Frontier and hotspot evolution in pulmonary embolism: A bibliometric analysis from 2004 to 2023

Abstract

Pulmonary embolism (PE) is the third most common cause of cardiovascular death worldwide after stroke and heart attack. Research priorities in each PE period over the past 2 decades remain largely unassessed. This study aims to summarize PE research hotspots and trends using a bibliometric approach and predict future research hotspots. Relevant publications between 2004 and 2023 were extracted from the Web of Science Core Collection. Microsoft Excel 2016 and the R Bibliometrix package were used for descriptive and bibliometric data analyses. Country collaboration was visualized using VOSviewer. CiteSpace was used for the keyword burst analysis and timeline view construction. In total, 25,487 PE-related articles were included in the analysis. The annual number of publications followed a general upward trend. The USA was the most prolific country, with 9568 publications. The timeline of reference indicated 9 clusters including “#1 submassive PE,” “#2 suspected PE,” “#3 venous thromboembolism,” “#4 direct oral anticoagulants,” “#5 acute PE,” “#6 cancer-associated thrombosis,” “#7 medical patient,” “#8 PE,” and “#9 knee arthroplasty.” The keyword “helical computed tomography” had the highest burst strength, followed by “unfractionated heparin,” “society” and “molecular weight heparin.” The keywords “direct oral anticoagulants,” “venous thromboembolism,” “american society,” “society,” “coronavirus disease 2019,” “infection,” “case report,” “pneumonia” and “clinical characteristics” appeared frequently in recent years. Research on PE has focused on risk factor exploration, PE prevention and diagnosis, and then shifted to the treatment strategies, guideline updates, and specific types of PE. The use and feedback of guidelines, better improvement of prognosis, and life quality for PE-specific types might be potential future hotspots.

1. Introduction

Pulmonary embolism (PE) is defined as the blockage of the pulmonary artery or branch due to substances, such as tumor, thrombus, fat, or air, and is mainly caused by the ascending clot in the deep vein of the lower extremities.^[1,2] PE ranks third among officially reported causes of mortality for individuals receiving hospitalized care,^[3] as well as the third leading cause of cardiovascular mortality globally, after stroke and heart disease.^[4] The annual PE incidence rate has been increasing, particularly among the elderly population, with rates reaching approximately 300 cases per 100,000 people annually.^[5] Despite the decline in mortality, 90-day and in-hospital mortality rates remain high at 18% and 13.9% respectively.^[6,7]

Over the last 2 decades, numerous studies have been conducted on PE risk factors, mechanisms, diagnosis, treatment, and prognosis.^[8] Despite numerous conventional reviews on the current state of PE research, presentations on the research focus during different PE periods remain limited.^[7,9,10] Bibliometric is a statistical and mathematical technique that analyzes publications from various quantitative and qualitative disciplines. Bibliometrics assesses the impact and partnerships of various nations, organizations, journals, and authors. Additionally, they conducted cluster and burst analyses of keywords and references to pinpoint active research areas, measure scholarly output, and evaluate trends in specific domains.^[11,12] This study summarized the changes in the PE research field over the past 20 years and potential directions to provide a reference for future research.

2. Materials and methods

2.1. Data sources

On July 26, 2024, literature retrieval on PE was performed using the Web of Science Core Collection (WoSCC) database, which is commonly used in bibliometric analysis.^[13] The search term was: (TS = [“PE*” OR “pulmonary artery embolism” OR “pulmonary thromboembolism” OR “tremendous pulmonary artery embolism” OR “pulmonary artery embolization” OR “pulmonary arterial embolization” OR “lung artery clots” OR “lung embolism”]) AND (FPY = 2004–2023). Only articles and review articles written in English were included. Then, publications with research areas unrelated to the topic were excluded. The search results were selected as “Full Record and Cited References” and saved as a .txt file for storage. Primary data analyzed in this study can be retrieved by replicating the search term in the database. Figure 1 depicts the literature inclusion and exclusion flow chart.

Figure 1.

Flow diagram of literature selection.

