A safety review of topical bovine thrombin-induced generation of antibodies to bovine proteins

Clinical Therapeutics/Volume 31, Number 4, 2009

A Safety Review of Topical Bovine Thrombin–Induced Generation of Antibodies to Bovine Proteins Frederick A. Ofosu, PhD1,2; Sheila Crean, MS3; and Matthew W. Reynolds, PhD3 1Department

of Pathology and Molecular Medicine, McMaster University, Hamilton, Ontario, Canada; Blood Services, Hamilton, Ontario, Canada; and 3United BioSource Corporation, Lexington, Massachusetts

2Canadian

ABSTRACT Background: Topical bovine thrombin has been used to accelerate attainment of hemostasis in the surgical setting for >60 years, and its immunogenicity has been widely reported. Although the development of antibodies is inherent in the introduction of any non–self-therapeutic protein such as bovine-sourced thrombin, there are questions about the relationship between the presence of antibodies to constituents of the therapeutic protein preparation and the occurrence of clinically relevant adverse events (AEs). Objective: This review examines the proposed mechanisms for the immunogenicity of topical bovine thrombin preparations and summarizes available evidence from randomized clinical trials, observational studies, and case reports to explore possible relationships between the reported immunogenicity of topical bovine thrombin and the occurrence of AEs. Methods: A search of MEDLINE (1966–August 2008) for studies published in English was conducted using the Medical Subject Heading terms surgery, antibodies, and hemorrhage, as well as equivalent key words for bovine, adverse events, and thrombin. For inclusion in the review, studies had to report clinical or laboratory safety data for patients exposed to topical bovine thrombin during surgery. Results: The evidence suggests that patients with repeated perioperative exposure to topical bovine thrombin have a 3- to 10-fold greater risk for development of antibodies to topical bovine thrombin than do patients with no history of surgery-related exposure to this agent. Early case reports associated the development of anti–bovine protein antibodies with bleeding and/or thrombotic complications. However, in one prospective, randomized controlled trial comparing topical bovine thrombin with topical recombinant human thrombin, 99.5% of patients in each treatment arm developed postoperative AEs. In anApril 2009

other, 54% and 55% of patients in the respective treatment arms developed postoperative AEs. In a prospective, randomized controlled trial that compared topical bovine thrombin and plasma-derived human thrombin, 95.5% of patients in each treatment arm developed postoperative AEs. Conclusions: Repeated perioperative exposure to topical bovine thrombin may increase both the prevalence and titers of antibodies to ≥1 protein contained in nonhomogeneous topical bovine thrombin preparations. However, the evidence reviewed does not support a definitive association between preoperative or postoperative generation of anti–bovine protein antibodies and an increased risk of AEs in surgical patients treated with topical bovine thrombin. (Clin Ther. 2009; 31:679–691) © 2009 Excerpta Medica Inc. Key words: surgical hemostasis, topical thrombin, recombinant proteins, immunogenicity, adverse events.

INTRODUCTION Topical bovine thrombin has been used to accelerate attainment of hemostasis in the surgical setting for >60 years.1 It is used alone or in combination with other drugs to control perioperative surface bleeding in cardiovascular, vascular, neurologic, orthopedic, gynecologic, and transplant surgeries.2–5 Because the use of topical bovine thrombin is often undocumented in patients’ medical records,6 it is difficult to accurately estimate the risk for adverse events (AEs) associated with clinical use of this agent. One study estimated that at least 1 million surgery patients in the United States are treated with topical bovine thromAccepted for publication September 29, 2008. doi:10.1016/j.clinthera.2009.04.021 0149-2918/$ - see front matter © 2009 Excerpta Medica Inc. All rights reserved.

679

Clinical Therapeutics bin each year.7 Exposure, particularly repeated exposure, of surgical patients to bovine thrombin preparations may result in the generation of anti–bovine protein antibodies that may also cross-react with some human coagulation factors,6–21 as most bovine thrombin preparations are not homogeneous.21,22 Generation of antibodies in patients treated with animal plasma– derived, human plasma–derived, or recombinant therapeutic proteins is a well-recognized phenomenon.23–25 Exposure of patients to topical bovine thrombin may lead to the generation of antibodies directed against ≥1 constituent of the preparation. Up to 70% of patients exposed to bovine thrombin more than once may develop antibodies to several of the proteins present in nonhomogeneous topical bovine thrombin preparations.4 It is not possible to accurately quantify antibody generation in surgical patients exposed or reexposed to bovine thrombin, however, as the methods used for this purpose are generally not comparable across studies. Furthermore, the purity of bovine thrombin preparations can differ markedly, with the bovine thrombin content varying from >90% to <30%.21,22 Estimated qualitatively, the factor V/Va content in topical bovine thrombin preparations can also vary widely.6,11–16,20–22,26–28 Bovine factor V, a highly immunogenic constituent, was likely to be present in topical bovine thrombin preparations in clinical use before 2000. Up to 50% of patients treated with topical bovine thrombin containing bovine factor V/Va develop anti–bovine factor V/Va antibodies that may also cross-react with and inactivate human factor V/Va.8,10–16,26–28 In published reports, the purity and/or concentrations of antigens and the specificity of the antibodies used in immunoassays have generally not been well described.6,18,28 Nonetheless, studies in patients exposed/reexposed to bovine thrombin have reported the generation of neutralizing and nonneutralizing antibodies that have varying affinities for bovine factor V/Va, bovine prothrombin, bovine thrombin, and bovine factor X, and may cross-react with the corresponding human clotting factors.6–9,12,15,20,21 Differences in the amino acid sequences of human and bovine clotting factors may result in the development of antibodies to bovine clotting factors in surgical patients exposed/reexposed to bovine thrombin. For example, the amino acid sequences of bovine and human thrombins differ by 36 residues. There are also minor differences in the patterns of N-glycosylation of 680

