A bibliometrics and visualization analysis of cannabidiol research from 2004 to 2021

Category analysis

The category of publications is an important indicator to reveal the hot spots and applications of CBD research. Figure 3A presented the top 10 most published categories in WOS SCI-E database. There are a total of 1271 publications divided into Pharmacology/Pharmacy, which account for 31.70% of the top 10 most published categories. Neurosciences is the second largest published category, with 651 publications (16.24%), followed by Clinical Neurology with 492 publications (12.17%), Psychiatry with 437 publications (10.90%), and Biochemistry Molecular Biology with 316 publications (7.88%). Subsequently, we performed the co-occurrence network of categories based on the dataset of WOS using Citespace software. 248 nodes and 1190 links were included in the network, and Figure 3B showed the top 10 most published categories. Each of the nodes represents a specific category, and the node size was used here to reflect the frequency of category occurrence. Betweenness centrality provides a method to quantify the importance of the node’s position in a network (Chen, 2006), and the nodes with purple trims indicate the high betweenness centrality. Figure 3B revealed that Pharmacology & Pharmacy was the highest co-occurrence category in the network, and its centrality was 0.14. Moreover, Chemistry presented the maximum centrality (0.26), suggesting that Chemistry is a pivotal point or tipping point in the network.

Co-authorship analysis

To investigate the co-authorship between different countries, institutions, and authors, we performed the collaboration network based on the WOS dataset using VOSviewer software. There are a total of 94 countries, 3,569 institutions, and 14,040 authors who have contributed to this topic. As shown in Figure 4, each node represents a different country/institution/author, the node size represents the number of documents, the thickness of connecting lines represents the strength of inter collaboration, and each color represents a cluster. Figure 4A presented the inter collaboration between different countries. Some of the 94 countries in the network were not connected, and the largest set consisted of 86 countries. The United States was the most productive and cooperative country, with 573 link strengths. The second most cooperative country was England, with a sum of 412 link strengths, and the third-highest number of documents. Italy, Spain, Canada, Germany, and Brazil also showed high productivity and cooperative strength in the network.

Figure 4B displayed the inter collaboration between different institutions. We picked out the top 100 most productive institutions to generate the partnership network, but two institutions were excluded due to a lack of cooperation with other institutions. The University of Sao Paulo was the most productive and cited institution, with the highest cooperative link strength. The University of Sao Paulo was the first and the largest modern comprehensive university in Brazil. It is also the most important scientific research center in Brazil. Table 1 submitted detailed information on the top 10 most cooperative institutions. Nine of the top 10 most cooperative institutions were from different research universities and only one was owned by a pharmaceutical company, indicating that the current study on CBD was still dominated by basic research. GW Res Ltd. established a world-leading position in the development of plant-derived cannabinoid therapeutics, which has developed the first FDA-approved CBD-based drug, Epidiolex for the treatment of Dravet syndrome and Lennox-Gastaut syndrome.

Figure 4C showed the inter collaboration between different authors. 106 authors met the threshold of 10 or more published articles, while only 86 authors were founded to connect with the other authors in the network. The most productive author was Guimaraes FS with a total of 51 documents, 2,406 citations, and 65 link strengths. However, the most cited author was Di Marzo V with a sum of 4,419 citations, 47 documents, and 53 link strengths. The most cooperative author was Huestis MA with a total of 98 link strengths, 45 documents, and 1,684 citations.

Funding agency contribution

The source of funding reflects, to a certain extent, the different levels of the country’s support for the field at different stages and the enthusiasm of the research community. Table 2 presented the top 10 most productive funding agencies for CBD research. United States Department of Health and Human Services was listed in the first, providing funds for 426 documents, which accounts for 11.98% of all the publications. The followed funding agencies were the National Institutes of Health (NIH)-USA and the European Commission with 422 and 360 documents, accounting for 11.87% and 10.13% of all the publications, respectively. Three of the top 10 most productive funding agencies were from the United States and sponsored a total of 1,071 publications, which accounts for 30.12% of all the publications. Additionally, the residual top 10 most productive funding agencies including three Brazil agencies, two England agencies, and one Canada agency, accounting for 12.19%, 4.81%, and 1.97% of all the publications, respectively.