2.2. Data analysis

Bibliometric analyses were conducted using Microsoft Excel 2016, CiteSpace (version 6.3 R1; developed by Chaomei Chen at Drexel University, Philadelphia), and VOSviewer (version 1.6.20; developed by van EckNJ at Leiden University, Leiden, the Netherlands). Microsoft Excel 2016 was used to quantitatively analyze publications (annual and cumulative numbers). VOSviewer was used to generate a visualization network map of countries,^[14] where the colors of the nodes and connecting lines represent different clusters.^[15] CiteSpace was used to perform burst analysis on keywords and to construct a timeline view of references.^[16]

The H-index and average per item were used in country, institution, and author analyses. Jorge Hirsch described the H-index in 2005 and was introduced as a scientific contribution metric corresponding to the number (H) of papers published that have been cited.^[17] The average per item is calculated by dividing the total number of citations by the number of publications, resulting in an average citation number of per publication.^[18]

3. Results

3.1. Growing publication trends

The final analysis included 25,487 publications (21,517 articles and 3970 review articles). Figure 2 illustrates the annual publication distribution on PE. The model fitting curve revealed that the annual publication quality in the previous 2 decades was consistent with linear growth (y = 65.957x + 581.8, R² = 0.8879).

Figure 2.

Annual publications and annual cumulative number of publications related to pulmonary embolism.

3.2. Analysis of countries/regions and institutions

Table 1 ranks the top 10 countries/regions in terms of publication number. Most of the articles on PE were published by researchers from the USA (9568 papers), followed by those in China (2436 papers) and Italy (1972 papers). These 3 countries accounted for 37.54%, 9.56%, and 7.74% of the total number of articles, respectively. However, the top 3 countries in terms of the average citation frequency were the Netherlands (75.28), Canada (66.89), and France (64.02). Canada (158) and the Germany (147) followed the USA, with the highest H-index of 232.

Table 1.

Top 10 productive countries related to pulmonary embolism.

Rank	Country	Count	Percentage (%)	H-index	Times cited	Average per item
1	USA	9568	37.54	232	381,593	39.88
2	China	2436	9.56	70	33,827	13.89
3	Italy	1972	7.74	145	103,202	52.33
4	United Kingdom	1821	7.14	140	97,918	53.77
5	Canada	1811	7.10	158	121,137	66.89
6	Germany	1721	6.75	147	104,701	60.84
7	France	1466	5.75	139	93,846	64.02
8	Japan	1278	5.01	66	26,016	20.36
9	Netherland	1199	4.70	139	90,264	75.28
10	Spain	1000	3.92	101	53,018	53.02

A collaborative analysis involving 91 countries publishing at least 5 articles among 142 countries/regions generated a clustering network map obtained through co-authorship analysis (Fig. 3).

Figure 3.

Collaboration network map of countries in the filed of pulmonary embolism.

Table 2 lists the top 10 institutions in terms of the publications number. Six of the top 10 research institutes publishing in PE research are from the USA, 3 from France, and 1 from Canada. Harvard University published the most research papers (1286), and obtained the highest H-index (121). Regarding the average citation frequency, McMaster University in Canada exhibited the highest frequency (92.08).

Table 2.

Top 10 productive institutions related to pulmonary embolism.

Rank	Institution	Country	Count	Percentage (%)	H-index	Times cited	Average per item
1	Harvard University	USA	1286	5.05	123	76,771	59.7
2	Harvard Medical University	USA	815	3.20	94	50,047	61.41
3	University of California System	USA	772	3.03	102	49,802	64.51
4	Mcmaster University	Canada	615	2.41	112	56,629	92.08
5	Assistance Publique Hospitaux Paris Aphp	France	610	2.39	103	50,930	83.49
6	Brigham Women Hospital	USA	575	2.26	91	42,123	73.26
7	Universite Paris Cite	USA	574	2.25	99	48,047	83.71
8	University of Texas System	France	509	2.00	71	27,550	54.13
9	Institut National De La Sante Et De La Recherche Medicale (Inserm)	France	504	1.98	86	36,581	72.58
10	Massachusetts General Hospital	USA	498	1.95	74	24,650	49.5

3.3. Analysis of authors

Authorship analysis gives significant information about the leading scientists and important research topics in a specific domain.^[11] Table 3 presents the top 10 authors in terms of publication number. Monreal M (Spain) was the most productive author (249 articles), followed by Huisman MV (Netherlands) (186 papers) and Konstantinides S (Germany) (184 papers). Goldhaber SZ (USA) had the highest H-index (113) and the highest average citation frequency (65.38).

