1 Introduction

Extracorporeal membrane oxygenation (ECMO) is a life-saving intervention for patients experiencing lung and heart failure [1, 2]. The blood pump, which substitutes the function of the heart and supports blood circulation for patients, is a key component of the ECMO [3, 4]. When the blood pump is activated, the rotor inside the blood pump rotates at high speed. Although this facilitates blood circulation, it also generates non-physiological shear stress (NPSS) owing to the rapid rotation [5]. The NPSS typically exceeds 100 Pa, surpassing the physiological range and can cause blood damage [6, 7]. This NPSS-induced blood damage can lead to clinical complications, such as hemorrhage, thrombus, and hemolysis [8,9,10]. These clinical complications are linked to red blood cell (RBC) damage directly or indirectly.

RBCs are important blood cells, constituting approximately 50% of the total blood volume [11], and play an important role in oxygen transport owing to their hemoglobin content [12,13,14]. However, under the influence of NPSS, the structure and function of RBCs can be damaged. When the NPSS surpasses the hemolytic threshold, RBCs sustain a damage, leading to the release of free hemoglobin (fHb). [15] explored the correlation between RBC mechanical sensitivity and the hemolytic threshold and found that impaired RBC deformability lowered the hemolytic threshold [16]. [17] utilized ektacytometry to analyze the mechanical sensitivity (MS) index and observed a significant impairment in the MS index of RBCs after exposure to 10 Pa above the sub-hemolytic threshold. [18] employed a Poiseuille shearing system and subjected RBCs to 125 Pa for 1.5 s in three duty cycles and found a decrease in the electrochemical charge of the RBC membrane along with increased RBC aggregation. [19] proposed that shear stress primarily induces hemolysis by damaging the RBC membrane, with accumulated shear stress leading to sudden RBC rupture during blood circulation [17, 18, 20]. Most previous studies have focused on NPSS-induced RBC damage, particularly hemolysis. However, hemolysis represents only one aspect of NPSS-induced blood damage. In clinical conditions, RBCs damaged by blood pumps release numerous biochemical factors, including hemoglobin, which can affect coagulation and the immune system. [21] reported that NPSS-exposed RBCs generate shear stress-induced RBC-derived microparticles (MPs). This contributes to increased coagulation and inflammation. Additionally, structural alterations in RBCs can result in the formation of RBC-derived extracellular vesicles (REVs). [22] investigated the interaction between REVs and monocytes and discovered that they stimulate endothelial cell activation, leading to increased von Willebrand factor (vWF) levels [23]. Plasma fHb and REVs induce thrombosis by scavenging plasma nitric oxide (NO) in vascular endothelial cells.

RBCs are the predominant cells in the blood; therefore, NPSS-induced RBC damage and subsequent interactions among blood components can trigger dysfunction in the coagulation and immune systems. This potentially exacerbates clinical complications, such as hemolysis and thrombosis. Therefore, understanding the various effects of NPSS on RBCs can help to determine the mechanisms underlying related complications. However, systematic research into device-induced RBC damage beyond hemolysis, including the correlation between hemolysis, biochemical parameters, and REVs, is lacking.

Therefore, this study aimed to comprehensively explore the effect of NPSS on RBCs and associated biochemical indicators under two distinct clinical support conditions provided by blood pumps. Using a commercial blood pump, different NPSS levels were generated. Computational fluid dynamics (CFD) methods were used to analyze the NPSS created via blood pumps under these two clinical conditions. Subsequently, morphological parameters, biochemical changes, plasma extracellular vesicles (EVs), and oxidative damage in RBCs were systematically compared and analyzed in these two loops.

2 Materials and methods

2.1 Studied pump

The blood pump used in this study was a Rotaflow pump (Maquet Inc., Germany), a centrifugal blood pump consisting of a volute and rotor. Part dimensions were measured using a ruler (Fig. 1a). The axial balance of the rotor was maintained through shaft support (Fig. 1b). In addition, a secondary flow (indicated by the blue arrows) was incorporated in the pump design to prevent blood stasis.

