Ventricular pump performance during hypocalcemia

Ventricular pump performance during hypocalcemia Clinical and experimental studies We have compared indices of ventricular function during rapid transfusion of citrated (1.5 ml/kg /min) or heparinized (1.5 ml/kg /min) autologous blood in six patients following discontinuation of cardiopulmonary bypass. Infusion of citrated blood was associated with a lowering of plasma ionized calcium concentration ([Ca++],from 0.90 ± 0.04 to 0.71 ± 0.4 mM, p < 0.001) and an increase in pulmonary artery balloon-occluded pressure (PAo,from 9.4 ± 2.6 to 15.5 ± 1.7 mm Hg , p < 0.01), without a change in left ventricular stroke work index, stroke index, or cardiac index. Transfusion of heparinized blood caused no change in plasma [Ca++]. A rise in PA o' which was similar in magnitude to that observed during citrated blood transfusion, was associated with increased left ventricular stroke work index, stroke index, cardiac index, and mean arterial pressure. Although data obtained during citrated blood transfusion suggest the presence of transient left ventricular dysfunction, its magnitude is not readily expressed in terms of ventricular function curves when accompanied by a simultaneous change in [Ca++]. To clarify this point, we have studied ventricular pump performance in the anesthetized closed-chest dog by volume loading during hypocalcemia, when mean arterial pressure, heart rate, and [Ca++] were in a steady state, both prior to and following beta blockade with propranolol. Function curves obtained during severe hypocalcemia ([Ca++] = 0.43 ± 0.02 mM) were shifted significantly to the right and downward, when compared to those obtained during normocalcemia ([Ca++] = 1.06 ± 0.03 mM). Hypocalcemia combined with beta blockade resulted in severe left ventricular failure, as demonstrated by a fiat ventricular function curve.

P. M. Stulz, M.D.,* D. Scheidegger, M.D.,** L. J. Drop, M.D., Ph.D., E. Lowenstein, M.D., and M. B. Laver, M.D., Boston, Mass.

T

he calcium ion has long been known to be essential for muscular contractile function. I. 2 On the basis of data obtained in the isolated perfused heart preparation,': 2 in the intact animal." 4 and in man," 5, 6 an association between decreased myocardial contractility

From the Anesthesia Services of the Masachusetts General Hospital and the Department of Anesthesia, Harvard Medical School, Boston, Mass. 02114. Supported in part by NIGMS Grant GM 15904-07. Dr. Scheidegger was the recipient of a grant from the Lichtenstein Stiftung, Basel, Switzerland. Received for publication Dec. 5, 1978. Accepted for publication April 17, 1979. *Present address: Chirurgische Universitatsklinik, 4004-Basel, Switzerland. **Present address: Service de Cardiologic, Hopital Cantonal, 1200--Geneva, Switzerland. Address for reprints: Dr. L. J. Drop, Department of Anesthesia, Massachusetts General Hospital, Boston Mass. 02114.

and low concentrations of calcium ion in plasma appears well established. However, data that quantitatively define the relationship between blood ionized calcium ([Ca++]) and ventricular pump performance (in terms of ventricular function curves) are incomplete. 5-7 We have studied the hemodynamic response to citrated blood transfusion in six patients subjected to coronary artery bypass or heart valve replacement. During rapid transfusion of citrated blood, calcium ion concentration in the recipient's blood is typically in an unsteady state," and since myocardial contractile force is critically dependent upon extracellular calcium ion sources;" 10 difficulties arise in the assessment of ventricular pump performance when such transient changes in [Ca++] appear. In view of these considerations, we studied ventricular performance (in terms of ventricular function curves) during steady-state hypocalcemia in the anesthetized closed-chest dog. Although previous data'': 11 have suggested that the hypocalcemia-induced

0022-5223/79/080185+ 10$01.00/0 © 1979 The C. V. Mosby Co.

1 85

The Journal of Thoracic and Cardiovascular

I 86 Stul; et al.

Surgery

Table I. Patient data: Biochemical values measured in arterial blood specimens withdrawn following cardiopulmonary bypass, but prior to the first autologous blood transfusion Biochemical variables prior to reinfusion of autologous blood Age (yr)

Sex

Weight (kg)

57 60

M M

56 74

58

M

63

58 43

M M

44

M

Bypass time (min)

Propranolol

MVR CABG

100 60 89

66 69

AVR CABG CABG CABG

52 120

79

CABG

58

None 160 mg (DIC 24 hr preop.) 40 mg (DIC 48 hr preop.) None 160 mg (DIC 48 hr preop.) 40 mg (DIC 24 hr preop.)

