Cardiopulmonary Bypass Circuit Treated With
Surface-Modifying Additives: A Clinical Evaluation
of Blood Compatibility
Y. John Gu, MD, PhD, Piet W. Boonstra, MD, PhD, Anthony A. Rijnsburger,
Johan Haan, BSc, and Willem van Oeveren, PhD
Department of Cardiothoracic Surgery, Thorax Center, University Hospital Groningen, Groningen, the Netherlands
Background. The cardiopulmonary bypass (CPB) cir-
cuit induces blood activation and a systemic inflamma-
tory response in cardiac surgical patients. The CPB circuit
treated with surface-modifying additive (SMA) has been
found to reduce blood activation by in vitro and ex vivo
experiments. This study evaluates the surface thrombo-
genicity and complement activation of SMA circuits
during clinical CPB.
Methods. Twenty patients undergoing coronary artery
bypass grafting were randomly divided into two groups.
In the SMA group (n 5 10), all blood-contacting surfaces
in the CPB circuit were treated or coated with SMA,
whereas in the control group (n 5 10) patients were
perfused with an identical circuit without treatment.
Results. During CPB, platelet count and b-thrombo-
globulin were found similar in both the SMA and the
control groups. Prothrombin activation indicated by frag-
ment F112 was found less in the SMA group (p < 0.05).
After CPB, platelet deposition on the CPB circuit was
significantly less (p < 0.05) in the SMA group than in the
P atients undergoing a cardiac operation with cardio-
pulmonary bypass (CPB) are still confronted with a
systemic blood activation caused by the contact of blood
with the CPB circuit [1–3]. This blood–surface interaction
induces a whole body inflammatory response and in-
creases postoperative morbidity such as bleeding com-
plication and organ dysfunction [4, 5].
In the past, a number of studies have shown that this
blood–surface interaction can be reduced partly by coat-
ing the surface with heparin [6 –9]. Recently, a surface-
modification technique called surface-modifying addi-
tive (SMA) has been introduced [10 –12]. The SMA
technology is based on a family of polysiloxane-
containing copolymers that can be either blended with
base polymer resins before processing or coated to
blood-contacting surfaces [10]. Initial investigations have
demonstrated that SMA-treated biomaterial surface re-
duces blood activation during in vitro and ex vivo exper-
Accepted for publication Dec 26, 1997.
Address reprint requests to Dr van Oeveren, Blood Interaction Research,
Dept. of Cardiothoracic Surgery, University Hospital Groningen, Hanze-
plein 1, 9700 RB Groningen, the Netherlands.
© 1998 by The Society of Thoracic Surgeons
Published by Elsevier Science Inc
control group as assessed by the labeled monoclonal
antibody against platelet glycoprotein IIIa. Complement
activation identified by C3a and terminal complex C5b-9
did not differ between the two groups, but C4a genera-
tion was less in the SMA group (p < 0.05). Leukocyte
activation identified by elastase and cytokine release
indicated by interleukin-8 were found uniformly in both
groups. Postoperatively, chest tube drainage, blood
transfusion, duration of ventilatory support, as well as
the intensive care unit and hospital stay were not signif-
icantly different between the two groups.
Conclusions. These preliminary clinical results suggest
that SMA inhibits platelet interaction with the biomate-
rial surface of the CPB circuit. Complement activation
assessed by the terminal complement complex is not
influenced by SMA. The clinical benefit of this surface-
modifying technique has yet to be assessed in a larger
population of patients undergoing cardiac operations.
(Ann Thorac Surg 1998;65:1342–7)
© 1998 by The Society of Thoracic Surgeons
iments [10 –12]. In this study, we examine whether the
SMA-treated CPB circuit improves the blood compatibil-
ity by modifying the surface thrombogenicity and com-
plement activation during clinical CPB in cardiac surgical
patients.
Material and Methods
Patients
This clinical evaluation was carried out on 20 patients
who underwent elective coronary artery bypass grafting.
