Organ Blood Flow Measured by Radionuclide-labeled Microspheres During Mechanical Ventricular Assistance

Transcription

1 Organ Blood Flow Measured by Radionuclide-labeled Microspheres During Mechanical Ventricular Assistance Alfio Raciti, M.D., Alan J. Greenfield, M.D., David B. Skinner, M.D.,* Michael H. Newman, M.D., and Robert B. Rutherford, M.D. ABSTRACT Regional blood flow distribution measured before, during, and after mechanical ventricular assistance (MVA) was employed for total circulatory support in 7 dogs. Techniques used were electromagnetic flowmeter measurements of cardiac output and radionuclide-labeled carbonized microsphere distribution after left atrial injection to determine organ perfusion. Average cardiac output during ventricular fibrillation and MVA was 66% of base line. The percentage of cardiac output reaching the heart, liver, intestines, and brain increased during MVA. The kidney and pancreas received a constant percentage of cardiac output during MVA compared to base line. Flow to the spleen and to the musculoskeletal and skin portions of the body decreased. Comparison of flows measured simultaneously in the renal artery by electromagnetic flow probe and microsphere distribution techniques yielded no significant differences between the two methods. These results confirm that MVA provides satisfactory blood flow to vital organs during total circulatory support and illustrate the value of the microsphere distribution technique in performing simultaneous measurements of blood flow to multiple organs. D uring mechanical circulatory support, blood flow distribution to various organs may differ from distribution during normal circulation. The flow reaching vital organs may determine the success of a method for circulatory support more than the total cardiac output and blood pressure. Mechanical ventricular assistance (MVA) described by Anstadt, Schiff, and Baue [I] provides a means of total circulatory support which appears promising for clinical application [8]. From the Department of Surgery, The Johns Hopkins University School of Medicine and The Johns Hopkins Hospital, Baltimore, Md. Supported by Grant No. HE from the National Institutes of Health, Department of Health, Education, and Welfare, and Contract No. DADA C-8070 from the U.S. Army Medical Research and Development Command. *John and Mary R. Markle Scholar in Academic Medicine. Accepted for publication July 2, Address reprint requests to Dr. Skinner, Blalock 626, The Johns Hopkins Hospital, Baltimore, Md VOL. 11, NO. 1, JANUARY, I971 43

2 RACITI ET AL. Previous studies of regional blood flow during MVA have been performed employing an electromagnetic flowmeter to measure the flow to one arterial system at a time, such as the coronary, renal, or carotid system [S, 101. To determine the overall effect of MVA on regional blood flow, flow to several organs was measured simultaneously by the radionuclide-carbonized microsphere distribution technique introduced by Rudolph and Heymann [6]. MATERIALS AND METHODS MECHANICAL VENTRICULAR ASSISTANCE Mechanical ventricular assistance is given by a cupshaped device held on the cardiac ventricles by suction applied to the apex. Systole and diastole are effected by alternating positive and negative pressures delivered through a sidearm opening into a space between the outer cup and an inner diaphragm that are joined at the apex and rim. A Silastic flange at the cup orifice seals to the heart in the region of the atrioventricular groove to prevent pressure from being applied to the atria. The delivery of positive pneumatic pressure and vacuum is controlled by solenoid valves, activated by an electronic programmer. More complete descrip tions of the device and its effect on the heart have been presented previously [I, 8, 101. MEASUREMENT OF REGIONAL B'L66D FLOW If radionuclide-labeled carbonized microspheres (15 p k 5 in diameter") are injected into the left atrium, it is postulated that their distribution throughout the body will be the same as that of the cardiac output. The microspheres will lodge in the first capillary bed that they reach. After the animal is killed, the specific radioactivity of the radionuclides incorporated into the microspheres can be determined for each individual organ and compared to the total amount of radioactivity injected in order to estimate the percentage of cardiac output reaching each organ. The product of this value and the cardiac output, which is mt&s,ured simultaneously with the microsphere injection, provides an estimate of the -actual flow rate to an organ. If three different radionuclides such as 141Ce, 51Cr, and 85%- are employed as labels, the distribution of blood flow can be measured on three different occasions in the same animal by using differential gamma spectrometry and formulas similar to those of Rudolph and Heymann [61. In the experiments reported below the percentage of total counts from each injection reaching the brain, lungs, heart, liver, kidneys, spleen, pancreas, and gastrointestinal tract was determined by the technique previously described in detail by Rutherford and his associates 171 from this laboratory. Radioactivity was measured at three different gamma spectrometer settings-from 50 to 220 kiloelectron volts (kev.), 220 to 400 kev., and 400 to 900 kev.-to separate the emission from each of the three radionuclides according to their different gamma ray energy levels. Results were analyzed on an IBM System 360 Model 40 computer using a program previously developed for this purpose 171. The percentage of counts from each radionuclide reaching each organ and the cardiac output determined simultaneously during each injection were used to calculate the flow per minute to each organ, and this value was expressed in terms of organ and animal weight. Portal blood flow was estimated by a summation of flow to the spleen, pancreas, and gastrointestinal tract. Total liver blood flow was calculated by the addition of portal to hepatic artery flow. 'Supplied by the Minnesota Mining and Manufacturing Co.. St. Paul, Minn. 44 THE ANNALS OH THORACIC SURGERY

