Prothrombin complex concentrate (Beriplex® P/N) for emergency anticoagulation reversal: a prospective multinational clinical trial

Authors


Ingrid Pabinger, Universitätsklinik für Innere Medizin I, Klinische Abteilung für Hämatologie und Hämostaseologie, Währinger Gürtel 18-20 A-1090, Wien, Austria.
Tel.: +43 1 40400 4952; fax: +43 1 40402 6930. E-mail: ingrid.pabinger@meduniwien.ac.at

Abstract

Summary. Background: Prothrombin complex concentrate (PCC) can substantially shorten the time needed to reverse antivitamin K oral anticoagulant therapy (OAT). Objectives. To determine the effectiveness and safety of emergency OAT reversal by a balanced pasteurized nanofiltered PCC (Beriplex® P/N) containing coagulation factors II, VII, IX, and X, and anticoagulant proteins C and S. Patients and methods: Patients receiving OAT were eligible for this prospective multinational study if their International Normalized Ratio (INR) exceeded 2 and they required either an emergency surgical or urgent invasive diagnostic intervention or INR normalization due to acute bleeding. Stratified 25, 35, or 50 IU kg−1 PCC doses were infused based on initial INR. Study endpoints included INR normalization (≤1.3) by 30 min after PCC infusion and hemostatic efficacy. Results: Forty-three patients, 26 requiring interventional procedures and 17 experiencing acute bleeding, received PCC infusions at a median rate of 7.5 mL min−1 (188 IU min−1). At 30 min thereafter, INR declined to ≤1.3 in 93% of patients. At all postinfusion time points through 48 h, median INR remained between 1.2 and 1.3. Clinical hemostatic efficacy was classified as very good or satisfactory in 42 patients (98%). Prompt and sustained increases in circulating coagulation factors and anticoagulant proteins were observed. One fatal suspected pulmonary embolism in a patient with metastatic cancer was judged to be possibly PCC-related. Conclusions: PCC treatment serves as an effective rapid hemorrhage control resource in the emergency anticoagulant reversal setting. More widespread availability of PCC is warranted to ensure its benefits in appropriate patients.

Introduction

Long-term coumarin oral anticoagulant therapy (OAT) is often prescribed for primary and secondary prevention of arterial and venous thromboembolism (e.g. in patients with atrial fibrillation, mechanical heart valves or deep vein thrombosis). The primary complication of OAT with coumarin is bleeding, which can be potentially fatal if manifested as intracranial hemorrhage (ICH) [1]. In large-scale epidemiological studies involving a total of 9065 patients receiving OAT, the annual incidence of major bleeding complications has ranged from 1.1% to 1.5% [2,3]. Sites accounting for the highest proportions of such complications were gastrointestinal (30–60%) and intracranial (17–30%) [2,3]. Half of the patients with OAT-associated ICH die within 30 days [4].

Rapid reversal of OAT is crucial in cases of acute major bleeding or emergency surgery with elevated international normalized ratio (INR) [5]. Coumarins inhibit the maturation of functional vitamin K-dependent coagulation factor (F) II, FVII, FIX, and FX, and hence cause a functional deficiency of these proteins. They also produce a functional deficit of anticoagulant proteins C and S.

In recent clinical guidelines, either prothrombin complex concentrate (PCC) or fresh frozen plasma (FFP) has been recommended for rapid reversal of OAT [5–10]. The most recent update from the British Committee for Standards in Haematology recommends that, for reversal of anticoagulation in patients with major bleeding, PCC should be administered in preference to FFP [10].

PCC can be administered rapidly without the need for matching the blood group or thawing the product, and has been shown in a number of studies to reverse warfarin-related coagulopathy more rapidly than FFP [11–13]. For an average adult weighing 70 kg, the recommended volume of FFP is 1050 mL [14], and for most patients approximately 2000 mL or more would be required as an optimal dose [15]. Transfusion of such large volumes of FFP may lead to fluid overload and pulmonary edema [16]. FFP increases the risk of acute lung injury in critically ill patients [17]. Moreover, unlike most FFP preparations, PCC is prepared using viral inactivation methods [18,19]. Finally, the concentration of the coagulation factors in FFP is less predictable, and some units do not contain sufficient levels of the four coagulation factors to correct the coagulation defect [20].

