- Email: [email protected]
Effect of cardiopulmonary bypass on the circulating level of soluble GMP-140
Effect of Cardiopulmonary Bypass on the Circulating Level of Soluble GMP-140 Hiroyoshi Komai, MD, and Sheila G. Haworth, FRep Vascular Biology and Pharmacology Unit, Institute of Child Health, London, England, and Cardiothoracic Unit, The Hospital For Sick Children, London, England
Soluble GMP-140 can prevent the adhesion of activated neutrophils to endothelium in vitro. Because cardiopulmonary bypass causes neutrophil-cendothelial interaction, the plasma level of soluble GMP-140 was measured using an enzyme immunoassay system in 32 children undergoing intracardiac repair of congenital heart disease. They had either a high, low, or normal pulmonary blood flow (n = 13, 12, and 7 respectively). Because activated platelets are a source of GMP-140, the plasma l3-thromboglobulin level was also measured. Blood was sampled before, during, and for 24 hours after cardiopulmonary bypass. Plasma levels of both soluble GMP-140 and l3-thromboglobulin increased after the onset of cardiopulmonary bypass in all patients but for both substances the increase was greater in those with a low
pulmonary blood flow ip < 0.05 for all comparisons). The sum total of soluble GMP-140 values during and after operation was correlated negatively with the preoperative mean pulmonary arterial pressure (p < 0.05 for all time intervals). GMP-140 level correlated with the plasma l3-thromboglobulin level (r = 0.5, P < 0.05). This work supports the contention that soluble GMP-140 is released from activated platelets during cardiopulmonary bypass, the level being particularly high in those who had intrinsically abnormal platelets preoperatively in association with a low pulmonary blood flow. Patients with a high pulmonary blood flow, who are more susceptible to endothelial cell injury, may be less well protected by soluble GMP-140. (Ann Thorae Surg 1994;58:478-82)
C
GMP-140 in the circulation is unknown [10], although we know that platelets release it when activated. Cardiopulmonary bypass is thought to be associated with activation of endothelium, leukocytes, and platelets. It can have a more injurious effect on infants and young children than on adults [1]. We measured the levels of soluble GMP-140 in 32 children with congenital heart disease who were undergoing intracardiac repair. The plasma J3-thromboglobulin level was also measured as a marker of platelet activation [12].
ard iop u lm onary bypass (CPB) causes transient multiple organ dysfunction [1]. Leukocyte-endothelial cell interaction is an important factor in the response to this injury, as has been emphasized in a recent study [2] on the effects of CPB in children. When leukocytes and endothelium are stimulated they each express specific molecules on their cell surface, known as adhesion molecules [3], that promote the adhesion of leukocyte to endothelium. Adhesion is essential for further leukocyte activity, such as release of lysosomal enzymes and infiltration into the perivascular tissue. Soluble GMP-140 can prevent adhesion of activated neutrophils to the endothelium in vitro. Soluble GMP-140 is the soluble component of GMP-140 [4], an adhesion molecule first identified in platelet a-granules.lt also is synthesized by endothelial cells [5]. GMP-140 is also known as P-selectin. Expression of GMP-140 on the cell surface of human umbilical vein endothelial cells is induced rapidly, within 3 minutes, by histamine, thrombin, phorbol myristate acetate, calcium ionophore, and activated complement [6, 7]. Both megakaryocytes and endothelial cells have a messenger RNA for GMP-140 that lacks the transmembrane domain [8, 9]. Recent studies [10] have shown that the soluble component of GMP-140 is detectable in normal human plasma. In vitro, this molecule can prevent the adhesion of activated neutrophils to the endothelium [11]. The origin of soluble Accepted for publication Dec 16, 1993. Address reprint requests to Prof Haworth, Vascular Biology and Pharmacology Unit, Institute of Child Health, 30 Guilford St, London, WC1N 1EH, England.
© 1994 by The Society of Thoracic Surgeons
Material and Methods Patients Thirty-two children undergoing elective open heart operations with CPB were studied. None suffered from extracardiac disease. They were divided into three groups according to the preoperative pulmonary blood flow (Qp), One group consisted of 13 children with a high pulmonary blood flow due to a left to right intracardiac shunt (high Qp), a second group of 12 children had a reduced pulmonary blood flow (low Qp), and a third group consisted of 7 children with aortic stenosis who had a normal flow (normal Qp). The diagnoses, age, sex, and preoperative cardiac catheterization data are shown in Table 1. Only 20 children required cardiac catheterization; treatment of the remaining patients was determined by cross-sectional echocardiography only. In the present study blood samples were taken only from lines inserted for: clinical management purposes. Permission to carry out this study was 0003-4975/94/$7.00
KOMAI AND HAWORTH SOLUBLE GMP-140 AND CARDIOPULMONARY BYPASS
Ann Thorac Surg 1994;58:478-82
Blood Sampling
Table 1. Patients' Preoperative Clinical Data Normal Qp
High Qp
LowQp
No. of patients Diagnosis
7 AS7
Sex ratio (M:F) Median age (y) Age range (y) PAP (mm Hg) Pp/Ps Qp/Qs % O 2 Sat
1:6 8.5 1.7-14.2 17.6 2: 2.5 0.3 2: 0.04 1 98.2 2: 0.7
13 ASD3 VSD4 AVSD5" TAPVR 1 1:3.3 1.9 0.4-14.1 27.9 2: 3.9 b 0.52: O.le 2.7 2: 0.6 c 95.82: 1.7
12 TF6 DORV2 TA 1 DlV3 1:0.5 3.8 1.1-16.4 13.72: 1.5 0.22:0.1 0.5 2: 0.1 82.02: 3.9 d
Characteristic
479
Three complete, 2 partial AVSDs. b Significantly greater than in the normal or low Qp groups (p < 0.05 and p < 0.01, respectively). C Significantly greater than in the normal or low Qp groups (p < 0.01 for both). d Significantly lower than in the normal or high Qp groups (p < 0.01 for both). a
ASD ~ secundum atrial septal defect; AVSD = AS = aortic stenosis; atrioventricular septal defect; DIV = double-inlet ventricle; DORV = double-outlet right ventricle; PAP = mean pulmonary arterial pressure (SEM); Pp/Ps = pulmonary to systemic systolic arterial pressure ratio; % O 2 Sat = arterial oxygen saturation (SEM); Qp = pulmonary blood flow; Qp/Qs = pulmonary to systemic arterial blood flow ratio (SEM); SEM = standard error of the mean; TA = tricuspid atresia; TAPVR = total anomalous pulmonary venous return; TF = tetralogy of Fallot; VSD = ventricular septal defect.
given by the Ethics Committee of The Hospital for Sick Children, London, England.
