Comparison of three blood-processing techniques during and after cardiopulmonary bypass

Comparison of Three Blood-Processing Techniques During and After Cardiopulmonary Bypass Robin G. Sutton, MS, John M. Kratz, MD, Francis G. Spinale, PhD, and Fred A. Crawford, Jr, MD Division of Cardiothoracic Surgery and Department of Extracorporeal Circulation Technology, Medical University of South Carolina, Charleston, South Carolina

The return of extracorporeal circuit blood at the termination of cardiopulmonary bypass is an important feature of blood conservation during open heart procedures. However, the relative benefits and disadvantages of different circuit blood salvage methods remain unclear. Accordingly, the purpose of this study was to examine whether quantifiable differences exist between three different circuit blood-salvaging techniques: direct infusion, centrifugation, and ultrafiltration. Sixty patients with very similar preoperative characteristics were randomly assigned to each of the three groups, and blood coagulation screens, plasma profiles, and respiratory function were determined at 20 minutes and at 6 and 18 hours after cardiopulmonary bypass. Early after cardiopulmonary bypass (20 minutes), the plasma colloid osmotic pressure and fibrinogen and platelet concentrations were significantly higher with ultrafiltration ( p < 0.05) versus those observed for the other two methods.

The plasma thromboplastin times were significantly ( p < 0.05) longer after cardiopulmonary bypass with centrifugation as compared to direct infusion and ultrafiltration. However, the coagulation profiles and plasma composition normalized by 18 hours after cardiopulmonary bypass with all three blood-salvaging methods. There were no significant differences in terms of blood utilization or chest tube drainage over the entire postoperative period among any of the circuit blood-salvaging methods. These results suggest that ultrafiltration of postcardiopulmonary circuit blood may preserve plasma colloid pressure and platelet concentration in the early postoperative period, but these differences do not persist. Thus, for routine cardiopulmonary bypass procedures, direct infusion, centrifugation, and ultrafiltration may all be satisfactory methods of circuit blood salvage.

A

pressure to remove plasma water from the cellular elements and plasma proteins. Assembled in parallel to the cardiopulmonary bypass circuit, this process can be used during cardiopulmonary bypass to remove excess plasma water due to the hemodilution brought about by cardioplegia, surgical irrigation, and crystalloid administration. An apparent advantage of the ultrafiltration method is that both the plasma proteins and the red blood cells are salvaged [2-61. However, disadvantages of this method can include: (1) exposure of the circuit blood to high transmembrane pressures, which can cause hemolysis; and (2) the heparin is concentrated, requiring additional protamine for reversal. In an attempt to determine the relative advantages of centrifugation and ultrafiltration, past studies have compared the use of these two blood-processing methods in patients undergoing cardiopulmonary bypass [7-141. In most of these studies, the results associated with ultrafiltration appeared to be better than those from centrifugation. However, interpretating the results from these past studies is complicated by a disparity in the parameters examined, reduced sample sizes, and diversity within the patient samples. Moreover, a prospective study that included direct infusion of the extracorporeal circuit blood for comparison purposes was rarely performed. As a result, the relative advantages and disadvantages of the direct infusion, centrifugation, and ultrafiltration of circuit

ppropriate salvage and processing techniques of the blood remaining in the extracorporeal circuit after cardiopulmonary bypass have historically been and continue to be a subject of great interest. In the early 1970s, blood remaining in the extracorporeal circuit was administered directly to the patient after cardiopulmonary bypass. This direct infusion of the large volume of diluted extracorporeal circuit blood depended on the patient having adequate cardiac and renal function to accommodate the increased volume load. To address this problem, the centrifugation technique was developed in the late 1970s [l]. In this method, the diluted blood remaining in the extracorporeal circuit is centrifuged, resulting in a red blood cell suspension relatively free of heparin, plasmafree hemoglobin, and other solutes. In direct contrast to the direct infusion method, blood salvaged by centrifugation resulted in a hematocrit value of 50% to 70%. However, a disadvantage to the centrifugation method is that only the red blood cells are retained, while the less dense components, such as plasma proteins, platelets, and coagulation factors, are discarded. A second means by which cardiopulmonary circuit blood is processed is by ultrafiltration [2-61. Ultrafiltration uses transmembrane Accepted for publication Dec 22, 1992. Address reprint requests to Mr Sutton, University of Iamwa Hospitals and Clinics, Iowa City, IA 52242.

