Apparent coagulopathy caused by infusion of shed mediastinal blood and its prevention by washing of the infusate

Apparent Coagulopathy Caused by Iinfusion of Shed Mediastinal Blood and Its Prevention by Washing of the Infusate Lee D. Griffith, MD, Glenn F. Billman, MD, Pat 0. Daily, MD, and Thomas A. Lane, MD Departments of Cardiothoracic Surgery and Pathology, University of California, San Diego; Department of Pathology, Children’s Hospital; and Blood Bank, Veterans Administration Medical Center, San Diego, California

We found that reinfusion of shed mediastinal blood (SMB) after a cardiac operation was associated with laboratory evidence of disseminated intravascular coagulation. In view of this, we compared the effect of infusing washed or unwashed SMB on the coagulation profiles and blood use of two serial groups of patients undergoing cardiopulmonary bypass. We found that the results of testing for fibrin degradation products converted from negative to positive in 17 of 20 patients who received unwashed SMB versus 1 of 14 patients who received washed SMB ( p < 0.0001). Other coagulation studies did not reveal disseminated intravascular coagulation in either group, nor were there differences in blood use between the two groups. The unwashed SMB

contained high titers; of fibrin degradation products (mean reciprocal titer = 354 f 161) compared with washed SMB (mean reciprocal titer = 34 2 18) ( p < 0.01). Based on the volume of SMB infused, the amount of fibrin degradation products in unwashed SMB was sufficient to account for the positive fibrin degradation product assays after infusion in this group. We conclude that infusion of unwashed SMB may confuse the interpretation of tests for diisseminated intravascular coagulation or fibrinolysis. 11s this could lead to unnecessary blood component use and is preventable by washing before infusion, we recommend that the routine infusion of unwashed SMB not longer be employed. (Ann Thorac Surg 1989;47:400-6)

H

ascertain if the coagulopathy could be prevented by washing the SMB before reinfusion.

omologous blood transfusion therapy entails substantial risks, including transfusion reactions, alloimmunization, and transmission of diseases such as hepatitis, cytomegalovirus, and acquired immunodeficiency syndrome due to human immunodeficiency virus. Cardiac surgical patients comprise a high-risk group in that they often require transfusion during hospitalization. Schaff and colleagues [ 11 reported an average homologous blood usage of 8.4 units per adult patient undergoing coronary artery bypass grafting in 1976. Since that time, autologous transfusion techniques, which include perioperative phlebotomy with hemodilution, salvaging of intraoperative blood loss by cell washers, and postoperative autotransfusion of shed mediastinal blood (SMB), have reduced transfusion requirements by 26% to 50% [2, 31. Early reports [4, 51 on the reinfusion of SMB warned of coagulopathies, but others [6] have demonstrated no adverse effects on the control of hemostasis. At our institution, evaluation of several patients who received SMB in the postoperative period revealed several in whom abnormal coagulation profiles developed. The association between the apparent coagulopathy and the administration of SMB stimulated us to investigate the effects of this product on the coagulation system and to Accepted for publication Sep 20, 1988 Presented in part at the Twenty-first Annual Meeting of the American College of Chest Physicians, Atlanta, GA, Oct 26-30, 1987. Address reprint requests to Dr Griffith, 2090 Redwood Dr, Santa Cruz, CA 95060.

0 1989 by The Society of Thoracic Surgeons

Material and Methods Two nonrandomized groups of patients were studied serially during the 17-month period between April 1985 and September 1986. Group 1 (unwashed SMB group) included 21 patients who received unmodified SMB collected by means of the Sorenson drainage system (Sorenson Research Corp, Salt Lake City, UT) with reinfusion through a 40-pm blood filter without further processing. Group 2 (washed infusate group) included 17 patients who similarly had the SMB collected through the Sorenson drainage system, but before reinfusion, the SMB was washed in a Haemonetics Cell Saver model CSII (Haemonetics Corp, Braintree, MA) using 1,000 mL of normal saline solution. The data base included the operation performed, the length of cardiopulmonary bypass (CPB), the total chest tube output at 24 hours, average hourly chest tube output before and after reinfusion of unwashed or washed SMB, and the volume of :shed blood infused. Hematological data included blood count, prothrombin time (PT), activated partial thromboplastin time (PTT), platelet count, fibrinogen level, protamine paracoagulation phenomenon, and analysis for fibrin degradation products (FDP). The FDP were measured by the Thrombo-Wellco latex agglutination method. (Burroughs Wellcome Co, Research Triangle Park, NC), which has a sensitivity of 2 pg/mL for 0003-4975/89/$3.50