Operation

(mg lday]

[Ca++] (mM)

I

rCa] (mM)

I

PI (mM)

I

TP (gllGO ml)

I I pH

P co, (mm Hg)

I

Po, (mm Hg)

0.93 0.96

2.31 2.37

1.01 0.75

2.8 3.2

7.51 7.46

32 38

437 387

1.08

1.73

1.07

3.1

7.51

30

523

0.98 1.08

2.15 2.02

0.97 0.82

2.8 2.7

7.38 7.36

43 48

448 425

0.95

1.73

0.94

3.1

7.33

42

362

Legend: MYR, Mitral valve replacement. AYR, Aortic valve replacement. CABG, Coronary artery bypass graft. DIC, Discontinued, [Ca++], Ionized calcium

concentration. [Cal, Total calcium concentration. PI, Inorganic phosphorusconcentration. TP, Total protein concentration.

changes in ventricular pump function may be more pronounced in the presence of beta blockade, ventricular function curves derived in the presence of both hypocalcemia and beta blockade are not available. This problem is of interest in patients who are chronically treated with propranolol up to the day of major surgery requiring massive bank blood replacement. We studied the influence of beta blockade on ventricular performance during hypocalcemia in the closed-chest dog by means of left ventricular function curves.

Methods Patients. Studies were conducted in six patients subjected to coronary artery bypass grafting or heart valve replacement. Pertinent details on these patients are shown in Table I. In four patients, propranolol had been administered prior to operation in a dose ranging from 40 to 160 mg per day, but this regimen had been discontinued at least 24 hours prior to operation. In each patient, anesthesia was produced with intravenous morphine (0.5 to 1.0 mg/kg) and diazepam (0.15 to 0.20 mg/kg); pulmonary ventilation was manually controlled with a mixture of nitrous oxide and oxygen (50%) such that arterial P0 2 was higher than 150 mm Hg and PC02 was in the range of 32 to 48 mm Hg. Radial arterial, pulmonary arterial, and right atrial pressures were recorded continuously from indwelling catheters as part of routine monitoring employed in patients subjected to cardiac operations. Mean pulmonary artery balloon-occluded pressure (P A.,) was utilized to estimate changes in left ventricular filling pressure. Prior to institution of cardiopulmonary bypass, two 450 ml aliquots of whole blood were withdrawn from

each patient, each aliquot being drained into a commercially available collecting bag which contained either a solution of heparin (2,250 IV) or acid-citratedextrose (ACD, Formula A VSP, 67.5 ml). During the withdrawal of the blood, Ringer's lactate and a commercial 5% albumin solution were infused intravenously, their rates of infusion being determined by PA., values obtained. Body temperature, monitored with an esophageal temperature probe (Yellow Springs Instrument Co., Yellow Springs, Ohio), ranged from 36° to 37° C. Approximately 25 minutes after cessation of cardiopulmonary bypass, autologous blood, preserved in either heparin or ACD solution, was reinfused " through a blood warmer in each patient at a rate of 1.5 ml/kg/rnin for 4 minutes in random order approximately 30 minutes apart. Prior to first transfusion, two successive values of cardiac output, PA.,, heart rate, and mean arterial pressure, obtained 8 to 10 minutes apart, showed a difference of less than 10%. A third set of control measurements was then taken by recording hemodynamic variables and withdrawal of arterial blood samples, following which transfusion was started. Hemodynamic measurements were made and arterial blood specimens were withdrawn immediately prior to transfusion, at 2 minute intervals during transfusion, and 5 and 10 minutes following termination of transfusion. Experimental study. Mongrel dogs of either sex (weight range 19.5 to 22.5 kg) were anesthetized with intravenous thiopental (25 mg/kg) and spontaneous inhalation of halothane (0.9% to 1.1%) in oxygen via a cuffed endotracheal tube. Oral intake had been unrestricted up to I hour prior to the time of induction of anesthesia; premedication was omitted. Body tempera-

Volume 78

Ventricular pump performance during hypocalcemia

Number 2 August. 1979

ture was monitored with an esophageal temperature probe and did not vary more than 0.5 C from 37.5 C in any experiment. In each animal, femoral arterial and venous cannulas and indwelling catheters were placed in the pulmonary artery (Edwards Laboratories, Inc., Santa Ana, Calif.) and in the right atrium. The position of the latter two catheters was confirmed later at autopsy. In each dog, a period of at least I hour elapsed between the time of thiopental infusion and the first set of measurements. Solutions. Group I. For the purpose of volume loading during normocalcemia, a normocalcemic dextran solution was prepared by adding calcium chloride to 6% dextran in 0.9% sodium chloride (mol wt 70.000) in a proportion sufficient to adjust [Ca""] to approximately I. IO mM. Group II. To determine the hemodynamic consequences of volume loading with transient hypocalcemia in the experimental animal, as may be observed in the patient during citrated blood transfusion, we added 67.5 ml of ACD solution to 450 ml of 6% dextran (mol wt 70.000) in 0.9% sodium chloride. Group III. To produce steady-state hypocalcemia as described below, we prepared a solution of citrate by adding trisodium citrate (14.37 gm), citric acid (13.493 grn), and dextrose (24.444 gm) to distilled water to a final volume of I L. The total citrate concentration in this solution corresponds to that present in ACD solution Formula A USP, but the proportion of trisodium citrate was selected such that Na" would range from 149 to 152 mM. In all solutions, potassium chloride was added to a final K+ concentration of 3.5 mM and pH was adjusted to 7.40 by addition of TRIS* buffer. These adjustments were necessary to identify effects of [Ca++] changes only. Experimental protocol. Group I. In six dogs, volume loading was performed with normocalcemic dextran solution. Hemodynamic measurements were made at end expiration, and arterial blood samples were withdrawn immediately prior to dextran infusion (2.5 ml/kg/min for 6 minutes), at 2 minute intervals during infusion, and 5 and 10 minutes following termination of infusion. Approximately 30 minutes following the last set of measurements, when all measured hemodynamic variables had returned to near control, beta blockade was produced by intravenous infusion of propranolol (0.5 mg/kg over a 5 minute period) and confirmed by a lack of heart rate response to an isoproterenol challenge (0. 91Lg/kg). After 0

*Tris(hydroxymethyl)aminomethane.

