Protective Effect of an Asanetlineous Re erfusion Solution on Myocardial Performance Fo E owing Cardioplegic Arrest Philippe Menasche, M.D., Christian Grousset, Ph.D., Georges de Boccard, M.D., and Armand Piwnica, M.D.

ABSTRACT This study assesses whether an appropri- ately designed asanguineous initial reperfusate effectively reduces the reperfusion injury following prolonged global ischemia and improves the recovery of cardiac perfor- mance after cardioplegic arrest. Forty-eight isolated per- fused working rat hearts underwent two hours of hy- pothermic (15" to 18°C) ischemic arrest followed by 30 minutes of normothermic reperfusion. During ischemic in- jury, multidose cardioplegia was delivered at 30-minute intervals. The reperfusion solution under study was in- fused during the last 3 minutes of ischemia, just prior to release of the aortic clamp. The usual hemodynamic vari- ables of this preparation (heart rate, aortic pressure, aortic flow, coronary flow, and stroke volume) were serially re- corded and expressed as percent of recovery of control values.

The influence of the concentration of Ca2+, pH, and buff- er was more specifically investigated. A reperfusate con- taining 1 mM of Ca2+ was found to result in higher post- ischemic hernodynamic values than a Ca2+-poor (0.25 mM) reperfusate. The best functional recovery was provided by an alkalotic (pH 7.70 at 28OC), glutamate-enriched initial reperfusate, which, by 30 minutes of reperfusion, yielded a 93.5 f 2.3% recovery of aortic flow versus 83.6 2 1.8% in the control group receiving unmodified reperfusion ( p < 0.01). We conclude that an appropriate composition of the initial reperfusate can improve the recovery of cardiac function significantly following two hours of cardioplegic k s t and that such an improvement can be achieved by an asanguineous reperfusate provided its composition is properly designed with respect to electrolytes, pH, and substrates.

Although many experimental studies have been devoted to the composition of cardioplegic solutions used to pro- tect the heart during aortic cross-clamping, less attention has been paid to the chemical composition of the initial reperfusate. This aspect of myocardial protection is

From the Cardiovascular Surgical Department and INSERM U-127, Hdpital Lariboisiere, Paris, France.

Accepted for publication June 24, 1983.

Address reprint requests to Dr. Menasche, Service de Chirurgie Cardio- Vasculaire, HBpital Lariboisiere, 2, rue Ambroise Pare, 75475 Pans, France.

likely to be clinically relevant, since the reperfusion in- jury reported to occur following temporary global isch- emia could contribute substantially to impair postopera- tive cardiac performance [l]. The main potential causes of this reperfusion damage include an increased uptake of Ca2+ [2], acidosis [3], edema [4], lack of substrates [5], and decreased oxygen extraction [6]. On the basis of these metabolic derangements, Follette and co-workers [l] showed that reperfusion damage could be largely avoided by appropriate modification of the initial blood reperfusate.

There are some limitations to the use of blood in this particular setting. Noteworthy among them are the complexity of adding intraoperatively the various com- pounds required to achieve the desired composition of the reperfusate and the difficulties of accurately control- ling the final chemical characteristics of such a modified blood. However, a crystalloid, sterilized, precontrolled, and readily available solution would obviate these prob- lems in the clinical setting. Consequently, the present study was undertaken to assess whether such an asan- guineous reperfusion solution, whose composition would be specifically designed to reduce reperfusion damage, could effectively improve the functional recov- ery of isolated perfused working hearts subjected to a prolonged period of global ischemia under cardioplegic protection.

Material and Methods Experimental Model The preparation is a modification of the isolated working rat heart model of global ischemia described by Neely and associates [7]. Briefly, the oxygenated perfusate medium enters the cannulated left atrium at a pressure of 20 cm H 2 0 and is passed to the left ventricle, from which it is spontaneously ejected (pacing was not used in this study) against a hydrostatic pressure of 100 cm H20. Coronary flow exits from the pulmonary artery and is pooled and recirculated with the aortic output. The aortic and coronary flow rafes are recorded at timed intervals. Aortic pressure is monitored by a pressure transducer connected to a Washington MD 200 mul- tichannel recorder. The cardioplegic and reperfusion so- lutions are infused through a sidearm on the aortic can- nula at a pressure of 60 cm H20.