Ann Thorac Surg 1989;47400-6

Table 1 . Comparison of Patient Groups

Variable Operation CABG Valve replacement Valve replacement + CABG Reoperation for CAD No. of vessels bypassed/pa tient Preoperative hernatocrit (76) CPB time (min) Chest tube drainage (totaV24 h) a

401

GRIFFITH ET AL AUTOTRANSFUSION OF SHED MEDIASTINAL BLOOD

Group 1 (Unwashed SMB) (n = 21)

Group 2 (Washed SMB) (n = 17)

17 3 0 1 4.5 ? 2.2

4.7 c 1.1

42 f 3 94 f 38 1,037 357

41 f 5 128 f 34” 1,380 ? 828

11

4 1 1

Significance: p < 0.005 compared with group 1.

CABG = coronary artery bypass grafting; disease; CPB = cardiopulmonary bypass; blood.

CAD = coronary artery SMB = shed mediastinal

FDP, chiefly the D and E fragments [7]. As normal serum values of FDP are less than 10 Pg/mL, normal serum diluted 1:5 gives a negative result. The patients’ sera were tested at two standard dilutions of 1:5 and 1:20, and the results were reported as either positive or negative at these concentrations. These data were collected on the patient’s arrival in the surgical intensive care unit and again one half hour after infusion of either unwashed or washed SMB. Patients were not transfused with other blood products between blood sampling before and after infusion of SMB. Blood counts were also obtained preoperatively and 24 hours postoperatively. The hematological analyses were performed in the Veterans Administration Medical Center Department of Laboratory Services using standard techniques. Samples of the unwashed or washed SMB were obtained before infusion into the patients, and were analyzed for hematocrit, free hemoglobin, and the titer of FDP. For ease of interpretation of the FDP assay, the reciprocal of the highest dilution (titer) giving a positive result is reported. Total blood product usage was tallied for the entire hospital stay. Mathematical analysis employed unpaired t testing except for comparisons between the preinfusion and postinfusion data sets of patients who received either unwashed or washed products, in which case the paired t test was used. McNemar’s test for matched pairs with a dichotomous outcome was utilized for the FDP assays, and significant differences were assumed to exist when the probability of a random occurrence was less than 5% ( p < 0.05). Data in the tables are presented as the mean & one standard deviation unless otherwise stated.

operation, or preoperative hematocrit. On the other hand, the duration of CPB was significantly longer in the group subsequently given washed SMB (group 2). There was no correlation between the duration of CPB and the clinical evaluation of bleeding, and the two groups had similar amounts of total chest tube drainage after operation. A mean volume of 405 -1- 130 mL of SMB was collected from and given to the group that received unwashed SMB (group 1)and a mean volume of 537 173 mL of SMB was collected in group 2 ( p < 0.01 versus group 1). After processing, a mean of 240 2 50 mL of washed SMB was given. The physical characteristics of unwashed and washed SMB are shown in Table 2. The SMB did not clot in any of the samples, and a small subset was tested to confirm nonclotting PT and PTT assays (n = 3 each), low fibrinogen levels (less than 50 mg/dL; n = 3), and decreased factor VIII levels (17.5% 2 1%;n = 4). The platelet count was also low (37,000 21,000 mL; n = 6). The entire series was not tested because results are well described in the literature [3, 61, our data correspond well to the previously reported results, and the focus of our study was on other factors. As expected, the hematocrit of the washed infusate was higher than that of the unwashed SMB because of red cell concentration during the washing process. In addition, both the free hemoglobin level and titer of FDP were significantly lower in washed SMB (see Table 2). Extended serum dilutions were performed to ascertain the concentration of FDP in the unwashed infusate. The FDP assays 161 were positive in SMB at reciprocal titers of 354 (median titer = 300). These values were significantly reduced to 34 2 18 (median titer = 40) by washing. The effects of an operation with CPB and infusion of unwashed versus washed SMB on chest tube drainage and laboratory tests of hemostasis are shown in Table 3. No significant difference was observed between the preoperative and postoperative values of hematocrit between the two groups or in the amount of chest tube drainage before or after infusion of SMB. There were also no significant differences between groups for initial values of PT, PTT, platelets, or fibrinogen measured on arrival in the intensive care unit. After infusion of unwashed SMB, however, there was a significant decrease in PT compared with the corresponding value before infusion ( p < 0.001).

