THE DESIGN AND USE OF A DYE-DILUTION TECHNIQUE FOR CONTINUOUS MEASUREMENT OF BLOOD FLOW IN THE FOREARM AND HAND OF MAN Thesis for the Degree of Ph‘. D. MICHIGAN STATE UNIVERSITY 7 ROGER ALLEN wounurs A 1968 I LIBRARY 7an ‘Michigat State University This is to certify that the thesis entitled The Design and Use of a Dye-Dilution Technique for Continuous Measurement of Blood Flow in the Forearm and Hand of Man presented by Roger Allen Wolthuis has been accepted towards fulfillment of the requirements for Ph. D. degree in Phxsiology e. 1/ - flow: (Kw , 0 rr A... "an“ 1U! D,’Ph. D. \. Date _ 0.169 Q— I"?! ABSTRACT THE DESIGN AND USE OF A DYEfDILUTION TECHNIQUE FOR CONTINUOUS MEASUREMENT OF BLOOD FLOW IN THE FOREARM AND HAND OF MAN by Roger Allen Wolthuis The continuous-infusion indicator—dilution technique has been used previously to measure blood flow in the forearm and hand of man. During brachial arterial indi— cator infusion, blood samples were taken from downstream veins and a contralateral artery. Samples were later analyzed for indicator concentration in order to calculate blood flow. Thus, with previous methods, measurements were discontinuous, and blood flow values were not immedi- ately available. The present study sought to improve the technique by developing a detection system which provides continuous and immediate blood flow measurements. A Gilson cuvette densitometer, modified for improved stability, was used to measure venous indocyanine green dye concentrations in the dog forelimb and human forearm during intrabrachial arterial dye infusion (.N2 mg/min). Of special concern was the use and accuracy of this instrument during very low cuvette blood withdrawal rates. In a series of eight dogs, dye was infused into the Roger Allen Wolthuis brachial artery of the isolated, pump-perfused forelimb while downstream dye-concentration plateaus were con— tinuously recorded by withdrawing venous blood through the cuvette at 1.5 ml/min. Forelimb flow rates were varied over the range 25 to 225 ml/min. Actual fore- limb flow was measured by the timed collection of total venous outflow in graduated cylinders. In a series of six dogs, dye and RISA were simultaneously infused into the brachial artery of the intact, naturally perfused forelimb while downstream dye-concentration plateaus were continuously recorded by withdrawing venous blood through the cuvette at 1.5 or 2.4 ml/min. Brachial artery inflow was measured independently by a rotameter; discontinuous RISA-dilution flows were calculated by the technique out— lined by Overbeck (1966). Forelimb blood flow was varied over the range 25 to 200 ml/min by adding either nor- epinephrine or acetylcholine to the infusion of indicator. Finally, in the forearms of eight men, dye and RISA were simultaneously infused into the brachial artery and down- stream dye concentrations were continuously recorded for periods up to two hours by withdrawing forearm venous blood through the cuvette at 1.5 or 2.4 ml/min. Blood flow was again varied over the range 13 to 345 ml/min by adding either nor-epinephrine or acetylcholine to the indicator infusion, or by having the subject exercise his hand. Indicator infusion followed the technique Roger Allen Wolthuis outlined by Overbeck (1966). In addition, quantitative measurements were made of systemic recirculating dye during continuous dye infusion. In 69 paired blood flow measurements from the isolated dog forelimb series, dye calculated flow differed from actual forelimb flow by a mean percent difference of —.5 i 10.3 (S.D.). In the naturally perfused dog forelimb, both indicator-dilution techniques over-estimated brachial arterial inflow, probably because of the presence of un- dyed forelimb collateral blood flow. However, there was good correlation (r = .97) between dye and RISA-dilution flow measurements at the higher cuvette blood withdrawal rate (2.A ml/min). In the human forearm studies, the correlation coefficient between dye-dilution and RISA- dilution calculated blood flow was .99 at a cuvette blood withdrawal rate of 2.A ml/min. In addition, during con- tinuous dye infusion for up to three hours, recirculating dye concentrations were both low (< .5 mg/L) and quite constant. The present study has demonstrated, for the first time, the practicality of using a cuvette densitometer at low cuvette blood withdrawal rates for continuously measuring forearm venous dye concentrations, and thus blood flow, during continuous dye infusion. This system provided measurements which were in close agreement with THE DESIGN AND USE OF A DYE-DILUTION TECHNIQUE FOR CONTINUOUS MEASUREMENT OF BLOOD FLOW IN THE FOREARM AND HAND OF MAN By Roger Allen Wolthuis A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Physiology 1968 ACKNOWLEDGMENTS The author wishes to express his sincere appreciation to both Dr. H. w. Overbeck and Dr. W. D. Collings for their assistance, concern, and encouragement during the course of his Ph.D. program. The author is also indebted to the members of his guidance committee, and the teaching staff in the Depart- ment of Physiology for their time and interest in his behalf. ii TABLE OF CONTENTS Page ACKNOWLEDGMENTS . . . . . . . . . . . ii LIST OF TABLES O O O O O O O O O O 0 Vi LIST OF FIGURES . . . . . . . . . . . vii Chapter I. SURVEY OF LITERATURE . . . . . . . 1 Introduction. . . . . . . . . 1 Theory of Indicator (Dye) Dilution Technique. . . . . . . . . 1 Historical Development of Indicator- Dilution Methods . . . . . . l2 DeveIOpment of the Constant Infusion Indicator-Dilution Technique . . 14 Constant Infusion Indicator-Dilution and Forearm Blood Flow . . . . l6 Indocyanine Green: Properties. . . 2O Indocyanine Green and Hepatic Removal . . . . . . . . . 25 Cuvette Densitometry . . . . . . 27 II. INSTRUMENTATION. . . . . . . . . 30 Introduction. . . . . . . . . 30 Stability of the Gilson Cuvette Densitometer. . . . . . . . 30 iii Chapter Page Sensitivity of the Gilson Cuvette to Changes in Blood Withdrawal Rates. . 32 Response of the Gilson Cuvette Densito- meter to Changes in Optical Density . 3A Sensitivity of the Gilson Cuvette Densitometer to Changes in Blood Hematocrit . . . . ._ . . . . A2 Design and Use of a Blood Mixing Chamber 0 o . c o o o o o o o LIB III. CONTINUOUS DYE-DILUTION FLOW MEASUREMENTS IN THE ISOLATED, PUMP-PERFUSED DOG FORELIMB . . . . . . . . . . . U9 Purpose of Investigation. . . . . . A9 Methods . . . . . . . . . . . A9 Results . . . . . . . . . . . 53 IV. CONTINUOUS DYE-DILUTION FLOW MEASUREMENTS IN THE INTEACT, NATURALLY PERFUSED DOG FORELIMB . . . . . . . . . . . 6O Purpose of Investigation. . . . . . 60 Methods . . . . . . . . . . . 60 Results . . . . . . . . . . . 67 V. CONTINUOUS DYE-DILUTION FLOW MEASUREMENTS IN THE HUMAN FOREARM AND HAND . . . . 7“ Purpose of Investigation. . . . . . 74 Methods . . . . . . . . . . . 7A Results . . . . . . . . . . . 82 VI. DISCUSSION AND CONCLUSIONS . . . . . . 91 Recirculating Dye . . . . . . . . 91 iv Chapter Page Actual Versus Calculated Blood Flow . . 92 Indicator Calculated Flow in the Intact Dog Forelimb . . . . . . . . . 93 Dye- Versus RISA-Dilution Calculated FlOWS o o o c o o o o o o o 9“ Measuring Blood Flow in the Forearm and Hand of Man . . . . . . . . . 95 Errors in Technique. . . . . . . . 96 BIBLIOGRAPHY . . . . . . . . . . . . . 98 LIST OF TABLES Table Page 1. Hepatic Removal of Injected Indocyanine Green Dye in Dogs and Humans . . . . 25 2. Arterial and Venous ICG Levels During Constant ICG Infusion . . . . . . 27 3. Data Showing the Relationship Between Size of Calibration Blood Aliquot and Their Reproductability . . . . . . . . A0 A. Mean Percent Change in Optical Density (Pen Deflection) Per 1 Percent De- crease in Blood Hematocrit. . . . . A2 5. Comparison of Actual and Dye-Dilution Calculated Blood Flows in the Isolated, Pump—Perfused Dog Forelimb Series . . 56 6. Measured Blood-Flow Values from the Intact, Naturally Perfused Dog Forelimb Series . . . . . . . . . . . 68 7. Comparison of Dye- and RISA-Dilution Calculated Flow Measurements from the Human Forearm and Hand . . . . . . 87 vi Figure 10. ll. l2. 13. LIST OF FIGURES Single-Circuit, Open Flow Model . Constant Infusion, Indicator-Dilution Curve Obtained in the Single-Circuit, Open Flow Model . . . . . . . Bolus Injection, Indicator-Dilution Curve Obtained in the Single-Circuit, Open Flow Model . . . . . . . . . . Multiple Circuit, Open Flow Model Single-Circuit, Closed Flow Model with Recirculation . . . . . . . . Constant Infusion, Indicator-Dilution Curve Obtained in the Single-Circuit, Closed Flow Model with Recirculation . Bolus Injection, Indicator-Dilution Curve Obtained with the Single—Circuit, Closed Flow Model with Recirculation . Semilogarithmic Plot of Downslope from Bolus Injection Curve . . . . . . Multiple Circuit, Closed Flow Model with Recirculation . . . . . . . . Structural Formula for Indocyanine Green Rate of Optical Deterioration of Dilute ICG Solution . . . . . . . Structural Formula of ICG Following Deterioration in Aqueous Solution . Change in Optical Density of the Same Blood-Dye Mixture at Various With— drawal Rates . . . . . . . vii Page 23 2A 29 Figure Page 1A. Gilson Cuvette Response to Blood-Dye Con- centrations in the Range 0-20 mg/L . . 36 15. Typical Calibration Curve from the Gilson cuvette c o o o o c o o I a o 39 16. Standard Curve for Converting "Seconds of Dye" to Blood Flow (ml/min) . . . . 39 17. Change in Optical Density (Pen Deflection) of the Same Blood—Dye Mixture at Various Blood Hematocrits . . . . . AA 18. Photograph of the disassembled BloOd Mix- ing Chamber 0 o o o o o o o o o LI7 19. The Isolated, Pump-Perfused Dog Forelimb Experimental Preparation . . . . . 51 20. A Typical Dye-Concentration Plateau from 'the Isolated, Pump-Perfused Dog Fore- limb Series. . . . . . . . . . 55 21. A Plot of Actual Versus Dye-Dilution Ca1- culated Blood Flows in the Isolated, Pump-Perfused Dog Forelimb. . . . . 59 22. The Intact, Naturally Perfused Dog Fore- limb Experimental Preparation. . . . 62 23. A Typical Dye-Concentration Plateau from the Intact, Naturally Perfused Dog Forelimb Series . . . . . . . . 66 2A. A Plot of Brachial Arterial (Rotameter) Versus RISA-Dilution Blood Flows in the Intact, Naturally Perfused Dog Fore- limb . . . . . . . . . . . . 7O 25. A Plot of Dye- Versus RISA-Dilution Cal- culated Flow in the Dog at Two Cuvette Blood Withdrawal Rates . . . . . . 73 26. Experimental Preparation for Continuous- Infusion, Indicator-Dilution Blood Flow Measurement in the Human Forearm and Hand. . . . . . . . . . . 78 viii Figure 27. A Typical Dye-Concentration Plateau from the Human Forearm and Hand . . . . 28. A Graph of Systemic ICG Dye Concentrations in Man Against Time During Continuous Intravascular ICG Dye Infusion. . . 29. A Plot of Dye- Versus RISA-Dilution Calcu- lated Flow in the Human Forearm and Hand at Two Cuvette Blood Withdrawal Rates . ix Page 81 8A 89 CHAPTER I SURVEY OF LITERATURE Introduction This chapter introduces theoretical and historical information pertinent to the present study. The section on theory, presented first, is designed to orient the reader and to make the succeeding historical review more meaningful. Theory of indicator-dilution techniques has been extensively reviewed elsewhere by Stephenous, 1948; Andres g£_al., 1954; Meier g£_§1., 1954; Zierler, 1958, 1962; Kupic §£_al,, 1966; Cropp e£_al,, 1966. Relevant theoretical considerations are summarized below without reference to these individual authors. Theory of Indicator (Dye) Dilution Technique A single—circuit, open flow model with a constant fluid flow is shown in Figure 1. An indicator, con- tinuously infused at the injection site, begins to appear at the outflow after a delay known as the traver- sal time. Continuous sampling for this indicator at the outflow site would produce the curve shown in Figure 2. The shape of the upslope portion of this curve is Reservoir Sampling site b -- / IInjection site \ Figure 1.--Single-circuit, open flew model. Plateau O _ Time tion at Sampling Site Indicator Concentra- Figure 2.