The embolic material responsible for MVO is heterogeneous, originating from rupture of upstream unstable atherosclerotic plaque, which not only releases plaque debris but also triggers the formation of platelet-neutrophil aggregates and sheds endothelial-derived microparticles (EMPs) and PMPs into the coronary microcirculation^[10,11]. A clinical study in ST-segment elevation myocardial infarction (STEMI) patients found that intracoronary levels of EMPs and PMPs are elevated and correlate with thrombus burden^[11]. These cellular and subcellular components physically occlude capillaries and arterioles, contributing directly to the “no-reflow” phenomenon^[5].

Endothelial dysfunction is a critical initiating event in the pathogenesis of microthrombosis [Figure 1]^[17]. Upon endothelial injury, the potent procoagulant protein TF is exposed, triggering the extrinsic coagulation cascade. In a study involving clinical cohorts and animal models^[12], elevated levels of trimethylamine N-oxide (TMAO), a gut microbiota-derived and oxidized metabolite, are associated with increased risk of major adverse cardiovascular events (MACE); the animal experiments further demonstrate that TMAO dose-dependently induces expression of TF and vascular cell adhesion molecule-1 (VCAM-1) in human microvascular endothelial cells, thereby enhancing arterial thrombosis potential. This endothelial disruption is often exacerbated by underlying conditions, such as hypertension, which promotes oxidative stress and reduces nitric oxide (NO) bioavailability, further predisposing the microvasculature to a prothrombotic status^[6]. In addition, damaged endothelium releases high mobility group box 1 (HMGB1) as a damage-associated molecular pattern (DAMPs), which further promotes inflammatory signaling (see Section “The HMGB1/TLR4/NF-κB pathway: a key signaling axis” and Figure 1).

Following endothelial injury, platelet adhesion is primarily mediated by the glycoprotein Ib-IX-V complex (GPIb-IX-V) through binding to vWF that anchors on the exposed subendothelial matrix [Figure 2]. This interaction is amplified by high shear stress. In a mouse model of venothrombosis, inflammatory mediators can induce CD40L-mediated release of ultra-large vWF that strings from venous endothelium and avidly binds to platelets^[18]. Subsequent platelet activation is reinforced by signaling pathways involving Toll-like receptor 4 (TLR4) and leads to the conformational activation of integrin αIIbβ3 (the GPIIb/IIIa receptor). Upon activation, this integrin binds to fibrinogen, forming a molecular bridge crosslinking adjacent platelets, thereby consolidating stable platelet aggregation^[19] [Figure 2]. In a mouse model of endotoxemia, Rumbaut et al. demonstrated that lipopolysaccharide (LPS) enhances formation of venular platelet thrombus via a TLR4-dependent mechanism^[20]. Platelet surface P-selectin (CD62P) is rapidly expressed upon activation but is transient and quickly shed. Importantly, hyperactive platelets are a hallmark of prothrombotic conditions like diabetes, significantly amplifying this response^[21].

Upon activation, platelets shed submicron vesicles known as PMPs. Experimental studies show that these vesicles carry platelet surface antigens such as GPIIb/IIIa, P-selectin, and CD40L (see Section “Platelets as bidirectional mediators of inflammation” for detailed signaling mechanisms), and expose phosphatidylserine on their outer leaflet [Figure 2]^[22]. The exposure of phosphatidylserine provides a high-affinity surface for the assembly of tenase and prothrombinase complexes, thereby accelerating thrombin generation and fibrin deposition. Beyond promoting thrombin generation, PMPs themselves can bind directly to fibrin fibers, as shown in an experimental thrombosis model where PMPs generated during thrombus formation were found to associate with local fibrin^[23]. Additionally, PMPs have been shown to increase platelet and fibrin deposition on arterial walls, thereby promoting mural thrombus formation^[24].

As depicted in Figure 3, there exists a multi-step cascade covering endothelial activation, platelet aggregation, neutrophil extracellular trap formation (NETosis), and thrombus formation, in which PMPs greatly amplify the coagulation cascade and consolidate the nascent platelet aggregate into a stable microthrombus. The prothrombotic role of PMPs is supported by human pathological evidence: autopsy studies of severe COVID-19 have identified vWF-rich thrombi together with PMPs in the pulmonary microvasculature, suggesting that PMPs cooperate with endothelial-derived vWF and NETs to propagate microvascular occlusion^[14,25]. Thus, PMPs function not merely as passive by-products of platelet activation but as active participants that promote stable microthrombus formation.

The process of microthrombosis is closely intertwined with MVL, creating a vicious cycle that worsens tissue perfusion. I/R injury damages endothelial cell-cell junctions and degrades the glycocalyx, a surface layer of cross-linked glycoproteins and glycolipids, leading to increased vascular permeability and interstitial edema^[5]. Tissue edema compresses microvessels externally, worsening luminal obstruction from microthrombi and contributing to the “no-reflow” phenomenon. In a mouse cardiac I/R model, the MVL-affected area exceeds the infarction and MVO regions, and MVL severity correlates with LV dilatation and functional decline^[4].

Activated platelets and PMPs directly modulate endothelial barrier function through multiple mechanisms. Shedding from activated platelets, PMPs carry surface molecules such as P-selectin, CD40L, and GPIIb/IIIa that can interact with endothelial cells^[26] (Figure 2, right portion depicting PMP surface molecules). Studies in human umbilical vein endothelial cells found that PMPs dose-dependently promoted tube formation, migration, and proliferation, effects that were mediated by upregulation of matrix metalloproteinase (MMP)-2 and MMP-9 via extracellular signal-regulated kinase signaling, and inhibition of these MMPs abolished the pro-angiogenic and pro-migratory effects of PMPs^[26].

The interaction between platelets and endothelial cells downregulates protein expression of tight junctions (claudin-5, occludin, zonula occludens protein-1) and adherens junctions (VE-cadherin), resulting in paracellular gap formation and hyperpermeability^[5,27]. In cardiac microvascular endothelial cells, I/R injury induces cytoskeletal reorganization and gap junction disassembly, leading to increased permeability and extravasation^[4]. Peptide hormone relaxin has been shown to counteract these detrimental effects: in vitro, it attenuates hypoxia-reoxygenation, H₂O₂, or cytokine-induced endothelial hyperpermeability independently of NO but with preservation of VE-cadherin; in a mouse cardiac I/R model, it reduces Evans blue extravasation by 56% in jeopardized myocardium, decreases leukocyte density and cytokine expression, and partially preserves LV contractile function while limiting chamber dilatation^[28].

Activated platelets also release vasoconstrictors such as thromboxane A2 and endothelin-1, exacerbating microvascular constriction and the “no-reflow” phenomenon^[2]. In the context of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection, pericyte dysfunction plays a critical role in MVL. Using a vascular organoid model, researchers demonstrated that pericytes preferentially take up SARS-CoV-2 viral particles due to high expression of angiotensin-converting enzyme 2 (ACE2), leading to pericyte detachment, damage and cell death, disruption of pericyte-endothelial cell crosstalk, microvascular endothelial hyperpermeability, and enhancement of platelet recruitment and vWF deposition in capillaries^[29]. Endothelial dysfunction in severe COVID-19 is further characterized by decreased expression of CD31 (an endothelial marker) and high expression of vWF, as observed in post-mortem cardiac tissues from 88 patients, directly linking viral infection to a prothrombotic endothelial phenotype^[30].

Platelets are not merely passive targets of inflammatory signals but also actively sense and amplify inflammation through reciprocal interactions with immune cells and the endothelium. Activated platelets function as genuine immune cells, expressing pattern recognition receptors such as TLR4 that can be directly engaged by pathogen-associated molecular pattern (PAMPs) or DAMPs ^[20]. In murine models, endotoxin (LPS) enhanced venular platelet thrombus formation via a TLR4-dependent mechanism, and this effect was independent of neutrophils, demonstrating that platelets can directly sense inflammatory signals without requiring intermediary immune cells^[20]. Upon activation, platelets release a wide array of inflammatory mediators stored in their α- and dense granules, including soluble CD40 ligand (sCD40L), interleukin-1β (IL-1β), platelet factor 4, and thrombospondin-1, which act on endothelial cells and leukocytes to augment inflammation^[31].