Surgical Management and CPB All medication was stopped at least 12 hours before the operation. Conventional CPB was used with one episode of aortic cross-clamping and multiple dose cardioplegia. After heparin (300 IU /kg) was given they underwent extracorporeal circulation at 2.4 L· m- 2 • min- 1 of perfusion flow in moderate to deep hypothermia (17° to 28°C). Aortic cross-clamping was followed by infusion of St. Thomas' cardioplegic solution. Cardiopulmonary bypass equipment consisted of a nonpulsatile roller pump (Dideco D701 [Mirandola, Italy] or Maxima membrane oxygenator [Medtronic, Anaheim, CAD and Dideco 752 or Sorin venous cardiotomy reservoir (Dideco). During CPB, the patients' hematocrit values were kept between 15% and 29%. Sodium nitroprusside was infused during the rewarming period and stopped just before the discontinuation of CPB. During and after the operation all patients were given packed red cells, crystalloid, colloid, or a combination of these for supplement of circulating blood volume. The duration of CPB ranged from 29 to 238 minutes (80.7 ::':: 13.2 minutes in those with a normal Qp, 82.7 ::':: 8.6 minutes in those with a high Qp, and 129.7 ::':: 17.1 minutes in those with a low Qp): duration of CPB for the low Qp group was significantly longer than that of the other groups (p < 0.05 for both). The aortic cross-clamping time varied from 10 to 107 minutes (41.4::':: 8.8 minutes in those with a normal Qp, 44.8 ::':: 6.6 minutes in those with a high Qp, and 62.0 ::':: 8.5 minutes in the low Qp group) with no significant differences between the groups.
Seven heparinized blood samples were taken from a radial or femoral arterial line (1) before CPB (just after heparinization and before cannulation), (2) within 5 minutes of removing the aortic cross-clamps, (3) having been off CPB for a few minutes, and (4) 20 minutes, (5) 3 hours, (6) 6 hours, and (7) 24 hours after discontinuing CPB. All blood samples were immediately centrifuged and plasma was kept frozen at -70°C until assayed.
Plasma GMP-140 Assay The soluble form of GMP-140 was measured by an enzyme immunoassay system using two different monoclonal antibodies specific for soluble GMP-140 (GMP-140 EIA kit; Takara Shuzo Co Ltd, Kyoto, Japan). Intraassay and interassay variations were 12.5% and 8.9%, respectively. In addition, the plasma {3-thromboglobulin was measured in the same samples taken from 19 patients (9 in the high Qp and 10 in the low Qp groups) as a marker of the activation and degranulation of platelet a-granules. Platelet-poor plasma was assayed with a radioimmunoassay system using a monoclonal antibody specific for {3-thromboglobulin ({3-thromboglobulin RIA kit; Kodak Clinical Diagnostics Ltd, Amersham, UK). The intraassay and interassay variations for this assay were 4.4% and 9.7%, respectively.
Statistical Analysis All results are expressed as the mean ::':: standard error of the mean. The plasma levels of both soluble GMP-140 and {3-thromboglobulin in each sample were standardized to the pre-CPB level using the hematocrit (Hct) value to mitigate the effect of hemodilution: Corrected soluble GMP-l40 or corrected (3-thromboglobulin = X
measured level
Hct before CBP/Hct when sample was taken)
Statistical evaluation of both components was performed using analysis of variance for the factorial data and Dunnett's t test for unpaired observations. To assess differences between the three groups in the level of each component during and after CPB, we calculated the sum total value of soluble GMP-140 and {3-thromboglobulin in each group and then compared them using analysis of variance. Correlations were analyzed using the Spearman rank correlation coefficient. Values were considered to be statistically significant when p was less than 0.05.
Results Soluble GMP-140 Levels Before CPB The mean level of soluble GMP-140 before CPB was significantly higher in the low Qp group than in either of the other groups (p < 0.05 for both comparisons), which were similar to each other (Fig 1). For all patients the soluble GMP-140 level before CPB was correlated negatively with the preoperative mean pulmonary arterial pressure (r = -0.5; P < 0.05) and positively with the preoperative hemoglobin level (r = 0.6, P < 0.01). There
480
KOMAI AND HAWORTH SOLUBLE GMP-140 AND CARDIOPULMONARY BYPASS
Ann Thorac Surg 1994;58:478-82
BOO
,.
3000
600
Total soluble GMP-140 (nglml)
Soluble
GMP-140 (nglml)
•• •• • • •• • • • •
4000
400
2000
1000 200
0 0
o --'---.---,---.---,---.----,-----,PRE
OAe
08
20mln
3H
6H
10
Fig 1. Mean plasma soluble GMP-140 levels (ng/mL) In patients with a low pulmonary blood flow (closed circles), a high flew (open circles), and a normal flow (open squares). Levels were obtained before bypass (PRE), immediately after the aortic cross-clamps were taken off (OAC), after coming off bypass (OB), and 20 minutes (20rnin),3 hours (3H), 6 hours (6H), and 24 hours (24H) after coming off bypass. (*p < 0.01; **p < 0.05; tp < 0.001 in comparison with the prebypass (PRE) data in each group).
were no significant correlations between the soluble GMP140 levels and either age or sex.
Soluble GMP-140 Levels During and After CPB In all groups of children the level of soluble GMP-140 increased after the onset of CPB (see Fig 1). After removal of the aortic cross-clamps the level increased progressively to reach a peak at 3 to 6 hours. In all three groups of children the level of soluble GMP-140 at each time of sampling from 20 minutes to 6 hours after CPB was significantly higher than it was before CPB. The total of soluble GMP-140 values during the operation (the sum total of values obtained at the three time intervals from removal of the aortic cross-clamps until 20 minutes after CPB) was significantly higher in the low Qp group compared with both the high and normal Qp groups (p < 0.01 for both comparisons), which were similar to each other (see Fig 1). After the operation (between 3 and 24 hours after CPB) the total soluble GMP-140 values of the low Qp group were higher than those of the high Qp group (p < 0.01), and there was no significant difference between the high and normal Qp groups in this regard. There were no significant correlations between the level of soluble GMP140 and the duration of either CPB or aortic cross-clamping time within any of the groups or in the total of 32 patients.
