The effects of calcium and magnesium in hyperkalemic cardioplegic solutions on myocardial preservation

J THORAC CARDIOVASC SURG 1989;98:239-50

The effects of calcium and magnesium in hyperkalemic cardioplegic solutions on myocardial preservation Sustained left ventricular pressure development during each inf~ion of a cold calcium-containing hyperkalemic cardioplegic solution has been observed in rat hearts. The present study was undertaken to relate such contraction (i.e., increase in resting pressure) to myocardial preservation and to the calcium and magnesium contents of a crystaUoid hyperkalemic cardioplegic solution. Isolated perfused rat hearts with a left ventricular isovolumic baBoon were arrested at 8° C by the fuDy oxygenated cardioplegic solution infused every 15 minutes for 2 hours. Cardioplegic solutions containing ionized calcium in concentrations of 0,0.1, or 1.2 mmoljL were each studied with (groups 2, 4, and 6) and without (groups 1, 3, and 5) the addition of magnesium (16 mmoljL). Hearts arrested by the cardioplegic solution with no calcium or magnesium (group 1) developed a pressure (averaged over the second to eighth inf~ion and expressed as percent prearrest left ventricular pressure) of 6.0% ± 0.4% during cardioplegic infusions. This solution maintained end-arrest myocardial adenosine triphosphate (13.1 ± 1.0 nrnoljmg dry weight) and phosphocreatine (21.7 ± 2.8 nrnoljmg dry weight) contents near the prearrest contents and preserved left ventricular function at 95% ± 3% of prearrest developed left ventricular pressure at 15 minutes of reperfusion at 37° C. Calcium (groups 3 and 5) increased pressure development during cardioplegic inf~ions (10.4% ± 0.5% and 15.1 % ± 0.9%), depleted adenosine triphosphate (7.2 ± 1.0 and 7.4 ± 0.9) and phosphocreatine (13.3 ± 1.8 and 10.7 ± 1.5), and depressed left ventricular functional recovery (71 % ± 1 % and 73 % ± 3 %). Magnesium alone (group 2) decreased pressure development during cardioplegic infusions (3.0 % ± 0.3 % ), maintained adenosine triphosphate (15.6 ± 0.9), augmented phosphocreatine (38.3 ± 1.2), and preserved left ventricular function (99 % ± 4 %). Magnesium added to calcium (groups 4 and 6) prevented the calcium-induced increased pressure development during cardioplegic infusions (4.0% ± 0.5% and 6.7% ± 0.6%), maintained adenosine triphosphate (13.6 ± 1.4 and 14.9 ± 0.7), augmented phosphocreatine (31.3 ± 1.6 and 32.2 ± 2.4), and ameliorated the depression of functional recovery (82 % ± 2 % and 86 % ± 2 %). These data suggest that left ventricular pressure development during arrest contributed to calcium-induced energy depletion and impairment of functional recovery and that these deleterious effects were inhibited by magnesium. The inhibitory effects of magnesium on left ventricular pressure development were rapidly reversed on reperfusion. The data support the addition of magnesium to calcium-containing cardioplegic solutions and perhaps to calcium-free solutions.

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

From the Department of Surgery, Massachusetts General Hospital, and the Harvard Medical School, Boston, Mass. Supported in part by National Institutes of Health Grant HL 12777. Received for publication Jan. 15, 1988. Accepted for publication Nov. 30, 1988. Address for reprints: Gillian A. Geffin, MB, BS, Department of Surgery, Massachusetts General Hospital, Boston, MA 02114.

Calcium, a normal extracellular constituent, is a component of many cardioplegic solutions. Those solutions containing blood necessarily include calcium,':' and calcium is added to certain crystalloid cardioplegic solutions whose formulations are based on the composition of extracellular fluid.r" The presence of calcium in cardioplegic solutions precludes the possibility of the calcium paradox,"? the severe necrosis that can occur on 239

The Journal of

240

Thoracic and Cardiovascular Surgery

GejJin et al.

Table I. Composition of solutions

Na+ (mEqjL) K+ (mEqjL) Ca2+ (mEqjL) Mg2+ (mEqjL) ci- (mEqjL) HCO,- (mEqjL) H,PO,- (mEqjL) 50/- (mEqjL) Glucose (mmoljL) Mannitol (mmoljL)

Krebs-Henseleit bicarbonate buffer

Base cardioplegic solution*

143.0 5.9 2.4 2.4 125.1 25.0 1.2 2.4 11.0

109.3 20.0

103.3 27.0

27.8 54.9

'Calcium chloride. magnesium sulfate. and sucrose were added to this solution in appropriate experimental groups and the solution was equilibrated with 98'if oxygen and 2'if carbon dioxide at 4 C before administration. as explained in the text.

readmission of calcium after a period of calcium-free perfusion. to On the other hand, although a cornerstone of myocardial preservation is the reduction of the energy requirements of the ischemic heart, calcium in certain solution formulations increases myocardial oxygen consumption t 1 and depletes energy stores. 12. 1J Magnesium appears to counteract these unwanted effects of calcium in cardioplegic solutions, improving myocardial preservation.P''"!" Recent work from our laboratory" provides additional evidence for the view'
measurement of ventricular pressure during arrest. Solutions with ionized calcium concentrations of 0, 0.1, and 1.2 mmoljL were studied, each solution with and without magnesium. The low ionized calcium concentration selected for study, 0.1 mmol/L, is the approximate concentration in the dilute blood cardioplegic solution that we' use clinically and is more than adequate to prevent the calcium paradox. 2o. 21 The high concentration, 1.2 mmol/L, is the normal plasma ionized calcium concentration." Magnesium was studied at a concentration of 16 mmol/L, near optimal in the St. Thomas' Hospital solution" and in the optimal range in our solution.l" All solutions were fully oxygenated before administration, since oxygenation of crystalloid cardioplegic solutions is of established benefit in experimentaJl J · 2J.27 and clinical" settings. Myocardial preservation was assessed by myocardial high-energy phosphate concentrations and left ventricular functional recovery.

