The Effect of Filtration on Residual Levels of Coagulation Factors in Plasma

Transcription

1 Coagulation and Transfusion Medicine / Filtration of Plasma The Effect of Filtration on Residual Levels of Coagulation Factors in Plasma Hiba S. Alhumaidan, MD, 1 Tracey A. Cheves, MT, 1 Stein Holme, PhD, 2 and Joseph D. Sweeney, MD 1 Key Words: Blood filtration; Clotting factors DOI: 1.139/AJCPRRESG7PGIAH5 Abstract blood components are commonly manufactured via filtration. There are specifications for the residual leukocyte content of any final cellular blood component but not for residual clotting factors. and nonleukoreduced platelet-poor products were manufactured from filtered vs unfiltered platelet-rich, respectively, using platelet leukoreduction filters. The leukoreduced showed lower levels of factor VIII (75% ± 16% vs 88% ± 18%, P.5), factor XI (86% ± 9% vs 96% ± 1%, P.1) and factor VII (87% ± 14% vs 98% ± 11%, P.1). No difference was seen with factor X, factor V, or fibrinogen. Plasma filtered through a whole blood filter showed a reduction in factor V (15% ± 12% vs 124% ± 1%, P.1) but a minimal reduction in factor VIII (8% ± 5% vs 82% ± 6%, P =.4). Filtration can alter the residual levels of clotting factors to a variable extent in manufactured, most noticeably factors V, VII, VIII, and XI. Whole blood donations undergo centrifugation and filtration during the manufacture of different components. 1 Prestorage leukoreduction by filtration is a common processing step that can be performed on whole blood (WB), intermediate products such as platelet-rich (), or the final blood product. 1,2-4 In recent years, universal leukodepletion of cellular products and, in some countries, acellular components has become a widespread practice. 5 Standards and specifications exist for the residual cellular content of the final leukoreduced blood component, though these differ slightly between the United States and Europe. 1,6,7 No current standards exist for residual leukocytes in leukoreduced products in the United States. However, the French standard is less than 1 6 /L. 8 No specific standards exist regarding the residual clotting factor content for the 2 licensed products manufactured from WB in the United States, fresh frozen (FFP), and 24-hour frozen (FP24), leukoreduced or otherwise. However, in Germany the requirement is that the contains 7% of the original level of factor (F) VIII:C after thawing and that the concentration of FVIII:C exceed.7 IU/mL 9. At this time, frozen and manufactured after storing WB or for 24 hours at room temperature (24RTFP) is an unlicensed product but could conceivably be licensed in the future. 1 Prestorage filtration of any of these products will occur if the whole blood is filtered, the is filtered using an in-line system, or if the platelet-poor (PPP) is filtered before freezing. Previous studies have examined the effect of whole blood or RBC filters on the clotting factor content of the final product, both immediately after thawing and after storage in the liquid state at 1 C to 6 C Reports on the effect of filters on proteins 11 Am J Clin Pathol 13;139: DOI: 1.139/AJCPRRESG7PGIAH5

2 Coagulation and Transfusion Medicine / Original Article are limited, 16,17 and to our knowledge, except for 1 report on a prototype filter, no other reports have described the effect of platelet filters. 19 In a previous report, we examined the stability of coagulation factors in leukoreduced FFP and compared it with that of leukoreduced 24RTFP. 1 The aim of this study was to compare prestorage leukoreduced manufactured using as an intermediate product with nonleukoreduced to examine whether leukofiltration had an effect on residual clotting factors and to systematically evaluate the difference between filtered through a WB filter and a platelet filter. Materials and Methods These studies were approved by the institutional research board of The Miriam Hospital (Providence, RI). WB donations were collected from donors who met all American Association of Blood Banks and US Food and Drug Administration (FDA) criteria into 5-mL containers using citrate phosphate double dextrose as anticoagulant. 