Changes in Platelet Counts in Patients Undergoing Extracorporeal Membrane Oxygenation Technique

Abstract

Platelet count disorders in patients on extracorporeal membrane oxygenation (ECMO) are intricate and involve complex mechanisms. This chapter aims to summarize knowledge on platelet count changes in the ECMO population. Thrombocytopenia occurs in about 21% during ECMO, declining over 2–3 days, persisting up to 7 days post-ECMO, with heparin-induced thrombocytopenia at 3.7%. Diagnosis relies on complete blood count with platelet count <150 G/L or rotational thromboelastometry (A5 EXTEM <35 mm and A5 FIBTEM ≥9 mm). Combination of thrombocytopenia and coagulation disorders may lead to bleeding (44.7%) or thrombosis (22.9%) in patients undergoing ECMO. Platelet testing every 6–12 hours is crucial, target in ECMO population for ≥100,000 × 10^9/L with bleeding and lower (≥50,000–100,000 × 10⁹/L) without. Up to 50% of ECMO patients may require platelet transfusions, as per recent studies.

1. Introduction

Extracorporeal membrane oxygenation (ECMO) serves as a transient mechanical support, witnessing a substantial surge in utilization over the past two decades. Notably, ECMO treatment is frequently accompanied by hemorrhagic and thromboembolic complications, contributing significantly to morbidity and mortality [1]. Exposure to foreign circuit surfaces and elevated shear stress amplifies platelet activation, heightening the predisposition to thrombosis [2, 3]. [4]. In the realm of anticoagulant therapies, essential for circuit patency and prevention of thrombotic complications unfractionated heparin (UFH) emerges as the predominant choice. However, this preference introduces potential complications, such as heparin-induced thrombocytopenia (HIT) and paradoxically, an escalated risk of thrombosis [5]. When compounded with the complexities of severe illness in critical care settings, these mechanisms collectively contribute to a downward trend in platelet count [6, 7].

2. Structure and physiology of platelet

2.1 Platelet physiology

Platelets, derived from megakaryocyte cells in the bone marrow and lungs, undergo continuous clearance and production, generating approximately 100 billion platelets daily (70 million/min) to maintain a physiological concentration of around 250 × 10⁹/L (with a range of 150–400 × 10⁹/L) [4]. Their normal lifespan is 10 days. The platelet role helps prevent blood loss and ensures vascular integrity [8]. Normal platelets participate in the balance of hemostasis through the following activities: activation, adhesion, secretion of active substances, and aggregation [9].

2.2 Platelet granules

Platelets contain three main secretory granules: alpha (α) granules, dense granules, and lysosomes (Figure 1). Each platelet typically holds between 50 to 80 alpha granules, packed with a variety of vital components, including fibrinogen, fibronectin, von Willebrand factor (VWF), platelet factor 4 (PF4), P-selectin, and plasminogen activator inhibitor-1 (PAI-1). Dense granules, with 3–8 per platelet, house adenosine triphosphate (ATP), adenosine diphosphate (ADP), P-selectin, calcium, and others. Lysosomes, ranging from 0 to 3 per platelet, contain acid hydrolases. Released stimuli from platelets enhance adhesion, recruiting more platelets and promoting thrombus growth through platelet aggregation [4, 10, 11].

Figure 1.

The ultrastructure of platelets. ADP, adenosine diphosphate; PDGF, platelet-derived growth factor; PF, platelet factor; VWF, von Willebrand factor [10].

2.3 Membrane glycoproteins

The platelet membrane comprises nine glycoproteins (GPI to GPIX), with the GPIb-IX-V complex serving as the primary receptor for von Willebrand factor (VWF), crucial for platelet tethering to subendothelial collagen during vascular injury [12]. This interaction initiates hemostasis and pathological thrombosis, facilitating subsequent fibrinogen binding to GPIIb/IIIa for platelet aggregation. Platelet adhesion through GPIb-IX-V is shear-dependent with high shear stress promoting a high-affinity ligand between VWF and the complex. Dysfunction in VWF or platelets increases bleeding risk due to impaired primary hemostasis [4, 10]. GPVI, a collagen receptor, triggers platelet activation and adhesion upon interaction with collagen. The reversible VWF-GPIb-IX-V interaction is insufficient for stable adhesion. [4, 13]. GPIIb/IIIa, a fibrinogen receptor, plays a central role in platelet adhesion. Inactivated GPIIb/IIIa binds immobilized fibrinogen, while activated GPIIb/IIIa captures soluble fibrinogen, leading to platelet aggregation [12, 14]. Thrombin, generated in response to tissue factor exposure during vessel injury, triggers inside-out activation of GPIIb/IIIa.

2.4 The role of platelets on the hemostatic coagulation system

Primary hemostasis involves a sequential three-step process aimed at forming a blood clot to seal a breach in a vessel wall: (1) platelet adhesion to the vessel wall, (2) platelet activation, and (3) the formation of platelet aggregates (Figure 2). Immediately following vessel damage, a mix of stationary (such as collagen) and mobile (such as thrombin, ADP, and TxA2) platelet stimulants gather at the site. Platelets kickstart adhesion to the subendothelial matrix proteins using their collagen receptors, wherein the level of von Willebrand factor (vWF) engagement hinges on the nearby blood flow’s shear rate (Figure 3). This orchestrated sequence of events is crucial for the rapid response to vascular injury, initiating the process of forming platelet aggregates and ultimately leading to the formation of a blood clot, effectively sealing the breach in the vessel wall [11, 15].

Figure 2.

