Research Trends and Hotspots of Q Fever Research: A Bibliometric Analysis 1990-2019

Q fever is a worldwide distributed zoonosis caused by Coxiella burnetii, a Gram-negative bacterium. Despite existence of large amount of research data on the developments related to Q fever, no bibliometric analysis of this subject is available to our knowledge. Bibliometric studies are an essential resource to track scholarly trends and research output in a subject. This study is aimed at reporting a bibliometric analysis of publications related to Q fever (2,840 articles published in the period 1990-2019) retrieved from Science Citation Index Expanded, an online database of Clarivate Analytics Web of Science Core Collection. Data was retrieved using keywords “Q fever” or “Coxiella burnetii” in title, abstract, and author keywords to describe important research indicators such as the kind and language of articles, the most important publications, research journals and categories, authors, institutions, and the countries having the most significant contribution to this subject. Finally, the emerging areas in field of diagnosis, host range, and clinical presentation were identified. Word cluster analysis of research related to Q fever revealed that major focus of research has been on zoonosis, seroprevalence, laboratory diagnosis (mainly using ELISA and PCR), clinical manifestations (abortion and endocarditis), vectors (ticks), and hosts (sheep, goat, and cattle). This bibliometric study is intended to visualize the existing research landscape and future trends in Q fever to assist in future knowledge exchange and research collaborations.


Introduction
Q fever is recognized as a global zoonotic disease that has been declared as potential bioterrorism category B select agent by the Centre for Disease Control and Prevention (CDC) [1]. It is regarded as a reportable disease in some countries [2]. This disease is caused by Coxiella burnetii, an obligate Gram-negative bacterium, which can infect human, various animals such as ruminants (cattle, goat, and sheep), pets, birds, ticks, and rarely reptiles and marine mammals [3]. This bacterium is secreted in birth products (such as placenta), urine, milk, and faeces [3,4].
Main route of transmission is inhalation of contaminated aerosols. However, ingestion of contaminated raw milk can at least cause seroconversion. Human-to-human transmission was described and might happen through contaminated blood transfusion, sexual contact, and exposure to contaminated birth products of women. Mainly, this disease is reported in humans having close contact with infected animals and their products [3].
Q fever can manifest as an acute or chronic disease. Acute infections are mostly asymptomatic (60%) or manifests as a flu-like and often self-limiting disease. Symptoms include but are not limited to flu-like symptoms, endocarditis, hepatitis, pneumonia, abortion, and premature fetal death in pregnant women and neuropathies [5]. Differentiation of acute from chronic Q fever solely on clinical manifestation may be misleading. Currently, acute and chronic forms are differentiated on the basis of different antibodies present in the sera of the patient. This demonstrates that presence of IgG to phase I indicates the chronic form while detection of IgG to phase II antigen demonstrates acute form [6]. In most of the cases, it is asymptomatic and therefore remains underreported. Different techniques are used for its diagnosis such as IgG-based serological assays and DNA-based molecular assays [5]. In symptomatic patients seeking medical advice, this disease can be treated through administration of antibiotics such as doxycycline [3].
Despite existence of large amount of research data on Q fever, to best of our knowledge, no bibliometric analysis of this topic is available. Bibliometrics makes it easy to investigate and decipher different developments on a subject to pursue the dynamics and evolution of scientific knowledge. Identifying future research directions based on a bibliometric analysis of the characteristics of available literature in a field reduces the error margin and thus improves the decision-making. These indicators may further be helpful for early career researcher to identify the latest developments of the topic.
The present bibliometric study is intended to analyze 3,673 Q fever-associated publications retrieved from the Science Citation Index Expanded (SCI-EXPANDED) database of Web of Science, the most-acknowledged database in bibliometric studies [7]. The retrieved data were analyzed according to language and type of publication, most productive authors, laboratories, countries, and scientific journals as well as the most cited articles. Based on this analysis, hotspots and recent trends in scientific developments pertinent to Q fever were identified and discussed in this study.

