Neonatal Hypoxic-Ischemic Encephalopathy: From the Limitations of Therapeutic Hypothermia to Precision Intervention

Abstract

Therapeutic hypothermia (TH) is the current standard of care for moderate-to-severe neonatal hypoxic-ischemic encephalopathy (HIE). However, its clinical application faces significant limitations, including a 44–55% rate of death or disability, a narrow 6‑h therapeutic window, inequitable global access, and inconsistent patient selection. These challenges highlight the need for emerging strategies that may complement or integrated with TH in the future. This review systematically synthesizes emerging therapeutic strategies beyond TH, encompassing four cutting-edge directions: first, targeting mitochondrial dysfunction and metabolic reprogramming; second, modulating neuroinflammation and immune responses; third, inhibiting programmed cell death pathways; and fourth, promoting neural repair and regeneration. We also discuss combination regimens, including sildenafil, resveratrol, and melatonin with magnesium, which have shown promise in preclinical or early clinical studies, while acknowledging that erythropoietin failed to show added benefit in Phase III trials. Importantly, most of the novel interventions discussed remain at preclinical or early‑phase clinical stages. To facilitate future translation, we systematically delineate the critical support infrastructure for HIE management in the precision medicine era: multimodal biomarkers and risk stratification, including neuroimaging, neuroelectrophysiology, fluid biomarkers, and clinical risk scoring models. By integrating these tools, it will be possible to rapidly identify the dominant injury pathway in each infant within the “golden 6-hour” window, implement individualized combination interventions, and promptly deploy neurorestorative strategies during the subacute phase. Realizing this vision will require global collaboration, standardized data sharing, and low-cost interventions tailored for LMICs. In conclusion, while precision medicine holds great promise for HIE, it is not yet an established replacement for therapeutic hypothermia; rather, it represents a scientifically rational and evolving strategy that requires rigorous clinical validation to improve long‑term outcomes for affected neonates.

Graphical Abstract

Introduction

Neonatal HIE is a severe brain injury disorder resulting from perinatal asphyxia and remains the leading cause of neonatal mortality and lifelong neurodevelopmental disabilities, including cerebral palsy, intellectual impairment, and epilepsy.1 This condition not only imposes substantial psychological and economic burdens on affected families but also represents a major challenge in global public health. Although significant progress has been made over the past decades in understanding the pathophysiological mechanisms underlying HIE, effective clinically available therapeutic options remain extremely limited. This stark reality urgently calls for a critical re-evaluation of the limitations of current therapies and an active exploration of more effective intervention strategies.

Since 2005, multiple multicenter randomized controlled trials have demonstrated that therapeutic hypothermia (TH), initiated within six hours of birth and maintained for 72 hours to lower the core body temperature to 33.5–34.5°C, significantly reduces the composite outcome of death or severe disability in term infants with moderate or severe HIE.2 The NICHD trial reduced the primary outcome from 62% to 44%;3 the CoolCap trial showed benefit in infants with less severe aEEG changes,4 but not in the most severe subgroup; and the larger TOBY trial confirmed these neuroprotective effects in 325 infants.5 On the basis of this robust evidence, TH has been adopted as the standard of care for HIE in international guidelines.2 As the current standard of care, TH provides the essential clinical foundation upon which emerging mechanism‑based adjunctive therapies should be developed. Nevertheless, TH is far from a panacea, and its clinical application faces several important limitations. First, even under optimal conditions, approximately 44–55% of infants treated with TH still die or survive with moderate-to-severe disability,6 suggesting an inherent ceiling effect for the neuroprotective efficacy of monotherapy. Second, in Low- and Middle-Income Countries (LMICs), where healthcare resources are scarce, large-scale clinical trials have failed to confirm the efficacy of TH; indeed, some studies have even observed an increased risk of mortality, a finding that starkly contrasts with results from high-income countries.7 This discrepancy underscores the context-dependent heterogeneity of TH’s effectiveness across different healthcare settings. Furthermore, the clinical benefit of TH for mild HIE remains controversial,8 and robust evidence regarding its safety and efficacy in preterm infants is still lacking. These limitations do not diminish the value of TH; rather, they underscore the urgent need to develop novel, effective, and broadly applicable strategies that build upon this established therapeutic foundation.

Fortunately, recent years have witnessed rapid advances in research methodologies, including single-cell sequencing, multi-omics technologies, and novel animal models, which have substantially deepened our understanding of the molecular mechanisms underlying secondary brain injury following neonatal HIE. The injury cascade involves multiple complex processes, such as energy metabolism failure, oxidative stress, excitotoxicity, neuroinflammation, and programmed cell death (eg, apoptosis, necroptosis, ferroptosis). Importantly, these pathological events have now been delineated at the level of subcellular organelles (eg, mitochondria) and specific signaling pathways. For instance, recent preclinical studies have revealed a critical role of cardiolipin remodeling in mitochondrial dysfunction in HIE and identified a novel astrocyte–neuron crosstalk mechanism mediated by thyroid hormone (T3).9,10 These findings provide abundant and specific intervention targets for the development of precision therapies. A deeper understanding of these novel molecular mechanisms is now opening avenues for integrating emerging mechanism‑based strategies alongside TH.

This review aims to systematically summarize the core challenges currently encountered in the clinical application of TH and, based on recent advances in basic and preclinical research, to explore potential future therapeutic strategies. We will focus on a range of emerging therapeutic directions beyond TH, which are designed to interrupt the injury cascade or promote neural repair at different levels. These cutting-edge strategies span diverse fields, including mitochondrial protection, modulation of inflammation, interventions targeting specific cell death pathways, and cell replacement along with regenerative medicine. Through a systematic synthesis and in-depth evaluation of these novel therapies, we hope to provide valuable insights into future research directions and clinical translation in this field, ultimately facilitating the evolution of HIE treatment from empirical cooling toward a multi‑target, mechanism‑based approach that complements the current standard of care.

Current Therapeutic Challenges

Although TH represents a milestone in the treatment of neonatal HIE, its clinical application and widespread adoption continue to face multiple challenges. Collectively, these limitations constrain its overall efficacy and generalizability, highlighting the urgent need to move toward more refined therapeutic strategies.

The “Ceiling Effect” of Efficacy and Remarkable Heterogeneity

Although TH has become the standard neuroprotective strategy for moderate-to-severe neonatal HIE and has been shown by numerous clinical studies to significantly improve survival rates and neurodevelopmental outcomes, its efficacy remains far from ideal, exhibiting a notable “ceiling effect” and interpatient heterogeneity.

The “Ceiling Effect”

Multiple long-term follow-up studies have revealed a sobering reality: even after receiving TH, infants with HIE remain at risk for a wide range of neurodevelopmental disorders as they grow. For example, a follow-up study of children treated with TH who did not develop cerebral palsy found that their communication skills—including structural and pragmatic language abilities—remained significantly poorer than those of healthy peers at 6–8 years of age, and that these deficits were closely associated with alterations in white matter microstructure.11 Broader evidence indicates that school-aged children following TH commonly exhibit cognitive dysfunction, executive function deficits, specific learning disabilities, and behavioral and psychological problems.12 Together, these findings point to a core limitation of TH: although it primarily acts on acute energy metabolism failure and excitotoxic injury, its capacity to modulate more persistent secondary injury mechanisms—such as chronic neuroinflammation, developmental apoptosis, and endogenous neurorepair processes—is very limited. Preclinical studies support this view; for instance, in animal models, the effect of TH on motor functional recovery is sex-dependent,13 suggesting that its neuroprotective action may be modulated by multiple biological factors. Thus, TH alone is insufficient to fully reverse the long-term neuropathological changes induced by HIE, underscoring the urgent need for adjunctive therapies targeting secondary injury pathways.