Journal contribution

A total of 3,555 publications were published in 999 different journals. Table 3 listed the top 10 most-cited journals in the CBD area. The most cited journal was the British Journal of Pharmacology with cited times of 9,906, and a sum of 95 papers were published. Epilepsia was listed second with 2,212 citations and 42 papers, followed by Neuropsychopharmacology with 2,093 citations and 21 papers. Four of the 10 listed journals were from England, other four journals were from the United States, and the rest were from Germany and Switzerlands. The top three journals with the highest impact factor (IF) were the British Journal of Pharmacology (8.739), Neuropsychopharmacology (7.885), and Epilepsia (5.866). In addition, the H-index and ISSN of the journal were also listed in Table 3.

Keywords co-occurrence and burst detection

The keyword plays an important role to highlight and emphasize the focus and core content of the whole article. The keyword co-occurrence network enables an overview of the core content of published articles, as well as reveals the connections between the content. There are a total of 11,116 keywords included in the 3555 publications, and 282 keywords meeting the frequency threshold of more than 20. Figure 5A displayed the co-occurrence network consisting of 282 keywords using the Vosviewer software. Each node represents a keyword, the node size indicates the frequency of the keyword, and each color reveals a cluster in the network. The top 15 keywords with the highest occurrence frequencies and strongest connection were cannabidiol (1,960, 11,444), cannabinoids (949, 5,747), cannabis (535, 3,336), endocannabinoid system (405, 2,750), delta (9)-tetrahydrocannabinol (394, 2,770), CBD (363, 2,393), marijuana (348, 2,236), THC (326, 2,275), delta-9-tetrahydrocannabinol (303, 2,204), and double-blind (282, 2,024), cannabinoid (248, 1,565), epilepsy (240, 1,565), expression (233, 1,376), in-vitro (214, 1,385), seizures (213, 1,316). 282 keywords in the network were divided into five clusters, and the detailed information was summarized as follows: 1) Mechanism was represented by red: cannabinoid receptor, CB1, CB2, inflammation, antioxidant, protein, and apoptosis; 2) Disease was represented by blue: schizophrenia, anxiety, psychosis, behavior, rat, prefrontal cortex, and cognitive impairment; 3) Treatment of epilepsy was represented by purple: epilepsy, Dravet syndrome, therapy, seizures, antiepileptic drugs, and children; 4) Clinical trial was represented by yellow: double-blind, neuropathic pain, efficacy, medical cannabis, safety, risk, and tolerability; 5) Synthesis and metabolism were represented by green: bioavailability, blood, plasma, chromatography, metabolism, pharmacokinetics, synthetic cannabinoids, extraction, identification, and quantification.

Brust detection served to recognize the emergent terms and sharp increases of interest (Chen, 2006), which significantly contributed to revealing the frontiers in one specific field. We performed the keyword burst detection based on the dataset of WOS using Citespace software, and the top 10 keywords with the strongest bursts were displayed in Figure 5B. Anandamide was the top 1 strongest burst keyword that emerged in 2004 and ended in 2015, revealing that anandamide was extensively studied and discussed during this period. The seizure was the most recent keyword with high burst strength that appeared in 2018. The bursts of the keyword cannabinoid receptor, constituent, in vivo, mice, and 5-ht1a receptor sustained at least 10 years.

