The 100 most-cited articles in COVID-19: a bibliometric analysis

Abstract Corona virus disease 2019 (COVID-19) pandemic, sparked by the emergence of a novel coronavirus in early 2020, has prompted a surge in published articles. This study aims to systematically analyse the characteristics and trends of impactful research in the field. The 100 most-cited publications associated with COVID-19 were identified by two independent reviewers using the ‘Web of Science’ database across all available journals up to the year 2023. Data collected include country, citation count, subject, level of evidence (using Oxford Centre for Evidence-Based Medicine System 2011), impact factor, funding, and study design. We identified 394 038 publications, and the 100 most-cited publications were ranked. These were cited by a total of 283 034 articles (median citation = 767), median impact factor of 66.9 and 72 articles with fundings. China (n = 44), USA (n = 19), and UK (n = 13) were the three highest contributors (n = 220 505). Most articles were level 5 evidence (n = 48), followed by level 3 (n = 28), 4 (n = 14), 2 (n = 7), and 1 (n = 3). The main subjects were mechanism of action and structures of SARS-CoV-2 virus (n = 18) and impact of COVID-19 on public health (n = 18). Publications in 2022 and 2023 predominantly focused on the impact of COVID-19. Majority of the highly cited studies were of low-to-moderate quality, with only 10 consisting of randomized controlled trials or systematic reviews with or without meta-analysis. These findings reflect a growing interest in understanding the impact of COVID-19 pandemic on public and mental health. This analysis found the potential for future double-blinded randomized controlled trials to validate existing findings.


I
n early 2020, the novel coronavirus SARS-CoV-2, known as COVID-19, emerged in Wuhan, China, causing a global public health emergency [1].Efforts to contain the virus through awareness, lockdowns, and testing have been widespread.Despite these measures, COVID-19 has had significant negative socioeconomic and health impacts [2].The global scientific community has responded to this public health emergency by participating in extensive research and publications to elucidate the viral mechanism, prevention techniques, update management strategies, epidemiological study, diagnostic tools, and risk stratification [3].According to Organization for Economic Co-operation and Development (OECD), research funding for COVID-19 has reached 7 billion dollars to address the abovementioned issues [4].
The COVID-19 outbreak has led to an exponential increase in published articles.� Skori� c et al. [3] found 5761 COVID-19-related articles on PubMed as of April 2020.This surge in publications reflects the urgent need for timely data in managing the pandemic.However, the rapid increase may compromise the quality of information due to expedited reviews.Therefore, we aim to conduct a comprehensive bibliometric analysis of the top 100 most-cited articles on COVID-19 to evaluate their characteristics, quality, and trends.

Methods
We have undertaken a comprehensive literature search to identify the 100 most-cited articles on COVID-19.We searched the online database-Web of Science (Clarivate Analytics, Philadelphia, Pa)-using the terms 'COVID-19' OR 'coronavirus-19' OR 'SARS-CoV 2' OR 'coronavirus disease 2019' as a 'topic' on 18 February 202318 February , 1930 (GMT zone).The period of the search includes all the available years up until 2023.
Of the publications gathered from Web of Science, we ranked these in descending order of 'times cited'.Articles with an equal number of citations were separated by the average number of citations per year, with the more recent articles ranking higher.
Two reviewers (Y.L.L. and N.L.) independently reviewed these articles to obtain the 100 most-cited papers relevant to 'Covid-19'.Any discrepancies were resolved by the consensus discussion with another author (Q.M.D.).Further doubts were resolved by examining the article's full text.A total of 123 papers were examined to provide 100 articles included in this study.Exclusion criteria include articles that were not related to COVID-19 and if the article was not in full-text form.Figure 1a shows a flowchart summarizing our methodology.
Following the screening stage, the data were independently extracted from the full texts by three authors (Y.L.L., N.L., and Q.M.D.).The data extracted were categorized into article title, year of publication, names of authors, source journal, the total number of citations, average number of citations per year, level of evidence, study design, funding status, conflict of interest, and the main subject of the article.We apply the Oxford Centre for Evidence-Based Medicine System 2011 to classify our level of evidence ranging from evidence level 1 (systematic review and meta-analysis), level 2 (randomized controlled trial), level 3 (non-randomized controlled cohort, retrospective and prospective cohort study), level 4 (caseseries, case-control), and level 5 (laboratory studies, questionnaires, non-clinical studies, case report) [5].These data were documented onto a standardized computer spreadsheet (Microsoft Excel, Version 16.71).
Another comprehensive literature search was carried out in April 2024 to assess for stability of the 100 most-cited articles one year following the initial search.The methodology is the same as the previous search with changes to the timeline of interest to include articles from inception to 2024. Figure 1b shows the flowchart summarizing our methodology.

