ORIGINAL ARTICLE Whole blood treated with riboflavin and ultraviolet light: quality assessment of all blood components produced by the buffy coat method Peter Schubert,1,2,3 Brankica Culibrk,1,2 Simrath Karwal,1,2 Katherine Serrano,1,2,3 Elena Levin,1,2,3 Daniel Bu,1,2 Varsha Bhakta,1,4 William P. Sheffield,1,4 Raymond P. Goodrich,5 and Dana V. Devine1,2,3

BACKGROUND: Pathogen inactivation (PI) technologies are currently licensed for use with platelet (PLT) and plasma components. Treatment of whole blood (WB) would be of benefit to the blood banking community by saving time and costs compared to individual component treatment. However, no paired, pool-andsplit study directly assessing the impact of WB PI on the subsequently produced components has yet been reported. STUDY DESIGN AND METHODS: In a “pool-and-split” study, WB either was treated with riboflavin and ultraviolet (UV) light or was kept untreated as control. The buffy coat (BC) method produced plasma, PLT, and red blood cell (RBC) components. PLT units arising from the untreated WB study arm were treated with riboflavin and UV light on day of production and compared to PLT concentrates (PCs) produced from the treated WB units. A panel of common in vitro variables for the three types of components was used to monitor quality throughout their respective storage periods. RESULTS: PCs derived from the WB PI treatment were of significantly better quality than treated PLT components for most variables. RBCs produced from the WB treatment deteriorated earlier during storage than untreated units. Plasma components showed a 3% to 44% loss in activity for several clotting factors. CONCLUSION: Treatment of WB with riboflavin and UV before production of components by the BC method shows a negative impact on all three blood components. PLT units produced from PI-treated WB exhibited less damage compared to PLT component treatment.

P

athogens and emerging infectious agents like viruses, bacteria, and protozoa are a constant threat to blood safety.1 Several technologies aiming for pathogen inactivation (PI) have been developed for the treatment of blood components2,3 and have become routinely used in various jurisdictions providing an increasing body of evidence on the clinical efficacy, safety, and cost–benefit ratio of PI.4 A challenge for the implementation of PI for whole blood (WB) donations is the cost and processing burden imposed by the need to treat finished components. Optimally, treatment of the WB unit before processing would reduce the impact on processing and likely overall cost of PI. Custer and colleagues5 modeled the cost-effectiveness of PI treatment of WB versus platelet (PLT) and plasma alone concluding that WB PI would likely be more costeffective because of the reduction of the threat in all blood

ABBREVIATIONS: APTT = activated partial thromboplastin time; BC = buffy coat; ESC = extent of shape change; FP = frozen plasma; PC(s) = platelet concentrate(s); PI = pathogen inactivation; PRT = pathogen reduction technology; PS = phosphatidylserine; PT = prothrombin time; WB = whole blood. From the 1Centre for Innovation, Canadian Blood Services, and the 2Centre for Blood Research and the 3Department of Pathology and Laboratory Medicine, University of British Columbia, Vancouver, British Columbia, Canada; 4McMaster University, Hamilton, Ontario, Canada; and 5TerumoBCT Biotechnologies, Lakewood, Colorado. Address reprint requests to: Dana Devine, Canadian Blood Services, UBC Centre for Blood Research, 2350 Health Sciences Mall, Vancouver, British Columbia, V6T 1Z3, Canada; e-mail: [email protected]. Received for publication March 5, 2014; revision received August 28, 2014, and accepted September 1, 2014. doi: 10.1111/trf.12895 © 2014 AABB TRANSFUSION **;**:**-**. Volume **, ** **

TRANSFUSION

1

SCHUBERT ET AL.

components. Such analyses can inform policy decisions regarding pathogen reduction technology (PRT) in the context of other initiatives designed to improve transfusion safety. Research studies of PI-treated PLT and plasma components consistently demonstrate a negative effect on in vitro quality variables triggering ongoing discussions of the risks and benefits of its application.6 Clinical studies generally indicate that current PI procedures leave PLT components sufficiently viable to achieve an effective transfusion;7,8 however, the diverse technologies currently on the market have different mechanisms of action. Pathogen-inactivated plasma seems to perform as well clinically as standard fresh-frozen plasma (FFP); however, most reported trials have been underpowered and the evidence for standard FFP being of clinical value is also weak in many cases.9 PI treatment of PLT concentrates (PCs) accelerates the development of the storage lesion,10,11 while coagulation proteins in pathogen-inactivated plasma exhibit a reduction in their activation.12 Ideally, further development of PI technologies will focus on minimizing the damage caused by these treatments. Research efforts focus on determining whether these changes impact transfusion outcome and on developing strategies to minimize PI-induced changes mainly through molecular level studies.13 Although PIs are currently licensed only for PLT and plasma products,4,9,14 methods for the treatment of red blood cells (RBCs) are not far from routine use.15 Currently, blood components are individually PI treated; however, it is desirable to carry out PI in one step before production of all three blood products. The riboflavin and ultraviolet (UV) light process for PI (Mirasol) has recently been adapted to the treatment of WB units16 and has been shown to be an alternative to gamma irradiation to prevent transfusion-associated graft-versus-host disease.17 Several studies have been conducted to assess the PI capacity of this riboflavin and UV light technology on WB using a variety of illumination doses ranging from 22 to 110 J/mLRBCs. Using 80 J/mLRBCs, the growth of several bacteria spiked into WB before illumination was stopped with 92% to 96% efficiency.18 At the same illumination dose, a variety of viruses including human immunodeficiency virus19 could be reduced below detection limit.16 Finally, 80 J/mLRBCs resulted in significant inactivation of parasites such as Trypanosoma cruzi,20 Babesia microti,21 and Plasmodium falciparum.22 Since these studies addressed the PI capacity, our study did not assess PI efficacy. To date, a single study has examined the quality of blood components at expiry produced from WB treated with riboflavin and UV light.23 Component production in that study used the standard American PLT-rich plasma method, and its design compared treated and untreated 2

