A comprehensive literature search of the WoSCC database was conducted on October 15, 2024. The primary search terms included “hypothermia” OR “therapeutic hypothermia” OR “cooling therapy” OR “neonatal hypoxic-ischemic encephalopathy” OR “neonatal HIE” OR “newborn hypoxic-ischemic encephalopathy” OR “perinatal asphyxia”. The search was restricted to English-language articles. The article type is limited to “Article”, excluding other types such as “Meeting”, “Case report” and “Review article”, in order to ensure that only published articles that have undergone rigorous and systematic scientific review are included, and to avoid the interference of secondary literature. To avoid the effect of database updates, all data results were downloaded on the same day. It is worth noting that as of the execution time of the search (October 15, 2024), although it was not possible to fully incorporate all the published literature in 2024, we still hoped to conduct a reliable and relatively complete quantitative characterization of the status of this research field in 2024. Therefore, the published literature of that year was retained for subsequent analysis. File formats include plaintext, BibTex, and tab-limited files. Finally, 1,165 records were identified for analysis. The detailed search and screening process is shown in Figure 1.

In this study, we imported the retrieved data into the bibliometrix (version 4.3.0) package in CiteSpace (version 6.3. R1), VOSviewer (version 1.6.20), R software (version 4.4.0), and KH Coder (version 3b07d) for analysis and visualization. VOSviewer was used to visualize knowledge graphs, including co-occurrence analysis and co-citation analysis of large-scale literature data, representing authors, journals, and other relevant information. It has been widely used in bibliometric analysis and research (5, 6). We also used CiteSpace (version 6.3. R1) to perform reference cluster analysis to identify developments and future trends in the research field (7).

Additionally, an international cooperation network among countries was mapped using an online scientometrics platform. The bibliometrix package in R software was used to generate histograms of citations, illustrating citation relationships between documents and revealing how knowledge is transferred within the field of study. By analyzing the chain of citations, we can trace the spread of important theories or discoveries.

We employed the latent Dirichlet allocation (LDA), a widely-used topic modeling method capable of efficiently processing unstructured textual data (8), to analyze large-scale document collections. Through LDA, we first generated a vocabulary of term functions by statistically analyzing word co-occurrence patterns across documents. We then quantified the association between documents and latent topics based on term frequency distributions within each text.

Raw data downloaded from WoSCC were imported into Microsoft Excel 2021 for the initial collation. Only article titles and abstracts were retained to form the original corpus. To ensure result reliability and validity, word frequency thresholds were established, generic stop words were removed, and specific stop words were defined. LDA-based topic modeling and word cloud mapping were performed using KH Coder software and Word Cloud. At the final stage, research topics were manually labelled based on the 10 most relevant articles and 20 key topic terms.

We used ChatGPT (Open AI, version 4.0) for part of the language polishing. The date of the visit is February 6, 2025. We have manually checked and guaranteed the accuracy of the AI polished content.

This study is a bibliometric analysis and therefore ethical approval is not applicable.

Our analysis of the annual number of publications and citations in the field from 1985 to 2024 (as of the time of the search execution: October 15, 2024) provides insight into the research trends and the importance of this area. As shown in Figure 2. Between 1985 and 2024, the number of publications and citations in the field of TH in neonatal HIE showed different trends. There is an increasing trend in the number of published papers, indicating that research in this field is gradually gaining attention. The number of citations also increases with the number of published papers, reflecting the high academic value and clinical significance of research results in this area.

As illustrated in Figure 3, the subject distribution analysis, based on the superimposed CiteSpace double charts, demonstrates the interdisciplinary nature of research on TH in neonatal HIE. The participation of different disciplines has provided a wealth of perspectives and methods for research in this area, contributing to a more comprehensive understanding and solution for the complex medical problems of neonatal HIE.

Figure 4A highlights significant differences in the number of publications in this field between different countries. The United States had the highest number with 361. Figure 4B shows the differences in the research activities of different institutions in this field. The University of Bristol led this process with 45 publications. Figure 4C shows that highly cited authors, such as Maier, Steven F. Watkins, and Linda R., each with 808 citations, have high visibility and influence in this field. Figure 4D illustrates that of the many journals, the New England Journal of Medicine leads with 4,472 citations.

We also summarized the total number of publications, number of citations, and average number of citations for countries, authors, institutions, and journals, and the results are shown in Table 1.

In the co-occurrence network shown in Figure 5A, the size and color of the nodes indicate the strength of the relationship between the frequency of keyword occurrences and the research topic. Notable themes included “Mild Systemic Hypothermia”, “Birth Asphyxia”, “amplitude-integrated” “electroencephalography”, “Near-infrared Spectroscopy”, “Childhood Outcome”, and “Prognostic Value”.

Time-dimensional network analysis of keyword co-occurrence using VOSviewer revealed the evolution of these research topics over different periods. Earlier studies, shown in blue and green, focused on topics such as “Therapeutic Hypothermia” and “Hypoxic-Ischemic Encephalopathy.” Over time, the research focus gradually shifted to more specific topics, such as “Mild Systemic Hypothermia,” “Oxidative Stress” and “Biomarkers,” indicating that they have become research hotspots in recent years. Further details are shown in Figure 5B.

