Table 2.

Experimental studies investigating changes in leukocyte numbers, phenotype, and function in the context of ECMO.

Study (year)	Model (species),System rep.(total n)	Pump system	ECMO membrane	Circuit tubing	Summary of leukocyte-related outcomes	Ref
Bergman et al. (1994)	Ex vivo (human), Neonatal (16)	Centrifugal pump	Silicone	PVC	Overall leukocyte number and deformability reduced significantly during the 72 h of ECMO compared to pre-ECMO levels. Leukocyte adhesion and clogging rate to the oxygenator membrane also increased.	(30)
Skogby et al. (1998)	Ex vivo (human), Neonatal (16)	AREC vs. centrifugal pump	Hollow-fiber (type unknown) vs. Silicone	Silicone vs. PVC	Two perfusion systems were compared: AREC perfusion system had lower neutrophil loss and CD11b expression during the 24 h of ECMO compared to centrifugal system. No differences were observed for expression of CD11a and ICAM-1, or plasma levels of cytokines and chemokines (IL-1β, IL-6, IL-8) between systems.	(31)
Graulich et al. (2000)	Ex vivo (human), Neonatal (6)	Nonocclusive roller pump	Silicone	PVC	Early activation of both neutrophils and monocytes was characterized by increased CD18 expression and CD62L shedding after 2–4 h of ECMO compared to baseline. After 8 h, expression of CD18 returned to baseline level. While the reduced CD62L expression, and increased plasma NE level persisted. No leukocyte loss, or lymphocyte phenotypic changes observed.	(32)
Adrian et al. (2003)	Ex vivo (human), Neonatal (8)	Roller pump	Silicone	Unknown	Three groups were compared: No S-nitrosglutenione (GNSO), low and high GNSO doses (no differences found). Regardless of GNSO, total leukocyte and neutrophil numbers reduced during the 24 h of ECMO. Increased CD18 and CD11b expression was observed at initiation of ECMO and reduced after 3 h in all groups. GSNO did not dampen ECMO-mediated inflammatory responses.	(33)
McILwain et al. (2010)	Animal (porcine), Neonatal (16)	Centrifugal pump	Microporous PP Hollow-fiber	Unknown	Increased CD18, CD35, CD62L, and CD11b expression on neutrophils following 2 h of VA-ECMO in comparison to sham controls. Leukocyte infiltration also elevated resulting in damage in both lungs and intestine after 8 h of ECMO. Paralleled by increase in plasma IL-1β, IL-6, IL-8, and TNF-α levels, and tryptase activity (associated with degranulation of mast cells).	(34)
Rungatscher et al. (2015)	Animal (rat), Unknown (20)	Roller pump	Hollow-fibee (type unknown) vs. No oxygenator	Unknown	The effect of extracorporeal circuit with and without oxygenator was compared. p38 MAPK and NF-κB phosphorylation (inflammatory pathways) significantly upregulated in mononuclear leukocytes and granulocytes after 1H of extracorporeal circulation in the presence of an oxygenator (“ECMO”). Paralleled by the rise in IL-6, TNF-α, and NE plasma levels. An increased neutrophil infiltration and severe lung injury scores was also reported.	(35)
Passmore et al. (2016)	Animal (ovine), Adult (43)	Centrifugal pump	PMP hollow-fiber	PVC	ECMO significantly reduced leukocyte numbers in circulation after 24 h in an ovine model with pre-existing lung injury compared to those with injury alone. The combination also promoted monocyte/macrophage and neutrophil MMP2 and MMP9 expression in lung tissues, and increased lung infiltration. All aligns with the increased lung oedema and inflammatory response reported.	(36)
Ki et al. (2019)	Ex vivo (human), Adult (10)	Centrifugal pump	PMP hollow-fiber	PVC	Two ECMO flow rates were compared: high (4 LPM) and low 1.5 (LPM). Both models showed ~50% reduction in monocyte numbers over 6 h of circulation, but not others. A rapid release of NE and MPO by activated neutrophils over time during high flow compared to low flow. Paralleled by elevated plasma levels of IL-1β, IL-6, and TNF-α. However, no differences in leukocyte subset activation (CD18, CD11b, CD66b, HLA-DR, CD25, and CD80 expression) were observed between groups.	(37)
Zhang et al. (2019)	Animal (porcine), Unknown (24)	Centrifugal pump	PMP hollow-fiber	PVC	Compared to conventional cardiac arrest post-resuscitation method, ECPR (VA-ECMO) increased spleen CD4+ T (helper) lymphocytes and CD4+/CD8+ T lymphocyte ratio, and enhanced T lymphocyte proliferation and low apoptotic rate. Paralleled by increased IL-2, IL-4 and IFN-γ levels and reduced ROS production in the spleen. All associated with increased survival.	(38)
Meyer et al. (2020)	Ex vivo (human), Neonatal (15)	Roller pump	Microporous PC hollow-fiber	Unknown	Three ECMO flow rates were compared: low (0.3 LPM), nominal (0.5 LPM), and high (0.7 LPM). Greatest leukocyte-derived EV released at high flow followed by low flow, peaked at 2–4 h of ECMO. Expression of tissue factor on leukocytes and their EV exhibited a similar trend, may be contributing to occlusive thrombosis. Akin to static control, nominal flow remained unchanged.	(39)

A structured search of EMBASE and PubMed/MEDLINE for all relevant publications was conducted. Search terms were “ECMO” OR “extracorporeal membrane oxygenation” OR “ECLS” OR “extracorporeal life support” OR “artificial heart/lung” AND “leukocytes” OR “white blood cells” OR “immune response.” Studies published specifically to report on leukocyte modulation during ECMO were included (see Supplementary Material for detailed literature search strategy).

AREC, Assistence respiratoire extracorporelle; CD, cluster of differentiation; CD62L, L-selectin; ECPR, extracorporeal cardiopulmonary resuscitation; EV, extracellular vesicles; GNSO, S-nitrosglutenione; IFN, interferon; IL, interleukin; LPM, liters per minute; MMP, matrix metalloproteinase; MPO, myeloperoxidase; NE, neutrophil elastase; NF-κB, nuclear factor kappa-light-chain-enhancer; p38 MAPK, p38 mitogen-activated protein kinase; PC, phosphorylcholine; PVC, polyvinylchloride; PMP, polymethylpentene; PP, polypropylene; ROS, reactive oxygen species; System rep, system represented; TNF, Tumor necrosis factor; VA, Veno-arterial; VS., Versus.