Extended ex vivo preservation of the heart and lungs

J

THORAC CARDIOVASC SURG

1990;100:687-98

Extended ex vivo preservation of the heart and lungs Effects of acellular oxygen-carrying perfusates and indomethacin on the autoperfused working heart -lung preparation The autoperfused working heart-lung preparation bas been proposed as a method for long-term heart-lung preservation. We investigated the effects of acellular oxygen-carrying perfusates (study 1) and the effect of donor pretreatment with indomethacin (study 2) on the working ex vivo heart-lung block. In study 1 perfusion with stroma-free hemoglobin resulted in significantly reduced survival (118 ± 46 minutes) compared with autologous blood (561 ± 125 minutes, p < 0.05) or perfluorocarbon (438 ± 114 minutes, p < 0.05). Decrease in survival with stroma-free hemoglobin perfusate is associated with a marked decrease in left ventricular performance and a significant increase in pulmonary vascular resistance. Perfusion with autologous blood is associated with a significant increase in pulmonary vascular resistance after 240 minutes of explantation, which is significantly delayed by perfusion with perfluorocarbon. Perfusion for 6 hours with blood pretreated with indomethacin (study 2) resulted in a decrease in the concentration of prostacyclin and thromboxane A2 metabolites but an increase in the prostaglandinjthromboxane A2 metabolite ratio. This is associated with abrogation of the increase in pulmonary vascular resistance (12,787 ± 1682 dynesjsecjcm- 5, T = 0; 13,134 ± 2654 dynesjsecjcm- 5, T = 360 minutes) observed in preparations perfused with autologous blood (1J,194 ± 1942 dynesjsecjcm- 5, T = 0; 24,768 ± 3325 dynesjsecjcm- 5, T = 360 minutes, p < 0.05). We conclude that alteration of the cellular and humoral components of autologous blood may prove advantageous for increasing the utility of the autoperfused working heart-lung preparation as a preservation technique.

Elliot Kaplan, MD,a James T. Diehl, MD,a Myron B. Peterson, MD, PhD,b Kenneth H. Somerville, BS,b Benedict D. T. Daly, MD,a Raymond J. Connolly, PhD,a Arniel G. Cooper, MD,c Shawn D. Seiler, BA,a and Richard J. Cleveland, MD, 8 Boston, Mass.

Lng and heart-lung transplantation are viable therapeutic alternatives for patients with certain end-stage cardiopulmonary diseases. A small pool of donor organs and the lack of a suitable method for long-term preservation (12 to 24 hours) of the heart-lung block have limited utilization of this procedure. The autoperfused working From the Department of Surgery, Tufts University School of Medicine, Division ofCardiothoracic Surgery, New England Medical Center,• the Department of Pediatric Critical Care, New England Medical Center,b and the Department of Pathology, Faulkner Hospital, Boston Mass. Supported in part by National Institutes of Health BSRG Grant S07RR05598-22 Read at the Fifteenth Annual Meeting of The Western Thoracic Surgical Association, Monterey, Calif., June 21-25, 1989. Address for reprints: James T. Diehl, MD, Department of Cardiothoracic Surgery, New England Medical Center Hospitals, 750 Washington St., #266, Boston, MA 02111.

12/6/21224

heart-lung preparation has been used sporadically as a method of short-term preservation. Further refinement in the technique is, however, necessary before this method can achieve widespread clinical application. Experimentally, use of the autoperfused working heart-lung preparation for organ preservation is limited by a pulmonary vasoconstrictive response leading to right-sided heart failure.'· 2 The pulmonary vasoconstrictive response may be attenuated by perfusion of the heart-lung block with leukocyte-poor blood 3• 4 or administration of indomethacin to the donor before harvest. 5 Both methods are presumed to work by altering circulating humoral mediators such as thromboxane A2 and prostacyclin. Circulating concentrations of vascular humoral mediators have not, however, been correlated with either pulmonary vascular resistance or cellular preservation in the autoperfused working heart-lung preparation. These experiments were performed to answer two 687

The Journal of Thoracic and Cardiovascular

6 8 8 Kaplan et a/.

Surgery

LV pressure transducer

Fig. 1. The autoperfused working heart-lung preparation. PA, Pulmonary artery.

questions. Study 1-the effects of acellular oxygen-carrying perfusates-was designed to determine if exchange transfusion with acellular oxygen-carrying perfusates alters the viability of the autoperfused working heart-lung preparation, pulmonary vascular resistance, or left ventricular (LV) performance. Study 2-e./fects of donor pretreatment with indomethacin-asks two related questions. First, does donor pretreatment with intravenous indomethacin alter eicosanoid concentration, pulmonary vascular resistance, or pulmonary compliance? Second, is there a relationship between eicosanoid concentrations and pulmonary vascular resistance or pulmonary compliance?

Methods and materials Male New Zealand white rabbits (>4 kg) were premedicated with intramuscular xylazine (0.1 mg/kg) and anesthe-