TRANSFUSION Volume **, ** **

units from different donors rather than using a pool-andsplit design. Herein we report for the first time comprehensive data comparing plasma and RBC components produced using the buffy coat (BC) production method with or without PI treatment of WB with the Mirasol process. For the PLT components, the PI impact was assessed by comparing the unit obtained from the PI-treated WB with direct illumination of the PLT unit produced from the untreated control study arm (Fig. 1).

MATERIALS AND METHODS Materials Common chemicals were purchased from Sigma-Aldrich (St Louis, MO) or Fisher Scientific (Ottawa, Ontario, Canada).

WB donation This study was approved by the research ethics board of Canadian Blood Services. Informed consent was obtained from all healthy volunteers before the donation. WB collection and component production were carried out by the netCAD development laboratory of Canadian Blood Services (Vancouver, British Columbia, Canada).

Study design and blood component production The general study design is outlined in Fig. 1. For each repetition of the experiment, eight ABO-matched WB units were pooled and then split into eight identical units. Six independent sets of experiments were conducted. Using a two-arm study design, four of these units were put through the Mirasol treatment process in which 35 mL of riboflavin solution (500 μmol/L) was added before light treatment. Illumination was carried out using the Mirasol device according to the manufacturer’s protocol (TerumoBCT, Lakewood, CO) using an irradiation dose of 80 J/mLRBC. UV illumination time was calculated by the device using the hematocrit (Hct), which was determined from the pooled unit using a Hct centrifuge (HAEMATOKRIT 210, Hettich Zentrifugen, Tuttlingen, Germany) according to the manufacturer’s instructions and the weight of the WB unit. To the other four units, 35 mL of saline was added to control for dilution. All units were held overnight on cooling plates for a minimum of 18 hours. The two sets of differently treated units were subjected to routine component processing using the BC method. This resulted in an untreated PC (BC/PC), RBCs, and frozen plasma (FP) units as well as the products prepared after the WB PI treatment (*), BC/PCWB*, RBCWB*, and FPWB*, respectively. To compare BC/PCWB* to a routine Mirasol-treated BC/PC, the untreated BC/PC was subjected to riboflavin and UV light treatment (BC/PC*) according to manufacturer’s instructions on Day 1.

WHOLE BLOOD PATHOGEN INACTIVATION

Fig. 1. Schematic overview of the study design. Eight ABO-matched WB units were pooled and split into eight identical units; all were kept on a cooling tray overnight (O/N). In a two-arm study (A, B), four of these units were treated with riboflavin and UV light (Mirasol) before O/N hold and four were processed untreated. Units from Study Arm A were produced into a BC/PCWB*, RBCWB*, and FPWB* and the respective comparator components (BC/PC), RBCs, and FP produced from the untreated WB units from Study Arm B. BC/PC from Study Arm B were treated with Mirasol after production. Sampling for quality analyses during their individual storage at 22, 4, and −80°C was performed as stated. The study has six replicates (n = 6).

Product storage and sample preparation PCs were stored in the Mirasol illumination bag at 20 to 24°C under constant agitation and samples were drawn on Day 1 (= day of production) and Days 2, 5, and 7 of storage. RBCs were stored at 4°C and samples were collected at Week 0 (= day of production) and Weeks 1, 3, and 6. Plasma units were frozen immediately after production at −80°C and were shipped frozen to CBS in Hamilton, Ontario. Plasma units were thawed at 37°C and samples were then analyzed without being refrozen. Sampling from storage bags was carried out aseptically in biosafety cabinets.

Analyses of PCs The PLT count and mean PLT volume were measured on a hematology analyzer (Siemens, Deerfield, IL). Blood gases (pCO2 and pO2) and metabolites (glucose, lactate, and pH) were quantified using a blood gas analyzer (Gem Premier

3000, Instrumentation Laboratories, Bedford, MA). If lactate levels were out of the range of the analyzer, the samples were diluted with phosphate-buffered saline (PBS) and retested. PLT activation was monitored by the expression of P-selectin (CD62P) on the PLT surface using flow cytometry. The PLT sample was diluted with PBS to a concentration of approximately 200 × 109/L and was incubated for 30 minutes with phycoerythrin-labeled antiCD62P (Beckman-Coulter, Mississauga, Ontario, Canada). In parallel, samples were treated with 10 μmol/L ADP before the antibody incubation to determine the degranulation response capacity of PLTs. The PLTs were analyzed on a flow cytometer (FACSCanto II, BD Biosciences, Mississauga, Ontario, Canada) defining PLT activation as the percentage of PLTs assessed that were positive for CD62P. ADP was used to compare PLT responsiveness obtained in the extent of shape change (ESC) assay. ESC was measured using an aggregometer (SPA-2000, Chronolog Corp., Volume **, ** **

TRANSFUSION

3

SCHUBERT ET AL.