human and bovine clotting factors.29–31 The xenogeneic Forssman disaccharide found in some bovine proteins also contributes to the immunogenicity of bovine thrombin preparations in humans.21,27,32 Bovine glycoproteins containing the Forssman antigen elicit immune responses in all patients exposed/reexposed to this antigen, which is normally synthesized by all nucleated mammalian cells except those of humans, apes, and Old World monkeys.27,32 Exposure of patients to bovine thrombin preparations containing galactose-α-1-3-galactose increases titers of the antiForssman antibodies that are normally found in most human plasma and are directed against this xenogeneic disaccharide.15,21 The lack of the Forssman disaccharide in bovine thrombin may explain the relatively low frequency of anti–bovine thrombin antibodies (<10%) in the plasma of patients exposed to topical bovine thrombin, as this disaccharide is highly immunogenic in humans.15,16,21,27,33 In addition to inducing the generation of anti-Forssman antibodies, bovine coagulation proteins may induce the generation of anti–bovine clotting factor antibodies that cross-react with human clotting factors (ie, autoantibodies to human coagulation factors) by breaking self-tolerance.34,35 Among factors that may interact with ≥1 constituent of topical bovine preparations to favor the breaking of self-tolerance are the generalized increased inflammation associated with major surgery and the transfusion of blood products.34,35 This review examines the proposed mechanisms for the immunogenicity of topical bovine thrombin preparations and summarizes available evidence from randomized controlled trials comparing the safety profile of topical bovine thrombin with placebo or recombinant human thrombin (rhThrombin), observational studies, and case reports to explore possible relationships between the reported immunogenicity of topical bovine thrombin and the occurrence of AEs.

METHODS A search of MEDLINE (1966–August 2008) for studies published in English was conducted using the Medical Subject Heading terms surgery, antibodies, and hemorrhage, as well as equivalent key words for bovine, adverse events, and thrombin. All Englishlanguage abstracts and papers describing studies, case reports, or case series in surgical patients treated with topical bovine thrombin that reported the incidence of antibodies and/or AEs were included in the review. Volume 31 Number 4

F.A. Ofosu et al. The reference lists of identified review articles were also examined for additional relevant publications. All included studies were reviewed and summarized with respect to patients and/or study populations, surgery type, topical bovine thrombin exposure, incidence of antibodies, and incidence of AEs.

RESULTS Randomized Controlled Trials Five randomized controlled trials were identified that reported safety data on bovine topical thrombin,20,36–39 some of which are summarized in the table. In a doubleblind study by Mathiasen and Cruz,36 70 children with obstructive sleep apnea who were undergoing traditional cold-steel adenoidectomy were randomized to receive FloSealtm (Baxter International Inc., Deerfield, Illinois), a bovine thrombin–containing product, or to undergo cauterization (standard treatment). The 35 patients treated with FloSeal attained postsurgical hemostasis significantly more rapidly than the 36 patients who received standard treatment (mean [SD], 0.6 [1.3] vs 9.5 [5.4] minutes, respectively; P < 0.001). FloSeal-treated patients also lost significantly less blood than patients who received standard treatment (2.5 [9.2] vs 29.4 [27.1] mL; P < 0.001). Over 7 days of follow-up, neither treatment was associated with any complications (hospitalization, postoperative hemorrhage, or need for blood transfusion or aspiration). In a study by Winterbottom et al,20 309 patients undergoing cardiac, vascular, or spinal surgery were randomized to receive FloSeal or Gelfoam® (Pfizer Inc., New York, New York). The thrombin component in both products was Thrombin-JMI® (Jones Pharma Incorporated, Bristol, Tennessee). The study reported total AEs but did not compare rates of AEs between treatment arms. Sixteen patients (5.2%) died, 15 (4.9%) developed anemia, 9 (2.9%) had excessive postoperative bleeding, 7 (2.3%) developed arterial thrombosis, and 5 (1.6%) developed congestive heart failure or stroke. This study found no significant association between postoperative bleeding complications and the presence of antibodies to bovine thrombin and/or bovine factor Va, either preoperatively or 6 to 8 weeks after surgery, in patients who had received topical bovine thrombin. Although 51 patients had anti–bovine thrombin antibodies and 82 had anti–bovine factor Va antibodies, none experienced a coagulopathy. Chapman et al37 conducted a double-blind clinical trial in which 206 patients were randomized to receive April 2009

topical bovine thrombin (Thrombin-JMI) and 205 were randomized to receive topical rhThrombin. Preoperatively, 10 of 200 evaluable patients (5.0%) who received topical bovine thrombin had antibodies to this agent, and 3 of 198 evaluable patients (1.5%) who received topical rhThrombin had antibodies to this agent. It is not clear why 3 patients who had never been exposed to rhThrombin had antibodies to this agent. The authors did not identify the antigens in the topical bovine thrombin preparation to which antibodies were generated. One month after surgery, 43 (21.5%) of the patients treated with topical bovine thrombin had developed anti–bovine thrombin antibodies, whereas no additional patients in the topical rhThrombin group had developed antibodies to human thrombin (P < 0.001). The majority of patients in both treatment groups experienced an AE (205/206 [99.5%] for bovine thrombin, 204/205 [99.5%] for rhThrombin), and 46 (22.3%) and 36 (17.6%) patients in the respective arms experienced a serious AE. Life-threatening and treatment-related AEs were reported in 4 (1.9%) and 2 (1.0%) patients, respectively, in the bovine thrombin arm and 10 (4.9%) and 7 (3.4%) patients in the rhThrombin arm. The incidence of AEs was not significantly associated with the presence of antibodies to bovine thrombin or rhThrombin. Weaver et al38 conducted a double-blind trial comparing the efficacy and safety profiles of topical bovine thrombin and topical rhThrombin in 164 patients undergoing peripheral arterial bypass surgery or arteriovenous graft procedures. Rates of AEs were comparable with topical bovine thrombin and topical rhThrombin (54% and 55%, respectively). Although both treatments were similarly well tolerated, rhThrombin had a superior immunogenicity profile. Specifically, no patients in the rhThrombin group developed antibodies to human thrombin, whereas 27% of patients who received bovine thrombin developed antibodies to bovine thrombin (P < 0.001). In a multicenter, double-blind study, Doria et al39 compared the efficacy and safety profiles of topical bovine thrombin and plasma-derived topical human thrombin in 152 and 153 patients, respectively, who were undergoing elective cardiovascular, neurologic, or general surgery. The 2 products had comparable AE profiles: 95.5% of patients in each treatment arm developed postoperative AEs. A larger proportion of patients receiving topical bovine thrombin developed antiproduct antibodies compared with those receiving 681

Clinical Therapeutics

682 Table. Randomized controlled trials reporting safety data and/or generation of antibodies after exposure to topical bovine thrombin. Authors

Bovine Thrombin Product*

Dorion et al4

Not reported

Stored blood samples and hospital records were used to identify 120 patients previously exposed to bovine thrombin (102 single exposures, 18 multiple exposures) and 114 patients unexposed to bovine thrombin