The perfusion fluid is Krebs-Henseleit bicarbonate buffer containing the following in millimoles: NaCl, 118; KC1, 4.7; NaHC03, 25; MgS04 * 7H20, 1.2; CaC12 2H20, 2.5; KH2P04, 1.2; and glucose, 11.1. The perfu-

222

223 Menaschk et al: Asanguineous Reperfusion Solution

sate is gassed with a mixture of 95% oxygen and 5% carbon dioxide. The pH of the perfusate is adjusted to 7.40 (at 37"C), and the partial pressure of oxygen is maintained over 600 mm Hg. Before use, the perfusate is filtered through a cellulose acetate filter with 5-p pores.

Experimental Time Course Male Sprague-Dawley rats weighing 300 gm were anes- thetized with diethyl ether and injected intravenously with 4,000 IU of heparin. The heart was promptly re- moved and connected to the aortic cannula. Retrograde perfusion (Langendorff) was initiated while the left atrium was cannulated, thus allowing conversion of the preparation to a working heart system for a 20-minute control period.

During this time, the stability of the preparation was confirmed and control values for heart rate (HR), peak aortic systolic pressure, aortic flow, and coronary flow of the working heart were serially established. The total cardiac output was calculated as the sum of aortic flow and coronary flow. Since pacing was not used, the car- diac output, although measured under futed conditions of preload and afterload, could not be considered as an accurate indicator of myocardial performance. Conse- quently, to eliminate the variability related to differences in HR between the experimental groups, the stroke volume (SV) was calculated for each individual heart at any preischemic and postischemic interval.

Global ischemia was achieved by clamping the can- nulas leading to the aorta and the left atrium. Subse- quently, ischemia was maintained for 120 minutes. The 4°C cardioplegic solution was delivered at the onset of ischemia and thereafter at 30-minute intervals. Each car- dioplegic infusion consisted of 30 ml delivered over a 3- minute period. The reperfusion solution under study (30 ml) was infused during the last 3 minutes of ischemia, immediately prior to removal of the aortic clamp. Initial reperfusion by the Krebs-Henseleit buffer was done in the Langendorff mode for 1 minute. Left atrial perfusion was then reinstituted, and the heart was allowed to beat on a working mode during a 30-minute recovery period.

The recovery rates of HR, peak aortic systolic pres- sure, aortic flow, coronary flow, cardiac output, and SV were recorded at 3, 5, 10, 20, and 30 minutes of aerobic reperfusion. Postischemic values were expressed as a percent of the control values. This allowed us to relate the functional recovery profile of each heart to the com- position of the reperfusion solution.

During the control and recovery periods, the hearts were perfused at 37°C. Throughout the period of isch- emia, the heart was kept in a sealed water-jacketed chamber at 15" to 18°C through a cooling circuit that was separate from the rest of the perfusion apparatus, the latter still maintained at 37°C. During the final phase of ischemia, the chamber containing the heart was progres- sively rewarmed so that the myocardial temperature av- eraged 28°C by the time the reperfusion solution was delivered. The temperature of the heart chamber was reset at 37°C when the preparation was again converted

to the working mode for the postischemic mea- surements.

Experimental Groups CONTROL. The control series consisted of multidose car- dioplegia with unmodified reperfusion. The composi- tion of the cardioplegic solution was as follows: Na+, 100 mM/L; K + , 4 mM/L; Ca2+, 0.25 mM/L; M g + , 13 mM/L; L-histidine, 10 gm/L; mannitol, 12 g d ; pH, 7.40 (20°C); osmolarity, 370 mOsm/L. To minimize the reperfusion edema [4], all perfusates were made hyperosmolar (370 mOsm/L.) by the addition of mannitol[8]. At the comple- tion of the two-hour period of arrest, the aortic clamp was released and the heart was reperfused with the standard Krebs-Henseleit buffer.