3 Organ Blood Flow During Ventricular Assistance EXPERIMENTAL PROTOCOL Seven dogs ranging between 41 and 53 pounds in weight were anesthetized with pentobarbital, 10 mg. per pound. Through a left thoracotomy, an electromagnetic flow probe was placed on the pulmonary artery and connected to a Biotronix Pulsed-logic flowmeter to measure cardiac output. End-diastolic flow was taken to represent zero. A polyethylene catheter was inserted into the left atrium to inject the microspheres. Blood pressure and central venous pressure were recorded. In 6 animals a separate laparotomy was made to place an electromagnetic flowmeter probe and distal occluder on the left renal artery. In 4 of the dogs a snug fit of the probe to the artery permitted stable zero flow values by vessel occlusion. In these animals nine measurements of renal flow were made simultaneously by the flowmeter and microsphere distribution techniques. Only those determinations with a stable vessel occlusion zero were included. After the dissection was completed, blood pressure and cardiac output were measured until these values had been stable for 30 minutes. The first dose of microspheres suspended in 0.1 to 0.2 ml. of 10% dextran was injected into the left atrium, and the catheter was flushed with 10 ml. of saline solution. Next, the heart was fibrillated and MVA was given for approximately two hours. When cardiac output and blood pressure were stable, the second injection of microspheres was made. Cardiac defibrillation was accomplished in 6 dogs, and the third injection of microspheres was made 30 minutes after defibrillation when cardiac output and blood pressure had again stabilized. The animals were sacrificed, and the proportion of each microsphere injection reaching various organs was determined. RESULTS During the base line microsphere injection cardiac output in the 7 dogs averaged 1,342 ml. per minute (standard deviation [SDI 255). During the injection after two hours of MVA, the output averaged 865 ml. per minute (SD 175), or 66% of the control level. Thirty minutes following defibrillation cardiac output averaged 1,180 ml. per minute (SD 475), or 85% of control level. The mean aortic blood pressure at the time of the base line injection averaged 108 mm. Hg (SD 18) in the 7 dogs. After two hours of MVA mean blood pressure averaged 88 mm. Hg (SD 131, or 76% of base line level. At the time of the third injection after defibrillation, mean blood pressure averaged 102 mm. Hg (SD 22). or 90% of control. Continuous monitoring of cardiac output and blood pressure during the injections revealed five occasions when blood pressure and cardiac output decreased for approximately one minute after the microsphere injection. This was accompanied by premature ventricular contractions in the four instances when this was noted during normal heart action. On two occasions systolic blood pressure dropped more than 25 mm. Hg and cardiac output transiently decreased nearly 50%. Analyses of microsphere recovery indicate that there were no significant differences among the three injections (Fig. 1). The percentages of cardiac output unaccounted for include microspheres which lodged in the skin and musculoskeletal portions of the animal as well as the loss in counts due to various technical factors. The decrease in radioactivity unaccounted for during MVA, while not statistically significant, suggests that a smaller proportion of the cardiac output reached the skin and musculoskeletal tissue of the dog during artificial circulation. The proportion of cardiac output reaching the lungs represents bronchial artery flow plus microspheres shunted past capillary beds elsewhere to reach the lungs through the pulmonary artery. The decrease in percentage of cardiac output reaching the lungs during MVA is of borderline significance (0.1 > p > 0.05), suggesting that shunting during artificial circulation was decreased (see Fig. 1). VOL. I I, NO. I, JANUARY,

