Basic characteristics
A total of 3187 samples were collected. After excluding the individuals due to failure of HIV-1 RNA extraction and amplification (n = 214), sequencing failure ((n = 86), duplicated individuals (n = 69), and sequences failing quality control ((n = 43), 2775 sequences were included for subtype identification. As shown in additional file 1, 138 sequences were identified as CRF119_0107. The prevalence of CRF119_0107 from 2019 to 2024 was as follows: 3.7% (13/351) in 2019, 7.1% (22/308) in 2020, 7.8% (39/499) in 2021, 5.0% (29/581) in 2022, 3.6% (18/506) in 2023, and 3.2% (17/530) in 2024.
Among the 138 CRF119_0107 individuals, the mean age was 26.5 ± 8.0 years, ranging from 16 to 61 years. The majority were male (99.3%), unmarried (86.2%), college-educated (75.4%), and infected through homosexual transmission (84.8%). Initial CD4 counts were primarily 200–499 cells/µL (65.2%), and initial VL before ART predominantly ranged 10,000–99,999 copies/mL (49.3%) (Table 1). Individuals were distributed across all 12 districts of Nanjing, with the highest proportions in Gulou District (23, 16.7%) and Jiangbei New District (23, 16.7%), followed by Jiangning (18, 13.0%), Lishui (16, 11.6%), Qixia (13, 9.4%), Yuhuatai (11, 8.0%), Xuanwu (10, 7.2%), Qinhuai (10, 7.2%), Jianye (5, 3.6%), Luhe (4, 2.9%), Gaochun District (3, 2.2%) and Pukou District (2, 1.4%).
Drug resistance of CRF119_0107
The CRF119_0107 exhibited a TDR rate of 15.9% (95%CI: 9.8–22.1%), with 15.2% (21/138) resistance to NNRTI drugs, and 0.7% (1/138) to NRTI drugs. NNRTI-associated mutations K103N/KN (15.2%) and V179D (11.6%), as well as NRTI-associated mutations M41ML (0.7%) and S68G/SN (1.4%) were detected. No PI-associated mutations were detected. The V179D mutation conferred potential resistance to efavirenz (EFV), etravirine (ETR), nevirapine (NVP), and rilpivirine (RPV), while the K103N/KN mutation led to high-level resistance to EFV and NVP. The M41ML mutation induced low-level resistance to zidovudine (AZT) and stavudine (D4T), and potential resistance to didanosine (DDI). The S68G/SN mutation showed no resistance to any ART drug (Table 2).
Characteristics and dynamic change of CRF119_0107 clusters
We performed a sensitivity analysis spanning a spectrum of GD thresholds ranging from 0.005 to 0.015 substitutions/site (Additional file 2). At a GD threshold of 0.005 substitutions/site, the total number of clusters reached a peak and the network had the strongest resolution ability. A total of 78 sequences formed 11 distinct clusters, and the clustering rate was 56.6% (95%CI: 48.2–64.9%). The majority of clustered individuals were male (98.8%), aged < 25 years (56.5%), unmarried (84.7%), college-educated (73.1%), and infected through homosexual transmission (85.9%). Cluster sizes ranged from 2 to 21 nodes, with a median degree of 3 (IQR: 1–6). Analysis of transmission routes identified 67 nodes infected via homosexual transmission, 5 via commercial heterosexual transmission, and 6 via non-commercial heterosexual transmission, with corresponding median degree of 4 (IQR: 1–6), 2 (IQR: 1–2), and 3 (IQR: 1.25–4), respectively.
Regarding the drug resistance, the network contained 20 TDR individuals, including 19 TDR individuals in cluster 1 induced by K103N/KN and 1 TDR individual in cluster 7 induced by M41ML. Of particular concern, cluster 1 gained annual additions of 2, 6, 5, and 5 TDR individuals from 2021 to 2024, and the K103 mutation persistently spread in this cluster (Fig. 1). In the network, the proportion of high-degree individuals among TDR individuals was as high as 65.0% (13/20), while the proportion was 43.1% (25/58) among non-TDR individuals, without statistical significance (χ2 = 2.854, P = 0.091). The median of initial VL among TDR individuals was 153, 000 (IQR: 35775–428000), significantly higher than that (i.e. 49931, IQR: 35775–428000) among non-TDR individuals (Z=−1.996, P = 0.046).