Fig. 1
figure 1

Model and methodology overview (a) Three-dimensional model representation of the Rotaflow blood pump. (b) Longitudinal section of blood pump and internal blood flow path. (c) Characteristics of the mesh used for simulation. (d) Experimental setup of the circulation loop utilized in the study

Full size image

2.2 Mesh generation

A hybrid mesh was generated using ANSYS mesh (ANSYS Inc., Canonsburg, PA, USA, Fig. 1c), comprising polyhedral cells. Near-wall regions were refined with six layers of prismatic mesh to maintain y + below 1. Mesh-independent verification was conducted with three mesh resolutions (3, 9, and 12 million cells). The mesh number had a negligible effect on key variables such as pressure head and mean shear stress. Approximately 12 million meshes were generated in all simulations.

2.3 Numerical method

The NPSS generated via the blood pump under two clinical conditions (pressure heads of 100 and 350 mmHg and a flow rate of 5 LPM) was analyzed via simulation method. Blood, treated as an incompressible Newtonian fluid in the high-speed blood pump [24, 25], had a density of 1055 kg/m3 and viscosity of 0.0035 Pa·s [26]. The Reynolds-averaged Navier equation was solved using ANSYS CFX software (ANSYS Inc.). The convection term was solved at high resolution, and the SST k-ω turbulence model [27, 28] was applied for steady-flow simulation. A frozen rotor interface linked the rotating area–the impeller, gaps, and secondary flow and stagnant area (inlet duct and volute). At the inlet, a boundary condition of a volume flow rate of 5 LPM was set. At the outlet, a static pressure boundary condition of 0 mmHg was applied. The required head pressures (100 and 350 mmHg) were obtained by adjusting the pump rotation speed. All the solid walls were assumed to be non-slip and adiabatic, with a convergence criterion set to 10–6.

The rotating speeds needed for the blood pump to achieve pressure heads of 100 and 350 mmHg, with a flow rate of 5 LPM, were determined through simulations and experiments at 2300 and 3700 rpm, respectively. The experimental and simulated rotational speeds were consistent, demonstrating the reliability and accuracy of the simulation method results.

2.4 NPSS

Viscous scalar shear stress (SSS) was derived from the simulated flow fields using the following formula [29,30,31]:

$$\tau ={\left[\frac{1}{6}\sum {\left({\tau }_{ii}-{\tau }_{jj}\right)}^{2}+\sum {\tau }_{ij}^{2}\right]}^{{~}^{1}\!\left/ \!{~}_{2}\right.}$$
(1)

where \(\tau\) is the shear tensor, which is calculated by multiplying the shear rate tensor with the blood viscosity; \({\tau }_{ij}\) is the shear stress tensor, \({\tau }_{ij}=\mu (\partial u\_i/\partial x\_j+\partial u\_j/\partial x\_i)\), \({\tau }_{ii}=-p+2\mu \partial u\_i/\partial x\_j-2/3\mu \nabla \bullet u\), \(\partial u\_i/\partial x\) _j is the derivative of the i-axis component of velocity with respect to the i-axis in the Cartesian coordinate system x; and ∑ represents the summation of the tensor analysis.

2.5 Experimental platform

The experimental platform was used to investigate RBC damage induced by NPSS under two clinical conditions (pressure heads of 100 and 350 mmHg and a flow rate of 5 LPM, Fig. 1d) [32]. The experimental platform included a reservoir, pump, flow sensor, pressure sensor, throttle valve, and tubes with connectors. All blood-contacting surfaces on the platform were constructed using biocompatible materials. The flow rate was measured using an inline ultrasonic flow sensor (Transonic, Inc., Ithaca, NY, United States), whereas the pressure head was monitored via two disposable pressure sensors (UTMD Inc., Utah, WM, United States) positioned near the inlet and outlet of the pump under investigation.

2.6 Blood sample collection

Ethical standards were reviewed and approved by the Ethics Committee of the School of Biological Science and Medical Engineering at Beihang University. Bovine blood, treated with heparin (40 IU/ml) for anticoagulation, was obtained from a slaughterhouse. The blood was filtered using medical gauze to eliminate clots and impurities, with the hematocrit (HCT) adjusted to 35% using 0.5% bovine serum albumin (BSA) [33, 34]. The two circulation loops were rinsed with phosphate-buffered saline (PBS). Subsequently, 900 mL of blood was gently loaded into each circulation loop, with all bubbles removed from the circuits. After loading blood into both circulation loops, 5-mL blood samples were collected from each circulation loop as base samples. Subsequently, 5-mL blood samples were obtained from each loop per hour during the 600-min circulation experiments. Eleven blood samples were collected from each loop in total, and the experiment was repeated five times.