Critical to the proinflammatory properties of platelets is platelet-derived CD40L that binds to CD40 receptors on endothelial cells and monocytes, triggering nuclear factor kappa-B (NF-κB)-dependent expression of cytokines, chemokines, and TF, thereby amplifying the inflammatory cascade together with a procoagulant state (Figure 3, left portion depicting CD40/CD40L axis)^[22]. In an experimental study using staphylococcal superantigen-like protein 5-induced platelet microparticles, CD40L on PMPs was shown to directly interact with CD40 on monocytes, leading to activation of the TNF receptor-associated factor 6/NF-κB signaling pathway and a robust production of IL-1β, tumor necrosis factor-α (TNF-α), monocyte chemoattractant protein-1, and MMP-9^[22]. Blockade of CD40L or CD40 significantly reduced these proinflammatory responses, highlighting the CD40/CD40L axis as a key driver of platelet-mediated inflammation (Figure 3, left portion depicting the CD40/CD40L axis).

P-selectin on activated platelets can bind to P-selectin glycoprotein ligand-1 (PSGL-1) on monocytes or neutrophils, forming platelet-leukocyte aggregates^[32]. Genetic deficiency of Mac-1 (CD11b/CD18) or mutation of its binding site on platelet GPIbα markedly delayed thrombosis in murine models without affecting hemostasis, indicating that platelet-leukocyte adhesion is a specific and potentially targetable component of the thromboinflammation cascade^[32]. These platelet-monocyte aggregates (PMAs) promote monocyte differentiation toward a proinflammatory phenotype and enhance the release of MMPs, such as MMP-9, which degrade the extracellular matrix and facilitate vascular remodeling and further inflammation^[33,34]. In patients with type 2 diabetes mellitus, platelet hyperreactivity is particularly pronounced, which contributes to an exaggerated I/R injury response and greater risk of both microvascular and macrovascular complications, underscoring the clinical importance of platelet-mediated inflammation in common metabolic diseases^[7,21].

Beyond these classical pathways, several non-canonical mechanisms have recently been identified to link platelet activation to microthrombosis. In platelets from patients with STEMI, stress-activated protein kinase-interacting protein 1 (Sin1), an essential component of mammalian target of rapamycin complex 2, was found to be phosphorylated at T86, and such phosphorylation correlated with no-reflow phenomena after revascularization^[35]. Importantly, platelet-specific Sin1 deficiency protects mice from ischemia-induced microvascular embolization and subsequent heart dysfunction, and Sin1-mediated outside-in signaling through integrin αIIbβ3 is a critical regulator of platelet-driven microthrombosis^[35]. In a rat model of type 2 diabetes, fibrinogen-like protein 2 is highly expressed in cardiac microvascular endothelial cells. Its upregulation is switched on by local or circulating TNF-α and is associated with cardiac hyaline microthrombosis and impaired diastolic and systolic function^[36]. In Kawasaki disease, thymic stromal lymphopoietin has been identified as a platelet-activating cytokine that induces mitophagy via a novel pathway and promotes thrombosis^[37], representing another non-canonical pathway by which inflammatory mediators directly drive platelet-mediated microvascular injury.

Circulating platelets are not a homogeneous population but exhibit considerable functional and maturational heterogeneity^[38]. Based on functional characteristics, platelets can be categorized into procoagulant, aggregatory, and secretory subtypes, each contributing distinctly to thrombus formation and inflammatory modulation^[38]. Among these, procoagulant platelets are particularly relevant to microthrombosis. Marked by sustained Ca²⁺ elevation and surface exposure of phosphatidylserine, procoagulant platelets provide a catalytic platform for tenase and prothrombinase complexes, thereby accelerating thrombin generation and fibrin deposition at sites of microvascular injury^[39]. This procoagulant phenotype positions these cells at the dynamic intersection between coagulation and inflammation, as their activation is potently modulated by inflammatory mediators, and they in turn release signals that perpetuate endothelial dysfunction^[39].

Platelets can also be classified by maturational stage into reticulated (immature) and non-reticulated (mature) platelets. Reticulated platelets (RPs)—newly released from megakaryocytes and rich in residual messenger RNA—exhibit a prothrombotic transcriptome with enrichment of receptors critical for platelet activation and aggregation^[40], and demonstrate heightened propensity for arterial thrombus incorporation under high-shear conditions comparable to those of stenotic coronary arteries^[41] (Figure 2, left portion depicting GPIb-IX-V/vWF interaction under high shear stress). Clinical studies have revealed that elevated RP content correlates with increased platelet reactivity and reduced efficacy of standard dual antiplatelet therapy in patients with CAD^[42], and that the immature platelet fraction functions as an independent long-term predictor of MACE in ACS patients^[43]. Multi-omic profiling of RPs has further identified selectively enriched glycoprotein VI (GPVI) and phosphatidylinositol 3-kinase (PI3K) signaling pathways, whose pharmacological inhibition effectively reduces RP-mediated aggregation and thrombus recruitment^[44].

The inflammatory mediators released by activated platelets, together with other inflammatory cytokines such as IL-1β and TNF-α, activate NF-κB in endothelial cells, leading to the upregulation of adhesion molecules including intercellular adhesion molecule-1 (ICAM-1), VCAM-1, and P-selectin^[6] (Figure 1, right portion showing NF-κB-driven upregulation of VCAM-1 and ICAM-1). This inflammatory activation creates a pro-adhesive surface that captures circulating platelets and leukocytes, initiating the thromboinflammatory cascade. TMAO directly promotes this prothrombotic transformation by dose-dependently inducing expression of TF and VCAM-1 in human microvascular endothelial cells. Notably, this effect persists despite antiplatelet therapy, indicating that TMAO-enhanced thrombosis is mediated at least in part through endothelial pathways beyond platelet inhibition^[12].

Hypertension and other metabolic risk factors exacerbate endothelial dysfunction through multiple molecular mechanisms. Increased reactive oxygen species (ROS) production via NADPH oxidase activation reduces NO bioavailability, uncouples endothelial nitric oxide synthase (eNOS), and promotes angiotensin II-mediated vasoconstriction, all of which shift the endothelium toward a proinflammatory and prothrombotic phenotype^[6]. The loss of normal antithrombotic and anti-inflammatory functions of endothelial cells leads to dysregulation of coagulation, complement activation, and platelet aggregation in the microvasculature, a state now recognized as central to thromboinflammation across a broad range of disorders including sepsis, I/R injury, and the antiphospholipid syndrome^[45]. In patients with antiphospholipid syndrome as an autoimmune disorder, persistent endothelial dysfunction, accelerated intimal hyperplasia, and secretion of inflammatory products are hallmark features that predispose to microvascular thrombosis and cardiovascular events^[46].

Beyond classical oxidative stress pathways, iron metabolism has emerged as a key regulator of inflammation and thrombosis. Iron overload promotes ROS generation via the Fenton reaction, activating NF-κB and exacerbating endothelial dysfunction^[47]. In this context, ferroptosis—an iron-dependent, lipid peroxidation-driven cell death—disrupts endothelial integrity, promotes phosphatidylserine externalization and expression of TF, thereby amplifying prothrombotic signals. Use of ferroptosis inhibitors restores endothelial homeostasis in thrombotic models^[48,49].

Neutrophils recruited to sites of microvascular injury undergo a specialized form of cell death called NETosis, releasing NETs composed of decondensed DNA, histones, and granule proteins^[17]. Acting as a physical scaffold, NETs capture and directly activate platelets, promote platelet adhesion and aggregation, and initiate the intrinsic coagulation pathway via factor XII, ultimately leading to enhanced fibrin deposition (Figure 3, lower right area depicting NETs and fibrin scaffold). In mouse models of myocardial I/R injury, NETs were identified within the microvasculature of the ischemic zone, and treatment with DNase to degrade NETs reduced MVO and improved myocardial perfusion^[50]. In sepsis and I/R injury models, similar benefits have been observed with DNase treatment, highlighting the therapeutic potential of targeting NETs in thromboinflammatory diseases^[45].