--Constant infusion, indicator-dilution curve obtained in the single-circuit, open flow model. primarily a function of three variables: (1) the degree of mixing of indicator with the fluid, (2) the linear velocity of the fluid, and (3) the volume of the system between the points of indicator infusion and sampling. The outflow indicator concentration eventually reaches a steady state described by the flat concentration plateau. This steady state represents an equilibrium in which the rate of indicator infusion is equal to the rate at which indicator is leaving at the outflow site. The relationship between indicator leaving at the out— flow site, indicator concentration at the outflow site, and fluid flow is given by equation 1: RATE OF INDICATOR LEAVING _ FLOW (ml/min) X CONCENTRATION OUTFLOW SITE (mg/min) ' OF INDICATOR AT OUTFLOW SITE (mg/ml) With indicator—infusion and fluid-flow rates constant, equation 1 may be re-written as follows: RATE OF INDICATOR = FLOW (ml/min) X CONCENTRATION OF INFUSION (mg/min) INDICATOR AT OUTFLOW SITE (mg/ml) Rearranging terms leads to the equation for flow: RATE OF INDICATOR INFUSION (mg/min) FLOW (ml/min) = O T TIO OF I D CATOR T OUTFLOW SITE (mg/ml) Instead of continuous indicator infusion, an indi— cator could be injected into the flowing fluid as a bolus. With reference to the model system in Figure l, a bolus injection of indicator will describe a time concentration curve at the sampling site similar to that shown in Figure 3. The character of this curve would be subject to the same variables already listed in connection with the constant infusion technique. If the amount of in- jected indicator were known, one could determine the (l) (3) tion at Sampling Site Indicator Concentra- Time Figure 3.-—Bolus injection, indicator-dilution curve obtained in the single-circuit, Open flow model. mean indicator concentration under the curve and calcu- late flow from the following relationship: AMOUNT OF INDICATOR INJECTED (mg) MEAN ONC NTRATION OF I DICATOR (4) UNDER CURVE (mg/ml) X TIME UNDER CURVE (min) . FLOW (ml/min) = The mean, time-dependent indicator—concentration obtained from a bolus injection is comparable to the steady state indicator concentration plateau observed with the con- stant infusion method. A multiple circuit, cpen flow model is shown in Figure 4. Indicator may now travel by more than one route to arrive at the outflow Site. The average time it takes all indicator particles to arrive at the out- flow site (mean transit time) may be greater than the mean transit time for a single-circuit, open flow model. Hence, in the multiple circuit, Open flow model, both Reservoir Sampling site (L A! II‘InjectiorinEEEEEEZ/P“ \\ site _ Figure 4.--Multiple circuit, Open flow model. the constant infusion and bolus injection indicator- dilution curves may differ in character from those Ob- tained with the single—circuit, open flow model. This change in character would be evident in the shape of the uplepe portion of the constant infusion curve, and, in the time interval Observed under the bolus injection curve. The relationship between mean transit time, fluid flow, and volume of the model is given in equation 5: _ FLOW (ml/min) X MEAN (5) VOLUME (m1) ' TRANSIT TIME (min) A third model, shown in Figure 5, differs from those considered above in that the fluid and indicator are allowed to recirculate. Recirculating indicator modifies the lepe of a constant infusion indicator dilution curve as illustrated in Figure 6. It may be seen that the initial portion of this curve may be II 49....Sampling site Pump .Injection '5 site .<__. ("Y") Figure 5.—-Single-circuit, closed flow model with recirculation. Plateau ampling Site Recirculation Indicator Concentra- tion at ‘2 IfiTime ' Figure 6.—-Constant infusion, indicator-dilution curve obtained in the singleacircuit, closed flaw model with recirculation (dashed line). identical’to that obtained in the open circuit models previously discussed. However, after an interval of time (a-b) in which indicator circulates once around the model, recirculating indicator is added to the indicators concentration plateau, and a constant rise in the curve is recorded. The dashed line in Figure 6 represents the actual concentration Of recirculating indicator in this example. Flow in this model may be estimated with equation 6, derived from equation 3. RATE OF INDICATOR INFUSION (mg/min) CON N A I N OF IN ICATOR AT SAMPLING SITE MINUS CONCENTRATION (6) OF RECIRCULATING INDICATOR (mg/ml) FLOW (ml/min) = The amount Of recirculating indicator can be determined by sampling the circulating fluid upstream to the indi— cator infusion site (e.g., point "Y" in Fig. 5). The interval a-b is independent of the placement of infusion and sampling sites in the single-circuit, closed flow model. However, in the multiple circuit, closed flow model to be discussed later, the interval a-b may be a function of the position of and distance separating the infusion and sampling sites. Figure 7 shows a bolus injection indicator-dilution curve Obtained in the single-circuit, closed flow model tion at Sampling Site Indicator Concentra- 0 Time Figure 7.--Bolus injection, indicator-dilution curve obtained with the single-circuit, closed flow model with recirculation. (Fig. 5), and illustrates the effect produCed by re- circulating indicator. Provided the volume Of the flow model is small, the downSlOpe of the bolus injection indicator-dilution curve does not return to the baseline. Instead, a secondary rise occurs concurrent with the appearance of indicator recirculation. Hamilton (1932) showed that the initial portion of the dilution curve downlepe becomes a straight line when replotted on semilogarithmic coordinates (Fig. 8). Assuming that the Log Indicator Concentration / Time Figure 8.--Semilogarithmic plot of downslope from bolus injection curve. initial logarithmic relationship between time and con- centration holds for the final portion of the downlepe as well, this technique disposes of the recirculating effect. Flow is then calculated by equation 4. Finally, a multiple.circuit, closed flow model is presented in Figure 9. This model approximates the‘ cardiovascular system of mammals. Indicator infusion Té—‘Sample 6-— Inject Pump ("X") ("Y") M T <——Inject . <——— Sample Figure 9.?-Multiple circuit, closed flow model with recirculation. and sampling sites may be variously located to measure either pump flow ("X")or flow through a single-circuit of the model ("Y"). The constant infusion and bolus injection indicator-dilution curves are similar to those previously shown in Figures 6 and 7, respectively. However, in this multiple circuit, closed flow system, the infusion and sampling Sites must be carefully chosen to minimize the recirculation effect. For example, the measurement of flow at Site("X")would be difficult be- cause of the quick circulation of indicator through the adjacent shunt. This shunt would cause recirculating indicator to appear earlier at the sampling site. On the other hand, the measurement of flow in a single— circuit ("Y") would be less affected by recirculating indicator because indicator is not shunted in this in- stance and must then travel the entire circuit before reappearing. 10 There are several basic assumptions underlying the use of the indicator-dilution technique. In the case of a non-diffusible indicator: (1) it is assumed that the distribution of transit times for indicator particles is the same as that for blood flowing through the given system; (2) it is assumed that complete mixing of indi— cator with blood occurs upstream to the sampling site. Complete mixing only occurs in the presence of turbulent flow. In the measurement of cardiac output, indicator is infused at or near the right heart and mixing of indicator and blood is caused by turbulent blood flow within the ventricles. Regional arterial blood flow, on the other hand, is normally laminar, not turbulent, flow. Thus, extraordinary measures must be used to produce mixing of indicator with the arterial blood in order to measure regional blood flow accurately; (3) it is assumed that infused indicator remains in the blood between the injection and sampling sites (non-diffusible indicator); (4) it is assumed that total flow through the system remains constant during the period between injection and sampling of indicator; (5) it is assumed that the volume of the system does not Change during the experimental period; (6) it is assumed that the distribution of transit times of indicator particles does not change during the measurement of flow; and (7) it is assumed that recircu— lation either does not occur or that it can be measured. 11 Blood flow in the mammal is pulsatile rather than constant. Cropp and Burton (1966) have studied the effect of varying or pulsatile flow on estimates of blood flow made by constant-infusion, indicator—dilution methods. They have shown that a . . . sampling site some distance down a flow stream may be exposed to a low indicator concen- tration (generated by systole) for a much longer time than the actual duration of systole. This is so because the segment of low concentration may move past the sampling site at low diastolic velocities. As,a result, the indicator detector at the distant sampling site may report a low indicator concentration for too long a time and a high indicator concentration for too short a time. This results in an over-estimate of the total flow by indicator-dilution techniques. These authors have called this effect the "distance distortion error." This type of error would occur if the sampling site were located in the same vessel as the injection site, i.e., downstream in the same artery during measurement of regional blood flow. According to CrOpp and Burton (1966), the distance distortion error is largely removed and flow is measured with great accuracy by the constant infusion indicator- dilution method where . . flow is pulsatile but of constant stroke volume and rate with a mixing region between in- fusion and sampling points. This mixing region acts like a low pass filter—~attenuating the high frequency oscillations in the concentration time curve produced by the heart without attenuating significantly those oscillations produced by gradual changes in flow. Use of such a mixing region precludes the possibility of obtaining information about instantaneous flow by indi- cator-dilution methods. 12 Thus, if the capillary bed acts as a satisfactory mixing region, sampling downstream to the capillaries of a vascular bed during continuous indicator infusion into the artery supplying the bed would probably eliminate the distance distortion error, but would only allow measurement of mean, not instantaneous, regional blood flow. Historical Development of Indicator— Dilution Methods In 1829, Hering reported a technique for the measurement of blood circulation time. He injected potassium ferricyanide into one jugular vein and noted the elapsed time before this indicator could first be detected in the contralateral vein. Sixty years later, as an improvement on Hering's technique, Stewart (1897) detected an intravenously infused salt solution by the change in electrical conductivity of arterial blood. Quantitative determination of the dilution of the salt comprised the original Stewart measure of cardiac output: Cardiac Output _ Rate of salt infusion (ml/min) (ml/min) ' Concentration of salt at sampling site (ml/ml) Henriques (1913) used injections of thiocyanate as his indicator. Thiocyanate was easily measured chemically, but Henriques found that it remained in the blood sufficiently long to produce a recirculation effect. l3 Henriques demonstrated that the indicator did not travel in blood as a square wave, as envisioned by Stewart, but rather in a dispersed form. This dispersion was re- flected at the sampling site by a slow rise and fall in the indicator concentration. Hamilton and his colleagues developed the bolus injection indicator-dilution technique with a classical series Of papers beginning in 1928 (1928a, 1928b, 1930, 1932). Their problem was twofold: (l) a solution of the curvilinear response to a bolus indicator injection, and (2) the separation of recirculating indicator from the primary dilution curve. In model systems simulating the. mammalian circulation, in heart-lung preparations, and in intact animals, they were able to show that an-intra— venous bolus injection of indicator (Vital Red or Evans Blue dyes), and semilogarithmic extrapolation of’the disappearance curve in arterial blood provides a measure of cardiac output by the following formula: Cardiac Output = £2; at I is the amount of indicator injected, 5 the average con- centration of indicator under the corrected curve (sum- mation of concentration values at one-second intervals divided by the duration of the curve in seconds), and t the duration of the dilution curve. 