To investigate the blood pump-induced generation of EVs, two additional circulation loops containing purified RBCs were established. Bovine blood was centrifuged in 50 mL tubes at 800 × g for 30 min at 20 °C, and this process was repeated five times to remove white blood cells and platelets. The HCT of the resulting purified RBC suspension was adjusted to 35% using 0.5% BSA. Subsequently, two experimental circulation loops were filled with 900 mL of pure RBC suspension. At 0 and 600 min, 50-mL blood samples were extracted from both circulation loops for a transmission electron microscope (TEM) assay.

2.7 Hematological analysis

Hematological analysis was conducted by collecting 200 μL of whole blood. RBC counts and morphological data, such as mean corpuscular volume (MCV) and RBC distribution width (RDW), were assessed using a hematology analyzer (BH-40vet, URIT, China).

2.8 Detection of free hemoglobin

Whole blood (1 mL) from collected blood samples was centrifuged at 2500 × g for 15 min at 25 ℃ to obtain plasma. The plasma was then diluted approximately 50-fold with 0.01 M PBS. Subsequently, 15 μL of the diluted plasma was mixed with 250 μL chromogenic agent (Nanjing Jiancheng Bioengineering Institute, China), whereas 15 μL of 0.01 M PBS was mixed with 250 μL chromogenic agent as a blank control. All samples were then incubated at 37 ℃ for 20 min. The fHb content was then determined via colorimetry at 510 nm using an automatic enzyme immunoassay analyzer (SpectraMax iD3; Molecular Devices, LLC, USA).

2.9 Detection of biochemical change

Whole blood (1 mL) from collected blood samples was centrifuged at 2500 × g for 15 min at 25 ℃ to obtain plasma for detection. The levels of lactic dehydrogenase (LDH), glucose, triglyceride (TG), cholesterol (CHO), and iron were measured using a colorimetric method with an automatic biochemical analyzer (AU2700, Japan). To ensure the accuracy of the detection results, the detection system was calibrated using standards (20,000,439, Leadman, China) matched with the system. Quality control serum (HN1530, Randox, UK) was used to confirm system control, with a variable coefficient of 0.3–0.5%.

2.10 Scanning electron microscope (SEM) detection

Whole blood samples of 100 μL were collected at 0 and 600 min from the two experimental circulation loops. First, each sample was fixed with 2.5% glutaraldehyde and stored at 4 ℃ for 24 h. Subsequently, the samples were rinsed thrice with PBS (0.1 M), fixed with 1% osmium solution for 1 h, and rinsed thrice with ddH2O. Next, the samples were centrifuged at 1500 g for 10 min to separate the plasma from blood cells, the supernatant was discarded, and cell pellets were suspended in a buffer solution. Thereafter, the cell pellets were dehydrated using 30%, 50%, 70%, and 90% alcohol for 3 min each, followed by dehydration with 100% alcohol for 5 min, and the process was repeated thrice. Next, the alcohol-containing blood cell liquid was dropped onto a cover glass, dried at a critical point, gilded using an ion sputtering instrument, and observed under a SEM (JSM-7900F, JEOL Ltd., Japan).

2.11 TEM detection of plasma vesicles

Whole blood samples (50 mL) and purified RBC samples were obtained from the two loops at 0 and 600 min. Initially, the collected samples were centrifuged at 600 × g for 15 min and 2500 × g for 15 min at 4 ℃ to isolate the supernatants. Subsequently, the supernatant was further obtained by centrifugation at 10,000 × g for 45 min at 4 °C using an ultracentrifuge. Finally, the sediment was dissolved in 200 μL of 0.01 M PBS and observed with TEM (JEM-1400, JEOL Ltd.) after negative staining.

