Bibliometric Analysis of Highly Pathogenic Avian Influenza Research Globally from 2003 to 2023

Abstract

Introduction: Despite decades of research on highly pathogenic avian influenza, especially H5N1, an understanding of the impact of research outputs on policy, the impact of funding and policies, collaboration between authors, and other bibliometric measurable indices is elusive. Methods: H5N1 research data were retrieved from the PubMed database using “H5N1”, “highly pathogenic”, and “influenza” as keywords, and combined with the Boolean operator, “OR”. The evaluated growth pattern of H5N1 research was analysed using Microsoft Excel, while the VOS viewer window version 1.6.15 was utilized in analyzing the keywords, authors’ inputs, and collaborations amongst research institutes/universities involved in H5N1 research. Results: A total of 8411 articles published within 2003 and 2023 were retrieved. The trend of the published articles indicates that 2003 to 2004 had the lowest number of articles, while the highest were observed for 2009 and 2010. Since then, there has been a gradual decline. The most dominant article type was original research article (89.80%), followed by reviews with 7.54%. The most productive countries were Japan, the United States of America, China, and Egypt in Africa. The top 15 authors accounted for 12.40% of the total published data. Keywords analysis revealed the emergence of newer keywords such as H5N8, wild birds, mammals, and Nigeria; however, as expected, influenza virus avian H5N1 subtype was the dominant keyword. Research institutes with backgrounds in agriculture and veterinary medicine dominated the top 15 organizations involved in H5N1 research. Conclusions: Our findings suggest that research aimed at developing therapeutics and vaccines, as well as continuous surveillance in high-risk areas is urgently needed. Furthermore, the findings provide baseline data for all the stakeholders involved in H5N1 research, which could inform future research, funding, and policies.

Introduction

The highly pathogenic avian influenza virus (HPAIV) H5N1 strain was initially detected in China in domestic waterfowl in 1996 [1,2]. Since then, the HPAIV H5N1 strain has acquired new hosts that include chicken [3,4], ducks [5], geese [1], turkeys [6], wild birds [7,8] and mammals [4,8,9]. So far, more than 48 mammalian species have been infected with the HPAIV H5N1 [10]. The spread of this virus and other avian viruses is in part due to its ability to evolve into multiple clades through heritable changes to its genome [11,12,13]. As the virus continues to evolve, it has acquired the potential to cross species barrier and infect humans and other mammals [2,9,10]. Infections in humans have been associated with varying severity taking a clue from the reported human cases in Vietnam and the United States of America (USA) [2,10].

Although human to human infections with HPAIV H5N1 subtype is rare, there has been a total of 882 cases that has been reported so far with a case fatality rate of 52% [12,13]. Despite the low number of human cases so far, the virus has great potential to cause a major pandemic [14] as evidenced by the 1918 Spanish flu that claimed an estimated 50 million human lives [15]. Given the public health significance of HPAIV H5N1, numerous research has been conducted covering key areas such as surveillance [14], outbreaks [1,3,5,6,7,9],vaccines development and clinical trials among others [16,17,18] and prevalence in wild birds [8]. Despite the seemingly increasing number of research on the subject, outbreaks of HPAIV H5N1 is repeatedly reported in various countries across all the continents of the world [12]. A recent situation report published by the Word Health Organization indicates that the virus has been detected in 67 countries so far [12].

With the possibility of a pandemic arising from the HPAIV, especially the H5N1 strain, there is a need to have a comprehensive understanding of the trajectory of the various research conducted globally in order to identify potential limitations or gaps that have to be addressed. To the best of our knowledge, this is the first bibliometric analysis that attempts to fill this research gap in relation to avian influenza. As a type of study, bibliometric analysis allows researchers to visualize research conducted in a particular field of study. It allows for the quantitative analysis of research outputs, and how avian influenza research were conducted in terms of the collaborations among the authors and institutions. Hence, we aimed to understand the trajectory and depth of research conducted on H5N1 and identify potential gaps or areas for future studies or inform policy development.