0

187

approximately 20 minutes, volume loading was repeated and measurements were made as described. To study the possible influence of repeated volume loading per se, we infused normocalcemic dextran solution (2.5 ml/kg/min for 6 minutes) and collected hemodynamic data in six dogs as described earlier. Approximately 30 minutes after the last set of data, hemodynamic variables had returned to near control, and volume loading was repeated. Group II. Citrated dextran solution was infused in six dogs at a rate of 2.5 ml/kg/min for 6 minutes by use of a variable speed, constant infusion pump (Harvard Apparatus Co., Inc., Dover, Mass.). Measurements were made and arterial blood samples were withdrawn as described in Group I. The rate of 2.5 ml/kg/min was such that the observed [Ca++] changes would be comparable to those observed during transfusion of citrated blood in patients. Approximately 30 minutes following the last set of measurements, when all measured hemodynamic variables had returned to near their respective control values, beta blockade was produced and confirmed as described. The citrated dextran solution was then infused again and measurements were made as described. Group III. The effects of hypocalcemia on ventricular pump performance in six dogs were studied by a volume load during steady-state hypocalcemia. We selected a [Ca""] value of approximately 60% below normal ([Ca++] = 0.43 ± 0.02 mM) in order to examine ventricular performance at the lower end of the hypocalcemia range encountered clinically. The hypocalcemia plateau was attained by rapid intravenous infusion of ACD solution (1.5 ml/kg over a 30 second period) and maintained by further ACD infusion at a slow rate using a variable speed, constant infusion pump (Harvard Apparatus Co., Inc., Dover, Mass.). After the hypocalcemia plateau had been present for approximately 25 minutes, a solution of 6% dextran (mol wt 70.000) in 0.9% sodium chloride was infused intravenously (2.5 ml/kg/min for 6 minutes) and measurements were made as described in Group I. Approximately 30 minutes after the last set of measurements, at which time all measured hemodynamic variables had returned to near their respective control values, beta blockade was instituted and confirmed as noted earlier. Volume loading then was repeated. Measurements. Both in patients and in the dogs, left ventricular filling pressure was estimated from PAo,13 which was determined at each measurement period by balloon inflation. All vascular pressures were recorded on an eight-channel polygraph (Sanborn Co., Waltham, Mass.) via appropriate transducers (Hewlett-

The Journal of

188

Stul; et al.

Thoracic and Cardiovascular Surgery

Table II. Selected biochemical variables during blood transfusion in patients Citrated blood

Heparinized blood

[Ca""] (mM)

[Cal (mM) P, (mM)

Control

End of infusion

Ten minutes following infusion

0.92 ± 0.08 2.17 ± 0.07 0.91 ± 0.05

0.91 ± 0.08 2.12 ± 0.11 0.92 ± 0.11

0.88 ± 0.04 2.09 ± 0.10 0.82 ± 0.06

Control

End of infusion

Ten minutes following infusion

0.90 ± 0.04 2.05 ± 0.04 0.81 ± 0.05

0.71 ± 0.04* 2.0 ± 0.04 0.83 ± 0.09

0.91 ± 0.04 2.11 ± 0.07 0.89 ± 0.05

Legend: At the end of transfusion of citrated blood, [Ca++] decreased without a change in [Ca]. [Ca++], Ionized calcium concentration. [Ca] , Total calcium concentration. P" Inorganic phosphorus concentration.

*p < 0.001.

Packard 267 BC) calibrated with a mercury manometer at frequent intervals. Mean pressures were determined by electronic integration. Cardiac output was determined by thermodilution':' by use of a commercially available computer (Edwards Laboratories, Inc., Santa Ana, Calif., Model 9510); results of two measurements were averaged. Lead II of the electrocardiogram was recorded to determine heart rate. Cardiac index, stroke volume, stroke index, left ventricular stroke work, left ventricular stroke work index, and systemic vascular resistance were calculated according to standard formulas 15 by means of a programmable computer (Hewlett-Packard Model 67). In samples of heparinized whole blood, [Ca""] was determined by a thermostated, automated, flowthrough electrode system (Orion SS 20) as previously described. 16, 17 Of particular value during the protocol of Group III was the fact that [Ca++] results were known within 3 minutes of withdrawal of the blood specimen, so that the intravenous infusion rate of citrate could be adjusted to achieve a steady-state hypocalcemia. Blood gas tensions and pH were measured by use of appropriate electrodes (Radiometer A/S, Copenhagen, Denmark), which were maintained at 37° C. Plasma total calcium ([Ca]) was measured by EDTA* titration using calcein as the indicator"; total protein by refractometry'"; and sodium and potassium concentration by flame photometry. Inorganic phosphorus (PI) was determined by spectrophotometry. 20 Ventricular function curves. Left ventricular function curves were constructed by plotting values of stroke volume against those of PAo.21 An important consideration leading to the use of this expression of ventricular pump performance was that, although mean arterial pressure was essentially constant during volume loading at the normocalcemia and hypocalcemia *Ethylenediaminetetraacetic acid.

