- Email: [email protected]
Inhomogeneous and complementary antegrade and retrograde delivery of cardioplegic solution in the absence of coronary artery obstruction
Inhomogeneous and complementary antegrade and retrograde delivery of cardioplegic solution in the absence of coronary artery obstruction Inhomogeneous delivery of cardioplegic solution may result in postischemic myocardial injury. This study compares the distribution of warm blood antegrade and retrograde cardioplegia to multiple discrete left ventricular myocardial regions in pigs with unobstructed coronary arteries. Cardioplegic solution was delivered antegradely and retrogradely at 150 ml/min, and flows to 1152 individual myocardial regions were determined twice for each route with four different radiolabeled microspheres. The antegrade system delivered greater flow to each gram of myocardium than did the retrograde system (1.37 ± 0.31 versus 0.39 ± 0.09 ml/gm per minute, p < 0.001). Flow to individual myocardial regions was significantly inhomogeneous for both antegrade and retrograde cardioplegia, but much more so for retrograde cardioplegia (coefficientof variation was 48 % ± 17% for antegrade cardioplegia and 106% ± 16% for retrograde cardioplegia; p < 0.001). The pattern of flow to individual myocardial regions was highly reproducible for a given route of delivery as confirmed by repeated measurements with different radioactive microsphere isotopes (correlation coefficients 0.88 ± 0.12 for ACt - AC2 and 0.84 ± 0.10 RC t - RC2), but antegrade cardioplegia and retrograde cardioplegia patterns were significantly different and therefore complementary (correlation coefficients 0.03 ± 0.04, P < 0.001). These findings support the routine combined use of antegrade cardioplegia and retrograde cardioplegia to enhance delivery of cardioplegic solution to aU regions of the heart and minimize the potential risk of postischemic myocardial dysfunction. (J THORAC CARDlOVASC SURG 1994;107:499504)
Gabriel S. Aldea, MD (by invitation), Di Hou, MD (by invitation), James D. Fonger, MD (by invitation), and Richard J. Shemin, MD, Boston, Mass.
Qtimal myocardial protection relies on adequate delivery of cardiopIegic solution to all areas of the heart. Despitetheir universal clinical use, patterns of delivery of antegrade (AC) and retrograde (RC) cardioplegia to small discrete left ventricular (LV) myocardial regions remain largely unknown. Understanding these patterns may result in improved strategies to enhance myocardial protection. We propose the following experimental hypotheses, which are based on previous work!": (1) the From the Department of Cardiothoracic Surgery, Boston University Medical Center, Boston, Mass. Read at the Seventy-third Annual Meeting of The American Association for Thoracic Surgery, Chicago, Ill., April 25-28, 1993. Addressfor reprints: Gabriel S. Aldea, MD, Department of Cardiothoracic Surgery, B 402, Boston University Medical Center, 88 East Newton St., Boston, MA 02118-2393. Copyright
1994 by Mosby-Year Book, Inc.
0022-5223/94 $1.00 +.10
12/6/51198
distribution of cardioplegic solution may differ with the route (AC or RC) of delivery, (2) delivery of cardioplegic solution to small myocardial regions may be significantly inhomogeneous even in the absence of coronary artery disease, and (3) patterns of AC and RC may be mismatched. To investigate these hypotheses, we designed an experimental model to study the distribution of cardioplegic solution in adult pigs with unobstructed coronary arteries using radioactive microsphere methodology. Materials and methods General preparation. All animals received humane treatment in accordance with the "Principles of Laboratory Animal Care" formulated by the National Society for Medical Research and the "Guide for the Care and Use of Laboratory Animals" prepared by the Institute of Laboratory Animal Resources and published by the National Institutes of Health (NIH Publication No. 86-23, revised 1985). Six adult pigs 499
5 00
The Journal of Thoracic and Cardiovascular Surgery February 1994
Aldea et al.
(mean weight 45.5 kg) were premedicated with intramuscular morphine sulfate (2 mg/kg), anesthetized with alpha-chloralose (75 mg/kg), and supported with positive-pressure ventilation. A median sternotomy was performed. After anticoagulation with heparin (10,000 U intravenously, activated clotting time> 600 seconds), the animals were supported by normothermic cardiopulmonary bypass (37° C) via the femoral artery and bicaval cannulation, and a warm antegrade (37° C) blood cardioplegic arrest (500 m!) was induced. Regional flow determination with radioactive microspheres. Flow distribution of cardioplegic solution was determined twice for each route (AC and RC) by injecting microspheres 15 /-Lm in diameter (New England Nuclear, Boston, Mass.) and containing one of four gamma-emitting nuclides (I53GD, 113S n, 57CO, and 46Sc). Two million microsphereswere injected over 90 seconds at each experimental condition, resulting in a minimal entrapment of 400 spheres per region studied. Care was taken to avoid microsphere aggregation with suspension in saline solution with surfactant (0.01 % Tween 80; ICI Americas, Wilmington, Del.) and vigorous vortex mixing for 5 minutes. After a steady state of flow of cardioplegic solution was achieved for each experimental condition, the microspheres were injected approximately 80 em proximal to the heart to allow adequate mixing within the delivery system, with a method previously described," Blood was collected continuously during and for 2 minutes after the microsphere injection from a dedica ted distal port of the cardioplegic solution delivery system. Previous methods rely on a timed collection of blood distal to the delivery system and measurements of right atrial and right ventricular blood (during retrograde delivery of cardioplegic solution) to serve as a reference samples for calculation of regional flow delivery." By collecting the reference sample from the distal delivery tubing itself, we eliminated the potential for sampling errors" and were able to use a single exact reference sample, thereby enhancing the accuracy of regional flow measurements. The timed withdrawal of blood cardioplegic solution provided a microsphere-containing reference sample collected at a known flow rate. This allowed calculation of the absolute delivery of cardioplegic solution to any tissue sample with the following relations: Csample Q sample = Q reference X -C--reference
where Qsample is the delivery of cardioplegic solution to the tissue sample, Qreference is the delivery to the reference sample, Csample is the radioactivity of the tissue sample, and Creference is the radioactivity of the reference blood sample. At the conclusion of each experiment, the animal was killed with pentobarbital, and the heart was removed. The LV free wall and septum were excised, weighed, flattened, and placed in 4% formalin for 5 days. The tissue samples were then sectioned into subendocardial, midwall, and subepicardial layers. Each layer was further subdivided into 64 regions (0.5 ern? and 0.6 gm, on average) in a manner that maintained their anatomic relationship. Radionuclide emissions for calculation of regional flow distributions of cardioplegic solution were counted as previously described by Baer and associates? with an NaI (Tl ) gamma-scintillation multichannel counter (Canberra Series 35 Plus model No. 4261; Canberra Industries Inc., Meriden, Conn.), a multichannel pulse-height analyzer (Norland Corp., Fort Atkinson, Wis.), and a NOVA-3 minicomputer (Data