0 1993 by The Society of Thoracic Surgeons

(Ann Thorac Siirg 1993;56:93843)

0003-4975/93/$6.00

SUTTONETAL CIRCUIT BLOOD SALVAGE METHODS

Ann Thorac Surg 1993;5693843

blood after cardiopulmonary bypass remain unclear. Accordingly, the overall goal of this project was to determine whether these are quantifiable differences among these three methods in the preparation and delivery of postcardiopulmonary bypass circuit blood.

Material and Methods To determine whether quantifiable and significant differences in the coagulation profiles, plasma composition, and respiratory function occur after the use of different modalities of extracorporeal circuit blood administration, 60 patients undergoing elective coronary bypass graft procedures were randomly assigned in a prospective manner to one of three groups: (1) direct infusion of post-cardiopulmonary bypass circuit blood with no processing; (2) centrifugation of the circuit blood before administration; and (3) ultrafiltration of the circuit blood before administration. Patient consent was not required because the procedures for post-cardiopulmonary bypass circuit salvage are standardized and no additional blood samples or laboratory tests were required .for the study. The patients selected represented nonemergent cases and had undergone no previous heart procedures. Patients who exhibited signs of renal dysfunction or bleeding disorders, and those on long-term antithrombotic or antiplatelet therapy were excluded from the study. An estimated cardiopulmonary bypass hematocrit of 18% or greater was established as a criterion for the patient to be entered into the study.

Cardiopulmonary Bypass The cardiopulmonary bypass protocol and equipment used were identical for all patients enrolled in the study. The cardiopulmonary bypass circuit consisted of a hollowfiber membrane oxygenator with a venous reservoir bag (16310 and 16322; Sarns 3M, Inc, Ann Arbor, MI), a filtered cardiotomy reservoir (H-4700; C.R. Bard, Inc, Tewksburg, MA), and an arterial line filter (EC3840; Pall Biomedical Products, Corp, Glen Cove, NY). The perfusion pump was controlled by a Computer Aided Perfusion System (Shiley Laboratories, Irvine, CA) and a custom tubing pack (Baxter Bentley Laboratories, Irvine, CA) during cardiopulmonary bypass. Blood flow (1.6 to 2.5 l/midm2) and blood pressure (50 to 70 mm Hg) were maintained within identical limits for the three groups. All patients were cooled to a moderate hypothermia (28”C), and myocardial protection was achieved by oxygenated crystalloid cardioplegia (pH, 7.70 to 7.77; oxygen tension, 480 to 560 mm Hg; Na+, 152 mEqL; K+, 24 mEq/L; Ca2+, 4.5 mEq/L; C1-, 156 mEqL; HC03, 5 mEq/L) delivered through the aortic root. The priming solution for the cardiopulmonary bypass circuit in all cases consisted of 500 mL of 6% hetastarch and 1,000 mL of lactated Ringers solution. No blood products or colloids were administered to the patient during or after cardiopulmonary bypass until 20 minutes after the appropriate circuit blood was administered. Blood for laboratory studies was drawn at that time, and afterward bank blood was administered as required.

939

Direct lnfusion of Circuit Blood In the direct infusion of circuit blood, no attempt was made to concentrate the patient’s blood volume during or after cardiopulmonary bypass. The post-cardiopulmonary bypass circuit blood was placed in a transfusion bag and administered to the patient. Diuretics were given as necessary. Protamine was administered through a separate line to neutralize the heparin given in the processed blood. The protamine dose was calculated by determining a ratio of the patient’s protamine dose to the patient’s volume, and applying that to the circuit volume.

