Oxygenation of cardioplegic solutions

Oxygenation of cardioplegic solutions

J THoRAc CARDIOVASC SURG 1987;94:614-25 Oxygenation of cardioplegic solutions Potential for the calcium paradox Oxygenation of crystalloid cardiopl...

1MB Sizes 0 Downloads 85 Views

J

THoRAc CARDIOVASC SURG

1987;94:614-25

Oxygenation of cardioplegic solutions Potential for the calcium paradox Oxygenation of crystalloid cardioplegic solutions is beneficial, yet bicarbonate-containing solutions equilibrated with 100% oxygen become highly alkaline as carbon dioxide is released. In the isolated perfused rat heart fitted with an intraventricular balloon, we recently observed a sustained contraction related to infusion of cardioplegic solution. In the same model, to record these contractions, we studied myocardial preservation by multidose bicarbonate-containing cardioplegic solutions in which first the calcium content and then the pH was varied. An acalcemic cardioplegic solution (Group 1) and the same solution with calcium provided by adding calcium chloride (Group 2) or blood (Group 3) were equilibrated with 100% oxygen. Ionized calcium concentrations were 0, 0.10 ± 0.06, and 0.11 ± 0.07 mmoljL and pH values were 8.74 ± 0.07,8.54 ± 0.08, and 8.40 ± 0.07, all highly alkaline. Hearts were arrested for 2 hours at 8° ± 2.5° C and reperfused for 1 hour at 37° C. At end-arrest, myocardial adenosine triphosphate was depleted in all three groups, significantly in Groups 2 and 3. In Group 1 the calcium paradox developed upon reperfusion, with contracture (left ventricular end-diastoUc pressure = 60 ± 7 mm Hg), creatine kinase release up to 620 ± 134 U fL, a profound further decrease in adenosine triphosphate to 1.9 ± 1.7 nmoljmg dry weight, and either greatly impaired or no functional recovery (17 % ± 10% of prearrest developed pressure). Three hearts in this group released creatine kinase during arrest and did not resume beating during reperfusion. In Groups 2 and 3, the calcium paradox did not occur; functional recovery was 61 % ± 4% and 71 % ± 9% at 5 minutes of reperfusion. In two additional groups (4 and ~ the pH of the acalcemic cardioplegic solution was decreased by equilibration with 2% and 5% carbon dioxide in oxygen to 7.53 ± 0.03 and 7.11 ± 0.02. Contractions during arrest were smaller than in Groups 1, 2, and 3; adenosine triphosphate was maintained during arrest; functional recovery was 101 % ± 3 % and 96 % ± 4 % at 5 minutes of reperfusion. We conclude that acalcemic solutions with carbon dioxide are superior to highly alkaline calcium-containing solutions. H oxygenation of cardioplegic solutions, of proved value, causes severe alkalinity, then calcium paradox may result even with hypothermia. This hazard is prevented by adding calcium or blood to the solution or carbon dioxide to the oxygen used for equilibration.

William G. Hendren, MD,* Gillian A. Geffin, MB, BS, Tim R. Love, MD, James S. Titus, Brian E. Redonnett, Dennis D. O'Keefe, MD, and Willard M. Daggett, MD, Boston, Mass.

In both experimental models and patients, oxygen supplied to the myocardium by cardioplegic solution From the Department of Surgery, Massachusetts General Hospital and Harvard Medical School, Boston, Mass. Supported in part by National Institutes of Health Grant HL 12777. Received for publication Sept. 25, 1986. Accepted for publication Nov. 17, 1986. Address for reprints: Willard M. Daggett, MD, Department of Surgery, Massachusetts General Hospital, Boston, MA 02114. *Present address: Department of Thoracic Surgery, The Emory Clinic, 1365 Clifton Rd., N.E., Atlanta, GA 30322.

614

(CS) during arrest is beneficial.!" Crystalloid solutions, fully oxygenated and cold, provide excellent myocardial protection." 3, 5, 6 Although vigorous oxygenation of a CS maximizes its oxygen content, equilibration of a bicarbonate-containing solution with 100% oxygen will decrease the dissolved carbon dioxide and result in a highly alkaline solution. The addition of carbon dioxide to the equilibrating gas mixture will increase the dissolved carbon dioxide and thereby lower the pH of the CS. The optimal pH of CS in general is uncertain'< and has not been determined for our solution. Calcium-free CSs provide successful myocardial protection in clinical use- 10-14 and their efficacy is supported

Volume 94 Number 4 October 1987

by certain laboratory studies.v" However, the safety of acalcemic CS has been questioned because of the potential for development of the calcium paradox," the severe myocardial cellular necrosis accompanied by contracture, leakage of enzymes and other cellular constituents, and failure to recover function, that occurs upon readmission of a calcium-containing perfusate after a period of calcium-free perfusion." Calcium is therefore added to many solutions, 18. 19 although hypothermia 20-24 and the provision of calcium by noncoronary collateral flow? minimize the risk of calcium paradox in the clinical setting. The present study was initially designed to determine, in the isolated rat heart," the effect of varying the percentage of carbon dioxide added to oxygen for equilibrating a bicarbonate-containing CS. Since we6 recently showed improved myocardial protection in the absence of calcium, we used an acalcemic CS to address this question. Preliminary experiments, however, demonstrated the calcium paradox on reperfusion after arrest by a CS equilibrated with 100% oxygen. Calcium was therefore added to the CS in additional experimen. tal groups, either as calcium chloride or in whole blood. A sustained contraction during the infusion of cold calcium-containing hyperkalemic CS has been recently observed in the isolated rat heart fitted with an intraventricular balloon." We chose the same preparation to relate CS calcium content and pH to the magnitude of this cardioplegic infusion-related contraction and to myocardial protection assessed by the preservation of myocardial high-energy phosphates and by left ventricular (LV) functional recovery.

