Heparin dosing and monitoring for cardiopulmonary bypass

J

THORAC CARDIOVASC SURG

1990;99:518-27

Heparin dosing and monitoring for cardiopulmonary bypass A comparison of techniques with measurement of subclinical plasma coagulation Subclinical plasma coagulation during cardiopulmonary bypass has been associated with marked platelet and clotting factor consumption in monkeys. To better define subclinical coagulation in man, we measured plasma fibrinopeptide A concentrations before, during, and after cardiopulmonary bypass. Patients were assigned to one of three groups of heparin management: group 1 (n = 10)-initial heparin dose 300 IU I kg, with supplemental heparin if the activated coagulation time feU below 400 seconds; group 2 (n = 6)-initial heparin dose 250 IU Ikg, with supplemental heparin if activated coagulation time was less than 400 seconds; and group 3 (n = 5)-initial heparin dose 350 to 400 IU /kg, with supplemental heparin if whole blood heparin concentration was less than or equal to 4.1 IU I ml, Activated coagulation time and heparin concentration were measured every 30 minutes during cardiopulmonary bypass, and fibrinopeptide A was measured at hypothermia, normothermia, and whenever activated coagulation time was less than 400 seconds. Quantitative and qualitative blood clotting competence was assessed after cardiopulmonary bypass, including mediastinal drainage for the first 24 hours. Fibrinopeptide A values were markedly elevated during cardiopulmonary bypass but were weU below the levels present before and after cardiopulmonary bypass. Fibrinopeptide A correlated inversely with heparin concentration during cardiopulmonary bypass (r = -0.46, p = 0.03~ but higher fibrinopeptide A levels during cardiopulmonary bypass did not correlate with postcardiopulmonary bypass coagulopathy. Group 3 patients received the highest heparin doses (p < 0.05) and had the greatest postoperative blood loss (p < 0.05). Protamine dose and heparin concentration during cardiopulmonary bypass correlated best with postoperative mediastinal drainage. Our findings support the foUowing conclusions: (1) compensated subclinical plasma coagulation activity occurs during cardiopulmonary bypass despite activated coagulation time greater than 400 seconds or heparin concentration greater than or equal to 4.1 IV Iml; (2) post-cardiopulmonary bypass mediastinal drainage correlates strongly with increased heparin concentration during cardiopulmonary bypass (p < 0.05) and protamine dose (p < 0.05); and (3) during cardiopulmonary bypass at both normothermia and hypothermia, activated coagulation times greater than 350 seconds result in acceptable fibrinopeptide A levels and post-cardiopulmonary bypass blood clotting.

Glenn P. Gravlee, MD, W. Scott Haddon, MD, Henry K. Rothberger, MD, Stephen A. Mills, MDCM, Anne T. Rogers, MBChB, Virgil E. Bean, MD, David H. Buss, MD, Donald S. Prough, MD, and A. Robert Cordell, MD, Winston-Salem. n.c.

From the Departments of Anesthesia. Cardiothoracic Surgery. Medicine (Section of Rheumatology), and Pathology. Bowman Gray School of Medicine of Wake Forest University. Winston-Salem.

N.C. Presented in part at the 1987 annual meeting of the American Society ~f Anesthesiologists. Atlanta. Ga., October I n7.

5I8

Received for publication Nov. 18. 1988. Accepted for publication May 3. 1989. Address for reprints: Glenn P. Gravlcc, M D, Department of Anesthesia, Bowman Gray School of Medicine. 300 S. Hawthorne Rd.. Winston-Salem. NC 27103.

12/1/14368

Volume 99 Number 3 March 1990

Anticoagulation with heparin has been an integral component of extracorporeal circulation for more than three decades, yet optimal dosing and monitoring schemes remain controversial. In 1975 Bull and colleagues] contributed importantly to heparin management by showing striking variability in the magnitude and duration of the anticoagulation response. These authors suggested routine monitoring of anticoagulation during cardiopulmonary bypass (CPB) with the activated coagulation time (ACT).2 Culliford and co-authors' subsequently revealed factors distorting the relationship between the ACT and blood heparin concentration. Foremost among these are hypothermia and hemodilution, which commonly accompany extracorporeal circulation. Because of this distortion, some authorsv" suggest monitoring heparin concentration during CPB, either with or without ACT monitoring.. For adequate anticoagulation during CPB, complete suppression of plasma coagulation activity appears desirable. The primate model of Young, Kisker, and Doty6 confirmed this hypothesis, applied the findings successfully to a small group of patients (six in number), and established that ACT values greater than 400 seconds suppressed production of fibrin monomer during CPR Fibrinopeptide A (FPA) represents another sensitive marker of thrombin-mediated conversion of fibrinogen to fibrin.7.8 We hypothesized that the presence of ACT prolongation during CPB would be more important than the contributory mechanisms of heparin, hypothermia, and hemodilution. Using plasma FPA levels as an indicator of subclinical coagulation activity, we compared different methods of heparin administration for CPB in 21 patients.