14 The multiplicity of present day techniques are an outgrowth of this early work. They have been reviewed elsewhere (Cournand, 1945; Hamilton, 1953; Dow, 1956; Schlenk, 1966). DeveIOpment of the Constant Infusion Indicator-Dilution Technique Holt (1944) reported the use of a constant infusion indicator-dilution method for estimating cardiac output in anesthetized dogs. He infused Evans Blue dye at a constant rate into the right atrium for 13 seconds and sampled femoral artery blood 8-23 seconds after dye in- fusion was begun. Reasonable indicator concentration plateaus were obtained at the sampling site, and re- circulating dye appeared no earlier than the 16th second. Hamilton and Remington (1947) reviewed the original work of Stewart and suggested that his constant infusion method was valid so long as recirculation of indicator did not occur. In a series of experiments with a model and in the dog, Hamilton and Remington showed that the slowest moving infused indicator particles appeared at the sampling site chosen by Hamilton and Remington much later than the fastest moving particles, and well after the appearance of recirculating indicator. Hence, the recirculation effect distorted the critical portions of the constant infusion, indicator-concentration plateau as obtained by these workers. They concluded that the l5 constant infusion technique could not accurately measure blood flow due to the presence of indicator recirculation. Howard, Hamilton and Dow (1953) studied indicator dilution curves made simultaneously by the bolus in— jection and the constant infusion techniques in a model and in the dog. Once again, their conclusion was that both techniques were valid only in the absence of re- circulation. However, in contrast, Shepherd, Bowers and Wood (1954, 1955) simultaneously determined cardiac out- put by three methods: direct Fick, bolus injection, and constant infusion indicator-dilution techniques. All three techniques gave the same cardiac output estimate. In 20 of 22 trials with the constant infusion technique, there was a 2-7 second indicator concentration plateau prior to the onset of indicator recirculation. The work of Shepherd et_al, (1954, 1955) led to the eventual acceptance of the constant infusion indicator- dilution technique. In addition, these authors demon- strated that the achievement of a constant infusion plateau, prior to recirculation, was dependent on the location of the infusion and sampling sites. Similarly, Peterson et_al. (1954) infused indicator at a constant rate near the root of the aorta and sampled downstream aortic blood. They obtained 7—10 second indicator con- centration plateaus prior to onset of recirculation and reported values in close agreement with direct Fick measurements. l6 Constant Infusion Indicator-Dilution and Forearm Blood Flow The successful development of a continuous-infusion indicator-dilution technique for measurement of cardiac output soon yielded new applications. This technique was first used to measure regional blood flow by Andres gt_§1. (1954). They infused Evans Blue dye into brachial artery and sampled the effluent blood from a deep and a supere ficial forearm vein, calculating forearm blood flow by the following equations: Rate of indicator infusion (mg/min) Forearm Plasma Flow = Mean concentration of dye in venous (L/min) plasma minus concentration of indi- cator in recirculating plasma (mg/L) Plasma Flow 1 - Hematocrit Forearm Blood Flow If complete mixing of infused dye and forearm blood had occurred, effluent blood in all forearm veins would have contained equal concentrations of the dye. However, Andres gt_§1. (1954), found that indicator concentrations in their forearm veins differed significantly. These authors suggested several explanations for their venous indicator concentration differences. Un- dyed blood from collateral arteries about the elbow might have disprOportionately diluted the brachial arterial blood-dye mixture downstream in the forearm vascular tree. This explanation was discarded because the authors 17 felt that total collateral arterial blood flow in the forearm was too low to explain the differences in indie cator concentration.. A second possible explanation might be an unequal escape Of indicator between the infusion and the sampling sites. However, Evans Blue dye was. shown to be rapidly bound by serum albumin which is essentially non—diffusible. High bifurcation of the brachial artery might have led to inadvertent dying Of. radial Or ulnar arterial blood exclusively, causing significant differences in venous indicator concentrations. It had been shown previously (Quain, 1854) that the inci- dence of anomalous brachial artery bifurcation in the human was 20 percent. Thus, 20 percent of the cases Of inadequate mixing could be explained on this basis. Finally, incomplete arterial mixing of indicator might have occurred, resulting in an unequal distribution of the indicator to downstream vascular beds. This was a real possibility, and these authors attempted_to improve intra-arterial indicator mixing by creating arterial turbulence with jet-injection of indicator. In 63 forearm experiments on 27 subjects, Evans Blue dye was infused at a constant rate into brachial arterial blood while two forearm veins were sampled simultaneously. Dye was infused via either a standard arterial needle or a "jet injector" needle. Despite theoretical indications and favorable data from model. 18 and dog studies, the jet injection did not improve indi- cator mixing in the forearm. In addition, the jet in- jector produced erythrocyte hemolysis and limb vasodil- ation. In a final group of 10 subjects, resting forearm blood flow was repeatedly measured during dye infusion through a standard needle. Statistical analysis of the data allowed the authors to conclude that sufficient mixing of indicator upstream to the sampling site (mainly in the capillary-venous portion of the vascular bed) was obtained in four-fifths of their subjects to permit the accurate measurement of forearm blood flow. Wahren (1965, 1966) also used the continuous in- fusion indicator-dilution technique to measure forearm blood flow. He infused indocyanine green dye (0.6-1.5 mg/min) at a high volume rate (34 ml/min) into the brachial artery for one-minute intervals, relying on mixing of dye and blood by a volume effect. Blood was sampled discontinuously in the radial artery and a deep vein. The degree of mixing of dye with blood was deter- mined by comparing dye concentrations in the radial artery with those from the deep vein. The recirculation dye level was interpolated from dye concentrations ob— tained immediately before and 15 seconds after the dye infusion period. 19 There was a large difference between radial artery and deep vein dye concentrations in 17 percent of the forearm flow determinations. These cases were attributed to undetected, high bifurcation of the brachial artery with consequent dying of radial or ulnar arterial blood exclusively. Equal dye concentrations at the two sampl- ing sites were found in 83 percent of the experiments, suggesting satisfactory mixing. In addition, simultaneous flow estimates by venous occlusion plethysmography agreed closely with flow estimates obtained by the indicator- dilution technique. Wahren (1966) reported that the high volume rate of indicator infusion apparently did not have an effect on pre-infusion flow patterns in the forearm. In pilot studies, deep and superficial vein oxygen saturation re- mained unchanged after infusion of dye began. Further- more, the decrease in venous hemoglobin concentrations, following onset of indicator infusion, supported the conclusion that infused dye was simply added to the initial forearm blood flow. Overbeck (1966) measured forearm and hand blood flow with the continuous infusion indicator-dilution 131 labeled human serum albumin (RISA) technique, using I as his indicator. He developed and used an arterial jet injector needle which distributed indicator laterally into the stream of arterial blood. The kinetic energy 20 of infusate was 4500 g. cm2 sec-2, substantially below the level reported by Andres to cause hemolysis. By comparing the effectiveness of this jet needle against standard needles in a model, in the isolated dog fore- limb,and in the human forearm, Overbeck was able to show statistically significant improvement in indicator mixing with his jet injector system. All of the techniques described above for measuring forearm and hand blood flow provided discontinuous, not continuous,measurements of blood flow. In addition, flow measurements were available only after completion of the experimental procedure, because blood had to be processed before indicator concentrations could be determined. Indocyanine Green: Properties Evans Blue dye (T-l824) and reduced hemoglobin both absorb light maximally at 640 millimicrons. Hence, a change in blood reduced hemoglobin concentration will interfere with the measurement of this dye by the in: give, whole blood cuvette densitometry technique. Fox gt_al, (1957), in conjunction with Heseltine and Brooker, reported the use of indocyanine green, a dye which maximally absorbs light in the infrared region (800 millimicrons). At this wavelength, light trans- mission by oxyhemoglobin and reduced hemoglobin is 21 identical. Indocyanine green dye (ICG) is marketed under the trade name Cardio-Green (Hynson, Westcott and Dunning, Baltimore, Maryland). The structural formula of ICG is given in Figure 10. It has a molecular weight of 775. Its light ab- sorption in plasma closely follows Beers' law; there is a linear increase in Optical density with increasing con- centrations Of homogeneous solutions of the dye., How- ever, in non-homogeneous solutions (i.e., blood), ICG (CH3)2 (Cili3)2 C ‘\C(=CH-CH) = CH-C’/// / 3 I i cuz CH2 I cuz CH2 cuz CF2 803Na SO3 Figure 10.--Structura1 formula for indocyanine green (Fox §£_gl., 1957). absorption does not closely follow Beers' law. There— fore, devices for detecting ICG in whole blood (densito- meters) usually contain a linearizing circuit. The effect of plasma protein binding, pH, tempera— ture and Chemical additives on ICG Optical activity have 22 been studied. Fox et_§l. (1960) reported a change in peak absorption of ICG (775 to 800 mu) as an aqueous solution of this dye is added to blood. Bassingthworghte gt_§l, (1960) were able to show that this change in ab- sorption is due to binding of the dye to a plasma pro- tein (primarily albumin). Stabilization of plasma ICG absorption at 800 mu occurs very rapidly. Because the dye is prepared initially in an aqueous solution, the effect of pH on Optical activity was studied. Peak absorption of the dye remained unchanged over a range of pH 6.0 to 9.0. Varying the temperature of an aqueous dye solution also failed to change ICG absorption characteristics. ICG itself failed to effect the pH of triple distilled water (pH5.5) in concentrations up to one milligram per milliliter. ICG is highly soluble in distilled water (Fox gt_§l., 1960), but poorly soluble in saline solutions. Once ICG is dissolved in distilled water, however, NaCl may be added up to 9 g percent without causing precipitation of ICG. Optical density of ICG decreases with increasing salinity, but this effect disappears on addition of dye and saline to plasma. Varying the pH of plasma (7.0 to 7.8), the plasma protein concentrations (10 to 100 percent of normal) and the plasma NaCl level by the addition of .3 to 1.5 g percent all fail to influence the Optical density of ICC in plasma. 23 While ICG was found to be stable in blood and plasma for up to eight days, Fox gt_§l. (1960) reported varying instability of ICG in aqueous solutions. They found dilute aqueous solutions of ICC increased in optical density initially, thereafter decreasing in activity. More concentrated solutions (1 mg/ml) ex- hibited lesser changes. Barbier and DeWeerdt (1964) reported a high rate of Optical deterioration of dilute. ICG solutions (2.5 mg percent). Their findings are illustrated in Figure 11. They also verified earlier 0.5004 0.475. 0.450. Optical Density 0.425‘ Y r . . 100 200 300 400 min Time Figure ll.