Centrifugation of Circuit Blood A centrifuge system (Cell Saver Plus; Haemonetics, Inc, Braintree, MA) was attached to the cardiotomy drain line of the extracorporealcircuit, and the volume was removed to approximate the amount of crystalloid cardioplegia given. Bowls were not washed during cardiopulmonary bypass. The post-cardiopulmonary bypass circuit volume was concentrated by the centrifuge and washed with 500 mL of 0.9% sodium chloride. The packed red blood cells were administered to the patient through a peripheral line. After cardiopulmonary bypass, no additional protamine was administered to the patients in the centrifugation group unless the activated clotting time was above baseline values. The volume was cautiously removed during cardiopulmonary bypass to avoid the need to replace it with crystalloid so that the patient could be safely weaned from cardiopulmonary bypass.

Ultrafiltration of Circuit Blood The ultrafiltration system was set up in parallel to the extracorporeal circuit using a hemoconcentrator (H4201; C.R. Bard) which was primed according to the manufacturer’s recommendations. During cardiopulmonary bypass, attempts were made to keep the amount of plasma water filtered equal to the amount of crystalloid cardioplegia infused. After cardiopulmonary bypass, the circuit volume was concentrated by approximately 50% using the hemoconcentrator, then placed in a blood transfer bag and administered to the patient through a peripheral intravenous line. Protamine was administered in fashion similar to that used in the direct infusion group. As in the centrifugation group, the volume was cautiously removed during cardiopulmonary bypass to prevent the need to replace it with crystalloid to safely wean the patient from cardiopulmonary bypass.

Data Analysis The variables measured and the times these data were collected for this study are outlined in Tables 1 to 5. Analysis of variance and covariance was performed on the observations obtained from the three groups using the BMDP computerized statistical package (88.2; BMDP Statistical Software, University of California Press, Berkeley, CA). A two-tailed a level of less than 0.05 was used throughout the analyses. Because certain variables may be affected by specific pretreatment variables, analysis of covariance was performed. The Scheff6 method and the Tukey-Kramer multiple-comparison methods were uti-

940

S~ONETAL CIRCUIT BLOOD SALVAGE METHODS

Ann Thorac Surg 1993;56:938-43

lized to compare treatment means. Based on the large number of variables collected in the study, and the relatively high degree of variance associated with some of these values, a power analysis was performed to ensure that the selected sample sizes were adequate to avoid significant type II errors. In all cases, the power analysis revealed a minimum of 0.85 for all variables tested. All data are presented as the mean f standard error of the mean.

Results Twenty patients were entered into each of the three circuit blood-salvaging protocols: direct infusion, centrifugation, and ultrafiltration. Three patients in the direct infusion group were excluded, 2 because they were given packed red blood cells during cardiopulmonary bypass, and a third patient because packed red blood cells were given before completion of the circuit blood administration. In the ultrafiltration group, 2 patients were excluded from further analysis because both were given blood before circuit blood administration. This left 18 patients in the ultrafiltration group, 20 patients in the centrifugation group, and 17 patients in the direct infusion group. No deaths or complications occurred in any of the patients enrolled in any of the three circuit blood administration protocols.

Table 1 . Patient Profile and Pre-Cardiopulmonary Bypass Data" Variable Age (Y) Weight (kg) Body surface area (mZ) Left ventricular ejection fraction Fibrinogen (mg/dL) Platelet count (I@/&) Dynamic lung compliance (mWmm Hg) Static lung compliance (mL/mm Hg) Colloid osmotic pressure (mm Hg) Arterial pOz (mm Hg) Arterial pC0, (mm Hg) Hematocrit (%) Sample size (n)

Direct Infusion

Centrifugation

Ultrafiltration

61 f 2 84f3 1.99 f 0.04

60 f 2 85 f 3 2.00 f 0.04

59 f 3 88 f 3 2.06 f 0.03

0.62 f 0.04

59 f 3

62 f 5

397 f 44

395 f 32

408 f 35

287 f 18

259 f 14

303 f 21

* 1.6

43.6 f 3.1

40.7 f 2.1

49.7f 3.4

49.8 f 2.8

51.0 f 2.2

16.7 2 0.5

16.5 f 0.4

16.7 f 0.4

388 f 35

376 f 31

371 f 21

36.6 f 1.8

33.3 f 0.9

33.6 f 0.9

35.6 f 1.1 17

32.4 f 1.0 20

34.1 f 1.1 18

39.9

Data presented as mean f standard error of the mean. pC0, = carbon dioxide tension; PO, = oxygen tension.