Remarkable Heterogeneity

Beyond the overall limitations on efficacy, the response to TH varies markedly among different patient subgroups. Studies have shown that the efficacy of TH differs according to sex and the pattern of brain injury (eg, basal ganglia-predominant versus watershed-predominant).14,15 Preclinical data further suggest that sex is an important determinant of the response to TH and adjunctive therapies, with female rats potentially exhibiting advantages in motor coordination recovery.13 This striking heterogeneity strongly indicates that the current “one-size-fits-all” approach to TH may not be optimal. Future directions urgently require a shift from uniform treatment protocols to more refined, stratified strategies based on biomarkers and injury phenotypes.16,17 For example, early identification of distinct injury patterns (using magnetic resonance imaging and electroencephalography features, among others) and biological markers (such as cardiolipin remodeling and neuroinflammatory status9) could enable the selection of subgroups most likely to benefit from TH and provide precise targets for combination therapies, thereby breaking through current therapeutic bottlenecks. In other words, understanding and leveraging this heterogeneity is a critical step in the transition from standardized treatment to individualized medicine.

A Strict and Narrow Therapeutic Time Window

The neuroprotective efficacy of TH is highly dependent on the timeliness of intervention, leading international guidelines to unanimously recommend initiating cooling within the golden 6-hour window after birth. However, this strict and narrow time window imposes exceptionally high demands on rapid identification, clinical decision-making, and interhospital transfer in real-world practice.2 In clinical settings, inherent delays in interhospital transfer, insufficient experience in recognizing neonatal HIE at primary hospitals, and the limited accuracy of neurological assessments within the first six hours of life collectively result in a substantial number of otherwise eligible infants missing the optimal treatment window.18 For instance, the French population-based LyTONEPAL cohort study revealed that nearly one-third of all TH-eligible infants did not receive timely treatment; notably, among infants born at non-cooling centres, the proportion of delayed TH initiation reached as high as 36.2%.18 Given this formidable challenge, some studies have begun exploring the feasibility of extending the therapeutic window to 6–12 hours post-birth. Nevertheless, current evidence suggests that delayed TH initiation yields suboptimal outcomes in infants with severe HIE, indicating that extending the treatment window may not benefit the most severely affected patients.19 Furthermore, even in centres equipped to deliver TH, the time elapsed across various steps—from birth to activation of the transport team, and from team activation to admission—has been closely correlated with significant delays in TH initiation.18 Collectively, these findings underscore the critical importance of optimising pre-hospital identification protocols, establishing efficient regional transport networks, and developing earlier and more precise adjunctive diagnostic tools to maximise TH efficacy and improve outcomes in HIE. Therefore, overcoming the biological constraints of the time window and addressing healthcare system barriers represent core challenges that urgently need to be resolved in the current clinical dissemination of TH.

Global Accessibility and Equity Challenges

Although TH represents the standard of care for moderate-to-severe HIE in high-income countries, its implementation is heavily dependent on expensive cooling devices, specialized neonatal intensive care teams, continuous neuromonitoring (eg, aEEG), and neuroimaging.2 However, even in high-income settings, the integration and widespread availability of these resources face considerable challenges; in LMICs, TH remains largely inaccessible.20 More alarmingly, a multicenter randomized controlled trial (HELIX) conducted in India, Sri Lanka, and Bangladesh revealed that TH not only failed to improve 18-month disability-free survival in infants with moderate-to-severe HIE but was also associated with higher mortality.20 This paradigm-shifting finding highlights the substantial risk of directly transplanting a high-income medical model into LMICs. The underlying reasons are likely multifactorial, encompassing distinct etiological spectra of HIE in LMICs (eg, a higher prevalence of intrauterine infection and chronic hypoxia rather than acute perinatal events), weak basic medical support, and suboptimal management of TH-related complications such as infections and hypoglycemia. Consequently, the development of low-cost, easy-to-administer, and intensive-care-independent alternative or adjunctive therapies is of paramount importance for the global fight against HIE.

Furthermore, even in relatively resource-rich settings such as the United States, significant racial and ethnic disparities exist in the application of TH. For example, a large retrospective cohort study from California demonstrated that Black infants have a 60% higher risk of developing HIE than White infants, and among those with moderate-to-severe HIE, Black infants are significantly less likely to receive TH than their White counterparts,21,22 Another study corroborated these findings, showing that Black infants have both a higher incidence and greater severity of HIE, whereas Native American infants are substantially less likely to receive TH.22 These deeply entrenched healthcare inequities further compound the fairness dilemma surrounding TH utilization, with underlying drivers including systemic barriers to care, implicit biases, and socioeconomic disparities—factors that urgently require policy-level attention and systematic intervention. Thus, whether across national borders or hidden within the racial divides of developed societies, these disparities profoundly underscore the formidable challenges that current TH-centered HIE treatment strategies face in achieving global health equity. In doing so, they also provide compelling sociological and epidemiological justifications for accelerating the development of precision-based and universally accessible therapeutic interventions.

Variability in Patient Selection Criteria and Associated Potential Risks

Although TH is recommended by international guidelines for the treatment of moderate-to-severe neonatal HIE,2 significant heterogeneity persists across clinical centers worldwide regarding the inclusion criteria for TH. These criteria broadly encompass perinatal events, blood gas analysis results, Apgar scores, neurological examinations, and amplitude-integrated electroencephalography (aEEG).23 This lack of standardization leads to a series of clinical dilemmas: on one hand, some infants who do not fully meet the optimal indications may be unnecessarily exposed to the potential risks of TH; on the other hand, some infants who might benefit from TH—particularly those with mild HIE—may be excluded from treatment.23,24 Notably, whether infants with mild HIE should receive TH remains a focal point of clinical debate, and there are substantial regional differences in the classification of HIE severity and the weight given to neuroelectrophysiological assessments, which directly affects the accuracy of epidemiological data and the reliability of cross-center comparative studies.23 Consequently, the heterogeneity in patient selection criteria in current clinical practice not only constitutes a major barrier to evidence-based medicine translation but also underscores the urgent need for a globally unified and precise stratification system for HIE severity.

Regarding the appropriateness of TH for infants with mild HIE, existing clinical evidence presents a marked contradiction. Specifically, some studies indicate that TH reduces abnormal findings on magnetic resonance imaging (MRI) in infants with mild HIE, particularly conferring protective effects in watershed regions.25 However, other large-scale population-based studies have failed to demonstrate significant short-term clinical benefits of TH for mild HIE, such as any notable difference in the incidence of neonatal seizures.24 This uncertainty in the evidence places clinicians in a difficult position when making decisions. Beyond the uncertainty surrounding indications, TH itself is not a risk-free intervention. Available studies suggest that TH may increase the risks of cardiac arrhythmias, coagulation dysfunction (which may in turn elevate the risk of intracranial hemorrhage), electrolyte disturbances (such as hypocalcemia, hypomagnesemia, and hyperphosphatemia), and thrombocytopenia.26,27 Therefore, in the absence of a clear risk–benefit assessment, the indiscriminate extension of TH to the mild HIE population carries potential hazards, and future research must focus on identifying mild HIE subtypes that are most likely to truly benefit from TH.

In specific clinical contexts, the risks associated with TH implementation are particularly pronounced. For example, when HIE is complicated by hemorrhagic shock resulting from severe anemia, the administration of TH may mask or exacerbate underlying intracranial hemorrhage. Case reports have indicated that although no hemorrhage was detected by bedside cranial ultrasound before the initiation of TH, subsequent computed tomography (CT) or MRI revealed extensive subdural hematomas, presenting a severe dilemma for clinical decision-making.26 Adding further complexity, there is substantial variability across centers in the adjunctive support strategies employed during TH, and these factors can exert a non-negligible impact on patient outcomes. Specifically, hemodynamic management is challenging because TH itself affects cardiac output and blood pressure; some infants may require inotropic support, yet there is no current consensus on optimal blood pressure targets or drug selection.28 These findings caution that the neuroprotective effects of TH are highly dependent on meticulous adjunctive management, and that any lapse in a single component could offset the benefits of cooling, particularly in critically ill infants with multiorgan injury.