*

*

*

Table 2 . Comparison of Unwashed and Washed Shed Mediastinal Blood

Variable

Group 1 (Unwashed SMB)

Hematocrit (%) Free hemoglobin (mg/100 mL) FDP titerb Volume SMB processed (mL) Volume SMB infused (mL)

25 298 354 405 405

9 164 2 161 2 130 2 130 -t

?

Group 2 (Washed SMB) 39 2 15“ 112 f 77“ 34 2 18” 537 2 173” 240 2 50 ~

Results

a

AS shown in Table 1, there were no differences between the two patient groups with respect to age, type of

FDP

~~

Significance: p < 0.01 compared with group 1. The FDP titer was the reciprocal value of highest dilution of sample resulting in a positive test for FDP (see Material and Methods). =

fibrin degradation products;

SMB

=

shed mediastinal blood.

402

GRIFFITH ET AL AUTOTRANSFUSION OF SHED MEDIASTINAL BLOOD

Ann Thorac Surg 1989;4740&6

Table 3. Comparison of Hematological and Hemostatic Data Before and After Infusion of Unwashed and Washed Shed Mediastinal Blood Group 1 (Unwashed SMB) Variable

Chest tube output (mL/h) Hematocrit (%) Platelets ( x i03/pL) PT (ratio)” PTT (ratio)”

Fibrinogen (mg/dL) Protamine paracoagulation phenomenon‘ FDP‘

-

Before

After

159 2 59

43 f 20

42 f 3 122 f 38 1.23 f 0.07 1.22 f 0.19

27 C 4 110 f 33 1.13 5 O . l l b 1.30 f 0.35

183 ? 43 0/20

195 f 67 2/20

2/20

Group 2 (Washed SMB)

19/20d

Before 203

?

After

100

62

5

49

41 2 5 126 f 54 1.23 2 0.12 1.30 2 0.40

27 f 4 118 f 42 1.21 C 0.16 1.48 f 0.53

205 t 64 0/14

217 t 84 0114

5/14

5/14

This ratio is expressed as the ratio of the patient‘s value over the control value. Significance: p < 0.001 compared with value obtained before infusion. Data are reported as the number of patients testing positive over the total number tested. Significance: p < 0.0001 compared with value obtained before infusion.

a

FDP

=

fibrin degradation products;

PT = prothrombin time;

FTT = activated partial thromboplastin time;