Methods Patient selection and management. After protocol approval was obtained from our human studies committee and informed consent from each patient, 21 patients undergoing cardiac operations were studied. Any adult requiring CPB was considered eligible unless one or more of the following exclusions applied: ( I) abnormal blood coagulation by clinical history or preoperative laboratory screening; (2) recent therapy with any of the followingdrugs: streptokinase within 10 days, aspirin within 10 days, warfarin within 5 days, or heparin within 3 days of operation; (3) reoperative cardiac procedure, (4) anticipated CPB duration exceeding 150 minutes, and (5) anticipated absence of CPB hypothermia below 30° C. All patients were anesthetized with benzodiazepines and high-dose opioids with muscle relaxants. Cardiopulmonary bypass was conducted with a hollow-fiber membrane oxygenator (Bentley, Irvine, Calif.) and systemic hypothermia to 25° to 29° C. The extracorporeal circuit was primed with 2000 ml of lactated Ringer's solution, 250 ml of 5% albumin in normal saline solution, 12.5 gm mannitol, and 10,000 IU heparin. Perfusion flows were adjusted to maintain a mixed venous oxygen saturation above 70% and mean arterial pressures between 40 and 90

Heparin dosing and monitoring for CPS

5 19

mm Hg. Cardiopulmonary bypass separation was accomplished after patients were rewarmed to a rectal temperature greater than or equal to 36 C. Preoperative testing included prothrombin time, activated partial thromboplastin time, platelet count, plasma fibrinogen concentration, ACT, and FPA level. Intraoperatively, radial arterial blood samples? were taken to measure ACT, whole blood heparin concentration (HC), and FPA at predetermined intervals during operation. After protamine was infused and a blood HC below 0.2 IU/ ml was attained, arterial blood was withdrawn for the following tests: ACT, prothrombin time, activated partial thromboplastin time, platelet count, and plasma concentrations of fibrinogen and fibrin/fibrinogen degradation products (FDP). The senior surgeon rated clinical coagulation as dry, intermediate, or wet (respective qualitative coagulopathy scores 0, I, and 2). Additional 25 to 50 mg doses of protamine were administered if ACT showed evidence of heparin rebound after return to baseline. Platelet concentrates were empirically transfused if a clinical coagulopathy existed despite a blood HC less than 0.2 IU /m!. Once the sternotomy was closed, mediastinal blood drainage was recorded hourly. Transfusion of homologous whole blood or red blood cells was not recorded after the patient left the operating room, because clinical transfusion practices varied among the surgical attending staff. Group assignment. Patients were placed in one of three groups. Group I included 10 patients whose initial dose of bovine lung heparin (Upjohn, Kalamazoo, Mich.) was 300 IU/kg. Subsequent doses of 50 to 100 IU/kg were given before or during CPB only if ACT fell below 400 seconds. Group 2 consisted of six patients whose initial heparin dose was 250 IU /kg. Subsequent heparin doses were administered by use of the same criteria as in group I. Group 3 included five patients whose initial heparin doses were 350 to 400 IU /kg. Additional heparin (100 IU /kg) was given if the blood heparin concentration was less than or equal to 4.1 IU /ml (Hepcon level of 3.0). Group I (December 1986 to February 1987) was studied before groups 2 and 3 (April to May 1987) to determine if the group I protocol would produce unacceptable subclinical coagulation activity during CPB. With confirmation of acceptable coagulation outcomes in group I, the initial heparin dose for group 2 was set at 250 IU /kg, with the intent to more often attain a pre-CPB ACT levelbetween 400 and 500 seconds. Initial protamine doses were determined by an algorithm based on blood HC, estimated blood volume, 10and a protamine/heparin neutralization ratio of 1.1 mg/IOO IU (HemoTec Inc., Englewood, Colo.). Coagulation testing. The ACT method used 2 ml of celiteactivated blood placed in a Hemochron 400 or 800 device (International Technidyne, Edison, N.J .). Heparin concentration was measured by automated protamine titration (HemoTec Inc.). Prothrombin time and activated partial thromboplastin time samples were collected in citrated tubes, centrifuged for plasma separation, and performed on an MLA Electra 700 (Medical Laboratory Automation, Pleasantville, N.Y.) automated photo-optical clot detector (normal prothrombin time, 11.0 to 13.4 seconds; normal activated partial thromboplastin time <31.0 seconds). Fibrinogen assay was based on clottable protein present in a diluted plasma sample to which excess thrombin (Dade Diagnostics lnc., Miami, Fla.) was added, producing a normal plasma fibrinogen range of 180 to 363 tng] d!. Serum FDP was measured on serum by latex agglutination by use of particles coated with antibodies ro fibrinogen and FDP 0

The Journal of Thoracic and Cardiovascular

5 2 0 Gravlee et al.

