Redo cardiac surgery: Late bleeding complications from topical thrombin-induced factor V deficiency

Redo cardiac surgery: Late bleeding complications from topical thrombin-induced factor V deficiency Bovine thrombin-induced factor V deficiency was thought to be a very rare acquired coagulopathy, with only three documented cases. We report a series of nine patients seen during a period of 32 months; these patients had normal preoperative coagulation profiles, and this unique coagulopathy developed 1 to 2 weeks after cardiovascular operations. The coagulopathy was characterized by a markedly elevated prothrombin time (40.9 ± 5.8 seconds), an elevated activated partial thromboplastin time (96.3 ± 12.2 seconds), a study positive for lupus anticoagulation (9/9), and markedly decreased levels of factor V (0.09 ± 0.03 U/m1) and factor XI (0.04 ± 0.02 U/m1), respectively. All patients had been exposed to commercially available bovine thrombin during prior cardiovascular or vascular operations and received a second bovine thrombin challenge during the latest procedure. Coagulopathic bleeding developed in four of the nine patients. Bleeding was unrelated to absolute fall in factor V level, but cessation of hemorrhage appeared to correlate with improvement in factor V level. Treatment with vitamin K, fresh frozen plasma, and platelet infusion were all unsuccessful in altering prothrombin time or factor V levels. Intravenous gamma globulin was used in three patients, two of whom were bleeding. All three patients showed a transient increase in factor V levels. Bleeding stopped in one of the two patients; the other continued to bleed and subsequently died. The third patient was treated prophylactically to increase factor V levels in preparation for flap reconstruction of his sternum. His factor V level increased from 0.26 to 0.49 U / mI, and he underwent the procedure without incident. Bovine thrombin-induced factor V deficiency may have been previously unrecognized. This deficiency should be suspected in patients who have undergone redo cardiovascular operations and in whom marked elevations in their prothrombin time occur 7 to 10 days after exposure to bovine thrombin. The resulting coagulopathy, although usually self-limited, has the potential to produce devastating bleeding complications. Intravenous gamma globulin (1 gm/kg during each of 2 days) has been used to increase factor V levels transiently but its role in therapy of this coagulopathy requires further investigation. (J THORAC CARDIOVASC SURG 1993;105:222-8)

Brian L. Cmolik, MD (by invitation)," Joel A. Spero, MD (by invitation)," George J. Magovern, MD,a and Richard E. Clark, MD,a Pittsburgh. Pa. With the assistance of Carol Reid, BS

Hemostasis after cardiovascular (CV) operations is dependent on many factors. Adequate neutralization of heparin, meticulous surgical technique, and the approFrom the Departments of Surgery" and Medicine." Allegheny General Hospital; Cardiovascular and Pulmonary Research Center, Allegheny-Singer Research Institute; Allegheny Campus, The Medical College of Pennsylvania, Pittsburgh. Pa. Read at the Seventy-second Annual Meeting of The American Association for Thoracic Surgery, Los Angeles, Calif., April 26-29, 1992. Address for reprints: Brian L. Cmolik, MD, Division of Thoracic Surgery. 320 E. North Ave., Pittsburgh, PA 15212. Copyright

1993 by Mosby-Year Book, Inc.

0022-5223/93/$1.00/+ 0.10

222

12/6/42460

priate use of blood products all contribute to a "dry" surgical field. Application of topical agents, including bovine thrombin, has been widely used in CV operations as an adjunctive measure to attain hemostasis. Some of the substances used to support hemostasis may represent a potential antigenic stimulus, and an appropriate immunologic response may occur. Acquired antibodies to specific blood coagulation factors have been reported in association with a variety of conditions, including infection, malignancy, pregnancy, and autoimmune disorders.l-' Acquired inhibitors to clotting factor V had been thought to be rare and, when present, have resulted in a bleeding diathesis ranging from

The Journal of Thoracic and Cardiovascular Surgery Volume 105, Number 2

mild to severe." A recent case report showed that a factor V inhibitor occurred after a CV operation and that topical bovine thrombin was the likely stimulus' In the mid 1980s antithrombin antibodies were reported to occur after CV operations. Their presence, however, was not associated with the development of coagulopathic bleeding. 6 During a 2'/z-year period we have observed nine patients in whom acquired inhibitors to factor V developed I to 2 weeks after redo CV operations. Five of these also had antithrombin antibodies. All nine were exposed to topical bovine thrombin at their original and subsequent CV procedure. Four of the nine patients had coagulopathic bleeding associated with the detection of antibodies to factor V.