The involvement of blood vessels, platelets, and blood coagulation in hemostasis and ADP, adenosine diphosphate [10].

Figure 3.

Adhesion and activation mechanisms support the hemostatic and prothrombotic function of platelets [15].

Platelets’ interaction with the damaged vessel wall relies on the GPIb-V-IX receptor complex, crucial for stable adhesion during primary hemostasis. Endothelial disruption, as in atherosclerotic plaque rupture, exposes vWF and collagen, initiating platelet adhesion. At low shear rates, platelets interact directly with the extracellular matrix via various receptors. In high shear stress areas, vWF immobilizes collagen, binding to GPIba receptors and unfolding, exposing sites for GPIb-IX-V complex, and enabling GPVI interaction with collagen and fibronectin [11, 15].

Initial platelet tethering to vWF-collagen complexes allows for activation by collagen, forming a platelet monolayer that supports adhesion. Released contents from platelet granules strengthen adhesion by forming cross-bridges between platelet GPIIb/IIIa receptors and endothelial integrins. Integrin-induced adhesion, via phospholipase C-dependent GTPase Rap1b stimulation, activates GPIIb/IIIa receptors. This explains enhanced platelet adhesion in high shear stress areas, emphasizing their role in arterial thrombosis and clot formation in stenotic lesions [11, 16].

Platelet activation through GPVI collagen receptors triggers calcium signaling, enhanced by released mediators, such as TxA2 and ADP. Externalized phospholipids during activation facilitate coagulation progression. Phosphatidylserine exposure, induced by collagen plus thrombin, enhances procoagulant transformation. Platelet-derived microvesicles contribute to clot-promoting activity. Common outcomes of receptor activation include amplification, additional platelet recruitment, and GPIIb/IIIa receptor activation for hemostatic plug formation. Understanding these processes sheds light on clot formation and potential therapeutic targets [11, 17].

3. ECMO associated coagulopathy

3.1 Initiation of ECMO: coagulation pathway activation and inflammatory response

Contact between blood and the artificial surfaces of an extracorporeal circuit triggers a multifaceted inflammatory reaction, intertwining both the coagulation and inflammatory pathways (Figure 4). This response leads to capillary leak, which can cause temporary dysfunction of every organ. Mechanisms that normally maintain physiologic homeostasis become disrupted, which results in increased acuity in the already physiologically compromised patient [9].

Figure 4.

Simplistic representation of the blood-surface interaction during ECLS. This shows the components relevant to thrombosis, and even though complement and leukocytes are considered to be involved with inflammation, they are also very relevant participants in thrombosis formation [9].

3.1.1 Pathophysiology of the blood/biomaterial surface interaction

Contact with synthetic, nonendothelial surfaces, along with factors such as shear stresses, turbulence, cavitation, and osmotic forces, directly induces damage to blood during extracorporeal membrane oxygenation (ECMO). This process leads to the progressive denaturation of plasma proteins and lipoproteins, causing an increase in plasma viscosity, protein solubility, and the generation of macromolecules. Red blood cells (RBCs) undergo reversible echinocyte changes during ECMO, but some are subject to hemolysis due to shear forces and activated complement. It is worth noting that platelets and white blood cells (WBCs) also experience injury during perfusion, but the consequences of their activation outweigh the effects of direct injury. The impact of shear rate on platelets and coagulation components is crucial, influencing platelet deposition and the generation of factor Xa through augmentation of the tissue factor (TF): VIIa complex. Lower shear rates, in contrast, lead to less platelet deposition but more fibrin deposition [9].

3.1.2 Platelets

When collagen is exposed due to blood vessel damage, von Willebrand Factor (vWF) is discharged, leading to platelet adhesion [9]. In environments of elevated shear stress, such as arteries and microvessels, high molecular weight vWF facilitates platelet adhesion by linking with collagen and the inactivated platelet via the platelet receptor glycoprotein (GP) Ib-IX-V. GPIb mediates platelet interactions with VWF. The adherence of platelets via GPIb to adsorbed vWF depends upon shear stress, which provides the conformational change in VWF needed to allow binding. In low shear stress conditions (large veins) platelet adhesion occurs through direct interaction via GP VI with collagen.

Platelet secretion occurs from intracellular granule collections and results in a number of actions due to agents released including:

• Increasing platelet adhesion and aggregation (released: adenosine diphosphate, vWF, fibrinogen, and thrombospondin).

• Participation in coagulation (released: fV, fibrinogen).

• Increased vascular tone and contraction (released: serotonin).

• Increased cell proliferation and migration (released: platelet-derived growth factor, PDGF)

• Fibrinogen binds to activated platelets through the receptor GPIIb-IIIa and acts as a bridge between platelets resulting in aggregation.

GPIIb-IIIa (CD41/CD61) is the dominant platelet receptor (Figure 5). While in a resting state, platelets possess inactive GPIIb-IIIa, resulting in a low affinity for adhering to adsorbed fibrinogen. Upon activation, platelets undergo conformational changes, exposing the high-affinity binding site of GPIIb-IIIa. This allows for the binding of soluble fibrinogen, triggering platelet aggregation and the formation of platelet-leukocyte aggregates. These aggregates are formed either by the cross-linking of two GPIIb-IIIa receptors or by the interaction of GPIIb-IIIa with Mac-1 on the leukocyte, facilitated by fibrinogen.

Figure 5.