Surgery

-

-....

-Go GROUP 1 ....... GROUP2 -a- GROUP3

800

U

Q,)

en

600

U

400

«

CPS 200 O...L-_~----r----~-----,r--------.--

caD

HEPARIN

CQ\JTFDL

WARM PROTAMINE

TIME Fig. 1. ACT by group at different times during surgery (mean ± SE). CPB is denoted by the box. Plus sign indicates a significant difference (p < 0.05) between group 3 and each of the other groups. Time explanations: Control, before anesthetic induction; Heparin. 5 minutes after heparin administration; Cold. lowest temperature during CPB; Warm. esophageal temperature >36 0 C; Protamine. 5 minutes after completing protamine infusion.

-E

5

-...-

3

« D.

2

--=> -a:Z

W ::I:

......

* 4

*

CPB

1 Q....L.---.-----.,....----r----D--co,o HEPARIN WARM PROTAMINE

TIME Fig. 2. Whole blood heparin concentrations by group at different times during surgery (mean ± SE). Times are explained in Fig. I. Asterisks denote p < 0.05 for group 2 versus group 3. Plus sign indicates p < 0.05 for group 2 versus each of the other groups.

(normal, 0 to 8 ~g/ml), with higher values quantified by serial dilutions. Platelet counts were performed on ethyenediaminetetraacetic acid-anticoagulated samples with Coulter S-plus (Coulter Electronics, Hialeah, Fla.). Fibrinopeptide A samples were collected by syringe from the arterial catheter after prewithdrawal of 5 ml, then placed

immediately in a prechilled tube containing EDT A, aprotinin, and a thrombin inhibitor. Samples were mixed, placed on ice, and transported for centrifugation at 4 0 C for 15 minutes at 2000 g. Supernatant plasma was then pipetted and frozen at -80" C for later analysis. In group I, FPA was determined by radioimmunoassay, as described by Nossel and colleagues,"

Volume 99 Number 3

Heparin dosing and monitoring for CPB

March 1990

50

-E

-« C)

...

-o- GROUP 1

40

-a-

GROUP 2 GROUP 3

30 CPS

t:

Q.

52 1

20

U.

10 0

CCNTFO...

SURGERY

HEPARIN

COLD

WARM PROTAMINE

TIME Fig. 3. Plasma fibrinopeptide A concentrations by group at different times during surgery (mean ± SE). Times are explained in Fig. I; Surgery time represents a poststernotomy FPA level before heparin administration. Plus sign indicates p < 0.05 for group 3 versus group I.

with iodine 125-labeled FPA and rabbit anti-FPA antibodies from a commercial kit (Mallinckrodt Inc., St. Louis, Mo.). Because this k"it was discontinued after completing group I FPA assays but before beginning groups 2 and 3 assays, the latter two groups required a different technique, a competitive enzymelinked immunoassay (Asserachrom FPA, Diagnostica Stago, Asnieres-Sur-Seine, France). This method also uses rabbit FPA antibody but measures light absorbance rather than radioactivity. Amiral, Walenga, and Farced" described the method and found nearly identical FPA levels by Asserachrom competitive enzyme-linked immunoassay and Mallinckrodt RIA methods in both the normal and abnormal ranges. Statistical analysis. Statistical analysis was performed with StatView software (Brainpower Inc., Calabasas, Calif.). Demographic comparisons among groups were performed with the x2 method or one-factor analysis of variance (ANOV A) where appropriate. One-factor ANOVA was used to compare groups for differences in heparin and protamine doses, blood clotting studies, and total of heparin and protamine. Two-factor repeated-measures ANOVA compared preoperative and post-CPB clotting studies by group, as well as FPA levels over time by group. If variation among the three groups was significant, differences between groups were tested with Fisher's test of least significant differences. Group comparisons of the surgeon's clinical impression of coagulopathy were performed with the x2 method. For all tests, p < 0.05 was considered significant. Data from the three groups were pooled to permit simple least-squares linear (Pearson product-moment) regressions of factors possibly contributing to mediastinal blood drainage at 8 and 24 hours, considering regression relationships significant if the regression line slope differed (p <0.05) from zero. If a parameter's linear correlation (r) with 8-hour or 24-hour blood loss exceeded 0.25, it was placed into a stepwise regression analysis. For heparin concentration, ACT, and FPA, only the strongest correlation among low, mean (or median), and high CPB values was placed into this analysis. Simple regressions among CPB

ACT, heparin concentrations, and FPA levels were also performed.