General, Boston, Mass.). All samples were counted for 3 minutes. The measured activity of each tissue sample was corrected for the differences in geometric configuration by applying an empirically derived relation (least-square linear fit) between sample weight and the measurement of known activities. All blood samples were the same height and were counted with the use of a calibration based on reference samples placed at the bottom of sample vials. Experimental protocol. After induction of warm blood AC arrest (37° C), cardioplegic solution was administered with a centrifugal pump through a vented multiple perfusion set. The cardioplegic solution was administered antegradely through a standard 16 gauge cannula and retrogradely through a 15F cannula with an autoinflating cuff (DLP Co., Grand Rapids, Mich.). Aortic root and coronary sinus pressures were simultaneously recorded, and cardioplegic solution flow rates were measured with an in-line electromagnetic flowmeter (Carolina Medical model70lD, series 600 probe, Carolina Medical Electronics, Inc., King, N.C.). During retrograde delivery of cardioplegic solution the caval cannulas were snared, and the right atrium was vented (opened). Cardioplegic solution was delivered antegradely and retrogradely at 150 nil/min. In six hearts, the delivery of cardioplegic solution to 1152 discrete myocardial regions was determined twice for each route of delivery with one of four different radioactive microspheres for a total of 4608 regional flow measurements. Previous work reported significant dislodgment, caused by AC, of radiolabeled microspheres entrapped within the myocardiam with RC, leading to serious errors in regional flow measurements." RC followed AC in all our animals in the experiment to prevent this problem. The sequence of radioactive isotope injection was randomized in each animal. Validation studies. In four preliminary experimental animals, complete coronary sinus occlusion with R C by the balloon was visually confirmed through the open right atrium. No coronary sinus back flow was noted, although thebesian vein flow was appreciated. In the same animals, we determined for each flow condition the proper number of injected radioactive microspheres to result in a minimal entrapment of 400 spheres in greater than 95% of all myocardial regions. All hearts evaluated in this study fullfilled these criteria. Proper mixing of microspheres and reproducibility of the technique were verified by sequentially injecting two different radiolabeled microspheres for each route of cardioplegic solution administration. These repeated injections are referred to as AC 1, AC2, RC), and RC2. For both AC and RC, the correlation coefficients (R2) were greater than 0.84. Consequently, errors resulting from clumping of microspheres, separation, calculation, and random distribution unrelated to flow could be excluded. We also excluded the possibility that the differences between AC and RC in patterns of delivery of cardioplegic solution to discrete myocardial regions were artificially caused by dislodgment of previously entrapped (by AC) radioactive microspheres; this was done by collecting the aortic effluent during retrograde delivery of cardioplegic solution, measuring its radioactive content, and comparing it with that of the LV. We found that less than 0.14% ± 0.24% of LV radioactivity was recovered in the aortic effluent during retrograde delivery of cardioplegic solution, which is too small a value to account for the significant differences in the patterns of cardioplegic solution flow between AC and RC.
The Journal of Thoracic and Cardiovascular Surgery
Aldea et al.
Volume 107, Number 2
l
1.5
125
c
c 100 !!
E 1.0
;: 75
I
'0
~ 0.5
II
E
50 1
~
!?
C
50
~
ii:
CD 25
o
o
A 0.0
ANTEGRADE
RETROGRADE
B
0 ANTEGRADE RETROGRADE
Fig. 1. Global LV flow (A) and inhomogeneity of flow ofcardioplegic solution to discrete myocardial regions (B) in six experimental animals. Inhomogeneity of regional flow is quatified by the coefficient of variation. *p < 0.001. Finally, toexclude the possible effects ofedema ontheexperimental findings, we minimized times for cardiopulmonary bypass andcardioplegic solution delivery toless than 15minutes in allouranimals in the experiment. Myocardial watercontent was evaluated in six control animals (no cardiopulmonary bypass or cardioplegic arrest)andcompared with oursixexperimental animals after cardioplegic arrest. Water content was measured bydessicating myocardial tissue ina vacuum oven for 72 hours." No differences were noted, and myocardial water content in the experimental animals waswithin 2% of matched control values (81.3% ± 1.3%). Numerical methods. Allcomparisons were madeby repeated measures analysis of variance and Newman-Keuls a posteriori test of significance. Statistical comparisons of the coefficients of variation (the ratio of the standard deviation of a distribution ofmeasurements toits mean) were performed with the technique of Lewontin, as described by Zar.lO R2 refers to a "simple" or Pearson product-moment correlation coefficient. All mean values are given ± one standarddeviation. Results Cardioplegia perfusion pressure. Cardioplegic solution was delivered antegradely and retrogradely at 150 ml/rnin, The mean cardioplegia perfusion pressure in our six experimental animals was 72.5 ± 7.0 mm Hg for AC and 24.7 ± 6.4 mm Hg for RC (p < 0.001). Flow. The mean LV flow for all six experimental animals was 1.37 ± 0.31 ml/gm per minute for AC and 0.39 ± 0.09 ml/gm per minute for RC (Fig. I, A). This difference, greater than threefold, between routes of administration was highly significant (p < 0.001). Inhomogeneity. Patterns of flow of cardioplegic solution to discrete myocardial regions in a representative experimental animal (No.4) are demonstrated in Fig. 2. Flow of cardioplegic solution for both routes of delivery was uneven (inhomogeneous), with some regions receiving flows sixfold to tenfold higher than others. This unevenor inhomogeneous flow of cardioplegic solution to specific myocardial regions was further characterized by
the coefficientof variation at the percentage of which is the standard deviation/mean X 100. The mean coefficient of variation for all six experimental animals was 48% ± 17%for AC and 106% ± 16%for RC (Fig. 1, B). This difference was also highly significant (p < 0.001). Correlation. Flows to individual myocardial regions measured with different radioactive micropshere labels were compared. With the use of linear regression analysis, each regional flow was compared with itself under different experimental conditions, and R 2 was obtained. Correlations of flowin a representative experimental animal (No.4) are shown in Fig. 3. AC and RC flow patterns measured with different radioactive isotopes were highly correlated (R2; 0.88 for AC I - AC 2 and 0.81 for RC I - RC 2). Thus, a region with a high flow during a particular route of cardioplegic solution delivery maintained its characteristic high flowwhen the measurement was repeated for the same route. However AC and RC patterns were mismatched (R 2 = 0.04). Thus, a region with low flow during antegrade delivery of cardioplegic solution had a different flow profile using retrograde delivery. These findings were consistent in all six experimental animals. R2 was 0.88 ± 0.12 for AC I - AC 2, 0.84 ± 10 for RC I - RC2, and 0.03 ± 0.04 when AC and RC were compared (Fig. 4). These highly significant differences (p < 0.001) suggest that although patterns of flow of cardioplegic solution are highly reproducible for a given route of administration, no correlation exists between AC and RC flow patterns, which confirms that they are mismatched. Discussion We designed an experimental model to study patterns of antegrade and retrograde delivery of normothermic blood cardioplegic solution to discrete myocardial regions. We specificallychose to study the regional distri-
The Journal of Tho racic and Cardiovascular Surgery February 1994
5 0 2 A ldea et al.