Antibodies to bovine thrombin were present in 12 of 120 patients (10%) exposed to bovine thrombin (95% CI, 4.6–15.4), including 5 of 102 (5%) with a single exposure and 7 of 18 (39%) with multiple exposures; rates not reported for unexposed patients; 2 serious bleeding episodes in exposed patients (2% of total exposed, 17% of those with antibodies to bovine thrombin); length of hospital stay significantly longer in those with bovine thrombin antibodies vs those without (23 vs 11 days, respectively; P = 0.007)

Ortel et al16

Thrombogen

151 Surgery patients (109 CABG, 42 valve surgery) exposed to bovine thrombin

Hemorrhagic complications: 11 patients with excessive postoperative bleeding or reoperation for bleeding, 5 with GI bleeding, 5 requiring transfusion support after postoperative day 2, 4 with recurrent episodes of epistaxis after hospital discharge, 1 with hemorrhagic stroke, 1 with hemorrhagic pericarditis; thromboembolic events: 7 with thromboembolic stroke, 2 with venous thromboembolism, 3 other (bypass occlusion, mesenteric thrombosis); wound complications: 10 with persistent wound drainage, 7 with wound infection, 5 deaths, 4 with mediastinitis

Patients

Safety Outcomes

(continued) Volume 31 Number 4

April 2009

Table (continued). Authors

Bovine Thrombin Product*

Patients

Safety Outcomes

FloSeal or Gelfoam (both containing Thrombin-JMI)

309 Surgery patients (cardiac, vascular, spinal) randomized to receive one or the other treatment

Total of 384 complications in 144 patients, of which 75 postoperative complications were blood related (not reported separately for the 2 products); hemorrhagic complications: 15 patients with anemia, 9 with excessive postoperative bleeding, 2 with a decreased platelet count, 1 with a transfusion reaction; thromboembolic complications: 7 with arterial thrombosis, 5 with congestive heart failure or stroke, 1 with deep venous thrombosis, 1 with hypercoagulation; wound complications: 1 abscess, 1 seroma; other serious complications: 16 deaths, 11 with infection, 4 with fever, 1 with renal failure; no association between presence/absence of bovine thrombin antibodies and occurrence of bleeding complications at 6- to 8-week follow-up; proportion of patients with blood-related complications was 32% in those with preoperative antibodies to bovine thrombin or bovine factor Va and 26% in those without (P = NS); corresponding proportions in those with and without postoperative antibodies to bovine thrombin or bovine factor Va were 20% and 22% (P = NS)

Carroll et al28

Fibrastat bovine glue

21 Cardiothoracic surgery patients, excluding those with history of sensitivity to bovine products or reaction to intradermal bolus challenge with fibrin glue

No patients developed skin reaction to intradermal skin test; no patients developed postoperative clinical signs or symptoms of an allergic reaction; no patients had postoperative bleeding complications (even at a time when anti–bovine fibrinogen antibody titers were high)

Chapman et al37

Thrombin-JMI vs rhThrombin

411 Surgical patients randomized to receive one or the other treatment

5.0% and 1.5% of the bovine thrombin and human thrombin groups, respectively, had elevated antibodies to either bovine thrombin or human thrombin at baseline; at 1-month follow-up, the respective rates were 21.5% and 1.5%; incidence of any AEs was almost 100% in both groups; incidence of safety events—any serious AE (22% and 18%), treatmentrelated AEs (1% and 3%), incision-site complications (63% and 63%), nausea (35% and 28%), procedural pain (34% and 29%), constipation (25% and 22%), pyrexia (19% and 20%), vomiting (15% and 13%), hypokalemia (14% and 9%), anemia (11% and 14%), peripheral edema (10% and 16%), insomnia (9% and 12%)

683

(continued)

F.A. Ofosu et al.

Winterbottom et al20

Authors

Bovine Thrombin Product*

Wai et al40

Thrombostat

65 Hemodialysis patients exposed to bovine thrombin and 46 unexposed controls (14 hemodialysis patients, 32 blood donors)

Significantly more patients had elevated IgE and IgG antibody levels among those exposed to bovine thrombin compared with those who were unexposed (IgE: 27.7% vs 4.3%, respectively; IgG: 26.2% vs 8.7%; both, P < 0.01); symptoms of rhinoconjunctivitis were more common in hemodialysis patients with an IgE titer than in those without such antibodies (26.3% vs 6.9%, respectively; P = 0.04), but no other differences in allergic symptoms were noted; this analysis was prompted by a bovine thrombin–related case of anaphylaxis

Sands et al41

Thrombogen

88 Hemodialysis patients (number treated with bovine thrombin not specifically reported)

In this cross-sectional study, 27 patients (30.7%) had elevated titers of antibodies to bovine thrombin; those with PTFE grafts or cuffed catheters were significantly more likely to have elevated antibody titers (P < 0.001); those with elevated antibody titers were more likely to have previousaccess thromboses or total thromboses (P < 0.01); rates of thromboses per 100 patient-years of follow-up were 3.1 in those with normal antibody levels and 45.2 in those with elevated antibody levels

Su et al42

Not reported

162 Cardiovascular surgery patients (151 exposed to bovine thrombin, 11 unexposed)

Thromboembolic complications in 12 patients treated with bovine thrombin (6 with stroke, 2 with graft stenosis, 1 with amaurosis fugax, 1 with mesenteric thrombosis, 1 with pulmonary embolism, 1 with thrombophlebitis); 4 of these events occurred within 48 hours after surgery; elevated antibody levels were not significantly related to risk of AEs

Patients

Safety Outcomes

Volume 31 Number 4

CABG = coronary artery bypass graft; GI = gastrointestinal; rhThrombin = recombinant human thrombin; AEs = adverse events; Ig = immunoglobulin; PTFE = polytetrafluoroethylene. *Thrombogen® is a trademark of Ethicon, Inc., Cornelia, Georgia. FloSealTM is a trademark of Baxter International Inc., Deerfield, Illinois. Gelfoam® is a trademark of Pfizer Inc., New York, New York. Thrombin-JMI® is a trademark of Jones Pharma Incorporated, Bristol, Tennessee. Fibrastat bovine glue® is a trademark of Fibratek, Longmont, Colorado. Thrombostat® is a trademark of Pfizer Inc.