Methods Hearts were obtained from male Sprague Dawley rats weighing 290 to 450 gm. The animals were heparinized (1,000 U intraperitoneally), anesthetized with 20 mg of pentobarbital intraperitoneally, and the heart was excised. The isolated heart was mounted on the apparatus developed by Yamamoto, Braimbridge, and Hearse" and then perfused via the aortic root at a pressure of 100 em H 20 by the method of Langendorff with KrebsHenseleit bicarbonate buffer (Table I) at 37° C, equilibrated with 95% oxygen and 5% carbon dioxide. The left atrium was opened and an LV apical drain was placed through the mitral anulus for dependent drainage. The pulmonary artery was incised to allow free drainage of coronary effluent and placement of a probe for measuring myocardial temperature. A water-filled balloon mounted on a catheter was inserted into the left

Calcium paradox and cardioplegia

615

Table I. Composition of solutions Krebs-Henseleit bicarbonate buffer

Na+ (mEqjL) K+ (mEqjL) Ca'" (mEqjL) Mg'" (mEqjL) ci (mEqjL) HC0 3- (mEqjL) H,PO,- (mEqjL) 80,-- (mEqjL) Glucose (mmoljL) Mannitol (mmoljL)

143.0 5.9 2.4 2.4 125.1 25.0 1.2 2.4 11.0

Basic CS 109.3 20.0

102.4 26.8

27.8 54.9

ventricle through the left atrium for measuring LV pressure under isovolumic conditions. The balloon volume was adjusted until a stable LV end-diastolic pressure between 8 and 12 mm Hg was obtained. The balloon itself generated no pressure at the volumes used. Coronary flow was measured by timed volumetric collection of effluent from the drained heart. Aortic pressure and LV pressure were measured (GouldStatham P23ID pressure transducers, Gould Inc., Cardiovascular Products, Oxnard, Calif.) and, together with LV end-diastolic pressure obtained by amplification, were recorded on a strip chart (Hewlett-Packard 7758B, Hewlett-Packard Company, Andover, Mass.). Prearrest control values of peak LV pressure, LV end-diastolic pressure, coronary flow, and heart rate obtained from the LV pressure recording were determined at the end of 20 minutes of Langendorff perfusion. LV developed pressure was calculated as peak LV pressure minus LV end-diastolic pressure. Hearts with a peak LV pressure less than 80 mm Hg, apparent coronary flow greater than 24 ml/rnin, or heart rate less than 250 beats/min were excluded from the study. At the end of the control period the heart was rendered globally ischemic by clamping the aortic perfusion line. Arrest was immediately induced by a hyperkalemic CS at 4° C, appropriate to the experimental group, which was infused at a constant pressure of 65 cm H 20 through a side arm on the aortic cannula. CS was infused every 15 minutes for 2 hours. The first infusion was terminated when the volume of coronary effluent reached 15 rnl and each subsequent infusion at 10 m!. During arrest the heart was maintained at 8° ± 2.5° C by a water-jacketed chamber surrounding it. The heart was then reperfused with Krebs-Henseleit buffer at 37° C for 1 hour. The recovery of LV developed pressure, LV end-diastolic pressure, and cor-

The Journal of

616

Hendren et al.

Thoracic and Cardiovascular Surgery

Table II. Experimental groups: CS modifications Group

Equilibrating gas

1.0, 2.0,-Ca 3.0,-BI 4.2% CO, 5.5% CO,

100% 0, 100% 0, 100% 0, 98% 0,/2% CO, 95% 0,/5% CO,

Pco/' pH*

(mm Hg)

Po;* (mmHg)

Ionized Ca++ (mmo/jL)

± ± ± ± ±

<8 <8 <8 31 ± I 78 ± 3

>800 >800 >800 >800 >800

0 0.10 ± 0.01 0.11 ± om 0 0

8.74 8.54 8.40 7.53 7.11

0.07 0.08 0.07 0.03 0.02

Hematocrit (%) 0 0 3.6 ± 0.2 0 0

Legend: Pco, Carbon dioxide tension. Po" Oxygen tension. BI, Blood. Values are means ± standard error of the mean (SEM).

'Measured at 37° C.

Table

m. Creatine kinase activity

in coronary effluent (UjL)

Arrest (cardioplegic infusion No.)

8

1 min

10 ± 1 10 ± 2

55 ± 25 21 ± 3

223 ± 56 17 ± 2t

12 ± 2 10 ± 0.2

24 ± 3 24 ± 3

16 ± It 19 ± It

± ± ± ± ±

Group 1. 0, 2. O,-Ca 3.0,-BI 4.2% CO, 5. 5% CO,

Reperfusion

I

4

I

395 12 15 38 25

45 It 2t

nt

3t:l:II

I

2 min 619 II 28 31 28

± ± ± ± ±

92 It 16t 8t 3t§

I

3 min 601 II 14 34 27

± ± ± ± ±

117 0.5t 2t 9t 2t§~

I

4 min 620 9 13 35 25

± ± ± ± ±

134 3* 2* 9*:1: 5*:1:

5 min

I

586 9 14 35 21

± ± ± ± ±

124 3* 2* 8':1: 3*:1:

Legend: Values are mean ± SEM. Each value was calculated from the results of 10 to 12 hearts during arrest or four to six hearts during reperfusion.

'p < 0.01 versus Group I. tp < 0.005 versus Group 1. tp < 0.05, Group 4 or 5 versus 2. §p < 0.005, Group 4 or 5 versus 2. lip < 0.05, Group 4 or 5 versus 3. '\[p < 0.005, Group 4 or 5 versus 3.

onary flow were measured at 5, 15, 30, 45, and 60 minutes of reperfusion. Coronary effluent from the first, fourth, and eighth CS infusions and effluent collected during each of the first 5 minutes of reperfusion were analyzed for creatine kinase activity with the Worthington stazyme kit (Cat. No. 27251, Cooper Biomedical, Malvern, Pa) and a Gilford Model 3402 spectrophotometer (Gilford, Oberlin, Ohio). In parallel groups of hearts at the end of the prearrest, arrest, and reperfusion periods, approximately 100 mg of ventricular tissue was excised, rapidly frozen by compression between paddles at -70° C, and divided into three. Each portion was assayed for ATP, ADP, AMP, and creatine phosphate by high-pressure liquid chromotographyand the results were averaged.v " Total adenine nucleotides were calculated as the sum of ATP, ADP, and AMP. Percentage heart water was determined by desiccation at 90° C to a constant weight of samples of the remainder of the ventricles. The composition of the basic CS is shown in Table I. The experimental groups differed with respect to calcium content and the pH of the CS, which was determined by the percentage of carbon dioxide added to oxygen for equilibration (Table II). In Group I (oxygen), the acalcemic CS was equilibrated with 100%