The red color of the cluster labels in Figure 5C indicates that the study time was closer to the present time. Using CiteSpace cluster analysis, we can clearly observe research hotspots and trends in the TH of HIE.

Figure 5D shows the top 25 keywords with the strongest citation bursts between 1985 and 2024, with “mild hypothermia” beginning to show citation bursts in 1998 and continuing to 2012 with a higher intensity. This suggests that mild TH has received widespread attention in this field, likely because it can reduce side effects, while still playing a role in the treatment of neonatal HIE.

In addition, after analyzing the occurrence frequency of keywords and their Total Link Strength value, it was found that research on the TH of HIE focused on perinatal asphyxia, hypoxic ischemic encephalopathy, different hypothermia treatment methods, therapeutic effects, prognosis evaluation, the particularity of the study object, and the application of detection technology. Further details are presented in Table 2.

VOSviewer's citation analysis in Figure 6 revealed key literature and author networks in the field, where, Toet et al. (48) published in Arch Dis Child Fetal Neonatal Ed, Rutherford, Rutherford et al. (49) published in Lancet Neurol (9), Shankaran et al. (50) published in Arch Dis Child Fetal Neonatal Ed (10), etc. have been frequently cited, showing the core position of these studies in the field of TH of neonatal HIE. Early studies, such as that by Sarnat Hb (51), published in Archives of Neurology (11), have constantly innovated research methods. Over time, our understanding of TH in neonatal HIE has continued to expand.

In this study, the LDA algorithm was used to model the topics of 1,165 articles, identifying 13 key topics based on widely recognized criteria. These topics were further categorized into four main research directions, with corresponding English terms summarized for each direction. These topics included “TH,” “nanoparticle,” “miRNAs” and “signaling pathways,” etc. The results are shown in Figure 7.

HIE is a brain damage caused by hypoxia and ischemia that occurs during the perinatal stage and affects the survival and development of newborns. Lack of oxygen impairs the energy metabolism of brain cells, which in turn triggers a chain reaction involving cell damage and death. It has been shown that in the presence of hypoxia, the supply of oxygen to brain tissue is insufficient, resulting in reduced adenosine triphosphate synthesis within cells, damage to cell membrane integrity, and ultimately cell death (12). Hypoxia can trigger a range of inflammatory responses, further exacerbating brain damage (13).

Currently, the research on biomarkers for HIE has shifted from single traditional markers (such as protein markers, cell damage/apoptosis markers, inflammation markers, oxidative stress markers, etc.) to multimodal combinations, and has explored new sources of biomarkers, such as microRNAs (miRNAs).

We have already known that miRNAs are non-coding RNAs composed of 22 nucleotides that regulate gene expression by cutting specific messenger RNAs (mRNAs) or inhibiting their translation. Since miRNAs can regulate more than 60% of protein-coding genes in the human genome, they play a crucial role in key cellular processes, including differentiation, proliferation, growth, and apoptosis (25). Additionally, miRNAs exhibit tissue- and time-specific expression in the nervous system, making them potential biomarkers and therapeutic targets (26).

Overexpression of miR-124 significantly improves the survival of neurons under hypoxic conditions and reduces the number of apoptotic cells (27). Additionally, miR-499-5p plays a protective role against hypoxic-ischemic injury by regulating C-reactive protein expression and inhibiting apoptosis in nerve cells (28). Another study confirmed that miR-363-3p alleviates hypoxic-ischemic injury by targeting DUSP5, demonstrating its potential as a therapeutic target (29). These findings suggest that miRNAs may become a key target in future therapeutic strategies for HIE, contributing to the advancement of personalized medicine.

TH is widely used in the management of HIE, primarily to reduce the metabolic rate of brain cells and mitigate hypoxia-induced cell damage. Studies have shown that Low temperatures can significantly inhibit apoptosis and inflammatory responses, thereby protecting the integrity and function of brain cells. Low temperatures can reduce the concentration of intracellular calcium ions, production of intracellular ROS, and damage caused by oxidative stress in cells (30). In addition, hypothermia reduces inflammation by regulating key signaling pathways such as NF-κB and MAPK signaling pathways, thereby promoting cell survival and functional recovery (31). In animal models, TH has been shown to be effective in alleviating brain tissue damage following hypoxic ischemia, reducing cell mortality, and improving the recovery of neurological function (32).

Mild systemic hypothermia is currently the most widely used and evidence-based treatment method for TH in clinical practice. It refers to a standardized neuroprotective treatment for neonates with moderate to severe HIE. It is initiated within 6 h after birth, using active cooling techniques to precisely control the core body temperature within the range of 33°C–34°C, and maintaining this temperature for 72 h. Subsequently, the body temperature is slowly restored (33).

