Photochemically treated fresh frozen plasma for transfusion of patients with acquired coagulopathy of liver disease

Although improvements in donor testing have increased the safety of blood components prepared for transfusions, risk from transfusion-transmitted pathogens persists.¹  Transfusion of blood components has been implicated in the transmission of viral, bacterial, and protozoan diseases.²  Pathogens such as West Nile virus have recently emerged as transfusion-transmitted infectious agents, demonstrating that new infectious agents continue to enter the donor population and may pose significant risks to recipients before they can be definitively identified and avoided by sensitive tests.³ 

Development of inactivation processes effective against a broad array of infectious pathogens could provide a measure of safety against transfusion-transmitted infectious agents not routinely screened for or detected by current assays. An ex vivo photochemical treatment (PCT) process using a synthetic psoralen, amotosalen hydrochloride (S-59), has been developed to inactivate pathogens and residual leukocytes in fresh frozen plasma (PCT-FFP).⁴  Preclinical testing indicates that the system is effective in the reduction of a broad range of pathogens in plasma. The pharmacokinetics of select coagulation factors in PCT-FFP have been shown to be comparable to conventional FFP (C-FFP).^5,6 

Patients with acquired coagulopathy (most frequently due to end-stage liver disease) constitute a significant population requiring treatment with FFP. Patients with these disorders may require large volumes of FFP during a single treatment episode, and many of these patients will require repeated FFP transfusions resulting in exposures to FFP from many different blood donors.^7,8  This study describes a prospective, randomized, controlled clinical trial in which the efficacy and safety of PCT-FFP was compared with C-FFP in patients with acquired, complex coagulopathy of liver disease.

This study was a 2-arm, randomized, controlled, double-blinded study conducted at 5 sites in the United States. Eligible patients included those of either sex, age 2 years or older, with an acquired coagulopathy requiring therapeutic or prophylactic FFP transfusion indicated for an invasive procedure or reversal of warfarin therapy. Patients with coagulopathy due to trauma were excluded. No concomitant therapies were specifically excluded with the exception of other investigational therapies within 30 days prior to randomization. After screening and giving informed consent, patients were randomized to receive either PCT-FFP or C-FFP during a 7-day treatment period. An FFP dose of 15 to 20 mL/kg was recommended, although the treating physician determined the final dose and rate of administration of study FFP. In some instances, liver transplantation patients received a continuous transfusion of FFP for the duration of surgery (typically 5 to 10 mL/kg/h).

All patients who received at least one 2-unit plasma transfusion (about 400 mL) of C-FFP or PCT-FFP were included in the intent-to-treat (ITT) population. The primary end point was the response of the prothrombin time (PT) and partial thromboplastin time (PTT) to the first study FFP transfusion. Secondary efficacy analyses included the recovery of factor VII 1 hour after the first transfusion, the correlation of the PT response to the factor VII level following the initial transfusion, an analysis of covariance to the response in PT and PTT, assessment of clinical hemostasis, the PT and PTT responses to all subsequent transfusions, use of blood components, and safety.

Clinical hemostasis was assessed during the 8 hours prior to the first study FFP transfusion and then for 8 hours following the first study FFP transfusion. Study personnel, blinded to treatment assignment, conducted hemostatic assessments using an established bleeding assessment scale, the World Health Organization (WHO) bleeding criteria, as described for a prior study of platelet transfusion efficacy.⁹ 

Blood samples were obtained for each patient within 8 hours before and 1 hour after the first FFP transfusion to assess the PT, PTT, S-59 concentration, serum chemistry, hematology, and factor VII activity. For patients treated with continuous rather than bolus FFP infusions, the end of the transfusion period was defined as an interval of 8 hours without FFP infusion. In these cases, samples for laboratory analyses were obtained within the first hour after cessation of the FFP infusion. PT and PTT values and S-59 samples were drawn before and after each subsequent study FFP transfusion. PT and PTT assays were performed using the methods in use at each study site clinical laboratory.

In addition, blood samples were obtained within 8 hours prior to the first transfusion and at the 6-week follow-up visit for assessment of antibody formation to potential S-59-associated plasma neoantigens. Sera for detection of potential antibodies to S-59-related neoantigens were tested using a previously described solid-phase immunoassay.^10,11  All adverse events were recorded according to the National Cancer Institute-Common Toxicity Criteria (NCI-CTC). Adverse events, including hemorrhagic events, were recorded for 2 weeks after the first transfusion, and major medical adverse events were recorded through 6 weeks after first transfusion. Adverse events were reported by study personnel and reviewed by the principal investigator at each study site.