-—Rate of Optical deterioration of dilute ICG solution (Barbier and DeWeerdt, 1964). reports that ICG is stable in plasma. Chromatographic work by these workers showed that the optical deterio- ration of ICG in aqueous solution is due to the formation 24 of a ring structure at the SO3- group, with concommitant 3 function and appearance of a C=0 character in infrared saturation of the unstable C=N group, loss of the SO spectra. The probable formula for the new compound is given in Figure 12. I . T2 O = IS = 0 CH2 CH2\ Cg 2 Figure 12.--Structural formula of ICG following deterioration in aqueous solution. Cobb (1965) has recently shown that various re- ducing agents markedly affect the spectrophotometric characteristics of ICG in both plasma and aqueous solu- tions. For example, heparin preparations containing sodium bisulfite lowered the optical density of ICG (1 mg/L) by 72 percent. Sodium metabisulfite, sodium borohydride and potassium borohydride also markedly alter the optical density of ICG at 800 mu. Other heparin preparations with different preservatives have no apparent effect on ICG optical activity. Finally, ICG has been found to be non-toxic in both acute and chronic toxicity studies in animals. No 25 untoward effects attributable to the use of this dye in man have been reported at blood concentrations as high as 5 mg/kg (Fox et a1., 1960). Indocyanine Green and Hepatic Removal Since the observation by Fox §t_al, (1957) that ICG is excreted into the bile, there has been considerable interest regarding the role of the liver in removing this dye from the blood. Table 1 presents data on hepatic removal of ICG when given in single dose injections. Hunton et_al. (1960) found that an increase in the load- ing dose of ICG decreased the plasma disappearance rate TABLE l.--Hepatic removal of injected indocyanine green dye in dogs and humans. Worker Species ICG Injected Disapfizigance Hunton (1960) dog 1 mg/kg average 7.6%/min " 9 ms/ks " 3.8%/min Ketterer (1960) " 0.5 mg/kg " 5.4%/min Wheeler (1958) " 1 mg/kg " 9.6%/min Hunton (1960) human .25 mg/kg " 25.9%/min Reemtsma (1960) " ? " 23.2%/min Cherrick (1960) " .5 mg/kg " 18.5%/min of the dye, but noted that the total quantity of dye be- ing removed per unit time was increased. It was noted 26 by others that the disappearance rate varied substanti— ally among subjects with apparently normal liver function. Hepatic removal appears to be the sole route for clearance of ICC from the blood. It has been shown that arterial concentrations do not differ from venous con- centrations of ICC across limb vascular beds (see Table 2), and excretion of ICG by the kidneys does not occur (Cherrick et_al,, 1960). Various workers (Wheeler et_al., 1958; Hunton gt_al,, 1960; Ketterer et_al., 1960) have reported recovering 92 to 97.5 percent of an injected ICG dose from the bile. There is little, if any, recycling of bile-excreted ICG via intestinal absorption (Wheeler gt_a_l_., 1958). Hepatic ICG extraction is a function of blood flow through the liver, and the ICG concentration in this per- fusing blood. As long as the extraction mechanism is not saturated and hepatic blood flow remains relatively constant, the rate of hepatic ICG extraction becomes a function of the systemic blood ICG concentration. An increase in systemic ICG concentration leads to an in- crease in the hepatic ICG extraction rate (Hunton et_a1., 1960). Thus, with continuous ICG infusion (i.e., with the continuous-infusion indicator-dilution technique), one Should obtain a steady state systemic ICG concen- tration (recirculating ICG) during which ICG infusion equals hepatic ICG extraction. Cherrick et a1. (1960) 27 primed one patient with 10 mg of ICG and followed this with a constant ICG infusion (0.5 mg/min) for 50 minutes. His plasma ICG levels during the infusion period are presented in Table 2. TABLE 2.--Arterial and venous ICG levels during constant ICG infusion (Cherrick et a1., 1960). Following Start of Brachial ' Antecubitol ICG Infusion Arterial ICG Venous ICG minutes mg/L mg/L 5 1.68 1.66 10 1.64 1.64 15 1.54 1.56 20 1.47 1.43 25 1.38 1.42 30 1.28 1.26 35 1.33 1.37 40 1.36 1.34 45 1.40 1.39 50 1.52 1.54 Cherrick et_al, (1960), concluded that there was no detectable arterial-venous ICG concentration differ— ence, and "detectable peripheral tissue uptake of ICC did not occur." It is apparent, in addition, that the concentration of circulating ICG remained quite con- stant during this infusion period. Cuvette Densitometry One of the earliest cuvette densitometer systems was reported by Friedlich, Heimbeck and Bing (1950). Their system consisted of a flow-through blood optical 28 chamber interposed between a filtered light source and a multiplier phototube. Following indicator injection, arterial blood was withdrawn from the subject through the cuvette, and indicator concentration changes were directly and continuously recorded. Dow (1956) has reviewed the history and use of cuvette densitometers elsewhere. The introduction of cuvette densitometry was of immediate value in the measurement of cardiac output by the dye-dilution technique. While cuvette blood withdrawal rates employed in this technique were relatively high (20 to 40 ml/min); inscription of the arterial dye—dilution curve took less than a minute and blood loss to the subject was minimal. The literature, on the other hand, contains little infor- mation about the use of cuvettes at relatively low blood withdrawal rates. Fox gt_§1. (1957), noted a relation- ship between lower cuvette blood withdrawal rates in the range 9 to 26 ml/min and decreasing cuvette frequency response. Sekelj et_al. (1967) reported the design of a cuvette insensitive to blood flow in the range 10 to 40 ml/min. Below this range, however, there was a sharp decrease in recorded Optical density with flow as shown in Figure 13. .l I'\) [Optical Density Withdrawal Rate ml/min 10 [:1 1 L J ' 54M: :3 a) N) r~ N 2100\\o ox . ~ - . . . . Ln :1- OO mtn: r~ O\ H N in Figure l3.-—Change in optical density of the same blood-dye mixture at various withdrawal rates (Sekelj et a1., 1967). CHAPTER II INSTRUMENTATION Introduction The Gilson cuvette densitometer (model DTL, Gilson Medical Electronics, Middleton, Wisconsin) was used in the present study. The performance of this instrument was evaluated in terms of the four characteristics Sekelj gt_al, (1967), considered optimum in the operation of a cuvette densitometer system. These are: (l) the instrument should be "stable and free of drift," (2) it should be "relatively insensitive to changes in the rate of blood withdrawal through the cuvette," (3) it Should be ”linear in response to changes in optical density over the range of dye concentrations employed," and (4) it Should be "insensitive to changes in blood hematocrit." The following four sections treat each of these criteria individually. Stability of the Gilson Cuvette Densitometer It was apparent early in the present study that the cuvette densitometer was electronically unstable, parti- cularly at low cuvette blood withdrawal rates (see 30 31 following section). For example, cuvette output drifted excessively during the withdrawal of undyed blood through the cuvette, producing an unstable baseline and hampering the accurate measurement of dye-concentration plateaus. Sensitivity of the cuvette also changed with time as evidenced by the difference between pre— and post—experi- ment cuvette calibration curves; this change in sensi- tivity introduced error into calculated blood flow measurements. Finally, the optical chamber nylon inter- facing was found to adsorb dye from perfusing blood—dye mixtures, resulting in an apparent increase in recorded Optical densities with time. The manufacturer of this instrument was notified of these findings and a modification program begun. In the densitometer, along with basic circuitry design changes, temperature sensitive components were carefully Shielded from heat-producing elements. In the cuvette, the filtered light source was moved away from the heat- sensitive, light—dependent resistor (e.g., photocell). The optical Chamber nylon interfacing was replaced with a non-adsorbing Teflon material. Because of the continuing modification of the cuvette densitometer during this study, quantitative measurements of instrumental stability were not deemed apprOpriate. A qualitative estimate of sensitivity stability may, however, be made for the system in its 32 present state by comparing the pre- and post—experiment cuvette calibration curves obtained in later experiments. In experiments 107, 108, and 109 (Chapter IV), there was a mean cuvette sensitivity shift of -2.2 percent (range -9 percent to +12 percent) during the four-hour interval between cuvette calibrations. This can be compared with a mean sensitivity shift of +6 percent per four—hour interval observed in the earlier experi- ments. Further improvement in instrumental stability is expected from the continuing modification program. Sensitivity of the Gilson Cuvette to Changes in Blood Withdrawal Rates According to the manufacturer, the Gilson cuvette is relatively insensitive to changes in blood with- drawal rates in the range 20 to 40 ml/min. Withdrawal rates within this range are suitable for discontinuous indicator-dilution measurement of cardiac output, as the inscription of the arterial dye—dilution curve takes less than a minute and blood loss to the subject is thus minimal. However, the present study required con- tinuous cuvette sampling of venous dye concentrations over periods of up to 120 minutes. This could be ’ accomplished only by using a substantially lower cuvette blood withdrawal rate, both because of limited venous blood flow in the vessel being sampled, and the necessity for minimizing blood loss in the subject. 33 At low blood withdrawal rates the Gilson cuvette displayed a sensitivity to rate of blood flow through the cuvette similar to that described by Sekelj §t_a1, (1967) (see Chapter I, Figure 13). During the sampling of blood containing 12 mg/L dye concentration, a change in blood withdrawal rate from 3.8 to 1.5 ml/min caused a 3 percent decrease in recorded optical density. A change in withdrawal rate from 1.5 to .76 ml/min produced a 10 percent drop in recorded Optical density. The mechanism of this flow dependency is not clear. It may be related to changes in erythrocyte orientation within the cuvette Optical chamber at low blood withdrawal rates. These changes in orientation could affect the apparent optical density of the perfusing blood. Cuvette flow dependency could also be due to variation in axial streaming at the lower withdrawal rates. In any case, with the use of this cuvette at withdrawal rates which make the cuvette flow—dependent, care must be exercised to maintain a constant blood withdrawal rate; variations in the recorded dye—concentration plateau might be an artifact induced by a change in the cuvette blood with— drawal rate. A cuvette blood withdrawal rate of 1.5 ml/min was used initially in the present study. The withdrawal rate was increased to 2.4 ml/min in later experiments in an effort to bring dye—dilution flow measurements 34 into agreement with flow measurements made by an inde- pendent technique. These studies are presented in subse- quent chapters. Response of the Gilson Cuvette Densitometer to Changes in Optical Densigy Cuvette Linearity Eleven experiments were conducted in which aliquots of outdated, citrated human blood were dyed to give known blood—dye mixtures over the range 0 to 20 mg/L. These mixtures were withdrawn (1.5 or 2.4 ml/min) through the cuvette and the response recorded. Figure 14 is a plot of the cuvette response from one of these experiments. The dotted line is an extension of the first four plotted points drawn to illustrate the shift in linearity. This Shift in linearity occurred in all eleven experiments, and is apparently a characteristic of the instrumentation. The linearity shifted at a dye concentration between 8 and 10 mg/L. The presence of a non—linear cuvette out- put requires a more rigorous calibration technique. Cuvette Calibration Fresh ICG dye was prepared for each experiment. A standard technique for calibrating the cuvette densito— meter was used for all in ELIE blood flow experiments. A 70 ml sample of well—heparinized blood was obtained; 30 ml were set aside as a blank—blood sample. The 35 Figure l4.