Recovery of platelet‐rich red blood cells and acquisition of convalescent plasma with a novel gravity‐driven blood separation device

Abstract Objectives Our objectives were to determine the separation characteristics and blood product quality of a gravity‐driven microfiltration blood separation system (HemoClear, The Netherlands). Background A range of centrifugal blood separation devices, including intraoperative cell salvage devices (cell savers) and apheresis machines, are available to assist in preparing both allogenic and autologous blood products. These devices are expensive to operate and require extensive training. Methods and Materials Nine whole blood units were collected under standard conditions and analysed for haematological parameters, thromboelastographic properties, platelet morphology and activation, and red blood cell (RBC) deformability and morphology. Three whole blood units were separated by means of the HemoClear device, into a liquid and cellular component. The cellular component was diluted with SAGM and cold stored for 14 days. To simulate cell salvage six whole blood units were diluted with isotonic saline, followed by multiple HemoClear separation rounds. Results The recovery of both RBCs (100 ± 1.6%) and white blood cells (99 ± 4.5%) after undiluted filtration were very high, while platelet recovery was high (83 ± 3.0%). During the filtration, and cold storage after filtration storage both the non‐deformable RBC fraction and the RBC maximum elongation remained stable. Parameters of thromboelastography indicated that platelets remain functional after filtration and after 7 days of cold storage. In the cell salvage simulation the total protein load in the cellular fraction was reduced by 65 ± 4.1% after one washing round and 84 ± 1.9% after two consecutive washing rounds. Conclusion The novel blood filter studied effectively separates whole blood into diluted plasma and platelet‐rich RBCs. Moreover, the device effectively washed diluted whole blood, driving over 80% of proteins to the liquid component.


| INTRODUCTION
Patient blood management is a cornerstone of healthcare. 1 Each year over 118 million blood donations are collected (WHO, Blood safety and availability factsheet 2 ), but autologous blood transfusion, including cell salvage (i.e., re-infusion of patients' own lost blood) is used increasingly. 3 Both in processing of allogeneic and autologous blood, centrifugation-based technology is the gold standard.
Based on differences in density and buoyancy, centrifugation-based technologies effectively separate or wash blood components; whole blood (WB) donations are separated to facilitate optimal storage of, and clinical need for, the individual blood components. 4 Moreover, use of apheresis blood component isolation is growing rapidly, and mostly relies on centrifugal technology as well. 5,6 Autologous shed blood is washed by so-called 'cell savers' in order to reduce non-cellular contaminants. Although robust and reliable, Various studies have indicated that washing of red blood cells by centrifugation-based technology reduces the blood cell quality.
Haemolysis and sublethal injury have been shown to occur immediately after washing and to continue during the afterward storage 8,9 indicating that the centrifugation induced shear stress increases RBCs fragility. In addition, the erythrocytes' ability to change shape, referred to as deformability, seems to be negatively affected by centrifugal washing. 10 Apheresis was found to cause platelet activation, and erythrocyte complement disposition and antigen alterations. 11,12 Aside from blood quality limitations, blood processing centrifuges with the necessary disposables represent high capital expenditure and operational cost. [13][14][15] Hence in the emerging world, where costeffectiveness is imperative, the centrifugal devices are largely inaccessible. 16  their temperatures to remain at 20-24 C until the start of the filtration study protocol. 18 The day of blood collection was designated as day 0 of the study; the filtration study protocol was initiated at around 16 h after collection.

| Separation of WB by means of a gravity-driven microfilter
Separation of WB was performed with the HemoClear device (HemoClear BV) ( Figure 1A). According to the device manufacturer, the device yields two filtration products; concentrated blood cells (referred to as cellular component) and the plasma (referred to as liquid component) ( Figure 1B). The filter was used in a two-blood bag system that was primed with isotonic saline prior to use. Due to cross-flow technology the RBCs can enter the filtration device from either of the two inlet ports ( Figure 1C). This feature allows for the filtration system to remain closed during consecutive filter rounds. Being disposable, a HemoClear system can be used to salvage multiple units of autologous blood provided that these were shed by the same patient. For the next patient a new HemoClear system should be used. Similarly, in this protocol a new HemoClear system was used for each new unit of (diluted) WB.
The device performance was evaluated in two protocols; a WB separation protocol and a washing protocol. The separation protocol ( Figure 1D) was performed with undiluted WB and was intended to separate the cellular component from the liquid component. The washing procedure was performed with diluted WB to mimic shed blood ( Figure 1E). This procedure was intended to remove unwanted solutes from the cellular component.

| Separation of WB
Three WB units were filtered through a HemoClear device by gravity (i.e., for each unit a new HemoClear device was used). Upon completion of the filter run, 100 ml of SAGM (Feresenius Kabi) was added to the cellular component by means of a backflush through the filter via an inlet in the liquid product line ( Figure 1D). Both the liquid and cellular components were analysed for composition and quality immediately after the separation procedure. The cellular component was stored for 7 days at 2-6 C and analysed again.

| Washing of WB
Six half units of WB (around 300 ml) were diluted with a calculated volume of 0.9% NaCl (around 300 ml) to achieve 600 ml of diluted WB with an haematocrit (Ht) of 20%.
The six diluted WB units were subjected to separation by six HemoClear devices, driven by gravity. Upon completion, 300 ml of 0.9% NaCl was added to the cellular component and a second filtration round was performed. Consecutively, the second cellular component was subjected (i.e., without fluid addition) to a third filtration round to concentrate the cellular compound.