plateaus, mean arterial pressures at both of these [Ca""] levels were not equal. Since our experiments were performed in the closed-chest dog, this variable was not readily controllable. Mean arterial blood pressure is a major determinant of left ventricular stroke work; 15 therefore, comparison of left ventricular stroke work-PA, relationships derived at normocalcemia or hypocalcemia would introduce a bias induced by mean arterial pressure. Sonnenblick and Downing'< demonstrated that ventricular afterload influences the stroke work-left atrial pressure relationship, the curve moving upward and to the left as mean arterial pressure increased. In contrast, the stroke volume--Pa, relationship was minimally affected by mean arterial pressure in the range of 75 to 120 mm Hg. On the basis of these considerations, we have examined ventricular pump performance in terms of the stroke volume- PAo relationship. To this end, curves obtained during hypocalcemia were compared to those during normocalcemia by examining values of stroke volume at high physiological values of PAo (9 and 12 mm Hg). In the case of normocalcemia, stroke volume was also examined at PAo = 6 and 10.5 mm Hg; during hypocalcemia, values of stroke volume were also examined for each animal at PAo = 15 and 17 mm Hg. In some experiments, these values were actually recorded; in others, values above and below these were observed. In that case a straight line was fitted to the experimental data points by means of the method of least squares. The curve was then used to determine, by interpolation, the value of stroke volume corresponding to the selected value of PAo as noted. Mean stroke volume values were computed at a given PAo and a straight line was fitted, The t test for paired data was applied to determine the significance of the difference between data obtained in patients or in animals at the different time intervals. The t test for unpaired observations was used to determine the significance of the difference between data obtained in two groups of animals. Values are given as

Volume 78

Ventricular pump performance during hypocalcemia

Number 2

189

August. 1979

• P< 005

........ Heparinized 0- -
MEAN PULMONARY ARTERY OCCLUDED PRESSURE (mmHg)

MEAN RIGHT ATRIAL PRESSURE (mmHg)

~

*" p<0025 u. p<001

6

t

P< 0.001

**.

.&

::[----'----'-L----H------'-----J----1 ~'-i-----

12t I 8f-I

1.:-.---==="-*-~=,__._

i.

r-

4L1_L-----LI__

~-~J

~--

···r 90= ~+------f----I

MEAN ARTERIAL BLOOD PRESSURE (mmHg)

110'-

j

!

70~-'------'-----'------------If__---'-----"-__1

3-

CARDIAC INDEX (L/mlfl/m:!)

--r-t------

2-

-§-- - - -

-----'------1

1 '----------'--_~_~_-----II--_----l_

STROKE VOLUME INDEX (ml/beat/m:!)

--r--+-----=--=-~=-~ L

- - _ . ----II

45·

LEFT VENTRICULAR STROKE WORK INDEX (g m/m:!)

35-

T

P 1"--r----------¥---" T

25-

1 _-------L

15'

1

_

_1

If

2000f-

SYSTEMIC rVASCULAR RESISTANCE 1600;"

(dyn sec cm' 5)

[Ca++} (mM)

MINUTES

t1~OO~ 1

o~

o5

--- ---

---...~--"'" _L_

'* ,

L

Control

~

-.-L

2

.J.~ I

_ _ --1f'

'

--§---- I

4

L...During---J

__

J

_

I

--:0--

f - - uu---.-L__ -- -

5

10

L...Following-J

VOLUME LOADING Fig. I. Data obtained during autologous blood transfusion (citrated versus heparinized) in patients following

cessationof cardiopulmonary bypass. During citrated blood transfusion, a transient decrease in [Ca++]appeared. mean ± SEM. Differences were considered significant if p < 0.05.

Results Patients. Hemodynamic variables recorded prior to transfusion of either citrated or heparinized blood were not statistically different, as can be seen in Fig. I. Some [Ca++] values obtained prior to blood transfusion were at the lower limit of normal (Table I).

During transfusion of citrated blood (Fig. I), PA" increased. However, changes in left ventricular stroke work index, stroke index, and cardiac index were insignificant, as were changes in mean arterial pressure and systemic vascular resistance. Approximately 10 minutes following termination of blood transfusion, [Ca++] and all measured hemodynamic variables were not significantly different from their respective control values.

The Journal of Thoracic and Cardiovascular Surgery

1 90 Stui; et al.