Methods Male Sprague Dawley rats weighing 290 to 540 gm were given 1000 units of heparin intra peritoneally and anesthetized 20 minutes later with pentobarbital in a dose of 60 mg/kg, given intraperitoneally. The heart was excised and placed in ice cold Krebs-Henseleit bicarbonate buffer (Table I), mounted on a perfusion apparatus," and perfused through the aortic root at a pressure of 100 cm H 20 with the same buffer at 37° C equilibrated with 95% oxygen and 5% carbon dioxide. A flanged polyethylene tube, inserted through the opened left atrium across the mitral valve, was pushed through the left ventricular apex; the flanged end was seated inside the apex and the tube cut short to form a drain for left ventricular thebesian flow. The pulmonary artery was incised to ensure free drainage of coronary venous effluent and for the insertion of a temperature probe into the right ventricle. A water-filled, balloon-tipped catheter was placed in the left ventricle through the left atrium. The balloon volume was adjusted until the left ventricular end-diastolic pressure was stable and between 8 and 12 mm Hg. The balloon itself generated no pressure at the volumes used. A water-jacketed chamber around the heart held myocardial temperature near 37° C. Coronary flow was measured by timed volumetric collection of effluent draining from the heart. Left ventricular full scale and diastolic pressures (measured through the balloon-tipped catheter) and aortic pressure were continuously recorded by pressure transducers (P23ID, Gould Inc., Oxnard, Calif.) on a strip-chart recorder (Hewlett-Packard, Andover, Mass.). Heart rate was obtained from the left ventricular pressure recording. Left ventricular developed pressure was calculated as the difference between left ventricular peak and end-diastolic pressures. Prearrest control values of the hemodynamic variables were measured after 20 minutes of perfusion. Hearts with a peak left ventricular pressure less than 80 mm Hg, apparent coronary flow less than 12 ml /rnin, or heart rate less than 250 beats/min were excluded from the study. In preparation for hypothermic arrest, a cold water-jacketed cardiac chamber replaced the warmed chamber and the aortic cannula was water-cooled. Then the aortic perfusion line was clamped, which rendered the heart globally ischemic, and arrest was immediately induced by infusing a cold hyperkal-

Volume 98 Number 2

Hyperkalemic cardioplegic solutions

August 1989

emic oxygenated cardioplegic solution at a pressure of 65 cm Hp through a side arm on the aortic cannula. Before and during arrest, the solution in a water-jacketed reservoir supplying the side arm was continuously equilibrated at 4° C with 98% oxygen and 2% carbon dioxide. This resulted in a pH of 7.35 to 7.50, a carbon dioxide tension of 30 to 35 mm Hg, and an oxygen tension greater than 800 mm Hg-the upper limit of the measurement-when measured at 3r C (Radiometer A/S. Copenhagen, Denmark) in samples of the solution drawn from a side arm on the delivery line just above the aortic cannula. The infusion pressure was controlled by setting the level of cardioplegic solution in the reservoir to a predetermined height above the heart before each infusion. An infusion of the solution was given every 15 minutes for 2 hours. In all hearts, the first infusion was ended when the volume of coronary drainage reached 15 ml and subsequent infusions at 10m\. The duration of each infusion was noted. Since the cardioplegic solution level changed little during each infusion, infusion pressure was almost constant. Coronary flow drained freely from the right side of the heart so that coronary venous pressure was near atmospheric. Therefore, mean coronary vascular resistance during each infusion was proportional to the reciprocal of coronary flow, that is, to infusion duration divided by infusion volume. During arrest, the heart was maintained at 8 ° ± 2.5 C. The heart was reperfused at 37° C for I hour with Krebs-Henseleit buffer during which time left ventricular peak and end-diastolic pressures, heart rate, and coronary flow were measured at 5, 15, 30, 45, and 60 minutes. Six experimental groups were studied, differing with respect to the calcium chloride (10% solution) and magnesium sulfate added to the base cardioplegic solution, whose composition is shown in Table I. Solutions with ionized calcium concentrations of 0, 0.1, and 1.2 mmol/L were each studied without magnesium sulfate (groups 1,3, and 5) and with magnesium sulfate in a concentration of 16 mrnol/L (groups 2, 4, and 6). Solutions without magnesium sulfate contained sucrose 32 rnmol/L instead, to give the same calculated osmolarity. Ionized calcium was measured by a Nova 2 calcium analyzer (Nova Biomedical, Waltham, Mass.). An additional group was perfused for 80 minutes with warm oxygenated buffer to provide control data for the same total duration of warm perfusion (20 minutes' prearrest plus 60 minutes' reperfusion) without intervening arrest. At the end of the prearrest, arrest, and reperfusion periods, experiments were terminated in parallel groups of hearts to assay myocardial contents of high-energy phosphates and percent heart water. After tissue adjacent to the ventricular drain was excised and discarded, approximately 100 mg of tissue was excised from the ventricular apex, rapidly frozen by compression between metal paddles precooled to -70° C, and then divided into three portions. Each portion was assayed for adenosine triphosphate (A TP), adenosine diphosphate (ADP), adenosine monophosphate (AMP), and phosphocreatine (PCr) by high-pressure liquid chromatography as previously described.P" and the average was expressed as nanomoles per milligram of dry weight. Total adenine nucleotides were calculated as the sum of ATP, ADP, and AMP. Percent heart water was assayed by desiccation at 90° C to constant weight in two samples taken from the remainder of the ventricles and the results were averaged. The following null hypotheses were tested: (I) that arrest and reperfusion had no effect; (2) that the addition of calcium to the CS at a concentration of either 0.1 or 1.2 mmol/L had 0

tt

mmHg ;----,

mMCa

o

::tl=_UP_f

'l :IL 'l

24 1

++

mMMg

0

_

3

40

0

01

a

01

16

12

a

12

16

r---1

,---,

5

40

0

r---1

::t~

Fig. l. The effect of calcium and magnesium in a hyperkalemic cardioplegic solution on left ventricular pressure during arrest. Chart recordings replayed from magnetic tape showing the induction of hypothermic arrest in one representative experiment in each group. Concentrations of magnesium and ionized calcium in this and the following figures refer to the cardioplegic solutions. The bracket above each recording indicates the timing of the first cardioplegic solution infusion. Left ventricular pressure was recorded by a pressure transducer attached to a water-filled balloon in the left ventricle (full details in text). (Prearrest peak left ventricular pressure is off-scale. The recordings have been slightly retouched in two respects: The prearrest tracings have been darkened, and some mechanical noise during arrest has been deleted). no effect in either the absence or presence of magnesium (group 3 or 5 versus I; group 4 or 6 versus 2); and (3) that the addition of magnesium to CSs containing calcium concentrations of 0, 0.1, or 1.2 rnmol/L had no effect (group 2 versus I; group 4 versus 3; group 6 versus 5). Myocardial concentrations of adenine nucleotides and PCr and percent heart water were subjected to a two-way analysis of variance (ANOYA) with experimental group and time (end arrest and end reperfusion) as the between-subjects factors (BMDP Statistical Software, Berkeley, Calif., program P7D). Only to enable comparison of

The Journal of Thoracic and Cardiovascular Surgery

2 4 2 Geffin et al.