oxygen. In Group 2 (oxygen-calcium), the CS was also fully oxygenated and 0.12 mmol of calcium chloride was added to each liter of CS to give an ionized calcium concentration of 0.10 ± 0.01 mmol/L, In Group 3 , (oxygen-blood), the CS was equilibrated with 100% oxygen and fresh heparinized whole rat blood was added to give a similar ionized calcium concentration, 0.11 ± 0.01 mmol/L, to the CS in Group 2 and a hematocrit value of 3.6% ± 0.2%. In. Groups 4 (2% carbon dioxide) and 5 (5% carbon dioxide), the acalcemic crystalloid CS was equilibrated with 2% and 5% carbon dioxide in oxygen, respectively, which produced progressively more acid solutions. The highly alkaline acalcemic CS, Group 1 (oxygen), may be compared first to the same CS with calcium added either as calcium chloride (Group 2) or in blood (Group 3) and second to the less alkaline acalcemic solutions (Groups 4 and 5). The sequence of experiments was determined by a randomized block design. Ionized calcium was measured by an ionized calcium analyzer (NOVA Biomedical, Waltham, Mass). A blood-gas analyzer (Radiometer, Inc., Copenhagen, Denmark) measured CS pH, carbon dioxide tension, and oxygen tension at 37° C. All rats received humane care in compliance with the "Guiding Principles in the Care and Use of Animals" approved by the Council of the American Physiological

Volume 94 Number 4 October 1987

Calcium paradox and cardioplegia 6 1 7

Fig. 1. Representative recordings displaying the end of the prearrest period, arrest interval, and first minutes of reperfusion for one heart in each group. The higher aortic root pressure, at the beginning and end of each strip, indicates perfusion with Krebs-Henseleit solution at 37° C. The square wave increases in aortic root pressure indicate cold CS infusions. Note a rise in LV pressure (LVPj during each cardioplegic infusion. On reperfusion the heart in Group I (oxygen), arrested by an acalcemic solution equilibrated with 100% oxygen, has an abrupt and sustained rise in LV diastolic pressure (LVDPj and does not resume beating.

Society (revised 1980) and the "Guide for the Care and Use of Laboratory Animals" (NIH Publication No. 78-83, revised 1985). Statistical analysis was undertaken by univariate analysis of variance (BMDP Statistical Software, Inc.,

Los Angeles, Calif.). When analysis of variance rejected the hypothesis of equal means among the groups or among experimental periods, the significance of the differences between specific pairs of means was tested by the t test. The following null hypotheses were

The Journal of Thoracic and Cardiovascular

61 8 Hendren et al.

50

40 ........ .....

Surgery

° °°

GfOIJP f: 0 too% 2 , mM CO"" 2: 0 too% 2 , O.t mM CO" J: • 100% 02• Blood 4: '" 9B% °2/2% CO2 ' OmM Co'" 5: 'l 95% 02! 5%co 2, mM Co'"

°

III

t:: 30

~ CI.. ~

§:

° °° 2,

mM CO""

2: 0

tOO% 2 . 0.1mM Co"

5:

95% 02I 5%CO 2,

° °

I

I

J: • 100%O2 • Blood 4: '" 98% O2 I 2%CO 2 , mM Co'"

70

'l

mM Co""

60

~

'-

G{()(JP t: 0 tOO%

50

20

-..J

10

0

2345678

20

CARDIOPLEGIC DOSE NUMBER

Fig. 2. Cardioplegic infusion-related contractions. Note that the pressure developed depends on both the CS calcium content and pH. The developed pressure during the first few infusions is highest in Group 2, hearts given CS containing calcium chloride. In hearts given acalcemic CS (Groups 1,4, and 5), larger contractionsoccur when the CS is most alkaline (Group I). % LVP, Difference between peak LVP during the contraction and prearrest LVEDP, expressed as a percentage of prearrest peak LVP. Bars indicate ± SEM. For clarity, only selected bars are shown in this and subsequent figures.

examined: (1) that the addition of calcium, blood, or carbon dioxide to a fully oxygenated acalcemic CS was without effect (Group I versus each of Groups 2 to 5); (2) that the addition of calcium as calcium chloride or blood was equivalent to the addition of carbon dioxide (Groups 2 or 3 versus 4 or 5); and (3) that in any group arrest or reperfusion was without effect. Probability values of less than 0.05 were taken to be statistically significant. Data are expressed as mean ± standard error of the mean (SEM).

Results Cardioplegic infusion-related contractions. A sustained contraction, measured as an increase in LV pressure, occurred during CS infusions. In all hearts the pressure rapidly rose upon arrest from LV end-diastolic pressure to reach a maximum during or at the end of the CS infusion (Fig. 1). When the infusion ceased, the pressure declined rapidly at first and then more slowly until the next CS infusion. The peak pressure was influenced by the calcium content and pH of the CS: During the first hour of arrest, the addition of calcium to CS as calcium chloride or blood increased the contrac-

10

o

Longendorff Con1rol ~

/

o

I

15

30

I

45

I

60

REPERFUSION (minutes)

Fig. 3. LV end-diastolic pressure (EDP) during reperfusion. Group I (oxygen) showan abrupt and sustainedrised in EDP. In this group, three hearts included in the data in the figure never resumed beating. In the remaining three hearts EDP was lower at 49.0 ± 11.4 mm Hg at 5 minutes of reperfusion and decreased to 29.3 ± 10.3 mm Hg at 60 minutes of reperfusion. The remaining groups have only small and transient increases in EDP above the prearrest values.

tion magnitude whereas the addition of carbon dioxide decreased it. Fig. 2 depicts the developed pressure' of the contractions during each CS infusion. In successive infusions this pressure was equal to or less than that of the preceding infusion except during the last few CS infusions in Groups 1 (oxygen) and 3 (oxygen-blood). The effects of calcium to increase and carbon dioxide to decrease contraction magnitude are confirmed in Fig. 2. Developed pressures in Group 2 (oxygen-calcium) were significantly higher than in Group 1 (oxygen) (p < 0.001, for each of the first five infusions). Developed pressures in Group 3 (oxygen-blood) were also initially higher than in Group I (oxygen), but the difference attained statistical significance only during the third infusion (p < 0.05). In some hearts in this group, however, CS delivery was slower and the accompanying contractions were prolonged (Fig. 1). Of the three groups given acalcemic CS, Group 1 (oxygen) had the largest contractions, the group with the most alka-

Volume 94 Number 4

Calcium paradox and cardioplegia

October 1987

120

° ° ° ° ° ° °

61Oi1p f: 0 100% 2 , mM Co" 2· 0 100% 2 . 0.1mM Co" 3: • 100% 2 • Blood 4: t:. 98% 2 / 2% CO2 , OmM Co" ;: '"

95% 2 / 5% CO 2 ,

61Oi1p f:

0

4:

t:.