| Volume
The volume of blood components was calculated from the net weight and the specific gravity: 1.026 g/ml for plasma, 1.100 for RBCs, and

| Platelet morphology
Platelet morphology was assessed by Kunicki morphology scoring. 21 The PC sample was fixed with 0.5% glutaraldehyde. Samples were analysed by phase contrast microscopy. The number of discs identified in a 100 cell count of fixed platelets under the microscope is multiplied by 4, spheres by 2, platelets with dendrites by 1, and balloons by 0, resulting in a maximal score of 400 for perfect discoid platelets.
PLT activation was detected using a flow cytometer (LSRII-HTS, BD Biosciences, Breda, the Netherlands)) after staining of PLTs with fluorescent CD62P-FITC (P-selectin, Beckman Coulter, Immunotech) as described before. 22 2.3.7 | Total protein Supernatant total protein was measured using the biuret method on Architect clinical chemistry analyser (Abbot, Abbot Park, IL, United States).

| Deformability
Referred to as the deformation index (DI), red blood cell deformability was defined as the ratio of the major axis length to the minor axis width. That is, the DI of a disc shaped red cell by definition equals

| Statistics
Results are expressed as mean values ± SD. Paired two-sided t-tests were performed to compare the WB measurements to the data acquired on the cellular and liquid components. Significance was defined as p < 0.05.
In line with RBC recovery, no significant difference was found between the haemolysis prior to (0.00 ± 0.01%), and after separation Platelet function was assessed using TEG as shown in Table 1. The R value (6.4 ± 0.5), that is, time until the first evidence of a clot, was reduced after filtration as compared to pre-processed blood (  3.2 | Washing of diluted WB

| Characteristics of cellular component
The washing procedure yielded a RBC recovery of at least 87 ± 6% (Hb recovery of at least 88 ± 6%), WBC recovery of at least 93 ± 7% and platelet recovery of at least 68 ± 10%.
Washing by one filtration round, two consecutive rounds or the entire washing protocol did not significantly increase haemolysis ( Figure 2B).

| Washing efficiency
Free total protein load was studied as an indication for the removal of extracellular solutes from the cellular component and washing effectivity ( Figure 5A).

| RBC deformability and morphology
The non-deformable RBC fraction as measured with ARCA, remained stable during the washing procedure ( Figure 5D). Similarly, the RBC

| Platelet function
Platelet function in the cellular component was assessed using TEG and are shown in Table 1

| DISCUSSION
The processing of WB units showed that the HemoClear device recovers high percentages of red blood cells, white blood cells and platelets. Cellular recoveries were higher in the separation protocol compared to the washing protocol. In the washing protocol, cellular recoveries in between washing rounds were measured while the filter and filtration lines still contained a fraction of the blood volume. This fraction of the blood cells was not included in the cellular count as determined in the blood bag. One way to correct for this seemingly lost volume is to measure the recoveries from the first washing round to the second round. This corrected calculation yielded RBC, WBC and platelet recoveries of 97.6 ± 9.6%, 93.0 ± 3.3% and 81.8 ± 12.9% respectively. (Figure 2A; Table S1). Red blood cell deformability and morphology, and platelet morphology and activation were not negatively affected in the neither separation nor washing protocol. Also the levels of free haemoglobin and potassium suggested minimal sublethal injury and haemolysis.
The unique cross-flow microfiltration technology on which the studied device's mechanism is based, allows for both highly specific separation, washing and concentration of blood cells. The HemoClear could support centrifugal devices in separation of WB, and cell salvage.
Moreover this device could be used to produce platelet-rich RBCs.

| Production of washed platelet-rich red blood cells
Intraoperative cell salvage has been shown to increase platelet transfusion requirements. 27,28 Cell salvage is performed to recuperate red blood cells from shed blood. The autotransfusion device RBC washing procedures not only remove unwanted components, such as proinflammatory substances, but also eliminate platelets, coagulation factors and plasma proteins.
Second generation cell salvage devices are enhanced with platelet sequestration features that enable WB fractionation into RBCs and platelet-rich plasma. However, evaluation of three autotransfusion devices showed that merely 50%-60% of platelets is recovered in the PRP with this enhanced function. 29

| Harvest of (convalescent) plasma
The total protein, free haemoglobin and potassium loads indicated that per HemoClear washing round about 65% of noncellular components is driven to the liquid component. Two consecutive washing rounds yielding 80%-90% of noncellular substances in the liquid component. Based on this finding we hypothesised usability of this device in the harvest of convalescent plasma. While finalising this manuscript, use of the device for the acquisition of anti-COVID-19 convalescent plasma was already studied by a group in Suriname. 33 Bihariesingh-Sanchit and colleagues applied a two round washing protocol, very similar to the washing protocol studied here, to isolate diluted convalescent plasma from recovered COVID19 patients in hospital.
Centrifugal apheresis is the main technology utilised in the collection of anti-COVID convalescent plasma. 34 Nevertheless, use of this technology in the emerging world is hindered by several barriers including lack of funds, no availability of apheresis kits and absence of technical expertise. 35 Possibilities to use the device in production of convalescent plasma remain under study.

CONFLICT OF INTEREST
Arno Nierich is the inventor of the HemoClear device, holds patent right to the device's technology and owns shares in HemoClear BV. Dion Osemwengie is employed by HemoClear BV.