tized with intramuscular acepromazine/ketamine mixture ( 1:10 by volume 0.3 ml/kg). Pentobarbital ( 10 mg/kg) was then infused via a lateral ear vein over 3 minutes. All animals were treated in accordance with National Institutes of Health guidelines for humane animal care. 6 Experimental protocol. The autoperfused working heartlung preparation was based on the experimental model of Muskett and associates 5 and is illustrated in Fig. 1. After anesthesia, a tracheostomy was performed and the endotracheal tube connected to a Harvard respirator (Harvard Apparatus Co., Inc., S. Natick, Mass.) set at 24 breaths/min with a tidal volume of 6 ml /kg. Intravenous heparin ( 1000 U /kg) was administered and a median sternotomy performed. The right innominate artery was ligated and a silk tie placed around the aortic arch distally. Tubing from a reservoir bag was placed into the ascending aorta via the right innominate artery and positioned distal to the coronary ostia. A second piece of tubing from the reservoir was placed into the right atrium via the right atrial appendage. Autoperfused working heart-lung preparations were placed into one of five groups. Study 1 consisted of groups I (autologous blood, control), II (perfluorocarbon, PFC), and III (bovine stroma-free hemoglobin, SFH). The perftuorocarbon used was FC-43 (Alpha Therapeutic Corp., Los Angeles, Calif.). Stroma-free hemoglobin perfusate was prepared by mixing fresh plasma from a donor rabbit 1:1 with bovine stroma-free hemoglobin (donated by Dr. Glenn Wegner). Groups IV (indomethacin control) and V (pretreated) composed study 2. Pretreatment for group V (pretreated) was an intravenous bolus (0.2 mg/kg) of indomethacin sodium trihydrate (Merck Sharp & Dohme, West Point, Pa.) infused 15 minutes before incision. Groups I (control, study 1), IV (indomethacin control, study 2), and V (pretreated, study 2) underwent aortic ligation with blood diversion to the reservoir. When the reservoir contained 100 to 150 ml of blood, the venae cavae were ligated and blood was returned to the heart via the right atrial catheter. In groups II (PFC) and III (SFH) (study I) the reservoir was filled with the appropriate acellular perfusate. The venae cavae were ligated and the heart allowed to eject until empty. As infusion of the perfusate into the right atrium was begun, the aorta was ligated and exchange transfusion performed. After aortic ligation the respiratory rate was decreased to 2 to 3 breaths/min. Inspired oxygen concentration for group II (perfluorocarbon, PFC) was 1.0, all other groups were ventilated with room air. The heart and lungs were removed en bloc, instrumented, and placed into a plastic bag containing 20 ml of saline. The plastic bag was placed into a 39o C water bath. Glucose concentration was determined every 4 hours and maintained greater than 80 mgjdl. Instrumentation. After explantation, Mikro-Tip pressure transducers (Millar Instruments Inc., Houston, Tex.) connected to a multichannel recorder (HP 77588, Hewlett-Packard Co., Palo Alto, California) were positioned in the pulmonary artery and the LV. Piezoelectric crystals (5 MHz, Triton Technologies, San Diego, Calif.) were attached to the epicardium of the heart aligned with the LV minor axis (study 1). Preparations in study 2 had a fluid-filled catheter introduced into the trachea distal to the endotracheal tube. Output from the multichannel recorder was digitalized and acquired at a rate of 50 Hz onto a personal computer (Data 286, Dataworld Inc., Pico Rivera, Calif.) utilizing Asystant software (Asystant Software Technologies, Rochester, N.Y.).

Volume 100

Preservation of heart and lungs 6 8 9

Number 5 November 1990

,.

I

100 -•-•-o-o-o-o

QJ

.D

I •-•• ~-If

·-·

75

i

50

·-··-· ...... .

Q_

<(

Ii

'+-

0

~

Group II Group Ill

• • 0-0

Q_

3:

•

6 -----· 6

I

> I

•

o-o

i

0

_j

0 - 0 Group I

25

0

I

~-6-6-6-6

o-o-o-o-o-• ••-o-o-o-o-o-o-o-o

I

0

4

*

8

12

16

20

24

28

Time from Explantation (Hours) Fig. 2. Survival curves for study I. APWHLP, Autoperfused working heart-lung preparation. *p groups I and II.

Physiologic· parameters Study 1: The effects ofacellular oxygen-carrying perfusates. Pulmonary artery pressure, LV pressure, LV minor axis diameter, and cardiac output were acquired. Pulmonary vascular resistance was calculated by the following formula: (mean pulmonary artery pressure - LV end-diastolic pressure)/ cardiac output X 80. Pressure and volume measurements were obtained continuously while loading conditions were being altered by changing the height of the reservoir bag. Pressure volume loops were created by plotting LV pressure versus LV diameter. LV end-systolic pressure - LV diameter at varying loading conditions were plotted and linear regression performed. For the measured ranges the slope obtained by linear regression represents a load-independent measurement of LV performance. 7• 8 Measurements were obtained at explantation (T = 0) and every hour subsequently until the autoperfused working heartlung preparation could no longer maintain a systolic pressure of 50 mm Hg or greater. Study 2: Effects of donor pretreatment with indomethacin. Tracheal pressure, cardiac output, pulmonary artery pressure, and LV pressure were obtained. Pulmonary vascular resistance was determined by the previously stated formula and pulmonary compliance was calculated by the following formula: Tidal volume/Change in tracheal pressure. Arterial blood samples (2 ml) were obtained, placed into glass tubes containing potassium EDTA* (0.06 ml of 7.5% solution) and sodium indomethacin (241-Lg/ml whole blood), and immediately placed in a 4° C ice bath. At the conclusion of the experiment, samples were centrifuged at 1900 g for 25 minutes at 4 o C. Plasma was separated by bulb aspiration, placed into polypropylene tubes, and frozen at -80° C until radioimmunoassay was performed. Measurements were taken at explantation (T = 0) and every hour subsequently for 6 hours. Biochemical parameters. We measured three eicosanoids in •Ethylenediaminetetraacetic acid.

