Clinical implications of ferroptosis in critical illness: a narrative review

Abstract

Despite major advances in care for patients admitted to the intensive care unit (ICU), mortality remains high and functional outcomes for ICU survivors are often poor. The persistent burden underscores the need for innovative, mechanism-based approaches. Precision medicine has emerged as a promising paradigm in critical illness, and regulated cell death (RCD) may offer a novel entry point for targeted interventions. Among the various RCD pathways, ferroptosis, an iron-dependent form of RCD driven by excessive lipid peroxidation within cellular membranes, has gained increasing attention. Its mechanistic distinctiveness from other RCDs and potential reversibility make it particularly relevant in acute, dynamic disease states. In this narrative review, we explore the role of ferroptosis in critical care, focusing on high-impact conditions such as sepsis, COVID-19, ischaemia–reperfusion injury, and neurological emergencies. We outline the molecular mechanisms of ferroptosis and discuss how it may contribute to organ dysfunction across systems. While most insights to date stem from preclinical models, emerging clinical data suggest translational potential. We highlight key studies, discuss current limitations in detection and therapeutic targeting, and consider how the evaluation of ferroptosis could help shape the future of precision medicine in the ICU.

Background

Despite significant advances in intensive care unit (ICU) management, mortality rates remain at approximately 20% in European ICUs, where survivors often experience limited functional outcomes [1]. This significant ICU mortality and morbidity highlight the urgent need for innovative approaches, as not all patients with the same disease respond equally to treatment [2], with precision medicine emerging as a promising paradigm in critical illness [3, 4].

Cell death is a fundamental process that shapes life from its very beginning. During embryonic development, for example, cell death sculpts tissues and organs, such as the separation of digits, ensuring proper anatomical formation [5]. To describe this orchestrated process of tissue formation, the term programmed cell death (PCD) was originally introduced [6]. Today, PCD refers to the physiological elimination of cells that occurs in the absence of external injury or stress, which plays essential roles in both development and tissue homeostasis [7]. Indeed, millions of cells in the human body die and are replaced every second to maintain balance and renewal [8]. However, a breakthrough came with the discovery that cell death is genetically regulated, a finding recognised by the 2002 Nobel Prize in Physiology or Medicine [9]. Since then, the term regulated cell death (RCD) has been adopted to describe cell death processes that depend on specific molecular signalling pathways and can be modulated pharmacologically or genetically [7]. Thereafter, the field of cell death has expanded far beyond the classic dichotomy of apoptosis versus necrosis [10, 11]. We now recognise a diverse landscape of regulated death programs, including necroptosis, pyroptosis, ferroptosis, and NETosis, each with distinct signalling pathways and immunological consequences [12]. Nevertheless, understanding the RCD mechanisms is increasingly relevant to intensive care medicine, where dysregulated cell death contributes to conditions such as sepsis, ischaemia–reperfusion injury, and multiple organ failure [2, 13, 14].

Over the past 15 years, ferroptosis, an iron-dependent form of RCD driven by lethal lipid peroxides, has attracted increasing attention. Growing evidence suggests that ferroptosis has a detrimental role in tissue injury following ischaemia-reperfusion events or iron overload [15, 16], as well as in neurodegeneration [17], while also serving as a novel anticancer strategy to “rust away” malignancies [18]. Moreover, in critical illness, ferroptosis appears to contribute to the pathophysiology of multiple organ dysfunction syndrome (MODS), a condition arising from severe systemic insults such as trauma, haemorrhage, surgery, or infection [19]. MODS is characterised by significant but potentially reversible organ dysfunction, with high mortality in the absence of timely intervention [20]. The dynamic interplay of inflammatory dysregulation, immunosuppression, and cellular death in MODS, and by extension critical illness, complicates research efforts and clinical outcomes [2, 21].

Traditional clinical trials on critical illness often face limitations, including population heterogeneity and methodological challenges, leading to inconclusive results [21]. Interestingly, the identification of a therapeutic target such as ferroptosis as a highly relevant mechanism in the pathogenesis of critical illness offers a novel perspective. By expanding the current set of biomarkers toward cell death mechanisms, and monitoring them, clinicians may apply predictive enrichment strategies to stratify patients more effectively and develop targeted interventions, enhancing the precision of critical care [2].

The purpose of this review is to explore the molecular mechanisms of ferroptosis and its relevance to critical illness (Fig. 1), with a focus on commonly encountered diseases such as sepsis, COVID-19, ischaemia-reperfusion injury, and neurological emergencies. While the current understanding is mostly based on preclinical data, we examine emerging clinical studies, emphasising how RCD and more specifically ferroptosis could help shape the future of precision medicine in critical care.

Fig. 1.

Overview of the types of critical illnesses in which ferroptosis might be involved. Colours indicate the strength of evidence for ferroptosis involvement in the underlying pathophysiology. Green indicates human data demonstrating an effect of a ferroptosis inhibitor. Orange indicates human data suggestive of ferroptosis involvement based on biomarker measurements, gene-expression analyses, or tissue staining. Red indicates evidence limited to preclinical models. AKI acute kidney injury, ARDS acute respiratory distress syndrome, CABG coronary artery bypass grafting

Search strategy

A systematic literature search was conducted in PubMed and Web of Science via the following search terms: (ferroptosis OR programmed cell death) AND (detection OR ischaemia-reperfusion OR traumatic brain injury OR polytrauma OR sepsis OR multiple organ failure OR ARDS OR respiratory failure OR liver failure OR COVID OR influenza OR cardiac arrest OR ICU OR critical illness). Reviews, systematic reviews, and meta-analyses published from 2015 to the present were included in the analysis. Additionally, original research articles, including clinical trials and observational studies in humans, published from 2020 to present, were prioritised to capture the most recent advancements in the field. Although the search strategy was intentionally broad, the synthesis prioritised studies with clinical relevance or translational potential. Mechanistic preclinical studies without a clear connection to critical illness in humans were not emphasised, but were incorporated where necessary to provide biological context and clarify underlying mechanisms.

Furthermore, to identify ongoing clinical trials, a systematic search was performed on ClinicalTrials.gov and the Clinical Trials Information System (CTIS). The following search terms were used: ferroptosis, deferoxamine, ferrostatin-1, liproxstatin-1, deferasirox, deferiprone, UAMC-3203, ciclopirox, CN128, vitamin E, BHT, XJB-5-131, CoQ10, D4-arachidonic acid, D10-docosahexaenoic acid, baicalein, PD 146176, AA-861, zileuton, FerroLoxin-1/2, and lipid peroxidation. All ongoing and recently completed trials were considered, regardless of phase, to capture the most up-to-date information on ferroptosis-targeting interventions.

Mechanisms of ferroptosis

Ferroptosis is an iron-dependent type of RCD characterised by the formation of toxic lipid peroxides (LOOHs) within cellular membranes [22]. Since the scientific understanding of oxygen-based metabolism, iron is known to play a central role in catalysing redox reactions. Although oxygen and iron are essential for cellular survival, their utilisation inevitably generates reactive oxygen species (ROS), such as LOOH, as byproducts of these redox reactions. While often viewed as pathological, ferroptosis may also play physiological roles in tumour suppression and immune surveillance [16]. In this section, we briefly outline the core mechanisms of ferroptosis to clarify its relevance in the pathogenesis of critical illness (Fig. 2), with more comprehensive details available elsewhere [23].