Clinical Therapeutics

684 Table (continued).

F.A. Ofosu et al. topical human thrombin (12.7% vs 3.3%, respectively; P < 0.01). All 5 studies cited a general concern about the likelihood of patients treated with topical bovine thrombin developing antibodies to bovine clotting factors that might neutralize their human counterparts. The studies that compared topical human thrombin and topical bovine thrombin supported the greater immunogenicity of bovine thrombin. However, none of the AEs in these studies were attributable to the development of anti–bovine protein antibodies. Thus, the safety profiles of topical bovine thrombin, plasmaderived thrombin, rhThrombin, and standard treatment (cauterization) were comparable in these randomized, double-blind clinical trials.

Case Reports Thirty case reports and case series involving patients treated with topical thrombin products were identified.* The first case of bovine thrombin–induced generation of antibodies that inactivated ≥1 human clotting factor was reported in 1989 by Flaherty et al.10 In 1994, Ortel et al6 described 17 patients (4 from their hospital and 13 reported by other investigators) exposed to topical bovine thrombin during surgery who subsequently developed neutralizing antibodies against several human coagulation factors. Among these patients, there were 6 cases of serious bleeding, 3 cases of minor bleeding, 1 thromboembolic event, and 4 deaths. A number of case reports have described AEs occurring in the surgical setting in association with the use of several formulations of bovine thrombin.12,13,15,26,40,41,43–54 These AEs were considered to be attributable to the generation of antibodies that inhibited ≥1 bovine clotting factor present in topical bovine thrombin preparations and that appeared to cross-react with their human coagulation factor homologues. Specifically, the anti–bovine protein antibodies in question crossreacted with ≥1 of the following human clotting factors: human thrombin, fibrinogen, prothrombin, factor V, factor VII, factor IX, and factor X.12,13,15,26,40,41,43–57 Some authors attributed the occurrence of minor to life-threatening bleeding to the cross-reacting anti– human clotting factor antibodies generated in certain patients.8,17,18,44,47,57 Given the foregoing observations, it appears that the specific topical bovine *References 6, 8–15, 17–19, 26, 40, 41, 43–57.

April 2009

thrombin preparations used in these studies were not homogeneous. There have been 5 case reports suggesting that generation of inhibitors of topical bovine thrombin that cross-react with human factor V or thrombin may cause other serious AEs, including death.13,18,26,47,57 Without careful consideration of the individual reports, concerns arising from them, added to concerns about the possible generation of neutralizing antibodies against bovine clotting factors (that may also inactivate human clotting factors), may lead to inappropriate withholding of anticoagulant therapy from patients who require it. Furthermore, the potential contributions of any postoperative anticoagulants given to patients to prevent postoperative thromboembolism (eg, heparin and/or oral anticoagulants) should be taken into consideration when using clotting tests to evaluate the generation of antibodies thought to prolong clotting test results. It should be kept in mind that no occurrences of thromboembolism, bleeding events, or coagulopathy reported in the double-blind comparisons of rhThrombin and bovine thrombin37,38 or plasma-derived human thrombin and bovine thrombin39 were considered related to the development of antibodies to topical rhThrombin or bovine thrombin.

Generation of Antibodies After Exposure to Topical Bovine Thrombin Thirteen studies have addressed the risk for development of antibodies against the glycoproteins present in topical thrombin preparations.† Some of these studies are summarized in the table. In the crosssectional blood surveillance study by Favaloro et al,54 the total number of patient samples tested was not provided. Several of these studies used nonstandardized assays to measure inhibitory antibodies, making it difficult to compare and/or quantify inhibitory antibodies across studies. Although Fastenau et al58 and Su et al,42 for example, reported mean antibody titers, most of the studies presented antibody outcomes as a binary function. Individual titers were often plotted against an antibody titer scale and were considered elevated if the score was 2 to 3 SDs from a mean pretreatment or control level. The number of times that patients were exposed to bovine thrombin was usually uncertain in these stud†References

4, 16, 20, 21, 28, 37, 39, 40, 42, 54–56, 58.

685

Clinical Therapeutics ies. In the retrospective studies,4,40,54,55 exposure to topical bovine thrombin was based on the detection of antibodies to topical bovine thrombin; a billing entry for bovine thrombin or a history of surgery was generally used as a surrogate for exposure to bovine thrombin. In some prospective studies,16,20,56,58 the amount of bovine thrombin to which particular patients were exposed was not stated. Furthermore, data on the amount of bovine thrombin received are almost impossible to obtain for some products that combine topical bovine thrombin with other hemostatic agents. A study of the nonuniformity of the protein content of bovine thrombin preparations found that 96% of the proteins in Thrombin-JMI consisted of α- and β-thrombin, whereas only ~25% of the protein content of 2 other marketed topical bovine thrombin preparations consisted of these forms of bovine thrombin.22 Overall, the 13 studies noted 22 separate antigens in bovine thrombin preparations. The most common antigens eliciting antibody production were prothrombin, thrombin, collagen, fibrinogen, and factors V, VII, IX, and X. Generation of anti–galactose α1,3 galactose antibodies was also reported.16,21 Fastenau et al58 studied the generation of antibodies to phosphatidylserine, phosphatidylethanolamine, and phosphatidylcholine to examine the hypothesis that antibodies could inhibit the binding of coagulation proteins to cell membranes and thereby inhibit their functions. It was found that all 6 patients exposed to bovine thrombin developed anticardiolipin immunoglobulin G (IgG), and 5 of the 6 patients developed antiphosphatidylethanolamine IgG. Five studies estimated the risk for the production of antithrombin antibodies subsequent to multiple exposures to bovine thrombin in the surgical setting.4,16,20,37,42 Using a hospital billing code for thrombin as a proxy marker of exposure, Dorion et al4 reported a global immunization prevalence estimate of 10% for all patients who were exposed to topical bovine thrombin at least once (95% CI, 4.6–15.4). In 2 cross-sectional studies in dialysis patients, estimates of the mean prevalence of antithrombin IgG were 27%40 and 29%.55 This high prevalence of anti–bovine thrombin antibodies in patients undergoing renal dialysis (a population likely to have repeated exposures to exogenous blood products) was consistent with the observation by Dorion et al that patients with >1 exposure to bovine thrombin were 8 times more likely to 686