a

Table 2 . Twenty Minutes After Circuit Blood Administration' Variable Plasma free hemoglobin (mg/dL) Colloid osmotic pressure (mm Hg) Arterial PO,(mm Hg) Arterial pC0, (mm Hg) Fibrinogen (mg/dL) Activated clotting time (s) Platelet count (l@/pL) Prothrombin time (s) Partial thromboplastin time ( s ) Thrombin time (s) Hematocrit (%) Sample size (n)

Direct Infusion Centrifugation Ultrafiltration

40 f 3

36 f 5

45 f 4

11.8 f 0.4 10.6 f 0.4

12.2 f O.Sb

278 f 30 332 f 27 36.2 f 1.5 34.8 f 1.4

302 f 24 34.5 2 1.3

196 f 16 105 f 4

191 f 16 111 f 5

248 2 21b 101 f 4

*

152 f 4 137 11 13.8 f 0.2 14.7 f 0.2' 30.0 f 1.6 36.2 ? 1.4'

197 f 23 13.7 f 0.3 30.6 f 1.5

25.4 f 0.8 43.5 f 8.2' 25.5 2 1.0 25.6 f 1.1 17 20

25.8 f 1.8 27.5 f 0.9 18

Data presented as mean & standard error of the mean. Ultrafiltration greater than centrifugation and direct infusion; p < 0.05. ' Centrifugation greater than ultrafiltration and direct infusion; p < 0.05. pC0, = carbon dioxide tension; PO, = oxygen tension.

a

Patient profile data and pre-cardiopulmonary bypass variables for the patients randomly assigned to the three different circuit blood infusion protocols are presented in Table 1. Patient age, weight, body surface area, and left ventricular ejection fraction were very similar among all three groups of patients. There was no significant difference in the blood plasma profile, hematocrit, or lung compliance tests among the three groups (p > 0.10), and there was no significant differences in the cross-clamp times between the direct infusion, centrifugation, or ultrafiltration groups (74 f 33, 59 f 26, 59 f 18 minutes, respectively; p = 0.128). The overall cardiopulmonary bypass time for the three groups was 92 31 minutes, and this also did not differ significantly among the three groups (p > 0.20). A summary of the data collected 20 minutes after the administration of circuit blood for the three groups is presented in Table 2. Significantlyincreased prothrombin and partial thromboplastin times were observed for the centrifugation group compared with the direct infusion or ultrafiltration groups (p < 0.05). In contrast, the mean platelet count was higher in the ultrafiltration group compared to either the direct infusion or centrifugation group (p < 0.05). Based on an analysis of covariance, a higher colloid osmotic pressure and fibrinogen concentration were observed in the ultrafiltration group compared to those observed for the other blood-salvaging methods (p < 0.05). A summary of the data collected 6 hours after cardiopulmonary bypass for the three treatment groups is presented in Table 3. Only two significant differences were observed among the three groups: the mean platelet

*

1993;5693843

Table 4 . Eighteen Hours After Circuit Blood Administration"

Table 3 . Six Hours After Circuit Blood Administration" Variable

941

SUTTONETAL CIRCUIT BLOOD SALVAGE METHODS

Ann Thorac Surg

Direct Infusion Centrifugation Ultrafiltration

Plasma free 36 f 5 hemoglobin (mg/dL) 14.2 f 0.4 Colloid osmotic pressure (mm Hg) 256 f 25 PO, (mm Hg) PC02 (mm Hg) 34.6 f 1.7 Dynamic lung 43.8 & 5.1 compliance (mL/mm Hg) Static lung compliance 49.6 & 3.8 (mL/mm Hg) Fibrinogen (mg/dL) 205 f 16 Activated clotting 104 f 5 time (s) Platelet count ( ~ O ~ / 1% ~ Lf ) 10 Prothrombin time (s) 13.2 f 0.2 Partial thromboplastin 29.0 f 1.5 time (s) Thrombin time (s) 29.4 f 4.1 Hemotocrit (%) 27.7 t 0.8 Sample size (n) 17