Fig. 2.

Metabolic pathways involved in ferroptosis. Adapted and simplified from Veeckmans et al. [23], with permission. Ferroptosis is tightly regulated by three key elements: redox, iron, and lipids. The different antioxidant defence mechanisms implicated in the ferroptosis pathway include the Xc -GSH-GPX4 pathway, transsulfuration pathway, mevalonate pathway, FSP1-vitamin K/CoQ₁₀ pathway, glutaminolysis, DHODH–CoQ₁₀H₂ pathway, and aldo-keto reductases (redox metabolism displayed in green). Fluctuations in the labile iron pool (Fe²⁺) are mainly controlled by TfR1, FPN, DMT1, NCOA4, NRF2, and HMOX1 (iron metabolism displayed in blue). Fatty acids enter the cell through CD36 and FATP. The peroxidation of PUFA-containing phospholipids (PUFA-PLs) within cellular membranes is regulated mainly by ACSL4 and LPCAT3 (lipid metabolism displayed in orange). An imbalance between the production of endogenous oxidants and antioxidants and the presence of excess free labile iron and oxidisable PLs acylated with PUFAs are both required for ferroptosis execution. 4-HNE 4-hydroxy-2-nonenal, ACSL acyl-CoA synthetase long-chain family member, CoQ₁₀ ubiquinone, CoQ₁₀H₂ ubiquinol, DHODH dihydroorotate dehydrogenase, DMT1 divalent metal transporter 1, FABPpm plasma membrane fatty acid binding protein, FATP fatty acid transport protein, Fe²⁺ ferrous iron, Fe³⁺ ferric iron, FPN ferroportin, FSP1 ferroptosis suppressor protein 1, Gln glutamine, GPX4 glutathione peroxidase 4, GSH glutathione, GSS glutathione synthetase, GSSG oxidised glutathione, HEP hepcidin, HEPH hephaestin, HMOX1 haem oxygenase 1, LPCAT3 lysophosphatidylcholine acyltransferase 3, MDA malondialdehyde, MUFA monounsaturated fatty acid, NCOA4 nuclear receptor coactivator 4, NRF2 nuclear factor E2-related factor 2, O₂ oxygen, PL phospholipid, PUFA polyunsaturated fatty acid, ROS reactive oxygen species, SFA saturated fatty acid, STEAP3 transmembrane epithelial antigen of the prostate 3, TCA cycle tricarboxylic acid cycle, TfR1 transferrin receptor 1, VK vitamin K, VKH₂ vitamin K hydroquinone

Execution

The execution of ferroptosis involves the degradation of cell membranes, which are composed of a phospholipid (PL) bilayer. PLs containing polyunsaturated fatty acids (PUFAs), in contrast to saturated or monounsaturated fatty acids, are vulnerable to peroxidation because of the presence of two or more adjacent double bonds. Therefore, PUFAs-containing PLs within cell membranes increase the susceptibility to oxidative damage [23].

Ferroptosis may be initiated through nonenzymatic reactions, such as the iron-catalysed Fenton reaction, or via enzymatic pathways mediated by lipoxygenase (LOX) or cytochrome P450 oxidoreductase. Both enzymatic and nonenzymatic reactions result in the formation of PL hydroperoxides (PLOOH), mainly in the phospholipids of cell membranes, triggering a self-propagating loop of lipid peroxidation. The process of lipid peroxidation subsequently generates toxic aldehydes, including 4-hydroxy-2-nonenal (4-HNE) and malondialdehyde (MDA), which cause biomolecular damage by forming covalent bonds with proteins and DNA, contributing to organ dysfunction [24]. Notably, the secondary toxic aldehydes are currently used as a means to detect ferroptosis, for instance in blood or for tissue staining [19, 25]. The resulting membrane thinning, curvature formation, and increased accessibility to oxidants ultimately lead to cell membrane rupture [26]. Nevertheless, the destructive cycle of cell membrane rupture can be halted only if ROS are neutralised by endogenous radical trapping agents (RTAs), such as vitamin A [27], vitamin E [28], oestrogen metabolites [29] or hydropersulfides [30]. Alternatively, detoxification of PLOOH by selenium-dependent glutathione peroxidase 4 (GPX4), an enzyme essential to the glutathione (GSH) system, can prevent ferroptosis [23]. Finally, glutamine, an amino acid, is known to induce ferroptosis through its use in the tricarboxylic acid cycle, leading to the production of ROS in the mitochondria [31]—a mechanism that may partly explain the increased mortality observed with glutamine supplementation in critically ill patients in the REDOXS trial [32]. In summary, the self-reinforcing loop of LOOH formation leads to peroxidation of the PUFAs in the cell membrane and subsequently, membrane rupture, which is the hallmark of ferroptotic cell death.

Regulation

Ferroptosis is tightly regulated by lipid metabolism, iron, and redox balance (Fig. 2). First, regarding lipid metabolism, fatty acid uptake is known to be facilitated through specific receptors , e.g. fatty acid transport protein (FATP), and contributes to the intracellular fatty acid pool [33]. Thereafter, PUFAs are synthesised and remodelled from the fatty acid pool by enzymes such as acyl-CoA synthetase long-chain family member 4 (ACSL4) and lysophosphatidylcholine acyltransferase 3 (LPCAT3). Interestingly, genetic deletions of ACSL4 and LPCAT3 are shown to suppress ferroptosis by depleting substrates required for LOOH formation [34].

Iron metabolism is undoubtedly another critical regulatory factor [35], and its dysregulation is associated with the severity of critical illness since the late 1980s [36]. Under physiological conditions, stable ferric iron (Fe³⁺) is imported into cells via transferrin receptor 1 (TfR1) in complex with transferrin. In the endolysosomal compartment, ferric iron is reduced to unstable ferrous iron (Fe²⁺) and subsequently transported to the cytosol. Cytosolic ferrous iron forms a labile iron pool (LIP) that can be stabilised by storage in ferritin. Ferritin degradation, which is mediated by nuclear coreceptor activator 4 (NCOA4) in a process known as ferritinophagy, releases iron back to the cytosol [37, 38]. Additional sources of ferrous iron include the degradation of haem via HMOX1, which is regulated by nuclear factor erythroid 2-related factor 2 (NRF2) [23]. Furthermore, when the systemic iron demand decreases, ferroportin (FPN), an iron export protein, is inhibited by hepatic hepcidin, limiting iron export from cells and increasing intracellular iron availability [39]. However, in the case of iron export, hephaestin converts Fe²⁺ back to Fe³⁺ at the outer side of the plasma membrane, enhancing membrane stability and supporting FPN function [40].

Finally, antioxidant systems play pivotal roles in ferroptosis regulation. GSH, a primary antioxidant, depends on intracellular cysteine for its synthesis. Depletion of cysteine, such as through inhibition of system Xc⁻ (SLC7A11), directly compromises the activity of the key regulator GPX4, thus permitting unregulated lipid peroxidation. However, lipid peroxidation can be counteracted by endogenous antioxidant systems, such as the reduction of ubiquinone (CoQ₁₀) via NADPH, which is mediated by ferroptosis suppressor protein 1 (FSP1) in the extramitochondrial membrane, or through DHODH in the mitochondrial membrane, which is independent of GPX4 or FSP1 [23, 41]. Interestingly, FSP1 also has the potential to reduce vitamin K to its hydroquinone, which acts as a potent radical trapping agent, thereby protecting against lipid peroxidation [42].