... 5000

SYSTEMIC VASCULAR 4000 RESISTANCE 3000 (dyn'sec'cm-5j 2000

1000

MEAN ARTERIAL BLOOD PRESSURE (mmHq) STROKE VOLUME (ml/beotJ

01>

HYPOCALCEMIA

NORMOCALCEMIA

Before Il Blockode

I SEM

Afler Il Blockode

0'6

~E

30 20

n=6

tr::-~ _'" I~

~-+--§--~ !--- !... " --1i,

'40~

120 100 80 60

• p
~-~--"'---i5!

~

~'

.. E~

••

... ...

-•••

7

~~_

**

I

6 CARDIAC OUTPUT (L/min)

5 4

~_.

3

t--+-+-~ HEART RATE

(opm )

l---l--t--J

a» -H-J (mu)

4

6

CONTROL

2

4

6

MINUTES

Fig. 2. Anesthetized closed-chest dogs. Hemodynamic function during volume loading at normocalcemia and hypocalcemia plateaus. At each [Ca++] level, mean arterial pressure and heart rate remained essentially unchanged, but "control" data points for mean arterial pressure and systemic vascular resistance were different. The shaded area represents the 95% confidence band of normal [Ca++]. During transfusion of heparinized blood, increases were noted in PA", left ventricular stroke work index, cardiac index, and stroke index. The change in systemic vascular resistance did not reach significance. Ten minutes following termination of transfusion, none of the measured hemodynamic variables was significantly different from control values. Heart rate remained essentially unchanged in each patient during transfusion of blood collected in either preservative. As shown in Table II, [Ca++] decreased at the end of citrated blood transfusion, returning to control 10 minutes following termination of transfusion, whereas [Ca++] remained unchanged throughout heparinized blood transfusion. Other biochemical values, blood gases, and pH remained unchanged. Experimental data. Values of mean arterial pres-

sure, cardiac output, heart rate, PA", and systemic vascular resistance that were measured prior to the onset of the experiment were similar in all three groups of animals. Hemodynamic variables recorded immediately prior to and during volume loading (Group III) at normocalcemia are shown in Fig. 2. Changes in mean arterial pressure were insignificant, and heart rate did not vary more than 3 beats per minute in any given experiment during the volume challenge. Cardiac output increased and systemic vascular resistance decreased. Ventricular function curves derived during normocalcemia are shown in Fig. 3. Data summarized in Table III demonstrate that stroke volumes determined at PA" values of 9 and 12 mm Hg, which were obtained during volume loading with normocalcemic dextran solution at a time

Volume 78

Ventricular pump performance during hypocalcemia

Number 2

August,

1979

• •• ••• ,

DEXTRAN + ACD !J-BLOCKADE

50

• Before After

o

MEAN PULMONARY ARTERY OCCLUDED PRESSURE (mm Hg J

CARDIAC OUTPUT (L/minJ

Slope=0

20

[Co](mM)

8

~~ D---iJ

to

1.10 (n=6)+,B Blockade 043 (n=6)+.B Blockade

l2

14

16

18

20

Fig. 3. Anesthetized closed-chest dogs. Ventricular function curves during normocalcemia and hypocalcemia, and the influence of beta blockade. Table III. Effects of repeated volume loading at a 90 minute interval in six dogs ,

First volume loading (ml) Second volume loading (ml)

:~~

2500

(mmHg)

I

.-

SYSTEMIC VASCULAR RESISTANCE 2000

MEAN PULMONARY ARTERY OCCLUDED PRESSURE

SV at PA o =

I'

t--t--t-t--'-""

..

3

ICIi](mM)

- - 110 (n=6) 043 (n=6) -

p<005 p<0025 I P
.. .. .-.- ...":' --- 1935t~' ,

4

--------i-----~-----~

I9 I

9 mmHg

SVatPA o = /2 mm Hg

44.39 ± 4.37* 45.08 ± 4.05*

52.09 ± 4.84* 53.62 ± 5.25*

Legend: Stroke volumes (S V) derived at filling pressures of 9 and 12 mm Hg during the first and second volume loadings were not statistically different. "Au, Mean pulmonary artery balloon-occluded pressure. 'Mean ± SEM.

interval of approximately 90 minutes, were not significantly different. Thus an influence of repeated volume loading per se at this time interval can be excluded. Hemodynamic consequences of transient hypocalcemia. As demonstrated in Fig. 4, volume loading with citrated dextran solution was associated with a progressive and significant decrease in [Ca++], which returned toward control 10 minutes following termination of infusion, During infusion, changes in cardiac output and stroke volume were insignificant despite a significant elevation of PAn. Mean arterial pressure and systemic vascular resistance progressively decreased. Hemodynamic function during sustained hypocalcemia. Hemodynamic effects of sustained hypocalcemia ([Ca++] = 0.43 ± 0.02 roM) were evaluated by comparison of data obtained in Group III immediately prior to volume loading (data points marked "con-

r k,'f"

..........~:

(dyn' sec·cm- 5J

MEAN ARTERIAL BLOOD PRESSURE (mm HgJ

.

..

..