Initial search in 2023
Collectively, we gathered a total of 394 038 publications from the Web of Science in 2023.Of these, the top 100 most-cited papers associated with the theme of 'Covid-19' were ranked.(Please refer to Supplementary Appendix S1, which displays complete citations provided for all of the 100 most-cited articles associated with COVID-19.) The paper with the highest citation number was cited by 16 248 articles, while the lowest citation number was 1508 as shown in Table 1.Collectively, the 100 most-cited articles on COVID-19 were cited by a total of 283 034 articles.The study had shown that the most common clinical presentation is fever followed by cough with a mean incubation period of four days.The most common radiological finding on computed tomography of the chest is ground glass changes.On the other hand, the lowest cited article in our study belonged to a narrative study conducted by Singhal [7]  Understandably that most of the articles were published in 2020 at the initial phase of the pandemic measuring up to 91 articles.Seven articles were published in 2021, while two articles were noted to be published in 2022.The 100 most-cited articles vary widely in terms of their locations with China producing the most papers (n ¼ 44) followed by the USA (n ¼ 19) and UK (n ¼ 13).Three papers were from an international collaboration, namely, 'The species Severe acute respiratory syndrome-related coronavirus: classifying 2019-nCoV and naming it SARS-CoV-2' by Coronaviridae Study Group of the International Committee on Taxonomy of Viruses [8], 'Clinical management of severe acute respiratory infection (SARI) when COVID-19 disease is suspected.Interim guidance' by World Health Organization (WHO) [9], and the effectiveness of ADZ1222 vaccine by Voysey [10].The geographical spread of the 100 mostcited publications was very diversified as demonstrated by Fig. 2, which includes Germany (n ¼ 5), India (n ¼ 4), Canada (n ¼ 2), France (n ¼ 2), Netherlands (n ¼ 2), Hong Kong (n ¼ 1), Italy (n ¼ 1), Taiwan (n ¼ 1), Iraq (n ¼ 1), Singapore (n ¼ 1), and Australia (n ¼ 1).
The majority of these articles were of evidence level 5 (n ¼ 48) as per Oxford Centre of Evidence-Based Medicine (OCEBM) (Fig. 3a).This was followed by evidence level 3 (n ¼ 28) which was predominantly cohort studies (n ¼ 24) and one non-labelled non-randomized clinical trial, with retrospective cohort studies contributing 18 articles while prospective contributed nine articles.One of the level 3 papers was an open-labelled non-randomized trial evaluating hydroxychloroquine and azithromycin [11] as a treatment.Fourteen articles were noted to be of level 4 evidence, which includes a mix of case series (n ¼ 9), cross-sectional study (n ¼ 4), and case-control study (n ¼ 1).Level 2 and level 1 evidence-based        articles were seven and three papers, respectively.All of the level 2 papers were randomized controlled trials.Two of the level 1 evidence-based articles were systematic reviews and meta-analyses, while the remaining paper was a systematic review assessing the impact of COVID-19 on the general public's mental health.