Methods

Source of data

The data utilized in this study were retrieved from the PubMed database. PubMed as a database is hosted by the National Library of Medicine (NLM) - National Center for Biotechnology Information (NCBI). The database holds over 37 million biomedical literature that is freely available to the scientific community.

Retrieval of studies from PubMed

All HPAIV H5N1 research conducted globally and indexed in PubMed between 2003 and 2023 was retrieved. The search was performed on 16 June 2024, and as previously reported [19]. In performing the search, the advanced search option (MeSH terms) was utilized. The descriptors "highly pathogenic avian influenza” [Title/Abstract], and "H5N1” [Title/Abstract] were applied to the "title” and "abstract” fields in the advanced search option. The descriptors were linked with the Boolean operator "OR”. The search duration was limited to 2003 and 2023, representing a 21-year period. The choice of the year 2003 was made to ensure that adequate and even coverage of the HPAIV H5N1 research in all the continents. Furthermore, the search was refined to include "retraction” and "expression of concern”.

Data screening criteria and protocol

Articles included in the bibliometric analysis were those that met the pre-set inclusion criteria to ensure that the included articles are relevant to the study. The inclusion criteria were year (2003 to 2023) and research focusing on H5N1 or HPAIV in animals and humans. In addition, articles were not restricted by type. On the other hand, articles that did not focus on H5N1 or H HPAIV or published outside the 2003-2023 period were excluded. The starting year was deliberately placed at 2003 to ensure an even spread of studies as the first outbreak on the African continent was reported in 2006 [20]. The articles were further screened first using titles and where the relevance or type of article could not be inferred from the titles, abstracts were examined. As a last resort, the full texts were also consulted where necessary [19,21]. We only considered articles that focused on HPAIV H5N1. We also disqualified articles that did not specifically focus on the HPAIV H5N1 but mentioned it in passing. Additionally, duplicate articles were eliminated. Two review groups comprising of two authors each selected the articles independently, and where there were differences of opinion, these were resolved by both groups coming to a consensus. During the screening, where there were differences in opinions on any type of article, a consensus was reached by all four researchers after thorough checking.

Social network analysis

The VOS viewer window version 1.6.15 was utilized to visualize the social network for the downloaded articles. The software was downloaded at https://www.vosviewer.com/ [21]. It was utilized to map the keywords, search words and collaborators from the retrieved data and presented as previously reported [19].

Bibliometric analyses of the data

Various bibliometric parameters namely authors and institution participation index, keywords analysis and co-authorship analysis were evaluated.

Authors and institution participation index analysis (AIPIA). We also evaluated the authors and institutions participation index from the retrieved articles. The evaluation of AIPIA reveals the number of articles published by the various authors as well as their affiliated institutions compared to the overall publication in the field of practice [19]. It provides at a glance their contributions to their field of study.

Keyword analysis. As previously reported [19],keyword analysis was performed to confirm the trend of the research and discussion on HPAIV (H5N1) conducted in 2003 to 2023. Analysis of keywords has the potential to also show trend in the research by revealing both old and new keywords utilized by the various authors in their published data.

Bibliometric mapping analysis. This was done using the co-authorship and the co-occurrence analyses to reveal the co-authorship interactions or collaborations among the various authors. This analysis was done using the VOS viewer software as previously reported [19,22,23,24].

Statistical analysis

The dataset obtained in this study was analysed in Microsoft Excel where basic descriptive statistics (bar charts and percentages) was utilized.

Results

The trend in HPAIV research from 2003 to 2023 is presented in Figure 1. The results indicate a total of 8,411 publications with 2009 and 2010 having the highest number of publications with 649 and 641, respectively. The year with the lowest number of publications was 2003 with 47. Starting from 2013, the number of HPAIV publications has been on the decline, reaching 297 in 2021, which is lower than that of 2006 which was 426 publications. A closer look at the trend of publications reveals that the rise in the number of publications to corresponds with the wave of the pandemics, especially for the 2003-2005, 2006-2010 and 2020-till date. However, this is not the case with the 2013-2017 wave, during which the number of publications witnessed a decline.