Variations in adjunctive support strategies during TH manifest in several specific areas. First, in respiratory management, particular attention must be paid to the potential harm of hypocapnia, as studies have shown a strong association with adverse outcomes in infants with HIE; therefore, hyperventilation should be strictly avoided during TH.29,30 Second, glycemic control is equally critical, as marked fluctuations (including both hyperglycemia and hypoglycemia) and high glycemic variability are associated with poor neurodevelopmental outcomes, necessitating meticulous management.31 Furthermore, the use of sedative and analgesic drugs lacks standardization across centers with respect to drug selection, dosage, and long-term effects on neurodevelopmental outcomes.32 Regarding the timing of enteral nutrition initiation, one randomized controlled trial indicated that early initiation of enteral feeding did not increase the risk of feeding intolerance compared with delayed feeding and might shorten hospital stays; nevertheless, further research is needed in this area.33 Collectively, respiratory, metabolic, circulatory, and nutritional management strategies during TH exhibit extensive practice heterogeneity, and this “therapeutic uncertainty” may be an important reason for the diminished overall efficacy of TH and for inter-center variability in outcomes.

Finally, the rewarming process following TH itself constitutes an independent risk period. Animal studies have shown that rewarming can induce cortical neuronal apoptosis via caspase-dependent pathways, and if rewarming is too rapid, this apoptotic effect may attenuate or even completely negate the neuroprotective benefits conferred by TH.34 Clinical studies further confirm that an increased risk of seizures during rewarming is significantly associated with poor outcomes in infants, indicating that rewarming rates must be tightly controlled in clinical practice and that patients’ physiological changes should be closely monitored.35 In summary, as detailed in this section—from the inconsistency in patient selection criteria and the contradictory evidence for treating mild HIE, to the risks inherent in TH itself and its adjunctive management, and finally to the potential injury during rewarming—the current TH treatment paradigm reveals substantial limitations. Collectively, these limitations point to a core issue: the existing “one-size-fits-all” approach to therapeutic hypothermia is no longer sufficient to meet clinical needs. Accordingly, there is an urgent need to transition toward the integration of biomarker‑guided precision approaches to complement the current standard of care, so as to maximize therapeutic efficacy while minimizing unnecessary risks (Figure 1).

Figure 1.

Limitations of Therapeutic Hypothermia in Neonatal HIE. This figure illustrates the four major categories of limitations associated with therapeutic hypothermia (TH), the current standard of care for moderate-to-severe neonatal HIE. These include: efficacy ceiling (44–55% still die or suffer disability); narrow 6-hour therapeutic window; inconsistent patient selection and risks (eg, arrhythmias, rewarming injury); and global accessibility and racial/economic disparities. All figures were created in BioRender. (2026) https://BioRender.com/vcdpdnn.

Future Therapeutic Strategies

Given the clinical limitations of therapeutic hypothermia, current research efforts have shifted toward developing combinatorial treatment strategies that can serve as adjuncts to hypothermia. These emerging approaches aim to precisely target multiple key nodes in the complex pathophysiological cascade of HIE, including mitochondrial dysfunction, neuroinflammation, distinct forms of cell death (eg, apoptosis, necroptosis, ferroptosis), and the promotion of neural repair and regeneration. Through multimodal combination interventions, it may be possible to overcome the efficacy bottlenecks of monotherapies and offer more comprehensive neuroprotection for neonates with HIE (Figure 2 and Table 1).

Figure 2.

Emerging Therapeutic Strategies Beyond Hypothermia. This figure categorizes future therapeutic strategies for neonatal HIE into four domains: (A) targeting mitochondrial dysfunction: cardiolipin remodeling (Preclinical), ketone bodies (Preclinical), GLP-1 agonists (Investigational); (B) modulating neuroinflammation: MG1-KLA (Preclinical), C5aR1 antagonist (Preclinical), sildenafil (Investigational); (C) inhibiting programmed cell death: ferroptosis and necroptosis (Preclinical); and (D) promoting neural repair: stem cells (Investigational, Phase I/II), p75ECD-Fc (Preclinical), melatonin (Investigational). All figures were created in BioRender. (2026) https://BioRender.com/vcdpdnn.

Table 1.

Level of Evidence for Major Therapeutic Strategies in Neonatal HIE

Therapeutic Strategy	Mechanism	Level of Evidence
Therapeutic hypothermia (TH)	Multitarget neuroprotection (energy metabolism, excitotoxicity)	Meta-analysis/Guideline-supported
Edaravone	Free radical scavenger (ferroptosis inhibition)	Preclinical
Necrostatin-1 (Nec-1)	RIPK1 inhibitor (necroptosis inhibition)	Preclinical
MG1-KLA peptide	Targeted depletion of pro-inflammatory microglia	Preclinical
PMX205 (C5aR1 antagonist)	Complement C5aR1 antagonism (anti-neuroinflammation)	Preclinical
p75ECD-Fc fusion protein	p75NTR pro-apoptotic signaling blockade	Preclinical
4-Octyl itaconate (4OI)	Nrf2 activator (antioxidant/anti-inflammatory)	Preclinical
Cardiolipin remodeling	Mitochondrial inner membrane stabilization	Preclinical
Exosomes	MSC paracrine mediators (cell-free repair)	Preclinical
GLP-1 receptor agonists (Exenatide, Semaglutide)	PI3K/AKT pathway activation (metabolic modulation)	Preclinical
DA5-CH (GLP-1/GIP dual agonist)	TLR2/NF-κB/NLRP3 pathway inhibition	Preclinical
Ketone bodies/Ketogenic diet	Alternative energy substrate; anti-inflammatory, antioxidant	Preclinical
TAK-242	TLR4 inhibitor (anti-neuroinflammation)	Preclinical
Vitamin D/VDR agonist	Antioxidant (GPX4 restoration, iron metabolism regulation)	Preclinical
Sildenafil	PDE5 inhibitor; PI3K/AKT/mTOR pathway activation	Phase I/II (repurposing)
Melatonin	Antioxidant; CDK5 inhibition; AQP4 polarity preservation	Phase I/II (repurposing)
Mesenchymal stem cells (MSCs)	Paracrine effects (immunomodulation, angiogenesis, etc.)	Phase I/II
Magnesium + Melatonin	Anti-excitotoxicity + antioxidant synergy	Phase I/II (small RCT)
Resveratrol	Antioxidant; improved cerebral oxygen delivery	Preclinical
Erythropoietin (EPO)	Anti-apoptotic; neurogenesis promotion	Phase III (negative)

Targeting Mitochondrial Dysfunction and Metabolic Remodeling

Mitochondria are central organelles in ischemia-reperfusion injury, and their dysfunction plays a critical role in the pathogenesis of neonatal hypoxic-ischemic brain damage.36 Specifically, following hypoxic-ischemic insult, the aberrant opening of the mitochondrial permeability transition pore (mPTP) rapidly leads to cellular energy depletion, a burst of reactive oxygen species (ROS), and the subsequent activation of multiple programmed cell death pathways. Therefore, a detailed understanding and targeted modulation of specific aspects of mitochondrial damage have become key directions for developing novel neuroprotective strategies.

Cardiolipin Remodeling

Cardiolipin is a phospholipid uniquely localized to the inner mitochondrial membrane and is essential for maintaining the structural integrity of respiratory chain supercomplexes, electron transfer efficiency, and mitochondrial dynamics. However, the acyl chains of cardiolipin are highly susceptible to oxidative stress, and abnormalities in its remodeling process can directly impair mitochondrial function. Recently, a cross-species study integrating in vitro, small animal, and large animal HIE models has systematically revealed, for the first time, significant pathological alterations in the cardiolipin profile following HIE injury: specifically, a marked increase in the ratio of monolysocardiolipin to cardiolipin, along with an accumulation of saturated cardiolipin species.9 Notably, this study further demonstrated that in mice with a conditional knockout of the gene encoding tafazzin—a key enzyme in cardiolipin remodeling—the pathological accumulation of monolysocardiolipin significantly exacerbates cerebral infarct volume following HIE. These findings not only establish a direct link between abnormal cardiolipin remodeling and the severity of HIE but also suggest that the monolysocardiolipin-to-cardiolipin ratio may serve as a novel biomarker of mitochondrial injury. Consequently, this provides a theoretical foundation for the development of drugs aimed at stabilizing mitochondrial membrane architecture or modulating cardiolipin composition.