Methodology
This study relied on the data obtained from Clarivate Analytics Web of Science Core Collection (WoSCC), a platform of the SCI-EXPANDED. All data were obtained on February 2021 by searching the database for articles related to Q fever with the following parameters: WoS TOPIC (title, abstract, author keywords, and KeyWords Plus): ("Q fever" or "Coxiella burnetii") and year . Use of quotation marks (" ") is essential to identify the exact searched terminology by avoiding the lemmatization and synonym features of WoSCC (by default, in search setting, these features are ON) [7]. Owing to this search feature, it was necessary to find different expressions, and therefore, Boolean operator "or" was used which ensured the appearance of at least one term ("Q fever" or "Coxiella burnetii") in the topic.
Database search by using this strategy yielded 4,270 documents including 3,242 articles. It is important to note that additional search terms can be obtained by KeyWords Plus which are extracted from article titles enlisted as references (in reference list or footnotes) in the Clarivate Analytics database which causes a substantial increase in title-word and author-keyword indexing [8]. SCI-EXPANDED has been widely used for bibliometric studies, but it is mainly designed to facilitate authors to search suitable literature but not meant for direct bibliometric analyses [9]. Therefore, to avoid inclusion of irrelevant documents to the desired ones, use of SCI-EXPANDED necessitate the accurate bibliometric analysis instead of relying only on KeyWord Plus [10]. Therefore, use of "front page" (that considers the abstract, manuscript title, and author keywords) has been suggested as used as a filter [11]. This filter has been employed in present study by considering only the search keywords which were found in their "front page." This modified method by considering "front page" as a filter resulted in 3,673 documents (86% of 4,270). These documents were used for further analysis.
Impact factor of journals (mentioned as IF 2019 ) was obtained from the Journal Citation Reports in 2019. After downloading, the data were arranged using Microsoft Excel 2016 as described elsewhere [12]. The term of "corresponding author" (RP) was used though it is designated as "reprint author" in SCI-EXPANDED database [13]. Single author in articles with unspecified authorship was both the first as well as corresponding author. Similarly, articles published by single institution were classified as the institution of the first author and corresponding author [13]. Moreover, in articles having multiple corresponding authors, we only considered the last corresponding author. Type of collaboration was assessed by using addresses of the authors. Geographic location of the countries was determined as described elsewhere [7].
Publication citations were assessed using following indicators: (i) C year . The number of citations from WoSCC in a particular year (e.g., C 2019 describes citation count in 2019) [13] (ii) TC year . The total citations from WoSCC received since publication year till the end of the most recent year (2019 in this study, TC 2019 ) [14] (iii) CPP year . Citations per publication (CPP 2019 = TC 2019 /TP), TP: total number of publications [13] 3. Results and Discussion 3.1. Characteristics of Document Types. It has been recently suggested to identify the characteristics of a document type on the basis of their citations per publication (CPP year = TC year /TP) and number of authors per publication (APP = AU/TP) [15]. Use of TC 2019 for CPP 2019 is advantageous owing to their invariability and ensured repeatability as compared to the citation index from WoSCC [16]. A total of 3,673 Q fever related documents published in SCI-EXPANDED were found among 12 document types which are listed in Table 1. This publication count includes 2,840 (77%) articles having 6.0 as APP (number of authors per publication) which was higher than most other document types except reprint with an APP value of 16. The percentage of articles focusing on Q fever (77%) was higher than other medical-related topics, for example, 70% in Ebola [17], cisplatin-based chemotherapy for small cell lung cancer (68%) [18], and acupuncture (66%) [19] but similar to 75% human papillomavirus [20] and 79% in asthma in children 2 BioMed Research International [21]. The document type "review" with 217 documents had the greatest CPP 2019 value of 53 followed by articles with CPP 2019 of 20. The CPP 2019 of the document type "review" was found to be 2.6 times of articles. Six of the top ten documents in TC 2019 were reviews in Q fever research including "Q fever" with TC 2019 of 1,220 [22], "Ticks and tickborne bacterial diseases in humans: An emerging infectious threat" with TC 2019 of 585 [23], "Antimicrobial mechanisms of phagocytes and bacterial evasion strategies" with TC 2019 of 485 [24], "Natural history and pathophysiology of Q fever" with TC 2019 of 434 [25], "Endocarditis due to rare and fastidious bacteria" with TC 2019 of 398 [26], and "Diagnosis of Q fever" with TC 2019 of 372 [27]. It is important to point out that documents could be categorized in two document types in WoS. For example, the highly cited document entitled "Comparative study of the clinical presentation of Legionella pneumonia and other community-acquired pneumonias" [28] was classified as "article" and "proceedings paper." Therefore, cumulative percentages exceed 100%.