1 The initial step A 8 hours FFP Pooled leukoreduced was the centrifugation of the WB within 8 hours by a soft spin (1,9g for 4.5 minutes) using a RC3BP Sorvall Centrifuge (Fisher Scientific, Pittsburgh, PA) to produce and an RBC concentrate stored in AS3. Two different manufacturing schemas were then used. Schema 1: Plasma Products was leukoreduced by inline filtration (Leukotrap RC-PL, Pall Medical, Covina, CA). 2 Two ABO-identical leukoreduced s were then pooled and subsequently divided into 2 equal aliquots yielding identical pairs of. One of the paired products was centrifuged within 8 hours on the day of collection by a hard spin centrifugation (2,35g for 8 minutes) to produce PPP and platelet concentrate. The PPP was then frozen at 18 C, yielding leukoreduced FFP. The second of the paired products was left unagitated at room temperature for an additional 18 to 24 hours. On the day after collection, this was centrifuged by a hard spin centrifugation to PPP and a platelet concentrate. The PPP was then frozen at 18 C, yielding leukoreduced 24RTFP. This schema is illustrated in Figure 1A. B 8 hours FFP Pooled nonleukoreduced 8-24 hours 24RTFP 8-24 hours 24RTFP Figure 1 A schematic representation of both manufacturing processes yielding leukoreduced (A) and nonleukoreduced (B) products. 24RTFP, frozen after 24-hour room temperature storage; FFP, fresh frozen ;, platelet-rich ;, whole blood donation. Am J Clin Pathol 13;139: DOI: 1.139/AJCPRRESG7PGIAH5 111

3 Alhumaidan et al / Filtration of Plasma Schema 2: Plasma Products In this schema, the was expressed without passing through a filter, but all other manufacturing aspects were identical to schema 1, yielding pairs of nonleukoreduced FFP and nonleukoreduced 24RTFP. This schema is illustrated in Figure 1B. Both schema 1 and schema 2 manufactured identical pairs of frozen products for comparative studies that were the subject of previous reports. 1, Frozen pairs from both schemas were stored at 18 C for a minimum of 54 days before testing for schema 1 and 36 days before testing for schema 2. Pairs were then thawed in a waterbath at 37ºC. Coagulation factor assays were performed on the day of thaw and on day 5 and day 7 of liquid storage at 1 C to 6 C after thawing. Plasma Filtration Studies Using WB and Platelet Filters Five pairs of nonleukoreduced FFP and 24RTFP product from schema 2 were used in this experiment after the had been frozen for at least 2 months. Paired units were thawed, pooled, and a sample taken from the pool (prefiltration sample). Immediately afterwards, the pool was divided into 2 identical aliquots. One aliquot was filtered using a whole blood filter (WBF3, Pall Medical) and the second aliquot using a platelet filter (LPS1, Pall Medical). After filtration, the liquid pairs were stored at 1 C to 6 C for 7 days. Postfiltration samples were taken on day (the day of thaw), day 3, day 5, and day 7 for selected Identical aliquots FFP Pooled nonleukoreduced clotting factor assays. The post-thaw processing and testing schema is illustrated in Figure 2. Coagulation Assays The clotting factor assays performed were prothrombin time (PT), activated partial thromboplastin time (APTT), FV, FVII, FVIII, FX, FXI, fibrinogen, antithrombin III (ATIII), protein C, and free protein S. The PT and aptt were measured using an automated coagulation analyzer and manufacturer s reagents in a device that uses optical endpoint for clot detection (MDA II, TCoag, Wicklow, Ireland). Assays for FV, FVII, FVIII, FX, and FXI were all conducted using a 1-stage chronometric assay with factor-deficient reagents (Precision BioLogic, Dartmouth, Nova Scotia, Canada) in the same analyzer. Fibrinogen was measured using the Clauss method (Fibriquik, TCoag). ATIII and protein C were measured using a chromogenic assay with reagents (Diagnostica Stago, Parsippany, NJ) in the MDA II. Free protein S assay was performed using an enzyme-linked immunosorbent assay (Diagnostica Stago). Statistical Analysis Results were entered into an Excel spreadsheet (Microsoft, Redmond, WA), exported into a statistical application (Epistat, Richardson, TX) and descriptive statistics derived. Data were analyzed using independent t tests or paired t tests as appropriate. Pearson correlation coefficient was 24RTFP Prefiltration sample Leukoreduction by filtration WB filter filter filter Postfiltration samples D, D3,, Figure 2 A schematic representation of the filtration experiment using whole blood and platelet filters yielding paired filtered products. 24RTFP, frozen after 24-hour room temperature storage; D, day of filtration; D3, day 3 of liquid storage after filtration;, day 5 of liquid storage after filtration;, day 7 of liquid storage after filtration; FFP, fresh frozen ;, platelet-rich ; WB, whole blood. 112 Am J Clin Pathol 13;139: DOI: 1.139/AJCPRRESG7PGIAH5