2.12 Detection of malondialdehyde (MDA), total antioxidant capacity (T-AOC), methemoglobin (metHb), and NO

For MDA and T-AOC content detection, 2 μL of RBCs were diluted with 200 μL of ddH2O. After mixing with the detection reagent (Nanjing Jiancheng Bioengineering Institute), the samples were analyzed at 530 nm and 593 nm using an automatic enzyme immunoassay analyzer (SpectraMax iD3; Molecular Devices).

For metHb detection, 5 μL of whole blood was collected. After mixing with the detection reagent (Nanjing Jiancheng Bioengineering Institute), the samples were analyzed at 630 nm using an automatic enzyme immunoassay analyzer (SpectraMax iD3; Molecular Devices).

For NO detection, 1 μL of whole blood was centrifuged at 2500 × g for 15 min at 25 ℃. Subsequently, 300 μL of supernatant plasma was mixed with 400 μL of the prepared solution (Nanjing Jiancheng Bioengineering Institute), and the detection was performed at 550 nm with the automatic enzyme immunoassay analyzer (SpectraMax iD3, Molecular Devices), followed by the calculation of NO content.

2.13 Statistical analysis

The experimental data were expressed as mean ± SE (standard error of the mean). Data were compared between the two groups hourly from 0 to 600 min. A two-way repeated-measures ANOVA with Sidak correction for multiple comparisons and a confidence interval of 95% was conducted to assess differences at a single time point between the groups. Pearson correlation coefficients were calculated for correlation studies using SPSS 17.0 statistical analysis software, and p < 0.05 was considered statistically significant.

3 Results

3.1 Shear stress distribution in studied pumps for the pressure head of 100 and 350 mmHg

Figure 2a and b illustrates the distributions of SSS at 50% blade height in the rotor and volute areas of the blood pump under the two working conditions. Notably, SSS levels at 350 mmHg and 5 LPM exceeded those at 100 mmHg and 5 LPM (mean values: 5.12 Pa at 100 mmHg and 8.47 Pa at 350 mmHg). Figure 2c and d displays the wall shear stress (WSS) on the rotor surface. The WSS at 350 mmHg significantly exceeded that at 100 mmHg (mean values: 47.41 Pa at 100 mmHg and 92.59 Pa at 350 mmHg). Of note, although the maximum shear stress can cause hemolysis, the volume proportion of maximum shear stress in the blood pump was very low. The maximum shear stress had a limited effect on hemolysis which have not been considered in this study [6].

Fig. 2
figure 2

CFD results (a) Mean shear stress under a flow rate of 5 LPM in the centrifugal pump was 5.12 Pa under the 100-mmHg condition. (b) Mean shear stress under a flow rate of 5 LPM in the centrifugal pump was 8.47 Pa under the 350-mmHg condition. (c) Mean WSS under a flow rate of 5 LPM in the centrifugal pump was 47.41 Pa under the 100-mmHg condition. (d) Mean WSS under a flow rate of 5 LPM in the centrifugal pump was 92.59 Pa under the 350-mmHg condition. The mean values of SSS were calculated over the entire pump domain and over the rotor wall

Full size image

3.2 RBC hematology and morphology analysis

Changes in RBC morphology was examined using a hematology analyzer and SEM. The hematology analyzer was employed to determine RBC count, MCV, and RDW from samples collected at 0 to 600 min under both 100 and 350 mmHg conditions. According to the hematological analysis (Fig. 3a–c), no significant changes were observed in RBC count, MCV, and RDW between 0 and 600 min under the two conditions (100 vs. 350 mmHg). SEM results similarly revealed no significant differences in RBC morphology over time (0 to 600 min) or between different conditions (100 vs. 350 mmHg, Fig. 3d–f).

Fig. 3
figure 3

Hematology and morphology analysis (a) RBC count; (b) MCV; (c) RDW; and (d), (e), and (f) morphological changes observed via SEM at a scale of 10 μm for the base group, 600 min for the 100-mmHg group, and 600 min for the 350-mmHg group, respectively. RBC, red blood cell; MCV, mean corpuscular volume; RDW, red cell distribution width; SEM, scanning electron microscopy

Full size image

3.3 Hemolysis evaluation

In this study, fHb levels were assessed via hemolysis analysis of the two loops, revealing an increase over time in both. Moreover, fHb concentrations were notably higher in the loop subjected to 350 mmHg than in that subjected to 100 mmHg (Fig. 4a and b). Significant increases in fHb were observed from 120 min (p < 0.05), intensifying at 420 min (p < 0.001) in the 100-mmHg loop, and from 60 min (p < 0.05), intensifying at 420 min (p < 0.001) in the 350-mmHg loop (Fig. 4b). Differences between these conditions became significant from 60 min onwards (p < 0.05, Fig. 4a). These findings suggest that the heightened hemolysis levels were associated with pressure head and circulation time.