Beriplex® P/N, a biochemically well-characterized balanced PCC derived from plasma screened by polymerase chain reaction (PCR) and prepared using pasteurization and nanofiltration, contains FII, FVII, FIX, and FX, as well as the anticoagulant proteins C and S [18]. In a clinical pharmacokinetic study of healthy volunteers, infusion of 50 IU kg−1 Beriplex® P/N increased median circulating coagulation factor and anticoagulant protein concentrations by 59–158% within 5 min [21]. Among patients requiring emergency surgery and those experiencing major bleeding, Beriplex® P/N has been shown to reverse anticoagulation by warfarin or phenprocoumon rapidly, effectively, and safely [22–26]. This PCC has also been found effective in managing patients with critical illness or severe liver disease involving depletion of vitamin K-dependent coagulation factors [27,28].

The present prospective multinational study was designed to evaluate the efficacy and safety of stratified Beriplex® P/N dosing in rapidly normalizing elevated INR and achieving hemorrhage control. Thus far, anticoagulant reversal has been most extensively investigated in patients treated with warfarin, and there continues to be a need for further data pertaining to other widely used coumarins (e.g. phenprocoumon). The present study was also intended to provide such data. Finally, in most prior investigations on OAT reversal follow-up has been short [23], and the present study is among the first to delineate the long-term time course of change in INR, coagulation factors, anticoagulant proteins, and thrombogenicity markers following PCC infusion.

Patients and methods

This single-arm prospective study was conducted from October 2005 to November 2006 at 15 centers in Austria, Germany, Hungary, Israel, Lithuania, the Netherlands, Poland, and Switzerland. The study was performed in accordance with the International Conference on Harmonization Good Clinical Practice guidelines, the 1996 Declaration of Helsinki, and standard operating procedures for clinical research and development of the study sponsor (CSL Behring GmbH, Marburg, Germany) and the Clinical Research Organization responsible for local management (Parexel International GmbH, Berlin, Germany). The study protocol was approved by the ethics committees or Institutional Review Boards at all participating centers and by the regulatory authorities of the corresponding countries.

Informed written consent was a prerequisite for study entry. If a candidate were unable to grant consent under emergency circumstances and neither commercial Beriplex® P/N nor a comparable licensed product were available, then consent could be secured from a legal representative or family member of the patient or, depending on local law, a physician other than a study investigator. One patient with traumatic subdural bleeding was unconscious upon admission, and informed consent was obtained in writing from a physician not involved in the study. All other study patients rendered informed written consent on their own behalf.

Based on previous clinical studies of Beriplex® P/N, a target enrollment of at least 40 patients was selected. This sample size was expected to ensure an exact 95% confidence interval width of <20% around the proportion of patients attaining INR normalization, on the assumption that this proportion would be at least 90%.

Patient selection

Patients ≥18 years old under OAT with an INR >2 were eligible for study entry if they required (i) an emergency surgical or urgent invasive diagnostic intervention or (ii) INR normalization as a result of acute bleeding. Exclusion criteria included: treatment with any other investigational drug within 30 days before study entry; exposure to whole blood, plasma or plasma fractions over the prior 2 weeks; hypersensitivity to PCC constituents; hereditary protein C deficiency; life expectancy <3 months; <2 weeks of stable oral anticoagulation among patients with a recent history of deep vein thrombosis or pulmonary embolism; and acute ischemic cardiovascular disorder, disseminated intravascular coagulation or sepsis. Pregnancy and current or planned breast feeding were also reasons for exclusion.

Endpoints

The primary study endpoint was normalization of INR (≤1.3) at 30 min after the end of PCC infusion. INR is the standard test routinely performed worldwide and recommended in clinical guidelines for monitoring patient response to OAT [5–10]. Prespecified secondary endpoints were clinical hemostatic efficacy in stopping acute bleeding or preventing major bleeding during interventional procedures and the response and in vivo recovery of FIX, FII, FVII, FX, and proteins C and S. Physician rating categories for clinical hemostatic efficacy were: very good, satisfactory, questionable, and none. In patients with acute hemorrhage, ratings of very good, satisfactory, questionable, and none, respectively, denoted: prompt cessation of existing bleeding and/or a rapid decline in INR; >1–2 h delay in bleeding cessation and INR decrease; >2 h delay in bleeding cessation and INR decrease; and absence of effect on bleeding and INR. The corresponding definitions among patients undergoing interventional procedures were: normal hemostasis during the procedure; mildly abnormal intraprocedural hemostasis as judged by quantity or quality of blood loss (e.g. slight oozing); moderate abnormality in intraprocedural hemostasis (e.g. controllable bleeding); and severe hemostatic abnormality during the procedure (e.g. severe refractory hemorrhage). Although a formally validated and widely accepted scale has yet to be established to assess bleeding among recipients of anticoagulant therapies [29], the rating system adopted in the present study is closely similar to the systems employed in prior clinical trials of PCC infusion for reversal of OAT [26,30].

