As the most common form of dementia, Alzheimer’s disease (AD) comprises 60–80% of all dementia cases. Worldwide, the numbers of patients with AD are expected to reach 65.7 million in 2030 and 115.4 million in 2050, with a rate of one new case every 3 s (Prince et al., 2013). As a fatal illness, AD’s average survival time ranges from 4 to 8 years for patients aged 65 years and older (Larson et al., 2004). Since AD is still not curable, early diagnosis and intervention are critical to slow the development of AD and improve the quality of life of patients. In addition to the classical biomarkers recommended by guidelines (Zou et al., 2023), many other biomarkers in cerebrospinal fluid (CSF) and/or serum/plasma, such as neurofilament light (NfL), glial fibrillary acidic protein (GFAP), and alpha-synuclein, were recently recommended for the early and differential diagnosis of AD (Teunissen et al., 2022; Johnson et al., 2023).

As an intermediate filament structural protein involved in cytoskeleton assembly and integrity, GFAP is highly expressed in activated glial cells (Ganne et al., 2022). Thus, GFAP serves as a biomarker of neurodegenerative diseases, including AD, dementia with Lewy bodies, and frontotemporal dementia. In recent years, an increasing number of studies have aimed to explore the effect of GFAP on the development, diagnosis, and prognosis of AD. GFAP in body fluids, including CSF and serum/plasma, is considered a good biomarker for the early diagnosis and prognosis of AD, especially blood GFAP, which is considered an excellent biomarker for large-scale screening (Heimfarth et al., 2022; Teunissen et al., 2022; Guo et al., 2023). Moreover, it was reported that GFAP plays important roles in AD development, such as neuropathic aggregate accrual, and could become a novel promising therapeutic target to alleviate, delay, and even prevent AD (Ganne et al., 2022). However, GFAP is not specific for AD but is related to many neurological diseases (Heimfarth et al., 2022). Thus, it is still urgent to further explore the relationship between GFAP and AD and clarify the effect of GFAP on the occurrence, development, diagnosis, and prognosis of AD.

Thus, we conducted the first bibliometric analysis to identify research trends and hotspots on GFAP within the area of AD to provide a comprehensive understanding of this topic and further guide future research directions.

To analyze the research trends and hotspots on GFAP within the area of AD, we conducted a bibliometric analysis and eventually included 2,269 articles and review publications. To our knowledge, this is the first bibliometric analysis to explore the research on GFAP and AD. Despite some minor fluctuations, the general trend for the number of publications was upward, especially from 2014, when the upward trend became successive and remarkable. Around this time, the diagnostic performance of GFAP in CSF and blood was explored based on biochip array technology (Rosén et al., 2011). Then, the most cited publication (Olsson et al., 2016) further evaluated the diagnostic performance of many CSF and blood biomarkers for AD by conducting a meta-analysis, including GFAP in CSF, which implied the importance of GFAP for the diagnosis of AD. However, related articles about GFAP in CSF for the diagnosis of AD were still limited at that time, and only two articles were included in the meta-analysis; thus, a null conclusion was reported. Moreover, the application of GFAP in serum/plasma in AD had not been comprehensively recognized at that time, thus this review did not include it. During the past few years, the diagnostic value of GFAP in body fluids, especially in blood, was further explored and it could be used for large-scale screening (Guo et al., 2023). Although astrocyte activation is not specific to AD pathophysiology, plasma GFAP was found to be more sensitive for the early diagnosis of AD than Aβ-PET, CSF and plasma p-tau181, and CSF total tau (Guo et al., 2023), and it linearly increased with increasing Aβ-PET burden (Asken et al., 2020). Blood GFAP levels can predict not only AD development in patients with mild cognitive impairment but also cognitive decline in cognitively unimpaired subjects (Verberk et al., 2020; Cicognola et al., 2021; Verberk et al., 2021). At the same time, more sensitive technology, such as ultrasensitive single-molecule array technology, was also developed (Rissin et al., 2010) and applied for the accurate measurement of blood GFAP (Bogoslovsky et al., 2017), which prompted its application. Under these circumstances, the diagnostic and prognostic performance of GFAP in CSF and serum/plasma for AD should be further evaluated and verified with better analytical assays in larger-scale cohorts, especially for different age groups and clinical contexts (Hansson et al., 2022). Moreover, new systematic reviews and meta-analyses based on more publications, which are currently lacking, need to be conducted.