Relationship Between Soluble GMP-140 Level and Preoperative Pulmonary Arterial Pressure Of the 32 patients investigated, only 20 had had a preoperative cardiac catheterization study. There were negative correlations between the total soluble GMP-140 and preoperative mean pulmonary arterial pressure during the operation itself, the postoperative period, and the whole study period (r = -0.6, -0.6, and -0.7, respectively; p <
30
•
40
50
preoperative mPAP (mmHg)
24H
Time
20
Fig 2. Relation between preoperative mean pulmonary arterial pressure (mPAP) (measured in 18 patients at cardiac catheterization, irrespective of magnitude of pulmonary blood flow) and sum total of soluble GMP-140 values during the whole study period (r = -0.7; P < O.OV.
0.01 for whole study period, p < 0.05 for the other study times) (Fig 2). In addition, the total soluble GMP-140 during the operation itself, during the postoperative period, and during the whole study period was correlated negatively with the preoperative pulmonary to systemic arterial pressure ratio (r = -0.8, -0.6, and -0.7, respectively; p < 0.05 for postoperative period; p < 0.01 for the other comparisons).
Plasma {3- Thromboglobulin Levels During and After CPB Preoperatively, the mean f3-thromboglobulin level was higher in patients with a low Qp than in those with high Qp, but not significantly so (Fig 3). The levels increased after onset of CPB (p < 0.05) and returned to the pre-CPB level 20 minutes after discontinuing CPB in both groups. The sum total of values for f3-thromboglobulin during the whole study period was significantly higher in the low Qp group than in the high Qp group (p < 0.05). There were no significant correlations between the level of plasma
100
80 f3-thromboglobulin
60
(ng/ml)
40 20
o...l-..--..--..---.---.---.---.-PRE
OAe
08
20min 3H
6H
24H
Time
Fig 3. Mean ptasma (3-thromboglobulin levels (ngjml.) in patients with low pulmonary blood flow (closed circles), and high flow (open circles). Time intervals as in Fig 1. (*p < 0.01; **p < 0.05.)
KOMAI AND HAWORTH SOLUBLE GMP-140 AND CARDIOPULMONARY BYPASS
Ann Thorac Surg 1994;58:478-82
400 300 Total soluble GMP-140 (ng/ml)
200 100
• • • • • • • '. • •• •• •• •
.
0+--.---.---.---.--.-----.---, 50 100 150 200 250 300 350 400 Total l3-thromboglobulin (ng/ml) Fig 4. Relation between sum total of soluble GMP-140 (ng/mL) values during the whole study period and sum total of f3-thromboglobulin, (nglml.) during the operation (r = 0.5; P < 0.05).
l3-thromboglobulin and the duration of either CPB or the aortic cross-clamping time.
Relationship Between Soluble GMP-140 Levels and Plasma f3- Thromboglobulin Levels The sum total of plasma l3-thromboglobulin levels during the operation was compared with the total soluble GMP140 during the whole study period. There was a significant positive correlation between them, as shown in Figure 4 (r = 0.5; p < 0.05).
Comment Preoperatively, the level of soluble GMP-140 was higher in children who had a low pulmonary blood flow than in those with either a normal or high flow, and it was correlated inversely with the mean pulmonary arterial pressure. Patients with the highest hemoglobin concentrations, reflecting the severity of chronic hypoxemia, had the highest plasma levels of soluble GMP-140. The source of the elevated level of soluble GMP-140 is unknown. However, platelet function can be abnormal in patients with cyanotic congenital heart disease [12-14], and in vitro studies indicate that soluble GMP-140 is released from activated platelets [10]. Degranulated platelets release a-granules and thromboglobulin probably is located within these granules [I2]. The plasma level of l3-thromboglobulin is a marker of platelet activation and in the present study the level was higher in those patients with a low pulmonary blood flow, although the difference was not statistically significant. Thus the abnormal platelets might be the source of the elevated level of GMP-140 found in the cyanotic patients with low pulmonary blood flow. Endothelium is another possible source, but there is no evidence that human endothelial cells that can express cell surface GMP-140 can release soluble GMP-140, nor do the findings in the present study suggest such endothelial release. Moreover, the level of soluble GMP-140 was elevated in patients with a low pulmonary blood flow (which does not damage the endothelium) and normal in those with a high pulmonary blood flow [15-17] (where the
481
endothelium is exposed to an increase in shear stress and is morphologically and functionally abnormal) . After the onset of cardiopulmonary bypass the level of soluble GMP-140 increased in all patients. In this study it obviously was not possible to distinguish between the effect of cardiopulmonary bypass itself and that of surgical trauma on the level of soluble GMP-140. However, we investigated this group of patients because soluble GMP140 is thought to reduce endothelial neutrophil interaction, and this interaction is promoted by cardiopulmonary bypass. The level of soluble GMP-140 increased in all patients irrespective of the preoperative pulmonary blood flow, but it was consistently higher in those who had had a low flow preoperatively and it remained higher in these patients after bypass was discontinued. Like prebypass plasma levels, the total soluble GMP level during and after the operation was correlated inversely with preoperative pulmonary arterial pressure. By demonstrating an increase in the expression of cell surface GMP-140, Rinder and associates [IS] showed that platelet activation occurs during open heart operations. The expression of GMP-140 increased within 10 minutes of starting bypass and returned to the prebypass level 18 hours after discontinuing bypass. Although the time course is not identical with that seen in the present study, it is consistent with our hypothesis that soluble GMP-140 is released from platelets activated during bypass. This hypothesis is supported by the observation that plasma l3-thromboglobulin levels were higher in children with a low Qp than those with a high Qp during CPB. The level of l3-thromboglobulin increased during CPB in all patients, as has been shown previously [19], but was particularly high in those with a low preoperative flow. Also there was a significant, positive correlation between the amount of GMP-140 released during the whole period of this study and the plasma l3-thromboglobulin level. Thus the excessively high levels of soluble GMP-140 seen during and after CPB in patients who had had a low pulmonary blood flow might be an exaggeration of the prebypass picture also indicative of excessive release from intrinsically abnormal platelets undergoing further activation on bypass. In speculating on the possible influence of soluble GMP140 on leukocyte-endothelial cell interaction, we do not know whether the soluble GMP-140 level we measured in our patients using an enzyme immunoassay system was biologically active. However, Dunlop and associates [10] showed that soluble GMP-140 extracted from human plasma was functional by demonstrating in vitro that it bound to the GMP-140 receptors on the neutrophils to prevent adhesion. Dunlop reported a normal human plasma level that was comparable with preoperative values in the present study [10]. Gamble and associates [11] showed that soluble GMP-140 purified from animal platelets inhibited the adhesion of neutrophils to the endothelium in vitro. It appeared to do so by inhibiting both GMP-140 -mediated adhesion and CD11 b / CD18-mediated adhesion, which is thought to be important in CPBmediated endothelial cell injury [18, 20]. Although the mechanism is unknown, a high soluble GMP-140 level may work as a feedback mechanism to prevent an extended
482
Ann Thorac Surg 1994;58:478-82
KOMAI AND HAWORTH SOLUBLE GMP-140 AND CARDIOPULMONARY BYPASS
inflammatory reaction. Gamble and associates [11] showed that a linear, dose-dependent increase in soluble GMP-140 induced inhibition of neutrophil adhesion onto the endotheliaI layer at a relatively high dose of 10 JLg/mL. The in vivo saturation of unstimulated neutrophil GMP-140 receptors is estimated to be only 20% to 40% nO] in the presence of a low plasma level of soluble GMP-140 (approximately 0.2 JLg/mL), suggesting the possibility of further receptor binding. Cardiopulmonary bypass is likely to have a more injurious effect on the pulmonary endothelium of patients whose endothelium was injured preoperatively by a high pulmonary blood flow [21, 22], but if the plasma level resembles the microvascular level then these vulnerable patients are less protected from neutrophil-endothelial cell interaction than those who had low preoperative flow. In conclusion, the level of soluble GMP-140 increases during and after CPB, more so in children who had low (compared with high) preoperative pulmonary blood flow and pulmonary arterial pressure. A possible source of soluble GMP-140 may be platelets activated by CPB; the particularly high levels in those who had had a low pulmonary blood flow are attributed to the abnormal platelets found in such patients. The present study showed a significant, positive correlation between soluble GMP140 and plasma {Hhromboglobulin level, a marker of platelet activation. Soluble GMP-140 is thought to help inhibit adhesion of neutrophils to the endothelium and therefore the most vulnerable patients (those with high preoperative blood flow) are the least well protected. This work was supported by the British Heart Foundation and the Wellcome Trust.
