Table 1.

Clinical studies investigating changes in leukocyte numbers, phenotype, and functions in the context of ECMO.

Study (year, type) Study population (total n)	Reported indications	Pump system	ECMO membrane	Circuit tubing	ECMO config.	Summary of leukocyte-related outcomes	Ref
Hocker et al. (1990, prospective obs.) Neonates (6)	MAS, PPHN, RDS, AFA	Unknown	Silicone	Unknown	VA	Neutrophil and monocyte numbers reduced significantly during the initial 2 h of ECMO, and remained low until decannulation. Compared to monocytes, neutrophil recovery was slow. Lymphocyte numbers were unchanged. Neutrophil CD11b/CD18 expression (activation) increased during the initial 2 h of ECMO, and returned to pre-ECMO level after 6–24 h. No ECMO-related changes were observed in neutrophil phagocytic function, or monocyte HLA-DR expression and LPS response when compared to pre-ECMO.	(13)
Zach et al. (1990, prospective obs.) Neonates (20)	Unknown	Vortex	Silicone	Unknown	VA	Neutrophil and lymphocyte numbers reduced substantially after ECMO cannulation. Lymphocyte numbers began to recover on day 5, while neutrophil numbers remained low throughout the assessed 7 ECMO days.	(14)
DePalma et al. (1991, prospective obs.) Neonates (25)	PFC, CDH, RDS, PH	Roller occlusion	Silicone	Unknown	VA	Significant drop in overall lymphocyte counts after ECMO cannulation compared to pre-ECMO, and recovered by day of decannulation. No changes were observed in B lymphocyte numbers, and T lymphocyte subset numbers. Expression of activation CD25 and HLA-DR surface marker also unchanged.	(15)
Plotz et al. (1993, prospective obs.) Neonates (10)	MAS, CDH, RDS, PH, GBSP sepsis	Roller	Silicone	PVC	VA, VV	Leukocyte numbers dropped on the day of ECMO cannulation compared to pre-ECMO. Slow recovery was observed after 48 h, and returned to pre-ECMO levels post-cannulation.	(16)
DePuydt et al. (1993, prospective obs.) Neonates (66)	Unknown	Unknown	Unknown	Unknown	Unknown	Neutrophil phagocytosis and intracellular killing indices of ECMO neonates were significantly higher against in vitro candida stimulation, compared to healthy neonates. The effects were independent of ECMO duration.	(17)
Fortenberry et al. (1996, prospective obs.) Neonates (15)	PPHN, MAS, GSBP sepsis, CDH and more	Roller	Unknown	Unknown	VA, VV	CD11b expression on neutrophils peaked at 15 min of ECMO. Began to reduce after 30 min and dropped to below the pre-ECMO expression after 24 h. A similar trend was also observed for NE and IL-8 plasma levels. Authors correlated neutrophil activation with early pulmonary deterioration. However, no reported changes in CD62L expression, IL-6 and TNF-α levels.	(18)
Kawahito et al. (1998, retrospective) Adults (16)	Cardiogenic shock	Unknown	Unknown	Unknown	VA	Lymphocytopenia occurred in both survivors and non-survivors after ECMO. While lymphocyte numbers recovered after weaning in survivors, they remained low in non-survivors. This was correlated with infection-mediated mortality after weaning of ECMO reported in 90% of non-survivors.	(19)
Hong et al. (2015, prospective obs.) Adults (23)	Cardiogenic shock	Centrifugal	Hollow-fiber	Unknown	VA	Sharp reduction in overall lymphocytes and CD4+ T (helper) lymphocytes after 2 h of ECMO compared to when ECMO was being cannulated. Authors associated overall lymphocytes at 2 h, T helper lymphocytes at 6H, and overall T lymphocytes at 24 h with increased mortality, alongside a number of other markers.	(20)
Ziemba et al. (2016, prospective obs.) Neonates, pediatrics (13)	Unknown	Unknown	Unknown	Unknown	VA, VV	LPS-induced HLA-DR expression on monocytes decreased at initiation of ECMO compared to healthy controls. Paralleled by low LPS-induced TNF-α and PHA-induced IFN-γ and IL-10 release. Low HLA-DR expression persisted over the entire duration of ECMO, and proposed to increase risk for secondary infection while on ECMO.	(21)