most-cited articles in COVID-19
The main subjects of the articles were well spread across a few categories.Papers assessing the core structure of the SARS-CoV-2 virus and their mechanism of action (n ¼ 18), as well as papers assessing the effect of COVID-19 on public health and public health response (n ¼ 18) represent the largest number.This is followed by papers investigating the clinical presentation and outcomes, representing 14 publications.Other main subjects of note in descending order were biochemical and immunological responses to COVID-19 infection (n ¼ 10), mental health (n ¼ 9), risk factors associated with an increased mortality rate (n ¼ 9), management of COVID-19 infection (n ¼ 8), complications of COVID-19 infection (n ¼ 6), and radiological diagnosis of COVID-19 infection (n ¼ 3) (Fig. 3c).The common themes for complications of COVID-19 investigated were neurological events (n ¼ 1), cardiovascular events (n ¼ 2), and thrombotic events (n ¼ 3).
Majority of these papers were published in New England Journal of Medicine (n ¼ 13), followed by Nature (n ¼ 7), and the third most frequently noted journal is shared between Cell (n ¼ 6) and Lancet (n ¼ 6).The impact factors for each of these journals in 2022 were 176.1, 69.5, 66.9, and 202.7, respectively.The journal with the highest impact factor noted in our study was Lancet (202.7), and the journal with the lowest impact factor found was Pediatria I Medycyna Rodzinna -Paediatrics and Family Medicine (0.1).The published journals for each article with their corresponding impact factors (as of year 2022) can be found in Table 1.Seventy-two papers (72%) were found to have fundings from external sources.Of this, 44 (61.1%) disclosed no specific conflicting of interests.From the 28 papers that disclosed conflict of interests, 2 articles were documented to have received support from the funders outside of the published papers, while 1 article received cash and stocks compensation.

Comparison search in 2024
The following search carried out in April 2024 gathered a total of 528 745 articles from Web of Science associated with the theme 'Covid-19'.These were ranked from highest to lowest based on the total number of citations (Supplementary Appendix S2).Collectively, the 100 most-cited COVID-19 articles in 2024 were cited by a total of 316 817 articles.
When compared to the 100 most-cited articles searched in 2023, one article was retracted due to concerns from the publisher, and 11 articles were replaced with different articles of higher citations.Seven articles remained unchanged in terms of ranking, 35 articles had their ranking downgraded, while 47 articles had their ranking upgraded.The highest most-cited article was Zhou et al. with a total citation of 16 610 (previously ranked two in 2023).While the article with the lowest citation in our list belongs to Wynants et al., entitled 'Prediction models for diagnosis and prognosis of COVID-19 infection: systematic review and critical appraisal'.Figure 3b shows the study design for both 2023 and 2024 searches.
Of the 11 new articles added, 3 were of level 1 evidence, 1 was level 2 evidence, 1 was level 3 evidence, 1 was level 4 evidence, and 5 were level 5 evidence (Fig. 3a).Three of these papers discussed about the public health aspect of the pandemic, which includes surveillances, prediction model, and global pandemic policies.Two of these papers discussed the complications associated with COVID-19 infection focusing on post-COVID syndrome and multisystem inflammatory syndrome in children, whereas one of the articles was a systematic review discussing the impact of COVID-19 infection on mental health.Other themes of discussion were risk factors (n ¼ 1), mechanism of action and structure (n ¼ 1), biochemical and immunological response (n ¼ 1), clinical outcomes (n ¼ 1), and management (n ¼ 1).Four these new articles were published in 2021, whereas seven of these articles were published in 2020 (Fig. 3d).
Of 47 articles that had their rankings upgraded compared to the list in 2023, majority of these articles were related to mental health (n ¼ 8, 88.9% of the previous mental health articles in 2023) followed by COVID vaccination (n ¼ 4, 80% of the previous COVID vaccination articles in 2023) and public health responses (n ¼ 5, 62.5% of the previous public health responses articles in 2023).Supplementary Appendix S3 tabulates the 100 most-cited articles searched in 2024, giving an overview of their current ranking versus their previous ranking, with newly added articles marked as 'NEW'.