The total number of published articles according to the type of article is presented in Supplementary Table S1. These included reports, reviews (ordinary and systematic), symposia, books/chapters, corrections, correspondences, original articles among others. From the 8,411 articles finally included in the study, original articles dominated with a total of 7,553, representing 89.80% followed by ordinary and systematic reviews that accounted for 7.54% (n=604). The top 15 most productive authors, H5N1 or HPAIV-related publications, total link strength, hi-index, and their affiliations are presented in Supplementary Table S2. The results further indicate that Kawaoka Yoshihiro, followed by Webster Robert were the top two most productive authors in terms of number of avian influenza virus-related publications indexed in PubMed, with 117 and 109 publications, respectively.

The collaborative research of the authors as well as the number of publications of the various authors is presented in Figure 2. The results, as shown in Figure 2, display a network of different bubble sizes and colors connected with various colored lines, both short and long. The size of the bubbles is proportional to the number of publications of the various authors. Based on the sizes of the bubbles, it is evident that Webster Robert and Kawaoka Yoshihiro were the top two authors in terms of publications related to avian influenza virus. The collaborative efforts between the various authors are indicated by the various connecting lines. The most collaborative authors are Webster Robert, Kawaoka Yoshihiro, Swayne David, Katz Jacquelin, Kida Hiroshi, and Fouchier Ron in no particular order as depicted by the various colored lines. The lead position by the top author could also be attributed to the fact that he has the highest affiliations.

Supplementary Table S3 indicates the distribution of the top 15 organizations/institutions/universities involved in avian influenza virus research around the world. The countries involved in the researches were Belgium (n=1), China (n=4), Egypt (n=2), Japan (n=1), the Netherland (n=1), Thailand (n=1), and USA (n=4). The top three research institutes were University of Tokyo Japan, Erasmus Medical Centre, Rotterdam Netherland, and St. Jude Children's Research Hospital Memphis, USA. Figure 3 shows the various participating institutions/organizations/universities. The lack of connecting lines indicates the absence of a strong collaboration between the various participating institutions/organizations/universities.

The visualization of the keywords utilized in HPAIV research globally, their occurrence and their connectivity with to one another is shown in Figure 4. On the other hand, Figure 5 not only indicates the keywords utilized but it goes further to show both the old and new keywords utilized in avian influenza virus research globally. From Figure 4 and Figure 5, influenza A virus, H5N1 subtype, sits at the center, and it has the largest bubble size which indicates that it is the most utilized keyword and this is evident by the fact that it had the highest occurrence. It was also the keyword that was connected the most with other keywords in the various studies. Other keywords as shown in Figure 4 and Figure 5 that were also frequently mentioned were chicken, phylogeny, avian influenza virus in bird, vaccine, and gender (female and male) to mention a few. As further shown in Figure 5, various new research keywords are emerging and these are displayed in yellow and these include H5N8, wild bird species, mammals, and Nigeria. The top 15 journals that published HPAIV H5N1 research is captured in Supplementary Table S4. The results indicate that these journals published at least 107 HPAIV H5N1 research papers. The top three journals were Journal of Virology, PLoS One and Vaccines, respectively.

Discussion

The present bibliometric study describes various studies conducted on HPAIV H5N1 globally from 2003 to 2023 indexed in PubMed. The trajectory of the research over the years revealed that 2003 had the fewest number of publications, followed by 2004 and 2005. However, in 2006, there was a spike in the number of publications, which may be attributed to the second global wave of HPAI outbreaks including the maiden incursion of the virus into Africa [25]. The jump in the number of publications observed in 2006 and 2007 can be explained in part to be due to the introduction of the H5N1 strain into the African continent, first in Nigeria and Egypt in 2006 [25]. The outbreak in Nigeria, which occurred in February 2006 in Kaduna, later spread to 25 states out of 36 in Nigeria killing at least 336,000 birds and persisting for 21 months [26]. Since then, various outbreaks have been reported worldwide [12,13]. The results further indicate that 2009 and 2010 witnessed the highest number of publications as scientific exploration to better understand the new emerging virus increased, and since then, the number of publications has been on a slight decline even though that of 2023 is almost ten times higher than that of 2003.