Havertown, PA) according to Holme24 and described in detail previously.25 Apoptosis was monitored by changes in the exposure of phosphatidylserine (PS) on the PLT surface using annexin V binding as described previously.26 All PLT units were tested for sterility after the sampling on Day 7 by plating an aliquot on a blood agar plate (VWR, Mississauga, Ontario, Canada) followed by incubation for 24 hours at 37°C. No units showed any growth.

Analyses of RBCs The RBC count, the mean corpuscle volume (MCV), and hemoglobin (Hb) concentration were measured on a hematology analyzer (ADVIA 120, Siemens). Metabolites (glucose and lactate) and potassium (K+) were quantified using a blood gas analyzer (Gem Premier 3000, Instrumentation Laboratories, Bedford, MA). pH was measured with a pH probe (Orion Ross Ultra Semi-Micro, Thermo Scientific, Waltham, MA). If glucose, lactate, or potassium levels were out of the range of the analyzer, the samples were diluted with saline and retested. The degree of hemolysis was determined by the Harboe method as published previously.27 The level of ATP in RBCs was quantified by HPLC after perchloric acid extraction of the RBCs. Quantification of RBC-derived microparticles, defined as events smaller than 1 μm, labeled with RPE-glycophorin A was conducted by flow cytometry against a known concentration of fluorescent beads using a flow cytometer (FACSCanto II, BD Biosciences). Bacterial testing after expiry using a bacterial detection system (BacT/ALERT, bioMérieux, Marcy l’Etoile, France) showed no units yielded results for bacterial growth.

Analyses of FFP The activity levels of Factor (F)V, FVII, FVIII, protein S, and α2-antiplasmin, as well as fibrinogen and the prothrombin time (PT) and the activated partial thromboplastin time (APTT) were determined as previously described,28-30 using an automated coagulation analyzer (STA Compact, Diagnostica Stago, Asnieres, France) and one-stage clotting assays; the fibrinogen determination employed a modified Clauss assay. Similarly, von Willebrand factor (VWF) activity levels were measured using an enzymelinked immunosorbent assay–type method exploiting a monoclonal antibody to an epitope specific to functional VWF (Corgenix, Broomfield, CT). FXIII activities were measured via an ammonia release assay (Technoclone GmbH, Vienna, Austria).

Statistical analysis First, normality of distribution of the data was tested (Minitab, Inc., State College, PA). When not normally distributed, Johnson transformation was applied to the data 4

TRANSFUSION Volume **, ** **

set being analyzed. Statistical analyses to determine differences between BC/PCWB* and BC/PC* as well as RBCWB* and RBC and FPWB* and FP of each blood component, respectively, were performed using a two-way analysis of variance (ANOVA) with repeated measures and a p value of less than 0.05 was considered significant. In case Johnson transformation was not possible, nonparametric analyses were carried out using the Kruskal-Wallis test. Post-ANOVA analysis to identify differences at each time point of the individual variables between BC/PCWB* and BC/PC* as well as RBCWB* and RBC and FPWB* and FP of each blood component, respectively, were determined with t tests using computer software (Excel, Microsoft, Redmond, WA), adjusting the p value for multiple comparisons. Sample size calculations assumed a significance level of 5% and a power of 80% to detect a potential difference in the in vitro quality variables between the treated and untreated samples for the RBC and plasma samples as well as between the PLT samples obtained from the WB illumination and routine illumination of the PLT unit produced in the study arm without WB treatment.

RESULTS This study was designed to directly compare plasma and RBCs derived from PI-treated and untreated WB as well as to compare BC-derived PCs derived from PI-treated WB to those subjected to PI treatment after production. The in vitro quality of the PCs derived from WB treatment versus direct treatment of BC PC from the control study arm was analyzed using a panel of variables (Table 1). The treatment mode had an impact on the PLT concentration leading to a lower PLT count in the PCs produced after WB treatment, which was maintained throughout the storage period. This reduced PLT count was in part due to a larger volume of the PLT units after addition of 35 mL of riboflavin in saline. In comparison to untreated samples, previous studies showed that these numbers are approximately 20% to 25% lower compared to untreated PLTs, which are in the range of 807 × 109 ± 45 × 109/L on Day 1 to 826 × 109 ± 48 × 109/L on Day 7 of storage.11 In accounting for this difference, therefore, an all-over reduced yield in the units derived from the WB illumination of approximately 5% was still observed. The mean PLT volume did not vary either with the PI treatment or with storage, but was bigger when compared to untreated PLTs (e.g., MPV of 7.9 ± 05 fL on Day 1 and 8.6 ± 0.3 fL on Day 7 of storage11). The pH was similar in both PC units within the first 4 days of storage, but became significantly different on Days 5 and 7 with the pH in the PC derived from the WB treatment being higher; however, the pH values of these treated units are within the range seen in untreated units.11 For glucose, a significant difference was detected both in the trend between the two study arms and over the

WHOLE BLOOD PATHOGEN INACTIVATION

TABLE 1. Summary of routine PLT quality measures using a variety of in vitro assays comparing PCs produced by the BC method* Assay PLT concentration (×109/L) Unit volume (mL) Yield (×109/unit) Mean PLT volume (fL) pH Glucose (mmol/L) Lactate (mmol/L) PLT activation (% positive for CD62P) Response to PLT activation (% positive for CD62P) ESC (%) Annexin V (% positive)

