| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ann Thorac Surg 2000;70:197-205
© 2000 The Society of Thoracic Surgeons
a Department of Surgery, Massachusetts General Hospital and Harvard Medical School, Boston, Massachusetts, USA
b Department of Cardiology, Massachusetts General Hospital and Harvard Medical School, Boston, Massachusetts, USA
c Department of Radiology, Massachusetts General Hospital and Harvard Medical School, Boston, Massachusetts, USA
Address reprint requests to Dr Torchiana, Department of Surgery, Massachusetts General Hospital BUL119, Fruit St, Boston, MA 02114
e-mail: torchiana.david{at}mgh.harvard.edu
 |
Abstract |
|---|
 Top Abstract IntroductionMaterial and methodsResultsCommentReferences |
|---|
Methods. Isolated cross-perfused canine hearts underwent cardioplegic arrest followed by 45 minutes of global ischemia at 10°C or 37°C, or 45 minutes of normothermic ischemia without prior cardioplegia. Left ventricular function was measured at baseline and during 2 hours of recovery. Metabolism was continuously evaluated by phosphorus-31 magnetic resonance spectroscopy.
Results. Adenosine triphosphate was 71% ± 4%, 71% ± 7%, and 38% ± 5% of baseline at 30 minutes, and 71% ± 4%, 48% ± 5%, and 39% ± 6% at 42 minutes of ischemia in the cold ischemia, warm ischemia, and normothermic ischemia without prior cardioplegia groups, respectively. Left ventricular systolic function, left ventricular relaxation, and high-energy phosphate levels recovered fully after cold cardioplegia and ischemia. Prior cardioplegia delayed the decline in intracellular pH during normothermic ischemia initially by 9 minutes, and better preserved left ventricular relaxation during recovery, but did not ameliorate the severe postischemic impairment of left ventricular systolic function, marked adenosine triphosphate depletion, and creatine phosphate increase. Left ventricular distensibility decreased in all groups.
Conclusions. When cardioplegia is followed by prolonged ischemia, better protection is provided by hypothermia than by normothermia. Prior cardioplegia confers little advantage on recovery after prolonged normothermic ischemia but delays initial ischemic metabolic deterioration, which would contribute to the safety of brief interruptions of warm cardioplegia.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentReferences |
|---|
We tested two hypotheses: (1) that as ischemia is prolonged, the reduced metabolic rate offered by cooling would give the advantage to hypothermic cardioplegia; and (2) that prior warm blood cardioplegia would mitigate the negative effects of normothermic ischemia. We used the isolated cross-perfused canine heart and phosphorus-31 (31P) magnetic resonance spectroscopy (MRS) to assess recovery from ischemia and to enable repeated nondestructive measurements of myocardial metabolism.
|
Material and methods |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentReferences |
|---|
When instrumentation was complete, the heart was supported by a pericardial sling within a water-jacketed Plexiglass cylinder. The anterior LV surface was uppermost and positioned to maintain contact, without restricting cardiac pulsation, with an MRS surface coil mounted within the cylinder. The cylinder was placed in the center of a wide-bore horizontal magnet [3].
Heparin (6,000 U) was administered to the support dog for anticoagulation, followed by 1,000 U/h. Indomethacin 50 mg, methylprednisolone 1 g, chlorpheniramine maleate 10 mg, and cimetidine 300 mg were administered intravenously, followed by a continuous infusion of 110 mg indomethacin, 1 g solumedrol, 600 mg cimetidine, and 20 mg chlorpheniramine maleate in 500 mL of normal saline solution at 50 mL/h to maintain hemodynamic stability of the support dog. Hetastarch was infused if needed. Chloralose (2 mg · kg-1 · min-1) and urethane (20 mg · kg-1 · min-1), administered intravenously, maintained anesthesia. Ventilation was adjusted, and sodium bicarbonate was added as necessary to maintain arterial PO2 above 200 mm Hg, and PCO2 and pH in the physiologic range. The heart was maintained at 37°C, or at 10°C during cold cardioplegic arrest and cold ischemia [3].