RETROGRADE
ANTEGRADE 16 14
12
12
10
10
~
u
4
0
1.5 lV Flow
0
2.5
2.5
Fig. 2. Patterns of antegrade and retrograde deliveryof cardioplegic solution to discrete myocardial regions in a singleanimal (No.4). The number of regions (count) with a characteristicflow (rnl/grn per minute) are demonstrated by this frequency histogram. 1.0
1.0
c-
E
E
!!
:! .... •;;
..
•· ;:
ACn. AC
A' • • •
~
0.1
E !!
...... ..e
:!
0.6 0.4 0.2
1.0
FlC n.FlC Fl' . 11.11
C
C
0.1
E E
0.1
0.8
:!
O.ll
!!
l!
..
;
II:
•
0.4
l! l!
OJ
II:
O ·~ .o
0.2
0.4
Anl eg. - - l
0.1
CmII""",-,
0.'
ACn. FlC Fl'.CUM
0
0.4
~ 0.2
..... .....-. 0
02
0.4
0.1
' CmII. ..,-,
0 .1
1.0
0.0
0
oJ'
0 .01-..:....~_i4J~~~'-
1.0
..-. .;1. •
0.5
\
~
~.
0 0 0
1.0
........ C- . . . .,
--.
1.5
2.0
Fig. 3. Comparison of correlation of flows to discrete myocardial regionsin a singleanimal (No.4). R 2 is the correlation coefficient.
1.0
N a:
c
GI
U :;:
GI
0
0
0.5
c 0
..
iii
*
GI
0
0
0.0
AC -A C
RC·RC
AC -RC
Fig. 4. Comparison of mean correlation coefficients of flow to discrete myocardial regions for all six experimental animals. *p
< 0.001
bution of warm blood cardioplegic solution (37° C) to eliminate the potential vasoconstrictive effects of a cold solution, whic h may further accentuate differences (inh omogeneity) in regional flows.I I. 12 Although this work centered on the study of blood cardioplegic solution, our
previous work demonstrated similar results when patterns of blood and crystalloid cardioplegic solution flows were compared.' Duration for cardiopulmonary bypass and delivery of cardioplegic solution were minimized to avoid the confounding effects of increasing edema on LV complian ce and regional cardioplegic solutio n flow distribution.l ' We conclude that the pattern of delivery of cardioplegic solution to small myocardial regions differ with the route of administration. For a given flow rate, AC delivers significantly more flow per gram of myocardium than does R C (Fig. I, A). This corroborates similar findings by Gates and assoc iates'< 15 of differences in AC and RC flow delivery in both piglet and explanted human hearts. Distribution of cardioplegic solution to individual myo cardial regions is inhomogeneous for both AC and R C. This inhomogeneity was greatest with RC (Fig. I , B). This profound inhomogeneity makes comparisons of endocardial/epicardial flow ratios (which imply homogenei ty of flow within each layer) less meaningful." Differences in the regional delivery of cardioplegic solution did not correlate with the gross coronary vascular anatomy (proximity to epicardial vessels) and varied among individual animals. Patterns of regional flow of cardioplegic solution to discrete myocardial regions were
The Journal of Thoracic and Cardiovascular Surgery Volume 107, Number 2
highly reproducible for a given route of delivery (Figs. 3 and 4). However, regional flow patterns with AC were significantly different compared with those with RC. Mismatched patterns of AC and RC solution delivery suggest that these modalities are complementary. Even in the absence of coronary artery obstruction, some regions received a flow of cardioplegic solution six to ten times higher than did others. This inhomogeneous delivery was previously noted to be most accentuated in the subendocardium.l" perhaps explaining the patchy nature of subendocardial infarction noted after ischemic revascularization.lvl'' Inhomogeneous delivery of cardioplegic solution to discrete myocardial regions would be further accentuated by the presence of coronary artery obstructions.P Although inhomogeneous delivery of cardioplegic solution may be sufficient to result in asynchronous electrical activity and an isoelectric electrocardiogram, individual regions may have ongoing metabolic activity that may result in a postischemic injury. Furthermore, the patchy nature of such damage may go undetected by global measurements of ventricular "wellbeing," such as ventricular function, 20, 21 because hypercontactile well-preserved areas can compensate for small areas of regional dysfunction.V Thus, preserved ventricular function and a normal electrocardiogram tracing after myocardial revascularization do not exclude limited regional myocardial damage. Optimal delivery of cardioplegic solution to all myocardial regions is particularly important in the presence of compromised ventricular function.I: 23 Reliance on a single route of delivery cardioplegic solution may result in inadequate regional myocardial protection, Because the minimal amount of cardioplegic solution necessary to establish a complete regional metabolic arrest is unknown, a generous amount of cardioplegic solution has to be administered to best assure protection of the entire myocardium and overcome the cardioplegic solution flow disparity to small myocardial regions. However, smaller volumes of cardioplegic solution could be used if AC and RC were used as complementary techniques to enhance the regional delivery of cardioplegic solution. Our findings may explain the experimental and clinical observations of the salutary effect of combining the use of A C and RC in patients with compromized LV function. 23, 24 We recommend the routine combined use of AC and RC, on the basis of our experimental results, to optimize delivery of cardioplegic solution to all regions of the heart and minimize the potential of postischemic myocardial dysfunction. REFERENCES I. Buckberg GD, Stategies and logic of cardioplegia delivery to prevent, avoid, and reverse ischemic and reperfu-
Aldea et al.