Characteristics of Publication Outputs.
To determine CPP 2019 , use of TC 2019 was found better than citation index from WoSCC directly because they are consistent and ensure repeatability [30]. To better understand the trends and impacts of publications in a particular research theme, Ho established a correlation between total articles (TP) in a year and their citations (CPP year = TC year /TP) by the decades [13] and years [18]. Recently, it has been applied in medical-related topics, for example, pain [31], Ebola [17], and dengue [32]. Between 1990 and 2019, 2,840 articles associated to Q fever were published. The mean value of TC 2019 was 20 with 1,903 as the maximal value for an article.  Figure 1, citations per publication related to Q fever articles attained a plateau in 10 years of publication. Similarly, one decade was taken to attain a plateau by articles related to dengue [32] and Ebola [17].  [29]. Total 615 journals published articles related to Q fever in 87 WoS categories in SCI-EXPANDED. Following five categories were highly productive with total number of 2,018 articles (71% of 2,836 articles) including (i) infectious diseases with 902 articles (32% of 2,836 articles), (ii) microbiology with 739 (26%) articles, (iii) veterinary sciences with 533 (19%), (iv) immunology with 499 (18%) articles, and (v) public, environmental, and occupational health with 350 (12%) articles. Similar to articles, journals can also be listed in more than one category in WoSCC like Clinical Infectious Diseases belongs to multiple categories (immunology and infectious 3 BioMed Research International diseases and microbiology). Therefore, cumulative percentage of categories exceeds 100% [34]. In total of 144 articles (rank 9th) were published in category of parasitology including 16 articles (rank 6th) in 2019.
In total, 2,840 Q fever-related articles were published in 615 journals including 539 listed in SCI-EXPANDED in 2019. Table 2

Publication Performances: Countries and Institutions.
In order to evaluate publication output of institutions and countries, Ho's group proposed six publication indicators including the total number of publications (TP), first-author publications (FP), single-author publications (SP), correspondingauthor publications (RP), independent publications (IP), and collaborative publications (CP) [29]. First and the corresponding authors are widely recognized for having most of the contribution in an article [35]. At the institutional level, the determined institution of the corresponding author might be the hosting institution or university of the study or origin of the paper [13]. There were 16 articles (0.56% of 2,840 articles) without affiliations in SCI-EXPANDED. Q fever articles (2,824) were published by authors affiliated from 121 countries. However, 2,216 articles (78% of 2,824 articles) were published from 65 countries by authors affiliated from single country, while remaining 608 articles (22% of 2824) were published by authors from 117 countries as internationally collaborative articles. This situation demonstrates that collaborative research is very limited and should be focused on future studies. Table 3 enlists the top 10 most productive countries with six publication indicators [36] and a citation indicator (CPP 2019 ). Use of CPP 2019 is advantageous due to its invariant nature as compared to the citation index from the WoSCC which is regularly updated [11]. The USA was the topranking country in the six publication indicators with TP of 633 articles (22% of 2,824 articles), IP of 434 articles (20% of 2,216 independent articles), CP of 199 articles (33% of 608 internationally collaborative articles), FP of 517 articles (18% of 2,824 first-author articles), RP of 506 articles (19% of 2,733 corresponding-author articles), and SP of 29 articles (32% of 92 single-author articles). France had the highest CPP 2019 of 30 while Germany, Spain, and Netherland had only 16. Among the 121 countries published Q fever articles, 56 countries (42% of 121 countries) had no single country articles while four countries (3.3%) had no internationally collaborative articles. Similarly, 32 (26%), 37 (31%), and 95 (79%) countries have no first-author, corresponding-author, and singleauthor articles, respectively. Trends in publication of the top five productive countries in 2019 are presented in Figure 2. Iran had TP of 46 (ranked 15th) and was a new member of productive countries with 11 articles in 2019.
With regard to institutions, 919 Q fever-related articles (33% of 2,824 articles) originated from single institutions while 1,905 articles (67%) were institutional collaborations. List of top 10 productive institutions and their characteristics  3.5. Publication Performances: Authors. For articles related to Q fever, average number of authors per Q fever article was 6.0 whereas maximum number of authors was 27 in one article. Of the 2,837 articles with author information, 407 (14%), 391 (14%), and 374 (13%) were written by groups of 4, 5, and 6 authors, respectively. Figure 3 shows the relationship among number of articles (TP) and citations per publication (CPP 2019 ) by number of authors in an article. There were two articles which were published by 24 authors [37,38] including the one with the highest CPP 2019 of 184 entitled "Complete genome sequence of the Q-fever pathogen Coxiella burnetii" having the highest TP 2019 of 364 [38]. Table 5 lists the top 10 productive authors, among whom, D. Raoult was the only one published single-author articles. In addition, T.J. Marrie published the most of