Treatment

Prior to PCC infusion, most patients received vitamin K at the standard dosage for each center. Further vitamin K dosages were given in accord with individual study center practice, for instance, every 12 h. One 25, 35, or 50 IU kg−1 body weight dose of PCC (Beriplex® P/N; CSL Behring) was administered to patients with baseline INR of 2–3.9, 4–6 or >6, respectively. Concomitant therapy with whole blood, plasma or plasma fraction within the first 30 min after PCC infusion was not allowed unless urgently required as judged by the attending investigator.

Assessments

Blood samples were drawn for determination of INR, coagulation factors (FIX, FII, FVII, FX), anticoagulant proteins C and S, and thrombogenicity markers prior to infusion and at intervals of 0.5, 1, 3, 6, 12, 24, and 48 h afterward. Baseline INR was to be assessed no more than 2 h prior to PCC infusion. Hematology parameters were determined at baseline and 1, 3, and 24 h, and vital signs at the same time points plus at 30 min. Viral exposure was evaluated at baseline and 7–10, 30, and 90 days postinfusion. The occurrence of any adverse events was monitored until the first virus safety follow-up at 7–10 days postinfusion and of serious adverse events up to 90 days. The relationship of adverse events to PCC infusion was determined by the attending investigator subject to the review and concurrence of the coordinating investigators (Profs. Pabinger and Ostermann).

Patient assessments included: medical history; thorough physical examination; determination of vital signs and hematology parameters; enzyme immunoassays of prothrombin activation fragments 1 + 2 (F1+2), thrombin–antithrombin complex (TAT), D-dimer, antibodies to human immunodeficiency virus (HIV)-1 and -2, hepatitis B core antigen (HBc) and hepatitis C virus (HCV), IgG/IgM antibodies against hepatitis A virus (HAV) and parvovirus B19 (B19V), and hepatitis B surface antigen (HBsAg).

Laboratory tests

The INR and the hematology parameters were measured at the local laboratories of the study centers. Central laboratories performed assays for coagulation factors, anticoagulant proteins and thrombogenicity markers (Laboratory Uwe Kalina, CSL Behring) and tests for viral exposure (Laboratory Prof. Seelig and colleagues, Karlsruhe, Germany).

Statistical analysis

Study data were analyzed using R version 2.5.0 statistical software (R Foundation for Statistical Computing, Vienna, Austria). The analysis was descriptive. The response and in vivo recovery of the coagulation factors and anticoagulant proteins were calculated using the maximum plasma level measured during the 3 h after PCC infusion. Calculated summary statistics consisted of the mean, median, standard deviation (SD), and interquartile range (IQR).

Results

Forty-four patients were enrolled. One patient was excluded from the analysis of results, because before PCC treatment he withdrew consent to participate in the study. The withdrawn patient was the only study enrollee at one center. Among the remaining 14 centers with at least one patient completing the study, eight centers accounted for 33/43 (77%) participating patients.

The baseline characteristics of the 43 study patients (intent-to-treat population) are summarized in Table 1 and indications for PCC administration in Table 2. The long-acting agent phenprocoumon was the oral anticoagulant medication in use by 17 patients (40%), while the remaining patients were under treatment with shorter-acting coumarins. Interventional procedures were performed in 26 patients, and 17 patients were experiencing acute bleeding.

Table 1.   Baseline patient data
Characteristic*Descriptive statistics
  1. INR, International Normalized Ratio; IQR, interquartile range; n, number of patients; SD, standard deviation. *All patients were Caucasian. †An additional indication in one of these patients was cardiac aneurysm. ‡One case each of: aortic bypass; myocardial ischemia; thrombophlebitis; vascular stent insertion; venous embolism; ill-defined disorder. §Four patients also received concomitant acetylsalicylic acid and two heparin. ¶Diagnosis prompting oral anticoagulation therapy.