32

If: 4

30

f

2

13.2 f 0.4

14.4 f 0.5

319 f 29 34.1 f 1.2 42.1 f 4.4

320 f 23 33.6 f 1.7 33.8 f 1.7

f 3.3

47.3 f 2.4

15 f5

257 f 26 97 f 3

43.7 210 99

f

151 5 9 14.3 f 0.3' 33.7 f 2.5 30.7 28.4

f 4.9

f

1.1

199 f 20b 13.2 f 0.3 30.5 f 1.7 38.1 28.1

20

7.4 f 0.8 18 f

_ _ _ _ _ _ _ ~

32 f 4 Plasma free hemoglobin (mg/dL) Colloid osmotic 15.0 f 0.4 pressure (mm Hg) 252.8 f 24.5 PO, (mm Hg) 40.8 f 1.2 PC02 (mm Hg) Dynamic lung 42.6 f 8.0 compliance (mUmm Hg) Static lung compliance 50.2 f 4.0b (mL/mm Hg) Fibrinogen (mg/dL) 304 f 23 Activated clotting 102 f 3 time (s) Platelet count (l@/pL) 165 f 12 Prothrombin time (s) 12.5 f 0.2 Partial thromboplastin 26.9 f 0.8 time (s) 22.4 f 0.8 Thrombin time (s) Hematocrit (%) 28.3 f 0.8 Sample size (n) 17

32

f4

14.6 f 0.5

28

f4

15.3 f 0.4

323.0 f 31.8 323.3 f 30.09 36.7 f 1.9 38.5 f 2.0 33.6 ? 3.9 33.3 t 2.4 37.8

2.5

38.8 f 3.3

273 f 16 101 f 4

318 f 27 104 f 4

160 -t 10 13.2 f 0.2 30.8 f 1.F

179 f 13 12.6 f 0.3 27.2 f 1.2

f

21.7 f 1.1 23.0 f 2.5 29.1 f 1.1 28.2 f 0.6 20 18

Data presented as mean f standard error of the mean. Direct Infusion is greater than ultrafiltration and centrifugation; p < 0.05. Centrifugationgreater than direct infusion; p < 0.1.

a

Data presented as mean k standard error of the mean. Ultrafiltration is greater than centrifugation and direct infusion; p < 0.05. Centrifugationis greater than ultrafiltration and direct infusion; p < 0.01. a

pC0, = carbon dioxide tension;

Direct Infusion Centrifugation Ultrafiltration

Variable

PO, = oxygen tension.

count was higher for the ultrafiltration group and the prothrombin time was greater for the centrifugation group (both p < 0.05). At 18 hours after cardiopulmonary bypass, higher prothrombin and partial thromboplastin times were found in the centrifugation group (Table 4) ( p < 0.05).Analysis of covariance revealed higher static lung compliance in the direct infusion group compared with the centrifugation and ultrafiltration groups ( p < 0.05). No differences were observed for data collected in the intensive care unit in terms of the mean duration of chest tube drainage, the patients' time on the respirator, or the total usage of blood products (Table 5). There was no significant difference in the hematocrit at the time of hospital discharge for the three groups.