BioMed Research International
single-author article with seven articles (7.2% of 97 singleauthor articles). Publication performance of authors was further analyzed. In recent years, the Y-index was suggested [13,33,34] to evaluate potential of publications and to characterize the scientific publications by authors, institutes and respective countries to the number of articles as first author (FP) and as corresponding author (RP). Recently, the Y -index has been demonstrated in research publications in medical research such as highly cited articles in health sciences and dentistry [39]. With two parameters (j, h), the Y -index can be helpful in visualizing and comparing among different author's publications and is described as: In Y-index diagram (Figure 4), the authors with higher j value are situated further away from origin of the polar coordinates (0, 0). Authors having similar number of articles as first and corresponding author would have h = 0:7854 (radian) and located in diagonal line. Furthermore, with more publications as corresponding author as compared to that as first author demonstrating h > 0:7854, the author would be represented in the upper left half quadrant of the Y-index diagram. However, an author having higher number of publications as first author than as corresponding author, with h < 0:7854, would be represented positioned in lower right half of the quadrant. The authors with h = 0 and j = number of first-author articles then would be positioned along x-axis of the diagram. However, when h = π/2 then j = number of corresponding-author articles, the author would be demonstrated along the y-axis of the diagram.
In total of 2,683 Q fever articles (95% of 2,840 publications), both as first and corresponding-authors in SCI-EXPANDED were extensively investigated based on Y -index. The 2,683 Q fever-related articles were contributed by 9,168 authors in which 6,934 authors (76% of 9,168 authors) had no firstor corresponding-author articles with Y − index = ð0, 0Þ; 439 (4.8%) authors published only corresponding-author articles with h = π/2; 99 (1.1%) authors published more corresponding-author articles with π/2 > h > 0:7854; 944 (10%) authors published the same number of firstand corresponding-author articles with h = 0:7854; 80 (0.87%) authors published more first-author articles with 0:7854 > h > 0; and 672 (7.3%) authors published only first-author articles with h = 0. In Figure 4, distribution of the Y-index (j, h) of the leading 17 potential authors with j ≥ 17 was demonstrated. Every point has a coordinate (j, h) that could symbolize a single-author or multiple authors, for example, L.M. Kampschreur and R. Sting with the same Y-index (18, 0.7854). D. Raoult (133, 1.435) published 284 Q fever-related articles which includes 16 articles as first-author and 117 articles as correspondingauthor with j value of 133 which is far away from the original (not in Figure 4). D. Raoult had the highest publication potential in Q fever research. It is worth mentioning that D. Raoult is a renowned infectious disease expert. He cultured 16% of the newly isolated bacteria from human [3]. Moreover, he has published ( Figure 4)

BioMed Research International
Kampschreur and Sting with a j of 18. Thus, the location on the graph along one of the curves or along a line from the origin represents different families of author publication potential or publication characteristics, respectively. It has been pointed out that with these data have a potential for bias in the analysis of authorship; it might attribute to different authors having the same name or the same author using different names over time [30].