Metabolic Modulators

Beyond directly preserving mitochondrial membrane architecture, modulating systemic or local metabolic states to enhance neuronal tolerance to hypoxia represents a promising therapeutic avenue. One such approach involves the ketogenic diet and ketone bodies, which—including β-hydroxybutyrate and acetoacetate—not only serve as alternative energy substrates efficiently utilized by neurons under hypoxic conditions but also exert anti-inflammatory, antioxidant, and anti-apoptotic effects through multiple pathways, such as inhibiting histone deacetylases and activating G protein-coupled receptors. A substantial body of preclinical studies consistently demonstrates that, in animal models of HIE, both preconditioning with a ketogenic diet and direct administration of ketone bodies confer significant neuroprotection;37,38 given their safety, low cost, and ease of implementation, ketogenic diet or ketone-based therapies hold considerable promise in LMICs as effective adjuncts to standard therapies like therapeutic hypothermia.

Another metabolic strategy gaining widespread attention is the use of glucagon-like peptide-1 receptor agonists (eg, exenatide, semaglutide), which are novel anti-diabetic agents with demonstrated neuroprotective effects in adult stroke models. Recent studies confirm that, in murine HIE models, administration of exenatide or semaglutide immediately following hypoxia-ischaemia significantly improves neuropathological scores, enhances survival, and restores motor function; mechanistic investigations indicate these protective effects involve activation of the PI3K/AKT signalling pathway and increased intracellular cAMP levels.39 The AMPK/SIRT1/PGC-1α signaling pathway, a central regulator of mitochondrial biogenesis and energy homeostasis, has been extensively implicated in neuroprotection and may represent a convergent mechanism through which metabolic modulators exert their beneficial effects.40 Moreover, a separate study shows that a novel GIP/GLP-1 dual receptor agonist (DA5-CH) attenuates HIE injury by suppressing TLR2/NF-κB/NLRP3 pathway-mediated neuroinflammation,41 collectively providing robust preclinical evidence for repurposing these well-tolerated agents as adjuncts to therapeutic hypothermia.

In summary, strategies targeting mitochondrial dysfunction and metabolic remodelling—such as cardiolipin repair, ketone body supplementation, and the use of GLP-1 receptor agonists—constitute a highly translatable therapeutic arsenal. Future research should further define the optimal timing, dosage, and long-term safety of these interventions when combined with therapeutic hypothermia, thereby paving the way for an era of personalized precision medicine.

Modulation of Neuroinflammation and Immune Responses

Post-resuscitation neuroinflammation is a major driver of secondary injury following HIE, characterized by persistent microglial activation, reactive astrogliosis, and infiltration of peripheral immune cells.42,43 These interconnected processes form a self-amplifying injury cascade that exacerbates neuronal damage and impedes neurorepair. Consequently, precisely targeting distinct nodes of neuroinflammation has emerged as a highly promising therapeutic strategy.

Targeting Microglia

As the resident immune cells of the central nervous system, microglia play a “double-edged sword” role in the pathogenesis of neonatal HIE. Upon hypoxic-ischemic challenge, they can polarize toward a pro-inflammatory (M1) phenotype, releasing copious inflammatory cytokines and thereby exacerbating brain injury. Therefore, selectively eliminating harmful pro-inflammatory microglial subsets or modulating their function has become a key intervention point. For instance, an innovative study exploited the ability of the MG1-targeting peptide to specifically bind to pro-inflammatory microglia. By conjugating MG1 with the pro-apoptotic peptide KLA, the researchers successfully constructed an MG1-KLA complex. This complex precisely recognizes and binds to pro-inflammatory microglia, inducing their apoptosis for selective elimination. In a mouse model of HIE, intracerebroventricular injection of MG1-KLA not only significantly reduced brain atrophy and improved motor function but also downregulated the expression levels of pro-inflammatory cytokines.44 This strategy of “precision targeting” of harmful cell subsets theoretically offers higher specificity and fewer systemic side effects compared with conventional broad-spectrum anti-inflammatory drugs, thereby opening new avenues for immunomodulatory therapy in HIE. Concurrently, inhibiting excessive complement system activation has also proven to be an effective intervention strategy. A critical step in the complement cascade is the binding of the anaphylatoxin C5a to its receptor C5aR1, which amplifies the neuroinflammatory cascade. A study conducted in a rat HIE model demonstrated that treatment with the C5aR1-specific antagonist PMX205 significantly improved fine motor function, balance, and exploratory behavior in animals, while also alleviating neuronal injury in the cortical region.45 Collectively, these targeted therapeutic strategies directed at microglia and the complement system, by specifically inhibiting distinct facets of neuroinflammation, show great promise as precision interventions for HIE.

Nevertheless, despite the encouraging outcomes achieved with complement-targeted strategies, this study also revealed an important limitation: although PMX205 exerted neuroprotective effects, injury to the hippocampal CA1 region persisted, and the therapeutic time window appeared to be narrow.45 These findings underscore the marked heterogeneity of complement cascade involvement across different brain regions and injury time points, suggesting the need for more refined intervention strategies that may target specific brain areas or injury stages. Moreover, although the MG1-KLA therapy targeting pro-inflammatory microglia has shown excellent results in preclinical models, its clinical translation still faces challenges such as drug delivery efficiency, long-term safety assessment, and differential efficacy among HIE subtypes. Therefore, future research should not only continue to explore additional targets for modulating microglial function—for example, by altering their activation status through metabolic or epigenetic modifications—but must also integrate advanced neuroimaging and biomarkers to achieve precise stratification of HIE patients. Such stratification will enable matching of individual pathological profiles with the most appropriate immunomodulatory regimens, ultimately facilitating the transition from the limitations of therapeutic hypothermia toward a new era of precision intervention.

Modulation of Other Inflammatory Pathways

Beyond directly targeting microglia and the complement system, modulating broader inflammatory signalling cascades has opened up multiple promising new strategies for the treatment of HIE. For instance, sildenafil, a well-established phosphodiesterase type 5 inhibitor, has demonstrated significant neurorestorative effects in a rat model of term neonatal HIE. Its core mechanism involves activating the PI3K/AKT/mTOR signalling pathway, thereby markedly reducing neuroinflammation, as evidenced by decreased numbers of reactive astrocytes and microglia, as well as reduced levels of key inflammatory cytokines such as tumour necrosis factor-alpha (TNF-α) and interleukin-1β (IL-1β). Ultimately, this improves white matter integrity and promotes oligodendrocyte survival and maturation.46 Similarly, the itaconic acid derivative 4-octyl itaconate (4OI) exerts comparable protective effects in a mouse model of HIE, primarily by activating the nuclear factor erythroid 2-related factor 2 (Nrf2) pathway in astrocytes, thereby achieving both antioxidant and anti-inflammatory effects, which effectively alleviates cognitive dysfunction induced by HIE.47 Collectively, targeting post-resuscitation neuroinflammation through diverse strategies has become a crucial direction for improving HIE outcomes. Future research should focus on elucidating the spatiotemporal dynamics of different inflammatory pathways across distinct stages and brain regions following HIE.

Building on this foundation, novel multi-target agents provide compelling evidence for HIE therapy. Notably, the dual agonist of glucagon-like peptide-1 (GLP-1) and glucose-dependent insulinotropic polypeptide (GIP)—two important incretin hormones—has shown remarkable neuroprotective potential. In a groundbreaking study, the novel GLP-1/GIP dual agonist DA5-CH exerted significant protective effects in both cellular and animal models of HIE by inhibiting TLR2/NF-κB/NLRP3 pathway-mediated neuroinflammation.41 This finding not only provides new scientific rationale for treating HIE with agents possessing pleiotropic metabolic regulatory functions but also heralds a future direction toward precision-based combination therapies. Consequently, future studies should strive to develop more precise and safer combination treatment strategies grounded in a thorough understanding of the spatiotemporal heterogeneity of inflammatory pathways, thereby maximally improving the long-term prognosis of neonates with HIE.