Despite the decline in the number of publications, the outbreaks of the HPAIV H5N1 subtype continue to rise across the different continents [12]. This can be explained by the fact that some of the studies appear to be classified and are not allowed to be published as evidenced by the various calls by researchers advocating for more research on HPAIV H5N1 and the lifting of the ban on certain publications [27,28]. The restriction on publications can also be attributed to biosecurity concerns expressed by experts and regulatory agencies, especially regarding gain of function experiments which were feared could be disastrous if unregulated and can speed up the potential of HPAI causing a pandemic [28,29,30]. It has been reported that there is a need to control H5N1 virus research and related pathogens that pose a national and global biosecurity concern; however, there is a significant risk of a pandemic from the HPAIV H5N1 [31]. At the moment, there is an estimated 52% mortality rate for the virus [12,32]. However, in the event of a pandemic, the mortality rate may be much higher [14].

An evaluation of the study types revealed that original research was the most dominant research type observed. Here, original research refers to studies conducted in either a wet-lab or dry-lab scenario on HPAIV H5N1. Other important study types were reviews (ordinary and systematic), letters to editors, and clinical trials. The dominance of original research articles corroborates a previous bibliometric study, which showed original articles accounting for 92% of their retrieved studies in PubMed on HIV/AIDS conducted in Nigeria. Similarly, the authors also reported a low level of systematic review as observed in our study compared to ordinary or narrative reviews. In this study, a total of 55 systematic reviews were retrieved representing less than 1% of the total studies. Systematic reviews represent the highest evidence in medicine and when well-written can provide data that can inform clinical practice, future research and policy [33]. There is a pressing need for systematic reviews, which will help enhance our understandings of the roles of wild birds, live bird market, wetland and mammals (pigs, goats, sheep, cattle) in the spread of HPAIV H5N1. Although there are a substantial number of studies bordering on vaccine, approximately 704, it is essential to leverage the technological advancements and infrastructure developed for COVID-19 vaccine production, which enabled vaccine development in record time [34,35]. Given that an outbreak of the virus appears to be inevitable in the future, there is a need to develop such capacities now in preparation for a potential pandemic from H5N1 or any other zoonotic virus with pandemic potential. Adequate clinical tests of present and future vaccines for efficacy, immune tolerance, immunological profiles and safety could be the turning point in the event of a future pandemic. This would reduce the hypothetical issues that surrounded various COVID-19 vaccines [34].

An analysis of the top 15 authors revealed that collectively, the top 15 authors had a total of 1043 publications, accounting for 12.40% of the total publications indexed in PubMed on HPAIV H5N1 globally. This implies that these authors were the most productive as far as HPAIV H5N1 is concerned. An examination of their affiliations of the top 15 authors revealed that USA, followed by China were the most frequent countries of the authors; however, Kawaoka Yoshihiro affiliated to University of Wisconsin-Madison/University of Tokyo was the most productive author with 172 publications. A recent ranking of the University of Tokyo revealed an overall ranking of 15 in academics further attesting to their visibility in research [36]. It is important to note that although other authors worldwide did not make the top 15 authors, their combined publications accounted for over 85% of the total publications on HPAIV H5N1, which is a good contribution. Furthermore, as revealed by the authorship network analyses, there is a gross absence of research collaboration among the participating universities. Collaborative research provides a platform for the exchange of ideas, skills and expertise, while facilitating access to funding [37]. Even though there is substantial evidence of collaborative research, some of the authors had limited collaborations. The invisibility of Africa authors among the top 15 authors does not indicate or suggest that HPAIV H5N1 research is not ongoing on the continent. This is further evidenced by the fact that among the top 15 research institutes/universities there were two African-based research institutes located in Egypt. Among the research organizations, veterinary and agriculture-based research institutes were most mentioned, probably as a result of the fact that HPAIV H5N1 is a zoonotic disease and can disrupt the food security chain network around the world [12,31]. Even though the National Veterinary Research Institute in Nigeria did not make the top 15 institution, it made the top 50 research institutes and was responsible for the bulk largest part the research on HPAIV coming from Nigeria. Another possible explanation for the low research input from African researchers may be due to lack of funding which has been a serious issue as far as HPAIV H5N1 is concerned [38]. Furthermore, their inability to publish their work in high impact research journals could also be a factor. It is important to note that authors from Nigeria have published widely on various avian influenza virus strains [4,8,25,26,39]. Securing more funding will certainly translate into more HPAIV H5N1-related research and preparedness. Overall, there is a need for more collaborative research among the various authors and the wider scientific community, and for countries to work together in all sectors to protect lives (animals and humans) from the virus [12].