An independent data and safety monitoring board reviewed all serious adverse events and the general conduct of the study. The investigational review boards of all participating institutions approved the study, and all patients provided informed consent documented in writing.

The investigational product, PCT-FFP (Cerus, Concord, CA), consisted of plasma photochemically treated using the synthetic psoralen (amotosalen HCl; 3-[(2-aminoethoxy)methyl]-2,5,9-trimethyl-7H-furo[3,2-g] [1] benzopyran-7-1 hydrochloride) formerly known as S-59. Single units of plasma used for this study were prepared from whole-blood collections or by approved plasmapheresis methods. Approximately 15 mL S-59 was mixed with approximately 250 mL plasma (final S-59 concentration, about 150 μM), illuminated with UVA light (320 to 400 nm) with a 3.0 J/cm² treatment, and transferred to an S-59 compound adsorption device (CAD). The CAD consisted of approximately 12.5 g cross-linked copolymer polystyrene divinylbenzene spherical adsorbent beads (Dowex Optipore L493; Dow Chemical, Midland, MI) enclosed in a medical grade polyester mesh pouch inside a PL 2410 (Baxter Healthcare, Deerfield, IL) plastic container. The treated plasma was incubated in the CAD at room temperature for 1 hour with agitation to reduce the residual levels of S-59 and transferred to a final storage container for freezing and storage within 8 hours of collection. The PCT-FFP units were frozen at -35°C or colder and stored at -18°C or colder for up to 1 year from the date of collection. PCT-FFP was prepared at 6 regional blood centers and shipped frozen from processing centers to patient-care sites as required for transfusion, thawed, and administered according to standard procedures at each clinical site.

Before freezing of PCT-FFP units, samples for coagulation testing were removed. Coagulation factor activities were measured using standard laboratory methods in samples obtained from 277 PCT-FFP units randomly selected from inventory. Reference ranges for C-FFP were determined by measuring coagulation factor activities using aliquots obtained from plasma units prior to PCT processing. The reference range was expressed as the 95% central interval (2.5th to 97.5th percentile) of the distribution for the values of untreated plasma. Antithrombotic protein activities were measured in 14 units of PCT-FFP. Reference range values for antithrombotic proteins were obtained from the literature. Mean coagulation factor and antithrombotic protein activities in PCT-FFP were within the reference ranges (Tables 1, 2).

The primary efficacy end points in this study were change in prothrombin time (ΔPT) and change in partial thromboplastin time (ΔPTT), adjusted for FFP dose and patient body weight, in response to the first study FFP transfusion. Changes in PT and PTT were adjusted for FFP dose (milliliters per kilogram) and patient weight using the following formulas¹³ : ΔPT_N (s·kg·mL^-1) = (PT_pre - PT_post)/(volume transfused/patient weight) and ΔPTT_N (s·kg·mL^-1) = (PTT_pre - PTT_post)/(volume transfused/patient weight).

Published data describing responses of the PT and PTT are limited. A previous study comparing the transfusion of solvent/detergent (SD)-treated plasma with C-FFP in patients with a prolonged PT indicated that plasma transfusion in patients with acquired coagulopathy resulted in mean decreases in the PT and PTT of 4.6 and of 4.8 seconds, respectively.¹⁴  Results from this published study were used to describe a clinically relevant definition of equivalence for response to FFP using PT and PTT responses not adjusted for FFP dose and patient weight and to determine the appropriate sample size and power for the an equivalence analysis of the PT and PTT response in this trial. The noninferiority margins for the change in PT and PTT were prespecified as a difference in reduction of -4.6 seconds for the PT and -4.8 seconds for the PTT. Based on these inferiority margins, a sample size of 120 patients, randomized in a 1:1 ratio, was estimated to provide more than 80% power to reject the null hypotheses of nonequivalence for the 2 primary end points (PT and PTT) at a 2-sided alpha level of 0.05. The hypotheses to be tested were as follows: (a) The absolute difference in mean PT reduction (PT_post - PT_pre), (PT_N), between test and control is greater than or equal to 4.6 seconds (H₀₁: |mean PT_N (test) - mean PT_N (control)| ≥ 4.6 s), and (b) the absolute difference in mean PTT reduction (PTT_post - PTT_pre), (PTT_N), between test and control is greater than or equal to 4.8 seconds (H₀₂: |mean PTT_N (test) - mean PTT_N (control)| ≥ 4.8 s).

For the coprimary end points, observation of a small P value (eg, P ≤ .025 for a .025 level of significance) would indicate rejection of the null hypotheses of nonequivalence and therefore confirm equivalence between PCT-FFP and C-FFP for the specified margins.