Discussion
This study aims to evaluate and summarize the COVID-19 research thus far, to provide insight into the evolution of research focus since the initial phase of the pandemic.This paper also aims to delineate the epidemiology of the 100 most-cited articles and to shed light upon less-covered themes.
Given the timeline of the COVID-19 pandemic, most of the research publications were groundbreaking to the scientific community in 2020, thus resulting in the rapid growth of COVID-19 related scientific studies during the initial phase of the pandemic.This would explain the abundance of publications in 2020 measuring up to 91 articles in our study.At the beginning of the pandemic, there remained an urgent demand to understand the viral structure and mechanism, pathology, clinical presentations, risk factors associated with poor prognosis, and diagnostic evaluation associated with COVID-19 infection, representing a total of 54 articles included in our study.Hence, most of these articles were published in 2020 (n ¼ 51).Only three papers were published in 2021 with one of the articles describing the 6-months consequences of patients discharged following a COVID-19 infection [12].This is in keeping with the progression of pandemic as long-term sequala of COVID-19 infection were starting to be recognized [13].
Summarizing the main subjects found, this study observed a comprehensive breadth of themes around COVID-19.From the 100 most-cited papers gathered, these studies covered management, impact on mental health, biochemical and radiological diagnosis, clinical presentations, clinical outcomes, complications, risk factors, origin, transmission route, core viral structures and genomes, the impact of COVID-19 pandemic on the environment, and COVID-19 vaccination.As mentioned above, there was a clear theme with papers published in 2020, but the same applies to papers published in 2021 and 2022.In 2021, more articles were observed to be investigating the different drug therapies available and also COVID vaccination.The remaining two papers in 2022 assessed the environmental impact of COVID-19 lockdown on the spatial distribution of aerosol [14], and fear of COVID-19 [15].This has shown that there is a move in terms of research demands to assess prevention, management of acute COVID-19 and long COVID-19 as well as the impact of COVID-19 lockdown on general public and the environment.
Mental health emerged as a significant subject within the literature, with nine articles in 2023 dedicated to this topic.Out of the nine articles discussing mental health, two papers were observed to focus their participants on public health workers, five papers focused on the general public, one paper focused on college students, and one paper provided a panel expert opinion on the research scope.In terms of health workers, the studies collectively found that female nurses and healthcare professionals were at high risk for anxiety, insomnia, and depression.The five articles on general public's mental health have confirmed the level of distress from COVID-19 pandemic and lockdown as well as risk factors and protective factors against mental health illness.Ahorsu [16] presents a development of a new scale which they termed 'Fear of Covid-19 Scale' (FC-19S).This was created following a rigorous literature review to create the first known psychometric assessment.This is the only patientreported outcome measurement (PROM) described in the current study, which can be utilized by any governing body when planning holistic goals for a COVID-19 fear-free public.
The global scale of COVID-19 pandemic has also shown how the scientific community were able to collaborate both nationally and internationally to bring about advances in medical knowledge and innovative treatments.Zyoud and Al-Jabi [15] demonstrated in their study that China, the USA, UK, and Italy were the leading contributors in early 2020.This was attributed to a high density of COVID-19 infection in the abovementioned countries.The findings were replicated in our study as well whereby most of the current articles were produced by China, the USA, UK, and Europe (in this study, represented by Germany, Italy, Netherlands, and France).
Of the 100 articles searched in 2023, 26 publications were multicentre collaborations.Three of the 26 articles were of international collaboration, respectively describing the discussion of SARS-CoV-2 virus naming [8], guidelines on management and triaging of COVID-19 infection [9], and the effectiveness of the AZD1222 vaccine [10].China produces the most national multi-centre collaborative publications totalling 10 papers, followed by UK (n ¼ 5) and USA (n ¼ 4).Other countries noted in our studies are France (n ¼ 1), Hong Kong (n ¼ 1), Italy (n ¼ 1), and Netherlands (n ¼ 1).The majority of these multi-centre articles investigate the clinical presentation and outcomes of COVID-19 infection (n ¼ 5), and risk factors associated with poor prognosis (n ¼ 5).This is followed by a discussion on management (n ¼ 4), COVID-19 vaccination (n ¼ 3), mental health (n ¼ 3), complications (n ¼ 2), biochemical and immunological responses (n ¼ 2), and public health (n ¼ 1).Concerning management as a main subject, all four articles were multi-centred randomized controlled trials investigating dexamethasone [17], remdesivir [18,19], and lopinavir-ritonavir [20].It is not unexpected that the large-scale RECOVERY randomized controlled trial investigating the effect of dexamethasone on mortality rate falls within the 10 most-cited articles.These large-scale randomized controlled trials are fundamental to a pandemic in exploring the necessary interventional measures.A meta-analysis carried out by Dechartres et al. [21] have demonstrated that results from smallerscale studies should be interpreted with caution as the impact of biases were higher.
This study has also shown that accelerated research was achievable given how multiple randomized controlled trials were approved and completed within a short period.Despite so, a systematic review carried out by Kudhail et al. [22] in 2022 demonstrated that these accelerated publications were shown to be at risk of bias.This, in particular, is due to the unblinded or single-blinded nature of some studies.Five out of seven randomized controlled trials included in our study are either observer blinded (n ¼ 2) or single-blinded (n ¼ 3), thus resulting in a higher risk of randomization sampling biases [23].
Subsequent search in 2024 showed more focus on the complications of COVID-19 infection as well as how the appropriate bodies can further reduce transmission risk through non-pharmacological interventions.This is notable based on the themes of the 11 newly added articles to the 100 most-cited list.Furthermore, it is reassuring to note that the impact of COVID-19 and lockdown on mental health is taken seriously by the scientific community.
The main limitations of this study include those that are inherent to most bibliometric analyses.Given that this study only utilizes 'Web of science' as the only source of articles, the breadth of publications included will be limited to the abovementioned database only.Second, bibliometric analysis is inherently a quantitative analysis that relies on the citation counts to decide the impact of an article.The underlying assumption drawn in such scenario would be the total number of citations of an article equal to the confidence level of other authors for this publication.This assumption ignores the possibility of authors citing the specific paper for their study design, or facts within their introduction [24].This study has the added benefit of looking at the changes to the 100 most-cited list by assessing the articles one year apart.Unfortunately, newer studies may be missed as they were not given sufficient time to gather citation counts.
Regardless of the abovementioned limitations, this study presented an extensive literature search into the 100 most-cited literature for COVID-19 while tracking for the stability of these articles one year down the line.The study analysed the current available articles and categorized them based on the time, level of evidence, study design, location of publication, and the study design.This study presented the temporal trend in emerging COVID-19 topics such as the impact of COVID-19 pandemic and lockdown on the environment, mental health, and long COVID-19 syndrome.All themes are the lingering impact of COVID-19.