4 RACITI ET AL. LUNG I i contml MVA FIG. 1. The percentages of radioactivity from each microsphere injection recovered in the lungs and recovered from all organs counted. Lung radioactivity represents bronchial artery flow plus the microspheres passing through or around capillary beds in the systemic circulation. The decreases in lung radioactivity or systemic shunting during MVA are of borderline significance (0.1 > p > 0.05). While not statistically significant, the increase in total counts recovered during MVA suggests that less flow reached the musculoskeletal and skin portions of the body, which were not counted separately. The percentages of cardiac output and flow (ml. per kilogram of body weight per minute) to the heart, brain, liver, intestines, spleen, and kidneys are shown in Figures 2, 3, and 4. The increase in the percentage of cardiac output reaching the heart and intestines during MVA is of borderline statistical sig- HEART INTESTINES WTWr 1 I I FIG. 2. The proportions of injected microspheres reaching the heart and gastrointestinal tract, Increases during MVA are of borderline significance (0.1 > P > 0.05). The pow values obtained when these proportions are multiplied by the actual cardiac output are shown in the lower portion of the figure. 46 THE ANNALS OF THORACIC SURGERY

5 Organ Blood Flow During Ventricular Assistance BRAIN SPLEEN I Control MVA & control MVA After DeFib C S ~ l V 2 n 1 'I 1. FIG. 3. The proportions of injected microspheres and calculated flow to the brain and spleen. The differences noted are not statistically significant. nificance by the t test (0.1 > p > 0.5). The remainder of the data suggests trends rather than statistically significant findings. The percentage of cardiac output reaching the heart, brain, intestines, and liver increased during MVA, while the distribution of cardiac output to the kidneys and pancreas remained unchanged and the proportion of minospheres reaching the spleen and musculoskeletal portions of the body decreased. Because of the diminished cardiac output during MVA there were no significant increases in actual flow to any of the organs during artificial circulation. Flow to the heart, brain, intestines, and liver remained almost unchanged, while flow to the kidneys averaged 64% and flow to the pancreas averaged 72% of base line levels. Flow to the spleen was approximately one-third of control level. After defibrillation the proportion of cardiac output and flow to the heart increased further. The proportions of cardiac output and flow to the portal LIVER KIDNEYS Atter Control MVA tlgi DaFib FIG. 4. The proportions of injected microspheres and calculated flow to the liver and kidneys. The differences noted are not statistically significant. VOL. 11, NO. 1, JANUARY,

6 RACITI ET AL. r =.865 p<.01 I I I IS0 Microsphoro Valuoi cc/min. FIG. 5. The linear correlation best fitting the points shown for simultaneous measurement of left renal flow by flowmeter and microsphere techniques is y = x. The shift in flow measured by the flowmeter from 173 to 112 ml. per minute during one microsphere injection is shown by the arrow. Using the latter value, the correlation coefficient for these points is 0.865, which indicates a significant correlation (p < 0.01). organs and liver after defibrillation were intermediate between base line and assisted circulation levels. The flow and percentage of cardiac output reaching the kidneys, brain, and pancreas were all diminished after defibrillation. In 4 dogs, nine simultaneous measurements of left renal artery flow by the flowmeter and the microsphere distribution technique were compared (Fig. 5). Mean flow measured by the flowmeter was 89.5 ml. compared to 92.3 ml. by the microsphere technique. The points best fitted the slope indicated in Figure 5. The correlation coefficient was 0.87, and the measure of certainty was The test for significance of the correlation coefficient gave a value of t = 4.6 (p < 0.01) [21. During one measurement, cardiac output and blood pressure dropped sharply after microsphere injection and caused a drop in renal flow, as measured by the flowmeter, from 173 to 112 ml. per minute. The latter value was employed in the calculations. There was no significant difference between flow to the undissected right kidney and flow to the left kidney by the microsphere technique. COMMENT The results suggest that blood flow to the vital organs is satisfactorily maintained during total circulatory support by mechanical ventricular assistance in spite of the decreases in cardiac output and blood pressure. Although the values are not statistically significant, the flow to such nonvital regions as the spleen and musculoskeletal system is decreased while flows to the heart, brain, liver, intestines, and kidneys remain unchanged or decrease slightly. 48 THE ANNALS OF THORACIC SURGERY