< 0.05 versus

these studies: 6-keto-prostaglandin F Ia• the stable metabolite of prostacyclin; thromboxane 8 2, the stable inactive metabolite of thromboxane A 2; and 11-dehydro-thromboxane 82, the stable inactive metabolite of thromboxane 8 2 • 6-Keto-prostaglandin F Ia and thromboxane B2 appear to be reliable indices of parent compound synthesis. However, both underestimate total parent production because of the formation of other metabolites. This method of analysis, however, is predictive of parent compound synthesis. Both these end products were quantified by standard double antibody radioimmunoassay as previously described. 9 The thromboxane 82 antibody was used at a dilution of I :4000 and detects 6 to 300 pg of thromboxane 8 2• Our 6-ketoprostaglandin F Ia antibody at a dilution of 1:5000 detects 10 to 270 pg of 6-keto-prostaglandin F 1a. Cross reactivities of these antibodies have been previously published. 9 Sampling technique can cause artificial generation of thromboxane 82 in some cases; thus we measured the 11-dehydro-thromboxane B2 metabolite, as well as parent thromboxane B2• 11-Dehydro-thromboxane 8 2 is a relatively permanent metabolite ofthromboxane 82 in blood and is formed from tissue-bound enzymes; thus artifactual generation of this compound does not occur with sampling. We used the technique of Westland and associates 10 with slight modifications and have been able to assay 11-dehydro-thromboxane 8 2 in plasma samples from humans, rabbits, and dogs. The range of sensitivity of this assay in these studies was 2.5 to 50 pg of 11-dehydro-thromboxane 82. All values are reported as picograms per milliliter of blood. Processing for histologic study. Specimens were obtained as described and stored in 10% formalin. Vacuum filtration processing was performed for 12 hours. Samples were embedded in paraffin, sectioned at 6 /-Lm, and stained with hematoxylin and eosin. Lung wet/dry measurements. At the conclusion of all experiments the left lung was weighed and placed in 60° C drying oven for 48 hours. The lung was then reweighed and the wet/dry weight ratio of the lung determined. This value was

690

The Journal of Thoracic and Cardiovascular Surgery

Kaplan et a!.

·-·

75

0 - 0 Group I

0 0 0

6-/J.

*

Group II Group Ill

1

X

I

*

50

1.{)

I

E

/J.

QJ

(f)

-......._ QJ

c

;/l

/j

u -......._ u (f)

0

25

>-

"0

0:::

e

>

Q_

0

0

60

d:>

·-----------120

T~l •

.......---:i.L

T

180

.l

•

240

300

T

T

.L

Time from Explantation (min)

Fig. 3. Pulmonary vascular resistance (PVR) X 1000 dynesjsecjcm- 5 over 300 minutes for study I. Group I, autologous blood; group II, PFC; group III, SFH. *p < 0.05 versus baseline. used as a measurement of pulmonary water content. In study 2 lung biopsy specimens were obtained hourly for histologic evaluation. Data analysis. Values are reported as mean ± standard error of the mean. Mean survival and lung wet/dry ratio were analyzed by the Kruskal-Wallis test .for overall comparison. Pairwise comparisons were made using the Mann-Whitney U test. Survival curves were analyzed with the product limit survival analysis and Mantel-Cox analysis. Comparisons within groups (pulmonary vascular resistance, pulmonary compliance, LV performance) underwent repeatedmeasures analysis of variance followed by F-test adjustment. Comparison with baseline values was made with the NewmanKeuls test for multiple comparisons. In all cases differences were considered significant at p < 0.05.

Results Study 1: The effects of acellular oxygen-carrying perfusates Survival. Group I (control) and group II (PFC) have similar survival times (group I [control], 561. ± 125 minutes; group II [PFC], 438 ± 114 minutes). Both survived significantly longer than group III (SFH, 118 ± 46 minutes; p < 0.05). Survival curves are illustrated in Fig. 2. Pulmonary vascular resistance. The pulmonary vascular resistance over 300 minutes is shown in Fig. 3. Baseline pulmonary vascular resistance in group I (control) is 7342 ± 1469 dynesjsecjcm- 5 and increases to 22,339 ± 5051 dynesjsecjcm- 5 (p < 0.05) 240 minutes after explantation. At 300 minutes pulmonary vascular resistance was also significantly increased. Group II (PFC) demonstrates an initial pulmonary vascular resistan~ of 7121 ± 1353 dynesfsecjcm- 5. This value did

not change significantly over 300 minutes of perfusion. Initial pulmonary vascular resistance in group III (SFH) was 14,693 ± 4768 dynesjsecjcm- 5• Within 60 minutes of explantation this value increased significantly (36,371 ± 10,916 dynesjsecjcm- 5; p < 0.02). Myocardial performance. In group I (control) the slope of the LV end-systolic pressurefLY diameter relationshipatexplantationis 76.5 ± 11.6mmHgjmm. This remains constant for 300 minutes of perfusion (Fig. 4). In group II (PFC) the slope starts at 39.1 ± 8.1 mm Hgjmm and demonstrates minor variation over the course of the experiment with no significant deviation from baseline. Group III (SFH) baseline slope is 50.3 ± 9.9 mm Hgjmm. This decreases within the first hour of perfusion to 32.3 ± 8.5 mm Hgjmm (p < 0.06) before failure of the preparation. Lung wetfdry ratios. The wet/dry ratio of the lungs was used as an indicator of pulmonary edema. Values obtained were compared with a value established in our laboratory for the wet/dry ratio for normal rabbits (normal value= 5.04 ± 0.19). The wetjdry ratios for the experimental groups (group I, [control], 9.04 ± 0.59; group II, [PFC], 8.06 ± 0.46; group III [SFH], 9.87 ± 0.34) were significantly greater than the normal ratio (p < 0.01). Histology. Pulmonary histologic studies from group I (control) reveal margination of white blood cells and occasional alveolar hemorrhage. Groups II and III show the presence of cellular material in the preparation. There is margination of retained white blood cells in addition to proteinaceous material and sporadic hemorrhage in the alveolar space. B.iopsy specimens from group III (SFH)

Volume 100

Preservation of heart and lungs 6 9 1

Number 5 November 1990

I ---0~1

~ 100

'-o> I

E

E c

75

1

0

·-g Q>

0:: 0

> _j

+

50

T

if)

> _j

0

o--_T

j

•1

25

1~

6

/•l--_T ___ I~t

~t/

'-Q_ w

o-oGroup I •-•Group II 1::.-t::.Group Ill

!

±

!

180

240

300

Q>

Q_

0 if)

0

0

60

120

Time from Explantation (min) Fig. 4. Slope of the LV end-systolic pressure/LV diameter (LVESP/LVD) relationship over 300 minutes for study I. Group I, autologous blood; group II, PFC; group III, SFH. tp < 0.06 versus baseline.

show an inc;rease in the thickness of arteriolar walls, consistent with vasoconstriction.