have high titers of anti–bovine thrombin antibodies than were patients exposed only once (39% vs 5%, respectively; P < 0.001). In 2 prospective studies, Winterbottom et al20 and Ortel et al16 described cardiac patients with previous surgical exposure to bovine thrombin at baseline and at 4 to 8 weeks after surgery. Overall, the risk of developing antibodies to bovine thrombin after surgery was comparable when this hemostatic agent was added to other formulations (18.0% for FloSeal,20 20.0% for Gelfoam,20 and 20.5% in a sternal paste, Gelfoam collar, or fibrin glue16). When only patients with suspected repeat exposures to topical thrombin were considered, the majority had antibodies to either the whole product or to individual components of the product (eg, antibodies to bovine thrombin [62.5%] and factor V [80.0%]). There are few estimates of the true incidence of development of antibodies to human thrombin after surgical exposure. In the studies reviewed, assays were conducted only in patients with confirmed elevations in antibodies to bovine thrombin. For example, Winterbottom et al20 measured autoantibodies to human thrombin in the 20% of patients with anti–bovine thrombin antibodies and found that 70% of these patients also had autoantibodies to human thrombin. Formulations of topical bovine thrombin used before 2003 probably contained trace amounts of bovine factors V and Va. The coagulation abnormalities associated with exposure to bovine thrombin have been most frequently attributed to anti–bovine factor V antibodies cross-reacting with and neutralizing the human homologue.13,14,26,47,50,52,53,57 Although human factor V is immunogenic, nearly all of the studies reviewed performed assays for the presence of antibodies to human thrombin only in patients already known to have mounted an antibody response to bovine thrombin or bovine factor V. It is not surprising that the incidence of anti–bovine factor V antibodies in this review ranged from 9.7% to 90.7%. The best estimates may thus be derived from the comprehensive evaluation by Su et al42 of stored clinical samples from the study by Ortel et al,16 which were drawn when IgG levels were likely to be highest. It was found that 91.4% of the samples reacted to a bovine thrombin preparation, and most of them had a positive antibody titer to bovine thrombin V and Va (79.5% and 89.4%, respectively). Nearly half of the samples (46.4%) had an antibody response to human Volume 31 Number 4

F.A. Ofosu et al. factor V. Importantly, 8.6% of the samples had a positive antibody titer to human factor V, even though none of these samples had detectable antibodies to bovine factor V.

Safety Studies Eight studies were identified that presented information on the generation of anti–bovine protein antibodies and/or the incidence of AEs in surgical patients who had single or multiple exposures to bovine thrombin.4,16,20,28,37,40,42,55 Some of these studies are summarized in the table. Carroll et al28 conducted a prospective, uncontrolled clinical study in 21 cardiothoracic surgery patients, excluding those with a positive skin test for bovine products. None of the patients developed any signs/symptoms of an allergic reaction postoperatively. In addition, none of the patients developed postoperative bleeding complications. Although the cross-sectional study in 88 hemodialysis patients by Sands et al41 did not specifically report patients’ previous exposure to bovine thrombin, 31% of patients were described as having high levels of antibodies to the constituents of bovine thrombin. A significantly higher rate of thrombosis occurred in patients with high levels of anti–topical bovine thrombin antibodies compared with those having normal antibody levels (13/27 vs 5/61, respectively; P < 0.001). Su et al42 explored the relationship between the occurrence of AEs and the presence of anti–bovine clotting factor antibodies in stored postoperative plasma samples from 162 cardiovascular surgery patients, including 151 patients with topical bovine thrombin exposure from the study by Ortel et al16 and 11 surgical controls who were naive to topical bovine thrombin. Twelve of the 151 exposed patients (7.9%) experienced thromboembolic complications; 4 of these events occurred within 48 hours after surgery. However, these thromboembolic events were not considered related to the generation of antibodies to bovine proteins. In a cross-sectional study, Dorion et al4 examined stored blood samples and hospital records from 120 patients previously treated with bovine thrombin. One hundred two patients had received bovine thrombin once, and 18 had received it multiple times. Of the 102 patients with only 1 previous exposure to bovine thrombin, 5% developed antibodies to bovine thrombin; of the 18 patients with multiple exposures, 39% developed such antibodies (P < 0.001). However, no information on the timing and number of previous April 2009

exposures to bovine thrombin was reported. The incidence of serious bleeding episodes in all patients exposed to bovine thrombin was 2%; in the 12 patients who developed antibodies to bovine thrombin, the incidence was 17%. In an uncontrolled clinical trial, Ortel et al16 studied 151 patients (109 undergoing coronary artery bypass graft [CABG] surgery and 42 undergoing valve surgery) who were exposed to bovine thrombin. They found that 7.3% of patients developed excessive postoperative bleeding or had a reoperation for bleeding, 3.3% required transfusion support after postoperative day 2, and 3.3% developed gastrointestinal bleeding. Whether the patients who developed these AEs were receiving any anticoagulant or antiplatelet medication was not stated. A further 4.6% of patients had a thromboembolic stroke, and 1.3% developed a venous thromboembolism. There were 5 deaths (3.3%). Eight patients who reported having surgery in the past 10 years had a significantly higher incidence of antibodies to bovine thrombin compared with normal donors recruited from laboratory and clinic personnel (63.0% vs 2.0%, respectively; P = 0.01). When Ortel et al16 examined the relationship between previous surgery and the preoperative presence of anti–bovine thrombin antibodies, they found that 27 of the 151 patients in the study had undergone surgery in the past 10 years. Of these, 9 (33.3%) had high preoperative titers of anti–bovine protein antibodies, compared with 6 of the 124 patients (4.8%) with no previous surgery. Thus, previous surgery that involved likely perioperative exposure to bovine thrombin was associated with increased titers of antibodies to bovine antigens. After surgery, the most common IgG antibodies against an exogenous coagulation protein were those directed against bovine factor V (80.7% overall) and bovine factor Va (90.7%). Fewer patients overall had developed antibodies to bovine prothrombin (47.0%) or thrombin (20.5%). Previous surgery was not associated with an increased risk for high titers of antibodies to bovine factor V, factor Va, or prothrombin, although it was associated with high titers of antibodies to topical bovine thrombin (P = 0.02). Four to 8 weeks after surgery in which they had a repeat exposure to topical bovine thrombin, 94% of the 105 patients with normal preoperative levels of anti–bovine thrombin antibodies had developed antibodies to ≥1 bovine coagulation protein. After surgery, 51% of patients had also devel687

Clinical Therapeutics oped antibodies against human coagulation proteins, predominantly anti–human factor V antibodies, anti– human thrombin antibodies, or both. In a logistic regression model, Ortel et al16 found that abnormal results on coagulation tests ex vivo were unrelated to the type of surgery performed, the amount of thrombin used, or the presence of high titers of antibodies to ≥1 bovine protein. Furthermore, the risk of adverse clinical outcomes was unrelated to previous surgery, use of higher amounts of thrombin perioperatively, the presence of antibodies to bovine or human coagulation proteins after surgery, or abnormal ex vivo coagulation test results. Patients with high preoperative titers of antibodies to ≥2 bovine proteins had a significantly higher risk for postoperative AEs compared with those without such titers (11/15 [73.3%] vs 46/136 [33.8%], respectively; P = 0.004). Patients with postoperative elevations in titers of ≥2 antibodies to bovine proteins had a significantly lower risk of AEs compared with those without such elevations (47/136 [34.6%] vs 10/15 [66.7%], respectively; P = 0.023).