Results Demographic. There were no significant differences among groups in mean age (58 to 62 years), height (169 to 176 em), weight (71 to 81 kg), body surface area (1.77 to 1.97 m-), duration ofCPB (97 to 112 minutes), or protamine dose (238 to 276 mg). Group 1 comprised seven patients with coronary artery bypass (CABG) and three patients with valve replacements; group 2 contained six patients with CABG, and group 3 contained four patients with CABG and one patient with valve replacement (differences not significant). Total heparin doses were 365 ± 65 (x 102 IV) in group 1, 298 ± 44 in group 2, and 423 ± 27 in group 3. Heparin monitoring and FPA concentrations. Figs. 1 to 3 plot ACT, heparin concentration, and FPA level by group at different points in the procedure. Analysis of variance showed significant variation in ACT among the groups after heparin administration (p = 0.(07) and after CPB rewarming (p = 0.036); group 3 differed significantly from the other groups at those times. Heparin concentrations varied significantly among groups after CPB rewarming (p = 0.(05), with group 2 differing from the other groups. Group variation in heparin concentration narrowly missed significance after initial heparin injection (p = 0.06) and during hypothermia (p = 0.08). Plasma FPA concentrations varied significantly during operation (p = 0.(001), with peaks occurring after ster-

The Journal of

5 2 2 Gravlee et al.

Thoracic and Cardiovascular Surgery

oQ)

(f)

.... U <

m a-

U Z

< W ~

900

••

800

•

•

700 600

•

500 400 300 200 100

y = 130.0694 + 133.3538x R = 0.51

a 2

3 4 5

MEDIAN CPS [HEPARIN] (IU/ml) Fig. 4. Scattergram between mean ACT and median heparin concentration during CPB; 95% confidence limits for the regression line slope are 25 to 241.

--g> <

40

y = 38.3278 - 7.1487x R = 0.46

E

• • •

30

au. 20 m a-

U Z

< W

10

• ••

~

•

o-t-___,r__'"""T"-~__r-_r_-r--__r-_r____,r____,

o

1

2

3

4

MEDIAN CPS [HEPARIN] (IU/ml)

5

Fig. S. Linear regression shows an inverse correlation between mean plasma FPA concentration and median heparin concentration during CPB; 95% confidence limits for regression line slope are -0.61 to -13.6.

notomy and protamine infusion (Fig. 3). Nineteen of 21 patients experienced their peak FPA level either at the sternotomy or the postprotamine sampling times. Average FPA levels during CPB stayed well below the peaks in all groups. Figs. 4 to 6 display regressions among ACT, FPA, and heparin concentration during CPB. To avoid distortion from unequal numbers of samples in different patients,

mean CPB values were used. Median CPB values were selected for heparin concentration, because the variability within individual patients was low and the concentration range tested was limited and discontinuous. Of the three simple regressions, the strongest association was between ACT and heparin concentration (r = 0.51, = 0.26, Y = 133x + 130, P = 0.02). Fibrinopeptide A was inversely related to heparin concentration (r =

r

Heparin dosing and monitoring for CPR

523

•• • y

=

683.968 - 5.904x R = 0.35

10

20

30

40

MEAN CPS FPA (ng/ml) Fig. 6. Regression relationship between mean ACT and mean plasma FPA concentration during CPS shows a weak, statistically insignificant association (p = 0.12); 95% confidence limits for the regression line slope are - 2.13 to 0.16.

-0.46,?- = 0.21,y = -7.lx + 38.3,p = 0.03) and ACT (r =-0.35,?- = 0.12,y = -0.20x + 251, P = 0.12); the slope of the latter relationship failed to reach significance. Nine CPB FPA values (in five patients) exceeded 20 ng] ml during CPB, coinciding with ACT values from 379 to 856 seconds and heparin concentrations from 2.0 to 4.1 IU /ml. Eight of these nine markedly elevated CPB FPA levels occurred when ACT exceeded 400 seconds and six when HC equalled or exceeded 3.4 IV/ml. Although ACTs of less than 400 seconds occurred eight times, only one coincided with an FPA level greater than 20

ng/rnl. Blood clotting after cardiopulmonary bypass. Table I compares indicators of post-CPB blood clotting by group. Analysis of variance demonstrated significant variation among the groups in 24-hour mediastinal drainage (p = 0.023), with group 3 having significantly greater blood loss than groups I and 2. Group 3 experienced the largest blood loss over the first 8 hours as well, but overall group variation did not reach significance at that interval. None of the other parameters listed in Table I showed differences among groups, but prothrombin time, activated partial thromboplastin time, fibrinogen, and platelet count all changed significantly between the preoperative and postprotamine measurements. Only one patient (in group I) in any group received treatment other than protamine for post-CPB bleeding; that patient received a IO-unit platelet concentrate. Pearson product-moment correlations relating 8-hour

