Demographic characteristics of the participants

From 1,004 HIV-1 positive plasma samples, 627 gag, 809 pol and 723 env sequences were successfully amplified and sequenced. Based on the genotypes of at least two genetic regions per sample, 833 samples were assigned with final genotyping results. The demographic characteristics of the 833 participants showed no statistical differences with those of the 171 failing to obtain genotypes (Supplementary Table 1).

Of the 833 individuals, 59.7% (497/833) were reported from four out of thirteen counties (Mengzi, Yuanyang, Kaiyuan and Gejiu). And 88.1% (734/833) were Honghe natives, 5.4% (45/833), 5.8% (48/833) and 0.7% (6/833) were from other cities in Yunnan Province, other provinces and foreign countries, respectively. The male to female ratio was 2.18:1. Age ranged from 16 to 89 years, with a median age of 47 years. The dominant ethnic groups were Han (40.5%, 337/833), Hani (23.2%, 193/833) and Yi (20.6%, 172/833). Unmarried, married and divorced/widowed persons accounted for 24.1% (201/833), 43.7% (364/833) and 32.2% (268/833), respectively. Those with a primary school education or less accounted for 66.0% (550/833). The main occupation was farmer (80.9%, 674/833). Of the participants, 93.3% (777/833) were infected through heterosexual contact.

Prevalence characteristics of HIV-1 genotypes

Of the samples genotyped, the main HIV-1 genotypes were CRF08_BC (56.7%, 472/833), URFs (17.8%, 148/833), CRF07_BC (13.9%, 116/833) and CRF01_AE (7.1%, 59/833). The remaining HIV-1 genotypes included CRF85_BC (2.2%, 18/833), CRF101_01B (1.1%, 9/833), Subtype C (0.7%, 6/833), CRF55_01B (0.5%, 4/833) and CRF64_BC (0.1%, 1/833). Among URFs, the most common recombinant form was BC recombinants (72.3%, 107/833); others included CRF01_AE/C (7.4%, 11/833), CRF07_BC/CRF08_BC (6.8%, 10/833), CRF01_AE/CRF07_BC (6.1%, 9/833), CRF01_AE/CRF08_BC (3.4%, 5/833), CRF01_AE/BC (2.0%, 3/833), CRF01_AE/B (0.7%,1/833), CRF101_01B/C (0.7%, 1/833), and CRF55_01B/C (0.7%, 1/833).

Spatial distribution characteristics of HIV-1 genotypes

Overall, the four main HIV-1 genotypes had a wide distribution, but primarily in eastern and central regions (Fig. 1). Specifically, CRF08_BC was concentrated in Mengzi City, Gejiu City and Yuanyang County, CRF07_BC was concentrated in Mengzi City, Kaiyuan City and Gejiu City, and CRF01_AE was concentrated in Mengzi City, while URFs were mostly found in Mengzi City, Kaiyuan City, Gejiu City and Yuanyang County. The other genotypes had a more restricted spatial distribution, with CRF85_BC mainly in Kaiyuan County and CRF101_01B mainly in Yuanyang County.

Spatial distribution of HIV-1 genotypes in Honghe Prefecture. The dot density maps for Subtype C, CRF01_AE, CRF07_BC, CRF08_BC, CRF55_01B, CRF64_BC, CRF85_BC, CRF101_01B and URFs. One dot represented 0.03% of the total cases subtyped. The areas outlined in red were the high-high aggregation areas analyzed by the local spatial autocorrelation analysis. Administrative boundary data on were downloaded from the National Catalogue Service for Geographic Information (https://www.webmap.cn). The map content approval number: Yun-S-(2024)14.

Global and local spatial autocorrelation analyses showed that CRF07_BC formed a high-high cluster area in Mengzi City and Kaiyuan City, URFs formed a high-high cluster area in Kaiyuan City.

HIV-1 molecular network analysis

The HIV-1 molecular network was constructed with pol sequences using a combination of phylogenetic tree analysis and genetic distance. Of 809 pol sequences, 344 segregated into 124 clusters. The number of sequences per cluster ranged from 2 to 19.

The multivariable logistic analysis showed that reporting area and education level were associated with the clustering rate (Table 1). Compared with the clustering rate in Kaiyuan County (42.5%, 45/106), which was equivalent to the general clustering rate in the whole prefecture (42.5%, 344/809), the clustering rate in Honghe County (19.0%, 4/21) was significantly lower (OR (95% CI): 0.271 (0.084–0.877)), the clustering rate in Luxi County was significantly higher (OR (95% CI): 2.279 (1.075–4.83)). And the clustering rates in the illiterate individuals (51.5%, 88/171) and those with primary school education (43.6%, 159/365) were significantly higher than that in those with a senior high school education or above (27.8%, 22/79).

Cross-regional HIV-1 transmission network

Among the 220 links found in the HIV-1 molecular network, 58.1% (128/220) of links connected nodes within the same county, while 41.8% (92/220) connected nodes from different counties (Fig. 2), which suggested the presence of cross-county transmission. With the number of links as the strength of connection between counties, Mengzi-Kaiyuan, Kaiyuan-Mile, Mengzi-Gejiu, Gejiu-Yuanyang formed strong connections between each other (Fig. 3).

Cross-country HIV-1 genetic transmission networks. (A) The HIV-1 molecular network coded by the reporting area of the individuals in the network. (B) The matrix of HIV-1 genetic links between counties. The cells on the diagonal represented the number of links in the same region. The other cells represented the number of links between different regions.

The visualization of cross-county transmission on the map of Honghe prefecture. The size of the dot represented the degree of cross-county transmission for the corresponding county, indicating its centrality. The thickness of the line represented the number of links between counties and indicated the strength of the link. The different colored areas represented the cohesive subgroups obtained using the faction analysis. The map content approval number: Yun-S-(2024)14.

To explore the internal associations and grouping of counties in the cross-regional transmission network, the cohesive subgroups were further analyzed. The result showed that there were three subgroups (Fig. 3). The subgroups I included Mengzi, Kaiyuan, Mile, Luxi, Gejiu, Yuanyang, Jinping and Lvchun. The subgroup II included Jianshui, Honghe and Shiping. The subgroup III included Hekou and Pingbian.

Among the 344 individuals in the network, 154 were involved in cross-county transmission. Multivariate logistic regression analysis suggested that age and ethnicity were associated with cross-county transmission (Table 2). Those aged 50–59 years (OR (95% CI): 2.103 (1.015–4.358)) and of ethnic minorities (OR (95% CI): 1.904 (1.026–3.532)) were more likely associated with cross-county transmission.

Prevalence of pretreatment HIV-1 drug resistance

Among the individuals with pol sequences, the overall prevalence of HIV-1 pre-treatment drug resistance was 6.2% (50/809). Further, the prevalence of PDR to NNRTIs, NRTIs and PIs were 5.2% (42/809), 0.9% (7/809) and 0.4% (3/809), respectively. The overall prevalence of PDR among all counties showed statistical difference (χ² = 23.789, P = 0.022), with those in Luxi (18.2%, 8/44), Hekou (14.3%, 3/21), and Yuanyang (10.9%, 13/119) ranking among the top three (Table 3). The prevalence of PDR to NNRTIs among all counties showed statistical difference (χ² = 22.920, P = 0.028), with those in Luxi (15.9%, 7/44) and Hekou (14.3%, 3/21) exceeding 10% (Table 3).