7.
8.
9.
10. 11. 12. 13. 14.
15. 16.
17.
References 1. Kirklin JK, Westaby S, Blackstone EH, Kirklin JW, Chenoweth DE, Pacifico AD. Complement and the damaging effects of cardiopulmonary bypass. J Thorac Cardiovasc Surg 1983;86: 845-57. 2. Finn A, Moat N, Rebuck N, Klein N, Strobel S, Elliott M. Changes in neutrophil CD11b/CDI8 and L-selectin expression and release of interleukin 8 and elastase in paediatric cardiopulmonary bypass. Agents Actions 1993;38(Special Conference Issue):C44-6. 3. Springer TA. Adhesion receptors of the immune system. Nature 1990;346:425-34. 4. Steinberg PE, McEver RP, Shuman MA, Jacques YV, Bainton DF. A platelet alpha-granule membrane protein (GMP-140) is expressed on the plasma membrane after activation. J Cell BioI 1985;101:880-6. 5. McEver RP, Beckstead JH, Moore KL, Marshall-Carlson L, Bainton DF. GMP-140, a platelet a-granule membrane protein, is also synthesized by vascular endothelial cells and is located in Weibel-Palade bodies. J Clin Invest 1989;84:92-9. 6. Hattori R, Hamilton KK, Fugate RD, McEver RP, Sims PJ. Stimulated secretion of endothelial von Willebrand factor is
18. 19.
20.
21.
22.
accompanied by rapid redistribution to the cell surface of the intracellular granule membrane protein GMP-140. J BioI Chem 1989;264:7768-71. Hattori R, Hamilton KK, McEver RP, Sims PJ. Complement proteins C5b-9 induce secretion of high molecular weight multimers of endothelial von Willebrand factor and translocation of granule membrane protein GMP-140 to the cell surface. J BioI Chem 1989;264:9053-60. Johnson GI, Bliss GA, Newman PJ, McEver RP. Structure of the human gene encoding granule membrane protein-140, a member of the selectin family of adhesion receptors for leukocytes. J BioI Chem 1990;265:21381-5. Johnson GI, Cook RG, McEver RP. Cloning of GMP-140, a granular membrane protein of platelets and endothelium: sequence similarity to proteins involved in cell adhesion and inflammation. Cell 1989;56:1033-44. Dunlop LC, Skinner MP, Bendall LJ, et aL Characterization of GMP-140 (P-selectin) as a circulating plasma protein. J Exp Med 1992;175:1147-50. Gamble JR, Skinner MP, Berndt MC, Vadas MA. Prevention of activated neutrophil adhesion to endothelium by soluble adhesion protein GMP-140. Science 1990;249:414-7. Ludlam CA, Moore S, Bolton AE, Pepper DS, Cash JD. The release of human platelet-specific protein measured by radioimmunoassay. Thromb Res 1975;6:543-8. Maurer HM. Hematologic effects of cardiac disease. Pediatr Clin North Am 1972;19:1083-93. Adatia I, Barrow SE, Strattow P, Ritter JM, Haworth SG. Abnormalities in thromboxane A 2 and prostacyclin biosynthesis in children with cyanotic congenital heart disease. Br Heart J 1993;69:179-82. Hall SM, Haworth SG. Onset and evolution of pulmonary vascular disease in young children: abnormal postnatal remodelling studied in lung biopsies. J Pathol 1992;166:183-93. Rabinovitch M, Bothwell T, Hayakawa BN, et aL Pulmonary artery endothelial abnormalities in patients with congenital heart defects and pulmonary hypertension: a correlation of light with scanning electron microscopy and transmission electron microscopy. Lab Invest 1986;55:632-53. Sole MJ, Drobac M, Schwartz L, Hussain MN, Vaughan-Neil EF. The extraction of circulating catecholamines by the lungs in normal man and in patients with pulmonary hypertension. Circulation 1979;60:160-3. Rinder CS, Bonan JL, Rinder HM, Mathew J, Hines R, Smith BR. Cardiopulmonary bypass induces leukocyte-platelet adhesion. Blood 1992;79:1201-5. Harker LA, Malpass TW, Branson HE, Hessel EA II, Slichter SJ. Mechanism of abnormal bleeding in patients undergoing cardiopulmonary bypass: acquired transient platelet dysfunction associated with selective a-granule release. Blood 1980; 56:824-34. Gu YJ, van Oeveren W, Boonstra PW, de Haan J, Wildevuur CRH. Leukocyte activation with increased expression of CR3 receptors during cardiopulmonary bypass. Ann Thorac Surg 1992;53:839-43. Komai H, Adatia IT, Elliott MJ, de Leval MR, Haworth SG. Increased endothelin-I plasma levels following cardiopulmonary bypass in pulmonary hypertensive congenital heart disease. J Thorac Cardiovasc Surg 1993 (in press). Gillis CN, Cronau LH, Greene NM, Hammond GL. Removal of 5-hydroxytryptamine and norepinephrine from the pulmonary vascular space of man: influence of cardiopulmonary bypass and pulmonary arterial pressure on these processes. Surgery 1974;76:608-16.