2: 0 3: •

140

mM Co"

~

s. '"

°° °

6 19

100% 2 , mM Co" 100% 2 . 0.1mM Co" 100% O2 • Blood 98% °212% CO2 , mM Co" 95% 2 / 5%CO 2 , mM Co"

° °

°

80

~ ~

Q::

60

8

40

~ ~

o

I

o

I

15

I

30

I

45

I

60

REPERFUS/ON (minutes)

Fig. 4. LV functional recovery during reperfusion. Note that there is poor functional recovery throughout the reperfusion period in Group I (oxygen). Data from three hearts in Group I that did not resume beating and had zero developed pressure are included in the figure. Function was severely impaired in the remaining hearts in Group I, which had a developed LVP (% prearrest) of 34% ± 13%, 38% ± 12%, 48% ± 11%, 51% ± 13%, and 47% ± 10% at 5, 15,30,45, and 60 minutes of reperfusion, respectively.

line of the three CSs: The contractions were significantly diminished throughout arrest by the addition of carbon dioxide. Creatine kinase. Table III displays creatine kinase measured in coronary effluent obtained during the first, fourth, and eighth CS infusions and during each of the first 5 minutes of reperfusion. Eight of the 12 hearts in Group 1 (oxygen) (three of the six subsequently reperfused) had substantial creatine kinase leakage by the eighth CS infusion, which suggested cellular damage during arrest. On reperfusion, there was substantial creatine kinase leakage in all six hearts, the amount peaking within the first 5 minutes. In several hearts in Group 1, the effluent was also assayed for concentrations of serum glutamic oxaloacetic transaminase and lactic dehydrogenase, both of which were elevated in a parallel manner. The addition of calcium to the CS in Group 2 (oxygen-calcium) prevented creatine kinase leakage during arrest and reperfusion. Red cells in the CS in Group 3 (oxygen-blood) interfered with the

1o

0

I

I

15

I

30

I

45

I

60

REPERFUS/ON (minutes)

Fig. 5. Coronary flow after reperfusion as a percent of prearrest control.

creatine kinase assay and accounted for elevated creatine kinase levels during arrest (not shown); the sanguineous CS itself, not infused, had a similar elevated level of creatine kinase activity. In Group 3 (oxygenblood), however, in which the CS also contained calcium, there was no significant creatine kinase leakage during reperfusion. In neither Group 4 (2% carbon dioxide) nor Group 5 (5% carbon dioxide), with the more acid acalcemic CSs, was there creatine kinase leakage during arrest, but in some animals in these groups the creatine kinase leakage during the first 5 minutes of reperfusion was above the normal range for the assay. Recovery of LV function. On reperfusion in Group 1 (oxygen), there was an immediate increase in diastolic pressure to 60 ± 7 rom Hg. Diastolic pressure remained high throughout the reperfusion period, which indicates contracture. In the remaining groups there was only a transient rise in LV end-diastolic pressure during early reperfusion (Fig. 3) (p < 0.05 versus Group 1). After reperfusion, hearts in Group 1 felt stiff and were paler and more yellow than hearts in the other groups. Fig. 4 shows LV developed pressure during reperfusion as a percent of its prearrest value. Three of the six hearts in Group 1 (oxygen), those with enzyme leakage

The Journal of

Hendren et al.

6 20

Thoracic and Cardiovascular Surgery

Table IV. Myocardial concentrations of adenine nucleotides and creatine phosphate (nmolfmg dry weight) Group

1.0, 2. O,-Ca

3.0,-B1 4.2% CO, 5. 5% CO,

n

ATP

6

12.6 ± 2.0

6 6 7 7 6

8.2±1.4 5.6 ± 0.7# 7.1 ±0.6# 13.2 ± 0.6t§~ 12.8 ± 1.0t§~

ADP

AMP

Prearrest 8.4 ± 0.6 5.3 2 hr arrest 6.6 ± 0.6 3.7 7.1 ± 0.2 8.8 8.0 ± 0.5 5.6 8.2 ± 0.4 3.2 8.3 ± 0.7 3.4

TAN

CP

± 1.7

26.1 ± 1.2

15.8 ± 5.0

± ± ± ± ±

18.6 ± 21.5 ± 20.6 ± 24.6 ± 24.6 ±

14.5 ± 13.5 ± 15.3 ± 21.7 ± 19.6 ±

0.5 1.2t# 0.7 0.5§11 0.5§11

1.8** 1.2# 0.9** 0.9~1

1.3'111

2.0 1.2 1.6 1.6 3.8

Legend: Values are mean ± SEM. ATP, Adenosine triphosphate. ADP, Adenosine diphosphate. AMP, Adenosine monophosphate. TAN, Total adenine nucleotides.CPo Creatine phosphate. 'p < 0.05 versus Group I. tp < 0.005 versus Group I. :j:p < 0.05, Group 4 or 5 versus 2. §p < 0.005, Group 4 or 5 versus 2. lip < 0.05, Group 4 or 5 versus 3. 'lIP < 0.005, Group 4 or 5 versus 3. #p < 0.05 versus prearrest. "p < 0.005 versus prearrest. ttp

< 0.05

versus 2 hour arrest. :j::j:p < 0.005 versus 2 hour arrest.

Table V. Heart water percent 2 hr arrest Group

n

1.0, 2. 02-Ca 3.0,-BI 4.2% CO, 5. 5% CO,

6

[

6

7 7 6

1 hr reperfusion

Heart water (%)

n

81.4 78.8 78.4 81.0 81.1

6 6 6 6 6

± 0.6 ± 0.3* ** ± 0.2t** ± 0.3§~ ± 0.5§~

I

Heart water (%)

85.8 82.0 82.8 82.5 83.3

± ± ± ± ±

0.7:1::1: 0.8tH 0.5tH 0.7t 0.2*H

Legend: Prearrest control. 81.1% ± 0.4%. Values are mean ± SEM. 'p :j:p

< 0.05

versus Group 1. tp < 0.005 versus Group I.