J:..._-

..... ----ri --

1500

120f 80

~'_T;*'•• ~---

__ ~

••

-~ -

15

rce") (mM/LJ

CONTROL 2

4

6

10

MINUTES LdlJring-----.J

VOLUME

LfollowingJ

LOAOJNG

Fig. 4. Hemodynamic consequences of infusion of citrated dextran in the normovolemic, anesthetized, closed-chest dog. In these normovolemic animals, systemic vascular resistance decreased when [Ca""] was decreased. Note the difference in response in patients (Fig. I), in whom decreased [Ca" "] was not associated with decreased systemic vascular resistance, presumably because of their blood volume status, which was likely to be reduced following cardiopulmonary bypass. The shaded area represents the 95% confidence band of normal [Ca++].

trol"), i.e., 25 minutes following onset of the hypocalcemic plateau, with those recorded at the corresponding time in Group I. As can be seen in Fig. 2, mean arterial pressure and systemic vascular resistance were lower during hypocalcemia; the differences in cardiac output and heart rate did not reach significance. PAn was higher during hypocalcemia (Fig. 3). Ventricular pump function during hypocalcemia. The mean values of stroke volume obtained at PAn values of 9 and 12 mm Hg were significantly below those observed during normocalcemia (p < 0.025 and p < 0.01, respectively), as shown in Fig. 3. As shown in Fig. 2, changes in mean arterial pressure were insignificant during volume loading. Heart rate did not

I 92 Stui; et al.

The Journal of Thoracic and Cardiovascular Surgery

vary more than 4 beats per minute in any experiment; mean heart rate remained unchanged. Combination of beta blockade and hypocalcemia. Hemodynamic function during transient hypocalcemia. As shown in Fig. 4, changes in mean arterial pressure and systemic vascular resistance observed during citrated dextran infusion following beta blockade were directionally similar to those prior to blockade, whereas there was a slightly higher PA., for a lower cardiac output (p < 0.025) when beta blockade was present. Control values of systemic vascular resistance obtained following beta blockade were higher (p < 0.025). Ventricular function during sustained hypocalcemia. When both beta blockade and hypocalcemia were present, augmentation of stroke volume was not apparent despite an almost twofold increase in PA." i.e., the function curves were flat (Fig. 3). Two dogs died within IO minutes of termination of volume loading; resuscitative efforts with rapid infusion of calcium chloride (up to 20 mg/kg) and isoproterenol were not successful. Blood gases and biochemical variables. Arterial blood gases and biochemical variables measured during collection of ventricular function curve data at a hypocalcemia plateau remained unchanged. During and following volume loading, values of arterial P0 2 were above 150 mm Hg in all animals; arterial Pco; ranged from 38 to 49 mm Hg and pH, from 7.31 to 7.36. Na" ranged from 149 to 153 mM and K+, from 3.9 to 4.1 mM. Total protein ranged from 5.5 to 5.9 gm/ 100 ml. Pre-existing beta blockade did not introduce significant alterations in any of the aforementioned changes. Total calcium values measured in Groups I and III were similar (l.95 ± 0.08 mM versus 2.02 ± 0.12 mM, respectively) although [Ca++] was decreased to approximately 60% below normal in the latter group. Inorganic phosphorus (PI) was lower during hypocalcemia (p < 0.025). Discussion The data demonstrate that a transient decrease in [Ca++] during transfusion of citrated blood in the human being is associated with left ventricular dysfunction of relatively short duration. These results, although qualitatively in agreement with previous data on the subject,"- 6 provide quantitative information on both the changes in indices of hemodynamic function and in [Ca!"] that appear following citrated blood transfusion. However, the magnitude of left ventricular dysfunction is not readily expressed in terms of left ventricular function curves when a transient decrease in [Cat"] is

present; in addition, hemodynamic consequences of transient hypocalcemia cannot be distinguished readily from those of volume loading. Although we obtained data during transfusion of heparinized blood in the same patients, randomizing the order in which blood preserved in either of the two preservatives was transfused, interpretation of data during transfusion of heparinized blood is difficult because of the observation of increased mean arterial pressure. Such differences in mean arterial pressure induce compensatory changes in myocardial contractility and impose limitations to evaluation of ventricular pump function." Our experimental data define the change in ventricular pump performance with institution of hypocalcemia in the normovolemic closed-chest dog. These observations extend the present knowledge of cardiovascular consequences of hypocalcemia in three ways: (I) They show a rightward and downward displacement of the ventricular function curve during hypocalcemia; (2) they demonstrate flat ventricular function curves in the presence of both sustained hypocalcemia and beta blockade with propranolol; and (3) they demonstrate a reduction of peripheral vascular resistance during both transient and sustained hypocalcemia. The ventricular function curve is a sensitive expression of ventricular pump performance. It is known that mean arterial pressure and heart rate are major determinants of ventricular function'"; hence these variables should not change while ventricular function curves are generated. In order for the influence of hypocalcemia upon ventricular pump performance to be determined, a steady-state hypocalcemia should be present while ventricular function curves are generated, because the intracellular calcium ion supply necessary for activation of the contractile elements is critically dependent upon extracellular sources. 9. 10 Furthermore, it is known that excitation-contraction coupling is a calcium-dependent process. 22, 24 Recent observations" suggest that the sensitivity of the excitation-contraction 'coupling mechanism to changes in [Ca++] is influenced by muscle length. Blood transfusion is likely to alter sarcomere length and, therefore, contractile function. These considerations impose limitations to the validity of ventricular function curves derived in the presence of alterations in mean arterial pressure, heart rate, and [Ca++]. It appears that these factors were not taken into account in previous studies. 3-8 The ventricular function curves in our experimental study were derived when [Ca++] was known, controlled, and in a steady state. Hypocalcemia was instituted at a level approximately 60% below normal to encompass the clinically encountered hypocalcemia