Citation Histories of the Ten Most Frequently Cited
Articles. Total citations are updated weekly on the WoSCC.
To improve bibliometric study, the total number of citations from the WoSCC since publication to the end of the most recent year of 2019 (TC 2019 ) was applied to improve the bias using data from WoS directly. Use of TC 2019 is advantageous because of their invariability and ensured repeatability than the citation index from WoSCC [11]. The 2,840 Q fever articles were selected with search keywords within title, abstract, and author keywords from SCI-EXPANDED in the last three decades. A total of 2,104 articles (74% of 2,840 articles); 2,601 articles (92% of 2,686 articles with abstract); and 1,310 articles (46% of 1,720 articles with     Table 6 shows the top 10 most frequently cited articles with search keywords in their title and author keywords. The citation histories of the Q fever articles ranked top 40th in both TC 2019 and C 2019 are shown in Figure 5. These articles contain search keywords in their title or author keywords, were not only highly cited with T C 2019 ≥ 143 but also high impact in the most recent year of 2019 with C 2019 ≥ 13. They were summarized as follows: 3.6.1. Comparison of Coxiella burnetii Shedding in Milk of Dairy Bovine, Caprine, and Ovine Herds [41]. This article was published by 12 authors from five institutes: National Institute of Agricultural Research (INRA) in France, Adiagene in France, SNGTV in France, La Condamine in France, and Le Bourg in France with TC 2019 of 157 (ranked 30th) and C 2019 of 16 (ranked 20th). This study demonstrates the route of shedding of Coxiella burnetii in cattle, sheep, and goats. The study revealed that the bacterium is mainly excreted through milk of infected cattle and goats while in sheep it was mainly through faeces and vaginal excretions. The different route of shedding of bacteria may explain the higher association of human outbreaks with sheep as compared to cattle and goats.