Factor VII increments (international units per deciliter) and incremental factor VII recovery (international units per deciliter per international units per kilogram) were determined and mean difference compared by t test. In addition, dose-adjusted PT and PTT responses for all FFP transfusions administered during the 7-day transfusion support period were compared using repeated measures of analysis of variance (ANOVA). ANOVA with treatment and study site in the model was used for continuous variables. The Fisher exact test was used for comparison of adverse events.

Collection of intraoperative surgical hemostatic observations and postoperative hemostasis, as measured by hemostatic scoring using a modified version of the WHO bleeding criteria as previously described⁹  and by review of the medical record of surgical procedures for estimated blood loss, was used to assess clinical hemostasis. The median posttransfusion bleeding grade following FFP transfusion was compared between treatment groups. The Wilcoxon rank sum test was used for comparison of hematology and serum chemistry parameters as well as grade 2 bleeding. All secondary end points were analyzed for differences between treatment groups at the P equals .05 significance level.

Approximately 250 patients with acquired coagulopathy were screened for enrollment; 132 gave informed consent and were randomized; and 121 patients received at least one FFP transfusion of the assigned study product. The 121 patients transfused with 2 or more units of FFP (about 400 mL) constituted the ITT analysis group (60 PCT-FFP, 61 C-FFP). The PCT-FFP and C-FFP groups exhibited similar demographics, respectively, for male sex (52%, 67%), white race (70%, 75%), age (60 ± 15, 61 ± 12), A-B-O blood type (40%, 12%, 48% versus 39%, 10%, 51%), and concomitant medications (data not shown).

Primary diagnoses at enrollment were similar between treatment groups (Table 3). Of these, 51 patients (42%) were transfused for hemostatic support during orthotopic liver transplantation (OLT); 22 (37%) received PCT-FFP, and 29 (48%) received C-FFP (P = .27). The most common non-OLT indications for FFP transfusion were minor surgical procedures (Table 4).

Overall patient population. The response of the PT to the first FFP transfusion was assessed in 58 (97%) PCT-FFP and 58 (95%) C-FFP patients. When changes in PT and PTT were adjusted by FFP dose and patient weight, no statistically significant differences in the correction of the PT and PTT were observed between PCT-FFP and C-FFP groups (Table 5).

The unadjusted correction of the PT was equivalent between the PCT-FFP and C-FFP groups following the first transfusion (Table 6). The mean difference for change of the PT (PCT-FFP minus C-FFP), not adjusted for FFP volume infused or body weight, between the treatment groups was 0.2 seconds (95% confidence interval [CI], -3.0 to 2.5 seconds; P = .002 by noninferiority). This observation was statistically significant for rejection of inferiority, thus confirming equivalence.

The PTT response to the first episode of FFP transfusion was assessed for 57 (95%) PCT-FFP and 56 (92%) C-FFP patients. The observed mean difference in the PTT response between the treatment groups without adjustment for FFP volume infused or body weight was 1.3 seconds (95% CI, -6.04 to 8.62 seconds; P = .348 by noninferiority). The 95% CI for the upper and lower limit of the difference in the PTT response was greater than the prespecified equivalence margin for the PTT (4.8 seconds). This was due to the large variance in the PTT response, and thus rejection of inferiority or equivalence could not be confirmed. However, the correction of PTT showed a trend slightly favoring the PCT-FFP group (Table 6).

OLT patient population. Because liver transplantation patients generally received larger volumes of FFP and had more substantial surgical procedures than nontransplantation patients, the OLT and non-OLT populations were analyzed separately. OLT patients (n = 51) had a mean difference in the absolute PT response between treatment groups of 1.7 seconds (P = .103 by noninferiority). The mean difference in the absolute PTT response between treatment groups was 4.1 seconds (P = .890 by noninferiority).

Non-OLT patient population. Non-OLT patients (n = 70) demonstrated a mean difference in absolute PT response between the treatment groups of 0.6 seconds (P = .042 by noninferiority). The mean difference in the absolute PTT response between treatment groups was 3.1 seconds (P = .907 by noninferiority).

Factor VII levels. The mean dose of factor VII transfused with the first transfusion was lower for the PCT-FFP group (1623 ± 1844 IU versus 2434 ± 3256 IU; P = .100). The mean increment for factor VII after the first study transfusion was 12.3 ± 26.3 IU/dL for test and 15.0 ± 31.43 IU/dL for control patients; P = .100. Incremental factor VII recovery was comparable between test and control patients (0.69 ± 1.80 IU/dL/IU/kg versus 0.48 ± 0.76 IU/dL/IU/kg; P = .422).