Soluble GMP-140 can prevent the adhesion of activated neutrophils to endothelium in vitro. Because cardiopulmonary bypass causes neutrophil-cendothelial interaction, the plasma level of soluble GMP-140 was measured using an enzyme immunoassay system in 32 children undergoing intracardiac repair of congenital heart disease. They had either a high, low, or normal pulmonary blood flow (n = 13, 12, and 7 respectively). Because activated platelets are a source of GMP-140, the plasma l3-thromboglobulin level was also measured. Blood was sampled before, during, and for 24 hours after cardiopulmonary bypass. Plasma levels of both soluble GMP-140 and l3-thromboglobulin increased after the onset of cardiopulmonary bypass in all patients but for both substances the increase was greater in those with a low
pulmonary blood flow ip < 0.05 for all comparisons). The sum total of soluble GMP-140 values during and after operation was correlated negatively with the preoperative mean pulmonary arterial pressure (p < 0.05 for all time intervals). GMP-140 level correlated with the plasma l3-thromboglobulin level (r = 0.5, P < 0.05). This work supports the contention that soluble GMP-140 is released from activated platelets during cardiopulmonary bypass, the level being particularly high in those who had intrinsically abnormal platelets preoperatively in association with a low pulmonary blood flow. Patients with a high pulmonary blood flow, who are more susceptible to endothelial cell injury, may be less well protected by soluble GMP-140. (Ann Thorae Surg 1994;58:478-82)
C
GMP-140 in the circulation is unknown [10], although we know that platelets release it when activated. Cardiopulmonary bypass is thought to be associated with activation of endothelium, leukocytes, and platelets. It can have a more injurious effect on infants and young children than on adults [1]. We measured the levels of soluble GMP-140 in 32 children with congenital heart disease who were undergoing intracardiac repair. The plasma J3-thromboglobulin level was also measured as a marker of platelet activation [12].
ard iop u lm onary bypass (CPB) causes transient multiple organ dysfunction [1]. Leukocyte-endothelial cell interaction is an important factor in the response to this injury, as has been emphasized in a recent study [2] on the effects of CPB in children. When leukocytes and endothelium are stimulated they each express specific molecules on their cell surface, known as adhesion molecules [3], that promote the adhesion of leukocyte to endothelium. Adhesion is essential for further leukocyte activity, such as release of lysosomal enzymes and infiltration into the perivascular tissue. Soluble GMP-140 can prevent adhesion of activated neutrophils to the endothelium in vitro. Soluble GMP-140 is the soluble component of GMP-140 [4], an adhesion molecule first identified in platelet a-granules.lt also is synthesized by endothelial cells [5]. GMP-140 is also known as P-selectin. Expression of GMP-140 on the cell surface of human umbilical vein endothelial cells is induced rapidly, within 3 minutes, by histamine, thrombin, phorbol myristate acetate, calcium ionophore, and activated complement [6, 7]. Both megakaryocytes and endothelial cells have a messenger RNA for GMP-140 that lacks the transmembrane domain [8, 9]. Recent studies [10] have shown that the soluble component of GMP-140 is detectable in normal human plasma. In vitro, this molecule can prevent the adhesion of activated neutrophils to the endothelium [11]. The origin of soluble Accepted for publication Dec 16, 1993. Address reprint requests to Prof Haworth, Vascular Biology and Pharmacology Unit, Institute of Child Health, 30 Guilford St, London, WC1N 1EH, England.
© 1994 by The Society of Thoracic Surgeons
Material and Methods Patients Thirty-two children undergoing elective open heart operations with CPB were studied. None suffered from extracardiac disease. They were divided into three groups according to the preoperative pulmonary blood flow (Qp), One group consisted of 13 children with a high pulmonary blood flow due to a left to right intracardiac shunt (high Qp), a second group of 12 children had a reduced pulmonary blood flow (low Qp), and a third group consisted of 7 children with aortic stenosis who had a normal flow (normal Qp). The diagnoses, age, sex, and preoperative cardiac catheterization data are shown in Table 1. Only 20 children required cardiac catheterization; treatment of the remaining patients was determined by cross-sectional echocardiography only. In the present study blood samples were taken only from lines inserted for: clinical management purposes. Permission to carry out this study was 0003-4975/94/$7.00
KOMAI AND HAWORTH SOLUBLE GMP-140 AND CARDIOPULMONARY BYPASS
Ann Thorac Surg 1994;58:478-82
Blood Sampling
Table 1. Patients' Preoperative Clinical Data Normal Qp
High Qp
LowQp
No. of patients Diagnosis
7 AS7
Sex ratio (M:F) Median age (y) Age range (y) PAP (mm Hg) Pp/Ps Qp/Qs % O 2 Sat
1:6 8.5 1.7-14.2 17.6 2: 2.5 0.3 2: 0.04 1 98.2 2: 0.7
13 ASD3 VSD4 AVSD5" TAPVR 1 1:3.3 1.9 0.4-14.1 27.9 2: 3.9 b 0.52: O.le 2.7 2: 0.6 c 95.82: 1.7
12 TF6 DORV2 TA 1 DlV3 1:0.5 3.8 1.1-16.4 13.72: 1.5 0.22:0.1 0.5 2: 0.1 82.02: 3.9 d
Characteristic
479
Three complete, 2 partial AVSDs. b Significantly greater than in the normal or low Qp groups (p < 0.05 and p < 0.01, respectively). C Significantly greater than in the normal or low Qp groups (p < 0.01 for both). d Significantly lower than in the normal or high Qp groups (p < 0.01 for both). a
ASD ~ secundum atrial septal defect; AVSD = AS = aortic stenosis; atrioventricular septal defect; DIV = double-inlet ventricle; DORV = double-outlet right ventricle; PAP = mean pulmonary arterial pressure (SEM); Pp/Ps = pulmonary to systemic systolic arterial pressure ratio; % O 2 Sat = arterial oxygen saturation (SEM); Qp = pulmonary blood flow; Qp/Qs = pulmonary to systemic arterial blood flow ratio (SEM); SEM = standard error of the mean; TA = tricuspid atresia; TAPVR = total anomalous pulmonary venous return; TF = tetralogy of Fallot; VSD = ventricular septal defect.
given by the Ethics Committee of The Hospital for Sick Children, London, England.
