Risk and Benefit of Low Systemic HeparinizationDuring Open Heart OperationsLudwig K. von Segesser, MD, Branko M. Weiss, MD, Miralem Pasic, MD,Eligio Garcia, BA, and Marko 1. Turina, MDClinic for Cardiovascular Surgery and Institute of Anesthesiology, University Hospital, Zurich, Switzerland

Heparin surface-coated perfusion equipment with im­proved thromboresistance was evaluated in 104 consecu­tive patients undergoing open heart operation in a pro­spective, randomized trial with low versus full systemicheparinization. Surgical procedures included coronaryartery revascularization in 47 of 54 (87%) for low versus44 of 50 patients (88%; not significant [NS» for full, valverepair/replacement in 8 of 54 (15%) for low versus 5 of 50patients (10%; NS) for full, left ventricular aneurysmrepair in 1 of 54 (2%) for low versus 2 of 50 patients (4%;NS) for full, and other in 3 of 54 (6%) for low versus 3 of50 patients (6%; NS) for full. Cross-clamp time was 39.2 ±10.7 minutes for low versus 39.5 ± 10.5 minutes for full(NS). Cardiopulmonary bypass time was 68.6 ± 20.1minutes for low versus 69.3 ± 16.6 minutes for full (NS).Lowest activated coagulation time during perfusion was255 ± 75 seconds for low versus 537 ± 205 seconds for full

Fu ll systemic heparinization to prevent activation of theplasmatic coagulation system has been the standard

practice during operation using cardiopulmonary bypassover the past decades. This approach is based on variousolder studies including the work of Young and colleagues[1], who had shown for a primate model that an activatedcoagulation time (ACT; Hemochron, International Techni­dyne, Edison, NJ) greater than 400 seconds suppressedproduction of fibrin monomer. Hence, nowadays mostmanufacturers of oxygenating devices recommend main-

For editorial comment, see page 285.

taining the ACT during perfusion greater than 480 sec­onds.

More recently, however, a number of authors questionedthese heparinization levels. Gravlee and colleagues [2]found that in the clinical setting compensated subclinicalplasma coagulation occurred during cardiopulmonary by­pass despite ACTs greater than 400 seconds on one handand, on the other, that ACTs greater than 350 secondsresulted in acceptable fibrinopeptide A levels. Higher ACTlevels, however, were associated with increased bleeding.

Presented at the Thirtieth Annual Meeting of The Society of ThoracicSurgeons, New Orleans, LA, Jan 31-Feb 2, 1994.

Address reprint requests to Dr von Segesser, Clinic for CardiovascularSurgery, University Hospital, Riimistrasse 100, CH-8091 Zurich, Switzer­land.

© 1994 by The Society of Thoracic Surgeons

(p < 0.0005). In the low group, the target activated coagula­tion time of more than 180 seconds was not reached duringperfusion in 4 of 54 patients (7%), the lowest value being164 seconds. No oxygenator failure occurred. Hospital mor­tality was 0 of 54 (0%) for low versus 1 of 50 patients (2%)for full (NS). Bleeding required surgical revision in 0 of 54(0%) for low versus 4 of 50 patients (8%) for full (p = 0.05).Drainage (24 hours) was 790 ± 393 mL for low versus 1,039± 732 mL for full (p < 0.025). Amount of packed homolo­gous red cells transfused (24 hours) was 300 ± 354 mL forlow versus 957 ± 596 mL for full (p < 0.0005). Baselinehematocrit of 43.3% ± 3.7% for low versus 43.0% ± 3.9%(NS) for full before operation moved to 28.9% ± 3.2% forlow versus 28.8% ± 3.2% for full (NS) at 24 hours. Lowsystemic heparinization during open heart operation re­sults in reduced blood loss and transfusion requirements.

(Ann Thorae Surg 1994;58:391-8)

Cardoso and colleagues [3] showed in membrane oxygen­ators and experimental set-ups that reduced heparin ad­ministration, enough to keep the ACT between 250 and 300seconds, was not related either to major coagulation factorsand platelet consumption or to derangements in the oxy­genator performance. Metz and Keats [4] found in a clinicalstudy that for ACT values around 300 seconds there wereno macroscopic clots in the perfusion circuit.