Pressure measurements
Measurement of coronary blood flow, coronary perfusion pressure, support dog aortic pressure, and myocardial temperature has been described [3]. Left ventricular pressure (LVP) was measured by a high-fidelity micromanometer-tipped catheter (SPC 350 MR, Millar Instruments, Houston, TX) within the balloon, and zero-corrected by referencing diastolic pressure measured through a saline-filled catheter [3]. All pressures were monitored and recorded on a strip chart recorder, and the LVP was recorded on magnetic tape [3].
Left ventricular function
Left ventricular systolic function was assessed from maximal systolic elastance (Emax), the slope of a line fit by least squares regression to the peak LVP-LV volume relation [10], obtained from function curves at constant coronary perfusion pressure and heart rate as described previously [9]. Left ventricular end-diastolic pressure (EDP) as a measure of LV chamber stiffness, and the LV relaxation constant, 
[11], obtained from the digitized, high-fidelity LVP data as previously described, were measured at constant LV volume, coronary perfusion pressure, and heart rate [9].
Chemical measurements
Blood samples for oxygen content determination were drawn anaerobically into glass syringes, capped, and held at 4°C until analysis. Blood gases and oxygen content, plasma ionized calcium, and serum sodium and potassium were measured as previously described [3].
Magnetic resonance spectroscopy
Phosphorus-31 spectra were acquired in a 4.7-T GE-Omega MRS system using a surface coil double-tuned to 1H and 31P as detailed previously [3, 9]. Before data acquisition, the field was shimmed to approximately 80 Hz using the 1H water signal. A 3-mm glass sphere filled with dimethyl phosphate containing 8 mmol/L chromium acetyl acetonate was fixed to the coil for use as an intensity standard. The radiofrequency pulse width was 40 milliseconds. Spectra were obtained continuously in approximately 3-minute blocks of nine acquisitions using a pulse delay of 20 seconds. This pulse delay has been demonstrated to be adequate for complete relaxation of all nuclei of interest at all temperatures from 10°C to 37°C. Data during baseline and recovery were summed over 15 minutes, and during cardioplegia or ischemia over 3 minutes. Peak areas for creatine phosphate (PCr), adenosine triphosphate (ATP), and the standard were obtained by fitting resonances to Lorentzian shapes using the NMR1 program (NMRI, New Methods Research, Inc, Syracuse, NY; Fig 1). For inorganic phosphate (Pi), peak intensity was used instead of peak area. Areas or Pi intensity was scaled with respect to the intensity standard and expressed as a percentage of their baseline values. B1-corrected peak areas were used to calculate the ratio of PCr to ATP. Intracellular pH (pHi) was calculated from the chemical shift of the Pi peak with respect to the PCr peak and corrected for myocardial temperature by the method of Kost [12]. When the Pi peak was indistinguishable from noise, as occurred later in cardioplegia and during the few minutes of cold ischemia, Pi was assigned a value of 0, and pHi could not be measured. During normothermic ischemia, PCr decreased and was assigned a value of 0 unless the peak in question was within the correct chemical shift range, its apex surpassed neighboring noise levels, and it resembled a Lorentzian shape. When PCr was 0, pHi was obtained from the chemical shift of the Pi peak with respect to the standard peak.
|
Experimental groups
Hearts were randomly assigned to one of three groups: group 1 (CI): 36 minutes of arrest at 10°C followed by 45 minutes of ischemia at 10°C; group 2 (WI): 36 minutes of arrest at 37°C followed by 45 minutes of ischemia at 37°C; or group 3 (I37): 45 minutes of ischemia at 37°C.
Protocol
Baseline
Coronary perfusion pressure was maintained at 86 to 94 mm Hg, and LV volume was adjusted until LVEDP was 6 to 10 mm Hg. When LV function was stable, control MRS and hemodynamic data were acquired, blood samples were drawn, and coronary blood flow was measured. Then LV function curves were obtained, and LV volume was returned to its baseline value.