Day BC/PC* 609 ± 57 399 ± 17 243 ± 30 9.4 ± 0.5 7.1 ± 0.0 15.3 ± 1.1 6.3 ± 0.3† 40.9 ± 10.0

1 BC/PCWB* 567 ± 35 413 ± 16 234 ± 22 9.3 ± 0.4 7.1 ± 0.0 15.4 ± 0.6 7.9 ± 0.6 37.2 ± 7.8

Day BC/PC* 622 ± 76 387 ± 15 241 ± 29 9.0 ± 0.5 7.2 ± 0.0 13.4 ± 0.6 7.3 ± 0.6† 35.9 ± 4.9

2 BC/PCWB* 571 ± 50 405 ± 18 232 ± 33 9.3 ± 0.5 7.3 ± 0.0 15.2 ± 0.5 8.8 ± 0.5 35.7 ± 4.1

Day BC/PC* 641 ± 59 391 ± 15 249 ± 15 8.9 ± 0.5 7.1 ± 0.1† 11.2 ± 0.8† 13.2 ± 0.7 53.5 ± 3.9

5 BC/PCWB* 584 ± 48 401 ± 19 234 ± 29 9.2 ± 0.4 7.2 ± 0.0 12.7 ± 0.8 13.7 ± 1.0 48.7 ± 1.1

Day BC/PC* 616 ± 40 397 ± 11 244 ± 20 9.0 ± 0.4 7.0 ± 0.0† 9.3 ± 0.7† 16.2 ± 1.1 64.6 ± 2.8†

7 BC/PCWB* 587 ± 97 409 ± 14 240 ± 38 9.2 ± 0.4 7.1 ± 0.0 11.0 ± 0.7 15.1 ± 0.2 54.0 ± 1.9

26.2 ± 7.7

31.4 ± 7.2

20.5 ± 4.2†

32.0 ± 5.2

15.3 ± 2.6†

23.8 ± 4.5

7.1 ± 2.4†

15.5 ± 3.3

16.5 ± 3.1 4.7 ± 2.7

20.5 ± 2.8 5.5 ± 2.3

14.2 ± 2.3† 5.7 ± 1.9

18.8 ± 1.5 4.9 ± 1.1

13.4 ± 1.7 13.2 ± 1.6†

17.6 ± 2.4 8.5 ± 1.8

10.8 ± 2.2† 17.6 ± 1.6†

15.3 ± 1.7 14.1 ± 1.2

* In a pool-and-split workflow, BC/PC* were subjected to PRT as BC PCs while BC/PC*WB was prepared following PRT treatment of the WB unit. Sample analyses occurred 1 day after production (Day 1) and after 2, 5, and 7 days of storage. Results are displayed as means of six replicates ±1 SD. † Significant difference between the two study arms (p < 0.05).

course of storage with storage-dependent differences achieving significance on Days 5 and 7. Contrary to this, the differences in the lactate levels were significant on Days 1 and 2 of storage. In comparison, in a previous study glucose and lactate levels in untreated PLTs changed during storage from Day 1 to Day 7 from 16.3 ± 0.9 to 12.2 ± 0.9 mmol/L and 6.5 ± 0.4 to 17.0 ± 7.6 mmol/L, respectively.11 PLT activation as measured by the surface expression of P-selectin showed a significant increase over the storage period in both study arms as well as between the two study arms. The responsiveness to ADP decreased significantly over the storage period and was different between the PCs derived from WB PI treatment and directly treated PC. This observation was mirrored in the responsiveness assay using ESC as the readout. To put these data into perspective, our previous study reported PLT activation of 14.3 ± 3.0% to 37.0 ± 5.8%,11 responsiveness to ADP of 58.2 ± 5.5% to 20.7 ± 4.3%, and ESC of 23.2 ± 3.9% to 14.4 ± 5.9% for Day 1 versus Day 7 stored samples, respectively.26 Finally, the development of apoptosis measured as PS exposure significantly increased throughout storage, however, less in the BC/PCWB* compared to the BC/PC*. Untreated samples had PS positivity of 0.54 ± 0.3% on Day 1 to 2.43 ± 0.03% on Day 7 indicating some contribution of apoptosis development from the riboflavin UV light treatment.11 For the RBC units, the concentration of the RBCs derived from the WB PI treatment were similar to the units not treated (Table 2). The MCV was similar within the first weeks of storage but became significantly larger in the units treated as WB. The pH did decrease throughout storage, but no difference between the two study arms could be detected. Similarly, the glucose level decreased during the 42-day storage period with slightly more