Comment A great deal of interest has been focused on the most appropriate and efficient means of blood conservation after cardiopulmonary bypass. In the present study, the effects of three different extracorporeal circuit bloodsalvaging methods were examined in patients with very similar preoperative characteristics. The most important findings from this study were: (1)centrifugation resulted in longer prothrombin and partial thromboplastin times, and (2) ultrafiltrationresulted in a higher post-cardiopulmonary bypass platelet count and colloid oncotic pressure. Immediately after cardiopulmonary bypass, the mean

pCOz = carbon dioxide tension;

PO, = oxygen tension.

hematocrit was higher using the ultrafiltration method. This was to be expected because ultrafiltration removes the volume and thereby concentrates the blood compo-

Table 5 . Postoperative Values After Different Blood Salvaging Methods" Variable

Dilutional

Time on respirator (h) 1-Hour chest tube drainage (mL) 6-Hour chest tube drainage (mL) Total chest tube drainage

25.5

Packed red blood cells administered (mL) Fresh frozen plasma administered (mL) Hematocrit (%) at discharge from hospital Sample size (n) a

Centrifugation

Ultrafiltration

19 f 2

18 f 1

146 f 22

155 f 19

180 f 25

323 f 43

411 f 52

380

1114 f 161

931 f 110

f5

385 f 74

87 t 36

31.8

Data presented as mean

f 0.6

17 k

546

f

115

120 f 44

32.7

f

1.1

20

standard error of the mean.

f 36

1024 f 109 325 f 72

90

32.7

f 47

f

18

1.9

942

SWITONETAL CIRCUIT BLOOD SALVAGE METHODS

nents. However, by 20 minutes after the infusion of circuit blood, there were no observable differences in the hematocrit values between any of the three groups. In addition, no differences in the hematocrit were noted at the subsequent data collection times. In a similar study conducted by Solem and colleagues [9], no differences in the hematocrit values were observed between ultrafiltration and centrifugation 4 hours after the administration of the circuit blood. In two separate studies carried out by Bolt and colleagues [lo, 131, similar findings were observed when centrifugation and ultrafiltration were compared. In a study performed by Zhou and associates [12], in which ultrafiltration and direct infusion were used, the hematocrit values were comparable to those in the present study. Twenty minutes after the circuit blood was administered, the fibrinogen concentration was higher in the ultrafiltration group. However, this significant difference did not persist at 6 and 18 hours postoperatively, although the p values for analysis of variance were 0.11 and 0.14, respectively. Solem and associates [9] observed no difference in the fibrinogen concentrations between ultrafiltration and centrifugation 4 hours after administration of the treated circuit blood. These past findings may not necessary conflict with those yielded by the present study because of the differences in the sample times. Bolt and colleagues [13], in comparing ultrafiltration to centrifugation, did not find a difference in the mean fibrinogen concentration at any time point chosen for measurement. However, using the data reported for this latter study, the calculated percent change from preoperative to postoperative values suggests a higher fibrinogen concentration on the first postoperative day. The platelet counts were significantly higher in the ultrafiltration group after the circuit blood was given, and this difference persisted at 6 hours postoperatively. Similar findings for ultrafiltration have been reported previously [13, 141. None of the patients in our study received platelets, thus the differences in the platelet counts among the three groups were not due to different infusion practices. The difference in the platelet count may stem from the discarding of platelets in the centrifugation group, and from dilution or blood loss in the direct infusion and centrifugation groups. Fluid relocation to the interstitial space in the ultrafiltration group may also explain the higher platelet count, but this seems unlikely because of the higher colloid osmotic pressure in this group. Because ultrafiltration concentrates but does not remove plasma proteins, the higher colloid oncotic pressure after administration of the circuit blood was to be expected. Interestingly, a persistent depression in the mean colloid oncotic pressure in the direct infusion group was not observed throughout the study period. A possible explanation for this is that blood remaining in the extracorporeal circuit and directly infused into the patient provides the same amount of proteins as that provided by ultrafiltration, but in a more dilute form. In a study conducted by Bolt and colleagues [lo], the colloid osmotic pressure was found to be higher when ultrafiltration was used. In a later study [13], these investigators reported a