Surgical Management and CPB All medication was stopped at least 12 hours before the operation. Conventional CPB was used with one episode of aortic cross-clamping and multiple dose cardioplegia. After heparin (300 IU /kg) was given they underwent extracorporeal circulation at 2.4 L· m- 2 • min- 1 of perfusion flow in moderate to deep hypothermia (17° to 28°C). Aortic cross-clamping was followed by infusion of St. Thomas' cardioplegic solution. Cardiopulmonary bypass equipment consisted of a nonpulsatile roller pump (Dideco D701 [Mirandola, Italy] or Maxima membrane oxygenator [Medtronic, Anaheim, CAD and Dideco 752 or Sorin venous cardiotomy reservoir (Dideco). During CPB, the patients' hematocrit values were kept between 15% and 29%. Sodium nitroprusside was infused during the rewarming period and stopped just before the discontinuation of CPB. During and after the operation all patients were given packed red cells, crystalloid, colloid, or a combination of these for supplement of circulating blood volume. The duration of CPB ranged from 29 to 238 minutes (80.7 ::':: 13.2 minutes in those with a normal Qp, 82.7 ::':: 8.6 minutes in those with a high Qp, and 129.7 ::':: 17.1 minutes in those with a low Qp): duration of CPB for the low Qp group was significantly longer than that of the other groups (p < 0.05 for both). The aortic cross-clamping time varied from 10 to 107 minutes (41.4::':: 8.8 minutes in those with a normal Qp, 44.8 ::':: 6.6 minutes in those with a high Qp, and 62.0 ::':: 8.5 minutes in the low Qp group) with no significant differences between the groups.
Seven heparinized blood samples were taken from a radial or femoral arterial line (1) before CPB (just after heparinization and before cannulation), (2) within 5 minutes of removing the aortic cross-clamps, (3) having been off CPB for a few minutes, and (4) 20 minutes, (5) 3 hours, (6) 6 hours, and (7) 24 hours after discontinuing CPB. All blood samples were immediately centrifuged and plasma was kept frozen at -70°C until assayed.
Plasma GMP-140 Assay The soluble form of GMP-140 was measured by an enzyme immunoassay system using two different monoclonal antibodies specific for soluble GMP-140 (GMP-140 EIA kit; Takara Shuzo Co Ltd, Kyoto, Japan). Intraassay and interassay variations were 12.5% and 8.9%, respectively. In addition, the plasma {3-thromboglobulin was measured in the same samples taken from 19 patients (9 in the high Qp and 10 in the low Qp groups) as a marker of the activation and degranulation of platelet a-granules. Platelet-poor plasma was assayed with a radioimmunoassay system using a monoclonal antibody specific for {3-thromboglobulin ({3-thromboglobulin RIA kit; Kodak Clinical Diagnostics Ltd, Amersham, UK). The intraassay and interassay variations for this assay were 4.4% and 9.7%, respectively.
Statistical Analysis All results are expressed as the mean ::':: standard error of the mean. The plasma levels of both soluble GMP-140 and {3-thromboglobulin in each sample were standardized to the pre-CPB level using the hematocrit (Hct) value to mitigate the effect of hemodilution: Corrected soluble GMP-l40 or corrected (3-thromboglobulin = X
measured level
Hct before CBP/Hct when sample was taken)
Statistical evaluation of both components was performed using analysis of variance for the factorial data and Dunnett's t test for unpaired observations. To assess differences between the three groups in the level of each component during and after CPB, we calculated the sum total value of soluble GMP-140 and {3-thromboglobulin in each group and then compared them using analysis of variance. Correlations were analyzed using the Spearman rank correlation coefficient. Values were considered to be statistically significant when p was less than 0.05.
Results Soluble GMP-140 Levels Before CPB The mean level of soluble GMP-140 before CPB was significantly higher in the low Qp group than in either of the other groups (p < 0.05 for both comparisons), which were similar to each other (Fig 1). For all patients the soluble GMP-140 level before CPB was correlated negatively with the preoperative mean pulmonary arterial pressure (r = -0.5; P < 0.05) and positively with the preoperative hemoglobin level (r = 0.6, P < 0.01). There
480
KOMAI AND HAWORTH SOLUBLE GMP-140 AND CARDIOPULMONARY BYPASS
Ann Thorac Surg 1994;58:478-82
BOO
,.
3000
600
Total soluble GMP-140 (nglml)
Soluble
GMP-140 (nglml)
•• •• • • •• • • • •
4000
400
2000
1000 200
0 0
o --'---.---,---.---,---.----,-----,PRE
OAe
08
20mln
3H
6H
10
Fig 1. Mean plasma soluble GMP-140 levels (ng/mL) In patients with a low pulmonary blood flow (closed circles), a high flew (open circles), and a normal flow (open squares). Levels were obtained before bypass (PRE), immediately after the aortic cross-clamps were taken off (OAC), after coming off bypass (OB), and 20 minutes (20rnin),3 hours (3H), 6 hours (6H), and 24 hours (24H) after coming off bypass. (*p < 0.01; **p < 0.05; tp < 0.001 in comparison with the prebypass (PRE) data in each group).
were no significant correlations between the soluble GMP140 levels and either age or sex.
Soluble GMP-140 Levels During and After CPB In all groups of children the level of soluble GMP-140 increased after the onset of CPB (see Fig 1). After removal of the aortic cross-clamps the level increased progressively to reach a peak at 3 to 6 hours. In all three groups of children the level of soluble GMP-140 at each time of sampling from 20 minutes to 6 hours after CPB was significantly higher than it was before CPB. The total of soluble GMP-140 values during the operation (the sum total of values obtained at the three time intervals from removal of the aortic cross-clamps until 20 minutes after CPB) was significantly higher in the low Qp group compared with both the high and normal Qp groups (p < 0.01 for both comparisons), which were similar to each other (see Fig 1). After the operation (between 3 and 24 hours after CPB) the total soluble GMP-140 values of the low Qp group were higher than those of the high Qp group (p < 0.01), and there was no significant difference between the high and normal Qp groups in this regard. There were no significant correlations between the level of soluble GMP140 and the duration of either CPB or aortic cross-clamping time within any of the groups or in the total of 32 patients.
Relationship Between Soluble GMP-140 Level and Preoperative Pulmonary Arterial Pressure Of the 32 patients investigated, only 20 had had a preoperative cardiac catheterization study. There were negative correlations between the total soluble GMP-140 and preoperative mean pulmonary arterial pressure during the operation itself, the postoperative period, and the whole study period (r = -0.6, -0.6, and -0.7, respectively; p <
30
•
40
50
preoperative mPAP (mmHg)
24H
Time
20
Fig 2. Relation between preoperative mean pulmonary arterial pressure (mPAP) (measured in 18 patients at cardiac catheterization, irrespective of magnitude of pulmonary blood flow) and sum total of soluble GMP-140 values during the whole study period (r = -0.7; P < O.OV.
