Pathogen reduction/inactivation of products for the treatment of bleeding disorders: what are the processes and what should we say to patients?

The process of separating whole blood donations into cellular components, with details of particular separation methods, is shown in Fig. 2; the different storage conditions for each blood-derived product are indicated in this figure. Whole blood donations from the UK are not processed in some countries (e.g. Australia, USA) due to the potential risk of prion contamination [26, 27].

Storage methods for whole blood (also known as ‘fresh’ blood) and PC vary. Whole blood can be stored for 24 h at a temperature of 20–24 °C [28‐30], but is no longer utilized for transfusion in resource-rich countries due to the potential risk due to bacterial transmission and post-transfusion hypervolemia. The process of separation of whole blood should therefore be started within 24 h of collection (as detailed in Fig. 2). PC may be stored for 4–5 days at 20–24 °C, buffered and under permanent motion to preserve cell viability [24, 31‐33]. It should be noted that storage of PC at this temperature will increase the risk of bacterial growth, as PC are at risk of bacterial contamination from donor bacteraemia or skin flora.

The initial step of whole blood processing by centrifugation involves the separation of cells based on their density, yielding RBC, white cells (buffy coat) and plasma [34]. Following centrifugation of whole blood, two methods are routinely used to continue the separation process—the buffy coat method and the platelet-rich plasma (PRP) method. The buffy coat method involves the separation of whole blood by high-speed centrifugation into RBC, white cells/platelets and plasma, while the PRP method involves the low-speed centrifugation of whole blood, leaving a larger proportion of platelets suspended in plasma. This platelet-rich plasma is then separated from the other blood components [16, 34].

Following the centrifugation of non-filtered whole blood, RBC and plasma are separated from the buffy coat, resulting in leukoreduction (i.e. total leukocyte number ≤0.1–1.2 × 10⁹ per bag). The leukocyte level of plasma and RBC can then be lowered further by an additional filtration step, after which they are referred to as leukodepleted (i.e. total leukocyte number <1.0 × 10⁶ per unit) [16]. The process of leukodepletion is mandatory in most resource-rich countries; the individual components prepared from leukodepleted whole blood do not require further leukocyte removal procedures [16].

Removal of leukocytes can reduce the risk of pathogen transmission, as leukocytes are potential carriers of pathogens such as cytomegalovirus (CMV), human T-lymphotropic virus type 1 (HTLV-1), proteinaceous infectious agents (prions) and bacterial species such as Listeria and Coxiella. The process of leukocyte removal also reduces the risk of transfusion-associated lung injury (TRALI) and alloimmunization to human leukocyte antigens (HLA) in recipients [5, 16].

Cell-free plasma is usually frozen at −25 °C within 24 h of collection and is then referred to as fresh frozen plasma. FFP may be stored for 36 months, during which time it can be unthawed (usually at 4 °C) for use in transfusion [24, 33]. FFP can also be quarantined for several months, at which point the donor is re-screened to confirm the absence of pathogens; this procedure is preferred in some blood banks (e.g. in Germany) as it overcomes the problem of donation in the diagnostic window period when infection is not yet detectable by antibody/nucleic acid testing [43] and to avoid loss of clotting factor activity by the inactivation process. Plasma-derived coagulation factors are prepared from pooled donations by a multi-step procedure including precipitation and chromatography. The prepared factors might be kept for up to 10 days at 1–6 °C, but are usually lyophilized and can be stably maintained in this state for 6–36 months at 4 °C, or for a shorter duration if stored at 25 °C [31, 32].

Aside from plasma-derived clotting factors, recombinant concentrates have been available for factors VIIa, VIII, IX and XIII since the early 90s [45‐50] and a recombinant von Willebrand factor (VWF) concentrate has recently been approved by the FDA [51]. Recombinant coagulation factors are stored in a lyophilized state and can maintain full activity for up to 30 months at 5 °C (reduced activity maintained up to 30 °C) [44]. Recombinant products are not derived from human blood, but are cloned proteins, maintained in the presence of stabilizers, such as albumin or other substances, that may be added during manufacture [52‐54]. Recombinant products are classed according to the level of exposure to human plasma or albumin during their preparation. The most recently produced recombinant products (‘third-generation’ products) are produced in the complete absence of blood-derived components [45‐47, 55].