Patients and methods Patient population. Between June 1989 and February 1992 at Allegheny General Hospital, Pittsburgh, Pennsylvania, nine patients were evaluated for prolonged prothrombin times (PTs) by the Hematology Service after they had CV operations. All had previous cardiac or vascular surgical procedures. Three patients underwent redo coronary artery bypass grafting, two had redo mitral valve replacements, two had redo aortic valve replacements, and one required aortic valve replacement and redo coronary artery bypass grafting. The final patient required replacement of the ascending aorta after having had previous abdominal aortic aneurysm repair (Table I). The four patients with prosthetic valves discontinued warfarin approximately I week before the redo procedure. All patients had received topical bovine thrombin to ensure hemostasis at both the original and the most recent CV operations. The dosage of thrombin used could not be quantified based on the records in the majority of the patients. In four patients the dose varied between 1,000 and 60,000 U. Two brands of patient-grade bovine thrombin were used in the most recent CV operations. Thrombostat (normal thrombin clotting time [TCT) 19.9 seconds) was provided by Parke-Davis, Morris Plains, New Jersey and Jones Pharmaceutical, St. Louis, Missouri, and Thrombinar (normal TCT 19.0 seconds), provided by Armour, Kankakee, Illinois, were both used as topical hemostatic agents during this time period. Laboratory procedures. Tests for PT, normal 10 to 13 seconds, activated partial thromboplastin time (APTT), normal 24 to 34 seconds, and TCTs normal II to IS seconds, were performed according to standard laboratory protocol. The plasma of all patients with prolonged TCTs was treated with Heparsorb (General Diagnostics,Organon Teknika Corp., Durham, N.C.), and all subsequent tests were performed on adsorbed plasma. If no heparin was detected, testing was performed on native plasma. Clotting factor assays were performed with commercial substrates in separate dilutions with one-stage test systems on a Coa-Screener (American Labor, Durham, N.C.). Factors II, V, and VII were assayed with human substrate (George King Bio-Medical, Inc., Overland Park, Kan.). Factors VIII, IX, X, and XII were assayed with human substrate and factor XI with bovinesubstrate, all of which were obtained from Dade, Baxter-Travenol, Miami, Florida. The normal range of concentrations in human beings for all clotting factors was 0.50

Cmolik et al.

223

Table I. Patient procedures Patient

Original procedure

1* 2*

6 7

Porcine MVR, 1980 Mechanical AVR/MVR, 1980 CABG,1981 Porcine MVR, 1979 Abdominal aneurysm, 1986 CABG,1990 CABG,1984

8* 9

AVR porcine, 1986 CABG,1981

3 4* 5

Subsequent procedure Mechanical MVR, 1989 Mechanical AVR, 1990 CABG,1991 Mechanical MVR, 1991 Ascending aortic replacement, 1991 CABG,1991 CABG and mechanical AVR, 1991 AVR porcine, 1992 CABG,1992

MVR, Mitral valve repair; A VR, aortic valve repair; CABG, coronary artery bypass graft. *Coagulopathic bleeding developed.

to 1.50 U Iml. All patients were evaluated for antiphospholipid antibodies by means of the platelet neutralization test of Triplett and colleagues." TCT tests were performed with reagentgrade bovinethrombin (normal TCT 21.5 seconds), coagulation factor-depleted human thrombin (normal TCT 16.7 seconds), and horse thrombin (normal TCT 24.4 seconds), all from Sigma Chemical Co., St. Louis, Missouri. Bovine thrombin, from General Diagnostics, Durham, North Carolina (normal TCT 15.7 seconds), and Dade, Miami, Florida, (normal TCT 14.2 seconds), was tested as well. Finally, thrombin time was also studied with Reptilase (ABC, Parsippany, N.J.) and atrox venom (Sigma Chemical Co., St. Louis, Mo.) (normals 24.4 and 6.2 seconds, respectively).