An additional dimension to this debate was added bythe increasing availability of heparin-coated componentsfor cardiopulmonary bypass with improved thromboresis­tance. Since the initial report on heparin bonding toartificial surfaces by Vincent Gott and colleagues [5] in1963, a number of heparin surface-coating techniques havebeen developed [6-8]. The fact that heparin surface-coatedcomplex devices like heat-exchanger/oxygenator struc­tures could be used in the experimental set-up with low orno systemic heparinization [9, 10] triggered progressiveintroduction of heparin-coated perfusion equipment intoclinical practice using varying heparinization regimens[11-13]. Various advantages have been reported for use ofheparin-coated equipment during cardiopulmonary by­pass with full systemic heparinization including reductionof complement activation as well as reduction of thrombinformation [14-18]. Under exceptional circumstances, suchas a major contraindication for systemic heparinization,clinical cardiopulmonary bypass also has been realizedwithout systemic heparinization [19,20].

0003-4975/94/$7.00

392 VON SEGESSER ET ALLOW SYSTEMIC HEPARINIZAnON

Ann Thorac Surg1994;58:391- 8

The present study was designed for evaluation of riskand benefit of low systemic heparinization during cardio­pulmonary bypass with heparin-coated perfusion equip­ment in a prospective, randomized trial. Perioperativeblood loss and transfusion requirements were the mainissues addressed.

Patients and Methods

Inclusion Criteria and RandomizationAfter approval of the protocol by the institutional reviewboard, 104 patients undergoing elective open heart opera­tion were selected for a prospective trial with low versusfull systemic heparinization in accordance to the followinginclusion criteria: body weight, more than 50 kg; expectedhematocrit during cardiopulmonary bypass, more than20% for patients with ejection fractions more than 0.45 ormore than 20% for those ejection fractions less than 0.45,allowing for crystalloid priming. Exclusion criteria werethe following: known coagulopathies; ongoing anticoagu­lation treatment; requirement of blood priming; emergencyoperation; and the expectation of volume overload oranemia requiring hemofiltration. Informed consent wasobtained from the patients to be included in the study, whowere assigned randomly to two groups.

FULL SYSTEMIC HEPARINIZATION GROUP (ACT MORE THAN 480

SECONDS). Perfusion was with heparin surface-coated car­diopulmonary bypass equipment and full systemic hepa­rinization. Heparin (Liquemin: Roche, Basel, Switzerland)loading dose was 300 IU /kg body weight; priming dosewas 5,000 IU /L; ACT was more than 480 seconds. Prota­mine (Protamin: Roche, Basel, Switzerland) dose wasequivalent to heparin loading dose; additional doses weregiven according to ACT.

LOW SYSTEMIC HEPARINIZATIN GROUP (ACT MORE THAN 180

SECONDS). Perfusion was with heparin surface-coated car­diopulmonary bypass equipment and low systemic hepa-

Fig 1. Arteriovenous shunt in the surgeon's field (3IB-inch arterialline and 112-inch venous line).

Fig 2. Venting of the arterial cannula through the shunt into the ve­nous line (during recirculation blood and bubbles are drained from thearterial cannula through the shunt and the venous line into the filterportion of the venous reservoir).

rinization. Heparin loading dose was 100 IU /kg bodyweight; priming dose was 1,000 IU /L; ACT was more than180 seconds. Protamine dose was equivalent to heparinloading dose; additional doses were given according toACT.

Randomization attributed 54 of 104 patients to lowsystemic heparinization (heparin loading dose of 100IV /kg body weight) and 50 of 104 patients to full hepa­rinization (heparin loading dose, 300 Ilf /kg). Surgicalprocedures included (either isolated or combined) coro­nary artery revascularization, valve repair/replacement,left ventricular aneurysm repair, and other open heartprocedures.

Anesthesia and Cardiopulmonary BypassStandard anesthetic (flunitrazepam, fentanyl, pancuro­nium) and monitoring techniques (electrocardiogram, cen­tral venous and arterial catheters, urinary catheter, temper­ature probes) were used in both groups of patients.Cardiopulmonary bypass was realized with identical hep­arin surface-coated perfusion equipment in both groups.Over the time frame of the study, the configuration of theperfusion circuit followed the progress in manufacturingtechniques and the resulting availability of more sophisti­cated heparin-coated perfusion components. The final cir­cuit, which was used in similar quantity in both analyzedgroups, included a heparin-coated two-stage venous can­nula, a heparin-coated venous line, a heparin-coated flex­ible venous reservoir, a heparin-coated silicone pump loop,a heparin-coated low-prime heat-exchanger/oxygenatorstructure (Univox-Cold: Baxter-Bentley, Irvine, CA), a he­parin-coated arterial filter, a heparin-coated arterial line, aheparin-coated arteriovenous shunt in the surgeon's field(Fig 1), a heparin-coated arterial cannula, and a heparin­coated cardiotomy reservoir (BCR 3500 Gold; Baxter­Bentley).