< 0.05 Group 4 or 5 versus 2. §p < 0.005 Group 4 or 5 versus 2. < 0.05 Group 4 or 5 versus 3. '\[p < 0.005 Group 4 or 5 versus 3. #p < 0.05 versus prearrest. "p < 0.005 versus prearrest. ttp < 0.05 versus 2 hour arrest. :j::j:p < 0.005 versus 2 hour arrest.

I"

during arrest, did not resume beating and so generated zero LV developed pressure. Hearts in this group recovered only 17% ± 10% of prearrest LV developed pressure at 5 minutes of reperfusion (including the three zero values). Function remained severely impaired: LV developed pressure was significantly less throughout reperfusion than in any other group (p < 0.01). The remaining four groups had good functional recovery. At 5 minutes ofreperfusion, Group 2 (oxygen-calcium) and Group 3 (oxygen-blood) recovered 66% ± 4% and 71 ± 9% of prearrest LV developed pressure, whereas Groups 4 (2% carbon dioxide) and 5 (5% carbon dioxide) recovered 101% ± 3% and 96% ± 4% of prearrest LV developed pressure, respectively (p < 0.004, Group 2 or 3 versus 4; p < 0.03, Group 2 versus 5). At 15 minutes, Groups 4 and 5 retained some advantage (p < 0.02, Group 2 or 3 versus 4; p < 0.03, Group 2 versus 5). After 45 minutes of reperfusion, LV developed pressure

was similar in Groups 2, 3, 4, and 5, averaging between 68% and 73% of prearrest LV developed pressure. Fig. 5 shows that coronary flow in Group 1 (oxygen) was below the prearrest value at 5 minutes of reperfusion, lower than in Groups 2, 4, and 5 (p < 0.002), and that it decreased progressively throughout reperfusion. In Groups 2, 4, and 5, coronary flow was initially above the prearrest value, which indicates vasodilation consistent with reactive hyperemia. Coronary flow in Group 3 (oxygen-blood) was initially intermediate between flow in Group I and in the remaining groups. Myocardial adenine nucleotides and creatine phosphate (Table IV). In Group 1 (oxygen) at end-arrest, total adenine nucleotides were decreased significantly from the prearrest value although the decreases in the individual nucleotides did not reach statistical significance. In Groups 2 (oxygen-calcium) and 3 (oxygenblood), both ATP (Fig. 6) and total adenine nucleotides

Volume 94 Number 4 October 1987

Calcium paradox and cardioplegia

ATP

n

ADP

AMP

TAN

621

CP

1 hr reperfusion 6 6 6 5 6

1.9 7.2 9.3 8.6 11.5

± ± ± ± ±

0.7:j::j: 0.9t IAt 0.5tH 0.9t§

2.6 ± 0.5:j::j: 7.8 ± 0.2t 704 ± 0.7t 8.3 ± OAt 804 ± 0.6t

decreased significantly during arrest. The addition of carbon dioxide to acalcemic CS, in Groups 4 (2% carbon dioxide) and 5 (5% carbon dioxide), maintained both ATP and total adenine nucleotides at prearrest levels. Moreover, at end-arrest ATP was higher and AMP lower in Groups 4 and 5 than in Groups 2 and 3. In Group 1 (oxygen) there were profound decreases in ATP, ADP, AMP, total adenine nucleotides, and creatine phosphate during reperfusion: ATP decreased from 8.2 ± 1.4 nmol/rng dry weight at end-arrest to 1.9 ± 0.7 nmol/rng (Fig. 6) and was below 1 nmoljmg in the three hearts that leaked creatine kinase during arrest. At the end of reperfusion ATP, ADP, total adenine nucleotides, and creatine phosphate were all significantly lower than in any other group. In Groups 2 and 3, there was no significant loss of adenine nucleotides during reperfusion. Although ATP (and creatine phosphate) declined during reperfusion in Group 4 (2% carbon dioxide), AMP increased and total adenine nucleotides were maintained. In Group 5 (5% carbon dioxide), with the most acidic of the acalcemic CS, adenine nucleotides and creatine phosphate did not alter significantly during reperfusion from their end-arrest values. Heart water decreased during arrest in hearts given calcium-containing CS (Groups 2 and 3) (Table V). Heart water increased in all groups during reperfusion, to the greatest extent in Group 1 (oxygen); in the three hearts that leaked creatine kinase during arrest, heart water exceeded 87% after reperfusion.

Discussion In this study, hearts that were arrested with an acalcemic CS equilibrated with 100% oxygen developed the calcium paradox upon reperfusion, manifested by (1) enzyme leakage" 22-24. 28.29 (2) sustained contracture with very high diastolic pressures,": 20. 22-24. 28.29 (3) severe impairment or no recovery of systolic function, 17.22-24.28.29

± ± ± ± 404 ±

1.8 6.2 4.9 6.6

0.5tt 1.1t 1.2* 0.7tH 1.0

6.3 21.2 21.7 23.5 24.2

± ± ± ± ±

1.6:j::j: OAt

i.st

0.9t 1.1t:j:

2.5 9.5 13.7 10.8 14.6

± ± ± ± ±

0.6:j::j: 1.3t tt 3.6* 2.9*tt 2.5t

(4) a severe decline in high-energy phosphates from end-arrest values,21.28 and (5) discoloration". The calcium paradox did not occur when small amounts of calcium, as calcium chloride or in blood, or carbon dioxide were added to the CS. The calcium paradox was first reported by Zimmerman and Hulsmann" in 1966 in the normothermic rat heart. Upon readmission of a calcium-containing perfusate after a few minutes of perfusion with a calcium-free solution, there was massive cellular damage characterized by the rapid onset of contracture, the loss of intracellular constituents, and irreversible mechanical failure. The calcium paradox occurs in all species tested so far 24. 28 and is probably the most severe form of acute myocardial necrosis that can be produced experimentally." The calcium paradox has been attributed to calcium washout from the sarcolemma leading to uncontrolled calcium influx on subsequent calcium repletion."" This may activate calcium-dependent enzymes, accounting for the loss of high-energy phosphates and necrosis.v" During calcium depletion, the glycocalyx splits and lifts and the intercalated disks partially separate but remain attached at the nexus junctions.v-":" There is also cellular sodium loading and some calcium loss." Ganote and Nayler" have postulated that on repletion, initial small calcium entry through physiologic channels triggers contracture, which then disrupts the weakened intercalated disks and allows massive enzyme release and additional uncontrolled calcium entry." This hypothesis is supported by enzyme release in the absence of calcium repletion when caffeine or dinitrophenol initiate contracture.v" We observed substantial creatine kinase release by the eighth CS infusion, before calcium repletion, in some hearts arrested with the very alkaline acalcemic solution. In these hearts the CS infusion-related contractions may have disrupted the cells and allowed immediate uncontrolled calcium entry upon reperfusion. These hearts did not resume beating and their degree of contracture and high-energy phos-

The Journal of

622

Hendren et al.