--Gilson cuvette response to blood-dye concentrations in the range 0-20 mg/L. 36 l I I I I I o o o o o o g: <_> an to e- N (WW) SNOIIOS'HBO N3d OAHBS ISO- I404 DYE in BLOOD (mg/L) 37 remaining blood was divided into four 10 ml aliquots. Dye was pumped ("infused") into these aliquots for 3, 6, l6, and 24 seconds respectively. The rate of dye infusion was the same as that used during subsequent brachial arterial dye infusion (.42 mg/min). The cali— bration aliquots were then drawn through the cuvette in sequence. The recorder response in millimeters of pen deflection was plotted on the abscissa. On the ordinate, concentration of dye was plotted in terms of "seconds of dye" according to the method used above in making the calibration aliquots. A typical cuvette-recorder cali- bration curve is shown in Figure 15. The non-linear cuvette response is quantitatively similar to that de— scribed in the previous section. A second standard curve was used to convert the seconds of dye value to actual blood flow (ml/min). Since 3 seconds of dye in 10 m1 of blood is equivalent to 60 seconds of dye in 200 m1 of blood, a 3 second dye concentration represented the concentration which would be present in blood flowing at 200 ml/min. In a similar manner 6, 12, and 24 seconds of dye represented 100, 50, and 25 ml/min blood flow, respectively. These points were found to fall on.a straight line when plotted on 2-cycle log—log paper with blood flow on the horizontal axis and seconds of dye on the vertical axis, as shown in Figure 16. 38 .AQHE\HEV soad pooan on :omp mo nocooom: wsflpao>soo pom o>a50 pamocmpmll.oa oaswflm .mppo>so comaflo can Soap o>aso COHOBAQHHOO ascedzeln.ma thmflm 39 con — CON r E_E\ :5 304... 000.5 00. oo 00 CV p — b _ ON _ O. 3A0 :10 SONOOBS ON. a 00— a AEEV zoFow-Ewo 2mm 00 00 0? 4 ON m. o. h. a. 3A0 30 SONOOB S 40 Calibration Error Two experiments were performed to determine the amount of error associated with the cuvette calibration procedure. Both citrated human and heparinized dog blood were used. In each experiment, a carefully pre- pared series of 50 and 5 ml blood aliquots were infused with dye (.42 mg/min) for 120 and 12 seconds respectively. The blood-dye samples were well mixed, and withdrawn through the cuvette in sequence. A mean response was determined for the various samples in each group, and the percent deviation from this mean was averaged to give mean percent error. Data from these studies are presented in Table 3. TABLE 3.--Data showing the relationship between size of calibration blood aliquot and their reproductability. Aliquot Size 50 ml blood 5 ml blood + 120 sec dye + 12 sec dye Human 2.3 mean % error 3.2 mean % error blood of 6 samples of 6 samples Dog 3.4 mean % error 4.6 mean % error blood of 7 samples of 7 samples It is apparent that there was a small reduction in error associated with the use of larger blood-dye calibration aliquots. Since the use of smaller blood calibration aliquots in man helped reduce overall experimental blood loss, a compromise calibration blood 41 aliquot size of 10 ml was selected. It is apparent that there was a slightly higher mean percent error associ— ated with calibrations using dog blood. This increased error may be related to the generally higher erythrocyte sedimentation rate of dog blood; at the low cuvette blood withdrawal rates used in this study, changes in the homogeneity of blood may be more readily detected by the cuvette. Blood Temperature Two in XIKQ experiments were conducted to determine whether or not variations in blood temperature were critical during the cuvette calibration procedure. Un- dyed blood was obtained from a dog and added to two pre- warmed beakers (37° C.), one containing dye. As soon as possible, these two blood samples were withdrawn in turn through the cuvette, and the responses recorded. Both the undyed and dyed blood aliquots were again withdrawn through the cuvette some time later, after they had reached room temperature (240 C.). The above procedure was repeated in each of two dogs; there was no apparent difference in cuvette response to the blood samples read immediately (approximately 330 C.) and the same samples read at room temperature. It was concluded that cali- bration of the cuvette could be carried out with blood at room temperatures. 42 Sensitivity of the Gilson Cuvette Densitometer to Changes in Blood Hematocrit Sixteen in 13333 experiments were conducted to determine the effect of changing blood hematocrit on cuvette sensitivity. Multiple aliquots of outdated, human citrated blood were carefully diluted with either human plasma, dextran, or isotonic saline, within the hematocrit range 50 percent to 30 percent. Dye was added to each aliquot (17 mg/L) and the various aliquots then withdrawn through the cuvette. The response of the cuvette to varying blood hematocrits is shown in Figure 17: The same data are presented in a somewhat different form in Table 4. TABLE 4.--Mean percent change in Optical density (pen deflection) per 1 percent dedrease in blood hematocrit. Range of Hematocrit Changes 50-45 45-40 40-35 35—30 Saline -.l2 -.19 -.37 -.68 Dextran- -.54 -.57 -.59 -.61 Plasma -.61 -.62 -.65 -.67 —v —1 It is apparent that saline dilution of blood causes a smaller decrease in Optical density over the hematocrit range 50 percent to 35 percent than do dextran and plasma dilutions. The decrease in blood Optical density with decrease in hematocrit becomes greater at the lower 43 Figure 17.--Change in Optical density (pen deflection) of the same blood—dye mixture at various blood hematocrits. SERVO PEN DEFLECTION (mm) l05‘ IOO- 95‘ 44 -t~ " SALINE DILUTIONS 85 I05“ IOO‘ 95 . 85 DEXTRAN DILUTIONS ." l05 ‘ IOO 4 a“ PLASMA DILUTIONS . . 56 4'5 43 3'5 3'0 25 BLOOD HEMATOCRIT (Vol 45 hematocrits, and is quite similar in magnitude for all three diluents over the hematocrit range 35 percent to 30 percent. Sekelj 33431. (1967) found, with their cuvette system, a slight linear increase in optical density as the hematocrit increased from 24 percent to 59 percent. Since they used their own cuvette and did not explain their technique, it is not possible to compare their results with the present findings. Apparently this hematocrit phenomenon is independent of the instrumen- tation used. Design and Use of a Blood Mixing Chamber The studies to be described in Chapters IV and V required an arterial blood mixing chamber to insure adequate mixing of arterially infused dye with the blood. A mixing chamber was designed and built to serve this purpose. The chamber consisted of a leucite plastic block with a cylindrical chamber into which was placed a Teflon coated magnetic stirring bar. Chrome fittings conducted blood into and out of the stirring chamber. Figure 18 is a photograph of the disassembled mixing chamber. In Operation, the mixing chamber was inserted in the arterial cannula upstream to the brachial artery. The chamber was placed above a magnetic stirrer. Dye was infused upstream to the Chamber which insured the complete mixing of dye with blood prior to entry into the brachial artery. 46 Figure l8.-—Photograph of the disassembled blood mixing chamber. 48 £2.2i12 tests were conducted to determine the hemolyzing effect, if any, of the mixing chamber on perfusing blood. The mixing chamber was interposed upstream to a pump in the pump-perfused forelimb Of a dog and pump perfusion pressure was recorded. There was no change in pump perfusion pressure whether the mixing chamber was turned on or off. Pump perfusion flows were varied over the range 25 to 200 ml/min. The limb vascular bed was responsive, however, to pro— ducts of hemolysis as demonstrated by a Sharp drop in pump perfusion pressure following the intra-brachial arterial injection of whole blood at a velocity which produced hemolysis of the injected blood. CHAPTER III CONTINUOUS DYE-DILUTION FLOW MEASUREMENTS IN THE ISOLATED, PUMP-PERFUSED DOG FORELIMB Purpose of Investigation This section of the investigation was designed to provide a direct comparison of dye-dilution calculated blood flow with actual blood flow. The isolated, pump- perfused dog forelimb was a suitable model for these studies in that both the duration of steady-state flow and the range of flow rates in the forelimb could be readily controlled and measured. By comparing dye- dilution calculated and actual flow measurements an estimate of the accuracy of the dye-dilution system could be made. Methods Figure 19 is a diagram of the experimental prepar- ation. Eight mongrel dogs (18-27 kg) were anesthetized with sodium pentobarbital (30 mg/kg), intubated and provided with external respiration. Except for the brachial artery and brachial and cephalic veins, the right forelimb was totally severed from the body at a 49 50 Figure l9.—-The isolated, pump—perfused dog forelimb experimental preparation. 51 m_o>mmmmmw. maozm> Inimofion. Aflmwr .. . . . O>mwm I > o a mmemzoemzuo 4<>>> \ PW...‘ mine/.8 n=22 NJI r 62¢.de ( _U mum—2410 “ms—Dd ms @222 Demure. 52 point midway between the Shoulder and elbow. Tourniquets were applied to the tissues of both stumps. The dog was systemically heparinized (10,000 U.S.P. units). The femoral artery was cannulated and the cannula connected to the blood mixing chamber (see Chapter II). A pressure- independent (0-300 mm Hg) Sigmamotor (Middleport, N. Y.) pulsatile blood pump was interposed between the mixing chamber and brachial artery so that the forelimb arterial inflow was solely supplied by pump. Pump flow was varied over the range 25-225 ml/min to simulate the expected range of blood flows in the resting human upper extremity. Total venous outflow from the severed limb drained separately from the cephalic and brachial veins into a single reservoir. Polyethylene tubing connected limb venous catheters to the dye cuvette, and, the dye cuvette to a Harvard (Model 600-910, Harvard Apparatus Co., Dover, Mass.) withdrawal pump. The cuvette response to venous blood and blood-dye mixtures was recorded on a Gilson servo polygraph. Venous reservoir blood was re- turned to the femoral vein by a second Sigmamotor pump. The Cuvette was calibrated before and after each experiment, and a mean of the two calibrations was used for calculating dye-dilution blood flow. The calibration procedure followed the protocol outlined in Chapter II. Cardio-Green dye was freshly prepared for each experiment. The dye was diluted with distilled water and isotonic saline to a concentration of 0.55 mg/ml. 53 The measurement of forelimb flow by the continuous— infusion dye-dilution technique was conducted in the following manner (refer to Figure 20). The arterial blood pump was set at a desired flow rate. Venous blood was withdrawn through the cuvette (1.5 ml/min) until a stable baseline response was observed on the polygraph recording. Arterial dye infusion was then begun (0.42 mg/min, 0.76 ml/min) upstream to the mixing chamber, and blood from the two venous cannulae was withdrawn in sequence through the cuvette until a dye-concentration plateau was recorded for each vein. The dye infusion was then stopped and limb venous dye concentration was allowed to return toward the baseline. Limb blood flow was simultaneously measured by timed collection of total venous outflow in graduated cylinders. The slope of the dye-concentration plateau(s) was taken as an indication of dye recirculation (see Figure 6, Chapter I). The baseline (dashed line) was drawn parallel to the plateau (Figure 20), in order to correct for this recirculation (see equation 6, Chapter I). The height of the dye-dilution plateau above the corrected baseline was used to calculate blood flow. Results A total of 69 paired blood flow measurements were made in eight dogs. The results are presented in Table 5. Overall, dye—dilution calculated forelimb flow 54 Figure 20.--A typical dye-concentration plateau from the isolated, pump-perfused dog forelimb series. [The base- line (dashed line) is drawn parallel to the dye-concentra- tion plateaus to correct for recirculating dye.] ‘— fl 55 Es}... at. 2069.2. 8. 20.3.12. co. no.5 455.52 451045 2.39 ‘ C .Illlllll llllllllllllllllllll Imidm; \x INI 45:35.5 03583.5 iI >o NAN >m .II.A|z_u> 023.124..." .8233 O mun: +44a4_a«_.naa....a.ati.t.++mmmmww+¢m¢mw.mammfimmmwmmamfi.ww m2... 56 TABLE 5.