Complete preinfusion and postinfusion data on FDP were available for 20 of the 21 patients in group 1 (unwashed SMB) and 14 of the 17 in group 2 (washed SMB) (see Table 3). In group 1, 2 of 20 patients had positive FDP tests before infusion of SMB (both were positive only at 1:5 dilutions). As shown in Figure lA, after infusion of unwashed SMB, 19 of 20 patients either had increases in FDP titer or converted from negative to positive FDP assays (10 of the patients with a previously negative assay had a positive assay at 1:5 dilution, 7 had a positive assay at 1:20 dilution, and the 2 who had a positive assay at 1:5 dilution converted to a positive assay at 1:20). Thus, in only 1 patient who received unwashed SMB did the assay remain negative for FDP. In comparison, 5 of 14 patients in group 2 had positive tests for FDP (at 1:5 dilution) before infusim. After the washed product was infused, only 1 patient whose assay was previously negative had a positive result (at a 1:5 dilution only), and 1 patient whose assay was positive for FDP before infusion converted to a negative FDP assay (Fig 1B). Analysis using McNemar’s test revealed a highly significant difference (2= 15.06, p < 0.0001) between the preinfusion and postinfusion results of tests for FDP in group 1 but not group 2 (2 = 0.50, p < 0.48). Table 4 shows predicted FDP assay results after autotransfusion of either unwashed or washed SMB in a hypothetical man weighing 70 kg. Infusion of unwashed SMB at a mean collection volume of 405 mL with an FDP titer positive at a mean serum dilution of 1:354 would give positive results in the FDP assay at both 1:5 and 1 :20 test dilutions. In contrast, infusion of a mean volume of 240 mL of washed SMB would not be expected to give rise to a positive test for FDP. Blood product administration is listed in Table 5. There were no significant differences between groups in the transfusion of packed red cells, fresh frozen plasma, or platelets. In the group receiving the washed infusate, 2 patients had preoperative hematocrits of 33% and 34%;

SMB = shed mediastinal blood.

these patients received 5 of the 17 units of blood used postoperatively. In the unwashed infusate group, only 1 patient had a preoperative hematocrit of 34%, and he received 2 of the 8 units of packed red blood cells in the postoperative period. The mean red cell usage for the combined groups was 0.7 unit per patient. Fresh frozen plasma was used in 4 patients in each group (mean, 0.5 unit per patient overall). Platelets were utilized in no patients and 3 patients receiving unwashed and washed SMB, respectively (mean, 0.5 unit per patient). Transfusion was avoided entirely in 26 (68%)of the 38 patients.

Comment Autotransfusion grew from the need to have a ready source of blood for infusion into rapidly hemorrhaging

Table 4. Predicted Fibrin Degradation Product Assay Results After Autotransfusion o,E Shed Mediastinal Blood“ Variable Volume infused (mL) FDP titer (reciprocal) Total FDP infused (pg) Plasma volume (mL) Predicted FDP (pg/mL) Predicted FDP in assayb 1 :5 Serum dilution 1 :20 Serum dilution

Group 1 (Unwashed SMB) 405 354 286,700 3,800 75 15 (+) 3.7 (+)

Group 2 (Washed SMB) 240 34 16,300 3,800 4.3

0.8 (-) 0.2 (-)

The model for this table was a patient weighing 70 kg with a postoperNumbers are predicted FDP in micrograms ative hematocrit of 27%. per milliliter after indicated serum dilution for FDP assay. Signs in parentheses are expected results of FDP assay with associated concentration of FDP. The FDP assay has a sensitivity of 2 &mL. SMB = shed mediastinal blood. FDP = fibrin degradation products; a

Ann Thorac Surg 1989;47400-6

POS NEG

1 yxH;

I kzJ; ,

XX2=15.06 2 = 15.06 p < 0.0001 0.0001

000 00

mw 1 :5

1:20

Serum FDP Titers

A

1 :5

1 :20

Serum FDP Titers

B Fig 1 . (A)Effect on serum fibrin degradation product fFDP) titers of infusion of unwashed shed mediastinal blood fSMB). Each symbol represents I study patient. The symbols are located in the quadrant that represents the serum FDP titer of the patient before infusion of the unwashed SMB. Arrows are drawn from blocks of patients whose serum titer of FDP changed after the infusion. The number of patients and the change in titer are indicated by the quadrant in which the point of the arrow is located. Note that the FDP titer of only I patient remained unchanged after infusion of unwashed SMB. (B) Effect on serum FDP titers of infusion of washed SMB. The symbols are the same as in (A). After infusion, only 2 patients changed FDP titer. One patient with a previously negative fNEG) assay for FDP had a positive fPOS) result, and 1 with a previously positive assay had a negative result. fNSD = no significant difference.)

