Professional Documents
Culture Documents
Vox Sanguin Mei 2021
Vox Sanguin Mei 2021
In this issue
The immune potential of ex vivo stored platelets: a review
Contents
Review Article 564 An economic reappraisal of hepatitis B virus testing strategy for
blood donors in Taiwan T. B. Matanhire & S.-W. Lin
477 The immune potential of ex vivo stored platelets: a review
B. Wood, M. P. Padula, D. C. Marks & L. Johnson Transfusion Medicine
574 Blood transfusion activity in a general hospital during the
Commentary COVID-19 pandemic K. Marín-Mori, I. González-Gascón y Marín,
M.-Á. Foncillas-García, C. Muñoz-Novas, M. Infante,
489 The interaction between Glycophorin A (GPA) and Band 3 in the
J. Churruca-Sarasqueta, E. Landete-Hernández, B. Bueno-García,
formation of the Wright b (Wrb) antigen S. Ekman, R. T. Barnard,
M. Duffort-Falco & J.-Á. Hernández-Rivas
R. Flower, A. Gould & X. T. Bui
581 Costs associated with transfusion therapy in patients with
493 Evaluation of SARS-CoV-2 antibody titers and potency for
myelodysplastic syndromes in Sweden: a nationwide
convalescent plasma donation: a brief commentary E. Wouters,
retrospective cohort study J. Zhao, T. Dahlén & G. Edgren
M. Steenhuis, H. Schrezenmeier, P. Tiberghien, H. Harvala,
H. B. Feys & E. van der Schoot Immunohematology
591 Clinical consequences of the extremely rare anti-PP1Pk
Original Papers isoantibodies in pregnancy: a case series and review of the
Donors and Donations literature P. Di Ciaccio, B. Cutts, T. I. Alahakoon, P. M. Dennington,
L. A. Soo & J. Curnow
497 Blood supply management in times of SARS-CoV-2 pandemic
– challenges, strategies adopted, and the lessons learned from the 601 Red cell genotyping of rare blood donors: donation behaviour and
experience of a hospital-based blood centre H. C. Pandey, data visualization W. Q. Anani, J. Gottschall & G. A. Denomme
P. Coshic, C. C S, P. J. Arcot & K. Kumar
504 Compliance and attitudes of blood donors following transitioning International Forum
from permanent to 12-month deferral of men who have sex with 609 International Forum on Transfusion Practices in Haematopoietic
men in Hong Kong J.-C. Lau, C.-K. Lee, C.-P. Chan, J.-S. Leung, Stem-Cell Transplantation: Summary P. Solves, M. Lozano,
C.-M. Poon & S.-S. Lee E. Zhiburt, J. Anguita Velasco, A. Maria Pérez-Corral,
S. Monsalvo-Saornil, S. Yamazaki, H. Okazaki, K. Selleng,
513 Unknown, so also unvalued? Blood donation awareness and
K. Aurich, W. Krüger, A. Buser, A. Holbro, L. Infanti, G. Stehle,
attitudes of potential donors of Dutch and African descent
L. Pierelli, A. Matteocci, L. Rigacci, K. M. K. De Vooght,
E. F. Klinkenberg, M. P. Fransen, W. L. de Kort, E. M. Huis in ’t Veld
J. H. E. Kuball, K. L. Fielding, D. A. Westerman, E. M. Wood,
& J. C. van Weert
C. S. Cohn, A. Johnson, M. B. C. Koh, D. Qadir,
Blood component collection and Production C. Cserti-Gazdewich, E. Daguindau, F. Angelot-Delettre,
524 Ovine red cell concentrates for transfusion research – is the P. Tiberghien, S. Wendel-Neto, R.-M. Fachini, S. Morton,
storage lesion comparable to human red cell concentrates? C. Craddock, M. Lumley, J. Antoniewicz-Papis, K. Hałaburda,
G. Simonova, R. Wellburn, Y. L. Fung, J. F. Fraser & J.-P. Tung M. Łe˛towska & N. Dunbar
533 The ’rejuvenating factor’ TIMP-2 is detectable in human blood e25 International Forum on Transfusion Practices in Haematopoietic
components for transfusion J. Hoefer, C. Dal-Pont, S. Jochberger, Stem-Cell Transplantation: Responses P. Solves, M. Lozano,
R. Fantin & H. Schennach E. Zhiburt, J. Anguita Velasco, A. Maria Pérez-Corral,
S. Monsalvo-Saornil, S. Yamazaki, H. Okazaki, K. Selleng,
540 A microfluidic analysis of thrombus formation in reconstituted
K. Aurich, W. Krüger, A. Buser, A. Holbro, L. Infanti, G. Stehle,
whole blood samples comparing spray-dried plasma versus fresh
L. Pierelli, A. Matteocci, L. Rigacci, K. M. K. De Vooght,
frozen plasma R. S. Bercovitz, C. S. Drew, C. L. Bushee,
J. H. E. Kuball, K. L. Fielding, D. A. Westerman, E. M. Wood,
M. A. Popovsky, K. D. Friedman & W. Q. Anani
C. S. Cohn, A. Johnson, M. B. C. Koh, D. Qadir,
547 Quality assessment of red blood cell suspensions derived from C. Cserti-Gazdewich, E. Daguindau, F. Angelot-Delettre,
pathogen-reduced whole blood I. Kumukova, P. Trakhtman, P. Tiberghien, S. Wendel-Neto, R.-M. Fachini, S. Morton,
N. Starostin, D. Borsakova, A. Ignatova & Y. Bayzyanova C. Craddock, M. Lumley, J. Antoniewicz-Papis, K. Hałaburda,
M. Łe˛towska & N. Dunbar
Transfusion-transmitted Disease and it Prevention
557 Anti-A and SARS-CoV-2: an intriguing association 613 Diary of Events
V. de Freitas Dutra, C. Bonet-Bub, A. P. H. Yokoyama, R. Achkar,
R. R. G. Machado, M. Assunção, G. Candelária, C. P. Soares,
R. M. Fachini, R. Fontão-Wendel, N. Hamerschlak, L. F. L. Reis,
D. B. Araujo, V. Nudelman, J. R. R. Pinho, L. V. Rizzo, A. M. Sakashita,
P. Scuracchio, E. L. Durigon, S. Wendel & J. M. Kutner t
ea
n lin
a lo vo
x
ou
rn
n al/
i sj jo ur
th m/
This journal is available online at Wiley Online Library. w
co
Vie y.
Visit wileyonlinelibrary.com to search the articles and r ar
register for table of contents e-mail alerts. lib
li ne
on
ley
wi
Vox Sanguinis (2021) 116, 477–488
© 2020 International Society of Blood Transfusion
REVIEW ARTICLE DOI: 10.1111/vox.13058
Abstract
Platelets are now acknowledged as key regulators of the immune system, as they
are capable of mediating inflammation, leucocyte recruitment and activation.
This activity is facilitated through platelet activation, which induces significant
changes in the surface receptor profile and triggers the release of a range of sol-
uble biological response modifiers (BRMs). In the field of transfusion medicine,
the immune function of platelets has gained considerable attention as this may
be linked to the development of adverse transfusion reactions. Further, compo-
nent manufacturing and storage methodologies may impact the immunoregula-
tory role of platelets, and an understanding of this impact is crucial and should
be considered alongside their haemostatic characteristics. This review highlights
the key interactions between platelets and traditional immune modulators. Fur-
ther, the potential impact of current and novel component storage methodolo-
Received: 18 September 2020,
gies, such as refrigeration and cryopreservation, on this functional capacity is
revised 14 November 2020,
accepted 2 December 2020,
examined, highlighting why further knowledge in this area would be of benefit.
published online 16 December 2020 Key words: platelet, storage, leucocyte, function, refrigeration, cryopreservation.
477
478 B. Wood et al.
Fig. 1 A simplified schematic depicting platelet–leucocyte interactions, which are mediated by the expression of surface receptors and release of soluble
factors. Platelet–leucocyte interaction can be triggered by toll-like receptor (TLR) recognition of 1. pathogen-associated molecular patterns (PAMPs e.g.
bacterial lipids) or 2. damage-associated molecular patterns (DAMPs e.g. HMGB1). This triggers platelet activation and release of 3. a-granules (a), 4.
dense granules (d) and 5. extracellular vesicles (EVs; brown lines). Degranulation of a-granules increases the surface expression of activation markers,
including CD40L (blue rectangle) and P-selectin (green rectangle) and 6. the release of a range of biological response modifiers (BRMs; e.g. HMGB1, IL-
27, RANTES, PF4; black lines) into the plasma which can influence surrounding leucocytes (purple receptors) and have been linked to inflammation and
adverse events. 7. Activated platelets can attach to neutrophils through P-selectin, CD40L and GPIba to PSGL-1, CD40 and Mac-1, respectively. Attach-
ment can also occur through platelet GPIIb/IIIa and leucocyte Mac-1 through a fibrinogen intermediate. This interaction generally increases the activa-
tion of both platelets and leucocytes. 8. The activated platelet can then release internal agonist stores (ADP, calcium) which lead the activation of
platelets in the surrounding area. 9. EVs can activate leucocytes either through transport of their internal contents (e.g. mitochondria, mtDNA, micro-
RNAs, proliferation factors and cytokines) or by receptor signalling. The activation of neutrophils can trigger 10. the generation neutrophil extracellular
traps (NETs) which are associated with the pathogenesis of TRALI. [Colour figure can be viewed at wileyonlinelibrary.com]
recognition of a range of signalling factors, including to external morphological changes [2,14]. Cytoskeletal
pathogen-associated molecular patterns (PAMPs), damage- rearrangement typically results in the release of alpha-
associated molecular patterns (DAMPs) and traditional granules (a-granules) and dense granules, dispersing a
platelet agonists (Fig. 1) [2,14,15]. The signals are then range of soluble factors into the circulation (Fig. 1) [16].
transduced across the cell membrane and propagated in Extracellular vesicles (EVs) are also released, which act as
the cytosol by downstream signalling, resulting in signifi- mobile signalling mediators (Fig. 1) [17–19]. In vivo these
cant changes to internal cytoskeletal structure and leading changes increase localized inflammation and encourage
the recruitment and activation of leucocytes, escalating platelets, a significant increase is observed following acti-
the immune response [20]. vation [19,24]. Platelet EVs express a range of surface
receptors including P-selectin and CD40L, which facilitate
attachment and activation of monocytes through Mac-1
Immunoregulatory surface receptors
and CD40, respectively [35]. Pro-coagulant platelets also
It has been established that platelets express a variety of release EVs rich in surface exposed phosphatidylserine,
surface molecules that enable them to moderate early which attach to monocytes and neutrophils in vivo,
immune responses to both infection and sterile/self-injury. although the specific mechanics of this interaction and its
Platelets express most of the known Toll-like receptors immunomodulatory effects are still under evaluation
(TLRs) on their surface membrane (TLR1, 2, 4, 5, 6), or [24,34]. Platelet EVs can also transport and deliver a range
internally within endosomes or platelet specific T-granules of platelet-derived immunomodulatory compounds to leu-
(TLR3, 7, 8, 9) [2,21,22]. In general, platelet TLRs recog- cocytes including receptors (CD40L), mitochondria and
nize and bind highly conserved PAMPs, including lipids, mitochondrial DNA (mtDNA), microRNAs, cytokines (IL-
lipoproteins, proteins or nucleic acids derived from bacte- 1b, MIP-3), proliferation factors (PF4, CXCL-7) and
ria, viruses, fungi or parasites, to initiate anti-microbial chemokines (RANTES; Fig. 1) [14,18,31,34,36]. As a result,
signalling [1,2,15,21,22]. TLRs can also bind a range of platelet EVs exhibit wide-ranging effects and can influ-
‘self’-derived DAMP molecules that can be secreted or ence the activation of most leucocyte populations,
released from damaged / apoptotic cells or accumulate in although interactions with neutrophils and monocytes are
platelet components during storage, which induces pro- more common [14,18,24,34,36]. EV-activated monocytes
inflammatory signals (Fig. 1; Tables 1 & 2) [2,8,18]. Plate- exhibit increased expression of several adhesion receptors
let TLR-mediated responses are facilitated through their (ICAM-1, LFA-1 and Mac-1), increasing their capacity to
interactions with leucocytes [8,15,23]. The binding of bind to endothelial cells [35,37,38]. Additionally, EV-
PAMPs and DAMPs to TLRs initiates distinctive signal monocyte interaction causes the release of a range of
transduction responses, which induce platelet activation, monocyte derived pro-inflammatory cytokines (IL-8, IL-
the release of biological response modifiers (BRMs) and 1b, TNF-a, MCP-1 and MMP-9) [35,37,38].
immunomodulatory EVs leading to leucocyte activation
and recruitment [8,15,24].
Platelet–leucocyte interactions
The initial response of platelets to tissue damage is the
Release of biological response modifiers
initiation of haemostasis via adhesion to sub-endothelial
Platelets possess a large internal store of soluble factors structures and subsequent activation of the coagulation
within a- and dense granules [16,25], which are released cascade; however, they also play a key role in recruitment
following platelet activation (Fig. 1). The platelet releasate and activation of leucocytes in order to contain and elim-
contains a range of BRMs, including adhesion molecules, inate any potential infiltrating pathogens [3]. Attachment
cytokines, chemokines and growth factors [16,25,26] to leucocytes requires platelet activation and degranula-
(Tables 1 & 2), which are able to modulate many aspects tion, which can be induced by stimulatory signals,
of the innate immune response following transfusion including haemostatic agonists (thrombin, collagen and
[7,27,28]. Specifically, granule release promotes leucocyte calcium), PAMPs/DAMPs or ex vivo storage [4,8,15,39,40].
activation (sCD40L, RANTES, PF4, NAP2), proliferation Specifically, platelet activation results in a-granule
(PF4, NAP2), adhesion (RANTES, IL-1b) and inflammation release, and the consequent localization of P-selectin,
(IL-1b, RANTES) [23,29,30], as well as amplifying the CD40L and phosphatidylserine on the platelet surface,
activation of the original platelet and any surrounding which facilitates leucocyte attachment (Figure 1 &
platelets (ADP, calcium; Fig. 1). Table 1) [3,41]. P-selectin is often the first point of con-
tact between platelets and leucocytes through binding to
its ligand, PSGL-1 [3,41]. While P-selectin/PSGL-1 bind-
Platelet extracellular vesicles as immune
ing alone is relatively weak, co-stimulation with CD40L
regulators
induces downstream signalling in leucocytes [3,30,41].
Platelets produce up to 90% of the EVs present in the cir- This promotes Mac-1 expression on myeloid cells,
culation of healthy subjects [31,32]. Initially, interest in enabling more permanent attachment directly to platelet
platelet-derived EVs was due to their haemostatic proper- GPIba (Fig. 1) [3,30] or through fibrinogen bound to pla-
ties [33]. However, it is now apparent that platelet EVs are telet GPIIb/IIIa [42]. Activated platelets also exhibit exter-
also key mediators of immune signalling [14,24,31,32,34]. nalized phosphatidylserine on the surface membrane,
While low numbers of EVs are released from resting which has been shown to facilitate platelet attachment to
Receptors Role in platelet-mediated immunity References Conventional Storagea Refrigerationb Cryopreservationc References
TLR 1, 2, 4, 5 & 6 • recognizes PAMPs (e.g. proteins and lipids) and [1,2,21,22] na na na na
DAMPs (e.g. HMBG1 and Histone H4)
• triggers platelet activation
• can lead to pro-inflammatory effects
• expressed on the surface membrane
TLR 3, 7 & 9 • recognizes nucleic acids and HMGB1 (TLR9) [1,2,21,22] na na na na
• triggers platelet activation
• expressed internally and on the surface membrane
of activated platelets
Receptors/factors • Role in leucocyte interaction
GPIIb/IIIa (Active • facilitates indirect attachment to leucocytes through [1,42] ↔ or ↓ ↑ ↔ or ↓ [40,48,59–61,70,74,80]
conformation) fibrinogen bound to Mac-1
GPIba • facilitates attachment to leucocytes through Mac-1 [1] ↓ ↓ ↓ [39,58–60,68,70,74,76,77]
Changes over storage are described as general trends and vary depending on the method of platelet manufacture, including collection method and plasma content.
Relative changes in factor expression/concentration are expressed as the following: increase = ↑; decrease = ↓; unchanged = ↔; na = no published data available.
DAMPs, damage-associated molecular patterns; PAMPs, pathogen-associated molecular patterns; TLR, toll-like receptor.
a
Conventionally stored platelets (20–24°C, with agitation) at day 5–7 compared to day 1.
b
Refrigerated platelets compared to conventionally stored platelets at same time point of storage.
c
Thawed platelets compared to conventionally stored platelets at day 1.
Biological Response
b c
Modifier Role in platelet-mediated immunity References Conventional Storagea Refrigeration Cryopreservation References
Table 2 (Continued)
Biological Response
b c
Modifier Role in platelet-mediated immunity References Conventional Storagea Refrigeration Cryopreservation References
• co-stimulatory molecule
EVs [1,3,17,19,31,47,72] ↑ ↔ or ↑ ↑ [26,52,59,68,72,73,80]
• remote signalling factors
• facilitate intracellular transfer of receptors,
cytokines and miRNAs
• activate leucocytes
Mitochondrial DNA [53] ↑ na na [53]
• pro-inflammatory signal
• recognized by TLRs
Relative changes in factor expression/concentration are expressed as the following: increase = ↑; decrease = ↓; unchanged = ↔; na = no published data available.
Changes over storage are described as general trends and vary depending on the method of platelet manufacture, including collection method and plasma content.
EV, extracellular vesicles; TLR, toll-like receptors.
a
Conventionally stored platelets (20–24°C, with agitation) at day 5–7 compared to day 1.
b
Refrigerated platelets compared to conventionally stored platelets at same time point of storage.
c
Thawed platelets compared to conventionally stored platelets at day 1.
macrophages and neutrophils through a range of poten- the initial composition and concentration of BRMs in the
tial surface receptors [39,43]. Notably, the formation of component [49].
platelet–leucocyte aggregates results in the activation and Platelet concentrates are currently stored for 5–7 days
degranulation of both cell types, increasing the local at room temperature (20–24°C), which leads to deleterious
haemostatic and inflammatory effects [3]. storage-related effects, known as the platelet storage
While platelet–leucocyte aggregates can be beneficial lesion (PSL) [40,47,50]. Progression of the PSL impacts
to the immune response, an overabundance is associated the platelet phenotype and function, the composition of
with pro-inflammatory disorders and adverse transfusion the supernatant of stored components (Tables 1 and 2;
events such as TRALI [2,3,5,44]. Activated platelets are conventional storage), and also affects the immunogenic-
able to attach to and circulate with all major leucocyte ity of the component, with older platelet units more likely
subgroups including neutrophils, monocytes, T cells and to induce adverse transfusion events [6].
B cells [44]. However, association with monocytes and Room temperature storage induces a gradual shift
neutrophils is much more common [30,44]. Notably, towards a more activated phenotype. This is observed by
polarized neutrophils, which display asymmetric receptor a loss of certain glycoproteins (GPIba, GPIV, GPVI),
expression, can actively search for activated platelets in increased surface expression of P-selectin, phos-
the bloodstream and attach through P-selectin/PSGL-1 phatidylserine and CD40L (Table 1) [40,47,51]. These are
[45]. The formation of platelet-neutrophil aggregates is all key mediators of platelet–leucocyte signalling (Fig. 1)
pro-inflammatory and a key step towards the generation [3,30]. There is also an accumulation of metabolic by-
of platelet-induced neutrophil extracellular traps (NETs) products, pro-inflammatory cytokines, mitochondrial
[4,5,45,46]. However, platelet–neutrophil binding alone is DNA (mtDNA) and EVs in the supernatant (Table 2)
usually insufficient to induce NET formation, which [7,49,51–56]. The initial concentration of many soluble
requires additional secondary signals such as pathogen factors is reduced by supplementation with PAS. How-
recognition via platelet TLR4, signalling from traditional ever, current data indicate that ex vivo storage can stimu-
platelet agonists (thrombin), soluble factors (sCD40L, late the release of BRMs from platelets several fold higher
IL-1b, HMGB1) or EVs [4,5,8,29,32]. Following receipt of than baseline, regardless of the starting concentration in
secondary signals, the neutrophil ejects its nucleus, plasma or plasma / PAS (Table 2) [7,26–28,36,54].
releasing a web of DNA coated in a range of anti- Conventional storage induces changes in platelet
microbial factors (Fig. 1) [4]. The induction of NET for- immune function over time, which is directly linked to
mation is tightly controlled due to its inflammatory and the incidence of adverse events. A retrospective study by
pro-coagulant properties, which can cause significant Losos et al. examined more than 50 000 leucoreduced
collateral tissue damage, particularly to the sensitive platelet transfusions and identified that the majority of
alveoli of the lung [4,5]. transfusion reactions occurred after 3 days of storage,
As platelet activation is key in mediating platelet with the likelihood of adverse events increasing per day
immune function, it is important to note that a progres- (OR, 130), 95% CI (112, 152) [6]. Consequently, the
sive degree of platelet activation occurs during ex vivo benefit of extending conventional platelet storage time
storage [40,47,48]. While the haemostatic impact of stor- must be weighed against the accumulation of BRMs in
age has been well characterized, the storage-related components and thus the likelihood of inducing adverse
effects on the immunological functions of platelets are events. The induction of adverse events has been associ-
still being unravelled. ated with platelet activation, high concentrations of
HMGB1, soluble CD40L, OX40L, IL-27, IL-1b, mtDNA and
mitochondria-carrying EVs in platelet components (Fig. 1
The effect of ex vivo storage on platelet
& Table 2) [7,18,27,29,53]. Further, DAMPs, such as
immune function
HMGB1 and mtDNA, may accumulate in platelet compo-
nents and have been implicated in the induction of
Conventional room temperature stored platelets
adverse events, including febrile non-haemolytic transfu-
While specific manufacturing methods vary between sion reactions (FNHTRs) and transfusion-related acute
countries, platelet components are produced either by lung injury (TRALI) [2,5,8,18,27,53]. Although the mecha-
pooling and isolating platelets from the WB of multiple nisms of TRALI are multi-factorial, transfusion of plate-
donors or by selectively isolating platelets from a single lets is thought to trigger leucocyte activation and
donor by apheresis [10]. Additionally, the solution in increase lung inflammation and NET formation, poten-
which the platelets are suspended can vary, consisting of tially via the action of activated platelets, TLR signalling,
100% donor plasma or a combination of plasma and up EVs and BRMs (Fig. 1) [5,8,18,29,45]. Further, platelet
to 70% platelet additive solution (PAS), which influences depletion and inhibition by aspirin have been found to
have a protective effect against acute lung injury in mice cryoprotectant, before concentrating the platelets and
[45,46]. removing excess DMSO prior to freezing at -80°C
[63,67,68]. This method allows for long-term storage for
potentially 2–4 years [69]. When required, cryopreserved
Refrigerated platelets
platelets can be thawed and reconstituted in saline,
There has been a resurgence of interest in platelet refrig- thawed plasma or additive solution [67,70,71].
eration (2–6°C) as a storage methodology since its discon- The cryopreservation process alters the surface phe-
tinuation in the 1970s, as refrigeration inhibits bacterial notype and increases degranulation and release of
growth and preserves haemostatic function [12,50,57]. phosphatidylserine-positive EVs (Table 2) [26,72]. In
Platelet refrigeration results in significant morphological terms of membrane changes, 50–70% of cryopre-
change and more pronounced expression of activation served platelets externalize phosphatidylserine (Table 1)
markers in comparison to room temperature stored compo- [39,73,74]. There is some evidence that phosphatidylserine
nents [58–61]. However, alterations to the immune charac- on cryopreserved platelets may be pro-inflammatory,
teristics of platelets induced by refrigeration are largely mediating interactions with macrophages [39,75]. Simi-
uncharacterized (Tables 1 & 2). Changes in the abundance larly, GPIba expression is reduced on up to 50% of cry-
of adhesion receptors have been observed during refrigera- opreserved platelets (Table 1) [39,68,74,76], which may
tion, including decreased GPIba, increased CD40L and P- affect their ability to bind to leucocytes. The concentra-
selectin and phosphatidylserine externalization [59–61] tion of soluble factors in the supernatant of cryopreserved
(Table 1), which could increase the likelihood of platelet components is significantly higher than conventionally
leucocyte–aggregate formation. Refrigeration appears to stored platelet components (Table 2), including RANTES
slow the rate of release of a-granule-associated soluble fac- and TGF-b1, which have pro- and anti-inflammatory
tors, compared to conventionally stored platelets (Table 2) effects, respectively [26,76,77]. Further, the EV concentra-
[26,54]. A study by Johnson et al. demonstrated that refrig- tion of cryopreserved platelet components is up to 15-
erated platelets stored for 14 days contained cytokine fold higher than conventionally stored platelets, and
levels comparable to day 5 room temperature stored plate- 10-fold higher than refrigerated platelets at the end of
lets [26]. The release of phosphatidylserine-positive EVs is storage, which contribute to their haemostatic function
also significantly increased during refrigerated storage [26,73]. Additionally, EVs from cryopreserved platelets
compared to conventional storage [52], and while they have been shown to possess an altered surface receptor
have been demonstrated to be haemostatic [26], the profile compared to EVs from fresh platelet components,
immunologic effects remain unknown. demonstrating increased expression of several adhesion
Research examining the immune effect of refrigerated receptors (GPIIb, GPIX and integrin associated protein;
platelets in vivo is limited to a single study in a mouse IAP) and exposure of phosphatidylserine [72]. Thus, cry-
model, whereby refrigerated platelets increased vascular opreserved platelets have the potential to be more
leakage compared to conventionally stored components immunogenic than conventional platelet units.
[62], which could facilitate increased leucocyte infiltration To date, relatively few studies have been conducted to
through endothelial cells. Preliminary safety data suggest examine the immunogenicity of cryopreserved platelets
that cold-stored platelets are not associated with more (Tables 1 & 2). In an in vitro human whole blood model
adverse events than conventionally stored platelets, at of transfusion, cryopreserved platelets suppressed the
least in patients undergoing cardiac surgery [13]. How- responsiveness of BDCA3+ myeloid-derived dendritic cells
ever, further studies are warranted, particularly in patient to stimulation by LPS and poly(I:C), producing less IL-8,
cohorts where altered platelet-mediated immunological TNF-a and IP-10, suggesting a potentially immunosup-
function may already be a risk factor for adverse events, pressive effect [78]. In contrast, studies using THP-1 cells
such as in trauma [24]. and a murine model of controlled haemorrhage suggest
that cryopreserved platelets may be pro-inflammatory, by
increasing leucocyte release of TNF-a, IL-1b and IL-6
Cryopreserved platelets
[39,79]. While these contradictory findings are likely due
Platelet cryopreservation, pioneered by Valeri et al. [63], to differences in experimental models, they highlight the
was used throughout the 1970s to supply autologous pla- need for further work in this area. Very few adverse
telets to alloimmunized leukaemia patients [64]. In recent events have been reported following transfusion of cryop-
years, an emphasis has been placed on using cryopre- reserved platelets [11,64,65,71]. However, larger studies
served platelets for the treatment of acute bleeding due to are required, as the number of cryopreserved platelets
their increased haemostatic effectiveness [11,65,66]. Plate- transfused is too low to provide confidence that they are
let cryopreservation involves adding 5–6% DMSO as a not associated with adverse events of low incidence.
Future areas of study immune cell. Ex vivo storage conditions have the capacity
to significantly alter the characteristics of platelet compo-
While the immune function of platelets is now acknowl-
nents, which may affect their immune function and the
edged, and studies have begun to consider the effect of
likelihood of adverse events. Consequently, the impact of
conventional and alternative storage methodologies, there
new manufacturing and storage methodologies on the
are significant gaps in our current understanding. Specifi-
immune characteristics of platelets should be considered
cally, the concentration of key BRMs that have been associ-
alongside their haemostatic function prior to implementa-
ated with adverse events (IL-8, IL-27, sCD40L and mtDNA)
tion.
and the expression of immune-related receptors in platelets
stored under novel conditions are largely uncharacterized,
particularly those involved in PAMP / DAMP recognition Acknowledgements
and leucocyte interaction (Table 1) [7,53]. As current data
Australian governments fund Australian Red Cross Life-
demonstrate differences in phenotype and releasate
blood to provide blood, blood products and services to
between platelets stored under different conditions [52,54],
the Australian community. The authors thank Chris Roan
there is the potential that this information could be used to
for his careful and critical reading of the manuscript.
better tailor transfusions to suit the specific condition and
risk factors of the recipients. Additionally, very few studies
have examined how stored platelets interact with recipient Conflicts of interest
leucocytes, which is believed to be a key factor in the inci-
The authors have no conflicts of interest to declare.
dence of adverse events [5,45]. This could be achieved by
the use of animal and in vitro models of transfusion.
Ultimately, clinical trials, which are in progress for both Author contributions
cold-stored and cryopreserved platelets, offer the perfect
All authors contributed to development of the concepts
opportunity to assess the immunological changes occurring
and design of the review article and critically reviewed
in the recipient post-transfusion and should be incorpo-
the manuscript. BW and LJ wrote the manuscript. BW
rated into the design of these trials. This knowledge would
prepared the figure.
aid in characterizing the pathogenesis of adverse events,
support the development of preventative strategies and
inform future changes to transfusion guidelines. Funding
Australian governments fund Australian Red Cross Life-
blood to provide blood, blood products and services to
Concluding remarks
the Australian community. BW is supported by a
Platelets are no longer recognized as simply mediators of Research Excellence Scholarship provided by the Univer-
haemostasis, but also as having a tangible role as an sity of Technology Sydney.
References
1 Stolla M, Refaai MA, Heal JM, et al.: 5 Caudrillier A, Kessenbrock K, Gilliss transfusion reactions. Transfusion
Platelet transfusion - the new BM, et al.: Platelets induce neutrophil 2014; 54:613–25
immunology of an old therapy. Front extracellular traps in transfusion- 8 Land WG: Transfusion-related acute
Immunol 2015; 6:1–10 related acute lung injury. J Clin Invest lung injury: the work of DAMPs.
2 Cognasse F, Nguyen K, Pauline D, 2012; 122:2661–71 Transfus Med Hemother 2013; 40:3–
et al.: The Inflammatory Role of Plate- 6 Losos M, Biller E, Li J, et al.: Pro- 13
lets via Their TLRs and Siglec Recep- longed platelet storage associated with 9 Cap AP, Spinella PC, Borgman MA,
tors. Front Immunol 2015; 6:1–15 increased frequency of transfusion- et al.: Timing and location of blood
3 Kral JB, Schrottmaier WC, Salzmann related adverse events. Vox Sang 2018; product transfusion and outcomes in
M, et al.: Platelet Interaction with 113:170–6 massively transfused combat casu-
Innate Immune Cells. Transfus Med 7 Hamzeh-Cognasse H, Damien P, alties. J Trauma Acute Care Surg
Hemother 2016; 43:78–88 Nguyen KA, et al.: Immune-reactive 2012; 73:S89–94
4 Clark SR, Ma AC, Tavener SA, et al.: soluble OX40 ligand, soluble CD40 10 Keitel S: Guide to the preparation, use
Platelet TLR4 activates neutrophil extra- ligand, and interleukin-27 are simulta- and quality assurance of blood compo-
cellular traps to ensnare bacteria in sep- neously oversecreted in platelet com- nents, 19th edn. Strasbourg, Council
tic blood. Nat Med 2007; 13:463–9 ponents associated with acute of Europe, 2017
11 Reade MC, Marks DC, Bellomo R, 22 Cognasse F, Hamzeh H, Chavarin P, activation of neutrophils in systemic
et al.: A randomized, controlled pilot et al.: Evidence of Toll-like receptor sclerosis. Sci Transl Med 2018; 10:
clinical trial of cryopreserved platelets molecules on human platelets. Immu- eaao3089
for perioperative surgical bleeding: the nol Cell Biol 2005; 83:196–8 33 Van Der Meijden PEJ, Van Schilf-
CLIP-I trial (Editorial, p. 2759). Trans- 23 Rossaint J, Margraf A, Zarbock A: gaarde M, Van Oerle R, et al.: Platelet-
fusion 2019; 59:2794-804. Role of platelets in leukocyte recruit- and erythrocyte-derived microparticles
12 Stubbs JR, Tran SA, Emery RL, et al.: ment and resolution of inflammation. trigger thrombin generation via factor
Cold platelets for trauma-associated Front Immunol 2018; 9:1–13 XIIa. J Thromb Haemost 2012;
bleeding: regulatory approval, accredi- 24 Vulliamy P, Gillespie S, Armstrong PC, 10:1355–62
tation approval, and practice imple- et al.: Histone H4 induces platelet bal- 34 Weiss R, Gr€ oger M, Rauscher S, et al.:
mentation-just the “tip of the iceberg". looning and microparticle release dur- Differential interaction of platelet-
Transfusion 2017; 57:2836–44 ing trauma hemorrhage. PNAS 2019; derived extracellular vesicles with
13 Strandenes G, Sivertsen J, Bjerkvig 116:17444–9 leukocyte subsets in human whole
CK, et al.: A pilot trial of platelets 25 Heijnen H, van der Sluijs P: Platelet blood. Sci Rep 2018; 8:6598
stored cold versus at room tempera- secretory behaviour: as diverse as the 35 Bei JJ, Liu C, Peng S, et al.: Staphylo-
ture for complex cardiothoracic sur- granules . . . or not? J Thromb Hae- coccal SSL5-induced platelet
gery. Anesthesiology 2020; 133:1173– most 2015; 13:2141–51 microparticles provoke proinflamma-
83. 26 Johnson L, Tan S, Jenkins E, et al.: tory responses via the CD40/TRAF6/
14 Brown GT, McIntyre TM: Lipopolysac- Characterization of biologic response NFjB signalling pathway in mono-
charide signaling without a nucleus: modifiers in the supernatant of con- cytes. Thromb Haemost 2016;
kinase cascades stimulate platelet ventional, refrigerated, and cryopre- 115:632–45
shedding of proinflammatory IL-1b– served platelets. Transfusion 2018; 36 Sprague DL, Elzey BD, Crist SA, et al.:
rich microparticles. J Immunol 2011; 58:927–37 Platelet-mediated modulation of adap-
186:5489–96 27 Cognasse F, Sut C, Hamzeh-Cognasse tive immunity: unique delivery of
15 Hamzeh-Cognasse H, Berthelot P, H, et al.: Platelet-derived HMGB1: CD154 signal by platelet-derived
Tardy B, et al.: Platelet toll-like recep- critical mediator of SARs related to membrane vesicles. Blood 2008;
tors are crucial sensors of infectious transfusion. Ann Transl Med 2019; 111:5028–36
danger moieties. Platelets 2018; 8:140–1 37 Nomura S, Tandon NN, Nakamura T,
29:533–40 28 Cognasse F, Sut C, Fromont E, et al.: et al.: High-shear-stress-induced acti-
16 Blair P, Flaumenhaft R: Platelet alpha- Platelet soluble CD40-Ligand level is vation of platelets and microparticles
granules: basic biology and clinical associated with transfusion adverse enhances expression of cell adhesion
correlates. Blood Rev 2009; 23:177–89 reactions in a mixed threshold and hit molecules in THP-1 and endothelial
17 Lindemann S, Tolley ND, Dixon DA, model. Blood 2017; 130:1380–3 cells. Atherosclerosis 2001; 158:277–
et al.: Activated platelets mediate 29 Khan SY, Kelher MR, Heal JM, et al.: 87
inflammatory signaling by regulated Soluble CD40 ligand accumulates in 38 Barry OP, Pratico D, Lawson JA, et al.:
interleukin 1b synthesis. J Cell Biol stored blood components, primes neu- Transcellular activation of platelets
2001; 154:485–90 trophils through CD40, and is a poten- and endothelial cells by bioactive
18 Marcoux G, Magron A, Sut C, et al.: tial cofactor in the development of lipids in platelet microparticles. J Clin
Platelet-derived extracellular vesicles transfusion-related acute lung injury. Invest 1997; 99:2118–27
convey mitochondrial DAMPs in pla- Blood 2006; 108:2455–62 39 Zhao J, Xu B, Chen G, et al.: Cryopre-
telet concentrates and their levels are 30 Jin R, Yu S, Song Z, et al.: Soluble served platelets augment the inflam-
associated with adverse reactions. CD40 ligand stimulates CD40-depen- matory response: role of
Transfusion 2019; 59:2403–14 dent activation of the b2 integrin phosphatidylserine- and P-selectin–
19 Aatonen M, Gr€ onholm M: Siljander Mac-1 and protein kinase C zeda mediated platelet phagocytosis in
PR-M: Platelet-derived microvesicles: (PKCf) in neutrophils: implications for macrophages. Transfusion 2019;
multitalented participants in intercel- neutrophil-platelet interactions and 59:1799–808
lular communication. Sem Thromb neutrophil oxidative burst. PLoS One 40 Shrivastava M: The platelet storage
Hemost 2012; 38:102–13 2013; 8:1–10 lesion. Transfus Apher Sci 2009;
20 Ali RA, Wuescher LM, Worth RG: Pla- 31 Baj-Krzyworzeka M, Majka M, Pratico 41:105–13
telets: essential components of the D, et al.: Platelet-derived microparti- 41 Rainger GE, Buckley C, Simmons DL,
immune system. Curr Trends Immunol cles stimulate proliferation, survival, et al.: Cross-talk between cell adhesion
2015; 16:65–78 adhesion, and chemotaxis of molecules regulates the migration
21 Thon JN, Peters CG, Machlus KR, hematopoietic cells. Exp Hematol velocity of neutrophils. Curr Biol
et al.: T granules in human platelets 2002; 30:450–9 1997; 7:316–25
function in TLR9 organization and 32 Maugeri N, Capobianco A, Rovere- 42 Patko Z, Csaszar A, Acsady G, et al.:
signaling. J Cell Biol 2012; 198:561– Querini P, et al.: Platelet microparti- Roles of Mac-1 and glycoprotein IIb/
74 cles sustain autophagy-associated IIIa integrins in leukocyte–platelet
of cryopreserved platelets. Transfusion and liquid platelets: potential clinical 80 Stolla M, Bailey SL, Fang L, et al.:
2018; 58:2645–56 implications. Transfusion 2014; Effects of storage time prolongation
75 Birge RB, Boeltz S, Kumar S, et al.: 55:838–47 on in vivo and in vitro characteristics
Phosphatidylserine is a global 78 Ki KK, Johnson L, Faddy HM, et al.: of 4° C–stored platelets. Transfusion
immunosuppressive signal in efferocy- Immunomodulatory effect of cryopre- 2020; 60:613–21
tosis, infectious disease, and cancer. served platelets: altered BDCA3+ den- 81 Rondina MT, Garraud O: Emerging
Cell Death Differ 2016; 23:962–78 dritic cell maturation and activation evidence for platelets as immune and
76 Crimmins D, Flanagan P, Charlewood in vitro. Transfusion 2017; 57:2878– inflammatory effector cells. Front
R, et al.: In vitro comparison between 87 Immunol 2014; 5:1–6
gamma-irradiated cryopreserved and 79 Zhao J, Sun Z, You G, et al.: Transfu- 82 Morrell CN, Aggrey AA, Chapman LM,
Day 7 liquid-stored buffy coat–derived sion of cryopreserved platelets exacer- et al.: Emerging roles for platelets as
platelet components. Transfusion bates inflammatory liver and lung immune and inflammatory cells. Blood
2016; 56:2799–807 injury in a mice model of hemorrhage. 2014; 123:2759–67
77 Johnson L, Reade M, Hyland R, et al.: J Trauma Acute Care Surg 2018;
In vitro comparison of cryopreserved 85:327–33
Glycophorin A (GPA) and Band 3 (also known as Anion using GROMOS [11], it was determined that although the
Exchanger 1) are the two most abundant integral proteins in majority of the monomeric extracellular domain appeared
the red blood cell (RBC) membrane [1]. GPA facilitates the to have no stable secondary structure, a stable b-hairpin
expression of Band 3 in the cell membrane [2], and the two was identified at residues Glu64 to His85. The Gly74 to
proteins are co-precipitated by an anti-Wrb antibody [3]. Phe87 region of GPA had the propensity to form a b-
Wrb, which is antithetical to Wra [4], is a high-frequency sheet as one half of this b-hairpin.
antigen, and antibodies against it can cause haemolytic To further investigate this, our group produced initial
transfusion reactions [5]. Although it has been observed that homology models, based on our GPA model, showing that
Glu658 of Band 3 is a required amino acid for the formation GP.Sch (Wrb + ve) displays a b-hairpin secondary struc-
of the Wrb antigen [6], a number of proposals have been put ture, in contrast to the highly homologous GP.Dantu
forward for the GPA site of interaction. (Wrb-ve) which lacks the b-hairpin (Figs. 1 & 2). Further-
Bruce et al. [6] proposed that the interaction occurs more, the b-sheet in GP.Sch, formed at 49VQLAH53, and
between residues 80–89 (formerly 61–70) of GPA, with b-sheet in GPA at 81VQL83 are consistent with the resi-
Arg80 (formerly Arg61) potentially forming a charge-pair dues proposed by Huang et al. [7] to be critical for Wrb
with Glu658 of Band 3. However, Huang et al. [7] have formation: 81VQL83 His85 and 87FSEP90 (formerly 62–64,
suggested that Arg80 is not crucial for the formation of 66 and 68–71). 87FSEP90 is the start of the GPA trans-
the antigen, and that Wrb is reliant on a longer sequence membrane domain, where Young et al. [12] showed that
87
in GPA: Val70 to Ile95 (formerly 51–76). This sequence is FSE89 is associated with enhancement of Band 3 anion
encoded by part of exon 3 and extends to the beginning transport capabilities.
of the transmembrane domain encoded in exon 5. Huang et al. [7] further postulated that stabilization of
Early computational prediction of Gly74 to Phe87 (for- Wrb is reliant upon parallel packing of the a-helical
merly 55-68), using a small GPA peptide, presented as an transmembrane domains of GPA and Band 3, forcing
a-helical region [8]. Poole et al. [9] showed that the alter- proximity between the extracellular domains. Molecular
ation of GPA Ala84 (formerly Ala65)?Pro resulted in dynamics simulations of Band 3 and GPA by Kalli and
formation of an aberrant Wrb. They concluded that Reithmeier [13] substantiate this finding. These simula-
because Proline is a helix breaker, this polymorphism dis- tions include only the transmembrane domain of both
rupted the proposed a-helical structure. When modelling proteins, with Arg80 (GPA) anchored to Glu658 (Band 3)
various small GPA peptides in our group, similar a-he- based on the findings by Bruce et al. [6]. Our simulations
lices were seen. However, when modelling the complete placed Arg80 on the end of a central b-sheet (Fig. 3B)
extracellular domain of the GPA monomer (performed rather than at the end of a flexible loop segment as done
using de novo techniques facilitated by the Robetta soft- by Kalli and Reithmeier [13]. With the addition of the
ware [10]), it appeared to be comprised of 3 to 5 b- extracellular domains and altered presentation of Arg80,
strands. After 200-ns molecular dynamics simulations it is possible that the molecular dynamics of GPA/Band 3
interactions would not mimic what is shown by Kalli and
Correspondence: Xuan T. Bui, ARC Training Centre for
Reithmeier. However, it is possible that the clustering of
Biopharmaceutical Innovation, Australian Institute for Bioengineering Band 3 shown in Kalli and Reithmeier’s simulations, and
and Nanotechnology, The University of Queensland, Australia, Building predicted by Huang et al. [7,13], is facilitated by residues
75, Cnr College Rd & Cooper Rd, St Lucia QLD 4072 in the transmembrane domain as opposed to anchoring
E-mail: x.bui@uq.edu.au by Arg80 and Glu658.
489
490 S. Ekman et al.
Fig. 1 Initial tertiary structural predictions of GPA, GPB and hybrid glycophorins GP.Sch, GP.Dantu, GP.SAT, GP.Hil and GP.Mur. Models were produced
using comparative modelling techniques under the parameters of the Robetta software [10] using a previously designed GPA extracellular domain mono-
mer model as a template (GPA monomer created using de novo modelling techniques through Robetta software). GPA exon 3 (or exon 3 insert) is
coloured in cyan, GPA exon 4 is coloured in dark blue, GPB exon 2 in orange, GPB pseudoexon 3 segment is coloured in red, GPB exon 4 is coloured in
magenta. GPA or GPB transmembrane (TM) and Wrb status included for each structure. The images were produced using PyMOL [15].
Fig. 2 Comparison of MNS variant allele sequences, presence of Wrb, b-hairpin and GPA transmembrane (TM). GPA exons presented in blue, and GPB
exons presented in green for GPA, GPB and hybrid glycophorins GP.Dantu, GP.Sch, GP.SAT, GP.Hil and GP.Mur. GPA, GPB, GP.Dantu, GP.Sch sequences
adapted from [7,14], GP.SAT sequence from [16], GP.MiV (GP.Hil) sequence from [17], other sequences. Wrb interaction site by Huang et al. and Hsu
et al. outlined in red [7,14].
An alternate site for Wrb expression was proposed by non-expressing cells, whilst expression of Wrb and Band
Hsu et al. [14] as their study showed that GPA expression 3 was increased in GP.Mur-positive cells. They concluded
levels remained consistent in GP.Mur expressing and that the increase in expression of Wrb was due to GP.Mur
Fig. 3 Comparison of proposed GPA amino acids required for Wrb formation. GPA and GP.Mur extracellular homodimer created using HADDOCK web
server [18] from GPA and GP.Mur monomers, respectively. GP.Mur monomer created from comparative modelling techniques using Robetta software [10]
with GPA monomer template. GPA transmembrane X-ray crystal structure PDB 5EH4 [19]. GP.Mur transmembrane domain created using comparative
modelling of GPA transmembrane. GPA extracellular domain and transmembrane domain homodimer coloured in green with Wrb interaction sites in red
A) Ridgwell et al. [8], B) Bruce et al. [6] (with essential amino acid Arg80 in blue), C) Huang et al. [7], D) Poole et al. [9] (with essential amino acid
Ala84 in blue), E) Our identified exon 3-4 b-hairpin (residues 64–84). GP.Mur extracellular domain and transmembrane domain homodimer in blue with
interaction sites in red F) Our identified exon 3-4 b-hairpin (residues 64–84) and G) Wrb interaction site proposed by Hsu et al. [14] (with Mur antigen
coloured green). Images were visualized using PyMOL [15].
interacting with Band 3 alongside GPA, hence its facilita- could be tested by repeating the experiments of Hsu et al.
tion of Wrb production. However, we propose that the [14] using transfected cells that express GP.Mur and not
increase in Band 3 expression, facilitated by GP.Mur, GPA, to determine whether there is still Wrb expression.
allows a greater proportion of GPA to form a complex If GP.Mur is shown to be directly involved in Wrb forma-
with Band 3, rather than GP.Mur itself forming the Wrb tion, this would be in contrast to the mechanism proposed
antigen. This is based on the report by Poole et al. [1] by Huang et al. [7].
that orientation plays a crucial role in Wrb expression, Hsu et al.’s Wrb antigen site was identified as the
indicating that not all available Band 3 proteins would region Asp46 to His60 (GP.Hil)/Ser62 (GP.Mur). This was
interact with all available GPA proteins. This hypothesis because Gp.Mur and Gp.Hil differ only in this region,
where GP.Hil has been shown to be Wrb-ve. However, this anion transport, yet is Wrb-ve [7,12]. Further research will
does not explain the expression of Wrb on GP.Sch be necessary to validate these conclusions and to eluci-
homozygous cells [7], which lack this region. Moreover, date the mechanism by which GP.Mur facilitates elevation
this region is located in a different equivalent region of of Wrb antigen expression.