Conclusion
The majority of articles included in this study fell within the range of level 3 to level 5 evidence, primarily consisting of experimental studies, literature reviews, case series, and cohort studies.Fortunately, the secondary searched in 2024 showed an increase in number of level 1 and 2 studies compared to 2023, which showed the move from a scientific community to collate stronger evidences.A few noteworthy high-level studies (level 1 and level 2) focused on new management strategies for COVID-19, the impact of the pandemic on the general public's mental health, and preventive measures to reduce transmission rates.This suggests a shift in research focus within the scientific community as we transition from the peak of the COVID-19 pandemic to a post-pandemic period, emphasizing the need to understand the long-term consequences of the pandemic and lockdown on the general population.Of note is the increasing focus on impact of COVID-19 and lockdown on mental health.It is worth noting that most of the randomized controlled trials in our study were either open-labelled or single-blinded, highlighting the potential for future studies to strive for double-blinded randomized controlled trials to validate the existing findings.
from India entitled 'A Review of Coronavirus Disease-2019 (COVID-19)'.The paper discussed the course of COVID-19 outbreak and the record of COVID-19 associated death as of 2020.Tanu also discussed the measures taken by different countries to reduce the spread of COVID-19 during the peak of the pandemic.

Figure 1 .
Figure 1.(a) Flowchart showing the methodology for article screening and data extraction for the 2023 search.(b) Flowchart showing the methodology for article screening and data extraction for the 2024 search 98 � 69 mm (300 � 300 DPI).

Figure 2 .
Figure 2. Choropleth map showing the countries contributing to the 100 most-cited articles (dark blue represents high-density of publications, while yellow represents low-density of publications).

Figure 3 .
Figure 3. (a) Number of articles based on the level of evidence as per Oxford Centre of Evidence-Based Medicine (OCEBM) classification for both 2023 and 2024 searches.(b) The study designs of the 100 most-cited articles for both 2023 and 2024 searches.(c) Main subjects of the 100 most-cited articles in 2023 (category based on years of publication).(d) Main subjects of the 100 most-cited articles in 2024 (category based on years of publication).
The average number of citations per article per year ranged from 5416 to 469.The highest cited article belonged to a retrospective cohort study conducted by Guan W-Jie et al. [6] from China entitled 'Clinical Characteristics of Coronavirus Disease 2019 in China'.The paper assessed the clinical presentation, incubation, and radiological diagnosis of COVID-19.

Table 1 .
List of 100 most-cited articles on COVID-19 for the