These patients had no evidence of severe heart failure or
renal or hepatic dysfunction, had no history of bleeding
diathesis, and had not used platelet-inhibiting drugs
within 3 days before the operation. Written informed
consent was obtained from each patient before operation
and the study protocol was approved by the medical
ethics committee of the University Hospital in Gro-
ningen. Patients’ demographic data in both groups are
summarized in Table 1.
Patients were randomly assigned to either the SMA
group (n 5 10), in which an SMA-treated circuit (Cobe
0003-4975/98/$19.00
PII S0003-4975(98)00223-9
Ann Thorac Surg GU ET AL 1343
1998;65:1342–7 SMA AND CPB CIRCUIT

Table 1. Patients’ Demographic Data (mean 6 standard was shorter than 400 seconds. Heparin was neutralized
deviation) by means of protamine chloride infusion (3 mg/kg) after
SMA the completion of cardiopulmonary bypass.
Group Control Group p
Variable (n 5 10) (n 5 10) Value Blood Compatibility Parameters
Age (y) 62 6 7 55 6 9 0.1000
Blood samples were taken from the indwelling radial
Length (cm) 175 6 12 175 6 7 0.9641 artery catheter after heparinization but before the start of
Weight (kg) 83 6 5 83 6 6 0.7966 CPB, 5 minutes and 30 minutes after the start of CPB, 5
Sex (M/F) 8/2 8/2 1.0000 minutes after release of the aortic cross-clamp, and at the
Body surface area (m2) 2.03 6 0.11 2.00 6 0.09 0.6178 end of CPB. Platelets, leukocytes, and sample hematocrit
Duration of bypass (min) 69 6 10 67 6 16 0.6556 were measured by an electronic automatic cell counter
Aortic cross-clamp (min) 41 6 10 39 6 9 0.7184 (Cell-Dyn 610; Sequoia-Turner Corp, Mountain View,
Type of grafting CA) in samples anticoagulated with citrate. For biochem-
LIMA 1 ... ical assays, plasma was obtained by centrifuging whole
LIMA/RIMA 3 3 blood at 1,000 g for 10 minutes and was stored at 280°C
LIMA/GEA 2 3 until further determinations. Platelet degranulation was
LIMA/RIMA/GEA 4 4 indicated by b-thromboglobulin, which was determined
by radioimmunoassay (Kodak Clinical Diagnostics Ltd,
GEA 5 gastroepiploic artery; LIMA 5 left internal mammary artery; Amersham, UK). Prothrombin activation was indicated
RIMA 5 right internal mammary artery; SMA 5 surface modifying
additive. by fragment F112, which was determined by enzyme
linked immunoassay (Behringwerke AG, Marburg, Ger-
many). Complement release products C3a and C4a were
determined by radioimmunoassay (Amersham Interna-
Cardiovascular Inc, Arvada, CO) was used, or the control
tional Inc, Amersham, UK), whereas the terminal com-
group, (n 5 10) in which an untreated but otherwise
plement complex C5b-9 was determined by enzyme
similar extracorporeal circuit was used. Anesthesia was
linked immunosorbent assay (Quidel, San Diego, CA).
induced and maintained by intravenous infusion of
Leukocyte elastase was quantitated in complex with
sufentanil citrate (1 to 3 mg/kg) and midazolam (0.05 to
0.1 mg/kg). Muscle relaxation was achieved with pancu- a1-proteinase inhibitor (Merck, Darmstadt, Germany) by
ronium bromide (100 to 140 mg/kg). Cefamandol (2 g) and enzyme-linked immunosorbent assay. Interleukin-8 as
dexamethasone (1 mg/kg) were administered after in- cytokine marker was also determined by enzyme-linked
duction of anesthesia. Anticoagulation was achieved by immunosorbent assay (Innogenetics, Zwijnwaarde, Bel-
intravenous administration of 300 IU/kg bovine lung gium).
heparin before CPB and 1,500 IU in the pump prime.