Interestingly, it is reported that the association of plasma GFAP with Aβ pathology appears stronger than that of CSF GFAP, which is different from other glial biomarkers (Pereira et al., 2021). A recent study found that plasma GFAP is more stable than CSF GFAP, without the effect of freeze–thaw cycles, making it a better biomarker for clinical application (Simrén et al., 2022). However, the reason why blood GFAP showed a stronger association with Aβ pathology is still unclear; thus, further work is still needed to investigate the mechanisms of GFAP release into the bloodstream under pathological conditions and clarify the key pathological substrates of the better separation of serum/plasma versus CSF in AD (Hansson et al., 2022). Moreover, even if the Alzheimer’s Association has recommended plasma GFAP as a promising and available biomarker since 2022 (Hansson et al., 2022; Teunissen et al., 2022), many problems still need to be solved, such as preanalytical confounders and biological factors. It was found that sample types with different anticoagulants could affect the concentration of blood GFAP, but centrifugation delay within 24 h, storage delay within 2 w at −20°C, centrifugation temperature and freeze–thaw cycle did not significantly affect plasma GFAP (Verberk et al., 2022). Additionally, significant differences were found for many AD-related biomarkers, including GFAP, between the postprandial group and fasting group, implying that the concentration of these biomarkers could be affected by food intake (Huber et al., 2023). Moreover, the concentration of blood GFAP is usually simultaneously measured with other biomarkers, such as amyloid and tau; thus, a standardized preanalytical process including sample drawing, treatment, and storage is still necessary (Verberk et al., 2022). Furthermore, it is important to evaluate the effects of clinical confounders and biological factors on the clinical diagnostic and prognostic performance of GFAP for AD (Hansson et al., 2022). Previous studies preliminarily found that impaired kidney function could affect the concentration of amyloid 1–42, p-tau181, and NfL but not plasma GFAP (Sedaghat et al., 2023; Zhang et al., 2023); however, more comprehensive studies with larger cohorts are still needed to explore the effect of more factors, such as peripheral neuropathies and body mass index, on blood GFAP for more accurate laboratory measurements and clinical interpretation (Hansson et al., 2022).

By this bibliometric analysis, it was also found that the USA, China, England, Japan, Spain, Germany, and Sweden contributed the most research on GFAP within the area of AD. The USA always took the lead in this area, but publications from China and England increased annually. As the top populous country with a rapidly increasing aging population, China has an increasing number of patients diagnosed with AD (Jia et al., 2020), and GFAP in CSF and serum/plasma has become a research hotspot, showing promising early diagnosis and prognosis value for AD. Of the top 10 prolific and highly cited institutions, three institutions are from the USA, and another three institutions are from Sweden, of which the University of Gothenburg contributed most publications in this area. Moreover, Henrik Zetterberg and Kaj Blennow, both professors at the University of Gothenburg, contributed the most prolific publications. As key leaders of the UK Dementia Research Institute and the Alzheimer’s Association, respectively, Henrik Zetterberg and Kaj Blennow focused on the fluid biomarkers of neurodegenerative diseases and made remarkable contributions in this area. Moreover, the Journal of Alzheimer’s Disease published most articles and reviews in this area, and the publications of Neurobiology of Aging had the highest average citations.

The discipline distribution analysis of publications revealed that “molecular, biology, genetics” and “molecular, biology, immunology” were the key research areas of 21 domain disciplines, and the published articles on “molecular, biology, immunology” were based on those on “molecular, biology, genetics,” which implied that “molecular, biology, genetics” had provided a foundation for this research area. Furthermore, the most published and hottest meaningful associated keywords included “oxidative stress,” “neuroinflammation” or “inflammation,” “microglia,” “neurodegeneration,” “hippocampus,” “mouse model,” “amyloid,” “immunohistochemistry,” of which “inflammation,” “cognitive impairment,” “neuroinflammation,” “tau,” and “dysfunction” were the hot associated keywords in the past 5 years. Because it is difficult to directly intervene in the human body and obtain human brain tissues, animal models especially mouse models are helpful and necessary for associated immunology and genetics studies (Qin et al., 2020), and immunohistochemistry remains the gold-standard technique for capturing the spatial expression patterns of astrocytes in postmortem tissue sections (Viejo et al., 2022). Since astrocytes and microglia could release proinflammatory cytokines and reactive oxygen, triggering neuroinflammation and oxidative stress (Novoa et al., 2022), it is not surprising that “inflammation,” “neuroinflammation,” and “oxidative stress” were hot keywords related to GFAP, which is an important biomarker for astrocyte activation. Moreover, the inflammatory response, oxidative stress, and glial activation (especially microglia and astrocytes) have been found as three main factors for AD progression based on both mouse models and human post-mortem brains (Companys-Alemany et al., 2022; Viejo et al., 2022). Furthermore, Aβ and tau are found able to activate astrocytes and microglia, and GFAP expression and protein concentrations were higher in areas surrounding dense Aβ plaques and increased with tau accumulation, implying the association of GFAP with amyloid and tau pathology (Simpson et al., 2010; Novoa et al., 2022). However, the results based on clinical individuals only showed a significant association between higher CSF and plasma GFAP levels and elevated Aβ-PET load but not for tau-PET load (Pereira et al., 2021; Ferrari-Souza et al., 2022). Moreover, it was found that glial biomarkers including GFAP increased with aging regardless of Aβ status and also in other neurological diseases besides AD (Milà-Alomà et al., 2020). Thus, the mechanisms of GFAP in the development of AD are complex due to the multiple functions of astrocytes, and more mechanism studies are still needed to clarify the specific causal relationship between GFAP and AD. Noticeably, all findings based on animal models need to be further verified based on human cohorts to ensure the clinical application.

In summary, the effects of GFAP on the occurrence, development, and prognosis of AD still need to be further researched, and our analysis could provide some guidance on further research trends. It should be emphasized that the GFAP change is not specific for AD but for many neurodegenerative diseases. Thus, GFAP in body fluid could only be used for the first screening but not a final diagnosis.

YZ: methodology, data curation, and writing – original draft preparation. LL, LG, and CCM: methodology and data curation. SY, XM, CHM, and JG: writing – reviewing and editing. LQ: conceptualization and methodology. All authors contributed to the article and approved the submitted version.