glucose consumption observed in the untreated units. Consequently, lactate production over the course of storage was not significantly different between the PI-treated RBCs and the untreated control. Hemolysis levels increased throughout storage with a significantly higher rate seen in the treated units. Between Day 21 and Day 42, the level in the treated units exceeded the 0.8% level, which is the standard for maximum allowable RBC hemolysis in Canada. The potassium levels increased during storage and were significantly elevated in the treated units on all sampling days. Similarly, ATP levels decreased throughout storage and were determined to be significantly lower in the illuminated RBCs. Finally, the RBC-derived microparticles displayed a large increase in the supernatant during storage, which was accelerated in the RBC units derived from the WB-treated units showing significant differences by Day 21 (Table 2). For the plasma units, the effect of WB PI treatment was determined in fresh-frozen units by comparing paired units derived from WB PI treatment with plasma prepared from untreated WB (Table 3). A panel of previously selected variables was used to assess plasma quality, comprising labile FV and FVIII, fibrinogen, and the two global hemostasis screening tests, the PT and the APTT,29 supplemented with a vitamin K-dependent factor (FVII), transglutaminase FXIII, VWF, and two factors previously reported to be reduced when solvent/detergent methods were used to reduce pathogens in plasma, protein S, and α2-antiplasmin.31 The activity of FV, FVII, FVIII, fibrinogen, the free form of protein S, and α2-antiplasmin were significantly lower in the treated units. Finally, significantly prolonged clotting times were observed for both PT and APTT assays in the treated plasma units complementing the reduced factor activities. Volume **, ** **

TRANSFUSION

5

SCHUBERT ET AL.

TABLE 2. Summary of routine RBC quality measures using a variety of in vitro assays comparing RBCs produced by the BC method* Week 0 Week 1 Week 3 Week 6 Assay RBCs RBCWB* RBCs RBCWB* RBCs RBCWB* RBCs RBCWB* 6.2 ± 0.3 6.3 ± 0.3 6.2 ± 0.2 6.2 ± 0.2 6.3 ± 0.2 6.1 ± 0.2 6.1 ± 0.2 6.0 ± 0.2 RBC concentration (×109/L) MCV (fL) 101.5 ± 0.8 100.2 ± 0.8 101.5 ± 1.7 101.5 ± 1.7 102.7 ± 1.5† 107.5 ± 1.3 107.5 ± 1.4† 113.1 ± 1.5 pH 7.13 ± 0.02 7.11 ± 0.03 7.00 ± 0.05 6.96 ± 0.06 6.80 ± 0.06 6.80 ± 0.07 6.60 ± 0.02 6.66 ± 0.08 Glucose (mmol/L) 25.4 ± 3.2 25.0 ± 4.1 21.6 ± 2.0 21.8 ± 1.7 18.5 ± 3.02 18.4 ± 1.7 14.3 ± 0.6 15.2 ± 0.9 Lactate (mmol/L) 3.8 ± 0.0 4.2 ± 0.1 8.8 ± 1.6 9.0 ± 1.6 14.4 ± 2.9 13.7 ± 1.8 17.6 ± 3.0 16.0 ± 2.2 Hemolysis (%) 0.04 ± 0.01 0.05 ± 0.01 0.06 ± 0.01† 0.40 ± 0.01 0.12 ± 0.03† 0.47 ± 0.06 0.27 ± 0.06† 1.04 ± 0.09 Potassium 1.1 ± 0.2† 1.7 ± 0.5 9.6 ± 0.2† 27.1 ± 1.3 20.1 ± 0.9† 35.4 ± 1.3 28.4 ± 1.4† 36.7 ± 1.3 (mmol/L) ATP (μmol/g Hb) 4.25 ± 0.36† 4.14 ± 0.31 4.51 ± 0.28† 4.15 ± 0.32 3.81 ± 0.29† 3.23 ± 0.28 2.67 ± 0.49† 2.00 ± 0.37 749 ± 220 765 ± 220 838 ± 318 2,227 ± 929 1,759 ± 512† 19,900 ± 4,819 6,957 ± 1,860† 89,809 ± 39,421 MP count (×106/L SN) * In a pool-and-split workflow, RBCWB* was prepared after WB PRT and RBCs is an untreated control. Sample analyses occurred one day after production (Week 0) and at 1, 3, and 6 weeks of storage. Results are displayed as means of six replicates ±1 SD. † Significant difference between the two study arms (p < 0.05). MP = microparticle; SN = supernatant.

TABLE 3. Summary of routine plasma quality measures using a variety of in vitro assays comparing FPs produced by the BC method* Assay FV (IU/mL) FVII (IU/mL) FVIII (IU/mL) FXIII (IU/mL) VWF (IU/mL) Protein S (IU/mL) α2-Antiplasmin (IU/mL) Fibrinogen (mg/mL) PT (sec) APTT (sec)

FP 0.81 ± 0.07† 1.03 ± 0.07† 0.83 ± 0.13† 1.12 ± 0.14 0.90 ± 0.18 0.84 ± 0.08† 0.88 ± 0.06† 2.62 ± 0.20† 13.17 ± 0.26† 34.17 ± 0.73†

FPWB* 0.60 ± 0.05 0.78 ± 0.05 0.47 ± 0.07 1.05 ± 0.15 0.87 ± 0.19 0.73 ± 0.04 0.79 ± 0.04 1.85 ± 0.14 15.19 ± 0.35 42.97 ± 1.00

Δ (%) −26 −24 −43 −6 −3 −13 −10 −29 +15 +26

* In a pool-and-split workflow, FPWB* was prepared after WB treatment with PRT while FP is an untreated control. Sample analyses occurred 2 months after being frozen at −80°C on day of production. Results are displayed as means of six replicates ±1 SD. † Significant difference between the two study arms (p < 0.05).