Understanding the regulatory axes—lipid metabolism, iron handling and redox balance—is clinically relevant because each represents a potential therapeutic target and corresponds to recognisable phenotypes in critical illness, as outlined in the following sections.

Detection methods

Detecting ferroptosis in clinical practice, particularly in the ICU, remains a major challenge. Current detection methods are designed primarily for research purposes and rely on tissue sampling, molecular profiling, and advanced analytics. At the bedside, clinicians must currently rely on indirect markers that suggest iron-dependent lipid peroxidation or oxidative stress, such as the dysregulation of systemic iron metabolism, rather than confirming ferroptosis itself.

In preclinical models, the role of ferroptosis is often demonstrated by reversal: inhibitors such as ferrostatin-1 (Fer-1) or analogues are used to show that blocking ferroptosis improves outcomes. This approach is by definition not feasible in clinical samples. Instead, surrogate markers are used to approximate ferroptotic activity [43].

Several surrogate markers have been proposed, with MDA being the most widely studied. As a reflection of ferroptotic cell death, MDA concentrations are typically measured via spectrophotometry after derivatisation with thiobarbituric acid or 1-methyl-2-phenylindole. While useful in population studies, MDA assays currently exhibit high variability owing to substantial interference from endogenous compounds and the inherent instability of MDA [44, 45]. Concentrations as low as 3.9 µM are reported in population studies, approaching assay detection limits [46]. However, in high oxidative stress conditions such as ischaemia-reperfusion injury, MDA levels peak shortly after reperfusion before declining, potentially offering a clinically relevant measurement window [47]. Further optimisation of assay protocols is essential to enable reliable patient-level stratification.

Routine ICU parameters such as serum iron and ferritin levels in peripheral blood provide insight into systemic iron metabolism, but do not directly reflect ferroptosis involvement and are not organ specific. Ferritin, while useful as a marker of iron dysregulation, lacks specificity: although elevated as an acute-phase reactant in critical illness, it is disproportionately elevated in conditions such as haemophagocytic lymphohistiocytosis independently of ferroptosis [48]. In contrast, the quantification of catalytic iron (Fe²⁺) offers a more direct assessment of the labile iron pool [49]. In septic patients, elevated labile iron levels have been associated with mortality and shock severity, suggesting clinical relevance [19]. Moreover, hepcidin is shown to be inversely correlated with catalytic iron, which may also be useful as a prognostic marker [50]. In summary, although iron metabolism is not a direct readout of ferroptosis, it remains closely linked to it, and iron dynamics will be essential to consider when moving toward clinical practice.

In addition to blood-based markers, tissue-based analyses remain the gold standard. Immunohistochemical staining for markers such as 4-HNE, oxidised phosphatidylcholine, TfR1, or MDA can visualise ferroptotic processes in situ [25, 51]. However, the invasive nature of tissue sampling makes staining techniques impractical for routine use in critically ill patients and lesion heterogeneity in the sampled tissue further hampers interpretation.

A more advanced approach focuses on lipidomics. Oxidative lipidomics, the study of oxidised PLs, offers high-resolution insight into ferroptosis-related lipid damage [43]. Oxidative lipidomic analyses rely on mass spectrometry and can detect low-abundance oxidised species with high sensitivity [52]. Owing to the novelty of this field, oxidative lipidomic analyses are generally labour intensive and not well suited for routine clinical testing. Nonetheless, mass spectrometry-based techniques including metabolomics and lipidomics, have made significant contributions to clinical diagnostics in other domains, suggesting translational potential [53]. However, translating omics technologies into critical care will require a solid understanding of lipid dynamics and the development of dedicated analytical pipelines.

Finally, gene expression profiling is another tool often used to infer ferroptosis. The downregulation of protective regulators such as NRF2 or GPX4 may indicate increased vulnerability. However, transcriptional changes reflect adaptive responses to oxidative stress and cannot confirm ferroptosis per se. As highlighted by Mishima et al., reproducible research in the ferroptosis field requires careful interpretation of gene signatures [43]. Similarly, methylome profiling could represent an alternative approach, compatible with rapid non-invasive cell-free DNA diagnostics [54].

While several methods suggest ferroptosis activity, none are currently validated for patient-level stratification in the ICU. Commonly used blood-based markers such as MDA, 4-HNE, catalytic iron and hepcidin lack specificity and largely reflect general oxidative stress, but may contribute to broader RCD-related biomarker patterns that track with illness severity. As a result, current blood biomarkers cannot be interpreted as standalone evidence of ferroptosis, and robust identification still requires multiple orthogonal lines of evidence combining biochemical, molecular and morphological readouts. Furthermore, clinically meaningful cut-offs have not yet been defined [55]. Bridging the gap to clinical application will require consensus on biomarker panels, clinically meaningful thresholds, and streamlined analytical platforms, ideally integrating oxidative lipidomics with established ICU parameters into a unified multimarker approach.

Pharmacological inhibition

The ferroptosis field is predominantly shaped by research into ferroptosis-inducing compounds, initially developed as cancer immunotherapies. The work on inducers has driven the parallel development of ferroptosis inhibitors, the therapeutically relevant counterpart for critical illness, where ferroptosis is a pathological process to be suppressed rather than exploited. Ferroptosis can be inhibited pharmacologically at multiple levels: iron availability, lipid peroxidation, and antioxidant defence, each corresponding to the regulatory axes outlined above. Here, we present a snapshot of the most relevant ferroptosis inhibitors relevant for critical illness: iron chelators, radical-trapping agents, and repurposed ICU drugs [56].

First, iron chelators such as deferoxamine or deferiprone prevent the formation of highly reactive hydroxyl radicals by chelating free iron, showing variable clinical improvement in ischaemia-reperfusion injury, as discussed below [57–59]. Although systemic iron depletion raises theoretical concerns, adverse effects such as clinically relevant anaemia are not consistently reported to date [58].

Second, radical-trapping agents interrupt lipid peroxidation, and thus the destructive cycle of membrane rupture. On the one hand, endogenous radical-trapping agents (RTAs) include vitamin E [28], CoQ₁₀H₂ [60] and vitamin KH₂ [42] (Fig. 2). On the other hand, synthetic lipophilic radical traps, including Fer-1 and liproxstatin-1 (Lip-1), directly neutralise toxic lipid radicals, inhibiting the peroxidation cascade in ferroptosis. Challenges with solubility and metabolic stability in vivo have led to the development of improved ferroptosis inhibitors, such as FXT-001 (formerly UAMC-3203), which is a more stable and soluble Fer-1 analogue. Additionally, inhibitors such as deuterated PUFAs and lipoxygenase inhibitors broaden the options for ferroptosis inhibition, although these options are controversial, as genetic targeting of LOXs failed to prevent ferroptotic cell death [23, 61]. Therefore, current ferroptosis inhibition strategies focus on the synthesis of potent radical-trapping agents [47, 62], with edaravone, mainly used in amyotrophic lateral sclerosis, as an FDA-approved example [63].