All-crystalloid priming with Ringer's lactate solutionwas used in both groups; the solution included heparin

Ann Thorac Surg1994;58:391-8

VON SEGESSER ET ALLOW SYSTEMIC HEPARINIZATION

393

Results

Fig 3. Lowest activated coagulation time (ACT) measured during car­diopulmonary bypass (mean:±: standard deviation). There is a signifi­cant difference between groups.

Data AnalysesSystat statistical package (Systat Inc, Evanston, IL ) wasused for analysis of data. Quantitative data are presentedin the form of mean ± standard deviation and are com­pared using Student's t test for paired or unpaired vari­ables where applicable. Fisher's exact test was used forcomparison of nonparametric data. The criterion of statis­tical significance taken was a probability value less than0.05.

ACT>480ACT >180o

200

400

600

800

s

Similar baseline characteristics were observed in bothgroups as shown in Table 1. There were no differences withregard to age, sex, body weight, and body surface area.Surgical procedures included (either isolated or combined)coronary artery revascularization in 47 of 54 (87%) for lowsystemic heparinization versus 44 of 50 patients (88%; NS)for full, valve repair/replacement in 8 of 5405%) for lowversus 5 of 50 patients 00%; NS) for full, left ventricularaneurysm repair in 1 of 54 (2%) for low versus 2 of 50patients (4%; NS) for full, and other in 3 of 54 (6%) for lowversus 3 of 50 patients (6%; NS) for full. Number ofcoronary artery bypass grafts was 3.0 ± 1.1 for low versus3.3 ± 1.0 for full (NS) and number of arterial grafts was1.1 ± 0.4 for low versus 0.9 ± 0.5 for full (NS). Aorticcross-clamp time was 39.2 ± 10.7 minutes for low versus39.5 ± 10.5 minutes for full (NS). Cardiopulmonary bypasstime was 68.6 ± 20.1 minutes for low versus 69.3 ± 16.6minutes for full (NS).

The mean level of the lowest ACT measured duringperfusion in each patient is shown in Figure 3 and was255 ± 75 seconds for low systemic heparinization versus537 ± 205 seconds for full (p < 0.0005). A scattergram ofthe lowest measured ACT is given in Figure 4. In thegroup perfused with low systemic heparinization, thetarget ACT of more than 180 seconds was not reached

1000

Table 1. Baseline Characteristics

Heparinization

Variable Low Full p Value

No. of patients 54/104 50/104 NSAge (y) 59:±: 12 60:±: 11 NSNo. of male patients 39/54 36/50 NSNo. of female patients 15/54 14/50 NSBody weight (kg) 71.1 :±: 15.9 72.1 :±: 11.4 NSBody surface area (rrr') 1.8 :±: 0.2 1.8 :±: 0.2 NSCoronary bypass grafting 47/54 (87%) 44/50 (88%) NSValve repair / replacement 8/5405%) 5/5000%) NSLeft ventricular aneurysm 1/54 (2%) 2/50 (4%) NS

repairOther open heart 3/54 (6%) 3/50 (6%) NS

proceduresAortic cross-clamp time 39.2:±: 10.7 39.5:±: 10.5 NS

(min)Pump time (min) 68.6 :±: 20.1 69.3 :±: 16.6 NS

NS = not significant.

5,000 IV /L in the group receiving full systemic hepariniza­tion and 1,000 IV/L in the group receiving low systemicheparinization. Cardiopulmonary bypass was started afterventing both the arterial (Fig 2) and the venous cannulasinto the venous line. Moderate hypothermia, pericardialcooling, and cold, potassium-rich cardioplegia were usedfor myocardial protection in similar fashion for bothgroups. Routinely, the pressure gradient over the heat­exchanger/ oxygenator structure was measured to havesome information about the blood path. To avoid stagna­tion of blood in the perfusion circuit, recirculation throughthe shunt in the surgeon's field (see Fig 1) was startedimmediately after weaning. Oxygenator sump blood wasretransfused directly to the patient in the group with fullsystemic heparinization, whereas it was retransfused con­centrated and washed in the group with low systemicheparinization to avoid potential transfusion of clots. Allperfusion systems were rinsed with clear fluid, carefullyanalyzed, and finally weighed (drained weight).

Volume SubstitutionBefore cardiopulmonary bypass, Ringer's lactate solutionand hydroxyethyl starch were used if required. Aftercardiopulmonary bypass, the hematocrit was maintainedgreater than 25%. Volume was substituted with Ringer'slactate solution up to 1,000 mL or hydroxyethyl starch upto 1,500 mL if required. If blood loss exceeded 400 mL inthe first hour or 200 mL in the following hour, ACT andcoagulation studies were performed before fresh frozenplasma (2 + 2 units) was given. If this was insufficient, 6units of platelets were given. Blood loss, transfusion re­quirements, and fluid balance were recorded. Myocardialinfarction in the perioperative period was diagnosed ac­cordingly to two or more of the following criteria: new Qwave in the electrocardiogram, creatine kinase values morethan three times the normal, and echocardiographic orscintigraphic evidence.