Thoracic and Cardiovascular Surgery

Croup f.' 100% 0" 0 mM Co"

0,.

e.

100% 0,,0.1 mM Ca" Blood 3.' 100% 4. 9B%0,/2% CO,. 0 mM ce5: 95% 0,/5% CO, • 0 mM Ca"

End Arrest 14

o

12345

1 hr. Reperfusion

I

I

12345

GROUP NUMBER

Fig. 6. MyocardialATP concentrations at the end of 2 hours of cardioplegic arrest and after 1 hour of reperfusion. Note the steep decline of ATP in Group 1 (oxygen) during reperfusion. phate depletion was the most severe. In the other hearts in this group, which resumed beating although severely damaged, some repair may have accompanied repletion. The calcium paradox did not occur in hearts given CS with added calcium or blood. The protection afforded by blood can be attributed to the provision of calcium, because minute amounts of calcium, as little as 25 or 50 J.Lmol/L, prevent the paradox.v" Bielecki" demonstrated that decreasing the pH of the calcium-free perfusate is protective, whereas increasing the pH has the opposite effect. We observed the calcium paradox in hearts arrested by an acalcemic, highly alkaline CS but not when the CS was acalcemic but had added carbon dioxide. As described by the HendersonHasselbalch equation, the pH of a solution containing a given amount of bicarbonate is determined by its carbon dioxide content. Equilibration of a bicarbonate solution with 100% oxygen, as in Group 1, drives off carbon dioxide and results in severe alkalinity. After equilibration with 100% oxygen, the pH of our CS, which contained a bicarbonate concentration of 27 mrnol/L, was 8.7 when measured at 37° C. This pH would be slightly higher if corrected to the 4 ° C at which the CS was delivered." The addition of carbon dioxide to the equilibrating gas made the CS more acid and prevented the paradox. Adding carbon dioxide does not decrease CS oxygenation to any important degree. The prolonged period of ischemia" or calcium depletion" and the hyperkalemic CS29 present in all experimental groups and known to predispose to the calcium paradox may

have contributed, together with severe alkalinity," to producing the paradox in Group 1. We found that hypothermia to 8° ± 2.5° C did not prevent the calcium paradox. Hypothermia is generally considered to be very protective, preventing damage to the intercalated disk and glycocalyx and slowing sodium/calcium exchange." Previous work has shown that the paradox occurs when periods of calcium depletion exceed 3 minutes at 37° C, whereas ultrastructure and function are preserved during normothermic repletion after 10 minutes' depletion at 4 ° C.20 Histologic damage and high-energy phosphate loss on calcium repletion are largely prevented at 30° c.2I When the calcium-free period is prolonged, however, the paradox can be induced at temperatures as low as 20° C." Mechanical failure is induced by a milder insult than is required for massive creatine kinase release." Baker, Bullock, and Hearse" confirmed and extended these observations, finding that creatine kinase release was greatly reduced by mild hypothermia but there was no functional recovery until the temperature during 10 minutes' depletion was lowered to 20° C; at 15° C and below, functional recovery was complete. The present study appears to be the first report of the fully developed calcium paradox at temperatures as low as 10° C, which indicates that deep hypothermia is not invariably protective. The present study confirms that infusion of cold hyperkalemic calcium-containing CS in the rat heart produces a prolonged contraction. This phenomenon is described in a companion paper in this JOURNAL by Torchiana and associates" and is akin to the potassium contracture described by Niedergerke." Adding calcium to the highly alkaline CS increased the contraction amplitude, which supports earlier observations that the phenomenon is calcium related.w" The 'occurrence of contractions with the highly alkaline but acalcemic solution may reflect calcium release from the sarcoplasmic reticulum, which has a decreased affinity for calcium at a more alkaline pH.38 The singular importance of the contraction may be that it is an energy-consuming event." Our data support this concept. Hearts given highly alkaline calciumcontaining CS (Groups 2 and 3) contrasted with hearts given the more acid acalcemic CS (Groups 4 and 5) in several respects. Myocardial ATP was depleted after 2 hours of arrest in Groups 2 and 3, the groups with the greatest contractions, whereas ATP was maintained at prearrest levels in Groups 4 and 5, where the contractions were small. Functional recovery was better in Groups 4 and 5, reflecting higher end-arrest ATP levels, a relationship observed before.s'? Calcium in CS may