Table 3.

Top 10 authors in number of publications related to PE.

Rank	Author	Count	Country	H-index	Average per item
1	Monreal	249	Spain	59	34.57
2	Huisman	186	Netherlands	71	63.67
3	Konstantinides	184	Germany	73	48.54
4	Goldhaber	180	USA	113	65.38
5	Jimenez	168	Spain	54	52.84
6	Le Gal	164	Canada	70	35.54
7	Klok	163	Netherlands	55	38.77
8	Ageno	161	Italy	75	35.68
9	Righini	142	Switherlands	54	44
10	Agnelli	139	Italy	90	55.67
10	Kline	139	Italy	64	43.39

3.4. Analysis of journals

Pulmonary embolism-related papers were published in 2362 kinds of journals. Table 4 presents the top 10 most prolific journals. Thrombosis Research published the most papers (836 papers). The Journal Chest achieved the highest impact factor (9.4998).

Table 4.

Top 10 journals in number of publications and co-citations related to pulmonary embolism.

Rank	Journal	Publications	Impact factor (2023)
1	Thrombosis Research	836	3.7001
2	Journal of Thrombosis and Haemostasis	592	5.5002
3	Thrombosis and Haemostasis	479	5.0002
4	Journal of Thrombosis and Thrombolysis	410	2.3000
5	Clinical and Applied Thrombosis-Hemostasis	342	2.3000
6	Medicine	269	1.3000
7	Chest	245	9.4998
8	American Journal of Emergency Medicine	233	2.7001
9	PLoS One	229	2.8997
10	Journal of Arthroplasty	204	3.4000

3.5. Analysis of keywords and references

Table 5 presents the top 20 keywords and the number of their occurrences. Apart from “PE,” the top 5 most frequent keywords were “deep vein thrombosis (DVT),” “venous thromboembolism (VTE),” “risk factors,” “thromboembolism,” and “management.”

Table 5.

Top 20 high-frequency keywords related to PE.

Rank	Keywords	Occurrences	Rank	Keywords	Occurrences
1	Pulmonary embolism	12,238	11	Computed tomography	1884
2	Deep venous thrombosis	6742	12	Complications	1658
3	Venous thromboembolism	6558	13	Molecular-weight heparin	1600
4	Risk factors	5692	14	Anticoagulant therapy	1566
5	Thromboembolism	4768	15	Outcomes	1444
6	Management	3501	16	Venous thrombosis	1431
7	Thromboembolism prophylaxis	3093	17	d-dimer	1422
8	Diagnosis	2847	18	Disease	1378
9	Prevention	2542	19	Therapy	1342
10	Mortality	2053	20	Guidelines	1214

PE = pulmonary embolism.

A keyword burst refers to many commonly mentioned keywords within a designated timeframe.^[19] The keyword burst reflects the research trend at a specific moment and may anticipate future research direction based on the hotspot duration.^[20] Figure 4 depicts the top 25 keywords with the strongest citation bursts. The keywords “helical computed tomography (CT)” and “spiral CT” had the highest burst strength, followed by “unfractionated heparin,” “society” and “molecular weight heparin.” The keywords “direct oral anticoagulants,” “VTE,” “american society,” “coronavirus disease 2019,” “infection,” “case report,” “pneumonia” and “clinical characteristics” appeared frequently in recent years.

Figure 4.

Top 25 burst keywords related to PE. *The length of the red line represents the burst duration. PE = pulmonary embolism.

The timeline view of references in Figure 5, depicts how themes were distributed and evolved over time. Cluster modularity (Q value) = 0.865 > 0.3 indicates a significant cluster structure, while cluster mean silhouette (S value) = 0.9421 > 0.7 indicates convincing clustering.^[11] There were 9 clusters including “#1 submassive PE,” “#2 suspected PE,” “#3 venous thromboembolism,” “#4 direct orai anticoagulant,” “#5 acute PE,” “#6 cancer-associated thrombosis,” “#7 medical patient,” “#8 PE,” and “#9 knee arthroplasty.”

Figure 5.

Timeline view of references related to pulmonary embolism.