Study 2: Effects of donor pretreatment with indomethacin Pulmonary vascular resistance. Groups IV and V have similar baseline pulmonary vascular resistance (group IV, indomethacin control, 13,194 ± 1942 dynesjsecjcm- 5; group V, pretreated, 12,787 ± 1682 dynesjsecjcm- 5; Fig. 5). Pulmonary vascular resistance gradually increases in group IV (indomethacin control), becoming and remaining significant 300 minutes after explantation {18,356 ± 1784 dynesjsecjcm- 5; p < 0.05). Group V (pretreated) shows no significant alterations in pulmonary vascular resistance. Pulmonary compliance. Groups IV and V demonstrate similar initial values for pulmonary compliance (group IV [indomethacin pretreated], 4.70 ± 0.56 mm Hgjml; group V [pretreated], 5.70 ± 1.16 mljmm Hg; Fig. 6). Group IV (indomethacin control) compliance decreases significantly 240 minutes after explantation (3.06 ± 0.16; p < 0.02) and remains significantly lower than baseline for the remainder of the experiment. Group V (pretreated) demonstrates no significant changes in pulmonary compliance. Eicosanoid synthesis. The two eicosanoids of interest in this study are thromboxane and prostacyclin. Thromboxane, primarily produced by platelets, is a potent vasoconstrictor and platelet aggregator. Prostacyclin is primarily produced by endothelial cells and is a potent vasodilator and antiaggregatory for platelets. 11 These metabolites of arachidonic acid have a short half-life, so

that the stable metabolites are measured to determine the presence of the parent compounds. The metabolite of prostacyclin that we measured was 6-keto-prostaglandin F 1"' and those of thromboxane A2 are thromboxane B2 and 11-dehydro-thromboxane B2. Although actual values from individual preparations show wide intragroup variation, four of five autoperfused working heart-lung preparations in group IV (indomethacin control) demonstrate a rise in thromboxane B2 at the end of the experiment (Table 1). Additionally, 11dehydro-thromboxane B2, a longer-lived metabolite of thromboxane B2, markedly increases by the end of the experiment in all five preparations (Fig. 7). The variation in circulating thromboxane B2 in individual preparations is not unexpected because these animals underwent surgical manipulation before initiation of the actual isolated preparation. This is most likely due to variable activation of platelets, which account for most of the circulating thromboxane B2 under these conditions. It is likely that a secondary rise in thromboxane B2 can also be attributed to synthesis within the pulmonary circulation. However, this should be independent of prior surgical manipulation. Although these two sites of production can be separated experimentally, we chose not to in this study because the measurement of the 11-dehydro-thromboxane B2 metabolite would yield a reliable index of thromboxane B2 production. Although thromboxane B2 can be converted in plasma to other metabolites, a significant amount of this compound is rapidly converted to 11-dehydro-thromboxane B2 and this provides an accurate index of parent thromboxane Bz and A 2synthesis. This is a relatively new

The Journal of Thoracic and Cardiovascular Surgery

6 9 2 Kaplan et a/.

o-o Group IV

,.--, 30

•

1.[)

I

*

-- • Group V

E u

'-...... u

~ 20 '-...... (/)

v c >, -o

'--""

0 0 0

. ·. T

•1

10

X

0:::

>

o._

0 60

0

120 180 240 300 Time from Explantation (min)

360

Fig. 5. Pulmonary vascular resistance (PVR) X 1000 dynesfsecjcm- 5 over 360 minutes for study 2. Group IV, autologous blood; group V, indomethacin-pretreated autologous blood. *p < 0.05 versus baseline.

0'> ------

.---·

8

0 - 0 Group IV

I

E E "---E v

6

u

c 0

Q_

E 0

u

>, '-

4

Group V

l

--

•I

1

l~:

0

c 0

E

l

l

•

*

*

1

g

0

240

300

1

·--o

::l

o_

2

0

60

120

180

Time from Explantation (min)

360

Fig. 6. Pulmonary compliance over 360 minutes for study 2. Group IV, autologous blood; Group V, indomethacinpretreated autologous blood. *p < 0.05 versus baseline.

assay method and has not been widely used in experimental preparations. Therefore we do not know the exact quantitative relationship of 11-dehydro-thomboxane B2 to the parent compound at this time. However, appearance of this metabolite should accurately reflect thromboxane B 2 parent synthesis. We believe that this is clearly demonstrated by our findings presented in Fig. 7. Since this metabolite is a longer-lived metabolite than throm-

boxane Bz and is not artifactually generated, we chose to use this metabolite when calculating the prostacyclin/ thromboxane ratio. As noted, 6-keto-prostaglandin F 1a values remain greater than 8000 pgjml in two of the five preparations and demonstrate a decrease in the other three (Fig. 8). These differences in control values are not unusual since prostacyclin is synthesized by endothelial cells. Because of the large endothelial surface area, sur-

Volume 100

Preservation of heart and lungs

Number 5 November 1990

E

D-DGroup IV

"---ry

•-•Group V

Q_

c

0

~

D

6000

2

D

C1>

D

c

u

c 0 u

693

4500

N

rn

C1>

c 0

X

0

3000

D

_o

E

2

£

I

2 "0

1500

•

>, ·£ (J)

"0

I

0

I

Ex plantation

6 hours post explantation

Fig. 7. 11-Dehydro-thromboxane B2 concentrations at explantation and 6 hours after explantation for group IV (autologous blood) and group V (indomethacin-pretreated). ,A..

D-OGroup IV

E "---o>

•-•Group V

Q_

c 0

6000

•

+J

0

0

1....

+J

c

Q.l

u

c

4500

0

u

LJ...

c "0

c

3000 0

0 o> 0

+J

f.{)

0.._ Q_

1500

--I

I

0

Q;

":£.