mediastinal drainage to factors potentially influencing post-CPB coagulopathy showed statistically significant correlations with median CPB heparin concentration (r = 0.64, P = 0.(02) and protamine dose (r = 0.50, P =0.02). Similar correlations occurred with 24-hour blood loss (protamine dose r = 0.63, P = 0.003, median CPB heparin concentration r = 0.58, P = 0.(07). Parameters that did not correlate significantly with postoperative blood loss included FPA, ACT, total heparin dose, CPB duration, protamine/heparin dose ratio, and all postprotaminecoagulation tests. Stepwise regression analysis for 24-hour mediastinal drainage accepted only protamine dose and CPB heparin concentration as independent positive correlators. Variation in those two parameters statistically explained 53% of the observed variation in 24-hour drainage. Discussion Heparin monitoring and fibrinopeptide A levels. All three heparin dosing protocols suppressed CPB plasma coagulation activity when compared with activity before and after bypass. With the exception of group 3 during hypothermia, FPA levels during CPB were indistinguishable among groups. Fibrinopeptide A levels correlated with ACT (r = -0.35, P = NS) and heparin concentration (r = -0.46, P = 0.03), but neither relationship suggests an ACT or HC threshold below which FPA substantially rises. Measuring fibrin monomers in rhesus monkeys during 2 hours ofCPB at different ACT levels,

The Journal of Thoracic and Cardiovascular Surgery

5 2 4 Gravlee et al.

Table I. Selected group comparisons ANOVA P values N*

Group 1

8-hour blood loss 21 407 ± 24-hour blood loss 20 659 ± Prothrombin time (sec) Before 20 11.9 ± After 21 15.2 ± Activated partial thromboplastin time (sec) Before 24.7 ± 20 After 21 33.4 ± Fibrinogen (rng/dl) Before 16 308 ± After 18 166 ± Platelet count (XIOJ/ltl) Before 263 ± 21 After 21 147 ± FDP (ltg/ml) After 17 13 ± Qualitative coagulopathy score" 0 6 1 1 2 3

Repeated measures

Group 2

Group 3

One-factor

209 270

401 ± 170 738 ± 299

653 ± 216 1104 ± 197

0.079 0.023t

0.0001

0.6 1.1

12.1 ± 0.7 14.6 ± 0.4

12.0 ± 0.2 15.1 ± 1.5

0.707 0.511

0.0001

1.7 6.5

26.2 ± 3.0 30.0 ± 3.0

25.8 ± 1.5 33.1 ± 4.2

0.353 0.435

0.0001

92 38

351 ± 57 188 ± 21

323 ± 15 159 ± 36

0.549 0.338

0.0001

64 40

271 ± 48 144 ± 53

306 ± 99 168 ± 88

0.527 0.753

9

29 ± 22

12 ± 9

0.128

5 1 0

2 2 0

NA NA NA

0.0001

0.213:j:

Data are presented as mean ± SD.

Missing values occurred as follows: 24-hour blood loss-t-one from gorup I; prothrombin time and activated partial thromboplastin tjme~one each from group I: fibrinogen before-four from group I. one from group 3; fibrinogen after-two from group 1. one from group 3 (16 before and after matched pairs): fibrinogen degradation products-three from group I. one from group 3; qualitative coagulopathy score-s-one from group 3. Before. Preoperative value; After. post-protamine value; NA. not applicable. "Number of patients. tp < 0.05 for group I versus group 3 and for group 2 versus group 3 by I test. tChi-square p value combining group and coagulopathy score.

Young, Kisker, and Doty" identified an ACT threshold (400 seconds) below which platelet and fibrinogen consumptionclearlywould haveproduced a post-CPB coagulopathy. They tested their results in five pediatric patients and found that CPB ACTs exceeding 400 seconds suppressed fibrin monomer formation. We could not find a sensitivity comparison of fibrin monomer and FPA levels forplasmacoagulation activity, butonereview suggests that FPA is more sensitive. II If so, this might explain the differentsubclinical coagulation findings betweenthe studiesby Young,Kisker, and Doty" and by us. Alternatively, man may havea lowerACT threshold for clinically important CPB plasma coagulation than monkeys. Young, Kisker, and Doty" apparently did not test ACT levels below 400 seconds in man. What constitutes unacceptable plasma conagulation activityduring CPB? Both operation and the postoperativestate causemarked FPA elevation evenwithoutCPB. Mammen and colleagues 12measuredFPA infive patients undergoing CPB and foundhigherlevels beforeand after CPB than during CPB. Their postthoracotomy FPA levels averaged98 ng/rnl, whereas CPB levels averaged 7.9