394 VON SEGESSER ET ALLOW SYSTEMIC HEPARINIZAnON

Ann Thorae Surg1994;58:391-8

Fig 4. Lowest activated coagulation time (ACT) measuredduring car- 500diopulmonary bypass (scatter plot). Arrows indicate ACT levels belowthe target of 180 seconds.

9,-------------------,

1500

1000

2000

o

I

patient nr50

• 0

o ACT>180

•

t

•

• • ACT>480

40302010

... . .....·0 •••• • •

a. •• ..-.. • • 0• • • 0 0 00

00 0 0_ 1\ 00 0 .°cP 0 o· 000 ~"o oo,..m-,., RJoo 00 0 0 cP 0

a cP 0 0 0 o " '""'-tJ 0

t t to

600

400

200

800

s1000 •

during perfusion in 4 of 54 patients (7%), the lowest valuesbeing 164, 170, 176, and 178 seconds. No device failureoccurred and a rinsed, heparin-coated heat-exchanger/oxygenator structure perfused with low systemic hepa­rinization is shown in Figure 5. There are no macroscopicred clots. Mean weights of the used heparin-coated heat­exchanger/ oxygenator structures are shown in Figure 6.The heparin-coated devices perfused with low systemicheparinization weighed 1,690 ± 16 g as compared with1,690 ± 26 g for full systemic heparinization (NS). Meanweight of heparin-coated cardiotomy reservoirs perfusedwith low systemic heparinization was 711 ± 52 g. Exclud­ing one cardiotomy reservoir weighing 870 g because ofexposure to protamine after weaning from cardiopulmo­nary bypass, the mean weight for cardiotomy reservoirsperfused with low systemic heparinization was 696 ± 12 gas compared with 681 ± 19 g for full systemic hepariniza­tion (Fig 7; NS).

Hospital mortality was 0 of 54 (0%) for low systemic

ACT>180 ACT>480

Fig 6. Oxygenator drained weight (mean::+: standard deviation).There is no difference between groups. (ACT = activated coagulationtime.)

heparinization versus 1 of 50 patients (2%) for full (NS)(Table 2). Bleeding required surgical repair in 0 of 54 (0%)for low systemic heparinization versus 4 of 50 patients(8%) for full (p = 0.05). Drainage (24 hours) was 790 ± 393mL for low systemic heparinization versus 1,039 ± 732 mLfor full (p < 0.025) as shown in Figure 8. The homologousblood products transfused in the two analyzed groups aredepicted in Figure 9. The amount of packed homologousred cells transfused (24 hours) was 300 ± 354 mL for lowsystemic heparinization versus 957 ± 596 mL for full (p <0.0005). The amount of fresh frozen plasma transfused (24hours) was 52 ± 218 mL for low systemic heparinization ascompared with 192 ± 396 mL for full (p < 0.01). Theamount of platelet concentrate transfused (24 hours) was7 ± 50 mL for low systemic heparinization as compared

9,-----------------

1000

800

600

400

200

oACT >180 ACT>480

Fig 5. Rinsed heparin-coated oxygenator perfused with low systemicheparinization. There are no visible macroscopic red clots.

Fig 7. Cardiotomy reservoirdrained weight (mean::+: standard devia­tion). There is no difference between groups. (ACT = activated coagu­lation time.)

Ann Thorac Surg1994;58:391-8

VON SEGESSER ET ALLOW SYSTEMIC HEPARINIZATION

395

Table 2. Outcome mll ~------------------,

1000

NS = not significant.

Variable

Patients transfusedPatients bleeding requiring

surgical revisionMyocardial infarctionHospital mortalityLate vein graft occlusion

Heparinization

Low Full p Value

25/54 (46%) 43/50 (86%) <0.000020/54 (0%) 4/50 (8%) 0.05

2/54 (4%) 2/54 (4%) NS0/54 (0%) 1/50 (2%) NS1/54 (2%) 0/50 (0%) NS

750

500

250

o

~ ACT>180

D ACT>480

Fig 9. Transfusion of blood products (mean ± standard deviation).DifferCllces bcttoccn groups are significant. (ACT = activated coagula­tion iime.)