Inhibition of Multiple Programmed Cell Death Pathways

The pathophysiology of neonatal HIE involves multiple complex programmed cell death mechanisms rather than a single pathway. Consequently, the targeted combinatorial inhibition of these death pathways has emerged as an important strategy to overcome the limitations of current therapies. Although TH remains the standard of care, its neuroprotective effects are incomplete. For infants with severe HIE, combining TH with drugs that target distinct cell death pathways may offer synergistic or additive neuroprotection.

Ferroptosis

Ferroptosis is a novel form of regulated cell death characterized by iron dependence and excessive accumulation of lipid peroxidation,48 and it plays a critical role in the pathological progression of neonatal HIE.49 Specifically, following neonatal HIE, disrupted iron metabolism and oxidative stress bursts collectively promote the occurrence of ferroptosis. Based on this mechanism, inhibiting key molecules in the ferroptosis pathway, such as the phospholipid hydroperoxide glutathione peroxidase GPX4, or employing iron chelators (eg, deferoxamine) and lipophilic antioxidants (eg, vitamin E, edaravone), has become a highly promising therapeutic direction.49,50 For instance, edaravone, a potent free radical scavenger, has demonstrated neuroprotective effects in preclinical models of neonatal HIE, with efficacy closely linked to the timing of administration—acute intervention with a short course (eg, two days) being more effective.51 Furthermore, exploration of agents that activate endogenous antioxidant pathways (eg, the Nrf2/HO-1 axis), such as the statin derivative 4-octyl itaconate, has shown potential to inhibit ferroptosis and mitigate neuroinflammation and oxidative damage by modulating microglial/astrocytic function.47 Collectively, as a key node in HIE injury, ferroptosis represents a robust theoretical foundation for the development of novel neuroprotective agents through precise targeted intervention.

Necroptosis

Distinct from ferroptosis, necroptosis is a form of programmed necrosis mediated by receptor-interacting protein kinases (RIPKs).52 Existing evidence suggests that necroptosis may play a particularly important role in severe HIE infants who respond poorly to TH. Notably, recent preclinical evidence indicates that in neonatal HIE models, combining TH with the necroptosis-specific inhibitor Necrostatin-1 (Nec-1) yields superior neuroprotective effects compared to TH alone. The mechanism involves Nec-1 specifically inhibiting activation of the RIPK1/RIPK3/mixed lineage kinase domain-like protein (MLKL) signaling pathway, thereby significantly reducing brain injury. Moreover, this combination therapy outperforms monotherapy in terms of brain volume preservation and neurological functional improvement.53 This finding carries important clinical translational implications, suggesting that for severe HIE infants with poor response to TH and consequently worse prognoses, combining TH with targeted inhibition of the necroptosis pathway may represent an effective “sensitizing” strategy to overcome the current therapeutic bottleneck of TH. Therefore, integrating TH with necroptosis inhibitors represents a highly promising direction to enhance the efficacy of the existing standard of care.

Cross-Regulation of Apoptosis and Autophagy

Beyond ferroptosis and necroptosis, classical apoptosis and autophagy also play complex and interactive roles in the pathological process of HIE. For example, studies have shown that bone marrow mesenchymal stem cells (BMSCs) can alleviate neuronal autophagic damage by downregulating SYNPO2 protein expression, thereby improving neurological dysfunction after HIE.54 Additionally, certain neuroprotective agents, such as the estrogen metabolite estetrol delivered via liposome encapsulation, can effectively reduce neuronal apoptosis and promote angiogenesis.55 These findings collectively suggest a complex crosstalk among different cell death pathways.56 Future therapeutic strategies should therefore not be confined to a single pathway but rather focus on the holistic regulation of the broader programmed cell death network. In this context, a deep understanding of the crosstalk mechanisms among apoptosis, autophagy, ferroptosis, and necroptosis will be pivotal for achieving precision intervention in HIE.

Promoting Neural Repair and Regeneration

In recent years, significant breakthroughs have been achieved in the therapeutic landscape of neonatal HIE. This section systematically elucidates the current state and future directions of emerging strategies, including stem cell-based therapies and neurotrophic factors.

Stem Cell-Based Therapies

Stem cell therapy represents one of the most translationally promising avenues in neonatal HIE. The proposed mechanisms of action have evolved beyond the initial “cell replacement” hypothesis to encompass multifaceted paracrine effects, including trophic support, immunomodulation, angiogenesis, and mitochondrial transfer.57,58 Current preclinical and clinical investigations have predominantly focused on two major categories: mesenchymal stem cells (MSCs) and umbilical cord blood mononuclear cells (MNCs). Preclinical studies have consistently demonstrated that transplantation of MSCs derived from human umbilical cord, bone marrow, or amniotic membrane effectively improves motor and cognitive functions in HIE animal models.59,60 Of particular note, a Phase I/II clinical trial conducted in Japan (the SHIELD trial) provided preliminary evidence for the safety and tolerability of CL2020, a product based on multilineage-differentiating stress-enduring (Muse) cells, in neonates with HIE, thereby furnishing crucial safety data for clinical application of MSCs.61 Although MNCs have garnered considerable attention due to their ease of procurement and lack of ethical concerns—and despite meta-analyses of preclinical studies showing that MNCs effectively improve cognitive and sensorimotor functions in HIE models—the translation from bench to bedside remains fraught with challenges.62 Notably, preclinical studies often employ substantially higher cell doses and typically do not combine MNCs with therapeutic hypothermia (the current standard of care), which likely accounts for the currently insufficient clinical evidence of efficacy.62 Consequently, the future trajectory of stem cell therapy lies in systematically optimizing cell dosage, route of administration, optimal therapeutic window, and combinatorial strategies with therapeutic hypothermia.

As a highly promising non-invasive delivery method, intranasal administration enables therapeutic agents or cells to bypass the blood–brain barrier and directly target the central nervous system, offering the advantages of operational simplicity and low equipment requirements—benefits particularly valuable in resource-limited settings. A comprehensive systematic review evaluated over 30 intranasal therapies in HIE animal models and concluded that cell-based therapies and certain growth factors hold substantial developmental promise.20 The same review also noted that translation from animal experiments to clinical application necessitates further investigations, particularly those incorporating comprehensive assessments of brain connectomics and neurobehavioral outcomes.20 To propel this field forward, future research must systematically address discrepancies in study design between preclinical and clinical studies—such as standardizing cell dosages, delineating optimal protocols for combination with therapeutic hypothermia, and leveraging advanced neuroimaging techniques to thoroughly assess therapeutic efficacy. However, several critical translational hurdles must be overcome before stem cell‑based therapies can be integrated into clinical practice. These include the risk of tumorigenicity with pluripotent stem cell‑derived products, immune compatibility of allogeneic cells in the neonatal host, manufacturing consistency with batch‑to‑batch reproducibility, and the current lack of long‑term safety data regarding cell persistence, biodistribution, and off‑target effects. Additionally, the optimal cell source, dosing regimen, and timing of administration relative to TH remain to be systematically determined. In summary, stem cell-based therapies, particularly when delivered via non‑invasive routes such as intranasal administration, represent a promising avenue for exploration that may eventually complement the current standard of care.

Neurotrophic Factors and Signaling Pathways

In addition to cell-based therapeutic strategies, molecular agents that directly target specific signaling pathways have shown remarkable neurorepair potential, offering promising adjunctive or alternative options to therapeutic hypothermia. For instance, the novel recombinant fusion protein p75ECD-Fc exerts its effects by blocking the pro-apoptotic signaling of the neurotrophin receptor p75NTR. In rodent models of HIE, p75ECD-Fc demonstrated even superior efficacy compared with hypothermia, not only significantly reducing brain injury and promoting neuronal survival and regeneration but also inducing the conversion of glial cells toward a protective phenotype (A2 astrocytes and M2 microglia), thereby exhibiting potent anti-inflammatory and neurorepair activities.63 Furthermore, breakthrough progress has been made in the study of the pleiotropic neuroprotective agent melatonin: recent findings reveal that melatonin enhances the interaction between aquaporin-4 (AQP4) and α-syntrophin by inhibiting CDK5 activity in a rat model of HIE, thereby preserving the polarized expression of AQP4 on astrocytic endfeet. This mechanism is essential for restoring the waste clearance function of the glymphatic–lymphatic system, ultimately alleviating brain edema and long-term brain atrophy.64 These discoveries not only consolidate the rationale for using melatonin as an adjunct to hypothermia but also highlight the substantial potential of targeting protein–protein interaction networks in neuroprotection.