There is strong clinical evidence for the association between NETosis and microthrombosis across diverse inflammatory conditions. In severe COVID-19, post-mortem analyses have demonstrated NET-rich thrombi occluding microvessels, with vWF and NETs synergistically contributing to thrombotic microangiopathy^[14]. This thromboinflammatory state extends to post-acute sequelae of COVID-19 (long COVID), where ongoing NET-driven pathology is strongly indicated by persistent neutrophil activation, elevated platelet-neutrophil aggregates, and circulating microclots^[51]. In addition to NETosis, SARS-CoV-2 infection directly amplifies endothelial injury through viral invasion of ACE2-expressing cells, triggering inflammation, microthrombosis, and angiogenesis^[52]. Consistent with this, post-mortem studies of COVID-19 patients consistently demonstrate small coronary vessel microthrombosis alongside a systemic inflammatory infiltrate dominated by CD3+/CD8+ cytotoxic lymphocytes and CD68+ macrophages^[53], underscoring the broad cardiovascular complications of this thromboinflammatory state^[54]. Beyond COVID-19, NETs have been implicated in atrial fibrillation—where levels of NETs in left atrial blood samples correlate with those of vWF and cardiac injury markers^[55]—as well as in vaccine-associated thrombosis, trauma-induced hypercoagulability, and complement-mediated thrombotic microangiopathy, where NETs cooperate with vWF and activated complement to promote MVO^[56-58]. In the failing heart due particularly to ischemic cardiomyopathy, endothelial vWF expression and Weibel-Palade bodies are markedly increased and correlate with coronary microthrombi^[59].

HMGB1 is a nuclear protein that functions as a DAMP when released from injured, necrotic, or activated cells. Once in the extracellular space, HMGB1 triggers robust inflammatory responses by binding to TLR4 and the receptor for advanced glycation end products (RAGE)^[60,61] (Figure 3, left-hand pathway depicting HMGB1/TLR4/NF-κB signaling). In a rat model of CME induced by LV injection of 4000 polyethylene microspheres (45 μm diameter) into the LV cavity, HMGB1 expression was significantly upregulated and translocated from the nucleus to the cytoplasm, accompanied by increased expression of TLR4 and NF-κB and a marked decrease in autophagic activity in myocardial tissue (predominantly in cardiomyocytes) as reflected by reduced LC3BII/LC3BI and Beclin-1 protein levels^[13].

Yang et al. reported in a CME model that pharmacological inhibition of HMGB1 using glycyrrhizin administered 30 min before CME markedly reduced HMGB1 nuclear export, suppressed TLR4 and NF-κB expression, and significantly restored autophagic activity^[13]. This restoration of autophagy was associated with substantial cardioprotection, including improved left ventricular ejection fraction (LVEF) and reduced microinfarct size (see Section “Myocardial level: microinfarction, contractile dysfunction, and biomarker elevation” for details). Critically, these protective effects were completely abrogated by pretreatment with the autophagy inhibitor 3-methyladenine, demonstrating that the beneficial actions of HMGB1 inhibition are mediated through restoration of autophagic flux rather than direct anti-inflammatory effects alone^[13] (Figure 3, left-hand pathway depicting HMGB1/TLR4/NF-κB signaling). It should be noted that the evidence for HMGB1/TLR4/NF-κB-mediated autophagy suppression in this model is derived from whole myocardial tissue; hence, it is likely that cardiomyocytes are the primary cell type for the observed autophagic changes^[13,62]. However, recent studies have established that HMGB1-TLR4 signaling is functionally active in microvascular endothelial cells, inducing proinflammatory and procoagulant phenotypes^[63], and that HMGB1 can regulate autophagy in cardiac microvascular endothelial cells under stress conditions^[64]. It remains to be determined whether HMGB1/TLR4/NF-κB directly suppresses autophagy in cardiac microvascular endothelial cells, and whether the observed effects are mediated indirectly via paracrine signals from neighboring cells. Endothelial-specific genetic deletion of HMGB1 or TLR4 in future studies will be essential to confirm the cell-autonomous function of this signaling axis and to dissect direct versus indirect regulatory mechanisms. In the failing human heart—independent of COVID-19—microthrombi are increased 7- to 9-fold compared to control hearts, and this is associated with a 3.5- to 5-fold increase in Weibel-Palade bodies and vWF expression, especially in ischemic cardiomyopathy. PECAM-1 is also upregulated but apparently insufficient to compensate for the prothrombotic shift^[59]. These findings underscore that the HMGB1-mediated inflammatory pathway and vWF-related prothrombotic pathway are conserved mechanisms of microthrombosis across different etiologies of heart failure^[59,60].

In macrophages, HMGB1 is known to promote the secretion of matrix vesicles via the RAGE/p38 MAPK/neutral sphingomyelinase 2 pathway, leading to ectopic mineralization and further vascular dysfunction^[65]. In RAW264.7 macrophages, HMGB1 stimulation significantly increased secretion of matrix vesicles that exhibited higher activity of tissue non-specific alkaline phosphatase and greater potential to initiate mineralization when tested in collagen-coated dishes or subcutaneous injection models in vivo^[65]. These findings indicate that HMGB1’s prothrombotic and proinflammatory roles extend beyond the immediate injury phase to include long-term vascular remodeling and calcification, the processes known to exacerbate atherosclerosis progression and heart failure.

The interdependent cycle of inflammation, platelet activation, endothelial barrier disruption, and thrombosis thus represents a central therapeutic target (see also the reciprocal interactions between microthrombosis and MVL in Section “Interactions of microthrombi with microvascular leakage”). Interruption of any node in this cycle - whether by antiplatelet agents (aspirin, P2Y12 inhibitors, GPIIb/IIIa antagonists), anti-inflammatory biologics (IL-1/IL-6 inhibitors, HMGB1 inhibitors such as glycyrrhizin), or targeted inhibition of neutrophil NETosis - holds promise for mitigating microvascular damage in AMI, COVID-19, and other thromboinflammatory disorders.

Microthrombosis directly contributes to myocardial injury by forming microinfarcts characterized by small, discrete areas of cardiomyocyte necrosis resulting from occlusion of the coronary microvasculature. In animal models of CME, this histologic damage translates into profound contractile dysfunction—marked by a substantial reduction in LVEF and progressive LV dilatation—along with elevated serum cardiac troponin I levels, confirming ongoing cardiomyocyte necrosis^[13]. Notably, HMGB1 inhibition with glycyrrhizin significantly attenuates both microinfarct burden and LV functional decline, whereas co-administration of the autophagy inhibitor 3-methyladenine abrogates these protective effects, indicating that autophagy restoration is a critical downstream mediator of HMGB1-driven myocardial injury^[13]. Using the same rat CME model, Su et al. further demonstrated that oxidative stress induced by USP16 S-nitrosylation disrupted glutathione homeostasis, and that blocking this modification alleviated cardiac dysfunction^[66]. In line with these preclinical findings, clinical observations show that the degree of contractile dysfunction correlates with the extent of MVO in STEMI patients undergoing primary PCI^[67].

Biochemical markers of myocardial injury, particularly cardiac troponin I, are significantly elevated in the setting of microthrombosis. In the CME rat model, serum cardiac troponin I levels increased markedly compared to sham controls, reflecting ongoing cardiomyocyte necrosis^[13]. In patients with STEMI, the presence of MVO detected by cardiac magnetic resonance (CMR) is associated with higher peak troponin levels, larger infarct size, and worse long-term clinical outcomes, including a higher incidence of MACE^[1,68]. Moreover, intracoronary levels of PMPs and EMPs are positively correlated with thrombus score and corrected TIMI frame count (a measure of coronary blood flow velocity) and independently associated with angiographic and electrocardiographic MVO, suggesting that microparticles not only serve as markers but may also directly participate in the pathogenesis of MVO^[11].