DISCUSSION This paired study reports the impact of PI by riboflavin and UV light applied to WB analyzing the in vitro quality of all blood components produced by the BC method monitored throughout storage at blood banking conditions. The design chosen was a pool-and-split model, which is intended to provide sufficient statistical power with a smaller number of units because of the removal of donorto-donor variation across the study arms. In general, the quality data collected on the RBC units derived from the WB PI (RBCWB*) comparing after treatment and upon 6 weeks of storage are in agreement with the study by Cancelas and colleagues.23 The effects of the treatment of RBCs with riboflavin and UV light treatment include a reduction in glucose consumption and consequently a lower lactate production, but this did not mani6

TRANSFUSION Volume **, ** **

fest as a negative impact on the pH. Potassium levels increased significantly upon treatment both in this study and in the prior report of Cancelas and colleagues.23 This effect is similar to that seen with gamma-irradiated RBCs.32 Increases in potassium release and microparticle production might be biochemically connected based on studies showing that Ca2+-activated efflux of potassium leads to PS exposure and subsequently to microparticle release.33 Furthermore, this observation suggests an impact of the PI treatment on the integrity of the RBC membrane. Declining ATP levels are indicative of the common RBC storage lesion, and lower ATP levels in the treated units support an acceleration of the lesion development. Finally, hemolysis levels rose above 0.8% at the end of a 42-day storage period, exceeding the standard used in Canada for RBC quality control. Hemolysis levels were acceptable up to Day 21, suggesting a reduced shelf life for PRT-treated RBCs prepared after overnight hold at room temperature. The decrease in pH and glucose and the increase in lactate were mainly accounted for by the storage lesion development in RBCs. On the other hand, PI treatment–dependent changes included increases in the RBC MCV, hemolysis levels, and potassium release as well as microparticle production. The impact of microparticles on transfusion outcome is under debate and any clinical effect of increased numbers of microparticles in treated RBCs needs to be assessed. PI treatment of the WB units produced a plasma product with a mean reduction in the activity of multiple coagulation proteins ranging from 10% (α2-antiplasmin) to 44% (FVIII); FXIII and VWF activities were unaffected. PI-associated losses in coagulation protein activity have been previously reported using the riboflavin and UV light treatment of plasma depending on the UV dose.34-36 Cancelas and coworkers noted no significant reduction in fibrinogen, FV, FVIII, or FXI activities after riboflavin and

WHOLE BLOOD PATHOGEN INACTIVATION

UV light treatment at up to 44 J/mLRBCs.23 In contrast, Pidcoke and coworkers37 reported losses in fibrinogen, FV, and FVIII attributable to riboflavin and UV light treatment of WB at 80 J/mLRBCs; protein C, antithrombin, and VWF were unaffected. Given that Bihm and coworkers12 reported 20% to 30% mean losses of multiple coagulation factor activities and riboflavin and UV light treatment of plasma at a UV dose of 6.24 J/mL, it appears that the cellular components of WB provide some degree of protection to plasma proteins from UV-associated damage until a threshold UV dose is crossed. The fact that our findings with plasma protein activity resemble those of Pidcoke and coworkers more than those of Cancelas and coworkers is entirely consistent with the higher doses of UV light employed in this study and in that of Pidcoke and coworkers. We can eliminate trivial explanations for the reduction in plasma protein activities that we observed, such as interference from riboflavin, because we employed the same assays as Coene and colleagues35 who specifically reported no spectral interference from riboflavin in photometric assays, only in fluorescence-based thrombin generation assays. While we did not specifically address the possibility of riboflavin interference in one stage clotting assays, the differential losses that we observed in coagulation-related protein activities also argue against this possibility. Among the coagulation proteins, the most noticeable activity losses found in this study were observed among the three factors considered to be the most labile in plasma (FV, FVII, and FVIII)38 as well as in fibrinogen, also observed in the study by Bihm and colleagues.12 The losses in FVIII activity that we observed brought the treated units to a residual activity level at or below the Canadian regulatory standard of at least 0.52 IU/mL in 75% of FP units tested, suggesting that making transfusable plasma by PI of WB might be problematic with respect to manufacturing a product that would meet regulatory requirements. In this regard, it should be noted that FP typically contains lower FVIII activity levels than FFP, due to the longer period from phlebotomy to freezing, and that this phenomenon cannot be avoided in the BC method of WB processing due to the overnight hold. Canadian regulatory requirements for FP are therefore more stringent than Council of Europe requirements for pathogen-reduced plasma, which dictate a 0.5 IU/mL mean FVIII activity and a retention of 60% of original fibrinogen activity.39 In this small study FP from PI-treated WB met the latter requirements. Finally, in vitro data obtained from PCs treated with PI (BC/PC*) were in agreement with our earlier studies.11,26 The main changes in quality triggered by the treatment are an increase in metabolic activity, increase in PLT activation and signs of apoptotic development as well as a reduction of PLT responsiveness beyond the effect of the PLT storage lesion. In comparison to PLTs treated using the currently approved procedure (BC/