The dose response of the factor VII level after the first study transfusion was analyzed with linear regression models, where the covariates were treatment (test versus control), indication for transfusion (OLT versus non-OLT), and the dose of factor VII transfused, and the dependent variable was the change in PT (dose adjusted or unadjusted) or the change in factor VII level. Analyses were performed for the modified intent to treat (MITT) population. As anticipated, there was a significant correlation between factor VII dose and unadjusted PT response (P = .043) and change in factor VII activity (P < .001) after the first transfusion. This relationship with factor VII dose was absent for the adjusted PT response (P = .862) because it already accounted for FFP volume, which highly correlated with factor VII dose.

Clinical hemostasis and blood-component usage. The maximal bleeding grade was assessed prior to and after the first study transfusion. Clinical hemostasis was comparable between PCT-FFP and conventional FFP (Table 7). Because very few patients had bleeding requiring red-cell transfusion (WHO grade 3) prior to the first study transfusion, the use of red cells is a surrogate indicator for hemostasis. In addition, a poor hemostatic response to study FFP transfusion would result in increased FFP use, thus indicating failure of clinical hemostasis. There were no differences between the PCT-FFP and conventional FFP treatment groups in use of FFP or other blood components (Table 8). As expected, OLT patients exhibited greater blood-component usage than non-OLT patients, but there was no difference in blood-component usage within these subgroups between PCT-FFP and C-FFP treatment groups (Table 8).

S-59 levels were measured prior to transfusion, 5 to 60 minutes after the first study transfusion, and immediately prior to the final study period transfusion to determine trough levels. For all test patients, the mean peak level was 5.4 ± 3.6 ng/mL, compared with OLT patients with a mean peak level of 8.3 ± 4.0 ng/mL, and non-OLT patients, with a mean peak level of 3.8 ± 2.0 ng/mL. For all patients, the mean trough level of S-59 was 2.0 ± 2.0 ng/mL and, despite larger exposures to FFP, mean trough S-59 levels were lower in OLT patients than non-OLT patients (1.4 ± 1.4 ng/mL versus 2.1 ± 22 ng/mL). For comparison, in a phase 1 study, healthy nonbleeding subjects transfused with 1000 mL PCT-FFP exhibited mean peak S-59 levels of 11.5 ± 0.8 ng/mL.¹⁵ 

Adverse events, by system organ class and preferred term, were reported with similar frequency between PCT-FFP and conventional FFP patients (Table 9). There were no statistical differences in hematology and chemistry clinical laboratory values between treatment groups (P > .05, Wilcoxon rank sum test). No antibodies against PCT-FFP were detected in any patients.

Six cases of hepatic artery thrombosis (HAT) were reported (2 in the PCT-FFP group and 4 in the C-FFP group). The incidence of this complication was within the range reported for liver transplantation.¹⁶  The first 4 cases of HAT occurred within 2 weeks of OLT surgery and the last transfusion of study FFP, while the remaining 2 cases of HAT occurred more than 4 weeks after the last administration of study FFP. Plasma samples collected before and after transfusion from these patients, as well as samples from PCT-FFP units, indicated that the antithrombotic proteins, antithrombin III and proteins C and S, were conserved comparably between test and control patients (Table 10). Anticardiolipin antibodies were not present in any of the 6 patients with HAT (Table 10). The clinical course for each patient with HAT was reviewed by the study's data and safety monitoring board, and none of the cases was attributed to use of study FFP.

There were 21 deaths in the overall study population (12 PCT-FFP, 9 C-FFP). All deaths were consistent with the typical course of hepatic failure, and none were assessed as related to the transfusion of study FFP.

Approximately 4 million units of FFP and 1 million units of cryoprecipitate prepared from FFP are transfused annually in the United States, and a similar number in Europe. Acquired coagulopathy often presents with potential hemorrhagic complications that require transfusion with FFP and frequently with large volumes. FFP transfusion continues to carry a risk of microbial transmission that is increased with exposure to large volumes of FFP.^17-20  To reduce these risks, a variety of methods have been developed to improve safety, including quarantine and pathogen inactivation treatment of plasma using solvent/detergent (SD-FFP), methylene blue (MB-FFP), and the amotosalen (S-59) photochemical process (PCT-FFP).^21,22 

Photochemical treatment of plasma with amotosalen and UVA light provides a pathogen inactivation method for single-unit preparation that conserves functional activity of both procoagulant and antithrombotic proteins.⁵  Preclinical studies have shown that this process inactivates enveloped viruses, both cell-free and cell-associated, some relevant nonenveloped viruses, and residual leukocytes in plasma.^4,23,24 