Fig. 4
figure 4

Hemolysis evaluation (a) Visualization of plasma of samples collected from the two loops over a 600-min circuit duration. (b) Changes in fHb levels in both loops. (c) Changes in iron levels in both loops. (d) Changes in LDH in both loops. fHb, free hemoglobin; LDH, lactic dehydrogenase. Significance: *p < 0.05, **p < 0.01, ***p < 0.001, #p < 0.05

Full size image

To further evaluate the capacity of RBCs to transport oxygen and the risk of rupture under the two conditions, plasma iron and LDH were measured. In the 350-mmHg condition, plasma iron levels increased significantly from 540 min (p < 0.05), whereas no significant change was observed in the 100-mmHg condition throughout the test period (Fig. 4c). As the main function of iron in hemoglobin is to bind and transport oxygen, the high NPSS generated by the high-pressure head may be the primary cause of iron precipitation and the reduced oxygen transport capacity of RBCs. LDH, primarily present in RBCs, showed no significant changes during circulation in the two loops (Fig. 4d).

3.4 Extracellular vesicle changes assessed using SEM and TEM

SEM was used to detect EVs from RBCs in three groups (base, 100-mmHg, and 350-mmHg groups at 600 min) in whole blood from the examined circulation loops (Fig. 5a). SEM measurements revealed that EVs budded and outflowed from the RBCs, with a higher abundance observed at 600 min than at 0 min in both loops.

Fig. 5
figure 5

Observation of EVs using SEM and TEM (a) Representative SEM images of RBCs in the three groups (base, 100-mmHg, and 350-mmHg groups at 600 min). MPs and exosomes are indicated by red and yellow arrows, respectively. (b) Representative TEM images of plasma EVs from whole blood in the three groups. MPs and exosomes are indicated with red and yellow arrows, respectively. (c) Representative TEM images of EVs from purified RBCs in the three groups. MPs and exosomes are indicated with red and yellow arrows, respectively

Full size image

TEM analysis was conducted to identify plasma MPs and exosomes in whole blood samples from the 100- and 350-mmHg groups at 0 and 600 min. Compared with that in the base group at 0 min, a greater presence of exosomes and MPs was observed in the 100- and 350-mmHg groups at 600 min (Fig. 5b).

Purified RBC suspensions were used to further explore whether plasma EVs originated from RBCs. Samples were obtained from the base, 100-mmHg, and 350-mmHg groups at 600 min. TEM analysis revealed an increase in plasma MPs at 600 min in both the 100- and 350-mmHg groups compared with that in the base group, whereas the number of exosomes decreased at 600 min in both groups compared with that in the base group (Fig. 5c).

3.5 Plasma MPs and exosomes evaluation

TEM results of the purified RBC circulation demonstrated a significant increase in MPs at 600 min, surpassing the count observed in the whole blood circulation loop (Fig. 6a). Conversely, the number of exosomes in the purified RBC circulation loop decreased significantly at 600 min, and was lower than that in the whole blood circulation loop (Fig. 6b). These findings suggest that shear-induced RBCs during blood pump-assisted blood circulation can lead to a substantial increase in MPs, whereas exosomes may primarily originate from damage to other cell types, such as white blood cells and platelets.