Results PT became elevated in nine patients (five men) I to 2 weeks postoperatively after little or no warfarin administration, and this prompted consultation from the Hematology Service. The average elevation in PT was 40.9 seconds (Table II). At the time of consultation, no patient had clinical signs of bleeding. The initial coagulation evaluation consisted of APTT, TCT, and plasma clotting factor (factors II, V, VII, X, XI, and XII) measurements. The average peak APTT was 96.3 seconds, and the average peak TCT was 43.9 seconds, with a range of 12.3 to greater than 150 seconds (see Table II). Factor II, VII, and X levels were in the normal range for the vast majority of the patients in the series. Only two patients had low levels of factors II and X, and one patient had a consistently low factor VII level (Table III). Factors V and XI were uniformly 'depressed in all patients in this series. Factor V levels averaged 0.09 U jml on initial evaluation, whereas factor XI levels were lower, averaging 0.04 U/rnl, These levels were monitored throughout the hospital course (see Table III). The plasma from these patients was tested with three

224

The Journal of Thoracic and Cardiovascular Surgery February 1993

Cmolik et at.

Table II. Coagulation profile Admission PT (10-13 sec)

Patient 1* 2* 3 4* 5 6 7 8* 9 Mean ± SEM

Peak PT (10-13 sec)

13.5 13.5 12.4 13.6 12.9 11.7 11.3 15.9 13.1 13.1 ± 0.4

69.6 64.2 47.6 30.6 43.1 21.7 3l.2 36.4 19.6 40.9 ± 5.8

Peak APTT (24-34 sec)

TCT (11-15 sec)

125 167 110 103 95 45 68 93 61 96.3 ± 12.2

15.6 83.0 23.1 150 12.3 13.7 14.2 33.4 50 43.9 ± 15.3

PT. Prothrombin time; APTT. activated partial thromboplastin time; TCT. thrombin clotting time; SEM. standard error of the mean.

'Coagulopathic bleeding developed.

Table III. Initial coagulation factor levels Patient 1* 2* 3 4* 5

6 7 8* 9 Mean ± SEM

F11 0.50-1.50 Ufml

FV 0.50-1.50 Utml

FVII 0.50-1.50 Ulml

FX 0.50-1.50 Ulml

FXI 0.50-1.50 Ulml

0.01 0.90 0.63 0.28 0.78 l.01 0.82 0.205 0.99 0.62 ± 0.12

0.01 0.05 0.01 0.07 0.02 0.28 0.08 0.05 0.26 0.09 ± 0.03

0.65 0.83 0.54 0.55 0.52 0.89

0.25 0.25 0.73 0.86 0.75 0.79 0.75 0.25 l.00 0.62 ± 0.09

0.01 0.01 0.01 0.01 0.01 0.19 0.03 0.06 0.05 0.04 ± 0.02

0.77 0.26 l.00 0.67 ± 0.07

Fll. Factor II; FV, factor V; FVll, factor VII; FX. factor X; FXf, factor FXI; SEM. standard error of the mean.

'Coagulopathic bleeding developed.

types of bovine thrombin, as well as reagent-grade, clotting factor-depleted human thrombin, to evaluate the conversionof fibrinogen to fibrin, The TCT, when the test was performed with bovine thrombin, was the most markedly abnormal in two patients in whom overt coagulopathic bleeding developed; bleeding time ranged from 150 to greater than 200 seconds (normal 14.2 to 21.5 seconds). When the plasma of these patients was tested with human thrombin, the results were similar, with elevated TCTs observed at 20.9 and 52.9 seconds (normal 16.7 seconds) (Table IV). Bleeding complications occurred in four of the nine patients I to 2 weeks after the CV operations. One patient had mild upper gastrointestinal tract bleeding, which abated without treatment. In a second patient a hemothorax developed, together with subsequent respiratory insufficiency, which was treated with closed tube thoracostomy. In a third patient a wound hematoma developed I week after an intraaortic balloon pump (IABP) was removed from the left femoral artery. She also bled from the tracheostomy site I week after that procedure. Both

wound sites were thoroughly explored, and the cause of the bleeding was deemed coagulopathic. The extensive nature of the hematoma at the IABP insertion site subsequently required evacuation and muscle flap closure. The fourth patient had severe epistaxis, which responded only to anterior nasal packing. A profound thrombocytopenia and a large hemothorax with persistent hypoxemia subsequently developed. For the hemothorax the patient was treated with closed tube thoracostomy. He died of intractable ventricular arrhythmias within 36 hours of the development of the hemothorax. Coagulopathic bleeding was believed to bear significant responsibility for his death. Treatment of patients who have a factor V deficiency included vitamin K, fresh frozen plasma, platelet pack infusion, and intravenous gamma globulin. No improvement in either the factor V level or the PT was noted in response to vitamin K, fresh frozen plasma, or the use of platelet pack infusion transfusions. Intravenous gamma globulin improved factor V levels in all three patients within 24 hours. Continued bleeding from the IABP

The Journal of Thoracic and Cardiovascular Surgery Volume 105, Number 2

Cmolik et al.