Gender, n (%)
 Female22 (51)
 Male21 (49)
Age (years), n (%)
 <60 6 (14)
 60–6911 (25)
 70–7918 (42)
 ≥80 8 (19)
Body mass index (kg m−2), mean (SD)26.7 (4.5)
Body weight (kg), mean (SD)76.3 (14.8)
INR, n (%)
 <426 (61)
 4–6 7 (16)
 >610 (23)
Indication for oral anticoagulation, n (%)
 Atrial fibrillation19 (44)
 Thrombosis 8 (19)
 Pulmonary embolism 4 (9)
 Heart valve replacement 3 (7)
 Myocardial infarction† 3 (7)
 Other‡ 6 (14)
Oral anticoagulant medication,§ n (%)
 Acenocoumarol17 (40)
 Phenprocoumon17 (40)
 Warfarin 9 (20)
Time since diagnosis¶ (years); median (IQR) 3.8 (0.6–8.2)
Table 2.   Indication for prothrombin complex concentrate administration
Indicationn (%)
  1. n, number of patients. *One case each of: trepanation; lumbar puncture; cardiac pacemaker implantation; hip fracture surgery. †One case in each of the following bleeding categories: bladder; hemarthrosis; intracranial; nasal; peritoneal; subdural.

Interventional procedure
 Incision and drainage5 (12)
 Vascular surgery5 (12)
 Appendectomy4 (9)
 Herniotomy/hernioplasty4 (9)
 Abdominal surgery2 (5)
 Cholecystectomy2 (5)
 Other*4 (9)
 All26 (60)
Acute bleeding
 Gastrointestinal8 (19)
 Subcutaneous/intramuscular3 (7)
 Other†6 (14)
 All17 (40)

PCC infusion

In accordance with the baseline INR value (Table 1), single PCC doses of 25, 35, and 50 IU kg−1 were selected, respectively, for 26, 7, and 10 patients. PCC was infused in a median volume of 90 mL (IQR 80–105 mL) over a median of 12 min (IQR 10–18 min). The median rate of infusion was 7.5 mL min−1 (IQR 5.8–8.2 mL min−1), equivalent to 188 IU min−1 (IQR 146–208 IU min−1). In six patients (14%) PCC was infused at ≥10 mL min−1 (≥250 IU min−1).

Vitamin K

Thirty-eight patients (88%) received concomitant vitamin K. The route of administration was intravenous (i.v.) in 33 patients, per oral (p.o.) in four patients, and subcutaneous (s.c.) in one patient, and the doses were 5, 10, and 20 mg in three, 31, and four patients, respectively. The baseline INR was ≤6 in all five patients who did not receive concomitant vitamin K, and acute bleeding was absent in four of the five.

INR

At 30 min postinfusion, INR declined to ≤1.3 in 40 patients (93%), as shown in Fig. 1. In the remaining three patients, INR was 1.4 at this time point.

Figure 1.

 International Normalized Ratio (INR) values at baseline and 30 min post prothrombin complex concentrate (PCC) infusion in individual patients (A) undergoing interventional procedures (open circles) or (B) experiencing acute bleeding (closed circles). Dashed lines show medians and dotted line an INR of 1.4.

Median INR declined to 1.2 at 30 min from 3.2 at baseline. Throughout the 48 h observation period INR remained stable, with the median value fluctuating between 1.2 and 1.3 at all postinfusion time points (Fig. 2). Median INR in the subset of patients who had been receiving long-acting phenprocoumon closely coincided with that of the entire study population at all postinfusion time points.

Figure 2.

 Time course of change in International Normalized Ratio (INR). Boxes span the interquartile range (IQR). Horizontal lines bisecting the boxes indicate median values and lower and upper error bars the 10th and 90th percentiles, respectively.

Hemostatic efficacy

Clinical hemostatic efficacy was very good in 40 patients (93%). In two patients, hemostatic efficacy was judged satisfactory. In one patient with a malignant bladder tumor, a rating of questionable was assigned because of persistent postinfusion bladder bleeding, probably from malignant epithelium. That patient entered the study with acute bladder bleeding and did not undergo an emergency surgical or urgent invasive diagnostic intervention.

Coagulation factors

Markedly increased plasma concentrations of FIX, FII, FVII, and FX were apparent at 30 min following PCC infusion (Fig. 3). Normal or nearly normal concentrations of all four coagulation factors persisted throughout the 48 h observation period. Coagulation factor response and in vivo recovery data are presented in Table 3. The median postinfusion FIX response was 1.29% (IU kg−1)−1. Corresponding medians for the other three coagulation factors ranged from 1.65% to 1.90% (IU kg−1)−1.

Figure 3.

 Time course of change in circulating concentrations of (A) factor (F) IX, (B) FII, (C) FVII, and (D) FX expressed as a percent of normal. Graphic conventions as in Fig. 2.