Thromboinflammatory processes that drive microthrombosis in the heart are not organ-specific, and similar mechanisms operate in other vascular beds, leading to multi-organ dysfunction in systemic conditions. Table 2 summarizes detailed mechanisms involved across organs. Exemplifying this systemic nature is Kawasaki disease, in which ultrastructural studies of endomyocardial biopsies have directly demonstrated platelet aggregation with thrombosis in coronary microvessels, along with endothelial injury, microvascular dilatation, and luminal stenosis—changes that persist for years after disease onset and correlate with myocardial fibrosis and dysfunction^[15]. These microvascular lesions are accompanied by systemic vascular dysfunction, including reduced coronary flow reserve and persistent endothelial impairment even in patients without coronary aneurysms^[16,69]. The persistent platelet activation state in Kawasaki disease—now recognized as a paradigm shift from merely increased platelet counts to functional activation^[70]—drives thromboinflammatory cascades through platelet-leukocyte aggregates, NLRP3 inflammasome activation, and IL-1β production, directly linking platelet biology to vascular injury and microthrombosis^[71].

In the kidney, hypertension induces microvascular remodeling and rarefaction, characterized by a decreased vessel density and increased wall-to-lumen ratio in resistance arterioles^[6]. These structural changes, combined with endothelial dysfunction and activation of the renin-angiotensin system, lead to glomerular hyperfiltration, proteinuria, and progressive chronic kidney disease. Microthrombi within the renal microcirculation have been documented in severe COVID-19 autopsies, contributing to the onset of acute kidney injury^[72]. In the brain, hypertension is a major risk factor for cerebral small vessel disease, which is characterized on MRI as white matter hyperintensities, lacunas, microhemorrhages, and microinfarcts^[6]. These microvascular lesions, driven by lipohyalinosis and fibrinoid necrosis of small arterioles, are associated with cognitive decline, dementia, and an increased risk of ischemic and hemorrhagic stroke. In the retina, hypertensive retinopathy - marked by increased vascular permeability, microhemorrhages, and cotton-wool spots - serves as a visible indicator of systemic microvascular damage and predicts cardiovascular and renal complications^[6].

Activation of the complement pathway differentially contributes to microvascular versus macrovascular thrombosis. In patients with complement-mediated thrombotic microangiopathy, higher urinary sC5b-9 levels and renal C5b-9 deposition correlate with renal microthrombosis, while plasma levels of C5a and renal expression of C5aR1 are associated with macrovascular thrombosis, suggesting that targeting individual complement components could yield more precise antithrombotic effects^[58]. In primary antiphospholipid syndrome, pulmonary microthrombosis, stroke and deep vein thrombosis are common clinical manifestations, and microvascular involvement frequently underlies multiorgan failure in the catastrophic variant^[73,74].

Pre-existing conditions such as hypertension, diabetes mellitus, hyperhomocysteinemia, and dyslipidemia profoundly amplify the risk and severity of microthrombosis.

Hypertension, as detailed in Section “Inflammation-induced endothelial dysfunction”, promotes endothelial dysfunction, vascular remodeling, and rarefaction, all of which predispose to microthrombosis. CD40 knockout mice developed profound pulmonary microthrombosis when subjected to pressure overload induced by transverse aortic constriction (TAC)^[75]. This prothrombotic phenotype was not attributable to differential LV dysfunction, as CD40 deficiency did not affect LV remodeling and function after TAC. Rather, it was accompanied by enhanced pulmonary oxidative stress, endothelial activation (upregulation of VCAM-1 and ICAM-1), and platelet hyperresponsiveness (augmented aggregation to thrombin, collagen, and ADP), indicating a direct prothrombotic effect of CD40 loss on the pulmonary microvasculature under pressure overload^[75].

Diabetes mellitus induces a state of platelet hyperreactivity, endothelial dysfunction, and impaired fibrinolysis, collectively known as a prothrombotic milieu^[7,8]. Diabetic patients have elevated levels of circulating PMPs and EMPs, which correlate with the presence and severity of microvascular complications such as retinopathy and nephropathy^[76,77]. In myocardial I/R injury, diabetes worsens microvascular damage through mechanisms involving enhanced oxidative stress, impaired mitophagy, and reduced endothelial barrier function^[21,78]. Clinically, diabetes has also been identified as an independent predictor of MVO after primary PCI^[79].

Hyperhomocysteinemia is a further underappreciated risk factor for coronary microvascular dysfunction due to several molecular mechanisms. First, hyperhomocysteinemia impairs annexin A2-dependent fibrinolysis: in hyperhomocysteinemic mice, tPA-dependent plasmin activation, an annexin A2-dependent process, is abolished, leading to perivascular fibrin accumulation and impaired angiogenesis^[80]. Second, during cardiac I/R, hyperhomocysteinemia synergizes with mobilized copper ions (Cu²⁺) to induce necroptosis of cardiac microvascular endothelial cells via peroxynitrite generation from NADPH oxidase and eNOS^[81]. Third, hyperhomocysteinemia activates the lncRNA MEG3/microRNA-223/NLRP3 inflammasome axis, leading to endothelial cell pyroptosis with reduced microvascular density^[82]. Fourth, hyperhomocysteinemia increases oxidative stress and thromboxane A₂ activity while reducing NO bioavailability, promoting vasomotor dysfunction and a prothrombotic state^[9]. Finally, hyperhomocysteinemia-induced microvascular dysfunction after AMI can be reversed by a protective transcriptional axis driven by the transcription factor SP1 (specificity protein 1), which activates STAT3/KCa3.1/eNOS signaling^[83].

Dyslipidemia is an important modifiable risk factor that predisposes to microthrombosis. Dyslipidemia induces a proinflammatory and prothrombotic phenotype through enhanced production of free radicals, leukocyte recruitment, and platelet adhesion^[84]. Oxidized low-density lipoprotein (oxLDL) directly damages the endothelial monolayer, facilitates thrombocyte adhesion, and promotes the formation of microthrombi. Moreover, hypercholesterolemia contributes to platelet hyperreactivity via direct oxLDL-platelet membrane interaction through scavenger receptors such as CD36, and oxLDL has been shown to induce platelet activation and arterial thrombosis^[85]. Among lipid abnormalities, lipoprotein(a) has recently emerged as an independent risk factor for coronary microvascular dysfunction. Acting as a carrier of oxidized phospholipids, elevated lipoprotein(a) promotes endothelial dysfunction, oxidative stress, and microthrombotic susceptibility through its antifibrinolytic effects^[86]. These findings suggest that lipoprotein(a) screening may be valuable in patients with microvascular angina. (Table 3 provides a summary of the shared and unique molecular pathways of these risk factors involved).

In summary, underlying diseases, as specified above, amplify microthrombosis through shared pathways of oxidative stress, endothelial dysfunction, platelet activation, and impaired endogenous protective mechanisms, making them critical determinants of clinical outcomes in patients with AMI, heart failure, and systemic inflammatory conditions.

Because PMPs are released by activated platelets and can be quantified in peripheral blood, they have been extensively investigated as circulating biomarkers of microthrombosis^[87]. Flow cytometric enumeration of PMPs (defined as CD41-positive, annexin V-positive particles < 1 μm) allows assessment of in vivo platelet activation and thrombogenicity. Studies on STEMI patients have consistently shown that intracoronary levels of PMPs exceed those in the systemic circulation and correlate with local thrombotic activity, as reflected by MVO indices (myocardial blush grade ≤ 2 and incomplete ST-segment resolution (STR)) and by SYNTAX score in those with ACS^[11,88]. A meta-analysis further confirmed that plasma PMP concentrations are significantly elevated in ACS patients compared to those with stable angina or healthy controls^[89], and elevated baseline PMP levels correlate with increased MACE risk during long-term follow-up, encompassing death, recurrent myocardial infarction (MI), and heart failure hospitalization. Elevated PMP levels have also been reported in hypertension^[90], diabetes^[76,77], and COVID-19^[91], where they correlate with disease severity and adverse outcomes. Despite their promise, routine clinical adoption of PMP quantification is hampered by pre-analytical variability, lack of standardized reference ranges, and limited sensitivity of conventional flow cytometers^[92].

