Two solvent/detergent‐treated plasma products with a different biochemical profile

The distribution of US‐produced solvent/detergent (S/D)‐treated plasma (Plas+SD) was discontinued in 2003 due to bleeding and thrombotic events. In this work, we were searching for explanations that might explain these severe side‐effects.


Introduction
Transfusion of human plasma is essential to treat coagulation disorders and to control bleeding. However, enveloped and non-enveloped viruses such as human immunodeficiency virus (HIV), hepatitis viruses (A, B, C and E), West Nile virus, parvovirus B19 (B19), cytomegalovirus and human T-cell lymphotrophic viruses are capable of being transmitted through plasma transfusion [1]. Although the implementation of HIV, HBV and HCV nucleic acid testing (NAT) has improved its safety [2], there are still cases reported in which viruses were transmitted through plasma transfusions [3][4][5]. Therefore, preventing pathogen transmission through plasma transfusion is of great importance. The solvent/detergent (S/D) inactivation method has been successfully used since 1985 in the manufacturing of coagulation factor concentrates [6]. Also, this inactivation technology has been applied to fresh frozen plasma (FFP) to reduce the risk of enveloped virus transmission through plasma transfusions [7,8]. Moreover, the virus validation studies showed very effective inactivation of different enveloped viruses, which led to a high safety margin [7,9].
As S/D-treated plasma is produced from a large donor pool, the transmission risk of a virus that is resistant to this inactivation procedure is increased. However, as discussed by Rollag and colleagues, neutralizing antibodies that occur in the plasma pools function as important barriers against many enveloped and non-enveloped viruses, which might contaminate the plasma pool [10].
Transmission of parasites [11], such as Plasmodium falciparum (causing malaria), Babesia microti (causing babesiosis), Trypanosoma cruzi (causing Chagas disease) as well as various bacteria [12] including Staphylococcus aureus and Escherichia coli, more often occur in cellular blood components. The likelihood of these pathogens to survive the plasma-freezing process that is performed after donation (at temperatures under -18°C) is very low. Furthermore, as S/D-treated plasma products are sterile filtered, bacteria and parasitic infections do not pose a risk [13].
As specific prion proteins, which are associated with Creutzfeldt-Jakob disease, can be transmitted through blood and blood products [14,15], a prion removal step was added to the manufacturing process of the S/D-treated plasma which is, in most countries, named OctaplasLG (second-generation product). Since several years, this product is manufactured by the company Octapharma and is available in Europe, as well as in Canada and the USA.
A S/D-treated plasma product, named Plas+SD, manufactured by V.I. Technologies, Inc. (VITEX) received approval in the USA already in 1998 [16] and in Canada in 1999 [17]. Importantly, the distribution of this product was, however, discontinued in both countries some years later due to bleeding and thromboembolic complications [18,19]. Furthermore, after administration of the firstgeneration OctaplasLG product (4 h S/D treatment), seven cases with venous thromboembolism (VTE) have been reported [20]. However, as the performed study was retrospective and other known risk factors for VTE were identified, the authors concluded that the number of events is insufficient to assess whether OctaplasLG was the risk factor for the observed VTE cases.
The purpose of this study was to compare the biochemical profiles and manufacturing processes of the two S/Dtreated plasma products Plas+SD and OctaplasLG, and to explain the thromboembolic events that occurred following administration of the Plas+SD product.

S/D-treated plasma products used for biochemical analysis
Because the Plas+SD product has been discontinued, only three units (blood group O) from different batches were obtained from VITEX. The products were manufactured in 2000/2001 and stored at ≤À18°C according to the manufacturer instructions. The units were tested for 19 different plasma parameters within the approved 1-year shelf life. The results were compared with the results of three batches blood group O plasma, which were produced in 2013/2014 from plasma collected in the USA according to the OctaplasLG process.

Downscaling of the Plas+SD manufacturing process
Three different units of blood group O plasma pools, produced in 2015, were used for the downscaling experiments of the Plas+SD manufacturing process. As a start volume of 500 ml plasma was used, this equals approximately 1:2000. Residual cells and cell fragments were removed by filtration through 1-lm filters. Calcium chloride at a final concentration of 2 mM was added to stabilize coagulation. S/D treatment was performed in the presence of 1% tri(n-butyl)phosphate (TNBP) and 1% Octoxynol for 4 h at 31°C. Residual S/D reagents were removed by oil extraction (5% soya bean oil for 15 min) and chromatography (100 ml column packed with preparative C18 silica, Waters Corporation). After filtration through 1-lm, 0Á45-lm and 0Á2-lm filters, respectively, ultrafiltration was performed using a 30K membrane (Pellicon XL, Merck Millipore).