Concurrently, other emerging molecular targets and signaling pathways are opening new avenues for precision intervention in HIE. For example, the GLP-1/GIP dual receptor agonist DA5-CH alleviates neuronal injury and cognitive dysfunction in HIE models by suppressing TLR2/NF-κB/NLRP3 pathway-mediated neuroinflammation.41 Notably, significant advances have also been made in precision therapies targeting specific immune cell subsets: a peptide therapeutic agent targeting pro-inflammatory microglia (eg, MG1-KLA) can specifically induce apoptosis of activated microglia, thereby effectively curbing neuroinflammation and improving both brain atrophy and motor function in HIE model mice.44 Collectively, from blocking pan-apoptotic signaling to modulating glial phenotypes and precisely eliminating pathogenic immune cells, these diverse molecular strategies highlight the expanding repertoire of potential therapeutic targets for HIE. However, it must be emphasised that the vast majority of these approaches have been evaluated only in preclinical models, and their clinical applicability, safety, and efficacy remain to be established through rigorous translational research.

Combination Pharmacotherapy

Given the extreme complexity of the pathophysiology of neonatal HIE—which involves multiple interconnected and cascading processes, including energy metabolism failure, oxidative stress, neuroinflammation, apoptosis, and even ferroptosis—single-target therapies are often insufficient to completely halt the progression of brain injury. This challenge is not unique to HIE; in other ischemia-related conditions, the concept of an “ischemic cascade network” has been proposed to underscore the necessity of multi-target synergistic interventions.65 Consequently, the use of combination therapies that target different pathways,66 as a strategy aimed at achieving synergistic efficacy and overcoming the limitations of monotherapy, has become a research hotspot in the field. However, not all combination regimens have shown the anticipated promise, and the clinical translation of erythropoietin (EPO) has been particularly contentious.67 EPO was once considered one of the most promising adjunctive therapies to TH. Early preclinical studies and some meta-analyses suggested potential improvements in mental and psychomotor developmental indices.68 Nevertheless, the largest Phase III clinical trial to date (the HEAL trial) and its subsequent analyses yielded disappointing results: the addition of high-dose EPO to TH did not improve disability-free survival at 18–22 months of age in infants with moderate-to-severe HIE, nor did it reduce brain injury observed on MRI.69,70 These negative findings pose a serious challenge to the routine clinical application of EPO and suggest that future studies require more refined designs, such as focusing on specific subgroups or exploring different dosing regimens. Therefore, there is an urgent need to further explore and validate other combination strategies with distinct mechanisms of action.

In contrast to the dilemma surrounding EPO, targeting neuroinflammation and promoting white matter repair have emerged as important alternative directions. In this context, sildenafil, a phosphodiesterase-5 inhibitor, has demonstrated robust white matter protective effects in preclinical studies using a term-equivalent rat model of HIE. The study found that sildenafil treatment significantly attenuated microglial and astrocytic activation, reduced levels of pro-inflammatory cytokines such as TNFα and IL-1β, and supported oligodendrocyte survival and maturation by restoring p-AKT levels and activating the PI3K/AKT/mTOR pathway, thereby improving white matter integrity in structures such as the corpus callosum and external capsule.46 These findings suggest that sildenafil combined with TH holds promise as a neuroreparative strategy, offering new hope for improving long-term neurological outcomes in HIE survivors. Concurrently, natural product-derived agents have also shown unique value in exploratory studies. For instance, resveratrol, a natural polyphenol, has garnered attention for its antioxidant properties. Using optical monitoring via diffuse reflectance spectroscopy, a study found that resveratrol (40 mg/kg) effectively inhibited the progression of brain injury in a neonatal rat HIE model, potentially through enhanced cerebral oxygen delivery during hypoxic stress, as evidenced by beneficial modulation of total hemoglobin and tissue oxygen saturation in cerebral blood flow.71 This non-invasive optical monitoring approach also provides a novel tool for real-time assessment of drug efficacy, highlighting the potential of multimodal monitoring in precision intervention.

Furthermore, preliminary clinical evidence supports the strategy of combining different drugs to synergistically reduce biomarkers of brain injury. Melatonin, recognized for its potent antioxidant and anti-inflammatory properties, is considered a promising agent for HIE. A randomized controlled trial further explored the potential of combination therapy, demonstrating that adding magnesium sulfate to melatonin therapy more effectively reduced serum levels of S100-B protein (a biomarker of brain injury) in infants with moderate HIE compared with melatonin alone.72 This finding provides initial clinical evidence for the synergistic neuroprotective effects of magnesium and melatonin, supporting the feasibility of combining antioxidant and anti-excitotoxic strategies, and offers a valuable reference for developing multi-target intervention regimens. In summary, despite the failure of EPO clinical trials to demonstrate added benefit, combination strategies targeting distinct pathological mechanisms—such as sildenafil for white matter repair, resveratrol for improving oxygen delivery, and magnesium combined with melatonin for synergistic antioxidant effects—have all shown promise in preclinical or clinical studies.

Moreover, emerging therapeutic targets are continually being identified. For example, intervention strategies targeting ferroptosis have shown potential; the vitamin D/vitamin D receptor (VDR) signaling pathway, by promoting antioxidant defense, restoring GPX4 activity, and regulating iron metabolism, demonstrates promise as an adjunct to therapeutic hypothermia.49 Concurrently, targeted therapies against specific inflammatory signaling pathways have advanced in preclinical models, including TAK 242, which exerts neuroprotective effects by inhibiting the TLR4/MyD88/TRIF/NF-κB signaling pathway,73 and DA5-CH, a GLP-1/GIP dual receptor agonist that acts through inhibiting the TLR2/NF-κB/NLRP3 pathway.41 Additionally, a peptide therapy targeting pro-inflammatory microglia (MG1-KLA) has been shown to significantly reduce brain atrophy and improve motor function in a mouse model of HIE by selectively inducing apoptosis,44 providing a novel paradigm for precision targeting of neuroinflammation. Collectively, these diverse emerging strategies point toward an increasingly complex multi-target therapeutic network; nevertheless, substantial challenges remain in translating these preclinical findings into clinical practice.

In conclusion, although the large-scale EPO trial failed to confirm added benefit, combination therapy strategies targeting distinct pathological mechanisms—such as sildenafil for white matter repair, resveratrol for improving oxygen delivery, and magnesium combined with melatonin for synergistic antioxidant effects—have all shown promise in preclinical or clinical studies. Future research should leverage advanced neuroimaging and biomarker technologies to more precisely screen patients, assess therapeutic efficacy, and explore additional multi-target combination interventions grounded in the complex pathophysiology of HIE, thereby moving toward the goal of integrating precision‑based, combinatorial adjunctive strategies that build upon the current standard of care.

Despite the considerable promise of the emerging therapeutic strategies discussed above, it is imperative to critically appraise the substantial challenges that must be overcome before these approaches can be translated from bench to bedside. Most of the interventions remain at the preclinical stage, with evidence derived predominantly from rodent models of HIE. The translational relevance of these findings is inherently constrained by well‑recognised species‑specific differences in brain development, metabolic rates, immune responses, and the timing of myelination between rodents and humans. For instance, the neonatal rodent brain is substantially more immature at birth than the term human brain, and the pattern of hypoxic‑ischaemic injury in the widely used Rice‑Vannucci model (unilateral carotid ligation followed by systemic hypoxia) does not fully recapitulate the global, multifocal injury seen in term neonatal HIE. Although large‑animal models more closely approximate human cerebral development and injury patterns, their higher cost, logistical complexity, and limited availability constrain their use in early‑stage translational research. However, it is critical need for rigorous validation across multiple species before progressing to clinical evaluation. Beyond the inherent limitations of experimental models, stem cell‑based therapies face a distinct set of safety, regulatory, and manufacturing hurdles that require particularly careful consideration. These translational barriers do not diminish the scientific promise of the strategies reviewed; rather, they underscore the importance of a measured, stepwise approach to clinical development. Future research must prioritise validation in large‑animal models, standardisation of cell products, rigorous long‑term safety monitoring, and careful patient stratification to ensure that the transition from preclinical promise to clinical reality is both safe and effective.