225

Table IV. TCTs in three patients who bled Reagent-grade Reagent-grade Reagent-grade Reagent-grade Reagent-grade Patient-grade Patient-grade bovine bovine human bovine Reptilase Atrox bovine bovine horse thrombin I thrombin 2 thrombin 3 thrombin thrombin thrombin 4 time venom thrombin 5

Normal PT

1 2 4

21,5 15.9 >200 >100

15,7 14.7 >150 >150

14,2 13.5 >180 >150

24.4 22.4 34.9 NO

16,7 14.7 20.9 52.9

19,9 14.4 240 >150

19.0 14.0 99 >150

24.4 21.7 26.1 24.4

6.2 8.2 Il.l 7.9

All studies in patients with prolonged thrombin clotting times (TCTs) were performed on plasma treated with triethyl-aminoethyl-cellulose (Heparsorb); bovine thrombin I was depleted of clotting factors while bovine thrombin 2 and 3 were not; reagent-grade human thrombin was also depleted of clotting factors. PT. Prothrombin time; ND, not done.

insertion site failed to respond to vitamin Kyfresh frozen plasma and platelet pack infusion in patient 2. In this patient, factor V levels increased transiently from 0.05 to 0.42 U jml within 24 hours of intravenous gamma globulin infusion, and her bleeding stopped. In patient 4 epistaxis developed that did not respond to vitamin Kyfresh frozen plasma and platelet pack infusions. He received intravenous gamma globulin after the unexpected development of severe thrombocytopenia (platelet count 5000j Ill) and a hemothorax. His factor V level increased from 0.20 to 0.40 U jml within 24 hours of intravenous gamma globulin administration. His thrombocytopenia persisted, however, and he died despite a seemingly adequate factor V level (0.40 U Iml). Patient 9 had a sternal nonunion after redo coronary artery bypass grafting, and this patient required plastic surgical flap reconstruction. Routine preoperative coagulation studies showed an elevated PT, and subsequent workup defined factor V deficiency. Up to the time of his flap reconstruction he had had no clinical signs of bleeding. He was treated with intravenous gamma globulin, and within 24 hours his factor V level increased from 0.26 to 0.49 U jml. Flap reconstruction of the sternum was done 4 days after intravenous gamma globulin administration, at which time his factor V level was 0.70 U jml. The procedure was successful, and no evidence of excessive bleeding was noted either intraoperatively or postoperatively. Eight of nine patients were discharged from the hospital. The average PT at the time of discharge was 19.4 ± 1.5 seconds. Factor V levels increased to 0.33 ± 0.10 U jml. Factor XI levels remained depressed at 0.06 ± 0.Q2U /rnl (Table V). Follow-up data available in four patients at 4 to 12 weeks after CV operations show an average PT of 15.5 ± 1.3 seconds and an average factor V of 0.66 ± 0.18 Ujml.

Discussion Intraoperative and postoperative bleeding associated with cardiovascular operations remains an ongoing problem, especially in patients undergoing "redo" procedures.

Perioperative bleeding has been most commonly attributed to acquired platelet dysfunction.v" and, in the majority of patients, adequate hemostasis is accomplished within the first 24 postoperative hours. We report a series of patients who experienced an only recently recognized coagulopathy, which appeared 8 to 12 days after operation and resulted in clinical bleeding in four of nine affected patients. In the latter 1980s reports began to appear of prolonged thrombin times after CV operations. Stricker and colleagues" reported three cases in which coagulation times became prolonged on screening tests a week or more after redo procedures. Two of their three patients died of what was probably a coagulopathy-related hemorrhage.? All three of their patients had prolonged PTs and APTTs, but clotting factor assays were not performed. Flaherty, Henderson, and Wener4 reported a similar picture in a patient after a CV operation and suggested that the prolonged TCT resulted from iatrogenic immunization of the patient to bovine thrombin used as a hemostatic agent during the operation. Subsequently, Zehnder and Leung' reported a case similar to two of the cases reported by Stricker and colleagues. Zehnder and Leung's patient had a profound coagulopathy after a CV operation and was shown to have a markedly decreased factor V level. Zehnder and Leung hypothesized that the patient's factor V deficiency resulted from a cross-reaction to an antibody elicited after exposure to bovine thrombin. They then proceeded to demonstrate bovine factor V in the topical thrombin preparation. There have since been other reports of a similar coagulapthy developing in other patients." 10. II Acquired inhibitors directed toward specific clotting factors are rare. The most common have been circulating anticoagulants directed toward factor VIII:C. 3 Demonstration of specific inhibitors of factor V have been even less common. A 1978 review identified only 12 reported cases in the world Iiterature.l? Nine of these occurred in the postoperative period and resulted in a mild to severe bleeding diathesis in seven of the nine affected individu-