50 3
sion damage, J THORAC CARDIOVASC SURG 1987;93:12739, 2, Austin RE Jr, Aldea GS, Coggins DL, et al. Profound spatial heterogeneity of coronary reserve, discordance between patterns of resting and maximal myocardial flow, Circ Res 1990;67:310-31. 3. Coggins DL, Flynn AE, Austin RE Jr, et aL Nonuniform loss of regional flow reserve during myocardial ischemia in dogs, Circ Res 1990;67:253-64. 4, Aldea GS, Austin RE Jr, Flynn AE, et al. Heterogeous delivery of cardioplegic solution in the absence of coronary artery disease. J THORAC CARDIOVASC SURG 1990;99:34553. 5. Franzen D, Conway RS, Zhang H, et al. Spatial heterogeneity of local blood flow and metabolite content in dog hearts, Am J Physiol 1988;254:H344-53, 6, Austin RE Jr, Hauck WW, Aldea GS, et aL Quantitating error in blood flow measurements with radioactive microspheres. Am J Physiol 1989;257:H 1523-33. 7. Partington MT, Acar C, Buckberg GD, et al, Studies of retrograde cardioplegia, L Capillary blood flowdistribution to the myocardium supplied with open and occluded arteries. J THORAC CARDIOVASC SURG 1989;97:605-12. 8. Heymann MA, Payne BD, Hoffman JIE, et al. Blood flow measurements with radionuclide-Iabeled particles. Prog Cardiovasc Dis 1977;20:55-79, 9. Baer RW, Payne BD, Verrier ED, et al. Increased number of myocardial blood flow measurements with radionuclidelabeled microspheres. Am J Phsiol 1984;246:H418-34, 10. Zar GH, Biostatistical analysis. 2nd ed. Engelwood Cliffs, N,J.: Prentice-Hall, Inc., 1984:122-6. II, Archie JP, Kirklin JW. Effect of hypothermia on O 2 consumption and coronary resistance. Surg Forum 1973; 24:186-8. 12. McDonell DH, Brazier JR, Cooper N, et al. Studies on the effects of hypothermia and regional myocardial blood flow. II. Ischemia during moderate hypothermia in continually perfused dogs. J THORAC CARDIOVASC SURG 1977;73:95101. 13. Follette D, Key K, Mulder D, et al. Prolonged safe aortic clamping by combining membrane stabilization, multidose cardioplegia, and appropriate pH reperfusion. J THORAC CARDIOVASC SURG 1977;74:683-94. 14. Gates RN, Laks H, Drinkwater DC, et al. The microvascular distribution of cardioplegic distribution in the piglet heart: retrograde versus antegrade delivery. J THORAC CARDIOVASC SURG 1993;105:845-53. 15. Gates RN, Laks H, Drinkwater DC, et al. Gross and microvascular distribution of retrograde cardioplegia in explanted human hearts. The 29th meeting of the Society of Thoracic Surgeons. Ann Thorac Surg 1993;56:410-7. 16. Bulkley BH, Hutchins GM. Pathology of coronary artery bypass surgery. Arch Pathol Lab Med 1978;102:27380. 17. Levine HD. Subendocardial infarction in retrospect: pathologic, cardiographic and ancillary features. Circulation 1985;72:790-800. 18. Buckberg GD, Fixler DE, Archie JP, et aL Experimental
504
19.
20.
21.
22.
23.
24.
Aldea et al.
subendocardial ischemia in dogs with normal coronary arteries. Circ Res 1972;30:67-81. Heinman FW, MacGregor DC, Wilson GJ, et al. Regional and transmural myocardial flow distribution in cold chemical cardioplegia: significance of critical artery stenoses. J THORAC CARDIOVASC SURG 1981;81:851-9. Flameng W, Borgers M, Dansnen W, et al. Ultrastructural and cytochemical correlates of myocardial protection by cardiac hypothermia in man. J THORAC CARDIOVASC SURG 1980;79:413-24. Rosenkrantz ER, Okamoto F, Buckberg GD, et al. Studies on controlled reperfusion after ischemia. II. Biochemical studies: failure of A TP levels to predict recovery of contractile function after controlled reperfusion. J THORAC CARDIOVASC SURG 1986;92:502-12. Lew WYW, Chen Z, Guth B, et al. Mechanisms of augmented shortening in nonischemic areas during ischemia in the canine left ventricle. Circ Res 1985;56:351-8. Misare BD, Krukenkamp IB, Lazer ZP, et al. Retrograde is superior to antegrade continuous warm blood cardioplegia for acute cardiac ischemia. Circulation 1992;86(Suppl):II393-7. Partington MT, Acar C, Buckberg GD, et al. Studies of retrograde cardioplegia. II. Advantages of antegrade/retrograde cardioplegia to optimize distribution in jeopardized myocardium. J THORAC CARDIOVASC SURG 1989;97:61322.
Discussion Dr. John H. Kennedy (Cambridge, England). Would you continue to classify what you are measuring as flow per gram of heart as such if there were spherules appearing in the coronary sinus after antegrade injection? Dr. Davis C. Drinkwater, Jr. (Los Angeles, Calif). I have one question about the technical aspect of RC. When RC is used clinically, it may result in much effluent and reflux back out of the coronary sinus and therefore may account for a considerable proportion of the total volume delivered. In your model, did you have a technique to occlude the coronary sinus totally during RC to avoid this leakage which might effect your findings? Dr. William D. Spotnitz (Charlottesville, Va.). I would like to support the data that you have presented. Recent data from our laboratory, presented at the American College of Surgeons,
The Journal of Thoracic and Cardiovascular Surgery February 1994
confirm the results that you have presented. I would also like to point out that if you use a metabolized product such as thallium, which can be taken up by the cells, you actually have an even greater discrepancy between AC and RC. I think radiolabeled microspheres as a method of measuring this discrepancy are problematic because they do not measure the loss of retrograde flow that occurs by collaterals, which are smaller then 10 !Lm. So the discrepancy that you have between AC and RC is even greater than that which you have already shown. I would like to know how, in aortic valve replacement or mitral valve replacement, when there is this large discrepancy, you would achieve good myocardial preservation with retrograde flow. In our own data, we found the same discrepancy that you have shown, in fact, an even greater one. However, the retrograde cooling was at least equivalent to that of antegrade and probably reflects many of the local effects of the transmission of cold solution. Dr. Aldea. To answer Dr. Kennedy's question, the entrapment rate with the antegrade route was 95% or more. The ventricles were vented, and therefore the microspheres were not recirculated. With the retrograde route, we had bicaval cannulation, and the left atrium was also vented; again, it was a single-shot effect. We were very careful to measure the effluent from the aortic root for radioactivity during retrograde injection to make sure that we did not dislodge any entrapped antegrade miscrospheres, and the average dislodgment of microspheres during retrograde injection was less than 0.2%. This rate would not account, therefore, for the measurements that we have made. I am familiar with the work at the University of California, Los Angeles. The measurements that we made were strictly cardioplegic solution flow delivery. We understand that there are some secondary effects of temperature and cooling. In this experiment, we were particulary interested in whether the patterns of cardioplegic solution delivery were matched. We did not measure infarct size or the actual protection. To answer the last question of Dr. Spotnitz, I think that we may sometimes be too cavalier when we assume that when we have good preservation of ventricular function at the end of an operation, as seen with an isolectric electrocardiogram, that all areas of the heart were truly protected. I think this work suggests that some small areas of the heart are not protected, even under the best of circumstances and with the use of global measurements such as LV function, electrocardiographic activity, or global function, such as cardiac output, may not particularly reflect these regional differences and sensitivities. These discrepancies may be particularly important when we are dealing with impaired LVs.