Measurement of plasma proteins
Although the sensitivity of the available test systems has improved over time, the principle of each test has not changed. All parameters were tested using validated test methods according to the valid European (EP) and/or US Pharmacopeia (USP) protocols in use at the time of testing.
The total protein content was measured by the BIURET method according to EP (Biuret reagent, Roche Diagnostics GmbH). Fibrinogen levels were measured by onestage clotting assay according to CLAUSS (Fibrinogen C, Instrumentation Laboratory).
Coagulation factors V (FV), IX (FIX) and XI (FXI) as well as protein S (PS) and activated partial thromboplastin time (aPTT) were all determined by one-stage clotting assays according to EP (i.e. by using FV-or FIX-deficient plasma from Precision Biologic, aPTT-SP Kit from Instrumentation Laboratory or STA-Protein S clotting kit from Diagnostica Stago). Factor VIII (FVIII), protein C (PC) and plasmin inhibitor (PI; also known as a 2 -antiplasmin) were quantified by chromogenic substrate assays according to EP (i.e. COAMATIC-FVIII or COAMATIC-Protein C from Chromogenix; HemosIL Plasmin Inhibitor from Instrumentation Laboratory).
The pH value (potentiometry; pH-meter 3110, WTW GmbH) and osmolality (freeze point; Osmomat auto, Ekomed GmbH) of the plasma were measured according to EP and USP methods.
Sodium, potassium, calcium and phosphate levels were determined by spectroscopy, while citrate levels were measured by HPLC; all methods according to EP.
Residual S/D reagents TNBP and Octoxynol were determined according to EP methods, by gas chromatography and HPLC, respectively.
Von Willebrand factor (VWF) multimer analyses were performed using 1Á2% agarose gel electrophoresis (Sigma-Aldrich Handels GmbH) with subsequent transfer to nitrocellulose membranes (GE Healthcare). For visualization of the VWF bands, membranes were incubated with polyclonal rabbit anti-human VWF antibody A0226, which was conjugated with horseradish peroxidase (DAKO). The film was developed with Super Signal West Pico Chemiluminescent Substrate (Pierce). Densitometric evaluation was performed using an automated imager reader. Normal Plasma Reference Standard (NP, American Diagnostica GmbH) was used as control for VWF multimer pattern.

Statistical analysis
All results are expressed as mean values AE 1 standard deviation. The Student's paired t-test was used to assess statistically significant differences in the activities/concentrations of the selected plasma proteins between the Plas+SD and OctaplasLG plasma. A P-value of <0Á05 was considered as statistically significant.

Biochemical analysis
The mean activity values, AE 1 standard deviation, of the biochemical analysis of the historical Plas+SD product (VITEX) as well as the recently produced OctaplasLG are shown in Table 1. The percentile mean difference of each parameter in the Plas+SD product compared to the Octa-plasLG product, as well as the p-values, is also included in this table.
From all parameters tested, no significant differences were found between the two S/D plasma products in fibrinogen, FVIII levels, aPTT or pH values.
A significant increase of more than 20% was found in the Plas+SD product in total protein, FV-and FXI levels, whereas a significant decrease of more than 40% was seen for FIX, PC, PS and PI. In addition, the osmolality was significantly lower in the Plas+SD product when compared to OctaplasLG.
Additionally, the mean of calcium and phosphate values were significantly increased, while sodium, potassium SD, standard deviation; n.a., not applicable. Mean levels and standard deviation are shown for three batches Plas+SD and OctaplasLG. Mean difference between Plas+SD and OctaplasLG is indicated in percentage and P-value. and citrate were significantly decreased at varying degrees in the Plas+SD in comparison with the Octa-plasLG.
Regarding the S/D reagents, Octoxynol was below detection limit in both products, whereas a small amount of TNBP was detectable in the Plas+SD, but was below the detection limit in OctaplasLG.
The mean values of all parameters which showed a difference between the two S/D plasma products are shown in Fig. 1.
In addition to the quantitative analysis of the coagulation factors and inhibitors, the VWF multimeric pattern was also analysed in Plas+SD and OctaplasLG and compared with the normal plasma reference standard (NP, Fig. 2). The loss of the high molecular weight VWF multimers was detected in the Plas+SD, whereas the VWF multimeric pattern found in OctaplasLG was comparable with the NP pattern.

Comparison of the two manufacturing processes
As the biochemical profiles of the two S/D plasma products were different, the single steps used for the two manufacturing processes were more closely examined ( Table 2). The main differences in the manufacturing processes (highlighted in bold) are the prolonged S/D treatment time and the shorter extraction time used for the removal of the S/D reagents, as well as the concentration step (ultrafiltration) utilized in Plas+SD when compared to the OctaplasLG production method.