226

The Journal of Thoracic and Cardiovascular Surgery February 1993

Cmolik et al.

Table V. Outcome data at end of hospitalization Patient

PT 10-13 sec

FV 0.50-1.50 U/ml

FXI 0.50-1.50 U/ml

1* 2* 3 4* 5 6 7 8* 9 Mean ± SEM

23.9 16.3 26.0 15.2 15.3 17.3 20.2 24.6 15.4 19.4±1.5

0.11 0.56 0.05 0.33 0.53 0.31 0.08 0.05 0.92 0.33 ± 0.10

0.01 0.01 0.01 0.09 0.12 0.19 0.03 0.06 0.03 0.06 ± 0.02

Outcome

Alive Alive Alive Died Alive Alive Alive Alive Alive

PT. Prothrombin time; FV. factor V; FXI. factor XI. SEM, standard error of the mean.

*Coagulopathic bleeding developed.

also In none of the surgical procedures was there any rationale for, or suggestion of, the use of bovine thrombin. I2, 13 Given the previous rarity of factor V inhibitors, nine cases of acquired inhibitors to clotting factor V identified in patients having cardiac operations at a single institution in a 32-month period (July I, 1989, to March 1, 1992) is noteworthy. All of our patients demonstrated decreases in clotting factor V and factor XI assays. Only factor XI assays were performed with bovinesubstrate, and it is possible that the strikingly severe and prolonged decreases of factor XI concentrations were a reflection of the assay method. Rapaport and colleagues 10 recently investigated the assay problems associated with the use of bovine reagents in two patients similar to ours who had markedly prolonged TCTs with bovine thrombin but whose TCTs were clearly less prolonged by human thrombin. In our patients, however, deficiencies were noted in an apparently random fashion in other clotting factor assays (factors II, X, and VII), which were assayed using human substrate. In all of our patients, the diminished factor V level clearly behaved in vivoand in vitro as though it was produced by an antibody because it failed to improve when fresh frozen plasma was given to the patient or added to the plasma fraction of the patient in vitro. Unlike the large majority of reported acquired clotting factor inhibitors, however, the factor V levelin our patients did not increase in the face of progressive dilution during factor assay. Therefore, the kinetics of the factor V inhibitor associated with bovine thrombin remain to be elucidated. Interestingly, those individuals with abnormal TCTs with human thrombin appeared to be the patients most likely to bleed. The development of hemorrhage in these patients was not predicted by the depth of the fall in factor V levels, although cessation of bleeding appeared to correlate with improvement in the factor V level.

Factor V deficiency in our patients resolved slowly after a period of time, and the factor V level generally began to improve within a few weeks of its discovery. The patients who bled after the operation, however, presented a therapeutic challenge. Not surprisingly, they did not increase their factor V level in response to administration of either vitamin K or platelet fresh frozen plasma.!" Platelet transfusion, reported to be effective in treating individuals with anti-factor V presumably by absorbing the antibody to factor V on the platelets.P also failed to produce any response. Intravenous gamma globulin has been effective in the treatment of patients with acquired inhibitors to factor VIII:c. 16 In all patients in whom it was used, the factor V level increased approximately two-fold or more within 24 to 48 hours after a dose of I gm/kg intravenous gamma globulin on each of two consecutive days. One patient's factor V level continued toward normal; in the others it fell during the next few days, but not to baseline. The patient who had acute thrombocytopenia received intravenous gamma globulin for presumptive idiopathic thrombocytopenic purpura and was one of the patients who showed an increase in factor V. It is tempting to postulate that the fall in platelet count may have been related to an antiplatelet factor V antibody.!? Zehender and Leung'' reported success with plasmapheresis in their patient, but we did not pursue this approach in our patients. The long-term course of these patients remains unknown. In no patient did the PT return to normal, remaining mildly prolonged as long as 3'12 months after the operation. Those individuals with mechanical valves were discharged and placed on a regimen of aspirin therapy that has not been changed to warfarin (Coumadin). The problem of assessing warfarin therapy in patients who begin with a prolonged PT is significant. Additionally, it is unclear to what, if any, degree the patients are