The keyword burst analysis and timeline view divided PE research into 2 phases. Phase I (2004–2014) is the period with “spiral CT,” “medical patients,” “suspected PE,” “VTE,” “unfractionated heparin,” “fatal PE,” “Factor V Leiden,” “low-molecular-weight heparin,” “emergency,” “initial treatment,” “outpatient,” “brain natriuretic peptide (BNP),” “knee arthroplasty” as main keywords. Phase II (2015–2023) is the period with “antithrombotic therapy,” “submassive PE,” “coronavirus disease 2019,” “direct oral anticoagulants (DOACs),” “American society,” “infection,” “cancer-associated thrombosis,” “pneumonia,” and “case report” as main keywords.

4. Discussion

This study presents a thorough overview of the current research status and transition of research hotspots in PE through bibliometric analysis. Annual publications on PE research shown an increasing trend over the past 20 years, indicating that researchers are focusing increasing attention on PE.

The USA, France, and Switzerland have significant contributions to PE research. The USA has contributed substantially to PE research with 6 of the top 10 most productive research institutions from the USA. Goldhaber SZ from the USA is the most influential author, with the highest H-index and the fourth-largest number of publications. France has 3 top 10 institutions, all ranked at the top of the citation times. Monreal M from Spain, Huisman MV from Netherlands, and Konstantinides S from Germany are the top 3 authors with the most publications. The journal analysis demonstrated that Thrombosis Research, Journal of Thrombosis and Haemostasis, and Thrombosis and Haemostasis are very concerned with PE research and deserves to be focused on by academics interested in this field.

The top 3 most frequent keywords in addition to “PE” were “DVT,” “VTE,” and “risk factors.” VTE comprises PE and DVT development, with different clinical presentations of the same condition occurring at different stages and locations.^[21] DVT refers to abnormal blood coagulation in the deep veins, which blocks the venous lumen and leads to abnormal venous reflux.^[22] A blood clot that dislodges from the vein wall, travels through the heart to the lungs and blocks the pulmonary artery, impairing oxygen and carbon dioxide exchange in the lungs, a condition known as PE.^[23]

The analysis of keywords burst and timeline view of references may be useful in grasping research focus in different time periods. The keyword burst analysis and the timeline view of references in this study divided PE research into 2 phases. Phase I (2004–2014) is the first period focusing on the risk factors and prevention, diagnosis, and initial treatment of PE. Phase II (2015–2024) is the second period focusing on treatment strategies, guideline updates and specific PE changes, including submassive PE, cancer-related PE and infectious-related PE.

4.1. Risk factors and prevention

Risk factors for PE can be broadly categorized into genetic and acquired. The former include deficiencies of anticoagulant proteins S and C, antithrombin, and numerous genetic variants, the most common of which are the Factor V Leiden and prothrombin 2010A mutations.^[24] Acquired factors include lower-limb orthopedic surgery such as knee/hip replacement, malignancy, pregnancy, and antiphospholipid syndrome. Low-molecular-weight heparin is widely used to prevent or treat DVT and PE.

4.2. Diagnosis

The first step in the diagnostic strategy is to estimate the PE clinical probability as low, moderate, or high. PE clinical probability can be identified using 2 validated structured scores: the Wells score and the revised Geneva score.^[25] Combined with the PE low initial probability, a regular D-dimer test can effectively exclude PE presence because of its remarkable sensitivity and negative predictive value that can reach up to 100%.^[3] Above-threshold D-dimer levels indicate the potential presence of PE, and further investigations in conjunction with CT pulmonary angiography (CTPA) or echocardiography are usually required.^[26] CT exhibits a remarkable level of sensitivity and specificity, and can visually demonstrate vessel morphology and the vessel blockage location and degree.^[27] By employing CT-adapted angiographic scores (Miller and Walsh scores) or specialized CT scores (Qanadli and Mastora scores), PE magnitude can be quantified using CTPA.^[28] CTPA has replaced ventilation-perfusion scintigraphy and catheter pulmonary angiography as the primary imaging technique and the gold standard for diagnosing PE.^[29] BNP level is a sensitive indicator of impaired cardiac function. Some studies have reported that higher N-terminal pro-BNP levels are associated with more severe PE, poorer outcomes, and mortality.^[30]