patients. Blundell [S] has been credited with the first literature report, in which he described autotransfusion in 10 women who had severe hemodynamic decompensation following postpartum hemorrhage. Except for sporadic reports in the literature, this treatment was overlooked until Klebanoff and Watkins (91 revised the technique for trauma victims during the Vietnam War. Citrate anticoagulant was added to a cardiotomy reservoir for collection of blood, which was reinfused rapidly with a roller pump. Because there were dramatic successes and very few side effects, the Bentley Corporation (Bentley Laboratories, Inc, Irvine, CA) marketed this device with some modifications as the first commercially available autotransfuser. The Bentley device was withdrawn from the market after several deaths occurred due to infusion of massive amounts of air from the pressurized cardiotomy chamber [6]. Several devices with modifications of this basic principle are still in use at a variety of institutions for trauma and vascular surgery (10, 111. Building on the successes of these devices, Symbas [12] began experiments and clinical studies on the autotransfusion of shed pleural blood obtained during chest trauma, his reasoning being that this blood was a readily

GRIFFITH ET AL AUTOTRANSFUSION OF SHED MEDIASTINAL BLOOD

403

available source of warm, compatible blood. He initially collected the blood in chest bottles with citrate anticoagulant, but after early experiments confirmed the blood was defibrinated and unclottable, he collected the blood in bottles with normal saline solution. However, a recent report by his group [13] stated that in rapid clinical bleeding, citrate is still added in the event the blood is not fully defibrinated. The gynecological specialty provides clinical cases of rapid hemorrhage from ruptured ectopic pregnancies. Carty and co-workers [14] analyzed peritoneal shed blood and found it to be depleted of fibrinogen and factors V, VIII, and X. Also, it was low in platelets and had high titers of FDP. As ruptured ectopic pregnancies can lead to disseminated intravascular coagulation (DIC) and high FDP levels, they conjectured that autotransfusion of peritoneal blood obtained during operation could precipitate this problem, and recommended washing the shed blood to avoid the possibility of DIC and platelet dysfunction. Reinfusion of SMB for cardiac surgical patients was a logical extension of the literature base built from experiences in trauma, vascular, and gynecological surgery. Schaff and colleagues [3] analyzed the constituents of SMB and found that this blood contained almost normal levels of factors VIII and IX, moderate platelet counts, and elevated levels of free hemoglobin and FDP. Comparing these findings with those in banked blood, they discovered the latter was acidotic, contained low levels of factors VIII and IX, and had minimal platelets and somewhat elevated levels of free hemoglobin. Schaff and co-workers concluded that the SMB was at least as good as banked blood. Clinical studies of SMB demonstrated no apparent hematological side effects and no difference in clinical bleeding between the groups receiving SMB or banked blood.

Table 5 . Record of Blood Product Utilization

Product Red blood cells Total units Patients tranfused Unitdpatient transfused Units/total patients Fresh frozen plasma Total units Patients transfused Unitslpatient transfused Units/total patients Platelets Total units Patients transfused Unitdpatient transfused Unitdtotal patients SMB = shed mediastinal blood.

Group 1 (Unwashed SMB) (n = 21)

Group 2 (Washed Overall SMB) (n = 17) (n = 38)

8 5 (24%) 1.6 0.4

17 7 (41%) 2.4 1.0

0.7

8 4 (19%) 2.0 0.2

12 4 (24%) 3.0 0.7

0.5

18 3 (18%) 6.0 1.1

0.5

404

Ann Thorac Surg 1989;4740&6

GRIFFITH ET AL AUTOTRANSFUSION OF SHED MEDIASTINAL BLOOD

Table 6 . Comparison of Postoperative Blood Loss and Homologous Transfusion Requirements for Adult Cardiac Surgical Patients

Reference

Year

Institution

Patients

Mean Blood Loss" (mL/patient)

Morgan et a1 [25] Kaplan et a1 [26] Moran et a1 [27]

1972 1977 1978 1979 1979 1983 1989

Peter Bent Brigham Hospital Emory University Hospital Loyola University (Chicago) Johns Hopkins Hospital Cleveland Clinic Massachusetts General Hospital University of California, San

50 60 25 271 59 168 38

1,055 ? 90 825 f 84 1,118 2 271 912 t 150 1,258 1,160 ? 89 1,174 2 98

Schaff et al [l]