Finally, several commonly used medications in the ICU are suggested to have protective properties. For example, sedatives such as dexmedetomidine, etomidate and propofol upregulate NRF2, which subsequently downregulates HMOX1, thereby inhibiting haem degradation and preventing iron accumulation in mouse models [64–66]. Propofol additionally enhances antioxidant capacity, as demonstrated in vitro [67]. Although the exact mechanisms remain incompletely understood and evidence is currently limited to preclinical models, repurposing approved therapies in clinical studies could accelerate research into ferroptosis in the ICU.

Clinical phenotypes

Sepsis

Sepsis remains a major challenge in critical care, affecting nearly 49 million people worldwide annually. Sepsis has in-hospital mortality rates exceeding 10%, which can reach 40% in patients with septic shock [68]. Although sepsis is defined as life-threatening organ dysfunction caused by a dysregulated host response to infection [69], the exact molecular mechanisms of sepsis remain incompletely understood, limiting current treatments to timely antibiotics, fluid resuscitation, and organ support [70]. Recent evidence suggests that ferroptosis is one of the mechanisms of organ failure in sepsis, therefore providing a therapeutic target [71].

Pathogenesis

Iron is indispensable for both the host and pathogen. Many bacteria rely on iron to proliferate, resulting in increased virulence [72]. To counteract proliferation, the human body limits free iron in plasma during infection, primarily through hepcidin-mediated downregulation of FPN. While protective in early sepsis, the response through hepcidin paradoxically leads to the accumulation of intracellular iron, particularly in mitochondria, where it catalyses lipid peroxidation and ferroptotic cell death [71, 73]. Hepcidin levels, however, fluctuate over time. Beyond a certain threshold, lower hepcidin levels are associated with increased mortality in critically ill patients, suggesting that iron is released as catalytic iron into the plasma, driving ferroptosis rather than containing it. Interpretation is further complicated by the iron-independent antimicrobial properties of hepcidin, meaning that elevated levels during infection may reflect host defence rather than iron dysregulation alone [50]. Infection-induced ferritinophagy via NCOA4 further amplifies intracellular iron overload. Moreover, ROS, which are abundant in sepsis, exacerbate iron overload, damage cellular membranes and perpetuate organ injury. Concurrent erythrocyte breakdown increases the iron burden, creating a self-reinforcing loop of oxidative stress and ferroptosis [74]. Additionally, iron accumulation in phagocytic cells may promote bacterial and fungal growth [75], and worsen the course of certain viral, bacterial and fungal infections [76–78]. Iron-driven injury contributes to the rapid progression of multiple organ failure observed in septic patients [79], but manifests differently across organ systems.

Heart

Septic cardiomyopathy (SCM), characterised by left ventricular systolic dysfunction, is a frequent complication of sepsis whose molecular mechanisms remain incompletely understood [80]. Preclinical models suggest a role for ferroptosis: elevated MDA and ROS, increased mitochondrial TfR1 expression, and upregulation of ferroptosis-related gene sets are observed in the septic heart, with Fer-1 treatment improving cardiac function [81, 82]. STAT3 inhibition, central to inflammatory signalling, is additionally shown to suppress GPX4, potentially increasing myocardial vulnerability to oxidative injury [83]. Interestingly, dexmedetomidine, widely used in ICUs, may attenuate ferroptosis by reducing HMOX1-driven iron accumulation in preclinical models, though whether this contributes to its clinical effects remains speculative [84].

Lungs

The lungs are particularly vulnerable to sepsis, with acute respiratory distress syndrome (ARDS) representing a frequent and severe complication [85, 86]. Beyond the well-described mechanisms of neutrophil influx and fibroblast activation leading to early fibrotic remodelling [87], ferroptosis has emerged as an additional pathogenic contributor. Preclinical LPS models consistently demonstrate elevated ROS, ACSL4 expression, and MDA levels alongside glutathione depletion. However, while Fer-1 attenuates the stress effects, it does so only partially, suggesting that ferroptosis contributes to, but does not fully account for, sepsis-induced lung injury [13, 84, 88].

Iron dysregulation amplifies this vulnerability at the alveolar level. A 2003 study demonstrated elevated haemoglobin, ferritin, and transferrin in bronchoalveolar lavage fluid of ARDS patients compared to healthy volunteers, reflecting disrupted pulmonary iron homeostasis and increased ferroptosis susceptibility [89]. Lactate, a metabolic marker of tissue hypoperfusion commonly elevated in sepsis, has additionally been shown to upregulate ACSL4 in alveolar epithelial cells via histone lactylation, suggesting a direct injurious role beyond its function as a metabolic marker [90]. Calcium signalling represents another regulatory axis: verapamil-mediated blockade of L-type calcium channels reduced calcium influx and iron overload in septic mice, attenuating ferroptotic lung injury [91]. Moreover, mild hypothermia is associated with GPX4 upregulation and reduced lung injury in mouse models, suggesting a possible protective axis, though clinical validation remains absent [92]. Whether redox imbalance, iron dysregulation, and metabolic reprogramming can be therapeutically targeted in septic lung injury and ARDS patients remains to be established.

Kidneys

Ferroptosis-related injury is not confined to the cardiopulmonary system. Acute kidney injury (AKI) is frequently encountered in septic ICU patients, and independently associated with a higher mortality [93]. Under physiological circumstances, iron is filtered and subsequently reabsorbed in the renal tubules. However, sepsis-induced ROS, haemolysis, and inflammatory mediators disrupt the balance between filtered and reabsorbed iron, leading to intracellular iron accumulation in renal tubular cells [73]. A similar pattern of iron deposition and oxidative injury is observed in chronic kidney disease, where it contributes to disease progression and therefore may involve ferroptosis [94].

Direct evidence from ferroptosis inhibition studies in septic AKI is more limited than in cardiac or pulmonary models. NADPH oxidase inhibition and Fer-1 both attenuate AKI in preclinical models, though primarily in diabetic and obese mice, highlighting the need for context-specific interpretation [95]. Sex differences add a further layer of complexity: 17β-oestradiol confers anti-ferroptotic protection, potentially explaining lower AKI rates observed in premenopausal women [29]. Preclinical proteomic analysis of renal tubular cells identifies mitochondrial oxidative stress as a key driver, with HMOX1 activation suggested as an initiating factor of iron deposition [96]. Context-dependent results in AKI underscore the importance of patient-specific factors in ferroptosis susceptibility.

Brain

The brain is another organ at risk, given its high metabolic demand and limited antioxidant reserve. Septic encephalopathy (SAE) affects nearly half of septic patients and encompasses a spectrum from delirium to coma. The pathophysiology involves neuroinflammation, glutamate-mediated excitotoxicity, oxidative stress, and blood–brain barrier disruption [97]. Preclinical CLP models show GPX4 inactivation and transferrin upregulation, with Fer-1 inhibition attenuating both ferroptosis and glutamate excitotoxicity [98]. Corroborating evidence from human tissue is limited but notable: paediatric SAE patients exhibit decreased GPX4 and SLC7A11 alongside increased LOX and ACSL4 transcription. Excess ferrous iron contributes to hippocampal injury, attenuated by deferoxamine in preclinical models [99]. Interestingly, low doses of acetaminophen are shown to reduce cognitive deficits in septic mice via mediation of the GPX4 pathway, raising the question of whether its neuroprotective effects are partly mediated through this pathway. However, in the same model acetaminophen also attenuated the broader neuroinflammatory response, illustrating the mechanistic crosstalk between ferroptosis and other inflammatory cell death pathways [100]. Neurological ferroptosis susceptibility appears region-specific, with the hippocampus emerging as particularly vulnerable, though direct evidence in adult ICU patients remains scarce.