I

lD

0

Ex plantation

6 hours post explontotion

Fig. 8. 6-Keto-prostaglandin F 1a concentrations at explantation and 6 hours after explantation for group IV (autologous blood) and group V (indomethacin-pretreated).

gical manipulation and perfusion techniques can markedly stimulate prostacyclin production for a number of reasons usually related to shear stress on endothelial cells. Although values greater than 8000 pgjml may actually have decreased over time, we thought that, as a matter of convenience, pursuing serial dilutions with measurements

of this magnitude would not have biologic relevance. That is, significant biologic effects from prostacyclin can be demonstrated when 6-keto-prostaglandin F I a metabolites reach concentrations of approximately 500 pgjml. To demonstrate our concept of 6-keto-prostaglandin F1a/ 11-dehydro-thromboxane B2 ratios, we chose to calculate

694

The Journal of Thoracic and Cardiovascular Surgery

Kaplan et a/.

6

0 - 0 Group IV •-•Group V

•

4 0

6 hours post explontotion

Ex plantation

Fig. 9. Ratio of 6-keto-prostaglandin F1a and 11-dehydro-thromboxane 82 at explantation and 6 hours after explantation for group IV (autologous blood) and group V (indomethacin-pretreated).

this ratio using concentrations that we were able to determine without serial dilutions (Fig~ 9). Values of thromboxane B2 in group V (pretreated) increase in one of the five preparations. In the other four preparations thromboxane B2 concentrations either decrease or remain the same (Table I). The 11-dehydrothromboxane B2 metabolite shows minor increases in all five preparations (Fig. 7). Two of five group V (pretreated) autoperfused working heart-lung preparations undergo a minor rise in 6-keto-prostaglandin F1m whereas there is a modest decrease in this metabolite in the other three preparations (Fig. 8). All five preparations in group V (pretreated) begin with a 6-keto-prostaglandin Fla/11-dehydro-thromboxane Bz ratio greater than 2 and decrease to a value greater than 0.60 (Fig. 9). Wet/dry ratio. There was no difference in the lung wetjdry ratio between either experimental group and normal rabbit lungs (group IV [indomethacin control], 5.48 ± 0.17; group V [pretreated]) 5.62 ± 0.23; normal 5.04 ± 0.19). Histology. Biopsy specimens obtained hourly reveal no alteration in pulmonary histologic features. Minimal leakage of proteinaceous material into alveolar spaces is observed. There is no alveolar hemorrhage and there is minimal margination of white blood cells. The histologic appearance of the lungs in the rabbit autoperfused working heart-lung preparation does not worsen within the time period studied.

Discussion Multiple causes have been proposed for the pulmonary vascular response seen in the autoperfused working heart-lung preparation. Previous studies have attempted to remove cellular elements from autologous blood4 or alter humoral mediators by inhibition of cyclooxygenase with indomethacin. 5 Our studies focused on two methods for altering the pulmonary vascular response. Study I diluted the blood cellular component by exchange transfusion with acellular oxygen-carrying perfusates. Study 2 inhibited eicosanoid production by pretreating the donor with intravenous indomethacin. Study 1: The effects of acellular oxygen-carrying perfusates LV performance. LV performance, measured by the slope of the LV end-systolic pressure/L V diameter relationship, is not altered by perfusion of the autoperfused working heart-lung preparation with autologous blood. This is consistent with results obtained by other groups. 2· 12· 13 Late dysfunction of the myocardium in the autoperfused working heart-lung preparation has been associated with substrate depletion, notably hypoglycemia.11· 12 Hypoglycemia was avoided by maintaining the glucose concentration greater than 80 mgjdl. This indicates that preservation of LV function can be maintained for at least 6 hours in an autoperfused working heart-lung preparation perfused with autologous blood. The effects of perfluorocarbons on LV performance

Volume 100 Number 5

Preservation of heart and lungs

November 1990

695

Table I. Thromboxane B 2 (pgjml): Study 2 Group IV: Control

Group V· Indomethacin pretreated

Time {min)

TSI

TS2

TS3

TS4

TS5

0 60 120 180 240 300 360

387 360 333 480 440 547 534

<80 93 93 93 106 <80 120

<80 <80 <80 <80 93 93 <80

280 200 173 280 187 160 320

<80 <80 <80 93 <80 <80 107

have not been investigated in the autoperfused working heart-lung preparation. Perfluorocarbons have been used as a perfusate in isolated heart models and found to be superior to isotonic salt solutions. 14• 15 However, comparisons with autologous blood perfusion have not been performed in a similar model. Our results indicate that perfusion of the autoperfused working heart-lung preparation by perfluorocarbon adequately preserves LV performance for up to 6 hours. Stroma-free hemoglobin perfusion of the autoperfused working heart-lung preparation results in a decrease in LV performance that approaches significance (p < 0.06). Our results are similar to those obtained in isolated heart preparations perfused with stroma-free hemoglobin. In these studies stroma-free hemoglobin has been associa ted with a marked increase in coronary vascular tone and decreased myocardial performance. 16• 17 Stroma-free hemoglobin is capable of binding to and inactivating endothelium-derived relaxing factor (EDRF). 18 EDRF activates guanylate cyclase increasing cyclic guanosine monophosphate. Increases in cyclic guanosine monophosphate inhibit the contractile process, resulting in vascular relaxation. 19 Inhibition ofEDRF could result in vasoconstriction in the autoperfused working heart-lung preparation. Vogel and colleagues 17 reported lack of vasoconstriction in rabbit Langendorff preparations perfused with bovine albumin, whereas perfusion with rabbit stroma-free hemoglobin resulted in coronary vasoconstriction. These observations suggest that activation of immune competent cells is not a mechanism for myocardial depression in the stroma-free hemoglobin-perfused working heart-lung preparation. Pulmonary vascular resistance. Pulmonary vascular resistance increases before failure of the autoperfused working heart-lung preparation perfused with autologous blood. Our results are in agreement with data obtained by other groups. 2• 5• 11 • 2 Cellular elements, metabolic alterations, and circulating mediatorsl.2 have been proposed as etiologic agents of the pulmonary vascular response.