aprotinin (not used here) is part of our standard practice.The higher difference in transfusion requirements as com­pared with the difference in blood loss may be due to thefact that transfusions are given only if the hematocrittrigger level described in the Methods section was reached.As previously demonstrated, reduced blood loss in perfu­sion with low systemic heparinization is attributablemainly to the improved function of the coagulation system[8-10, 12, 13, 21, 22]. Reduced platelet depletion [8, 10]during perfusion with low systemic heparinization andheparin-coated equipment as well as better preservation ofplatelet function [22] contribute to this finding. Further­more, more efficient neutralization of circulating heparinwith protamine has been demonstrated for perfusion withlow systemic heparinization as ACT, thrombin time, andprothrombin time returned to normal values earlier [13,21]. Interestingly, reduced production of n-dimers also wasobserved in the groups perfused with low systemic hepa-

with 36 ± 116 mL for full (p < 0.05). Transfusion ofhomologous blood products was necessary in 25 of54 (46%) for low systemic heparinization as comparedwith 43 of 50 patients (86%) for full (p < 0.00002). Base­line hematocrit of 43.3% ± 3.7% for low systemic heparin­ization versus 43.0% ± 3.9% (NS) for full before operationdecreased to 28.9% ± 3.2% for low versus 28.8% ± 3.2% forfull (NS) at 24 hours postoperative (Fig 10). Perioperativeinfarction according to the criteria defined above wasobserved in 2 of 54 (4%) for low as compared with 2 of 50patients (4%) for full (NS). Vein graft occlusion withoutmyocardial infarction was observed during late follow-upin 1 of 54 (2%) for low versus 0 of 50 patients (0%) forfull (NS).

Comment

Low systemic heparinization during cardiopulmonary by­pass with improved perfusion equipment in combinationwith adapted surgical techniques allows performance ofopen heart operation with reduced blood loss and trans­fusion requirements. It has to be stated here that the bloodloss reported for the control group does not represent theblood loss usually observed in our clinical practice as

red cells frozen plasma platelets

ACT>480

ACT>180

24 hbaseline

o

40

60

20

% ,-------------------------,

o

1500

500

ml , --,

1000

Fig 8. Chest tube drainage over 24 hours (mean ± standard devia­tion). Difference between groups is significant. (ACT = activated co­agulation iime.)

ACT>180 ACT>480Fig 10. Hematocrit before operation versus 24 hours after operation.There is no difference between groups despite less transfusions for pa­tients perfused with low systemic heparinization (see Fig 9). (ACT =

activated coagulation iimc.)

396 VON SEGESSER ET ALLOW SYSTEMIC HEPARINIZATION

Ann Thorac Surg1994;58:391-8

Table 3. Checklist for Perfusion With Reduced Systemic Heparinization, Potential Problems, and Recommended Action

Topic

Antithrombin III levelACT

Begin of CPB

Cardiotomy reservoirAir block in venous linePump stop during CPBCirculatory arrestWeaning from CPBMediastinal shed blood recovery

after protamine applicationRetransfusion of oxygenator

sump bloodUnforeseen surgical or technical

problem

Potential Problem

LowInadequate increase after loading dose

(lack of AT III?)Potential clot in cannulas

Potential occlusionPotential clot formation during pump stopPotential clot formationPotential clot formationPotential clot formationPotential clot formation

Potential clot formation

New priorities

Recommended Action

Addition of AT III or fresh frozen plasmaAddition of AT III or fresh frozen plasma

Venting of cannulas into venous line and venousreservoir

Avoid aspiration of clots, change cardiotomyRecirculation through shunt in the surgeon's fieldMore flow or more heparinFull systemic heparinizationRecirculation through shunt in the surgeon's fieldRed cell washing device

Red cell washing device

Full systemic heparinization

AT III = antithrombin Ill; CPB = cardiopulmonary bypass.

rinization as compared with full systemic heparinization.This difference probably was due to the fact that at thattime mediastinal shed blood recovery was performed witha red cell washing device in the group with low systemicheparinization as no cardiotomy reservoirs with improvedthromboresistance were available. Kunz and colleagues[22) studied coagulation parameters in patients undergo­ing coronary artery revascularization using either bubbleor membrane oxygenators with full systemic hepariniza­tion in comparison with heparin-coated membrane oxy­genators with low systemic heparinization. Patients of thelatter group had not only superior platelet aggregabilityinduced by adenosine diphosphate, but also bleedingtimes significantly closer to normal after perfusion,whereas there were longer bleeding times for perfusionboth with bubbler and membrane oxygenators using fullsystemic heparinization. This finding is in agreement withprevious reports [23). It has to be remembered here thatdetermination of the standardized bleeding time is some­what cumbersome, but in the surgical environment, thebleeding time is probably the most powerful test to deter­mine the existence of coagulation disorders. Consideringthis and other data it appears that reduced blood loss andconsecutively reduced transfusion requirements mainlyare due to the limited use of anticoagulant during theperfusion procedure, which by definition is closer to thenormal situation and allows also for limited use of prota­mine, which has some anticoagulant potential by itself.Some contribution of the heparin surface coating withregard to reduction of the whole body inflammatoryresponse also can be expected [14, 16-18). Reduced bloodcell adhesion to heparin-coated artificial surfaces [24) maybe one explanation for these findings.