The keyword analysis revealed as expected the use of numerous keywords. Keywords are an important aspect of research as they aid in the screening of articles and also at a glance provide the readers with valuable insight into papers [40]. The most frequently used keyword was influenza A virus, H5N1 subtype. This highlights the centrality of H5N1 in avian influenza virus research, as it remains the strain most implicated in avian influenza virus outbreaks, despite the fact that other subtypes have been reported, such as H5N8 [41], H5N6 [42], H5N2 [43], and LPAI H9N2 [44,45]. Its connectivity with other keywords in the keyword analysis further indicates the diverse research conducted around the HPAIV H5N1 strain. These keywords were chicken, phylogeny, and vaccine among others. Its connection with chicken, wild birds and mammals indicates outbreak and detection of the virus in various hosts or species as revealed by various studies [4,8,39]. Combined, all the potential species where HPAIV H5N1 strain can be detected (chicken, ducks, wild birds, domestic birds, poultry, ostrich, turkey and waterfowl) had more mentions than all other keywords apart from HPAIV H5N1 strain and the list of species continues to increase [8]. In a recent study, a prevalence rate of 11.1% was observed in apparently healthy diverse wild bird species in Nigeria [8]. As one of the keywords, Nigeria as a country was mentioned, and this could be attributed to the fact that, together with Egypt, both countries continue to record the highest number of outbreaks of HPAIV H5N1 on the continent in poultry settings [4,20]. The inter-connectivity of the avian influenza with wild birds, chicken, turkey, mammals, poultry and zoonosis goes to attest to the ability of the virus to not just cause infection in birds other than chicken but also in mammals and animals [45].

The findings in the study have several implications for clinical practice and public health. As can be observed from the retrieved articles, there is a need to prioritize research aimed at developing therapeutics or vaccines for the H5N1 virus. There is also a need to enhance surveillance systems in high-risk areas such as Nigeria that emerged as one of the new keywords. This is particularly important due to the fact that Nigeria as a country had in the past reported outbreaks of viruses with pandemic potential, including the Ebola virus. Furthermore, the findings suggest the need to develop and implement comprehensive evidence-based guidelines aimed at controlling and preventing avian influenza outbreaks in animals and a potential spillover to humans.

The present study had some limitations. First, only PubMed was used. Thus, it may not truly represent the totality of studies conducted on the HPAIV H5N1 subtype globally. Second, the study focused on only one of the subtypes of HPAI. Third, studies conducted in 2024 were excluded.

Conclusions

Despite the highlighted limitations of the study, this study has shown the trend of HPAIV H5N1 research globally from 2003 to 2023. Our findings have also shown that the HPAIV H5N1 research in 2003 remains the lowest in terms of number, probably due to the absence of a major outbreak as of 2003. Our findings further indicate that the period 2006 to 2017 had the highest number of publications with 2009 and 2010 topping the chart of publications. Worrisome is the fact that the volume of research began to drop from 2018 to 2023. In addition, the study has also shown the top 15 authors and institutions and how they collaborate as far as HPAIV H5N1 research is concerned. Also revealed were the various keywords, both old and new, which are involved in HPAIV H5N1 around the world. Our findings suggest that further research is necessary to develop therapeutics and vaccines, as well as to maintain continuous surveillance in high-risk areas. Furthermore, the findings provide baseline data for all the stakeholders involved in H5N1 research, which could inform future research, funding, and policies.