Cardioplegia
Just before cardioplegic arrest, LV volume was decreased until LVEDP was 0 to 5 mm Hg, and pacing was discontinued. Immediately before and during cold cardioplegia, cold water was supplied to the water jacket cylinder and perfusion line heat exchanger. Cardioplegia was administered when myocardial temperature reached 32°C or immediately on arrest, if that was earlier. Blood from the arterial reservoir and the modified Fremes’s solution were continuously mixed in a ratio of 4 parts blood to 1 part solution and delivered to the heart through the coronary perfusion line. Arrest was induced by infusing the high-K+ blood-cardioplegia mixture at approximately 100 mL/min for 5 minutes followed immediately by a continuous infusion of low-K+ blood cardioplegia mixture for a total period of 36 minutes of arrest. Flow rates were checked by timed volumetric collections of coronary drainage and adjusted if necessary to maintain perfusion pressure above 30 mm Hg during cold or below 125 mm Hg during warm blood cardioplegia. A hemofiltrator was used to minimize hemodilution caused by cardioplegia [3].
Ischemia
After the baseline period in I37 or cardioplegic arrest in groups WI and CI, global ischemia was instituted for 45 minutes at 37°C or 10°C, as appropriate, by clamping the coronary perfusion line. During ischemia, blood was continuously circulated through the coronary arterial reservoir so that reperfusion was effected with fresh arterial blood.
Recovery
Immediately before resuming coronary perfusion and again 2 minutes later, lidocaine (1 mg/kg intravenously) was administered to the support dog. Coronary perfusion pressure was returned to 87 to 93 mm Hg. Electrical defibrillation was performed if necessary—after CI when myocardial temperature reached 32°C. Temporary atrioventricular block occurred in some hearts, requiring ventricular pacing; atrial pacing was resumed as soon as possible. After 45 minutes of recovery, the LV balloon was gradually inflated to baseline volume. At 60, 90, and 120 minutes of recovery, MRS data, hemodynamic data, and blood samples were collected, followed by a function curve, as at baseline. Finally, the position of the balloon within the LV was confirmed, and the LV was weighed.
Calculations
Left ventricular volume was calculated as the volume in the LV balloon plus balloon displacement volume. Coronary vascular resistance was calculated as coronary perfusion pressure divided by coronary flow per 100 grams of LV. Downstream pressure was assumed to be atmospheric because the right heart was drained by wide-bore tubing that opened to air at the level of the heart over the coronary venous reservoir. Myocardial oxygen consumption (M
O2) was calculated as the product of coronary blood flow per 100 grams of LV and the difference between coronary arterial and venous oxygen contents. Total blood cardioplegia volume was estimated as the product of the volume of crystalloid cardioplegic solution administered multiplied by 5 (because the volume of crystalloid cardioplegic solution represents one fifth the volume of blood-crystalloid cardioplegia), and normalized to an LV weight of 100 g.
Functional recovery
Left ventricular pressure, Emax, and during recovery were expressed as percentages of their respective baseline values. The change in EDP was calculated as LVEDP during recovery minus baseline LVEDP. Delay in pHi decline was calculated from linear regression of paired values of pHi and ischemic duration; time to reach a given value (x) of pHi equals (x - intercept)/slope. The difference in such times between groups WI and I37 indicates the delay attributable to prior warm blood cardioplegia.
Statistical analysis
Data were subjected to repeated measures analysis of variance. If the F statistic for the group or interaction between group and time was significant, groups CI and I37 were each compared with group WI by unpaired Student’s t test. Baseline data were contrasted with subsequent measures by paired Student’s t test if the F statistic for time or the interaction between group and time was significant. The incidence of ventricular fibrillation among the groups was compared by Fisher’s exact test. Continuous variables measured only once were compared by Student’s t test. Values of p less than 0.05 were taken to indicate statistical significance. Data are expressed as mean ± standard error of the mean.
