Over recent years, significant changes in human immunodeficiency virus (HIV) transmission modes have occurred in China, and the proportion of sexually transmitted infections has increased greatly (1). In addition, both the number of reported cases of people living with HIV/AIDS (PLWHA) and the distribution of diverse HIV transmission routes display regional characteristics (2). However, the nationwide distribution of heterosexual transmission subtypes remains unclear, including commercial heterosexual transmission, non-marital non-commercial heterosexual transmission, and marital heterosexual transmission (3).
Fully understanding the spatial characteristics of heterosexual transmission could establish a basis to formulate targeted regional prevention and control measures. For this purpose, this study used data from the Chinese HIV/AIDS Comprehensive Response Information Management System (CRIMS) to analyze the spatial distribution characteristics of the two main modes of heterosexual transmission, including transmission through commercial heterosexual contact (CHC) and non-marital non-commercial heterosexual contact (NMNCHC).
Newly identified cases of HIV infection were reported through the web-based CRIMS by individuals from local CDCs and medical institutions (4). Data usage from CRIMS was authorized by the National Center for AIDS/STD Control and Prevention of China CDC. All newly identified PLWHA in 2018 in the CRIMS were included in our study.
The frequency distribution and the constituent ratio of PLWHA infected through various transmission routes and the subgroups of those infected through heterosexual transmission were analyzed in each provincial-level administrative divisions (PLADs). Moreover, Moran’s I index was utilized to explore the clustering patterns of PLWHA identified in 2018 in China, with the assumption that cities throughout the country were not significantly different from each other with respect to the spatial distribution of infection. Subsequently, Z-scores with corresponding
P-values were presented. Statistically significant Moran’s I indexes revealed the presence of overall nationwide clustering. After a national level of clustering was confirmed, local clustering analysis was required to detect specific clustering areas. Hotspot analysis (Getis-Ord Gi*) in ArcGIS (version 13.0, Esri Inc, Redlands, CA, USA) was conducted to determine areas with significantly high clusters and areas with significantly low clusters, using 3 levels of confidence intervals (5).
In this study, PLADs with large numbers of newly identified PLWHA normally have high proportions of heterosexual transmission. In 2018, heterosexual transmission accounted for 71.5% of all transmission. Among the top 10 PLADs with the highest numbers of reported PLWHA, the proportions of heterosexual transmission were higher than average in 7 PLADs, including Guizhou (92.5%), Guangxi (91.2%), Yunnan (89.8%), Xinjiang (85.6%), Chongqing (83.2%), Sichuan (83.0%), and Hunan (78.6%), while those of Henan (67.3%), Guangdong (62.4%), and Zhejiang (58.9%) were below average. In addition to these top 10 PLADs, the proportions of heterosexual transmission in most other PLADs were below average, except for Xinjiang Production and Construction Corps (XPCC) (87.8%) and Jiangxi (84.4%), where the numbers of cases were relatively smaller than those of other regions (Figure 1A, Figure 1B).
Figure 1. Frequency distribution and proportion of newly identified PLWHA transmitted through different transmission routes and through different types of heterosexual transmission by PLADs. (A) Frequency distribution of newly-identified PLWHA transmitted through different transmission routes by PLADs; (B) Proportion of newly-identified PLWHA transmitted through different transmission routes by PLADs; (C) Frequency distribution of newly-identified PLWHA transmitted through different types of heterosexual transmission by PLADs; (D) Proportion of newly-identified PLWHA transmitted through different types of heterosexual transmission by PLADs.
Abbreviations: PLADs=provential-level administrative divisions; MHC=marital heterosexual contact; CHC=commercial heterosexual contact; NMNCHC=non-marital non-commercial heterosexual contact; PLWHA=people living with HIV/AIDS; XPCC=Xinjiang Production and Construction Corps.
In general, the proportion of newly identified PLWHA transmitted through NMNCHC (49.0%) was higher than through CHC (38.0%) in heterosexual transmission, and the proportion of NMNCHC-related transmission was higher than that of CHC in most PLADs and regions in 2018. Among the top 10 PLADs with the highest number of reported PLWHA, the ratio of NMNCHC to CHC transmission was greater than 1 in 7 PLADs, with Xinjiang having the highest ratio (5.94∶1), followed by other PLADs such as Guizhou (2.74∶1), Yunnan (2.19∶1), Henan (1.27∶1), Guangdong (1.18∶1), Sichuan (1.05∶1), and Hunan (1.03∶1). Those with a ratio less than 1 included Zhejiang (0.9∶1), Chongqing (0.87∶1), and Guangxi (0.56∶1) (Figure 1C, Figure 1D).
General spatial autocorrelation was conducted for newly identified PLWHA transmitted through CHC and NMNCHC in China in 2018. Moran’s I indexes were greater than 0 and the Z-values were greater than 1.96 for both transmission routes (P<0.0001). These results confirmed the potential existence of nationwide HIV/AIDS epidemic clustering for PLWHA transmitted through CHC and NMNCHC (Table 1).
Transmission routes Moran's I Z-value P-value CHC 0.2916 22.33 <0.0001 NMNCHC 0.1847 15.32 <0.0001 Abbreviations: CHC=commercial heterosexual contact; NMNCHC=non-marital non-commercial heterosexual contact; PLWHA=people living with HIV/AIDS.
Table 1. Results of general spatial autocorrelation for the newly identified PLWHA reported as transmitted through CHC and NMNCHC in 2018 in China.
In 2018, 373 cities reported PLWHA transmitted through heterosexual contact, among which cases transmitted through NMNCHC were reported in 371 cities, while PLWHA transmitted through CHC were reported in 360 cities. Furthermore, 13 cities had identified more than 500 cases of PLWHA reported as transmitted via CHC, located in Chongqing, Sichuan, Guangxi, and Yunnan. In contrast, 20 cities had more than 500 PLWHA reported as being transmitted through NMNCHC, including Chongqing, Sichuan, Yunnan, Xinjiang, Guizhou, Guangdong, and Beijing.
The Getis-Ord Gi* statistics revealed that at the 95% confidence level, hotspots of PLWHA infected through CHC were mainly discovered in Sichuan, Chongqing, Guizhou, Guangxi, northeast Yunnan, west Guangdong, west Hunan, southwest Hubei, southeast Gansu, and southwest Shaanxi, whereas cold-spots were observed in large areas of northeastern and northern China, comprised of more than 10 PLADs such as Shandong, Hebei, Tianjin, and Beijing. Hotspots for PLWHA infected through NMNCHC were detected mainly in Sichuan, Chongqing, Yunnan, Guizhou, and the western regions of Guangxi, while cold-spots were observed in eastern Hebei, Tianjin, Shandong, Anhui, northern Jiangsu, eastern Henan, and a small part of western Jilin.