On September 06, 2022, we conducted an advanced search on the Web of Science Core Collection (WoSCC). Search formulas were “TS = (depress* OR depression OR depressed OR depressive disorder OR mood disorder)” AND “TS = (HIV OR HIV AND AIDS OR human-immunodeficiency-virus OR acquired immune deficiency syndrome),” to identify publications primarily focused on HIV in the field of depression. The date ranged from 1999/01/01 to 2022/09/06, language type was set as English, and the document type was limited to articles and reviews. First, two authors (Xiaoyu Du and Qian Zhang) independently searched the WoSCC and then screened the title, abstract, or full manuscript to identify whether these publications met this research topic. Second, the other two senior scholars (Jing Liu and Jia Chen) read and evaluated the discrepancies between the former authors, and discussed with them until a consensus was reached. All publications were exported as the format of “Full Record and Cited References,” and then used for bibliometric parameters extraction.

Firstly, we used VOSviewer (version 1.6.11, Leiden University, Leiden, Netherlands) to extract the following bibliometric parameters, mainly including the following elements, e.g., title, keywords, authors, institutions, countries/regions, published journals, publication year, total citations (TC), citations per publication (CPP) and cited references. Then we used Microsoft Excel 2010 (Redmond, Washington, United States) to identify the publication trend, the top contributors (e.g., prolific authors, institutions, and countries, journals). And we mapped the collaboration network between authors, institutions, and countries by VOSviewer, to demonstrate their scientific influence in this field. CiteSpace (Version 6.1. R1) were used to map the co-cited references network and identify references with highly citation burst. Co-occurrence of author keywords and keywords with citation burst were used to visualize the knowledge evolution, hot topics and potential research frontiers in this filed. In the map of VOSviewer and CiteSpace, the node size positively related to the number of publications, and the line width positively associated with their link strength.

Figure 1 shows the study design and analytic approach. After restrictions on the document types (original article and review) and the language (English), 8,190 publications were included in the final analysis. In total, 30,026 authors, 6,582 institutions, 140 countries/regions, and 1,308 journals participated in these publications.

Figure 2 displays the publication trend. From 1999 to 2021, the number of publications roughly presents a steadily increasing trend, and 388 publications as of the first 9 months in 2022. In addition, we presented the publication trend of the top three prolific countries.

Figure 3 displays the top 10 most prolific countries and the collaborative network between them. The United States ranked first with 5,622 publications (68.6%) and 191310 TC, followed by South Africa (771 publications, 18755 TC) and the United Kingdom (657 publications, 26550 TC; Figure 3A). As for CPP, Australia ranked first with 44.3 CPP, followed by the United Kingdom (40.4 CPP) and the United States (34.03 CPP). To present the international collaborative network, we utilized the module of co-authorship-country in VOSviewer. With at least 30 publications, a total of 42 prolific countries/regions formed a cooperative network. Of which, the United States, South Africa, the United Kingdom, Canada and China are large nodes with relatively thicker links. The United States had the highest degree of co-operation total link strength (TLS = 45,077) and co-operated with 39 prolific countries. Of which, South Africa (TLS = 15,833), and the United Kingdom (TLS = 13,858) had a closer academic cooperation with the United States.

Figure 4 displays the most prolific institutions which published at least 70 articles. The top 10 prolific institutions published 42.9% of publications. University Calif San Francisco (United States) ranks first with 508 publications, followed by University Calif Los Angeles (United States) with 408 publications, Johns Hopkins University (United States) with 380 publications. In terms of citations, University Calif San Francisco (United States) ranks first with 23709 TC and 46.7 CPP, followed by Harvard University (United States) with 22206 TC and 68.3 CPP, and University Calif Los Angeles (United States) with 18107 TC and 44.4 CPP. We used VOSviewer to visualize the institution collaborative network. As shown in Figure 4B, the top 60 prolific institutions were grouped into four clusters. The biggest cluster was institutions that primarily came from the United States (e.g., University Calif San Francisco, University Calif Los Angeles, Johns Hopkins University, University Alabama Birmingham). The red, blue, and yellow clusters contain mainly United States institutions, and the green cluster contains mainly African institutions. However, United States institutions and African institutions are mainly associated with Columbia University.