0.01 for whole study period, p < 0.05 for the other study times) (Fig 2). In addition, the total soluble GMP-140 during the operation itself, during the postoperative period, and during the whole study period was correlated negatively with the preoperative pulmonary to systemic arterial pressure ratio (r = -0.8, -0.6, and -0.7, respectively; p < 0.05 for postoperative period; p < 0.01 for the other comparisons).
Plasma {3- Thromboglobulin Levels During and After CPB Preoperatively, the mean f3-thromboglobulin level was higher in patients with a low Qp than in those with high Qp, but not significantly so (Fig 3). The levels increased after onset of CPB (p < 0.05) and returned to the pre-CPB level 20 minutes after discontinuing CPB in both groups. The sum total of values for f3-thromboglobulin during the whole study period was significantly higher in the low Qp group than in the high Qp group (p < 0.05). There were no significant correlations between the level of plasma
100
80 f3-thromboglobulin
60
(ng/ml)
40 20
o...l-..--..--..---.---.---.---.-PRE
OAe
08
20min 3H
6H
24H
Time
Fig 3. Mean ptasma (3-thromboglobulin levels (ngjml.) in patients with low pulmonary blood flow (closed circles), and high flow (open circles). Time intervals as in Fig 1. (*p < 0.01; **p < 0.05.)
KOMAI AND HAWORTH SOLUBLE GMP-140 AND CARDIOPULMONARY BYPASS
Ann Thorac Surg 1994;58:478-82
400 300 Total soluble GMP-140 (ng/ml)
200 100
• • • • • • • '. • •• •• •• •
.
0+--.---.---.---.--.-----.---, 50 100 150 200 250 300 350 400 Total l3-thromboglobulin (ng/ml) Fig 4. Relation between sum total of soluble GMP-140 (ng/mL) values during the whole study period and sum total of f3-thromboglobulin, (nglml.) during the operation (r = 0.5; P < 0.05).
l3-thromboglobulin and the duration of either CPB or the aortic cross-clamping time.
Relationship Between Soluble GMP-140 Levels and Plasma f3- Thromboglobulin Levels The sum total of plasma l3-thromboglobulin levels during the operation was compared with the total soluble GMP140 during the whole study period. There was a significant positive correlation between them, as shown in Figure 4 (r = 0.5; p < 0.05).
Comment Preoperatively, the level of soluble GMP-140 was higher in children who had a low pulmonary blood flow than in those with either a normal or high flow, and it was correlated inversely with the mean pulmonary arterial pressure. Patients with the highest hemoglobin concentrations, reflecting the severity of chronic hypoxemia, had the highest plasma levels of soluble GMP-140. The source of the elevated level of soluble GMP-140 is unknown. However, platelet function can be abnormal in patients with cyanotic congenital heart disease [12-14], and in vitro studies indicate that soluble GMP-140 is released from activated platelets [10]. Degranulated platelets release a-granules and thromboglobulin probably is located within these granules [I2]. The plasma level of l3-thromboglobulin is a marker of platelet activation and in the present study the level was higher in those patients with a low pulmonary blood flow, although the difference was not statistically significant. Thus the abnormal platelets might be the source of the elevated level of GMP-140 found in the cyanotic patients with low pulmonary blood flow. Endothelium is another possible source, but there is no evidence that human endothelial cells that can express cell surface GMP-140 can release soluble GMP-140, nor do the findings in the present study suggest such endothelial release. Moreover, the level of soluble GMP-140 was elevated in patients with a low pulmonary blood flow (which does not damage the endothelium) and normal in those with a high pulmonary blood flow [15-17] (where the
481
endothelium is exposed to an increase in shear stress and is morphologically and functionally abnormal) . After the onset of cardiopulmonary bypass the level of soluble GMP-140 increased in all patients. In this study it obviously was not possible to distinguish between the effect of cardiopulmonary bypass itself and that of surgical trauma on the level of soluble GMP-140. However, we investigated this group of patients because soluble GMP140 is thought to reduce endothelial neutrophil interaction, and this interaction is promoted by cardiopulmonary bypass. The level of soluble GMP-140 increased in all patients irrespective of the preoperative pulmonary blood flow, but it was consistently higher in those who had had a low flow preoperatively and it remained higher in these patients after bypass was discontinued. Like prebypass plasma levels, the total soluble GMP level during and after the operation was correlated inversely with preoperative pulmonary arterial pressure. By demonstrating an increase in the expression of cell surface GMP-140, Rinder and associates [IS] showed that platelet activation occurs during open heart operations. The expression of GMP-140 increased within 10 minutes of starting bypass and returned to the prebypass level 18 hours after discontinuing bypass. Although the time course is not identical with that seen in the present study, it is consistent with our hypothesis that soluble GMP-140 is released from platelets activated during bypass. This hypothesis is supported by the observation that plasma l3-thromboglobulin levels were higher in children with a low Qp than those with a high Qp during CPB. The level of l3-thromboglobulin increased during CPB in all patients, as has been shown previously [19], but was particularly high in those with a low preoperative flow. Also there was a significant, positive correlation between the amount of GMP-140 released during the whole period of this study and the plasma l3-thromboglobulin level. Thus the excessively high levels of soluble GMP-140 seen during and after CPB in patients who had had a low pulmonary blood flow might be an exaggeration of the prebypass picture also indicative of excessive release from intrinsically abnormal platelets undergoing further activation on bypass. In speculating on the possible influence of soluble GMP140 on leukocyte-endothelial cell interaction, we do not know whether the soluble GMP-140 level we measured in our patients using an enzyme immunoassay system was biologically active. However, Dunlop and associates [10] showed that soluble GMP-140 extracted from human plasma was functional by demonstrating in vitro that it bound to the GMP-140 receptors on the neutrophils to prevent adhesion. Dunlop reported a normal human plasma level that was comparable with preoperative values in the present study [10]. Gamble and associates [11] showed that soluble GMP-140 purified from animal platelets inhibited the adhesion of neutrophils to the endothelium in vitro. It appeared to do so by inhibiting both GMP-140 -mediated adhesion and CD11 b / CD18-mediated adhesion, which is thought to be important in CPBmediated endothelial cell injury [18, 20]. Although the mechanism is unknown, a high soluble GMP-140 level may work as a feedback mechanism to prevent an extended
482
Ann Thorac Surg 1994;58:478-82
KOMAI AND HAWORTH SOLUBLE GMP-140 AND CARDIOPULMONARY BYPASS
inflammatory reaction. Gamble and associates [11] showed that a linear, dose-dependent increase in soluble GMP-140 induced inhibition of neutrophil adhesion onto the endotheliaI layer at a relatively high dose of 10 JLg/mL. The in vivo saturation of unstimulated neutrophil GMP-140 receptors is estimated to be only 20% to 40% nO] in the presence of a low plasma level of soluble GMP-140 (approximately 0.2 JLg/mL), suggesting the possibility of further receptor binding. Cardiopulmonary bypass is likely to have a more injurious effect on the pulmonary endothelium of patients whose endothelium was injured preoperatively by a high pulmonary blood flow [21, 22], but if the plasma level resembles the microvascular level then these vulnerable patients are less protected from neutrophil-endothelial cell interaction than those who had low preoperative flow. In conclusion, the level of soluble GMP-140 increases during and after CPB, more so in children who had low (compared with high) preoperative pulmonary blood flow and pulmonary arterial pressure. A possible source of soluble GMP-140 may be platelets activated by CPB; the particularly high levels in those who had had a low pulmonary blood flow are attributed to the abnormal platelets found in such patients. The present study showed a significant, positive correlation between soluble GMP140 and plasma {Hhromboglobulin level, a marker of platelet activation. Soluble GMP-140 is thought to help inhibit adhesion of neutrophils to the endothelium and therefore the most vulnerable patients (those with high preoperative blood flow) are the least well protected. This work was supported by the British Heart Foundation and the Wellcome Trust.