A controlled, randomized, crossover study in healthy subjects treated with warfarin demonstrated that PCT-FFP provided an equivalent in vivo response to C-FFP, as assessed through postinfusion kinetics of factor VII and recovery of factors II, VII, IX, and X.⁶  A single-arm, open-label study in patients with congenital factor deficiencies (I, II, V, VII, X, XI, XIII, and protein C) demonstrated posttransfusion recoveries and support of hemostasis during invasive procedures comparable to that observed with C-FFP.⁵ 

The current study demonstrated no difference in correction of the PT or PTT on a dose- and weight-adjusted basis. Without adjustment for FFP dose and patient weight, equivalence was observed between PCT-FFP and C-FFP for correction of the PT in patients with acquired coagulopathy due to liver disease either for minor invasive procedures or liver transplantation. The improvement seen in PTT with PCT-FFP was greater than with the C-FFP group; however, statistical equivalence was not demonstrated. Although neither the PT nor the PTT responses corrected to within the reference range after transfusion with either PCT-FFP or C-FFP, both assays demonstrated partial correction, clinical hemostasis was sustained, and the PT and PTT did not deteriorate during the period of support. Similar observations have been reported for a comparative study of C-FFP and SD-FFP in patients with liver disease and liver transplantation.²⁵ 

Evaluation of changes in coagulation assays is difficult in the setting of liver transplantation due to the nonsteady state induced by large-volume blood loss. For non-OLT patients with less active bleeding, the performance of PCT-FFP for the PT response was statistically better when compared with transfusion with conventional FFP. Comparable increases in factor VII levels and in factor VII incremental recovery were observed between the PCT-FFP and conventional FFP treatment groups. These findings are consistent with the pharmacokinetic study of factor VII in nonbleeding subjects⁶  and indicate that PCT-FFP can be dosed similarly to C-FFP. In this study, factor VII dose, the PT response, and the change in factor VII levels were significantly correlated. This is relevant information given the complex coagulopathy of patients with severe liver disease and the dependence of the PT on other coagulation factors with longer half-lives.²⁶ 

An important indication of the clinical efficacy of FFP transfusion is by assessment of clinical hemostasis and the use of FFP and other blood components to support patients during minor and major invasive procedures. The PCT-FFP and conventional FFP treatment groups appeared comparable with regard to clinical hemostatic assessments, transfusion of additional blood components, and hemorrhagic adverse events. The amount of blood-component usage by patients undergoing OLT was comparable between treatment groups and was consistent with previously reported blood-component transfusion use in liver transplantation.^25,27 

Adverse events reported in this trial were consistent with the spectrum of clinical events described in association with severe underlying liver disease.^28,29  HAT is a recognized postoperative complication associated with OLT surgery.¹⁶  It has been reported that solvent/detergent-treated plasma has diminished levels of some antithrombotic proteins.^30,31  Because of concerns regarding this complication, the levels of antithrombotic proteins in PCT-FFP and in the plasma of postoperative patients were measured during this study. The incidence of HAT was comparable between treatment groups, and the levels of antithrombotic proteins in PCT-FFP and C-FFP were similar, as were the levels in patient plasma.

Studies of FFP transfusion for patients with coagulopathy of liver disease have been conducted for SD-FFP but were smaller in scope.^14,25  Randomized trials of MB-FFP have not been reported for patients with acquired coagulopathy of liver disease.³²  However, in some countries, experience has been reported to indicate increased usage of MB-FFP, red cells, and cryoprecipitate after widespread usage of MB-FFP.³³  The present study in conjunction with the study of warfarin reversal and factor VII kinetics and PCT-FFP transfusion of patients with congenital factor deficiencies provides further characterization of PCT-FFP for support of patients with acquired coagulopathy of liver disease. A recently completed randomized, controlled study of PCT-FFP for therapeutic plasma exchange of patients with TTP indicated therapeutic efficacy of that disorder as well.³⁴  Thus, PCT-FFP offers a potential means to further improve the safety for each major clinical indication of FFP transfusion.

Special acknowledgments for conduct of the study are extended to study site coordinators: Ahmed Aldilaimi, Yu Ming Huang, Lin Maslow, and Mary Meester. We acknowledge the efforts of Marye Ellen Valentine for coordination of the study cites and FFP supplies. The authors acknowledge the extensive efforts of Maureen Conlan and Jin-Syng Lin for review of the statistical analyses. The data and safety monitoring board was composed of Marc Shuman, MD (University of California, San Francisco, CA); Margaret Rick, MD (Clinical Center, National Institutes of Health, Bethesda, MD); William Swaim, MD (University of Minnesota, Minneapolis); and Joel Verter, PhD (George Washington University, Washington, DC).