to 8.3 ng/rnl. Davies, Sobel,and SalzmanI 3 found postincision FPA levels exceeding 100 ng/ml in 11 of 15 patientswho had cardiac operations and a mean level of 37 ± 25 ng/rnl late in CPB. They attributed the latter FPA levels to inadequate anticoagulation, judging from 54 of 73 CPB ACTs that fell below 400 seconds. The thromboplastic nature of operative trauma is well established, and it appears reasonable to assumethat safe anticoagulation exists whenCPB FPA levels are below those produced by operation alone,even if CPB FPA levels far exceed normal.FPA levels greater than 20 ng/rnl during CPB did not lead to post-CPB coagulopathy, although these levels were not always sustained. We suggest that considerable subclinical plasmacoagulation normally occurs during CPB withoutimplying inadequate anticoagulation or heralding post-CPB coagulopathy. The lowest CPB FPA levels were found in group 3 during hypothermia (Fig. 3). Because thisoccurrence coincided with the highestCPB ACTs (p = NS) and heparin concentrations (p < 0.05comparedwithgroup2), it might be assumed that a regimen of a higherdoseof heparin better protects against clotting factor consumption

Volume 99 Number 3 March 1990

and post-CPB coagulopathy. Higher postoperative blood loss (p < 0.05) in group 3 contradicts this theoretical advantage. Nevertheless, logic dictates that there must be some level of sustained CPB FPA elevation that would produce clinically important clotting factor consumption and post-CPB coagulopathy. Surely this would be a time-related phenomenon, and perhaps our CPB durations were insufficient to manifest it. Hofman and Straub'" measured FPA levels in a patient who had therapeutic defibrination with snake venom. Baseline FPA levels of 1.8 ng/rnl rose to 280 to 304 ng/rnl 30 to 360 minutes after venom injection. Corresponding fibrinogen levels were 10 to 76 mg/dl, Kockum'? reported a fibrinogen concentration of 300 rng/dl despite an apparently sustained FPA level of 96.0 ng/rnl, Nossel and colleagues- calculated that a 24-hour total FPA production of 0.3.6 to 0.60 mg represented lysis of 40 to 67 mg of fibrinogen, representing just 2% to 3% of normal daily catabolism. These studies suggest that a sustained FPA level of 16 to 25 ng/rnl would match normal fibrinogen turnover. Thus, despite being many times the upper limit of "normal," the CPB FPA levels we report prove insufficient to dangerously deplete fibrinogen over 2 hours. Postoperative bleeding. Group 3 experienced the largest 24-hour mediastinal drainage (p < 0.05). This group received the largest amounts of heparin and sustained the highest CPB ACTs and heparin concentrations. Why might this regimen predispose patients to increased bleeding? Heparin's effects on coagulation are complex, involving several levels of the intrinsic and extrinsic coagulation cascades, tissue plasminogen activator, and platelet activation." Protamine does not neutralize all of heparin's anticoagulant actions, notably its action on factor X, which smaller heparin molecules preferentially inhibit. 17 Commercial heparin solutions contain a broad spectrum of molecular weights; thus higher doses should increase the amount of circulating heparin that cannot be pharmacologically neutralized. Unfortunately, traditional clotting times and protamine titration heparin assays do not reliably detect this effect. Furthermore, although heparin's large molecular size and marked polarity prohibit widespread extravascular distribution, some tissue distribution does occur. IS. 19 This tissue heparin reservoir should enlarge in proportion to blood heparin concentration, theoretically predisposing to postoperative heparin rebound. The regressions performed with 8-hour and 24-hour blood loss showed that only CPB heparin concentration(s) and protamine dose correlated significantly with blood loss. Traditional coagulation tests correlated insignificantly with blood loss, as did ACT and FPA levels during CPB. The correlation between protamine dose and