Management of HIE in the Precision Medicine

As understanding of the pathophysiology underlying neonatal HIE continues to deepen, future therapeutic strategies are increasingly moving towards individualised approaches. Central to this paradigm shift is the integration of multimodal biomarkers to enable early, accurate risk stratification of affected infants, thereby guiding critical treatment decisions. Such decisions include whether to initiate TH, whether to combine TH with other interventions, and when to counsel families regarding prognosis or consider adjusting life-support measures (Figure 3 and Table 2).

Figure 3.

Precision Medicine Infrastructure for HIE Management. This figure outlines the core components of precision medicine–based HIE management: multimodal biomarkers (MRI, DTI, aEEG, cardiolipin, GFAP, UCH-L1, microRNAs); clinical risk scores; early stratification within the golden 6-hour window; and individualized combination therapy targeting dominant injury pathways. All figures were created in BioRender. (2026) https://BioRender.com/vcdpdnn.

Table 2.

Biomarkers for Precision Management of Neonatal HIE

Biomarker	Biological Rationale	Stage of Development	Clinical Applicability	Utility for Stratification and Decision-Making
Conventional MRI (T1/T2, DWI)	Detects injury patterns (basal ganglia/thalamic, watershed, hemispheric extensive) correlated with long-term outcomes	Clinically validated	Routinely used in TH era; predictive value unaffected by TH	Predicts death or IQ <70 at 6–7 years
DTI (Diffusion Tensor Imaging)	Assesses microstructural integrity of corticospinal tracts and optic radiations	Early clinical	Superior accuracy to conventional scoring; measurable within 4–7 days	Predicts neurodevelopmental outcomes at 24 months
CMRO2 (via TRUST MRI)	Measures cerebral metabolic rate of oxygen; reflects metabolic demand	Early clinical	Non-invasive assessment; correlates with 18-month outcomes	Potential biomarker for individual response to TH
Conventional EEG	Abnormal background activity predicts death or severe disability	Clinically validated	Bedside, continuous monitoring; combined with MRI improves prediction	Guides TH decision-making and prognostication
Quantitative EEG (qEEG)	Reduced background power (delta band) in temporal/parietal lobes reflects injury severity	Early clinical	Enables rapid, round-the-clock bedside decision tools (up to 98% accuracy)	Distinguishes mild from moderate/severe HIE; guides TH initiation
Cardiolipin profile (MLCL/CL ratio)	Reflects mitochondrial dysfunction and oxidative damage to inner mitochondrial membrane	Preclinical	Measurable in brain tissue; correlates with infarct volume	Potential biomarker for mitochondrial injury severity
GFAP (Glial Fibrillary Acidic Protein)	Astrocyte activation and glial injury marker	Clinically validated	Measurable in serum/CSF; associated with HIE severity and adverse prognosis	Prognostic biomarker for poor outcomes
UCH-L1 (Ubiquitin C-terminal Hydrolase L1)	Neuronal injury marker	Clinically validated	Measurable in serum/CSF; associated with HIE severity	Prognostic biomarker for injury severity
S100B	Astroglial injury marker	Clinically validated	Measurable in serum; elevated in severe HIE	Prognostic biomarker
Tau protein	Axonal injury marker	Clinically validated	Measurable in CSF; associated with adverse prognosis	Prognostic biomarker
NSE (Neuron-Specific Enolase)	Neuronal injury marker	Clinically validated	CSF levels remain predictive even during TH	Predicts brain injury severity and 2-year outcomes
sLOX-1 (Soluble Lectin-like Oxidized LDL Receptor-1)	Endothelial injury/inflammatory marker	Early clinical	Measurable within 6 hours of birth; reflects HIE severity	Predicts poor outcomes following TH
MicroRNAs (miR-374a-5p, miR-376c-3p)	Gene expression regulators; reflect injury pathways	Preclinical/Early clinical	Measurable in umbilical cord blood; multicentre validation completed	Enables early HIE identification and TH decision-making
Multidimensional risk models	Integrates Apgar score, EEG seizures, MRI findings, clinical exam	Clinically validated	Combines multiple clinical parameters for predictive modeling	Predicts cognitive, language, motor outcomes at 24 months
Early admission/discharge models	Uses routine clinical data to predict death or neurodevelopmental impairment	Clinically validated	AUC >0.85; applicable at early timepoints	Early identification of high-risk infants for targeted interventions

Neuroimaging

In the current era of TH, magnetic resonance imaging (MRI) remains the gold standard for assessing the extent of brain injury and predicting long-term outcomes. Injury patterns revealed by conventional T1-/T2-weighted imaging and diffusion-weighted imaging (DWI) are closely correlated with long-term outcomes. For example, a long-term follow-up study based on the NICHD Neonatal Research Network trial demonstrated that specific MRI injury patterns (eg, basal ganglia/thalamic injury, hemispheric extensive injury) effectively predict the risk of death or an IQ <70 at 6–7 years of age, with predictive value that is unaffected by whether the infant received TH.15 Furthermore, more advanced diffusion tensor imaging (DTI) studies have shown that, as early as the first 4–7 days after birth, microstructural integrity abnormalities in the corticospinal tracts and optic radiations not only effectively predict neurodevelopmental outcomes at 24 months of age but also exhibit accuracy superior to conventional basal ganglia/thalamic scoring.74 Notably, emerging MRI techniques such as T2-relaxation-under-spin-tagging (TRUST) enable non-invasive assessment of the cerebral metabolic rate of oxygen (CMRO₂).75 Studies have found that CMRO₂ is significantly reduced during TH, whereas CMRO₂ levels after rewarming are positively correlated with cognitive and motor scores at 18 months of age, positioning this metric as a potential physiological biomarker for assessing individual responses to TH.76 Collectively, from conventional MRI to DTI and cerebral metabolic imaging, multiparametric imaging assessments are providing an increasingly rich set of objective evidence for precise prognostication in HIE.

Neuroelectrophysiology

Early electroencephalography (EEG) holds irreplaceable value in HIE assessment, particularly regarding TH decision-making and prognostication. Multicentre studies have confirmed that abnormal background EEG is a powerful predictor of death or severe disability, and that combining EEG with MRI can construct highly reliable predictive models.77,78 At the clinical decision-making level, quantitative analyses based on early conventional EEG (eg, delta-band power) have already enabled the development of simple decision tools that distinguish mild HIE from moderate-to-severe HIE requiring TH with up to 98% accuracy, strongly supporting the feasibility of rapid, round-the-clock bedside decision-making.79 Concurrently, quantitative EEG (qEEG) analyses have revealed that background electrical activity is significantly reduced in the temporal and parietal lobes of infants with moderate-to-severe HIE, most notably in the delta band, providing objective neuroelectrophysiological metrics for assessing brain injury.80 Of particular importance, electrographic seizures occurring during the rewarming period have been found to be independently associated with an increased risk of death or moderate-to-severe disability at 18–22 months of age, underscoring the critical need to continue EEG monitoring throughout the rewarming phase.35 Therefore, extending EEG from traditional qualitative assessment towards quantitative analysis and dynamic monitoring represents a key step in optimising clinical management pathways for HIE.