Sepsis in trauma patients

Trauma patients represent a unique context in which ferroptosis and sepsis may be intertwined. Post-traumatic sepsis remains a clinical challenge, often driven by immunosuppression and persistent organ dysfunction. In a prospective study of 59 severe polytrauma patients, CD4 + T-cell lymphopenia was associated with ferroptosis via downregulation of the cysteine-GSH-GPX4 antioxidant axis, potentially offering an early signal of sepsis risk, though replication in larger cohorts is needed [101]. Because the concept of sepsis after trauma is complex—organ dysfunction may stem from the trauma itself or from an ICU-acquired infection such as aspiration pneumonia [79]—ferroptosis-related markers may help distinguish progressing organ failure from infection-driven sepsis, though their clinical utility in this context remains to be established.

More than ferroptosis alone

Although ferroptosis is prevalent in sepsis and closely linked to organ dysfunction and mortality, to date its inhibition alone has proven insufficient in models of sepsis [19]. ICU patients with simultaneous signatures of ferroptosis and pyroptosis, a form of RCD driven by inflammasome activation, have the lowest survival rates. The high mortality associated with combined RCD suggests that patient stratification based on combined biomarker signatures may be necessary to identify those most likely to benefit from targeted intervention, whether against ferroptosis, pyroptosis, or both [55].

COVID-19

COVID-19 provides another clinical context in which ferroptosis may play a pathogenic role. Since the first case in 2019, SARS-CoV-2 has led to more than 776 million confirmed cases and approximately 7 million deaths worldwide [102]. While the majority of infections are mild [103], a subset of patients develop severe disease characterised by hyperinflammation, ARDS, and MODS [104]. Current treatments remain largely supportive, although corticosteroids, IL-6 inhibitors and convalescent plasma have demonstrated clinical benefit [51, 105]. Emerging evidence implicates ferroptosis especially in the pathophysiology of severe COVID-19 [103, 106–108].

Pathogenesis

Three principal mechanisms link ferroptosis to SARS-CoV-2 infection. First, viruses induce iron dysregulation via IL-6-mediated hepcidin upregulation, which inhibits ferroportin and promotes intracellular iron sequestration, similar to sepsis pathogenesis. Concurrently, TfR1 expression is increased, facilitating Fe³⁺ uptake [103, 107]. Stimulation of TfR1 is not unique to SARS-CoV-2 and has been observed in other viral infections, including influenza [109]. Additionally, viral interaction with ACE2 receptors on erythrocyte precursors, resulting in a haemoglobin drop associated with the hypoxic problems in COVID infection, may release free iron, contributing to hyperferritinaemia and promoting ferroptotic cell death [110, 111]. Second, GSH depletion, frequently observed in COVID-19 patients, impairs GPX4 activity, leading to ROS accumulation. Third, mitochondrial dysfunction induced by viral replication further exacerbates oxidative stress, increasing sensitivity to ferroptotic cell death and RCD pathways more broadly [103, 107]. The mechanisms of iron dysregulation, GSH depletion and mitochondrial dysfunction may converge to cause tissue-specific injury, as illustrated by proposed ferroptotic damage to sinus node pacing cells leading to bradycardia in COVID-19 [108].

Clinical evidence

Early in the pandemic, post-mortem analysis of a fatal COVID-19 case associated with myocarditis and MODS at our institution revealed immunohistochemical evidence of lipid peroxidation in cardiac and renal tissues, providing first histological evidence of ferroptosis in COVID-19. Ferroptosis was nonetheless demonstrable despite the inability to evaluate lung tissue due to extensive haemorrhage [25, 107].

Qiu et al. subsequently reported elevated ferroptosis markers, including serum ferritin, TfR1, and MDA, in lung autopsies of COVID-19 patients. Interestingly, necroptosis and pyroptosis marker levels were not significantly altered, suggesting a degree of pathway specificity rather than generalised cell death activation. Among the genetic markers assessed, ferritin light chain 1 (FTL) and TfR1 were the most strongly upregulated, while GPX4 and FSP1 showed no significant changes [51]. Jiang et al. confirmed the limited role of GPX4, although they identified FSP1 suppression as a potential driver of ferroptosis [112]. In contrast, GPX4 seems to be downregulated in a subset of patients with critical COVID-19 [113]. The discrepant GPX4 findings across studies likely reflect patient heterogeneity and disease stage, underscoring the difficulty of using single markers to confirm ferroptosis in clinical samples. The predominance of iron dysregulation markers in the absence of necroptotic or pyroptotic signatures nonetheless suggests a driving role for ferroptosis during COVID-19 progression.

Despite the above findings, ferroptosis is not universally present across critically ill COVID-19 patients. In a cohort of 120 ICU COVID-19 patients clustered based on ferroptosis and pyroptosis markers, a subgroup with elevated MDA and free iron levels was associated with increased mortality. Notably, 40% of patients lacked defining ferroptosis biomarkers, suggesting that universal ferroptosis inhibition is unlikely to benefit all critically ill COVID-19 patients and reinforcing the need for biomarker-based patient selection [114].

Ischaemia–reperfusion and organ transplantation

Ischaemia-reperfusion (I/R) involves restricted blood supply followed by reoxygenation, causing oxidative stress and cell death [115]. I/R remains a significant clinical challenge in settings ranging from cardiac surgery and stroke to solid organ transplantation. Observations made in I/R implicate ferroptosis as a key mechanism of I/R-induced cell death, suggesting therapeutic targets not yet exploited in clinical practice [116–118]. Although lipid peroxidation is the common endpoint shared with sepsis, the triggers differ: I/R is characterised by a predictable, biphasic timeline that may offer specific therapeutic windows. For clarity, I/R injury is discussed as either acute (e.g. stroke) or anticipated (e.g. transplantation). However, the distinction between the two is clinically useful, but not always that strict, as e.g. in CABG where ischaemic injury and procedural reperfusion both happen.

Pathogenesis

I/R comprises two sequential phases. During ischaemia, reduced ATP production shifts metabolism to anaerobic pathways, disrupting ionic homeostasis and causing cellular oedema, inflammation, and iron accumulation through ferritin degradation. Upon reperfusion, the sudden influx of oxygen reacts with accumulated iron to generate ROS that overwhelm antioxidant defences, driving endothelial dysfunction, DNA damage, and ferroptotic cell death [119, 120]. Importantly, although ferroptosis is seen during ischaemia [47], it occurs predominantly during the reperfusion phase rather than during ischaemia itself [121], a temporal specificity that defines the window for preventive intervention.

Acute I/R

Acute I/R injury affects heart, kidney and brain through a shared pathophysiological mechanism, yet manifests in distinct clinical contexts: myocardial infarction and cardiac surgery in the heart, cardiopulmonary bypass-associated AKI in the kidney, and stroke or post-cardiac arrest brain injury in the brain. Although timely reperfusion through thrombolysis or interventional or surgical treatments is key [122, 123], reperfusion itself may exacerbate damage [124].