Liu et al. (2017, prospective obs.) Adults (51)	ARDS ± pulmonary hypertension, arrhythmia, and more	Unknown	Unknown	Unknown	VA, VV	High IL-10 levels pre- and 24 h after initiation of ECMO were linked to a poorer prognosis, which was associated with slow or delayed recovery of CD14+CD16+ and CD14+TLR4+ monocytes and regulatory T lymphocyte numbers.	(22)
Wilm et al. (2017, prospective obs.) Adults (41)	Cardiac and respiratory failure	Unknown	PMP hollow-fiber	Unknown	VA, VV	Leukocyte adhesion onto oxygenators increased in younger (30–47 years old) patients with higher leukocyte counts and was reduced in older patients (61–71 years old). No negative effects were reported for either population.	(23)
Bredthauer et al. (2017, prospective obs.) Adults (38)	Cardiac and respiratory failure	Unknown	PMP hollow-fiber	Unknown	VA, VV	Leukocyte adhesion differs between two different ECMO systems. A greater coverage (88%) of small single-cell leukocytes was reported for Xenios AG Hilite7000LT membrane oxygenators compared to Maquet (-Getinge) PLS membrane oxygenators (33%). No significant difference in coagulation disorder was reported between systems.	(24)
Francischetti et al. (2017, retrospective) Adults, pediatrics (166)	Sepsis, ARDS, MAS, CHF and more	Unknown	Unknown	Unknown	VA, VV	Significant reduction in overall lymphocyte numbers after ECMO initiation was linked to mortality, alongside a number of laboratory values.	(25)
Santiago-Lozano et al. (2018, Retrospective) Pediatrics (100)	Sepsis, cardiac disease, respiratory failure	Unknown	Unknown	Unknown	VA	No changes in overall leukocyte numbers between ECMO patients with or without subsequent infection were observed. Paralleled by lack of changes in C-reactive protein. ECMO-related leukocyte reduction was observed, but not discussed within the publication.	(26)
Ortega et al. (2019, prospective obs.) Neonates, pediatrics (25)	Septic shock, cardiogenic shock, PPHN, CDH, MAS, cardiac arrest, respiratory failure and more	Centrifugal	Unknown	Unknown	VA, VV	Activated CD4+ T (helper) lymphocytes expressing CD161 reduced, and T-cell autoreactivity to CNS antigens increased with ECMO support compared to sick controls. Further, ECMO patients presented with acquired brain injury and cerebral autoregulation impairment were reported with increased CD8+ (cytotoxic) T lymphocyte and B lymphocyte selective autoreactivity to CNS antigens. Increased CD161+ cytotoxic T, B, and NKT lymphocytes in these patients were also observed, paralleled by significant IL-6 and IL-8 plasma levels. No significant changes were observed in other studied leukocyte populations between the patient groups.	(27)
Sargin et al. (2019, retrospective) Adults (119)	Post-cardiotomic shock	Unknown	Unknown	Unknown	VA	A significant NLR increase in the first 3 days of ECMO was observed in patients who developed renal failure. These ECMO patients had a comparatively higher NLR to those who did not.	(28)
Han et al. (2020, prospective obs.) Unknown (57)	Unknown	Unknown	Unknown	Unknown	Unknown	ECMO reduced monocyte capacity to secrete IL1β, IL-6, and TNF-α compared to healthy controls.	(29)

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).

AFA, amniotic fluid aspiration; ARDS, acute respiratory distress syndrome; CD, cluster differentiation; CDH, congenital diaphragmatic hernia; CHF, chronic heart failure; CNS, central nervous system; GBSP, Group B Streptococcus;, H, hours; HLA-DR, human leukocyte antigen-DR isotype; IL, interleukin; IFN, interferon; LPS, lipopolysaccharide; MAS, meconium aspiration syndrome; Min, minutes; NKT, natural killer T lymphocyte; NLR, neutrophil-to-lymphocyte ratio; Obs., observational; PHA, phytohemagglutinin; PLS, permanent life support; PFC, persistent fetal circulation; PH, pulmonary hypoplasia; PMP, polymethylpentene; PPHN, persistent pulmonary hypertension of the newborns; PVC, polyvinylchloride; RDS, respiratory disease syndrome; VA, veno-arterial; VV, veno-venous.