INFUSION

A

~

INFUSIONS 2 to 8

1

B

-....J

~

E3 ~

Q.)

IIIID

iY6JI

40

OmM Mg+<" 16 mM Mg++

II')

~ S

~

~

~

~

~

20

~

'-

0

~ ::::s

12

~ :s -,

~

h::

~ ~

~

~

~

C

8

(.,)

Q.)

~

4

~

~ ~

0

Co++ (mMJ Fig. 2. A, Maximum left ventricular developed pressure (LVP) during the first cardioplegic infusion. The difference between peak pressure during the infusion and the prearrest left ventricular end-diastolic pressure is expressed as a percent of prearrest left ventricular pressure. B, Maximum left ventricular developed pressure expressed as in A and averaged over the second to eighth cardioplegic infusions. C, Duration of initial cardioplegic infusion divided by infusion volume (15 ml). This variable reflects the mean coronary vascular resistance during the infusion as explained in the text. D, Infusion duration divided by infusion volume (10 ml) averaged over the second to eighth cardioplegic infusions. In A, B, C, and D, each mean represents 12 to 23 hearts. Error bars are standard errors in this and the following figures.

prearrest values of these variables with values obtained at end arrest, the ANOVA was repeated with the prearrest data entered as a third level of the time factor in each group. Hemodynamic variables were subjected to repeated-measures ANOVA (P4V). When there were significant interactions between the main effects, data at single levels of each main effect were subjected to one-way ANOVA (P7D). When the ANOVAs rejected the hypothesis of equal means, the significance of the differences between specific pairs of means was tested by appropriate t tests. Differences were considered significant at p < 0.05. Data are expressed as mean ± standard error of the mean. 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 Society (revised 1980) and the "Guide for the Care and Use of Laboratory Animals" published by the National Institutes of Health (NIH Publication No. 85-23, revised 1985).

Results Fig. 1 shows representative left ventricular pressure recordings during arrest by each of the cardioplegic solutions. In each heart, diastolic pressure increased as peak left ventricular pressure decreased and continued to rise after beating ceased. A comparison of hearts given the magnesium-free solutions (groups 1, 3, and 5) shows that, after arrest, the maximum pressure and the rate of pressure development during infusion were related in a dose-dependent manner to calcium concentration of the solution. In groups 3 and 5, pressure remained substantially above its prearrest diastolic level for a considerable time after the infusion stopped; with a calcium concentration of 1.2 mmol/L, pressure was highest for the longest time. Comparisons between

Volume 98 Number 2

Hyperkalemic cardioplegic solutions

August 1989

20

50,---------------------,

I

.0mMMg"

I

.0mMMg*

~ 16mMMg"

Q..

......

"<

8

243

~ 16mMMg*

I

I 4

o

o

0.1

12

Co" (mM)

o

01

i 2

Co" (mMJ

Fig. 3. Myocardial ATP contents at 2 hours' arrest.

Fig. 4. Myocardial per contents at 2 hours' arrest.

recordings from groups 1 and 2, 3 and 4, and 5 and 6 show that the addition of magnesium to each of the three solutions decreased the peak pressure and the rate of pressure development and resulted in a more rapid return to, or below, the prearrest diastolic pressure. The dose-related effects of calcium on pressure development, however, remain discernible in the presence of magnesium. Fig. 2, A and B. depicts group mean values of left ventricular developed pressure as a percent of prearrest pressure during the first infusion of cardioplegic solution and averaged over the second to the eighth infusion. The addition of calcium alone, 0.1 or 1.2 mmol/L, to the cardioplegic solution increased developed pressure significantly in the first and subsequent infusions (p < 0.005, groups 3 or 5 versus group I). Developed pressures were highest with a calcium concentration of 1.2 mmol/L. When magnesium was added to each of the three solutions, the developed pressure decreased significantly (p < 0.005, group 2 versus 1, group 4 versus 3, group 6 versus 5). Developed pressure was higher in the first than in subsequent infusions and, for the most part, decreased in successive infusions in all groups. Fig. 2, C shows for the initial infusion that the duration per milliliter of cardioplegic solution infused, at near constant pressure, was prolonged by calcium at either concentration (p < 0.005, group 3 or 5 versus 1), which indicates increased mean coronary vascular resistance. Also, as reflected by the infusion duration per

milliliter, magnesium in the absence of calcium in the cardioplegic solution had no effect on mean coronary vascular resistance during the first cardioplegic infusion, but prevented the increase in resistance with a calcium concentration of 0.1 mmol/L and diminished the increase with a concentration of 1.2 mrnol/L (p < 0.005, group 4 versus 3, group 6 versus 5). In subsequent infusions (Fig. 2, D), differences between groups were less marked or absent. Myocardial ATP and PCr contents at end arrest are depicted in Figs. 3 and 4. Myocardial contents of nucleotides and PCr, including statistics, are in Table II. ATP after 2 hours of arrest remained at the prearrest level in hearts protected by the acalcemic solution without magnesium (group 1). The addition of calcium reduced end arrest ATP significantly. When magnesium was included in the cardioplegic solution, ATP at end arrest was greater than before arrest, a significant effect with the acalcemic solution and the solution containing a 1.2 mmol/L concentration of calcium. There were no significant differences among the groups at end arrest in myocardial AMP and ADP, but total adenine nucleotides were decreased by the addition of calcium without magnesium, significantly so at the higher calcium concentration. Myocardial PCr remained at the prearrest level after 2 hours of arrest by the acalcemic cardioplegic solution without magnesium. Addition of calcium to this solution significantly decreased PCr at end arrest (Fig. 4, Table II). The addition of magnesium to all three solu-

The Journal of Thoracic and Cardiovascular Surgery

Geffin et al.