Thurer et al [24] Johnson et a1 [23] Griffith et al

[present report]

a

Data are shown as the mean

No. of

Homologous Blood" (mL/patient) 3,500 ? 220 2,346 ? 232 2,175 ? 175 1,290 ? 91 707 447 f 60 300 t 270

Diego ?

the standard error of the mean

Bell [15] thought that SMB was of benefit over banked blood because of the presence of factors VIII and IX along with moderate numbers of platelets. However, recent work by Symbas's group [13] involving shed pleural blood showed marked decreases in the levels of factors VIII and V, moderate reductions in factors VII, X, XI, and XII, almost total absence of platelets and fibrinogen, and high levels of FDP. Only the factor IX level was near normal. These findings are remarkably similar to the findings of Carty and co-workers [14] in peritoneal blood. In addition, several authors [16, 171 reported that platelets obtained from autotransfusate do not aggregate normally to the usual stimuli of adenosine diphosphate, epinephrine, or collagen. Furthermore, radioactive serotonin studies and measurements of P-thromboglobulin levels showed that platelet release had taken place. This may result in platelet dysfunction (17, 181. Together these data indicate that there are few platelets remaining in SMB and that they do not have normal function. This study was prompted by a major difference between our experience with infusion of SMB and the data reported by Schaff and associates [3]. We found that infusion of unwashed SMB in cardiac surgical patients was routinely associated with the development of a positive test for FDP, thus raising the possibility of DIC. Consequently, one of the chief reasons we performed this study was to investigate the effect of the infusion of SMB on hemostasis; specifically, to ascertain whether the infusion of unwashed SMB was associated with DIC. Taken together, our clinical and laboratory data do not suggest the occurrence of DIC in these patients. This conclusion is supported by several lines of evidence, including the lack of a significant decrease in postinfusion fibrinogen or platelets, failure of the PT or PTT to become prolonged, and the negative protamine paracoagulation phenomenon in this group. In addition, the unwashed SMB group did not have increased bleeding after infusion. On the other hand, in nearly all patients who received unwashed SMB, tests became positive for FDP. Thus, our studies strongly suggest that the positive FDP assays in these patients were not due to DIC, but were a direct result of the infusion of high titers of FDP in the unwashed SMB. This is supported by our calculations of expected FDP

titers, based on the known levels of FDP in the infusates (see Tables 2, 4). Washing the SMB both reduced the high levels of FDP (and free hemoglobin) and prevented the conversion of serum FDP assays after autotransfusion. The induction of high levels of FDP in patients who receive unwashed SM13 may be clinically significant. First, elevated levels of FDI' may be additive to the effects of CPB in producing platelet dysfunction and may increase the likelihood of a bleeding diathesis. Levels of FDP greater than 25 pg/mL have been associated with serious clinical bleeding [19], possibly because of inhibition of platelet adhesiveness to connective tissue. These clinical findings are accompanied by laboratory evidence of decreased platelet aggregation in the presence of FDP by thrombin, adenosine diphosphate, and collagen stimuli [20, 211. In this study, serum levels of FDP after transfusion of unwashed Sh4B were as high as 75 pg/mL (see Table 4), which is well above the level at which platelet dysfunction has been reported. Second, FDP may impair fibrin polymerization [22]. We were unable to demonstrate a clinical problem of increased bleeding from these theoretical disadvantages of infusing high titeris of FDP. However, differences in blood loss or the use of blood products could easily be obscured by the significantly longer CPB time of the group receiving washed shed mediastinal blood, the fact that shed blood was not indexed to body weight, and the relatively small sample numbers. Although the differences in CPB time and the trend toward higher red cell use in this group may be in part due to a change in the surgical team, which included only the addition of a more inexperienced cardiotlnoracic surgical resident at the time studies in this group were undertaken, all other aspects of the study including operative technique, sample collection, and SMB washing were performed by one of us (L.D.G.) to decrease variability between groups. Differences in clinical bleeding are difficult to demonstrate in any study because of patient variability. Table 6 adds our data regarding blood loss and replacement to a previously published table by Schaff and co-workers [l] and the experience of Johnson and associates [23] and Thurer and colleagues [24]. The mean blood loss in our patients was 1,200 mL per 24 hours, which compares