Notably, even before the clarification of ferroptosis as a distinct cell death pathway, residual iron accumulation after myocardial infarction was already linked to impaired left ventricular remodelling [124]. During ischaemia, ferritin degradation may initiate the iron-catalysed Fenton reaction [125]. When blood flow is restored, murine models exhibit increased ACSL4, iron, and MDA alongside reduced GPX4 [121]. In response to cellular hypoxia, hypoxia-inducible factor 1 upregulates TfR1, thereby enhancing iron uptake in the reperfused myocardium [126]. Importantly, the cell death response in reperfusion, whether ferroptosis or another RCD, is cell type dependent [127], underscoring the complexity of targeting a single pathway.

The temporal specificity of RCD is confirmed in cardiac surgery. A recent study identified ferroptosis signatures post-cardiopulmonary bypass, including elevated COX-2 mRNA, lipid peroxidation and fibrosis [128]. Strikingly, even prior to the formal identification of ferroptosis, intravenous deferoxamine during CABG reduced oxygen radical production, particularly in patients with a reduced ejection fraction [129]. Similar oxidative patterns are observed in PCI, although without a reduction in infarct size [130]. Importantly, ferroptotic injury during cardiac surgery is not confined to the myocardium: free iron release following cardiopulmonary bypass exacerbates tubular cell peroxidation, contributing to postoperative AKI [131]. Both established drugs and experimental inhibitors modulate ferroptosis pathways in these settings. Preclinical studies show that dexmedetomidine [132], propofol [67] and etomidate [115] modulate NRF2 signalling, suppressing lipid peroxidation and iron accumulation, and thereby attenuating myocardial injury and fibrosis [67]. Fer-1 additionally reduces cardiomyocyte cell death and blocks neutrophil recruitment after reperfusion in cardiac transplants [133].

A similar pattern of iron-driven injury is observed in the brain. Iron accumulation in ischaemic areas post-reperfusion has been documented in both patients and animal models, raising concern about oxidative damage in neuronal membranes, which are rich in PUFAs and therefore highly susceptible to lipid peroxidation [134, 135]. In murine I/R models, elevated hepcidin and reduced ferroportin levels point to cerebral iron dysregulation [136], whereas increased NCOA4 expression suggests augmented ferritinophagy [137]. Direct evidence in human stroke patients remains limited, largely owing to the ethical and technical constraints of post-reperfusion tissue sampling [138]. Nonetheless, deferoxamine has demonstrated neuroprotective effects in moderate to severe ischaemic stroke, reducing infarct volume with an acceptable safety profile [139–141].

Of particular relevance to the ICU, post-cardiac arrest brain injury is the leading cause of death after successful resuscitation and a major determinant of long-term neurological outcome [142]. The global I/R insult during resuscitation creates a biochemical environment highly permissive to ferroptosis. In murine asphyxia models, progressive hippocampal iron accumulation, blood-brain barrier disruption, and mitochondrial changes consistent with ferroptosis develop within 24 h [143]. Large-animal cardiac arrest models corroborate these findings: in pigs, hippocampal Fe²⁺ elevation, glutathione depletion, and increased ROS after return of spontaneous circulation were mitigated by ferroptosis inhibitors [144].

An important limitation of preclinical ferroptosis research in acute I/R is the timing of inhibitor administration: most studies administer ferroptosis inhibitors prior to the ischaemic period [121, 132], a scenario that is clinically unrealistic in unplanned events such as myocardial infarction or cardiac arrest. Translational relevance therefore depends on demonstrating efficacy when inhibitors are administered at the onset of reperfusion—a window that is both clinically feasible and mechanistically justified, given that ferroptosis occurs predominantly during reperfusion rather than ischaemia.

Anticipated I/R

In acute I/R, the inability to administer ferroptosis inhibitors before the ischaemic period limits translational applicability. Solid organ transplantation overcomes this constraint, as the timing of I/R is known in advance and preventive strategies are feasible.

In lung transplantation, I/R injury is a major contributor to primary graft dysfunction, affecting up to 30% of recipients [145]. Reperfusion triggers ROS production, disrupts alveolar epithelial and endothelial integrity, and impairs barrier function [146]. Ferroptosis appears early in the reperfusion phase: ACSL4 knockdown limits lipid peroxidation and protects epithelial cells in vitro and in vivo [147, 148], and Lip-1 administered during cold ischaemia significantly reduces tissue damage and inflammation in murine models, with greatest efficacy when administered before reperfusion [149]. Additionally, ferroptosis pathway enrichment is demonstrated in porcine I/R models, bridging the gap between rodent and human physiology [150]. Dexmedetomidine may further inhibit ferritinophagy via NRF2 upregulation, suggesting an additional protective axis [151].

Human data illustrate the interplay of multiple RCD pathways: in human lung transplants, apoptosis and necroptosis genes are upregulated at reperfusion after cold storage, whereas ferroptosis-related genes are not [152]. During ex vivo lung perfusion (EVLP), however, all RCD pathways are activated. Notably, ferroptosis-related genes are predominantly enriched in lungs donated after brain death, whereas lungs donated after circulatory death—where greater reperfusion injury is expected given the warm ischaemia time—show minimal ferroptosis-related gene expression prior to EVLP, attributed to the neurogenic inflammatory response associated with brain death [153]. Interestingly, ferroptosis-related gene expression correlates with early recipient outcomes only in lungs donated after circulatory death, despite their lower baseline expression, suggesting that even low-level ferroptotic activity may be clinically consequential in already-injured grafts [153].

The liver presents a similar picture. In liver transplantation, the severity of I/R injury correlates with early allograft dysfunction [154], and ferroptosis occurs immediately after reperfusion, triggering secondary necrotic cell death and necroinflammation [155]. Elevated preoperative ferritin in paediatric donors is identified as an independent predictor of hepatic I/R damage [156], and ferroptosis-related genes are upregulated in pre-resection tissue of hepatectomy patients, suggesting pre-existing susceptibility [157]. Moreover, older and steatotic grafts appear to exacerbate ferroptotic injury, making ferroptosis inhibition potentially relevant to strategies aimed at expanding the donor pool in the context of persistent organ shortages [158, 159]. Beyond hepatocyte injury, ferroptosis inhibition may also protect intrahepatic bile duct cells, with implications for biliary complications such as non-anastomotic strictures in DCD grafts [160, 161]. Preclinical models further demonstrate that ferroptosis inhibition via Fer-1 reduces hepatic injury more effectively than inhibitors of other cell death pathways [159, 162].

In the kidney, the evidence is more limited but mechanistically consistent. Administration of the Fer-1 analogue FXT-001 prior to ischaemia significantly reduces renal tissue damage and serum creatinine levels after reperfusion in male mice, whereas necroptosis inhibition does not provide similar protection, suggesting a heightened sensitivity of renal tubules to ferroptotic stress [29, 163]. Nonetheless, necroptosis may act synergistically with ferroptosis in some contexts, underscoring the complexity of RCD in renal injury [164]. During the acute phase of kidney transplantation, ferroptosis-related genes are upregulated [165], and in the post-transplantation phase, ferroptosis contributes to allograft dysfunction, as evidenced by mitochondrial atrophy and reduced cristae in renal tubular cells [119]. Importantly, ferroptosis inhibition must be considered alongside immunosuppressive regimens: deferoxamine may reduce cyclosporine-induced nephrotoxicity, a clinically relevant interaction given that cyclosporine remains a cornerstone of post-transplant immunosuppression [166].

