Statistical and plotting process

The included literatures were systematically analyzed by Excel 2016 and CiteSpace V (Drexel University, Philadelphia, USA). We used Excel 2016 and GraphPad Prism 8 (University of California San Diego, San Diego, USA) to display the trend in the number of articles published by year and the distribution of articles by journal. ArcGIS was used to draw the heatmap of the top 10 countries’ total number of published articles between 2004 and 2020 and their average impact factor (IF). CiteSpace V was used to perform a co-occurrence analysis and visualize the collaboration networks of the authors/institutes/countries/keywords. Author co-citation analysis (ACA) and reference co-citation analysis (RCA) were also performed using CiteSpace V, and a related knowledge framework was constructed. Burst keyword detection was also performed to investigate recurrent new keywords. The 50 most cited or found articles were selected to create the individual network in a 1-year interval. Moreover, log-likelihood ratio weighting was used to analyze the contents of each cluster (20,21).