GPA compared to numerous other proposals for the Wrb
interaction site (Fig. 3).
Conflicts of interest
In conclusion, the 87FSE89 segment of the transmem-
brane domain is likely associated with both anion trans- The authors declare no financial or commercial conflict
portation in Band 3 and the formation of the Wrb of interest.
antigen. It is also highly likely that an extracellular
domain addition is required for Wrb antigen formation
Funding
including the residues 49VQLAH53 which have been pre-
dicted by our molecular dynamic simulations to be one Elements of this research utilized the facilities, and the sci-
half of a b-hairpin structure. The requirement for both entific and technical assistance of the National Biologics
the GPA transmembrane domain and the extracellular Facility (NBF) at The University of Queensland. NBF is
domain b-strand would explain why the hybrid gly- supported by Therapeutic Innovation Australia (TIA). Fund-
cophorin GP.Dantu (containing GPB extracellular domain ing supported by the Australian Research Council through
with no b-hairpin, and the GPA transmembrane domain) the Industrial Transformation Research Program (ITRP)
can enhance Band 3 expression and increase levels of scheme.
References
1 Poole J: Red cell antigens on band 3 8 Ridgwell K, Tanner MJ, Anstee DJ: The Molecular organization of the Wright
and glycophorin A. Blood Rev 2000; Wrb antigen, a receptor for Plasmod- (Wr) blood group antigen. PLoS Com-
14:31–43 ium falciparum malaria, is located on a put Biol 2018; 14:e1006284
2 Groves JD, Tanner MJ: Glycophorin A helical region of the major membrane 14 Hsu K, Lin Y-C, Lee T-Y, et al.: Mil-
facilitates the expression of human sialoglycoprotein of human red blood tenberger blood group antigen subtype
band 3-mediated anion transport in cells. Biochem J 1983; 209:273–6 III (Mi.III) supports Wrb expression.
Xenopus oocytes. J Biol Chem 1992; 9 Poole J, Banks J, Bruce LJ, et al.: Gly- Vox Sang 2011; 100:389–94
267:22163–70 cophorin A mutation Ala65 –>; Pro 15 DeLano WL: Pymol: An open-source
3 Telen M, Chasis J: Relationship of the gives rise to a novel pair of MNS alle- molecular graphics tool. CCP4 News-
human erythrocyte Wrb antigen to an les ENEP (MNS39) and HAG (MNS41) lett Protein Crystallogr 2002; 40:82–92
interaction between glycophorin A and and altered Wrb expression: direct evi- 16 Huang C, Reid M, Okubo Y, et al.:
band 3. Blood 1990; 76:842–8 dence for GPA/band 3 interaction nec- Glycophorin SAT of the human ery-
4 Wren MR, Issitt PD: Evidence that Wra essary for normal Wrb expression. throcyte membrane is specified by a
and Wrb are antithetical. Transfusion Transfus Med 1999; 9:167–74 hybrid gene reciprocal to glycophorin
1988; 28:113–8 10 Kim DE, Chivian D, Baker D: Protein Dantu gene. Blood 1995; 85:2222–7
5 Daniels G: Diego blood group system; structure prediction and analysis using 17 Johe K, Vengelen-Tyler V, Leger R,
human blood groups. Oxford: Wiley- the Robetta server. Nucleic Acids Res et al.: Synthetic peptides homologous
Blackwell, 2013; 336–53. 2004; 32:W526–W31 to human glycophorins of the Mil-
6 Bruce L, Ring S, Anstee D, et al.: 11 Schmid N, Eichenberger AP, Choutko tenberger complex of variants of
Changes in the blood group Wright A, et al.: Definition and testing of the MNSs blood group system specify the
antigens are associated with a muta- GROMOS force-field versions 54A7 epitopes for Hil, SJL, Hop, and Mur
tion at amino acid 658 in human ery- and 54B7. Eur Biophys J 2011; 40:843 antisera. Blood 1991; 78:2456–61
throcyte band 3: a site of interaction 12 Young MT, Tanner MJA: Distinct regions 18 van Zundert GCP, Rodrigues JPGLM,
between band 3 and glycophorin A of human glycophorin A enhance human Trellet M, et al.: The HADDOCK2.2
under certain conditions. [see com- red cell anion exchanger (Band 3; AE1) web server: user-friendly integrative
ments] Blood 1995; 85:541–7 transport function and surface traffick- modeling of biomolecular complexes.
7 Huang C, Reid M, Xie S, et al.: Human ing. J Biol Chem 2003; 278:32954–61 JMB 2016; 428:720–5
red blood cell Wright antigens: a 13 Kalli AC, Reithmeier RAF: Interaction 19 Trenker R, Call ME, Call MJ: Crystal
genetic and evolutionary perspective of the human erythrocyte Band 3 structure of the glycophorin A trans-
on glycophorin A-band 3 interaction. anion exchanger 1 (AE1, SLC4A1) membrane dimer in lipidic cubic phase.
Blood 1996; 87:3942–7 with lipids and glycophorin A: J Am Chem Soc 2015; 137:15676–9
1
Transfusion Research Center, Belgian Red Cross-Flanders, Ghent, Belgium
2
Department of Immunopathology, Sanquin Research and Landsteiner Laboratory Academic Medical Centre, Amsterdam, Netherlands
3
Department of Transfusion Medicine, Ulm University, Ulm, Germany
4
urttemberg – Hessen and
Institute for Clinical Transfusion Medicine and Immunogenetics, German Red Cross Blood Transfusion Service Baden-W€
University Hospital Ulm, Ulm, Germany
5
Etablissement Francßais du Sang, La Plaine St-Denis, France
6
UMR1098 RIGHT, INSERM, Etablissement Francßais du Sang, Universite de Franche-Comte , Besancßon, France
7
Microbiology Services, NHS Blood and Transplant, London, UK
Severe acute respiratory syndrome coronavirus 2 (SARS- respectively [5,6]. These assays utilize live SARS-CoV-2
CoV-2) is the causative agent of the ongoing COVID-19 virus and, hence, require a biosafety level 3 (BSL-3) facil-
pandemic. It is responsible for more than 1 million deaths ity. In addition, it is time-consuming (5–7 days). Further-
worldwide already [1]. Because preventive and anti-viral more, the output data cannot be compared among
treatment options are still limited, COVID-19 convalescent laboratories because different assay readouts (e.g. virus
plasma (CPP) has been suggested as a potential therapy concentration or % inhibition) and protocols are currently
[2–4]. being used. In addition, an international standard is not
‘Convalescent’ implies that anti-SARS-CoV-2 antibod- yet available. Blood establishments may choose to partner
ies are present in plasma collected from individuals with a virology laboratory that can perform viral neutral-
recovered from COVID-19. However, the dose and nature ization on donor samples. Alternatively, other assays are
of antibodies required to effectively interfere with a available using pseudoviruses (i.e. a recombinant virus
SARS-CoV-2 infection is unclear. Most ongoing observa- expressing a SARS-CoV-2 protein) that require lower bio-
tional studies and prospective clinical trials currently safety levels [7].
focus on neutralizing antibodies (nAbs) that interfere with Anti-SARS-CoV-2 antibody titers can also be measured
viral binding to host cells, but non-neutralizing antibod- using immunoassays such as enzyme-linked (ELISA) and
ies might mediate a therapeutic effect as well. These and chemiluminescent immunoassays (CLIA) which are based
other unknowns highlight the importance of testing CCP on biochemical detection of antibody binding to viral
efficacy in randomized trials. This commentary conse- proteins. Recently, the FDA suggested that all putative
quently does not claim to provide evidence on how to CCP donations should be tested in the Ortho VITROS
select potent CCP, but does want to provide an opinion- SARS-CoV-2 IgG CLIA-based test and donations with a
based discussion on how to investigate CCP potency. signal to cut-off of 12 or higher to be qualified as a high
The antibody level in CCP varies greatly between titer plasma [8]. In contrast, European blood establish-
donors. Therefore, it is required to measure antibody titer ments are using a variety of commercial immunoassays
and/or to assess the neutralization potency of CCP. The (Table 1), making it more difficult to compare data across
current gold standard for the latter is in vitro viral neu- the region. Sensitivities and specificities of the commer-
tralization like in the plaque reduction neutralization test cial assays presented in Table 1 can differ from those
(PRNT) or microneutralization (MN) assay. Both measure provided by the respective manufacturers. Thresholds,
the ability of nAbs to prevent infection in vitro calculated sensitivities and specificities may change depending on
either as a reduction in the formation of plaques or as the sample size, the timing post-symptom onset and the sero-
inhibition of viral infectivity in a cell monolayer, prevalence in the population [9,10].
Immunoassays allow the detection of total or isotype-
specific antibody binding the spike (S), receptor binding
Correspondence: Hendrik B. Feys, Transfusion Research Center, Belgian domain of spike (RBD) or nucleocapsid (N) proteins. In
Red Cross-Flanders, Ghent, Belgium our opinion, immunoassays for IgG targeting RBD are
E-mail: hendrik.feys@rodekruis.be most likely to be relevant because (i) most potent
*These authors contributed equally to this work.
This is an open access article under the terms of the Creative Commons Attribution‐NonCommercial‐NoDerivs License,
which permits use and distribution in any medium, provided the original work is properly cited, the use is non‐commercial and 493
no modifications or adaptations are made.
494
References
IgG is efficiently transported across the epithelial lung
[22]
[23]
[22]
[24]
[25]
barrier [11] and (iii) IgG has a longer half-life. Finally,
immunoassays are compatible with BSL-1 facilities, do
not require sophisticated technology and may be emu-
10 432/10 453
lated on robots to increase throughput.
Sample size
500/540
184/185
(90.0–94.7)c
(71.0–89.0)a
S-RBD
S-RBD
ELISA
CLIA
CLIA
CLIA
d
a
Fig. 1 Release of neutralizing CCP units using varying assay thresholds. [Colour figure can be viewed at wileyonlinelibrary.com]
no reduction in disease progression nor mortality [20] but international standards are anticipated to be made avail-
also did not determine nAbs levels upfront. Post hoc analy- able by the WHO in December 2020, which will facilitate
sis showed that the median titer of nAbs in this study was such direct comparisons [18].
low. Together with the scientific rationale of biochemical Although the observational studies are suggestive for
interference with viral binding, we suggest that CCP selec- CCP efficacy, hard evidence is lacking. Additional studies
tion is based on medium to high signal thresholds (i.e. the are required, but IgG levels obtained by ELISA seem to
top 30–40% of donations containing Abs). This selection correlate well with virus neutralization titers. This indi-
strategy may change in the future once the minimal effec- cates that an ELISA/CLIA assay can be used to select CCP
tive dose of nAbs has been established and high-through- donors, also in the light of the urgency. However, stan-
put standardized assays that can reliably predict viral dardization of ELISAs will be essential.
neutralization potency are available.
As mentioned previously, standardization or calibration
Acknowledgements
of these immuno- and neutralization assays to allow
comparison of data across studies has not yet been per- We thank Gaia Mori and Catherine Hartmann for the
formed. In this context, the European Commission and expert co-ordination of the SUPPORT-E project.
the European Blood Alliance (EBA) recently launched a
joint initiative to support high-quality clinical evaluation
Conflict of interests
of CCP. This SUPPORT-E consortium (Supporting high-
quality evaluation of COVID-19 convalescent plasma The authors declared no potential conflicts of interest
throughout Europe) [21] will investigate the relationship with respect to the research, authorship and/or publica-
between (i) donor and donation parameters, (ii) antibody tion of this article.
content and nature and (iii) clinical outcome of CCP
recipients in EU cohorts. The consortium will also provide
Funding
support for testing and distributes calibration standards
among participating blood establishments in the EU to This study was supported by the European Commission
allow cross border standardization of assays. In addition, (SUPPORT-E, grant number 101015756).
References
1 Coronavirus/COVID-19 Worldwide 3 Chai KL, Valk SJ, Piechotta V, et al.: with COVID-19: Initial Three-Month
Cases Live Data & Statistics. 2020. Convalescent plasma or hyperimmune Experience. medRxiv 2020.
https://covidstatistics.org/ (Last immunoglobulin for people with COVID- 5 Wolfel R, Corman VM, Guggemos W,
accessed October 22nd 2020). 19: a living systematic review. Cochrane et al.: Virological assessment of hospi-
2 Casadevall A, Joyner MJ, Pirofski LA: Database Syst Rev 2020; 10:CD013600 talized patients with COVID-2019. Nat-
A randomized trial of convalescent 4 Joyner MJ, Senefeld JW, Klassen SA, ure 2020; 581:465–9
plasma for COVID-19-potentially hope- et al.:Effect of Convalescent Plasma on 6 Conzelmann C, Gilg A, Gross R, et al.:
ful signals. JAMA 2020; 324:455–7 Mortality among Hospitalized Patients An enzyme-based immunodetection
assay to quantify SARS-CoV-2 infec- Assessment of methods available for 20 Agarwal A, Mukherjee A, Kumar G,
tion. Antiviral Res 2020; 181:104882 antibody detection and their correla- et al.: Convalescent plasma in the
7 Thompson CP, Grayson N, Paton R, tion with neutralising antibody levels. management of moderate covid-19 in
et al.: Detection of neutralising anti- Transf Med 2020 adults in India: open label phase II
bodies to SARS-CoV-2 to determine 14 Peterhoff D, Gluck V, Vogel M, et al.: multicentre randomised controlled
population exposure in Scottish blood A highly specific and sensitive sero- trial (PLACID Trial). BMJ 2020; 371:
donors between March and May 2020. logical assay detects SARS-CoV-2 m3939
Euro Surveill 2020; 25(42):2000685 antibody levels in COVID-19 patients 21 Commission supports crucial research
8 U.S. Department of Health and Human that correlate with neutralization. on convalescent plasma to treat
Services: Investigational COVID-19 Infection 2020; 1–8 COVID-19. September 11th 2020.
Convalescent Plasma: Guidance for 15 Jahrsdorfer B, Kroschel J, Ludwig C, https://ec.europa.eu/info/news/commis
Industry. September 2020. https:// et al.: Independent side-by-side vali- sion-supports-crucial-research-convale
www.fda.gov/regulatory-information/ dation and comparison of four sero- scent-plasma-treat-covid-19-2020-se
search-fda-guidance-documents/inve logical platforms for SARS-CoV-2 p-11_en (Last accessed October 22nd
stigational-covid-19-convalescent-pla antibody testing. J Infect Dis 2020; 2020).
sma(Last accessed October 22nd 2020). 16:jiaa656 22 GeurtsvanKessel CH, Okba NMA, Igloi
9 Mathur G, Mathur S: Antibody testing 16 COVID-19 Convalescent Plasma Trans- Z, et al.: An evaluation of COVID-19
for COVID-19. Am J Clin Pathol 2020; fusion. 2020. https://ec.europa.eu/hea serological assays informs future diag-
154:1–3 lth/blood_tissues_organs/covid-19_en nostics and exposure assessment. Nat
10 Theel ES, Harring J, Hilgart H, et al.: (Last accessed October 21th 2020). Commun 2020; 11:3436
Performance characteristics of four 17 Harvala H, Mehew J, Robb ML, et al.: 23 Harritshoej LH, Gybel-Brask M, Afzal
high-throughput immunoassays for Convalescent plasma treatment for S, et al.: Comparison of sixteen sero-
detection of IgG antibodies against SARS-CoV-2 infection: analysis of the logical SARS-CoV-2 immunoassays in
SARS-CoV-2. J Clin Microbiol 2020; first 436 donors in England, 22 April sixteen clinical laboratories. medRxiv
58:e01243–20 to 12 May 2020. Euro Surveill 2020; 2020: 2020.07.30.20165373.
11 Kim KJ, Malik AB: Protein transport 25(28):2001260 24 Ainsworth M, Andersson M, Auckland
across the lung epithelial barrier. Am 18 FDA calculator to select a COVID-19 K, et al.: Performance characteristics
J Physiol Lung Cell Mol Physiol 2003; Antibody Test for your community. of five immunoassays for SARS-CoV-
284:L247–59 2020. https://www.cdc.gov/corona 2: a head-to-head benchmark compar-
12 Luchsinger LL, Ransegnola B, Jin D, virus/2019-ncov/downloads/lab/fda-ca ison. Lancet Infect Dis 2020; 20:1390–
et al.: Serological assays estimate lculator.pdf (Last accessed October 400
highly variable SARS-CoV-2 neutral- 22nd 2020). 25 Muench P, Jochum S, Wenderoth V,
izing antibody activity in recovered 19 Salazar E, Christensen PA, Graviss EA, et al.: Development and validation of
COVID19 patients. J Clin Microbiol et al.: Treatment of Coronavirus Dis- the Elecsys anti-SARS-CoV-2
2020; 58:e02005-20–e02020 ease 2019 Patients with Convalescent immunoassay as a highly specific tool
13 Harvala H, Watkins N, Ijaz S, et al.: Plasma Reveals a Signal of Signifi- for determining past exposure to
Convalescent plasma therapy for the cantly Decreased Mortality. Am J SARS-CoV-2. J Clin Microbiol 2020;
treatment of patients with COVID-19: Pathol 2020; 190:2290–303 58:e01694–20
Background
Maintaining an adequate and safe blood supply is one of
the primary aims of blood centres across the world. Blood
*Correspondence: Hem Chandra Pandey, Department of Transfusion
Medicine, All India Institute of Medical Sciences, New Delhi, supply chain is affected by a number of factors which
India-110029. may operate independently or may influence each other.
E-mail: pandeyhemc@gmail.com Three main factors which govern the blood supply
497
498 H. C. Coshic et al.
include the availability of blood donors, the usage of data from the period starting from the day when first case
blood and blood components in the centre as well as the of COVID-19 was reported in the country (i.e. 29 Jan
availability of critical supplies required for collection and 2020) till the end of the first lockdown in India (i.e. 14
testing of blood. The extent to which any event affects April 2020) was included. Ethical permission to collect
blood supply will thus depend on the model of blood and analyse the data was obtained from institute ethics
transfusion services being followed and the responses of committee.
the centre/country to such an event. Our institute is a tertiary level research institute and
Blood transfusion services in India are decentralized receives referral for complex cases from all over the
with around 3108 licensed blood banks collecting around country. The hospitals attached to the institute provide
122 million units against an estimated clinical demand services catering to a number of specialities ranging from
of 146 million units of blood [1, 2]. Most of the blood basic medical/surgical services to complex procedures
centres are hospital based with the blood supply mainly including bone marrow transplants, organ transplants etc.
dependent on replacement donors or blood collected from There are three hospital-based blood banks with the other
donors at off-site voluntary blood donation drives. Each two blood banks providing services to trauma patients
of these centres maintains their own blood stocks and and cardio/neuro patients respectively. Our blood bank
processes blood based on daily demands with a minimal supplies blood and blood component to all other speciali-
buffer stock to tide over acute disruptions in blood supply ties other than those served by the other two blood banks
arising out of local epidemics, disasters or other such and a good coordination exists between the three blood
local events. Majority of blood transfusions are being banks in case of shortage of blood in one of the centres.
done for correction of anaemia due to nutritional causes
approximating to 39%, for haemato-oncologic needs
Study time periods
approximating to 37% and 3%–4% in thalassaemia
patients [2, 3]. To add to this, most blood banks are For the ease of describing the supply and demand
dependent on critical supplies and reagents which origi- dynamics, we divided the study period into three distinct
nate in developed countries with only a minority of blood periods. The first time period (Time Period A) included
banking reagents being manufactured in our country. the blood stock data from 29 Jan 2020 to 26 Feb 2020.
Despite these limitations, blood centres across the coun- During this period, the blood bank supply and demand
try fare quite well in situations causing small local disrup- were unchanged and not affected by the pandemic with
tions in supply and demand by coordinating with each only a single case reported in the country and thus this
other and by following government regulations favouring data served as a baseline to compare the stock and
transport of blood across different centres with well-de- demand during the period of lockdown. The blood centre
fined guidelines [4]. In addition to disrupting the medical used to follow first-in-first-out (FIFO) policy with certain
services, the 2019-20 corona virus (SARS-CoV-2) pan- exceptions such as the provision of blood within 7 days
demic has challenged the blood centres across the world to of collection to thalassaemia patients etc.
maintain a balance between blood supply and demand in The second time period (Time Period B) included blood
an unprecedented way [5–7]. A significant reduction in stock data from 27 Feb 2020 to 21 March 2020. This time
blood donation was also observed during previous SARS period differed from the first time period by the fact that
and MERS-CoV outbreaks due to pan-lockdown, panic, our centre implemented steps to increase the buffer stock
cessation of off-site blood donation drives and disruption (safety stock) of red cells to approximately two weeks
of transport services [8–10]. During the past three months instead of 1 week. This decision to increase the buffer
since the first case of SARS-CoV-2 was reported in India stock was an internal administrative decision taken by
on 29 January 2020, blood centres have adopted a number the blood centre last year as part of increased readiness
of strategies to tide over this crisis. The purpose of the pre- towards acute red cell shortage in summer season and
sent study is to share our experiences and the effect of var- was not in response to the impending SARS-CoV-2 pan-
ious policies adopted by our blood centre from time to time demic. Multiple high-volume off-site voluntary blood
and the lessons learned from it for future pandemics. donation drives were planned by the blood centre at reg-
ular intervals to achieve the target of maintaining a
higher buffer stock. FIFO policy was continued during
Material and methods
this period also.
The third time period (Time Period C) included blood
Study setting
stock data from 22 March 2020 to 14 April 2020. This
The study was done in Main Blood Bank, Department of time period marked the start of the pan-country lockdown
Transfusion Medicine, AIIMS, New Delhi and blood stock till the intended period of closure of the first lockdown.
Table 1 Existing disaster plan vs revised/additional strategies to maintain adequate blood supply during SARS-CoV-2 lockdown
• Maintain buffer stock to 1-week level • Maintain buffer stock to 2-week level
• Defer elective surgeries depending on the blood need • Defer/limit routine transfusions in addition to postponing
• Maintain a list of voluntary blood donors elective surgery
• Maintain reagent/critical supply stock to at least • Maintain a registry of blood donors from within the institute
1 month staff or neighbouring areas
• Arrange blood units from centres unaffected by the • Maintain inventory of Single Donor Platelet collected from
disaster voluntary donors
• Maintain reagent/critical supply stock to at least 2 months
• Daily prospective audit of blood requests
Effect of lockdown on blood supply and demand during cell issue of around 133 units (IQR = 124–157) (Fig. 1). A
this period and various strategies to maintain an adequate buffer stock (safety stock) of around 900 red cell units
and safe blood supply were reflected by this time period. was available during this time period which was sufficient
Changes in the existing disaster management plans were to sustain blood centre activities for a period of one week
done to cope up with the disrupted supply chain and their if blood supply is disrupted and demand remained same.
effects on blood supply were analysed (Table 1). FIFO There was no discard of blood units due to outdate during
policy was implemented strictly and no exceptions to this this time period.
policy were followed. The daily stock of whole blood derived/ random donor
platelet (RDP) during time period A was around 185 units
(IQR = 156–226) with a collection of around 101 units
Data collection and analysis
(IQR = 87–151) and a daily issue of 128 units
The data were collected using blood bank management (IQR = 110–137) (Fig. 1b). Thus, on an average a total of
software reports for each day of operation. The data was 70 platelet concentrate units remained in the buffer stock.
entered in an excel sheet and analysed to calculate the The data does not include apheresis platelet concentrates
daily stock as well as buffer stock. Blood collection, blood as it is prepared on demand and directed to the patient
issue and stock were calculated and represented as med- for whom it was donated.
ian (interquartile range, IQR), and a comparative analysis
of these parameters was done for the three time periods
Time period B – planned increase to counter
described above. The effect of implementing different
summer season decline in inventory
strategies during the third time period was also studied.
The total number of blood units collected during this time
period were 4350 units over a period of 24 days. As evi-
Results
dent from Fig. 1a, the daily stock of red cell units
The blood supply and demand dynamics are discussed increased to 1596 units (IQR = 1480–1765) despite the
under three different time periods as described in the daily collection remaining almost same at 106 units
methods section. This follows a detailed day-by-day (IQR = 80–140). This was the result of the high-volume
description of the various effects of COVID-19 pandemic off-site blood donation drive of >2000 blood units during
on blood stock dynamics. this time period and is represented as outlier in Fig. 1a.
The daily issue of red cells remained at the same level
132 units (IQR = 118–153) which resulted in an increase
Time period A – normal functioning of the blood
in the safety stock to 1600 red cell units equivalent to a
centre
stock necessary to keep the blood centre functioning for
The total number of blood units collected during this time 12 days. Three red cell units were discarded due to out-
period were 3452 units over a period of 29 days. The date during this period.
daily red cell stock during time period A was around Due to the effect of the high-volume drive the RDP
1058 units (IQR = 985–1103) with a daily red cell collec- stock also increased to 223 units (IQR = 163–269) though
tion of around 116 units (IQR = 93–159) and a daily red the RDP prepared daily during this time period remained
Fig. 1 Box plot showing a. Red Blood Cell (RBC) stock, collection and issue b. Platelet concentrate (RDP, Random donor platelet) stock, collection and
issue before and during COVID-19 pandemic.
almost same at 98 units (IQR = 76–126). The RDP issued red cells also decreased and as low as eight units was col-
daily decreased to 106 units (IQR = 85–120) (Fig. 1b). As lected on one of the days. Blood was transferred in and
a result, the average buffer stock of RDPs almost doubled out of the blood centre to/from other centres to maintain
to 120 units equivalent to a stock required for 1 day. sufficient stock at all times and to prevent the units from
expiring (Fig. 2).
In contrast to the red cell stock, the RDP stock was
Time period C – effect of the COVID-19 pandemic
maintained only for an initial period of 7 days after
on the blood supply and demand
which the stock decreased drastically and was maintained
The total number of blood units collected during this time at around 50 units daily. The reduction in the demand
period were 523 units over a period of 24 days (Fig. 2). was proportionate to the reduction in blood collection
There was a drastic fall in the daily red cell collection to which was not the case for red cells. The blood centre
around 26 red cell units (IQR = 12–30) which was 1/6th also collected additional single donor platelets from vol-
to 1/9th of the collection during the time period A and B untary blood donors who were called by the blood centre
respectively. On the other hand, the red cells issued dur- in anticipation of the decreased platelet supply (not
ing this time period decreased to around 69 red cell units shown in diagram).
(IQR = 61–76) (Fig. 1). Thus, the daily red cell stock
dropped to around 654 red cell units (IQR = 498–915).
Discussion
There was no discard of blood units due to outdate during
this time period. Maintaining an adequate blood inventory is a dynamic
Similarly, the RDP prepared and issued also decreased process and requires that blood centres keep an eye on
to 23 units (IQR = 10–29) and 21 units (IQR = 16–36) various aspects of blood inventory management. Blood
respectively with a daily stock of 73 (IQR = 61–99) units being a perishable product with a limited shelf life (for
(Fig. 3). RBCs with SAGM – 42 days, RDP – 5 days) pose a num-
The buffer stock of red cells decreased gradually from ber of challenges for inventory management. A low
1233 red cell units on the day when lockdown started to inventory exposes to an acute shortage during times of
317 red cells at the end of study period. Five blood dona- disaster whereas an overfilled inventory may lead to
tion drives needed to be cancelled in view of the lock- expiry of the blood components. The risk of expiry is
down and no further blood donation drives were more pronounced with platelet concentrates as the expiry
scheduled during the period of study. The collection of period is just 5 days.
Fig. 2 Dynamics of red cell stock during COVID-19 lockdown. Arrows away from the buffer stock represent red cell units issued to other centres, arrows
towards buffer stock represent red cells received from other blood centres and broken arrows towards buffer stock represent cancelled off-site blood
donation drive along with the expected number of units to be collected as strikethrough numbers.
Fig. 3 Dynamics of RDP stock during COVID-19 lockdown. Arrows towards buffer stock represent RDP received from other blood centres with number
of units in boxes.
COVID-19 outbreak resulted in disruption of various resulted in reduced movement of individuals and thus
aspects of the blood supply dynamics. The lockdown reduced the availability of the blood donors. The fear of
imposed by the government to contain the spread of virus acquiring infection during commute to the blood centre
and during the process of blood donation also added to day use and we rationed the platelet inventory strictly
the reduced number of blood donors presenting to the to therapeutic use in bleeding patients as well as pro-
blood centre in the initial period of lockdown. It also led phylactically when the platelet counts were <10 000/µl.
to cancellation of the already planned blood donation All the prophylactic transfusion requests were accepted
drives due to restrictions in movement of blood bank staff when a haematologist had been consulted for the same.
to the site of blood donation drive and restrictions on In addition, voluntary donors from the staff were moti-
public gathering. These all led to a sharp fall in the blood vated to donate SDP to be used in case RDPs were not
collection during initial stages of lockdown which available.
reduced further as the lockdown progressed. Although maintaining a blood inventory is a dynamic
Outpatient services as well as hospital admissions for process and is affected by a number of factors, following
non-emergent services were also curtailed to prevent the the basic principles of inventory management and disas-
spread in community in the initial period of lockdown. ter planning is a pre-requisite to tide over crises caused
However, the reduction in demand and supply was not by COVID-19 like situations. Formulating and imple-
proportionate as evident at our centre. This was due to menting local mitigation strategies and following them
the fact that the transfusion requirements of routine strictly can go a long way in maintaining blood inven-
surgical patients form only 20% of our blood requests tory during a pandemic [12]. The present study demon-
and almost 40–50% of patients are being transfused at strates that maintaining a buffer stock of blood and
our centre at the emergency department or for transfu- blood components, strict adherence to the transfusion
sion dependent anaemia such as thalassaemia or anae- triggers, good coordination with the clinical staff and a
mia post-chemotherapy. An additional unplanned prospective review of blood transfusion requests to
increase in transfusion requirement occurred due to ensure rational blood transfusion were some of the steps
shifting of the patients from trauma care hospital which helped us to successfully maintain transfusion
attached with our institute to the main hospital. The requirements in the initial phases of the COVID-19 pan-
trauma centre used to have a separate blood centre in demic. In addition to these, our experience also high-
pre-COVID times whose services were now diverted to lights the importance of maintaining a registry of
complete the transfusion requirements of COVID patients voluntary blood donors from around the neighbouring
due to the transformation of trauma centre into a areas of blood centre. In our case, it were the staff mem-
COVID-care hospital. bers who were easily accessible, and the registry helped
An adequate buffer stock equivalent to 1–2 weeks us in coordinating with them to maintain our inventory
requirement of red cells helped us in maintaining emer- especially for completing the requests of platelet concen-
gent transfusions despite sharp decrease in blood supply. trates or some of the O negative transfusions required
To prevent unnecessary wastage of blood due to further for intrauterine/ exchange transfusions. To conclude
decrease in the requirement in later stages of lockdown planning is an important part of blood inventory man-
we strictly implemented the policy of FIFO with exception agement which provides time to adapt and plan further
only for intrauterine transfusions and exchange transfu- even in case of future pandemics.
sions. There was no increase in discard of blood due to
outdating despite reduction in the requirement from the
Acknowledgement
non-COVID-19 times and having a higher buffer stock
due to strict implementation of FIFO policy. This was in None.
contrast to some reports where blood centres reported
increased discard of blood components due to decreased
Conflicts of interest
requirements during the lockdown [11]. Major user
departments were requested to adhere to the rational use The authors declare that they have no conflict of interest
of the blood components and a prospective review of regarding the submitted article.
each blood request was started. An adequate buffer stock
also provided us enough time to plan activities and mobi-
Author contribution
lize voluntary blood donors from our internal registry
which mainly comprised of institute staff. HCP: Designed the study, analysed data, wrote the manu-
In contrast to the red cell inventory, it is the platelet script and approved the final version. PC: Designed the
inventory which pose a major challenge due to a lim- study and revised the manuscript. CCS: Data collection
ited shelf life of 5 days, it is not possible to maintain a and analysis. PJA: Wrote the manuscript and revised the
sufficient buffer stock. At the start of lockdown, we had manuscript. KK: Wrote the manuscript and revised the
maintained a stock which was sufficient for a single manuscript.
References
1 CDSCO, Ministry of Health and Family Blood Banks, available at http://nac 8 Shan H, Zhang P: Viral attacks on the
Welfare, Government of India. Regula- o.gov.in/sites/default/files/MoHFW% blood supply: the impact of severe
tory requirements of blood and/or its 20Gazette%20Notification%20-% acute respiratory syndrome in Beijing.
components including blood products. 20Dt.%2003rd%20April%202017.pdf Transfusion 2004; 44:467–9
Available at https://cdsco.gov.in/openc [Last accessed on 2020 June 08]. 9 Teo D: Blood supply management dur-
ms/opencms/system/modules/CDSCO. 5 Yahia AIO: Management of blood sup- ing an influenza pandemic. ISBT Sci
WEB/elements/download_file_division. ply and demand during the COVID-19 Ser 2009; 4:293–8
jsp?num_id=MzY3 [Last accessed on pandemic in King Abdullah Hospital, 10 Kwon SY, Lee EH, Kim HS, et al.: Mid-
2020 June 08]. Bisha, Saudi Arabia. Transfus Apher dle East Respiratory Syndrome Coron-
2 National Aids Control Organisation, Sci 2020:102836. http://dx.doi.org/ avirus (MERS-COV) outbreak in South
Report on National Estimation of Blood 10.1016/j.transci.2020.102836. [Epub Korea: risk management at the Korean
Requirement in India. Available at ahead of print]. Red Cross Seoul Nambu Blood Centre
http://naco.gov.in/sites/default/files/ 6 Leung JNS, Lee CK: Impact of the (Oral abstracts). Vox Sang 2015; 109
Final%20Estimation%20Report%20of% COVID-19 – a regional blood centre’s (Suppl. 2):18
20Blood%20Requirement%20in%20Ind perspective. ISBT Sci Ser 2020. http:// 11 Raturi M, Kusum A: The blood supply
ia%20%281%29.pdf [Last accessed on dx.doi.org/10.1111/voxs.12558. [Epub management amid the COVID-19 out-
2020 June 08]. ahead of print]. break. Transfus Clin Biol 2020;
3 Aggarwal R, Prakash A, Aggarwal M: 7 Pagano MB, Hess JR, Tsang HC, et al.: 27:147–151. [Online ahead of print].
Thalassemia: an overview. J Sci Soc Prepare to adapt: blood supply and 12 Stanworth SJ, New HV, Apelseth TO,
2014; 41:3–6 transfusion support during the first 2 et al.: Effects of the COVID-19 pan-
4 MoHFW - Gazette Notification reg. weeks of the 2019 novel coronavirus demic on supply and use of blood for
Transfer of Blood & Blood Compo- (COVID-19) pandemic affecting Wash- transfusion. Lancet Haematol 2020; 7
nents to other Blood Banks & Blood ington State. Transfusion 2020; (10):e756–e764.
Donation Camps by Private Hospital 60:908–911
Background and Objectives Blood safety hinges not just on the scientific ratio-
nale for deferral period but potential donors’ compliance with the prevailing pol-
icy. This study aimed to investigate donors’ awareness, attitudes and compliance
with the two-phased policy implementation of time-limited deferral for men who
have sex with men (MSM) in Hong Kong.
Materials and Methods Three rounds of questionnaire survey were conducted
between July 2017 and June 2019 covering the periods of pre-implementation
(Round A), post-implementation without and with pre-donation questionnaire
revision (Round B and C). Chi-square test and multivariable regression analysis
were performed.
Results Of 3085 donors recruited, 968, 1036 and 1081 completed the surveys in
Round A, B and C, respectively. The non-compliance rate of MSM remained
stable at 06% (3/497), 04% (2/551) and 05% (3/587) among male donors in
Round A, B and C, respectively. Two MSM donors from Round C complying with
the prevailing policy were identified. About two-thirds (607%) of respondents
from Round B and C were unaware of the policy change. Overall, over 80% were
either neutral or positive about the change.
Conclusion Our study showed a consistently low non-compliance rate of MSM
over the three periods. The generally high level of acceptance of time-limited
deferral among donors lends support to science-based policy development to pro-
tect blood safety. The identification of compliant MSM donors suggests that the
12-month deferral is effective and acceptable to MSM. With a deferral period far
Received: 28 April 2020,
exceeding the window period, it is a step towards a more equitable policy.
revised 15 October 2020,
accepted 15 October 2020, Key words: blood donation, blood safety, donor deferral, men who have sex with
published online 16 November 2020 men, risk behaviours, transfusion-transmitted infections.
504
Transitioning to 1-year deferral of MSM donors 505
permanent deferral has been applied globally for decades, policy could minimize non-compliance [8], with similar
which led to numerous legal and ethical debates on per- results cited in a French study [9]. In Canada, however,
ceived discrimination in the community. Without an no change of compliance was seen when the deferral
international consensus on safe deferral period, time-lim- period was reduced to 5-year, then 1-year [10]. In Eng-
ited deferral policy with varied length was introduced in land, Scotland and Wales, following implementation of
some countries, following extensive review on scientific 3-month deferral, assessment of surveillance data showed
evidence in conjunction with increasing sensitivity of that there was no increase in the number of non-compli-
blood screening technology in recent years. Countries like ant HIV positive donors [11]. Understanding factors asso-
Australia, Belgium, Finland, Germany, Norway, Ireland, ciated with non-compliance could inform strategy to
Switzerland and the United States have all implemented reduce its occurrence. A number of studies have reported
12-month deferral policy. Japan stipulated a 6-month that non-compliance could arise from the underestima-
deferral and Taiwan a 5-year deferral. New Zealand incre- tion of risk linked with condom use [9,12], not perceiv-
mentally changed the deferral period from 10 years to ing one’s sexual behaviour as risky [8,9,13], confidence
5 years (in 2009) and to 1 year (in 2014). Some nations, in blood screening in detecting infection [9,12], concerns
for example, England, Scotland, Wales and Canada, have about confidentiality [12], and considering it unfair if
further shortened from 12 to 3 months’ deferral, while excluded from donation [9,12]. In Hong Kong, the per-
the Netherlands, Denmark and France have reduced to manent deferral policy has been in place since 1985,
4 months. In 2020, a decision of reducing the 1-year which was recently replaced by a newly introduced 12-
deferral period to 3 months since the last male-to-male month deferral period. In this study, we investigated the
sex was made both in the United States and Northern Ire- non-compliance of MSM donors following the implemen-
land. The policy change was a response to blood shortage tation of the new policy, and assessed donors’ awareness
caused by the coronavirus disease 2019 (COVID-19) in and attitudes towards the new policy. Results of the
the United States [2], and following recommendation by study could be useful for supporting the evaluation of
the Advisory Committee for the Safety of Blood, Tissues MSM donor deferral policy, improving its implementa-
and Organs in Northern Ireland [3]. tion and informing further policy development for
Despite the availability of highly effective HIV screen- enhancing blood safety.
ing, there remains the probability of infections from
donated blood escaping detection during the window
Men who have sex with men deferral policy
period. In this connection, self-deferral of donors with
change in Hong Kong
recent risk of HIV exposure could function as a comple-
mentary strategy to protect blood safety. Non-compliance In Hong Kong, blood donation is exclusively managed
of HIV-infected donors would potentially lead to by the Hong Kong Red Cross Blood Transfusion Service
increased risk of transfusion-related infections. The (BTS). All potential donors are requested to complete a
change of deferral period targeting male donors with his- health screening questionnaire (HSQ) with items on
tory of sex with another man (MSM) (hereafter referred deferrable behaviours including male-to-male sex. The
as MSM donors) was in line with the strategy of restrict- policy of permanently deferring MSM donors was
ing donation by the length of window periods instead of replaced by time-limited deferral in two phases. In the
permanently. Blood safety hinges however not just on first phase, the change of practice from permanent to
scientific rationale for selected deferral periods but one-year deferral starting from 25 September 2017 was
potential donors’ compliance with the prevailing policy, announced through the media, without alteration of the
and that the implementation of any new deferral policy administered HSQ. Prospective male donors indicating a
needs to be done with caution to ensure communication ‘Yes’ answer for the question of ‘If you are male, have
is appropriate and compliance is high. International posi- you ever had oral or anal sex with a man?’ was inter-
tion papers recognize the importance of donor compli- viewed by a nurse in donor centre to confirm if sex had
ance in the efficacy of deferral policy [4], and the use of occurred in the preceding 12 months. An MSM was
donor behavioural-based screening as an additional layer allowed to donate blood if his last sexual activity took
to protect blood safety [5]. Studies from the United place over one year ago. Starting from 1 January 2019,
States examining non-compliance rate in blood donors Phase 2 was implemented with a change of wording in
showed that a proportion of MSM disclosed donating the HSQ. Eligibility determination has since been framed
blood in spite of the implementation of the permanent by the question of ‘In the past 12 months - if you are
deferral policy [6–7]. Recent study in Australia showed male, have you had oral or anal sex with a man?’ An
high compliance of MSM with the time-limited deferral MSM would be eligible to donate blood if a ‘No’ answer
policy, suggesting that perceived equitability of the new is indicated.
Table 1 Characteristics of donors in Round A, B and C the respondents from these two rounds were not aware of
the policy change that has already been implemented.
Round A Round B Round C
Multivariable regression analysis shows that males (ad-
(N = 968) (N = 1036) (N = 1081)
justed odds ratio [aOR]: 137, 95% CI: 114–163), non-
N % N % N % student donors (aOR: 144, 95% CI: 105-198), having
attained post-secondary education or above (aOR: 141,
Male 497 513 551 532 587 543 95% CI: 117-171), and repeat donors (aOR: 183, 95%
Age
CI: 131–256) were significantly associated with a higher
16–29 437 451 385 372 358 331
level of awareness of policy change. Of the seven MSM
30–39 231 239 290 280 258 239
donors identified in Round B and C, all except one
40–49 176 182 201 194 247 228
50 or over 124 128 160 154 218 202
claimed to have read the HSQ before donation, four
Permanent residents 960 992 1023 987 1066 986 reported that they were aware of the change.
Occupation status Donors’ attitudes towards policy change were evaluated
Student 175 181 134 129 120 111 with data from the surveys as shown in Fig. 1. Overall,
In employment 686 709 812 784 865 800 more than 80% were either neutral or positive about pol-
Not employed 107 111 90 87 96 89 icy change at the three rounds of survey. A higher pro-
Education portion (about half) held positive attitude at Round A, as
Secondary or below 390 403 348 336 404 374 compared with that in Round B and C. Positive attitude
Post-secondary 218 225 229 221 213 197
was significantly associated with age <30 (aOR: 231,
Tertiary or above 360 372 459 443 464 429
95% CI: 162–328) (Table 2). Reasons for supporting or
Repeat donor 881 910 944 911 972 899
opposing the policy change among donors are shown in
Types of donor centrea
Large centresb 725 749 736 710 748 692
Table 3. Two of the most common reasons for supporting
Small centresb 243 251 300 290 333 308 the policy change after implementation were its capability
of ‘reducing stigmatization’ (469%), and ‘effectively
a
Donor centres that have enrolled 20 000 or more donors were classified reducing the number of HIV positive donors’ (451%).
as ‘large’ contrasting ‘small’ centres with less than 20 000. An apheresis Before policy implementation (Round A), the main rea-
centre was excluded. sons for the opposition were its ‘inability to reduce
b
Four and 3 donor centres were included in the category of ‘larger cen-
stigmatization’ (537%), followed by ‘lack of supporting
tres’ and ‘smaller centres’, respectively.
scientific evidence’ (443%) and ‘inability to reduce the
*P < 005.
number of HIV positive donors’ (409%). After implemen-
**P < 001.
***P < 0001; X2 = chi-square statistics.
tation, the main reasons became ‘inability to reduce the
number of HIV positive donors’ (662%). Such concern
was more prevalent in females (752%) than males
497, 95% CI: 000–128) in Round A, 04% (2/551, 95% CI: (592%). Around three-quarters of respondents (755%;
000–087) in Round B and 05% (3/587, 95% CI: 000–109) 800/1060) did not give any comments because of ‘unfa-
in Round C. Two male donors from Round C reported having miliarity with the issue’. Within the subgroup of those
male-to-male sex over 12 months ago and were therefore expressing ‘unfamiliarity with the issue’, 712% (274/385)
classified as being compliant with the prevailing policy of in Round B and 760 % (315/415) in Round C responded
deferral. Among the eight non-compliant MSM, half (4/8) that they were not aware of the policy change.
reported consistent condom use for sex. Five had only one As shown in Fig. 2, the preferred interval of deferral
male partner in the preceding 12 months, while the rest varied considerably among all donors. The plurality of
reported multiple male partnership (2 with 2–5 partners, and donors did not offer any views (277%, 387%, 445%).
1 with over 10 partners). There were a few inconsistent The most preferred interval of donors was one year
answers regarding sex partnership including 3 who reported (202%, 174%, 15%), while a consistent proportion of
having male partners which could not be clearly defined as <10% preferred ‘permanent deferral’ (79%, 87%, 91%).
either regular or non-regular. About half of the donors (47%) were uncertain about the
length of the window period for HIV tests, while most
(377% of the remainder) perceived that it should be
Donors’ awareness of policy change and their
between 3 and 6 months (Fig. 3). By dichotomising the
attitudes
perceived length of window period and preferred deferral
Donors’ awareness of the change of policy from perma- interval using 3 months and 1 year as the respective cut-
nent to 12-month deferral was examined in Round B and off, a significantly higher proportion of those with a long
C of the surveys. About two-thirds (607%; 1286/2117) of perceived window period preferred longer deferral
Table 2 Factors associated with the awareness and attitudes on policy change in Round B & C
a
0 = Unaware, 1 = Aware.
b
0 = Negative, 1 = Positive (excluding the 1060 responses of ‘Neutral’).
*P < 005.