Ann Thorac Surg 1989;47:40&6

favorably with that of the other groups. The amount of homologous blood transfused postoperatively has steadily decreased over the years from 3,500 mL per patient to our unusually low value of 300 mL per patient in this small number of study patients (our value includes all blood components). Part of the reason for this decrease is an awareness of the hazards of blood transfusion and the acceptance of lower postoperative hematocrits. In the past, some institutions [3] routinely maintained hematocrits at 35% in the postoperative period; now, the majority accept 30%. We allow the hematocrit to reach 22% without transfusion in patients without infarction, neurological problems, or congestive heart failure. The success or failure of postoperative blood salvage should be judged largely on its influence on the use of homologous blood. Blood use in our patients, as a group, was very low (due in part to the routine use of intraoperative blood salvage), consequently making it more difficult to demonstrate differences in blood use. In this study, the mean equivalent of a blood unit we infused from SMB (and thus the amount of homologous transfusion avoided) was approximately 140 mL of packed red cells, which is the equivalent of slightly more than 0.5 unit of blood. The maximum amount of homologous blood saved in any patient was approximately 1.5 units (3 patients). In reviewing our administration of other blood products, platelet usage was found to be appropriate in the 3 patients to whom platelets were administered. It is also possible that the fresh frozen plasma could have been eliminated as recommended by Collins 1281, as PTs are prolonged in the majority of patients after bypass and are not easily remedied by infusion of this product. The observed shortening of the PT after infusion in the group receiving unwashed SMB is not readily explainable in this study and does not fit with previously published results [6]. Further observations will be needed to determine if this is a reproducible difference. In summary, this study compares some hematological effects of infusing washed or unwashed SMB into patients in the early postoperative period after CPB. The infusate was analyzed for hematocrit, FDP titers, and free hemoglobin. No significant differences were noted between groups in levels of fibrinogen, platelet counts, PT, PTT, or hematocrit at 24 hours except for an unexplained significant decrease in the postinfusion PT in the group receiving unwashed infusate. This study, which involved relatively small numbers of patients, did not document an increase in either DIC or clinical bleeding in patients receiving unwashed SMB. It was clear, however, that infusion of this product produced elevated serum levels of FDP. The high titers of FDP in the unwashed SMB consistently converted the serum test for FDP to positive, whereas washing the SMB significantly decreased the FDP titers and prevented the conversion of the serum FDP test to positive. In a bleeding patient, the conversion of serum tests for FDP from negative to positive from infusion of unwashed SMB can confuse subsequent laboratory tests for DIC and might delay appropriate therapy or lead to empirical blood component transfusions. This abnormality and a review of the literature suggest

CRlFFlTH ET AL AUTOTRANSFUSION OF SHED MEDIASTINAL BLOOD

405

that potential but yet unproven hazards from the autotransfusion of unwashed SMB may outweigh the benefits. Until more specific tests are routinely available to rule out DIC, we recommend that strong consideration be given to washing SMB. Supported by the Veterans Administration.