Platelet adhesion to the damaged blood vessel wall and to each other (aggregation). The binding of the platelet glycoprotein GPIb, which forms a complex with GPV and GPIX to exposed long multimers of von Willebrand factor (VWF), leads to platelet adhesion to the subendothelium. It also exposes the GPIIb/IIIa binding sites, leading to further binding to VWF and so further platelet adhesion to the subendothelium. The GPI-IIa and -VI sites form direct attachments to exposed subendothelial collagen. Platelets also attach to each other (platelet aggregation) via binding to VWF and to fibrinogen by the GPIIb/IIIa receptors [10].

During procedures, such as CPB and ECLS, as well as with vascular access devices, platelet activation and adhesion occur, contributing to ongoing coagulation stimulation. Adherent platelets and platelet microparticles enhance coagulation activity. In ECLS, platelet adhesion and aggregation can reduce platelet counts, and despite minimal adhesion, microemboli formation may still happen. Over time in ECLS, adherent platelets detach, leaving platelet membrane fragments circulating. Platelet pool during ECLS comprises fewer normal platelets, more activated ones, and larger ones from bone marrow release. Bleeding times lengthen despite normal-looking platelets as ECLS extends beyond 24 hours due to ongoing platelet consumption [9].

3.2 How does the process of coagulation activation affect platelets?

Platelets adhere to the circuit surfaces and become activated, which leads to platelet aggregation and further activation of the coagulation system. Platelet activation and consumption occurs upon ECLS initiation causing decreases in platelet number and function within the first hour of ECLS. Platelet activation and consumption continue throughout the course of ECLS often requiring regular platelet transfusions [9]. The high molecular weight polymers of vWF under the effect of shear stress created by the rotation of the centrifugal pump will be fragmented and damaged, thereby causing acquired von Willebrand disease. Because of this reason, the adhesion ability of platelets is damaged, causing bleeding [18]. vWF levels will begin to increase again approximately 1 day after decannulation [19]. The complement system activation during ECMO contributes to associated coagulopathy. Circulating anaphylatoxins interact with coagulation cascade proteases, platelets, and endothelial cells, increasing TF expression and vWF production, and contributing to coagulation and platelet disorders [19, 20].

3.3 Effects of anticoagulants

Due to the heightened risk of thrombosis following contact with the artificial surfaces of the ECMO system, the administration of anticoagulants is imperative to mitigate potential thrombotic events at various points along the circuit [21]. Presently, unfractionated heparin (UFH) remains the predominant anticoagulant employed during ECMO due to its rapid onset of action and reversible effects with protamine. UFH exerts its anticoagulant properties by augmenting the activity of antithrombin (AT) through binding mechanisms. UFH is administered via intravenous infusion route. [21, 22].

3.4 How does coagulopathy differ between veno-venous (VV-ECMO) and venoarterial (VA-ECMO)?

Several investigations have unveiled notable distinctions in coagulation markers and platelet activity between veno-venous (V-V) and venoarterial (V-A) ECMO recipients. Elevations in fibrinogen levels and fibrin polymerization are specifically observed in V-V ECMO patients, fostering a prothrombotic milieu. This occurrence can be rationalized by the underlying pathophysiology of acute respiratory distress syndrome (ARDS), which prompts intractable hypoxemia necessitating V-V ECMO intervention. ARDS hallmark feature is an inflammatory-coagulation imbalance, instigating the intricate cascade of immunothrombosis, thereby influencing relevant coagulation parameters. Additionally, heightened GP IIb/IIIa activation is noted in V-V ECMO cases compared to V-A ECMO, fostering platelet activation and, consequentially, augmenting the immunothrombosis cascade [19, 23].

4. Test for diagnosis thrombocytopenia

4.1 Complete blood count

The total analysis of complete blood count with platelet count <150 G/L is the most important test to diagnose thrombopenia [10]. In addition, viscoelastic hemostatic assays (ROTEM, TEG) can be used. However, it is often used primarily as a point-of-care in diagnosing the cause of bleeding [24].

4.2 Rotational thromboelastometry

4.2.1 Principles

ROTEM is a comprehensive whole blood viscoelastic hemostasis analyzer, offering four distinct measurement channels for specific assays. This multifaceted approach enhances diagnostic accuracy compared to single-assay systems. The versatility of the ROTEM system permits not only the prompt detection of coagulopathies but also enables differentiation between various types, such as hypofibrinogenemia and thrombocytopenia. Designed to inform tailored hemostatic interventions in bleeding patients, the ROTEM system encompasses activated assays including EXTEM, FIBTEM, and APTEM for extrinsic activation, INTEM and HEPTEM for intrinsic activation, ECATEM for ecarin activation, and NATEM for nonactivated assessment [25].

In EXTEM assays, coagulation is initiated via the extrinsic pathway, predominantly relying on factors VII, X, V, II, and fibrinogen for initial thrombin generation and subsequent clot formation. FIBTEM, a variant of EXTEM, introduces a potent platelet inhibitor to specifically evaluate clot strength based solely on fibrinogen levels and polymerization, allowing for differentiation between thrombocytopenia and hypofibrinogenemia. The discrepancy in clot strength between EXTEM and FIBTEM facilitates the assessment of the platelet component, sometimes termed PLTEM. APTEM, another extrinsically activated assay, integrates an antifibrinolytic agent to assess the efficacy of antifibrinolytic therapy. In INTEM assays, clot formation predominantly relies on factors XII, XI, IX, VIII, X, V, II, and fibrinogen. Similar to EXTEM, clot firmness in INTEM reflects contributions from both platelets and fibrin. Unlike extrinsically activated assays, INTEM lacks a heparin inhibitor. However, a modified version, HEPTEM, incorporating additional heparinase, is employed alongside INTEM to detect any residual heparinization or protamine overdose. Other channels, such as NATEM and ECATEM, offer additional insights into coagulation dynamics [25].