DISCUSSION The results of published randomized, double-blind comparisons of bovine thrombin and placebo (ie, thrombinfree hemostats) support the tolerability of bovine thrombin.20,36–39 A randomized comparison of bovine thrombin, plasma-derived human thrombin, and rhThrombin found that for equivalent efficacy, the 3 active hemostats had comparable safety profiles.36–39 None of the randomized studies reviewed found a significant association between postoperative generation of antibodies to bovine proteins and any of the observed AEs. It is likely that a substantial portion of the immune response to antigens present in topical bovine thrombin is directed against glycoprotein contaminants that have the highly immunogenic Forssman antigen in ≥1 of the carbohydrate side chains found in nonhuman mammalian glycoproteins.21 Several case reports have attributed significant AEs to the use of bovine thrombin.* The reported AEs were generally seen in patients who had had multiple surgical exposures to bovine thrombin. The main reason for attributing AEs to repeated exposures to bovine thrombin is the generation of antibodies against bovine proteins, principally bovine factor V/Va and, less *References 4, 8, 10–15, 17–19, 26, 40, 47, 57.

688

often, bovine thrombin, that also cross-react with human factor V/Va and human thrombin, respectively. The production of neutralizing antibodies in patients with repeated exposures to plasma-derived and/or recombinant human factor VIII in patients with severe hemophilia A who synthesize very low (<1% of normal) or no detectable levels of factor VIII protein has been well described.59–65 In these patients, large factor VIII gene deletions, splice-site mutations, nonsense mutations, and factor VIII intron-1 or intron-22 inversions result in nondetectable levels of factor VIII protein in plasma. Hemophilia A patients with these mutations would, therefore, be expected to lack innate immunity to factor VIII, and >40% of these patients develop factor VIII–neutralizing antibodies as a result of their regular periodic treatment with factor VIII concentrates.59–62,64,65 Other mutations found in a second group of patients with severe hemophilia A include small deletions/insertions, missense mutations, and splice-site mutations; these patients are able to synthesize some defective factor VIII molecules. The prevalence of neutralizing anti–factor VIII antibodies in these patients, who may be innately tolerant to factor VIII, is <10%.60,62,64,65 It is important to note that overall, only a minority of patients with severe hemophilia A who receive periodic lifelong infusions of factor VIII (plasma derived or recombinant) develop neutralizing anti–human factor VIII antibodies.59–65 Because there are no reported cases to date of patient reexposures to topical plasma-derived human thrombin or topical rhThrombin, the overall immunogenicity of these topical agents in humans is not yet known. However, there are numerous reports of immune responses associated with multiple patient exposures to recombinant human thrombopoietin,66,67 erythropoietin,68 insulin,69 and other recombinant human hematopoietic growth factor factors.70–72 It is important to remember that patients treated with these recombinant proteins normally synthesize the proteins in question.66–72 Thus, it should be evident that generation of antibodies against therapeutic proteins after repeated exposures is relatively common in association with the appropriate therapeutic use of several approved plasma-derived and recombinant human proteins.4,23,24,59–72 With the exceptions of recombinant thrombopoietin, which is not yet approved, and recombinant factor VIII, reports of the production of neutralizing antibodies directed against recombinant human therapeutic proteins emerged only after the products had been approved for clinical use. Volume 31 Number 4

F.A. Ofosu et al.

CONCLUSIONS It has been reported that ≥30% of patients with previous perioperative exposures to bovine thrombin to facilitate hemostasis develop antibodies to bovine proteins when reexposed to bovine thrombin in subsequent surgeries. Repeated perioperative exposure to bovine thrombin may increase both the prevalence and titers of antibodies to ≥1 protein present in nonhomogeneous topical bovine thrombin products. However, the studies reviewed did not suggest that high preoperative or postoperative levels of anti– bovine protein antibodies are necessarily linked to an increased risk of AEs in surgical patients treated with bovine thrombin.

ACKNOWLEDGMENTS This review was funded by a research grant from King Pharmaceuticals, Cary, North Carolina, to United BioSource Corporation. Matthew Reynolds, Sheila Crean, and Frederick Ofosu are consultants for King Pharmaceuticals.

REFERENCES 1. Jackson MR, MacPhee MJ, Drohan WN, Alving BM. Fibrin sealant: Current and potential clinical applications. Blood Coagul Fibrinolysis. 1996;7:737–746. 2. Alving BM, Weinstein MJ, Finlayson JS, et al. Fibrin sealant: Summary of a conference on characteristics and clinical uses. Transfusion. 1995;35:783–790. 3. von Oppell UO, Zilla P. Tissue adhesives in cardiovascular surgery. J Long Term Eff Med Implants. 1998;8:87–101. 4. Dorion RP, Hamati HF, Landis B, et al. Risk and clinical significance of developing antibodies induced by topical thrombin preparations. Arch Pathol Lab Med. 1998;122:887–894. 5. Lundblad RL, Bradshaw RA, Gabriel D, et al. A review of the therapeutic uses of thrombin. Thromb Haemost. 2004; 91:851–860. 6. Ortel TL, Charles LA, Keller FG, et al. Topical thrombin and acquired coagulation factor inhibitors: Clinical spectrum and laboratory diagnosis. Am J Hematol. 1994;45: 128–135. 7. Lawson JH. The clinical use and immunologic impact of thrombin in surgery. Semin Thromb Hemost. 2006;32(Suppl 1):98–110. 8. Cmolik BL, Spero JA, Magovern GJ, Clark RE. Redo cardiac surgery: Late bleeding complications from topical thrombin-induced factor V deficiency. J Thorac Cardiovasc Surg. 1993;105:222–227; discussion 227–228. 9. Diez-Martin J, Sikkink RA, Gilchrist GS, et al. Development of anti-bovine thrombin antibodies following neurosurgical procedures. Br J Haematol. 1990;74:369–370.