Chemical/dynamic treatments rely upon the interaction of a chemical agent with nucleic acids; this interaction can be triggered by UV light (e.g. amotosalen [S59] or riboflavin) or adaptation of pH and glutathione (e.g. amustaline [S303], a nucleic acid alkylator similar to amotosalen [62, 63]). This interaction causes irreversible damage to nucleic acids, thereby reducing the level of viable pathogenic contaminants [64, 65]. Photochemical (PhC) methods employ chemical agents that form a permanent linkage with RNA/DNA when exposed to light, whereas photodynamic (PhD) methods use chemical agents that simultaneously damage DNA and impede DNA repair processes [42, 63, 65, 66]. PhC and PhD technologies vary in their ability to reduce the level of pathogen contamination. They are generally effective against enveloped viruses, bacteria and protozoans, but are less reliable for the inactivation of non-enveloped virus; bacterial spores and prions are not affected by these methods. The formation of relevant neoantigens in plasma proteins has not been observed after PhC/PhD treatment [67‐72].

These methods are used exclusively for plasma-derived purified products; after filtration with a 220-nm or 15- to 50-nm filter (filtration/nanofiltration, respectively), these products are considered to be free from bacteria and protozoans such as Plasmodium, Trypanosoma or Leishmania. However, contamination with small viruses and prions remains a primary safety concern [19].

There is currently no widely available reduction/inactivation method that is capable of completely removing prions from blood-derived products while maintaining functionality, but potential alternative solutions are already being implemented. Leukoreduction of red blood cells for transfusion was introduced in the UK in 1999 to reduce the risk of variant Creutzfeldt-Jakob Disease (vCJD) transmission by blood transfusion. Animal studies have shown that leukoreduction provides a high, but not complete, reduction of the risk of prion transmission by blood [98]. It is also noteworthy that the four cases of transfusion-associated vCJD transmission in the UK all occurred in recipients of non-leukoreduced RBC concentrates. More recent efforts to increase the efficiency of prion removal in blood have focused on the development of prion removal filters. However, most of these filters have been evaluated on blood spiked with hamster brain infected with a scrapie-derived prion strain (263K) [98‐100]. It is highly unlikely that the physicochemical properties of these brain-derived prion spikes mimic those of prions in the blood of individuals infected with vCJD. A recent evaluation of a prion filter on blood from primates endogenously infected with vCJD found some evidence of efficacy, but vCJD transmission occurred in one primate following transfusion of ‘prion filtered’ leukoreduced RBC [99]. Further improvement and evaluation of prion filters is required, but as yet there is no initiative to introduce the use of currently available prion filters in the UK or elsewhere. The incidence of vCJD cases in the UK is declining [101], and thus the risk of prion transmission by transfusion is very low. Any residual risk might be further reduced by incorporating recent advances in vCJD prion detection in a test for blood donor screening [102, 103].

The overall risk of pathogen transmission is reduced when pathogen reduction/inactivation treatments are applied to blood- and plasma-derived products. However, other risks are also associated with these treated products: key concerns include neoantigen formation, loss of product functionality/activity and the potential link between product use and adverse events in some patients [19, 64].

Neoantigens may be formed in treated blood-derived products due to changes in the three-dimensional structure of proteins, or from cellular debris. When denatured blood-derived cellular products and plasma-derived products are transfused into patients, there is a risk that the patient may develop antibodies to the neoantigens present and ultimately will be unable to maintain the course of treatment. Antibodies to therapeutic proteins are referred to as inhibitors. As a result of forming inhibitors towards a modified structure of a therapeutic protein (i.e. one which has undergone pathogen inactivation processing), there is a risk that patients may also mount an immune response against unmodified versions of the same protein, thus preventing effective replacement therapy [104], as observed, for example, in FVIII-treated patients with haemophilia A [2].

The risk of inhibitor formation with recombinant clotting factor concentrates was recently assessed in a study in 251 previously untreated patients, when patients were treated with either plasma-derived FVIII concentrate containing VWF, or recombinant FVIII to compare the risk of inhibitor formation between the two different types of FVIII products [105]. Patients receiving recombinant FVIII displayed significantly higher levels of inhibitor development compared with patients who received plasma-derived FVIII/VWF (cumulative inhibitor incidence of 44.5% vs. 26.8%, respectively; high-titre inhibitor incidence in 28.4% and 18.6%, respectively). However, the difference appeared to be maximally related to transient inhibitors: when the latter were removed from the analysis, the difference in inhibitor formation was no longer statistically significant. In addition, the results of this study differ from previously published studies which showed lower/minimal increase in inhibitor formation with recombinant FVIII [106, 107]. It is thought that differences in trial design may have played a role in this discrepancy between current and previous results.

Another limitation of pathogen inactivation processes is the reduction in the activity of products. For example, SD treatment can cause a reduction in FVIII activity in plasma-derived products; FVIII treated with SD (Octaplas®, Octapharma) has a FVIII activity level ~10–20% lower than that observed in FFP. Process validation studies conducted in three European blood centres showed that PhC-treated plasma displayed a ~26% reduction in FVIII activity compared with that of FFP [94]. The activity level of other coagulation factors was also reduced, but to a lesser extent with clotting factor activity generally reduced by less than 20% [108]. Compared with FFP, SD-treated plasma displayed reduced activity levels for VWF and Protein S (67–76% and 35% reduction, respectively), while MB-L-treated plasma displayed lower fibrinogen activity levels (approximately 20% reduction) [94, 104, 109].