°

TS6

56 124 127 115 60 62 126

TS8

358 98 575 495 197 407 340

TS/l

195 160 <65 <65 <65 <65 70

TSJ2

71 157 93 <65 <65 74 67

TSI4

<65 <65 <65 <65 <65 85 <65

Decreasing the blood-borne cellular elements by exchange transfusion and perfusion with perfluorocarbon delays the onset of pulmonary vasoconstriction. Microemboli, alteration in mediator concentrations, and decreased effectiveness of retained activated cells may play a role in this delayed response. Hypoxic vasoconstriction does not appear to be a significant factor because LV performance is not altered, indicating the absence of significant myocardial ischemia. Perfusion of the autoperfused working heart-lung preparation with stroma-free hemoglobin results in a rapid rise in pulmonary vascular resistance. This is associated with a decrease in LV performance. EDRF is also produced by pulmonary artery endothelial cells. 21 Inactivation ofEDRF by binding with stroma-free hemoglobin could lead to the increase in pulmonary vascular resistance. Since LV performance and pulmonary vascular resistance deteriorate simultaneously, hypoxic pulmonary vasoconstriction cannot be ruled out as a mechanism for failure of the preparation. Whether pulmonary hypertension results in decreased LV preload or decreased myocardial performance results in pulmonary hypertension cannot be determined from this study. Extravascular lung water and histology. Groups I (control), II (PFC), and III (SFH) all demonstrate an increase in extravascular lung water as measured by lung wet/dry ratio. This increase in extravascular lung water could be due to either an increase in hydrostatic pressure or an alteration in permeability as a result of endothelial damage. Histologic examination, although inconclusive, favors an increase in hydrostatic pressure as the causative factor. In the three groups, there is occasional but not consistent alveolar hemorrhage. Additionally, little endothelial damage is noted in capillaries. Experiments designed to determine the presence of a transcapillary protein leak (i.e., infusion of albumin labeled with radioactive isotopes) would be helpful.

Study 2: Effects of donor pretreatment with indomethacin

The Joutna1 of Thoracic and Cardiovascular Surgery

6 9 6 Kaplan et a/.

Pulmonary vascular resistance and eicosanoids. Perfusion of the autoperfused working heart-lung preparation with indomethacin-pretreated autologous blood attenuates the pulmonary vascular response both in our studies and in others. 5 Thromboxane A 2 and prostacyclin are both products of cyclooxygenase conversion of arachidonic acid. Indomethacin will alter not only levels of individual eicosanoids but also the prostacyclin/thromboxane A 2 ratio, which we believe reflects a "net physiologic index." Although shown to be important in the pathology of renal ischemia (increase in the prostacyclin/thromboxane A 2 ratio results in decreased renal damage from ischemia), 22 a relationship between the prostacyclin/ thromboxane A 2 ratio and the physiologic response of the autoperfused working heart-lung preparation has not been previously shown. However, it is known that increases in thromboxane systhesis can be directly eorrelated to rises in pulmonary vascular resistance in intact animals exposed to venovenous bypass or endotoxin infusion. 23 · 24 These studies also demonstrated that the pulmonary vascular pressure increase was essentially ablated with concomitant increases in prostacyclin synthesis. Thus it appears that increasing prostacyclin/thromboxane ratios were important in these experiments. In this study the prostacyclin/thromboxane A2 ratio was reflected in the ratio of the stable metabolites 6-ketoprostaglandin F 1a and 11-dehydro-thromboxane B2. Treatment with indomethacin lowered the concentration of all metabolites, but the prostacyclin/thromboxane A 2 ratio was higher than in group IV (indomethacin control). Additionally, the ratio declines more rapidly in group IV (indomethacin control) (Fig. 9). This observation correlates with the maintenance of normal pulmonary vascular resistance in the indomethacin-treated group (group V). Thus the prostacyclin/thromboxane ratio appears to be important, with higher ratios being associated with normal pulmonary vascular resistance in the autoperfused working heart-lung preparation after 6 hours of perfusion with indomethacin-pretreated autologous blood. Pulmonary compliance. Thromboxane A 2 is a potent stimulus for airway constriction and can also alter capillary permeability in intact animal systerns. 25 · 26 The origin of thromboxane A2 is unknown, but secretion could result from platelets or leukocytes activated by the extracorporeal circuit27· 28 or from synthesis by lung tissue. Indomethacin pretreatment inhibits cyclooxygenase, decreasing the amount of thromboxane A2. This decrease in thromboxane A 2secretion is associated with maintenance of normal pulmonary compliance in the rabbit autoperfused working heart-lung preparation.

Extravascular lung water and histology. There is no increase in extravascular lung water in group IV (indomethacin control). Also, there is no alteration in histologic appearance in biopsy specimens obtained from either group. Lack of increase in wet/dry ratio hints that the increase in extravascular lung water observed in the perfusate study was due to increased vascular hydrostatic pressure, not to an increase in interstitial oncotic pressure caused by capillary leak. Conclusions 1. Perfusion of the autoperfused working heart-lung preparation with stroma-free hemoglobin results in a marked decrease in viability when compared with perfusion with autologous blood or perfluorocarbon. This decrease in viability is associated with both a decrease in LV performance and an increase in pulmonary vascular resistance. 2. Perfusion with autologous blood or perfluorocarbon maintains normal LV performance. 3. Perfusion with perfluorocarbon delays the increase in pulmonary vascular resistance when compared with perfusion with autologous blood. This change is not associated with an increase in the viability of the autoperfused working heart-lung preparation. 4. Pretreatment of the donor with indomethacin results in lower concentrations of the metabolites of thromboxane A 2 and prostacyclin and an increase in the prostacyclin/thromboxane A 2 ratio. 5. Increasing the prostacyclin/thromboxane A2 ratio abrogates both the decrease in pulmonary compliance and the increase in pulmonary vascular resistance observed in the autoperfused working heart-lung preparation perfused with untreated autologous blood. Alteration or blockade of specific humoral mediators may prove advantageous for increasing the·utility of the autoperfused working heart-lung preparation as a preservation technique for heart, lung, and heart-lung transplantation. We would like to thank Constance L. Otradovec, MA, Department of Epidemiology, Francis Stern USDA Human Nutrition Research Center, for assistance with statistical analysis and Lori Hayes for preparation of histologic samples. REFERENCES I. Muskett AD, Burton NA, Gay W A Jr, Miller M, Rabkin MS. Preservation in the rabbit autoperfusing heart-lung preparation: a potential role for indomethacin. Surg Forum 1986;37;252-4. 2. Kontos GJ, Borkon AM, Adachi H, et al. Successful