PC*), the quality of PCs derived from the WB treated with riboflavin and UV light (BC/PCWB*) was less affected by storage and showed significantly improved quality on Day 7 of storage for pH, glucose, PLT activation, ESC, and annexin V binding. This trend is similar to the situation for PLT-rich plasma-derived PLTs on Days 1 and 5 after WB treatment and production.23 A potential explanation for this improved quality of the PC derived from the WB treatment compared to the treated PC (BC/PC*) could be a partial absorption of UV light by Hb. The riboflavin and UV light system applies UV light at a wavelength of 280 to 360 nm and the absorption spectrum of Hb exhibits a high extinction coefficient in this range peaking at about 410 nm possibly leading to a less intense treatment dose. Taken together, the data presented here show that PCs derived from WB PI treatment are less negatively impacted than by treatment of the finished component. The RBCs exhibit reductions of in vitro quality triggered by the riboflavin treatment of the WB; the clinical impact of these changes is currently being assessed through in vivo radiolabel and recovery studies in the United States (IMPROVE II; ClinicalTrials.gov Identifier NCT01907906) and patient clinical trials with WB products stored for up to 21 days post-PI in Western Africa (AIMS Study; ClinicalTrials.gov Identifier NCT02118428). Whether the activity loss of plasma coagulation factors is relevant with respect to the ability of transfused plasma to oppose coagulopathy and reduce bleeding needs to be determined by in vivo studies or clinical trials. Several studies have been reported showing that inactivation levels of white blood cells, bacteria, parasites, and viruses with the WB process are the same compared to the PLT and plasma system.16,18,20,22,40 From these data, it can be concluded that WB treatment with riboflavin and UV light might be a practical alternative for blood banks to obtain safe blood components in a one-step procedure, which should also contribute to a cost reduction compared to individual component treatment.

ACKNOWLEDGMENTS The authors thank the Canadian Blood Services Development Laboratory (netCAD) for providing processed blood components and Dr Susanne Marschner and Janna Mundt (both of TerumoBCT) for arranging disposables. We thank Dr Geraldine Walsh, Canadian Blood Services scientific writer, for proofreading and editorial contribution. This study was supported in part by a grant from Health Canada and Canadian Blood Services to DVD. The views expressed herein do not necessarily represent the view of the federal government. PS and DVD designed the study; PS, BC, SK, KS, EL, DB, and VB performed the research; PS, KS, EL, WPS, and RPG analyzed and interpreted the data; and PS, WPS, and DVD wrote the manuscript. Volume **, ** **

TRANSFUSION

7

SCHUBERT ET AL.

summary of an NHLBI workshop. Transfusion 2009;49:

CONFLICT OF INTEREST RPG is an employee of TerumoBCT. The disposables and the instrumentation for conducting the pathogen reduction by the Mirasol process were provided without charge by TerumoBCT. TerumoBCT was not involved in the study design and provided no editorial control over the research or manuscript. This study was supported by TerumoBCT with funds solely for reagents. Otherwise, none of the authors have any competing financial interest.

REFERENCES

1262-8. 14. Rock G. A comparison of methods of pathogen inactivation of FFP. Vox Sang 2011;100:169-78. 15. Wagner SJ. Developing pathogen reduction technologies for RBC suspensions. Vox Sang 2011;100:112-21. 16. Goodrich RP, Doane S, Reddy HL. Design and development of a method for the reduction of infectious pathogen load and inactivation of white blood cells in whole blood products. Biologicals 2010;38:20-30. 17. Fast LD, Nevola M, Tavares J, et al. Treatment of whole blood with riboflavin plus ultraviolet light, an alternative to gamma irradiation in the prevention of transfusionassociated graft-versus-host disease? Transfusion 2013;53: 373-81.

1. Glynn SA, Busch MP, Dodd RY, et al. Emerging infectious agents and the nation’s blood supply: responding to potential threats in the 21st century. Transfusion 2013;53:

18. Doane S, Keil SD, Gilmour DI, et al. Treatment of whole blood with riboflavin and UV light: effectiveness against clinically relevant bacterial contamination. Transfusion

438-54. 2. Lozano M, Cid J. Pathogen inactivation: coming of age. Curr Opin Hematol 2013;20:540-5. 3. Prowse CV. Component pathogen inactivation: a critical review. Vox Sang 2013;104:183-99. 4. Seltsam A, Muller TH. Update on the use of pathogenreduced human plasma and platelet concentrates. Br J Haematol 2013;162:442-54. 5. Custer B, Agapova M, Martinez RH. The cost-effectiveness of pathogen reduction technology as assessed using a multiple risk reduction model. Transfusion 2013;50: 2461-73. 6. Hervig T, Seghatchian J, Apelseth TO. Current debate on pathogen inactivation of platelet concentrates—to use or not to use? Transfus Apher Sci 2010;43:411-4. 7. Cazenave JP, Folléa G, Bardiaux L, et al. A randomized controlled clinical trial evaluating the performance and safety of platelets treated with MIRASOL pathogen reduction

2013;53:203A. 19. Keil SD, Rapaport R, Doane SK, et al. Viral reduction of intracellular HIV using the Mirasol system for whole blood. Vox Sang 2012;103:144. 20. Tonnetti L, Thorp AM, Reddy HL, et al. Evaluating pathogen reduction of Trypanosoma cruzi with riboflavin and ultraviolet light for whole blood. Transfusion 2012;52:40916. 21. Tonnetti L, Thorp AM, Reddy HL, et al. Riboflavin and ultraviolet light reduce the infectivity of Babesia microti in whole blood. Transfusion 2013;53:860-7. 22. El Chaar M, Atwal S, Freimanis GL, et al. Inactivation of Plasmodium falciparum in whole blood by riboflavin plus irradiation. Transfusion 2013;53:3174-83. 23. Cancelas JA, Rugg N, Fletcher D, et al. In vivo viability of stored red blood cells derived from riboflavin plus ultravio-

technology. Transfusion 2010;50:2362-75. 8. Seghatchian J, Hervig T, Putter JS. Effect of pathogen inac-

9. 10.