In these series, the initial reperfusate consisted of an asanguineous solution (reperfusion solution) whose ba- sic composition was as follows: Na+, 100 mM/L; K + , 4 mM/L; Mg+, 13 mM/L; and mannitol, which was added until the osmolarity reached 370 mOsm/L. The influence of three variables was specifically investigated: concen- tration of Ca2+, pH, and buffer.

To study the influence of the concentration of Ca2+, the concentration of ionic Ca2+ in the reperfusion solu- tion was lowered to 0.25 mM/L in one series of hearts. This low Ca2+ reperfusate was compared with a Ca2+- rich (1 mM/L) initial reperfusate.

To study the influence of the pH, the initial reperfu- sate pH was adjusted to 7.70 at 28°C and compared with a pH 7.40 reperfusate. To test the potential influence of the buffer at a given pH, four series of hearts were evalu- ated: pH 7.40, histidine buffer; pH 7.40, glutamate buff- er; pH 7.70, histidine buffer; and pH 7.70, glutamate buffer. The choice of histidine (10 gm/L) and glutamate (2,942 gm/L) as the reperfusate buffers is discussed later.

S ta tis tical Analysis Eight hearts were studied under each of the experimen- tal conditions. Statistical analyses were performed using the unpaired t test with p = 0.05 as the limit of significance. Results are expressed as the mean & the standard error of the mean.

Results On resumption of flow, all hearts defibrillated spontane- ously within a few seconds. As a whole, postischemic values for HR and coronary flow were neither sig- nificantly different from controls nor different between all the groups. On the other hand, the recovery of peak aortic systolic pressure, cardiac output, and SV was closely related to the composition of the initial re- perfusate.

Inpuence of Concentration of Ca2+ The results are depicted in Figure 1 and Table 1. There were two principal findings. Compared with multidose cardioplegia alone, the addition of an initially Ca2+-rich reperfusate resulted in higher postischemic values of

MULTIDOSE CARDIOPLEGIA AND REPERFUSION SOLUTION.

224 The Annals of Thoracic Surgery Vol 37 No 3 March 1984

' 0 recovery of aortic (low

100 I 9oc

501/ 40 t C P

-4- cP+ normo Ca" reperf. 11 mMI

-0- cP+ Ca2+ poor reperf. 1 0 2 5 m M 1

10

Y 1 1 I I I I I * 0 :3 5 10 15 20 25 30

Post-ischemic time lmnl

Fig I . Influence of the reperfusate ionic calcium on recowry of aortic flow after ischemia. (CP = cardioplegia; reperf. = reperfusate.)

peak aortic systolic pressure but without a concomitant increase in aortic flow or SV. Conversely, an initially Ca2+-poor reperfusion solution yielded a significantly lower recovery of aortic flow and SV than either multi- dose cardioplegia alone or multidose cardioplegia fol- lowed by a Caz+-rich reperfusate.

Influence of pH and Buffer To more sensitively assess any potential difference be- tween the pH 7.40 and the pH 7.70 reperfusates, we deliberately kept the Ca2+ concentration at a low level (0.25 mM) in these series. The primary importance of the buffer was demonstrated by the finding that the two histidine-containing reperfusates, respectively buffered to pH 7.40 and pH 7.70, resulted in lower postischemic values of aortic flow and SV than multidose cardioplegia alone, whereas the two glutamate-containing reperfu- sates, similarly buffered to pH 7.40 and pH 7.70, achieved a markedly better recovery than cardioplegia with unmodified reperfusion (Table 2; Fig 2).

Table 1 . Influence of Reperfusate Ionic Calcium on Postischemic Recovery of Peak Aortic Systolic Pressure, Heart Rate, and Stroke Volume"

Aortic Pressure Heart Rate Stroke Volume

Experimental Group Control Percent Control Percent Control Percent (N = 8) (mm Hg) Recovery (beatslmin) Recovery (mu Recovery

1: Multidose cardioplegia 190.8 f 4.7 92.3 f 1.0 270 f 9 101.9 f 1.6 0.26 f 0.01 88.8 f 2.2 2 Multidose cardioplegia + 147.4 f 2.0 95.5 f 0.8' 286 f 7 100.8 f 1.1 0.30 f 0.01 86.8 f 1.4

3: Multidose cardioplegia + 145.2 f 1.4 90.1 f 1.2' 295 f 5 99.4 f 1.1 0.29 f 0.01 82.5 f 1.6b normo CaZ+ reperfusion

Caz+-poor reperfusion

'Data for percent of recovery were obtained at 30 minutes of reperfusion. bSigruficance: p < 0.05 versus Group 1. 'Sigtuficance: p < 0.01 versus Group 2.