Figure 5 displays the top contributing authors (at least 30 papers) and collaborative map between them. Safren, Steven A. (University of Miami, United States) with 140 publications, Mayer, Kenneth H. (Fenway Health, United States) with 103 publications, and Pence, Brian W. (University of North Carolina, United States) with 82 publications were identified as the top three most prolific authors. In terms of citations, Safren, Steven A. also ranked first with 5251 TC and 37.5 CPP, followed by Mayer, Kenneth H. with 3784 TC and 36.7 CPP and Mimiaga, Matthew J. (University of California Los Angeles, United States) with 2953 TC and 36.9 CPP. We used VOSviewer to demonstrate the inter-authors cooperative network and their active time. Figure 5B displayed that a total of 53 authors have more than 30 publications in this filed. The largest cluster consists mainly of Safren, Steven A. Mimiaga, Matthew J., Pence, Brian W., and others, mainly from the United States. A cluster led by Peltzer, K, Jones, D L. Rodriguez, V J., who are primarily from South Africa and the United States, focuses on PLWH in South Africa. The cluster formed by Li xiaoming, Zhao guoxiang, Zhao junfeng, are all researchers from China. Additionally, we used a timeline view to present the active time of different authors in this field.

Table 1 presents the top 10 most prolific journals and co-cited journals. AIDS Care (n = 683 publications) was the top prolific journal, followed by AIDS Behavior (n = 615 publications) and JAIDS-J Acquired immune deficiency syndromes (n = 224 publications). However, in terms of CPP, AIDS (n = 153 publications) has recorded the highest CPP (n = 46.4 CPP), followed by JAIDS-J Acquired immune deficiency syndromes (n = 224 publications, 41.3 CPP) and AIDS Patient Care and STDS (n = 198 publications, 39.6 CPP). Co-cited journals were journals frequently cited in a series of selected publications. Of which, AIDS Behavior (n = 12,302 co-citations), AIDS Care (n = 11,426 co-citations), and AIDS (n = 9,775 co-citations) ranked the first three.

Citation is a quotation from one article to another article. The number of citations to the literature has been used as an indicator of the value, interest and utility of different articles in terms of the advancement of knowledge (González-Alcaide et al., 2016). Table 2 covered the top 20 most cited publications in this field. Among them, six publications are review, and the remaining publications are original articles. The most cited article (6677 TC) was published by Mathers, CD et al. in PLos Med., entitled “Projections of global mortality and burden of disease from 2002 to 2030” (Mathers and Loncar, 2006). In this article, the authors project disease-related mortality and major disease burden by 2030 based on model and baseline data. Specifically mentioned that HIV/AIDS become the leading disease burden by 2015 in middle- and low-income countries. Additionally, Seven of the top eight articles are studies on the global burden of disease, and all suggest that HIV/AIDS is a leading cause to the global burden of disease (Lopez et al., 2006; Mathers and Loncar, 2006; Murray et al., 2012; Vos et al., 2012; Global Burden of Disease Study, C, 2015; Disease et al., 2016; GBD 2019 Diseases and Injuries Collaborators, 2020).

Co-citation is defined as two publications being cited by a third at the same time (González-Alcaide et al., 2016). The analysis of co-citations is the identification and quantification of the frequency of co-occurring literature in the reference list in a research field. This helps us to identify the most influential literature in the field. The co-citation analysis was performed with CiteSpace and VOSviewer to examine the evolution of scientific paradigms in depression-related research on HIV/AIDS field. Figure 6A showed the publications have more than 20 citations. Large nodes indicate they received more citations, the red nodes represent an outstanding contribution to this field’s development. They connected important references in sub-domains of this field. For example, “Leserman J, 2008, Psychosom Med, V70, P539, (Leserman, 2008)” entitled “Role of depression, stress, and trauma in HIV disease progression.” This article bridges the knowledge between the cluster# 16 (hiv/aids) and cluster# 27 (fatigue). Moreover, CiteSpace also identified the influential co-cited references in different time. Articles published around 2000 were centered on topics about “symptoms,” “disease progression,” and “induction of drug-metabolism.” Articles published around 2010 were centered on topics about “adherence,” “fatigue,” “Africa,” and “transgender.” Finally, articles published around 2020 were centered on topics about “stigma,” “group support psychotherapy,” “cognition,” and “MSM.” Figure 6B lists these important articles and their active time. Furthermore, we identified several articles received lots of attentions from 2020 to 2022, including Cohen MS, 2016, New Engl J Med, V375, P830, (Cohen et al., 2016), which found early initiation of ART led to a sustained decrease in genetically linked HIV-1 infections in sexual partners; Rueda S, 2016, BMJ Open, V6, P0 (Bernard et al., 2017), which meta-analyzed the prevalence and factors of depression among PLWH in sub-Saharan Africa; Turan B, 2017, AMJ Public Health, V107, P863 (Turan et al., 2017), which developed a conceptual framework of HIV-related stigma to explore the four ways in which HIV-related stigma differently affects the health of patients.