April 2009

10. Flaherty MJ, Henderson R, Wener MH. Iatrogenic immunization with bovine thrombin: A mechanism for prolonged thrombin times after surgery. Ann Intern Med. 1989;111:631–634. 11. Zehnder JL, Leung LL. Development of antibodies to thrombin and factor V with recurrent bleeding in a patient exposed to topical bovine thrombin. Blood. 1990;76: 2011–2016. 12. Cruickshank MK, Black J, Wall W. Development of a factor V and thrombin inhibitor following partial hepatic resection and the use of topical thrombin concentrate. Clin Lab Haematol. 1994;16:187–190. 13. Israels SJ, Israels ED. Development of antibodies to bovine and human factor V in two children after exposure to topical bovine thrombin. Am J Pediatr Hematol Oncol. 1994;16:249–254. 14. Muntean W, Zenz W, Finding K, et al. Inhibitor to factor V after exposure to fibrin sealant during cardiac surgery in a two-year-old child. Acta Paediatr. 1994;83:84–87. 15. Chouhan VD, De La Cadena RA, Nagaswami C, et al. Simultaneous occurrence of human antibodies directed against fibrinogen, thrombin, and factor V following exposure to bovine thrombin: Effects on blood coagulation, protein C activation and platelet function. Thromb Haemost. 1997;77:343–349. 16. Ortel TL, Mercer MC, Thames EH, et al. Immunologic impact and clinical outcomes after surgical exposure to bovine thrombin. Ann Surg. 2001;233:88–96. 17. Christie RJ, Carrington L, Alving B. Postoperative bleeding induced by topical bovine thrombin: Report of two cases. Surgery. 1997;121:708–710. 18. Kajitani M, Ozdemir A, Aguinaga M, et al. Severe hemorrhagic complication due to acquired factor V inhibitor after single exposure to bovine thrombin product. J Card Surg. 2000;15:378–382. 19. Poynton AR, Nelson MC, McCance SE, et al. Bovine thrombin induces an acquired coagulopathy in sensitized patients undergoing revision spinal surgery: A report of two cases. Spine. 2003;28:E221–E223. 20. Winterbottom N, Kuo JN, Nguyen K, et al. Antigenic responses to bovine thrombin exposure during surgery: A prospective study of 309 patients. J Appl Res Clin Exp Ther. 2002;2:1–12. 21. Schoenecker JG, Hauck RK, Mercer MC, et al. Exposure to topical bovine thrombin during surgery elicits a response against the xenogeneic carbohydrate galactose alpha1-3 galactose. J Clin Immunol. 2000;20:434–444. 22. Nelson PA, Powers JN, Estridge TD, et al. Serological analysis of patients treated with a new surgical hemostat containing bovine proteins and autologous plasma. J Biomed Mater Res. 2001;58:710–719. 23. Hermeling S, Crommelin DJ, Schellekens H, Jiskoot W. Structure-immunogenicity relationships of therapeutic

689

Clinical Therapeutics

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

proteins. Pharm Res. 2004;21:897– 903. Wadhwa M, Bird C, Dilger P, et al. Strategies for detection, measurement and characterization of unwanted antibodies induced by therapeutic biologicals. J Immunol Methods. 2003;278:1–17. Astermark J. Overview of inhibitors. Semin Hematol. 2006;43(Suppl 4):S3– S7. Rapaport SI, Zivelin A, Minow RA, et al. 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. Schoenecker JG, Johnson RK, Fields RC, et al. Relative purity of thrombinbased hemostatic agents used in surgery. J Am Coll Surg. 2003;197: 580–590. Carroll JF, Moskowitz KA, Edwards NM, et al. Immunologic assessment of patients treated with bovine fibrin as a hemostatic agent. Thromb Haemost. 1996;76:925–931. Nilsson B, Horne MK III, Gralnick HR. The carbohydrate of human thrombin: Structural analysis of glycoprotein oligosaccharides by mass spectrometry. Arch Biochem Biophys. 1983;224:127–133. Mizuochi T, Yamashita K, Fujikawa K, et al. The carbohydrate of bovine prothrombin. Occurrence of Gal beta 1 leads to 3GlcNAc grouping in asparagine-linked sugar chains. J Biol Chem. 1979;254:6419–6425. Nelsestuen GL, Suttie JW. The carbohydrate of bovine prothrombin. Partial structural determination demonstrating the presence of -galactose residues. J Biol Chem. 1972;247:6096–6102. Galili U, Shohet SB, Kobrin E, et al. Man, apes, and Old World monkeys differ from other mammals in the expression of alpha-galactosyl epitopes on nucleated cells. J Biol Chem. 1988;263:17755–17762. Eckhardt AE, Goldstein IJ, Nelsestuen GL. Bovine prothrombin does

690

34.

35.

36.

37.

38.

39.

40.

41.

42.

not contain alpha-D-galactopyranosyl groups. Arch Biochem Biophys. 1980;199:635–636. Ahrens N, Pruss A, Kähne A, et al. Coexistence of autoantibodies and alloantibodies to red blood cells due to blood transfusion. Transfusion. 2007; 47:813–816. Garratty G. Autoantibodies induced by blood transfusion. Transfusion. 2004;44:5–9. Mathiasen RA, Cruz RM. Prospective, randomized, controlled clinical trial of a novel matrix hemostatic sealant in children undergoing adenoidectomy. Otolaryngol Head Neck Surg. 2004;131:601–605. Chapman WC, Singla N, Genyk Y, et al. A phase 3, randomized, doubleblind comparative study of the efficacy and safety of topical recombinant human thrombin and bovine thrombin in surgical hemostasis. J Am Coll Surg. 2007;205:256– 265. Weaver FA, Lew W, Granke K, et al, for the Study Investigators. A comparison of recombinant human thrombin to bovine thrombin as a hemostatic ancillary in patients undergoing peripheral arterial bypass and arteriovenous graft procedures. J Vasc Surg. 2008;47:1266–1273. Doria C, Fischer CP, Wood CG, et al. Phase 3, randomized, doubleblind study of plasma-derived human thrombin versus bovine thrombin in achieving hemostasis in patients undergoing surgery. Curr Med Res Opin. 2008;24:785–794. Wai Y, Tsui V, Peng Z, et al. Anaphylaxis from topical bovine thrombin (Thrombostat) during haemodialysis and evaluation of sensitization among a dialysis population. Clin Exp Allergy. 2003;33:1730–1734. Sands JJ, Nudo SA, Ashford RG, et al. Antibodies to topical bovine thrombin correlate with access thrombosis. Am J Kidney Dis. 2000; 35:796–801. Su Z, Izumi T, Thames EH, et al. Antiphospholipid antibodies after

43.