Table 2 . Influence of Reperfusate pH and Buffer on Postischemic Recovery of Aortic Flow and Stroke VolumL

Aortic Flow Stroke Volume

Experimental Group (N = 8)

Control Percent Control Percent ( d m i n ) Recovery (d) Recovery

1: Multidose cardioplegia 49.6 f 2.5 83.6 f 1.8 0.26 f 0.01 88.8 f 2.2 2 Multidose cardioplegia +

histidine reperfusate, pH 7.40 65.9 f 3.7 72.3 f 3.0' 0.29 f 0.01 82.5 f 1.6'

3: Multidose cardioplegia + 64.7 f 4.7 82.4 f 1.2 0.31 f 0.01 84.0 f 3.0

4: Multidose cardioplegia + 64.1 f 1.5 87.1 f 2.7 0.34 f 0.01 94.6 f 2.1

5: Multidose cardioplegia + 59.7 f 4.3 93.5 f 2.3b 0.30 f 0.01 94.3 f 2.6

histidine reperfusate, pH 7.70

glutamate reperfusate, pH 7.40

glutamate reperfusate, pH 7.70

'Data for percent of recovery were obtained at 30 minutes of reperfusion. bSignificance: p < 0.01 versus Group 1. CSigtuficance: p < 0.05 versus Croup 1.

225 Menascht! et al: Asanguineous Reperfusion Solution

5

":I 10

J ' - 0 3 5 10 15 20 25 30

Po51 is(:Iietiii(: titiit: 1 inn)

Fig 2. Influence of the reperfusate (reperf.) buffer on the postischemic recovery of aortic flour. (CP = cardioplegia.)

However, within each set of buffers, the results tended to favor the use of an alkalotic pH on the follow- ing grounds.

1. In the histidine-buffered series, the pH 7.40 reperfu- sate yielded a significantly poorer recovery of aortic flow and SV than the control group (cardioplegia alone). Raising the pH to 7.70 resulted in a relatively improved recovery reflected by the absence of a significant difference in the postischemic values of SV between pH 7.70 buffered reperfusion and cardiople- gia with unmodified reperfusion (see Table 2).

2. In the glutamate-buffered series, the recovery of aortic flow was not significantly different between cardioplegia alone and cardioplegia followed by pH 7.40 reperfusion, but it was significantly improved in hearts that received the pH 7.70 reperfusate. The pro- tective effect of an alkalotic pH was further evidenced by the finding that although there was no significant difference in the postischemic values of SV between unmodified reperfusion and pH 7.40 reperfusion, there was significant improvement in SV during the 10 initial minutes of reflow after pH 7.70 buffered reperfusion relative to cardioplegia alone (at 3 min- utes of reperfusion, p < 0.02; at 5 and 10 minutes of reperfusion, p < 0.01).

These higher values of SV were recorded without con- comitant differences in the percent of recovery of HR measured at the same time intervals: at 3 minutes of reperfusion, HR recovered by 95.5 2 3.3% in the control group versus 89.9 * 4.1% in the pH 7.70 glutamate- reperfused group; at 5 and 10 minutes of reperfusion, the percentages of recovery for HR were, respectively, 94.1 2 3.7% and 98.1 * 3.1% after multidose cardiople- gia and 92.3 2 4.1% and 97.0 2 2.1% after alkalotic glutamate-enriched reperfusion. None of these differ- ences were significant. Thus the higher postischemic re-

covery of SV in the pH 7.70 glutamate-reperfused group cannot be related to the intrinsic inotropic effects of HR but most likely reflects a better preservation of myocar- dial contractility.