--Comparison of actual and dye-dilution calculated blood flows in the isolated, ' pump-perfused dog forelimb series. ‘ IDve - Actual I Dog No. Actual Flow cgiidiiégdigion 1 Actual x 100) % Difference ml/min ml/min 40 . 92 97.5 + 5.9 . 37.5 42.7 +13.8 127 129 + 1.5 187.5 168 -10.4 24.5 26 + 6.1 41 93 97 + 4.3 134 128 — 4.4 56.3 56.5 + 0.3 167 168 + 0.5 22 23 + 4.5 200 193 - 3.5 39 41 + 5.1 42 110 119 + 8.1 138 ~ 149 + 7.6 22.1 20.5 - 7.2 91 78 -l4.2 162 108 -33-3 56 51.5 - 8.0 190 138 -27.3 37.5 36 - 4.0 129 122 - 5.4 165 . 185 +12.1 165 185 +12.l 165 191 +15.7 43 89.7 88.5 - 1.3 - 119.7 122 + 1.9 36.2 ‘37.3 + 3.0 137 132 - 3.6 . 179 170 - 5.0 44 109 108 - 0.9 21.8 24 +10.0 53.7 49 - 8.7 199.2 209 + 4.9 73.2 71 - 3.0 150.3 144 - 4.1 39.3 37 - 5.8 194 191 - 1.5 45 111 110 - .9 148 145 - 2.0 71.4 68 - .7 39.3 35 -10.9 188 170 - 9.5 21 5 21.5 0 87 87 0 53 5 53 - 1.1 210 180 -14.2 46 107 123 +14.9 180 9 206 +13.8 37.9 35.3 - 6.8 139.5 112 -l9.7 51 47 - 7.8 84 7 62.5 -26.2 19 4 21 + 8.2 79 1 97 +22.6 47 81.7 77 - 5.7 97.9 91 - 7.0 111.5 111 - 0.4 131 145 +10.6 21.3 22 5 + 5.6 199.5 206 + 3.2 38.4 39.5 + 2.8 63.4 64 5 + 1.7. 150 162 + 8.0 150 173 +15.3 150 162 + 8.0 28 29.7 + 6.0 28 26.8 - 4.2 28 30 + 7.1 57 differed from actual forelimb flow by a mean difference of -0.5% i 10.2 (S.D.). The correlation coefficient is .97. The data are also presented graphically in Figure 21. In this plot, the center solid line is the line of identity; points on this line represent a dye-dilution calculated flow equal to actual flow. The two lines on either side of identity represent a 10 percent error for dye-dilution calculated flows. It is clear that there was greater variability of dye-dilution calculated flows at high forelimb flow rates. This variability at higher flow rates occurred primarily in measurements made in two dogs, numbers 42 and 46, experiments which were technically very satisfactory; no justification for discarding these two experiments was found. In general, the lower dye- concentration plateaus obtained during high blood flow rates may be less accurate than plateaus during low flows because of the lower instrumental signal to noise ratio. 58 Figure 21.—-A plot of actual versus dye-dilution calcu- lated blood flows in the isolated, pump—perfused dog forelimb. 59 BESS. 30...... 300.5 mi... 44.2.04 can com o... om. om. co. m.» om mm \ X \ \\ o .mN an... .nx: .mm. .00. .0». .DNN MOB OEIV'IOO'IVO 3A0 (”AU/IL“) CHAPTER IV CONTINUOUS DYE-DILUTION FLOW MEASUREMENTS IN THE INTACT, NATURALLY PERFUSED DOG FORELIMB Purpose of Investigation This section of the investigation was designed to test whether or not the addition of indicator to brachial arterial blood and the sampling of downstream indicator concentrations allowed accurate measurement of brachial arterial blood flow in the intact dog forelimb. The intact dog forelimb is more similar to the forearm and hand of man than is the isolated dog forelimb. Thus, inferences derived from this natural preparation may be more rationally extended to man. In addition, by in- fusing two indicators simultaneously (ICG and RISA) it was possible to compare the technically more difficult dye-dilution measurement system against the technically simpler RISA-dilution measurement system. Thus, the RISA-dilution system was used as an internal standard. Methods Figure 22 is a diagram of the experimental prepar- ation. Six mongrel dogs (18—27 kg) were anesthetized with sodium pentobarbital (30 mg/kg), intubated and 60 61 Figure 22.--The intact, naturally perfused dog forelimb experimental preparation. 62 m_0>m_wmwm diam mDOZm> maozm> . ms‘ rarefied “WW .. IE. o>mmm ! mmkmiotmzmo @551 ...<.S> tmmm2o II >OIVI>0I7I>OI70I>TIAI )QI IIII ONJQSCm 2.”) .d: o .83 MI... 0 .:.E\ .E. .. on 55:33.. .. 2. -30: 453mb: .. 8. 45.635 0 I3 .0: \ .55 usuoucm g2ub¢< 67 Cephalic and brachial venous outflow were measured by timed collection of blood in graduated cylinders. Brachial arterial inflow was Obtained directly from the rotameter recording. The slope of the dye-concentration plateau(s) was taken as in indication of indicator re- circulation (see Figure 6, Chapter I). The baseline (dashed line, see Figure 23) was drawn parallel to the plateaus to correct for this recirculation (see equation 6, Chapter I). The height of the indicator-concentration plateau above the corrected baseline was used to calcu- late blood flow. The equation for RISA-dilution calcu- lated flow is given as follows: Flow ._ CPM RISA infusedgper minute into brachial artery (ml/min) - CPM/ml venous blood - CPM/m1 arterial blood In the case of dissimilar indicator—concentration plateaus in the two outflow veins, flow was calculated independently for each plateau, and a mean of these two flow rates re- corded as calculated flow. Results A total of 33 multiple flow measurements was ob— tained from six dogs, and the results are presented in Table 6. Figure 24 Shows the relationship between RISA- dilution calculated flow and brachial arterial flow. The SIOpe of the regression line (shown dashed) for RISA 68 TABLE 6.--Measured blood-flow values from the intact, naturally perfused dog forelimb series. HIS —Dilution Dye—Dilution Brachial Total Dog No. Calculated Calculated Arterial Flow Forelimb Venous Flow Flow (Rotameter) Outflow ml/min ml/min ml/min ml/min 104 130 145 --- -—- 132 149 121 127 168 178 167 164 87 96 87 86 48 51 42 46 20 22 22 20 105 162 113 122 188 138 140 171 77 68 96 28 37 106 79 109 54 72 . 122 174 104 108 170 200 145 150 38 2 27 37 65 C2 45 5. 107 127 143 100 115 149 227 129 140 169 255 153 175 47 53 35 “3 115 134 7 105 60 58 35 43 108 163 140 140 158 210 300 173 192 176 150 138 159 125 126 98 114 87 95 50 73 71 54 64 159 160 132 140 109 119 101 91 112 188 163 140 157 160 157 ‘20 136 110 119 78 99 69 Figure 24.--A plot of brachial arterial (rotameter) versus RISA—dilution blood flows in the intact, naturally perfused dog forelimb. The solid line is the line of identity; the dashed line is the calculated regression line for RISA dilution flow values. - 7O g o / 2 / E200“ / / 1 o / o O J / I50i u. ,’° / D o 2 _||00‘ ,’ 0/ EI] 751 p or 0 °’ 50~ °,;o a ,6 E 25* 25 50 75 '00 IZO I50 I75 200 BRACHIAL ARTERIAL FLOW (mu/mm 71 is 1.08 and the y— intercept is +13.8 ml/min. For clarity, the dye—dilution calculated flows were omitted from this figure. It is apparent that, at least in the intact dog forelimb, the isotopic indicator—dilution method did not measure brachial artery blood flow, but rather some flow in excess of brachial artery flow. Dye—dilution calculated flows in four dogs were obtained using a cuvette blood withdrawal rate of 1.5 ml/min. In two dogs (108 and 109) the cuvette was used at a withdrawal rate of 2.4 ml/min. Figure 25 shows the relationship between dye- and RISA-dilution calcu- lated flows at the two cuvette blood withdrawal rates. A higher correlation was obtained between dye- and RISA- dilution calculated flows at the 2.4 ml/min withdrawal rate (r = .97) than for a similar comparison at the 1.5 ml/min rate (r = .93). However, statistically these correlations were not significantly different. There are several indications for the presence of collateral arterial blood blow in this forelimb prepar- ation. First, with few exceptions, dye- and RISA- dilution calculated flow were higher than measured brachial artery inflow, suggesting that the indicator in brachial artery blood was diluted by indicator—free collateral blood flow. Second, total cephalic and brachial limb venous outflow was almost always greater than measured brachial arterial inflow. 72 Figure 25.--A plot of dye— versus RISA-dilution calcu— lated flow in the dog at two cuvette blood withdrawal rates. DYE CALCULATED FLOW (ml/Mm 73 d 0 20° oI.5 ml /min WITHDRAWAL RATE (30068) 02.4 ml/min WITHDRAWAL RATE (20065) l75- ° l50‘ l25- I004 75- 50- 25- 1 l I I I I I l 25 so 75 ICC 125 ISO I75 200 RISA CALCULATED FLOW (ml/MIN.) CHAPTER V CONTINUOUS DYE-DILUTION FLOW MEASUREMENTS IN THE HUMAN FOREARM AND HAND Purpose of Investigation This section of the investigation was designed to determine (1) the stability of systemic concentrations of ICG (i.e., recirculating ICG) during continuous intra- vascular ICG infusion in man, (2) the practicality of the present continuous-infusion dye-dilution technique for measuring blood flow in the human forearm and hand, and (3) the degree of agreement between blood flow in the human forearm and hand as measured simultaneously by the dye- and RISA—dilution techniques. Methods The experimental procedure was directly supervised by a licensed physician, H. W. Overbeck, M. D.‘ All sub~ jects participating in this study were selected by him and informed of the purposes, hazards, and procedures of the experiment. Written consent was obtained from all subjects. These healthy male volunteers, aged 21-30, 74 75 were studied in the resting, post—absorptive state at an ambient temperature maintained near 76° F. The sub— ject was made comfortable in the supine position, and his arms supported at a 45 degree angle from the long axis of the body. Recirculating dye concentrations were measured in eleven subjects during continuous ICG infusion. Dye was infused (OJHBmg/min) into either the antecubital vein or brachial artery of one forearm for varying lengths of time. In some.instances a loading dose (.2 to .6 mg/L blood volume) of dye was given to the subject prior to the dye infusion period. Arterial or venous blood was sampled at varying intervals from the opposite forearm. Blood samples were heparinized and centrifuged. Plasma from each sample was carefully diluted (1:7) with saline. The optical density of each diluted sample was measured against an undyed plasma sample on a Beckman DU spectro— photometer. A calibration curve, obtained by measuring the optical density of knoWn plasma-dye mixtures, allowed the conversion of sample Optical density to its corres- ponding plasma-dye concentration (mg/L). Forearm and hand blood flow was measured in eight subjects. Under sterile conditions, lB—gauge thin wall hypodermic needles were inserted in an upstream direction into the basilic, cephalic and/or antecubital veins of one forearm (designated ipsilateral forearm) distal to 76 the elbow. Thin wall, sterile Teflon cannulae were inserted through these needles, and the tips were posi- tioned to insure an adequate venous blood supply to the cuvette. The 18—gauge needles were then withdrawn from the veins and the Teflon cannulae connected to the dye cuvette by sterile silastic tubing. Under local procaine (procaine hydrochloride, Cutter Laboratories) anesthesia, the brachial arteries of both forearms were cannulated in an upstream direction with 20-gauge Riley arterial needles. A jet injector needle (modi- fied by Overbeck), requiring a volume flow rate of 2.7 ml/min, was inserted through the ipsilateral Riley needle until the tip of the jet needle lay within the lumen of the brachial artery. A blood reflux through the jet needle indicated that its tip lay within the arterial lumen. The ipsilateral Riley needle was then withdrawn. Venous cannulae and arterial needles were kept patent by intermittent flushing with a heparinized isotonic sodium chloride solution. The experimental preparation is shown in Figure 26. Cardio—Green dye (50 mg) with RISA (50 uc/subject) was prepared in sterile isotonic solutions. Indicator in saline alone was infused intra-brachial arterially dur- ing resting blood flow measurements. Indicator in saline with either nor-epinephrine (LeVOphed—Winthrop Laboratories, 0.10 y/min infusion rate) or acetylcholine 77 Figure 26.——Experimenta1 preparation for continuous- infusion, indicator—dilution blood flow measurement in the human forearm and hand. 78 Im€ca>dom GE“ 95.. Q ‘ ZO_mDuz_ .... U 03:53:. I cwhuzotmzmo _- I ' \ max—Mg £22 v 44223:...) 