244

Table II. Myocardial contents of adenine nucleotides and phosphocreatine (nmoljmg dry weight) Group

CS Cae+

CS Mg'+

(mmol/L)

(mmolfl.)

n

ATP

ADP

AMP

8.0 ± 0.5

5.6 ± 0.8

per

TAN

Prearrest 17

11.4 ± 0.8

25.0 ± 1.0

16.9 ± 2.2

End 2 hr arrest 1 2 3 4 5 6

0 0 0.1 0.1 1.2 1.2

0 16 0 16 0 16

9 8 6 6 8 12

13.1 15.6 7.2 13.6 7.4 14.9

± ± ± ± ± ±

1.0 0.9* 1.0*§ 1.4~

0.9*§ 0.7*~

8.2 7.1 7.0 8.5 7.3 8.2

± ± ± ± ± ±

0.7 0.4 0.9 0.6 0.2 0.3

4.5 3.1 6.0 5.3 5.7 4.3

± ± ± ± ± ±

0.9 0.4 1.8 0.9 0.8 0.4

25.8 25.7 20.2 27.4 20.4 27.5

± ± ± ± ± ±

1.3 1.0 2.9 1.8 0.7*§

5.1 5.3 4.4 5.9 5.6 5.1

± ± ± ± ± ±

0.6 0.8 0.9 1.8 0.8 0.6

21.1 21.5 20.0 23.2 18.9 23.9

± ± ± ± ± ±

0.7* 1.4* 1.9 2.6 0.6

0.7~

21.7 38.3 13.3 31.3 10.7 32.1

± ± ± ± ± ±

1.4t~r

12.4 11.0 15.3 13.8 11.0 17.5

± ± ± ± ± ±

1.7* I.3t 1.6 2.7t 1.9 2.8t

2.8 1.1t~

1.8:j: 1.6t~::t

1.5§

End / hr reperfusion I 2 3 4 5 6

0 0 0.1 0.1 1.2 1.2

0 16 0 16 0 16

12 7 6 6 9

II

9.1 8.6 9.1 9.7 6.7 11.0

± ± ± ± ± ±

1.0* 0.7t 1.5 1.7 0.9 0.9t

6.9 7.6 6.5 7.6 6.7 7.8

± ± ± ± ± ±

0.3 0.6 0.9 0.9 0.4 0.4

1.1*~

Values are means ± standard error of the mean. CS. Cardioplcgic solution; ATP. adenosine triphosphate; ADP. adenosine diphosphate; A\1P. adenosine monophosphatc: TAN. total adenine nucleotidcs: PCr. phosphocreatine. •p < 0.05 2-hour arrest versus prcarrest or l-hour reperfusion versus 2-hour arrest. tp < 0.005 2-hour arrest versus prearrest or I-hour reperfusion versus 2-hour arrest. :j:p < 0.05. group 3 or 5 versus group I. §p < 0.005. group 3 or 5 versus group 1. ~p < 0.005. group 2 versus I, group 4 versus 3. or group 6 versus 5. rp < 0.05. group 4 or 6 versus group 2.

tions increased end-arrest PCr substantially above the prearrest level; these increases were all highly significant. In the groups in which A TP and PCr were maintained or increased during arrest, they decreased during reperfusion; total adenine nucleotides showed a similar pattern (Table II). After 1 hour of reperfusion, there were no significant differences among the groups in adenine nucleotides or PCr. There were marked differences among the groups in functional recovery during early reperfusion. In hearts protected by the cardioplegic solution without calcium or magnesium (group 1), developed pressure (Table III, Fig. 5) did not differ significantly from its prearrest value at 5 or 15 minutes of reperfusion but then declined. When magnesium was added to the acalcemic solution (group 2), this level of recovery was sustained for 30 minutes of reperfusion. The addition of calcium in the absence of magnesium substantially depressed the recovery of developed pressure during initial reperfusion, equally at either calcium concentration. Magnesium decreased the magnitude of this depression; developed pressure attained its prearrest level at 5 minutes of reperfusion when magnesium was added to the solution with a 1.2 mmol/L calcium concentration, but, unexpectedly, only 84% of the prearrest level with a 0.1 mmol/L calcium concentration. The function of hearts

with greater developed pressure during early reperfusion deteriorated later, whereas that of hearts with an initially lower developed pressure improved; by 60 minutes of reperfusion, there were no significant differences among the groups. In the control group, developed pressure at 60 minutes of perfusion was 91% ± 2% (n = 6) of its value at 20 minutes, and at 80 minutes of perfusion it was 82% ± 1% (n = 5), similar to results in group 2 for comparable total durations of warm perfusion (Table III). Left ventricular end-diastolic pressure was above or at the prearrest pressure during early reperfusion and declined to or below the prearrest pressure by 60 minutes of reperfusion (Table III). At 5 and 15 minutes of reperfusion, end-diastolic pressure was significantly increased above prearrest levels in hearts in group I. The addition of either magnesium alone (group 2) or calcium alone at the higher concentration (group 5) further increased end-diastolic pressure substantially and significantly at 5 minutes of reperfusion. This increase was absent or significantly diminished with those solutions containing both magnesium and calcium. During reperfusion, heart rate remained close to its prearrest value in all groups. Coronary flow at constant perfusion pressure was significantly higher than prearrest flow at 5 minutes of reperfusion, except in hearts given the cardioplegic

Volume 98 Number 2

Hyperkalemic cardioplegic solutions

August 1989

24 5

100 ~

""-

~ ~

90

~

~

Q::

80

~

~

70

~

o (mM)

o Group 1

0

60

• Group 2

0

16

o Group 3

!j

50

• Group 4

0.1 0.1

0 16

~

(; Group 5

1.2

40

•

1.2

0 16

-...J

~

~

()

:::>:

0

Group6

15

30

45

60

MINUTES OF REPERFUSION Fig. S. Recovery of left ventricular function during reperfusion after 2 hours of cardioplegic arrest.

solution contammg a calcium concentration of 1.2 mmol/L without magnesium (Table III). By 15 minutes of reperfusion, coronary flow had declined to significantly below the prearrest level in all groups except in hearts given the cardioplegic solution with magnesium and no calcium. However, the decline in coronary flow was significantly less at this time in the hearts given calcium-containing solutions if magnesium had been included as well. By 30 minutes of reperfusion there were no significant differences in coronary flow among the groups. Discussion