Mechanism and meaning of parameters of some channels of ROTEM system. The meaning of the parameters is in the Figure 6:

Figure 6.

ROTEM® trace (“TEMogram”) displaying the clinically most important parameters and their informative value. FDPs = fibrinogen split products. Courtesy of Klaus Görlinger, TEM international [25].

4.2.2 The stages of clot formation

According to Figure 6, the clot formation process consists of three stages: (1) Coagulation activation and clot polymerization, (2) Clot firmness, and (3) Clot lysis. Each stage is characterized by distinct parameters. Coagulation time (CT) and clot formation time (CFT) represent stage (1) and are directly influenced by factors such as coagulation factors, anticoagulants, fibrin degradation products (FDPs), and tissue factor expression. Stage (2) pertains to the mechanical strength of the blood clot, contingent upon factors such as platelet count and function, fibrin concentration and polymerization capacity, and the influence of factor XIII and colloids. In cases of severe bleeding, expedited decision-making is crucial. Clot firmness amplitudes at 5- or 10-minute post-coagulation time (A5 and A10) exhibit a strong correlation with the maximum clot firmness (MCF), facilitating rapid assessment of clot strength. Additionally, the A5 and A10 values from EXTEM and INTEM assays are closely associated with platelet count and fibrinogen concentration, while those from FIBTEM correlate well with plasma fibrinogen levels. Furthermore, the difference between A10 and A5 values (PLTEM) from EXTEM and FIBTEM assays demonstrates a notable correlation with platelet count, aiding in comprehensive hemostatic evaluation. The parameters of stage (3), including maximum lysis (ML) and the lysis indices at 30 and 60 minutes (LI30 and LI60), offer insights into the activity of fibrinolytic enzymes, inhibitors of fibrinolysis, and the activity of factor XIII [25].

According to ROTEM protocol [25], a diagnosis of thrombocytopenia is established when A5 EXTEM <35 mm and A5 FIBTEM ≥9 mm. Notably, beyond diagnosing hemostatic disorders such as increased fibrinolysis, reduced platelet count, or significant deterioration in platelet quality, the regimen effectively identifies abnormalities in fibrin polymerization, deficiencies in vitamin K-related coagulation factors, and hypercoagulation. ROTEM, in comparison to standard coagulation tests, proves to be more proficient in diagnosing coagulation disorders, distinguishing between heparin-induced complications and deficiencies in the endogenous coagulation pathway. Additionally, it facilitates the specific diagnosis of protamine overdose when employed to counteract the anticoagulant effects of heparin-induced bleeding, especially in patients undergoing extracorporeal membrane treatments, such as CPB and ECMO [26, 27].

5. Changes of platelet count

Thrombocytopenia definition was according to agreed guidelines criteria: mild (100–150 x10⁹/L), moderate (50–100 x 10⁹/L), and severe (< 50 x10⁹/L) [10, 28, 29]. Federica Jiritano performed a systematic review and meta-analysis revealed a significant prevalence of thrombocytopenia, reaching up to 21% in patients undergoing extracorporeal membrane oxygenation (ECMO). The lowest prevalence of thrombocytopenia was 1.9%, while the highest was 80% [30].

Platelet count tends to decline over the initial 2–3 days, persisting up to 7 days after ECMO implantation [30]. Potential causes for a rapid decrease in platelet count include the patient’s primary disease, toxic drug effects, anticoagulation (heparin-induced thrombocytopenia), and the specific ECMO system used. Malfertheiner et al. conducted a randomized controlled trial, which revealed that the reduction in platelet count is not influenced by the specific ECMO technology employed [31].

Although the overall prevalence of thrombocytopenia is comparable between patients on venoarterial (V-A) ECMO (23.2%) and those on veno-venous (V-V) ECMO (25.4%), there is speculation regarding potential differences in the underlying mechanisms [9]. Patients undergoing V-A ECMO, representing those with cardiorespiratory failure, may have predisposing conditions for pre-ECMO thrombocytopenia, especially in cases of refractory cardiac arrest. Post-cardiotomy ECMO is also more prone to developing thrombocytopenia. Hemodynamic differences in perfusion rates and pressures during V-V ECMO may contribute to variable shear stress, leading to platelet activation and aggregation. Thus, platelet impairment may vary based on different ECMO configurations, such as central vs. peripheral cannulation [9, 30, 32].

Surprisingly, the duration of ECMO does not exhibit a significant relationship with the occurrence of thrombocytopenia, contradicting some previous reports. Despite efforts to control confounding factors, studies by Abrams [33] and others failed to demonstrate an association between the number of days on ECMO and worsening thrombocytopenia. In contrast to these findings, Weingart [34] and Panigada [35], their findings indicated a correlation between platelet count and ECMO duration. Nevertheless, it is important to consider potential confounding factors such as patient illness, baseline platelet count, and the onset of hepatic or renal failure, which could impact the results. The endothelialization process occurring after a few days on ECMO surfaces, particularly in the oxygenator, may mitigate steady thrombocytopenia and platelet dysfunction following the initial inflammatory response. New platelets generated after 8–9 days (platelet lifespan) may no longer be affected by the inflammatory and coagulative cascade triggered by ECMO [30].

In comparison to cardiopulmonary bypass (CPB), where patients undergo a relatively short treatment duration (2–16.5 hours), those treated with extracorporeal membrane oxygenation (ECMO) experience a longer duration of therapy (average 170.4 hours, with some patients up to 1008 hours). Thrombocytopenia, both moderate and severe, occurs more frequently in ECMO patients and begins to recover following ECMO cessation [36].