Volume 78 Number 2 August, 1979

Ventricular pump performance during hypocalcemia

range. In addition, mean arterial pressure and heart rate remained essentially unchanged during volume loading. It would be attractive to postulate a family of related ventricular function curves, each of which may be detennined by a [Ca++] value in the range examined in this study. Data shown in Figs. 2 and 4 demonstrate that decreased [Ca++], produced in the normovolernic, closed-chest animal, was associated with changes in both ventricular and peripheral vascular function. During infusion of citrated dextran (Fig. 4), cardiac output remained unchanged despite increased PAo and decreased mean arterial pressure, indicating left ventriclar dysfunction. In addition, systemic vascular resistance decreased, and this systemic vascular resistance appeared to be the principal cause of decreased mean arterial pressure. An effect of hypocalcemia on the peripheral vasculature was also suggested by data presented in Fig. 2. Despite the fact that pre- volume loading values of cardiac output during sustained hypocalcemia tended to be lower than those at the corresponding time during norrnocalcernia, systemic vascular resistance was significantly lower under the former conditions. This finding is striking since, in the presence of a reduced cardiac output, systemic vascular resistance would be expected to rise in order to maintain arterial blood pressure. Our data are consonant with data by Bunker and associates," who found that, following trisodium citrate infusion in the intact anesthetized dog, total peripheral resistance decreased in four of five dogs whereas changes in cardiac output were variable. In view of the importance of the calcium ion for maintenance of both ventricular and peripheral vascular function, calcium replacement may be appropriate in patients to correct a hypocalcemic state. However, it should be emphasized that in our patient study a decrease in [Ca++] by approximately 21% appeared at a citrated blood transfusion rate of 1.5 ml/kg/min for 4 consecutive minutes. Blood transfusion at a lower rate being more common, calcium replacement therapy may not be routinely required. The state of beta adrenergic activity has been proposed as an important determinant of the hemodynamic response to hypocalcemia. Smith and Hurley!' and Bunker and associates" have suggested that the hemodynamic consequences of hypocalcemia may be more pronounced following beta blockade. Four of our patients had received propranolol, but therapy had been discontinued at least 24 hours prior to operation. Data on these patients appear together with data on those who did not receive beta blocking drugs, since the observed changes in hemodynamic variables during tran-

193

sient hypocalcemia were similar in direction and magnitude and because the degree of beta blockade in these patients was uncertain. Data shown in Fig. 3 illustrate the importance of beta adrenergic activity for ventricular pump performance when sustained hypocalcemia is present. In the presence of hypocalcemia alone, the slope of the ventricular function curve was 1.53, whereas in the presence of both hypocalcemia and beta blockade the slope was zero. That irreversible myocardial failure can occur under the latter conditions is best illustrated by our observation that resuscitative efforts with large doses of calcium chloride were not successful in two of six dogs that died within 10 minutes of volume loading. The severe pump failure undoubtedly occurred secondary to the presence of both beta blockade and sustained hypocalcemia, since the possibility of an effect of repeated volume loading per se can be excluded (Table III). The correction of hypocalcemia in the presence of beta blockade appears to be justified. Calcium chloride, approximately 5 mg/kg, may be infused intravenously, and repeated doses may be necessary to achieve a normal [Ca++] level. Guidance by repeated [Ca++] measurements is considered extremely helpful. It may be remembered, however, that conditions underlying the need for beta blocking drugs, especially myocardial ischemia, may alter the response to changes in [Ca++]. Furthermore, the beneficial or possible adverse effects of increased [Ca" "] above normal upon ventricular perfonnance in the presence of ischemia have not been established. The severe left ventricular dysfunction observed with a combination of beta blockade and sustained hypocalcemia may be explained on the basis of a severe intracellular calcium ion deficit. First, if movement of calcium ions across the cell membrane is dependent on a concentration gradient;" the available calcium ion concentration within the myocardial cell is likely to be reduced. Myoplasmic [Ca++] may be decreased even further in the presence of beta blockade because of inhibition of lipid-facilitated transport of calcium across the cell membranes. Woolley and Campbell'" have demonstrated that lipids extracted from smooth muscle promote the transport of calcium ions from aqueous media through a lipid-soluble phase. Experiments by Naylor'? have shown that ventricular muscle extracts of various animal species have a similar function in heart muscle cells. Naylor proposed that propranolol impedes calcium transport from the sarcoplasmic reticulum through the lipid-containing membranes such that myoplasmic concentrations of the ion remain below a critical level required for initiation of contraction.