Volume 100 Number 5 November 1990

extended cardiopulmonary preservation in the autoperfused working heart-lung preparation. Surgery 1987; 102:269-7 6. 3. Kontos GJ, Borkon M, Baumgartner WA, et al. Neurohumoral modulation of the pulmonary vasoconstrictor response in the autoperfused working heart-lung preparation during cardiopulmonary preservation. Transplantation 1988;45:275-9. 4. Kontos GJ, Borkon AM, Adachi H, et al. Leukocyte depletion ameliorates the pulmonary vasoconstrictor response in the autoperfused working heart-lung preparation. Surg Forum 1986;27:255-6. 5. Muskett A, Burton NA, Grossman M, Gay W A. The rabbit autoperfusing heart-lung preparation. J Surg Res 1988; 44:104-8. 6. Guide for the care and use of laboratory animals. NIH publication No. 80-23, revised 1978. 7. Suga H, Sagawa K. Assessment of absolute volume from diameter ofthe intact canine left ventricular cavity. J App Physiol 1974;36:496-9. 8. Badellino MM, Deeb GM, Horowitz J, eta!. Pressure volume relationships as a determinant of myocardial ventricular function in an autoperfusing heart-lung preparation. Transplant Proc 1988;20(suppl1):820-2. 9. Hales CA, Sonne L, Petersen MB, Kong D, Miller M, Watkins WD. Role of thromboxane and prostacyclin in pulmonary vasomotor changes after endotoxin in dogs. J Clin Invest 1981;68:497-505. 10. Westland P, Kumlin M, Nordenstrom A, Granstrom E. Circulating and urinary thromboxane B2 metabolites in the rabbit: 11-dehydro-thromboxane B2 as parameter of thromboxane production. Prostaglandins 1986;31 :413-43. 11. Murphy RC. Biosynthesis and metabolism. In: Watkins WD, ed. Prostaglandins in clinical practice. New York: Raven Press, 1989:1-20. 12. Kontos GJ, Borkon AM, Baumgartner WA, et al. Improved myocardial and pulmonary preservation by metabolic substrate enhancement in the autoperfused working heart-lung preparation. J Heart Transplant 1988;7:140-4. 13. Robiscek F, Masters TN, Duncan GD, Denyer MH, Rise HE, Etchison M. An autoperfused heart-lung-preparation: metabolism and function. Heart Transplant 1985;4:334-8. 14. Chemmitius JM, Burger W, Bing RJ. Crystalloid and perfluorochemical perfusates in an isolated working heart preparation. Am J Physiol 1985;249:H285-92. 15. Deutschman W, Linder E, Deutschlander N. Perftuorochemical perfusion of the isolated guinea pig heart. Pharmacology 1984;28:336-42. 16. Saraudean S, Fallon ST, Austen WG, Goodman AJ. Stroma free hemoglobin solution for perfusion of the isolated lamb heart at 38° C. Trans Am Soc Artif Intern Organs 1978;24:261-9. 17. Vogel WM, Dennis RC, Cassidy G, Apstein CS, Valeri CR. Coronary constrictor effect of stroma free hemoglobin solutions. Am J Physioll986;251:H413-20. 18. Martin W, Villani GB, Jothanandain D, Furchgott RF.

Preservation of heart and lungs 6 9 7

19. 20.

21.

22.

23.

24.

25.

26.

27.

28.

Selective blockage of endothelium dependent and glycerol trinitrate induced relaxation by hemoglobin and methylene blue in rabbit aorta. J Pharmacal Exp Ther 1985;32:70816. Vanhoutte PM. The endothelium -modulataer of vascular smooth muscle tone. N Eng J Med 1988; 319:512-3. Miyamato Y, Lajos TZ, Bhayana JN, Bergsland J, Celik CF. Physiologic constraints in autoperfused heart lung preservation. J Heart Transplant 1987;6:261-6. lgnarro LJ, Burns RE, Buga GM, Wood KS. Endotheliumderived relaxing factor from pulmonary artery and vein possess pharmacologic and chemical properties identical to those of nitric oxide radical. Circ Res 1987;61:866-79. Klausner JM, Paterson IS, Kobzik L, et al. Vasodilating prostaglandins attenuate ischemic renal injury only if thromboxane is inhibited. Ann Surg 1989;209:219-24. Petersen MB, Huttemeier PC, Zapol WM, Martin EG, Watkins WD. Thromboxane mediates acute pulmonary hypertension in sheep extracorporeal perfusion. Am J Physioll982;243:471-9. Huttemeier PC, Watkins WD, Petersen MD, Zapol WM. Acute pulmonary hypertension in lung thromboxane released following endotoxin infusion in normal and leukopenic sheep. Circ Res 1982;50:688-94. Svensson J, Strandberg K, Tuvemo TK, Hamberg M. Thromboxane A2: Effects on airway and vascular smooth muscle. Prostaglandin 1977;14:425-36. Garcia-Szeabo RR, Petersen MB, Watkins WD, Bizios R, Kong DL, Mallack AB. Thromboxane generation after thrombin: protective effect of thromboxane synthetase inhibition on lung fluid balance. Circ Res 1983;53:214-22. Davidson EM, Doig MY, Ford-Hutchinson A W, Smith MJ. Prostaglandin and thromboxane production in rabbit polymorphonuclear leukocytes and rat macrophages. Adv Prost Thrombox Res 1980;8:1661-3. Watkins WD, Peterson MB, Kong DL, Kono K, Buckley MJ, Philibin OM. Thromboxane and prostacyclin changes during cardiopulmonary bypass with and without pulsatile flow. J THORAC CARDIOVASC SURG 1982;84:250-6.