Gabriel S. Aldea, MD (by invitation), Di Hou, MD (by invitation), James D. Fonger, MD (by invitation), and Richard J. Shemin, MD, Boston, Mass.
Qtimal myocardial protection relies on adequate delivery of cardiopIegic solution to all areas of the heart. Despitetheir universal clinical use, patterns of delivery of antegrade (AC) and retrograde (RC) cardioplegia to small discrete left ventricular (LV) myocardial regions remain largely unknown. Understanding these patterns may result in improved strategies to enhance myocardial protection. We propose the following experimental hypotheses, which are based on previous work!": (1) the From the Department of Cardiothoracic Surgery, Boston University Medical Center, Boston, Mass. Read at the Seventy-third Annual Meeting of The American Association for Thoracic Surgery, Chicago, Ill., April 25-28, 1993. Addressfor reprints: Gabriel S. Aldea, MD, Department of Cardiothoracic Surgery, B 402, Boston University Medical Center, 88 East Newton St., Boston, MA 02118-2393. Copyright
1994 by Mosby-Year Book, Inc.
0022-5223/94 $1.00 +.10
12/6/51198
distribution of cardioplegic solution may differ with the route (AC or RC) of delivery, (2) delivery of cardioplegic solution to small myocardial regions may be significantly inhomogeneous even in the absence of coronary artery disease, and (3) patterns of AC and RC may be mismatched. To investigate these hypotheses, we designed an experimental model to study the distribution of cardioplegic solution in adult pigs with unobstructed coronary arteries using radioactive microsphere methodology. Materials and methods General preparation. All animals received humane treatment in accordance with the "Principles of Laboratory Animal Care" formulated by the National Society for Medical Research and the "Guide for the Care and Use of Laboratory Animals" prepared by the Institute of Laboratory Animal Resources and published by the National Institutes of Health (NIH Publication No. 86-23, revised 1985). Six adult pigs 499
5 00
The Journal of Thoracic and Cardiovascular Surgery February 1994
Aldea et al.
(mean weight 45.5 kg) were premedicated with intramuscular morphine sulfate (2 mg/kg), anesthetized with alpha-chloralose (75 mg/kg), and supported with positive-pressure ventilation. A median sternotomy was performed. After anticoagulation with heparin (10,000 U intravenously, activated clotting time> 600 seconds), the animals were supported by normothermic cardiopulmonary bypass (37° C) via the femoral artery and bicaval cannulation, and a warm antegrade (37° C) blood cardioplegic arrest (500 m!) was induced. Regional flow determination with radioactive microspheres. Flow distribution of cardioplegic solution was determined twice for each route (AC and RC) by injecting microspheres 15 /-Lm in diameter (New England Nuclear, Boston, Mass.) and containing one of four gamma-emitting nuclides (I53GD, 113S n, 57CO, and 46Sc). Two million microsphereswere injected over 90 seconds at each experimental condition, resulting in a minimal entrapment of 400 spheres per region studied. Care was taken to avoid microsphere aggregation with suspension in saline solution with surfactant (0.01 % Tween 80; ICI Americas, Wilmington, Del.) and vigorous vortex mixing for 5 minutes. After a steady state of flow of cardioplegic solution was achieved for each experimental condition, the microspheres were injected approximately 80 em proximal to the heart to allow adequate mixing within the delivery system, with a method previously described," Blood was collected continuously during and for 2 minutes after the microsphere injection from a dedica ted distal port of the cardioplegic solution delivery system. Previous methods rely on a timed collection of blood distal to the delivery system and measurements of right atrial and right ventricular blood (during retrograde delivery of cardioplegic solution) to serve as a reference samples for calculation of regional flow delivery." By collecting the reference sample from the distal delivery tubing itself, we eliminated the potential for sampling errors" and were able to use a single exact reference sample, thereby enhancing the accuracy of regional flow measurements. The timed withdrawal of blood cardioplegic solution provided a microsphere-containing reference sample collected at a known flow rate. This allowed calculation of the absolute delivery of cardioplegic solution to any tissue sample with the following relations: Csample Q sample = Q reference X -C--reference
where Qsample is the delivery of cardioplegic solution to the tissue sample, Qreference is the delivery to the reference sample, Csample is the radioactivity of the tissue sample, and Creference is the radioactivity of the reference blood sample. At the conclusion of each experiment, the animal was killed with pentobarbital, and the heart was removed. The LV free wall and septum were excised, weighed, flattened, and placed in 4% formalin for 5 days. The tissue samples were then sectioned into subendocardial, midwall, and subepicardial layers. Each layer was further subdivided into 64 regions (0.5 ern? and 0.6 gm, on average) in a manner that maintained their anatomic relationship. Radionuclide emissions for calculation of regional flow distributions of cardioplegic solution were counted as previously described by Baer and associates? with an NaI (Tl ) gamma-scintillation multichannel counter (Canberra Series 35 Plus model No. 4261; Canberra Industries Inc., Meriden, Conn.), a multichannel pulse-height analyzer (Norland Corp., Fort Atkinson, Wis.), and a NOVA-3 minicomputer (Data
General, Boston, Mass.). All samples were counted for 3 minutes. The measured activity of each tissue sample was corrected for the differences in geometric configuration by applying an empirically derived relation (least-square linear fit) between sample weight and the measurement of known activities. All blood samples were the same height and were counted with the use of a calibration based on reference samples placed at the bottom of sample vials. Experimental protocol. After induction of warm blood AC arrest (37° C), cardioplegic solution was administered with a centrifugal pump through a vented multiple perfusion set. The cardioplegic solution was administered antegradely through a standard 16 gauge cannula and retrogradely through a 15F cannula with an autoinflating cuff (DLP Co., Grand Rapids, Mich.). Aortic root and coronary sinus pressures were simultaneously recorded, and cardioplegic solution flow rates were measured with an in-line electromagnetic flowmeter (Carolina Medical model70lD, series 600 probe, Carolina Medical Electronics, Inc., King, N.C.). During retrograde delivery of cardioplegic solution the caval cannulas were snared, and the right atrium was vented (opened). Cardioplegic solution was delivered antegradely and retrogradely at 150 nil/min. In six hearts, the delivery of cardioplegic solution to 1152 discrete myocardial regions was determined twice for each route of delivery with one of four different radioactive microspheres for a total of 4608 regional flow measurements. Previous work reported significant dislodgment, caused by AC, of radiolabeled microspheres entrapped within the myocardiam with RC, leading to serious errors in regional flow measurements." RC followed AC in all our animals in the experiment to prevent this problem. The sequence of radioactive isotope injection was randomized in each animal. Validation studies. In four preliminary experimental animals, complete coronary sinus occlusion with R C by the balloon was visually confirmed through the open right atrium. No coronary sinus back flow was noted, although thebesian vein flow was appreciated. In the same animals, we determined for each flow condition the proper number of injected radioactive microspheres to result in a minimal entrapment of 400 spheres in greater than 95% of all myocardial regions. All hearts evaluated in this study fullfilled these criteria. Proper mixing of microspheres and reproducibility of the technique were verified by sequentially injecting two different radiolabeled microspheres for each route of cardioplegic solution administration. These repeated injections are referred to as AC 1, AC2, RC), and RC2. For both AC and RC, the correlation coefficients (R2) were greater than 0.84. Consequently, errors resulting from clumping of microspheres, separation, calculation, and random distribution unrelated to flow could be excluded. We also excluded the possibility that the differences between AC and RC in patterns of delivery of cardioplegic solution to discrete myocardial regions were artificially caused by dislodgment of previously entrapped (by AC) radioactive microspheres; this was done by collecting the aortic effluent during retrograde delivery of cardioplegic solution, measuring its radioactive content, and comparing it with that of the LV. We found that less than 0.14% ± 0.24% of LV radioactivity was recovered in the aortic effluent during retrograde delivery of cardioplegic solution, which is too small a value to account for the significant differences in the patterns of cardioplegic solution flow between AC and RC.
The Journal of Thoracic and Cardiovascular Surgery
Aldea et al.
Volume 107, Number 2
l
1.5
125
c
c 100 !!
E 1.0
;: 75
I
'0
~ 0.5
II
E
50 1
~
!?
C
50
~
ii:
CD 25
o
o
A 0.0
ANTEGRADE
RETROGRADE
B
0 ANTEGRADE RETROGRADE
Fig. 1. Global LV flow (A) and inhomogeneity of flow ofcardioplegic solution to discrete myocardial regions (B) in six experimental animals. Inhomogeneity of regional flow is quatified by the coefficient of variation. *p < 0.001. Finally, toexclude the possible effects ofedema ontheexperimental findings, we minimized times for cardiopulmonary bypass andcardioplegic solution delivery toless than 15minutes in allouranimals in the experiment. Myocardial watercontent was evaluated in six control animals (no cardiopulmonary bypass or cardioplegic arrest)andcompared with oursixexperimental animals after cardioplegic arrest. Water content was measured bydessicating myocardial tissue ina vacuum oven for 72 hours." No differences were noted, and myocardial water content in the experimental animals waswithin 2% of matched control values (81.3% ± 1.3%). Numerical methods. Allcomparisons were madeby repeated measures analysis of variance and Newman-Keuls a posteriori test of significance. Statistical comparisons of the coefficients of variation (the ratio of the standard deviation of a distribution ofmeasurements toits mean) were performed with the technique of Lewontin, as described by Zar.lO R2 refers to a "simple" or Pearson product-moment correlation coefficient. All mean values are given ± one standarddeviation. Results Cardioplegia perfusion pressure. Cardioplegic solution was delivered antegradely and retrogradely at 150 ml/rnin, The mean cardioplegia perfusion pressure in our six experimental animals was 72.5 ± 7.0 mm Hg for AC and 24.7 ± 6.4 mm Hg for RC (p < 0.001). Flow. The mean LV flow for all six experimental animals was 1.37 ± 0.31 ml/gm per minute for AC and 0.39 ± 0.09 ml/gm per minute for RC (Fig. I, A). This difference, greater than threefold, between routes of administration was highly significant (p < 0.001). Inhomogeneity. Patterns of flow of cardioplegic solution to discrete myocardial regions in a representative experimental animal (No.4) are demonstrated in Fig. 2. Flow of cardioplegic solution for both routes of delivery was uneven (inhomogeneous), with some regions receiving flows sixfold to tenfold higher than others. This unevenor inhomogeneous flow of cardioplegic solution to specific myocardial regions was further characterized by
the coefficientof variation at the percentage of which is the standard deviation/mean X 100. The mean coefficient of variation for all six experimental animals was 48% ± 17%for AC and 106% ± 16%for RC (Fig. 1, B). This difference was also highly significant (p < 0.001). Correlation. Flows to individual myocardial regions measured with different radioactive micropshere labels were compared. With the use of linear regression analysis, each regional flow was compared with itself under different experimental conditions, and R 2 was obtained. Correlations of flowin a representative experimental animal (No.4) are shown in Fig. 3. AC and RC flow patterns measured with different radioactive isotopes were highly correlated (R2; 0.88 for AC I - AC 2 and 0.81 for RC I - RC 2). Thus, a region with a high flow during a particular route of cardioplegic solution delivery maintained its characteristic high flowwhen the measurement was repeated for the same route. However AC and RC patterns were mismatched (R 2 = 0.04). Thus, a region with low flow during antegrade delivery of cardioplegic solution had a different flow profile using retrograde delivery. These findings were consistent in all six experimental animals. R2 was 0.88 ± 0.12 for AC I - AC 2, 0.84 ± 10 for RC I - RC2, and 0.03 ± 0.04 when AC and RC were compared (Fig. 4). These highly significant differences (p < 0.001) suggest that although patterns of flow of cardioplegic solution are highly reproducible for a given route of administration, no correlation exists between AC and RC flow patterns, which confirms that they are mismatched. Discussion We designed an experimental model to study patterns of antegrade and retrograde delivery of normothermic blood cardioplegic solution to discrete myocardial regions. We specificallychose to study the regional distri-
The Journal of Tho racic and Cardiovascular Surgery February 1994
5 0 2 A ldea et al.