Platelet Binding on Cardiopulmonary Bypass Circuit
Cardiopulmonary Bypass After the termination of CPB, both parts of the arterial
The extracorporeal circuit consisted of roller pumps and venous tubing were gently flushed with saline solu-
(Stöckert Instrumentation, Munich, Germany) and a mi- tion, and collected for determination of platelet binding.
croporous polypropylene membrane oxygenator (CML Platelet binding was quantified by Eu-labeled antibody
Duo, Cobe Laboratories Inc, Lakewood, CO). In the SMA directed to the platelet GPIIIa receptor (M753, Dakopatts,
group, all the blood-contacting components in the extra- Glostrup, Denmark) bound to the tubing surface. A
corporeal circuit, including oxygenator, venous reservoir, control antibody directed against mouse antigen with
tubing, connectors, and cannula were treated with SMA. similar amount of label was used as a reference. Also,
Neither cardiotomy reservoir nor an arterial line filter aspecific antibody binding was tested on both nonused
was used in the circuit, and all the cardiotomy suction SMA and control tubing. After exposure to the antibody
blood was discarded according to the method described the tubing was washed with saline solution and filled
previously [13]. The extracorporeal circuit was primed with enhancement solution to release the Eu-label for
with a pentastarch solution (HEAS 2.5%; Fresenius, Bad counting in a fluorometer (Delfia, LKB Wallac, Turku,
Homburg, Germany). Standard hemodilution technique Finland).
was used including a dilution of the circulating blood
volume to a hematocrit of approximately 20% to 25%. Statistics
During bypass the pump flow was set at 2.4 L z m22 z Clinical data are expressed as mean and standard devi-
min21 and patients were cooled to a nasopharyngeal ation and the data of blood tests are expressed as mean
temperature of 28°C. Before the last anastomosis was and standard error of the mean. All values of blood tests
finished, patients were rewarmed to 37°C. The mean during CPB were corrected for hemodilution by hemat-
arterial pressure was maintained at 50 to 60 mm Hg ocrit. Statistical tests were performed with the StatView
during bypass. Anticoagulation during bypass was mon- software (Brain-power Inc., Calabasas, CA). Student’s t
itored with the activated clotting time (Hemochron 800; test or Mann-Whitney test was used for analysis of
International Technidyne Corp, Edison, NJ). Additional difference between the two groups. A p value less than
heparin was administered if the activated clotting time 0.05 was considered statistically significant.
1344 GU ET AL Ann Thorac Surg
SMA AND CPB CIRCUIT 1998;65:1342–7

Table 2. Postoperative Observations (mean 6 standard spasm of the arterial graft, respectively. There was no
deviation) statistical difference between the two groups regarding
Control
the postoperative chest tube drainage, red blood cell
SMA Group Group p transfusion, duration of postoperative ventilatory sup-
Variable (n 5 8)a (n 5 10) Value port, and hospital stay (Table 2).
Chest drainage (mL/24 h) 926 6 444 823 6 245 0.5408
Effect of Surface-Modifying Additive on Platelets,
RBC transfusion (mL/24 h) 350 6 418 311 6 369 0.8413
Prothrombin, and Surface Thrombogenicity
Urine output (mL/24 h) 3,262 6 1,097 3,472 6 1,058 0.7848
Ventilation (h) 11.1 6 2.9 13.4 6 4.3 0.2233 Platelet count dropped at 5 minutes on CPB in both the
ICU stay (days) 1.0 6 0.0 1.0 6 0.0 1.0000 SMA and the control groups (Table 3). The count was
Hospital stay (days) 7.4 6 1.9 7.6 6 2.5 0.8369 lowest at 30 minutes under hypothermia and recovered
toward the end of CPB without any difference between
a
Data do not include 2 patients with postoperative complications (see text the two groups (Fig 1A). Platelet degranulation indicated
for details).