The dynamics of thrombocytopenia progression during ECMO (Figures 7 and 8): Platelet count may decrease from the onset, even before the initiation of ECMO therapy. Over the initial 7 days, a predominant reduction in platelet count, ranging from moderate to severe thrombocytopenia, is observed, followed by gradual improvement over the course of ECMO treatment. In the later days of the ECMO procedure, the average platelet count tends to be higher across all categories—severe reduction, moderate reduction, mild reduction, or normal—with the majority having platelet counts exceeding 100 G/L [28].

Figure 7.

Changes of platelet count (stratified by severity upon admission (before VA-ECMO) over time. VA-ECMO, venoarterial extracorporeal membrane oxygenation [28].

Figure 8.

The evolution of thrombocytopenia over time in patients experiencing thrombocytopenia while on venoarterial extracorporeal membrane oxygenation (VA-ECMO). Panel (A) displays the total patient count on the y-axis. In panel (B), a stacked bar chart depicts the daily distribution of thrombocytopenia severity among patients. “No.” stands for “number” [28].

Heparin-induced thrombocytopenia (HIT) occurs in 0.5–5% of heparin-treated individuals with clinical suspicion arising when there is a ≥ 50% decrease in platelet count or a total count of 100 x 10⁹/L during heparin therapy [37]. Thrombocytopenia is typically moderate, with median platelet counts ranging from 50 to 80 x 10⁹/L, and nadir counts rarely below 20 x 10⁹/L [38]. The incidence of HIT in patients treated with ECMO using heparin is about 3.7%. Diagnostic tests for heparin-induced thrombocytopenia (HIT) in patients on ECMO are infrequently conducted, leading to a substantial underestimation of the actual HIT incidence in this population. In some cases, suspicion of HIT prompts clinicians to initiate alternative anticoagulants, such as Argatroban or Bivalirudin, particularly in conjunction with low platelet counts [30]. In another study by author Matthias Lubnow et al., it was observed that confirmed HIT occurred more frequently in VV-ECMO compared to VA-ECMO (3.9 vs. 1.7%, p = 0.173). Greater occurrences of different complication types were observed in the HIT-confirmed cohort compared to the ECMO-control group. Nonetheless, there was no notable variance in inhospital mortality rates between the two cohorts (31 vs. 41%, p = 0.804) (Table 1) [38].

Assay	Activators and additives	Clinical comments
ROTEM® delta assays
EXTEM	CaCl2 + recombinant tissue; factor + polybrene	Deficiency of factors of the extrinsic pathway; VKAs (coumadin/warfarin); indication for PCC administration
FIBTEM	CaCl2 + recombinant tissue; factor + cytochalasin D + polybrene	Fibrin polymerization; dose calculation for fibrinogen concentrate or cryoprecipitate
AFTEM	CaCl2 + recombinant tissue; factor + aprotinin/tranexamic; acid+polybrene	Verifying the effect of antifibrinolytic drugs; differential diagnosis to clot retraction and FXIII deficiency (in combination with EXTEM)
INTEM	CaCl2 + ellagic acid	Deficiency of factors of the intrinsic pathway; heparin and protamine effects (in combination with HEPTEM)
HEPTEM	CaCl2 + ellagic acid+heparinase	Heparin and protamine effects (in combination with INTEM)
ECATEM	CaCl2 + ecarin	Direct thrombin inhibitors (e.g., hirudin, argatroban, bivalirudin. dabigatran); not sensitive to heparin; actually only available in Europe
NATEM	CaCl2	Tissue factor expression on monocytes; other anticoagulants (e.g., LMWH)

Table 1.

ROTEM® delta and ROTEM® [25].

6. Bleeding and thromboembolic complications

Patients undergoing extracorporeal membrane oxygenation (ECMO) are often critically ill, heightening the risk of bleeding complications [39]. The incidence of bleeding during ECMO ranges from 20.8 to 39.6%, while thrombosis formation occurs in 10 to 46.1% of patients, depending on circuit type and patient age in various centers [40].

Both bleeding and thrombotic complications are linked to increased morbidity and mortality, with bleeding complications being more prevalent overall compared to thrombotic complications. According to the extracorporeal life support organization (ELSO) 2016 statistics [41], the Table 2 below illustrates the rates of potential complications during ECMO treatment. Bleeding and thrombosis occur frequently at various locations in the body, as well as within the ECMO circuit, as outlined in the table. Bleeding at cannulation or surgical sites is most commonly observed in both veno-venous (VV) and venoarterial (VA) ECMO. However, severe bleeding events, such as intracranial hemorrhage and pulmonary hemorrhage, have also been reported.

	Neonate; (%)	Pediatric; (%)	Adult; (%)
Respiratory
Mechanical: pump malfunction	1.6	2.2	1.5
Mechanical: oxygenator failure	5.7	10.6	9.1
Cannula hemorrhage	7.9	18.3	13.2
Surgical hemorrhage	6.3	12.6	10.5
Pulmonary hemorrhage	4.5	8.1	6.1
CNS hemorrhage	7.6	6.4	3.9
CNS infarction	6.8	4.2	2.0
Renal failure	7.8*	12.9*	9.3†
Hyperbilirubinemia	7.3	5.2	8.7
Infection	5.8	16.8	17.5
Cardiac
Mechanical: pump malfunction	1.5	1.8	0.8
Mechanical: oxygenator failure	6.1	7.2	6.6
Cannula site hemorrhage	10.7	15.6	18.5
Surgical site hemorrhage	29.3	28.9	20.2
Pulmonary hemorrhage	5.2	5.3	3.1
CNS hemorrhage	11.3	5.3	2.2
CNS infarction	3.4	5.0	3.8
Renal failure	12.3*	7.2*	12.3†
Hyperbilirubinemia	4.9	7.2	12.2
Infection	7.1	11.0	13.0

Table 2.