Heparin dosingand monitoring for CPB 525

postoperative bleeding is consistent with some previous studies. Protamine adversely affects ACT, activated partial thromboplastin time, platelet count, and platelet aggregation in dogs.i" and the study by Mammen and colleagues'? in man revealed marked post-CPB platelet functional compromise after protamine dosing. Guffin and co-workers-' demonstrated that higher protamine doses increased postoperative blood loss. Clinical implications for heparin dosing and monitoring. What is the appropriate ACT range for CPB? Most published recommendations are empirical. Bull and colleagues 1 reported an impression that ACT levels below 300 seconds were associated with visible clots in the bypass circuit, whereas Young, Kisker, and Doty" documented the desirability of a 4oo-second threshold in monkeys. Jobes and colleagues'? identified an ACT level of 300 seconds that ensured an empirically determined, safe CPB heparin concentration level of 2 IV/ml. Dauchot and colleagues-' reported on 22 adults with mean CPB ACTs of 360 ± 32 seconds at the onset of hypothermia and 267 ± 30 seconds after rewarming, apparently without producing obvious clots or post-CPB coagulopathy. Culliford and co-workers.' performed post-CPB scanning electron microscopy on arterial filters in patients perfused with ACTs ranging from 249 to 1800 seconds (mean 674 ± 25 seconds) and found minimal debris deposition. Neither of the latter two studies systematically evaluated post-CPB coagulation, but they suggest an ACT safety threshold of 300 seconds or less. ACTs of 170 to 240 seconds have been used for prolonged extracorporeal membrane oxygenation.ev-" Heiden and colleagues" reported high incidences of bleeding complications (25%) and thrombosis/embolism (25%) in that setting, but they (perhaps mistakenly) did not attribute these to inadequate anticoagulation. After Esposito and co-workers/" demonstrated a poor correlation between heparin concentration and ACT during CPB, some authorsv 5 attached importance to heparin concentration monitoring during CPB. Might a heparin concentration threshold be more important than an ACT threshold? Although our study sought to evaluate this, the lack of association between CPB FPA levels and post-CPB coagulation outcomes compromises clinical interpretation. We found heparin concentrations as low as 2.7 IV/ml during hypothermia, with simultaneous FPA levels less than 10 ng/rnl. Conversely, heparin concentrations greater than 4.0 IV/ml coincided with FPA levels greater than 25 tig]mI. Our findings suggest neither a heparin concentration threshold that confers protection against clotting factor consumption, nor that any advantage accrues by monitoring heparin concentrations to guide CPB heparin dosing. Relating hypothermic ACTs

The Journal of Thoracic and Cardiovascular

5 26 Gravlee et al.

to FPA levels also proves unrewarding, but our data support the safety of ACTs greater than 350 seconds at both normothermia and hypothermia. This impression is largely derived from coagulation outcomes. Our findings suggest that excessive CPB heparin concentrations increase postoperative blood loss. That it is better to give too much than too little heparin has long served as a clinical axiom. Although we do not challenge that view, we submit that excessive heparin administration for CPB increases post-CPB bleeding despite a strict protocol for heparin neutralization and for the diagnosis and treatment of heparin rebound. High CPB heparin concentrations and the associated need for higher protamine doses independently contribute to post-CPB blood loss. It is difficult to assign any benefit to ACT values greater than 500 seconds, although normothermic ACTs between 350 and 500 seconds often exceed 500 seconds during hypothermia. We see no harm in this, but we cannot demonstrate any need to increase the target ACT range to accommodate hypothermia, or to anticipate rewarming with additional heparin when hypothermic ACTs exceed 400 seconds. In conclusion, normal plasma coagulation activity occurs during CPB, but it appears to be compensated and does not correlate with post-CPB bleeding. Activated coagulation times exceeding 350 seconds during CPB produce plasma coagulation activity below that engendered by operation alone, and hypothermia does not increase the minimum safe ACT threshold. Higher CPB heparin concentrations are associated with increased post-CPB bleeding; thus excessive heparin should be avoided. We suggest an ACT range of 350 to 500 seconds during CPB; higher ACT values are acceptable but inessential during hypothermia. We gratefully acknowledge the secretarial assistance of Debra Richards, the technical assistance of Linda Tarleton, Pat Boese, and Carmen Scott, and Jeannie Kiger's assistance in acquiring postoperative data. Wealso thank Drs. Brian Horan, G. A. Harrison, Anthony Dodds, and Joyce Low (St. Vincent's Hospital, Sydney, Australia) for reviewing the manuscript. REFERENCES I. Bull BS, Korpman RA, Huse WM, Briggs BD. Heparin therapy during extracorporeal circulation: I. Problems inherent in existing heparin protocols. J THORACCARDlOVASC SURG 1975;69:674-84. 2. Bull BS, Huse WM, Brauer FS, Korpman RA. Heparin therapy during extracorporeal circulation: II. The use of a dose-response curve to individualize heparin and protamine dosage. J THORAC CARDIOVASC SURG 1975;69:685-9. 3. Culliford AT, Gitel SN, Starr N, et al. Lack of correlation between activated clotting time and plasma heparin during cardiopulmonary bypass. Ann Surg 1981;193:105-11.