by b-thromboglobulin release increased in both groups
ICU 5 intensive care unit; RBC 5 red blood cell. during CPB and peaked at 5 minutes after release of the
aortic cross-clamp (see Table 3; Fig 1B). Prothrombin
Results fragment F112 was significantly different at the baseline
and increased during CPB in both groups and peaked at
Patients the end of CPB (see Table 3). The F112 generation was
All operations were uneventful with an average bypass significantly lower in the SMA group than in the control
time of 69 6 10 minutes in the SMA group and 67 6 16 group during the late period of CPB (Fig 1C). After CPB,
minutes in the control group (Table 1). In the SMA group, platelet deposition identified by glycoprotein IIIa binding
2 patients had to be reoperated on within a few hours was significantly less on the SMA circuit than on the
after the first operation because of surgical bleeding and control circuit (p , 0.05; Fig 2), whereas the aspecific

Table 3. Blood Compatibility Parameters in Both the Surface-Modifying Additive and Control Groups
Before
Parameter CPB 5 Min of CPB 30 Min of CPB 5 Min of Declamp End of CPB
9
Leukocyte count (310 /L)
SMA 6.2 6 0.6 5.5 6 0.6 4.0 6 0.4 8.2 6 1.9 12.7 6 2.2
Control 5.6 6 0.7 5.5 6 1.1 4.8 6 1.2 7.7 6 1.8 11.0 6 1.6
Platelet count (3109/L)
SMA 165 6 14 157 6 17 128 6 10 177 6 19 191 6 23
Control 192 6 33 163 6 30 155 6 33 180 6 30 198 6 38
b-Thromboglobulin (ng/mL)
SMA 336 6 30 741 6 103 1,325 6 187 1,660 6 339 1,338 6 251
Control 446 6 56 611 6 58 2,292 6 481 2,773 6 729 2,300 6 725
Prothrombin F112 (nmol/L)
SMA 1.69 6 0.2a 2.07 6 0.3 2.17 6 0.2 2.29 6 0.2 2.38 6 0.3
Control 1.20 6 0.1 1.58 6 0.2 2.20 6 0.2 4.60 6 1.7 5.83 6 2.2
Complement C3a (ng/mL)
SMA 505 6 76 1,440 6 191 2,073 6 109 2,236 6 122 2,145 6 117
Control 690 6 160 1,371 6 232 2,179 6 151 2,184 6 88 2,153 6 113
Complement C4a (ng/mL)
SMA 2,720 6 545 2,391 6 545 3,637 6 1247 3,477 6 728 4,047 6 998
Control 1,714 6 366 2,236 6 469 2,255 6 824 3,993 6 1001 4,536 6 1892
Complement C5b-9 (ng/mL)
SMA 250 6 47 457 6 78 1,032 6 168 1,510 6 111 1,445 6 136
Control 184 6 12 412 6 108 1,205 6 182 1,605 6 82 1,554 6 74
Elastase (ng/mL)
SMA 110 6 37 94 6 34 142 6 53 321 6 93 353 6 88
Control 74 6 19 55 6 12 121 6 34 255 6 56 339 6 79
Interleukin-8 (pg/mL)
SMA 0.2 6 0.1 0.4 6 0.3 1.7 6 0.9 3.9 6 2.3 3.7 6 2.6
Control 0.3 6 0.2 1.0 6 0.7 1.4 6 0.8 2.8 6 1.6 5.0 6 2.6
a
p , 0.05 comparison between the two groups.
CPB 5 cardiopulmonary bypass; SMA 5 surface modifying additives. Values are expressed as mean 6 standard error in blood or plasma
concentrations.
Ann Thorac Surg GU ET AL 1345
1998;65:1342–7 SMA AND CPB CIRCUIT

plement activation identified by C3a did not differ be-

tween the two groups toward the end of CPB (see Table
3; Fig 3A). However, C4a generation was less pronounced
in the SMA group after release of the aortic cross-clamp
(see Table 3; Fig 3B). The terminal complement complex
C5b-9 was highest after release of the aortic cross-clamp
without difference between the two groups (see Table 3;
Fig 3C).