Adverse events during extracorporeal life support (ECLS) by age and indication [41].

^*

>1.5.

^†

>3.0 mg/dl.

Renal failure: serum creatinine.

hyperbilirubinemia: total bilirubin >2 mg/dl or indirect bilirubin >15 mg/dl.

CNS, central nervous system; ECLS, extracorporeal life support.

According to statistics from patients undergoing ECMO without anticoagulation, as reported by Sven R. Olson, MD, and colleagues in a systematic review involving 201 patients undergoing treatment for acute respiratory distress syndrome or cardiogenic shock, and the median duration of anticoagulant-free ECMO was 4.75 days. Incidences of circuit thrombosis and patient thrombosis were observed in 27 (13.4%) and 19 (9.5%) patients, respectively. Any bleeding and “major” or “severe” bleeding events were documented in 66 (32.8%) and 56 (27.9%) patients, respectively. The mortality rate was 19%, with 40 patients succumbing to the complications of ECMO therapy [42].

Previously several studies have demonstrated significant differences between V-V and V-A ECMO patients in terms of complications [9]. Despite the development of a subsequent prothrombotic state, V-V ECMO is linked to a higher incidence of intracranial hemorrhage compared to V-A ECMO, as evidenced by studies conducted by ELSO [43]. Contrary to some findings suggesting that systemic thromboembolism and bleeding in V-A ECMO patients may elevate the risk of intracranial hemorrhage, and other reports indicate that there is no significant discrepancy in the incidence of intracranial hemorrhage between V-V and V-A ECMO support [44]. But in recent 2022 study, including 358 patients [45], reported 44.7% hemorrhagic complications (26.8% minor and 17.9% major, with the vast majority within the first three days) and 22.9% thrombotic complications (15.6% venous and 11.2% arterial). There was no difference in these events between the groups in regard to ECMO support indication. This study addressed predictors of mortality in ECMO, and a multivariable analysis showed that hemorrhage has a hazard ratio of 1.74 (95% confidence interval 1.24–2.43, p = 0.001) for mortality [45, 46].

Certainly, a patient requiring ECMO is typically in a critical condition, often associated with severe coagulation disorders, thrombocytopenia, and dysregulation caused by the underlying illness [9]. Therefore, we will explore potential causes of thrombocytopenia in emergency critical care patients. Before diagnosing thrombocytopenia attributed to ECMO-related factors, it is crucial to investigate other potential primary or secondary causes in ICU patients. This is especially pertinent for those who already had thrombocytopenia before ECMO initiation.

6.1 Specific characteristics of thrombocytopenia in the ICU

Thrombocytopenia in ICU patients is often complex and multifactorial. While an initial identifiable cause for the drop in platelet count may be apparent, several other contributing factors may coexist [29]. Over time, different reasons for thrombocytopenia may emerge; for instance, drug-induced thrombocytopenia may persist if complications, such as sepsis develop, introducing additional mechanisms for platelet reduction.

6.2 Mechanisms of thrombocytopenia in the intensive care unit (ICU)

From a practical standpoint, thrombocytopenia may stem from five distinct mechanisms: diminished platelet production, increased platelet destruction, enhanced aggregation, dilution, or sequestration [29].

6.2.1 Diminished platelet production

Thrombocytopenia in ICU patients is a complex condition with various potential causes [29]. Bone marrow suppression leading to decreased platelet production can result from drugs (antibiotics, proton pump inhibitors, and diuretics), infections, and nutritional deficiencies such as vitamin B12, folate, and copper deficiencies. Additionally, bone marrow infiltration by metastases or hematological disorders, such as leukemia can contribute to thrombocytopenia. Excessive alcohol intake has a direct toxic effect on megakaryocytes, reducing platelet production.

Viral infections, both acute (e.g., rubella, mumps, varicella, CMV, Epstein-Barr virus) and chronic (e.g., hepatitis C, HIV), can induce thrombocytopenia. Bacterial and fungal infections initially cause thrombocytosis, followed by thrombocytopenia due to platelet aggregation rather than production issues.

6.2.2 Increased platelet destruction

Thrombocytopenia resulting from increased platelet destruction can be attributed to either immune-mediated or nonimmune mechanisms [29]. Immune-mediated thrombocytopenia is characterized by the production of antibodies against platelets, triggered by factors such as viral infections (e.g., EBV, CMV), certain drugs (e.g., heparin-induced thrombocytopenia), posttransfusion reactions, as seen in posttransfusion purpura (PTP) or vasculitides. On the other hand, nonimmune thrombocytopenia stems from physical platelet destruction, which is a mechanism specific to mechanical devices such as ECMO, CPB, hemodialysis, or IABP.

6.2.3 Enhanced platelet aggregation

In bacterial sepsis, thrombocytopenia primarily results from increased platelet aggregation due to activation of the inflammatory response system [29]. Increased platelet consumption is observed in acute and chronic disseminated intravascular coagulation (DIC), thrombotic microangiopathic disorders (e.g., TTP, HUS), HELLP syndrome in pregnant women, or massive pulmonary embolism and thrombotic storm.