Surgery

4. Umlas J, Gauvin G, Taff R. Heparin monitoring and neutralization during cardiopulmonary bypass using a rapid plasma separator and a fluorometric assay. Ann Thorac Surg 1984;37:301-3. 5. Saleem A, Shcnaq SS, Yawn DH, Harshberger K, Diemunsch P, Mohindra P. Heparin monitoring during cardiopulmonary bypass. Ann Clin Lab Sci 1984;14:474-9. 6. Young JA, Kisker T, Doty DB. Adequate anticoagulation during cardiopulmonary bypass determined by activated clotting time and the appearance of fibrin monomer. Ann Thorac Surg 1978;26:231-40. 7. Amiral J, Walenga JM, Fareed J. Development and performance characteristics of a competitive enzyme immunoassay for fibrinopeptide A. Semin Thromb Hemost 1984; 10:228-42. 8. Nossel JL, Yudelman I, Canfield RE, et al. Measurement of fibrinopeptide A in human blood. J Clin Invest 1974; 54:43-53. 9. Clapham MCC, Willis N, Mapleson WW. Minimum volume of discard for valid blood sampling from indwelling arterial cannulae. Br J Anaesth 1987;59:232-5. 10. AllenTH, Peng MT,Chen MP, HuangTF,ChangC, Fang HS. Prediction of blood volume and adiposity in man from body weight and cube of height. Metabolism 1956;5:32845. II. Walenga JM, Hoppensteadt D, Emanuele RM, Fareed J. Performance characteristics of a simple radioimmunoassay for fibrinopeptide A. Semin Thromb Hemost 1984; I 0:219-

27. 12. Mammen EF, Koets MH, Washington BC, et al. Hemostasis changes during cardiopulmonary bypass surgery. Semin Thromb Hemost 1985; II :281-92. 13. Davies GC, Sobel M, Salzman EW. Elevated plasma fibrinopeptide A and thromboxane B2 levels during cardiopulmonary bypass. Circulation 1980;61 :808-14. 14. Hofman V, Straub PW. A radioimmunoassay technique for the rapid measurement of human fibrinopeptide A. Thromb Res 1977;11:171-81. 15. Kockum C. Radioimmunoassay of fibrinopeptide A-clinical applications. Thromb Res 1976;8:225-36. 16. Farced J. Heparin, its fractions, fragments and derivatives: some newer perspectives. Semin Thromb Hemost 1985; II :1-9. 17. Racanelli A, Fareed J, Walenga JM, Coyne E. Biochemical and pharmacologic studies on the protamine interactions with heparin, its fractions and fragments. Semin Thromb Hemost 1985; II: 176-89. 18. Gravlee GP, Angert KC, Tucker WY, Case LD, Wallenhaupt SL, Cordell AR. Early anticoagulation peak and rapid distribution after intravenous heparin. Anesthesiology 1988;68: 126-9. 19. deSwart CAM, Nijmeyer B, Roelofs JMM, Sixma JJ. Kinetics of intravenously administered heparin in normal humans. Blood 1982;60: 1251-8. 20. Kresowik TF, Wakefield TW, Fessler RD, Stanley Jc. Anticoagulant effects of protamine sulfate in a canine model. J Surg Res 1988;45:8-14.

Volume 99 Number 3 March 1990

21. Guffin AV, Dunbar RW, Kaplan JA, Bland JW. Successful use of a reduced dose of protamine after cardiopulmonary bypass. Anesth Analg 1976;55: 110-3. 22. Jobes DR, Schwartz AJ, Ellison N, Andrews R, Ruffini RA, Ruffini JJ. Monitoring heparin anticoagulation and its neutralization. Ann Thorac Surg 1981;31:161-6. 23. Dauchot PJ, Berzina-Moettus L, Rabinovitch A, Ankeney JL. Activated coagulation and activated partial thromboplastin times in assessment and reversal of heparin-induced anticoagulation for cardiopulmonary bypass. Anesth Analg 1983;62:710-9. 24. Hill JD, Dontigny L, deLeval M, Mielke CH. A simple

Heparin dosing and monitoring for CPB 5 2 7

method of heparin management during prolonged extracorporeal circulation. Ann Thorac Surg 1974;17: 129-J4. 25. Heiden D, Mielke CH, Rodvien R, Hill JD. Platelets, hemostasis, and thromboembolism during treatment of acute respiratory insufficiency with extracorporeal membrane oxygenation: experience with 28 clinical perfusions. J THORAC CARDIOVASC SURG 1975;70:644-5. 26. Esposito RA, Culliford AT, Colvin SB, Thomas SJ, Lackner H, Spencer Fe. The role of the activated clotting time in heparin administration and neutralization for cardiopulmonary bypass. J THORAC CARDIOVASC SURG 1983; 85:174-85.