The Journal of

194 Stulr et al,

Finally, our observations (Table II) demonstrate that acute disturbances in ionized calcium homeostasis may not be apparent from total calcium measurement. This problem has been encountered in the past": 29 and emphasizes the need for repeated [Ca! "] measurement whenever a derangement in calcium equilibrium is to be evaluated. We wish to express our gratitude to Mrs. A. M. Scheidegger and Ms. C. Balkas for technical assistance, to Ayerst Laboratories for providing propranolol, and to Prof. W. Rutishauser for reviewing the manuscript.

2

3

4

5

6

7

8

9

10

REFERENCES Ringer S: A further contribution regarding the influence of different constituents of blood on contractions of the heart. 1 Physiol 4:29-31, 1883 McLean FC, Hastings AB: A biological method for the estimation of calcium ion concentration. 1 Bioi Chern 107:337-351, 1934 Bunker IP, Bendixen HH, Murphy AI: Hemodynamic effects of sodium citrate. N Engl 1 Med 266:373-377, 1962 Corbascio AN, Smith NT: Hemodynamic effects of experimental hypercitremia. Anesthesiology 28:510-516, 1967 Cooper N, Brazier lR, Hottenrott C, Mulder DG, Maloney IV, Buckberg GD: Myocardial depression following citrated blood transfusion. Arch Surg 107:756763, 1973 Olinger GN, Hottenrott C, Mulder DG, Maloney IV, Milker 1, Patterson RW, Sullivan SF, Buckberg GD: Acute clinical hypocalcemia depression of the myocardium during rapid blood transfusion and postoperative hemodialysis. J THoRAc CARDIOVASC SURG 72:503-512, 1976 Howland WS, Schweitzer 0, Carlon GC, Goldiner PL: The cardiovascular effects of low levels of ionized calcium during massive transfusion. Surg Gynecol Obstet 145:581-586, 1977 Hinkle IE, Cooperman LH: Serum ionized calcium changes during citrated blood transfusion in anesthetized man. Br J Anaesth 43: I 108- I II I, 1971 Chidsey CA: Calcium metabolism in the normal and failing heart. Hosp Pract 7:65-75, 1972 Langer GA, Brady AJ: Calcium influences in the mammalian ventricular myocardium. 1 Gen Physiol 46:703720, 1963

Thoracic and Cardiovascular Surgery

II Smith NT, Hurley El: Citrate infusion in dogs following autotransplantation. Arch Surg 98:44-48, 1969 12 Hallowell P, Bland JHL, Buckley MJ, Lowenstein E: Transfusion of fresh autologous blood in open-heart surgery. 1 THoRAc CARDIOVASC SURG 64:941-948, 1972 13 Buchbinder N, Ganz W: Hemodynamic monitoring. Invasive techniques. Anesthesiology 45: 146-155, 1976 14 Ganz W, Swan HJC: Measurement of blood flow by thermodilution. Am 1 Cardiol 29:241-246, 1972 15 Yang SS, Bentivoglio LG, Maranhao V, Goldberg H: From Cardiac Catheterization Data to Hemodynamic Parameters, Philadelphia, 1971, F. A. Davis Company, pp 157-210 16 Drop LJ, Fuchs C, Stulz PM: Determination of blood ionized calcium in a large segment of the normal adult population. C1in Chim Acta 89:503-510, 1978 17 Madsen S, Olgaard K: Evaluation of a new automatic calcium ion analyzer. Clin Chern 23:690-694, 1976 18 Bett 1M, Frazer GP: A rapid micromethod for determining serum calcium. Clin Chim Acta 4:346-356, 1959 19 Sunderman FW: A rapid method for estimating serum proteins. 1 Bioi Chern 153: 139-145, 1944 20 Chen PS, Toribara TY, Warner H: Microdetermination of phosphorous. Anal Chern 28: 1756-1761, 1956 21 Cohen IN, Franciosa JA: Vasodilator therapy of cardiac failure. N Engl 1 Med 297:27-31, 1977 22 Sonnenblick E, Downing SE: Afterload as a primary determinant of ventricular performance. Am J Physiol 204:604-610, 1963 23 Sarnoff sr, Mitchell IH, Gilmore IP, Remensnyder IP: Homeometric autoregulation in the heart. Circ Res 8: 1077-1091, 1960 24 Reuter H: Exchange of calcium ions in the mammalian myocardium Circ Res 34:599-605, 1974 25 Ter Keurs HEDl, Rijnsburger WH, Blott R, Nagelsmit MJ: Length-dependent activation of rat cardiac muscle (abstr). Trans Eur Soc Cardiol 1:23, 1978 26 Woolley DW, Campbell NK: Tissue lipids as ionexchangers for cations and the relationship to physiological processes. Biochim Biophys Acta 57:384-386, 1962 27 Naylor WG: The effects of pronethalol and propranolol on lipid facilitated transport of calcium ions. 1 Pharmacol Exp Ther 153:479-484, 1966 28 Drop LJ, Laver MB: Low plasma ionized calcium and response to i.v. calcium replacement therapy in critically ill man. Anesthesiology 43:300-306, 1975 29 Bristow MR, Schwartz HD, Binetti G, Harrison DC, Daniels lR: Ionized calcium and the heart. Circ Res 41:565-581, 1977