Fig. 6
figure 6

Plasma MPs and Exosomes evaluation (a) Comparison of MP numbers in the whole blood and purified RBC samples in the three groups. The number of MPs was counted in five images for the three groups. (b) Comparison of exosome numbers in the whole blood and purified RBC samples in the three groups. The number of exosomes was counted in five images for the three groups. MPs, microparticles; TEM, transmission electron microscope. Significance: *p < 0.05, **p < 0.01, ***p < 0.001, #p < 0.05

Full size image

3.6 Oxidative stress evaluation

To assess whether oxidative stress damages the membrane and hemoglobin of RBCs and whether it leads to variations in the antioxidant capacity of RBCs, MDA, metHb, and T-AOC levels were analyzed throughout the cycle. The balance between pro-oxidants and antioxidants determines the oxidative status of cells. MDA and metHb represent membrane lipid and hemoglobin oxidative damage, respectively, whereas T-AOC indicates the resistance level of RBCs to oxidative damage. No significant changes were observed in MDA, metHb, or T-AOC (Fig. 7a–c), suggesting that oxidative stress may not play a significant role in blood pump-induced RBC damage.

Fig. 7
figure 7

Evaluation of oxidative stress (a) Changes in MDA levels during circulation time. (b) Changes in metHb levels during circulation time. (c) Changes in T-AOC during circulation time. MDA, malondialdehyde; metHb, methemoglobin; T-AOC, total antioxidant capacity

Full size image

3.7 Analysis of TG, CHO, glucose, and NO levels

To comprehensively assess the damage to the RBCs, TG, CHO, glucose, and NO levels were measured from 0 to 600 min under the two conditions. TG and CHO are primary components of the RBC membrane. TG levels significantly increased from 120 min (p < 0.05), intensifying at 420 min (p < 0.001) in the 350-mmHg circulation loop (Fig. 8a), whereas they remained relatively stable in the 100-mmHg circulation loop. A statistically significant difference between the two loops was observed at 180 min (p < 0.05, Fig. 8a). Plasma CHO increased from 300 min in the 350-mmHg circulation loop (p < 0.05, Fig. 8b), whereas no significant change was observed in the 100-mmHg circulation loop (Fig. 8b). However, no statistically significant difference was observed in plasma CHO levels between the two circulation loops (Fig. 8b). These findings suggest that membranous structures can be released into the plasma owing to cellular damage.

Fig. 8
figure 8

Analysis of TG, CHO, glucose, and NO (a) TG, (b) CHO, (c) glucose levels, and (d) NO levels changed in the 100- and 350-mmHg circulation loops and compared between the two circulation loops. TG, triglyceride; CHO, cholesterol; NO, nitric oxide. Significance: *p < 0.05, **p < 0.01, ***p < 0.001, #p < 0.05

Full size image

Further studies revealed that plasma glucose levels declined from 180 min (p < 0.05), intensifying at 480 min (p < 0.001) in the 100-mmHg circulation loop (Fig. 8c). In the 350-mmHg circulation loop, the decreasing trend in glucose became significant (p < 0.05) from 120 min and intensified at 240 min (p < 0.001, Fig. 8c). A significant difference between the two circulation loops was observed at 600 min (p < 0.05, Fig. 8c). Owing to the absence of mitochondria, RBCs consume glucose via anaerobic glycolysis during metabolism. A persistent decline in plasma glucose levels suggests an increase in RBC metabolism. Plasma NO levels decreased from 420 min (p < 0.05) at 100 mmHg to 360 min (p < 0.05) at 350 mmHg (Fig. 8d). Statistically significant differences in plasma NO levels between these two conditions were observed at 240, 360, and 420 min (Fig. 8d), suggesting significant consumption of plasma NO during blood pump operation.

3.8 Correlation analysis of fHb, RBC parameters, and biochemistry markers

Pearson’s correlation coefficients were employed for correlation studies involving 13 biomarkers. In the 100-mmHg group (Fig. 9a), MCV, RDW, iron, LDH, T-AOC, TG, and CHO exhibited the strongest positive correlations with fHb. Conversely, metHb, glucose, and NO demonstrated the most negative correlations with fHb. RDW, indicative of RBC volume and size heterogeneity, exhibited the strongest positive correlation with MCV, iron, LDH, and TG (triglyceride) while displaying a negative correlation with glucose. Furthermore, plasma iron and LDH, reflecting RBC rupture, displayed the strongest positive correlations with T-AOC, TG, and CHO while exhibiting the most negative correlations with glucose and NO. In the 350-mmHg group (Fig. 9b), RDW, iron, LDH, TG, and CHO exhibited the strongest positive correlations with fHb. Conversely, RBC count, glucose, and NO displayed the most negative correlations with fHb. RDW demonstrated the strongest positive correlations with fHb, MCV, iron, LDH, TG, and CHO and the most negative correlations with glucose. Furthermore, plasma iron and LDH displayed the strongest positive correlations with TG and CHO and the most negative correlations with glucose. Changes in the levels of the tested biomarkers were interrelated, suggesting that RBC damage leads to changes in plasma and cellular composition. Notably, the correlations between variables may vary under different pressure heads.