**P < 001.
***P < 0001.
interval (X2 = 7742, P = 0005). Results of spearman cor- about the deferral policy, which was increasingly reported
relation also showed that the perceived length of window in the media, though the effect of having more men
period was only weakly correlated with the preferred becoming eligible for blood donation may play a role. A
deferral interval (r = 0183, P < 0001). high level of compliance is important in enforcing donor
deferral policy such that the residual risk of HIV infection
from transfusion can be minimized. A modelling study
Discussion
predicted a small increase in risk with a change from life-
In this study, we compared the non-compliance rate of time to 12-month deferral for MSM in the UK if compli-
MSM towards donor deferral before and after the imple- ance was unchanged [18]. Studies in United States and
mentation of the 12-month deferral policy in Hong Kong, Canada [19], and Italy [13] provided evidence of
a city in the Asia Pacific region where similar studies are increased safety as a result of improved compliance fol-
scarce. Using a unified recruitment method and a stan- lowing the adoption of a time-limited deferral policy,
dard questionnaire, we showed a consistently low non- despite the theoretical risk of ‘relaxing’ the criteria. Addi-
compliance rate among MSM at <1% both in the pre-im- tionally, authors of an Australia study concluded that
plementation and 2 post-implementation periods of the there was no increase in prevalent infection after compar-
new policy, compared to previous local studies with ing the prevalence of HIV from donors before and after
results ranging from 12–22% [14–16], and similar to the 12-month deferral policy [20]. Our results gave a
that of 023% in Australia [8], and 08% to 10% in same level of compliance with deferral criteria among
Canada [17]. The slightly lower non-compliance rate MSM, both before and after the policy change, suggesting
could be related to the increased awareness of MSM that such change is unlikely to impact blood safety
Pre-implementation Post-implementation
(Round A) (Round B & C)
unfavourably. Despite the small number of MSM donors deferral of MSM in Hong Kong. Some donors, such as
identified, their self-reported inconsistency of condom females, those with lower educational levels were less
use pointed to the needs for safer sex promotion. knowledgeable about the change of deferral policy. The
One small yet notable finding in our third survey fol- time period for collecting data for Round B and C was
lowing implementation of the 12-month deferral of MSM around 3–4 months after Phase 1 and 2 policy implemen-
was the identification of two male donors with male-to- tation, which was relatively short in length and may
male sex over 12 months before donation. This study did explain the low level of awareness recorded as it takes
not inquire about the timing of sex before their previous time for potential donors to be aware of the changes.
donations when the permanent deferral policy was in Although the new policy primarily affects selection of
place, but the fact that both were repeat donors with past donors from the MSM community, there is a need for
history of multiple donations did imply that they had publicity and education to enhance public confidence on
escaped the scrutiny of the permanent deferral mecha- blood safety and ensuring good communication with
nism in the past. Their detection at the second phase of donors to encourage self-deferral. Our results suggested
policy change could be by chance, but might also be that females were more likely than males in having con-
explained by their awareness of the new policy and there- cern of the new policy’s inability to reduce the number of
fore their willingness to self-declare in confidence. HIV positive donors. One possible explanation may be
Should they present for donation during the first phase, that females were less aware of the policy change com-
one’s sexual history would be collected at face-to-face pared to their counterparts and that may have potentially
interview after indicating a history of male-to-male sex led to misconception. Overall, the proportion of respon-
on the HSQ. The specific inclusion of the deferral period dents holding positive attitudes outnumbered those hold-
on the HSQ in the second phase eliminated the need for ing negative attitudes towards the policy change.
follow-up interview – a procedure that is subject to However, we also question whether attitudes might have
stigma. The stigma attached to disclosure of one’s MSM shifted over time as indicated by our data, with a slight
status might have led to the common practice of non- reduction in the proportion of those expressing support,
compliance [21–22]. The perception of permanent deferral and an increase of those holding neutral attitude after
as unjust discrimination [12] and the need to disclose implementation. The shift of the proportion of positive
male-to-male sex even for a single event [23] could have attitudes in Round B and C might be due to the smaller
led to higher rate of non-compliance with the permanent number of young donors, as younger age was associated
deferral than 12-month deferral policy. Enrolling poten- with positive attitude towards the policy. As ‘inability to
tial donors who are likely to increase their compliance reduce the number of HIV positive donors’ emerged as
because of the new policy provides a basis for policy the major concern after implementation of policy change,
enforcement aimed at blood safety. some donors remained unassured if blood safety could be
Clearly, our data shows a lack of awareness among maintained or improved. It is important to explain to the
donors even after the implementation of the 12-month wider community that the policy change represented only
one aspect of ensuring blood safety, amidst all procedures tested HIV positive per year was small, and there was a
including state-of-art donor screening. declining trend in the last two years, the phenomenon of
In countries like Japan and South Africa, the intervals which was probably unrelated to the implementation of
for deferring MSM has further been shortened to the new donor deferral policy. These small numbers (4 in
6 months [5], and in the UK decreased to 3 months with 2017, 2 in 2018 and 0 in 2019) did not allow us to draw
results suggesting no additional risk to blood safety [11]. correlation with the donor deferral policy and its changes.
The adoption of a shorter deferral period could be seen as Tracking the number of HIV positive MSM donors could
a realistic direction now that universal application of test- nevertheless contribute to evaluate and inform the impact
ing methodology like nucleic acid testing (NAT) has nar- of policy changes for achieving blood safety.
rowed the window period to a week or less, which has
minimized recipients’ risks of acquiring transfusion-trans- Conclusion
mitted HIV infection [24–25]. In our study, the preferred
This study showed a consistently low non-compliance
interval of deferral was 1 year. One possible explanation
rate among MSM at <1% both in the pre-implementation
could be the donors’ lack of awareness of window period
and 2 post-implementation periods of the new time-lim-
and thus higher acceptance of the length of 1-year for
ited deferral policy. The survey results showed that the
deferral to protect blood safety. The correlation between
adopted questions appeared to differentiate non-compli-
the length of the perceived window period and preferred
ant and compliant donors per the 1-year deferral policy.
deferral interval was however weak. Further investigation
The generally high level of acceptance of time-limited
is needed to determine whether the public’s awareness of
deferral among donors lends support to science-based
the length of the window period would improve the
policy development to protect blood safety. Communica-
acceptability of further shortening the deferral period as
tion of the deferral criteria and its rationale is also related
regards male-to-male sex. Separately, reduction of defer-
to public acceptability of the change of deferral policy.
ral period plus improved compliance would be beneficial
Our study however showed a low level of policy aware-
in expanding donor pool on top of ensuring safety of
ness among donors. Improved communication of the pol-
blood transfusion. It is important to conduct further
icy and its rationale would be important in increasing
review of deferral criteria while taking into account the
public confidence and sustaining their acceptance towards
updated epidemiological evidence both at the global and
policy change. The identification of compliant donors
local level in the refinement of deferral policy.
with history of male-to-male sex suggests that the 12-
month deferral is effective and acceptable to MSM. With
Limitations a deferral period far exceeding the window period, it is a
step towards a more equitable policy.
This study has some limitations. Firstly, the study was
based on a survey of blood donors who were selected
based on their availability and time-location sampling.
Acknowledgements
Donors recruited by BTS’s mobile services which visit other The authors thank the staff of the Hong Kong Red Cross
locations in the territory have not been included. While Blood Transfusion Service for supporting the conduction
the recruitment of donors was done targeting an estimated of the survey. This work was supported by a grant from
proportion of attenders by centre, the ultimate distribution the AIDS Trust Fund (ATF MSS261R). Li Ka Shing Insti-
remained highly variable by age, employment status and tute of Health Sciences, The Chinese University of Hong
education level. Generalizability of the results to the entire Kong, is acknowledged for rendering technical support in
blood donor populations in Hong Kong should be cau- conducting the studies.
tioned. The lack of standardization by sample weighting
may also explain the differences in attitudes, such as dif- Author contributions
ferent compliance response proportions, across the three
rounds of the survey. Secondly, all data collected were S-SL: conceptualizing the study. JYCL and C-PC: acquisi-
self-reported that could have potentially carried socially tion and analysis of data. JYCL: drafting the article. All
desirable bias. The use of tablet computer, however, should authors: revising it critically and giving final approval of
have minimized such bias by the sense of privacy and the version to be published.
confidentiality provided. Donors might feel easier to
answer truthfully when asked questions related to sensitive Data Sharing and Data Accessibility
topics such as sexual behaviours and illicit drug injection,
as the staff of the donor centres had no access to the com- All data generated or analysed during this study are
pleted questionnaire. Finally, the number of MSM donors included in this published article.
References
1 World Health Organisation (WHO): selection criteria – Complidon 2017, 18 Davison KL, Conti S, Brailsford SR:
Blood safety and availability. Avail- France. Transfusion 2020a; 60a:73–83. The risk of transfusion-transmitted
able from: https://www.who.int/news- 10 O’Brien SF, Osmond L, Fan W, et al.: HIV from blood donations of men
room/fact-sheets/detail/blood-safety- Compliance with time-based deferrals who have sex with men, 12 months
and-availability. [Last accessed 2 for men who have sex with men. after last sex with a man: 2005–2007
September 2019] Transfusion 2019; 59:916–920 estimates from England and Wales.
2 U.S. Food & Drug Administration 11 GOV.UK. Public Health England. Safe Vox Sang 2013; 105:85–88
(FDA). FDA Statement. Coronavirus supplies 2018: Monitor, inform pro- 19 Germain M, Remis RS, Delage G: The
(COVID-19) update: FDA provides gress. Available from: https:// risks and benefits of accepting men
updated guidance to address the www.gov.uk/government/publications/ who have had sex with men as blood
urgent need for blood during the pan- safe-supplies-annual-review/safe-sup donors. Transfusion 2003; 43:25–33
demic. Available from: https:// plies-2018-monitor-inform-progress 20 Seed CR, Kiely P, Law M, et al.: No
www.fda.gov/news-events/press-an [Last accessed 10 June 2020] evidence of a significantly increased
nouncements/coronavirus-covid-19-up 12 Grenfell P, Nutland W, McManus S, risk of transfusion-transmitted human
date-fda-provides-updated-guidance- et al.: Views and experiences of men immunodeficiency virus infection in
address-urgent-need-blood-during-pa who have sex with men on the ban on Australia subsequent to implementing
ndemic [Last accessed 9 June 2020] blood donation: a cross sectional sur- a 12-month deferral for men who
3 Department of Health. Rules on blood vey with qualitative interviews. BMJ have had sex with men (CME). Trans-
donation set to change. Available 2011; 343:d5604 fusion 2010; 50:2722–2730
from: https://www.health-ni.gov.uk/ne 13 Suligoi B, Pupella S, Regine V, et al.: 21 Duquesnoy A, Danic B, Santos A,
ws/rules-blood-donation-set-change Changing blood donor screening crite- et al.: Context and social perceptions
[Last accessed 31 Aug 2020] ria from permanent deferral for men of blood donation in donors found
4 Goldman M, Shih AW-Y, O’Brien SF, who have sex with men to individual positive for human immunodeficiency
et al.: Donor deferral policies for men sexual risk assessment: no evidence of virus in France. Transfusion 2017;
who have sex with men: past, present a significant impact on the human 57:2240–2247
and future. Vox Sang 2018; 113:95–103. immunodeficiency virus epidemic in 22 Levy I, Olmer L, Livnat Y, et al.: Atti-
5 Benjamin RJ, Bianco C, Goldman M, Italy. Blood Transfus 2013; 11:441–448 tudes, perceptions and knowledge
et al.: Deferral of males who had sex 14 Lee SS, Cheung EKH, Leung JNS, among men who have sex with men
with other males. Vox Sang 2011; et al.: Non-compliance to infectious towards the blood donation deferral
101:339–367 disease deferral criteria among Hong policy in Israel. PLoS One 2017; 12:
6 Belanger GA, McFarland W, Raymond Kong’s blood donors. Vox Sang 2017; e0170364
HF, et al.: If the permanent deferral 112:425–433 23 Sanchez AM, Schreiber GB, Nass CC,
were lifted would men who have sex 15 Wong HTH, Lee SS, Lee CK, et al.: et al.: The impact of male-to-male
with men want to donate blood, and if Failure of self-disclosure of deferrable sexual experience on risk profiles of
so, who would be eligible? Transfusion risk behaviors associated with transfu- blood donors. Transfusion 2005;
2013; 53:2729–2733 sion-transmissible infections in blood 45:404–413
7 Liszewski W, Terndrup C, Jackson NR, donors. Transfusion 2015; 55:2175– 24 Roth WK, Busch MP, Schuller A,
et al.: The beliefs and willingness of 2183 et al.: International survey on NAT
men who have sex with men to com- 16 Lee CK, Lee KC, Lin CK, et al.: Donors’ testing of blood donations: expanding
ply with a one-year blood donation perspectives on self-deferral of men implementation and yield from 1999
deferral policy: a cross-sectional study. having sex with men from blood to 2009. Vox Sang 2012; 102:82–90
Transfusion 2017; 57:2234–2239 donation. Transfusion 2013; 53:2441– 25 Kleinman S, Busch MP, Korelitz JJ,
8 Seed CR, Lucky TT, Waller D, et al.: 2448 et al.: The incidence/window period
Compliance with the current 12-month 17 Goldman M, Yi QL, Ye X, et al.: Donor model and its use to assess the risk of
deferral for male-to-male sex in Aus- understanding and attitudes about transfusion-transmitted human
tralia. Vox Sang 2014; 106:14–22 current and potential deferral criteria immunodeficiency virus and hepatitis
9 Sauvage C, Spinardi R, Pelat C, et al.: for high-risk sexual behavior. Transfu- C virus infection. Transfus Med Rev
Noncompliance with blood donor sion 2011; 51:1829–1834 1997; 11:155–172
Abstract
Background and objectives Many Western countries face a shortage of African
blood donors, while their specific blood groups are needed to transfuse chronic
transfusion patients of similar ethnic background. Blood donation awareness and
attitudes greatly impact the decision to become a blood donor, but how they are
related and differ across ethnic groups is understudied. This study investigated
blood donation awareness and attitudes of individuals of Dutch and African des-
cent in the Netherlands.
Materials and methods Survey data of 257 African and 152 Dutch non-donors
measuring donation awareness (i.e. being familiar with the Dutch blood bank
organization and knowing others who donated blood), cognitive (evaluative
judgements) and affective (emotional reactions) attitudes were included. t-Tests,
chi-square tests, linear and logistic regressions were conducted to study differ-
ences and associations between donation awareness and attitudes.
Results African individuals were less often aware of the Dutch blood bank orga-
nization (43%; p < 005) or others who donated blood (51%; p < 005) than
Dutch individuals (55% and 68%, respectively). African individuals had lower
cognitive donation attitudes compared with Dutch individuals (p < 0001), but no
differences were found for affective attitudes (p = 055). High donation awareness
was associated with higher cognitive (p < 0001) and affective (p < 005) dona-
tion attitudes among African minorities, but not among Dutch individuals.
Conclusion The lower donation awareness and cognitive attitudes of African
minorities should be taken into consideration in donor recruitment. Raising
Received: 19 June 2020,
awareness through effective communication strategies might be essential in the
revised 14 October 2020,
accepted 21 October 2020,
donor decision making process of this target group.
published online 7 November 2020 Key words: donor motivation, donor recruitment, donors.
Introduction
Blood donors of African descent are greatly underrepre-
sented in many Western high-income countries [1]. This
Correspondence: Elisabeth F. Klinkenberg, PhD-student at Department presents an important healthcare problem as blood groups
of Donor Medicine Research, Sanquin Research, Plesmanlaan 125, 1066 are different across ethnic groups. Especially, people of
CX Amsterdam, The Netherlands Sub-Saharan African/Black descent have extended
E-mail: l.klinkenberg@sanquin.nl
This is an open access article under the terms of the Creative Commons Attribution License, 513
which permits use, distribution and reproduction in any medium, provided the original work is properly cited.
514 E. F. Klinkenberg et al.
behaviour [22–25] and seem to play an important part background and university students, residing in the
particularly in the Precontemplation stage in behaviour Netherlands [29,30]. Participants in the Motive-study
change [18,26]. There is evidence that African minorities were recruited among the general population through an
have a slightly more negative attitude towards blood ISO-certified research company (Panelclix) and among
donation compared to White majority populations in students through Tilburg University. Participants recruited
Western countries [1]. Various studies in the United States through Panelclix were compensated with ‘Clix’, which
showed that African Americans less often believe it is can be traded for a small amount of money. Participants
important to donate blood (e.g. because of the perception recruited through the university were compensated with
that their blood is not wanted or will be wasted) and that course credits. Next to data originating from the Motive-
they are generally more afraid of needles and pain as study, additional data were also collected in 2018 among
compared to White Americans [1,27,28]. In our qualita- social media users (regardless of ethnic background) using
tive interview study among African migrants living in the a shortened version of the survey used in the Motive-
Netherlands, the overall general opinion towards giving study. An overview of the recruitment of the samples can
blood was positive, but participants reacted more hesi- found in Table 1. Adequate Dutch or English language
tantly when asked about their own attitude to donate proficiency was a requirement for participation. Social
blood [10]. media users were recruited predominantly via Facebook
It is expected that it could be beneficial to design and WhatsApp, using convenience- and snowball-sam-
recruitment campaigns and their target group with the pling. The surveys were shared on the social media
knowledge from the TTM and TPB in mind. Based on account of the Dutch blood bank organization Sanquin
these models and the literature, we expect that an assess- and also via Promoted Posts to reach social media users
ment of the level of and interplay between awareness and who do not ‘like’ or follow this Facebook page. Partici-
attitudes in our potential donor recruitment target group pants recruited through social media were compensated
will benefit the development of future donor recruitment with a voucher for a large online store. To our knowl-
strategies. Partly due to being unaware, individuals with edge, no recruitment campaigns aimed at the target group
low awareness or in the Precontemplation stage may that could interfere with the results were rolled out at the
experience resistance to behaviour change or believe the times of data collection.
costs of behaviour change are higher than the benefits For the analyses in the present study, we selected par-
[26]. Therefore, when the target audience becomes well ticipants between the ages of 18 and 65, being the age
informed, the gained awareness might alleviate cognitive range individuals can register as a blood donor in the
and affective unfamiliarity surrounding blood donation Netherlands. We only included participants who had
and act as a prerequisite for positive attitude change and never donated blood but were not definitely excluded for
a move to the next step in the TTM. blood donation and who were either of Dutch or African
Based on the studies and mechanisms discussed above, ethnic background. In total of the initial 672 participants
the following hypotheses were developed: combined, 3 (04%) were excluded for being either too
H1: African ethnic minorities have a lower blood dona- young or too old for this study’s purpose, 65 (97%) were
tion awareness compared with people of Dutch descent in excluded for being of different ethnic background than
the Netherlands. African or Dutch, 133 (198%) for having donated blood
H2: African ethnic minorities have a lower blood dona- previously, 23 (34%) for being definitely excluded to
tion attitude compared with people of Dutch descent in donate blood and 39 (58%) because of missing values on
the Netherlands. at least one of the variables included in the analyses of
H3: A higher blood donation awareness is associated this study. The final sample included 257 people with a
with higher blood donation attitudes for African ethnic Sub-Saharan, Afro-Surinamese or Afro-Caribbean (Afri-
minorities and people of Dutch descent. can) background (63%) and 152 people of Dutch back-
ground (37%).
Ethical approval for the study was granted by the Ethi-
Methods
cal Advisory Board of Sanquin and the Ethics Committee
of Tilburg University. The project protocol this study is
Design and participants
part of was also reviewed by the Medical Ethics Review
For this study, we made use of various survey data. The Committee of the Academic Medical Center Amsterdam
Motivations to Give, Donate and Share study (Motive- (now Amsterdam UMC – location AMC), but was waived
study; https://www.sanquin.org/research/donor-studies- from requiring medical ethical approval because the pro-
projects/motive-study/index) consists of stratified online tocol did not fall under the Medical Research Involving
survey data collected in 2018 among people of African Human Subjects Act of the Dutch law.
March–April 2018
Measures
Online, Qualtrics
All ethnicities
Social media
Ethnicity and background characteristics
n = 231
African background in this study refers to individuals of
10 min
n = 97
Dutch
n=2
sub-Saharan descent, which is defined as individuals
Yes
who reported they had at least one parent originating
from sub-Sahara Africa, Surinam or the Caribbean and
who indicated they were of sub-Saharan, African-Suri-
namese or African-Caribbean descent. Dutch background
n = 64
Donation awareness
Donation awareness was measured with two separate
questions. ‘Have you ever heard of the Dutch blood bank
organization, Sanquin?’ and ‘Do you personally know
someone who has donated blood or is currently a blood
donor?’ with ‘yes’ (a score of 1) and ‘no’ (a score of 0)
Panelclix and social media
18–65 years
namely . . .’ option.
4. Eligible age range
3. Type of survey
Type of sample
Donation attitudes
7. Recruitment
8. Language
adjusted from eight weeks in the original scale to than men in both groups, namely 63% (n = 96) in the
12 months, as the blood donation procedure in the Dutch group and 72% (n = 184) in the African group, but
Netherlands is more lengthy compared to donating at the the difference between the two ethnic groups was not sig-
blood drives in the United States (e.g. you cannot donate nificant (X2(1) = 31, p = 008). The overall age was rela-
at your first visit) and a blood donor in the Netherlands tively lower in the Dutch background group (M = 257,
donates on average only once or twice annually [29]. A SD = 105) compared with the African group (M = 359,
factor analysis verified that donation attitudes could be SD = 125; t (3618) = -89, p < 0001). Also, the Dutch
divided into cognitive attitude (evaluative judgements group was significantly higher educated compared with
towards blood donation) (a = 0915) and affective attitude the African group (X2(2) = 358, p < 0001). Among the
(emotional reactions towards blood donation) (a = 0863), African individuals, 48% (n = 124) were first-generation
which both accounted for 750% of the total variance in migrants and 52% (n = 133) were second-generation
the six items. When analysed separately for participants migrants (see Appendix A for the demographic compar-
of African and Dutch background, exactly the same two ison of the three study samples).
factors arose and the Cronbach alphas were still good
(cognitive attitude: a = 0927 African background,
Differences in donation awareness and donation
a = 0852 Dutch background; affective attitude:
attitudes
a = 0888 African background, a = 0821 Dutch back-
ground). Both cognitive attitude (useless/useful, pointless/ Slightly more than half of the Dutch individuals knew
worthwhile and the wrong thing to do/the right thing to Sanquin (55%; n = 83) (Table 3). This percentage was sig-
do) and affective attitude (unpleasant/pleasant, unenjoy- nificantly lower among the African individuals (43%;
able/enjoyable, frightening/not frightening) are con- n = 111) (adjusted odds ratio; AOR = 063, 95% CI [040,
structed from the mean of three 7-point bipolar 099], p < 005). Of those participants who were familiar
statements. On both scales, a score of 0 refers to the low- with Sanquin (n = 194), the Dutch participants (n = 83)
est possible attitude, while a score of 6 refers to the high- most often heard of it through family members (39%;
est possible attitude. n = 32) and the African participants (n = 111) most often
heard of it through friends (23%; n = 26) (Figure 1). A
relatively large part of the African individuals also
Analyses
became familiar with the blood bank through work and/
Statistical software (SPSS, version 23, Chicago, IL) was or colleagues (20%; n = 22). In the ‘Other’ category, indi-
used to examine the descriptive properties of the data viduals of Dutch background most often mentioned that
and test our hypotheses. t-Tests and chi-square tests were knowing Sanquin is part of their general knowledge,
performed to test differences in demographic characteris- whereas individuals of African background most often
tics between the two ethnic groups. The differences in mentioned they knew Sanquin via the Internet. More than
donation awareness and donation attitude between partic- two-thirds of the Dutch individuals knew someone per-
ipants of African and Dutch ethnic background were sonally who has donated blood (68%; n = 104), whereas
tested using logistic and linear regression analyses and about half of the African individuals mentioned this
controlled for age, gender and educational level to find (51%; n = 131) (AOR = 056, 95% CI [035, 090],
answers for hypotheses H1 and H2. Multivariate linear p < 05). Of those participants who personally knew one
regression analyses tested the associations of donation or more other blood donors (n = 235), this was in most
awareness on cognitive and affective attitude. We made cases a friend or acquaintance (63% (n = 66) of the Dutch
two separate (stratified) models for those of African and individuals and 62% (n = 81) of the African individuals)
Dutch descent (H3). These sub-group analyses were con- (Figure 1). In the ‘Other’ category, individuals of African
trolled for age, gender and educational level, plus migrant background mentioned for instance parents in law, the
generation in the African sub-sample. partner and a neighbour. When combining the two dona-
tion awareness measures, we notice that relatively most
individuals of Dutch background belong to the ‘high
Results
awareness’ category (45%, n = 70), while relatively least
individuals of African background fit in this category
Background characteristics of samples and Dutch
(28%, n = 72). H1 can be accepted as less African indi-
and African individuals
viduals were aware than Dutch individuals.
Table 2 shows the sociodemographic background charac- Dutch individuals showed an overall higher cognitive
teristics of the 152 participants of Dutch and the 257 par- attitude (M = 47, SD = 12) compared with African indi-
ticipants of African descent. There were more women viduals (M = 39, SD = 17). After controlling for age,
gender and educational level, this difference remained associated with a higher cognitive attitude (B = 098, t
significant (B = -090, t(403) = -517, p < 0001). (249) = 361, p < 0001), whereas this was not the case
Regarding affective attitude, we found no significant dif- for Dutch individuals who showed no significant associa-
ferences between the Dutch (M = 27, SD = 14) and Afri- tion between awareness and cognitive attitude (B = 036,
can individuals (M = 29, SD = 15; B = 010, t t(145) = 149, p = 014).
(403) = 059, p = 055). Thus, H2 could only be partly Regarding the models for affective attitude, scoring
accepted as there was a significant difference in cognitive high on awareness was significantly associated with a
attitude, but not in affective attitude. higher affective attitude among the African background
group (B = 056, t(249) = 223, p < 005), but not in the
Dutch background group (B = 017, t(145) = 110,
The associations between donation awareness on
p = 59).
donation attitudes
Based on our results, H3 was accepted for the African
Table 4 summarizes the results for the multivariate linear individuals as a higher awareness was associated with
regression analyses on donation attitudes. Among the higher cognitive and affective attitudes, but rejected for
African individuals, high awareness was significantly the Dutch individuals.
Table 3 Differences in awareness and attitudes between Dutch and African individuals (N = 409)
All tests are adjusted for gender, age and educational level. Dutch individuals were the reference group.
*p < 005.
***p < 0001.
Figure 1 Specified blood bank awareness (n = 194) and knowing another donor (n = 235), divided by ethnic background.
Table 4 Multivariate linear regressions on cognitive and affective attitude for individuals of Dutch (n = 152) and African background (n = 257)
Intercept 457 (056) 814 377 (054) 702 314 (072) 435 303(050) 612
Blood donation awareness
Not aware Ref. Ref. Ref. Ref.
Low awareness (knows either the blood bank or blood donor) 031 (025) 125 028 (025) 111 036 (032) 110 008(023) 033
High awareness (knows both the blood bank and blood donor) 036 (024) 149 098 (027)*** 361 017 (031) 054 056(025)* 223
Female 065 (019)*** 343 -001 (024) -003 -002 (024) -008 -002(022) -007
Age in years 001 (001) 079 -001 (001) -074 0004 (001) 032 001(001) 084
Second migrant generation n.a. -011 (023) -048 n.a. -016(021) -076
Educational level
Low Ref. Ref. Ref. Ref.
Middle -045 (050) -090 -002 (035) -006 -081 (064) -127 -055(033) -169
High -085 (048) -179 003 (037) 007 -078 (061) -126 -060(034) -179
distributed online, which can be an obstacle for first-gen- their social networks [9,10,37]. This is essential informa-
eration African migrants. We made this choice because tion for the development of future blood donor recruit-
Dutch or English fluency is required to donate blood in ment interventions.
the Netherlands. Additionally, the blood donor registra-
tion process is also online, thus basic Internet competen-
Acknowledgements
cies are essential for blood donors. The current sample
might not be generalizable to all African migrants living This work was support by Sanquin Research under the
in the Netherlands, but is generalizable to the potential internal grant: PPOC-14-25. We thank all participants for
donor population of African descent. their contribution to this study. JvW and EK conceived
The results of this study have valuable implications for the initial idea of the research. EK, MF and EHitV
future research and blood donor recruitment. Awareness designed the survey. EK led the survey data collection,
raising might be a promising first step to help African conducted the analyses and wrote the manuscript, while
minorities contemplating about blood donation. In the MF, WdK, EHitV and JvW critically reviewed and revised
Netherlands, Sanquin is the only blood bank organiza- the manuscript.
tion. If a person does not recognize the name or logo, he
or she will not notice a blood collection centre nearby
Conflict of interest
and it also makes it difficult to register as a donor, as the
person does not know where to do so. We did find that a The authors declare that they have no competing inter-
higher level of awareness – both being familiar with the ests.
blood bank organization and personally knowing a blood
donor – to have a positive association with attitudes. The
Data availability statement
importance of a social network needs to be highlighted in
this, as most individuals became familiar of blood dona- Data of the Motive-study or the particular data set used
tion through family members and friends, as also was for this study are available upon request by contacting
emphasized in other studies [36,37]. Since ethnic minori- Elisabeth Klinkenberg (L.Klinkenberg@Sanquin.nl). More
ties often have less blood donors in their social network, information on the Motive-study can be found here:
it is of importance to reach the target communities first https://www.sanquin.org/research/donor-studies-projec
as to enable blood donation awareness to spread through ts/motive-study/index
References
1 Klinkenberg EF, Huis In ‘t Veld EMJ, alloimmunization in sickle cell disease: optimizing blood matching strategies:
De Wit PD, et al.: Blood donation bar- pathophysiology, risk factors, and balancing patient complications
riers and facilitators of Sub-Saharan transfusion management. Blood 2012; against total costs incurred. Front Med
African migrants and minorities in 120:528–37 2018; 5:199
Western high-income countries: a sys- 6 Van Dongen A, Mews M, de Kort 9 Piersma TW, Klinkenberg EF: The rela-
tematic review of the literature. Trans- WLA, et al.: Missing Minorities – A tion between blood donor recruitment
fusion Med 2019; 29(Suppl 1):28–41 survey based description of the current and donor diversity and loyalty in the
2 Reid ME, Lomas-Francis C, Olsson ML: state of minority blood donor recruit- Netherlands. ISBT Sci Series 2018;
The blood group antigen facts book, ment across 23 countries. Divers Equal 13:384–93
3rd edn. San Diego (CA), Academic Health Care 2016; 13:138–45. 10 Klinkenberg EF, Huis In ‘t Veld EMJ,
Press, 2002 7 Statistics Netherlands: CBS Statline - De Wit PD,, et al.: Barriers and moti-
3 Weatherall D: The inherited diseases of Bevolking; generatie, geslacht, leeftijd vators of Ghanaian and African-Suri-
hemoglobin are an emerging global en herkomstgroepering, 1 januari. namese migrants to donate blood.
health burden. Blood 2010; 115:4331– 2016. http://statline.cbs.nl/Statweb/ Health Soc Car Community 2019;
6 publication/? 27:748–56
4 Zalpuri S, Zwaginga JJ, le Cessie S, DM=SLNL&PA=37325&D1=0&- 11 Lemmens KP, Abraham C, Ruiter RA,
et al.: Red-blood-cell alloimmuniza- D2=a&D3=0&D4=0&D5=0-5,65,75- et al.: Modelling antecedents of blood
tion and number of red-blood-cell 76,163,251,253&D6=a&HDR=G5,T,G3, donation motivation among non-
transfusions. Vox Sang 2012; G2,G4&STB=G1&VW=T (Last accessed donors of varying age and education.
102:144–9 19-06-2020). Brit J Psychol 2009; 100:71–90
5 Yazdanbakhsh K, Ware RE, Noizat-Pir- 8 Van Sambeeck JHJ, De Wit PD, Luken 12 Trevethan R: Deconstructing and
enne F: Red blood cell J, et al.: A conceptual framework for assessing knowledge and awareness in
public health research. Front Public 21 Breckler SJ: Empirical validation of 30 Klinkenberg EF, Fransen MP, De Kort
Health 2017; 5:194 affect, behavior, and cognition as dis- WLAM, et al.: Blood donation among
13 Gillespie TW, Hillyer CD: Blood donors tinct components of attitude. J Person individuals of African origin in the
and factors impacting the blood dona- Soc Psychol 1984; 47:1191–205 Netherlands: How are barriers and
tion decision. Transfus Med Rev 2002; 22 France JL, Kowalsky JM, France CR, facilitators associated with intention?
16:115–30 et al.: Development of common met- Blood Transfus-Italy 2020. Advance
14 Polonsky MJ, Renzaho AM, Brijnath rics for donation attitude, subjective online publication.
B: Barriers to blood donation in Afri- norm, perceived behavioral control, 31 Statistics Netherlands: Wat verstaat
can communities in Australia: the role and intention for the blood donation het CBS onder een allochtoon?, n.d.
of home and host country culture and context. Transfusion 2014; 54:839–47 https://www.cbs.nl/nl-nl/faq/specifiek/
experience. Transfusion 2011; 23 Ajzen I: The theory of planned behav- wat-verstaat-het-cbs-onder-een-alloc
51:1809–19 ior. Organ Behav Hum Dec 1991; htoon (Last accessed 19-06-2020).
15 Lemmens KP, Abraham C, Hoekstra T, 50:179-211. 32 Statistics Netherlands: Inpassing van
et al.: Why don’t young people volun- 24 De Vries H, Dijkstra M, Kuhlman P: het Nederlandse onderwijs in ISCED
teer to give blood? An investigation of Self-efficacy: the third factor besides 2011. 2011. https://www.cbs.nl/nl-nl/
the correlates of donation intentions attitude and subjective norm as a pre- onze-diensten/methoden/classificaties/
among young nondonors. Transfusion dictor of behavioural intentions. onderwijs-en-beroepen/isced/niveau-
2005; 45:945–55 Health Educ Res 1988; 3:273–82 isced-2011 (Last accessed 19-06-
16 Flora JA, Maibach EW, Maccoby N: 25 Conner M, Godin G, Sheeran P, et al.: 2020).
The role of media across four levels of Some feelings are more important: 33 UNESCO Institute for Statistics: Inter-
health promotion intervention. Annu cognitive attitudes, affective attitudes, national Standard Classification of
Rev Public Health 1989; 10:181–201 anticipated affect, and blood donation. Education ISCED 2011. Quebec, Mon-
17 James AB, Schreiber GB, Hillyer CD, Health Psychol 2013; 32:264–72 treal, 2012
et al.: Blood donations motivators and 26 De Vries H, Mudde AN, Dijkstra A, 34 Veldhuizen I, Ferguson E, De Kort W,
barriers: a descriptive study of African et al.: Differential beliefs, perceived et al.: Exploring the dynamics of the
American and white voters. Transfus social influences, and self-efficacy theory of planned behavior in the
Apher Sci 2013; 48:87–93 expectations among smokers in vari- context of blood donation: does dona-
18 Prochaska JO, DiClemente CC: The ous motivational phases. Prev Med tion experience make a difference?
transtheoretical approach; in Norcross 1998; 27:681–9 Transfusion 2011; 51:2425–37
JC, Goldfried MR (eds): Handbook of 27 Schreiber GB, Schlumpf KS, Glynn SA, 35 Wevers A, Wigboldus DH, de Kort WL,
Psychotherapy Integration (2nd ed). et al.: Convenience, the bane of our et al.: Characteristics of donors who
New York, Oxford University Press, existence, and other barriers to donat- do or do not return to give blood and
2005:141-71. ing. Transfusion 2006; 46:545–53 barriers to their return. Blood Trans-
19 Burditt C, Robbins ML, Paiva A, et al.: 28 Mathew SM, King MR, Glynn SA, et al.: fus-Italy 2014; 12(Suppl 1):s37–43
Motivation for blood donation among Opinions about donating blood among 36 Misje AH, Bosnes V, Gasdal O, et al.:
African Americans: developing mea- those who never gave and those who Motivation, recruitment and retention
sures for stage of change, decisional stopped: a focus group assessment. of voluntary non-remunerated blood
balance, and self-efficacy constructs. J Transfusion 2007; 47:729–35 donors: a survey-based questionnaire
Behav Med 2009; 32:429–42 29 Sanquin.org: Motivations to Give, study. Vox Sang 2005; 89:236–44
20 McVittie C, Harris L, Tiliopoulos N: "I Donate and Share (Motive-study). 37 Lemmens KP, Abraham C, Ruiter RA,
intend to donate but . . .": non-donors’ 2019. https://www.sanquin.org/researc et al.: Identifying blood donors willing
views of blood donation in the UK. h/donor-studies-projects/motive-study/ to help with recruitment. Vox Sang
Psychol Health Med 2006; 11:1–6 index (Last accessed 19-06-2020). 2008; 95:211–7
Appendix A
Part of Motive-study
Separate study
Study African background (n = 248) University students (n = 64) Social media users (n = 97) Significance testing
Characteristic n (%)/M (SD) n (%)/M (SD) n (%)/M (SD) F (df)/X2 (df)
NOTES
*p < 0.05.
**p < 0.01.
***p < 0.001.
Background and objectives Sheep are increasingly being used as a large in vivo
animal model of blood transfusion because they provide several advantages over
small animals. Understanding the effects of storage duration on ovine (ov) red
cell concentrates (RCCs) and how these changes compare with stored human (hu)
RCCs is necessary to facilitate clinical translation of research findings.
Materials and methods OvRCCs (n = 5) collected and processed in standard
human blood collection packs, and equivalent huRCCs provided by Australian
Red Cross Lifeblood (n = 5), were stored at 2–6°C for 42 days, with samples col-
lected weekly. Haemolysis index was determined by measuring supernatant hae-
moglobin concentration. Biochemical parameters were evaluated using a blood
gas analyser. Energy metabolites and biologically active lipids were measured
using commercial assays. Osmotic fragility was determined by lysis in various
saline concentrations. Extracellular vesicles were characterized by nanoparticle
tracking analysis.
Results Ovine red blood cells (RBCs) are double in number, smaller in size and
more fragile than human RBCs. Haematological values were unchanged through-
out storage. In contrast, biochemical and metabolic values, and haemolysis index
in three of the five ovRCCs exceeded huRCCs licensing criteria by day 42. Accu-
mulation of extracellular vesicles and biologically active lipids was comparable
between huRCCs and ovRCCs.
Conclusion This study documents similarities and differences in the storage
lesion of ovRCCs and huRCCs. This new information will guide the design of
ovine transfusion models to enhance translation of findings to human transfusion
Received: 14 April 2020,
settings.
revised 2 September 2020,
accepted 30 September 2020, Key words: sheep, storage lesion, haemolysis, osmotic fragility, extracellular vesi-
published online 26 October 2020 cles, biologically active lipids.
524
Storage lesion of sheep red cell concentrates 525
primate-specific enzyme-linked immunosorbent assays 7.05, GraphPad Software Inc, USA) with significance set
(ELISA) and controls (Roche, Mannheim, Germany) as per at values of P ≤ 005.
the manufacturer’s instruction.
Results
Assessment of osmotic fragility
RBC quality assessment and biochemical and
Osmotic fragility of ovRCCs and huRCCs at day 1 and day
metabolic changes
42 was determined by measuring lysis in a range of buf-
fered saline (NaCl) concentrations, as described previously Ovine RBCs were significantly smaller in size and double
[19]. Briefly, RCC samples were suspended in water (100% the number of human RBCs. The measured haematologi-
lysis control) or 1 ml of buffered NaCl solutions ranging cal values in both ovRCCs and huRCCs were unchanged
from 03–09% (w/v), incubated for 30 min at room tem- throughout storage (Table 1).
perature followed by centrifugation (10009 g for 5 min). Haemolysis of both huRCC and ovRCC increased dur-
Supernatants were collected, and haemolysis was mea- ing storage (Fig. 1A). In ovRCC, there was a significant
sured with a spectrophotometer (Synergy H1 Hybrid increase on day 21 that continued to day 42. The
Plate Reader, Millennium Science, BioTek) at a wavelength haemolysis index of all five huRCCs met the less than
of 540 nm. Results were normalized to the 100% lysis 08% specification by CoE until the end of shelf-life;
control, and osmotic fragility curves were plotted. however, this was true for only two of the five ovRCCs.
The pH of huRCCs decreased steadily throughout storage
(Fig. 1B). In contrast, the pH of ovRCCs remained
Characterization of extracellular vesicles (EVs)
unchanged. Potassium levels in both human and ovine
Concentration and size distribution of EVs in RCC super- RCCs increased during storage, but the increase was
natants were determined by nanoparticle tracking analysis modest in ovRCCs (Fig. 1C). Sodium levels in huRCCs
(NTA) using a NanoSight NS300 (Malvern Technologies, decreased steadily during storage (Fig. 1D). In contrast,
Malvern, UK). Frozen samples were thawed and diluted in in ovRCCs there was an initial decrease followed by an
02 µm filtered phosphate-buffered saline adjusted to 80– increase on day 14 which remained relatively stable
150 particles/frame. Ten experimental replicates of each thereafter (Fig. 1D). At day 1, ovRCCs had higher glu-
sample were analysed under constant flow conditions at cose (Fig. 1E) and lower lactate (Fig. 1F) levels than
room temperature. Concentration and size distribution of huRCCs, but during storage, glucose levels decreased
MVs were determined using NTA 3.2 software with detec- and lactate levels in both increased. Increased lactate
tion threshold of 5. levels were evident earlier in storage in huRCCs com-
pared to ovRCCs (day 14 vs. day 28). At day 1, ovRCCs
had higher ATP (Fig. 1G) and lower 2,3-DPG (Fig. 1H)
Quantification of biologically active lipid
concentrations compared to huRCCs. ATP and 2,3-DPG
mediators
concentrations in both huRCCs and ovRCCs decreased
The concentration of 12-hydroxyeicosatetraenoic acid throughout storage. In huRCCs, ATP decreased from day
(HETE) and 15-HETE was determined by using commer- 35 (P = 00287) and 2,3-DPG decreased from day 14
cially available non-primate-specific ELISAs (Enzo Life (P = 00109). In ovRCCs, ATP decreased from day 1
Sciences, Australia) according to the manufacturer’s throughout storage (P < 00001), while 2,3-DPG
instructions. All samples were run in duplicate. decreased at day 14 (P = 00272).
Table 1 Haematological values of ovine and human RCCs stored for 42 days
RBC [91012/l]
ovRCC 124 – 07 126 – 06 129 – 06 124 – 07 125 – 09 126 – 06 125 – 06 P < 00001
huRCC 56 – 03 57 – 03 58 – 03 57 – 03 56 – 03 58 – 03 59 – 03
Hb [g/l]
ovRCC 153 – 8 155 – 7 160 – 9 153 – 9 154 – 9 153 – 9 153 – 8 P = 00449
huRCC 173 – 12 173 – 12 177 – 17 173 – 14 169 – 13 175 – 14 179 – 14
Hct [%]
ovRCC 045 – 003 045 – 003 046 – 003 044 – 003 044 – 003 044 – 003 044 – 003 P = 00091
huRCC 050 – 003 051 – 003 052 – 004 051 – 003 050 – 003 052 – 003 053 – 003
MCV [fl]
ovRCC 362 – 33 355 – 30 356 – 31 352 – 30 351 – 29 351 – 29 352 – 29 P < 00001
huRCC 894 – 26 893 – 28 893 – 29 894 – 25 895 – 24 901 – 24 906 – 31
MCH [pg]
ovRCC 124 – 07 123 – 07 125 – 07 124 – 07 124 – 08 122 – 08 122 – 06 P < 00001
huRCC 308 – 15 305 – 16 306 – 20 306 – 16 302 – 16 304 – 19 306 – 19
MCHC [g/l]
ovRCC 343 – 14 347 – 11 351 – 13 352 – 11 354 – 9 347 – 9 348 – 11 ns
huRCC 345 – 7 342 – 8 343 – 13 342 – 8 337 – 10 338 – 12 337 – 10
Values are shown as mean – SD. P-value obtained using a two-way ANOVA comparing ovRCC vs. huRCC throughout storage.
ovRCC, ovine red cell concentrates; huRCC, human red cell concentrates; D, day; RBC, red blood cell; Hb, haemoglobin; Hct, haematocrit; MCV, mean cor-
puscular volume; MCH, mean corpuscular Hb; MCHC, mean corpuscular Hb concentration; ns, not significant.
Fig. 1 Biochemical and metabolic changes of ovine and human RCCs stored for 42 days. Haemolysis [A] in stored ovRCCs and huRCCs was determined
by measuring supernatant haemoglobin concentration, and the percentage of haemolysis was calculated. pH [B], concentrations of potassium [C] and
sodium [D] and levels of glucose [E] and lactate [F] were measured with blood gas analyser. Concentrations of ATP [G] and 2,3-DPG [H] were measured
using ELISA. Results are shown as mean – SD (n = 5, respectively). *P < 005 vs. day 1 of ovRCC and #P < 005 vs. day 1 of huRCC, using one-way
ANOVA. P-value obtained using a two-way ANOVA comparing ovRCC vs. huRCC throughout storage. ATP, adenosine-triphosphate; 2,3-DPG, 2,3-diphospho-
glyceric acid; K+, potassium; nd Na+, sodium. [Colour figure can be viewed at wileyonlinelibrary.com]
Fig. 3 Accumulation of extracellular vesicles and biologically active lipids in ovine and human RCC supernatants throughout 42 days of storage. Mean
particle size [A] and concentration [B] of extracellular vesicles were determined by nanoparticle tracking analysis. Concentrations of 12-HETE [C] and
15-HETE [D] were measured with ELISA. Results shown are mean – SD (n = 5, respectively). *P < 005 vs. day 1 of ovRCC and #P < 005 vs. day 1 of
huRCC, using one-way ANOVA. P-value obtained using a two-way ANOVA comparing ovRCC vs. huRCC throughout storage. 12-HETE = 12-hydroxyeicosate-
traenoic acid and 15-HETE = 15-hydroxyeicosatetraenoic acid. [Colour figure can be viewed at wileyonlinelibrary.com]
huRCCs in this study. In contrast, modest increases in characterized by the latter. It also worth noting that
both sodium and potassium were seen during ovRCC stor- the smaller EVs reported in our study (100–200 nm)
age. This may be in part due to the unchanged pH and were more numerous than larger EVs previously
reduced lactate accumulation in ovRCCs. Additionally, reported using flow cytometry (>~500 nm) [36]. Overall,
Australian Merino sheep, as were used in this study, have while ovRCCs have a lower concentration of EVs than
fewer sodium–potassium pumps on their RBCs [27], huRCCs, similarities in the mean size and increased
resulting in lower potassium and higher sodium levels concentration throughout storage suggest sheep may
than humans [27, 28]. still be a viable model to study the impact of EVs in
The osmotic fragility of RBCs is determined by mem- RCC transfusion. However, EVs may need to be concen-
brane integrity, cell size and shape, with spherical cells trated and added to ovRCCs in order to better replicate
more fragile than normal discocyte cells [29]. Therefore, the human setting.
as ovine RBCs are smaller and more spherical [30] than Changes to membrane integrity can also influence
human RBCs, the observation of increased osmotic fragi- the release and production of the biologically active
lity of ovRCCs compared to huRCCs was expected, and lipids. HETEs, such as 12-HETE and 15-HETE, accumu-
consistent with other reports [31–33]. No significant dif- late during storage of huRCCs as a result of peroxida-
ferences were found in the osmotic fragility of either tion of lipids in the RBC cell membrane [39]. Both 12-
ovine RBCs or human RBCs stored for day 1 or day 42. HETE and 15-HETE play a role in mediating inflamma-
This finding was unexpected, as earlier studies reported tion, cell proliferation and apoptosis [40], and have
association between huRBCs osmotic fragility and pro- been associated with adverse transfusion outcomes,
longed storage duration [29, 33]. Whether similar to other including TRALI [39]. Both 12-HETE and 15-HETE were
quality parameters it is due to the different manufactur- detected in ovRCCs and huRCCs at comparable concen-
ing methods [34] is unclear and needs further investiga- trations; however, storage-related accumulation was
tion. only observed for 15-HETE. Further investigation is
Extracellular vesicles are shed from cell membranes required before ovine models can be used to study the
and play a role in adverse transfusion events including effects of these biologically active lipids in pre-clinical
coagulopathies, inflammation and transfusion-related transfusion studies.
acute lung injury (TRALI) [35–37]. While the size of In summary, sheep RBCs differ from human RBCs in
EVs in ovRCCs is similar to that of humans, the con- size and metabolic characteristics. This study documents
centration of EVs differed, with initial concentrations in some similarities but importantly also defines the differ-
huRCCs approximately four times higher than in ences between the storage lesion of human and ovine
ovRCCs. While a number of different factors may have RCCs. OvRCCs stored for 21 days develop a storage lesion
contributed to this difference, it has been shown that similar to that of huRCCs stored for 42 days. These limi-
compared to human RBCs, ovine RBCs produce long tations need to be considered when designing ovine
membrane extensions when they undergo stress [38]. in vivo transfusion models for evaluating the effects of
These extensions do not result in the budding of the stored RCC transfusion in various clinical situations such
membrane as seen in human RBCs, rather they retract as trauma or sepsis.
then collapse on the membrane surface, and therefore,
ovine RBCs do not release EVs in the same way [38].