Comment
The SMA treatment of biomaterial surface has been
shown to improve blood compatibility by in vitro and ex
vivo experiments [10 –12]. In this pilot study on 20 cardiac
surgical patients we aimed to examine both the surface
thrombogenicity and complement activation during clin-
ical cardiopulmonary bypass. Results revealed that plate-
let deposition on the CPB circuit was significantly less in
the SMA group than in the control group as assessed by
the labeled monoclonal antibody against platelet glyco-
protein IIIa. During CPB, the pattern of platelet activation
tends to be modified by SMA treatment as indicated by a
less pronounced release of b-thromboglobulin. More-
over, activation of prothrombin was also reduced as
indicated by reduced generation of fragment F112. Com-
plement activation identified by C3a and terminal com-
plex C5b-9 did not differ between the two groups,
whereas C4a generation was found less in the SMA
group, suggesting that SMA did not influence the alter-
native pathway, but only the classic pathway of comple-
ment activation in clinical CPB patients.
Determination of platelet membrane glycoprotein IIIa
antigen bound to biomaterial surface has been reported
to reflect platelet deposition to the extracorporeal circuit
[14, 15]. When platelets become activated, glycoprotein
IIIa together with glycoprotein IIb serves as receptor for
platelet interaction with other platelets or with thrombo-

Fig 1. Platelets (A), b-thromboglobulin (B), and prothrombin frag-

ment F112 (C) expressed as percentage of initial (%) in both the
surface-modifying additive (SMA) and control groups. (CPB 5 car-
diopulmonary bypass; X-off 5 release of the aortic cross-clamp; *p
, 0.05 for comparison between the two groups.)

antibody binding was similar on both the SMA and

control tubing before blood contact.

Effect of Surface-Modifying Additive on Leukocyte and

Complement Activation
Leukocyte count increased in both the SMA and control
groups at the end of CPB without significant difference
between the two groups (see Table 3). The leukocyte
release product elastase increased during CPB and Fig 2. Platelet glycoprotein IIIa (GPIIIa) binding on either the arte-
peaked at the end of CPB in both groups (see Table 3). rial or venous tubing of the extracorporeal circuit. Significantly less
Also, there was no difference of elastase between the two GPIIIa was bound on surface-modifying additive (SMA)-treated
groups. Interleukin 8 increased during CPB without any tubings after the circuit had been used for cardiopulmonary bypass
difference between the two groups (see Table 3). Com- (*p , 0.05 for comparison between groups.)
1346 GU ET AL
SMA AND CPB CIRCUIT
Fig 3. Complement C3a (A), C4a (B), and the terminal complement
complex C5b-9 (C) expressed as percentage of initial (%) in both the
surface-modifying additive (SMA) and control groups. (CPB 5 car-
diopulmonary bypass; X-off 5 release of the aortic cross-clamp; *p
, 0.05 for comparison between the two groups.)
genic surfaces through adhesive proteins [16]. On the
surface of the extracorporeal circuit, glycoprotein IIIa was
found together with other plasma proteins after the
circuit was exposed to human blood in a simulated
extracorporeal circuit [14]. In the current study after
exposure to clinical CPB, we observed that the surface
binding of glycoprotein IIIa antigen was significantly
reduced in the CPB circuit that was treated with SMA.
This was particularly obvious on the tubing located at the
arterial side, suggesting that the antithrombogenic effect
of SMA is most effective under high flow circumstances
[12].
The mechanism of antithrombogenicity of the SMA
Ann Thorac Surg
1998;65:1342–7
circuit seems to be mainly based on the effect of limiting
platelet interaction with the surface and subsequently
reducing platelet activation. This is in contrast to the
heparin-coated material, which still allows platelet bind-
ing. Furthermore, in comparison with heparin coating,
the SMA-treated surface is supposed to have a longer
half-life than the heparin-coated surface, because the
former is not sensitive to biological degradation. This
assumption is supported by the recent results obtained
on vascular access grafts [17] and on long-term observa-
tions on ventricular assist device using this technology
[18].