4 Coagulation and Transfusion Medicine / Original Article used in univariant analysis. Statistical significance was defined as a P value less than or equal to.5. Results Sixty WB donations were processed using both manufacturing schemas for a total of 1 products. Ten paired products (FFP and 24RTFP), 5 group O and 5 group A, were chosen at random from both schema 1 and schema 2, then thawed and tested. Table 1 shows the comparison of leukoreduced products (both FFP and 24RTFP) manufactured using schema 1 with nonleukoreduced products manufactured using schema 2 (both FFP and 24RTFP). The Table 1 Coagulation Assay Results in the Filtered (Schema 1) and Nonfiltered (Schema 2) Plasma Units a data from both product types were combined for clarity after the initial assessment showed a similar effect of filtration on both FFP and 24RTFP. On day (day of thaw), all clotting factor levels, with the exception of FX, FV, and fibrinogen, appeared reduced in the filtered. The aptt was longer in the filtered, reflecting these changes. With liquid storage, there was a large decline (%-3%) in factor FV and FVIII:C in both product types, as illustrated in Figure 3, with lesser declines in the other clotting factors (5%-1%). The pattern of differences between the product types persisted on both day 5 and day 7, with statistical significance seen with FVIII, FXI, FV, and ATIII on either of these days and on day 5 with FVII, which may be an anomalous result related to the phenomenon of cold activation of FVII. Day Day 5 Day 7 Assays Filtered Plasma Nonfiltered Plasma Filtered Plasma Nonfiltered Plasma Filtered Plasma Nonfiltered Plasma PT ( sec ) 13.8 ± ±.5 16 ± ± ± ± 2.1 APTT (21-33 sec ) 33.9 ± ± 2.4 b 37.5 ± ± 2.7 b 39 ± ± 3.6 b Fibrinogen (15-48 mg/dl) 258 ± ± ± ± ± ± 27 FV (6%-16%) 122 ± ± ± 11 1 ± 19 b 74 ± 1 84 ± 19 b FVII (5%-15%) 87 ± ± 11 b 75 ± ± 6 76 ± ± 51 c FVIII (5%-18%) 75 ± ± 18 c 51 ± 9 61 ± 1 b 47 ± 9 59 ± 9 b FX (8%-15%) 96 ± 1 95 ± 9 91 ± ± ± ± 13 ATIII (8%-13%) 1 ± 6 16 ± 8 b 96 ± 6 13 ± 8 b 94 ± 6 12 ± 9 b FXI (6%-16%) 86 ± 9 96 ± 1 b 84 ± 8 97 ± 11 b 88 ± ± 11 c APTT, activated partial thromboplastin time; ATIII, antithromboplastin III; FV, factor V; FVII, factor VII; FVIII, factor VIII; FX, factor X; FXI, factor XI; PT, prothrombin time. a Coagulation results are from the day of thaw (day ) and day 5 and day 7 of liquid storage (n = ). Data were analyzed using independent t tests. Data are the mean ± 1 standard deviation. Reference ranges are given in parenthesis. b P.1, filtered vs unfiltered products. c P.5, filtered vs unfiltered products. A Factor V (%) B Factor VIII (%) Filtered Nonfiltered Day Day 5 Day 7 Filtered Nonfiltered Day Day 5 Day 7 Figure 3 A graphic representation of the decline in factor V (A) and factor VIII (B) in both filtered and nonfiltered thawed product. Day, day of thaw; day 3, day 3 of liquid storage; day 7, day 7 of liquid storage. Am J Clin Pathol 13;139: DOI: 1.139/AJCPRRESG7PGIAH5 113

5 Alhumaidan et al / Filtration of Plasma Table 2 Coagulation Assay Results in Pre- and Postfiltered Paired Plasma Products a Postfiltration Day Assays Prefiltration WB Filter Filter FV (6%-16%) 126 ± ± 12 b 124 ± 1 b FVIII (5%-18%) 82 ± 6 8 ± 5 76 ± 4 c Fibrinogen (15-48 mg/dl) 316 ± ± ± 14 Protein C (78%-17%) 19 ± ± ± Protein S (6%-1%) 77 ± 8 64 ± 4 c 64 ± 8 c FV, factor V; FVIII, factor VIII;, platelet-rich ; WB, whole blood. a Coagulation assays were conducted using or filter (n = 5). Data were analyzed using paired t tests. Data are the mean ± 1 standard deviation. Reference range for normal values are given in parenthesis. b P.1, comparing both filters. c P.1, comparing pre- and postfiltration products. A Factor V (%) C Free Protein S (%) Prefiltration D D3 8 6 filter filter Table 2 shows the results of