Storage duration was associated with increased EV con- Acknowledgements
centration in ovRCCs, but not in huRCCs. This contrasts
The authors thank Kristen Glenister for technical assis-
with previous reports of EV accumulation in huRCCs
tance. JFF received funding from the Office of Health and
[36], but might simply reflect the small sample size in
Medical Research, Queensland Health. Australian govern-
our study and the fact that one of the five huRCC
ments fund the Australian Red Cross Lifeblood to provide
units had remarkably different results from the other
blood, blood products and services to the Australian com-
four units, resulting in wider error bars and no clear
munity.
differences in the post-tests. Another possible explana-
tion for the discrepancy might relate to differences in
the techniques used to measure EVs: nanoparticle track-
Conflict of interest
ing analysis vs. flow cytometry; with the former pro-
viding data on smaller EVs that could not be The authors have no conflict of interest.
References
1 Yoshida T, Prudent M, D’Alessandro A: glucose solution for 35 days. J Vet erythrocyte enzyme abnormalities of
Red blood cell storage lesion: causes Intern Med 1997; 11:5–8 glycolysis. Blood 2005; 106:4034–
and potential clinical consequences. 12 Solomon SB, Wang D, Sun J, et al.: 4042
Blood Transfus 2019; 17:27–52 Mortality increases after massive 22 Sousa RS, Barreto RA Jr, Sousa IK,
2 Lacroix J, Hebert PC, Fergusson DA, exchange transfusion with older stored et al.: Evaluation of hematologic,
et al.: Age of transfused blood in criti- blood in canines with experimental blood gas, and select biochemical
cally ill adults. N Engl J Med 2015; pneumonia. Blood 2013; 121:1663– variables in ovine whole blood stored
372:1410–1418 1672 in CPDA-1 bags. Vet Clin Pathol 2013;
3 Heddle NM, Cook RJ, Arnold DM, 13 Ferreira RRF, Graca RMC, Cardoso IM, 42:27–30
et al.: Effect of short-term vs. long- et al.: In vitro hemolysis of stored 23 Koch JH: Glucose-6-phosphate dehy-
term blood storage on mortality after units of canine packed red blood cells. drogenase and potassium concentra-
transfusion. N Engl J Med 2016; J Vet Emerg Crit Care 2018; 28:512– tion in sheep erythrocytes. Nature
375:1937–1945 517 1963; 200:1334–1335
4 Cooper DJ, McQuilten ZK, Nichol A, 14 Patel NN, Lin H, Jones C, et al.: Inter- 24 Suzuki T, Agar NS, Suzuki M: Red cell
et al.: Age of red cells for transfusion actions of cardiopulmonary bypass metabolism: a comparative study of
and outcomes in critically ill adults. N and erythrocyte transfusion in the some mammalian species. Comp Bio-
Engl J Med 2017; 377:1858–1867 pathogenesis of pulmonary dysfunc- chem Physiol B 1984; 79:515–520
5 Steiner ME, Ness PM, Assmann SF, tion in Swine. Anesthesiology 2013; 25 Vraets A, Lin Y, Callum JL: Transfu-
et al.: Effects of red-cell storage dura- 119:365–378 sion-associated hyperkalemia. Transfus
tion on patients undergoing cardiac 15 Mock DM, Lankford GL, Burmeister Med Rev 2011; 25:184–196
surgery. N Engl J Med 2015; LF, et al.: Circulating red cell volume 26 Eaton JW, Brewer GJ, Beck CC, et al.:
372:1419–1429 and red cell survival can be accurately ATP and potassium content of sheep
6 Pavenski K, Saidenberg E, Lavoie M, determined in sheep using the [14C]- erythrocytes. Biochem Biophys Res
et al.: Red blood cell storage lesions cyanate label. Pediatric Res 1997; Commun 1967; 28:898–903
and related transfusion issues: a Cana- 41:916–921 27 Dunham PB, Hoffman JF: The number
dian Blood Services research and 16 Council of Europe: Guide to the prepa- of Na + : K + pump sites on red blood
development symposium. Transfus ration, use and quality assurance of cells from HK and LK lambs. Biochim
Med Rev 2012; 26:68–84 blood components, 19th edn. Strass- Biophys Acta 1971; 241:399–402
7 Makley AT, Goodman MD, Friend LA, bourg, France, Council of Europe Pub- 28 Dunham PB: Ion transport in sheep
et al.: Murine blood banking: charac- lishing, 2017 red blood cells. Comp Biochem Physiol
terization and comparisons to human 17 Baron DM, Yu B, Lei C, et al.: Pul- Comp Physiol 1992; 102:625–630
blood. Shock 2010; 34:40–45 monary hypertension in lambs trans- 29 Blasi B, D’Alessandro A, Ramundo N,
8 d’Almeida MS, Jagger J, Duggan M, fused with stored blood is prevented et al.: Red blood cell storage and cell
et al.: A comparison of biochemical by breathing nitric oxide. Anesthesiol- morphology. Transfus Med 2012;
and functional alterations of rat and ogy 2012; 116:637–647 22:90–96
human erythrocytes stored in CPDA-1 18 Simonova G, Tung JP, Fraser JF, 30 Zura Zaja I, Vince S, Poljicak Milas N,
for 29 days: implications for animal et al.: A comprehensive ovine model et al.: A new method of assessing
models of transfusion. Transfus Med of blood transfusion. Vox Sang 2014; sheep red blood cell types from their
2000; 10:291–303 106:153–160 morphology. Animals (Basel) 2019;
9 Baek JH, D’Agnillo F, Vallelian F, 19 Veale MF, Healey G, Sparrow RL: 9:1130
et al.: Hemoglobin-driven pathophysi- Effect of additive solutions on red 31 Perk K, Frei YF, Herz A: Osmotic fra-
ology is an in vivo consequence of the blood cell (RBC) membrane properties gility of red blood cells of young and
red blood cell storage lesion that can of stored RBCs prepared from whole mature domestic and laboratory ani-
be attenuated in guinea pigs by hap- blood held for 24 hours at room tem- mals. Am J Vet Res 1964; 25:1241–
toglobin therapy. J Clin Invest 2012; perature. Transfusion 2011; 51(Suppl 1248
122:1444–1458 1):25S–33S 32 Coldman MF, Gent M, Good W: The
10 Price GS, Armstrong PJ, McLeod DA, 20 Hess JR, Sparrow RL, van der Meer osmotic fragility of mammalian ery-
et al.: Evaluation of citrate-phosphate- PF, et al.: Red blood cell hemolysis throcytes in hypotonic solutions of
dextrose-adenine as a storage medium during blood bank storage: using sodium chloride. Comp Biochem Phys-
for packed canine erythrocytes. J Vet national quality management data to iol 1969; 31:605–609
Intern Med 1988; 2:126–132 answer basic scientific questions. 33 Zehnder L, Schulzki T, Goede JS,
11 Wardrop KJ, Tucker RL, Mugnai K: Transfusion 2009; 49:2599–2603 et al.: Erythrocyte storage in hyper-
Evaluation of canine red blood cells 21 van Wijk R, van Solinge WW: The tonic (SAGM) or isotonic (PAGGSM)
stored in a saline, adenine, and energy-less red blood cell is lost: conservation medium: influence on
cell properties. Vox Sang 2008; routine storage of packed red blood accumulate during the routine storage
95:280–287 cells: potential risk for prothrombotic of prestorage leukoreduced red blood
34 Shih AW, Apelseth TO, Cardigan R, post-transfusion complications. cells and cause acute lung injury.
et al.: Not all red cell concentrate Pathology 2017; 49:62–69 Transfusion 2011; 51:2549–2554
units are equivalent: international sur- 37 McVey M, Tabuchi A, Kuebler WM: 40 Powell WS, Rokach J: Biosynthesis,
vey of processing and in vitro quality Microparticles and acute lung injury. biological effects, and receptors of
data. Vox Sang 2019; 114:783–794 Am J Physiol Lung Cell Mol Physiol hydroxyeicosatetraenoic acids (HETEs)
35 Koshiar RL, Somajo S, Norstrom E, 2012; 303:L364–381 and oxoeicosatetraenoic acids (oxo-
et al.: Erythrocyte-derived microparti- 38 Backman L, Jonasson JB, Horstedt P: ETEs) derived from arachidonic acid.
cles supporting activated protein C- Phosphoinositide metabolism and Biochim Biophys Acta 2015;
mediated regulation of blood coagula- shape control in sheep red blood cells. 1851:340–355
tion. PLoS One 2014; 9:e104200 Mol Membr Biol 1998; 15:27–32
36 Aung HH, Tung JP, Dean MM, et al.: 39 Silliman CC, Moore EE, Kelher MR,
Procoagulant role of microparticles in et al.: Identification of lipids that
This is an open access article under the terms of the Creative Commons Attribution License, 533
which permits use, distribution and reproduction in any medium, provided the original work is properly cited.
534 J. Hoefer et al.
of the presence of age-dependent, blood-borne factors is therefore possible that TIMP-2 might affect cognitive
which affect neurogenesis, cognitive functions and ageing capabilities and/or motor functions in recipients if trans-
when transferred to other individuals. ferred during blood transfusion.
On the basis of these facts, it can be hypothesized that As stated above, it has been shown that TIMP-2 levels
such ‘ageing’ or ‘rejuvenating’ factors are present also in differ between young and old mice [3]. In humans, Rossig-
blood components for transfusion and may be transferred nol et al. demonstrated age-dependent changes. It has also
to recipients during transfusion, which might affect the been demonstrated that TIMP-2 is present in freshly drawn
recipient‘s cognitive capabilities. Indeed, it is known that a human plasma [12,13]. However, until now, it has neither
considerable proportion of patients experience a temporary been assessed if TIMP-2 is detectable in human blood com-
or permanent decline in cognitive performance after major ponents processed and stored for transfusion, nor if TIMP-2
surgery, a phenomenon called postoperative cognitive dys- correlates with age or gender in adult human beings.
function (POCD) [4]. The causal relationship of blood trans- In the present study, we assessed for the first time if
fusion and such disturbances has not been clarified so far. blood products for transfusion (FFP, EC, PI-PC) contain
However, intraoperative transfusion of erythrocyte concen- detectable TIMP-2 levels and if TIMP-2 concentration
trates (>3 units) has been described as an independent risk varies in relation to the donor‘s age or gender.
factor for the development of POCD [5]. Therefore, it is
essential to investigate whether factors associated with
Material and methods
cognitive function, neurogenesis and ageing are present in
blood components processed and stored for transfusion. This trial was performed as a descriptive, single-centre,
In a previous study, we were able to demonstrate that non-clinical pilot study. All experiments have been con-
the ‘ageing factor’ eotaxin-1 is detectable and stable in ducted in accordance with the Helsinki Declaration. The
ready-to-use blood products for transfusion. Most impor- study was approved by the Ethics Committee of the Medical
tantly, we discovered a gender-independent increase of University of Innsbruck (Ethical vote number: 1083/2018).
eotaxin-1 with rising age of donors in both fresh-frozen
plasma (FFP) and erythrocyte concentrates (EC), while
Investigational product
eotaxin-1 was subject to only minor fluctuations within
one donor over a longer period of time [6]. This is signifi- For the present study, we analysed leukocyte-depleted
cant as high eotaxin-1 levels are associated with blood products (FFP [n = 30], EC [n = 12], and PI-PC
decreased cognitive functions, although these were not [n = 12]) from volunteer donors aged 18 to 63. At the
examined in our previous study [1,2,6]. It might be possi- University Hospital of Innsbruck, the suitability for vol-
ble that besides eotaxin-1 also other factors associated unteer blood donation is assessed on the basis of strin-
with cognitive functions and ageing are detectable in gent, standardized health criteria. After completing a
human blood components processed for transfusion. medical questionnaire on their general health status,
TIMP-2 has originally been described as an endogenous recent diseases, medication and travelling, all persons are
inhibitor of matrix-metalloproteinases (MMPs). It is physically examined, including measurements of blood
involved in MMP-2 activation and contributes to the pro- pressure, heart rate, temperature and haemoglobin con-
tection of the extracellular matrix (ECM) from proteolysis centration. The donated blood is further tested for neop-
[7]. It has also been reported that TIMP-2 is involved in terin as well as transfusion-transmissible infections such
neuronal differentiation and can act as a key effector of as HIV, hepatitis, syphilis and parvovirus B19. In case
the pro-neurogenic response to an inducing stimulus any disease or abnormality is detected, the donated blood
[8,9]. Moreover, treatment of aged mice with either is discarded. This procedure is to ensure all donors are in
umbilical cord plasma or purified TIMP-2 led to signifi- good health at the time of blood donation.
cant improvements in multiple cognitive measures such After donation, the blood components used for this
as contextual memory, learning and nesting, as well as study were stored under standard conditions at the
increased synaptic plasticity in the hippocampus [3]. In Central Institute for Blood Transfusion & Immunological
addition, TIMP-2 knockout mice display motoric dysfunc- Department at the University Hospital of Innsbruck,
tions such as dyskinesia, trembling and a splayed, short- Austria, after having undergone the following procedures:
ened gait [10]. This phenotype has mechanistically been
linked to altered formation and maintenance of neuro-
Preparation of erythrocyte concentrate
muscular junctions (NMJ), the deterioration of which is a
common sign of ageing [11]. Taken together, these obser- Leukocyte-depleted ECs were prepared from single-donor
vations suggest that TIMP-2 influences cognitive as well whole blood donations by cell separation. The preparation
as motoric functions, which are both reduced in ageing. It was conducted in accordance with the guidelines of the
Council of Europe. One hundred millilitres of additive The kit contained a freeze-dried calibrator for solubiliza-
solution (SAGM) were added in order to increase shelf tion in sample buffer and for generating a seven-point
life. Each 280 ml EC package had a haematocrit of 50–70 standard curve. Signals were detected using ChemiDoc
vol% and contained less than 1*106 leukocytes. ECs are chemiluminescence imaging system (Biorad), and TIMP-2
stored for a maximum of 42 days at 4°C. was quantified with Q-View Software (Quansys) in rela-
tion to an 8-point standard curve obtained by serial dilu-
tion of recombinant TIMP-2. All samples were diluted
Preparation of fresh-frozen plasma
1:20 with sample dilution buffer and measured in dupli-
Leukocyte-depleted FFPs were prepared from single-donor cates. Plasma from fresh residual umbilical cord blood
whole blood donation by cell separation. FFPs at a core (n = 3, 2 ml each) was used as a positive control.
temperature of -30°C can be stored up to 2 years.
Statistical methods
Preparation of pathogen-inactivated PC (PI-PC)
Statistical analyses were performed in SPSS and Graph-
Leukocyte-depleted PCs were prepared by single-donor Pad Prism. Differences between groups were analysed
apheresis. Each 300 ml PI-PC contained 2–4*1011 plate- using Mann–Whitney test. All differences highlighted by
lets, suspended in approximately 35% plasma and 65% asterisks were statistically significant as encoded in fig-
SSP + platelet additive solution. For safety reasons, patho- ures (*P < 005, **P < 001, ***P < 0001). Data are pre-
gen inactivation treatment of all PCs was performed with sented as Box–Whiskers Plots showing median and
amotosalen/UVA treatment (INTERCEPT blood system, minimum to maximum values. For comparison of TIMP-2
CERUS corporation, Amersfoort, the Netherlands). PI-PCs levels in relation to age, donors were divided into two
were stored under continuous agitation in gas-permeable groups, representing the younger and the older 50% each.
sterile plastic bags at room temperature (20–24°C). Under
these conditions, the maximum storage time is 7 days.
Results
Fig. 1 TIMP-2 is detectable in fresh-frozen plasma (FFP), pathogen-inactivated platelet concentrate (PI-PC) and erythrocyte concentrate (EC) (a,c,d).
Umbilical cord plasma serves as a positive control. TIMP-2 in FFP is comparable in male and female donors and independent of age (a,b). TIMP-2 in PI-
PC is comparable in male and female donors and independent of age (c,d). TIMP-2 in EC is rather low concentrated but comparable in male and female
donors and independent of age (e,f). Data are presented as Box–Whiskers Plot, showing median and minimum to maximum values. *P < 005;
**P < 001; ***P < 0001; Mann–Whitney test.
differences in TIMP-2 in FFP donated by younger (21– younger (24–45 years) and older (46–63 years) individu-
45 years) or older (46–63 years) individuals could be als (Fig. 1e), as well as in samples from male and female
detected. However, umbilical cord plasma exhibited sig- donors (Fig. 1f).
nificantly higher TIMP-2 levels compared with plasma
from adult donors (Fig. 1a). Moreover, TIMP-2 levels in
Tissue inhibitor of metalloproteinases 2 is stable
FFP from male and female donors were comparable
in FFP during storage
(Fig. 1b). Similarly, the median TIMP-2 concentration did
not differ in PI-PC from younger (18–35 years) and older Tissue inhibitor of metalloproteinases 2 was detectable
(36–61 years) donors (Fig. 1c), nor in PI-PC samples from in FFP after 15 days and 42 days of storage at -20°C.
male or female individuals (Fig. 1f), although variability As shown in Fig. 2, no major differences in TIMP-2
was higher in the older age group and in male donors. As levels were found. Median TIMP-2 levels were 65 ng/
mentioned above, the median TIMP-2 concentration in EC mL (range: 444–873) after 15 days of storage and
was significantly lower compared with FFP or PI-PC, and 606 ng/mL (range: 535–656) after 42 days of storage
again, TIMP-2 levels were comparable in EC samples from (Table 2).
References
1 Villeda SA, Luo J, Mosher KI, et al.: 7 Clark IM, Swingler TE, Sampieri CL, 13 Rossignol P, Cambillau M, Bissery A,
The ageing systemic milieu negatively et al.: The regulation of matrix metallo- et al. Influence of blood sampling pro-
regulates neurogenesis and cognitive proteinases and their inhibitors. Int J Bio- cedure on plasma concentrations of
function. Nature 2011; 477:90–94 chem Cell Biology 2008; 40:1362–1378 matrix metalloproteinases and their
2 Bettcher BM, Fitch R, Wynn MJ, et al.: 8 Perez-Martınez L, Jaworski DM: Tissue tissue inhibitors. Clin Exp Pharmacol
MCP-1 and eotaxin-1 selectively and inhibitor of metalloproteinase-2 pro- Physiol 2008; 35:464–469
negatively associate with memory in motes neuronal differentiation by act- 14 Lema^ıtre V, D’Armiento J: Matrix met-
MCI and Alzheimer’s disease dementia ing as an anti-mitogenic signal. J alloproteinases in development and
phenotypes. Alzheimer’s Dementia Neurosci 2005; 25:4917–4929 disease. Birth Defects Res Part C
Diagnosis Assess Dis Monit 2016; 9 Sinno M, Biagioni S, Ajmone-Cat MA, Embryo Today Rev 2006; 78:1–10
3:91–97 et al.: The matrix metalloproteinase 15 Baba Y, Yasuda O, Takemura Y, et al.:
3 Castellano JM, Mosher KI, Abbey RJ, inhibitor marimastat promotes neural Timp-3 deficiency impairs cognitive
et al.: Human umbilical cord plasma progenitor cell differentiation into function in mice. Lab Invest 2009;
proteins revitalize hippocampal func- neurons by gelatinase-independent 89:1340–1347
tion in aged mice. Nature 2017; TIMP-2-dependent mechanisms. Stem 16 Ould-yahoui A, Tremblay E, Sbai O,
544:488–492 Cells Dev 2013; 22:345–358 et al.: A new role for TIMP-1 in mod-
4 Monk TG, Price CC: Postoperative cog- 10 Jaworski DM, Soloway P, Caterina J, ulating neurite outgrowth and mor-
nitive disorders. Curr Opin Crit Care et al.: Tissue inhibitor of metallopro- phology of cortical neurons. PLoS One
2011; 17:376–381 teinase-2(TIMP-2)-deficient mice dis- 2009; 4:e8289
5 Zhu S-H, Ji M-H, Gao D-P, et al.: play motor deficits. J Neurobiol 2005; 17 Holten-Anderson MN, Brunner N,
Association between perioperative 66:82–94 Christensen IJ, et al.: Levels of tissue
blood transfusion and early postopera- 11 Rudolf R, Khan MM, Labeit S, et al.: inhibitor of metalloproteinases-1 in
tive cognitive dysfunction in aged Degeneration of neuromuscular junc- blood transfusion components. Scand
patients following total hip replace- tion in age and dystrophy. Front J Clin Lab Invest 2009; 62:223–230
ment surgery. Upsala J Med Sci 2013; Aging Neurosci 2014; 6:99 18 Gresele P, Falcinelli E, Sebastiano M,
119:262–267 12 Hopps E, Presti RL, Montana M, et al.: et al.: Chapter four matrix metallopro-
6 Hoefer J, Luger M, Dal-Pont C, et al.: Study of the correlations among some teinases and platelet function. Prog
The “Aging Factor” Eotaxin-1 (CCL11) parameters of the oxidative status, Mol Biol Transl 2017; 147:133–165
is detectable in transfusion blood gelatinases, and their inhibitors in a 19 Tayebjee MH, Lip GY, Blann AD,
products and increases with the group of subjects with metabolic syn- et al.: Effects of age, gender, ethnicity,
donor’s age. Front Aging Neurosci drome. Mediat Inflamm 2014; diurnal variation and exercise on cir-
2017; 9:402 2014:510619 culating levels of matrix
metalloproteinases (MMP)-2 and -9, 20 Bonnema DD, Webb CS, Pennington 21 Momi S, Falcinelli E, Giannini S,
and their inhibitors, tissue inhibitors WR, et al.: Effects of age on plasma et al.: Loss of matrix metalloproteinase
of matrix metalloproteinases (TIMP)-1 matrix metalloproteinases (MMPs) and 2 in platelets reduces arterial thrombo-
and -2. Thromb Res 2005; 115:205– tissue inhibitor of metalloproteinases sis in vivo. J Exp Medicine 2009;
210 (TIMPs). J Card Fail 2007; 13:530–540 206:2365–2379
Background Prompt resuscitation with plasma and other blood products reduces
trauma-related morbidity and mortality. Standard storage and preparation tech-
niques for frozen plasma limit its utility in the pre-hospital setting. Plasma can
be dehydrated using hot air (spray-dried plasma), stored at room temperature and
rehydrated quickly for use. The spray-dry process decreases high-molecular-
weight multimers of von Willebrand factor compared with conventional plasma.
The objective of this study was to compare platelet adhesion and thrombus for-
mation in a microfluidic perfusion assay facilitated by spray-dried compared with
frozen plasma using a non-inferiority design.
Study Design and Methods Whole blood was centrifuged to obtain red cell concen-
trate, and a platelet pellet that was suspended in either spray-dried or frozen plasma to
create recombined whole blood. Platelets were fluorescently labelled, and samples were
flowed through a collagen-coated microchannel. Surface area coverage by platelets and
thrombi was analysed and compared between each spray-dried and frozen plasma pair.
Results Compared with whole blood samples containing frozen plasma, samples
with spray-dried plasma had similar surface area coverage of platelets and
thrombi after 180 s of flow. Even when diluted with von Willebrand factor-free
plasma, there was no reduction thrombus formation.
Conclusion Spray-dried plasma is not inferior in supporting haemostasis com-
pared with fresh frozen plasma in a paired analysis. It offers advantages with
respect to portability and ease of preparation over frozen plasma in the pre-hos-
Received: 4 August 2020,
pital setting. This study supports development of clinical studies to evaluate the
revised 20 October 2020,
accepted 21 October 2020,
efficacy and safety of spray-dried plasma in trauma patients.
published online 5 December 2020 Key words: component processing, plasma, spray drying, transfusion, trauma.
540
Spray-dried vs. fresh frozen plasma 541
Unlike plasma in a frozen or liquid state, anhydrous volume at 10 and 30 min and but was still inferior to that
plasma or plasma stored as a powder has an extended of controls [18].
shelf life, can be reconstituted quickly, tolerates warmer The impact of the decreased HMWM of VWF in SpDP
ambient temperatures and can be easily transported in on restoring and maintaining platelet adhesion and
the pre-hospital setting [7,8]. Lyophilized plasma was thrombus formation is unknown. Therefore, in order to
developed in the 1930s [9] and widely used in World War simulate the worst-case scenario of a potential patient
II [7]. Its use fell out of favour after infectious disease with VWD being resuscitated with SpDP, we tested its
risks were identified, and it is not currently FDA in vitro efficacy in a subject with T3VWD. The primary
approved. However, the French armed services have a objective of this study was to compare in vitro thrombus
long history of using lyophilized plasma [10]. While the formation in whole blood that has been reconstituted
infectious disease risk has been mitigated, it still requires with either SpDP or FFP. Secondary objectives were to
specialized manufacturing at offsite locations. evaluate SpDP and FFP simulate conditions of VWD by
Single donor spray-dried plasma (SpDP) removes water diluting SpDP and FFP with plasma from a subject with
by passing plasma microdroplets over heated (up to T3VWD to evaluate efficacy of SpDP and FFP of support-
150°C) flowing air into a collection bag [8]. It takes ing thrombus formation in conditions with decreased
approximately 25 min to spray dry a single donor unit of total VWF.
plasma. This spray-dried plasma powder can be reconsti-
tuted in five minutes with sterile water. Compared with
Materials and methods
FFP, SpDP has ≥80% of all pro- and anti-coagulant fac-
tors with the exception of Factor VIII, which is approxi-
Replacing whole blood plasma with SpDP or FFP
mately 75% [11]. Additionally, endothelial inhibitory
effects on inflammation and vascular permeability were Using a protocol approved by the local institutional
equivalent in models comparing SpDP with FFP [12]. review board, whole blood from consented healthy adult
These results suggest that coagulation is not affected by volunteers was collected into 32% sodium citrate (Bec-
the reduced coagulation factor levels, likely because they ton, Dickinson and Company, Franklin Lakes, NJ, USA).
are still above the critical level of 30% [13,14]. Addition- Donor FFP was collected and shipped to the Velico (Bev-
ally, both pro- and anti-coagulation proteins are erly, MA, USA) for SpDP manufacturing [8,11,19]. Fol-
decreased, which may enable an appropriate haemostatic lowing collection, the whole blood (WB) was centrifuged
balance. at 200 g for 12 min after which the platelet rich plasma
The coagulation factor that was most affected by the (PRP) was removed from the red blood cell concentrate
spray-drying process is von Willebrand Factor (VWF); the (RBCs). The PRP was diluted with 10 ml of 1,4-Piper-
high-molecular-weight multimers (HMWM) are affected azinediethanesulfonic acid (PIPES, Millipore Sigma, St.
due to the heat drying process and are more susceptible Louis, MO, USA) saline solution, and 10 µl of prosta-
than smaller-sized multimers. HMWM are the most active glandin E1 (PGE-1, Santa Cruz Biotechnology, Dallas,
forms of VWF and a lack of HMWM can impair VWF TX, USA) was added (final concentration = 1 µM). The
binding to glycoprotein (GP) Ib and collagen, which is platelet solution was then centrifuged at 1900 g for
essential for platelet adhesion [15]. von Willebrand dis- 8 min after which the supernatant was completely
ease (VWD) is an inherited bleeding disorder character- removed and discarded. The platelets were resuspended
ized by qualitative or quantitative defects in VWF, which in 400 µl of either SpDP or FFP to form a platelet con-
is associated with defective platelet adhesion and aggre- centrate (PC). The RBCs, PC and SpDP or FFP were then
gation [16]. While VWD has an estimated prevalence of mixed together to make recombined whole blood (rWB,
1% in the population, Type 3 VWD (T3VWD), which is SpDP-WB or FFP-WB). To control for the effect of
autosomal recessive and characterized by a total or near- haematocrit in the microfluidic model, whole blood
total absence of VWF both in the plasma and cellular reconstitution was designed to achieve a target haemat-
compartments, has an estimated prevalence of 05–3/ ocrit and platelet count [20]. The SpDP-WB and FFP-WB
1 000 000 [17]. Compared with normal controls, patients samples were made to have a platelet count 150 000/µl–
with T3VWD have decreased thrombus formation under 275 000/µl and a haematocrit of 33–40%. Platelet count
flow conditions as measured by the Total Thrombus-For- and haematocrit were similar in the preparation of the
mation Analysis System (T-TAS, Fujimori Kogyo, Yoko- SpDP/FFP pairs.
hama, Japan) [18]. After receiving a prophylactic dose of SpDP was rehydrated with de-ionized water, and
VWF-Factor VIII concentrate (median dose of VWF: 27 sodium hydroxide was added to bring it to a pH between
U/kg, range: 15–35 U/kg), thrombus formation improved 64 and 67. Six SpDP and FFP pairs were created and
with respect to time to thrombus initiation, thrombus represented a batch. In order to create reconstituted whole
blood, red blood cells and platelets from 16 donors and open source software platform as an image-processing
one type 3 vWD patient were combined with a batch programme designed for scientific multidimensional
(paired SpDP and FFP). Each was tested at least twice images to perform a wide variety of dimensional analy-
with repeated measures accounted for in the statistical ses.
analysis. Each specific batch was reconstituted per manu- Images from the SpDP-FFP pairs were processed identi-
facturer’s specifications to ensure parity with the FFP unit cally in order to calculate the SA of the platelets and
that was made from the same donation. The volume of thrombi (Fig. 1). Fluorescent images were background
rehydration fluid was dictated by the manufacturer to corrected, and threshold levels were set for the GPIIbIIIa
produce a product an osmotically similar to the paired positive platelets in each rWB sample using provided
FFP unit. ImageJ thresholding algorithms. Images were then con-
To dilute SpDP and FFP with VWF-free plasma, whole verted to binary masks, which resulted in individual pla-
blood was collected from an individual (blood type O) telets and thrombi. Surface area covered by platelets and
with T3VWD. The plasma from a subject with T3VWD thrombi as well as the number of objects were obtained
(T3VWD-P) was obtained via two centrifugation steps from the ‘Analyze Particles’ toolbox in ImageJ. Images
(300 g and 1900 g) that were performed to isolate platelet processing and analysis for each image was performed by
rich and platelet poor plasma, respectively. This T3VWD- at least two researchers blinded to the samples.
P was subsequently frozen and stored at -20°C and
thawed in a 37°C shaking water bath prior to use. The
Statistical analyses
SpDP or FFP was diluted 25, 50 or 75% using the
T3VWD-P. The geometric mean of the SA from the still images was
calculated, because it is less influenced by outliers and
variability. The geometric mean was also used in the cal-
Microfluidic perfusion assay
culation of the ratio paired t test, calculating the differ-
Platelets in either unmanipulated WB, SpDP-WB or FFP- ence between SpDP and FFP as SpDP/FFP as compared
WB samples were labelled with a GPIIbIIIa-specific mono- to SpDP—FFP. The pre-determined margin of non-inferi-
clonal antibody AP2 (Hybridoma Core, BloodCenter of ority was 20% (SpDP/FFP >08) based on the FDA’s tar-
Wisconsin [BCW], Milwaukee, WI, USA) conjugated to get of SpDP being – 20% of control. A total of 16 pairs
Alexa Fluor (AF) 488 (Life Technologies Corp. Carlsbad, were needed to have a power of 09 at the one-sided
CA, USA). While AP2 can inhibit platelet aggregation at alpha of 005. Prism 60h (GraphPad, La Jolla, CA, USA)
concentrations between 6–10 µg/ml, [21,22] the concen- software was used for the analyses, and PASS15 (NCSS,
tration used in these experiments was 06 µg/ml, which LLC, Kaysville, UT, USA) was used for the power calcula-
we found to have no effect on whole blood platelet tion.
aggregometry, light transmission aggregometry or adhe-
sion under flow (data not shown).
Results
The microchannels (Vena8Fluoro + chips, Cellix,
Dublin, Ireland) were coated overnight at 4°C with type I
SpDP compared with FFP
collagen (Chrono-log Corp, Havertown, PA, USA) 50 µg/
ml diluted in 20 mM glacial acetic acid. WB, SpDP-WB Six batches of SpDP/FFP were evaluated using 17 sub-
and FFP-WB were flowed through the channels using the jects. There was no statistical difference between the
VenaFlux (Cellix) microfluidic platform at arterial shear SpDP/FFP pairs (P = 07558). The mean ratio of SpDP/
rate of 1500/s for 180 s. Platelet adhesion and aggrega- FFP was 121 with a 95th per cent confidence interval
tion in traumatic injury models can be confounded by (CI) of 084–157, which means that the there is a 95%
lower shear rates (capillary and venous) in a microfluidic chance that the ‘true’ ratio is greater than the 08 ratio
device since platelet adhesion may occur independent of that had been a priori determined. Thus, SpDP would be
the model. Thus, arterial shear rates were selected to pro- considered not inferior to the FFP with respect to sup-
vide the most representative flow rates in trauma and porting platelet adhesion and thrombus formation in
decrease the likelihood of platelet adhesion naturally whole blood samples (Fig. 2). A two-way ANOVA analysis
occurring [23,24]. The surface area (SA) coverage was demonstrated batch (SpDP and FFP pairs) did not signifi-
evaluated by taking 15 still images over the length of the cantly affect platelet adhesion and thrombus formation in
microfluidic channel following completion of flow. the surface area calculation.
Images of adherent platelets and thrombi were captured Of the 17 samples run, 4 (235%) were below the 08
in order to calculate SA along the length of the channel. ratio (range 026–066). Two of the four samples less than
Images were processed using ImageJ [25]; ImageJ is an 08 were from the same SpDP sample. Figure 1 is a
Fig. 1 Representative still images of thrombus formation in the microfluidic perfusion channels after 180 s of flow. (a-c) recombined whole blood (rWB)
made with fresh frozen plasma (FFP). (d-f) rWB made with spray-dried plasma (SpDP). Images (a) and (d) are the originals; images (b) and (e) have been
background corrected using the same Image J algorithm; images (c) and (f) have been converted to a binary mask using the Image J algorithm. Surface
area (SA) coverage by platelets and thrombi computed by Image J using the binary image. SA in rWB-FFP = 136 495 and in rWB-SpDP = 157 682 (ratio
SpDP/FFP = 1.16).
Fig. 2 Ratio of SPDP/FFP surface area coverage of platelets and thrombi after flow through microfluidic channels. Symbols represent the different
batches (SPDP-FFP pairs) reconstituted with unique red blood cell and platelet donors. A repeated symbol represents the same batch with a different
donor. The solid grey line indicates ratio of 1, and the grey dashed line is thea prioridetermined level of non-inferiority (08). The solid black line repre-
sents the mean of the data and is above the threshold for non-inferiority; SpDP is non-inferior to FFP in platelet adhesion and thrombus formation.
haemostasis in individuals with decreased levels of VWF thrombus formation when diluted 50% by T3VWD-P.
as well as FFP. This donor had normal levels of VWF; however, it is
Lyophilized plasma has been used for decades in the possible that the VWF multimer distribution made this
battlefield; however, it is limited to production by spe- donor’s plasma more susceptible to degradation during
cialized companies rather than at blood centres. SpDP the drying process. A potential mitigation for this issue
can be made rapidly from single donor plasma units as production of SpDP moves forward would be the
and the spray-drying process can be done at local screening of donor multimers.
blood centres ad hoc. As with other plasma-based and In combination with other pre-clinical studies [12,19]
cellular products, SpDP is subject to the inter-donor that have shown SpDP to be equivalent to FFP
variability, which may have an impact on its efficacy. with respect to thrombin generation and endothelial
One batch of the SpDP (denoted by the ■ in Figs. 2 interactions, this study supports the development of in-
and 3) was consistently inferior compared to FFP pro- human studies to evaluate the efficacy and safety of
duced from the same donor. Additionally, FFP from this SpDP in preventing and reversing trauma-related coag-
same donor demonstrated a minimal support of ulopathy.
Fig. 3 Thrombus formation in samples using SpDP or FFP diluted with plasma from an individual with Type 3 von Willebrand Disease (VW factor-free
plasma, T3VWD-P). (a) Thrombus formation in samples of recombined whole blood (rWB) made with SpDP or FFP diluted by 50% with plasma T3VWD-P.
The solid grey line indicates ratio of 1, and the grey dashed line is thea prioridetermined level of non-inferiority (0.8). (b) Thrombus formation in rWB
made with SpDP diluted with 25, 50 and 75% T3VWD-P. Both 75 and 50% SpDP had the same surface area coverage as 100% SpDP, whereas 25% SpDP
was decreased (n = 3).
References
1 Eastridge BJ, Mabry RL, Seguin P, 5 Hess JR, Brohi K, Dutton RP, et al.: Forces Health Service. Transfusion
et al.: Death on the battlefield (2001– The coagulopathy of trauma: a review 2013; 53(Suppl 1):65S–71S
2011): implications for the future of of mechanisms. J Trauma 2008; 11 Liu QP, Carney R, Sohn J, et al.: Sin-
combat casualty care. J Trauma Acute 65:748–754 gle-donor spray-dried plasma. Trans-
Care Surg 2012; 73:S431–S437 6 Brohi K, Singh J, Heron M, et al.: fusion 2019; 59:707–713
2 Tisherman SA, Schmicker RH, Brasel Acute traumatic coagulopathy. J 12 Wataha K, Menge T, Deng X, et al.:
KJ, et al.: Detailed description of all Trauma 2003; 54:1127–1130 Spray-dried plasma and fresh frozen
deaths in both the shock and traumatic 7 Pusateri AE, Given MB, Schreiber MA, plasma modulate permeability and
brain injury hypertonic saline trials of et al.: Dried plasma: state of the science inflammation in vitro in vascular
the Resuscitation Outcomes Consor- and recent developments. Transfusion endothelial cells. Transfusion 2013; 53
tium. Ann Surg 2015; 261:586–590 2016; 56(Suppl 2):S128–S139 (Suppl 1):80S–90S
3 Borgman MA, Spinella PC, Perkins JG, 8 Booth GS, Lozier JN, Nghiem K, et al.: 13 Rizoli SB, Scarpelini S, Callum J,
et al.: The ratio of blood products Spray: single-donor plasma product et al.: Clotting factor deficiency in
transfused affects mortality in patients for room temperature storage. Transfu- early trauma-associated coagulopathy.
receiving massive transfusions at a sion 2012; 52:828–833 J Trauma 2011; 71:S427–S434
combat support hospital. J Trauma 9 Flosdorf EW, Mudd S: Procedure and 14 Tanaka KA, Key NS, Levy JH: Blood
2007; 63:805–813 apparatus for preservation in “lyo- coagulation: hemostasis and thrombin
4 Holcomb JB, Tilley BC, Baraniuk S, phile” form of serum and other biolog- regulation. Anesth Analg 2009;
et al.: Transfusion of plasma, platelets, ical substances. J Immunol 1935; 108:1433–1446
and red blood cells in a 1:1:1 vs a 29:389–425 15 Schneppenheim R, Budde U: von
1:1:2 ratio and mortality in patients 10 Sailliol A, Martinaud C, Cap AP, et al.: Willebrand factor: the complex molec-
with severe trauma: the PROPPR ran- The evolving role of lyophilized ular genetics of a multidomain and
domized clinical trial. JAMA 2015; plasma in remote damage control multifunctional protein. J Thromb
313:471–482 resuscitation in the French Armed Haemost 2011; 9(Suppl 1):209–215
16 Leebeek FW, Eikenboom JC: Von properties. Transfusion 2019; 59:714– induced platelet aggregation. Thromb
Willebrand’s disease. N Engl J Med 722 Res 1989; 54:253–260
2016; 375:2067–2080 20 Bercovitz RS, Brenner MK, Newman 23 Ruggeri ZM, Mendolicchio GL: Adhe-
17 Rodeghiero F, Castaman G, Dini E: DK: A whole blood model of thrombo- sion mechanisms in platelet function.
Epidemiological investigation of the cytopenia that controls platelet count Circ Res 2007; 100:1673–1685
prevalence of von Willebrand’s dis- and hematocrit. Ann Hematol 2016; 24 Philipose S, Konya V, Sreckovic I,
ease. Blood 1987; 69:454–459 95:1887–1894 et al.: The prostaglandin E2 receptor
18 Agren A, Holmstrom M, Schmidt DE, 21 Sheu JB, Ko WC, Hung WC, et al.: Inter- EP4 is expressed by human platelets
et al.: Monitoring of coagulation fac- action of thrombin-activated platelets and potently inhibits platelet aggrega-
tor therapy in patients with von Wille- with extracellular matrices (fibronectin tion and thrombus formation. Arte-
brand disease type 3 using a and vitronectin): comparison of the activ- rioscler Thromb Vasc Biol 2010;
microchip flow chamber system. ity of Arg-Gly-Asp-containing venom 30:2416–2423
Thromb Haemost 2017; 117:75–85 peptides and monoclonal antibodies 25 Rasband WS: ImageJ. Bethesda, MD:
19 Meledeo MA, Liu QP, Peltier GC, et al.: against glycoprotein IIb/IIIa complex. J NIH, 1997–2015. https://imagej.nih.
Spray-dried plasma deficient in high- Pharm Pharmacol 1997; 49:78–84 gov/ij/. [Last accessed on 2017
molecular-weight multimers of von 22 Fukuyama M, Sakai K, Itagaki I, et al.: November 15]
Willebrand factor retains hemostatic Continuous measurement of shear-
547
548 I. Kumukova et al.
eliminate donor white blood cell (WBC) side effects [9– Government Medical agency. Healthy donors who agreed
11]. Irradiation prevents transfusion-associated graft ver- to participate in the study gave WB as per standard col-
sus host disease, but does not affect other side effects of lection practice (volume of 450 – 20 ml in bags with
WBC presence such as febrile non-haemolytic transfusion 63 ml of citrate-phosphate-dextrose anticoagulant (IMU-
reactions and post-transfusion immunomodulation [12]. FLEX-WB-RP, Terumo, Belgium)). To determine RBCS
The only effective method for the prevention of infec- storage results, we randomized 50 WB units into two
tious disease transmission in the seronegative period is groups: pathogen-reduced and the gamma-irradiated con-
the reduction of pathogens in blood products. Such a sys- trol group. All were leukofiltered one hour after the dona-
tem is also potentially effective for the prevention of tion (IMUFLEX-WB-RP, Terumo, Belgium). WB units
untestable and new diseases and the adverse effects of assigned to the study group (PR-RBCS; n = 25) were
donor WBC [12–14]. Several pathogen reduction tech- pathogen-reduced at room temperature with RF + UV
nologies (PRT) are used in routine practice in many coun- (Mirasol PRT, Terumo BCT, Lakewood, CO, USA) as per
tries for the treatment of platelet products and plasma. the treatment protocol provided by the manufacturer. WB
However, it was difficult to develop a PRT for the pro- was transferred to an illumination bag with 35 ml of
cessing of red blood cell blood products, because of the riboflavin solution (500 lM) added and then exposed to
physical characteristics of red blood cells (RBC) All the ultraviolet light (wavelength 280–315 nm) for 42–64 min
same RBC-containing blood products are the most used in depending on the volume of the product (total dose 80 J/
clinical practice [13]. That is why there is great interna- ml RBC). WB was then transferred to a storage bag and
tional interest in the area of red cell pathogen reduction. separated manually into RBC and plasma after a hard
One PRT for whole blood (WB) can provide three spin centrifugation at 5000 g for 15 min at 4°C. The stor-
pathogen-reduced blood products. It is based on the com- age bags of the PRT system have a non-standard size,
bined action of riboflavin and ultraviolet (RF + UV) and therefore, they do not fit into the automatic extrac-
(Mirasol PRT, Terumo BCT, Lakewood, CO, USA). tor. One hundred ml of SAGM (Terumo, Tokyo, Japan)
Studies to evaluate the reduction of pathogen in WB was added to the RBC to prepare RBCS. From the units of
treated by RF + UV have shown promising results [14– WB assigned to the control group, RBCS was prepared
16], without toxicity in vitro and in animals [17,18]. according to a standard procedure (hard spin 5000 g,
Studies in humans also showed no safety problem and 15 min, t = 4°C) with separation using an automatic
transfusion criteria met FDA requirements [19,20]. plasma extractor (NOVOmatic LMB Technologie, Ger-
There has been extensive experience with RF + UV many) into RBC and plasma. RBC was suspended in
technology on platelets and plasma. In addition, pub- 100 ml of SAGM (saline-adenin-glucose-mannitol) (Ter-
lished data describe the components derived from WB umo, Japan) and immediately subjected to c-irradiation
treated by RF + UV [21,22]. Our interest focused on (Gammacell 3000 Elite, Best Theratronics, Canada) at a
pathogen-reduced red blood cell suspension (RBCS). dose of 25 Gy. RBCS were stored at a controlled tempera-
Although some have studied RF + UV-treated whole ture of 4 – 2°C for 28 days. Samples were taken from all
blood, our results show how the technology affects red WB samples (before treatment) after leukofiltration, as
blood cells and may interest others considering imple- well as from RBCS on day 0, 7, 14, 21 and 28 of storage.
mentation of the technology. We conducted a randomized On day 28, only bacterial culturing was performed.
prospective parallel study. The study included 2 stages. In The residual WBC count in the filtered WB and RBCS
the first stage, we determined the biochemical and meta- was calculated using the BD LeucocountTM Kit (BD Bio-
bolic parameters during the storage of the RBCS obtained sciences, USA) on a BD FACSCantoTM II flow cytometer
from RB + UV-treated WB (PR-RBCS) and compared these (BD Biosciences, USA) with BD FACS Diva Software. Hae-
data with gamma-irradiated RBCS (GI-RBCS), which was moglobin concentration (Hb) and haematocrit (Ht) of
the control group. In the second stage, we examined the RBCS were assessed on the Sysmex KX-21N (Sysmex Cor-
viability and proliferation of lymphocytes in pathogen-re- poration Japan). Potassium concentration was determined
duced WB (PR-WB) compared with gamma-irradiated (GI- on the Cobas c311 (Roche/Hitachi). The pH, glucose and
WB) and untreated WB (UT-WB). lactate concentrations were assessed using an
ABL800FLEX (Radiometer, Denmark). The free Hb was
assessed in supernatant, obtained from a sample after
Materials and methods
centrifugation at 1400 rpm for 2 min and measured using
The study protocol was reviewed and approved by the a HemoCue Plasma/Low Hb (HemoCue AB, Angelholm,
local Ethical Committee of Dmitry Rogachev National Sweden). Haemolysis was calculated by:
Medical Research Center of Pediatric Hematology Oncol- %Haemolysis = [(1-Ht) 9 Free Hb g/dL) 9 100]/ Total
ogy and Immunology and the applicable Russian Hb concentration (g/dL).