In clinical applications, TH has been shown to significantly improves the prognosis of HIE. Clinical trials have reported that after 72 h of TH, infants exhibit improved neurological function scores, increased survival rates, and better long-term neurodevelopmental outcomes (34). Additionally, TH has been found to reduce the incidence of epilepsy and the risk of long-term neurological damage in newborns (35). Despite these benefits, the efficacy of TH may be influenced by several factors, including the timing of treatment initiation, the duration of hypothermia, and the overall health status of the newborn (32).

Although TH has shown some efficacy in treating HIE, approximately 29% of newborns treated with this treatment still develop severe neurological developmental disorders (36, 37). Clinical studies have shown that while TH can significantly reduce mortality and the risk of major disability, it cannot completely eliminate the long-term complications associated with HIE, such as epilepsy, motor impairment, and cognitive impairment (14, 38). Several studies have also found that newborns may still experience a persistent neuroinflammatory response even after TH treatment, which may be an important factor in poor outcomes (36, 39). TH may also cause the newborn's body temperature to become too low, thereby triggering complications such as arrhythmia and coagulation dysfunction (31). In low-income countries, TH implementation is even less satisfactory due to the scarcity of medical resources. This further highlights the importance of seeking alternative or complementary treatments (40). Therefore, relying solely on TH is not sufficient to meet the needs of all patients. It is necessary to explore other drugs or interventions to enhance its effectiveness (39).

Emerging treatments for HIE are still in the exploratory stage.

Creatine is an important component of cellular energy metabolism. By participating in the creatine-phosphate system, it helps maintain the level of adenosine triphosphate, thereby supporting neuronal function and reducing damage related to energy depletion (41, 42). Creatine supplementation has been proven to protect brain tissue by reducing oxidative stress, inhibiting inflammatory pathways, and decreasing cell apoptosis (42). However, current data from related studies are still insufficient.

Nanoparticles typically range in size from 1 to 100 nm, which are tiny dimensions that allow them to penetrate cell membranes and enable efficient drug delivery within cells (43).

One study showed that the use of magnetic nanoparticles enabled local heating by an external magnetic field at low temperatures, thus enhancing the efficacy of drugs (44). Therefore, the use of nanoparticles can enhance the drug delivery efficiency at low temperatures and help the drug cluster better in the damaged area. In future, the combination of nanoparticles and TH may produce synergistic effects, providing new ideas and possibilities for the treatment of HIE.

Most emerging therapies are still in the research phase and far from becoming routine clinical treatments.

At present, the research on HIE mainly focuses on a series of new targets and emerging treatments, which will promote the early diagnosis of the disease, the precision of treatment, and the assessment of prognosis.

In terms of the practical implications of advancing HIE research, it is crucial to understand how the most active authors or institutions influence future work, collaborations or funding directions. On the one hand, by analyzing bibliometric data, one can identify researchers and institutions with significant influence in this field, thus providing a basis for future collaborations. In addition, these active authors are often able to attract more grant support, thereby facilitating larger scale and more innovative research projects.

HIE is not merely a medical issue; This is a complex challenge, involving multiple external factors. Among the external factors, underfunding often leads to delays in infrastructure construction (45, 46), while a lack of necessary human resources and training often results in ineffective infrastructure utilization. This, in turn, affects the quality of care (9). In addition, the inadequacy of regulatory policies can also lead to inconsistencies in clinical practice, making it difficult to extend best practices to all medical institutions. Therefore, in order to optimize treatment outcomes for patients with HIE, it is necessary to start from a systemic perspective and consider and continuously improve these key external factors comprehensively to ensure that every newborn receives timely and effective intervention (16, 47).

The study combines multidimensional bibliometric methods (coauthor/co-occurrence/co-citation analysis) with topic modeling, supported by a robust dataset (1,165 WoSCC records), to systematically map HIE research trends. It bridges theoretical insights (e.g., mild systemic hypothermia, oxidative stress, biomarkers and signaling pathways) to clinical priorities, emphasizing actionable strategies for optimizing TH in neonatal care.

However, this study has some limitations that warrant consideration. First, the research relied primarily on bibliometric analyses and lacked integration with wet-lab experiments, which could provide a more comprehensive understanding of the biological mechanisms underlying TH in neonatal HIE. Second, the sample size of the reviewed literature, although substantial, may not have encompassed all relevant studies, potentially leading to a skewed understanding of the field. Third, this study included only original articles published in English, which may have led to the omission of key regional studies published in non-English languages. There is a possibility that the data set over-represents the perspectives and research activities in high-income countries and does not fully reflect the true research trends and challenges in the global context, especially in lower-income countries with high HIE burdens. Fourth, relying on citation metrics (such as the number of citations) to assess influence has inherent biases. These metrics mainly reflect visibility and attention within the mainstream (especially English) academic communication system, which itself is biased towards high-income countries and English-language publications. Fifth, reliance on title and abstract analysis may miss key details in full-text articles. Finally, selection bias between databases (e.g., WoS vs. PubMed) affecting data representativeness.