However, if less anticoagulation is equivalent to morecoagulation, cardiopulmonary bypass cannot be realized instandard fashion even if heparin-coated components areused throughout the perfusion circuitry. Although it has

been demonstrated in the past that heparin-coated cardio­pulmonary bypass equipment remains fully functionalover hours and even days with low or no systemic hepa­rinization [8-13, 19,20], this holds true only for continuousflow. In contrast to full systemic heparinization (ACT morethan 480 seconds), where significant coagulation of stag­nant blood in general can be prevented (recirculationthrough the oxygenator is recommended as soon as thepatient is weaned from cardiopulmonary bypass), thiscannot be taken for granted if systemic heparinization isreduced. However, the increased risk of coagulationcaused by reduced systemic heparinization can be com­pensated for to some degree by reduction of phases withlow flow. Hence, for perfusion with low systemic hepa­rinization, we recommend to separate three significantlydifferent device flow patterns:

1. Phases with full pump flow (blood flows of 2 L/min ormore for an adult-size oxygenator). Perfusion with lowsystemic heparinization (ACT more than 180 seconds)can be realized with minimal risk of device failure overmany hours.

2. Phases with low pump flow (blood flows between 1 and2 L/min for an adult-size oxygenator) can be used withlow risk of device failure over shorter perfusion periods(minutes). Higher ACTs are recommended if longerlow-flow perfusion periods are expected.

3. Phases with minimal pump flow (blood flows less than1 L/min for an adult-size oxygenator) should be toler­ated for minimal duration only (seconds) and are bestavoided. Full systemic heparinization is recommendedfor circulatory arrest.

Decreased risk of clot formation can be expected duringperfusion with higher ACTs, high pump flows, and hypo­thermia, whereas increased risk of coagulation has to beexpected during perfusion with lower ACTs, low pump

Ann Thorne Surg1994;58:391-8

vorc SECFSSFR ET ALi.ow SYSTLivlJC IIEl'ARINIZATI(lf\j

397

flows, and normothermia. Potential problems during car­diopulmonary bypass with low systemic heparinizationare summarized in Table 3.

Because the antithrombotic activity of surface-boundheparin is antithrombin III (AT III) dependent, adequateAT III levels have to be secured. Transfusion of AT III orfresh frozen plasma is recommended if low AT III levelsare detected or inadequate ACT response is observed afterheparin loading. After cannulation of the aorta, blood maystay in the arterial cannula for some time. Venting of thearterial cannula before starting cardiopulmonary bypassallows this stagnant volume to drain into the filter systemof the venous reservoir (see Figs 1, 2). Clots occurring inthe surgical field before systemic heparinization should notbe aspirated into the cardiotomy system.

Air block in the venous line can be handled immediatelyby clamping of the arterial and venous cannulas andrecirculation through the arteriovenous shunt in the surgi­cal field (see Fig 2). Pump stop should be avoided duringperfusion with low systemic heparinization. Most surgicalsituations where pump stop is used also can be handledwith low pump flow (I L/min), which is much safer.Likewise, circulatory arrest cannot be recommended dur­ing perfusion with low systemic heparinization and is bestavoided. After weaning from cardiopulmonary bypass,recirculation is started immediately through the shunt inthe surgeon's field. This allows going back on bypass afterventing of the cannulas into the venous reservoir. Pumpsuction should not be used for mediastinal shed bloodrecovery after protamine application. Because protaminealso reacts with bound heparin [25], contamination of thecardiopulmonary bypass system may preclude its furtheruse. Direct retransfusion of oxygenator sump blood ispossible immediately after cardiopulmonary bypass. Lateron, retransfusion of washed and concentrated red cellsprobably is better. If unforeseen problems, either operationor device related, occur during perfusion with low sys­temic heparinization, the priorities change and immediatefull systemic heparinization is recommended.

This work was supported in part by the Swiss National Founda­tion for the Development of Scientific Research (grant numbers3.896-0.86, 32-26271.89, and 32-31045.91).

References

1. Young JA, Kisker CT, Doty DB. Adequate anticoagulationduring cardiopulmonary bypass determined by activated clot­ting time and appearance of fibrin monomer. Ann Thorac Surg1978;26:231-40.