These overall results demonstrate that the beneficial effects of the reperfusate are not merely due to a wash- out of the coronary bed prior to resumption of aortic flow. They are also closely related to the chemical com- position of the reperfusate.

Comment Influence of Concentration of Ca'+ One finding of our study was that after a two-hour pe- riod of cardioplegic ischemic arrest, an initial reperfusate containing an almost physiological level of free Ca2+ (1 mWL) resulted in significantly better functional recov- ery, as assessed by the postischemic values of aortic flow and SV, than a Ca2+-poor reperfusate (0.25 mM/L). These results are consistent with those of other inves- tigators who used an isolated feline [9] or rat heart [lo] preparation and found that the postischemic recovery of myocardial performance is not improved by lowering the concentration of ionized Ca2+ during the early phase of reperfusion.

However, these data are at variance with those re- ported by Follette and colleagues [l]. They advocated the use of an initially Ca2+-poor reperfusate (0.50 mM/L) on the basis of an increased uptake of Ca2+ by the in- jured myocardium during the postischemic period (21, this uptake leading to intramitochondrial Ca2+ precipi- tates and myofibrillar contraction bands [ 1 I].

This discrepancy between the data reported could be related to species differences with respect to the ability of mitochondria to accumulate Ca2+ [lo]. However, we think that the technique of myocardial protection used during the ischemic episode is likely to influence the extent of the Ca2+-mediated reperfusion damage greatly. Follette and co-workers [ l ] used topical hy- pothermia, a technique that does not prevent severe ul- trastructural damage [12] and marked shifts in intracel- lular electrolytes during reflow [13]. In that setting, the expected increase in membrane Ca2+ permeability [ 141 could account for the beneficial effects of a Ca2+-poor reperfusate allowing a reduced Ca2+ load to be available for cellular influx.

Conversely, cardioplegia has been reported to pre- serve the ability of the ischemic cell to control ion gra- dients [13, 151. In the present study, we used a M g + - rich (13 mM), Na+-rich (100 mM), Cat+-poor (0.25 mM) cardioplegic solution. This combination has previously been shown to afford substantial preservation of the high-energy phosphate compounds after elective isch- emic arrest (161. The availability of sufficient levels of adenosine triphosphate (ATP) is expected to support the required rate of the ion pumps, especially with respect to Ca2+, during reperfusion [17] and hence to obviate the need for a Ca2+-poor reperfusate. These data em- phasize that the optimal concentration of Ca2+ during the initial phase of reperfusion should be closely related

226 The Annals of Thoracic Surgery Vol 37 No 3 March 1984

Table 3. Influence of Reperfusate pH on Early Postischemic Recozwy of Coronary Flow

Experimental Group Control (N = 8) (ml/rnin) 3 Minutes 5 Minutes 10 Minutes

104.0 2 3.1 1: Multidose cardioplegia 20.5 2 1.9 105.8 5 2.9 105.3 2 2.5 2: Multidose cardioplegia + 20.1 lr 1.3 124.7 2 9.8" 120.7 5 7.8 109.9 2 5.3

3: Multidose cardioplegia + 21.1 ? 1.5 130.1 lr 7.1b 125.1 2 5.5h 108.1 2 4.6

Percent Recovery of Coronary Flow

histidine reperfusate, pH 7.70

glutamate reperfusate, pH 7.70

"Significance: I J < 0.05 versus Group 1 . 'Significance: 11 < 0.01 versus Group 1 .

to the composition of the cardioplegic solution used dur- ing the preceding period of ischemia [18, 191.

Influence of pH In each series of buffers, the best recovery of function was provided by the alkalotic reperfusate. The pH value of 7.70 at 28°C (the temperature of the heart while the reperfusion solution was being delivered) is at the upper limits of the pH range reported in cold-blooded animals that experience natural pH adjustments as a function of temperature fluctuations [20]. This value is in agreement with those reported to be successful by Follette and co- workers [l].