This study shows that the addition of magnesium to a calcium-containing hyperkalemic cardioplegic solution improves myocardial preservation. End-arrest myocardial high-energy phosphates and early functional recovery were both substantially decreased by adding calcium to a magnesium-free hyperkalemic cardioplegic solution. Functional recovery was impaired to the same extent by 0.1 and 1.2 mmol/L concentrations of calcium. Magnesium prevented calcium-induced ATP depletion, elevated end-arrest PCr above its prearrest content, and substantially ameliorated the impairment of functional recovery. Peak pressure development during cardioplegic infusions was increased by calcium in a dosedependent manner, an effect that was prevented or substantially diminished by magnesium. These observations suggest that consumption of energy by contractile

activity during arrest, manifest as increased resting pressure, contributes to impaired early recovery. The benefits of adding magnesium to a calcium-free hyperkalemic cardioplegic solution were less marked, possibly since the calcium-free solution without magnesium itself provided good protection. With this latter cardioplegic solution, ATP and PCr were not depleted during arrest and left ventricular developed pressure did not fall significantly below prearrest levels during the first 15 minutes of reperfusion. The addition of magnesium increased end-arrest ATP and PCr above their prearrest levels and prolonged the maintenance of developed pressure at prearrest levels to 30 minutes of reperfusion; at 60 minutes of reperfusion, developed pressure had declined no more than in control hearts that underwent warm perfusion for the same total duration without intervening arrest. Magnesium, however, exacerbated the otherwise modest increase in diastolic pressure, indicating increased ventricular stiffness, that occurred transiently during early reperfusion after arrest with the acalcemic cardioplegic solution. Pressure development in the rat heart on infusing a calcium-containing hyperkalemic cardioplegic solution was recently reported by Torchiana and associates. IS Similar pressure development occurs in canine hearts." As in the present study and previous studies by us? and others," tension development was related to the solution calcium concentration. IS Both the relation of developed pressure to cardioplegic solution composition and ultra-

The Journal of Thoracic and Cardiovascular Surgery

246 Geffin et at.

Table

m. Hemodynamic variables cs

Graup

Ca l+ (mmolfl.)

CS Mg'+

(mmolil.i

0

2

3

4

5

6

0

0.1

0.1

1.2

1.2

0

16

0

16

0

16

Reperfusion

n

12 devP

7

6

6

9

II

5 min

Prearrest

EDP CF devP EDP CF devP EDP CF devP EDP CF devP EDP CF devP EDP CF

104±4 9.2 ±0.4 17± I 99±5 9.6±0.6 19±2 112±4 10.5 ±0.3 18± I 107±4 10.3±O.2 19±1 IOO±5 9.5 ± 0.5 18± I 109±4 8.6 ±O.4 19± I

%devP

97±2

~EDP

1.7 ±0.7*

O/OCF o/odevP ~EDP

O/OCF 'JodevP ~EDP

%CF o/odevP ~EDP

O/OCF o/odevP ~EDP

%CF o/odevP ~EDP

O/OCF

122 ± 4t 98±3 6.9 ± 2.2*11 117 ± 5* 55 ±4t§ 2.7± 1.1 119±5* 84±2 t'll# 0.2± 1.0** 106± 2* 51 ± 3t§ 9.1 ± I.3t§ 106±6 96±3'll 3.0±1.5'll 110±4*

/5 min

95±3 3.5±1.4* 91 ±3* 99±4 7.5 ± 3.4 95±4 71 ± It§ 1.7 ± 0.8 76 ± 3t§ 82 ± 2~1** 0.5±0.8 89 ± 3*11 73 ± 3t§ 3.4± 1.2* 81 ±4 tt 86 ± 2t'll** 2.1 ± 1.2 87 ± 3t

30 min

85 ± 3t 1.3±1.4 81 ±3t 94± 511 4.0±3.0 85 ±4* 75 ± 3 tt -0.7±0.7 71 ±2t 76 ± 2t** -1.5 ±0.6 78 ±6* 73 ± 2 t§ O.O±O.4 72±3t 83 ±2 tll= -O.6±O.7 83 ± 3t

45 min

83 ± 3t 0.5 ± 1.3 80± 3t 88 ±4* 2.9±2.9 84±4* 74±3t -1.0±0.7 71 ±4t 72 ± 3t** -2.3±0.6* 76±6* 71 ± 2tt -1.1 ±O.2t 75 ± 4t 78± 3t= -1.3 ±O.4* 79±4t

60 min 76± 2t 0.1 ± lA 77±3t 82± S' 2.4 ± 2.9 77±3 t 71 ±3t -1.7 ±O.6* 69±4t 67 ±4t -2.3±O.4t 73±6* 72±2 t -1.8 ±O.2 t 76± st 73 ±4t -1.8±O.4 t 7g ±

st

Values are means ± standard error of the mean. CF. Coronary flow (rnl/rnin): CS. cardioplegic solution; devP. developed left ventricular pressure (rnrn Hgj: FUP. left ventricular end-diastolic pressure (0101 Hg): 'if.devP, percent prearrcst developed left ventricular pressure; '?,CF. percent prearrest coronary flow: .lEDP. ditfercncc between EDP during reperfusion and prearrest EDP (0101 Hg) . •P < 0.05 versus prearrest. tp

< 0.005 versus prearrest.

;J:p < 0.05. group 3 or 5 versus group I. §p < 0.005, group 3 or 5 versus group I.

lip < 0.05. group

2 versus I, group 4 versus 3, or group 6 versus 5. 1[p < 0.005, group 2 versus 1, group 4 versus 3, or group 6 versus 5.

< 0.05, group 4 or 6 versus group 2. **p < 0.005, group 4 or 6 versus group 2.