Fig. 9
figure 9

Correlation analysis of fHb, RBC parameters, and biochemistry markers (a) and (b) Correlations among biochemistry markers examined in the 100- and 350-mmHg circulation loops, respectively. All values are Pearson’s correlation coefficients. RBC, red blood cell; MCV, mean corpuscular volume; RDW, red cell distribution width; LDH, lactic dehydrogenase; TG, triglyceride; CHO, cholesterol; NO, nitric oxide. Significance: *p < 0.05, **p < 0.01, ***p < 0.001

Full size image

4 Discussion

Device-induced blood damage is inevitable in patients supported with ECMO, often resulting in severe complications such as thrombosis and bleeding. Among blood constituents, RBCs are the most abundant cells, making device-induced RBC damage a frequent occurrence in patients receiving blood pump support. Previous studies on blood pump-induced RBC damage primarily focused on assessing fHb concentration changes to evaluate the extent of damage to the RBCs. However, the fHb release does not fully reflect RBC damage. Notably, leakage of RBC intracellular hemoglobin through membrane pores, without cell rupture, is a common occurrence [35]. Our results showed that the mean SSS and WSS levels generated by the blood pump were 5.12 Pa and 47.41 Pa for the 100-mmHg loop and 8.47 Pa and 92.59 Pa for the 350-mmHg loop, respectively. The difference in shear stress-induced under the two pressure head levels results in variations in the damage to the RBCs. In the 350-mmHg loop, plasma fHb significantly increased at 60 min, which was significantly earlier than observed in the 100-mmHg loop. Although plasma iron levels significantly increased at 350 mmHg, lactate dehydrogenase levels remained unchanged in either group. Despite a substantial release of hemoglobin into the plasma due to shear stress, the RBC count, MCV, and RDW exhibited no significant changes between the two groups, suggesting a minimal alteration in the shape and number of RBCs. Furthermore, SEM results demonstrated the budding or release of EVs from RBCs at 600 min in both circulation loops. TEM analysis revealed a significant increase in MPs in both purified RBC suspension circulation loops (100 and 350 mmHg), along with a significant rise in exosomes in both whole blood circulation loops compared with that in the base group. Owing to higher shear stress levels, plasma TG concentration was significantly higher in the 350-mmHg group than in the 100-mmHg group. Plasma CHO levels increased significantly earlier in the 350-mmHg group than in the 100-mmHg group. Additionally, plasma glucose and NO levels decreased earlier in the 350-mmHg group than in the 100-mmHg group. Pearson’s correlation analysis revealed significant associations of changes in plasma TG, CHO, glucose, and NO levels with fHb levels in both groups. This suggests that RBC damage induces alterations in multiple plasma biomarkers, potentially altering their functions. In addition, MDA, metHb, and T-AOC levels reflecting oxidative stress in RBCs remained relatively stable. These findings indicate that although high shear stress induces RBC damage, it does not significantly alter their shape and quantity. A multifaceted approach is required to understand RBC damage, as fHb alone cannot fully capture the extent of damage to the RBCs.