Beyond PMPs, other platelet-related parameters have demonstrated prognostic utility. Admission mean platelet volume (MPV), a surrogate of platelet size and reactivity, independently predicts Killip class ≥ 2^[93]. Mechanistically, AMI triggers the splenic release of larger, more reactive platelets; splenectomy attenuates the rise in MPV, reduces cardiac platelet and leukocyte accumulation, and limits the infarct size^[94]. In patients with coronary artery calcification, the PDW/HDL-C ratio (where PDW reflects platelet size heterogeneity) has been proposed as a simple, inexpensive prognostic index^[95]. Additionally, immature platelet fraction, which quantifies circulating RPs, has emerged as an independent long-term predictor of MACE in a 5-year ACS cohort^[43], supporting its utility for risk stratification in microthrombosis.

Upon activation, platelets secrete or shed soluble forms of membrane proteins that can be measured in plasma or serum. sP-selectin (also known as sCD62P) is derived mainly from activated platelets and is elevated in ACS, atrial fibrillation, and peripheral artery disease. In a prospective cohort of 733 patients undergoing PCI, elevated sP-selectin levels were associated with a significantly increased risk of long-term MACE after a mean follow-up period of 9.7 years^[96]. Furthermore, in patients with coronary chronic total occlusion (CTO) undergoing technically challenging PCI, higher Japanese CTO scores and longer procedure duration were associated with significantly elevated platelet P-selectin positivity 2 days after the intervention and increased sP-selectin levels 3-6 months later, indicating sustained platelet activation^[97]. Similarly, sCD40L is predominantly released from activated platelets and has potent proinflammatory properties by inducing endothelial cells to express adhesion molecules and TF^[46]. In the setting of CME and I/R, increased plasma levels of sCD40L correlate with platelet activation and MVO. Consistent with this, a proof-of-concept study in STEMI patients showed that higher levels of sCD40-L, hs-CRP, and VCAM-1 within aspirated coronary thrombi were associated with MVO^[79]. Thrombospondin-1, another α-granule protein, is involved in platelet aggregation, angiogenesis, and TGF-β activation, and has been explored as a biomarker of adverse cardiac remodeling. While these soluble markers are easier to measure by ELISA and are commercially available, they are less specific for microthrombosis than cell-associated or microparticle-based markers because they can be released by other cells and are affected by renal function and sample handling.

PMAs are a sensitive indicator of in vivo platelet activation and cell-cell crosstalk. Both cell types form primarily through binding via platelet P-selectin to monocyte PSGL-1, with additional stabilization by Mac-1 (CD11b/CD18) on leukocytes interacting with platelet GPIbα^[32]. Elevated PMAs are present in patients with ACS, diabetes, and post-coronary stenting, and they have been linked to a prothrombotic and proinflammatory phenotype. In patients with COVID-19, massive image-based single-cell profiling revealed the presence of excessive circulating platelet-neutrophil aggregates and PMAs in nearly 90% of patients, and the concentration of these aggregates was strongly associated with disease severity, mortality, respiratory failure, and markers of endothelial dysfunction^[98]. Platelet-neutrophil aggregates are similarly elevated in STEMI, sepsis, and severe COVID-19 and are being explored as flow cytometric biomarkers for NETs-mediated microthrombosis, although pre-analytical challenges remain to be addressed, particularly the need for immediate sample processing to avoid spontaneous in vitro aggregate formation (see Sections “Platelets as bidirectional mediators of inflammation” and “Neutrophil extracellular traps promote platelet aggregation and fibrin deposition” for mechanistic details). The persistence of platelet-leukocyte aggregates has also been documented in long COVID, suggesting ongoing thromboinflammation^[51]. The time course of PMA formation is rapid. In AMI patients, circulating PMAs increase as early as 1-3 h post-MI and return to baseline within 24 h; antiplatelet therapy with clopidogrel or P2Y12 inhibitors effectively reduces PMA formation, leukocyte infiltration into the infarcted myocardium, and expression of inflammatory mediators^[94,99]. From a clinical perspective, detection of PMAs by flow cytometry is more robust than single platelet activation markers, as the larger aggregates facilitate gating and reduce the confounding effects of platelet fragmentation. Importantly, this advantage does not obviate the need for careful sample handling to prevent in vitro aggregation artifacts. The requirement for immediate sample processing and specialized equipment still limits its routine use.

Beyond flow cytometric enumeration of PMAs, recent work has leveraged platelet transcriptomics to capture the thromboinflammatory axis at a molecular level. In a longitudinal cohort study^[100], Beitzen-Heineke et al. defined high vs. low monocyte platelet aggregates (MPA) status based on consistently elevated MPA levels over two time points^[100]. High MPA individuals showed increased expression of platelet activation markers (P-selectin, PAC-1, CD40L) and enriched immune signaling pathways. Using the platelet transcriptome, the authors derived a 42-gene thromboinflammation platelet signature (TIPS) that correlated strongly with MPA levels, was reproducible over time, and was significantly elevated in patients with COVID-19 and in those with MI. Importantly, in a prospective cohort of patients undergoing lower extremity revascularization (median follow-up 18 months), higher TIPS independently predicted future cardiovascular events after adjusting for age, sex, race, and ethnicity. Notably, TIPS was modifiable by ticagrelor but not by aspirin. These findings extend the concept of PMAs from a cellular aggregate to a transcriptomic signature, offering a novel tool for cardiovascular risk stratification and highlighting the potential of platelet RNA-based diagnostics.

Flow cytometric assessment of platelet surface receptor expression provides a direct snapshot of platelet activation status (for the mechanistic roles of these receptors, see Section “Platelet adhesion, activation, and aggregation”). Activated GPIIb/IIIa, detected by PAC-1 antibody that recognizes the active conformation of integrin αIIbβ3, is a highly specific marker of ongoing platelet aggregation. In patients with ACS, early administration of eptifibatide (a GPIIb/IIIa antagonist) rapidly inhibited PAC-1 binding and ADP-induced aggregation, but did not affect CD62P or platelet-leukocyte aggregate formation, indicating that different activation pathways are not equally blocked by a single antiplatelet agent^[101]. The clinical relevance of GPIIb/IIIa inhibition as a strategy to reduce MVO was subsequently evaluated in the REVERSE-FLOW trial (see Section “Antiplatelet therapy” for detailed results and discussion). Platelet surface P-selectin (CD62P) is another standard activation marker, but it is transiently expressed and rapidly shed, making it less reliable than PAC-1 for detecting ongoing aggregation. TLR4 expression on platelets is upregulated in inflammatory states. In a study of 40 MI patients and 40 matched controls, Barillà et al. demonstrated that platelet TLR4 is activated in MI patients, as evidenced by increased TIRAP phosphorylation and TIRAP-MyD88 interaction, and that circulating LPS may serve as a trigger for this activation; ex vivo inhibition of TLR4 reduced collagen-induced platelet aggregation, suggesting that TLR4 contributes to platelet hyperreactivity after MI^[102]. Although flow cytometry-based assays for these surface markers are highly informative in research settings, their clinical application is hindered by the need for rapid sample processing, antibody panels, and technical expertise.

Despite their promise, the clinical translation of platelet-related biomarkers is hindered by profound pre-analytical variability. Sampling timing is critical: PMAs peak at 1-3 h post-MI and return to baseline within 24 h^[94], while MPV decreases by approximately 16% within the first 24 h^[93]. Sampling site also matters, as intracoronary levels of PMPs and EMPs are significantly higher than peripheral venous levels^[11,88]. Processing variables, including anticoagulant type, centrifugation speed, and time to analysis, further affect reproducibility^[103]. These factors, combined with the lack of standardized protocols across studies, have prevented the establishment of uniform clinical detection criteria and validated reference ranges.