Across these organ systems, Veeckmans et al. recently demonstrated that I/R injury triggers an early lipid peroxidation wave in human transplants, and that ferroptosis inhibition using FXT-001 during machine perfusion suppresses propagation and improves liver and lung graft function in human and porcine perfusion models [47]. If validated in larger trials, EVLP-based ferroptosis inhibition may represent the first clinically feasible application of targeted ferroptosis therapy, transforming graft preservation from a passive to an active intervention.

Remote organ injury

I/R injury not only damages the affected organ, but may propagate ferroptotic injury to distant organs through circulating mediators, a phenomenon with direct relevance to the development of MODS in critical illness. In preclinical studies, kidney I/R injury triggers the release of DAMPs into the circulation, inducing RCD in the lung microcirculation and contributing to increased mortality [167]. Similarly, rat models of cold hepatic I/R injury demonstrate concurrent ferroptotic damage in intestinal tissue, suggesting a liver–intestine ferroptotic axis [168].

Remote ischaemic preconditioning (RIPC)—a technique in which brief, repetitive cycles of limb ischaemia induced by blood pressure cuff inflation activate endogenous organ-protective signalling pathways—offers a clinically established example of inter-organ signalling: in two randomised clinical trials, RIPC prior to CABG significantly reduced the incidence of AKI [169, 170], and preclinical studies suggest that attenuation of ferroptosis and mitochondrial dysfunction in renal tubular cells may contribute to this protection [171]. Notably, propofol, described elsewhere as an anti-ferroptotic agent via NRF2 signalling, appears to inhibit RIPC, suggesting that its net effect on ferroptosis in the perioperative setting may be context-dependent and warrants further investigation [169, 170].

The heart–kidney and liver–intestine axes described above suggest that ferroptosis is not confined to the primary site of injury, but may propagate across organ systems. This capacity to amplify systemic dysfunction may partly explain the progression from single-organ dysfunction to MODS in critically ill patients, suggesting that therapeutic strategies targeting ferroptosis may need to account for inter-organ propagation rather than focusing on single organs.

Neurological emergencies

Beyond I/R injury, ferroptosis plays a role in three neurological emergencies frequently encountered in the ICU: spinal cord injury, traumatic brain injury, and subarachnoid haemorrhage. Despite their distinct mechanisms, iron overload and oxidative membrane damage are shared pathogenic features.

Spinal cord injury

Spinal cord injury (SCI) occurs in two distinct phases. The primary phase results from mechanical trauma, causing irreversible neuronal and axonal damage, blood–spinal cord barrier disruption, and intramedullary haemorrhage. The secondary phase is a dynamic and potentially reversible cascade involving oedema, inflammation, mitochondrial impairment, and excitotoxicity—and therefore has become the primary focus for therapeutic intervention [172–174].

Ferroptosis has gained attention particularly within this secondary phase. Preclinical studies demonstrate that hyperglycaemia exacerbates ferroptotic damage in SCI models, an effect reversed by Fer-1, which is clinically relevant given the frequency of stress hyperglycaemia in critically ill patients [175]. Fer-1 additionally inhibits reactive astrocyte and microglial activation and reduces white matter injury through protective effects on oligodendrocytes [176], while FXT-001 and edaravone suppress microglial activation and promote motor function recovery [63, 177]. Finally, iron-chelating agents such as deferoxamine improve hindlimb function in vivo, though meta-analytic data remain limited by study heterogeneity [178]. Despite promising preclinical results, clinical translation remains pending, and the optimal safety profile, timing of administration, and integration into multimodal neuroprotective strategies have yet to be established.

Traumatic brain injury

As in SCI, ferroptosis may contribute to secondary injury in traumatic brain injury (TBI), a leading cause of death and disability after trauma [179, 180]. Iron accumulation driven by erythrocyte breakdown, blood-brain barrier disruption, and dysregulation of iron metabolic proteins closely resembles the pathogenesis observed in SCI [181]. A notable feature of TBI is the elevation of PUFAs in neuronal membranes, increasing vulnerability to lipid peroxidation. Adiponectin, an anti-inflammatory adipokine associated with improved neurological outcomes, appears protective against lipid-driven neurotoxicity: adiponectin-knockout models exhibit elevated PUFA levels and increased ferroptotic injury [182]. Furthermore, the temporal dynamics of iron metabolism differ across brain cell types, with neurons showing rapid TfR1 upregulation and microglial activation contributing to neuroinflammation under excessive iron loading [181, 183].

Among neuroprotective strategies, NRF2 pathway activation has been explored using dexmedetomidine and ketamine, both showing efficacy in animal models [179, 184]. A 2017 clinical trial reported that deferoxamine enhances haematoma absorption and reduces oedema [185], and Fer-1 and Lip-1 have shown potential in reducing neuronal degeneration and improving motor and cognitive outcomes preclinically [179]. Given the significant impact of secondary injury on functional outcomes and the clearly defined time-dependent therapeutic window, TBI patients may represent a relevant population for future clinical application of ferroptosis inhibitors.

Subarachnoid haemorrhage

Subarachnoid haemorrhage (SAH), most commonly caused by cerebral aneurysmal rupture, triggers a cascade of cerebral injury. The abrupt rise in intracranial pressure compromises cerebral perfusion, causing early ischaemia. Delayed cerebral ischaemia, affecting approximately 30% of patients within 3-7 days, has long been attributed to vasospasm, though recent evidence suggests vasospasm alone cannot fully account for the secondary injury burden [186].

Lipid peroxidation in cerebrospinal fluid has been linked to poor outcomes for more than two decades, suggesting oxidative membrane damage beyond purely vascular mechanisms [187]. As in TBI, iron overload arises from haemoglobin degradation, and hepcidin appears to modulate post-haemorrhagic iron handling, implicating ferroptosis as a downstream contributor [186, 188, 189]. A prospective study of 210 SAH patients found that reduced SLC7A11 and GPX4 expression alongside elevated ACSL4 independently predicted complications and poor prognosis [190], while elevated PUFA levels in plasma and CSF further reinforce a ferroptosis-associated profile [191].

Deferoxamine reduces vasospasm, ferritin levels, and neuronal death in preclinical models, potentially complementing nimodipine-based vasospasm management rather than replacing it. Classic ferroptosis inhibitors show protection against neuroinflammation, albeit preclinically [192, 193]. A Cochrane review of iron chelation in intracerebral haemorrhage, sharing pathophysiological features with SAH, found no significant benefit on survival or neurological outcomes in two randomised trials [59], urging caution in extrapolating preclinical findings to clinical practice.

Across SCI, TBI, and SAH, iron accumulation and lipid peroxidation drive secondary neurological injury through mechanisms overlapping with ferroptosis. Ferroptosis inhibition shows preclinical promise, though clinical validation remains limited to early-phase studies.

Future perspectives

How should we move forward? Despite compelling preclinical evidence implicating ferroptosis across diverse clinical phenotypes, translation to the bedside remains limited. Clinical studies have largely been restricted to iron chelation, and no trial has yet prospectively enrolled patients based on ferroptosis biomarker stratification while using approved ICU medications specifically targeting ferroptosis. Several ongoing clinical trials are evaluating agents with possible anti-ferroptotic properties on hard clinical outcomes, as summarised in Table 1.