11.

12.

13.

8

tivation on the storage lesion in red cells and platelet concentrates. Transfus Apher Sci 2011;45:75-84. Prowse C. Properties of pathogen-inactivated plasma components. Transfus Med Rev 2009;23:124-33. Picker SM, Steisel A, Gathof BS. Effects of Mirasol PRT treatment on storage lesion development in plasma-stored apheresis-derived platelets compared to untreated and irradiated units. Transfusion 2008;48:1685-92. Schubert P, Culibrk B, Coupland D, et al. Riboflavin and ultraviolet light treatment potentiates vasodilatorstimulated phosphoprotein Ser-239 phosphorylation in platelet concentrates during storage. Transfusion 2012;52: 397-408. Bihm DJ, Ettinger A, Buytaert-Hoefen KA, et al. Characterization of plasma protein activity in riboflavin and UV light-treated fresh frozen plasma during 2 years of storage at −30 degrees C. Vox Sang 2010;98:108-15. Klein HG, Glynn SA, Ness PM, et al. Research opportunities for pathogen reduction/inactivation of blood components: TRANSFUSION Volume **, ** **

24.

25.

26.

27.

28.

let light-treated whole blood. Transfusion 2011;51: 1460-8. Holme S. In vitro assays used in the evaluation of the quality of stored platelets: correlation with in vivo assays. Transfus Apher Sci 2008;39:161-5. Levin E, Culibrk B, Gyongyossy-Issa MI, et al. Implementation of buffy coat platelet component production: comparison to platelet-rich plasma platelet production. Transfusion 2008;48:2331-7. Schubert P, Coupland D, Culibrk B, et al. Riboflavin and ultraviolet light treatment of platelets triggers p38MAPK signaling: inhibition significantly improves in vitro platelet quality after pathogen reduction treatment. Transfusion 2013;53:3164-73. Han V, Serrano K, Devine DV. A comparative study of common techniques used to measure haemolysis in stored red cell concentrates. Vox Sang 2010;98:116-23. Bhakta V, Jenkins C, Ramirez-Arcos S, et al. Stability of relevant plasma protein activities in cryosupernatant plasma units during refrigerated storage for up to 5 days postthaw. Transfusion 2013;54:418-25.

WHOLE BLOOD PATHOGEN INACTIVATION

29. Sheffield WP, Bhakta V, Jenkins C, et al. Conversion to the buffy coat method and quality of frozen plasma derived from whole blood donations in Canada. Transfusion 2010; 50:1043-9. 30. Sheffield WP, Bhakta V, Talbot K, et al. Quality of frozen transfusable plasma prepared from whole blood donations in Canada: an update. Transfus Apher Sci 2013;49:440-6.

with three methods of pathogen reduction. Transfusion 2014;54:1321-31. 36. Ettinger A, Miklauz MM, Hendrix BK, et al. Quality of proteins in riboflavin and UV light-treated FFP during 1 year of storage at −18 degrees C. Transfus Apher Sci 2012;46: 15-8. 37. Pidcoke HF, McFaul SJ, Ramasubramanian AK, et al.

31. Heger A, Romisch J, Svae TE. Stability of solvent/detergent-

Primary hemostatic capacity of whole blood: a compre-

treated plasma and single-donor fresh-frozen plasma during 48 h after thawing. Transfus Apher Sci 2005;33:

hensive analysis of pathogen reduction and refrigeration effects over time. Transfusion 2013;53(Suppl 1):

257-67. 32. Serrano K, Chen D, Hansen AL, et al. The effect of timing of gamma-irradiation on hemolysis and potassium release in leukoreduced red cell concentrates stored in SAGM. Vox Sang 2014;106:379-81. 33. Morrell CN. Immunomodulatory mediators in platelet transfusion reactions. Hematology Am Soc Hematol Educ Program 2011;2011:470-4. 34. Benjamin RJ, McLaughlin LS. Plasma components: properties, differences, and uses. Transfusion 2012; 52(Suppl 1):9S-19S. 35. Coene J, Devreese K, Sabot B, et al. Paired analysis of plasma proteins and coagulant capacity after treatment

137S-49S. 38. Kakaiya RM, Morse EE, Panek S. Labile coagulation factors in thawed fresh frozen plasma prepared by two methods. Vox Sang 1984;46:44-6. 39. Council of Europe. Guide to the preparation, use and quality assurance of blood components. Strasbourg: Directorate for the Quality of Medicines & Health Care of the Council of Europe, 2011. 40. Tonnetti L, Proctor MC, Reddy HL, et al. Evaluation of the Mirasol pathogen [corrected] reduction technology system against Babesia microti in apheresis platelets and plasma. Transfusion 2010;50:1019-27.

Volume **, ** **

TRANSFUSION

9

Whole blood treated with riboflavin and ultraviolet light: quality assessment of all blood components produced by the buffy coat method.

Pathogen inactivation (PI) technologies are currently licensed for use with platelet (PLT) and plasma components. Treatment of whole blood (WB) would ...
207KB Sizes 0 Downloads 16 Views