RETROGRADE
ANTEGRADE 16 14
12
12
10
10
~
u
4
0
1.5 lV Flow
0
2.5
2.5
Fig. 2. Patterns of antegrade and retrograde deliveryof cardioplegic solution to discrete myocardial regions in a singleanimal (No.4). The number of regions (count) with a characteristicflow (rnl/grn per minute) are demonstrated by this frequency histogram. 1.0
1.0
c-
E
E
!!
:! .... •;;
..
•· ;:
ACn. AC
A' • • •
~
0.1
E !!
...... ..e
:!
0.6 0.4 0.2
1.0
FlC n.FlC Fl' . 11.11
C
C
0.1
E E
0.1
0.8
:!
O.ll
!!
l!
..
;
II:
•
0.4
l! l!
OJ
II:
O ·~ .o
0.2
0.4
Anl eg. - - l
0.1
CmII""",-,
0.'
ACn. FlC Fl'.CUM
0
0.4
~ 0.2
..... .....-. 0
02
0.4
0.1
' CmII. ..,-,
0 .1
1.0
0.0
0
oJ'
0 .01-..:....~_i4J~~~'-
1.0
..-. .;1. •
0.5
\
~
~.
0 0 0
1.0
........ C- . . . .,
--.
1.5
2.0
Fig. 3. Comparison of correlation of flows to discrete myocardial regionsin a singleanimal (No.4). R 2 is the correlation coefficient.
1.0
N a:
c
GI
U :;:
GI
0
0
0.5
c 0
..
iii
*
GI
0
0
0.0
AC -A C
RC·RC
AC -RC
Fig. 4. Comparison of mean correlation coefficients of flow to discrete myocardial regions for all six experimental animals. *p
< 0.001
bution of warm blood cardioplegic solution (37° C) to eliminate the potential vasoconstrictive effects of a cold solution, whic h may further accentuate differences (inh omogeneity) in regional flows.I I. 12 Although this work centered on the study of blood cardioplegic solution, our
previous work demonstrated similar results when patterns of blood and crystalloid cardioplegic solution flows were compared.' Duration for cardiopulmonary bypass and delivery of cardioplegic solution were minimized to avoid the confounding effects of increasing edema on LV complian ce and regional cardioplegic solutio n flow distribution.l ' We conclude that the pattern of delivery of cardioplegic solution to small myocardial regions differ with the route of administration. For a given flow rate, AC delivers significantly more flow per gram of myocardium than does R C (Fig. I, A). This corroborates similar findings by Gates and assoc iates'< 15 of differences in AC and RC flow delivery in both piglet and explanted human hearts. Distribution of cardioplegic solution to individual myo cardial regions is inhomogeneous for both AC and R C. This inhomogeneity was greatest with RC (Fig. I , B). This profound inhomogeneity makes comparisons of endocardial/epicardial flow ratios (which imply homogenei ty of flow within each layer) less meaningful." Differences in the regional delivery of cardioplegic solution did not correlate with the gross coronary vascular anatomy (proximity to epicardial vessels) and varied among individual animals. Patterns of regional flow of cardioplegic solution to discrete myocardial regions were
The Journal of Thoracic and Cardiovascular Surgery Volume 107, Number 2
highly reproducible for a given route of delivery (Figs. 3 and 4). However, regional flow patterns with AC were significantly different compared with those with RC. Mismatched patterns of AC and RC solution delivery suggest that these modalities are complementary. Even in the absence of coronary artery obstruction, some regions received a flow of cardioplegic solution six to ten times higher than did others. This inhomogeneous delivery was previously noted to be most accentuated in the subendocardium.l" perhaps explaining the patchy nature of subendocardial infarction noted after ischemic revascularization.lvl'' Inhomogeneous delivery of cardioplegic solution to discrete myocardial regions would be further accentuated by the presence of coronary artery obstructions.P Although inhomogeneous delivery of cardioplegic solution may be sufficient to result in asynchronous electrical activity and an isoelectric electrocardiogram, individual regions may have ongoing metabolic activity that may result in a postischemic injury. Furthermore, the patchy nature of such damage may go undetected by global measurements of ventricular "wellbeing," such as ventricular function, 20, 21 because hypercontactile well-preserved areas can compensate for small areas of regional dysfunction.V Thus, preserved ventricular function and a normal electrocardiogram tracing after myocardial revascularization do not exclude limited regional myocardial damage. Optimal delivery of cardioplegic solution to all myocardial regions is particularly important in the presence of compromised ventricular function.I: 23 Reliance on a single route of delivery cardioplegic solution may result in inadequate regional myocardial protection, Because the minimal amount of cardioplegic solution necessary to establish a complete regional metabolic arrest is unknown, a generous amount of cardioplegic solution has to be administered to best assure protection of the entire myocardium and overcome the cardioplegic solution flow disparity to small myocardial regions. However, smaller volumes of cardioplegic solution could be used if AC and RC were used as complementary techniques to enhance the regional delivery of cardioplegic solution. Our findings may explain the experimental and clinical observations of the salutary effect of combining the use of A C and RC in patients with compromized LV function. 23, 24 We recommend the routine combined use of AC and RC, on the basis of our experimental results, to optimize delivery of cardioplegic solution to all regions of the heart and minimize the potential of postischemic myocardial dysfunction. REFERENCES I. Buckberg GD, Stategies and logic of cardioplegia delivery to prevent, avoid, and reverse ischemic and reperfu-
Aldea et al.