VOSviewer and CiteSpace were used to conduct keyword co-occurrence analysis to present major hotspots and frontiers in this field. To count the number of keyword occurrences more accurately, we used a thesaurus (Supplementary Table S1) to merge keywords with similar meanings. For example, acquired immunodeficiency syndrome was replaced by aids, and human immunodeficiency virus was replaced by HIV. Finally, 83 keywords appeared more than 40 times were classified into five clusters (Figure 7A). The top three most occurred keywords are “HIV” (n = 2,703), “depression” (n = 1,605), and “mental health” (n = 641). Further, to decipher potential research frontiers in this field, we used the overlay map of keywords occurrence by VOSviewer and keyword citation burst by CiteSpace. Frequently used keywords can be detected during a particular period. Keyword citation burst is an increase in the frequency and number of keyword occurrences within a certain period of time (Li et al., 2020). Figure 7B displayed the potential hot keywords (yellow) identified by VOSviewer, including “medication adherence,” “suicide,” “discrimination,” “social support,” “self-efficient,” and so on.” Figure 7C identified the evolution of keywords from 1999 to 2022. The top 42 keywords with the strongest citation bursts are displayed, which show that “hiv/aid” is the keyword with the highest burst strength (n = 49.9). In the early stage, “stress,” “quality of life,” “cocaine” received lots of attention. And then keywords like “social support,” “adherence,” “risk behavior” got much attention. In recent years (2010-now), keywords have received more attention (e.g., “sexual risk behavior,” “medication adherence,” “suicide” and “condom use”), which suggests that these keywords are likely to become frontier hot topics in the future.

In this study, 8,190 publications were included in the final analysis. We observed from Figure 2 that the trend of publications was slowly increasing from 1999 to 2021, which suggests that there is an increasing interest among scholars in different disciplines to focus on depression in HIV/AIDS-related research. It is worth noting that during the COVID-19, the number of publications remains high, with 697 publications in 2020 and 692 publications in 2021. Although the topics are still dominated by Public Environmental Occupational Health, Infectious Diseases, and Psychiatry in the last 3 years of publications, a portion of the studies have focused on the impact of pandemic outbreaks on PLWH. Many scholars became concerned about the impact of SARS-CoV-2 and lockdown on mental health of PLWH during COVID-19. First, patients with a deficient immune system will be more vulnerable to being affected by SARS-CoV-2 (Zhao et al., 2020), which aggravates their anxiety and worries in patients with AIDS. Second, the lockdown and social isolation increase the economic burden for HIV patients and their family members (Jones et al., 2021). Third, the COVID-19 pandemic caused extensive clinic pauses and high-contagious SARS-COV-2 may lower the patients’ visit rate to the hospital (López and Solis-Soto, 2022). Jamile Ballivian conducted a cross-sectional study in Argentina and found decline in medication adherence, disruption to mental health services and substance abuse treatment. During lockdown (Ballivian et al., 2020). Prior study evidenced that depression among HIV patients is associated with lower adherence of ART and health outcomes (Mayston et al., 2012). Therefore, pandemic outbreak is a huge challenge for public health, especially for special vulnerable populations like PLWH. The mental health of PLWH deserves focused attention during COVID-19.

Figures 3A,B showed that the United States dominated the field of AIDS-related depression, with the highest number of publications (68.6% of the total) and at the core of cooperation between countries/regions. The United States is at the center of the map and is connected to many countries, demonstrating the depth and breadth of the collaboration between U.S. researchers and South Africa, the United Kingdom, Canada, and China. Previous bibliometric analysis also reported that the United States plays the leading role in other areas of HIV (Liang et al., 2021), such as HIV dementia (Hussain et al., 2022), AIDS-related stigma and discrimination (Sweileh, 2019), and gerontology research in the field of HIV (Hoang et al., 2022), etc. South Africa ranked second in the number of publications. This result corresponds to the regional distribution of the PLWH. In 2021, 20.6 million people were living with HIV in Eastern and Southern Africa, nearly half of the total global PLWH (World Health Organization, 2021). Interestingly, when it comes to CPP, Australia ranked first with 44.3 CPP. A potential explanation for its high citations is that many of its’ publications published in Lancet, a very authoritative journal. The top most cited articles, all collaborations between the University Queensland, University Washington, and Harvard University, were published in Lancet. They analyzed the global burden of disease and mortality, which suggested that HIV/AIDS and depression were the top 10 leading diseases for global disease burden in recent decades (Lopez et al., 2006; Murray et al., 2012; Vos et al., 2012; Disease et al., 2016).