6.2.4 Dilution

Dilutional thrombocytopenia is frequently associated with excessive crystalloid/colloid infusions, or in cases of blood loss requiring blood volume replacement with blood products without adequate platelet supplementation. This is also one of the mechanisms that can be encountered in ECMO with dilution of the priming fluid [29].

6.2.5 Sequestration

Platelet sequestration due to splenomegaly is observed in conditions such as cirrhosis, congestive heart failure, portal hypertension, infections, and myeloproliferative disorders. Under normal circumstances, around one-third of the total platelet mass is housed in the spleen, and a proportion that can escalate to 90% in cases of extensive splenomegaly [29]. Patients with this condition exhibit a lower propensity for bleeding despite thrombocytopenia as platelets can promptly enter circulation in response to bleeding and maintain their normal lifespan.

6.3 Approach to diagnosis of thrombocytopenia in intensive care unit (ICU)

See Figure 9.

Figure 9.

Several approaches to the diagnosis of thrombocytopenia in the ICU. ITP, immune thrombocytopenia; HIT, heparin-induced thrombocytopenia, PTP, posttransfusion purpura; PT, prothrombin time; APTT, activated partial thromboplastin time; DIC, disseminated intravascular coagulation [6].

6.3.1 Heparin-induced thrombocytopenia (HIT)

The pathophysiology of HIT (as illustrated in Figure 10) stems from the formation of an immunocomplex comprising an autoantibody targeting platelet factor 4 (PF4)—heparin complex. This complex binds to the surface of platelets and monocytes, triggering their activation via cross-linking of FcgIIA receptors. Activation of platelets and monocytes leads to the onset of severe arterial and venous thrombosis, with a mortality rate of 20%. Without prompt recognition, HIT incidence in the ICU is rare, comprising less than 1% of cases of thrombocytopenia [46, 47]. It often develops after 5–10 days of heparin exposure, with earlier onset in 30% of cases and delayed onset in 10–15%. Diagnosis is often a clinical emergency, relying on clinical findings and pretest probability assessments like the “4Ts” scoring system. Heparin should be stopped, and an alternative anticoagulant initiated until confirmatory test results are available.

Figure 10.

Pathogenesis of HIT. HIT, heparin-induced thrombocytopenia; IgG, immunoglobulin G; PF4, platelet factor 4 (Table 3) [37].

4Ts	Points
2	1	0
Thrombocytopenia	Platelet count decrease > 50% and platelet nadir > 20 × 10^9/L	Platelet count decrease 30-50% or platelet nadir 10–19 × 109/L	Platelet count decrease < 30% or platelet nadir < 10 × 10^9/L
Timing of platelet count decrease	Clear onset between d 5 and 10 or platelet decrease ≤ 1 d (prior heparin exposure within 30 d)	Consistent with d 5–10 decrease, but not clear (eg, missing platelet counts); onset after d 10; or decrease ≤ 1 d (prior heparin exposure 30–100 d ago)	Platelet count decrease < 4 d without recent exposure
Thrombosis or other sequelae	New thrombosis (confirmed); skin necrosis; acute systemic reaction postintravenous unfractionated heparin bolus	Progressive or recurrent thrombosis; non-necroteing erythematous) skin lesions; Suspected thrombosis (not proven)	None
Other causes for thrombocytopenia	None apparent	Possible	Definite

Table 3.

The 4Ts scoring system.

The 4Ts score is the sum of the values for each category: score 1–3 = low HIT probability; score 4–5 = intermediate HIT probability; score 6–8 = high HIT probability [7].

6.3.2 Disseminated intravascular coagulation (DIC)

DIC is a common condition in ICU patients that can be caused by a variety of causes or mechanisms but is particularly common in patients with sepsis or septic shock. Disseminated intravascular coagulation (DIC) is a condition associated with various diseases, marked by the activation of coagulation within blood vessels leading to the formation of microvascular clots. It presents with thrombocytopenia, consumption of clotting factors, diverse bleeding manifestations, and damage to vital organs (Figure 11) [7].

Figure 11.

The pathogenesis of DIC [10].

Tests of hemostasis in DIC, including low platelet count, low fibrinogen concentration, prolonged thrombin time, and high fibrin degradation products, such as D-dimers, are found in serum and urine. Besides, the PT and APTT are prolonged in the acute syndromes [10].

There is no single laboratory test for accurate and precise diagnosis of DIC. The International Society on Thrombosis and Hemostasis (ISTH) committee has proposed a scoring system using the parameters reported in Table 4 in which a score ≥ 5 is consistent with overt DIC.

Parameter	Points
	0	1	2	3
Platelet count	> 100 × 10^9/L	< 100 × 10^9/L	< 50 × 10^9 L
Elevated fibrin degradation products	No increase		Moderate increase	Strong increase
Prothrombin time at upper limit of reference range	< 3 s	> 3 s	> 6 s
Fibrinogen level	> 1 g/L	< 1 g/L

Table 4.

Diagnostic score for the diagnosis of overt DIC [48].

An overall score of five or more is compatible with overt DIC. A score less than five is suggestive of non-overt/low-grade DIC. Adapted trom Taylor et al. [24].

In a study by Hyoung Soo Kim et al., 37 patients undergoing V-V ECMO were investigated, revealing that the pre-ECMO DIC score served as a significant predictor of hospital mortality in individuals with septic shock. Over the course of 48 hours of ECMO support, survivors and non-survivors exhibited notable differences in coagulation profiles, DIC scores, and lactate levels. Moreover, the combined variable of pre-ECMO DIC score and lactate level demonstrated the highest area under the curve (AUC) value, with combined scores exceeding 9.35 potentially indicating an unfavorable hospital outcome [49].