79 (acetylcholine chloride, Merck, 6 y/min infusion rate) was infused.intrabrachial arterially to vary the resting blood flow rate. Forearm flow was also varied by having the subject exercise his hand. In every case, dye was infused at the rate of .42 mg/min. The measurement of blood flow by the continuous- infusion indicator-dilution technique was conducted in the following manner (see Figure 27). Indicator infusion was begun into the contralateral arterial needle. Ipsilateral venous blood was withdrawn through the cuvette (1.5 or 2.4 ml/min) until a stable baseline re- sponse, indicating a constant level of recirculating dye, was observed on the polygraph recording. This usually required approximately 5 minutes. Indicator infusion was then changed to the ipsilateral arterial jet needle. Ipsilateral venous dye concentrations in the two down- stream veins were recorded by withdrawing venous blood through the cuvette in sequence. (The sudden decrease in optical density, which is shown during switching from one vein to the other, was artifactual.) Two ml samples of venous and contralateral arterial blood were obtained during the recorder inscription of the dye-concentration plateau. These blood samples were used for calculating RISA-dilution blood flow. Indicator infusion was then returned to the contralateral artery. 80 Figure 27.-—A typical dye-concentration plateau from the human forearm and hand. [The baseline (dashed line) is drawn horizontally because recirculation is constant in the human]. 81 is}... «3 >522 45:22; 355.2528 .2 228...... 55032. 332.53 Soomuz 53 >522 292... $4.53 5;. .2523 44554.9... 0» + + .6522. «9203:. :35“ 20.5.5058: 2.2.268 Al mz_._um< >< . Anemia; 26> .8208..." m an... “2:. D o3<§uo {J :o 777 82 The dye cuvette was calibrated after the experiment, following the protocol outlined in Chapter II. The height of the dye-concentration plateau above the baseline was used to calculate blood flow. Results de Recirculation Figure 28 is a plot of recirculating ICG dye con- centrations against time in three subjects. In subject B.S., blood-ICG concentrations varied between 0.32 and 0.52 mg/L during a continuous three-hour ICG infusion period (.42 mg/min). In subjects J.R. and L.R., ICG was infused for 30—minute periods with a 30-minute rest between infusions. Blood-ICG concentrations in these subjects were within the range 0-0.5 mg/L. Similarly, in eight other studies the level of recirculating ICG dye was low (< 0.5 mg/L) and quite constant with time. These data suggest that, under these conditions, a steady state is reached in which liver extraction equals dye infusion during the ICG infusion period. It is apparent that a steady state was rapidly achieved whether or not the subject was initially given a loading dose of dye. The concentration of recirculating dye (< 0.5 mg/L) should be compared to the variations in dye con- centration observed during dye-dilution blood flow measurements. These computed values range from 2.1 mg/L 83 Figure 28.—~A graph of systemic ICG dye concentrations in man against time during continuous intravascular ICG dye infusion. 84 EEBENE 20.8.12. .96 § 855.5 ms: ow. 0m. 0.! ON. 00. Om Om o¢ ON . P p p, p b - .m .._ :1. Ix... . . ‘ 1 .fi A>m SEN V 25.5 .>.m._\uENV «Eta 3.0._\oEt. «Etn— E g E 1.4 guesses .5... .fl .1\\\\\\\\\\‘“\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\‘\\\\\\‘\\‘\\\\\\‘\\\\\\‘\‘\\\\\\\\\\\\\\\““‘. .m.m 71,721} r 3am ._\uE Q. 25.51 I szrfii loo 0 0.. NOIlVHiNBONOO 3A0 (I/fiw) 85 for a flow of 200 ml/min to 16.8 mg/L for a flow of 25 ml/min. With recirculating dye concentrations varying, at most, in the range 0-0.5 mg/L, the maximum error introduced by changes in recirculating dye during the recording of dye-concentration plateaus is 23 percent 0.5 m /L 0.5 m /L ( . mg L x 100) and 3 percent (1678—gg7f x 100) for flows of 200 and 25 ml/min respectively. Practicality of the Technique in Humans Blood flow measurement was attempted in the fore- arms and hands of nine subjects. In one subject the forearm venous blood flow was inadequate for this techni- que and the experiment was terminated. In three other subjects, blood flow measurements were partially success- ful: in two subjects (203 and 204) by choice only one vein was sampled for indicator concentrations; in a third subject (214) only one vein would supply the re— quired blood flow. In the remaining five subjects, two forearm veins were continuously and sequentially sampled for indicator concentrations during continuous indicator infusion. The most common technical problem encountered with the present technique was that of maintaining an adequate venous blood flow to the dye cuvette. Frequent manipulation of the venous cannulae was necessary to insure a continued blood supply. Blood clotting in the blood withdrawal tubing was usually prevented by 86 frequent flushing with isotonic saline. Total blood loss to the subject was less than 300 ml per experiment. Comparison of Dye- and RISA- Dilution Flow Measurements A total of 27 paired human forearm and hand blood flow measurements were made in eight subjects. The data are presented in Table 7. Dye-dilution calculated flows were obtained in subjects 203 and 204 using a cuvette blood withdrawal rate of 1.5 ml/min. In the remaining subjects the cuvette was used at a withdrawal rate of 2.4 ml/min. The relationship between dye- and RISA~dilution calculated flows, and cuvette blood with- drawal rate is shown in Figure 29. A higher correlation was obtained between dye- and RISA—dilution calculated flows at the 2.4 ml/min withdrawal rate (r a .99) than for the similar comparison at the 1.5 ml/min rate (r = .97). (Statistically, these correlations were not significantly different.) Apparently, dye—dilution calculated flows agreed better with RISA-dilution calculated flows at the higher cuvette blood withdrawal rate. There is good evidence that subject 212 had an anomalous high brachial artery bifurcation, and that indicator was infused into the radial rather than the brachial artery. Thus basilic venous indicator concen— trations were only slightly higher than recirculating indicator concentration--most of the indicator was 87 'DABLE 7.--Comparison of dye- and RISA-dilution calculated flow measurements from the human forearm and hand. Subject Dye Calculated RISA Calculated Number Flow Flow ml/min ml/min 203 42 50 48 55 162 138 194 161 204 131 111 137 99 243 175 116 87 210 118 136 150 154 345 317 211 100 100 195 200 120 134 114 115 212 (C.V. flow only) 36 38 45 44 27.5 31 13.2 13.5 214 (B.V. flow only) 35 38 75 63 215 126 118 225 236 216 75 89 142 165 65 69 46 48 88 Figure 29.—-A plot of dye— versus RISA—dilution calcu- lated flow in the human forearm and hand at two cuvette blood withdrawal rates. DYE CALCULATED FLOW (ml/MIN.) 250- 225 5 I75‘ l50‘ I25 " I00“ 75‘ 25* 89 ol.5 ml/ MIN. WWW“ RATE 02.4 Iii/m. WITHORAWAL RATE 2'5 5'0 7‘5 IOO I25 g0 I‘l/S ZOO 225 20 RISA CALCULATED FLOW (ml/MIN.) 9O distributed to the vascular bed drained by the cephelic vein. During nor-epinephrine infusion in this subject, only the radial portions of the forearm and hand visibly blanched, suggesting that there is a clear demarcation between the distribution of radial and ulnar arterial blood, at least in this subject. CHAPTER VI DISCUSSION AND CONCLUSIONS Recirculating Dye Constant—infusion indicator-dilution methods for studying blood flow have been complicated by the presence of changing concentrations of recirculating indicator. The accumulation of indicator in the blood necessitated an additional measurement and appropriate correction in the calculation of blood flow (refer to equation 6, page 7). The present study used a dye (ICG) which is rapidly extracted from the blood by the liver. It was hoped that use of a low dye infusion rate, coupled with good hepatic removal, would produce low and constant recirculating dye levels in the blood during constant- infusion dye-dilution blood-flow measurements. This was not the case in the dog forelimb studies presented in Chapters III and IV; dye extraction by the liver may have been impaired by anesthesia, shock, or undetected liver disease. However, dye recirculation was measured in the human during continuous dye infusion and was found to be both low in concentration and quite con- stant with time. During blood flow measurements in the human, therefore, recirculating dye was essentially 91 92 constant, adding a constant increment of optical density to the dye-concentration plateau. Thus, dye recircu- lation need be measured only at intervals and the measure- ment of regional blood flow in the human is greatly simplified. The constancy of recirculating ICG levels is therefore a major advantage of the measurement techni- que presented in this paper. Actual Versus Calculated Blood Flow The isolated dog forelimb series (Chapter III) was a preliminary study to determine if the cuvette densito- meter dye—dilution method measured blood flow with reason- able accuracy. Dye-dilution calculated flows showed good agreement with actual flows-—a significant finding in View of the fact that cuvette blood withdrawal rates were well below those ordinarily used. The reported accuracy was lower than that found by Overbeck (1966) in a similar series of experiments in the isolated, pump- perfused dog forelimb using RISA for discontinuous, indicator—dilution blood-flow measurements. In 30 paired measurements, he found that 77 percent of RISA- dilution calculated blood flow values were within : 5 percent of actual forelimb flows. However, subsequent improvements in instrumentation have probably increased the accuracy of the cuvette densitometer dye—dilution method to a level equal to that of the RISA-dilution 93 method. This View is supported by data from the intact dog and human limbs comparing RISA—dilution and dye- dilution measurements. Indicator Calculated Flow in the Intact Dog Forelimb The isolated dog forelimb had only one arterial inflow, namely, that supplied by the blood pump. Such F“ a preparation was quite dissimilar from the human fore- arm where it is possible that more than one artery supplies blood to the limb. Studies in the intact, naturally perfused dog forelimb, then, were designed to determine which blood flow parameter is measured during the infusion of indicator into the brachial artery of a naturally perfused limb; inferences drawn from these studies could be more rationally extended to measurements made in the human forearm. In the intact dog forelimb, dye-dilution measured a flow greater than brachial arterial flow. The over—estimation of brachial arterial flow probably indicates that undyed blood was entering the forelimb through collateral vessels, diluting dyed blood from the brachial artery. The presence of these collateral vessels has been confirmed anatomically. In addition, venous outflow from the limb usually exceeded brachial arterial inflow, also arguing for the presence of collateral blood flow. Interestingly, the RISA regression line was nearly parallel to the line of flow 94 identity (see Figure 24, Chapter IV), suggesting that the amount of collateral blood flow was almost constant over the range of limb blood flows studied (if the pro— portion of total collateral blood flow entering the sampled limb veins remained constant--an assumption which seems justified). The apparently constant rate of this collateral dilution would also seem to indicate that the anastomosis occurred on the venous, not arterial, side of the limb vascular bed. If anastomosis occurred on the arterial side, one might expect that dilution by collateral blood flow would increase during vasoconstriction and decrease during vasodilation, rather than remain fairly constant. It would seem logical to extend these results in the intact dog forelimb to the forearm and hand in man. It follows that indicator- dilution measurement of blood flow in the limb of man may represent measurement of brachial arterial blood flow plus a constant increment. Dye— Versus RISA—Dilution Calculated Flows Because of its relative technical simplicity, and thus potentially greater accuracy, the RISA-dilution technique was used as an internal standard for measur- ing blood flow in the present studies. Simultaneously measured dye—dilution blood flows were compared with RISA-dilution calculated flows in order to estimate further the accuracy of the present dye technique. 