Various forms of RBC damage include hemoglobin release [36, 37], membrane alterations (including pore formation in the membrane) [38], inner structure damage [39], and functional trauma [12]. The increase in free hemoglobin caused by blood pumps has been previously reported, and the findings are consistent with the results of this study [40]. However, fHb release does not fully explain the RBC damage mechanism. In this study, fHb increased significantly with circulation time, followed by a subsequent increase in plasma iron levels. As the main function of iron is to bind to oxygen, increasing plasma iron levels could lead to a decreased capacity to convey oxygen in RBCs [41]. Apart from gradual rupture, the hematological analysis revealed no significant changes in RBC count or morphology over the 600-min circulation period. Plasma LDH [42,43,44,45], a marker of membrane rupture, was not extensively released into plasma, supporting the conclusion that RBC membrane rupture is not noticeable at both circulation loops. Notably, SEM analysis findings also supported this conclusion. Besides shear-induced RBC damage, oxidative stress can exacerbate damage to the RBCs and increase sensitivity to shear stress [46,47,48,49,50]. Markers such as MDA, metHb, and T-AOC reflect oxidative damage to the RBC membrane [49], hemoglobin [51], and antioxidative capacity [46], respectively. However, our study did not find significant changes in oxidative stress [49, 52]. Plasma glucose levels were assessed to gauge RBC metabolism, which is crucial for anaerobic glycolysis and the pentose phosphate pathway [53, 54]. Plasma glucose affects cell metabolism and protects RBCs from oxidative stress [55]. Under shear stress conditions in the two circulation loops, RBC metabolism accelerated, was evidenced by early significant decreases in both circulation loops [21, 56]. Furthermore, TG and CHO increased in both circulation loops, and their increase significantly correlated with fHb levels [57,58,59,60].

The SEM analysis revealed a significant increase in REVs in samples from both circulation loops at 600 min compared with that in the base group. Typically, REVs, including exosomes and MPs, exhibit different mechanisms of formation and secretion. In this study, TEM findings revealed a significant production of a large number of exosomes and MPs at 600 min in both whole blood circulation loops, with exosomes outnumbering MPs. In normal physiological circulation, exosomes and MPs are typically fewer, whereas under thrombotic and inflammatory conditions, their numbers increase. Therefore, blood-pump-assisted circulation may elevate thrombosis and inflammation risks due to increased EVs. To further understand the source of plasma EVs, two circulation loops at 100 and 350 mmHg were established using purified RBCs and plasma. After 600 min, samples were examined via TEM. A comparison between the whole blood circulation loop and the purified RBC suspension circulation loop revealed a higher quantity of MPs in the latter, whereas exosomes were fewer. This finding suggests that device-induced RBC damage may primarily generate MPs. In contrast, exosomes might be mainly produced from damage to other cells, such as white blood cells and platelets.

This study had certain limitations. First, although plasma EVs can exert various effects and significantly affect physiological functions, their specific effect on ECMO circulation requires further study. Second, our assessment of oxidative damage to RBCs was limited; hence, comprehensive experiments are required to further investigate oxidative damage. Third, considering the higher complexity of in vivo physiological environments compared to controlled in vitro experimental settings, the implications of this study require validation through future in vivo experiments. Forth, the impact of the resistance clamp in hemolysis needs to be considered in blood pump-induced blood damage study [61]. Finally, although different turbulence models can predict the main flow structure, the subtle flow details predicted by different turbulence models might be different. The SST k-ω turbulence model might affect the prediction of flow at the rotor and tongue, but not the prediction of shear stress generation. It is now common to use the SST k-ω turbulence model for shear stress prediction in blood pump studies, and the experimental values are in good agreement with the corresponding numerical results [4]. Therefore, the results obtained in this study are reliable and valid. In addition, some studies have also shown that a steady-state simulation method is feasible for continuous blood pumps.

5 Conclusion

The study findings suggest that NPSS generated during blood pump operation primarily damages RBCs, with oxidative stress playing a lesser role in RBC damage. Plasma fHb alone is not sufficient to gauge blood pump-induced RBC damage, which should be comprehensively assessed considering structural, functional, and metabolic aspects. In both the 100- and 350-mmHg circulation loops, plasma fHb increased over time, with a significantly earlier increase in the 350-mmHg group. Although iron-carrying hemoglobin exiting RBCs may alter their function, their shape and number do not change significantly. Shear stress prompts the release of MPs from RBCs into the plasma, accompanied by varying increases in plasma TG and CHO levels. RBCs respond to different shear stress by adjusting their metabolism through glucose consumption. Increased damage to the RBCs may lead to thrombosis and inflammatory responses via plasma NO consumption. There is a significant correlation between hemolysis and plasma TG, CHO, glucose, and NO levels. A comprehensive understanding of RBC damage can guide clinical management of ECMO-related complications and improve patient prognosis.