Table 3.   Response and in vivo recovery
ParameterMedian (interquartile range)
Response [% (IU kg−1)−1]In vivo recovery [% (IU mL−1)−1]
Factor IX1.29 (1.01–1.60)51.3 (41.0–65.1)
Factor II1.90 (1.48–2.28)74.3 (61.6–87.7)
Factor VII1.65 (1.15–1.94)58.7 (50.6–72.6)
Factor X1.84 (1.59–2.15)74.9 (64.4–81.5)
Protein C2.05 (1.75–2.32)81.6 (74.7–91.4)
Protein S2.33 (1.89–2.93)95.1 (76.8–111.2)

Anticoagulant proteins

Circulating concentrations of anticoagulant proteins C and S displayed temporal changes closely matching those of the coagulation factors (Fig. 4). After prompt major increases at 30 min, normal or almost normal concentrations were sustained for the rest of the observation period. The median response exceeded 2.0% (IU kg−1)−1 for both proteins C and S (Table 3).

Figure 4.

 Time course of change in circulating concentrations of (A) protein C and (B) protein S expressed as a percent of normal. Graphic conventions as in Fig. 2.

Thrombogenicity markers

Plasma F1+2 concentration exhibited an immediate but transient increase to a median of 6.37 nmol L−1 (IQR 5.07–8.52 nmol L−1) at 30 min compared with the baseline median of 0.19 nmol L−1 (IQR 0.11–0.29 nmol L−1), as shown in Fig. 5. The increase had largely abated by 6 h, and circulating levels remained stable thereafter. The temporal plasma F1+2 profile was similar among patients undergoing procedural interventions and in the group with acute bleeding. The increase likely reflected infused exogenous F1+2 rather than an endogenous prothrombotic state, as the administered PCC contains F1+2 at a mean concentration of 1.00 μmol L−1 (SD 0.21 μmol L−1, = 8). While the final PCC product is devoid of detectable thrombin activity, some thrombin generation does occur during the manufacturing process.

Figure 5.

 Time course of change in circulating concentrations of (A) prothrombin activation fragments 1 + 2 (F1+2), (B) thrombin–antithrombin complex (TAT) and (C) D-dimer. Values below assay sensitivity imputed as upper limit of detection ×0.5. Graphic conventions as in Fig. 2.

A more prolonged postinfusion rise was observed in circulating TAT concentration (Fig. 5), which did not subside until after 6 h. The median peak level was higher in the acute bleeding group (20.6 μg L−1; IQR 9.54–50.5 μg L−1) than in the interventional procedure group (7.28 μg L−1; IQR 2.96–12.1 μg L−1). Postinfusion changes in plasma D-dimer concentration (Fig. 5) and hematology parameters (Table 4) were relatively minor.

Table 4.   Hematology parameters
Time (h)Median (interquartile range)
Hemoglobin (mmol L−1)HematocritPlatelets (×109 L−1)
Baseline7.17 (5.88–8.68)0.36 (0.28–0.42)252 (198–287)
 16.83 (4.96–8.07)0.34 (0.24–0.39)235 (191–304)
 36.64 (5.30–8.19)0.33 (0.26–0.40)228 (176–260)
246.52 (6.03–7.73)0.33 (0.29–0.39)208 (161–246)

Safety

In no patient was there evidence of viral transmission. In 25 patients (58%) adverse events occurred, including two suspected thromboembolic complications.

In six patients the adverse events were classified as serious, and three of these patients died as a result. One serious adverse event was categorized as possibly related to PCC administration, while all other serious and non-serious events were judged to be unrelated.

The single possibly related serious event was a suspected pulmonary embolism resulting in the death of a 70-year-old man who entered the study requiring reversal of phenprocoumon anticoagulation to control acute bleeding resulting from perforation of stomach cancer. The patient was at increased risk for thrombosis because of the presence of metastatic gastrointestinal cancer and atrial fibrillation. This patient also displayed evidence of coagulation activation before PCC treatment, namely, marked baseline plasma elevations in both TAT (35.8 μg L−1) and D-dimer (2905 μg L−1). After the initial 50 IU kg−1 PCC infusion and 10 mg concomitant vitamin K i.v., the INR of this patient diminished from 3.1 to 1.3. INR rebounded to 2.2 after 48 h, and subsequently daily vitamin K treatment was started. Although no bleeding was observed, INR rebound prompted infusion of 20 IU kg−1 Beriplex® P/N on day 4. Two hours thereafter the patient experienced shortness of breath and died. An autopsy was not performed; however, on clinical grounds the cause of death was judged to be pulmonary embolism.