#p

sonic measurements of myocardial length" indicate that pressure during cardioplegic infusions results from myocardial force generation and not solely from the hydrostatic effect of the infusions themselves. Pressure development during cardioplegic infusions resembles potassium contracture." Generally, pressure declines between such infusions (Fig. 1). We previously showed, however, that if energy production is limited by omitting either oxygen or substrate" or both" from a cardioplegic solution containing calcium, ischemic contracture, characterized by sustained pressure development (see Fig. 5 of reference 15) and profound highenergy phosphate depletion." develops after repeated cardioplegic infusions. Arrest for a similar period after a single infusion of such a cardioplegic solution, in contrast, does not cause ischemic contracture. 15 Since ischemic or hypoxic contracture is thought to result when ATP is depleted to the point that there is insufficient ATP for actin-myosin cross-bridge dissociation," these observations suggest that repeated infusions of calcium-containing cardioplegic solutions consume energy. Both myothermal measurements" and the rapid

force regeneration after quick releases of papillary muscles in potassium contracture but not in hypoxic contracture 19 are consistent with rapid cross-bridge cycling in potassium contracture and very slow cycling of the "rigor bonds" of hypoxic contracture. Accordingly, the transient pressure development stimulated by cardioplegic infusions is distinct from energy depletion contracture; infusion-related pressure development appears to contribute to rather than result from energy depletion, whereas ischemic contracture results from energy depletion and, once initiated, further depletes the heart of energy." Hearse, Stewart, and Braimbridge," in the isolated rat heart, found the addition of magnesium to a cardioplegic solution containing potassium (J 6 mmol/ L) and calcium (1.2 mmol/L) improved postischemic aortic flow, the optimal magnesium concentration being 15 mmol/L, Arrest was maintained at 28° C for 70 minutes or 37° C for 30 minutes after a single dose of cardioplegic solution. Also, the addition of magnesium in a concentration of 16 mmol/L to a solution containing potassium (5.9 mmol/L) and calcium (1.2 mmol/L)

Volume 98 Number 2 August 1989

decreased creatine phosphokinase leakage during reperfusion at any sodium level between 30 and 150 mmolj L. 16 An increase in potassium concentration to 16 mmoljL decreased the protective effect of magnesium except in that important sodium concentration range between 90 and 120 mrnol/L, where the protective effect was unchanged." Pernot and associates" found that decreasing the calcium concentration in a hyperkalemic cardioplegic solution from I to 0.25 mmoljL improved high-energy phosphate preservation and functional recovery in isolated rat hearts. Substitution of magnesium for hyperkalemia in the presence of a 0.25 mmoljL calcium concentration enhanced functional recovery further." It was hypothesized that magnesium prevents calcium influx, thereby diminishing diastolic tone and energy consumption during arrest, and also prevents cellular potassium and magnesium loss.": 14. 16 Torchiana and associates '5 showed that magnesium inhibits calcium-induced pressure development (i.e., diastolic tone) during a single cardioplegic infusion. The present study extends these observations, relating the effectsof multidose magnesium cardioplegia on pressure development to biochemical preservation and functional recovery. In patients undergoing coronary artery bypass, Engelman and colleagues" compared St. Thomas' Hospital solution containing a 15 mmol/L potassium concentration and a 15 mmoljL magnesium concentration to a hyperkalemic (potassium 25 mmoljL) cardioplegic solution without magnesium. Both solutions contained calcium. Lactate production was measurable during arrest only with the cardioplegic solution without magnesium, but differences between these solutions were considered clinically unimportant. Calcium has a central role in excitation-contraction coupling and the activation of ion pumps, consuming ATP. 33.34 Calcium also stimulates intramitochondrial dehydrogenases and may thereby enhance citric acid cycle flux and oxidative phosphorylation, providing ATP for the increased consumption." On the other hand, calcium uptake by mitochondria can decrease ATP production." Furthermore, mitochondria overloaded by calcium during ischemic cellular damage have a decreased ability to produce high-energy phosphates." However, the present study and previous observations demonstrate that the net effect of calcium during hypothermic, hyperkalemic arrest is energy depleting.": 13. 15 Of the cations, the total intracellular concentration of magnesium, approximately 17 mmoljkg cell water, is second only to that of potassium; however, the greater part is bound to proteins and nucleotides, so that the concentration of the free ion, the physiologically active moiety, is far lower.16 Intracellular free magnesium

Hyperkalemic cardioplegic solutions

247

concentration, which is difficult to measure, is probably below 1 mmol/L, close to the extracellular concentration and in a range in which small changes can regulate the activities of many enzymes." Acute changes in the intracellular free ion concentration are presumably buffered by the large chelated pooI.38-40 Total myocardial magnesium content is depleted by ischemia," possibly a result of the depletion of adenine nucleotides that would dechelate magnesium, transiently increasing the free ion concentration with consequent passive loss from the leaky ischemic cells." Since adenine nucleotides enter cellular reactions as the magnesium chelates, the prevention of magnesium loss during cardioplegic arrest may be essential to subsequent recovery of function." The effects of magnesium in cardioplegic solutions have been ascribed in large part to actions at the sarcolemma (the prevention of calcium entry and of potassium and magnesium loss), chiefly because the effects are rapid and exchange between intracellular and extracellular magnesium is relatively slow; the time for half of the total intracellular magnesium to exchange is about 180 minutes in the beating rat heart at 37° C.14.16.40 Since recent measurements with a magnesium ion-selective microelectrode show that intracellular free magnesium increases by a substantial fraction and then plateaus within seconds of elevating extracellular magnesium," the possible effects of increased intracellular free magnesium resulting from high magnesium concentrations in cardioplegic solutions deserve consideration. Magnesium appears to counteract the actions of calcium in excitation-contraction coupling," thereby diminishing energy consumption: Magnesium blocks the influx of calcium through the slow channel," opposes the release of calcium from the sarcoplasmic reticulum," and decreases the contractile response of the myofibrils at any cytosolic calcium concentration." Magnesium also activates the calcium ATPase of the sarcoplasmic reticulum" which, by removing calcium from the cytosol, decreases diastolic tone, in keeping with our observations during cardioplegia. Magnesium also stimulates mitochondrial respiration and may protect ATP production during mitochondrial calcium uptake." Accordingly, the observations that cardioplegic solutions containing magnesium inhibit pressure development during arrest and maintain or enhance highenergy phosphate levels in general reflect the known actions of the ion and particularly its interactions with calcium. The transient increase in diastolic pressure during reperfusion after the infusion of the cardioplegic solution containing magnesium but not calcium remains unexplained. The inhibitory effects of magnesium on pressure development were rapidly reversed when ionic concentrations in the perfusate were returned to normal

The Journal of Thoracic and Cardiovascular Surgery

2 4 8 Geffin et al.