2. Gravlee GP, Haddon WS, Rothberger HK, et al. Heparindosing and monitoring for cardiopulmonary bypass. J ThoracCardiovasc Surg 1990;YY:518-27.

3. Cardoso PFG, Yamazaki F, Keshavjee 5, et al. A reevaluationof heparin requirements for cardiopulmonary bypass. J ThoracCardiovasc Surg lYYl;101:153-60.

4. Metz 5, Keats AS. Low activated coagulation time duringcardiopulmonary bypass does not increase postoperativebleeding. Ann Thorac Surg 1990;49:440-4.

5. Gott VL, Whiffen jo. Darton RC Heparin bonding on colloidalgraphite surface. Science 1Y63;142:12Y7-8.

6. Larrn 0, Larsson R, Olsson I'. A new non-thrombogenicsurface prepared by selective covalent binding heparin via amodified reducing terminal residue. Biomater Med DevicesArtif Organs 1983;11: 161-73.

7. Kim SW, Ebert CD, Lin jY, McRen JC Nouthrombogenicpolymers-i-phannaccutical approaches. ASAIO j lY85;6:76-87.

8. Von Segesser LK, Turina Ml. Heparin coated hollow-fiberoxygenator without systemic heparinization in comparison toclassic bubble and membrane oxygenators. AmSECT Proceed­ings, March 11-14, IY88, Anaheim, CA:76-83.

Y. Von Segesser LK, Turina ML Cardiopulmonary bypass with­out systemic heparinization: performance of heparin coatedoxygenators in comparison to classic membrane and bubbleoxygenators. J Thorne Cardiovasc Surg I Y89;Y8:386-·%.

10. Von Segesser LK, Weiss 13M, Turina ML Perfusion withheparin coated equipment: potential for clinical use. SenunThorac Cardiovasc Surg 1YYO;2:373- 80.

11. Bindslev L, Eklund], Norlander 0, et al. Treatment of acuterespiratory failure by extracorporeal carbon dioxide elimina­tion performed with a surface heparinized artificial lung.Anesthesiology 1987;67: 117-20.

12. Von Segesser LK, Weiss 13M, Garcia E, et al. Reduced bloodloss and transfusion requirements with low systemic hepa­rinization: preliminary clinical results in coronary artery re­vascularization. Eur j Thorne Cardiova-.c Surg lYY0;4:639-43.

13. Von Segesser LK, Weiss 13M, Garcia E, Turina ML Reductionand elimination of systemic heparinization during cardiopul­monary bypass. j 1 horae Cardiovasc Surg IY')2;I03:790-Y.

14. Mollnes TE, Vidern V, Cotze 0, et al. Formation of C5a duringcardiopulmonary bypass: Inhibition by precoating with hep­arin. Ann Thorac Surg lY91;52:Y2-7.

15. Pradhan M], Fleming [S, Nkere UU, et ..I. Clinical experiencewith heparin-coated cardiopulmonary bypass circuits. Perfu­sion lY91;6:235-42.

16. Gu Y], van Oeveren W, van del' Kamp KWII], et al. Heparin­coating of extracorporeal circuits reduces thrombin formationin patients undergoing cardiopulmonary bypass. Perfusion1991;6:221-5.

17. Videm V, Svennevig JL, Fosse E, et al. Reduced complementactivation with heparin-coated oxygenator and tubings incoronary bypass operations. j Thorac Cardiovasc Surg 1992;103:806-13.

18. Gu YJ, van Oeveren W, Akkerman C, et al. Heparin-coatedcircuits reduce the inflammatory response to cardiopulmo­nary bypass. Ann Thorac Surg 1YY3;55:Y 17-22.

19. Von Segesser LK, Garcia E, Turina MI. Perfusion withoutsystemic heparinization for rewarming in accidental deephypothermia. Ann Thorne Surg IYY1;52:560--1.

20. Dowling RD, Brown ME, Whittington RO, et al. Clinicalcardiopulmonary bypass without systemic heparinization.Ann Thorac Surg 1993;56:1176-8.

21. Von Scgesscr LK, Weiss BM, von Felter, E, Turina Ml. Coag­ulation patterns in perfusion with low systemic hoparinization. J Extracorp Techno! lYY2;23:80-5.

22. Kunz M, von Segesser LK, Turina MI. Reduced coagulationdisorders, blood trauma and immunologic response aftercardiopulmonary bypass with heparin coated equipment.J Cardiovasc Surg (in press).