Body Fluid Biomarkers

In the transition from therapeutic hypothermia toward a new paradigm of precision intervention, the discovery and application of body fluid biomarkers play a critically important role. As a direct reflection of mitochondrial dysfunction, the cardiolipin profile exhibits unique clinical potential. Preclinical studies have demonstrated that the ratio of monolysocardiolipin to cardiolipin (MLCL/CL) is significantly elevated in brain tissue following hypoxic-ischemic injury, and this ratio correlates strongly with cerebral infarct volume, suggesting its promise as a novel biomarker reflecting the extent of mitochondrial damage.9 Concurrently, the clinical value of protein-based biomarkers has been extensively validated in the era of therapeutic hypothermia. Multiple studies have consistently shown that levels of glial fibrillary acidic protein (GFAP), ubiquitin C-terminal hydrolase L1 (UCH-L1), S100B and Tau protein in the serum or cerebrospinal fluid of affected neonates are closely associated with the severity of HIE and adverse prognosis.81 Of note, even in the context of therapeutic hypothermia, cerebrospinal fluid levels of neuron-specific enolase (NSE) remain effective in predicting the severity of brain injury and neurodevelopmental outcomes at two years of age.82 Furthermore, soluble lectin-like oxidized low-density lipoprotein receptor-1 (sLOX-1) has also emerged as a biomarker with considerable potential; levels measured within six hours after birth not only reflect HIE severity but also predict poor outcomes following therapeutic hypothermia.83 Therefore, integrating lipidomic and proteomic biomarkers may enable more precise, multi-dimensional assessment of hypoxic-ischemic encephalopathy and guide individualized neuroprotective strategies.

Beyond the aforementioned markers, non-coding RNAs—key regulators of gene expression—also demonstrate substantial promise for the early diagnosis of HIE. A multicentre validation study confirmed that changes in the expression of specific microRNAs (eg, miR-374a-5p, miR-376c-3p) in umbilical cord blood enable early identification of HIE and inform decision-making regarding therapeutic hypothermia.84 Moreover, by combining metabolomics with miRNA analysis, further research has revealed that specific metabolites (eg, xanthine, betaine) and miRNA profiles are significantly correlated with scores on the Bayley Scales of Infant Development (third edition, BSID-III) at 24 months of age, indicating that multi-omics integrative analysis may serve as an effective auxiliary tool for predicting medium- to long-term neurodevelopmental outcomes.85 In summary, from mitochondrial lipid metabolites reflecting core pathological mechanisms, through clinically well-validated protein markers, to the promising frontier of non-coding RNAs, a multi-dimensional, multi-omics biomarker framework is progressively being established. This framework provides a robust scientific foundation for overcoming the efficacy limitations of therapeutic hypothermia and achieving early precise diagnosis as well as individualized prognostic assessment in neonatal HIE.

Clinical Risk Scores

To effectively integrate multi-dimensional clinical information into decision-support systems, researchers have devoted substantial efforts to developing various risk-scoring tools. For instance, one study successfully constructed a risk-scoring model to predict cognitive, language and motor outcomes at 24 months of age by integrating multi-dimensional indicators including the 5-minute Apgar score, electroencephalography (EEG)-confirmed seizures, abnormal magnetic resonance imaging (MRI) findings and clinical physical examination.86 In parallel, another multicentre study developed models for predicting death or neurodevelopmental impairment using routine clinical data collected early after admission and at discharge. These models achieved an area under the curve (AUC) exceeding 0.85, thereby providing a powerful tool for early identification of high-risk infants and the implementation of targeted interventions.78 The development and application of such risk stratification tools mark a critical step forward in clinical management: a transition from experience-based heuristic judgment to data-driven precise prediction and individualized treatment.

Conclusions and Future Perspectives

In summary, the treatment of neonatal HIE is at a critical historical juncture, undergoing a profound evolution from a standardised “monotherapy” paradigm toward the integration of precision‑based combination adjunctive strategies. Although TH remains the current standard of care, its inherent limitations—including suboptimal efficacy in low-income countries and limited benefits in severe or preterm cases—are powerfully driving the search for more robust, universally applicable, and precisely targeted interventions. However, it must be emphasized that most emerging approaches discussed in this review remain at preclinical or early‑phase clinical stages, and none has yet demonstrated superiority over TH in adequately powered randomized trials. The future therapeutic landscape will completely move beyond the era of “one-size-fits-all” treatment and instead embrace a new era of diversified therapies grounded in individual pathophysiological mechanisms.

First, future therapeutic strategies must initiate precise combination interventions within the “golden 6-hour” window of TH. The key to achieving this goal lies in the rapid identification of dominant injury pathways in affected infants using electroencephalography, multimodal neuroimaging, and biomarkers. For example, in cases involving cardiolipin remodeling disorders and mitochondrial dysfunction, combination with cardiolipin remodeling agents could be considered. Against ferroptosis as a key mechanism, ferroptosis inhibitors could be deployed. Moreover, given that neuroinflammation persists through the secondary and tertiary phases of injury, peptide-based therapies targeting microglia (eg, MG1-KLA) or immunomodulators show substantial promise. Specifically, sildenafil alleviates neuroinflammation and promotes oligodendrocyte recovery via the PI3K/AKT/mTOR pathway; melatonin preserves aquaporin-4 (AQP4) polarity and glymphatic function by inhibiting CDK5 activity; and the novel GLP-1/GIP dual receptor agonist DA5-CH exerts anti-inflammatory effects by suppressing the TLR2/NF-κB/NLRP3 pathway. Additionally, neurotrophic factors such as brain-derived neurotrophic factor (BDNF) promote neuronal survival by upregulating Stx1b and inhibiting VDAC1. These diverse pathological mechanisms offer a wealth of precisely targetable nodes for future combination therapies.

Second, for infants who already suffer from significant brain injury, neurorepair strategies should be promptly implemented during the subacute phase. Stem cell therapy, particularly mesenchymal stem cell (MSC) transplantation, has demonstrated strong neurorepair potential by modulating autophagy (eg, downregulating SYNPO2) and attenuating inflammatory responses. Clinical studies have shown that combining MSCs (eg, Temcell) with TH is safe and feasible, although definitive efficacy requires larger-scale validation. Furthermore, intranasal administration—a noninvasive, convenient strategy that directly targets the central nervous system—holds great promise for both cell-based and pharmacological delivery, especially in resource-limited settings. Exosomes, as key paracrine mediators of MSCs, also represent a promising cell-free therapeutic strategy. Collectively, these findings point to a broad horizon for promoting long-term recovery through the reconstruction of neurovascular unit function.

Finally, realizing this promising vision will require intensive global collaboration and profound reform of healthcare systems. Standardized, shared clinical registry databases must be established to harmonize data collection and facilitate cross-border, multi-center research collaborations. The current marked heterogeneity in data collection across global centers represents a major obstacle to large-scale collaboration; therefore, the promotion of common data elements is critically important. In LMICs in particular, there is an urgent need to validate low-cost, easily scalable interventions such as ketogenic diets, intranasal drug delivery, or telemedicine-guided TH. Moreover, educational outreach to primary hospitals, the application of telemedicine, and the development of EEG-based automated decision-support systems will help reduce regional disparities in healthcare access, ensuring that more infants can equitably benefit from medical advances. The successful integration of precision medicine approaches will ultimately depend on our ability to integrate multi-dimensional mechanisms into a cohesive and actionable framework. In conclusion, by integrating continuously emerging new targets from basic research, innovative combination strategies from clinical translation, and systemic dissemination and collaboration at the public health level, we can significantly reduce the global burden of HIE and secure a healthier future for these vulnerable new lives (Figure 4).

Figure 4.

Future Roadmap: From Empirical Cooling to an Integrated Precision Adjunctive Framework. This figure presents a temporal roadmap for the evolution of neonatal HIE treatment. The left panel represents the current era (eg, empirical TH with ceiling effect, rigid 6-hour window, selection risks, and ineffective/increased mortality in LMICs), and the right panel depicts the future paradigm: rapid biomarker-based stratification, combination therapy targeting specific pathways (eg, ferroptosis, necroptosis, neuroinflammation), and neurorepair strategies (stem cells, exosomes). Global collaboration and LMIC-adapted interventions (eg, ketogenic diet, intranasal delivery) are highlighted as enabling factors. All figures were created in BioRender. (2026) https://BioRender.com/vcdpdnn.

Funding Statement

This study was supported by the National Natural Science Foundation of China (Nos. 81901933) and Yancheng Medical Science and Technology Development Plan Project (No. YK2023034).

Disclosure

The authors report no conflicts of interest in this work.