Among the remaining five patients with serious adverse events, the events were judged to be unrelated to PCC infusion in all cases. Two of the five expired: a 79-year-old male due to cardiovascular causes, and a female aged 75 years as a result of acute respiratory failure, renal failure, and empyema. The three survivors were a female 74 years old who developed left crus peripheral arterial embolism and middle right cerebral artery embolism, a 70-year-old man with gastric cancer diagnosed after study admission, and a female aged 67 years who experienced duodenal ulcer hemorrhage.

The 79-year-old male non-survivor presented with traumatic subdural bleeding after a fall and received a single 25 IU kg−1 PCC infusion. The patient had been under OAT because of pulmonary embolism 7 months prior. Atrial fibrillation was a co-morbidity. During the course of the hospital stay he was febrile (body temperature 40 °C) and exhibited septic circulatory dysfunction, which was managed with vasopressive medication. Cardiovascular failure was diagnosed approximately 24 h after PCC infusion, at which time INR was 1.5 and D-dimer had decreased to 2617 μg L−1 from a markedly elevated preinfusion baseline value of 4900 μg L−1. D-dimer declined further to 1100 μg L−1 at 48 h. There were no signs of neurological recovery, and accordingly the treating physicians abstained from any further intensification of therapy. The patient died in cardiovascular failure 6 days after PCC infusion. A postmortem examination revealed some coronary sclerosis but no conspicuous cardiovascular findings indicative of a thromboembolic event.

The 75-year-old female non-survivor required urgent surgical pleural effusion drainage. Multiple co-morbidities and risk factors existed, namely, chronic heart failure, cardiomyopathy, prior pulmonary embolism, non-insulin-dependent diabetes mellitus, hypertension, stable angina pectoris, past malignant corpus uteri neoplasm, and morbid obesity. She received a single 25 IU kg−1 PCC infusion. Acute respiratory failure, renal failure, and empyema developed the next day, and the patient expired on the second postinfusion day. INR was 1.5 at 12 h after infusion and 1.6 at 24 h. D-dimer levels at 12 h (162 μg L−1) and 24 h (165 μg L−1) displayed negligible change from the baseline value (155 μg L−1). No postmortem examination was performed. According to the responsible investigator, a working diagnosis of severe infection was founded upon the biochemical profile of the pleural fluid, which was compatible with empyema (exudate with leukocytosis and pH 7.1). There were no signs of acute myocardial ischemia. There was additionally no clinical suspicion of pulmonary embolism and no significant rise in D-dimer. Heparin had also been administered, and INR 48 h after PCC infusion was 1.8, indicating that the patient was well anticoagulated.

The 74-year-old female survivor had been prescribed OAT to prevent thromboembolic complications of atrial fibrillation. Her history included arterial embolism and phlebothrombosis. Reversal of OAT was needed to manage acute left crus bleeding, and a single 50 IU kg−1 PCC dose was infused. At 6 h after PCC infusion, a transient D-dimer increase was observed to 463 μg L−1 from 118 μg L−1 at baseline; however, the D-dimer level promptly subsided to 165 μg L−1 by 12 h. Concomitant heparin was commenced the day after infusion and continued for 6 days. OAT was restarted 6 days post-PCC infusion. INR was maintained in a therapeutic range of 2–3 from 4 days after infusion onward. Left crus peripheral arterial embolism and middle right cerebral artery embolism developed respectively 4 and 7 days postinfusion.

One suspected thromboembolic complication, namely, mild leg pain and edema in a patient with a history of deep vein thrombosis, was classified as non-serious and not related to PCC. Color duplex sonography did not reveal any evidence of a recent thrombotic process. Deep vein thrombosis had been originally diagnosed in this patient 5 years prior to study entry. The location of the previous thrombosis was not documented. The symptoms resolved without sequelae through application of compressive bandages alone. The leg symptoms and history of deep vein thrombosis had aroused the clinical suspicion of a thromboembolic event, which was not, however, confirmed by sonographic imaging.

Discussion

This prospective study demonstrated prompt INR reduction after Beriplex® P/N infusion accompanied in nearly all cases by clinical hemostatic efficacy. In OAT reversal, the goal has been to restore INR to the normal level or slightly above [12,31–33]. In the present study, Beriplex® P/N infusion was successful in rapidly returning INR to 1.4 or below in all patients. These results were accomplished in a patient population with pretreatment INR values ranging widely from 1.9 to 17.4. Furthermore, Beriplex® P/N provided effective clinical hemorrhage control in 98% of patients. In another study of emergency OAT reversal with Beriplex® P/N, all eight patients achieved an INR of 1.0–1.4 by 10 min postinfusion [26]. In a third study involving Beriplex® P/N infusion for warfarin reversal, INR decreased within 20 min to below 1.3 in 79% of 42 patients with pretreatment INR as high as 27.6, while the postinfusion INR range was 1.3–1.9 among the remaining 21% of patients [23]. Other studies have also demonstrated the utility of PCC infusion for precise INR targeting [11,12,30].