44.

45.

46.

47.

48.

49.

50.

51.

52.

surgical exposure to topical bovine thrombin. J Lab Clin Med. 2002;139: 349–356. Adams JD, Jones S, Brost BC. Development of antibodies to topical bovine thrombin after abdominal hysterectomy. A case report. J Reprod Med. 2001;46:909–912. Berguer R, Staerkel RL, Moore EE, et al. Warning: Fatal reaction to the use of fibrin glue in deep hepatic wounds. Case reports. J Trauma. 1991;31:408–411. Berruyer M, Amiral J, Ffrench P, et al. Immunization by bovine thrombin used with fibrin glue during cardiovascular operations. Development of thrombin and factor V inhibitors. J Thorac Cardiovasc Surg. 1993;105:892–897. Caers J, Reekmans A, Jochmans K, et al. Factor V inhibitor after injection of human thrombin (tissucol) into a bleeding peptic ulcer. Endoscopy. 2003;35:542–544. La Spada AR, Skålhegg BS, Henderson R, et al. Brief report: Fatal hemorrhage in a patient with an acquired inhibitor of human thrombin. N Engl J Med. 1995;333:494–497. Rothenberg DM, Moy JN. Anaphylactic reaction to topical bovine thrombin. Anesthesiology. 1993;78:779– 782. Sarfati MR, Dilorenzo DJ, Kraiss LW, Galt SW. Severe coagulopathy following intraoperative use of topical thrombin. Ann Vasc Surg. 2004;18: 349–351. Spero JA. Bovine thrombin-induced inhibitor of factor V and bleeding risk in postoperative neurosurgical patients. Report of three cases. J Neurosurg. 1993;78:817–820. Streiff MB, Ness PM. Acquired FV inhibitors: A needless iatrogenic complication of bovine thrombin exposure. Transfusion. 2002;42:18– 26. Tarantino MD, Ross MP, Daniels TM, Nichols WL. Modulation of an acquired coagulation factor V inhibitor with intravenous immune

Volume 31 Number 4

F.A. Ofosu et al.

53.

54.

55.

56.

57.

58.

59.

60.

61.

globulin. J Pediatr Hematol Oncol. 1997;19:226–231. Zumberg MS, Waples JM, Kao KJ, Lottenberg R. Management of a patient with a mechanical aortic valve and antibodies to both thrombin and factor V after repeat exposure to fibrin sealant. Am J Hematol. 2000;64:59–63. Favaloro EJ, Posen J, Ramakrishna R, et al. Factor V inhibitors: Rare or not so uncommon? A multi-laboratory investigation. Blood Coagul Fibrinolysis. 2004;15:637–647. Sands JJ, Nudo SA, Moore KD, Ortel TL. Antibodies to prothrombin, factor V, and beta2-glycoprotein I and vascular access thrombosis. ASAIO J. 2001;47:507–510. Bänninger H, Hardegger T, Tobler A, et al. Fibrin glue in surgery: Frequent development of inhibitors of bovine thrombin and human factor V. Br J Haematol. 1993;85:528– 532. Muntean W, Zenz W, Edlinger G, Beitzke A. Severe bleeding due to factor V inhibitor after repeated operations using fibrin sealant containing bovine thrombin. Thromb Haemost. 1997;77:1223. Fastenau DR, Wagenknecht DR, Hormuth DA, McIntyre JA. Left ventricular assist system recipients exposed to bovine thrombin preparations have a higher frequency of antiphospholipid antibodies than nonexposed recipients. ASAIO J. 2001;47:537–540. Roberts HR, Cromartie R. Overview of inhibitors to factor VIII and IX. Prog Clin Biol Res. 1984;150:1–18. Boekhorst J, Lari GR, D’Oiron R, et al. Factor VIII genotype and inhibitor development in patients with haemophilia A: Highest risk in patients with splice site mutations. Haemophilia. 2008;14:729–735. Goudemand J. Inhibitor development in haemophilia A: The role of von Willebrand factor/factor VIII concentrates. Haemophilia. 2007;13 (Suppl 5):47–51.

April 2009

62. Schwaab R, Brackmann HH, Meyer C, et al. Haemophilia A: Mutation type determines risk of inhibitor formation. Thromb Haemost. 1995;74: 1402–1406. 63. Kaveri SV, Dasgupta S, Andre S, et al. Factor VIII inhibitors: Role of von Willebrand factor on the uptake of factor VIII by dendritic cells. Haemophilia. 2007;13(Suppl 5):61– 64. 64. Kemball-Cook G, Tuddenham EG, Wacey AI. The factor VIII Structure and Mutation Resource Site: HAMSTeRS version 4. Nucleic Acids Res. 1998;26:216–219. 65. Oldenburg J, El-Maarri O, Schwaab R. Inhibitor development in correlation to factor VIII genotypes. Haemophilia. 2002;8(Suppl 2):23–29. 66. Basser RL, O’Flaherty E, Green M, et al. Development of pancytopenia with neutralizing antibodies to thrombopoietin after multicycle chemotherapy supported by megakaryocyte growth and development factor. Blood. 2002;99:2599–2602. 67. Frost H. Antibody-mediated side effects of recombinant proteins. Toxicology. 2005;209:155–160.

68. Casadevall N, Nataf J, Viron B, et al. Pure red-cell aplasia and antierythropoietin antibodies in patients treated with recombinant erythropoietin. N Engl J Med. 2002;346: 469–475. 69. Van Haeften TW. Clinical significance of insulin antibodies in insulintreated diabetic patients. Diabetes Care. 1989;12:641–648. 70. Freund M, von Wussow P, Diedrich H, et al. Recombinant human interferon (IFN) alpha-2b in chronic myelogenous leukaemia: Dose dependency of response and frequency of neutralizing anti-interferon antibodies. Br J Haematol. 1989;72: 350–356. 71. Steis RG, Smith JW II, Urba WJ, et al. Resistance to recombinant interferon alfa-2a in hairy-cell leukemia associated with neutralizing antiinterferon antibodies. N Engl J Med. 1988;318:1409–1413. 72. Giovannoni G, Munschauer FE III, Deisenhammer F. Neutralising antibodies to interferon beta during the treatment of multiple sclerosis. J Neurol Neurosurg Psychiatry. 2002; 73:465–469.

Address correspondence to: Matthew W. Reynolds, PhD, United BioSource Corporation, 430 Bedford Street, Suite 300, Lexington, MA 02420. E-mail: [email protected] 691