These studies were performed in accordance with institutional guidelines and the National Institutes of Health "Guide for the Care and Use of Laboratory Animals" (NIH publication 85-23, revised 1985).
|
Results |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentReferences |
|---|
Baseline
Left ventricular weight and baseline volume, as well as baseline hemodynamic and metabolic variables, did not differ significantly among the groups (Table 1). At baseline, the hematocrit was 35.1% ± 1.0%, serum K+ 3.2 ± 0.1 mEq/L, and plasma ionized calcium 1.10 ± 0.02 mmol/L. These variables did not differ among groups nor change significantly from their baseline values during recovery.
|
|
O2 declined dramatically during arrest at either temperature and by 30 minutes of arrest was significantly lower at 10°C than 37°C (Fig 2). During the first 3 minutes of cardioplegic arrest when cooling contracture occurred, Pi increased transiently to a maximum of 229% ± 76% in CI, but decreased to 28% ± 18% in WI (p < 0.04), whereas PCr was 87% ± 5% in CI and 106% ± 4% in WI (p < 0.02). Inorganic phosphate then declined progressively in both groups. From 15 minutes of arrest, Pi did not differ significantly between CI and WI, later becoming undetectable in all but four hearts in CI and most in WI, thereby restricting pHi measurement. In CI, pHi increased to 7.85 ± 0.10 in those hearts in which it could be measured (Fig 2). Creatine phosphate increased slightly but significantly during cardioplegia at 37°C (Figs 2, 3). At 30 minutes, the small difference in PCr between the groups was also significant (Fig 2).
|
Magnetic resonance spectroscopy data for representative hearts are displayed in Figure 3, and mean results during ischemia, in Figure 4. Creatine phosphate declined only slowly during ischemia at 10°C, with more than half of baseline PCr remaining at the end of ischemia (Fig 4). Creatine phosphate fell more rapidly in WI, to less than 10% of baseline by 21 minutes of ischemia, and 0 by the end of the ischemic period. Adenosine triphosphate was equally well protected in CI and WI during the first 30 minutes of ischemia, declining at a similar rate to 71% ± 4% and 71% ± 7%, respectively (Fig 4). Thereafter the ATP decline in WI accelerated to reach 48% ± 5% by the end of the ischemic period, whereas it was maintained at 71% ± 4% in CI.
|
Figure 5 shows means of measurements of ATP during ischemia plotted against simultaneous mean measurements of PCr. Data from the two groups undergoing normothermic ischemia, WI and I37, form a single relationship, whereas data obtained during 10°C ischemia (CI) form a second distinct relationship. These relationships were fitted by linear regression to data in which PCr exceeded 6%. At 37°C, the theoretic point (100,100) is included, and ATP = 68.0 + 0.274 x PCr (r2 = 0.818; n = 13). At 10°C, ATP = 55.7 + 0.241 x PCr (r2 = 0.591; n = 15). The intercepts of these lines differ (p < 0.0001), but not the slopes. Adenosine triphosphate was lower at 10°C than at 37°C for any given value of PCr. Adenosine triphosphate declined at approximately one quarter the rate of PCr until PCr fell below 6%; then ATP declined precipitously but not to 0 as PCr declined further. By the end of ischemia, ATP was 39.3% ± 5.5% and 47.6% ± 4.8% in I37 and WI, respectively (not significant), whereas PCr was less than 1% in both groups.
|
A comparison of I37 to WI (Fig 4) shows that the declines in ATP, PCr, PCr/ATP, and pHi during ischemia at 37°C are significantly slowed by prior cardioplegia, but as ischemia is prolonged, the protective effect of prior cardioplegia is lost; by the end of ischemia, WI and I37 did not differ significantly with respect to any of these variables (Fig 4).
Recovery
Ventricular fibrillation followed reperfusion in 4 of 9 hearts in I37, 1 of 7 in WI, and 8 of 8 in CI (I37 versus WI, not significant; CI versus WI, p = 0.0014). Transient atrioventricular block was apparent in 6 of 7 in CI, 4 of 7 in WI, and 3 of 9 hearts in I37 (not significant). Aortic pressure was stable in all three groups, except for a significant but small decline during recovery in CI (137 ± 3 mm Hg at baseline, 130 ± 5 at 2 hours of reperfusion).