7.
8.
9.
10. 11. 12. 13. 14.
15. 16.
17.
References 1. Kirklin JK, Westaby S, Blackstone EH, Kirklin JW, Chenoweth DE, Pacifico AD. Complement and the damaging effects of cardiopulmonary bypass. J Thorac Cardiovasc Surg 1983;86: 845-57. 2. Finn A, Moat N, Rebuck N, Klein N, Strobel S, Elliott M. Changes in neutrophil CD11b/CDI8 and L-selectin expression and release of interleukin 8 and elastase in paediatric cardiopulmonary bypass. Agents Actions 1993;38(Special Conference Issue):C44-6. 3. Springer TA. Adhesion receptors of the immune system. Nature 1990;346:425-34. 4. Steinberg PE, McEver RP, Shuman MA, Jacques YV, Bainton DF. A platelet alpha-granule membrane protein (GMP-140) is expressed on the plasma membrane after activation. J Cell BioI 1985;101:880-6. 5. McEver RP, Beckstead JH, Moore KL, Marshall-Carlson L, Bainton DF. GMP-140, a platelet a-granule membrane protein, is also synthesized by vascular endothelial cells and is located in Weibel-Palade bodies. J Clin Invest 1989;84:92-9. 6. Hattori R, Hamilton KK, Fugate RD, McEver RP, Sims PJ. Stimulated secretion of endothelial von Willebrand factor is
18. 19.
20.
21.
22.
accompanied by rapid redistribution to the cell surface of the intracellular granule membrane protein GMP-140. J BioI Chem 1989;264:7768-71. Hattori R, Hamilton KK, McEver RP, Sims PJ. Complement proteins C5b-9 induce secretion of high molecular weight multimers of endothelial von Willebrand factor and translocation of granule membrane protein GMP-140 to the cell surface. J BioI Chem 1989;264:9053-60. Johnson GI, Bliss GA, Newman PJ, McEver RP. Structure of the human gene encoding granule membrane protein-140, a member of the selectin family of adhesion receptors for leukocytes. J BioI Chem 1990;265:21381-5. Johnson GI, Cook RG, McEver RP. Cloning of GMP-140, a granular membrane protein of platelets and endothelium: sequence similarity to proteins involved in cell adhesion and inflammation. Cell 1989;56:1033-44. Dunlop LC, Skinner MP, Bendall LJ, et aL Characterization of GMP-140 (P-selectin) as a circulating plasma protein. J Exp Med 1992;175:1147-50. Gamble JR, Skinner MP, Berndt MC, Vadas MA. Prevention of activated neutrophil adhesion to endothelium by soluble adhesion protein GMP-140. Science 1990;249:414-7. Ludlam CA, Moore S, Bolton AE, Pepper DS, Cash JD. The release of human platelet-specific protein measured by radioimmunoassay. Thromb Res 1975;6:543-8. Maurer HM. Hematologic effects of cardiac disease. Pediatr Clin North Am 1972;19:1083-93. Adatia I, Barrow SE, Strattow P, Ritter JM, Haworth SG. Abnormalities in thromboxane A 2 and prostacyclin biosynthesis in children with cyanotic congenital heart disease. Br Heart J 1993;69:179-82. Hall SM, Haworth SG. Onset and evolution of pulmonary vascular disease in young children: abnormal postnatal remodelling studied in lung biopsies. J Pathol 1992;166:183-93. Rabinovitch M, Bothwell T, Hayakawa BN, et aL Pulmonary artery endothelial abnormalities in patients with congenital heart defects and pulmonary hypertension: a correlation of light with scanning electron microscopy and transmission electron microscopy. Lab Invest 1986;55:632-53. Sole MJ, Drobac M, Schwartz L, Hussain MN, Vaughan-Neil EF. The extraction of circulating catecholamines by the lungs in normal man and in patients with pulmonary hypertension. Circulation 1979;60:160-3. Rinder CS, Bonan JL, Rinder HM, Mathew J, Hines R, Smith BR. Cardiopulmonary bypass induces leukocyte-platelet adhesion. Blood 1992;79:1201-5. Harker LA, Malpass TW, Branson HE, Hessel EA II, Slichter SJ. Mechanism of abnormal bleeding in patients undergoing cardiopulmonary bypass: acquired transient platelet dysfunction associated with selective a-granule release. Blood 1980; 56:824-34. Gu YJ, van Oeveren W, Boonstra PW, de Haan J, Wildevuur CRH. Leukocyte activation with increased expression of CR3 receptors during cardiopulmonary bypass. Ann Thorac Surg 1992;53:839-43. Komai H, Adatia IT, Elliott MJ, de Leval MR, Haworth SG. Increased endothelin-I plasma levels following cardiopulmonary bypass in pulmonary hypertensive congenital heart disease. J Thorac Cardiovasc Surg 1993 (in press). Gillis CN, Cronau LH, Greene NM, Hammond GL. Removal of 5-hydroxytryptamine and norepinephrine from the pulmonary vascular space of man: influence of cardiopulmonary bypass and pulmonary arterial pressure on these processes. Surgery 1974;76:608-16.