The Journal of Thoracic and Cardiovascular Surgery Volume 105, Number 2

self-anticoagulated. We are pursuing the answers to these questions. We believe that factor V deficiency resulting from the use of topical bovine thrombin will be seen with increasing frequency. Patient's undergoing CV operations, particularly those undergoing redo CV procedures, represent a growing population of potential candidates for development of this coagulopathy. The clinical spectrum of the factor V deficiency may vary from no evidence of bleeding to a life-threatening coagulopathy. Factor V levels in themselves are not good predictors of propensity for bleeding. TCTs that are prolonged when tested with human as well as bovine thrombin may be of value in predicting patients who are likely to bleed. The use of vitamin K, fresh frozen plasma, and platelets is ineffective in raising the level of factor V or correcting the PT in this deficiency. Intravenous gamma globulin was effective in increasing factor V levels in the three patients in whom it was used, but it stopped coagulopathic bleeding in only one of the two patients. Thus its role in the management of factor V deficiency remains to be defined. Since the recognition of these cases of acquired factor V deficiency, an attempt has been made to be more selective in the use of topical bovine thrombin at our institution. We would like to thank Dinean Samuel and Nancy Lynch for their help in preparing and editing this manuscript. REFERENCES I. Aliaga JL, deGarcia J, Vidal R, Pico VM, Flores P, Sampol G. Acquired factor V deficiency in a patient with pulmonary tuberculosis. Eur Respir J 1990;3:109-10. 2. Velez-Garcia E, Fradera J. Transfusion induced acquired immunodeficiency syndrome (AIDS) with Kaposi's sarcoma in a patient with congenital factor V deficiency. Aids Res 1984;1:401-6. 3. Shapiro SS, Hultin M. Acquired inhibitors to blood coagulation factors. Semin Thromb Hemost 1975;1:336-84. 4. Flaherty MJ, Henderson R, Wener MH. Iatrogenic immunization with bovine thrombin: a mechanism for prolonged thrombin times after surgery. Ann Intern Med 1989;111: 631-4. 5. Zehnder JL, Leung LLK. Development of antibodies to thrombin and factor V with recurrent bleeding in a patient exposed to topical bovine thrombin. Blood 1990;76:2011-6. 6. Stricker RB, Lane PK, Leffert JD, Rodgers GM, Shuman MA, Corash L. Development of antithrombin antibodies followingsurgery in patients with prosthetic cardiac valves. Blood 1988;72:1375-80. 7. Triplett DA, Brandt JT, Kaczor D, Schaeffer J. Laboratory diagnosis of lupus inhibitors: a comparison of the tissue thromboplastin inhibition procedure with a new platelet neutralization procedure. Am J Clin Pathol 1983;79:67882.

Cmolik et at. 2 2 7

8. Woodman RC, Harker LA. Bleeding complications associated with cardiopulmonary bypass. Blood 1990;76:168097. 9. Harker LA, Malpass TW, Branson HE, Hessel EA, Slichter SJ. Mechanism of abnormal bleeding in patients undergoing cardiopulmonary bypass: acquired transient platelet dysfunction associated with selective a-granule release. Blood 1980;56:824-34. 10. Rapaport SI, Zivelin A, Minow RA, Hunter CS, Donnelly K. Clinical significance of antibodies to bovine and human thrombin and factor V after surgical use of bovine thrombin. Am J Clin Pathol 1992;97:84-91. 11. Nichols WL, Daniels TM, Fisher PK, Owen WG, Figueroa PI, Pineda AA. Inhibitors of coagulation factor V and thrombin associated with surgical use of topical bovine thrombin or fibrin glue. Blood 1991;78(suppllO):241. 12. Feinstein DJ. Acquired inhibitors of factor V. Thromb Haemost 1978;39:663-74. 13. Nesheim ME, Nichols WL, Cole TL, et al. Isolation and study of an acquired inhibitor of human coagulation factor V. J Clin Invest 1986;77:405-15. 14. Tsuda H, Yumi M, Hara T, et al. A case of congenital factor V deficiency combined with multiple congenital anomalies: successful management of palatoplasty. Acta Haematol 1990;83:49-52. 15. Chediak J, Ashenhurst JB, Garlick I, Desser RK. Successful management of bleeding in a patient with factor V inhibitor by platelet transfusions. Blood 1980;56:835-41. 16. Sultan Y, Kazatchkine MD, Nydegger U, Rossi F, Dietrich G, Algiman M. Intravenous immunoglobulin in the treatment of spontaneously acquired factor VIII:C inhibitors. Am J Med 1991;91(suppl 5A):35-9. 17. Grigg AP, Dauer R, Thurlow PJ. Bleeding due to an acquired inhibitor of platelet associated factor V. Aust N Z J Med 1989;19:310-4.