Comparable precision has not been possible with FFP infusion. In a controlled study of emergency OAT reversal, the INR range attained by FFP infusion (1.6–3.8) was far broader than that by PCC (1.0–1.4) [11]. A recent randomized trial of patients undergoing cardiac surgery has also shown PCC to be more rapid and effective than FFP for OAT reversal [31]. In some cases, INR correction by FFP was inadequate.

The optimum dose of PCC for OAT reversal has not been established. Whether or not dosage should be individualized has been debated. A uniform standard dose of 30 IU kg−1 concomitant with 5 mg intravenous vitamin K has been advocated for use in OAT patients with major hemorrhage [25]. Among 19 patients receiving that dose of Beriplex® P/N, the median attained INR was 1.4 with a range of 1.2–2.2. In the present study, a narrower range (1.0–1.4) was produced by dosing on the basis of initial INR value, an approach also adopted in a previous study of Beriplex® P/N [23]. Additionally, individualized PCC dosing based on initial INR was more effective than a standard uniform dose for attaining target INR in a recent randomized trial [34].

A major advantage of PCC over FFP is increased speed of OAT reversal. For instance, Beriplex® P/N is stored at room temperature and hence suited well for fast reconstitution. PCC infusion can be performed rapidly, allowing prompt INR normalization and undelayed stabilization of patients. Numerous studies of PCC in OAT reversal have shown target INR to be reached within 10–15 min [11–13,26,30,34,35]. In one recent report, the target range was achieved as early as 3 min [33]. By contrast, INR normalization with FFP required a median of 30 h among patients with warfarin-associated ICH [16]. This protracted time period was attributed by the investigators to delay in initiating FFP administration and the lengthy infusion process.

In PCC, the concentration of the vitamin K-dependent coagulation factors is approximately 25 times that in plasma, and consequently a markedly lower fluid volume is needed for OAT reversal as compared with FFP [15]. In the present study, the median PCC volume infused was 90 mL over a median of 12 min. For comparable therapeutic effect, an FFP infusion would surpass 2000 mL [15].

This is the first major clinical study of OAT reversal in which all three of the most common oral anticoagulant drugs were well represented. While warfarin is used almost exclusively for OAT in the UK and North America, acenocoumarol and phenprocoumon are popular elsewhere, especially in Europe. In Germany, 95% of patients requiring OAT are treated with phenprocoumon [14]. Between the three agents, marked differences in half-life exist that might theoretically affect the prospects for sustained OAT reversal or necessitate a modified dosing strategy. The average half-lives of acenocoumarol, warfarin, and phenprocoumon (11, 40, and 140 h, respectively) span a 13-fold range [36]. Nevertheless, in this study OAT reversal succeeded without exception among recipients of all three anticoagulant medications. Furthermore, during the 48 h observation period after Beriplex® P/N infusion, in most cases with co-administration of vitamin K, median INR remained in the desired target range even in the patients who had been receiving long-acting phenprocoumon.

Beriplex® P/N is a balanced PCC containing both coagulation factors and anticoagulant proteins. This balance may contribute to its demonstrated low thrombotic risk [21–26,37]. In addition to FII, FVII, FIX, and FX, major constituents of Beriplex® P/N include proteins C and S. The observed coagulation factor and anticoagulant protein responses following Beriplex® P/N infusion among patients requiring OAT reversal in the present study were similar to those of healthy volunteers in a recent pharmacokinetic study [21].

Thrombogenicity marker elevations have been frequently observed [23,26,30,35,38]. In the present study, a transient plasma F1+2 increase probably reflected the F1+2 content of the PCC [21]. Circulating TAT concentration also exhibited a transitory rise. Only minor changes were noted in plasma D-dimer concentration, thus indicating the absence of any persistent coagulation activation following Beriplex® P/N administration.

While it is imperative that anticoagulation be reversed as quickly as possible in potentially life-threatening bleeding or emergency interventions, this population of patients is at high risk for thrombotic events, necessitating caution that the reversal be carefully controlled. Thromboembolic events have seldom been reported in patients receiving PCC for reversal of OAT [13,19,30,33,34,39]. In the present study, one possibly PCC-related thromboembolic event occurred after a second infusion of PCC on the fourth day in a patient who had also received multiple doses of vitamin K. Underlying thrombotic risk factors in this patient were metastatic gastrointestinal cancer and atrial fibrillation, and pretreatment thrombogenic marker abnormalities indicated activation of coagulation prior to the first PCC infusion.