The results of the present study support other experi- mental data [l, 211 suggesting that raising the pH reper- fusate is justified to counteract intracellular acidosis that develops during ischemia and may be exacerbated dur- ing reflow [3, 22].* This makes the pH appropriate for recovery of cell processes [23]. Further, our data confirm that an elevated reperfusate pH has the additional ad- vantage of reducing coronary vascular resistance [23], since alkalotic reperfusates were found to increase early postischemic coronary blood flow at a constant perfu- sion pressure (Table 3).

Influence of Buffer Several buffers can be used to adjust the pH reperfusate. Follette and co-workers [ l ] advocated the use of TRIS or tris(hydroxymethy1)aminomethane on the basis that be- ing nonionized, TRIS passes through cell membranes more readily than ionic buffers [24]. However, TRIS has been reported to be toxic in a number of tissues [25].

Our choice of histidine and glutamate was based on the assumption that they might act as both buffers and substrates. This concept was derived from biochemical data demonstrating an extraglycolytic production of high-energy phosphates from amino acids in the anoxic mammalian heart [26, 271. In fact, Rau and associates [28] demonstrated that only amino acids that accumulate in the cytosol as aspartate or glutamate are effective in

'Duval-Amould M, Ingwall JS, Menasche P, et al: Beneficial effects of nifedipine on cardiac metabolism and function after cardioplegic arrest: a phosphorus-31 nuclear magnetic resonance study. (submitted for publi- cation, 1984)

reducing ischemic damage. This observation probably accounts for the better results yielded by our glutamate- containing reperfusion solution (see Fig 2).

Previous experimental reports [29, 301 have already emphasized the beneficial effects of adding glutamate to a cardioplegic solution. There is strong evidence that glutamate acts principally by enhancing substrate sup- ply through the malate-aspartate shuttle [28] (Fig 3). Such a mechanism could be critically important during the early reperfusion period for two reasons.

1. Glycolysis and oxidation of fatty acids have been shown to remain inhibited during the first minutes of reoxygenation [5] so that the production of ATP from the citric acid cycle is substrate limited [31].

2. Tissue glutamate levels are expected to be low as a consequence of the preceding period of ischemia [32].

Card ioplegic Reoxygenat ion The maintenance of cardioplegia during the early phase of reoxygenation is based on the assumption that the avoidance of an energy-consuming mechanical activity could allow the aerobically produced ATP to be elec- tively channeled toward an intracellular reparative pro- cess (331. This concept has been validated by the experi- ments of Follette and co-workers [MI, who reported on the use of a cardioplegic blood reperfusate.

However, there are some concerns about the use of blood as the vehicle for the reperfusate components. Five of the most important are as follows:

1. The complexity of extemporaneously adding multi- ple compounds to achieve the proper composition with respect to electrolytes, substrates, pH, and os- molarity

2. The questionable reproducibility of this final compo- sition from one patient to another

3. A potentially increased risk of sepsis subsequent to these additional intraoperative manipulations

4. The unfavorable shift of the oxyhemoglobin dissocia- tion curve under hypothermic conditions

5. The systemic recirculation of the blood reperfusate containing the metabolic end-products resulting from the preceding period of ischemia

227 MenaschC et al: Asanguineous Reperfusion Solution

CYTOSOL MITOCHONDRIA

I NAD-, /MdI Ma1 \ \

Fig 3. The malate (Mal)-aspartate (Asp) shuttle. (Triose P = triose phosphates); F'yr = pyruvate; coA = coenzyme A; NADH = nicofinumide-adenine dinucieotide; OAA = omloacetate; Glu = glutamate; aKg = a-keto-glutarate; Succ = succinute; aKg Dhase = a-keto glutarate dehydrogenuse.)

Although the present study shows that increased pro- tection can be afforded by an appropriately designed crystalloid, hypoxic (partial pressure of oxygen equal to 100 mm Hg) reperfusate, we also have investigated asanguineous oxygenated reperfusion solutions that would have the major advantages of being accurately controlled, readily available, and easy to handle in a clinical setting. Effective oxygenation of such crystalloid perfusates can be provided by perfluorochemicals. The addition of these oxygen-carrying compounds in cardio- plegic solutions has proven to be successful experimen- tally [35], and our preliminary results with their use in the setting of reperfusion solutions appear equally promising [36].