Discussion Dr. L. Henry Edmunds, Jr. (Philadelphia, Pa.). Antibodies to the various coagulation proteins are really rare except in patients with hemophilia who have antibodies to factor VIII. Both bypass and stored blood reduce the concentrations of factor V and factor VIII, but not to the point at which a bleeding diathesis occurs. It appears from your presentation that you did not have any bleeding at the time of the operation. Apparently the antibody to factor V was not actually circulating in sufficient concentration during operation to increase early postoperative or perioperative bleeding. All the bleeding occurred late; this suggests a delayed hypersensitivity reaction and that reuse of bovine thrombin causes antibodies to form. I am surprised tha t your thrombin time test is prolonged. That is a simple test. All you do is add thrombin to plasma and see how long it takes to clot. I have always interpreted it as a measure of the concentration of fibrinogen. As far as I can tell, you did not measure fibrinogen in this study. Do you think that the fibrinogen concentration is low, and, if so, why not give cryoprecipitate to restore normal fibrinogen levels and thrombin times? I am also surprised that fresh frozen plasma did not work.

2 2 8 Cmolik et at.

Since these are late complications, you should be able to administer both plasmapheresis and fresh frozen plasma. If you gave enough fresh frozen plasma, the factor V levelsshould rise. You have brought to us a new disease and a new concern and we thank you for it. Dr. G. Frank O. Tyers (Vancouver, B.C., Canada). My understanding is that bovine topical thrombin is in fact a mixture of coagulants and that some of the manufacturers' products contain only a small percentage of thrombin and that is probably the origin of this "acquired immune disease." We have seen a few patients with this problem as well, and our hematologists have been concerned that the condition is affecting only the prothrombin time and that the patients are not actually anticoagulated. Did any of your patients have mechanical valves? Dr. Crnolik. Yes. There were, I think, five who had mechanical valves. Dr. Tyers. Our patients too had PTs that were at a frightening level for months after the operation. We have had one patient with a mechanical valve who had an embolic episode despite a high PT, and we are concerned that this condition may affect the tests more than it affects the patient's actual coagulation status. You have seen early bleeding problems, but have you seen late thrombotic problems? Dr. Eric Rose (New York, N.y.). To amplify that concern, the functional tests that you described, the PTs and the thrombin times are done in most laboratories with the use of bovine reagents. Have you repeated those examinations with human

The Journal of Thoracic and Cardiovascular Surgery February 1993

reagents used to do the PTs and the thrombin times? I would suspect that even after administration of human fresh frozen plasma, if the antibodies are present and you are still using bovine reagents, you may still see an elevation that is spurious. Dr. Crnolik. The fibrinogen level was measured postoperatively in all nine patients and was normal. The development of an antibody to coagulation factors was supported by the observations that the PT did not correct with the administration of fresh frozen plasma. We also conducted in vitro mixing experiments. The addition offresh frozen plasma to the serum of those patients who bled did not result in normal thrombin times. Thrombin times were tested with human, horse, and bovine thrombin. The thrombin times, specifically in the three patients who bled, were prolonged when tested with all types of thrombin. That is why we think a prolonged thrombin time may be a predictor of coagulopathic bleeding. We share the concern about anticoagulation in patients with mechanical valves. All patients who have mechanical valves have been discharged on a regimen of aspirin therapy, and, to my knowledge, in none has therapy been changed to warfarin at this time. The current method of assaying effectiveness of warfarin therapy may not reflect the degree of anticoagulation of the patient. The PT remains slightly elevated and factor V levels become normal. The significance of slightly elevated PTs is unclear at this time. We have not observed any thrombotic complications in these patients as yet.