Volume 94 Number 4

Calcium paradox and cardioplegia

October 1987

contribute to energy depletion by its effect on the contraction and ensuing relaxation" and by other actions.":" Although CS infusion-related contractions hve not been observed in man, we" have demonstrated a similar event in the canine heart. The optimal pH for CS is not known. Proponents of alkaline CS7argue that for optimal enzyme function the pH of CS, like the pH of blood,should remain about 0.6 pH units above neutrality, which is 6.8 at 37° C and 7.3 at 8° C.36 Certain studies show enhanced protection, however, by relatively acidic CS.8,9 In the present study, the more acidic calcium-free CSs were superior to the more alkaline calcium-containing CSs, The severealkalinity may have accentuated the effects of calcium, because hydrogen ions appear to oppose the actions of calcium at the sarcolemma and within the myocyte.": 42, 43 Only hearts giventhe most acid CS, Group 5 (5% carbon dioxide), showed no depletion of highenergy phosphates during arrest or reperfusion. Although Tyers" cautioned against calcium-free CS in 1975, there are data which argue for and against the inclusion of calcium in CS, Calcium-free CSs at varying degrees of hypothermia have proved deleterious in someI 8, 24,29, 44-47 but not a1l 6,15 laboratory studies, The isolated rat heart at 28° C was poorly protected by the Bretschneider and Kirsch solutions, both calcium-free, compared to the calcium-containing St. Thomas' Hospital solution." Bretschneider solution, formulated with histidine, was deleterious in the isolated rat heart at 20° C44,45 and in the dog heart at the higher temperatures studied, although the dog heart recovered from 2 hours of perfusion when the temperature was lowered to 15° C.24 Myocardial necrosis, attributed to the calcium paradox, in the canine heart in situ followed a 4V2 hour arrest at 27° C with an acalcemic CS only when the CS had both a high potassium content and histidine buffered at a high pH,46 Also in the canine heart in situ, Jacocks and colleagues" found no difference in protection by a bicarbonate-buffered CS with or without calcium, whereas Heitmiller and associates," after prolonged arrest (5 hours) at 10° C with precautions to minimize the provision of calcium by noncoronary collateral flow,' found that LV function declined during extended reperfusion when a calcium-free CS was used, In clinical studies, calcium-free CSs are accepted as satisfactory."?" Indeed, omitting calcium can improve myocardial protection: In the isolated rat heart arrested at 8° C, an acalcemic bicarbonate-buffered CS equilibrated with 98% oxygen/2% carbon dioxide provided better functional recovery and energy protection than the same solution with added calcium." These advan-

623

tages of omitting calcium were revealed when the paradox was prevented by the combination of deep hypothermia and a relatively low pH. In some hearts arrested by the same acalcemic CS with added carbon dioxide in the present study, despite excellent functional recovery there was some creatine kinase leakage during reperfusion. This argues for the inclusion of trace amounts of calcium in any perfusate. In summary, less alkaline acalcemic CSs provide superior myocardial protection to highly alkaline calcium-containingsolutions. This may be related to superior energy preservation associated with the smaller cardioplegic infusion-related contractions. Although oxygenation of CSs is highly desirable, equilibration of a bicarbonate-containing solution with 100% oxygen results in severe alkalinity. If a calcium-free CS is equilibrated with oxygen without any added carbon dioxideand is made extremely alkaline thereby, there is the potential for the calcium paradox developing upon reperfusion even with hypothermia during arrest. The hazard of paradox may be averted, and the proved benefits of oxygenation retained, by the addition of a small amount of calcium or blood to the solution, or carbon dioxide to the equilibrating gas mixture. We gratefully acknowledge the high-pressure liquid chromatographic analyses byAlvin G, Denenberg, James E, Vath, and Douglas Malnati, other chemical analyses by Carmelo Bondi and Richard P, Wawrzynski, technical assistance by Matthew Parker, and preparation of the manuscript by Mary Chasse and Judy Feiner. REFERENCES Bodenhamer RM, DeBoer LWV, Geffin GA, et al. Enhanced myocardial protection during ischemic arrest: oxygenation of a crystalloid cardioplegic solution, J THORAC CARDIOVASC SURG 1983;85:769-80, 2, Flaherty JT, Jaffin JH, Magovern GJ Jr, et al. Maintenance of aerobic metabolism during global ischemia with perfluorocarbon cardioplegia improves myocardial preservation, Circulation 1984;69:585-92. 3. Guyton RA, Dorsey LMA, Craver JM, et al. Improved myocardial recovery after cardioplegic arrest with an oxygenated crystalloid solution. J THORAC CARDIOVASC SURG 1985;89:877-87. 4. Novick RJ, Stefaniszyn HJ, Michel RP, Burdon FD, Salerno TA. Protection of the hypertrophied pig myocardium: a comparison of crystalloid, blood, and Fluosol-DA cardioplegia during prolonged aortic clamping. J THORAC CARDIOVASC SURG 1985;89:547-66. 5. Randolph JD, Toal KW, Geffin GA, et al. Improved myocardial preservation with oxygenated cardioplegic solutions as reflected by on-line monitoring of intramyocardial pH during arrest. J Vase Surg 1986;3:216-25. 6. Boggs BR, Torchiana OF, Geffin GA, et al. Optimal 1.

624

The Journal of Thoracic and Cardiovascular Surgery

Hendren et al.

myocardial preservation with an acalcemic crystalloid cardioplegic solution. J THORAC CARDIOVASC SURG 1987; 93:838-46. 7. Buckberg GD. A proposed "solution" to the cardioplegic controversy. J THORAC CARDIOVASC SURG 1979;77:'~3­ 15. 8. Nugent WC, Levine FH, Liapis CD, LaRaia PJ, Tsai CH, Buckley MJ. Effect of the pH of cardioplegic solution on post-arrest myocardial preservation. Circulation 1982;(Pt 2):68-72. 9. Bernard M, Menasche P, Canioni P, et al. Influence of the pH of cardioplegic solutions on intracellular pH, high-energy phosphates, and postarrest performance. J THORAC CARDIOVASC SURG 1985;90:235-42. 10. Adappa MG, Jacobson LB, Hetzer R, Hill JD, Kamm B, Kerth WJ. Cold hyperkalemic cardiac arrest versus intermittent aortic cross-clamping and topical hypothermia for coronary bypass surgery. J THORAC CARDIOVASC SURG 1978;75:171-8. II. Roberts AJ, Spies SM, Sanders JH, et al. Serial assessment of left ventricular performance following coronary artery bypass grafting: early postoperative results with myocardial protection afforded by multidose hypothermic potassium crystalloid cardioplegia. J THORAC CARDIOVASC SURG 1981;81:69-84. 12. Jacocks MA, Fowler BN, Chaffin JS, et al. Hypothermic ischemic arrest versus hypothermic potassium cardioplegia in human beings. Ann Thorac Surg 1982;34:157-65. 13. Fremes SE, Weisel RD, Mickle DAG, et al. Myocardial metabolism and ventricular function following cold potassium cardioplegia. J THORAC CARDIOVASC SURG 1985; 89:531-46. 14. Preusse CJ, Winter J, Schulte HD, Bircks W. Energy demand of cardioplegically perfused human hearts. J Cardiovasc Surg 1985;26:558-63. 15. Jacocks MA, Weiss M, Guyton RA, et al. Regional myocardial protection during aortic cross-clamp ischemia in dogs: calcium-containing crystalloid solutions. Ann Thorac Surg 1981;31:454-63. 16. Tyers GFO. Metabolic arrest of the ischemic heart. Ann Thorac Surg 1975;20:91-4. 17. Zimmerman ANE, HiiIsmann We. Paradoxical influence of calcium ions on the permeability of the cell membranes of the isolated rat heart. Nature 1966; 211:646-7. 18. Jynge P, Hearse DJ, Braimbridge MV. Myocardial protection during ischemic cardiac arrest: a possible hazard with calcium-free cardioplegic infusates. J THORAC CARDIOVASC SURG 1977;73:848-55. 19. Yamamoto F, Braimbridge MV, Hearse DJ. Calcium and cardioplegia: the optimal calcium content for the St. Thomas' Hospital cardioplegic solution. J THORAC CARDIOVASC SURG 1984;87:908-12. 20. Holland CE, Olson RE. Prevention by hypothermia of paradoxical calcium necrosis in cardiac muscle. J Mol Cell Cardiol 1975;7:917-28. 21. Bulkley BH, Nunnally RL, Hollis DP. "Calcium para-