Complement activation is known to occur in cardiac
surgical patients and to be caused by blood contact with
the foreign surfaces in the CPB circuit [1, 3]. Previously,
we and other investigators [6, 7, 9] have demonstrated
that patients perfused with a heparin-coated CPB circuit
had a reduced complement activation during coronary
artery bypass grafting. In the current study, complement
activation represented by C3a and C5b-9 was found
similar in both the SMA and control groups during the
entire period of CPB. However, C4a generation was
significantly less in the SMA group, particularly during
the late period of CPB, suggesting that SMA may have
modified complement activation through the classic
pathway in these patients. The mechanism by which the
SMA treatment inhibits the classic complement pathway
remains uncertain, particularly by the fact that it ap-
peared during the late part of CPB. In general, the
alternative pathway is considered to be the main com-
plement pathway activated by the CPB circuit [19, 20],
although a decrease in C4 and an increase in C4a during
CPB indicating classic pathway complement activation
was also found during CPB with untreated circuits [21,
22]. Protamine is regarded as the main trigger activating
the classic complement pathway [23]. However, in this
study the difference of C4a appeared before protamine
administration. Nevertheless, the clinical impact of this
difference in C4a seems minor because a more harmful
complement activation product, the terminal comple-
ment complex C5b-9 [24], was not reduced by SMA.
Although data from this study indicate that the SMA
treatment modified the surface thrombogenicity of the
CPB circuit, the patient’s hemostatic status remained
unchanged. Probably the beneficial effect of a more
biocompatible circuit does not appear to be significantly
related to the improvement of clinical hemostasis when
evaluated in a relatively low number of patients having a
relatively short duration of CPB. Furthermore, all pa-
tients enrolled in this study received dexamethasone
preoperatively. Dexamethasone may have affected the
body inflammatory response uniformly in patients per-
fused either with the SMA circuit or with the control
circuit, so that their clinical outcome after operation was
without difference.
We conclude that SMA treatment tends to inhibit
platelet interaction with the biomaterial surface of CPB
circuit. Complement activation assessed by the terminal
complement complex is not influenced by SMA. Further
study on a larger patient population is necessary to
Ann Thorac Surg
1998;65:1342–7
evaluate whether this surface-modifying technique con-
tributes to the improvement of clinical outcome in car-
diac surgical patients.
References
1. Kirklin JK. Prospects for understanding and eliminating the
deleterious effects of cardiopulmonary bypass [Editorial].
Ann Thorac Surg 1991;51:529–31.
2. Edmunds LH Jr. Blood–surface interaction during cardiopul-
monary bypass. J Cardiol Surg 1993;8:404–10.
3. Van Oeveren W. Leukocyte and platelet activation during
extracorporeal circulation. Cells Materials 1994;4:187–95.
4. Butler J, Rocker GM, Westaby S. Inflammatory response to
cardiopulmonary bypass. Ann Thorac Surg 1993;55:552–9.
5. Taggart DP, El-Fiky MM, Carter R, Bowman A, Wheatley DJ.
Respiratory dysfunction after uncomplicated cardiopulmo-
nary bypass. Ann Thorac Surg 1993;56:1123– 8.
6. Videm V, Svennevig JL, Fosse E, Semb G, Østerud A,
Mollnes TE. Reduced complement activation with heparin-
coated oxygenator and tubings in coronary bypass opera-
tions. J Thorac Cardiovasc Surg 1992;103:806–13.
7. Gu YJ, van Oeveren W, Akkerman C, Huyzen RJ, Boonstra
PW, Wildevuur CRH. Heparin-coated circuits reduce the
inflammatory response to cardiopulmonary bypass. Ann
Thorac Surg 1993;55:917–22.