There have been various reports linking pathogen inactivated products with adverse events. As mentioned, use of MB-L-treated plasma has been associated with post-transfusion anaphylactic reaction, and a study of plasma use in one hospital over several years showed that patients who received MB-L-treated plasma required 56% more plasma in total than patients who received non-inactivated plasma; the increased need for plasma in this instance was described as an adverse event, but could also potentially indicate reduced therapeutic efficacy of the MB-L-treated plasma [104]. MB-L-treated plasma, but not SD-treated plasma, may also contain residual RBC or cell fragments, which increase the risk of red cell alloimmunization; it is still unclear whether the presence of these residual elements is influenced by the nature of plasma processing prior to inactivation [104, 110‐112]. Concerns have previously been raised over reports of a potential link between SD-treated plasma and increased risk of venous thromboembolism, potentially caused by low levels of Protein S, leading to an increased risk of clot formation [94, 113]. However, a recent study, and the resumption of SD-treated plasma use in the USA, indicates that the association of a potential link is still uncertain [114]. There has been no reported association between SD treatment and TRALI, which can be a major cause of transfusion-related mortality using plasma components containing leukocyte-directed alloantibodies (5–25% of cases are fatal) [20]. This is thought to be due to low levels of human leukocyte antigen antibodies, the key mediators of TRALI, in SD-treated plasma caused by the pooling and dilution process [104].

There have been several studies investigating the cost-effectiveness of pathogen reduction/inactivation. In one report (based on the UK system), the costs per life year saved for patients aged ≤60 years using three pathogen inactivation systems (involving amotosalen/UVA, riboflavin/UVB or UVC treatment) were estimated to range from £3.4–9.1 million [122]. A retrospective Spanish study showed that universal implementation of pathogen reduction (amotosalen/UVA system) for platelet preparation from 2008 onwards extended platelet storage time and reduced the cost associated with out-of-date units, resulting in a saving of 13.8% of the budget compared with the initial combined cost of platelet production/storage/wastage for one institution [123]. In a Belgian study, implementation of an amotosalen-based pathogen reduction system for platelets was found to be cost-effective compared with the initial national system (measured in terms of lifetime costs and quality adjusted life years (QALYs)). A range of incremental cost-effectiveness ratios were observed (in absence of emerging pathogens) from 3,459,201 €/QALY to 195,364 €/QALY), and the mean threshold of emerging infection risk for the pathogen reduction system ranged from 1/1079 to 1/2858 transfusion [124]. In addition, a model that was developed from a Dutch study found that implementation of pathogen inactivation as a routine procedure would be a cost-effective measure (measuring cost-effectiveness in net costs per life year gained (LYG)); net costs per LYG with pathogen inactivation were estimated at €554,000 in the baseline-weighted average over three patient groups (90% simulation interval €354,000–1092,500). A sensitivity analysis indicated that cost-effectiveness in this model was insensitive to viral risks and indirect costing, but highly sensitive to the assumption of a requirement for excess transfusions required and also discounting of life years gained. The study authors concluded that, in view of the relatively high (and internationally accepted) net costs per LYG for blood transfusion safety interventions, the data generated by this model indicate that pathogen inactivation may be cost-effective. However, the study made the assumption that the pathogen inactivation method would be 100% effective, rendering this result uncertain [125]. In Germany (for example), approximately 20 million € (50 € (assumed price) for amotosalen/UVA treatment for 400,000 PC) would have to be paid in order to prevent 6 instances of TTBI [9]. In the USA, the cost-effectiveness of pathogen inactivation was assessed at five institutions, comparing pathogen inactivation costs with the costs of tests/procedures currently performed; total potential cost savings with the implementation of pathogen inactivation were estimated at $141.65 per apheresis platelet unit [126]. In the UK, approximately £1.0–3.2 million would have to be spent in order to save 0.9–1.3 LYG per unit when using amotosalen/UVA inactivation to reduce bacterial contamination of platelets [122].

Another potential cost-saving approach would be to prepare blood components from inactivated whole blood, treated by an agent such as riboflavin/UVB; studies have shown that whole blood treated with this method is safe for transfusion use, as are its components [59, 65, 127]. This approach may have merit in countries which rely on whole blood as their main transfusion source. However, it should be noted that the process of pathogen inactivation of whole blood by riboflavin/UVB has not yet been optimized, and that some non-enveloped viruses (e.g. hepatitis A virus) may be poorly inactivated. Therefore, the overall cost-effectiveness of this approach (when the potential cost of treating patients for TTI is taken into account) has yet to be determined.