Examination for sterility was carried out after treat- group, 10 ml for gamma irradiation and remaining vol-
ment on day 0 and day 28 using BACTEC Peds Plus/F ume for pathogen reduction), and thus, three groups were
culture vials (Becton Dickinson) and a BACTEC-9050 bac- formed: untreated (UT-WB, n = 35), gamma-irradiated
teriological analyser (Becton Dickinson). (GI-WB, n = 35) and pathogen-reduced samples (PR-WB,
The osmotic resistance was measured using hypotonic n = 35). Gamma irradiation and pathogen reduction were
sodium chloride solutions with different osmolality as performed on the day of preparation as described above.
described in Shcherbachenko I [23] on a plate spectropho- For assessment of proliferative activity, 1 9 107
tometer Anthos Zenyth 340rt (Biochrom, United King- mononuclear cells were isolated on a density gradient
dom). The osmotic resistance (Hc50) was characterized by and treated by CFSE (Carboxy-fluorescein-succinimidyl-
the osmolality of the solution, at which 50% of the origi- ester) in 1 ml of RPMI (Roswell Park Memorial Institute)
nal RBC were lysed. The width of the osmotic fragility medium, for 5 min at room temperature in the dark. The
distribution (WD) was characterized by the difference of amount of the dye was 05 ll. Then, the cells were
the osmolality, at which 10% and 90% of the original washed twice with PBS with 10% foetal bovine serum
RBC were lysed. (FBS) (STEMCELL Technologies, USA). One hundred
microlitres labelled cells were added at a number of
5 9 104 cells with the addition of 100 ll RPMI (negative
Adenosine triphosphate (ATP) measurement
control) or 10 lg/ml phytohemagglutinin (PHA).
Perchloric extracts were obtained: 025 ml of WB or mix- Untreated WB served as a positive control. The incubation
ture of 0125 ml of RBCS and 0125 ml of PBS were lasted three days. To determine the lymphocytes viability,
mixed with 05 ml of 05 M perchloric acid, and then, 7-AAD was added to the CFSE-labelled cells, which pene-
70 ll of a 5M K2CO3 was added. The sample was cen- trates apoptotic cells. The measurements were carried out
trifuged for 5min at 1000g. The supernatant was removed on a BD FACSCanto II flow cytometer (Becton Dickinson,
and frozen at 80°C. ATP was measured as described in USA) using Diva software.
Hans A [24] on the plate spectrophotometer. Statistical analysis was performed using XLSTAT for
Microsoft Excel 2010 and QI Macros 2018. The normality
of the distribution was estimated by the Shapiro–Wilk
Reduced glutathione (GSH) measurement
test. If the data were not normal distributed, the Box–Cox
Samples (WB or RBCS) were fast frozen and stored at - transformation was applied. The level of significance in
80°C. Then, samples were thawed and protein-free all calculations was taken to be a = 005.
extracts were obtained: to 100 ll of a sample, 1 ml of In the first stage of the study, WB and RBCS parame-
distilled water and 1 ml of meta-phosphoric acid ters were compared using the unpaired Student t-test or
(167 mg/ml meta-phosphoric acid, 2 mg/ml EDTA, the Mann–Whitney z-test, when data cannot be trans-
300mg/ml NaCl) were added. Extracts were mixed thor- formed. Comparison of lymphocytes proliferation and
oughly for 2min and centrifuged for 5 min at 1000 g. viability was performed using the Kruskal–Wallis H-test.
The supernatant was removed, and measurement was per-
formed as described in Pradedova E [25] on a plate spec-
Results
trophotometer.
RBC parameters
Phosphatidylserine expression (PS) measurement
Whole blood prior to treatment had comparable values in
Samples were diluted 1:1000 with HEPES buffer (10 mM both groups (Table 1). The residual WBC count after
HEPES; 150 mM NaCl; 25 mM CaCl2; pH 74; Sigma- leukofiltration in all samples was lower than 1 9 106/
Aldrich, USA)). 2 µL of Annexin V (Sony Biotechnology, Unit and did not differ between the compared groups
USA) was added to the 40 µL mixture and incubated for both in WB and in RBCS on day 0. Volume, Hb and Ht of
10 min at room temperature. Then, the samples were WB and RBCS met the requirements of the orders of the
diluted by adding a 360 µL buffer and analysed on a Ministry of Health and the Government of the Russian
Novocyte (Acea Bioscience, USA). Federation in relation to donated blood products. The
For the second stage of the study for determination of mean – SD of all the studied values of RBCS at different
the proliferation and viability of lymphocytes, we used storage days are presented in Table 2.
another 35 samples of unfiltered WB obtained from The Hb in the RBCS did not change during storage and
healthy donors (450 ml of WB (–10%) + 63 ml of CPD). did not differ between groups. The Ht was significantly
Leukofiltration was not performed in this stage. Each increased during storage in the GI-RBCS on day 14
sample was divided into three aliquots (10 ml untreated (p < 00001) and in the PR-RBCS on day 7 (p = 002),
Table 1 Initial variables of leukofiltered WB samples from study (PR) and control (GI) groups before treatment.1
Volume, ml 5093 – 84 5012 – 402 097 pH 706 – 006 705 – 004 066
Residual WBC, x106/Unit 015 – 019 012 – 017 047 К+, mmol/L 32 – 03 31 – 03 034
Hb, g/L 1238 – 106 1223 – 127 090 Glucose, mmol/L 209 – 09 208 – 13 090
Ht, % 373 – 28 369 – 31 060 Lactate, mmol/L 42 – 09 40 – 07 026
Free Hb, g/L 00 – 01 00 – 01 091 ATP, mmol/L RBC 102 – 023 100 – 028 076
Haemolysis, % 001 – 002 001 – 003 1 GSH, mmol/L RBC 258 – 041 249 – 049 062
H50, mOsm/L 1371 – 59 1349 – 51 032 PS expression, % 028 – 025 027 – 020 0.99
WD, mOsm/L 417 – 147 407 – 209 077
p-values were obtained by the Student t-test or Mann–Whitney U-test depending on the distribution law.
WBC = white blood cell, Hb = haemoglobin concentration, Ht = haematocrit, Free Hb = free haemoglobin; Hc50 = osmotic resistance, WD = width of
the osmotic fragility distribution, K+ = concentration of extracellular potassium, ATP = adenosine triphosphate, GSH = reduced glutathione, PS = phos-
phatidylserine.
1
For all parameters, the mean – standard deviations are presented.
Table 2 Variables changes during storage in study (PR-RBCS) and control (GI-RBCS) groups.1
Volume, ml 3072 – 167 3000 – 144 2972 – 167 2908 – 148 2922 – 167 2858 – 148 2872 – 167 2808 – 148
Hb, g/L 1978 – 117 1979 – 98 1983 – 117 1986 – 95 1990 – 104 1988 – 98 1992 – 117 1983 – 94
Ht, % 590 – 21 595 – 23 595 – 23 612 – 27* 619 – 23 644 – 30* 633 – 24 665 – 31*
pH 700 – 005 700 – 005 677 – 004 676 – 021* 667 – 003 663 – 004* 657 – 005 655 – 003
К+, mmol/L 20 – 02 20 – 03 450 – 41 452 – 48 516 – 53 517 – 51 651 – 46 647 – 42
Free Hb, g/L 03 – 02 01 – 01 09 – 03 09 – 06 15 – 04 22 – 10* 20 – 06 44 – 22*
Haemolysis, % 008 – 013 002 – 003 019 – 006 017 – 009 028 – 007 038 – 012* 036 – 010 072 – 037*
ATP, mmol/L RBC 112 – 038 119 – 033 109 – 026 114 – 023 108 – 024 102 – 021 090 – 026 081 – 024
GSH, mmol/L RBC 233 – 035 234 – 035 238 – 052 227 – 035 216 – 039 210 – 030 198 – 034 196 – 033
PS expression, % 022 – 011 023 – 017 029 – 022 024 – 016 045 – 042 032 – 027 021 – 014 051 – 05*
Glucose, mmol/L 283 – 12 269 – 11* 237 – 14 220 – 15* 204 – 14 190 – 15* 179 – 20 168 – 18*
Lactate, mmol/L 34 – 05 31 – 05 123 – 15 95 – 20* 188 – 22 154 – 29* 220 – 19 199 – 31*
H50, mOsm/L 1388 – 43 1403 – 51 1364 – 49 1422 – 71* 1386 – 49 1484 – 82* 1412 – 54 1553 – 84*
WD, mOsm/L 328 – 95 363 – 91 604 – 19 563 – 221 645 – 154 688 – 223 737 – 233 745 – 177
p-values were obtained by the Student t-test or Mann–Whitney U-test depending on the distribution law.
Hb = haemoglobin concentration, Ht = haematocrit, K+ = concentration of extracellular potassium; Free Hb = free haemoglobin, ATP = Adenosine
triphosphate, GSH = reduced glutathione, PS = phosphatidylserine, Hc50 = osmotic resistance, WD = width of the osmotic fragility distribution.
1
For all parameters, the mean – standard deviations are presented.
*Differences between the study and the corresponding control group are statistically significant.
compared with on day 0. Significant differences between equal or exceeded 08%. In the GI-RBCS, haemolysis
the groups were detected from day 7 (p = 003) and per- was below the threshold level for the entire observa-
sisted until the end of the observation period (on day 14 tion period.
p = 0002, on day 21 p = 00002). The extracellular potassium concentration increased
The free Hb and the haemolysis level (Fig. 1A) sig- significantly over storage, but did not differ between
nificantly increased over storage in both groups and groups. The pH progressively decreased in both groups.
significantly differed between the groups from the Statistically significant differences between groups in pH
14th day of storage (for free Hb, p = 00003, for were found on days 7 (p = 002) and 14 (p = 0002) of
haemolysis, p = 00004). Following 21 days of storage storage with lower level in the PR-RBCS, but by the 21st
in 9 PR-RBCS (36%), percentage of haemolysis was day these differences were levelled (p = 01).
Figure 1 Changes during the storage of the haemolysis level (A), osmotic resistance (B) and phosphatidylserine expression (C) of pathogen-reduced (PR)
and gamma-irradiated (GI) samples in the first stage of the study. The boundaries of the box are the 25th and 75th percentiles, respectively. The line in
the middle of the box is the median. The ends of the whiskers are the edges of a statistically significant sample (no outliers). Crosses indicate outliers.
*Differences between the groups are statistically significant (p < 005). p-values were obtained by the Student t-test or Mann–Whitney U-test depend-
ing on the distribution law. [Colour figure can be viewed at wileyonlinelibrary.com]
The extracellular glucose concentration at day 0 was p = 0006). In GI-RBCS, changes in the Hc50 during the
significantly higher in the control group (p = 00002). storage did not reach statistical significance compared
These differences persisted throughout storage (day 7 with day 0 after treatment. In the PR-RBCS, Hc50 signifi-
p < 00001; day 14 p = 0001; day 21 p = 003). Lactate cantly worsened on day 14 compared to days 0 and 7
concentration was significantly lower in the PR-RBCS (p = 0005), and between 14th and 21st days of storage
from day 7 (p < 00001), and these differences persisted (p = 0005). The Hc50 in the PR-RBCS were significantly
until the end of the follow-up period (day 14 p < 00001; worse than Hc50 in the GI-RBCS from day 7 (p = 0002).
day 21 p = 0006). However, the WD did not differ significantly between
Immediately after treatment, the Hc50 (Fig. 1B) of groups in any of the points. Immediately after treatment,
RBCS was significantly worse in both groups compared to the WD did not change in either WB or RBCS, although
WB (in the study group, p = 0046; in the control group, on the day 7 a significant increase in the WD occurred
(p < 00001) in both groups of RBCS. Subsequently, the enhances damage to blood components [26]. All changes
WD in the PR-RBCS increased significantly between the in RBCS found in our study are typical of the red cell
7th and 14th day of storage, and in the GI-RBCS between storage lesion.
7th and 21st day. The reason for the initially low concentration of glucose
Immediately after treatment, the GSH concentration in the PR-RBCS was the use of manual separation, which
decreased and the ATP concentration increased in both uses less preservative solution, although the differences in
groups RBCS compared to WB. In the PR-RBCS, these the volume of RBCS units did not reach statistical signifi-
changes reached statistical significance (for ATP p = 005, cance (p = 01) due to the small sample size. Our study, as
for GSH p = 003), but not in the GI-RBCS. ATP and GSH well as others [19,22], showed a decrease in glucose con-
concentrations did not differ between groups at any sumption in PR-RBCS, in contrast to control. The lactate
point. Interestingly, a statistically significant decrease of production in the control was higher. The high production
the ATP and GSH compared with the values on day 0 in of lactate should theoretically lower the pH in this group
the PR-RBCS was observed from day 14 of storage (for (control) lower than in the experimental group. But that
ATP p = 002; for GSH p = 0014) and in the GI-RBCS did not happen. It is not just lactate that affects pH. In the
from day 21 (for ATP p = 0007, for GSH p = 0001). experimental group, the pH was lower at the middle of
PS expression (Fig. 1C) did not significantly change storage (day 7 and day 14), but the differences between the
after both treatments compared to WB. In the GI-RBCS, a groups disappeared towards the end of storage.
significant increase in PS expression was observed on the All anticoagulants and preservative solutions are acidic
14 day, compared to day 0 (p = 0016), but in the PR- (CPD pH = 56; SAGM pH = 57). Therefore, after dona-
RBCS only on 21 days (p = 0008). Statistically signifi- tion, the pH of venous blood decreases which leads to
cant intergroup differences were found only on the 21st increase in ATP production [27]. This explains the higher
day of storage with higher level in the PR-RBCS ATP concentration in RBCS of both groups on day 0
(p = 0003). compared to WB. During storage, ATP production
All samples of both groups were negative for the bac- decreases, and its concentration in the RBC progressively
terial culture on day 0 and day 28. drops. Although the intergroup differences in ATP con-
centration did not reach statistical significance, the study
group had a more pronounced reduction in ATP than the
The results of the lymphocytes viability and
control group. The probable reason is decreased glycoly-
proliferation
sis. Faster depletion of ATP in the study group is impor-
Lymphocyte viability was significantly reduced during tant: as the concentration of ATP affects the recovery and
storage in all three groups (Fig. 2). PR-WB and GI-WB survival of RBC after transfusion [19].
showed significantly lower viability of lymphocytes com- Storage lesion leads to the spherocytic and echinocytic
pared to UT-WB (p < 005). Although throughout the fol- morphology of RBC. These changes showing resistance of
low-up period the lymphocytes viability after the RBC packaging in microhematocrit tube during centrifuga-
treatment by PRT was lower than after gamma irradia- tion and does not reflect actual haematocrit of the donor
tion, no significant differences between the treatment on the day of blood donation [28]. Thus, the increase of the
methods were found (p > 005). haematocrit found in our study may indicate more pro-
The spontaneous proliferative activity of untreated nounced changes in the morphology of PR-RBCS. We do
lymphocytes (Fig. 3) was 12%, which could be sporadic not believe that the introduction of a slightly lower volume
proliferation. After stimulation by PHA, the number of of SAGM significantly affected the haematocrit, if only in
proliferating cells increased to 460%. In gamma-irradi- part, because significant differences in Ht between groups
ated samples, spontaneous proliferation of lymphocytes were found from 7 days of storage.
did not significantly differ from untreated samples Hc50 demonstrates RBC resistance to osmotic stress.
(p > 005) and amounted to 10%. PHA stimulation The WD characterizes homogeneity of the RBC population
induced proliferation in only 53% of the cells in this by its ability to withstand hypo-osmotic stress. The
group. In the pathogen-reduced samples, spontaneous and greater WD the less homogeneity of the RBC in osmotic
stimulated proliferation of lymphocytes did not exceed resistance. Both treatments significantly influenced the
01% of the cells. Hc50, but in the PR-RBCS the deterioration was more
pronounced, which indicates less stability and a greater
tendency to haemolysis pathogen-reduced RBC in the
Discussion
early days of storage. The results of the changes in the
The use of additional methods for the processing of blood WD demonstrate a deeper damaging effect of PRT on the
products such as irradiation and reduction of pathogens mechanisms of cell protection from osmotic stress.
Figure 2 The viability of lymphocytes at 0, 1, 2, 3 days of storage, depending on the type of treatment. Values indicate the mean percentage of viable
lymphocytes. The ends of the whiskers show standard deviation. After both treatments (gamma irradiation and pathogen reduction) viability of lympho-
cytes statistically significantly lower than in untreated samples (p < 005). No significant differences between the treatment methods in lymphocyte via-
bility (p > 005)). p-values were obtained by the Kruskal–Wallis H-test.
GSH performs an antioxidant and detoxification func- of PR-RBCS will be no worse than that after GI-RBCS of
tion in RBC [29]. Both treatments (gamma irradiation and corresponding shelf life.
PRT) increase peroxidation and decrease glycolysis, which The greater haemolysis in the PR-RBCS compared with
leads to a drop in the concentration of GSH. In the PR- untreated, and GI-RBCS has been described in many stud-
RBCS, the peroxidation processes are more pronounced ies [30,31]. It was the indicator that limited shelf life in
than in untreated RBCS [30]. But in our study, as in Qadri our study. We used a haemolysis level below 08% of
et al. 2017 [31], differences in GSH concentration RBC to define the shelf life of PR-RBCS as 14 days.
between groups did not reach statistical significance, Our study showed that the PR-RBCS can be used for
although the dynamics of decline was more pronounced transfusions for 14 days of storage, during which it
in PR-RBCS. demonstrates quality.
The mean binding of annexin V did not exceed 1% in Suppression of lymphocytes proliferation and viability
both groups. Compared with the day 0 in the GI-RBCS, demonstrate that this PRT can serve as an alternative to
increase of the PS expression on RBC was observed ear- irradiation of blood products.
lier than in the PR-RBCS (day 14). Also, on the 14th day Perhaps modification of the RF + UV system will lead
of storage, significant differences were found in haemoly- to an extension of the storage time and the quality of the
sis, which was higher in the PR-RBCS. Possibly, the dif- PR-RBCS [19]. If the manufacturer could produce storage
ferences in the PS expression was due to pathogen- bags to fit automatic fractionation devices, this would be
reduced RBC (probably more fragile populations) under- helpful to transfusion personnel.
going more destruction while the remaining more stable
cells had a lower expression of PS than the gamma-irra-
Conflicts of interest
diated RBC, in which more cells survived to day 14.
According to our data, the incidence of post-transfu- The authors declare that they have no conflicts of
sion hyperkalemia and its adverse effects after transfusion interest.
Not-stimulated PHA-stimulated
(b)
(a)
Untreated WB
(d)
(c)
Gamma-irradiated WB
(e)
(f)
Pathogen-reduced WB
Figure 3 Proliferation of lymphocytes isolated from untreated WB (A, B), gamma-irradiated WB (C, D) and pathogen-reduced WB (E, F), non-stimulation
(A, C, E) and stimulated by PHA (B, D, F), after incubation for three days. In gate, P2 indicates the percentage of proliferating cells. The spontaneous pro-
liferative activity of lymphocytes in untreated (A) and gamma-irradiated (C) samples did not significantly differ (p > 005). PHA-stimulated lymphocytes
proliferation was significant higher in untreated samples (B) than in gamma-irradiated (D) (p < 005). Spontaneous (E) and PHA-stimulated (F) lympho-
cytes proliferation in RF + UV-treated samples were significant lower than in untreated and in gamma-irradiated samples (p < 005). p-values were
obtained by the Kruskal–Wallis H-test.
References
1 Stramer SL, Dodd RY: Transfusion- allergic reactions to RBCs and platelets 20 Cancelas JA, Rugg N, Fletcher D,
transmitted emerging infectious dis- after conversion to universal prestor- et al.: In vivo viability of stored red
eases: 30 years of challenges and pro- age leukoreduction. Transfusion 2004; blood cells derived from riboflavin
gress. Transfusion 2013; 53:2375–83 44:16–24 plus ultraviolet light-treated whole
2 Busch MP: Transfusion-transmitted 12 Ohto H: Gamma radiation does not blood. Transfusion 2011; 51:1460–8
viral infections: building bridges to prevent transfusion induced HLA 21 Hervig T, Braathen H, Jaboori AA,
transfusion medicine to reduce risks alloimmunization. Transfusion 1997; et al.: Platelet recovery and survival
and understand epidemiology and 37:878–9 after whole blood treated with mirasol
pathogenesis. Transfusion 2006; 13 Seltsam A: Pathogen inactivation of pathogen reduction. Transfusion 2016;
46:1624–1640 cellular blood products - an additional 56:3A–262A
3 Pruett CR, Vermeulen M, Zacharias P, safety layer in transfusion medicine. 22 Schubert P, Culibrk B, Karwal S, et al.:
et al.: The use of rapid diagnostic tests Front Med (Lausanne) 2017; 4:219. Whole blood treated with riboflavin
for transfusion infectious screening in https://doi.org/10.3389/fmed.2017.00219 and ultraviolet light: quality assess-
Africa: a literature review. Transfus 14 Yonemura S, Doane S, Keil S, et al.: ment of all blood components pro-
Med Rev 2015; 29:35–44 Improving the safety of whole blood- duced by the buffy coat method.
4 Glynn SA, Busch MP, Dodd RY, et al.: derived transfusion products with a Transfusion 2015; 55:815–23. https://
Emerging infectious agents and the riboflavin-based pathogen reduction doi.org/10.1111/trf.12895
nation’s blood supply: responding to technology. Blood Transfus 2017; 23 Shcherbachenko I, Lisovskaya I,
potential threats in the 21st century. 15:357–364. https://doi.org/10.2450/ Tikhonov V: Oxidation-induced cal-
Transfusion 2013; 53:438–54 2017.0320-16 cium-dependent dehydration of nor-
5 Stramer SL, Hollinger FB, Katz LM, 15 Goodrich RP, Doane S, Reddy HL: mal human red blood cells. Free
et al.: Emerging infectious disease Design and development of a method Radical Res 2007; 41:536–45
agents and their potential threat to for the reduction of infectious patho- 24 Hans A: Adenosine-5 -triphosphate:
transfusion safety. Transfusion 2009; gen load and inactivation of white determination with phosphoglycerate
49(Suppl 2):1S–29S. https://doi.org/10. blood cells in whole blood products. kinase; in Bergmeyer H-U (ed): Meth-
1111/j.1537-2995.2009.02279.x Biologicals 2010; 38:20–30 ods of Enzymatic Analysis (Second
6 Lanteri MC, Kleinman SH, Glynn SA, 16 Cap AP, Pidcoke HF, Keil SD, et al.: Printing, Revised), Academic Press,
et al.: Zika virus: a new threat to the Treatment of blood with a pathogen 1965:539–558, ISBN 9780123956309.
safety of the blood supply with world- reduction technology using ultraviolet Available at: https://doi.org/10.1016/
wide impact and implications. Trans- light and riboflavin inactivates Ebola B978-0-12-395630-9.50112-2.
fusion 2016; 56:1907–14. https://doi. virus in vitro. Transfusion 2016; 56 25 Pradedova E, Nimaeva O, Putilina T,
org/10.1111/trf.13677 (Suppl 1):S6–15. https://doi.org/10. et al.: Determination of glutathione
7 Harrington T, Kuehnert MJ, Kamel H, 1111/trf.13393 and its redox status in isolated vac-
et al.: West Nile virus infection trans- 17 Doane SK, Yonemura S, Hovenga N, uoles of red beetroot cells. J Stress
mitted by blood transfusion. Transfu- et al.: Evaluation of the acute toxicity Physiol Biochem 2016; 12:87–107
sion 2003; 43:1018–22. https://doi. of red blood cells derived from ribo- 26 Hess J: Red cell storage. J Proteom
org/10.1046/j.1537-2995.2003.00481.x flavin and UV light-treated whole 2010; 73:368–373
8 Gallagher L, Ganz P, Yang H: Advanc- blood in a canine red blood cell 27 Simon TL, McCullough J, Snyder EL,
ing risk assessment for emerging exchange model. Transfusion 2016; et al.: Strauss RG Rossi’s Principles of
infectious diseases for blood and blood 56:193A Transfusion Medicine. Oxford, UK:
products: proceedings of a public 18 Goodrich RP, Murthy KK, Doane SK, Wiley-Blackwell, 2016:760
workshop. Transfusion 2013; 53:455– et al.: Evaluation of potential immune 28 Mustafa I, Marwani A, Mamdouh N,
463. response and in vivo survival of ribo- et al.: Time dependent assessment of
9 Akahoshi M, Takanashi M, Masuda M, flavin-ultraviolet light-treated red morphological changes: leukodepleted
et al.: A case of transfusion-associated blood cells in baboons. Transfusion packed red blood cells stored in
graft-versus host disease not prevented 2009; 49:64–74. https://doi.org/10. SAGM. Biomed Res Int 2016;
by white cell-reduction filters. Trans- 1111/j.1537-2995.2008.01940.x 2016:4529434. https://doi.org/10.
fusion 1992; 32:169–72 19 Cancelas JA, Slichter SJ, Rugg N, 1155/2016/4529434
10 Fisher M, Chapman JR, Ting A, et al.: et al.: Red blood cells derived from 29 Dumaswala U, Wilson M, Wu Y, et al.:
Alloimmunisation to HLA antigens whole blood treated with riboflavin Glutathione loading prevents free radi-
following transfusion with leucocyte- and ultraviolet light maintain ade- cal injury in red blood cells after stor-
poor and purified platelet suspensions. quate survival in vivo after 21 days age. Free Radical Res 2000; 33:517–
Vox Sang 1985; 49:331 of storage. Transfusion 2017; 57: 529
11 Paglino JC, Pomper GJ, Fisch GS, 1218–1225. https://doi.org/10.1111/trf. 30 Chen D, Schubert P, Devine D: Identi-
et al.: Reduction of febrile but not 14084 fication of potential protein quality
markers in pathogen inactivated and 31 Qadri S, Chen D, Schubert P, et al.: lesion and promotes eryptosis. Trans-
gamma - irradiated red cell concen- Pathogen inactivation by riboflavin fusion 2017; 57:661–673
trates. Clin Appl Thromb Hemost and ultraviolet light illumination
2017; 11:7–8 accelerates the red blood cell storage
Araci M. Sakashita,1 Patrıcia Scuracchio,3 Edison Luiz Durigon,2,4 Silvano Wendel3 & Jose M. Kutner1
1
Hospital Israelita Albert Einstein, S~ao Paulo, Brazil
2
Department of Microbiology, Institute of Biomedical Sciences, University of Sao Paulo, S~
ao Paulo, Brazil
3
Hospital Sırio Liban^e s, S~ao Paulo, Brazil
4
Scientific Platform Pasteur USP, S~ao Paulo, Brazil
557
558 V. de Dutra et al.
in blood group antigen expressions and presence or symptoms and needed hospitalization. Blood type infor-
absence of anti-A or -B provide strong defensive lines mation was available in the electronic chart. Samples for
against infection [8]. In persons with O blood type, char- anti-NP and nAbs were drawn by the time of admission.
acterized by the absence of A or B antigens, stimulation In order to avoid blood group bias from repeat donors,
by microbiota having glycan motifs similar to A or B given that blood type O (‘universal donor’) is usually
antigens, leads to a natural production of anti-A and over-represented, our control group (CG) comprised 2212
anti-B [9]. Type B individuals also produce anti-A but in first-time voluntary healthy blood donors from Hospital
lower titres [10]. Interestingly, in 2003, during the SARS- Israelita Albert Einstein blood bank database, who
CoV outbreak in China, the O blood group was considered donated whole blood from August to October 2019,
protective against infection—OR = 018 (004–081) [11]. before the COVID-19 outbreak in Brazil.
Later, a cell-binding assay showed that either a mono- Subjects of blood types A and AB were grouped as
clonal or human natural anti-A could inhibit the SARS- ‘without anti-A’ (A/AB group), whereas those with O and
CoV S protein/ACE2 interaction [12]. At a cellular level, B were named ‘with anti-A’ (O/B group).
this supports the idea that the type O protection against
SARS-CoV involves the antibodies rather than the anti-
Samples tests
gens.
Based on these findings, our study aimed to analyse Blood typing was done by the automated analyser, gel
the association of SARS-CoV-2 infection with the pres- technique, Erytra Eflexis (Grifols, Barcelona, Spain) and
ence of anti-A (In types O and B) or its absence (in types IH-platform (Biorad, Creisser, Switzerland).
A and AB), related to the production of antibodies to RT-PCR: A real-time reverse transcriptase polymerase-
SARS-CoV-2 nucleoprotein (NP; IgA, IgM and IgG) and chain-reaction technique confirmed SARS-CoV-2 diagno-
neutralizing antibodies (nAb). sis from naso-oropharyngeal swab specimens. Molecular
tests were based on Corman et al. [13], with five copies/
reaction of sensitivity.
Methods and materials
Neutralizing antibodies (nAbs) and anti-nucleocapsid
(anti-NP) antibodies (IgM, IgG and IgA): We used the
Ethics statement
cytopathic effect-based virus neutralization test (CPE-
The study was approved by both hospitals’ Institutional based VNT), which was carried out with SARS-CoV-2
Review Boards (IRB) and the Brazilian Commission on (GenBank: MT350282.1), as previously described [14,15].
Ethics and Research (CONEP) under requests CAAE Anti-NP was determined according to Wendel et al. [15].
32558220.0.0000.0071 and CAAE 30259220.4.2001.5461. Both methods have been described in more detail else-
All patients and COVID-19 convalescent plasma donors where [15].
provided written informed consent.
Data analysis
Subjects
Age was compared between control, CCPD and CIP
We analysed a retrospective cohort of 430 people with groups using the ANOVA test. For sex and blood type
COVID-19 [268 convalescent plasma donors (CCPD) and analysis we used the chi-square test. CCPD and CIP
162 inpatients (CIP)] from both hospitals. All patients and groups were classified as COVID-19 individuals’ group
plasma donors had a previous diagnosis confirmed by and compared to the CG.
RT-PCR. For further analysis regarding the presence or absence
The CCPD group comprised convalescent patients who of ‘circulating anti-A’ and its association with COVID-19,
had had mild symptoms (no hospitalization during their we merged our study population into two groups: one
COVID-19 evolution). Eligibility criteria required a posi- ‘with circulating anti-A’; including types O and B (O/B
tive diagnostic test by naso-oropharyngeal swab (NOS) group)” and another ‘without circulating anti-A’; includ-
RT-PCR and resolution of symptoms for at least 14 days. ing types A and AB (A/AB group).”
The candidates were then tested for SARS-COV-2 by RT- For adjusted models, we applied multiple logistic
PCR either on peripheral blood or NOS swab. If the RT- regression. The Mann–Whitney test and Spearman’s test
PCR was negative, the plasma was collected, and antibod- were used to compare anti-A presence/absence and anti-
ies to SARS-CoV-2 nucleoprotein (anti-NP - IgA, IgM and NP (IgM, IgG and IgA) among groups.
IgG) and neutralizing antibodies (nAb) were measured. Data were organized in Microsoft Excel 2010 and were
The CIP group was composed of patients with a posi- analysed using Statistical Package for Social Sciences
tive SARS-CoV-2 RT-PCR, who had moderate to severe (SPSS) (Chicago, IL) or GraphPad Prism version 8.0 for
Windows, GraphPad Software (La Jolla, CA, USA). A P- Table 2 Multivariate logistic regression for COVID-19 individuals
value <005 was considered significant. (n = 430) and blood donors (CG) (n = 2212)
95% CI
Results
Variable OR Lower Upper P
We had age, sex and blood type of all 430 COVID-19
individuals (268 CCPD and 162 CIP) and 2212 healthy Age (years) 106 105 106 <0001
Gender (male) 127 102 159 0035
volunteer blood donors (control group: CG) However, as
Anti-A (O/B) 062 050 078 <0001
they were from a retrospective cohort, the anti-NP and
nAbs were available in only 295 of the COVID-19 per- CG, control group; CI, confidence interval, OR, odds ratio.
sons. Table 1 shows the distribution, mean age, sex and Age and male gender were positively related to COVID-19 (OR = 106,
blood type among groups. Although blood type O most 95% CI: 105–106; P < 0001 and OR = 127, 95%CI: 102–159;
frequent among blood donors, type A was more common P = 0035 respectively). The presence of circulating anti-A (O/B group)
in the COVID-19 group. There was no statistical differ- showed a protective factor to COVID-19 (OR = 062, 95% CI; 050–078;
ence in blood type distribution between CCPD and CIP. P < 0001).
Age and male sex were positively related to COVID-19
(OR = 106, 95% CI: 105–106; P < 0001 and OR = 127,
95%CI: 102–159; P = 0035 respectively). However, the Additionally, we analysed whether anti-NP was associ-
presence of circulating anti-A (O/B group) showed a pro- ated with the presence of anti-A. We included 295 sub-
tective effect against COVID-19 (OR = 062, 95% CI: jects from the COVID-19 group (148 from O/B group and
050–078; P < 0001), as shown in Table 2. 147 in the A/AB group). Figure 2 shows the distribution
In order to evaluate the association of circulating anti- of anti-NP between the groups. The O/B group showed
A with COVID-19 severity, we compared CCPD and CIP to lower median IgM, IgG and IgA levels than for A/AB
the CG. Belonging to the O/B group was protective only (016 vs. 019; P = 003, 211 vs. 255; P = 002, 023 vs.
for CCPD, as shown in Fig. 1 (OR = 060, 95% CI: 045– 032; P = 003, respectively).
075 P < 0001). Figure 3 shows the nAb distribution analysis. As the
As type O persons usually have higher anti-A titre, we samples’ value varies from <1:20 to >1:5120, we divided
also performed a sub-analysis for O vs. B blood, as shown the nAb titres into two groups: <320 and ≥320 (320 as
in Table 3. the middle cut-off point for the reaction range). Patients
Age was positively related to COVID-19 (OR = 105, 95% of types O or B showed a lower trend of neutralizing anti-
CI: 104–106; P < 0001). Sex did not show any association body value and lower frequencies when the titres were
to COVID-19 between these groups. However, a higher titre higher than 320 (chi-square test = 699, P = 0008). The
of circulating anti-A (O group) was protective against results are shown in Table 4.
COVID-19 (OR = 066, 95% CI: 046–095; P = 0026). There was an evident linear correlation between anti-
NP (IgA, IgG and IgM) and nAbs according to the Spear-
man’s correlation test (all P < 00001), as shown in Fig. 4.
Table 1 Demographic data from subjects included in the study
We identified a better correlation between IgG and nAbs
Blood donors (CG) CCPD CIP for both O/B and A/AB groups (r = 0687 in O/B and
Variable (N = 2212) (N = 268) (N = 162) P r = 0640 in A/AB). IgM and IgA did not show such good
correlations (O/B group IgA r = 0593; IgM r = 0430 and
Agea (years; 375 – 120 368 – 81 693 – 157 <0001
A/AB group IgA r = 0555; IgM r = 0457).
Mean – SD)
Gender, n (%)b
Female 1004 (454) 103 (384) 59 (364) 0012 Discussion
Male 1208 (546) 165 (616) 103 (636)
Blood group, n (%)b The ABO system can be associated with the inflammatory
A 785 (355) 128 (478) 70 (432) <0001 response and has a varied geographical frequency, with
AB 84 (38) 11 (41) 7 (43) growing evidence that it can affect the predisposition to
O 1117 (505) 103 (384) 59 (364) certain diseases, such as thrombosis or H. pylori infection
B 226 (102) 26 (97) 26 (161) [7]. A meta-analysis recently determined the odds of
SARS-CoV-2 positive individuals of having a specific
CCPD, COVID-19 convalescent plasma donors; CG, control group; CIP,
blood type compared with controls. The association of
COVID-19 inpatients; N, number of participants.
a
ANOVA.
SARS-CoV-2 with blood type A was significant with a
b
Chi-Square. pooled OR of 123 (95%CI: 109–140), although the
Fig. 1 Odds ratio (OR) for CIP and CCPD compared to blood donors (CG),
considering O/B vs. A/AB groups. Legend: O/B group was identified as a
protective factor only for CCPD (OR = 060, 95% CI: 045–075), whereas
for CIP this effect was not identified (OR = 085, 95% CI: 045–125).
Multivariate logistic regression; CCPD, COVID-19 convalescent plasma
Fig. 2 COVID-19 individuals (n = 295) IgM, IgG, and IgA distribution
donors; CIP, COVID-19 inpatients; *P < 0001.
between A/AB and O/B groups. Legend: O/B group showed an IgM, IgG
and IgA median level lower when compared to A/AB (016 vs. 019;
Table 3 Multivariate logistic regression analysis for O vs. B blood groups, *P = 003, 211 vs. 255; #P = 002, 023 vs. 032; &P = 003, respec-
comparing COVID-19 patients and blood donors (CG) tively).Mann–Whitney test.
Fig. 3 Distribution of the frequencies of COVID-19 individuals (n = 295) neutralizing antibodies between A/AB and O/B groups. Legend: O/B group
showed a neutralizing antibody lower value trend and lower frequencies when the titres were higher than 320.
Table 4 Distribution of COVID-19 individuals (n = 295) neutralizing anti- respiratory failure in COVID-19, due to the rs657152 A/C
body (nAb) titres in two groups: <320 and ≥320
single nucleotide polymorphism at 9q34.2. Also, some
nAb titres
procoagulant markers are associated with genetic varia-
tion at the ABO locus, and this region could have a role
ABO Group <320 ≥320 v2 P OR (95% CI) in modifying genes [21].
Curiously, when we evaluated specific immunoglobulin
O/B 97 (655) 51 (345) 699 0008 053 (033- 085)*
production, we found a statistically significant difference
A/AB 74 (503) 73 (497)
for IgM, IgG and IgA results, with median values lower in
O/B COVID-19 individuals had a lower chance of having values ≥320. the O/B group. Similarly, neutralizing antibody titres were
*054 (034; 087), when adjusted by age and gender. lower in the O/B group.
This is one of the few studies that analyse a possible
[12,19]. Genetic research also has increased in this area. correlation between humoral response for SARS-CoV-2
Genome-wide association analysis in an Italian-Spanish and ABO group. In that way, even though we did not
group showed that A-positive people are at higher risk of observe a statistical difference between the groups
Fig. 4 Linear correlation among COVID-19 individuals (n = 295) immunoglobulins and neutralizing antibody titres (Log2), into the group A/AB (a) and
O/B (b). Legend: Spearman correlation for A/AB group (left) and for O/B group (right). A better correlation between IgG and neutralizing antibodies was
found in both O/B and A/AB (r = 0687 in O/B and r = 0640 in A/AB). IgM and IgA did not show such good correlations (O/B group IgA r = 0593; IgM
r = 0430 and A/AB group IgA r = 0555; IgM r = 0457).
CCPD and CIP concerning severity, the profile of was no statistical difference between COVID-19 inpatients
increased humoral immune response in A/AB patients is and COVID-19 convalescent plasma donors, suggesting
an intriguing question and may be related to other fac- that blood types do not associate with COVID-19 severity.
tors, such as the clinical evolution and progression of Moreover, COVID-19 individuals from types O and B had
the disease. Our study did not correlate clinical status lower titres of neutralizing antibodies and lower levels of
and level of IgM, IgG, IgA anti-NP or nAb titres; thus, IgM, IgG and IgA anti-nucleocapsid antibodies than did
this parameter correction for ABO status may be biased. the types A and AB.
IgA is an immune barrier and can probably neutralize
SARS-CoV-2 before the virus reaches and binds to the
Acknowledgements
epithelial cells. This has taken on increased importance
recently. IgM and IgG levels have a potential role in the C.P.S. is funded by Grant 2018/23680-0 (Fundacß~ao de
evaluation of severity and prognosis of COVID-19 [22]. In Amparo a Pesquisa do Estado de S~ao Paulo); D.B.A. by
37 patients, IgA and IgG levels were markedly higher Grant 88 887.131387/2016-00 (Coordenacß~ao de Aper-
(P < 0001) in patients with severe disease compared with feicßoamento Pessoal de Nıvel Superior - CAPES), R.R.G.M.
mild disease, while there was no difference in IgM level by Grant 2017/24769-2 (Fundacß~ao de Amparo a Pesquisa
[23]. Additionally, in a two-year prospective study, IgG do Estado de S~ao Paulo) and E.L.D. by Grants 2016/
antibody and NAb titres were positively correlated in 20045-7 and 2020/06409-1 (Fundacß~ao de Amparo a Pes-
SARS-CoV [24]. However, we did not find any research quisa do Estado de S~ao Paulo). This project was partially
correlating IgM and IgA, especially in SARS-CoV-2 infec- supported by the initiative ‘Todos Pela Saude - Fundacß~ao
tions. Future longitudinal studies can show if blood Itau para Educacß~ao e Cultura’.
groups interfere with protection antibodies, including the
long-term immune response.
Conflict of interests
The impact of anti-A isohemagglutinin titres in the
SARS-CoV-2 infection and its association with neutraliz- The authors declare no conflict of interests.
ing antibodies is not clear [25]. In our retrospective study,
we could not measure the quantitative effect of anti-A.
Author contributions
Further studies are required to evaluate both anti-A iso-
hemagglutinin and neutralizing antibody titres, and their Conceptualization – VFD, CBB, APHY, SW, JMK; Investi-
role in SARS-CoV-2 infection. gation – VFD, CBB, APHY, NH, JRRP, RFW, RMF, RRGM,
GC, AS, RA, PS, DBA, CPS, ELD, MA, VN, SW, JMK. For-
mal analysis – VFD, CBB, APHY, SW, JMK; Resources –
Conclusion
LFLR, LVR; Writing – VFD, CBB, APHY, SW, JMK. Project
Blood types O and B, which produce anti-A, showed a administration and funding acquisition – LFR, LVR.
protective effect against SARS-CoV-2 infection. There
References
1 Dong E, Du H, Gardner L. An interac- 2020;2020.03.11.20031096. https://doi. 9 Breiman A, Ruv€en-Clouet N, Le Pendu
tive web-based dashboard to track org/10.1101/2020.03.11.20031096 J. Harnessing the natural anti-glycan
COVID-19 in real time. Lancet Infect 5 Zaidi FZ, Zaidi ARZ, Abdullah SM, immune response to limit the trans-
Dis 2020;20:533–4. et al. COVID-19 and the ABO blood mission of enveloped viruses such as
2 Frater JL, Zini G, d’Onofrio G, et al. group connection. Transfus Apher Sci SARS-CoV-2. PLoS Pathog 2020;16:
COVID-19 and the clinical hematology 2020;59:102838. e1008556.
laboratory. Int J Lab Hematol 6 Gerard C, Maggipinto G, Minon J-M. 10 Kang SJ, Lim YA, Baik SY.
2020;2:11–8. COVID-19 & ABO blood group: Comparison of ABO antibody titers on
3 Caussy C, Pattou F, Wallet F, et al. another viewpoint. Br J Haematol the basis of the antibody detection
Prevalence of obesity among adult 2020;39:62–3. method used. Ann Lab Med
inpatients with COVID-19 in France. 7 Franchini M, Bonfanti C. Evolutionary 2014;34:300–6.
Lancet Diabetes Endocrinol aspects of ABO blood group in humans. 11 Cheng Y, Cheng G, Chui CH, et al.
2020;8:562–4 Clin Chim Acta 2015;444:66–71. ABO blood group and susceptibility to
4 Zhao J, Yang Y, Huang H-P, et al. 8 Yamamoto F, Cid E, Yamamoto M, severe acute respiratory syndrome.
Relationship between the ABO Blood et al. An integrative evolution theory JAMA 2005;293:1450–1.
Group and the COVID-19 susceptibil- of histo-blood group ABO and related 12 Guillon P, Clement M, Sebille V, et al.
ity. medRxiv. Published online genes. Sci Rep 2014;4:1–12. Inhibition of the interaction between
the SARS-CoV Spike protein and its meta-analysis. PLoS One 2020;15(9): 21 Ellinghaus D, Degenhardt F, Bujanda
cellular receptor by anti-histo-blood e0239508. L, et al. Genomewide association study
group antibodies. Glycobiology 17 Dai X. ABO blood group predisposes of severe Covid-19 with respiratory
2008;18:1085–93. to COVID-19 severity and cardiovas- failure. N Engl J Med 2020;383:1522–
13 Corman VM, Landt O, Kaiser M, et al. cular diseases. Eur J Prev Cardiol 34.
Detection of 2019 -nCoV by RT-PCR. 2020;27:1436–7. 22 Hou H, Wang T, Zhang B, et al. Detec-
Euro Surveill 2020;25:1–8. 18 Zietz M, Tatonetti NP. Testing the tion of IgM and IgG antibodies in
14 Araujo DB, Machado RRG, Amgarten association between blood type and patients with coronavirus disease
DE, et al. SARS-CoV-2 isolation from COVID-19 infection, intubation, and 2019. Clin Transl Immunol 2020;9:1–
the first reported patients in Brazil death. medRxiv 2020;2020.04.08. 8.
and establishment of a coordinated 20058073. https://doi.org/10.1101/ 23 Yu H, Sun B, Fang Z, et al. Distinct
task network. Mem Inst Oswaldo Cruz. 2020.04.08.20058073 features of SARS-CoV-2-specific IgA
2020;115. https://doi.org/10.1590/ 19 Arend P. How SARS-CoV-2 (COVID- response in COVID-19 patients. Eur
0074-02760200342 19) may invade the human body via Respir J 2020;56:2001526.
15 Wendel S, Kutner JM, Machado R, ABO(H) blood group-determining car- 24 Liu W, Fontanet A, Zhang PH, et al.
et al. Screening for SARS-CoV-2 anti- bohydrates: a. perspective. 2020;2 Two-year prospective study of the
bodies in convalescent plasma in Bra- (May):17. https://doi.org/10.20944/PRE humoral immune response of patients
zil: preliminary lessons from a PRINTS202005.0097.V1 with severe acute respiratory syn-
voluntary convalescent donor pro- 20 Novaretti MCZ, Dorlhiac-Llacer PE, drome. J Infect Dis 2006;193:792–5.
gram. Transfusion 2020. https://doi. Chamone DAF. Estudo de grupos 25 Daniele F. Anti-A Isohemagglutinin
org/10.1111/trf.16065 sang€uıneos em doadores de sangue titers and SARS-CoV2 neutralization:
16 Golinelli D, Boetto E, Maietti E, et al. caucas oides e negr
oides na cidade de implications for children and conva-
The association between ABO blood S~ao Paulo. Rev Bras Hematol Hemoter lescent plasma selection. Br J Haema-
group and SARS-CoV-2 infection: a 2000;22:23–32. tol 2020;190:e148–50.
Abstract
Background and objectives Taiwan is among the few hepatitis B virus (HBV)
high-endemic countries that implement universal mini-pool nucleic acid testing
(MP-NAT) and hepatitis B surface antigen (HBsAg) testing together with confir-
matory individual donor nucleic acid testing (ID-NAT) for its blood supply since
2013. The aim of this study was to reappraise the value of HBsAg test in Tai-
wan’s HBV testing strategy.
Materials and methods A Markov model was constructed, and cost-effectiveness
analysis was conducted in order to reappraise the existing HBV screening strat-
egy in Taiwan.
Results The incremental cost-effectiveness ratio (ICER) for the current testing
strategy in Taiwan was estimated to be $US 443 154 per quality-adjusted life
year (QALY) gained. This is almost six times the willingness-to-pay (WTP) thresh-
old that reflects local preferences.
Conclusion Universal HBsAg and MP-8-NAT together with confirmatory ID-NAT
testing prevents a significant amount of HBV infections from entering the Taiwan
Received: 30 June 2020,
blood supply. However, this comes at a disproportionate increase in cost.
revised 11 November 2020,
accepted 15 November 2020, Key words: hepatitis B virus, blood safety, HBsAg test, NAT testing, cost-effec-
published online 5 December 2020 tiveness analysis, Markov model.