Downscaling of the Plas+SD manufacturing process
The data of the downscale experiments of the Plas+SD manufacturing process are shown in Fig. 3. Compared to the amount detected in the initial plasma pool, a decline of the mean activity levels to various extents was found for all parameters tested following the S/D treatment and removal steps. Accordingly, the aPTT was prolonged.
Fibrinogen (data not shown) was the only parameter that was increased during the concentration step (ultrafiltration) to the same extent as measured in the plasma pool. Compared to the plasma pool levels, an increase was shown in the amount of total protein, FIX and PC, whereas a reduced amount was still detected for FV,  FVIII, FXI, PS and PI (Fig. 3). aPTT was still prolonged when compared to the plasma pool (data not shown).
A possible influence of the S/D treatment and ultrafiltration steps performed during Plas+SD manufacturing on the VWF multimer pattern was investigated by densitometric analysis. The results of this analysis are shown in Fig. 4 and demonstrate that the S/D treatment had no impact on the VWF multimer pattern, whereas the ultrafiltration step contributed to the significant loss of the high molecular weight VWF multimers.

Discussion
The main limitation of this study is the sample size. However, the biochemical data are supported by a previous study performed by Solheim and Hellstern [21], who investigated 12 units of Octaplas (the first-generation product from Octapharma) and 8 units of Plas+SD, which showed lower concentrations of citrate, PS and PI in the historical Plas+SD product. Similar results were found by Salge-Bartels and colleagues [22], who compared coagulation factors and inhibitors in 8 units Octaplas and 2 units Plas+SD. Furthermore, Dr. Bernard Horowitz kindly provided unpublished data showing the mean values of 124 units Plas+SD and found levels of total protein (61Á7 mg/ml), fibrinogen (2Á6 mg/ml), FV (0Á82 IU/ml), FX (1Á08 IU/ml), FXI (0Á93 IU/ml), FXIII (1Á1 IU/ml), as well as single measurements of PC (0Á36 IU/ml), PS (0Á15 IU/ml) and PI (0Á25 IU/ml), which are comparable with the results of our study.
A better understanding of which manufacturing steps might have contributed to the different biochemical profiles of the two S/D-treated plasma products is provided by a stepwise comparison of the two manufacturing processes, as well as the downscaling experiments performed using Plas+SD. Within the Plas+SD manufacturing process, the S/D treatment time was much longer (4 h) and the S/D reagent (TNBP) removal time was much shorter (15 min) when compared to the Octa-plasLG process (1-1Á5 h S/D treatment time and 60-70 min TNBP removal time). Compared to the start material (plasma pool), the prolonged S/D treatment time used for Plas+SD led to a reduction of all protein levels tested except VWF. The lack of impact of the S/D treatment on high molecular weight VWF multimers has already been shown earlier [23]. Furthermore, previous studies already demonstrated that a shorter S/D treatment time used for OctaplasLG almost doubles PI levels and also increases PS levels [23,24], as it has been shown that S/D treatment using TNBP and Triton X-100 (Octoxynol) is responsible for the loss of activity of these plasma inhibitors [25,26].  The concentration step (ultrafiltration) implemented into the Plas+SD manufacturing process, which is not included in the OctaplasLG production process, compensated for the reduced activity of fibrinogen, FV, FVIII and FXI as well as for the aPTT prolongation, which occurred following S/D treatment leading to increased levels of total protein, FIX and PC activities. In addition, after the ultrafiltration step the mean values of PS activities were further reduced and significant losses of the high molecular weight VWF multimers were detected.
The very low PI activity in the Plas+SD product might have contributed to the reported bleeding events, as it has been earlier demonstrated that an increased risk of bleeding is associated with very low amounts of PI resulting in uncontrolled plasmin-mediated breakdown of the fibrin clot [27], and can lead, under certain clinical conditions, to unregulated excessive fibrinolysis [21].
Furthermore, the reduced high VWF multimers in the Plas+SD product might have contributed to a defective haemostasis especially as mainly the high VWF multimers are crucial in supporting platelet adhesion and aggregation [28].
The thrombotic events which occurred following the administration of the historical Plas+SD product might be related to the significantly reduced amount of PS activity in the final product, as it has already been shown that the loss of functional PS is associated with an increased risk for arterial/venous thrombosis [21,22,29]. In addition, the low citrate concentration in the Plas+SD product may have increased the risk for clot formation, especially in combination with high levels of ionized calcium. A previous study has demonstrated that a final citrate concentration lower than 10 mM led to a marked rise in fibrinopeptide A, indicating coagulation activation and fibrin formation [30].
In conclusion, although both manufacturing processes comprise an S/D treatment step for inactivation of enveloped viruses, the biochemical profile of the historical Plas+SD product is significantly different in regard to the levels of coagulation factors and inhibitors when compared to the currently available OctaplasLG. The imbalance of coagulation factors and inhibitors in the historical Plas+SD final product has the potential to be clinically relevant and might have been responsible for the adverse events that occurred with this product.