22.

23.

24.

25. 26.

27.

28.

29.

30.

31. 32.

33.

34.

35.

36. 37.

dox" and the effect of varied temperature on its development: a phosphorus nuclear magnetic resonance and morphologic study. Lab Invest 1978;39:133-40. Boink ABTJ, Ruigrok TJC, de Moes D, Maas AHJ, Zimmerman ANE. The effect of hypothermia on the occurrence of the calcium paradox. Pfliigers Arch 1980; 385:105-9. Baker JE, Bullock GR, Hearse DJ. The temperature dependence of the calcium paradox: enzymatic, functional and morphological correlates of cellular injury. J Mol Cell Cardiol 1983;15:393-411. Gebhard MM, Bretschneider HJ, Mezger VA, Preusse CJ, Spieckerrnann PG. Principles to avoid the Ca++ paradox in myocardial protection. In: Isselhard W, ed. Myocardial protection for cardiovascular surgery. Munich: Pharmazeutische Verlagsgesellschaft, 1981:7482. Neely JR, Rovetto MJ. Techniques for perfusing isolated rat hearts: methods in enzymology. 1975;39:43-60. Torchiana DF, Love TR, Hendren WG, et al. Calciuminduced ventricular contraction during cardioplegic arrest. J THORAC CARDIOVASC SURG 1987;94:606-13. DeBoer LWV, Ingwall JS, K10ner RA, Braunwald E. Prolonged derangements of canine myocardial purine metabolism after a brief coronary artery occlusion not associated with anatomic evidence of necrosis. Proc Nat! Acad Sci USA 1980;77:5471-5. Hearse OJ, Humphrey SM, Boink ABTJ, Ruigrok TJe. The calcium paradox: metabolic, electrophysiological, contractile and ultrastructural characteristics in four species. Eur J Cardiol 1978;7:241-56. Jynge P. Protection of the ischemic myocardium: calciumfree cardioplegic infusates and additive effects of coronary infusion and ischemia in the induction of the calcium paradox. Thorac Cardiovasc Surg 1980;28:303-9. Ruigrok TJe. Cardioplegia and the calcium paradox. In: Caldarera CM, Harris P, eds. Advances in studies on heart metabolism, Bologna, Italy. CLliEB, 1982;35460. Ganote CE, Nayler WG. Contracture and the calcium paradox. J Mol Cell Cardiol 1985;17:733-45. Muir AR. The effects of divalent cations on the ultrastructure of the perfused rat heart. J Anat 1967;101 :23961. Crevey BJ, Langer GA, Frank JS. Role of Ca 2+ in maintenance of rabbit myocardial cell membrane structural and functional integrity. J Mol Cell Cardiol 1978; 10:108I-100. Ganote CE, Grinwald PM, Nayler WG. 2,4-0initrophenol (ONP)-induced injury in calcium-free hearts. J Mol Cell Cardiol 1984;16:547-57. Bielecki K. The influence of changes in pH of the perfusion fluid on the occurrence of the calcium paradox in the isolated rat heart. Cardiovasc Res 1969;3:268-71. Rahn H. Body temperature and acid-base regulation. Pneumonologie 1974;151:87-94. Niedergerke R. The potassium chloride contracture of the

Volume 94 Number 4 October 1987

38.

39.

40. 41.

42.

heart and its modification by calcium. J Physiol 1956; 134:584-99. Nakamaru Y, Schwartz A. The influence of hydrogen ion concentration on calcium binding and release by skeletal muscle sarcoplasmic reticulum. J Gen Physiol 1972; 59:22-32. Reibel OK, Rovetto MJ. Myocardial ATP synthesis and mechanical function following oxygen deficiency. Am J Physiol 1978;234:H620-4. Katz AM. Excitation-contraction coupling. In: Physiology of the heart. New York: Raven Press, 1977:137-59. Lehninger AL. Ca 2+ transport by mitochondria and its possible role in the cardiac contraction-relaxation cycle. Circ Res 1974;34,35(Suppl III):III-83-90. Langer GA. The effect of pH on cellular and membrane calcium binding and contraction of myocardium: a possible role for sarcolemmal phospholipid in EC coupling. Circ Res 1985;57:374-82.

Calcium paradox and cardioplegia

625

43. Fabiato A, Fabiato F. Effects of pH of the myofilaments and the sarcoplasmic reticulum of skinned cells from cardiac and skeletal muscles. J Physiol 1978;276:23355. 44. Ruigrok TJC, deMoes D, Borst C. Bretschneider's histidine-buffered cardioplegic solution and the calcium paradox. J THORAC CARDIOVASC SURG 1983;86:412-7. 45. Hendriks FFA, Jonas J, van der Laarse A, Huysmans HA, van Rijk-Zwikker GL, Schipperheyn JJ. Cold ischemic arrest: comparison of calcium-free and calciumcontaining solutions. Ann Thorac Surg 1985;39:312-7. 46. del Nido PJ, Wilson GJ, Mickle DAG, et al. The role of cardioplegic solution buffering in myocardial protection. J THORAC CARDIOVASC SURG 1985;89:689-99. 47. Heitmiller RF, DeBoer LWV, Geffin GA, et al. Myocardial recovery after hypothermic arrest: a comparison of oxygenated crystalloid to blood cardioplegia. The role of calcium. Circulation 1985;72(Pt 2):11241-53.