References 1. Schaff HV, Hauer J, Gardner TJ, et al. Routine use of autotransfusion following cardiac surgery: experience in 700 patients. Ann Thorac Surg 1979;27:49>9. 2 . Hallowell P, Bland JHL, Buckley MJ, et al. Transfusion of fresh autologous blood in open-heart surgery. J Thorac Cardiovasc Surg 1972;64:941. 3. Schaff HV, Hauer JM, Bell WR, et al. Autotransfusion of shed mediastinal blood after cardiac surgery. J Thorac Cardiovasc Surg 1978;75:632. 4. Duncan SE, Edwards WH, Dale WA. Caution regarding autotransfusion. Surgery 1974;76:1024. 5. Stillman RM, Wrezlewicz WW, Stanczewski B, et al. The haematological hazards of autotransfusion. Br J Surg 1976;63: 651. 6. Thurer RL, Hauer JM. Autotransfusion and blood conservation. Curr Probl Surg 1982;19:121. 7. Drewinko B, Surgeon J, Cobb P, et al. Comparative sensitivity of different methods to detect and quantify circulating fibrinogedfibrin split products. Am J Clin Path01 1985;84:58. 8. Blundell J. Experiments on the transfusion of blood by the syringe. Med-Chir Trans 1818;9:56. 9. Klebanoff G, Watkins D. A disposable autotransfusion unit. Am J Surg 1968;116:475. 10. Glover JL, Broadie TA. Intraoperative autotransfusion. World J Surg 1987;11:60. 11. Toomasian JM, Schneiderman G, DeSmet GM, Bartlett RH. Evaluation of a new blood autotransfusion device. J Thorac Cardiovasc Surg 1986;92:936. Autotransfusion from hemothorax: experimen12. Symbas ."I tal and clinical studies. J Trauma 1972;12:689. 13. Napoli VM, Symbas PJ, Vroon DH, Symbas PN. Autotransfusion from experimental hemothorax: levels of coagulation factors. J Trauma 1987;27:296. 14. Carty MJ, Barr RD, Ouna N. The coagulation and fibrinolytic properties of peritoneal and venous blood in patients with ruptured ectopic pregnancy. J Obstet Gynaecol Br Comm 1973;80:701. 15. Bell WR. Hematologic aspects of autotransfusion. In: Hauer JM, Thurer RL, Dawson RB, eds. Autotransfusion. New York: Elsevier/North-Holland, 1981:6. 16. Moore EE, Dunn EL, Breslich DJ, et al. Platelet abnormalities associated with massive autotransfusion. J Trauma 1980;20: 1052. 17. Raines J, Buth J, Brewster DC, Darling RC. lntraoperative autotransfusion: equipment, protocols, and guidelines. J Trauma 1976;16:616. 18. Harker LA, Malpass TW, Branson HE, et al. Mechanism of abnormal bleeding in patients undergoing cardiopulmonary bypass: acquired transient platelet dysfunction associated with selective alpha granule release. Blood 1980;55:824. 19. Marder VJ, Machett MO, Sherry S. Detection of serum fibrinogen and fibrin degradation products. Comparison of six techniques using purified products and application in clinical studies. Am J Med 1971;51:71. 20. Barnhart MI, Cress RL, Riddle SM. Influence of fibrinogen

406

21.

22. 23.

24.

Ann Thorac Surg 1989;4740&6

GRIFFITH ET AL AUTOTRANSFUSION OF SHED MEDIASTINAL BLOOD

split products on platelets. Thromb Diath Haemorrh 1967;17: 78. Kopec M, Budzynski SS, Wegrzyrowicz Z, et al. Studies on the mechanism of interference by fibrinogen degradation products with platelet function: role of fibrinogen in the platelet atmosphere. Thromb Diath Haemorrh 1966;15:476. Larrieu MJ, Rigollot C, Marder VJ. Comparative effects of fibrinogen degradation fragments D and E on coagulation. Br J Haematol 1972;22:719. Johnson RG, Rosenkrantz KR, Preston RA, Hopkins C, Daggett WM. The efficacy of postoperative autotransfusion in patients undergoing cardiac operations. Ann Thorac Surg 1983;36:173-9. Thurer RL, Lytle BW, Cosgrove DM, Loop FD. Autotransfu-

25. 26. 27.

28.

sion following cardiac operations: a randomized, prospective study. Ann Thorac Surg 1979;27:50&7. Morgan A, Anderson W, Bevilacqua R, et al. Effects of computer-controlled transfusion on recovery from cardiac surgery. Ann Surg 1972;178:391. Kaplan JA, Cannarelh C, Jones EL, et al. Autologous blood transfusion during cardiac surgery: a re-evaluation of three methods. J Thorac Cardiovasc Surg 1977;74:4. Moran JM, Babka R, ljilberman S, et al. Immediate centrifugation of oxygenator contents after cardiopulmonary bypass: role in maximum blood conservation. J Thorac Cardiovasc Surg 1978;76:510. Collins JA. Recent developments in the area of massive transfusion. World J Surg 1987;11:75.