95 In both intact dog forelimb and human forearm, dye-dilution calculated flows tended to be higher than RISA-dilution calculated flows when a cuvette blood withdrawal rate of 1.5 ml/min was used. Increasing the cuvette blood withdrawal rate to 2.4 ml/min brought the two indicator-dilution measurements into better agree— ment, although this could not be statistically proven. An increase in the accuracy of dye—dilution calculated flows at a higher cuvette withdrawal rate is reasonable in view of the fact that the cuvette becomes less flow dependent, and thus more stable, at higher withdrawal rates. Overall, the good correlation between these two independent indicator-dilution techniques suggests that both techniques measure blood flow with equal accuracy in these applications. Measuring Blood Flow in the Forearm and Hand of Man Blood flow measurements in man are usually made by indirect techniques. Such techniques must provide reliable measurements without compromising the well— being of the subject. The present study has described a dye—dilution technique for measuring forearm and hand blood flow which meets both criteria. Less than 300 m1 of blood are withdrawn from the subject during the procedure, and the principle complaint of subjects tested is the necessity for remaining supine and quiet 96 during the experiment. The constant level of recircu- lating dye eliminates the need for the sampling of contralateral arterial blood, and thus the necessity for contralateral arterial puncture. In addition, the present technique uniquely provides immediate and con- tinuous blood flow measurements. Errors in Technique Instrumental Instability ' k Data on instrumental instability have been cited in Chapter II. While the present modification program has provided a reasonable degree of cuvette stability, further improvements are necessary. In addition, a cuvette densitometer system with a linear response would be highly desirable. Stability of Flow Through the Cuvette The sensitivity of the present cuvette densitometer system to changes in blood withdrawal rate has previously been described (see Chapter II). This phenomenon is a potential source of error in the present technique. Changes in cuvette blood withdrawal rate, caused by either a venous obstruction or a blood clot in the withdrawal cannula, are usually readily detected by a shift in the recording baseline during the withdrawal of undyed blood. However, a similar change in with- drawal rate during the recording of a dye concentration 97 plateau produces a change in the plateau which may be obscured by changes in forearm blood flow and vise versa. The sampling of blood from veins with a good blood supply and careful placement of the venous cannula both help to insure that blood flow through the cuvette is constant during the measurement of venous dye con- centrations. Monitoring withdrawal pressures would also help to detect changes in cuvette withdrawal rate. Blood Changes The relationship between cuvette output and blood hematocrit has been discussed in Chapter II. In the studies in man, systemic blood hematocrit decreased less than 2 percent during the experimental measure- ment of forearm and hand blood flow thus adding in- significant error. Changes in other blood constituents and their effect on cuvette output have not been studied. Agents which change the shape of red cells, such as hypo- or hyper-tonic solutions, might alter the optical density of blood and interfere with flow measurements by this system. I... . BIBLIOGRAPHY Andres, R., K. L. Zierler, H. M. Anderson, W. N. Stainsby, G. Cader, A. S. Ghrayyib, and J. L. Lilienthal. Measurement of blood flow and volume in the forearm of man; with notes on the theory of indicator-dilution and on production of tur- bulance, hemolysis and vaso dilatation by intra— vascular injection. J. Clin. Invest. 33:482, 1954. Agrifoglio, G., G. D. Thorburn, and E. A. Edwards. Measurement of blood flow in the human lower extremity by indicator-dilution method. Surg. Gynec. Obstet. 113:641, 1961. Baltzan, M. A., R. Andres, G. Cader, and K. L. Zierler. Effects of epinephrine on forearm blood flow and metabolism in man. J. Clin. Invest. 44:80, 1965. Barbier, F., and G. A. De Weerdt. Chromatography and I. R. spectrography of indocyanine green. Clinica Chemica Acta. 10:549, 1964. Bassingthwaighte, J. B., and A. W. T. Edwards. Dye-dilution curves from the pulmonary—artery and aorta sampling sites compared with the same curves sampled at the femoral artery. (Abstr.) Fed. Proc. 19:118(Mar), 1960. Cherrick, G. R., S. W. Stein, C. M. Leevy, and C. S. Davidson. Indocyanine green: observations on its physical prOperties, plasma decay and hepatic extraction. J. Clin. Invest. 39:592, 1960. Cobb, L. A. Effects of reducing agents on indo— cyanine green dye. Am. Heart J. 70:145, 1965. Cournand, A. F. Symposium on cardiac output. Federation Proceedings. 4:183, 1945. Cropp, G. J. A., and A. Burton. Theoretical con— siderations and model experiments on the validity of indicator-dilution methods for measurements of variable flow. Circ. Res. 18:26, 1966. 98 10. ll. 12. 13. l4. 15. l6. l7. l8. 19. 99 Dow, P. Estimations of cardiac output and central blood volume by dye dilution. Physiol. Revs. 36:77, 1956. Folse, R. Application of the sudden injection dye dilution principle to the study of the femoral circulation. Surg. Gynec. and Obstet. 120:1194, 1965. Fox, I. J., L. G. S. Brooker, D. W. Heseltine, H. E. Essex, and E. H. Wood. A tricarbocyanine dye for continuous recording of dilution curves in whole blood independent of variations in oxygen satur- ation. Proc. Staff Meet. Mayo Clinic. 32:478, 1957. Fox, I. J., W. F. Sutterer, and E. H. Wood. Dynamic response characteristics of systems for continuous recording of concentration changes in a flowing liquid. J. Applied Physiol. 11:390, 1957. Fox, K. J., and E. H. Wood. Indocyanine green: physical and physiologic properties. Proc. Staff Meet. Mayo Clinic. 35:732, 1960. Fox, I. J., and E. H. Wood. Blood flow measurements by dye dilution techniques. In Glasser, 0. (ed.): Medical Physics, Chicago, Year Book Medical Pub- lishers, Inc., 1960, vol. 3, pp. 155-156. Friedlich, A., R. Heimbecker, and R. J. Bing. Device for continuous recording of concentration of Evans Blue dye in whole blood and its application to determination of cardiac output. J. Appl. Physiol. 3(1):12-2l, 1950. Greenfield, A. D. M., R. J. Whitney, and F. J. Mowbray. Methods for the investigation of peri— pheral blood flow. Brit. Med. Bul. 19:101, 1963. Hamilton, W. F., J. W. Moore, J. M. Kinsman, and R. G. Spurling. Simultaneous determination of the pulmonary and systemic circulation times in man and of a figure related to the cardiac output. Am. J. Physiol. 84:338, 1928a. Hamilton, W. F., J. W. Moore, J. M. Kinsman, and R. G. Spurling. Simultaneous determination of the greater and lesser circulation times, of the mean velocity of blood flow through the heart and lungs, of the cardiac output and an approximation of the amount of blood actively circulating in the heart and lungs. Am. J. Physiol. 85:372, 1928b. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 100 Hamilton, W. F., J. W. Moore, J. M. Kinsman, and R. G. Spurling. Blood flow and intrathoracic blood volume as determined by the injection method and checked by direct measurements in perfusion experiments. Am. J. Physiol. 93:654, 1930. Hamilton, W. F., J. W. Moore, J. M. Kinsman, and R. G. Spurling. Studies on the circulation IV. Further analysis of the injection method and of changes in hemodynamics under physiological and pathological conditions. Am. J. Physiol. 99:534, 1932. Hamilton, W. F., and J. w. Remington. Comparison of time concentration curves in actual blood, of diffusible and nondiffusible substances when in- jected at constant rate and when injected in- stantaneously. Am. J. Physiol. 148:35, 1947. Hamilton, W. F. The physiology of cardiac output. Circulation. 8:527, 1953. Henriques, V. Uber die verteilung des blutes vom linken herzen swischen dem herzen and dem ubrigen organismus. Biochem. Z. 56:230, 1913. Hering, E. Experiment on the velocity of the cir- culatory motion of the blood and on the quickness with which secretions are formed. Edinburgh Phil. Trans. 6:78, 1829. Hobbs, J. T., and E. A. Edwards. Femoral artery blood flow, limb blood volume and cardiac output through continuous recording indicator-dilution curves. Ann. Surg. 158:159, 1963. Holt, J. P. The effect of positive and negative intrathoracic pressure on cardiac output and venous pressure in the dog. Am. J. Physiol. 142:594, 1944. Howard, A. R., w. F. Hamilton, and P. Dow. Limitations of the continuous infusion method for measuring cardi- ac output by dye dilution. Am. J. Physiol. 175: 173, 1953. Hunton, D. B., J. L. Bollman, and H. N. Noffman. Hepatic removal of indocyanine green. Proc. Staff Meet. Mayo Clinic. 35:752, 1960. Ketterer, S. G., B. D. Wiegand, and E. Rapaport. The excretion of indocyanine green and its use in the estimation of hepatic blood flow. Clin. Res. 7:71, 1959. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 101 Ketterer, S. G., B. D. Wiegand, and E. Rapaport. Hepatic uptake and bilary excretion of indocyanine green and its use in estimation of hepatic blood flow in dogs. Am. J. Physiol. 199:481, 1960. Krasavage, W. J., and S. M. Michaelson. Indocyanine green plasma half time clearance (t%) in normal beagles. Proc. Soc. for Exper. Biol. and Med. 119:215, 1965. Kupic, E. A., T. Neyazaki, R. Sugden, and L. A. Sapirstein. Theoretical limitations in the indi- cator dilution technique in coronary blood flow measurement. Invest. Radiol. 1:129—138, 1966. Marchall, H. W., I. J. Fox, F. S. Rodich, and E. H. Wood. Method for simultaneous measurement of total and lower body blood flow using indicator- dilution techniques. Proc. Staff Meet. Mayo Clinic. 35:774, 1960. Meier, P., and K. L. Zierler. 0n theory of indi- cator-dilution method for measuring blood flow and volume. J. Appl. Physiol. 6:731, 1954. Moore, J. W., J. M. Kinsman, W. F. Hamilton, and R. G. Spurling. Studies on the circulation 11. Cardiac output determinations, comparison of the injection method with the direct fick procedure. Am. J. Physiol. 89:322, 1929. Overbeck, H. W. Indicator-dilution measurement of blood flow in the upper extremity of man: develOp- ment and use of a method improving mixing of indicator and vasoactive agents with brachial arggrial blood. Graduate College, U. of Oklahoma, 19 . Peterson, L. H., M. Helrich, L. Green, C. Taylor, and G. Choquette. Measurement of left ventricular output. J. Appl. Physiol. 7:258, 1954. Quain, R. The anatomy of the arteries of the human body and its application to pathology and operative surgery. London, Taylor and Walton, 1844. Reemtsma, K., G. C. Hottinger, A. C. DeGraff, and 0. Creech. Studies of hepatic function and blood flow using indocyanine green. Clin. Res. 8:77, 1960. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 102 Schlenk, W. G. Methods for measurement of blood flow. A current appraisal. J. Surg. Res. 6:361, 1966. Sekelj, P., A. Oriol, N. M. Anderson, J. Morch, and M. McGregor. Measurement of indocyanine green dye with a cuvette oximeter. J. Appl. Physiol. 23(1):114-120, 1967. Shepherd, J. T., D. Bowers, and E. H. Wood. Measure- ment of cardiac output in man by the injection of dye at a constant rate into the right ventricle or pulmonary artery. Am. J. Physiol. 179:673, 1954. Shepherd, J. T., D. Bowers, and E. H. Wood. Measurement of cardiac output in man by injection of dye at a constant rate into the right ventricle or pulmonary artery. J. Appl. Physiol. 7:629, 1955. Stephenous, J. I. Theory of measuring blood flow by dilution with an indicator. Bull. Math Bio— physics. 10:117, 1948. Stewart, G. N. The output of the heart. J. Physiol. 22:159, 1897. Stewart, G. N. The output of the heart in dogs. Am. J. Physiol. 57:27, 1921. Wade, 0. L., J. M. Bishop and K. W. Donald. Cardiac output and regional blood flow. Blackwell Scien- tific Publications, Oxford, 1962. Wahren, J. A dye dilution method for the determi- nation of brachial artery blood flow during forearm exercise in man. Acta Physiol. Scand. 64:477, 1965. Wahren, J. A dye dilution method for the determi- nation of brachial artery blood flow during rhythmic exercise. Acta Physiol. Scand. 67:8upplementum 269:15, 1966. Wheeler, H. 0., W. I. Cranston, and J. I. Meltzer. Hepatic uptake and bilary excretion of indocyanine green in the dog. Proc. Soc. Exp. Biol. Med. 99:11, 1958. Wright, H. P. The measure of blood flow. Ann. Royal College Surg. England. 37:292, 1965. 103 53. Zierler, K. L. A simplified explanation of the theory of indicator-dilution for measurement of fluid flow and volume and other distributive phenemon. Johns Hopkins Hospital Bulletin. 103:199, 1958. 54. Zierler, K. L. Circulation times and the theory of indicator-dilution methods for determining blood flow and volume. Handbook of Physiology, Circulation vol. 1, Hamilton, W. E. (ed.), 1962. 1in Am 5" R”)! 31293 03 78 431 IIIHIIIIIHIIHII EIIIIIIIII