4.5% plained by the listed mechanisms. 100 NaCl 0.9% NaCl 5.8% Looking at the contrast media Diatrizoate acts like a salt 80 NaCl 20% solution due to the chemical structure as a high concen trated ionic sodium diatrizoate concentration and by this resistivity ( % of control ) control of % ( resistivity 60 reduces resistivity nonlinearly. All nonionic CM increase resistivity markedly, exceeding 40 the simple dilution effect obtained by isotonic nonionic glucose solution. This may in part result from their hyper 20 osmolality and in part from their extremely poor conduc tivity. Additionally further conductivity influencing prop 0 erties of solvents or additives like ion absorbing qualities 0 10 20 30 may be existing. reference solution amount Surprisingly the isotonic CM Iodixanol exhibits the highest ( Vol% ) increase of resistivity.

Fig. 2 Resistivity changes in blood by addition of the ref 4.2 Relevancy for Application erence solutions up to 28.6 volume percent. The nonionic The dilution curve amplitude needed for evaluation of in glucose solutions cause slight and fairly linear increases dicator dilution curves in EIT should reach resistivity of resistivity. The ionic sodium chloride solutions in gen changes of more than 10%. This requirement results from eral decreased resistivity determined by their sodium and the biological induced impedance changes by cardiac re chloride ion content. The isotonic NaCl solution (0.9%) lated oscillations with amplitudes of about 3% which in shows a linear decrease. Both hypertonic NaCl solutions contrast to ventilation related signals (about 15%) can not reduce resistivity powerful showing a nonlinear concen be stopped and superpose the recorded indicator dilution tration vs. resistivity plot. curves. The resulting limited quality of dilution curves recorded by EIT will allow simplified evaluation procedures estab lished in radiology for estimation of tissue perfusion by

dynamic Computed Tomography (CT) or Magnetic Reso resistivity increment by CM. In conclusion CM should be nance Tomography (MRT) imaging using respective con evaluated further on in view of EITbased perfusion as trast media. Descriptive lung perfusion parameters like sessment in patients including a possible need of dose ap time to peak (TTP), maximum slope or peak enhancement plication exceeding the contrast media need in radio may help to determine relative perfusion distribution in graphic angiography. view of detection of regional ventilation/perfusion mis The experimental setup of the present study may be useful match [7]. Application of evaluation procedures according for screening other solutions in view of a possible applica the mean transit time concept of Zierler [8] requires long tion in electrical impedance indicator dilution technique. period recording of dilution curves without interference by heart action and requires known indicator input functions. 6 References Using a 20% NaCl solution the generation of dilution curves with maximum amplitudes of about 20% resistivity [1] Hellige G and Hahn G: Cardiacrelated impedance change can be obtained by 0.5% indicator content in blood. changes obtained by electrical impedance tomogra Using the resistivity data directly a nonlinearity error will phy: an acceptable parameter for assessment of pul result. The amount can be estimated from the conductivity monary perfusion? Critical Care 2011, 15: 430 and resistivity functions calculated by the experimentally [2] Brown BH, Leathard A, Sinton A, McArdle FJ, Smith determined data. In case of a 20% resistivity change is less RW, Barber DC. Blood flow imaging using electrical then 10%. Redoubling the amplitude of the curve up to re impedance tomography. Clin Phys Physiol Meas sistivity changes of 40% in view of better signal quality 1992, 13: 175–179. needs application of higher volumes for increasing the in [3] Frerichs I., Hinz J., Herrmann P., Weisser G., Hahn dicator content. The resulting linearity error is about 30%. G., Quintel M., Hellige G.: Regional lung perfusion as Using the 5.8% NaCl nearly four times higher solution determined by electrical impedance tomography in volumes are needed. The linearity errors are comparable to comparison with electron beam CT imaging. IEEE 20% NaCl. Besides the linearity problem possible physio Trans Med Imaging 2002, 21(6), 646652. logical interactions caused by an osmolality of about 6600 [4] Hellige G, Dittmar J, Just A, Hahn G: Monitoing of mOsm/kg and a sodium concentration of about 3300 pulmonary perfusion distribution by electrical imped mEq/L must be considered even when the intravascular ance tomography. Biomed Tech 2010, 55(Suppl 1):31 dilution is taken into account. Changes of the so 34. [5] Borges JB, SuarezSipmann F, Bohm SH, Tusman G, dium/calcium relation in blood may disturb contractility Melo A, Maripuu E, Sandström M, Park M, Costa and electrophysiology of the heart when the steep NaCl EL,Hedenstierna G, Amato M: Regional lung bolus after lung passage reaches coronary perfusion. A perfusion estimated by electrical impedance properly evaluation of possible side effects of hypertonic tomography in a piglet model of lung collapse. J Appl NaCl boli in critically ill patients is required. Physiol 2012, 112: 225–236. [6] Hellige NC, Hahn G, Hellige G: Comment on Borges Using the clinically approached nonionic radiographic et al. “Regional lung perfusion estimated by electrical contrast media the 20% resistivity change needs a peak impedance tomography in a piglet model of lung col content of about 8% CM in blood. Doses applied for lapse”. J Appl Physiol 2012, 112: 225–236. clinical angiography of the pulmonary circulation seem to [7] Attenberger U I, Ingrisch M, Busing K, Reiser M, be too low and too slowly infused for reaching sufficient Schoenberg S O, Fink, C: Magnetic resonance imag indicator curve maxima. The possibility of increasing the ing of pulmonary perfusion. Technical requirements CM amount and infusion velocity again must be critically and diagnostic impact. Radiologe 2009, 49(8):73947. evaluated. An advantage of using CM as impedance indi [8] Zierler KL; Theoretical basis of indicatordilution cators would be the possibility to validate the EIT ap methods for measuring flow and volume. Circ Res proach by recording dilution curves by EIT and CT in par 1962, 10: 339407. allel.

5 Conclusions Hyperosmolar NaClsolutions effect impressive changes of resistivity but may imply errors in the evaluation of indicator dilution curves by nonlinearity. Possible side effects by osmotic mechanisms or electrolyte level shifts must be taken into account. The resistivity changes by the approved nonionic radio graphic CM are not complicated by linearity problems. A comparison with the simple blood dilution effect of iso tonic nonionic glucose indicates additional mechanisms of

528