To address these barriers and facilitate clinical application, the following strategies are recommended. (1) Standardized pre-analytical protocols should be adopted, specifying anticoagulant (citrate preferred), centrifugation protocol (e.g., 2,500 g for 15 min twice for platelet-free plasma), and time to analysis (fresh samples preferred)^[103]; (2) Stratified reference ranges should be established for each biomarker, separately defined by sampling site (intracoronary vs. peripheral) and sampling time relative to symptom onset and reperfusion, using large, well-characterized cohorts; (3) MVO risk stratification could categorize patients into low-, intermediate-, and high-risk groups. This can be achieved by developing risk prediction models that incorporate multiple biomarkers and potentially integrate other readily available parameters such as D-dimer to provide individualized risk scores (e.g., via logistic regression with nomogram presentation). Prospective validation in multicenter cohorts will be essential to confirm the clinical utility of these biomarkers and define their role in clinical decision-making.

Antiplatelet agents are the cornerstone of pharmacological intervention against microthrombosis in ACS and following PCI. Aspirin irreversibly inhibits cyclooxygenase-1, reducing thromboxane A2 production and attenuating platelet aggregation. P2Y12 inhibitors (clopidogrel, prasugrel, ticagrelor) block the ADP receptor, thereby suppressing platelet activation. In a randomized, blinded trial of ACS patients undergoing stenting, prasugrel and ticagrelor achieved stronger platelet inhibition than clopidogrel and were associated with transient improvements in microvascular reactive hyperemia measured by laser Doppler, suggesting a direct beneficial effect on microvascular function^[104]. However, this microvascular effect was lost one day after stenting, indicating that sustained protection may require more prolonged or additional strategies. GPIIb/IIIa antagonists (e.g., eptifibatide, abciximab) block the final common pathway of platelet aggregation. Early administration of eptifibatide rapidly and profoundly inhibited platelet aggregation and reduced the expression of activated GPIIb/IIIa (PAC-1 binding) as well as CD51/61 and CD42b^[101]. Nevertheless, platelet-leukocyte aggregate formation continued to rise within 24 h despite eptifibatide therapy, suggesting that GPIIb/IIIa blockade alone does not fully disrupt the inflammation-platelet cycle. The clinical impact of GPIIb/IIIa inhibition on MVO was evaluated in the REVERSE-FLOW trial^[105], which randomized 120 AMI patients with angiographic MVO after PCI to bailout GPIIb/IIIa inhibitor (eptifibatide or tirofiban) versus optimal medical therapy. Although the primary endpoint of the infarct size was not reduced, the treatment group showed significantly fewer patients with CMR-derived MVO (74.5% vs. 92.2%; P = 0.017) and a smaller MVO extent (2.1% vs. 3.4% of the LV; P = 0.025), but with an increased risk of non-life-threatening bleeding (22.6% vs. 6.9%; P = 0.016). This dissociation, with improvement in a mechanistic surrogate without translating into infarct size limitation, might be attributable to a few reasons. GPIIb/IIIa inhibitors were administered as bailout therapy after reperfusion, a time point at which the cellular events driving cardiomyocyte necrosis had already occurred. MVO reflects microvascular perfusion defects rather than myocardial necrosis per se, and improvement in MVO may not reverse established infarct size. Indeed, in a mouse model of cardiac I/R of 1 h/24 h, the size of MVO was much smaller than that of the infarct size (36% vs. 47%) of the risk zone^[4]. Further, the increased bleeding risk may partially offset the microvascular benefits. Thus, antiplatelet strategies aimed at myocardial salvage may require earlier administration or combination with agents targeting the inflammation-platelet axis, rather than bailout inhibition after MVO has already developed. Current guidelines recommend dual antiplatelet therapy for secondary prevention of arterial thrombosis in ACS; however, direct clinical evidence remains lacking on reduction in MVO or MVL by specifically antiplatelet regimens. Beyond their antithrombotic action, antiplatelet drugs may exert additional anti-inflammatory benefits: in a mouse MI model, clopidogrel, but not aspirin, significantly reduced cardiac rupture incidence (from 45% to 10%) and decreased intramural platelet thrombus formation, accompanied by suppressed systemic and cardiac inflammatory responses^[99,106].

Another potential therapeutic avenue targets the splenic release of platelets. In a study by Gao et al., AMI triggered the release of larger, more reactive platelets from the spleen; splenectomy attenuated the rise in MPV, reduced circulating platelet-monocyte conjugation as well as cardiac accumulation of platelets and leukocytes, limited infarct size, and alleviated cardiac dilatation and dysfunction^[94]. Importantly, interventions that are not directly antiplatelet - such as β-blockers or ACE inhibitors - also provided cardioprotection, suggesting that modulating splenic platelet release could be a novel therapeutic strategy^[94].

Looking forward, novel antiplatelet agents targeting platelet receptors distinct from conventional pathways are being explored. A recent review highlights potential benefits of inhibitors against GPVI, protease-activated receptor 4, GPIb-IX-V, and serotonin receptor 2A, as well as P-selectin inhibitors, which have shown efficacy in reducing microvascular thrombosis in sickle cell disease and could translate into benefits in ACS^[107]. Notably, patients with elevated RPs may derive particular benefit from these emerging strategies, as RPs show reduced responsiveness to P2Y12 inhibitors and aspirin, correlating with persistent thromboinflammatory activity despite dual therapy^[42]. Multi-omic profiling of RPs has further identified selectively enriched GPVI and PI3K signaling, while pharmacological inhibition (e.g., glenzocimab or PI3K blockade) effectively reduces RP-mediated aggregation and thrombus recruitment^[44]. However, translating these insights into clinical practice must contend with the persistent bleeding risk that undermines the utility of intensified antiplatelet regimens, particularly in patients at high bleeding risk^[45]. In this context, the natural compound formononetin has recently been shown to inhibit platelet activation and NETosis via the CD36/extracellular signal-regulated kinase 5 pathway, and attenuate myocardial I/R injury without prolonging bleeding time, suggesting a potentially safer antiplatelet strategy^[108].

Given the central role of inflammation in driving microthrombosis, targeting inflammatory cytokines and their signaling pathways has emerged as a promising therapeutic approach. In the ASSAIL-MI trial, a double-blind, randomized study of 199 STEMI patients, intravenous administration of tocilizumab (an interleukin-6 receptor inhibitor) prior to primary PCI improved the myocardial salvage index measured by CMR compared to placebo^[109]. However, a subsequent sub-study found that tocilizumab did not significantly affect circulating levels of endothelial- or platelet-derived CXC-chemokines (CXCL4, CXCL5, CXCL6, CXCL7, CXCL12), and none of these chemokines were associated with the infarct size, MVO, or troponin T levels^[110]. Notably, heparin administration in this trial significantly influenced chemokine levels, highlighting a potential confounding factor in biomarker studies. These findings suggest that the beneficial effects of tocilizumab in STEMI are not mediated through changes in these specific CXC-chemokines, and the exact mechanisms remain to be elucidated. Mechanistic sub-studies have revealed that tocilizumab attenuates neutrophil counts and suppresses the formation of NETs^[111], and reshapes monocyte transcriptome profiles toward a cardioprotective phenotype, potentially through downregulation of SOCS3, with in vitro evidence showing direct limitation of cardiomyocyte apoptosis^[112]. These immunomodulatory effects appear to protect the jeopardized myocardium within the area at risk, thereby improving myocardial salvage index. The reduction in final infarct size, however, may require larger sample sizes to achieve statistical significance.