Ann Thorac Surg 1993;5693843

similar difference in the plasma protein levels between an ultrafiltration group and a centrifugation group that persisted at 5 hours after cardiopulmonary bypass. Similarly, Zhou and associates [12] also reported a higher mean colloid osmotic pressure when ultrafiltration was used. To examine more closely the potential effects of differences in the colloid oncotic pressure among the three blood-salvaging methods, dynamic and static lung compliance were measured. At 18 hours postoperatively, the static lung compliance was significantly higher in the direct infusion group, but the reason for this finding remains unclear. However, examination of the data set for outliers in the direct infusion group did not reveal justifiable reasons to remove any measurements from the data set. There were no other differences in the lung compliance measurements taken throughout the study points. In addition, there were no differences in oxygen tension or carbon dioxide tension suggesting adequate gas exchange occurred with all three blood-salvaging methods. In a past study, Bolt and colleagues [lo], reported that ultrafiltration decreased the extravascular lung water content during cardiopulmonary bypass and lowered pulmonary vascular resistance. In the present study, only minor differences in the plasma oncotic pressures and lung compliance were observed among the three groups. However, results from past reports [lo, 12, 131 and from the present study do suggest that ultrafiltration may confer improved oncotic pressures and lung compliance, which may be an important consideration in patients with underlying pulmonary disease preoperatively. The mean prothrombin and partial thromboplastin times were longer in the centrifugation group compared to the other circuit blood administration methods. This may be due to a reduced fibrinogen content or the presence of circulating antithrombins such as heparin. Because the fibrinogen level was within normal limits, it would appear that higher amounts of circulating heparin remained in the centrifugation group. Giving additional protamine to the patients in the centrifugation group might have neutralized the circulating heparin, but this was not done in the present study, as the activated clotting times were not significantly different in these patients. The activated clotting time is a nonspecific test for the coagulation factors involved in fibrin formation, and is primarily used in monitoring heparin therapy during cardiopulmonary bypass. Thus, prolonged prothrombin and partial thromboplastin times in the centrifugation group compared with the ultrafiltration and direct infusion groups may not be prevented by conventional protamine reversal and activated clotting time calculations. The components of the cardiopulmonary bypass circuit can contribute to hemolysis when the end-product is plasma-free hemoglobin. The major sources of hemolysis during cardiopulmonary bypass are the shear stresses and turbulent flow arising primarily from roller pump suction. The plasma-free hemoglobin concentrations were not different among any of the groups at any time point when measured. Similar to the findings in the present study, Bolt and colleagues [13] found no differences in the

SUITONETAL CIRCUIT BLOOD SALVAGE METHODS

Ann Thorac Surg

1993;56938-43

plasma-free hemoglobin concentration. In contrast, Nakamura and associates [14] found higher plasma-free hemoglobin concentrations when ultrafiltration was used. These investigators concluded that ultrafiltration should not be used in patients with impaired renal function. Based on these past findings and those from the present study, we suggest that ultrafiltration is the best choice for cardiopulmonary bypass circuit blood management. However, if the plasma-free hemoglobin concentration is unusually high due to bank blood usage, prolonged cardiopulmonary bypass times, or the continued use of high suction pressures, then it may be appropriate to consider alternative blood-salvaging methods. The duration of chest tube drainage in the intensive care unit and the amount of fresh frozen plasma administered were not significantly different for any of the three circuit blood-salvaging methods. Moran and colleagues [l] reported decreased blood loss and bank blood usage associated with centrifugation use. However, other studies [13, 141 revealed no differences in either blood loss or bank blood usage in association with various bloodsalvaging methods. In the present study, analysis of covariance was performed on all intensive care values. Although it did not reach significance (p = 0.08), packed red blood cell administration was higher in the centrifugation group. However, packed red blood cell usage in the intensive care unit represents clinical decisions that are based on a wide variety of patient variables. Bank blood usage is decided at the attending physician’s discretion, and may be affected by blood loss before admission into the intensive care unit and the patient‘s general coagulation status at chest closing, neither being assessed. Thus, the differences in banked blood usage in this study and past studies may be due simply to inherent differences in infusion practices. Finally, with increasing attention to the rising cost of health care, it is appropriate to consider the relative cost of each of the circuit blood-salvaging methods. The direct infusion method is the least expensive, in that transfusion bags cost approximately $3.00 each. The ultrafiltration method requires an ultrafiltrator as well as the container system, and at our institution costs approximately $85.00 per patient. Finally, the centrifugation system has an initial up-front cost of approximately $50,000, after which the cost for centrifugation disposables is approximately $75.00 per patient. Accordingly, the most cost-effective approach for returning cardiopulmonary bypass circuit blood to the patient is direct transfusion. However, the hemodynamic considerations and coagulation characteristics of each patient ultimately dictate which circuit salvage method is most appropriate. In summary, in a sample of adult patients undergoing coronary artery bypass grafting with cardiopulmonary