564
Economic reappraisal of HBV blood donor tests 565
algorithms [9]. The current practice for the detection of principles or policy recommendations for transfusion
HBV in Taiwanese blood donors involves the use of uni- practitioners especially in high-endemic areas.
versal HBV-NAT and HBsAg testing. In Taiwan, anti-HBc
testing is not universal because deferring anti-HBc-posi-
Materials and methods
tive units critically affects blood supply at a high cost in
medium to high-endemic areas [10]. When the first gen-
Blood donor examination
eration of HBV-NAT technology was licensed, questions
were raised whether it should replace HBsAg for screen- Taiwan has two centralized testing laboratories in
ing blood donors [11]. Yet, most developed countries that charge of the nationwide blood examination operations,
added universal HBV-NAT to their testing strategies at Taipei and Kaohsiung blood centre. Whole blood
retained the HBsAg test. Lack of relevant scientific evi- donations were screened for HBV between 1 June and
dence made it difficult for the regulatory bodies and 31 November 2017 (six months) at Taipei blood centre.
experts to support its discontinuation at the time [12,13]. These donations were tested for HBsAg using Freedom
More recently, the value of retaining HBsAg screening EVOlyzer and HBV-DNA using Procleix Ultrio Plus
where universal HBV-NAT and anti-HBc tests are used assay (Grifols, Emeryville, CA, USA) on the fully auto-
has become a subject of active discussion by several mated TIGRIS system.
researchers, with current evidence indicating diminutive Presently, every donated whole blood unit is univer-
value in its continued use. Although discontinuing such a sally tested for HBsAg and HBV-DNA using the testing
seemingly redundant screening test might result in cost algorithm (see Fig. 1) similar to the one described in the
reduction, the existence of credible ethical and scien- pilot study for introducing NAT for screening blood
tific objections makes dropping it highly challenging donors in Taiwan [20].
[10,14–16]. Primarily:
In spite of the considerations to discontinue HBsAg (i) HBsAg seronegative and seropositive samples are both
screening test as discussed above, the test remains valu- tested for HBV-DNA.
able in high-endemic areas where the use of anti-HBc is (ii) MP-8-NAT is universally used to test for HBV-DNA,
not universal. Taiwan is among the few high-endemic and reactive samples are confirmed with ID-NAT.
countries that implement universal HBV-NAT and HBsAg (iii) HBsAg seropositive samples that are unreactive to
testing for its blood supply since 2013 [17]. In a previous MP-8-NAT are tested for HBV-DNA with ID-NAT.
study, a majority of Taiwanese blood donors were found (iv) ID-NAT reactive samples that test positive for both
to be under 30 years old [18] and according to another HBV-DNA and HBsAg are considered as confirmed
study, the anti-HBc-positive rate in this population cases of HBV infection.
appears to have decreased over the past 30 years due to (v) ID-NAT reactive samples that test positive for HBV-
the extensive infant vaccination programme [19]. There- DNA and negative for HBsAg are considered as con-
fore, giving rise to the question whether Taiwan could firmed cases of HBV infection.
adopt the blood donor policy for low-endemic countries (vi) Donors with samples that test HBsAg-positive and
(excluding the blood donors who are anti-HBc-positive in are nonreactive to MP-NAT and ID-NAT are consid-
order to eliminate the potential sources of occult HBV ered infectious (according to the blood bank policy
infection). since there is no universal anti-HBc testing).
This study aimed to substantiate the value of the (vii) HBsAg seronegative and MP-NAT nonreactive dona-
HBsAg test in Taiwan’s HBV blood donor testing strategy. tions are classified as negative, and their blood is
More specifically, an HBV Markov model was constructed permissible for donation.
and cost-effectiveness analysis was conducted in order to In order to substantiate the value of HBsAg in the
reappraise the existing HBV screening strategy in Taiwan. existing testing algorithm, four blood donation screening
The integrated decision analytic model presented in this strategies were considered as follows:
paper is adjustable and can readily be reused to perform (1) No intervention (reference/base case)
economic evaluations of blood donor diagnostic tests for (2) HBsAg (EIA)
other related TTIs (see Supporting Information). (3) NAT (MP-8-NAT and ID-NAT)
First, it is expected that the results of this study can (4) HBsAg (EIA) plus NAT (MP-8-NAT and ID-NAT)
provide feedback on the performance of the existing HBV To begin with, the residual risk of posttransfusion
screening strategy in Taiwan. Second, provide recommen- infection for each screening strategy was determined and
dations for maintaining, modifying, or changing the then cost-effectiveness analysis was performed to deter-
existing strategy. Finally, provide useful insights, guiding mine the optimal strategy. Readers are referred to [21] for
Figure 1 General schematic diagram of the HBV blood screening algorithm at Taipei blood centre.
a concise summary of health economic study types rele- the disease progression were incorporated to resemble the
vant to blood safety intervention assessments. natural history of an HBV infection [25]. The Markov model
consists of seven states as depicted in Fig. 2.
All blood donors are tested in the initial state. It is
Threat of infection and recipient exposure
assumed that diagnostic window period transmissions to
Due to the transient nature of HBV, residual risk estima- transfusion recipients can occur from this initial state. It
tion for HBV is more intricate as compared to that of HIV is important to note two things for this Markov model.
and HCV [22]. This value is critical in the analysis of the First, the acute infection state is temporary because it has
cost-effectiveness model and it mainly depends on the short-term effects. In that regard, it is a tunnel state with-
incidence rate, prevalence and window period estimates. out a self-loop arrow, which means that a patient cannot
The residual risk estimates were obtained using the inci- spend no more than a single cycle in the acute infection
dent rate-WP risk day equivalent model [22,23] (see Sup- state [26]. Second, the Markovian memoryless property is
porting Information for more details). Over 80% of the assumed to those recipients whose acute infection
donations in Taiwan are from repeat donors with an aver- resolves and acquires lifelong immunity. State transition
age annual donation frequency of 175. probabilities for the model were synthesized from litera-
A recent population-based cohort study estimated ture and are provided in Table 1.
mean age of blood transfusion patients in Taiwan to be
584 years [24]. We considered the cohort that included
Testing costs, disease burden costs and health
this demographic group and other limited populations of
utilities
susceptible potential recipients of blood products in
which the vaccine nonresponse rate is substantial. Each state of the Markov model is associated with its
own costs and health utilities. The cost for HBV blood
screening comprised of reagent, equipment, personnel
Disease progression model
and other ancillary costs. Costs were estimated at Taipei
In order to estimate the costs and effects of an HBV transfu- blood centre. Disease burden costs for the chronic phases
sion-transmitted infection, a Markov disease model was of hepatitis B were obtained from a Taiwanese study
constructed. Stages that have the most significant impact on which was conducted from a national health insurance
Figure 2 Markov state diagram for the natural history of HBV infection (for state transition probabilities, see Table 1).
perspective [30]. The disease burden costs were adjusted consider potential transmission from the blood recipients
for inflation using the consumer price index for Taiwan. to their partners.
Due to the spontaneous nature in which acute hepatitis B To ensure that the analysis captured important differences
resolves, no healthcare costs were attached to the acute in costs and outcomes, the model was simulated using a life-
state [31]. The health-related quality of life (HRQoL) of time horizon for 35 cycles, with each cycle analogous to a
the recipients who get posttransfusion chronic hepatitis B year. The simulation was run for a cohort of 288 947. Half-
is an essential parameter in determining the cost-effec- cycle correction was not applied to the model as it often
tiveness of the blood safety interventions. The HRQoL produces wrong results for discrete Markov models [32].
values used for the Markov model were obtained from Costs and health effects were discounted at a rate of 35%
related literature. The testing costs, disease burden costs per annum for all the Markov state values.
and health utilities are summarized in Table 2. All costs Deterministic sensitivity analysis was conducted to
are denominated in United States dollars. investigate the effect of varying specific input parameters.
Probabilistic sensitivity analysis (PSA) was also performed
in order to ascertain the overall parameter uncertainty.
Economic evaluation
Appropriate distributions were assigned for all the input
Cost-effectiveness analysis was performed using the Mar- parameters and fitted using the mean and standard devia-
kov model to compare the four blood donor screening tion (Table 2). The PSA was repeated for 1000 simulations.
strategies under consideration. The estimated residual risk
value for each strategy was used to determine the number
Computational techniques
of recipients who could get infected with hepatitis B from
transfusion. This study considered HBV transmissions The health economic model for donor blood testing was
from the blood donors to the blood recipients but did not implemented in RStudio using the ‘heemod’ package [33],
and the results were post-processed to give graphical for the current strategy is almost six times more than soci-
summaries by interfacing model results with the global etal WTP threshold. The addition of MP-NAT to HBsAg sig-
environment of the ‘BCEA’ package [34]. nificantly increased the estimated total cost by over 70%.
However, there was slight increase in the estimated net
effectiveness.
Results
Residual risk and infectious Window Period days Deterministic sensitivity analysis
The residual risk estimates were found to be 176, 779 Deterministic sensitivity analyses showed that ICER values
and 384 per million donations for ‘HBsAg’, ‘NAT’ and were more sensitive to the residual risk estimates, transition
‘HBsAg plus NAT’, respectively. The lengths of the probability from acute infection to chronic hepatitis B, test-
infectious window periods are presented in Table 3. ing costs and the discount rate, respectively. The tornado
plot for the variables which had the most impact on the
ICER values is shown in Fig. A1 (see Appendix 1). Very
Cost-effectiveness summary
low residual risk estimates significantly increased the ICER
The cost-effectiveness results for the evaluated strategies value, whilst higher estimates decreased the ICER value. On
are summarized in Table 4. The incremental cost-effective- the other hand, very high discount rates significantly low-
ness ratio (ICER) for the current testing strategy in Taiwan ered the ICER value, whilst lower discount rates increased
was estimated to be $US 443 614 per QALY, whilst the the ICER value. For example, increasing the discount rate
willingness-to-pay (WTP) threshold that reflects local pref- to 85% per annum decreased the ICER value twofold and
erences is $US 70 000 per QALY [35]. Therefore, the ICER this could lead to bias against the HBsAg intervention.
Interventions Estimated total costs (US$) Cost difference (US$) Effectiveness (QALYs) Effectiveness difference (QALYs) ICER (US$/QALY)
Figure 3 Cost-effectiveness acceptability curves and frontier. The overlaid grey line shows the effectiveness frontier curve.
speculate that implementing universal ID-NAT could fur- that HBsAg is the preferred choice as indicated on the
ther reduce the risk of OBI in Taiwan with very limited CEAF. Additionally, occult hepatitis B remains the main
added efficiency and at a much higher cost. safety risk for blood transfusion in Taiwan. Excluding anti-
There are some limitations to this study and to our HBc-positive donors might help to eliminate potential
model. The donor test results that were utilized for analy- sources of occult HBV infection; however, the main draw-
sis were from a 6-month period. Obtaining test data that back for anti-HBc testing is that it results in the loss of a
covers a longer period would be beneficial in refining the high percentage of non-infectious carriers thereby seri-
residual risk estimates. Similar to previous studies [37], ously affecting the blood supply. Therefore, the HBsAg test
some of the input parameters that populate the model are remains valuable in high-endemic areas, especially in those
single study-based estimates (e.g. HRQoL values). There with blood operators that do not afford to use NAT.
are limited published data on region-specific health state
utilities for HBV. This could result in the increase of
Acknowledgements
uncertainty in such parameter estimates. Despite these
limitations, our model utilized some specific inputs based This study was partially supported by the Ministry of
on Taiwanese published data, for example the disease Science and Technology of Taiwan under the grant num-
burden costs. Previous studies in Taiwan have mostly ber MOST 103-2410-H-011-012-MY3 and MOST 106-
focused on efficacy research of adding NAT to HBsAg 2410-H-011-004-MY3. Any opinions, findings, and con-
screening in order to increase the safety against TTIs clusions or recommendations expressed herein are those
[20,40]. The results of our study further demonstrate the of the authors and do not necessarily reflect the views of
benefit of implementing such a screening strategy in the sponsors.
regions having significant OBI carriers such as Taiwan by
conducting cost-effectiveness analysis.
Conflicts of interest
The authors declare that they have no conflicts of interest
Conclusion
relevant to the manuscript submitted to Vox Sanguinis.
To conclude, our study showed that the strategy of using
universal HBsAg and MP-8-NAT together with confirma-
Authors contributions
tory ID-NAT prevents a significant amount of HBV infec-
tions from entering the Taiwan blood supply. However, this All authors listed meet the authorship criteria and con-
comes at a disproportionate increase in cost. We also found tributed to research design, data acquisition, analysis and
interpretation, writing, critical review and approval of the Preparation: Tapiwa Blessing Matanhire. Review: Shi-
manuscript. Here we indicate the specific contributions Woei Lin, Tapiwa Blessing Matanhire. Visualization:
made by each author. Conceptualization/Methodology: Tapiwa Blessing Matanhire. Final approval of version for
Shi-Woei Lin, Tapiwa Blessing Matanhire. Acquisition, submission: Shi-Woei Lin, Tapiwa Blessing Matanhire.
Analysis and/Interpretation of data: Tapiwa Blessing Supervision/Project Administration/Funding Acquisi-
Matanhire, Shi-Woei Lin. Writing-Original Draft tion: Shi Woei Lin.
References
1 Hou J, Liu Z, Gu F: Epidemiology and screening of blood donors? Transfus based decision-making for blood
prevention of hepatitis B virus infec- Clin Biol 2004; 11(1):26–32 safety. Transfusion 2015; 55(8):2039–
tion. Int J Med Sci 2005; 2(1):50–7 12 Coste J, Reesink HW, Engelfriet CP, 47
2 Domanovic D: Assessing the risk of et al.: Implementation of donor 22 Lelie N, Vermeulen M, Van Drimmelen
transfusion-transmitted emerging screening for infectious agents trans- H, et al.: Direct comparison of three
infections. ISBT Sci Ser 2016; 11 mitted by blood by nucleic acid tech- residual risk models for hepatitis B
(Suppl. 1):68–75 nology : update to 2003. Vox Sang virus window period infections using
3 Chen DS: Fighting against viral hep- 2005; 88:289–303 updated input parameters. Vox Sang
atitis: Lessons from Taiwan. Hepatol- 13 Kuhns MC, Busch MP: New strategies 2020; 115(3):133–45.
ogy 2011; 54(2):381–92 for blood donor screening for hepatitis 23 Weusten J, Vermeulen M, DrimmelenH
4 Lin CK, Leung JNS, So BKL, et al.: B virus: nucleic acid testing versus Van, et al.: Refinement of a viral
Donor selection for blood safety: is it immunoassay methods. Mol Diagn transmission risk model for blood
still necessary? ISBT Sci Ser 2014; 9 2006; 10(2):77–91 donations in seroconversion window
(1):26–9 14 Seed CR: Value of retaining HBsAg phase screened by nucleic acid testing
5 Lieshout-Krikke RW, Zaaijer HL, Van donor screening where HBV NAT and in different pool sizes and repeat test
DeLaar TJW: Predonation screening of anti-HBc donor screening apply. ISBT algorithms. Transfusion 2011; 51
candidate donors and prevention of Sci Ser 2018; 13(1):70–5 (1):203–15
window period donations. Transfusion 15 Dodd RY: Epidemiology, performance 24 Lin SY, Hsu WH, Lin CC, et al.: Asso-
2015; 55(2):373–8 characteristics, or both? Transfusion ciation of transfusion with risks of
6 O’Flaherty N, Ushiro-Lumb I, Pomeroy 2017; 57(1):1–2 dementia or Alzheimer’s disease: a
L, et al.: Transfusion-transmitted hep- 16 Kramer K, Verweij MF, Zaaijer HL: Are population-based cohort study. Front
atitis B virus (HBV) infection from an there ethical differences between stop- Psychiatry 2019; 10:1–10
individual-donation nucleic acid (ID- ping and not starting blood safety 25 Pan CQ, Zhang JX: Natural history
NAT) non-reactive donor. Vox Sang measures? Vox Sang 2017; 112 and clinical consequences of hepatitis
2018; 113(3):300–3 (5):417–24 B virus infection. Int J Med Sci 2005;
7 Galel SA, Simon TL, Williamson PC, 17 Roth WK: History and future of 2(1):36–40
et al.: Sensitivity and specificity of a nucleic acid amplification technology 26 Sonnenberg FA, Beck JR: Markov
new automated system for the detec- blood donor testing. Transfus Med models in medical decision making: a
tion of hepatitis B virus, hepatitis C Hemother 2019; 46(2):67–75 practical guide. Med Decis Mak 1993;
virus, and human immunodeficiency 18 Tsai SL: Challenge of donor recruit- 13:322–38
virus nucleic acid in blood and plasma ment. ISBT Sci Ser 2009; 4:302–6 27 Liaw YF, Chu CM: Hepatitis B virus
donations. Transfusion 2018; 58:649– 19 Ni YH, Chang MH, Jan CF, et al.: Con- infection. Lancet 2009; 373:582–592
659 tinuing decrease in hepatitis B virus 28 Wiens A, Lenzi L, Venson R, et al.:
8 Kupek E: Residual risk of hepatitis-B- infection 30 years after initiation of Economic evaluation of treatments for
infected blood donations. ISRN Infect infant vaccination program in Taiwan. chronic hepatitis B. Brazilian J Infect
Dis 2013; 2013:1–15 Clin Gastroenterol Hepatol 2016; 14 Dis 2013; 17(74):418–26
9 Vermeulen M, Lelie N: The current sta- (9):1324–30 29 Lang HC, Wu JC, Yen SH, et al.: The
tus of nucleic acid amplification tech- 20 Li L, Chen PJ, Chen MH, et al.: A pilot lifetime cost of hepatocellular carci-
nology in transfusion-transmitted study for screening blood donors in noma: A claims data analysis from a
infectious disease testing. ISBT Sci Ser Taiwan by nucleic acid amplification medical centre in Taiwan. Appl Health
2016; 11(Suppl. 2):123–128 technology: detecting occult hepatitis Econ Health Policy 2008; 6(1):55–65
10 Candotti D, Laperche S: Hepatitis B B virus infections and closing the 30 Hsieh CR, Kuo CW: Cost of chronic
virus blood screening: Need for reap- serologic window period for hepatitis hepatitis B virus infection in Taiwan. J
praisal of blood safety measures? C virus. Transfusion 2008; 48:1198– Clin Gastroenterol 2004; 38(10 Suppl
Front Med 2018; 5:1–10 2106 3):S148–52
11 Busch MP, Should HBV: DNA NAT 21 Custer B, Janssen MP: Health eco- 31 Jindal A, Kumar M, Sarin SK: Man-
replace HBsAg and/or anti-HBc nomics and outcomes methods in risk- agement of acute hepatitis B and
reactivation of hepatitis B. Liver Int gained: what is the threshold of cost (HBV) infection through blood trans-
2013; 33(SUPPL. 1):164–75 effectiveness? Health Econ 2010; fusion from blood donors with occult
32 Barendregt JJ: The half-cycle correc- 19:422–37 HBV infection. Clin Infect Dis 2011;
tion: Banish rather than explain it. 36 WHO. Global Health Sector Strategy 52(5):624–32
Med Decis Mak 2009; 29(4):500–2 on Viral Hepatitis 2016–2021. 2016:1– 39 Dodd RY, Nguyen ML, Krysztof DE,
33 Filipovic-Pierucci A, Zarca K, Durand- 56 Available from: http://www.who. et al.: Blood donor testing for hepatitis
Zaleski I: Markov models for health int/hepatitis. B virus in the United States: is there a
economic evaluation: the R Package 37 Borkent-Raven BA, Janssen MP, Van case for continuation of hepatitis B
heemod. R package version 0.12.0. DerPoel CL, et al.: Cost-effectiveness of surface antigen detection? Transfusion
34 Baio G, Berardi A: BCEA: Bayesian additional hepatitis B virus nucleic acid 2018; 58(9):2166–70
cost effectiveness analysis. R package testing of individual donations or mini- 40 Yang MH, Li L, Hung YS, et al.: The
version 2.3-1.1. pools of six donations in the Nether- efficacy of individual-donation and
35 Shiroiwa T, Sung YK, Fukuda T, et al.: lands. Transfusion 2009; 49(2):311–9 minipool testing to detect low-level
International survey on willingness to 38 Yuen MF, Wong DKH, Lee CK, et al.: hepatitis B virus DNA in Taiwan.
pay (WTP) for one additional QALY Transmissibility of hepatitis B virus Transfusion 2010; 50:65–74
Supporting Information
Additional Supporting Information may be found in the online version of this article:
ICER (US$/QALY)
Figure A1 Tornado plot. The chart shows the sensitivity of the parameters around the ICER value of 386 585 between the HBsAg and the NAT strate-
gies.
Societal WP value
$80 000 $160 000 $240 000 $320 000 $6 40,000 $1 280 000 $2 560 000 $5 120 000
No Intervention 0 0 0 0 0 0 0 0
HBsAg 4009 8074 12 138 16 203 32 462 64 981 130 017 260 091
NAT 3899 7993 12 086 16 179 32 553 65 300 130 794 261 781
HBsAg plus NAT 3848 7953 12 057 16 162 32 581 65 418 131 092 262 441
The number in bold shows the highest net monetary value at that particular WTP threshold value, and the corresponding strategy is optimal at that
threshold.
Supplementary Material R code for the economic evaluation of hepatitis B virus testing strategies
& Jose-Angel Hernandez-Rivas 1,2
1
Transfusion Service and Hematology Department, Hospital Universitario Infanta Leonor, Hospital Virgen de la Torre, Madrid, Spain
2
Universidad Complutense de Madrid, Madrid, Spain
3
Intensive Care Unit, Hospital Universitario Infanta Leonor, Hospital Virgen de la Torre, Madrid, Spain
4
Internal Medicine Department, Hospital Universitario Infanta Leonor, Hospital Virgen de la Torre, Madrid, Spain
Background The COVID-19 outbreak has affected almost all hospital depart-
ments, including transfusion services. However, the demand for transfusions in a
general hospital designated to deal with COVID-19 patients has not been anal-
ysed before.
Study Design and Methods A retrospective study was conducted to evaluate
blood transfusion practices from 15 March to 14 April 2020 at Hospital Universi-
tario Infanta Leonor (Madrid, Spain). During this month, with few exceptions, the
hospital became a ‘COVID-19’ centre. In addition, transfusion rates during this
time frame and the same period over the last 4 years were compared.
Results From 15 March to 14 April 2020, only 254 blood components were
transfused, resulting in a 493% reduction over the previous year. Interestingly,
in critically ill patients, the red blood cell (RBC) transfusion/bed ratio signifi-
cantly decreased during this period (092) compared to the same ratio over the
past 4 years (270) (P = 002). Of note, 106 blood components (95 RBC; 11 plate-
let concentrates) were transfused to only 36 out of 1348 COVID-19 patients
(27%). The main reason for RBC transfusion in COVID-19 patients was a previ-
ous underlying disease (44%) followed by bleeding (25%) and inflammatory
anaemia (25%).
Conclusion This is the first study to report a decrease in blood transfusions dur-
ing the COVID-19 pandemic in a general hospital and especially in the intensive
care unit. The results of this study suggest that COVID-19 does not generally
Received: 27 May 2020,
induce transfusion requiring anaemia, being the main causes for transfusion in
revised 24 September 2020,
accepted 13 October 2020,
these patients underlying conditions or bleeding.
published online 2 November 2020 Key words: COVID-19, patient blood management, transfusion.
574
Transfusion during COVID-19 crisis 575
patients causing more than 740 000 deaths. Thus, Spain April 2020 at the Transfusion Department of the Hospital
has been severely hit by the pandemic and harbours one Universitario Infanta Leonor (Madrid, Spain). This is a
of the highest disease burden worldwide with more than second-level general hospital with 362 beds and teaching
300 000 affected patients and more than 28 000 deaths activities in some departments. The haematology service
[2]. Madrid has been the epicentre of the Spanish out- treats patients with acute leukaemia and performs, in col-
break, with around 30% of the confirmed cases in Spain laboration with a third level hospital, autologous bone
[3]. marrow transplantations (patient stays al our centre since
The new SARS-CoV-2 produces mild-to-moderate day + 1 until the discharge date). The intensive care unit
symptoms in approximately 85–90% of cases that do not (ICU) under normal conditions is a general facility in
require hospital admission. Accordingly, most COVID-19 charge of non-surgery critically ill patients and has a
patients do not develop significant anaemia nor thrombo- capacity of eight beds. Cardiovascular surgery and venti-
cytopenia [4,5]. However, 10% of patients may suffer a lation with extracorporeal oxygenation membranes
more severe clinical course (bilateral pneumonia, respira- (ECMO) is not available at this centre.
tory failure, septic shock, coagulopathy and multiple Our centre has a disaster programme with a decision-
organ failure, among others). This phase can lead to making committee. However, a specific COVID-19 surge
patient’s death in the context of systemic inflammatory plan for the pandemic was established from 30 January.
response syndrome associated with cytokine storm [6]. This plan has been evolving over time. As the pandemic
Among critical COVID-19 patients, a decrease in haemo- began, all surgeries and regular medical appointments
globin and platelet levels is common. Indeed, thrombocy- were cancelled or referred to other centres, including
topenia has been associated with a worse outcome [4]. deliveries, with the exception of oncological and haema-
Furthermore, this patients might suffer a coagulation dis- tological treatments and a non-COVID-19 ward with 30
order that results in a high incidence of thrombotic beds. In addition, general measures to prevent SARS-coV-
events, which is also related to prognosis and mortality 2 were adopted, such as the creation of COVID-19 free
[7]. This coagulopathy is characterised by hypercoagula- circuits and the restriction of visitors, among others.
bility with high levels of D-dimers and fibrinogen and The hospital has an active patient blood management
might resemble, in some cases, disseminated intravascular (PBM) programme with trimestral meetings composed of
coagulopathy (DIC). However, based on currently avail- surgical, medical and intensive care unit specialists and
able knowledge, an increment in haemorrhagic risk has nurses. Indications for blood transfusion are in accor-
not been reported before [8]. dance with the recommendations of the Spanish blood
In our region, the blood collection centres collect and transfusion society, based on patient´s clinical condition
distribute blood to hospital transfusion departments. One and suggesting the threshold for transfusion of red blood
of the problems associated with the COVID-19 crisis has concentrates (RBC) in a haemoglobin level of 7 g/dl. Dur-
been the decrease in blood donations, due to the large ing the pandemic, the transfusion service did not imple-
number of infected people and the social restrictions. ment a specific blood conservation strategy.
Nowadays, scarce data are available about the amount of We then described the blood usage for this 4-week per-
blood components consumed during the pandemic [9–12] iod (15 March to 14 April 2020) and the same 4-week
and the transfusion requirements for COVID-19 patients period on the last four years in all hospitalized patients.
[13]. Very recently some studies pointed out a decrease of We selected that month since it reflects the period with
blood transfusion among COVID-19 patients in compar- largest number of COVID-19 patients admitted to the hos-
ison with non-COVID-19 patients and a decrease in blood pital (Fig. 1), and at that point, the centre became a
components usage during the pandemic relative to the ‘COVID-19’ Hospital.
weeks before and after [13–15]. We also compared transfusion rates at the ICU during
In this setting, the aims of this study are (1) to evaluate this period with the rates of the last four years. As
the transfusion of blood components at our Hospital COVID-19 brought a high demand for ICU and our ICU
between 15 March and 14 April 2020; (2) to compare the triplicated its capacity for non-surgery patients, we calcu-
results with the same period of time during the last lated the transfusion ratio dividing the number of prod-
4 years and (3) to analyse the transfusion requirements in ucts used during this month by the median of occupied
COVID-19 patients. beds per day.
Clinical data were obtained from electronic medical and
hospital records (SELENE SP12) and the transfusion data
Material and methods
from the e-PROGESA computer system (version 5.0.3.).
We conducted a single-centre retrospective study evaluat- The local Ethics Committee at Hospital Universitario
ing blood transfusion requirements from 15 March to 14 Infanta Leonor approved the study.
Fig. 1 Number of patients with COVID-19 diagnosis admitted from 1 March to 30 April 2020 in bars chart. [Colour figure can be viewed at wileyonline
library.com]
Statistical analysis was performed using SPSS Table 1 Distribution of transfused blood components per year at the
21.0 software package (SPSS, Chicago, IL, USA). Mann– hospital in the last 4 years (2016–2019)
Whitney tests were used to compare quantitative
Blood Components 2016 2017 2018 2019 Total (mean)
variables. Statistical significance was defined as a
<005. RBC 5109 5159 5209 5079 5139
PC 543 513 437 458 488
FFP 308 329 422 251 328
Results Total 5960 6001 6068 5788 5983
By 14 April, 1348 adult patients with a confirmed SARS-
FFP, fresh-frozen plasma; PC, platelet concentrate; RBC, red blood cell.
COV-2 infection by a pharyngeal smear RT-PCR had been
admitted to our centre in the previous 30 days. Forty-
nine of them were transferred to the ICU (36%). Mortal- Table 2 Comparative transfused blood components: 15 March—14 April
ity among critically ill patients with COVID-19 was during the previous 4 years and the COVID-19 2020 outbreak
653% (32/45). In addition, 210 (1558%) deaths occurred
due to COVID-19 during this month of follow-up for the Blood Components 2016 2017 2018 2019 2020
other hospitalized patients. Median length of stay was
RBC 453 391 436 385 225
7 days for the whole cohort and 9 days for the ICU PC 51 54 34 33 29
patients. FFP 32 32 18 83 0
Total 536 477 488 501 254
The use of blood products during the COVID-19 FFP, fresh-frozen plasma; PC, platelet concentrate; RBC, red blood cell.
pandemic
Around 6000 blood components (5000 red blood concen-
trates [RBC]; 650 platelet concentrates [PC]; 350 fresh-
frozen plasma [FFP] units) are transfused annually at our
Transfusion in critically ill patients during the
hospital. The distribution of transfusions in the past four
COVID-19 pandemic
years is shown in Table 1. Between 15 March and 14 From 15 March to 14 April, the ICU tripled its capacity
April 2020, the total number of blood components trans- with around 26 admitted patients, when in normal condi-
fused decreased dramatically compared to the same per- tions approximately 8 beds are available.
iod of previous years. As Table 2 illustrates, only 254 All patients admitted at the ICU at this moment were
blood components were transfused, which results in a COVID-19 patients. We observed that the transfusion ratio
493% reduction over the previous year: 225 RBC (416% during the evaluated period was 092 for RBC (24 RCB/26
reduction); 29 PC (1212% reduction); and 0 FFP units beds), 004 for PC (1 PC/26 beds) and 0 for FFP units. The
(100%). ratio of transfused components during the last 4 years in
the same period was significantly higher for RBC (ra- drastic decline or absence of elective surgeries, paediatric
tio = 27, 100 RBC/37 beds) (P = 002), and for FFP units care, other medical admissions and obstetric care. Conse-
(ratio = 135, 50 FFP/37 beds) (P = 0012). It was also quently, blood use decreased at these institutions
higher for PC (ratio = 049, 18 PC/37 beds) although not [13,17,18]. The development and implementation of
reaching significance (P = 01; Table 3). patient blood management programmes during the pan-
demic might have also have played an important role in
the decline in transfusions [19].
COVID-19 patients who required transfusion
In this study, we confirm a significant decrease of
From 15 March to 14 April, 106 blood components (95 around 50% in total blood components transfusion during
RBC; 11 CP) were transfused to only 36 of 1348 COVID- the pandemic in a ‘COVID-19 Hospital’. It should be noted
19 patients (27%). Characteristics of these 36 patients are that in the period of time studied, our institution multi-
shown in Table 4. Median age was 75 years (34–94), and plied by 25 the number of conventional beds and by 3
611% of the patients were male (22/36). Interestingly, all the number of ICU beds compared with usual occupation.
of them had underlying chronic medical conditions Similar results have been observed in recent reports from
including 47% (17/36) of the patients with cancer (8/36 Maryland, Singapore, China or Chicago with a decrease
onco-haematological; 9/36 solid tumour). The indications on blood product transfusion that ranged between 12 and
for transfusion were underlying medical condition 42% [13–15,20]. We observed that RBC transfusion
(N = 16, 44%), bleeding (N = 10, 28%) and inflammatory decreased around 41%, platelet transfusion decreased
anaemia (N = 10, 28%). Detailed information about around 12% and plasma units were not transfused during
underlying medical conditions is summarized in Table 4. this period, as compared to the same period in previous
Of the 10 patients transfused due to inflammatory anae- years. Other groups have also showed a minimal use of
mia, 5 of them had a documented bacterial infection FFP among COVID-19 patients [14,15,20]. Even though
which led to septic shock. specially critically ill COVID-19 patients develop coagu-
Out of the 36 transfused patients, 23 (64%) died due to lopathy, it is usually associated with thrombotic events,
COVID-19 complications. elevation of D-dimer and fibrinogen and usually does not
impact in transfusion needs [8].
Furthermore, it is worth emphasizing that most of the
Discussion
blood component transfusions performed to non-COVID-
The COVID-19 outbreak has become the most important 19 patients were requested by Haematology and Oncology
health, economic and social problem in recent decades. Departments. Due to the SARS-CoV-2 complications, ICUs
Spain is now probably the European country in the world have been adapted to increase their usual capacity to
with more COVID-19 cases related to its population [3]. address the need for invasive mechanical ventilation [21].
Since its onset in December 2019, some groups have Similarly, from 15 March to 14 April, the ICU of our hos-
reported a decline in the number of donations, blood sup- pital tripled its capacity. We observed that, despite the
plies and concerns about blood safety [9–11]. However, to increase in critically ill patients, the ICU RBC transfusion/
date, no evidence of SARS-CoV-2 transmission by blood bed ratio was significantly lower than it was on the pre-
products has been published [16]. On the other hand, vious 4 years (P = 002, Table 3). Prior studies have
many of the hospitals in countries particularly affected shown that 95% of ICU patients are found to have anae-
by the pandemic have become ‘COVID-19’ centres, with a mia and 85% of those with a 7-day ICU admission or
longer received at least 1 RBC unit [22–24]. However, we
Table 3 Number of transfused blood components at the Intensive Care
hypothesize that the shift in practice (almost all admitted
Unit in the last four years vs 2020 (period 15 March—14 April)
and all critically ill patients were COVID-19 at this time,
Ratio last Ratio cancellation of elective surgery and non-essential care)
four years: N N Blood could have been involved in the decline in blood transfu-
Blood Blood Components/N Components/N sion. One limitation of our study is that at our hospital,
Components Beds ICU Beds ICU 2020 P Value ventilation with extracorporeal oxygenation membranes
(ECMO) is not available, and patients who require this
RBC 270 092 002
modality are referred to other centres. This ventilation
PC 049 004 01
modality is known to increase bleeding, with the need for
FFP 135 0 001
around 2–4 units of packed red blood cells and 4–8 units
Total 468 100
of platelet concentrate daily [25,26]. Therefore, the results
FFP, fresh-frozen plasma; ICU, Intensive Care Unit PC; platelet concen- of our work cannot be generalized to other centres that
trate; RBC, red blood cell. use ECMO ventilation.
Blood component
Transfusion (During
Patient Sex/Age Previous Disease ICU ABO/Rh Group Reason for transfusion hospitalization) Death
Table 4 (Continued)
Blood component
Transfusion (During
Patient Sex/Age Previous Disease ICU ABO/Rh Group Reason for transfusion hospitalization) Death
Metastatic Prostate
Cancer
35 M/68 HTN Yes A Rh+ Inflammatory anaemia 1 RBC Yes
AF
36 F/56 DM Yes A Rh + Inflammatory anaemia 2 RBC Yes
Hypothyroidism 1 PC
AA, aplastic anaemia; AF, atrial fibrillation; AML, acute myeloid leukaemia; APS, Antiphospholipid Syndrome; CLL, chronic lymphocytic leukaemia, CRF,
chronic renal failure; DM, diabetes; ET, essential thrombocythemia; HL, Hodgkin lymphoma; HTN, hypertension; MDS, Myelodysplastic Syndrome; SLE, sys-
temic lupus erythematosus.
Next, we observed that only 27% (36) of the patients products has been observed, which can help us to design
admitted with COVID-19 were transfused. Of note, all of our transfusion policy (blood procurement, processing
them had previous medical conditions including a high and use) and be prepared to undertake possible further
percentage of patients with an active neoplasm (44%). outbreaks of the disease in the next few months. Most of
When we analysed the reason for transfusion, we observed COVID-19 patients did not require blood products despite
that a very low proportion of patients were transfused due their critical condition.
to the virus-induced inflammatory state, which seems very
interesting, and has not been reported so far. Only 10
patients received RBC due to their critically ill condition, Acknowledgements
and in 5 of them, another bacterial infection was docu- We thank the technicians of the Transfusion Service for
mented. The need for blood transfusion in COVID-19 their technical assistance.
patients might be also considered a poor prognostic factor,
as 64% (23/36) of the COVID-19 transfused patients died.
This is not surprising, as the need of blood components Conflicts of interest
might reflect a serious underlying condition [27].
In this study, we analysed blood usage during the per- The authors declare no conflict of interests.
iod of 30 days in which the pandemic most hardly
affected our Hospital. However, as shown in Fig. 1
Author contributions
COVID-19 patients were admitted to the centre before and
after those dates. Therefore, another limitation of the KMM, JAH and IGGM conceived the idea, collected, anal-
study is that some patients, especially those with a pro- ysed the data and wrote the manuscript. MAFG, CMN, MI,
longed critical condition might be under-represented. JCHS, ELH, BBG, MDF collected, analysed the data and
In conclusion, during the first wave of COVID-19 pan- reviewed the manuscript. JAH provided critical revisions
demic, a very significant decrease in transfused blood to the manuscript.
References
1 Lai C-C, Shih T-P, Ko W-C, et al.: Sev- 3 Ministerio de Sanidad, Consumo y in the classification and prognosis
ere acute respiratory syndrome coron- Bienestar Social: Professionals - evaluation of COVID-19: a retrospec-
avirus 2 (SARS-CoV-2) and Situaci
on actual Coronavirus [Inter- tive cohort study. Lancet Haematol
coronavirus disease-2019 (COVID-19): net]. Available from: https://www. 2020; 7:e671–e678
the epidemic and the challenges. Int J mscbs.gob.es/en/profesionales/salud 5 Fan BE, Chong VCL, Chan SSW, et al.:
Antimicrob Agents 2020; 55:105924 Publica/ccayes/alertasActual/nCov- Hematologic parameters in patients
2 WHO: Coronavirus Disease (COVID-19) China/situacionActual.htm [Last with COVID-19 infection. Am J Hema-
Dashboard [Internet]. Available from: accessed on 2020 August 13] tol 2020; 95:E131–E134
https://covid19.who.int [Last accessed 4 Liao D, Zhou F, Luo L, et al. Haemato- 6 Zhou M, Zhang X, Qu J: Coronavirus
on 2020 August 13] logical characteristics and risk factors disease 2019 (COVID-19): a clinical
update. Front Med 2020; 14: 13 Barriteau CM, Bochey P, Lindholm PF, Zhejiang province China. Vox Sang
126–135 et al.: Blood transfusion utilization in 2020; 115:502–506
7 Jin S, Jin Y, Xu B, et al.: Prevalence hospitalized COVID-19 patients. Trans- 21 Baron DM, Franchini M, Goobie SM,
and impact of coagulation dysfunction fusion. 2020. https://doi.org/10.1111/ et al.: Patient blood management dur-
in COVID-19 in China: a meta-analy- trf.15947 ing the COVID-19 pandemic: a narra-
sis. Thromb Haemost 2020. https://doi. 14 Murphy C, Jackson B, Fontaine M: tive review. Anaesthesia 2020;
org/10.1055/s-0040-1714369 Tools for rapid analysis of blood usage 75:1105–1113
8 G€orlinger K, Dirkmann D, Gandhhi A, and inventory during the COVID-19 22 Corwin HL, Parsonnet KC, Gettinger A:
et al.: COVID-19 associated coagulopa- pandemic. Transfusion 2020. https://d RBC transfusion in the ICU. Is there a
thy and inflammatory response: what oi.org/10.1111/trf.15996 reason? Chest 1995; 108:767–771
do we know already and what are the 15 Fan BE, Ong KH, Chan SSW, et al.: 23 Vincent JL, Baron J-F, Reinhart K,
knowledge gaps? Anesth Analg 2020; Blood and blood product use during et al.: Anemia and blood transfusion
131:1324–1333 COVID-19 infection. Am J Hematol in critically ill patients. JAMA 2002;
9 Raturi M, Kusum A: The blood supply 2020; 95:E158–E160 288:1499–1507
management amid the COVID-19 out- 16 Cai X, Ren M, Chen F, et al.: Blood 24 Jandu AS, Vidgeon S, Ahmed N:
break. Transfus Clin Biol 2020; transfusion during the COVID-19 out- Anaemia and transfusion triggers in
27:147–151 break. Blood Transfus. 2020; 18:79–82 critically ill patients - What we have
10 Pagano MB, Hess JR, Tsang HC, et al.: 17 Goh KJ, Wong J, Tien J-CC, et al.: learnt thus far. J Intensive Care Soc
Prepare to adapt: blood supply and Preparing your intensive care unit for 2019; 20:284–289
transfusion support during the first 2 the COVID-19 pandemic: practical 25 Koeckerling D, Pan D, Mudalige NL,
weeks of the 2019 novel coronavirus considerations and strategies. Crit Care et al.: Blood transfusion strategies and
(COVID-19) pandemic affecting Wash- Lond Engl 2020; 24:215 ECMO during the COVID-19 pandemic.
ington State. Transfusion 2020; 18 Stanworth SJ, New HV, Apelseth TO, Lancet Respir Med 2020; 8:e40
60:908–911 et al.: Effects of the COVID-19 pan- 26 Kowalewski M, Fina D, Słomka A,
11 Garraud O: COVID-19: is a paradigm demic on supply and use of blood for et al.: COVID-19 and ECMO: the inter-
change to be expected in health care transfusion. Lancet Haematol 2020; 7: play between coagulation and inflam-
and transfusion medicine? Transfus e756–e764 mation-a narrative review. Crit Care
Clin Biol 2020; 27:59–60 19 Shander A, Goobie SM, Warner MA, Lond Engl 2020; 24: 205
12 Chang L, Yan Y, Wang L: Coronavirus et al.: Essential role of patient blood 27 Kamper-Jørgensen M, Ahlgren M,
disease 2019: Coronaviruses and blood management in a pandemic: a call for Rostgaard K, et al.: Survival after
safety. Transfus Med Rev 2020; action. Anesth Analg 2020; 131:74–85 blood transfusion. Transfusion 2008;
34:75–80 20 Wang Y, Han W, Pan L, et al.: Impact 48:2577–2584
of COVID-19 on blood centres in
Abstract
Background and objectives Blood transfusion is a cornerstone therapy for many
patients with myelodysplastic syndromes (MDS), but ranges from few to no trans-
fusions to intensive transfusion therapy. To date, no large studies have described
transfusion use or costs for patients with MDS, accounting for the range of dis-
ease severity.
Materials and methods A nationwide cohort study was conducted with all
patients diagnosed with MDS in Sweden between 2008 and 2017, based on the
Swedish MDS register and the Swedish part of the Scandinavian Donations and
Transfusions Database 3 (SCANDAT3-S). Patients were followed from diagnosis
until death, emigration, allogeneic hematopoietic stem cell transplantation or end
of follow-up. Average cumulative transfusion count and costs over time were
calculated, stratified by the revised international prognostic scoring system
(IPSS-R) and age at diagnosis. Costs calculations used data on incident transfu-
sions and laboratory testing and were divided into: direct material costs, direct
labour costs and laboratory costs.
Results In total, 2311 patients were included in the cohort. In the first four years
after diagnosis, patients in the very low IPSS-R category received on average 25
red cell (95% confidence interval, 20–32) and 4 (3–7) platelet transfusions. Con-
versely, patients in the very high-risk category received on average 171 (135–
200) red cell and 66 (51–78) platelet transfusions. Correspondingly, transfusion
costs ranged from $8805 ($6482–$11 625) to $80 106 ($61 460–$95 792).
Conclusion Transfusion count and costs for patients with MDS increased mark-
Received: 1 September 2020,
edly with IPSS-R risk category, but were similar across age groups. Transfusion
revised 24 October 2020,
accepted 2 November 2020,
costs were considerable for the highest IPSS-R risk categories.
published online 18 November 2020 Key words: myelodysplastic syndromes, blood transfusion, transfusion costs.
This is an open access article under the terms of the Creative Commons Attribution-NonCommercial License,
which permits use, distribution and reproduction in any medium, provided the original work is properly cited and 581
is not used for commercial purposes.
582 J. Zhao et al.
conversion from SEK to USD was done using mid-2018 at the time of their diagnosis (N = 93), a total of 2311
conversion rate of 8998 SEK to 1 USD. patients remained for analysis.
Baseline characteristics of the study cohort, stratified
by IPSS-R score, are presented in Table 1. There were 423
Statistical analysis
(183%), 741 (321%), 463 (200%), 364 (158%) and 320
For both transfusions and transfusion costs, the number (138%) patients in the very low, low, intermediate, high
of persons in the study was calculated per day, to and very high prognostic categories, respectively. The
account for the high mortality of the MDS population. proportion of females ranged from 348% in the very
Patients who emigrated, underwent hematopoietic stem low-risk group to 463% in the very high-risk group, and
cell transplantation or died were excluded at the end of the mean age from 744 years (standard deviation [SD],
the day. The daily mean number of transfusions was 95) in patients with low risk, to 710 years (SD, 128) in
calculated as the number of transfusions in each day patients with very high risk. The median length of fol-
divided by the number of people alive at the beginning low-up was progressively shorter with higher IPSS-R
of the day; these were accrued to form the expected score, ranging from 24 years (interquartile range [IQR],
cumulative number of transfusions. The expected cumu- 12–45) in patients with very low-risk disease, to
lative transfusion cost was calculated in a similar fash- 05 years (IQR, 03–10) in patients with very high-risk
ion. Both the cumulative number of transfusions and disease.
the ensuing cumulative cost were calculated per day The average cumulative number of transfusions strati-
since diagnosis, stratified by IPSS-R category and age fied by IPSS-R at diagnosis is presented in Table 2. Fur-
at diagnosis (<65 years, 65–80 years, >80 years). Costs ther stratification by age at diagnosis, is presented in
were also calculated separately for only red cell trans- Figure 1. Frequency of red cell and platelet transfusions
fusions. increased with higher IPSS-R category, whereas plasma
For all analyses, 95% confidence intervals were calcu- transfusions were rare in all IPSS-R categories. Amongst
lated using bootstrapping, with 10,000 runs. The pre- patients in the very low category, those >80 years at
sented confidence intervals for transfusion costs were diagnosis received more red cell transfusions but fewer
modified in such a way that the lower range for nurse platelet transfusions compared to those <65 years, on
administration times was used for the lower bound of the average. Interestingly, the number of transfusions was
confidence interval and the upper range for nurse admin- otherwise similar between age groups within the same
istration times for the upper bound of the confidence IPSS-R prognosis category. In the first 4 years after diag-
interval. The figures are censored when there were less nosis, patients in the very low-risk category overall
than 10 persons in study. received 25 red cell (95% confidence interval [CI], 20–32),
The proportion of transfusions conducted inpatient, and 4 platelet transfusions (95% CI, 3–7), whereas
including confidence intervals, was calculated using patients younger than 65 years at diagnosis received on
logistic regression, with time since diagnosis modelled as average 16 red cell (95% CI, 9–25) and 9 platelet (95% CI,
a restricted cubic spline (with knots at 0, 3, 7, 180, 360, 3–16) transfusions. Conversely, patients in the very high-
720, 1460 days). risk category had on average 171 red cell (95% CI, 135–
All data processing and statistical analyses were per- 200) and 66 platelet transfusions (95% CI, 51–78) 4 years
formed using SAS Statistical Analysis Software, version after their diagnosis. More than half of patients younger
9.4 (Cary, NC, USA). The creation of the SCANDAT3-S than 65 years of age in the intermediate, high and very
database and the conduct of this study was approved by high categories underwent HSCT (53%, 53% and 52%,
the regional ethics committee in Stockholm, Sweden and respectively), and very few remained in the study after
the Swedish Ethics Review Authority (ref. nr. 2018/167- 2 years since diagnosis (22, 5 and 2 patients, respec-
31, 2019-02636). tively). Amongst patients older than 80 years at diagnosis
in the high and very high-risk categories, few remained
alive after two years after diagnosis; however, none
Results
underwent HSCT.