the filtration studies using the WB and platelet filters. Both filters caused a small decrease in protein S, and the platelet filter caused a small decrease in FVIII:C. A statistically insignificant reduction in fibrinogen was evident, which was similar to the differences observed in Table 1. Of particular note was the % to 3% reduction in FV after filtration using the, which was not seen with the platelet filter. During liquid storage to day 7, fibrinogen remained stable but levels of FV, FVIII:C, and protein C declined. These changes are illustrated in Figure 4. FV showed a gradual decrease of about % to day 7 after filtration, but the difference between the 2 filtered products reduced with liquid storage was not statistically significant after day 3. FVIII:C showed a more abrupt decline between day and day 3, with little or no change after day 3. Protein C was stable until day 3, and B D Factor VIII (%) Protein C (%) Prefiltration D D filter filter Prefiltration D D3 Prefiltration D D3 Figure 4 A graphic representation of clotting factors V (A), VIII (B), free protein S (C), and protein C (D) before and after filtration and up to 7 days of liquid storage. D, day of filtration; D3, day 3 of liquid storage after filtration;, day 5 of liquid storage after filtration;, day 7 of liquid storage after filtration;, platelet-rich ; WB, whole blood. 114 Am J Clin Pathol 13;139: DOI: 1.139/AJCPRRESG7PGIAH5

6 Coagulation and Transfusion Medicine / Original Article then declined by about 1%. Protein S showed a decline of approximately 13% after filtration for both filter types, but levels increased after liquid storage to prefiltration levels by day 5 to day 7, possibly representing dissociation of free protein S from protein S bound to C4B-binding protein. Discussion Our results show that leukoreduction filters have a variable effect on the residual clotting factor content of manufactured products. No statistically significant differences were observed with FX, fibrinogen, or protein C; small changes were observed with ATIII and protein S; and larger differences were observed with FVII, FXI, and FVIII. In the case of FV, a significant decrease was observed only with the. The mechanism of these changes is unclear but is unlikely to be solely because of simple protein attachment to a foreign surface. The proprietary coating of these polyester filter surfaces is probably a contributing factor, which confers a degree of selectivity in clotting factor adherence. Nine previous reports, mostly from Europe, have addressed changes in clotting factors after the leukofiltration of either WB or. Heiden et al 11 studied the effect of s from 5 different manufacturers, 4 of which were polyester filters and 1 was a polyurethane filter. No statistically significant reductions in clotting factors were observed after filtration. Kretzschmar et al, 12 using a polyurethane WB filter, documented a postfiltration increase in aptt and an associated decline in FVIII:C, but this was only significant after the was stored at room temperature for 18 hours before processing. Williamson et al, 13 using s, assessed filtration after either warm (room temperature) or cold (1 C-6 C) storage and showed a reduction of FV (3%) with filtration. A reduction in FVIII was observed with WB filtration only after warm storage for 2 to 4 hours before processing. Solheim et al 14 used a and demonstrated a lower but more stable FV and FVIII in the filtered product after thawing and attributed this to platelet retention on the with subsequent activation and binding of FV. Runkel et al 15 studied 2 different s, a positively charged polyester filter and an uncharged polyester filter. Postfiltration fibrinogen, FV, and FVIII:C levels were measured and were found to be not different from those of a control population, but no prefiltration samples were tested. Cardigan et al 16 evaluated the effect of 5 s and 2 filters on the residual clotting factors of leukoreduced FFP. Statistically significant losses in factors V, VIII, IX, XI, and XII were observed with 2 of the 5 s. One of the s appeared to cause an increase in FV, another an increase in fibrinogen; both were attributed to release from platelet alpha granules. The filters showed a variable reduction in