There were no important changes in LV function during the course of reperfusion, so only data at 2 hours are presented (Fig 6). Both Emax and MO2 recovered to baseline levels in CI but remained substantially depressed in WI and I37. Left ventricular end-diastolic pressure was significantly increased in all groups, with no significant differences between CI or I37 and WI. At 2 hours, was prolonged in WI, significantly more so in I37, but only insignificantly in CI. At 2 hours of reperfusion, PCr at 106% in CI was not significantly above its baseline value, but was substantially increased to 136% and 137% in WI and I37, respectively. Adenosine triphosphate was 96% ± 6% in CI but only 65% ± 4% in WI and 58% ± 6% in I37. Intracellular pH returned to baseline levels in all groups.
|
|
Comment |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentReferences |
|---|
O2 during recovery in hearts subjected to warm as compared with cold ischemia presumably largely reflects the decreased Emax in these hearts. Comparison of WI and I37 demonstrates that prior warm continuous blood cardioplegia slows metabolic deterioration during the early phase of normothermic ischemia (Figs 3, 4). Persistent contraction was detectable for 9 minutes into ischemia in hearts without prior cardioplegia; the associated energy consumption must contribute to the more rapid decline in high-energy phosphates [5]. In addition, the altered cellular environment produced by cardioplegia infusion, specifically the lowered ionized calcium and increased magnesium, are protective [14–16], reducing cellular calcium influx and its associated energy demands. However, when ischemia continued for 45 minutes, prior cardioplegia conferred little advantage on metabolic or functional recovery. In hearts subjected to normothermic ischemia, prior cardioplegia significantly enhanced recovery of LV relaxation, but neither diminished the severe LV systolic functional impairment nor prevented the ATP depletion and PCr excess (Fig 6), which typify ischemic damage.
Metabolic requirements are reduced by cold [3, 7, 8], except for the transient increased energy demand of cooling contracture suggested by our data in rat [17, 18] and canine hearts [3]. The initial transient surge in Pi and concomitant decline in PCr observed in the present study during cold but not warm arrest at the time of cooling contracture are consistent with hydrolysis of high-energy phosphates to meet the energy demands of cooling contracture. The pH increase with cold, observed previously [3, 6], is chiefly the result of temperature-related physicochemical changes in cell buffers.
In early ischemia, there is only limited ATP production by glycolysis [19], but PCr hydrolysis by creatine kinase buffers ATP concentrations [19, 20]. Our data illustrate this buffering action (Figs 4, 5); ATP declined initially at a similar rate in WI as in CI, but fell faster in I37 in the absence of cardioplegic protection (Fig 4). The differences in rates of ATP decline in WI and CI became apparent only when PCr had fallen to very low concentrations. The difference in ATP concentrations between WI and I37 is striking for the first 30 minutes of ischemia (Fig 4). When PCr was nearly exhausted in WI, ATP then rapidly declined to similar concentrations as in the normothermic hearts unprotected by cardioplegia.
A comparison of the present results to those of our previous study, using the same preparation, of alternating ischemic and cardioplegic periods [9] illustrates the protective value of repeated interruptions of ischemia [21]. In our former study [9], after the third 15-minute ischemic episode during cardioplegia (a total of 45 minutes’ ischemia), PCr, ATP, and pHi were 11% ± 2%, 91% ± 1%, and 6.80 ± 0.03, in contrast to 0.7% ± 0.7%, 48% ± 5%, and 6.36 ± 0.03, respectively, after 42 minutes of uninterrupted warm ischemia (WI) here. This protective effect may result from recovery, particularly of PCr, during the intervals of cardioplegic reperfusion and to downregulation of 5'-nucleotidase, the enzyme that hydrolyzes adenosine monophosphate [22].
In a prior study of prolonged ischemia by Whitman and colleagues [6], crystalloid-perfused rabbit hearts were subjected to 1 hour of ischemia at temperatures from 7°C to 37°C after a 2-minute infusion of oxygenated K+ cardioplegia. Hearts were arrested in cold saline solution before perfusion was initiated, a method involving a period of ischemia during preparation—which we avoided—and possibly consequent ischemic preconditioning [23]. The declines in PCr and ATP were substantially slower than we observed at comparable temperatures, which might be related to ischemic preconditioning or species or other methodologic differences. There was initial transient alkalinization during ischemia attributable to the creatine kinase reaction, observed by us previously [9] but not in the present study. Whitman and colleagues [6] also found the loss of ATP to be faster from the outset at 37°C than at 7°C or 17°C.