We identified a total of 2858 patients with a diagnosis of The average cumulative cost of transfusions stratified
MDS between 2008 and 2017 from the Swedish MDS reg- by IPSS-R category is presented in Table 3 and further
ister. After excluding patients who had an unknown date stratified by age group at diagnosis in Figure 2. Trends in
of diagnosis (N = 1), no recorded IPSS-R score (N = 449), transfusion costs are congruent with transfusion trends
a diagnosis date after date of death (N = 1), a prior allo- and increased with increasingly severe IPSS-R category at
genic hematopoietic stem cell transplantation (HSCT) diagnosis. Across all age groups, transfusion costs were
(N = 3), or who were not recorded to be living in Sweden highest on average for the very high-risk category,
Very low risk Low risk Intermediate risk High risk Very high risk
No. subjects (% of total) 423 (183) 741 (321) 463 (200) 364 (158) 320 (138)
Females, N (%) 147 (348) 310 (418) 179 (387) 149 (409) 148 (463)
Age at diagnosis, N (%)
<65 years 63 (149) 116 (157) 93 (201) 92 (253) 81 (253)
65–79 years 232 (548) 398 (537) 267 (577) 188 (516) 165 (516)
≥80 years 128 (303) 227 (306) 103 (222) 84 (231) 74 (231)
Mean, years (SD) 742 (95) 744 (95) 719 (106) 703 (129) 710 (128)
Length of follow-up, N (%)
<1 year 169 (400) 346 (467) 319 (689) 321 (882) 297 (928)
1–4 years 164 (388) 274 (370) 120 (259) 39 (107) 21 (66)
≥5 years 90 (213) 121 (163) 24 (52) 4 (11) 2 (06)
Median, years (IQR) 24 (12–45) 21 (11–38) 12 (04–14) 08 (04–14) 05 (03–10)
Hematopoietic transplant during follow-up, N (%) 11 (26) 41 (55) 69 (149) 62 (170) 53 (166)
Table 2 Cumulative average number of transfusions over time, stratified by IPSS-R category at diagnosis.
Duration of follow-up
amounting to $80 106 (95% CI, $61 460–95 792), and given inpatient during the first months after diagnosis
lowest in the very low-risk category, amounting to $8805 that stabilized to around 30% after about half a year.
(95% CI, $6482–11 625) after four years. Costs for red
cell transfusion support only is presented in Appendix 4.
Discussion
The proportion of transfusions given in an inpatient
setting is shown in Figure 3. Trends were similar for all In this nationwide cohort study of transfusion patterns
IPSS-R groups, with a higher proportion of transfusions and costs for patients with HSCT-na€ıve MDS, average
Figure 1 Cumulative mean number of RBC and PLT transfusions over time, stratified by IPSS-R and age at diagnosis. [Colour figure can be viewed at
wileyonlinelibrary.com]
Table 3 Total costs of transfusion therapy during follow-up, stratified by IPSS-R category at diagnosis
Duration of follow-up
costs for transfusions in the first 4 years after diagnosis As far as we know, this is the largest and most detailed
were found to be as high as $80 106 for patients in the study of transfusion costs in this patient group. Our
very high IPSS-R prognostic category and as low as method for calculating costs expands on previously
$8805 for patients in the very low prognostic category. reported methods, by differentiating between inpatient
Figure 2 Cumulative mean cost of transfusions in USD over time, stratified by IPSS-R and age at diagnosis. *Upper and lower confidence interval uses
the upper and lower bound of nurse administration time, respectively.
and outpatient transfusions, as well as incorporating incorporate the high mortality of the MDS patients by
results from antibody screening to include costs for cross- calculating transfusion counts and costs daily. Interest-
matching and antibody panels. Furthermore, we ingly, transfusion incidence and cost were mainly driven
by IPSS-R risk category, whereas it was generally similar typically indicated if ferritin levels are above 1500 lg/L,
across age groups within the same IPSS-R category. or after approximately 25 red cell transfusions [11]. More
In a previous study, blood product costs for MDS indirect costs such as cost of hospital facilities or equip-
patients were calculated to exceed 20 000 USD in the first ment were not included as they are largely determined by
2 years after diagnosis [6]. However, there are several dif- local circumstance, such as cost of rent, resource utiliza-
ferences that lead to differential results in this study. tion rates, and alternative costs for resources and assets.
Firstly, the previous study included only blood product Secondly, the cost algorithms assume that all first-time
costs and did not include costs for laboratory testing, positive antibody screens lead to one antibody panel and
labour or materials. Secondly, the previous study did not cross-matching for subsequent units, disregarding more
stratify on disease severity and was an average across dif- complex scenarios with positive direct antiglobulin tests,
ferent age groups. Thirdly, the previous study calculated or need for additional antibody panels or blood group
average costs based on the number of patients at baseline genotyping. Thirdly, we do not account for the need of
which may lead to a downward bias in costs, particularly phenotype-matched blood products, which are typically
if mortality is high. In this study, we compensated for at least twice as expensive. Our projected costs therefore
mortality by using the number of patients at risk of trans- likely underestimate true costs, but the extent of the
fusions, updated daily, as the denominator. Finally, the underestimation is difficult to assess without more
previous study used data from 2000 to 2010, whereas this detailed clinical data and may vary according to local cir-
study uses more contemporary data from 2008 to 2017. cumstances.
As such, the figures are not directly comparable, and this At the same time, we were not able to differentiate
study offers a more accurate characterization of transfu- transfusions necessitated by the disease per se from trans-
sion incidence and costs for patients with MDS. Further- fusions driven by other indications such as high-dose
more, whilst we have earlier reported that transfusion chemotherapy. Whilst we did censor follow-up upon
costs were markedly higher in older patients above undergoing transplantation, the available data do not reli-
65 years at diagnosis, we show here that this is no longer ably capture other instances of chemotherapy and there-
the case after incorporating patient mortality and disease fore does not allow the same strategy to remove
severity [6]. transfusions due to cytopenia following chemotherapy,
The major strengths include the use of incident trans- including patients that may have transformed to acute
fusions, as well as transfusion laboratory tests and results myeloid leukaemia. We speculate that transfusions given
from a nationwide transfusion database, together with for such indications would predominantly be seen in the
data on IPSS-R prognosis categories from nationwide higher risk categories and that interpretation of blood use
quality registers, to calculate transfusion costs. Further- and costs in these groups must consider such effects,
more, using national health and population registers, we whilst likely having less impact on lower risk categories.
were able to identify the date for migration, HSCT and In conclusion, in this nationwide study on costs associ-
death. This allowed us to use the average transfusion ated with blood transfusions in patients with MDS, higher
count and cost on a daily basis, as weekly or monthly risk IPSS-R prognosis category at diagnosis was associ-
time-periods would likely underestimate the results due to ated with significant costs. On average in the first
the high mortality. 4 years, transfusion costs ranged from $8805 to $80 106.
There are several limitations to the study. Firstly, costs
are limited to blood product costs, direct labour costs and
Acknowledgements
additional transfusion-related laboratory costs. We did
not incorporate societal costs in a wider sense (e.g. costs We are greatly indebted to all the blood banks in Sweden
for donor transportation and absence from work), or costs for both collecting and contributing data to this study.
for possible transfusion complications. Costs for relevant Furthermore, we would like to thank Roland Fiskesund
medications, such as erythropoietin analogs and iron MD PhD (Region Stockholm) for information on costs
chelation, which may affect or be necessitated by long- associated with transfusions. JZ and GE performed the
term transfusions therapy, are also not included. Follow- data collection. JZ, TD, and GE designed the study, per-
ing Nordic MDS group guidelines, iron chelation is formed the analyses, and wrote the manuscript.
References
1 Cancer TIAfRo: WHO Classification of Lymphoid Tissue. Lyon, France, World 2 Greenberg PL, Attar E, Bennett JM,
Tumours of Haematopoietic and Health Organization, 2008 et al.: Myelodysplastic syndromes:
clinical practice guidelines in oncol- 6 Zhao J, Ryden J, Wikman A, et al.: 9 Zhao J, Rostgaard K, Hjalgrim H,
ogy. J Natl Compr Cancer Netw 2013; Blood use in hematologic malignan- et al.: The Swedish Scandinavian
11:838–74 cies: a nationwide overview in Sweden donations and transfusions database
3 Haase D, Stevenson KE, Neuberg D, between 2000 and 2010. Transfusion (SCANDAT3-S) – 50 years of donor
et al.: TP53 mutation status divides 2018; 58:390–401 and recipient follow-up. Transfusion;
myelodysplastic syndromes with com- 7 Moreno Berggren D, Folkvaljon Y, n/a.https://onlinelibrary.wiley.com/d
plex karyotypes into distinct prognostic Engvall M, et al.: Prognostic scoring oi/full/10.1111/trf.16027.
subgroups. Leukemia 2019; 33:1747–58 systems for myelodysplastic syn- 10 Stokes EA, Wordsworth S, Staves J,
4 Greenberg PL, Tuechler H, Schanz J, dromes (MDS) in a population-based et al.: Accurate costs of blood transfu-
et al.: Revised international prognostic setting: a report from the Swedish sion: a microcosting of administering
scoring system for myelodysplastic MDS register. Br J Haematol 2018; blood products in the United Kingdom
syndromes. Blood 2012; 120:2454–65 181:614–27 National Health Service. Transfusion
5 Malcovati L, Hellstr€ om-Lindberg E, 8 Ludvigsson JF, Otterblad-Olausson P, 2018; 58:846–53
Bowen D, et al.: Diagnosis and treat- Pettersson BU, et al.: The Swedish per- 11 Kittang AO, Cavelier L, Dybedal I,
ment of primary myelodysplastic syn- sonal identity number: possibilities et al.:Guidelines for the diagnosis and
dromes in adults: recommendations and pitfalls in healthcare and medical treatment of Myelodysplastic Syn-
from the European LeukemiaNet. Blood research. Eur J Epidemiol 2009; drome and Chronic Myelomonocytic
2013; 122:2943–64 24:659–67 Leukemia. Nordic MDS Group, 2019.
NOTES
*Using mid-2018 SEK to USD exchange rate of 89984.
**Sum of salary employer and employers’ social security contribution.
Appendix 4 Total costs of red cell transfusion therapy during follow-up, stratified by IPSS-R
category at diagnosis.
Duration of follow-up
NOTES
*Using mid-2018 SEK to USD exchange rate of 89984.
Background The absence of the red cell antigens P, P1 and Pk, known as ‘p’, rep-
resents an extremely rare red cell phenotype. Individuals with this phenotype
spontaneously form anti-PP1Pk isoantibodies, associated with severe haemolytic
transfusion reactions, recurrent spontaneous abortion and haemolytic disease of
the fetus and newborn (HDFN).
Methods We report a series of four successful pregnancies in three women with
anti-PP1Pk isoantibodies, one complicated by HDFN, another by intrauterine
growth restriction, all managed supportively. We also review the literature
regarding the management of pregnancy involving anti-PP1Pk isoimmunization.
Results The literature surrounding anti-PP1Pk in pregnancy is limited to a very
small number of case reports. The majority report management with therapeutic
plasma exchange (TPE) with or without intravenous immunoglobulin. The rela-
tionship between titre and risk of pregnancy loss remains unclear, though a his-
tory of recurrent pregnancy loss appears important. Although a positive cord
blood direct antiglobulin test is frequently noted, clinically significant HDFN
appears uncommon, though possible.
Conclusion Early initiation of TPE in high risk patients should be strongly con-
sidered. If possible, pregnancies should be managed in a high-risk obstetric or
maternal fetal medicine service. The fetus should be monitored closely with inter-
val fetal ultrasound and middle cerebral artery peak systolic volume Doppler to
screen for fetal anaemia. Timely sourcing of compatible blood products is likely
to be highly challenging, and both directed and autologous donation should be
contemplated where appropriate. The International Red Cell Donor Panel may
Received: 28 September 2020,
also provide access to compatible products.
revised 12 November 2020,
accepted 15 November 2020, Key words: haemolytic disease of the fetus and newborn, immunohaematology,
published online 16 December 2020 patient blood management, RBC antigens and antibodies, transfusion strategy.
591
592 P. Di Ciaccio et al.
are a mixture of IgG and IgM isoantibodies against P1, P dependent on the transferase encoded by the B3GALNT1
and Pk. This occurs as early as infancy [5,6]. Although gene (Fig. 1). The p phenotype arises in the context of
anti-P1 is generally a weak cold antibody without clinical homozygous mutation of the A4GALT gene, which apart
significance, anti-P and anti-Pk are potent haemolysins from coding for P1 and Pk, is also involved in upstream
associated with severe haemolytic transfusion reactions construction of the P antigen (Fig. 1). Over 30 different
[7]. polymorphisms in A4GALT resulting in the p phenotype
Anti-PP1Pk antibodies are also associated with recur- have been reported, the majority being single nucleotide
rent spontaneous abortion, particularly in the first trime- variants or small indels causing a variety of missense or
ster [3,8–10]. P and Pk in particular are expressed at high nonsense mutations [18]. Defects in B3GALNT1 cause the
levels on the placenta, and antibodies to both of these P1k and P2k phenotypes, which are even rarer than p
antigens are deleterious to placental trophoblasts. It is (Table 1) [1,19–22].
likely that both the IgG and IgM components are patho- We present a series of four successful pregnancies in
logical [3,9,11–15]. Although occurring only in a minor- three patients whose management at our institutions was
ity of cases, haemolytic disease of the fetus and newborn complicated by the presence of anti-PP1Pk isoantibodies.
(HDFN) has also been reported [16,17].
The synthesis of both P1 and Pk is dependent upon the
Case 1
action of alpha 1-4-galactosyltransferase, encoded by the
A4GALT gene which catalyses the addition of galactose A 22-year-old woman of Indian descent (Tamil Nadu
to lactosylceramide glycosphingolipid. The P antigen region) was referred at five weeks of gestation due to the
belongs to the separate globoside (GLOB) family, presence of anti-PP1Pk antibodies, detected during
Fig. 1 Formation of the P1, Pk and P antigens by sequential additions of monosaccharides, with catalysing enzymes. Simplified from Kazcmarek et al.
(2014) [42]. [Colour figure can be viewed at wileyonlinelibrary.com]
Table 1 Phenotypes and corresponding antibodies. Adapted from Hell- and four cryopreserved compatible red cell units in Japan.
berg et al. 2013 [1] Arrangements were made for the collection and transfer
of fresh red cells from these two donors prior to the
Antigens Present
anticipated time of delivery.
Phenotype Prevalence on red cells Isoantibodies
Although elective induction of labour was planned
P1 20–90% P1, Pk, P - during week 39 of gestation, the patient entered sponta-
P2 10–80% Pk, P Anti-P1 neous labour at 38 + 0 weeks, delivering by emergency
p null Rare - Anti-PP1Pk caesarean section (CS) for fetal compromise demonstrated
P1k Rare P1, Pk Anti-P by tachycardia on cardiotocography. The estimated blood
P2k Rare Pk Anti-P, P1 loss was 300 ml, and maternal transfusion was not
required. The newborn had Apgar scores of nine at 1 min
and nine at 5 min, and weighed 3140 g (within normal
routine antenatal screening. This was her first known range) [24]. The labour and delivery occurred prior to the
pregnancy, and there was no history of blood product arrival of the p red cells in Australia.
transfusion. Her red cell phenotype was A1, R1R1, K-k+, The newborn’s haemoglobin was initially normal at
Fy(a + b-), Jk(a-b+), MSs, p, Le(a-b-). The paternal pheno- 165 g/l. The cord blood group was B positive and the
type was P1+, and the couple were first cousins. There direct antiglobulin test (DAT) with monospecific antihu-
was no family history of the p phenotype, transfusion man globin reagents was positive for IgG and negative
reactions or miscarriages. for complement, using column agglutination. A panag-
Anti-PP1Pk antibodies were detected by room tempera- glutinin was eluted, reacting against all cells, including
ture saline indirect antiglobulin testing (IAT), polyethy- group O cells. This was consistent with anti-PP1Pk,
lene glycol (PEG) IAT and papain IAT. The initial though a concurrent anti-B could not be definitively
anti-PP1Pk titre was 1:8 by saline IAT tested against a excluded due to insufficient sample for further testing.
pool of PP1Pk cells at the Australian Red Cross Lifeblood The newborn was moderately jaundiced with a serum
(Lifeblood) Red Cell Reference Laboratory. All future titra- bilirubin (SBR) of 74 µmol/l at 6 h of age, initially man-
tions were performed centrally at this laboratory and aged with phototherapy. A fall in haemoglobin (133 g/L;
tested in parallel with the previous sample with titres normal range 142–240 g/l) and rise in SBR to a peak of
peaking at 1:64. Anti-K could not be definitively 119 µmol/l was consistent with evolving haemolysis and
excluded due to the lack of K + p reference cells. managed with IVIg 1 g/kg and ongoing phototherapy.
Genotyping revealed homozygosity for the Both mother and newborn were discharged six days after
NM_017436.7:c.752C> T point mutation in the A4GALT birth and remained well six months later.
gene. This mutation results in the missense mutation
p.Pro251Leu. This is a previously reported mutation in
A4GALT, known to result in the p phenotype in the
Case 2
homozygous state [23]. A 31-year-old woman of Burmese descent presented at
Given the presence of anti-PP1Pk, the decision was 6 weeks of gestation in the context of known anti-PP1Pk
made to observe the pregnancy in a high-risk maternal antibodies. She had two previous pregnancies. The first
fetal medicine setting, with close monitoring every pregnancy ended in miscarriage early in the first trimester
4 weeks of the antibody titre, as well as fetal well-being with anti-PP1Pk first detected at the time of fetal loss. For
by serial ultrasound, including middle cerebral artery the second pregnancy, thrice weekly therapeutic plasma
peak systolic velocity (MCA PSV), which was suggestive exchange (TPE), in conjunction with intravenous
of mild fetal anaemia (MCA Vmax 626 cm/s). The patient immunoglobulin (IVIg), was administered; however, the
had a normal full blood count. Vitamin B12 and folate pregnancy miscarried at eight weeks. At the time of fetal
stores were replete. There was a mild iron deficiency (fer- demise, the titre was 1:4. The maternal phenotype was B,
ritin 14 µg/l; normal range 20–220 µg/l), corrected with R1R2, K-k+, Kp(a-b+), Fy(a + b-), Jk(a + b-), Mss, p, Le(a-
oral supplementation. b+), Lu(a-). The paternal phenotype was P1+, with no
An exhaustive search for potential p blood donors was consanguinity.
undertaken. There were no compatible donors in Australia The initial titre of anti-PP1Pk in this third pregnancy
or New Zealand. Compatible donors from the patient’s was 1:2, and only trace after treatment with dithiothreitol
immediate family could not be identified. Autologous col- (DTT), suggesting a mix of IgG and IgM. No other anti-
lection was not offered due to safety concerns during bodies were detected after alloadsorption, by either saline
pregnancy. An international search via the International IAT or PEG IAT. Maternal full blood count and haema-
Rare Donor Panel (IRDP) located two potential p donors, tinic indices were normal.
The pregnancy was managed expectantly, with serial The DAT was negative. There was mild jaundice with the
antibody titres, as well as serial fetal ultrasound, includ- SBR rising to 220 µmol/l; however, treatment with pho-
ing MCA PSV, which remained within normal limits. The totherapy was not required.
patient was offered TPE and IVIg, however declined. She
was prescribed aspirin 100 mg daily until week 36 for
Case 3
placental protection. The anti-PP1Pk titre showed minimal
fluctuation, peaking at 1:8 at week 24, falling back to 1:2 A 28 year-old-lady of Burmese descent, the sister of Case
at week 36. 2, presented at 12 weeks of gestation. This was her first
No compatible donors or cryopreserved units in Aus- pregnancy, and anti-PP1Pk antibodies were detected on
tralia or New Zealand were identified, nor were there any routine antenatal screening. The maternal red cell pheno-
eligible directed family donors. A search via the IRDP type was O, R1R1, K-k+, Fy(a+,b-), Jk(a + b+), p, MSs.
located two potential p donors in Japan. Arrangements The paternal phenotype was P1+, with no consanguinity.
were made for the collection and importation of fresh red The initial titre of anti-PP1Pk was 1:4. No other anti-
cells from these two donors prior to the anticipated date bodies were detected by saline IAT or PEG IAT after
of delivery. alloadsorption, though definitive exclusion of anti-K was
The patient underwent elective CS at 37 + 6 weeks for not possible due to the absence of K + p reference cells.
a breech presentation. There was a mild postpartum The antibody titre was monitored throughout the preg-
haemorrhage, with 500 ml estimated blood loss. The new- nancy, remaining at 1:4.
born weighed 2668 g (normal for gestational age), with The patient was offered TPE and IVIg, however
Apgar scores of eight at 1 min and nine at 5 min [24]. declined. No compatible donors or cryopreserved units in
Cord blood showed haemoglobin of 162 g/L, blood group Australia or New Zealand were available, nor were there
B positive and a negative DAT. There was no neonatal any suitable family donors. An international search iden-
jaundice. Both mother and newborn were discharged well tified a potential donor in the United States; however,
after routine postoperative care. donation could not be arranged prior to delivery.
Nine months following this first successful pregnancy, Although MCA PSVs remained within normal limits
the patient re-presented for antenatal care with her fourth throughout pregnancy and there was no evidence of pre-
pregnancy at 16 weeks of gestation. Anti-PP1Pk was eclampsia or infection, the fetus demonstrated intrauter-
detected by saline IAT at 22°C and 37°C and tube low- ine growth restriction (IUGR) and was consequently
ionic IAT at her initial attendance, but its titre was not induced at 36 + 6 weeks and delivered by vaginal ven-
able to be estimated due to insufficient blood sample vol- touse delivery. There was no antenatal or postnatal haem-
ume. No additional alloantibodies were detected by saline orrhage. The newborn was small for gestational age at
or low-ionic IAT following alloadsorption. 2248 g and demonstrated Apgar scores of nine at 1 min
The patient attended several different institutions for her and nine at 5 min [24]. Cord blood showed a normal hae-
ongoing antenatal care. Once again TPE was offered and moglobin of 156 g/l, blood group B positive and a nega-
declined by the patient. She was prescribed aspirin 100 mg tive DAT, suggesting placental toxicity rather than
daily until week 36. The anti-PP1Pk titre was 1:2 after treat- haemolysis as the cause of IUGR. The newborn developed
ment with DTT at 25 weeks and was unchanged at 33 weeks. mild jaundice and was treated successfully with pho-
Fetal development was normal throughout the pregnancy. totherapy alone.
As was the case with the patient’s previous pregnancy,
no compatible donors in Australia or New Zealand were
Review of the literature
identified. However, the two red cell units originally
imported fresh from p donors in Japan and frozen upon A review of the literature was undertaken to evaluate the
arrival in Australia prior to the delivery date for her pre- available evidence regarding the impact of anti-PP1Pk
vious pregnancy, remained available in the cryopreserved antibodies in pregnancy. A keyword search of the MED-
rare blood inventory at Lifeblood. LINE database was undertaken, limiting to English lan-
The patient was scheduled for a repeat elective CS in guage. Further, we interrogated the references of these
week 38 of gestation for breech presentation; however, articles to identify further additional relevant articles.
she entered spontaneous labour and underwent emer- The identified literature was limited to case reports and
gency CS at 35 + 0 weeks for breech presentation and one case series of two patients (Table 2). Cases were
placental abruption. Her preoperative haemoglobin was included if authors reported a case of the management of
124 g/l, and perioperative blood loss was minimal. The a pregnancy with anti-PP1Pk, or a case of presentation
newborn weighed 2740 g (normal for gestational age) with fetal demise. Overall, 19 pregnancies complicated by
[24]. Apgar scores were nine at 1 min and nine at 5 min. anti-PP1Pk, in 16 different women, were identified.
Miscarriage Maternal
Reference Year Maternal Age Ethnicity history Phenotype Titre Management Fetal Outcome Complications
Rock [22] 1985 28 Kuwaiti 13 (1st P1k 1:32 (1:2 to 1:4 on TPE Delivered 33 weeks
Shirey [9] trimester) TPE) BW 1930 g
Positive DAT Jaundice
(phototherapy)
Hanafusa [32] 2006 36 Japanese 2 (2nd p 1:256 (<1:16 on TPE (double Term TPE catheter
trimester) TPE) filtration TPE) BW 2946 g infectionOccluded
(Continues)
Table 2 (Continued)
Miscarriage Maternal
Reference Year Maternal Age Ethnicity history Phenotype Titre Management Fetal Outcome Complications
BW, birth weight; CS, caesarean section; DAT, direct antiglobulin test; HDFN, haemolytic disease of the fetus and newborn; IVIg, intravenous immunoglobulin; IUGR, intrauterine growth restriction; TPE, thera-
peutic plasma exchange.
Reports of women or families simply with history of HDFN in either. The anti-PP1Pk titre in both of these
early-term miscarriage and anti-PP1Pk were not included. pregnancies was 1:16 at maximum [14].
One case concerned anti-P in a P1k woman.
Ten of the 16 women had a history of recurrent mis-
Discussion
carriage (defined as two or greater), typically in the first
trimester, though occasionally in the second. This history Anti-PP1Pk isoimmunization in pregnancy is very rare,
was unreported in three women. One patient had experi- and the associated published literature very limited. It is
enced 13 consecutive first-trimester miscarriages [22]. important that a multidisciplinary approach is provided
Three fetal losses in total were reported amongst the to such patients, ideally within a specialist maternal fetal
reviewed cases. One presented for the first time with a mis- medicine service.
carriage with a titre of 1:64, another patient miscarried Even though a positive DAT is frequently noted on
despite TPE from week nine and a third miscarried despite newborn cord blood, clinically significant HDFN appears
TPE commenced at week five [4,10,25]. Of note, either fetal uncommon, with only two definite cases reported, in
distress, anaemia or intrauterine growth restriction (IUGR) addition to Case 1 in our series [16,17]. Nevertheless, it is
was noted in four pregnancies that did not end in demise wise to monitor for fetal anaemia with serial MCA PSV.
[16,26,27]. A positive neonatal DAT was noted in five preg- There is, however, a clear link between anti-PP1Pk and
nancies, two of which occurred in jaundiced but not anae- recurrent miscarriage, due to direct placental toxicity [12,32].
mic neonates. Two cases of severe HDFN were noted [17]. The cornerstone of management appears to be TPE. Where
One woman suffered severe postpartum haemorrhage due there is a history of recurrent first-trimester miscarriage, the
to placenta accreta, and required transfusion of autologous literature suggests that early initiation of TPE between weeks
and directed red cells [28]. five and eight is important, with continuation at least into
There was high variability in whether titres were the third trimester, and possibly up to delivery [30]. Aggres-
reported before or after treatment with DTT, or whether sive regimes, typically three times a week, have been most
this was performed at all. Given this, as well as the fact commonly employed. It may be possible to taper the therapy
that both IgG and IgM isotypes are likely pathological, if antibody titres are low and well-controlled.
titres before DTT treatment, where stipulated by the The American Society for Apheresis (ASFA) supports
authors, are referred to [14]. the use of TPE for alloimmunization in pregnancy,
The peak titre, which was at the time of diagnosis in though does note that decision-making should be highly
all cases, ranged between 1:16 and 1:256, with a median individualized based on the low-quality of evidence.
titre of 1:128. There was no significant correlation ASFA Guidelines suggest commencement early in gesta-
between miscarriage and titre, though not all cases tion, exchanging one to 15 TPVs [33].
reported titres. For the two fetal losses where a titre was It should be noted, however, that TPE is highly
reported, they were 1:64 and 1:128, respectively. resource intensive, may not be available in all settings,
A variety of management strategies were employed, and is associated with certain risks, including catheter
most consistently TPE, which was administered in 12 of infection, hypotension and hypocalcaemia [34]. Our series
19 pregnancies. Most commonly, the replacement fluid of patients demonstrates four successful pregnancies
was human albumin, with approximately one total without TPE; however, one newborn had mild-to-moder-
plasma volume (TPV) exchanged in each session. Consis- ate HDFN, and another IUGR and fetal distress.
tently, TPE was started early in the first trimester, com- It is likely relevant that our patients presented with low
monly the fifth or sixth week of gestation, at a frequency titres of anti-PP1Pk relative to those reported in the litera-
between once to five times per week. ture. Of the reported cases involving the use of TPE, base-
Intravenous immunoglobulin was given in three preg- line titres ranged from 1:32 to 1:512. The majority resulted
nancies, in all cases concurrently with TPE [27,29,30]. One in viable births where the titre was maintained at <1:32. It
pregnancy was managed with oral dydrogesterone, due to is difficult, if not impossible, to be confident that there is a
the resource-poor setting, with delivery of a normal infant ‘safe’ titre. Notably, Case 2 in our series had previously
[31]. One patient was managed with random plasma infu- miscarried despite TPE and IVIg, with a titre of 1:4 at that
sions, presumably in an effort to adsorb anti-PP1Pk; how- time. Conversely, the titre in Case 1, a successful pregnancy
ever, both her pregnancies were complicated by IUGR and without specific intervention, incremented as high as 1:64
fetal distress, requiring emergent CS in prematurity [26]. during the pregnancy. Conceivably, a history of early mis-
Two pregnancies in two members of an Amish family carriage is more relevant in identifying the women who
were managed conservatively, both resulting in term will benefit most from TPE. In our series, cases 1 and 3 had
deliveries of a normal infant. One infant had a positive no history of pregnancy loss, and Case 2 had only two
DAT and the other did not, though there were no signs of early-term miscarriages, the second despite TPE therapy.
The role of IVIg is poorly defined in anti-PP1Pk-associ- consequent incompatibility with local thawing equipment.
ated pregnancies, and its antenatal use has only been Postpartum, patients with rare red cell phenotypes should
reported in conjunction with TPE. IVIg has been used for be encouraged to become blood donors.
several decades in cases of severe alloimmunization in Autologous donation and cryopreservation postnatally,
pregnancy [35]. Proposed mechanisms include inhibition in anticipation of future possible pregnancies, should be
of endogenous IgG production, saturation of Fc receptors strongly considered. Family members should be screened
on effector cells and placental tissue, and increased cata- as potential directed donors, and the paternal phenotype
bolism of endogenous IgG [36–38]. Its use in the context defined.
of anti-PP1Pk should be individualized. Limitations of the literature may include a publication
Another challenge is the identification and supply of bias of cases where interventions such as TPE were suc-
compatible red cells for transfusion. Management of a cessfully utilized, and of women with more extensive his-
pregnant patient with a rare red cell phenotype requires tories of miscarriage. Another difficulty in interpreting
careful assessment of the risks of antenatal, perinatal and the literature is inconsistency in whether and how anti-
postpartum haemorrhage, the anticipated need for trans- body titres are reported.
fusion and availability of compatible red cell components.
Ideally this information should be documented in a trans-
Conclusion
fusion plan communicated to all members of the treating
team. If the neonate is considered at risk of HDFN, plans We present a series of three patients with the very rare p
to access compatible red cells or IVIg may be required. phenotype, complicated by anti-PP1Pk alloimmunization
If available, fresh or thawed deglycerolized phenotype- in pregnancy. We describe a total of four successful
matched red cells can be used. If fresh phenotype-matched pregnancies in these patients with a conservative
red cells are not available, management of potential massive approach to management, which may be possible in
postpartum haemorrhage should include consideration of cases with very low antibody titre. Treatment should be
both transfusion and obstetric issues. Confirmation of a individualized, however, and strong consideration be
planned delivery time during working hours is sensible so given to early initiation of TPE, especially in women
that specialist services are readily available. Patient blood with a previous history of miscarriage. Management
management strategies during pregnancy including iron within a high risk fetal maternal medicine unit is recom-
repletion should be encouraged [39]. If no blood is available, mended, where possible, and monitoring for IUGR and
patients should be managed as per protocols developed for fetal anaemia is essential.
patients who decline transfusion [40].
Autologous collection during pregnancy is not rou-
Acknowledgements
tinely available in Australia but may be available in cen-
tres elsewhere. Typing of first-degree relatives should be Australian Federal and State Governments fund Aus-
considered; phenotype-matched siblings should be con- tralian Red Cross Lifeblood to provide blood products and
sidered for directed donations and future autologous and/ services to the Australian community. The authors would
or allogeneic donation. In our series, directed donation like to thank the staff of the Lifeblood Red Cell Reference
was precluded in cases 1 and 3 by ABO incompatibility Laboratories for their valuable assistance with the testing
and in Case 2 by low donor body weight (42 kg). of these samples.
Rare red cell units can also be accessed by the Interna-
tional Rare Donor Panel (IRDP) [41]. Lifeblood currently
Funding
has one group O positive donor on its panel who is p.
Both fresh or thawed deglycerolized red cells may be There are no sources of specific funding for this research.
accessed from overseas. However, obtaining frozen red
cells is often not feasible due to difficulties ensuring the
Conflict of interests
cold chain during long-distance transport, as well as
regional variability in packs used for storage and The authors declare no conflict of interests.
References
1 Hellberg A, Westman JS, Thuresson B, 2 Race RR, Sanger R: Blood Groups in 3 Cantin G, Lyonnais J: Anti-PP1Pk and
et al.: P1PK: the blood group system Man. 6th edn. Oxford, Philadelphia, early abortion. Transfusion 1983;
that changed its name and ex- PA: Blackwell Scientific Publica- 23:350–351
panded. Immunohematology 2013; tions; distributed by J. B. Lippincott; 4 Yoshida H, Ito K, Kusakari T, et al.:
29:25–33 1975 Removal of maternal antibodies from
a woman with repeated fetal loss due literature. J Gynecol Obstet Biol Reprod rare P blood group with anti-PP1Pk.
to P blood group incompatibility. (Paris) 1993; 22:393–397 Geburtshilfe Frauenheilkd 1991;
Transfusion 1994; 34:702–705 17 Levene C, Sela R, Rudolphson Y, et al.: 51:710–713
5 McNeil C, Trentelman EF, Thomas M: Hemolytic Disease Of The Newborn 28 Conlan S: Managing a pregnancy in
Anti-Tja-another family example. Vox Due To Anti-PPIPk (Anti-Tja). Trans- the presence of the rare blood group
Sang 1957; 2:114–116 fusion 1977; 17:569–572 antibody PP1Pk. Aust N Z J Obstet
6 Reeves HM, Cary V, Mino MA, et al.: 18 ISBT: Names for P1PK (ISBT 003) Gynaecol 2004; 44:479–480
Unexpected non-maternally derived Blood Group Alleles. ISBT. https:// 29 Heim MU, Weidemann R, Brehm G,
anti-PP1P(k) in an 11-week-old www.isbtweb.org/fileadmin/user_upload/ et al.: Plasmapheresis and/or inter-
patient. J Pediatr 2017; 181:302–305 files-2015/red%20cells/blood%20group mediate high dosage immunoglobulin
7 Poole J, Daniels G: Blood group anti- %20allele%20terminology/allele%20 therapy–effective measures in threa-
bodies and their significance in trans- tables/003_P1PK_Alleles_v3.1.pdf [Last tened premature labor and intrauterine
fusion medicine. Transfus Med Rev accessed 28 April 2020] hemolysis in anti-PP1Pk (-Tja)? Beitr
2007; 21:58–71 19 Kaczmarek R, Buczkowska A, Mikola- Infusionsther 1990; 26:302–306
8 Levine P, Koch EA: The rare human jewicz K, et al.: P1PK, GLOB, and 30 Fernandez-Jimenez MC, Jimenez-
isoagglutinin anti-Tja and habitual FORS blood group systems and GLOB Marco MT, Hernandez D, et al.:
abortion. Science 1954; 120:239–241 collection: biochemical and clinical Treatment with plasmapheresis and
9 Shirey RS, Ness PM, Kickler TS, et al.: aspects. Do we understand it all yet? intravenous immunoglobulin in preg-
The association of anti-P and early abor- Transfus Med Rev 2014; 28:126–136 nancies complicated with anti-PP1Pk
tion. Transfusion 1987; 27:189–191 20 Steffensen R, Carlier K, Wiels J, et al.: or anti-K immunization: a report of
10 Shechter Y, Timor-Tritsch IE, Lewit N, Cloning and expression of the histo- two patients. Vox Sang 2001; 80:117–
et al.: Early treatment by plasma- blood group Pk UDP-galactose: 120
pheresis in a woman with multiple Ga1beta-4G1cbeta1-cer alpha1, 31 Mohd Azri MS, Kunasegaran K, Azrina
abortions and the rare blood group p. 4-galactosyltransferase. Molecular A, et al.: Successful pregnancy out-
Vox Sang 1987; 53:135–138 genetic basis of the p phenotype. J come in malaysian woman with rare p
11 A4GALT alpha 1,4-galactosyltrans- Biol Chem 2000; 275:16723–16729 phenotype and anti-PP1P(k) antibody.
ferase (P blood group). Gene. https:// 21 Suzuki N, Yamamoto K: Molecular Indian J Hematol Blood Transfus
www.ncbi.nlm.nih.gov/gene/53947 cloning of pigeon UDP-galactose: 2014; 30:405–408
[Last accessed 28 April 2020] beta-D-galactoside alpha1, 4-galac- 32 Hanafusa N, Noiri E, Yamashita T,
12 Lindstrom K, Von Dem Borne AE, tosyltransferase and UDP-galactose: et al.: Successful treatment by double
Breimer ME, et al.: Glycosphingolipid beta-D-galactoside beta1,4-galactosyl- filtrate plasmapheresis in a pregnant
expression in spontaneously aborted transferase, two novel enzymes catalyz- woman with the rare P blood group
fetuses and placenta from blood group ing the formation of Gal alpha1-4Gal and a history of multiple early miscar-
p women. Evidence for placenta being beta1-4Gal beta1-4GlcNAc sequence. J riages. Ther Apher Dial 2006; 10:498–
the primary target for anti-Tja-anti- Biol Chem 2010; 285:5178–5187 503
bodies. Glycoconj J 1992; 9:325–329 22 Rock JA, Shirey RS, Braine HG, et al.: 33 Padmanabhan A, Connelly-Smith L,
13 Kojima Y, Fukumoto S, Furukawa K, Plasmapheresis for the treatment of Aqui N, et al.: Guidelines on the Use
et al.: Molecular cloning of globo- repeated early pregnancy wastage of Therapeutic Apheresis in Clinical
triaosylceramide/CD77 synthase, a gly- associated with anti-P. Obstet Gynecol Practice - Evidence-Based Approach
cosyltransferase that initiates the 1985; 66:57S–60S from the Writing Committee of the
synthesis of globo series glycosphin- 23 National Center for Biotechnology American Society for Apheresis: The
golipids. J Biol Chem 2000; 275: Information. ClinVar; 2019. [Last accessed Eighth Special Issue. J Clin Apher
15152–15156 2 May 2020] https://www.ncbi.nlm.nih. 2019; 34:171–354
14 Benidt GR, Jaben EA, Winters JL, gov/clinvar/variation/VCV000002693.1 34 Shemin D, Briggs D, Greenan M: Com-
et al.: Identification of anti-PP1P(k) in 24 Joseph FA, Hyett JA, Schluter PJ, plications of therapeutic plasma
a blood donor and her family: a case et al.: New Australian birthweight cen- exchange: a prospective study of
report following her pregnancy and tiles. Med J Aust 2020; 213:79–85 1,727 procedures. J Clin Apher 2007;
review. Transfus Apher Sci 2010; 25 Weiss DB, Levene C, Aboulafia Y, et al.: 22:270–276
43:369–374 Anti-PP1Pk (anti-Tja) and habitual abor- 35 Scott JR: Revisiting the use of intra-
15 Lopez M, Cartron J, Cartron JP, et al.: tion. Fertil Steril 1975; 26:901–903 venous immune globulin (IVIG) for
Cytotoxicity of anti-PP1Pk antibodies 26 Haentjens-Verbeke K, Dufour P, Vina- Kell alloimmunization. Am J Obstet
and possible relationship with early tier D, et al.: Anti-TJa alloimmunization Gynecol 2018; 219:223–224
abortions of p mothers. Clin Immunol (anti-PP1Pk): two consecutive pregnan- 36 de la Camara C, Arrieta R, Gonzalez
Immunopathol 1983; 28:296–303 cies of an anti-TJa-carrying patient. A, et al.: High-dose intravenous
16 Haentjens-Verbeke K, Dufour P, Vina- Fetal Diagn Ther 1996; 11:120–125 immunoglobulin as the sole prenatal
tier D, et al.: Anti-Tja (PP1Pk) isoim- 27 Strowitzki T, Wiedemann R, Heim MU, treatment for severe Rh immunization.
munization. A case, a review of the et al.: Pregnancy with an extremely N Engl J Med 1988; 318:519–520
37 Urbaniak SJ, Duncan JI, Armstrong- 5 Obstetrics and Maternity. Canberra: 42 Kaczmarek R, Buczkowska A,
Fisher SS, et al.: Transfer of anti-D anti- National Blood Authority (Australia), Mikołajewicz K, et al.: P1PK, GLOB,
bodies across the isolated perfused 2015 and FORS blood group systems and
human placental lobule and inhibition 40 Kidson-Gerber G, Kerridge I, Farmer S, GLOB collection: biochemical and
by high-dose intravenous immunoglob- et al.: Caring for pregnant women for clinical aspects. Do we understand it
ulin: a possible mechanism of action. Br whom transfusion is not an option. A all yet. Transfus Med Rev 2014;
J Haematol 1997; 96:186–193 national review to assist in patient 28:126–136
38 Yu Z, Lennon VA: Mechanism of intra- care. Aust N Z J Obstet Gynaecol 43 Taniguchi F, Horie S, Tsukihara S,
venous immune globulin therapy in anti- 2016; 56:127–136 et al.: Successful management of a
body-mediated autoimmune diseases. N 41 Nance S, Scharberg EA, Thornton N, p‐incompatible pregnancy using dou-
Engl J Med 1999; 340:227–228 et al.: International rare donor panels: ble filtration plasmapheresis. Gynecol
39 National Blood Authority: Patient a review. Vox Sang 2016; 110:209– and Obstet Invest 2003; 56:117–
Blood Management Guidelines: Module 218 120.
Abstract
Background The continual identification of rare blood among donors is critical
to support national programs like the American Rare Donor Program (ARDP).
Some blood centres require consent from donors to be registered with a national
registry. This situation provides an opportunity to determine whether a donor’s
willingness to register is associated with a change in donation behaviour.
Methods Rare donors were identified by molecular typing. The average number
of donations per year was compared for each donor prior to and after receiving a
consent letter. Donors were categorized as either accepting or declining the
request. Non-parametric t tests compared the statistical significance within and
between categories. Rare types were overlaid with consensus data to look for
trends using data visualization techniques.
Results A total of 270 molecularly typed rare donors received letters over
4 years. Half of the donors (132, 49%) agreed to participate in the ARDP. Overall,
donation frequency increased after the letter when enrolled. Both Caucasian and
non-Caucasian donors increased their donations after enrolling providing an
additional 159 red blood cell units over 3 years. Declining participation did not
change donation frequency. Data visualization showed that enrolled donors were
more affluent, high school and college educated, and lived in their home for
longer periods of time.
Conclusion A donor’s willingness to enrol in the ARDP was associated with a
post-response increase in donation frequency. New interventions to reach non-
Caucasian donors may be a prerequisite to increase donation frequency and a
Received: 10 August 2020,
willingness to be a rare blood donor.
revised 27 October 2020,
accepted 9 November 2020, Key words: donor recruitment, rare blood, mass-scale red cell genotyping, donor
published online 5 January 2021 geo mapping.
601
602 W. Q. Anani et al.
and non-Caucasian donors secondary to sample size of visualized with 3D Map (Microsoft Excel 2016, version
minority donors. Instead, all minority donors were 1803) [24]. The donor and extracted population data were
grouped as non-Caucasians and analysed as one group. superimposed on geographic maps by zip code.
Fig. 1 Rare donors displayed a variety of desirable antigen-negative genotypes sorted by prevalence.
Fig. 2 Rare donors were found in all donor age groups by decade. The largest single group of donors were ages 20–29 with a predominance of Afri-
can-American donors. Overall, rare donors were most likely African American.
060–40) and after (10, range 060–33) the letter was Of non-Caucasian donors, 49 (29%) agreed to participate
sent. The number of donations before and after respond- in the ARDP with a median of 13 (range 040–30) dona-
ing was statistically insignificant (P = 052). In total, the tions per year before and 17 (range 050–48) donations
two groups’ donation patterns were statistically different per year after the letter was sent (Fig. 3b). The 121 non-
after receiving the ARDP letter (P = 00004). Caucasians declining ARDP participation had a median of
The data were then categorized by Caucasian and non- 10 (range 067–40) and 10 (range 060–30) donations
Caucasian groups to look for further trends. Of Caucasian before and after the letter was sent, respectively. The num-
donors, 83 (83%) agreed to participate in the ARDP with ber of donations before and after responding within each
a median of 21 (range 016–45) donations per year group was statistically significant for donors declining
before and 25 (range 060–57) donations per year after (P = 0015) and participating (P = 00090) in the ARDP.
the letter was sent (Fig. 3b). The 17 Caucasian donors The two groups’ donation patterns were statistically differ-
declining ARDP participation had a median of 13 (range ent after receiving the ARDP letter (P < 00001).
060–37) and 12 (range 060–33) donations per year Overall, irrespective of ethnicity, donors enrolling in
before and after the letter was sent, respectively. The the ARDP resulted in an increase in the median donation
number of donations before and after responding within rate by 04. Compared with donors not enrolling in the
each group was only statistically significant for donors rare donor program, the median donation rate increase in
participating in the ARDP (P = 0050 vs <00001). The Caucasian donors increased red blood cells collected
two groups’ donation patterns were statistically different from 523 to 623 units, leading to an additional 100 red
after receiving the ARDP letter (P = 00024). blood cell units over three years. Non-Caucasian donors
Fig. 3 (a) Overall donation patterns in response to a rare donor letter. Donors not registering to be a rare donor did not change their frequency signifi-
cantly. In contrast, donors that responded favourably increased the number of donations per year significantly. (b) When the donation patterns were sep-
arated, Caucasian that declined participation (n = 17) in the rare donor program was not significant in comparison with non-Caucasian donors
(n = 121) that decreased their donation frequency. Both groups had statistically significant increases in donations following enrolment in the rare donor
program (Caucasian n = 83 and non-Caucasian n = 49).