upon reperfusion: Systolic function equaled prearrest levels at 5 minutes of reperfusion with the cardioplegic solution containing magnesium but no calcium. The calcium-induced increase in mean coronary vascular resistance during cardioplegic infusions may reflect extravascular compression by ventricular contraction, as suggested previously 15; both the increased resistance and the ventricular contractions were inhibited by magnesium. Direct effects of the ions on vascular smooth muscle may contribute to their effects on resistance during cardioplegic infusions.v-? At 5 minutes of reperfusion, coronary flow at constant perfusion pressure was above its prearrest value in all groups but the one given magnesium-free cardioplegic solution containing a 1.2 mmol/L concentration of calcium. Later during reperfusion, coronary flow fell below its prearrest value in all groups. This decline was more marked and occurred earlier after arrest by cardioplegic solutions containing calcium but no magnesium; it was diminished or delayed by the addition of magnesium. Coronary flow at this time may largely reflect the metabolic demand of pressure development, but residual effects of the cardioplegic solution ions on vascular smooth muscle cannot be excluded. Cardioplegia of the crystalloid-perfused, isolated rat heart differs greatly from the clinical situation. Further, with an isovolumic balloon distending the arrested left ventricle, myocardial contractile activity is expressed chiefly as pressure development that presumably incurs an energy cost above that of the same contractile activity in the empty ventricle and so might exaggerate differences between cardioplegic solutions in end-arrest A TP and PCr. In this model, although cardiac function deteriorates during reperfusion in the best protected groups, it improves in the least well protected groups. Vigorous contractions observed in the best protected groups during early reperfusion may have consumed excessive energy and thereby, or as a result of mechanical damage to the subendocardium contracting on the balloon, contributed to the later deterioration in these hearts. However, the decline in the percentage prearrest developed pressure in the group with the best functional recovery (group 2) was similar to that in control hearts undergoing warm perfusion for the same total duration without intervening arrest. Groups with the weakest contractions during early reperfusion showed some subsequent improvement. In this regard, Follette and associates" have suggested that the resumption of mechanical activity during postischemic reperfusion may divert energy from reparative work and that this can be prevented by cardioplegic reperfusion, thus sustaining the benefits of good preservation during arrest.

This study demonstrates benefits of including magnesium in a hyperkalemic cardioplegic solution. Although the limitations of the model and species differences may limit the clinical applicability of the findings/"" deep hypothermia during arrest, multidose administration, and full oxygenation of the cardioplegic solution are features of clinical practice. Magnesium was of particular benefit when calcium was present, even in a very low concentration (as in our dilute blood cardioplegia.'). The addition of magnesium allows the inclusion of calcium to preclude the unlikely but potentially hazardous calcium paradox, while avoiding the deleterious effects of calcium. During arrest by multidose calcium-free cardioplegic solutions, myocardial extracellular ionized calcium concentrations may vary regionally and with time because of uneven perfusion and washout with calcium-free cardioplegic solution alternating with calcium-containing blood provided by noncoronary collateral flow." Our observations and these considerations support including magnesium in hyperkalemic cardioplegic solutions that contain calcium and perhaps in those solutions without calcium. We gratefully acknowledge the high-pressure liquid chromatographic analyses by Alvin G. Denenberg, James E. Yath, and Douglas Malnati, other chemical analyses by Carmela Bondi and Richard P. Wawrzynski, technical assistance by Matthew Parker, and preparation of the manuscript by Cindy L. Getherall, Cathyleen Stone, and Emily C. Burton. REFERENCES I. Follette OM, Mulder DG, Maloney JY, Buckberg GO. Advantages of blood cardioplegia over continuous coronary perfusion or intermittent ischemia: experimental and clinical study. J THORAC CARDlOYASC SURG 1978;76:60419. 2. Fremes SE, Christakis GT, Weisel RD, et al. A clinical trial of blood and crystalloid cardioplegia. J THORAC CARDIOVASC SURG 1984;88:726-41. 3. Daggett WM, Randolph JD, Jacobs M, et al. The superiority of cold oxygenated dilute blood cardioplegia. Ann Thorac Surg 1987;43:397-402. 4. Tyers GF, Manley NJ, Williams EH, Shaffer CW, Williams DR, Kurusz M. Preliminary clinical experience with isotonic hypothermic potassium-induced arrest. J THORAC CARDIOVASC SURG 1977;74:674-81. 5. Conti YR, Bertranou EG, Blackstone EH, et al. Cold cardioplegia versus hypothermia for myocardial protection: randomized clinical study. J THORAC CARDIOVASC SURG 1978;76:577-89. 6. Hearse OJ, Braimbridge MY, Jynge P. Principles of formulation and administration. In: Protection of the ischemic myocardium: cardioplegia. New York: Raven Press, 1981:302-4. 7. Jynge P, Hearse OJ, Braimbridge MY. Myocardial protection during ischemic cardiac arrest: a possible

Volume 98 Number 2

Hyperkalemic cardioplegic solutions

August 1989

hazard with calcium-free cardioplegic infusates. J THORAC CARDIOVASC SLRG 1977:73:848-55. 8. Ruigrok T J'C, Cardioplegia and the calcium paradox. In: Caldarera CM, Harris P, eds. Advances in studies on heart metabolism. Bologna, Italy: CLUEB, 1982:353-60. 9. Hendren WG, Geffin GA, Love TR, et al. Oxygenation of cardioplegic solutions: potential for the calcium paradox. J THORAC CARDIOVASC SURG 1987;94:614-25. 10. Zimmerman ANE, Hiilsmann We. Paradoxical influence of calcium ions on the permeability of the cell membranes of the isolated rat heart. Nature 1966; 211:646-7. II. Gebhard MM, Bretschneider HJ, Gersing E, et al. Calcium-free cardioplegia-pro. Eur Heart J 1983; 4(suppl H):151-60. 12. Pernot AC, Ingwall JS, Menasche P, et al. Evaluation of high-energy phosphate metabolism during cardioplegic arrest and reperfusion: a phosphorus-31 nuclear magnetic resonance study. Circulation 1983;67: 1296-1303. 13. Boggs BR, Torchiana DF, Geffin GA, et al. Optimal myocardial preservation with an acalcemic crystalloid cardioplegic solution. J THORAC CARDIOVASC SURG 1987; 93:838-46. 14. Hearse DJ, Stewart DA, Braimbridge MY. Myocardial protection during ischemic cardiac arrest: the importance of magnesium in cardioplegic infusates. J THORAC CARDIOVASC SURG 1978:75:877-85. 14a. Reynolds TR, Geffin GA, Titus JS, et al. Myocardial preservation related to the magnesium content of hyperkalemic cardioplegic solutions at 8° e. Ann Thorac Surg 1989;47:907- I3. 15. Torchiana DF, Love TR, Hendren WG, et al. Calciuminduced ventricular contraction during cardioplegic arrest. J THORAC CARDIOVASC SURG 1987;94:606-13. 16. Hearse DJ, Braimbridge MY, Jynge P. Components of cardioplegic solutions. In: Protection of the ischemic myocardium: cardioplegia. New York: Raven Press, 1981:230-42. 17. Niedergerke R. The potassium chloride contracture of the heart and its modification by calcium. J Physiol (Lond) 1956;134:584-99. 18. Holubarsch C, Alpert NR, Goulette R, Mulieri LA. Heat production during hypoxic contracture of rat myocardium. Circ Res 1982;5 I:777-86. 19. Holubarsch e. Force generation in experimental tetanus, KCl contracture, and oxygen and glucose deficiency contracture in mammalian myocardium. Pflugers Arch. 1983;396:277-84. 20. 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. 21. Crevey BJ, Langer GA. Frank JS. Role of Ca H in maintenance of rabbit myocardial cell membrane: structural and functional ir.tegrity. J Mol Cell Cardiol 1978; 10:1081-1100. 22. Drop LJ, Tochka LN, Misiano DR. Comparative evaluation of two calcium ion-selective electrode systems, and