In summary, our results show that the recovery of cardiac function following a two-hour period of cardio- plegic arrest can be improved significantly by an asan- guineous cardioplegic initial reperfusate that contains an almost normal concentration of Ca2+ (1 mM), is al- kalotic, and is supplemented with glutamate. Although care should be taken when extrapolating the results ob- tained in our experimental model to the setting of car- diac operations on human beings, preliminary trials in patients undergoing valve replacement tend to confirm that the functional preservation afforded by cardioplegia can be substantially enhanced by an appropriate crystal- loid reperfusion solution [37].

Supported in part by grants from the Federation de Cardiologie and the UER Lariboisiere-St. Louis. We express our gratitude to Richard E. Clark, M.D., Bethesda, MD, for his assistance in preparing this paper.

References 1. Follette DM, Fey K, Buckberg GD, et al: Reducing post-

ischemic damage by temporary modification of reperfusate calcium, potassium, pH, and osmolarity. J Thorac Cardio- vasc Surg 8 2 2 1 , 1981

2. Schaper J, Hehrlein F, Schlepper M, Thiedemann KU: UI- trastructural alterations during ischemia and reperfusion in human hearts during cardiac surgery. J Mol Cell Cardiol 9:175, 1977

3. Hearse DJ, Humphrey SM, Bulloch GR The oxygen paradox and the calcium paradox: two facets of the same problem. J Mol Cell Cardiol 10641, 1978

4. Engelman RM, Chandra R, Baumann, G, Goldman RA: Myocardial reperfusion, a cause of ischemic injury during cardiopulmonary bypass. Surgery 80:266, 1976

5. Mochizuki S, Neely JR Energy metabolism during reperfu- sion following ischemia. J Physiol 76:805, 1980

6. Lucas SK, Kanker KR, Schaff HV, et al: Reduced oxygen extraction during reperfusion: a consequence of global isch- emic arrest. J Surg Res 28:434, 1980

7. Neely JR, Liebermeister N, Battersby EJ, Morgan HE: Effect of pressure development on oxygen consumption by the isolated rat heart. Am J Physiol212:804, 1967

8. Foglia RP, Steed DL, Follette DM, et al: Iatrogenic myocar- dial edema with potassium cardioplegia. J Thorac Cardio- vasc Surg 78217, 1979

9. Bixler TJ, Gardner TJ, Flaherty JT, et al: Effects of changes in calcium concentration during post-ischemic reperfusion on ventricular function and myocardial edema and metabo- lism. Surg Forum 28:277, 1977

10. Watts JA, Koch CD, Lanoue KF: Effects of Ca2+ antagonism on energy metabolism: CaZ+ and heart function during ischemia. Am J Physiol 238:909, 1980

11. Jennings RB, Ganote CE: Structural changes in myocar- dium during acute ischemia. Circ Res 34:Suppl3156, 1974

12. Flameng W, Borgers M, Daenen W, et al: St. Thomas car- dioplegia versus topical cooling: ultrastructural and bio- chemical studies in humans. Ann Thorac S,urg 31:339, 1981

13. Mavroudis C, Ebert PA: The effects of potassium cardiople- gia and hypothermia on the distribution of water and ions in the canine heart. Surgery 83:476, 1978

14. Reimer KA, Jennings RB, Hill ML: Total ischemia in dog hearts, in vitro: 11. High energy phosphate depletion and associated defects in energy metabolism, cell volume regu- lation and sarcolemmal integrity. Circ Res 4991, 1981

228 The Annals of Thoracic Surgery Vol 37 No 3 March 1984

15. Kay HR, Levine FH, Fallon JT, et al: Effect of cross- clamping time, temperature and cardioplegic agents on myocardial function after induced arrest. J Thorac Cardio- vasc Surg 76:590, 1978

16. Pemot AC, Ingwall JS, Menasche P, et al: Limitations of potassium cardioplegia during cardiac ischemic arrest: a phosphorus 31 nuclear magnetic resonance study. Ann Thorac Surg 32:536, 1981

17. Nayler WG, Ferrari R, Williams A: Protective effect of pre- treatment with verapamil, nifedipine and propranolol on mitochondria1 function in the ischemic and reperfused myocardium. Am J Cardiol%:242, 1980.