FVIII, FIX, and FXI. Of note, there was no effect on FV using the 2 filters. Chabanel et al 17 studied 4 different filters and the effect on manufactured from filtered WB or nonfiltered WB. Plasma filtered from nonfiltered WB had a higher FVIII and von Willebrand factor; a mild reduction in FVIII, FIX, and FXI was observed with 1 of the 4 filters; and a significant decrease was observed in FIX and FXI with another. Snyder et al 18 studied only postfiltration samples on the day of processing and 6 and 12 months later. No time-related differences with frozen storage were seen, and because prefiltration samples were not taken, no filter effect could be assessed. To our knowledge, the study by Dzik et al 19 is the only previous report on the effect of filtration on clotting factors, and this report found no reduction in the concentration of clotting factors. The results are clearly heterogeneous, which is likely related to the type of product filtered (WB or ), the type of filter used, and the time or temperature of product storage before processing, in addition to possible unknown variability related to the filtration process. Our results show similarity to these previously published studies. Factor X and protein C are both vitamin K dependent glycoproteins, and the oligosaccharide chains could cause an electrostatic repulsion from the negatively charged coated polyester surfaces, which restricts adherence. FXI is largely bound to high-molecular-weight kininogen, which has an affinity for negatively charged surfaces, and which explains the loss of this protein. FVII attaches to tissue factor in vivo; it is possible that cell membrane fragments containing tissue factor or surface-attached monocytes expressing tissue factor could explain any decrease observed. No clear explanation can be found for the loss of FVIII, but this appears more variable and is inconsistent. The loss of FV with the is more intriguing because the difference in FV activity appears to decrease with liquid storage (Figure 4). FV is known to exist in as 2 isoforms that differ in glycosylation in the C2 domain at position N The glycosylated form is designated FV1 and the nonglycosylated form FV2. FV1 has decreased binding affinity for negatively charged phospholipids but also less prothrombinase activity. Alteration in the ratio of FV1 to FV2 by a relatively selective binding of the FV2 isoform to the could explain the difference. If the FV1 isoform is more stable ex vivo, this difference would be attenuated with liquid storage, as observed. Our study has some important limitations. First, we studied only filters from a single manufacturer and the results may not be generalizable to other filter types. Second, our assays were limited and not all clotting factors or anticoagulant proteins were studied. Third, we studied PPP filtration through WB and platelet filters, which would not occur in practice. Nevertheless, our results are similar to those of previously published reports, thus lending credence to the data Am J Clin Pathol 13;139: DOI: 1.139/AJCPRRESG7PGIAH5 115

7 Alhumaidan et al / Filtration of Plasma with the exception of protein S, which has shown postthaw instability in other reports. 22 These observations may have manufacturing and clinical implications. From a manufacturing perspective, any reduction in FVIII:C could have implications because FVIII:C is still acquired from in the manufacture of -derived concentrates. From a clinical perspective, the reduction in FV with the and the reduction in FXI with the platelet filter may have implications for dosing. No concentrate is available for either of these factors in the United States; is used as a source of FV in both hereditary FV deficiency and the acquired coagulopathies of liver disease, disseminated intravascular coagulation, and hemodilution coagulopathy and as a source of FXI in the treatment of FXI deficiency (hemophilia C). Although dosing is inexact, the type of product transfused (filtered vs nonfiltered) does not appear to have received any attention in the clinical outcome of trials involving transfusion. 