7 Organ Blood Flow During Ventricular Assistance Coronary flow during MVA measured by the microsphere method correlated closely with results obtained previously using a flowmeter probe on a coronary artery or by collecting coronary sinus flow directly. Values obtained by these techniques averaged 86 to 110% of control levels [lo] compared to an average value of 92% by the microsphere technique. Renal blood flow during MVA as measured by the microsphere technique was less than that found previously by the flowmeter technique, in which values ranged between 90 and 113% of control renal flow [8]. Average cardiac output during MVA found in the previous flowmeter studies was approximately 80% of control levels, compared to 66% in these experiments. Blood pressure was similarly better maintained. This appeared to be the most likely explanation for the decrease in renal blood flow noted in the present experiments. The microsphere distribution technique has the advantage of measuring flow throughout the body simultaneously. Disadvantages of the technique are apparent. Sampling of regional flow distribution can be made on only three occasions. Alterations in cardiac output and blood pressure may occur; presumably they are due to microsphere emboli changing cardiac action, as manifested by premature ventricular contractions, and peripheral resistance, as suggested by the change seen on one occasion when artificial heart action was provided by MVA. By slowly injecting the microspheres and limiting their number, the hemodynamic effect of the injection can be eliminated [7]. The 0.87 correlation coefficient between the microsphere and flowmeter techniques suggests that the microsphere method is satisfactory for determining regional blood flow distribution. Other experimental evidence for the validation of this method has been presented by Rudolph and Heymann [6], Neutze, Wyler, and Rudolph [5], Forsyth and his colleagues [3], and Kaihara and associates 141. While no one method of measuring regional blood flow is completely satisfactory, the values obtained in the present experiments correlate well with previous observations that the function of vital organs is well maintained during prolonged periods of MVA [9]. The values also indicate that this method for total circulatory support provides satisfactory flow to vital organs, although total cardiac output and blood pressure are somewhat diminished. REFERENCES 1. Anstadt, G. L., Schiff, P., and Baue, A. E. Prolonged circulatory support by direct mechanical ventricular assistance. Trans. Amer. SOC. Artif. Intern. Organs 12:72, Documenta Geigy, Scientific Tables (5th ed.). Basle, Switzerland: J. R. Geigy, Forsyth, R. P., Nies, A. S., Wyles, F., Neutze, J., and Melmon, K. L. Normal VOL. 1 1, NO. 1, JANUARY,

8 RACITI ET AL. distribution of cardiac output in the unanesthetized, restrained rhesus monkey. J. Appl. Physiol. 25:736, Kaihara, S., Van Heerden, P. D., Migita, T., and Wagner, H. N., Jr. Measurement of the distribution of cardiac output. J. Appl. Physiol. 25:696, Neutze, J. M., Wyler, R., and Rudolph, A. M. Use of radioactive microspheres to assess the distribution of cardiac output in rabbits. Amer. J. Physiol. 215:486, Rudolph, A. M., and Heymann, M. A. Methods for studying distribution of blood flow. Circ. Res. 21:163, Rutherford, R. B., Fukushima, T., Greenfield, A. J., Haalebos, M. M. P., and Zuidema, G. D. Acute and chronic effects of vagotomy on fasting and postprandial hepatosplanchnic blood flow and its distribution. Curr. Topics Surg. Res. 2:137, Skinner, D. B., Anstadt, G. L., and Camp, T. F., Jr. Applications of mechanical ventricular assistance. Ann. Surg. 166:500, Skinner, D. B., Newman, M. H., and Squire, R. B. Preservation and trans 10. plantation of organs maintained in vivo for 24 hours by mechanical ventricular assistance. J. Surg. Res. 10:253, Skinner, D. B., Raciti, A., Sabatier, H. L., Newman, M. N., Simon, A. L., and Conti, C. R. Surgery for acute myocardial infarction: Coronary flow and heart work during total circulatory support. Surgery 68: 128, THE ANNALS OF THORACIC SURGERY