4.3. Treatment strategies and guidelines

Clinical therapy for PE has been a research focus. Anticoagulation is the cornerstone of PE treatment to prevent thrombus extension and recurrence and to allow natural fibrinolysis. It is the only treatment required for most patients with PE.^[10] Thrombolytic therapy is the initial treatment of choice for hemodynamically unstable patients with acute PE.^[31] Treatment strategies for PE include catheter-directed thrombolysis, surgical embolectomy, and vena cava filters.^[32]

The need to reduce the occurrence of bleeding events has motivated the advancement of catheter-based approaches to pulmonary reperfusion in PE. These strategies use lower-dose fibrinolytic regimens to expedite embolus removal.^[33] DOACs consist of dabigatran etexilate, a direct thrombin inhibitor, as well as apixaban, edoxaban, and rivaroxaban, which are 3-factor Xa inhibitors.^[34] Compared with vitamin K antagonists, DOACs have more predictable pharmacokinetics, fewer food and drug interactions, and do not require regular monitoring.^[35] In the last decade, DOACs have played a significant role in the paradigm change in thrombosis treatment. In numerous countries, DOACs are the most often prescribed anticoagulants, replacing vitamin K antagonists.^[36] Recently, the latest guidelines for PE diagnosis and management released by the European Society of Cardiology in partnership with the European Respiratory Society, have gained much attention.^[32]

4.4. Pulmonary embolism with special conditions

Pulmonary embolism management in patients with special conditions such as the coexistence of pregnancy, infections or tumors, would be complex. Over the past few years, much direction has been directed at these special PE.

There are numerous case reports of PE in patients with coronavirus disease 2019.^[37–39] A systematic review of 36 research papers comprising over 11,000 patients with COVID-19 revealed a pooled occurrence rate of 17.0% for VTE and 7.1% for PE.^[40] COVID-19 may lead to thrombosis through 2 separate yet interconnected mechanisms: a hypercoagulable condition that causes large-vessel thrombosis and thromboembolism and direct damage to blood vessels and endothelial cells resulting in microvascular thrombosis in situ.^[41]

According to the guidelines, acute PE can be categorized into 3 groups: massive/high-risk, submassive/intermediate-risk, and low-risk PE, based on hemodynamic status and the existence of right ventricular enlargement or strain 2 to 4.^[42] A significant subgroup of individuals with hemodynamic stability but signs of right ventricular dysfunction or strain called submassive PE.^[43] There remains a controversy regarding the proper use of thrombolytic treatment for this subset.^[44] Systemic thrombolytic therapy reduces mortality in patients with massive PE. Nevertheless, using this treatment in submassive PE has not proven a distinct advantage, as the positive effects on hemodynamics are counteracted by an increased risk of bleeding.^[45]

Pulmonary embolism is a common complication in cancer patients and the second leading cause of mortality, after the tumor.^[46] The direct vascular toxicity of anticancer therapy and endothelial cell destruction by platinum-based chemotherapy may explain the higher occurrence of VTE in patients receiving these medications, complicating anticancer therapy delivery.^[47] Clinicians face difficulties in managing cancer-associated thrombosis due to drug interactions, liver dysfunction, and elevated risk of bleeding associated with certain cancer types.^[48]

4.5. Limitations

First, the accuracy of the results may have been impacted, as this study only gathered documents from the WoSCC database and focused on reviews and original articles written in English. Second, bibliometric analysis findings may vary from the actual research activity, potentially leading to the omission of recently published, valuable literature with limited citations. Furthermore, the use of identical abbreviations by some authors and the possibility of multiple expressions of certain keywords may have compromised accuracy despite efforts to correct them.

5. Conclusion

Over the past 2 decades, research on PE has focused first on evaluating the associated hereditary and acquired risk factors and their diagnosis. Later on, it was mainly focused on the shift in treatment modalities, updating of relevant guidelines, and treating specific types of PE (cancer-associated, submassive, and infectious PE). The use and feedback of PE-related guidelines, better improvement of prognosis, and life quality for PE-specific types might be potential future hotspots.

Author contributions

Data curation: Kaiyao Huang, Jianghua Zhu, Ke Sun, Heng Wang, Yun Peng.

Writing – original draft: Kaiyao Huang.

Writing – review & editing: Bibo Peng.