8. Boonstra PW, Gu YJ, Akkerman C, Haan J, Huyzen R, van
Oeveren W. Heparin coating of an extracorporeal circuit
partly improves hemostasis after cardiopulmonary bypass.
J Thorac Cardiovasc Surg 1994;107:289–92.
9. Øvrum E, Mollnes TE, Fosse E, et al. Complement and
granulocyte activation in two different types of heparinized
extracorporeal circuits. J Thorac Cardiovasc Surg 1995;110:
1623–32.
10. Tsai CC, Deppisch RM, Forrestal LJ, et al. Surface modifying
additives for improved device-blood compatibility. ASAIO J
1994;40:M619 –24.
11. Eberhart R, Jessen M, Chao R, et al. Platelet response to
membrane oxygenator treatment with heparin, SMA and
nitric oxide. Artif Organs 1995;19:1047.
12. Spijker HT, Haan J, Graaff R, Boonstra PW, van Oeveren W.
Surface modifying additives reduce thrombogenicity: an in
vitro and clinical evaluation. Cells Materials (in press).
13. De Haan J, Boonstra PW, Monnink S, Ebels T, van Oeveren
GU ET AL 1347
SMA AND CPB CIRCUIT
W. Retransfusion of thoracic wound blood during cardiopul-
monary bypass impairs hemostasis. Ann Thorac Surg 1995;
59:901–7.
14. Gluszko P, Rucinski B, Musial J, et al. Fibrinogen receptors in
platelet adhesion to surfaces of extracorporeal circuit. Am J
Physiol 1987;252:H615–21.
15. Wenger RK, Lukasiewicz H, Mikuta BS, Niewiarowski S,
Edmunds LH Jr. Loss of platelet fibrinogen receptors during
clinical cardiopulmonary bypass. J Thorac Cardiovasc Surg
1989;97:235–9.
16. Nachman RL, Leung LLK. Complex formation of platelet
membrane glycoproteins IIb and IIIa with fibrinogen. J Clin
Invest 1982;69:263–9.
17. Allen RDM, Yuili E, Nankivell BJ, Fransic DMA. Australian
multicentre evaluation of a new polyurethane vascular ac-
cess graft. Aust NZ J Surgery 1996;66:742– 6.
18. Farrar DJ, Hill JD, Pennington DG, et al. Preoperative and
postoperative comparison of patient with univentricular and
biventricular support with the Thoratec ventricular assist
device as a bridge to cardiac transplantation. J Thorac
Cardiovasc Surg 1997;113:202–9.
19. Van Oeveren W, Kazatchkine MD, Descamps-Latscha B, et
al. Deleterious effects of cardiopulmonary bypass. A pro-
spective study of bubble versus membrane oxygenation.
J Thorac Cardiovasc Surg 1985;89:888–99.
20. Mollnes TE, Videm V, Gotze O, Harboe M, Oppermann M.
Formation of C5a during cardiopulmonary bypass: inhibi-
tion by precoating with heparin. Ann Thorac Surg 1991;52:
92–7.
21. Collett B, Alhaq A, Abdullah NB, et al. Pathways to comple-
ment activation during cardiopulmonary bypass. Br Med J
1984;289:1251– 4.
22. Tamiya T, Maeo Y, Okada T, Ogoshi S, Fujimoto S, Yasui H.
Significance of the concentrated red cell and albumin prim-
ing method with particular reference to anaphylatoxin gen-
eration. J Thorac Cardiovasc Surg 1992;103:78– 86.
23. Kirklin JK, Chenoweth DE, Naftel DC, et al. Effect of prota-
mine administration after cardiopulmonary bypass on com-
plement, blood elements and the hemodynamic state. Ann
Thorac Surg 1996;41:193–9.
24. Salama A, Hugo F, Heinrich D, et al. Deposition of terminal
C5b-9 complement complexes on erythrocytes and leuko-
cytes during cardiopulmonary bypass. N Engl J Med 1988;
318:408–14.