Table 1.

Ongoing clinical trials investigating agents with suggested anti-ferroptotic potential relevant for critical care

Investigators	Pathology	Intervention	Control	Ref
Ischaemia–reperfusion injury
Indiana University School of Medicine	Acute myocardial infarction	Deferiprone	placebo	[196]
George Washington University	Cardiac surgery associated AKI	CoQ10 + glutathione	placebo	[197]
Nanjing Medical University	Cardiac and aortic surgery	CoQ10	placebo	[198]
Neurological emergencies
University of Michigan	Subarachnoid haemorrhage	Deferoxamine	placebo	[199]
University of Illinois at Chicago	Spontaneous intracerebral haemorrhage	Minimally invasive surgery + deferoxamine	standard care	[200]
The George Institute	Intracerebral haemorrhage	(1) Deferoxamine (2) Colchicine (3) Deferoxamine + colchicine	placebo	[201]
Duke University	Subarachnoid haemorrhage	Deferiprone	placebo	[202]

^*The search terms used are listed in the Search strategy section

A key obstacle to clinical progress is the heterogeneity of ferroptosis signatures among patients. Meaningful trials will require well-defined inclusion criteria based on validated ferroptosis biomarkers. The recently published ImmunoSep trial, in which septic patients were stratified according to their immunogenic profile and assigned to targeted immunomodulatory therapy, demonstrated that biomarker-based enrichment is feasible in the ICU [194], a model that ferroptosis research should aim to replicate. Current biomarker assays, however, lack the sensitivity, specificity, and bedside applicability required for routine patient stratification. One ongoing observational study is prospectively assessing ferroptosis signatures in a broad cohort of critically ill patients [195], and its results will be important for defining clinically meaningful thresholds. A multimarker approach integrating ferroptosis, pyroptosis, and necroptosis signatures alongside clinical markers of organ dysfunction is likely necessary.

For specific ferroptosis inhibitors, the most feasible entry point is through populations with well-defined, homogeneous ferroptosis signatures and predictable therapeutic windows. Solid organ transplantation remains the clearest candidate: the pathophysiology is delineated, the timing of reperfusion is anticipated, and early human perfusion data already support ferroptosis inhibition as a tractable intervention [47]. Elective cardiac surgery and acute ischaemic stroke may represent additional contexts where inhibitor administration can be timed to reperfusion. Among available compounds, FXT-001 has demonstrated efficacy in human perfusion models but has not yet entered formal clinical trials, while edaravone—already approved for clinical use in other indications—represents a more immediately accessible candidate given its established safety profile. Once robust patient stratification is achievable, the field will be better positioned to evaluate whether agents with anti-ferroptotic properties can improve outcomes in the broader critically ill population.

Conclusion

Since ferroptosis was conceptualised in 2012, a growing body of evidence indicates that this iron-dependent RCD is a significant pathological factor in critical illness. Numerous studies have explored its biochemical pathways and preclinical strategies for inhibition. The next step is translating the preclinical findings into the hospital setting, particularly in the ICU, to gain a clinical perspective.

Key questions remain unresolved. For example, how does ferroptosis interact with other cell death pathways? Sepsis, for instance, involves a complex interplay of multiple cell death mechanisms. Furthermore, what proportion of ICU patients exhibit signs of ferroptosis and how can we correctly measure them? Prior to initiating clinical trials with ferroptosis inhibitors, comprehensive and robust biomarker profiling is essential to enable predictive enrichment and avoid ineffective outcomes from generalised treatment approaches. Additionally, timing may be critical: understanding biomarker dynamics in relation to clinical parameters could clarify optimal intervention windows.

While many challenges persist, it is clear that ferroptosis and, by extension RCD, are becoming integral to ICU research. The immediate priority is stratifying patients based on clinical and biochemical parameters to ensure targeted treatment approaches. Improving our understanding of molecular cell death mechanisms could represent a significant step toward precision medicine in critical care.

Abbreviations

4-HNE

4-Hydroxy-2-nonenal

ACE

Angiotensin-converting enzyme

ACSL4

Acyl-CoA synthetase long-chain family member 4

AKI

Acute kidney injury

ARDS

Acute respiratory distress syndrome

ATP

Adenosine-triphosphate

BALF

Bronchoalveolar lavage fluid

BSCB

Blood-spinal cord barrier

CABG

Coronary artery bypass graft

CLP

Caecal ligation puncture

CoQ₁₀

Ubiquinone

CSF

Cerebrospinal fluid

EVLP

Ex vivo lung perfusion

Fe²⁺

Ferrous iron

Fe³⁺

Ferric iron

Fer-1

Ferrostatin-1

FPN

Ferroportin

FSP1

Ferroptosis suppressor protein 1

FTL

Ferritin light chain 1

GPX4

Glutathione peroxidase

GSH

Glutathione

HMOX1

Haem oxygenase 1

ICU

Intensive care unit

IL

Interleukin

I/R

Ischaemia-reperfusion

LPCAT3

Lysophosphatidylcholine acyltransferase 3

LIP

Labile iron pool

Lip-1

Liproxstatin-1

LOOH

Lipid peroxides

LOX

Lipoxygenase

LPS

Lipopolysaccharides

MDA

Malondialdehyde

MODS

Multiple organ dysfunction syndrome

NADPH

Nicotinamide adenine dinucleotide phosphate

NCOA4

Nuclear coreceptor activator 4

NRF2

Nuclear factor erythroid 2-related factor 2

PCI

Percutaneous coronary intervention

PL

Phospholipid

PLOOH

Phospholipid hydroperoxides

PTGS2

Prostaglandin-endoperoxide synthase 2

PUFA

Polyunsaturated fatty acids

RCD

Regulated cell death

ROS

Reactive oxygen species

RTA

Radical trapping agents

SAE

Sepsis-associated encephalopathy

SAH

Subarachnoid haemorrhage

SCI

Spinal cord injury

SCM

Septic cardiomyopathy

SLC7A11

Solute carrier family 7 member 11

STAT3

Signal transducer and activator of transcription 3

TBI

Traumatic brain injury

TfR1

Transferrin receptor 1

Author contributions

All the authors have read and approved the final article. They agree to be accountable for all aspects of the work and confirm that all those entitled to authorship are listed as authors. TS, TVB, PGJ conceptualisation; TS investigation; TVB, PGJ supervision; TS, GV visualisation; TS writing—original draft; CW, MW, GV, SL, EH, DDB, TVB, PGJ writing—review and editing. The authors thank Ine Koeken for her assistance with visualisation.

Funding

TS and CW are supported by BOF-IMPULS from the University of Antwerp. GV received funding from the Fund for Scientific Research (FWO) Flanders (1SH9524N). TVB lab is supported by Strategic Basic Research Foundation Flanders, IRONIX, S001522N, Excellence of Science MODEL-IDI and CD-INFLADIS; Consortium of Excellence at the University of Antwerp INFLA-MED; Industrial Research Fund, BOF-SEP and BOF-IMPULS from the University of Antwerp; Foundation against Cancer F/2022/2067.

Availability of data and materials

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

TVB owns patents related to ferroptosis inhibitors and is the founder of FiriX Therapeutics. All the other authors declare that they have no financial or non-financial interests related to this manuscript.