50 3
sion damage, J THORAC CARDIOVASC SURG 1987;93:12739, 2, Austin RE Jr, Aldea GS, Coggins DL, et al. Profound spatial heterogeneity of coronary reserve, discordance between patterns of resting and maximal myocardial flow, Circ Res 1990;67:310-31. 3. Coggins DL, Flynn AE, Austin RE Jr, et aL Nonuniform loss of regional flow reserve during myocardial ischemia in dogs, Circ Res 1990;67:253-64. 4, Aldea GS, Austin RE Jr, Flynn AE, et al. Heterogeous delivery of cardioplegic solution in the absence of coronary artery disease. J THORAC CARDIOVASC SURG 1990;99:34553. 5. Franzen D, Conway RS, Zhang H, et al. Spatial heterogeneity of local blood flow and metabolite content in dog hearts, Am J Physiol 1988;254:H344-53, 6, Austin RE Jr, Hauck WW, Aldea GS, et aL Quantitating error in blood flow measurements with radioactive microspheres. Am J Physiol 1989;257:H 1523-33. 7. Partington MT, Acar C, Buckberg GD, et al, Studies of retrograde cardioplegia, L Capillary blood flowdistribution to the myocardium supplied with open and occluded arteries. J THORAC CARDIOVASC SURG 1989;97:605-12. 8. Heymann MA, Payne BD, Hoffman JIE, et al. Blood flow measurements with radionuclide-Iabeled particles. Prog Cardiovasc Dis 1977;20:55-79, 9. Baer RW, Payne BD, Verrier ED, et al. Increased number of myocardial blood flow measurements with radionuclidelabeled microspheres. Am J Phsiol 1984;246:H418-34, 10. Zar GH, Biostatistical analysis. 2nd ed. Engelwood Cliffs, N,J.: Prentice-Hall, Inc., 1984:122-6. II, Archie JP, Kirklin JW. Effect of hypothermia on O 2 consumption and coronary resistance. Surg Forum 1973; 24:186-8. 12. McDonell DH, Brazier JR, Cooper N, et al. Studies on the effects of hypothermia and regional myocardial blood flow. II. Ischemia during moderate hypothermia in continually perfused dogs. J THORAC CARDIOVASC SURG 1977;73:95101. 13. Follette D, Key K, Mulder D, et al. Prolonged safe aortic clamping by combining membrane stabilization, multidose cardioplegia, and appropriate pH reperfusion. J THORAC CARDIOVASC SURG 1977;74:683-94. 14. Gates RN, Laks H, Drinkwater DC, et al. The microvascular distribution of cardioplegic distribution in the piglet heart: retrograde versus antegrade delivery. J THORAC CARDIOVASC SURG 1993;105:845-53. 15. Gates RN, Laks H, Drinkwater DC, et al. Gross and microvascular distribution of retrograde cardioplegia in explanted human hearts. The 29th meeting of the Society of Thoracic Surgeons. Ann Thorac Surg 1993;56:410-7. 16. Bulkley BH, Hutchins GM. Pathology of coronary artery bypass surgery. Arch Pathol Lab Med 1978;102:27380. 17. Levine HD. Subendocardial infarction in retrospect: pathologic, cardiographic and ancillary features. Circulation 1985;72:790-800. 18. Buckberg GD, Fixler DE, Archie JP, et aL Experimental
504
19.
20.
21.
22.
23.
24.
Aldea et al.
subendocardial ischemia in dogs with normal coronary arteries. Circ Res 1972;30:67-81. Heinman FW, MacGregor DC, Wilson GJ, et al. Regional and transmural myocardial flow distribution in cold chemical cardioplegia: significance of critical artery stenoses. J THORAC CARDIOVASC SURG 1981;81:851-9. Flameng W, Borgers M, Dansnen W, et al. Ultrastructural and cytochemical correlates of myocardial protection by cardiac hypothermia in man. J THORAC CARDIOVASC SURG 1980;79:413-24. Rosenkrantz ER, Okamoto F, Buckberg GD, et al. Studies on controlled reperfusion after ischemia. II. Biochemical studies: failure of A TP levels to predict recovery of contractile function after controlled reperfusion. J THORAC CARDIOVASC SURG 1986;92:502-12. Lew WYW, Chen Z, Guth B, et al. Mechanisms of augmented shortening in nonischemic areas during ischemia in the canine left ventricle. Circ Res 1985;56:351-8. Misare BD, Krukenkamp IB, Lazer ZP, et al. Retrograde is superior to antegrade continuous warm blood cardioplegia for acute cardiac ischemia. Circulation 1992;86(Suppl):II393-7. Partington MT, Acar C, Buckberg GD, et al. Studies of retrograde cardioplegia. II. Advantages of antegrade/retrograde cardioplegia to optimize distribution in jeopardized myocardium. J THORAC CARDIOVASC SURG 1989;97:61322.
Discussion Dr. John H. Kennedy (Cambridge, England). Would you continue to classify what you are measuring as flow per gram of heart as such if there were spherules appearing in the coronary sinus after antegrade injection? Dr. Davis C. Drinkwater, Jr. (Los Angeles, Calif). I have one question about the technical aspect of RC. When RC is used clinically, it may result in much effluent and reflux back out of the coronary sinus and therefore may account for a considerable proportion of the total volume delivered. In your model, did you have a technique to occlude the coronary sinus totally during RC to avoid this leakage which might effect your findings? Dr. William D. Spotnitz (Charlottesville, Va.). I would like to support the data that you have presented. Recent data from our laboratory, presented at the American College of Surgeons,
The Journal of Thoracic and Cardiovascular Surgery February 1994
confirm the results that you have presented. I would also like to point out that if you use a metabolized product such as thallium, which can be taken up by the cells, you actually have an even greater discrepancy between AC and RC. I think radiolabeled microspheres as a method of measuring this discrepancy are problematic because they do not measure the loss of retrograde flow that occurs by collaterals, which are smaller then 10 !Lm. So the discrepancy that you have between AC and RC is even greater than that which you have already shown. I would like to know how, in aortic valve replacement or mitral valve replacement, when there is this large discrepancy, you would achieve good myocardial preservation with retrograde flow. In our own data, we found the same discrepancy that you have shown, in fact, an even greater one. However, the retrograde cooling was at least equivalent to that of antegrade and probably reflects many of the local effects of the transmission of cold solution. Dr. Aldea. To answer Dr. Kennedy's question, the entrapment rate with the antegrade route was 95% or more. The ventricles were vented, and therefore the microspheres were not recirculated. With the retrograde route, we had bicaval cannulation, and the left atrium was also vented; again, it was a single-shot effect. We were very careful to measure the effluent from the aortic root for radioactivity during retrograde injection to make sure that we did not dislodge any entrapped antegrade miscrospheres, and the average dislodgment of microspheres during retrograde injection was less than 0.2%. This rate would not account, therefore, for the measurements that we have made. I am familiar with the work at the University of California, Los Angeles. The measurements that we made were strictly cardioplegic solution flow delivery. We understand that there are some secondary effects of temperature and cooling. In this experiment, we were particulary interested in whether the patterns of cardioplegic solution delivery were matched. We did not measure infarct size or the actual protection. To answer the last question of Dr. Spotnitz, I think that we may sometimes be too cavalier when we assume that when we have good preservation of ventricular function at the end of an operation, as seen with an isolectric electrocardiogram, that all areas of the heart were truly protected. I think this work suggests that some small areas of the heart are not protected, even under the best of circumstances and with the use of global measurements such as LV function, electrocardiographic activity, or global function, such as cardiac output, may not particularly reflect these regional differences and sensitivities. These discrepancies may be particularly important when we are dealing with impaired LVs.