Reference co-citation analysis and timeline map

Reference citation and co-citation frequencies could clearly illustrate the intellectual concerns in a field. In general, an article with a high co-citation frequency suggests it has a high citation frequency. Table 4 listed the top 10 most cited articles from 2004 to 2021 in the CBD field. The most cited reference was published by Ryberg et al. (2007), which revealed that the orphan receptor GPR55 was a novel cannabinoid receptor. A review published by Pertwee (2008) summarized the diverse CB1 and CB2 receptor pharmacology of three plant cannabinoids, indicating that CBD displayed high potency as a CB2 receptor antagonist. The reference published by Russo (2011) reviewed the pharmacology effects of CBD in anti-oxidant, anti-anxiety, anti-depressant, anti-convulsant, and so on, and the possible mechanisms were also discussed. Devinsky et al. (2017) conducted a double-blind trial of cannabidiol for drug-resistant seizures in 120 individuals, revealing that CBD exhibited greater effects in decreasing the convulsive-seizure frequency than placebo. This research was published in the New England Journal of Medicine with the highest H-index (1030) in the top 10 most cited articles. Moreover, Leweke et al. (2012) demonstrated that CBD alleviated the psychotic symptoms of schizophrenia by enhancing anandamide signaling. Mechoulam and Parker (2013) summarized the actions of the endocannabinoid system on anxiety, depression, neurogenesis, reward, cognition, learning, and memory. Notably, three of the top 10 most cited articles attached concerns to the pharmacology effects of CBD on seizures, and these articles were published by the team of Devinsky et al. (2014), Devinsky et al. (2016), Devinsky et al. (2017), another three articles focus on the mechanisms and molecular targets of CBD pharmacology effects (Ryberg et al., 2007; Pertwee, 2008; Mechoulam and Parker, 2013).

Next, we performed the co-citation network of reference based on the WOS dataset using the Citespace software. 1,248 nodes and 6,077 links were founded in Figure 6A, and the top 10 most co-cited articles were highlighted in the network. The top 2 most co-cited articles were published by Devinsky et al. (2016), Devinsky et al. (2017), with 354 and 288 co-citations, respectively. The former article assessed the safety, tolerability and efficacy of CBD in patients already receiving stable doses of antiepileptic drugs, indicating that CBD might decrease the frequency of seizure frequency with sufficient safety in children and young adults with highly treatment-resistant epilepsy (Devinsky et al., 2016). The latter focused on drug-resistant epilepsy in Dravet syndrome and evaluated the efficacy and safety of CBD in it, suggesting that cannabidiol reduced the frequency of convulsive seizures in children and young adults with Dravet syndrome but was associated with adverse events including somnolence and elevation of liver enzyme levels (Devinsky et al., 2017). The followed most co-cited articles were published by Thiele et al. (2018) with 238 co-citations, and Devinsky et al. (2018) with 222 co-citations, both of them explored the effects of CBD on seizures in the Lennox–Gastaut syndrome through the clinical trial. Laprairie et al. (2015) illustrated that Cannabidiol is a non-competitive negative allosteric modulator of CB1 receptors. Besides, the side effects, toxicity, and drug-drug interaction of CBD therapies were also discussed based on clinical data and animal studies (Geffrey et al., 2015; Whiting et al., 2015; Iffland and Grotenhermen, 2017). In summary, the clinical trials of CBD on seizures have absorbed great interest from researchers, and investigated the mechanism of the CBD effect helps to better elucidate the pharmacology value of CBD. In addition, there are serious concerns about the issues of safety and side effects of CBD, more trials still need to be conducted for solving these problems.

Clusters of co-cited articles provided a method to detect the transition of research frontiers in an area (Chen, 2006). Figure 6B displayed the clusters of co-cited articles using Citespace software. The modularity Q and silhouette values of the clusters were 0.791 and 0.9081, respectively, which indicated the structure of the clusters was significant and all of the clusters were convincing. All of the co-cited articles were assigned to 18 different clusters, and the clusters of the same color represent co-citations made within the same time slice. The first cluster of co-cited articles was epilepsy, and the followed were anxiety, inflammatory bowel disease, cocaine, multiple sclerosis, HPLC, basal ganglia, cannabis, and oxidative stress. Figure 6C provided a timeline view of the top 10 clusters, and detailed information about these clusters was listed in Table 5. The early studies focused on the pharmacology effects of CBD on multiple sclerosis (2004), inflammatory bowel disease (2008), and anxiety (2009), as well as the regulatory effect of CBD on the endocannabinoid system and the serotonin system in the basal ganglia (2002). In 2012, the anti-oxidative effects of CBD attracted the attention of researchers. In 2014, researchers were interested in exploring the values of cannabis. Notably, the effects of CBD in the treatment of epilepsy have been enthusiastically investigated in 2016, and great progress has also been made in drug development. The most recent themes in the CBD area were cocaine and HPLC. CBD could be used as a potential therapeutic drug for the treatment of cocaine use disorders (Mahmud et al., 2017; Luján et al., 2020).