Discussion Dr. W. Gerald Rainer (Denver, Colo.). Dr. Kaplan, I have great respect for the enormous amount of work that you have accomplished here. It has incredible practical applications toward the efforts of extending the viability of transplant systems. First, I would like to ask you a couple of questions. One has to do with the inspired oxygen concentration (Fio2) of 1.0, used in your indomethacin control group, whereas you used an Fio2 of room air in the other groups. Might this not have had some effect on the parameters that you were measuring regarding the lung physiology? Dr. Kaplan. The Fio2 in all groups, except for the perfluorocarbon preparations, was 0.21. Only the perftuorocarbon group had an Fio 2 greater than room air. In these preparations the Fio2 was maintained at 1.0. Oxygen-carrying capacity for per-

6 9 8 Kaplan et a/.

fluorocarbon does not follow sigmoidal binding. It is dependent on Henry's law, thus requiring a high Fio2. This requirement for a high Fio2 is one of the drawbacks to the clinical use ofperfluorocarbons. Dr. Rainer. You just opened up the next question that I had, because one of the qualifications for my being here is the poor result we have had in attempting to do the same thing with other forms of perfluorocarbon. We have used perfluoro-octylbromide in attempts to prolong the survival oflungs, and it has an oxygen-carrying capacity five times greater than that of the perftuorocarbon that you are using. Might this have allowed you to use the Fio 2 of room air if you had had a higher oxygen-carrying capacity in your perftuorocarbon? Dr. Kaplan. It is possible that might have allowed us to use a lower Fio 2. The reason for using perftuorocarbons was to investigate removal of cellular elements from the perfusate. We believe that, given the requirment of an Fio2 of 1.0 for perfluorocompound, it is not clinically applicable for preservation of the lung at this time. Dr. Rainer. One other question I ask you concerns Fig. 4. For your control group, the LV end-systolic pressure/L V diameter slope is said to remain constant for 300 minutes, whereas it actually started at around 75 at T 0, drops to around 50, and then ends up at about 60 at 300 minutes. Dr. Kaplan. Although there is a decrease, it never achieved statistical significance. There is, however, a trend toward a decrease in myocardial function at that point. Dr. Rainer. We are to beindebted to you and your group for this work for several reasons: one is the elimination of the stroma-free hemoglobin and a second, in addition to using an acellular perfusate, is the use of other adjuvants, indomethacin being only one of them, to counteract some of the metabolites that you mentioned. Dr. William Y. Moores (Del Mar, Calif). Past experience with stroma-free hemoglobin solutions in rabbit models has been problematic. In fact, your ability to demonstrate any function at all with a rabbit model is encouraging. We have done most of our work in a swine model and have found the stroma-free hemoglobin solutions to be efficacious in this application. Other investigators, however, trying to duplicate those studies in rabbit models have not been successful. I think, therefore, that before you come to the conclusion that stroma-free hemoglobin solution is not efficacious, it is important to look at it in more detail and in other animal models as well. The other point that needs to be made is that stroma-free hemoglobin solution is not a single entity with uniform properties. It is a group of biologic products, which adds to the difficulty in obtaining a uniform product. There are many different formulations. The manufacturing of stroma-free hemoglobin solutions is a rapidly evolving field with second- and third-generation solutions being produced. I would, therefore, urge cau-

The Journal of Thoracic and Cardiovascular Surgery

tion in coming to the conclusion that stroma-free hemoglobin solutions are not efficacious. Dr. Kaplan. I am aware that stroma-free hemoglobin has been used in other models. There are several reports in the literature with rabbit and lamb models that have demonstrated a profound vasoconstrictive effect with stroma-free hemoglobin. We believe that a vasoconstrictive response is occurring, explaining both the decrease in LV function and the increase in pulmonary vascular resistance. We realize that further work is being done and results are difficult to interpret at this point. Dr. Alden H. Harken (Denver, Colo.). I want to expand on the observations of a couple of the previous discussants. We have been examining rats and rabbits in a more acute model than yours. However, in looking at the difference between crystalloid and blood perfused, our initial bias was that the oxygen delivery was tremendously important. We then progressively decreased the hemoglobin concentration in the reperfusate and found that we got just as much protection from a hemoglobin concentration of 2 gm%. Jamie Brown and Mike Grosso in our laboratory then treated red cells with carbon monoxide and found the same kind of protection, indicating that perhaps it was not oxygen delivery. We then looked at the antioxidant balance, and this is the basis for my question. I recommend that you look at the prostacyclin-thromboxane relationship, but would ask whether there is some other kind of antioxidant in the blood. We think. it is glutathione. Have you looked in this preparation at the antioxidant levels (GSH:GSSG ratios) in the tissue at various phases throughout this perfusion process? Dr. Kaplan. The physiologic alterations are dependent on more than prostaglandin synthesis and extraction. At present, we are in the midst of several studies using this model. One of the investigations is directed at the effect of various antioxidants, such as catalase and superoxide dismutase. This work is in progress at present, and we hope to be reporting the results in the near future. Dr. Harken. In that heart model, we actually blocked catalase and superoxide dismutase, and that did not seem to hurt the results. We do think it is the antioxidant that is operative. To expand, just one more brief comment. We also have a lung model, a crystalloid kind of Langendorf perfusion model. Again, I think your pinpointing the oxidant:antioxidant system is important. However, we have found blood to be a potent hydrogen ion buffer and, in pulmonary perfusion, we have found maintaining acid-base status absolutely stable was critical to the maintenance of pulmonary vascular resistance. I did not catch your maintenance or your evaluation of hydrogen ion concentration in the pulmonary perfusion component of this model. Dr. Kaplan. We performed random blood gas measurements and found that the pH ranged from 7.3 to 7.45. Dr. Harken. Because the pH is a log funci.ion, we think 7.3 to 7.45 may be a big difference.