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33. 34. 35.

36.

37.

24 9

their utility for monitoring steady-state changes in [Ca H ] . Clin Chern 1982;28:129-33. Bodenhamer RM, DeBoer LWY, Geffin GA, et al. Enhanced myocardial protection during ischemic arrest: oxygenation of a crystalloid cardioplegic solution. J THORAC CARDIOVASC SURG 1983;85:769-80. 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. 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. Oguma F, Imai S, Eguchi S. Role played by oxygen in myocardial protection with crystalloid cardioplegic solution. Ann Thorac Surg 1986;42: 172-9. Coetzee A, Kotze J, Louw J, Lochner A. Effect of oxygenated crystalloid cardioplegia on the functional and metabolic recovery of the isolated perfused rat heart. J THORAC CARDIOVASC SURG 1986;91 :259-69. Yamamoto F, Braimbridge MY, Hearse DJ. Calcium and cardioplegia: the optimal calcium content for the St. Thomas' Hospital cardioplegic solution. J THORAC CARDIOVASC SURG 1984;87:908-12. DeBoer LWY, Ingwall JS, Kloner RA, Braunwald E. Prolonged derangements of canine myocardial purine metabolism after a brief coronary artery occlusion not associated with anatomic evidence of necrosis. Proc Natl Acad Sci USA 1980;77:5471-5. Hedberg PS, Torchiana DF, Reynolds TR, Geffin GA, O'Keefe DD, Daggett WM. Reversal of calcium-induced ventricular contracture during cardioplegic arrest. Clin Res 1987;35: 106A. Hearse DJ, Garlick PB, Humphrey SM. Ischemic contracture of the myocardium: mechanisms and prevention. Am J Cardiol 1977;39:986-93. Engelman RM, Rousou JH, Lemeshow S, Dobbs W A. A clinical comparison of potassium and magnesium-potassium crystalloid cardioplegia: metabolic considerations. J Cardiovasc Surg 1985;26:343-50. Katz AM. Excitation-contraction coupling. In: Physiology of the heart. New York: Raven Press, 1977:137-59. Carafoli E. The homeostatis of calcium in heart cells. J Mol Cell Cardiol 1985;17:203-12. Denton RM, McCormack JG, Thomas AP. Mechanisms whereby insulin and other hormones binding to cell surface receptors influence metabolic pathways within the inner membrane of mitochondria. Ann NY Acad Sci 1986;488:370-84. Sordahl LA. Effects of magnesium, ruthenium red and the antibiotic ionophore A-23 I87 on initial rates of calcium uptake and release by heart mitochondria. Arch Biochem Biophys 1974;167:104-15. Nayler WG, Ferrari R, Williams A. Protective effect of pretreatment with veraparnil, nifedipine and propranolol on mitochondrial function in the ischemic and reperfused myocardium. Am J Cardiol 1980;46:242-8.

The Journal of

250

Geffin et al.

38. Garfinkel L, A1tschuld RA, Garfinkel O. Magnesium in cardiac energy metabolism. J Mol Cell Cardiol 1986; 18:1003-13. 39. Fry CH. Measurement and control of intracellular magnesium ion concentration in guinea pig and ferret ventricular myocardium. Magnesium 1986;5:306-16. 40. Shine K. Myocardial effects of magnesium. Am J Physiol 1979;237:H413-23. 41. Shen AC, Jennings RB. Myocardial calcium and magnesium in acute ischemic injury. Am J Pat hoi 1972;67:41740. 42. Lansman JB, Hess P, Tsien RW. Blockade of current through single calcium channels by Cd H, MgH, and Ca H: voltage and concentration dependence of calcium entry into the pore. J Gen Physiol 1986;88:321-47. 43. Meissner G, Henderson JS. Rapid calcium release from cardiac sarcoplasmic reticulum vesicles is dependent on Ca H and is modulated by MgH, adenine nucleotide, and calmodulin. J Bioi Chem 1987;262:3065-73. 44. Fabiato A, Fabiato F. Effects of magnesium on contractile activation of skinned cardiac cells. J Physiol (Lond) 1975;249:497-517.

Thoracic and Cardiovascular Surgery

45. Hasselbach W, Fassold E, Migala A, Rauch B. Magnesium dependence of sarcoplasmic reticulum calcium transport. Fed Proc 1981;40:2657-61. 46. Rasmussen H, Barrett PQ. Calcium messenger system: an integrated view. Physiol Rev 1984;64:938-84. 47. Altura BM, Altura BT. New perspectives on the role of magnesium in the pathophysiology of the cardiovascular system. II. Experimental aspects. Magnesium 1985;4:24571. 48. Follette OM, Fey K, Buckberg GO, et al. Reducing postischemic damage by temporary modification of reperfusate calcium, potassium, pH, and osmolarity. J THORAC CARDIOVASC SURG 1981;82:221-38. 49. Fabiato A, Fabiato F. Calcium-induced release of calcium from the sarcoplasmic reticulum of skinned cells from adult human, dog, cat, rabbit, rat, and frog hearts, and from fetal and new-born rat ventricles. Ann NY Acad Sci 1978;307:491-522. 50. Buckberg GO. A proposed "solution" to the cardioplegic controversy. J THORAC CARDIOVASC SCRG 1979;77:80315.