6.3.3 Briefly the mechanism, causes of thrombocytopenia are summarized in the figure below

See Figure 12.

Figure 12.

Factors contributing to thrombocytopenia during extracorporeal membrane oxygenation [30].

7. Treatment

The below tables list general guidelines for a recommended schedule of anticoagulation laboratories and goals for blood product replacement for ECLS. The patient’s clinical condition and underlying pathophysiology should dictate adjustments of laboratory determinations and blood product administration. Laboratory testing can be decreased, and blood product transfusion reduced in patients who have reached a stable clinical status (Tables 5 and 6) [9].

Laboratory Test	Frequency
Activated clotting time (ACT)	Every hour to every 2 hours
Activated partial thromboplastin clotting time (aPTT)	Every 6 hours to every 12 hours
Anti-factor Xa assay	Every 6 hours to every 12 hours
Platelets	Every 6 hours to every 12 hours
International normalized ratio (INR)	Every 12 hours to every 24 hours
Fibrinogen	Every 12 hours to every 24 hours
Complete blood count (CBC)	Every 12 hours to every 24 hours
Antithrombin level	Once per day as needed
Plasma free hemoglobin	Everyday
Thromboelastography/thromboelastometry	Once per day as needed for bleeding or thrombotic complications

Table 5.

Schedule of anticoagulation laboratories in patients undergo ECLS [24].

Goal	Product to Transfuse
Platelets	≥100,000 × 10^9/L (bleeding patient)	Platelets 10 ml/kg (max 2 units)
	≥50,000–100,000 × 10^9/L (nonbleeding patient)
INR	<1.5 (bleeding patient)	Fresh frozen plasma 10 ml/kg (max 2 units)
	<3 (nonbleeding patient)
Fibrinogen	>1.5 g/L (bleeding patient or before surgical intervention)	Cryoprecipitate 1 unit/5 kg (max 6 units)
	>1 g/L (nonbleeding patient)
Hemoglobin	>70–90 g/L (consider higher goal for neonates and children with cyanotic congenital heart disease or lower goal for stable, adult patients)	Packed red blood cells 10 ml/kg (max 2 units)
Antithrombin	>50–80% (>0.5–0.8 U/ml), consider AT replacement if on maximum dose of UFH and unable to obtain anticoagulation goals	AT concentrate: Thrombate III dose (IU) =; [desired AT – current AT] x weight (kg) 1.4

Table 6.

Goals for blood product replacement for ECLS [24].

The platelet test should be performed every 6–12 hours, and the target of platelet count in ECMO-undergo patients with platelet bleeding ≥100,000 × 109/L, and when there is no bleeding, the target is lower (≥50,000–100,000 × 109/L). Platelet products include those manufactured from whole blood (whole blood-derived platelets, random donor platelets, and platelet concentrates) and apheresis (apheresis platelets, single donor platelets, and plateletpheresis) donations. Per AABB Standards, 90% of whole blood-derived platelets must contain >5.5 x 10¹⁰ platelets, while 90% of apheresis platelets must contain >3.0 x 10¹¹ platelets. As a result, 4–6 units of whole blood-derived platelets must be pooled to make a therapeutic dose. In adults, transfusion of an apheresis platelet should increase the platelet count by 30,000-60,000/ml. In neonates and pediatric patients, a 10 ml/kg dose should increase the platelet count by 50,000-100,000/ml [9]. In recent study showed that platelet transfusions were needed in up to 50% of the patients underwent ECMO [30].

Furthermore, the ROTEM test serves as a valuable tool for addressing coagulation disorders in bleeding patients. In 2019, several guidelines were introduced, specifically focusing on the point-of-care management of bleeding, particularly in cardiovascular patients. These guidelines encompassed recommendations for platelet transfusion protocols and the management of heparin and protamine overdoses. In cases of thrombocytopenia observed on ROTEM, specific interventions are recommended based on defined parameters. A5EX = 23–30 mm: 1 pooled or apheresis platelet concentrate. A5EX = 15–22 mm: 2 platelet concentrates. A5EX < 15 mm: 2 platelet concentrates + fibrinogen substitution (Figure 13) [25, 26].

Figure 13.

Evidence-based algorithms for ROTEM (A5)-guided bleeding management in cardiovascular surgery. CT: Coagulation time, A5: Amplitude of clot firmness 5 min after CT, A5_EX: A5 in EXTEM, A5_FIB: A5 in FIBTEM, ACT: Activated clotting time, CT_IN: CT in INTEM; CT_EX: CT in EXTEM, CT_HEP: CT in HEPTEM, CT_FIB: CT in FIBTEM, ML: Maximum lysis (within 1 h run time), PCC: Prothrombin-complex-concentrate, FFP: Fresh frozen plasma [26].

8. Conclusions

Throughout the course of ECMO (extracorporeal membrane oxygenation) therapy, a spectrum of coagulation disorders may manifest, with a prevalent occurrence being a decline in platelet count. This phenomenon often arises from the orchestrated interplay of various physiological mechanisms. Diagnostic modalities encompass the assessment of platelet count through complete blood count or via rotational thromboelastometry (ROTEM) testing. Clinical complications commonly associated with coagulation disorders encompass hemorrhage and thrombosis, both exerting a substantial impact on the overall prognosis of the patient. Hence, meticulous monitoring and proficient management of diminished platelet count assume paramount significance.