Glucagon-like peptide-1 receptor agonists (GLP-1 RAs) have emerged as a class of agents with anti-inflammatory and vascular protective effects beyond glycemic control. The SELECT trial demonstrated that semaglutide 2.4 mg once weekly reduced the risk of MACE by 20% in patients with established cardiovascular disease and overweight or obesity but without diabetes^[113], while the SURPASS-CVOT trial showed that tirzepatide was non-inferior to dulaglutide with respect to cardiovascular outcomes in patients with type 2 diabetes and atherosclerotic cardiovascular disease^[114]. Mechanistically, GLP-1 receptor activation attenuates platelet aggregation and thrombosis through cAMP/PKA-dependent signaling^[115], reduces microvascular thrombosis and nitro-oxidative stress in endotoxemic models^[116], and modulates prothrombotic and atherosclerotic mechanisms associated with epicardial fat, neutrophils, and endothelial cells^[117]. Although their specific impact on coronary microvascular dysfunction requires further investigation, GLP-1 RAs represent a promising class of agents that target both metabolic dysregulation and thromboinflammation.

HMGB1/TLR4/NF-κB pathway has shown compelling preclinical results. In a rat model of CME, treatment with the HMGB1 inhibitor glycyrrhizin (10 mg/kg, intravenous) significantly reduced microinfarct size by ~20% and improved LVEF by ~18.5% compared to untreated controls, effects that were mediated through restoration of protective autophagy^[13]. Blockade of HMGB1 also suppressed TLR4 and NF-κB expression, reducing the proinflammatory and prothrombotic milieu. Glycyrrhizin has a well-established safety profile, making it an attractive candidate for clinical translation. Despite these encouraging preclinical data, several translational barriers must be considered. Glycyrrhizin can cause dose-dependent mineralocorticoid-like effects, including hypertension, hypokalemia, and fluid retention, which may limit its safety window in patients with cardiovascular disease, particularly those with heart failure or renal impairment^[118]. Additionally, glycyrrhizin doses above 90 mg/kg have been shown to produce significant hemorrhagic effects in animal models^[119], raising concerns about bleeding risk. The optimal dosing regimen in humans also remains undefined, as most preclinical studies employed pre-treatment protocols not directly translatable to the clinical setting. These considerations underscore the need for careful dose-finding studies and bleeding risk monitoring before clinical translation can be pursued.

Another promising anti-inflammatory and microvascular protective agent is relaxin. Relaxin directly inhibits platelet aggregation and suppresses platelet release from megakaryocytes^[120,121], and mitigates I/R-induced endothelial barrier disruption and inflammation. In a mouse cardiac I/R model, relaxin reduced the no-reflow zone of the LV by 38% and Evans blue extravasation by 56%, increased open capillaries, reduced leukocyte density and cytokine gene expression, mitigated I/R-induced loss of VE-cadherin, and at 2 weeks partially preserved LV contractile function while limiting chamber dilatation^[28]. A recent systematic review further supports that pretreatment with relaxin mitigates coronary microvascular injury induced by I/R^[122]. Other anti-inflammatory strategies include IL-1β inhibition (e.g., canakinumab), which has been shown to reduce cardiovascular event rates in the CANTOS trial, although its effects on MVO specifically have not been fully characterized. Complement inhibitors and neutrophil elastase inhibitors are also being explored to prevent NETosis-mediated microthrombosis^[17]. However, for NETs-targeted strategies, additional translational challenges exist. DNase I has a short plasma half-life, which may limit its therapeutic efficacy^[123]. Systemic NETs degradation could release histones and other pro-inflammatory contents, potentially exacerbating inflammation^[124]; and the balance between NETosis inhibition and preservation of neutrophil host defense raises concerns about infection risk^[125]. Safety concerns also extend to bleeding, as DNase I has been associated with lung hemorrhage in murine models and with hemorrhagic transformation when administered with delay in ischemic stroke models^[126,127]. These safety and pharmacokinetic considerations highlight the need for rigorous risk-benefit assessment before clinical translation.

Beyond their primary lipid-lowering action, statins exert clinically important antithrombotic effects that are particularly relevant to microvascular protection. The clinical significance of these effects is underscored by the observation that long-term nonadherence to statin therapy is associated with significantly increased mortality. In patients with ACS, nonadherence is associated with a nearly threefold increase in mortality (14.9% vs. 4.9%) compared with adherent patients^[128]. Among survivors of MI, adherence of less than 50% is associated with significantly increased risks of both all-cause mortality (HR 1.58) and cardiovascular mortality (HR 1.60)^[129]. Furthermore, in a large Korean multicenter registry of AMI survivors, postdischarge statin withdrawal was associated with higher mortality than continued statin treatment (HR 3.45), primarily as the result of increased cardiac mortality (HR 4.65)^[130]. These findings highlight the critical need to communicate the full spectrum of statin benefits beyond cholesterol lowering to improve patient persistence, particularly in patients at high risk for MVO.

In addition to lowering lipid levels, statins exert multiple pleiotropic effects that directly modulate the thrombotic and inflammatory milieu. Statins improve endothelial function by enhancing NO bioavailability and reducing oxidative stress^[131]. They also directly inhibit platelet activation, attenuate vascular inflammation, stabilize atherosclerotic plaques, and downregulate TF expression, leading to reduced thrombin generation and enhanced fibrinolysis^[132]. Collectively, these interrelated actions contribute to improved microvascular perfusion and have been associated with reduced no-reflow incidence after AMI. Emphasizing these antithrombotic and pleiotropic benefits in clinical practice may help improve both patient adherence and microvascular outcomes.

An alternative pharmacological strategy to directly dissolve microthrombi is the local delivery of low-dose thrombolytics into the infarct-related artery. Using an in vitro microfluidic model of MVO, a short 90-s exposure to microdoses of alteplase reduced microthrombus volume by up to 75%, at a dose 1,300-fold lower than conventional intravenous thrombolysis^[133]. The STRIVE trial, the first randomized controlled trial testing this approach in patients, was a multicenter, double-blind, placebo-controlled study that enrolled 210 STEMI patients with a large thrombus burden. Patients received either alteplase 10, 20 mg, or saline placebo directly into the culprit artery after reperfusion. The primary composite outcome (MACE at 30 days, myocardial blush grade 0/1, distal embolization, or incomplete STR) occurred in 53.3% of both alteplase groups versus 52.9% of the placebo group (RR 1.00; 95%CI: 0.76-1.31; P > 0.99). Results were neutral for all individual components and both doses. A trend toward more ventricular fibrillation was observed with alteplase (10.2% vs. 1.4%; P = 0.06), while bleeding events were rare^[134]. The neutral outcome of STRIVE does not support routine use of intracoronary low-dose alteplase in STEMI.

The discrepancy between promising in vitro results and the clinical failure highlights the gap between preclinical models and human pathophysiology. Unlike the relatively uniform fibrin-rich thrombi generated in microfluidic systems, MVO in patients is driven by a heterogeneous mixture of fibrin-rich thrombi, platelet aggregates, NETs, and endothelial swelling—components largely resistant to alteplase. Moreover, drug delivery to already obstructed microvessels is technically challenging, and the injection itself may have direct irritant effects on ischemic myocardium, as suggested by the trend toward increased episodes of ventricular fibrillation. Administration after reperfusion occurs after microthrombi have become structurally stabilized, further limiting the window for effective thrombolysis. Patient selection based on angiographic thrombus burden also does not reliably identify those in whom MVO is primarily fibrin-driven. The outcomes of STRIVE are consistent with prior neutral trials, including T-TIME^[135] and STREAM^[136], and collectively argue that targeting fibrin alone is insufficient to overcome the complex pathophysiology of MVO, reinforcing the need for strategies simultaneously addressing multiple components of the thromboinflammatory network.

Non-drug strategies that leverage the heart’s endogenous protective mechanisms offer a non-pharmacological approach to reducing microthrombosis. Ischemic preconditioning (IPC) - brief, repetitive episodes of I/R before a sustained ischemic insult - has been shown to attenuate MVO and MVL in animal models. A key mediator of IPC is macrophage migration inhibitory factor (MIF): in a mouse cardiac I/R model, IPC’s protective effects (reduced infarct size, myocyte apoptosis, ROS production, and inflammatory infiltration) were completely lost in MIF knockout mice, and IPC required MIF to activate the reperfusion injury salvage kinase (RISK) pathway and AMP-activated protein kinase^[137]. IPC, applied at the onset of reperfusion, has been shown to be cardioprotective in some animal studies, but has yielded inconsistent results in clinical trials^[138], possibly due to differences in protocols or patient populations studied.