18. Nayler WG, Yepez CE, Poole-Wilson PA: The effect of p- adrenoreceptor and Caz+-antagonist drugs on the hypoxia- induced increase in resting tension. Cardiovasc Res 12:666, 1978

19. Shine KI, Douglas AM, Ricchiuti NV: Calcium, strontium and barium movements during ischemia and reperfusion in rabbit ventricle: implications for myocardial preservation. Circ Res 43:712, 1978

20. White FN: A comparative physiological approach to hy- pothermia. J Thorac Cardiovasc Surg 82:821, 1981

21. Bixler TJ, Gardner TJ, Flaherty JT, et al: Beneficial effects of alkalotic reperfusion following ischemic cardiac arrest. J Surg Res 24488, 1978

22. Cobbe SM, Poole-Wilson PA: Tissue acidosis in myocardial hypoxia. J Mol Cell Cardiol 12761, 1980

23. Rovetto MJ, Whitmer JT, Neely JR: Comparison of the ef- fects of anoxia and whole heart ischemia on carbohydrate utilization in isolated working rat hearts. Circ Res 32:699, 1973

24. Beierholm EA, Grantham RN, OKeefe DD, et al: Effects of acid-base changes, hypoxia and catecholamines on ven- tricular performance. Am J Physiol228:1555, 1975

25. Gillespie JS, McKnight AT Adverse effects of TRIS hydro- chloride, a commonly used buffer in physiological media. J Physiol (Lond) 259561, 1976

26. Berkowitz S, Perille T, Lesch M: Anaerobic amino-acid me-

tabolism as a potential energy source in mammalian heart. Clin Res 26:219, 1978

27. Penney DG, Cascarano J: Effects of glucose and tricarboxy- lic acid cycle intermediates on metabolism and physiologi- cal performance. Biochem J 118221, 1970

28. Rau EE, Shine IU, Gervais A, et al: Enhanced mechanical recovery of anoxic and ischemic myocardium by amino acid perfusion. Am J Physiol 236:873, 1979

29. Engelman RM, Dobbs WA, Rousou JH, Meeran MK My- ocardial high-energy phosphate replenishment during isch- emic arrest: aerobic versus anaerobic metabolism. Ann Thorac Surg 33:453, 1982

30. Robertson JM, Vinten-Johansen J, Buckberg GD, et al: Pro- longed safe aortic clamping (4 hours) with cold glutamate enriched blood cardioplegia. Circulation WSuppl4147,1981

31. McGovem G, Jacobus WE, Gardner TJ, Flaherty JT Irre- versible damage induced by substrate deprivation during early post-ischemic reflow. Circulation 64:Suppl4:274,1981

32. Taegtmeyer H: Metabolic adaptation to myocardial hy- poxia: increased synthesis and release of succinic acid. Cir- culation 55, 56:Suppl286, 1977

33. Danforth WH, Naegles S, Bing RJ: Effect of ischemia and reoxygenation on glycolytic reactions and adenosine triphosphate in heart muscle. Circ Res 8:965, 1960

34. Follette DM, Steed DL, Foglia RP, et al: Reduction of post- ischemic myocardial damage by maintaining arrest during initial reperfusion. Surg Forum 28281, 1977

35. Kanter KR, Jaffin JH, Ehrlichman RJ, et al: Superiority of perfluorocarbon cardioplegia over blood or crystalloid car- dioplegia. Circulation 64:Suppl2:75, 1981

36. Menasche P, Fauchet M, Lavergne A, et al: Applications of a perfluorocarbon to heart ischemia. In Bolin RB, Geyer RP, Nemo GJ (eds): Advances in Blood Substitute Research. New York, Liss, 1983, pp 363-372

37. Menasche P, Dunica S, Kural S, et a1 An asanguineous reperfusion solution: an effective adjunct to cardioplegic protection in high risk valvular surgery. J Thorac Cardio- vasc Surg (in press, 1984)