There has been some concern that excessively rapid infusion of concentrated coagulation factors in PCC might increase the risk of thromboembolic events. A PCC infusion rate limit of 1 mL min−1 has been advised [40]. However, such a slow rate, and the attendant inevitable delay in OAT reversal, appears to be unnecessary. In a recent clinical study of Beriplex® P/N, the median infusion rate was 17.0 mL min−1, and there were no thromboembolic events [26]. A more modest median rate (7.5 mL min−1) was employed in the present study, and the only possibly PCC-related thromboembolic complication occurred in a patient receiving the initial PCC infusion at a rate (4.7 mL min−1) within the slowest quartile.

Another issue with plasma-derived preparations such as PCC is possible viral transmission. Beriplex® P/N is prepared from plasma screened by PCR and subjected to the dual viral elimination procedures of pasteurization and nanofiltration [18,41]. This study supports previous evidence of an excellent viral safety profile for Beriplex® P/N.

PCC has long been employed to provide rapid and effective reversal of coumarin anticoagulation. Yet this resource appears to be underutilized [42]. As noted by one author, hospitals with anticoagulant clinics and emergency departments often do not have at their disposal ‘the most effective antidote to one of the most commonly used and dangerous drugs’ [43]. Routine stocking of PCC by all hospitals managing anticoagulated patients has been advocated [43]. In conclusion, more widespread availability would place in the hands of clinicians a valuable tool, which, on the basis of the present findings, allows rapid and precise attainment of target INR and undelayed hemorrhage control, in both OAT patients requiring an interventional procedure and those with active bleeding.

Addendum

The members of the Beriplex® P/N Anticoagulation Reversal Study Group are:

J. Barth, Specialty Clinics, Bergmannstrost Medical Clinic, Halle/Salle, Germany.

B. Brand-Staufer, Clinical Department of Hematology, University Hospital Zürich, Zürich, Switzerland.

B. Brenner, Thrombosis and Hemostasis Unit, Rambam Medical Center, Haifa, Israel.

R. Germann, Department of Anesthesiology and Intensive Care, County Hospital Feldkirch, Feldkirch, Austria.

R. Griniute, Kaunas Medical University Clinic, Kaunas, Lithuania.

S. Haertel, Clinical Research and Development, Hemophilia/Critical Care, CSL Behring GmbH, Marburg, Germany.

U. Kalina, Clinical Research and Development, Hemophilia/Critical Care, CSL Behring GmbH, Marburg, Germany.

R. Kätzel, Institute for Transfusion Medicine and Clinical Hemostaseology, Municipal Clinic ‘St Georg’ Leipzig, Leipzig, Germany.

G. Kekstas, Centre of Anesthesiology, Intensive Therapy and Pain Management, Santariskiu Clinics, Vilnius University Hospital, Vilnius, Lithuania.

S. Knaub, Clinical Research and Development, Hemophilia/Critical Care, CSL Behring GmbH, Marburg, Germany.

S. Middeldorp, Department of Vascular Medicine, Academic Medical Center, Amsterdam, the Netherlands.

A. Nagy, Department of Surgery, Csolnoky Ferenc County Hospital, Veszprém, Hungary.

A. Oláh, Department of Surgery, Petz Aladár Teaching Hospital, Györ, Hungary.

H. Ostermann, Department of Hematology and Oncology, Medical Clinic III, University Hospital Munich - Großhadern, Ludwig Maximilian University, Munich, Germany.

I. Pabinger, Department of Internal Medicine, Division of Clinical Hematology and Hemostaseology, Medical University Vienna, Vienna, Austria.

T. Retteghy, Institute of Traumatology and Emergency, Budapest, Hungary.

J. Szmidt, Department of General, Vascular and Transplant Surgery, Medical University of Warsaw, Warsaw, Poland.

A. Tiede, Department of Hematology, Hemostaseology and Oncology, Center for Internal Medicine, Medical College Hannover, Hannover, Germany.

Disclosure of Conflict of Interests

I. Pabinger and H. Ostermann have received honoraria for their roles as coordinating investigators of the study. U. Kalina and S. Knaub are employees of CSL Behring. The remaining authors state that they have no conflict of interest.

Ancillary