As shown in Figure 4A, all the top 10 prolific institutions were from the United States, corroborating the dominance of the United States in the field of HIV-related depression. Figure 4B illustrates the collaboration between the different institutions. Geographical distribution can affect the scientific cooperation between authors, institutions, and even countries (Baghaei Lakeh and Ghaffarzadegan, 2017). Researchers tend to choose the closer one as their collaboraters because of similar culture background. The institutions from the United States collaborate relatively more intensive and extensive. The cooperation of these institutions has significantly contributed to the number and impact of U.S. publications. Figure 4B also showed a very close collaboration between South Africa and the United States, mainly between Columbia University and University Cape Town. They conducted a cross-sectional study from October 2004 to December 2005, include 465 participants, to analyze the high prevalence of depression, anxiety, and substance-related disorders, and validate some brief rating scales which could be used to identify psychopathology in HIV-positive individuals (Myer et al., 2008).

Interestingly, although Harvard Univ was not one of the top three productive institutions, it ranked first regarding average citations (CPP = 68.3). We reasoned that this may be attributed to the top prolific author, Safren, SA (University of Miami, United States). He has established a close collaboration with Mayer, KH. and Mimiaga, MJ, and Pence, BW, all of them ranked the top 10 prolific authors in Figure 5A. They focused on depression and other psychological problems in HIV patients and revealed a significant positive association between depression and ART nonadherence in PLWH (Gonzalez et al., 2011; Uthman et al., 2014). Besides, Safren, SA and Mayer, KH focused on the effect of cognitive-behavioral therapy to enhance medication adherence and reduce depression among PLWH (Safren et al., 2009). Besides research team from the United States, a research team from China, represented by Li, Xiaoming also attracted our attention. They were very active in recent years (Figure 5B, colored yellow). They focused on the depression and social support among children affected by parental HIV in China (Ezell et al., 2020; Wu et al., 2022; Jiang Y. et al., 2022; Jiang Y. P. et al., 2022), the relationship between depression, HIV stigma and sexual risk factors among men who have sex with men (MSM) in China (Sun et al., 2020; Mi et al., 2022). Since the outbreak of the COVID-19, they also paid attention to the stress and coping strategies among HIV healthcare providers (Mi et al., 2021).

Analysis of keywords can help to quickly understand the knowledge network of a field. As shown in Figures 7B,C, in the early stages, the main keywords include “cocaine,” and “quality of life.” Heaton et al. examined the relationship between neuropsychological deficits and quality of life in HIV-infected in 2004 (Heaton et al., 2004); Sadek et al. assessed the impact of neuropsychological deficits and depression on methamphetamine abuse and quality of life of PLWH in 2007 (Sadek et al., 2007); and Mindt et al. conducted a similar study in Spain in 2003 (Mindt et al., 2003). They provide clinical evidence for future research on HIV-associated neuropsychiatric deficits. In the last few years, “sexual minority,” “alcohol abuse,” and “suicidal ideation” have attracted the attention. PLWH are at high risk for suicidal ideation and behavior. In order to reduce the risk of suicide among PLWH, many scholars have conducted research on the protection and risk factors of suicidal ideation. Casale found that stigma is a highly significant risk factor among adolescents living with HIV, and social support can buffer the adverse effects of stigma (Casale et al., 2019). Wang et al. demonstrated, in the form of a structural equation model, that stigma is the most influential factor in the accumulation of suicidal ideation among AIDS patients in China, while self-esteem and depression are also important risk factors for suicidal ideation (Wang et al., 2018). They all agree that improving tangible social support for PLWH can mitigate the risk of suicidal ideation in future research.

In this study, we visualized the knowledge networks and research paradigms of depression-related research among HIV/AIDS. The results showed 83 keywords appeared more than 40 times in Figure 7A. “HIV” and “depression” are the most frequently occurring keywords. The six largest clusters mentioned below represent the whole study field.

This study has several limitations that should be acknowledged. First, all the literatures in this bibliometric analysis were collected from the WOSCC, which only represents one aspect of this topic. Restricted by data format for bibliometric tools, the publications in other databases (Pubmed, Scopus, Google Scholar etc.) were not included in this study. Second, the indicators for the quality of an article, an author, or an institution only included the number of publications, TC, and CPP. Other metrics are widely used in other studies, such as H-index, Impact Factor, CiteScore, etc. Third, due to the included criteria of language and publication types, only articles or reviews in English were included in this analysis. However,“English” remains the most common language used in published academic literatures worldwide, future research should integrate multiple non-English language (e.g., Chinese, German) publications to provide a more comprehensive analysis of this field.