943

bypass, the ultrafiltration of circuit blood resulted in improved colloid osmotic pressures, fibrinogen concentrations, and platelet concentrations in the early postcardiopulmonary bypass period. However, no clinically significant differences were observed among direct infusion, centrifugation, or ultrafiltration at longer postoperative intervals. Interestingly, there were no apparent differences between the direct infusion method and either the centrifugation or ultrafiltration method. Findings from the present study suggest that direct infusion, centrifugation, and ultrafiltration all continue to be appropriate methods for extracorporeal circuit blood salvage.

References 1. Moran JM, Babka R, Silberman S, et al. Immediate centrifugation of oxygenator contents after cardiopulmonary bypass. J Thorac Cardiovasc Surg 1978;76:510-7. 2. Darup J, Bleese N, Kalmar P, et al. Hernofiltration during extracorporeal circulation. Thorac Cardiovasc Sure; 1979;27 227-30. 3. Henderson LW, Besarab A, Michaels A, Bluemle LW Jr. Blood purification by ultrafiltration and fluid replacement (diafiltration). Trans Am SOCArtif Internal Organs 1967;13: 216-26. 4. Klinesberg PL, Kam CA, Johnson DC, Cartmill TB, Brown JH. Hematocrit and blood volume control during cardiopulmonary bypass with the use of hemoconcentration. Anesthesiology 1984;6047840. 5. Osipov VP, Lurie GO, Khodes MY, Mikhailov Y, Fadejeva NV. Hernoconcentrationduring open heart operations. Thorac Cardiovasc Surg 1985;33:815. 6. Magilligan DJ, Oyama C. Ultrafiltration during cardiopulmonary bypass: laboratory evaluation and initial clinical experience. Ann Thorac Surg 19&1;3733-9. 7. Hopeck JM, Lane RS, Schroeder JW. Oxygenator volume control by parallel ultrafiltration to remove plasma water. J Extracorp Technol 1981;13:267-71. 8. Hoeck JM, Kalshoven JD, Wilds SL, Dearing JP. A comparison of two methods of post-bypass hemocbncentration. J Extracorp Technol 1982;14431-6. 9. Solem JO, Tengborn L, Steen S, Luhrs C. Cell saver versus hemofiltrator for concentration of oxygenator blood after cardiopulmonary bypass. Thorac Cardiovasc Surg 1987;35 42-7. 10. Boldt J, Kling D, von Bormann B, Scheld HH, Hempelmann G. Extravascular lung water and hemofiltration during complicated cardiac surgery. Thorac Cardiovasc Surg 1987;35: 161-5. 11. Zhou JL, Guan HP. Ultrafiltration and two-hemodilution during excessive hemodilution cardiopulmonary bypass. Chin Med J 1988;101:807-10. 12. Zhou JL, Gong QC, Guan HP. Effect of ultrafiltration on plasma colloid oncotic pressure and red cell volume during cardiopulmonary bypass. J Cardiovasc Surg 1989;30:40-1. 13. Boldt J, Kling D, von Bormann B, Zuge M, Scheld H, Hempelmann G. Blood conservation in cardiac operations: cell separation versus hemofiltration. J Thorac Cardiovasc Surg 1989;97832-40. 14. Nakamura Y, Masuda M, Toshima Y, et al. Comparative study of Cell Saver and ultrafiltration nontransfusion in cardiac surgery. Ann Thorac Surg 1990;49973-8.

-