23 In conclusion, leukoreduction filters influence the protein content of the manufactured product. This aspect appears to have received minimal attention in the United States and may be important for the future design and overall performance characteristics of these filters. Furthermore, the change in FV and FVIII between day 5 and day 7 is minimal and the extension of liquid storage to day 7 has the potential to reduce wastage of thawed product. From the 1 Blood Bank and Transfusion Medicine Research Unit, The Miriam Hospital, Providence, RI; and 2 Pall Medical, Covina, CA. This study was supported by a grant from Pall Medical, Covina, CA. Dr Sweeney has received honoraria and grant support from Pall Medical; Dr Holme is an employee of Pall Medical. Address reprint requests to Dr Sweeney: Coagulation and Transfusion Medicine, The Miriam Hospital, 164 Summit Ave, Providence, RI 296; jsweeney@lifespan.org. References 1. Roback JD, Grossman BJ, Harris T, et al, eds. AABB Technical Manual. 17th ed. Bethesda, MD: American Association of Blood Banks; Sweeney JD, Holme S, Heaton WA, et al. White cell reduced platelet concentrates prepared by in-line filtration of platelet rich. Transfusion. 1995; 35: Rebulla P, Porretti L, Bertolini F, et al. White cell-reduced red cells prepared by filtration: a critical evaluation of current filters and methods for counting residual white cells. Transfusion. 1993;33: Riggert J, Simson G, Dittmann J, et al. Prestorage leukocyte depletion with in-line filtration of whole blood in comparison with blood component leukocyte depletion. Vox Sang. 1995;69: Sweeney JD. Universal leukoreduction of cellular blood components in 1? Am J Clin Pathol. 1;115: American Association of Blood Banks. Standards for Blood Bank and Transfusion Services. 27th ed. Bethesda, MD: American Association of Blood Banks; European Committee (Partial Agreement) on Blood Transfusion. Guide to the Preparation, Use and Quality Assurance of Blood Components. 16th ed. Strasbourg, France: Council of Europe Publishing; Masse M. Universal leukoreduction of cellular and components: process control and performance of the leukoreduction process. Transfus Clin Biol. 1;8: Madla W, Alt T, Jungk H, et al. Fresh frozen quality: relation to age and gender of blood donors. Vox Sang. 12; 12: Alhumaidan H, Cheves T, Holme S, et al. Stability of coagulation factors in prepared after a 24-hour room temperature hold. Transfusion. 1; 5: Heiden M, Salge U, Henschler R, et al. Plasma quality after whole-blood filtration depends on storage temperature and filter type. Transfus Med. 4;14: Kretzschmar E, Kruse F, Greiss O, et al. Effects of extended storage of whole blood before leucocyte depletion on coagulation factors in. Vox Sang. 4;87: Williamson LM, Rider JR, Swann ID, et al. Evaluation of and red cells obtained after leucocyte depletion of whole blood. Transfus Med. 1999;9: Solheim BG, Flesland O, Brosstad F, et al. Improved preservation of coagulation factors after pre-storage leukocyte depletion of whole blood. Transfus Apher Sci. 3;29: Runkel S, Bach J, Haubelt H, et al. The impact of two whole blood inline filters on markers of coagulation, complement and cell activation. Vox Sang. 5;88: Cardigan R, Sutherland J, Garwood M, et al. The effect of leucocyte depletion on the quality of fresh-frozen. Br J Haematol. 1;114: Chabanel A, Sensebé I, Masse M, et al. Quality assessment of seven types of fresh-frozen leucoreduced by specific filtration. Vox Sang. 3;84: Snyder EL, Whitley P, Kingsbury T, et al. In vitro and in vivo evaluation of a whole blood platelet-sparing leukoreduction filtration system. Transfusion. 1;5: Dzik WH, Cusack WF, Gacek MJ. Preparation of white cell-reduced platelet concentrates from whole blood during component preparation. Transfusion. 1991;31: Alhumaidan H, Cheves T, Holme S, et al. Comparison of clotting factors in fresh frozen and 24 hour room temperature hold frozen [abstract]. Transfusion. 1;5(S2):8A. 21. Cramer TJ, Gale AJ. The anticoagulant function of coagulation factor V. Thromb Haemost. 12;17: Thiele T, Kellner S, Hron G, et al. Storage of thawed for a liquid bank: impact of temperature and methylene blue pathogen inactivation. Transfusion. 12;52: Stanworth SJ, Brunskill SJ, Hyde CJ, et al. Is fresh frozen clinically effective? A systematic review of randomized controlled trials. Br J Haematol. 4;126: Am J Clin Pathol 13;139: DOI: 1.139/AJCPRRESG7PGIAH5