Remote ischemic conditioning (RIC) - transient I/R of a limb (e.g., via blood pressure cuff inflation/deflation) - is more clinically feasible and has been shown to reduce MVO and myocardial edema. Its mechanisms involve circulating humoral factors, including microparticles and exosomes, that modulate endothelial function and platelet reactivity^[139], though the precise mediators and their clinical relevance remain poorly defined. In STEMI patients, RIC before primary PCI reduced the infarct size and edema measured by CMR in some studies^[140,141]. However, the LIPSIA CONDITIONING trial reported an improvement in myocardial salvage index without significant reductions in infarct size or MVO^[142], and two large multicenter trials (CONDI-2/ERIC-PPCI) did not show improved clinical outcomes^[143].

The inconsistency between positive preclinical findings and neutral large-scale trials highlights several translational challenges specific to RIC. First, preclinical studies are typically conducted in young, healthy, genetically homogeneous animals, whereas STEMI patients usually present with multiple comorbidities (diabetes, hypertension, advanced age) and concomitant medications (antiplatelet agents, statins, β-blockers) that can blunt the cardioprotective signaling pathways activated by RIC^[144]. Second, the timing and delivery of RIC in clinical practice are logistically difficult to standardize; heterogeneity in ischemia duration, anesthetic exposure, and the precise timing of conditioning stimuli relative to reperfusion may all contribute to variable efficacy^[144]. Third, the robust mechanistic rationale for RIC, rooted in well-characterized signaling pathways such as the RISK pathway and the survivor activating factor enhancement pathway, may not be fully translational when these pathways are already impaired by comorbidities or pharmacological interference^[145]. These considerations suggest that RIC, as currently applied, may require optimized patient selection, refined conditioning protocols, or combination with pharmacological agents that sensitize cardioprotective pathways, rather than uniform application across all STEMI patients.

As reviewed above, extensive clinical findings underscore a recurring theme: single-agent strategies, although targeting distinct components of the thromboinflammatory cascade, have consistently failed to translate mechanistic benefits into improved clinical outcomes. This sobering reality is reflected in the current guideline landscape. The ESC 2023 guidelines explicitly identify MVO associated with primary PCI as an unmet clinical need and call for preclinical and clinical research to evaluate cardioprotective therapies^[146]. With respect to specific interventions, GP IIb/IIIa inhibitors should be considered (Class IIa, Level C) in the setting of no-reflow or thrombotic complications during PCI, while routine thrombus aspiration is not recommended (Class III, Level A)^[146]. The 2025 ACC/AHA guidelines similarly discourage routine manual thrombus aspiration (Class III, Level A)^[147]. Notably, no specific recommendations exist for anti-inflammatory or NETosis-targeted therapies targeting MVO, reflecting the gap between promising trial results and robust clinical evidence.

Microthrombosis originates from endothelial injury and is sustained by a positive feedback loop between platelet activation and inflammation, leading to MVO, MVL, and multi-organ dysfunction. Endothelial dysfunction, characterized by enhanced oxidative stress and upregulation of adhesion molecules, reduces NO bioavailability and creates a prothrombotic cellular surface. Platelets then adhere via GPIb-IX-V/vWF, become activated, and release inflammatory mediators that further amplify endothelial activation and recruit leukocytes, forming platelet-leukocyte aggregates and NETs. The HMGB1/TLR4/NF-κB pathway suppresses protective autophagy in myocardial tissue, predominantly in cardiomyocytes, exacerbating myocardial injury. Together, these mechanisms result in microinfarcts, contractile dysfunction, elevated cardiac troponin, and - through shared vascular beds - damage to the kidney, brain, and retina, especially in patients with underlying hypertension, diabetes, or hyperhomocysteinemia.

Platelet-related biomarkers, including PMPs, PMAs, sP-selectin, and activated GPIIb/IIIa, enable detection of microthrombosis and correlate with cardiac no-reflow and adverse outcomes. However, this class of biomarkers lacks routine clinical validation due to pre-analytical variability, instrument sensitivity limits, and lack of standardized reference ranges. Existing antiplatelet therapies reduce thrombotic events but do not fully disrupt the inflammation-platelet cycle, and carry a risk of bleeding. This incomplete efficacy may be partly explained by the existence of functionally and maturationally distinct platelet subpopulations that exhibit differential responses to standard antiplatelet agents and contribute to persistent thromboinflammatory activity despite therapeutic interventions. Anti-inflammatory agents show promise in preclinical and early-phase clinical studies but have not yet demonstrated consistent benefit in large clinical trials taking MVO as the endpoint. The failure of intracoronary thrombolysis in the STRIVE trial further illustrates the gap between preclinical promise and clinical reality. Mechanical conditioning is safe and inexpensive, but its inconsistent efficacy based on large trials highlights the need for optimized protocols and better patient selection. Collectively, current therapies are limited by their inability to adequately counteract the multifaceted nature of microvascular injury, and effective strategies have yet to achieve durable suppression of this self-perpetuating cycle.

Early identification of patients at risk for microthrombosis requires validated, clinically deployable biomarker panels. Future work should focus on constructing composite risk scores integrating: (1) platelet activation markers (PMPs, sP-selectin), (2) endothelial injury markers (EMPs, vWF), and (3) NETosis markers (cell-free DNA, MPO-DNA complexes). These panels should be validated in prospective cohorts and combined with culprit artery blood sampling during PCI to capture local thrombotic activity. Integration with non-invasive imaging, including CMR for MVO or intramyocardial hemorrhage, and real-time myocardial contrast echocardiography, would further enhance risk stratification and enable early identification of patients who might benefit from targeted adjunctive therapies.

Multi-target therapeutic strategies are needed to address the multifactorial nature of MVO in the setting of AMI with clinically relevant therapeutic regimens. Testable approaches include: (1) combination regimens pairing P2Y12 inhibitors with anti-inflammatory agents to simultaneously block platelet activation and inflammation, (2) HMGB1/TLR4 pathway inhibitors such as glycyrrhizin, which require dose-finding studies to define safety windows given their mineralocorticoid-like and hemorrhagic effects at higher doses, (3) NETosis-targeted agents including DNase I and neutrophil elastase inhibitors, and (4) natural compounds with multi-target effects that warrant investigation in preclinical models of MVO and, if successful, early-phase clinical trials^[148].

To bridge the gap between preclinical promise and clinical translation, human-relevant microphysiological systems should be developed. Coronary microvasculature-on-a-chip models incorporating patient-derived endothelial cells, pericytes, and circulating blood components under controlled flow conditions can model key features of MVO, including platelet adhesion, NETosis, and fibrin deposition. These platforms enable: (1) high-throughput screening of candidate drugs and drug combinations, (2) patient-specific assessment of thrombotic risk and therapeutic response, and (3) mechanistic dissection of the thromboinflammatory cascade in a controlled, reproducible environment. Standardization and validation against clinical outcomes will be essential for their integration into drug development pipelines.

Finally, large-scale, well-characterized clinical cohorts are needed to address critical knowledge gaps in special populations. Priority research directions include: (1) prospective studies in patients with post-infection sequelae to clarify the role of persistent microvascular thrombosis and NETosis in driving cardiovascular symptoms, (2) cohort studies in patients with comorbidities that amplify microthrombosis risk, including hypertension, diabetes and chronic kidney disease, to validate biomarker panels and risk scores in these high-risk groups, and (3) randomized controlled trials in high-risk MVO populations testing whether adjunctive therapies targeting specific nodes of the thromboinflammatory network improve clinical outcomes beyond current standard of care, with mechanistic sub-studies to confirm target engagement and validate biomarkers as surrogate endpoints.