Fig. 4 Rare donors were mapped by zip code with and overlaid with census data to look for commonalities among donors. Reference maps (a, b, c)
display donors by ethnicity in pie charts, age in bar charts and census data in colour gradients. Greater than 98% of the donor data were captured in
the map screenshots. Non‐Caucasian donors lived in defined zip codes and were younger on average than Caucasian donors. Non‐Caucasian donors had
lower incomes, lived in their homes for less time and were less likely to hold a high school education or Bachelor’s degree (d, e, f). The size of the pie
chart represents the total number of donors in the zip code. [Colour figure can be viewed at wileyonlinelibrary.com]
address cannot reach the intended recipient and donors only mailing letters [34]. By identifying donors less likely
are lost to follow-up. Engaging younger donors with or unfavourably respond to standard methods, such as
phone applications [32] and improved marketing strate- relatively short-term geographic residence, the monetary
gies [33] can improve the conversion of first-time to donor acquisition cost can be redirected to a more appro-
repeat donors. Our data suggest frequent donors are more priate method. Improving the first interaction with a rare
likely to consent when approached to participate in the donor can be crucial to encouraging life-long donation
ARDP. The overlying census data with our donor geogra- frequency [35]. Once established, a program could easily
phies do not provide causality of why a donor may not expand to identify and target uncommon donors such as
respond. These data instead highlight areas for further R0R0/R0r to donate more frequently, which are also in
investigation (e.g. messaging donors from a particular high demand due to guideline recommendations.5
community or testing new methods of communication to We demonstrated that rare donors enrolling in the
effect a change in donor frequency). The retrospective ARDP increase their average number of donations per
nature of this study has some limitations. The baseline year. The increased donations would result in a meaning-
donation frequency of donors enrolling in the ARDP was ful rise in rare units available for a blood centre. Addi-
higher than donors declining participation. This study tionally, rare donors not responding as favourably to
was not designed to ascertain differences in donor char- letters and phone calls were more likely non-Caucasian.
acteristics. Due to the limited number of rare donors, eth- These donors tended to be less wealthy, less educated
nicities and lifetime number of donations could not be and more transient making contact potentially more dif-
completely stratified. The passive nature of collecting data ficult. However, contact with these donors did not result
precluded varying the communication method to observe in a decrease in donations. By applying geo mapping
an effect. Thus, we cannot conclude which contact more broadly, other valuable donors could be located
method would be appropriate for unresponsive donors. and more effectively, targeted with messaging that res-
Regardless, the financial contraction of the blood onates better based on various donor groups and demo-
industry requires more efficient use of the current graphics.
resources and optimizing collection strategies. Direct costs
to recruit rare red blood cell donors are unknown, but in
Conflict of interest
2007, it was estimated that selective recruitment of family
members for HLA matched haematopoietic stem cell The authors declare that they have no conflicts of interest
screening cost approximately €7 per new donor found for related to this manuscript.
References
1 Yazer MH, Anani WQ, Denomme GA, 6 Fasano RM, Meyer EK, Branscomb J, 12 Paccapelo C, Truglio F, Villa MA,
et al.: Trends in antigen-negative red et al.: Impact of red blood cell antigen et al.: Rare donor program in Italy.
blood cell distributions by racial or matching on alloimmunization and trans- Immunohematology 2016; 32:47–8
ethnic groups in the United States. fusion complications in patients with 13 Soeker R: Rare donor program at Wes-
Transfusion 2018; 58:145–50 sickle cell disease: a systematic review. tern Province Blood Transfusion Service
2 Flickinger C: REGGI and the American Transfus Med Rev 2019; 33:12–23 in South Africa. Immunohematology
rare donor program. Transfus Med 7 Thornton NM: The United Kingdom 2016; 32:57–8
Hemother 2014; 41:342–5 national rare donor panel. Immunohe- 14 Moghaddam M: National rare donor
3 Meny GM, Flickinger C, Marcucci C: matology 2016; 32:67–9 program in Iran. Immunohematology
The American rare donor program. J 8 Goldman M, St Croix L: Rare donor 2016; 32:27–8
Crit Care 2013; 28: 110 e9–e18 program: Canadian Blood Services. 15 Price CL, Boyd JH, Watkins AR, et al.:
4 Seltsam A, Wagner FF, Salama A, Immunohematology 2016; 32:15–6 Mailing of a sickle cell disease educa-
et al.: Antibodies to high-frequency 9 Peyrard T: The French national rare tional packet increases blood donors
antigens may decrease the quality of blood program. Immunohematology within an African American com-
transfusion support: an observational 2016; 32:23–5 munity. Transfusion 2006; 46:1388–
study. Transfusion 2003; 43:1563–6 10 Yahalom V, Finkel L, Shinar E, et al.: 93
5 Compernolle V, Chou ST, Tanael S, The Israeli rare donor blood program. 16 Bachegowda LS, Timm B, Dasgupta P,
et al.: Red blood cell specifications for Immunohematology 2016; 32:11–2 et al.: Impact of predictive scoring
patients with hemoglobinopathies: a 11 Flickinger C: The American rare donor model and e-mail messages on African
systematic review and guideline. program. Immunohematology 2016; American blood donors. Transfusion
Transfusion 2018; 58:1555–66 32:71–3 2017; 57:1515–21
17 Ertas N: Millennials and volunteering: inder.census.gov (Last accessed 1/27/ transfusion management. Transfusion
sector differences and implications for 2019 2019). 2002; 42:1435–41
public service motivation theory. Pub- 24 Au C, Rischpater R: Power Map for 30 Price CL, Johnson MT, Lindsay T, et al.:
lic Adm Quarterly 2016; 40:517–58 Excel; Microsoft Mapping: Geospatial The Sickle Cell Sabbath: a community
18 Paulin M, Ferguson RJ, Schattke K, Development in Windows 10 with Bing program increases first-time blood donors
et al.: Millennials, social media, proso- Maps and C#. Berkeley, CA, Apress, in the African American faith commu-
cial emotions, and charitable causes: 2015:159–65 nity. Transfusion 2009; 49:519–23
the paradox of gender differences. J 25 Wagner FF, Doescher A, Bittner R, 31 Cooper RA, Cooper MA, McGinley EL,
Nonprofit Public Sector Market 2014; et al.: Extended donor typing by et al.: Poverty, wealth, and health care
26:335–53 pooled capillary electrophoresis: utilization: a geographic assessment. J
19 Scholz C: Generation Y and blood impact in a routine setting. Transfus Urban Health 2012; 89:828–47
donation: the impact of altruistic help Med Hemother 2018; 45:225–37 32 Gemelli CN, Carver A, Garn A, et al.:
in a darwiportunistic scenario. Trans- 26 Yee MEM, Josephson CD, Winkler AM, Evaluation of the impact of a personal-
fus Med Hemother 2010; 37:195–202 et al.: Red blood cell minor antigen ized postdonation short messaging ser-
20 Flegel WA, Gottschall JL, Denomme mismatches during chronic transfusion vice on the retention of whole blood
GA: Implementing mass-scale red cell therapy for sickle cell anemia. Trans- donors. Transfusion 2018; 58:701–9
genotyping at a blood center. Transfu- fusion 2017; 57:2738–46 33 Boenigk S, Leipnitz S: Acquiring
sion 2015; 55:2610–5; quiz 09 27 Chapuy CI, Nicholson RT, Aguad MD, potential blood donors in large cities:
21 Shaz BH, James AB, Demmons DG, et al.: Resolving the daratumumab a preference-based donor segmenta-
et al.: The African American church as interference with blood compatibility tion study. J Nonprofit Public Sector
a donation site: motivations and barri- testing. Transfusion 2015; 55:1545–54 Market 2016; 28:364–93
ers. Transfusion 2010; 50:1240–8 28 Anani WQ, Marchan MG, Bensing KM, 34 Schmidt AH, Stahr A, Baier D, et al.:
22 Shaz BH, Demmons DG, Hillyer KL, et al.: Practical approaches and costs Selective recruitment of stem cell
et al.: Racial differences in motivators for provisioning safe transfusions dur- donors with rare human leukocyte
and barriers to blood donation among ing anti-CD38 therapy. Transfusion antigen phenotypes. Bone Marrow
blood donors. Arch Pathol Lab Med 2017; 57:1470–9 Transplant 2007; 40:823
2009; 133:1444–7 29 Shirey RS, Boyd JS, Parwani AV, 35 Sojka BN, Sojka P: The blood dona-
23 Bureau USC: American Community et al.: Prophylactic antigen-matched tion experience: self-reported motives
Survey, 2015 American Community donor blood for patients with warm and obstacles for donating blood. Vox
Survey 5-Year Estimates. http://factf autoantibodies: an algorithm for Sang 2008; 94:56–63
609
610 International Forum
20–30
60–70
all hospitals. In fact, the restrictive blood transfusion
125
190
100
100
116
60
33
15
30
55
32
10
strategy is widely extended among centres. Only hospitals
from Japan, Australia, Germany and France use the 2
RBC units transfusion policy. The 1 RBC transfusion pol-
icy has been implemented in many transplantation units
20–30
60–80
most restrictive transfusion policy: a haemoglobin thresh-
250
120
158
350
110
22
45
20
30
32
24
21
old of 70 g/L and 1 RBC unit.
Platelet prophylactic transfusion threshold is mostly
10 9 109/L in the absence of bleeding risk factors, and
(usually the next day), then they perform the platelet count
Minneapolis
Sao Paulo
Besancon
Moscow
Toronto
London
Madrid
Tokyo
Rome
City
Basel
Germany
Australia
Canada
France
Japan
Brazil
Spain
Italy
USA
UK
Country Russia Spain Japan Germany Switzerland Italy Netherlands Australia USA UK Canada France Brazil
CS, steroids; Dara, daratumumab; EPO, erythropoietin; HSCT, haematopoietic stem cell transplantation; IS, immunosuppressive agents; NR, not reported; PE, plasma exchange; PLT, platelets; PRCA, pure red cell
aplasia; RP, random platelets in plasma; RP-AS, random platelets in additive solution; RP-AS-PR, random platelets in additive solution, pathogen reduced; RTX, rituximab; SDP, single donor platelets in plasma;
SDP-AS, single donor platelet in additive solution; SDP-AS-PR, single donor platelet in additive solution pathogen reduced.
*Incidence is provided as percentage of major/bidirectional ABO incompatible HSCT.
International Forum 611
612 International Forum
complication with 24-h continuous infusion of platelets, as topoietic-cell-transplantation. [Last accessed 10 December
previously published [14]. 2020].
Most centres do not apply any special measures to 8 Booth GS, Gehrie EA, Bolan CD, et al. Clinical guide to ABO-
detect early immunohaematological complications in incompatible allogeneic stem cell transplantation. Biol Blood
Marrow Transplant 2013;19:1152–8.
HSCT. Only 5 centres perform DAT systematically in the
9 Raimondi R, Soli M, Lamparelli T, et al. ABO-incompatible
post-transplantation period: weekly (Russia); two times a
bone marrow transplantation: a GITMO survey of current
week for 3 weeks 1 time a week for 2 weeks and monthly practice in Italy and comparison with the literature. Bone
in ABO incompatible transplantations (Italy); at day + Marrow Transplant 2004;34:321–9.
100 after HSCT (Australia); once a week in minor ABO 10 Worel N ABO-mismatched allogeneic hematopoietic stem cell
compatible HSCT (Brazil); and unspecified in ABO minor transplantation. Transfus Med Hemother 2016;43:3–12.
and mixed HSCT (Madrid). 11 Berger MD, Gerber B, Arn K, et al. Significant reduction of
Pure red cell aplasia (PRCA) is related to ABO major red blood cell transfusion requirements by changing from a
incompatible HSCT and characterized by anaemia, reticulo- double-unit to a single-unit transfusion policy in patients
cytopenia and the only absence or near absence of erythroid receiving intensive chemotherapy or stem cell transplanta-
progenitors, having excluded other causes, such as infec- tion. Haematologica 2012;97:116–22.
12 Karafin MS, Blagg L, Tobian AA, et al. ABO antibody titers
tions, haemolysis, disease relapse or drug toxicity. Patients
are not predictive of hemolytic reactions due to plasma
with PRCA can become RBC transfusion dependent for a
incompatible platelet transfusions. Transfusion
long period, suffering an important long-term iron over- 2012;52:2087–93.
load, alloimmunization and transfusion reactions [15]. Pure 13 Josephson CD, Castillejo M, Grima K, et al. ABO-mismatched
red cell aplasia incidence ranges from 0% (Russia) to 15% platelet transfusions: Strategies to mitigate patient exposure
(Switzerland) of ABO incompatible HSCT, and this variabil- to naturally occurring hemolytic antibodies. Transfus Apher
ity is also shown in the literature and could be explained by Sci 2010;42:83–8.
the use of different transplantation platforms, among others 14 Cid J, Guijarro F, Carbasse G, et al. 24-h continuous infusion
[16]. Treatment options are variable, with RBC transfusion of platelets for patients with platelet transfusion refractori-
being the main support approach. Six centres also consider ness. Br J Haematol 2018;181:386–9.
immunosuppressive treatment, and apheresis and daratu- 15 Bolan CD, Leitman SF, Griffith LM, et al. Delayed donor red
cell chimerism and pure red cell aplasia following major
mumab are proposed by one centre (Canada).
ABO-incompatible nonmyeloablative hematopoietic stem cell
transplantation. Blood 2001;98:1687–94.
References 16 Hirokawa M, Fukuda T, Ohashi K, et al. PRCA Collaborative
1 D’Souza A, Lee S, Zhu X, et al. Current use and trends in Study Group. Efficacy and long-term outcome of treatment
hematopoietic cell transplantation in the United States. Biol for pure red cell aplasia after allogeneic stem cell transplan-
Blood Marrow Transplant 2017;23:1417–21. tation from major ABO-incompatible donors. Biol Blood Mar-
2 Passweg JR, Baldomero H, Basak GW, et al. The EBMT activ- row Transplant 2013;19:1026–32.
ity survey report 2017: a focus on allogeneic HCT for nonma-
lignant indications and on the use of non-HCT cell therapies. Guest Editors
Bone Marrow Transplant 2019;54:1575–85. Dr. Pilar Solves
3 Yuan S, Yang D, Nakamura R, et al. RBC and platelet transfu- Hospital Universitari i Politècnic La Fe
sion support in the first 30 and 100 days after haploidentical Valencia, Spain
hematopoietic stem cell transplantation. Transfusion Email: solves_pil@gva.es
2019;59:3371–85.
4 Cohn CS. Transfusion support issues in hematopoietic stem Dr. Miquel Lozano
cell transplantation. Cancer Control 2015;22:52–9. Hospital Clı́nic Universitari
5 Kaufman RM, Djulbegovic B, Gernsheimer T, et al. Platelet Barcelona, Spain
transfusion: a clinical practice guideline from the AABB. Ann Email: mlozano@clinic.cat
Inter Med 2015;162:205–13.
6 Estcourt LJ, Birchall J, Allard S, et al. Guidelines for the use International Forum Editor
of platelet transfusions. Br J Haematol 2017;176:365–94. Dr. Nancy Dunbar
7 Negrin RS. Hematopoietic support after hematopoietic cell Dartmouth Hitchcock Medical Center
transplantation. UpToDate2019. Available from: https://www. Lebanon, NH, USA
uptodate.com/contents/hematopoietic-support-after-hema Email: nancy.m.dunbar@hitchcock.org
e25
e26 International Forum
Question 5
Question 8
Regarding platelet transfusion protocol:
We did not have pure red cell aplasia.
• Prophylactic.
• 10 000/µL if stable and 20 000 if bleeding risk fac-
References tors.
1 Akselrod BA, Balashova EN, Bautin AE, et al. Clinical guideli- • Indistinctly. HLA single donor to alloimmunization
nes for red blood cell transfusion. Russian Journal of Hematol- pts. Yes, the platelets are suspended.
ogy and Transfusiology (Gematologiya i Transfuziologiya). • No.
2018; 63:372-435.https://doi.org/10.25837/HAT.2019.62.39.006
• No.
2 Cid J, Guijarro F, Carbasse G, et al. 24-h continuous infusion of
platelets for patients with platelet transfusion refractoriness. Br
• HLA match single donor.
Question 8
Question 5
Approximately 15%, depending on the number of ABO
We transfuse random platelets prophylactically (transfu-
transplants with major incompatibility. Periodic transfu-
sion threshold is 10 000–20 000/ll). Platelets are sus-
sion mainly. In some very complicated and individual
pended in plasma and ACD-A liquid. In case of platelet
cases, we have used rituximab and also careful review of
transfusion refractoriness, we monitor corrected count
immunosuppression.
increment (about 60 min after transfusion). Additionally,
anti-HLA antibodies and anti-HPA antibodies are mea-
Javier Anguita Velasco
Hospital General Universitario Gregorio Mara~
non sured.
Madrid, Spain
Email: javier.anguita@salud.madrid.org Question 6
Ana Maria Perez-Corral No.
Hospital General Universitario Gregorio Mara~
non
Madrid, Spain
Email: apcorral@salud.madrid.org Question 7
No.
Silvia Monsalvo Saornil
Hospital General Universitario Gregorio Mara~
non
Madrid, Spain Question 8
Email: silvia.monsalvo@salud.madrid.org
About 10–15% of recipients who underwent major ABO-
mismatched HSCT developed PRCA. We continue red
blood cell transfusion until recovery.
Sho Yamazaki & Hitoshi Okazaki
Sho Yamazaki
Tokyo University of Tokyo Hospital
Tokyo, Japan
Email: shiyamazaki-ky@umin.ac.jp
Question 1
We performed 22 autologous and 33 allogeneic HSCT in Hitoshi Okazaki
University of Tokyo Hospital
2019 (approximately 50–60 per year).
Tokyo, Japan
Email: okazakih-ky@umin.ac.jp
Question 2
In Japan, all red blood cells and platelets are irradiated.
- X-ray. Kathleen Selleng, Konstanze Aurich & William Kr€
uger
- 15–50 Gy.
Germany
Question 3
Question 1
We perform a CBC count every day until engraftment in
both cases. Our centre is a University hospital including a haema-
tology department with an outpatient clinic and a spe-
cial unit for transplant patients. The collection of
Question 4
autologous and related donor stem cells is performed in
- We transfuse two units of red blood cells to maintain collaboration of the haematology clinic, the paediatric
a haemoglobin level of 7g/dl. haematology clinic and the transfusion medicine depart-
- One unit of red blood cells is prepared from 200 ml ment. Allogeneic stem cell transplants from unrelated
of whole blood in Japan. donors are imported. We perform between 20–30 autol-
- We switch to donor type blood group when anti- ogous and 20–30 allogenic stem cell transplantations
donor isoagglutinins are undetectable and RBCs of per year.
donor type are present.
References Question 4
1 Treleaven J, Gennery A, Marsh J, et al. Guidelines on the use
of irradiated blood components prepared by the British Com- Transfusion protocol:
mittee for Standards in Haematology blood transfusion task - 70 g/l.
force. Br J Haematol 2011;152:35–51 - 1 RBC unit.
2 Bundes€arztekammer. Querschnitts-Leitlinien zur Therapie mit - We change RBC blood groups to donor type usually
€
Blutkomponenten und Plasmaderivaten Deutscher ArzteVerlag; after 6 months. AB0 antigens must be donor type, and
2009. Available from: https://www.bundesaerztekammer.de/
also RH, K, JK, FY and MNS (extended phenotype). The
fileadmin/user_upload/downloads/QLL_Haemotherapie_2014.
DAT should be negative. Criteria for RBC engraftment
pdf.
are an absolute reticulocyte count of >30 9G/l and the Management information:
absence of RBC transfusion requirements. The detec- (i) red blood cell transfusion (Hb threshold 70 g/l).
tion of donor AB0 blood group is not part of the crite- (ii) we try to taper immunosuppression as fast as possi-
ria. (Neutrophil engraftment is defined as the first of ble.
three consecutive days of an absolute ANC count of (iii) daratumumab and/or rituximab in cases not
>0.5 G/l; PLT engraftment is defined as PLT > 20 G/l resolving spontaneously or not responding to the
without transfusions in the last 7 days). reduction of immunosuppression.
Question 5 Reference
1 Cid J, Guijarro F, Carbasse G, et al. 24-h continuous infusion
Regarding platelet transfusion protocol: of platelets for patients with platelet transfusion refractori-
- Prophylactic. ness. Br J Haematol 2018;181:386–9.
- 10 G/l in stable clinical situation, 20 G/l when there
are risk factors for bleeding such as fever, aGvHD,
30 G/l in case of systemic anticoagulation due to Andreas Buser
thrombosis. Regional Blood Transfusion Service, Swiss Red Cross
- We use both single donor and buffy coat pooled PLT, Basel, Switzerland
all are pathogen reduced using the Intercept Blood Email: andreas.buser@usb.ch
system from CERUS (since 2011) all are in platelet
additive solution (Intersol). No, considering also the Andreas Holbro
suspension in platelet additive solutions (PAS). Regional Blood Transfusion Service, Swiss Red Cross
Basel, Switzerland
- Yes. 15–60 min post-transfusion.
Email: andreas.holbro@usb.ch
- Assess if non-immunological reason for refractoriness,
check for HLA –Antibodies (Luminex). If HLA –Ab are Gregor Stehle
present, we try to prevent transfusion of single donor Regional Blood Transfusion Service, Swiss Red Cross
PLT bearing the relevant antigens, in special cases we Basel, Switzerland
transfuse HLA-matched platelet concentrates. Email: gregor.stehle@usb.ch
- In case of no success, we use PLT drip infusion (de-
scribed by J.Cid and M. Lozano): [1] adult doses are
divided into 4 split doses, and each dose will be con-
Luca Pierelli, Antonella Matteocci & Luigi Rigacci
tinuously infused over a 4–5 h period of time in
order to have a constant (small) amount of PLT in
the circulation supporting endothelial integrity.
Italy
Question 6
Yes. Question 1
- Before HSCT, at day 30 and at day 180.
We perform 20 autologous and 15 allogeneic haemopoi-
- More frequently in case of pure red cell aplasia or
etic transplantations per year.
relapse.
Question 2
Question 7
We irradiate all blood products for autologous and allo-
Yes, we use DAT and Elution techniques (incl. elution
geneic recipients except platelet concentrates (both from
against A and/or B panel cells).
single donors and from pooled buffy coats) for which
irradiation has been replaced by Amotosalen/UV patho-
Question 8 gen inactivation.
Irradiation is applied using a dedicated Caesium-137
The PRCA incidence is about 20% of the HSCT with
source and the energy of 2500 cGy.
major or bidirectional AB0 incompatibility (about 3% of
all allogeneic HSCT).
Question 3
Question 7
All patients, after autologous or allogeneic transplanta-
The basic immunohematologic monitoring of ABO-in-
tion, have a daily complete blood count and leucocyte
compatible recipients includes also direct/indirect
differential till full haemopoietic recovery.
antiglobulin tests and isoagglutinin titration. The direct
antiglobulin test is performed 2 times a week for 3 weeks,
Question 4 1 time a week for 2 weeks and then monthly until the
DAT will be negative. We also test LDH and haptoglobin.
The haemoglobin threshold to perform red blood cell trans-
fusion in stable allografted or autografted patients is 80 g/
l. In routine practice, we use a single-unit red blood cell Question 8
transfusion policy. In ABO-incompatible transplants, we
In our practice, the incidence of pure red cell aplasia in ABO-
change the red blood cell group of RBC concentrates to
incompatible haemopoietic transplantation is around 5%. In
donor cell group when stable full engraftment is obtained
general, patients are supported with carefully cross-matched
and monitored: engraftment is in general defined by
red cell transfusions and corticosteroids. During treatment
molecular full-donor chimerism and, for red blood cells, by
and support, immunohematologic monitoring includes
globular/plasmatic tests (including titrations) combined
direct/indirect antiglobulin tests, serological and genomic
with negative results of direct/indirect antiglobulin tests
blood group typing as well as isoagglutinin titration.
and, finally, with the ABO genotyping confirmation.
Luca Pierelli
Question 5
Sapienza University-San Camillo Forlanini Hospital
We use a prophylactic strategy for platelet transfusion. In Rome, Italy
stable patients, the prophylactic threshold is 10 9109/L in Email: luca.pierelli@uniroma1.it
both autologous and allogeneic transplantation. We trans-
Antonella Matteocci
fuse both single donor and pooled buffy-coat platelet con-
San Camillo Forlanini Hospital
centrates. Our platelet concentrates are produced in platelet
Rome, Italy
additive solution with a poor plasma content. In any case, Email: matteocci31@gmail.com
in most of our transfusion episodes, we release ABO identi-
cal/compatible platelet concentrates for all patients, includ- Luigi Rigacci
ing haemopoietic transplant recipients. We monitor platelet San Camillo Forlanini Hospital
transfusion efficacy with a complete blood count the morn- Rome, Italy
ing after transfusion. We do not perform routinely blood Email: lrigacci@scamilloforlanini.rm.it
counts 15–60 min after transfusion.
In recipients with serologically detected immune plate-
let refractoriness by ELISA (anti-HPA platelet antibodies) Karen M. K. De Vooght & Jurgen H. E. Kuball
and Luminex (anti-HLA platelet antibodies) tests, if the
patient is bleeding and the transfusion is urgent, we
assign fresh pooled buffy-coat platelet concentrates, and The Netherlands
otherwise, we perform platelet cross-match with SPRCA
technology on ABO, rarely HPA/HLA, compatible single Question 1
donor concentrates. For these patients, we monitor rou-
In our academic hospital, we perform about 60–80 autol-
tinely blood counts 60 min after transfusion.
ogous and 60–70 allogeneic HSCT’s in adults per year.
Question 6 Question 2
We routinely monitor isoagglutinin titres during ABO-in-
After an autologous HSCT, patients have an indication for
compatible haemopoietic transplantation. Isoagglutinin
irradiated blood products for a period of 1 year. Patients
titres are performed prior transplantation, before and after
undergoing an allogenic HSCT receive irradiated blood
plasma-exchange in case the recipient undergo isoagglu-
products 4 weeks before until 5 years after transplanta-
tinin removal. After transplantation, isoagglutinin titres
tion. In our country, institutes normally not irradiate
are performed every week until engraftment and then every
products themselves. This is performed by Sanquin, the
15 days, with ABO blood group checked every month.
only institute (foundation) in the Netherlands allowed to screening, and in case of bleeding, the patient will receive
and responsible for the blood supply. HLA-A and HLA-B typed (identical) single donor platelet
-X-ray is used as source of irradiation product(s) on a trial basis. In case of severe platelet refrac-
-at an energy dose of 25–50 Gy. toriness and absence of HLA class 1 antibodies, we screen
for HPA antibodies (following an HPA typed single donor
platelet product transfusion).
Question 3
We perform a complete blood count in autologous and
Question 6
allogeneic HSCT every 2nd day, until repopulation. In
case of pre-treatment with anti-thymocyte globulin Isoagglutinin titres are not determined routinely in patients
(ATG), a complete blood count is performed daily. after allogeneic HSCT. They do are determined before HCST
in case of major ABO blood group antagonism. Further-
more, we perform ABO typing (antigen and antibody) after
Question 4
HSCT on a regular basis, and anti-A and/or anti-B aggluti-
In stable patients, an haemoglobin threshold of 8 g/dl is nation strengths are sometimes used to evaluate aplastic
used for red blood cell transfusion. We use an one red anaemia due to major ABO blood group antagonism.
blood cell transfusion policy, in the outpatient ward occa-
sionally a two red blood cell unit transfusion policy.
Question 7
Patients receiving a non ABO identical transplant, receive
both patient and donor ABO compatible red cell units start- After HSCT, ABO blood group typing (antigen and anti-
ing from two weeks before HSCT. We do not change to body) and an irregular antibody screen is performed on a
donor ABO compatible units after engraftment. However, regular basis. At first every three days, thereafter with
we do change the patient’s ABO blood group to the donor each visit to the outpatient department, until engraftment
blood group, once we are sure engraftment is reached. For is reached. The direct antiglobulin test is not performed
that, the patient must be transfusion independent for three routinely, however, is done whenever there are discrepan-
months and his/her blood group must match the donor cies in the ABO blood group typing or when the irregular
ABO blood group, based on antigen typing, without dis- antibody screening is positive. Of course, the direct
crepancies. For treatment purposes, engraftment is deter- antiglobulin test is also performed in case of clinical signs
mined by white cell donor chimerism detection. of (auto-immune) haemolytic anaemia.
Question 5 Question 8
In our institution, a prophylactic platelet transfusion proto- We see about three cases of pure red cell aplasia in HSCT
col is used in case of an autologous HSCT. The prophylactic patients per year. In case of major ABO blood group, antago-
platelet transfusion trigger is >10 9 109/l both in patients nism anti-A and/or anti-B agglutination strengths can be
undergoing allogeneic as autologous HSCT. The standard used for monitoring. Patients receive red blood cell and plate-
platelet product for this purpose is a 5-donor random plate- let transfusions when needed. Further treatments of AA are in
let product containing PAS-E/plasma. In case of transfu- line with the Dutch guideline for AA (ATGAM, allo SCT).
sion refractoriness with proven HLA class 1 antibodies,
there is an indication for an HLA-A and HLA-B typed Karen M. K. De Vooght
(identical) single donor platelet product, again suspended University Medical Center Utrecht
in PAS-E/plasma. Because we use platelets suspended in Utrecht, the Netherlands
PAS-E/plasma, iso-agglutinins concentrations are lowered. Email: k.devooght@umcutrecht.nl
Therefore, we don’t apply any measure in case of a minor
Jurgen H. E. Kuball
ABO-incompatible platelet transfusion. We systematically
University Medical Center Utrecht
monitor the efficacy of platelet transfusions in patients Utrecht, the Netherlands
undergoing HSCT by a post-transfusion platelet count Email: J.H.E.Kuball@umcutrecht.nl
about an hour after finishing the transfusion. In case of
transfusion refractoriness, we first check for the presence
of HLA class 1 antibodies (by use of a Luminex technique).
Katherine L. Fielding, David A. Westerman & Erica M. Wood
In case they are present, the patient has an indication for
HLA-A and HLA-B typed (identical) single donor platelet
products. Waiting for the result of the HLA class 1 antibody
persistence of granulopoiesis and thrombopoiesis, our inci- transplants could be further broken down to include 54
dence of post-transplant red cell aplasia is approximately marrow transplants, 44 cord blood transplants and 27
1–3%. peripheral blood stem cell transplants.
Our general approach to treatment is supportive care
with transfusions and EPO and waiting for resolution.
Question 2
Question 8 Question 3
Transient hypoplasia post-major ABO-incompatible trans- A complete blood count occured daily during HSCT
plant more common (about 10%). Actual red cell aplasia admissions, with the exception of reduced-frequency test-
much less common. Managed with transfusion support ing scenarios (to every other day, in lower-volume tubes),
until the aplasia/hypoplasia recovers. in patients with stronger blood conservation mandates
(i.e. transfusions relatively or absolutely contraindicated).
Mickey B. C. Koh More recently, with the shift of autologous SCT to the
St George’s University Hospital outpatient setting, CBC frequencies have decreased. At
London, UK least one count check is prescribed in the first week post-
Email: mickey.koh@stgeorges.nhs.uk discharge, and patients are seen 3–10 times in the first
30 days after discharge with laboratory testing at those
Dara Qadir
times.
St George’s University Hospital
London, UK
Email: Dara.qadir@stgeorges.nhs.uk Question 4
The haemoglobin threshold for red blood cell transfusion
in stable, non-bleeding inpatients during the aplasia per-
Christine Cserti-Gazdewich iod is single unit red blood cell orders at ≤70–75 g/l.
We are able to change the red blood cell group of RBC
concentrates to the donor’s group when engraftment
Canada
occurs. This occurs when the forward group reflects
Question 1 exclusively donor erythropoiesis, without confounding by
transfusion, and by demonstration of the absence of anti-
The University Health Network performs HSCT at Princess graft isoagglutinin activity (on reverse type or by DAT).
Margaret Cancer Centre, one of its three main (acute care) However, we tend not to encode the group change for a
hospitals. Annually, 350 autologous and 190 allogeneic variety of reasons:
HSCTs are performed (540). (a) Grouping samples in the post-engraftment period,
after sufficiently lengthy transfusion-free periods,
Question 2 may not be drawn (due to the associated absence of
need for transfusion); and
Though we are in medical agreement with the principles (b) Experience with a number of ‘impermanence anec-
of Canada’s National Advisory Committee on Blood and dotes’ (i.e. disease relapses and/or evidence of
Blood Products (https://www.nacblood.ca/resources/guide delayed recurrences of host erythropoiesis), perhaps
lines/downloads/Recommendations_Irradiated_Blood_Com as a function of the increased proportion of reduced
ponents.pdf_), which specify 3–6 months after autologous intensity/non-myeloablative conditioning regimens,
HSCT and for as long as immune suppressant therapies with reversion to the hybrid type designation and
and/or (chronic) graft versus host disease may be present permissible component-specific ABO exposure truth
after autologous HSCT (with annual reviewer thereof). At tables.
our institution, the instruction to irradiate invariably
remains in place indefinitely for the following reasons:
(a) Disease relapses or de novo indications, which
themselves are indications for irradiation, have
Question 5
arisen after the NAC term conditions or intervals In autologous HSCT, the platelet transfusion protocol is
have elapsed, and permissive on the use of prophylactic platelet transfu-
(b) it has not been feasible to expeditiously obtain suf- sions, that is it is not restricted to treatment only. The
ficiently indubitable case details on the NAC term Platelet Transfusion Requirements in Hematopoietic
conditions or intervals, as patients may have health Transplantation (PATH) trial (https://clinicaltrials.gov/ct2/
show/NCT02650791), which is investigating outcomes by anticipated needs are reported a week in advance by the
either strategy, is active at our centre. Both approaches most responsible physician. An active list of sensitized
can be seen in audits by virtue of enrolments or clinician patients with platelet transfusion needs is coordinated in
equipoise (with the institutional policy permitting pro- the blood transfusion laboratory and amended with clini-
phylaxis at platelet counts of ≤10 for auto- or allo-HSCT). cian input and pertinent emerging laboratory results.
In times of (and in accordance with extent of) platelet For the latter, serologic profiles predicting for refrac-
shortage, stricter triggering is enforced. toriness come to the transfusion physician’s attention for
Our adult dose platelet concentrates are predominantly clinical review. Retesting the antibody repertoire is
buffy coat pools (four whole blood donor derived) with encouraged quarterly and/or when patients exhibit refrac-
single donor apheresis platelets. The ratio delivered to our toriness to matched products (should this be sooner [or
centre by Canadian Blood Services is presently 10:1, later/after missed rescreening opportunities]).
which has been rising from previous ratios (3:1) over the
last several years [1].
Question 6
Apheresis platelet stocks are enriched for those that
have been selected by HLA- (and/or HPA-)-type in the We do not routinely measure isoagglutinin titres routinely
care of sensitized and otherwise platelet transfusion in patients after allogeneic HSCT, though we provide this
refractory patients. service on request. Requests are usually limited to those
Our platelet concentrates are suspended in the plasma cases with suspected pure red cell aplasia/delayed red cell
of a contributing male donor in pooled concentrates. engraftment from ABO major mismatch.
In the case of minor ABO-incompatible platelet trans-
fusion, a room temperature (immediate spin) 1:50 titre is
Question 7
performed on group O platelets if this is the only avail-
able type left in inventory for non-O recipients. The pro- The transfusion laboratory does not prescribe or enforce a
duct is unmodified if titre negative. If titre positive, the prospective schedule of monitoring by direct antiglobulin
platelet undergoes a plasma volume reduction (PVR). tests or other measures to detect early immunohemato-
We advise monitoring of post-transfusion platelet logic complications after ABO-incompatible HSCT. How-
counts (15–60 min post) in ‘high-use alert cases’ (three ever, the transplant teams consult closely with the
consecutive days with ≥1 platelet transfusion per day), or transfusion laboratory medical directors to obtain inves-
in the evaluation of immune-selected single donor tigative advice and interpretation.
apheresis platelets (to assess the in vivo efficacy of prod-
ucts which had been selected to circumvent the last
Question 8
known array of HLA- and/or HPA-specific antibodies).
Our hospital transfusion laboratory receives its testing We do not have information on our incidence of pure red
and product support through Canadian Blood Services, cell aplasia. Management is on a case-by-case basis, from
with a local (Toronto-based) HLA Histocompatibility labo- antigen-negative chronic transfusion support to apheresis,
ratory, a national (Winnipeg-based) Platelet and Leuko- intensified immune suppressant therapy and off-label
cyte Reference laboratory, and the associated daratumumab in rare instances.
apheresis/donor management programmes.
Platelet transfusion refractoriness is managed by iden-
Reference
tifying suspicious cases (with the aforementioned [auto-
1 Cserti-Gazdewich C, Escorcia A, Pendergrast J, et al. Platelet
mated] high-use alert placed in the patient’s electronic transfusion reactions do not occur more often in recipients
medical record), and a clinical policy which supports transfused with apheresis versus buffy coat platelet concen-
immune refractoriness testing and the coordinated provi- trates. Transfusion 2016;56:3144–6.
sion of selected platelets. Suitable options for the latter
reflect a hierarchy of potentially available ‘matches,’ (i.e.
HLA-identical, HLA-‘similar’ [software-modelled/struc- Christine Cserti-Gazdewich
turally-predicted tolerance], HLA-‘negative’ [with respect University Health Network/University of Toronto
to the patient’s last-determined specificities] or ‘low MFI’ Toronto, Ontario, Canada
[HLA-mismatched with respect to the lowest mean fluo- Email: Christine.Cserti@uhn.ca
rescence intensity antibody specificities on PRA]).
Due to the preparation time entailed in sourcing and
delivering preferred platelets (especially for the lower fre- Etienne Daguindau, Fanny Angelot-Delettre & Pierre Tiberghien
quency profiles commanded by certain cases), the
plasma depletion or RBC depletion (in case of minor or provide this special care throughout the life of these
major ABO mismatch, respectively). patients.
To red blood cell concentrate (RBC), the source of irra-
diation is caesium-137 and the dose is of at least
Question 8
2500 cGy into the centre of the component.
The incidence of pure red cell aplasia (PRCA) after allo- To platelet components and fresh frozen plasma, since
geneic HSCT in our centre is around 5%. If PRCA persists March 2017, all of these components are treated to patho-
for more than 6 months, we consider the use of rituximab gen inactivation by INTERCEPT (from CERUS) and we
or more recently daratumumab. Otherwise, we wait for considered that the UV irradiation of this procedure is an
spontaneous improvement. effective prophylaxis for Transfusion-Associated Graft-vs-
Host Disease (TA-GVHD) too.
References
1 Giraud C, Thibert JB, Desbrosses Y, et al. Transfusion in Question 3
autologous and allogenic hematopoietic stem cell transplant:
Guidelines from the Francophone Society of Bone Marrow The routine is to perform a complete blood count every
Transplantation and Cellular Therapy (SFGM-TC). Bull Cancer day during HSCT until the platelet and neutrophil
2019;106:S52–s8. engraftments are established. No differentiation is made
2 Transfusion de plaquettes: produits, indications. Recommanda- in monitoring patients undergoing autologous or allo-
tions de la Haute Autorite de Sante et de l’AgenceNationale geneic HSCT.
de Securite du Medicament. 2015:(accessed 30/04/20).https://
The platelet engraftment is considered when the
www.has-sante.fr/upload/docs/application/pdf/2015-11/rec
patient’s count is at least 20 000/µl after three consecu-
ommandations_-_transfusion_de_plaquettes.pdf
tive days without platelet transfusion, and the neutrophil
engraftment is defined when the patient presents an abso-
Etienne Daguindau
lute neutrophil count of more than 500/µl across three
University Hospital of Besancon consecutive days too.
Besancon, France
Email: edaguindau@chuesancon.fr
Question 4
Fanny Angelot-Delettre Regarding the red blood cell transfusion protocol:
Etablissement Francais du Sang Bourgogne Franche-Comte - For stable and hospitalized patients, the red blood cell
Besancon, France transfusion has been considered to a haemoglobin
Email: fanny.delettre@efs.sante.fr threshold of 7.0–8.0 g/dl, unless the patient is symp-
tomatically anaemic or has other comorbidities.
Pierre Tiberghien
- We are adhering to a restrictive strategy and our pol-
Etablissement Francais du Sang
La Plaine St-Denis, France
icy is to transfuse one red blood cell unit and evaluate
Email: pierre.tiberghien@efs.sante.fr the clinical and laboratory response of the patient, as
long as he is stable, before indicating the second unit.
- The transfusion strategy in ABO-mismatched cases
must consider both the blood group systems of the
Silvano Wendel-Neto & Roberta Maria Fachini
recipient and the donor.
In case of major or bidirectional ABO-mismatched trans-
Brazil plants, transfusion of blood group O RBCs is necessary until
anti-donor isohemagglutinins (IHA) are undetectable in
Question 1 two consecutive blood samples of the recipients.
At our institution, if a patient is independent of RBC
The Sırio-Liban^es Hospital performs an average of 24 transfusion for 90 days and no incompatible isohemagglu-
autologous HSCT and 32 allogeneic HSCT per year. tinins against the new RBC phenotype are detected in two
consecutive blood samples, then the patient’s native blood
Question 2 type is switched to the donor type for future transfusion.
3 months post-transplant (6 months if total body irradia- trajectory of the count, frequency of visits and comor-
tion was used in conditioning). After allogeneic stem cell bidities of the patient.
transplant, irradiated components should be given from Donor group red cells are transfused following full
the time of initiation of conditioning chemoradiotherapy conversion to the donor group. This is established by the
and continued for the duration of graft-versus-host dis- (1) Disappearance of any dual population on forward and
ease (GvHD) prophylaxis, that is usually for 6 months reverse ABO grouping by indirect antiglobulin testing
post-transplant, or until lymphocytes are >1 9 109/l. If (2) Negative direct antiglobulin test (DAT) with
chronic GvHD is present or if continued immunosuppres- polyspecific anti-human globulin
sive treatment is required, irradiated blood components As an additional measure, at one of the hospital sites,
should be given indefinitely. donor group components are not transfused for the first
In practice, patients are identified as requiring irradi- year after transplant, and until full or near full chimerism
ated components and we do not proactively remove this is confirmed.
requirement. Few autologous recipients go on to have
significant transfusion requirements, and many allogeneic
Question 5
SCT recipients have other indications to need lifelong
irradiated components, for example anti-thymocyte glob- A prophylactic treatment strategy for both autologous
ulin, alemtuzumab or fludarabine given during condition- and allogeneic HSCT is used, with a threshold of
ing. 10 9 109/l in the absence of fever, bleeding or planned
Irradiation is undertaken by NHS Blood and Transplant procedures. Thresholds are in keeping with BSH guideli-
usually using a gamma irradiator or on occasion an x-ray nes for platelet transfusion [3]. Single donor apheresis
irradiator, if the component is irradiated at a site other platelets and pooled platelets are used interchangeably.
than our regional blood centre. These meet the specifica- Apheresis platelets are collected solely from male donors,
tion defined by the Joint United Kingdom (UK) Blood or female donors who have been tested and are negative
Transfusion and Tissue Transplantation Services Profes- for human leucocyte antigen (HLA) and human platelet
sional Advisory Committee, that is a minimum of 25 Gy, antigen (HPA) antibodies. Apheresis platelets are sus-
but with no part receiving more than 50 Gy [2]. pended in plasma. Pooled platelets are suspended in 60–
65% additive solution and 30–35% plasma (from a male
donor) [4].
Question 3
For minor ABO-incompatible platelet transfusions,
Platelet counts are performed daily for inpatients. Patients ‘high titre negative’ units are given (high titre is defined
remain in hospital until they demonstrate marrow as >1/128) [4]. Pooled platelets are given in this scenario
engraftment. Following engraftment, counts are done at where possible due to their lower plasma volume.
least weekly until the 100th day post transplant, or on Post transfusion increments are only undertaken if
every patient contact, if required more frequently for there is a concern regarding poor increments, or after
clinical reasons. transfusion of HLA-selected platelets in order to guide
Autologous transplants may be performed as an outpa- further donor selection. Platelet counts are routinely
tient but all patients are admitted at the onset of neu- taken within 24 h of most transfusions by virtue of the
tropenia until neutrophil recovery at which stage patient being in hospital during the hypoplastic phase.
neutrophile and platelet counts are measured until neu- When increments are checked, these are performed within
trophil engraftment documented according to EBMT crite- 15–60 min of the end of the transfusion.
ria. Following discharge from hospital, blood counts are Platelet refractoriness is managed by first demon-
done at every clinical review, many of which are per- strating true refractoriness, investigating and then pro-
formed at referring centres, the frequency of which is viding appropriate support. In the first instance,
guided by clinical circumstances. refractoriness is demonstrated with increments taken
within 24 h of a single unit of ABO identical platelet
transfusion on two occasions. If the rise in platelet
Question 4
count is <10 9 109/l on both occasions, the patient is
The haemoglobin threshold is 70 g/l for stable patients considered refractory. Causes of platelet refractoriness
during the period of aplasia, in the absence of significant (including immune and non-immune) are considered
cardiac disease. Single unit transfusions are used for and addressed where possible. Patients with no reversi-
inpatients unless they are scheduled for imminent dis- ble cause and clinical suspicion of immune refractori-
charge in which case two units may be transfused. For ness are tested for HLA antibodies. If no HLA
outpatients, two units may be given depending on the antibodies are detected, HPA antibodies are tested for.
Question 1
Question 8
At the Institute of Hematology and Transfusion Medicine
Pure red cell aplasia is only rarely observed after allo- (Institute), approximately 60–65 autologous and 65–70
geneic HSCT. Its incidence is not monitored routinely allogeneic transplants are performed per year.
in our institution. There are some patients with major
ABO incompatibility who have a significant red cell
Question 2
requirement for a prolonged period; we investigate in
order to exclude PRCA and parvovirus infection, and to At the Institute, every patient who requires multiple trans-
confirm trilineage engraftment, and then treat with fusions, every potential transplant recipient as well as
erythropoietin. Patients with evidence of immune transplant recipient is transfused with irradiated blood
haemolytic anaemia are treated with steroids and ritux- components until transfusions are required. In the case of
imab. HSCT patients, cellular blood components are irradiated for
at least 12 months after autologous donation and indefi-
nitely following allogeneic transplantations.
References Currently, caesium (Cs) is used as source of irradiation.
1 Treleaven J, Gennery A, Marsh J, et al. Guidelines on the use of The dose delivered to the blood component is no less
irradiated blood components. Br J Haematol 2010;152:35–51.
than 25 Gy at each site of the component, but does not
2 Joint United Kingdom (UK) Blood Transfusion and Tissue
exceed 40 Gy. Once in 3 years, the distribution of the
Transplantation Services Professional Advisory Committee:
Guidelines for the Blood Transfusion Services in the United
delivered dose is mapped. Mapping shows that the central
Kingdom, ed 8. London, TSO, 2013. dose is 32 Gy.
3 Estcourt L, Birchall J, Allard S, et al. Use of platelet transfu- For microbiological safety and prophylaxis against
sions. Br J Haematol 2016;176:365–94. graft-versus-host disease, most platelet concentrates are
4 NHSBT Portfolio of Blood Components and Guidance for pathogen reduced instead of being subjected to irradia-
their Clinical Use: SPECIFICATION SPN223/10. Available at: tion. If FFP is indicated, it is also administered after
https://nhsbtdbe.blob.core.windows.net/umbraco-assets-corp/ pathogen reduction or after quarantine.
16494/spn223v10.pdf Last accessed 10/05/2020.
4.5.2021 IPFA/PEI – The International Workshop on Surveillance and Screening of Blood-borne Pathogens
13–15.5.2021 The Canadian Society for Transfusion Medicine (CSTM) are holding their annual scientific conference virtually in 2021.
5–9.6.2021 ISBT In Focus, the 31st regional congress of the ISBT, will be a virtual event in 2021
17.9.2021 11th BIC International Conference – Advances in Haemostasis and Bleeding Disorders
613