Brucellosis and Q Fever Seroprevalences of Nomadic
Pastoralists and Their Livestock in Chad [42]. This article was published by seven authors from four institutes: Swiss Tropical Institute in Switzerland, Laboratoire de Recherches Vétérinaires et Zootechniques de Farcha in Chad, Direction de la Planification de la Formation in Chad, and Institute of Veterinary Bacteriology in Switzerland with TC 2019 of 154 (ranked 32nd) and C 2019 of 11 (ranked 47th).
This study investigated the association of seropositivity of Q fever in humans and animals in Chad. The authors reported that livestock remained a primary source of Q fever in humans that may be due to consumption of contaminated raw milk or through handling of placenta from infected animals.  Here, a short summary of these articles is given. Livestock plays a key role in maintenance and transmission of C. burnetii. However, the route of shedding of bacteria may vary depending on the specie of the animal. Infected cattle and goats shed the bacterium mainly through milk while infected  (6) [44]. Once infected, the goats may experi-ence reproductive problems and shed bacterium in at least two consecutive kidding seasons [45]. The vaginal secretions remain the suitable sample for detection of C. burnetii via PCR. On the other hand, milk and faeces may contain certain PCR inhibitory substances. By neutralizing the inhibitory PCR substances in faecal and improved DNA purification from milk samples, the efficacy of PCR may further be improved [43]. Additionally, molecular detection of C. burnetii DNA using PCR in clinical samples such as vaginal secretions, faeces, and milk of infected ewes highlights its diagnostic and disease transmission potential [43]. Guatteo et al. used real-time PCR for the detection of C. burnetii DNA from milk, faeces, and vaginal secretions of naturally infected cows. They highlighted that only 6% of infected cows shed the bacterium simultaneously through milk, faeces, and vaginal secretions. They also inferred that sampling strategy should consider prevalence and types of samples for serology and shedding [4]. Therefore, identification of DNA from different sample types should be preferred for better diagnosis of the disease. Guatteo et al. further demonstrated that infected cows may be sporadic or persistent shedder of C. burnetii. The persistent shedders are highly seropositive as compared to sporadic shedders. Therefore, antibody monitoring through serological assays such as ELISA could be an effective tool to identify the persistent and heavy shedders [46]. Furthermore, Schelling et al. demonstrated that human Q fever infection is associated with consumption of raw milk and direct contact with placenta of infected animals [42]. The association of human outbreak with infected animals was further reported while investigating an outbreak of Q fever in Netherland [47]. The higher incidence of Q fever in human was reported in areas with higher sheep densities and wind speeds which may speculate the higher aerosol transmission [48].  (5) Investigation of a Q fever outbreak in a rural area of The Netherlands Netherlands [47] 103 (88) 217 (5) Goats may experience reproductive failures and shed Coxiella burnetii at two successive parturitions after a Q fever infection France [45] 110 (87) 217 (5) Effect of vaccination with phase I and phase II Coxiella burnetii vaccines in pregnant goats France [52] TC Word distribution in article title, abstract, author keywords, and KeyWords Plus represents the most important information in conveying the findings of a study. Therefore, an analysis of word distribution can be very useful to evaluate the trends in a particular research field [49]. In the last decade, to determine research foci and trends during different years, Ho's group proposed distributions of article titles and abstracts, author keywords, and KeyWords Plus [49]. These analyses could minimize various limitations, such as the incomplete meaning of single words in article title and abstract, small sample size for author keywords, and the indirect relationship between KeyWords Plus and the research topics [50]. Therefore, these four kinds of words (in article titles, article abstracts, author keywords, and Key-Words Plus) were examined during the designated period to show the rough trends while minimizing the year-to-year fluctuations [49]. Distribution of words in article titles, article abstract, author keywords, and KeyWords Plus allowed us to visualize the major trends and foci in research related to Q fever over the years.
The 20 most frequently used author keywords of four subperiods (1990s, 2000s, and 2010s) are listed in Table 7. The most frequently used author keywords, except for the searching words, "Q fever" and "Coxiella burnetii," were seroprevalence, zoonosis, and ELISA. Overall, this table depicts that the most used words are related to the major focus of researchers working in the following fields related to Q fever: zoonosis/zoonoses (transmission of an infectious disease from animal to human), seroprevalence (measure of the disease burden through serological investigation), laboratory diagnosis (ELISA and PCR), clinical manifestations (abortion, endocarditis), vector (ticks), and hosts (sheep, goat and cattle). These topics are discussed below along with the results of word cluster analyses.
A word cluster analysis was also performed to identify the potential research hotspots in this topic. For this, synonymic expressions/words from words analyses denoting to a particular term were summed up. Our findings revealed that major focus of research about Q-fever dealt with its diagnostics, determination of host range, and clinical manifestation of the disease. We further investigated the keywords used in diagnostics, host range, and clinical manifestation that could facilitate to have an idea the most used keywords during different periods in each discipline.

Conclusion
This bibliometric analysis gives an insight on the developments in the discipline of Q fever, and it also provided the details of most influential publications, institutes, countries, and authors. Last decade showed the highest number of publications. Additionally, publications relevant to Q fever were published from the USA, Europe, and Australia. That could be helpful for researchers to collaborate with the relevant research groups and can further help post-docs or PhD for those pursuing their careers in Q fever research. PLoS One remained the most productive journal followed by "Infection and Immunity" and "Clinical and Infectious Diseases." This demonstrates the importance of these journals in research relevant to Q fever and can be helpful for authors looking to pub-lish similar research. Furthermore, research trends and hot areas including the clinical presentation, diagnosis, and host of Q fever were identified. Endocarditis, abortion, and pneumonia remained the most commonly used keywords to demonstrate clinical presentation. The use of PCR and ELISA, techniques used for the diagnosis, were most used over the recent years. In the end, we identified keywords pertinent to species infected with Q fever. This showed that human remained the most commonly used keyword followed by cattle and sheep. The presence of Q fever in human as well as various animal species such as cattle, sheep, and goat demonstrates its zoonotic importance further highlights the importance of collaborative effort from human physicians and veterinarians under the umbrella of one health to mitigate Q fever.

Data Availability
The data that support the findings of this study are available from the corresponding author upon reasonable request.