23. Edmunds LH. Blood-surface interaction during cardiopulmo­nary bypass. J Cardiovasc Surg IY93;8:404-IO.

24. Borowiec JW, Bylock A, van der Linden L Thelin S. Heparincoating reduces blood cell adhesion to arterial filters duringcoronary bypass: A clinical study. Ann Thorac Surg lYY3;55:1540-5

25. Von Segesser LK, Gyurech DD, Schilling Jj, Marquardt K,Turina Ml. Can protamine be used during perfusion withheparin surface mated equipment> ASAIO J ILJY3;3LJ:M190-4.

398 VON SEGESSER ET ALLOW SYSTEMIC HEPARINIZAnON

Ann Thorac Surg1994;58:391-8

DISCUSSION

DR 1. HENRY EDMUNDS, JR (Philadelphia, PAl: I am delightedto discuss this fine paper so nicely presented by Professor vonSegesser. I apologize to him that I have not had access to themanuscript and I did not know that he used a Duraflo IIheparin-coated circuit, but I do not think that is important to mycomments.

As all of us know, heparin is a prerequisite for cardiopulmonarybypass. You simply cannot perfuse blood extracorporeally with­out heparin. Thrombin is very difficult to measure directly, butrecently new markers of thrombin formation have been developedand these indicate that thrombin is formed during every perfu­sion. F1.2 is a fragment that is produced when prothrombin iscleaved to produce thrombin. Its presence indicates that thrombinhas been produced and is circulating. Fibrinopeptide A is afragment produced when thrombin cleaves fibrinogen to formfibrin. Thrombin-antithrombin III complex, or TAT, is producedby the combination of thrombin and antithrombin III. An increasein one or more of these three markers indicates formation ofthrombin.

Brister and colleagues have shown that during cardiopulmo­nary bypass with activated clotting times over 400 seconds,thrombin is continually formed directly in proportion to theduration of bypass. They measured F1.2 and showed a progres­sive increase in plasma F1.2 with increased duration of bypass.This increase also has been shown by another group. Ten yearsago Davies and Salzman showed that fibrinopeptide A wasproduced during cardiopulmonary bypass. Thus, with full-doseheparin and activated clotting times of more than 400 seconds,thrombin is produced during cardiopulmonary bypass.

Thrombin is not a nice enzyme to circulate. It is a very potentserine protease, and in addition to converting fibrinogen to fibrin,it activates platelets and endothelial cells directly to producetissue-plasminogen activator.

When I recently reviewed heparin-coated surfaces, the datashowed that the Carmeda heparin-bound surface reduces plateletadhesion and decreases thromboxane B2 production in vitro.

Animal studies are hampered by the fact that most immuno­chemical assays do not cross-react in animals. Most animal studiesonly show whether or not macroscopic clots have been produced.

In patients, most of the experience until recently has been inpatients with trauma, when the system has been used without anoxygenator, during aortic resections and during prolonged bypassfor respiratory insufficiency. In these circumstances all reportsindicate that you can use a heparin-bound surface with either low

or no-dose heparin safely. All reports lack data regarding throm­bin formation.

One of my colleagues recently had a 56-year-old patient whohad endocarditis of the mitral valve. The patient deterioratedhemodynamically, and because he had two recent cerebral in­farcts, my colleague wanted to use a system without heparin,although the chance of increasing the neurologic injury was only10% with full heparinization.

He replaced the valve with a prosthesis using the Carmedacircuit with half-dose standard heparin and activating clottingtimes around 200 seconds. As the patient was rewarmed theechocardiogram showed a large left atrial thrombus. He gavefull-dose heparin, opened the atrium, and took out a golf-ballsized clot. Happily, the patient survived.

In summary, I would like to emphasize that until we know whatis happening to thrombin during cardiopulmonary bypass, wewill be unwise to reduce the amount of systemic heparinization.

I enjoyed the paper and I think this is a very provocative subjectworth considerable discussion.

DR JOHN P. JUDSON (Harrisburg, PAl: I have a brief questionon cerebrovascular events in this series. Was there any differencein the two routines?

DR VON SEGESSER: I can readily answer the second commentfirst. There was no difference with regard to cerebrovascularaccidents in the two groups because this did not occur in eithergroup.

I thank Dr Edmunds for his wise comments, drawing ourattention that during perfusion with full or low systemic hepa­rinization, the coagulation system is not fully under control.Heparin does not control the thrombocyte or platelet activity, and,indeed, there may be problems that we cannot recognize readilywhen we run standard cardiopulmonary bypass.

I have seen clots in patients who have been perfused with fullsystemic heparinization; therefore, I am not surprised that one canfind some in patients who were perfused with low systemicheparinization. In my view, the coagulation system is a balanceand it can be the shifted to the side with more coagulation, butthen we have it everywhere and we have to take appropriatemeasures. I think the way we have proceeded we were able tohandle this situation, and we look forward to getting moreexperience in that field.