It is possible for temperature to change the creatine kinase equilibrium, thereby altering the PCr/ATP ratio. The enthalpy of the reaction relates the reaction equilibrium constant (Keq) to temperature. The measured enthalpy of -16 kJ · K-1 · mol-1 [24] corresponds to a 1.86 times increase in the Keq when the temperature declines from 37° to 10°C. From the Keq at 37°C [25], adjusted for the decline in temperature, the baseline reactant concentrations [25], and the observed increase in pH to 7.85 at 10°C, and a measured value of the ratio of 2.41 during arrest at 37°C, the PCr/ATP ratio is estimated to be 2.64 during cold cardioplegic arrest. The induction of arrest without a change in temperature also increases the ratio [26]. However these effects are not observable with the variance of the present data.
Our preparation allows the evaluation in a blood-perfused large heart of the myocardial effects of different temperatures without consideration of potentially differing systemic effects. Left ventricular function is evaluated while its main determinants, coronary perfusion pressure, LV volume, and heart rate are controlled. However, clinical application of the results is limited by possible species differences, and an absence of innervation and coronary collateral flow in our model.
In clinical practice, preexisting ischemic injury, valvular heart disease, and ventricular hypertrophy may increase the vulnerability of the human myocardium, and obstructive coronary artery disease and retrograde administration may contribute to suboptimal cardioplegic delivery. These factors reduce the safety margin and tolerance for ischemia during cardiac operation. The need to interrupt the delivery of cardioplegia varies according to the nature of the procedure: in typical practice, hypothermic cardioplegia is given at 20- to 30-minute intervals. From our observations in this and previous studies, 30 minutes of normothermic ischemia produces a profound metabolic decline, even with antecedent cardioplegic perfusion. To provide an adequate safety margin in the operating room, we try to limit interruption of warm cardioplegia to only 5 and always less than 10 minutes. In any procedure requiring longer intervals, hypothermia is a necessary adjunct. Whether any intermediate temperature can lessen the deleterious effects of cold while maintaining the safety margin is not addressed by our study but appears to be an attractive alternative.
|
Acknowledgments |
|---|
|
References |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  H. Yamamoto, K. Magishi, K. Goh, T. Sasajima, and F. Yamamoto Cardioprotective effects of normothermic reperfusion with oxygenated potassium cardioplegia: a possible mechanism Interactive CardioVascular and Thoracic Surgery, October 1, 2009; 9(4): 598 - 604. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. A. G. Louagie, J. Jamart, M. Gonzalez, E. Collard, S. Broka, L. Galanti, and A. Gruslin Continuous cold blood cardioplegia improves myocardial protection: a prospective randomized study Ann. Thorac. Surg., February 1, 2004; 77(2): 664 - 671. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. S. Rhodes, K. M. Ropella, S. H. Audi, A. K. S. Camara, L. G. Kevin, P. S. Pagel, and D. F. Stowe Cross-bridge kinetics modeled from myoplasmic [Ca2+] and LV pressure at 17{degrees}C and after 37{degrees}C and 17{degrees}C ischemia Am J Physiol Heart Circ Physiol, April 1, 2003; 284(4): H1217 - H1229. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Q. Chen, A. K. S. Camara, J. An, M. L. Riess, E. Novalija, and D. F. Stowe Cardiac preconditioning with 4-h, 17{degrees}C ischemia reduces [Ca2+]i load and damage in part via KATP channel opening Am J Physiol Heart Circ Physiol, June 1, 2002; 282(6): H1961 - H1969. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Q. Chen, A. K. S. Camara, J. An, M. L. Riess, E. Novalija, and D. F. Stowe Cardiac preconditioning with 4-h, 17{degrees}C ischemia reduces [Ca2+]i load and damage in part via KATP channel opening Am J Physiol Heart Circ Physiol, June 1, 2002; 282(6): H1961 - H1969. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
-adenosine triphosphate peak (D), 
-adenosine triphosphate peak (E), and ß-adenosine trisphosphate peak (F). Adenosine triphosphate is determined from the ß-adenosine triphosphate peak. (PPM = parts per million.)