Four large clusters (Clusters 1–4) with 21, 14, 11, and 10 nodes were identified, accounting for 70.6% (56/78) of all the clustering individuals. These clusters were exclusively male, dominated by homosexual transmission (85.8%, 48/56), with minor contributions from commercial (8.9%, 5/56) and non-commercial (5.4%, 3/56) heterosexual transmission. Geographically, cluster 1 spanned eight districts, with 57.2% (12/21) in Lishui District, and 61.5% (8/13) of the new clustering nodes were also distributed in Lishui District. Clusters 2, 3, and 4 spanned nine, seven, and six districts, respectively (Fig. 1).
Molecular network characteristics of newly reported HIV CRF119_0107 individuals in Nanjing. (A): Demographic characteristics; (B): Drug resistance; (C): Dynamic pattern. NRTI, nucleoside reverse transcriptase inhibitor; NNRTI, non-nucleoside reverse transcriptase inhibitor; Sensitive, sensitive to antiretroviral drugs
The analysis of dynamic change showed 48 nodes entering the network during 2019–2021 and 30 during 2022–2024. From 2022 to 2024, new clustering nodes across 11 clusters exhibited a median of 2 (IQR: 0.5–2). Clusters 1 and 3 were classified as high-growth cluster, gaining 13 and 7 new clustering nodes, respectively, and presenting over 2 nodes/year growth. Clusters 6, 7, and 8 were classified as stable clusters without new clustering nodes, whereas clusters 9, 10, and 11 were identified as emerging clusters (Fig. 1). The time intervals from diagnosis to initiating ART and VL at follow-up among individuals within high-growth clusters and stable clusters were further analyzed. In cluster 3, one individual did not initiate ART. Only 61.9% (13/21) and 50% (5/10) of treated individuals in clusters1 and 3 initiated ART within 30 days. All 7 individuals in three stable clusters initiated ART within 30 days, with 1 individual initiating ART within 7 days. Furthermore, after 3 months of ART, the VL of all 7 individuals remained undetectable during follow-up.
Factors influencing HIV-1 CRF119_0107 molecular transmission network
The χ² test revealed statistically significant differences in clustering rates based on initial CD4 counts and TDR status (P < 0.05). Multivariate analysis revealed that individuals with initial CD4 counts of 200–499 cells/µL (OR = 7.58, 95% CI: 1.95–29.45) and ≥ 500 cells/µL (OR = 26.50, 95% CI: 5.20–135.05) had a higher risk of clustering compared to those with CD4 counts < 200 cells/µL. Additionally, TDR individuals (aOR = 9.66, 95% CI: 1.32–70.56) were more likely to enter the network than non-TDR individuals(Table 1).
Spatial analysis of HIV-1 CRF119_0107 molecular network
Clustered individuals were distributed across all 12 districts of Nanjing, with more clustering individuals observed in Lishui (13), Gulou (12), Jiangbei New District (10), and Jiangning (10) and higher clustering rates observed in Lishui (81.3%), Luhe (75.0%) and Qinhuai (70.0%). The standardized clustering rate was calculated for spatial autocorrelation analysis based on the age composition of individuals in each district. Global Moran’s I value revealed no significant spatial autocorrelation (I = −0.121, P = 0.774), indicating a random distribution pattern at the district level (Fig. 2).
Spatial clustering and spatial transmission graph of the molecular transmission network of CRF119_0107. (A) Spatial clustering analysis, (B) Spatial transmission analysis between districts. Each line represents the HIV transmission links between two districts, with the line width representing the number of links between two districts
Spatial transmission patterns were further analyzed by visualizing The transmission network using intensity matrices and spatial transmission graphs. Inter-district Links accounted for 83.9% (260/310) of all the 310 Links, while intra-district Links constituted only 16.1% (50/310). Moreover, inter-district transmission predominated across all 12 districts, whereas Yuhuatai and Lishui District demonstrated mixed spatial connection pattern due to the exhibition of 38.5% and 34.0% intra-district transmission. (Fig. 3; Table 3). The strength of the connections between districts varied. Strong inter-district transmission linkage was also observed between geographically non-adjacent districts (between Qixia and Lishui) except for geographically adjacent districts (between Gulou and Qixia, Lishui and Jiangning) (Figs. 2 and 3).
Intensity matrices of transmission links observed in the HIV CRF119_0107 transmission network between 12 districts of Nanjing. The color of the grid cell at the intersection of two districts represents the number of linkages between the people living with HIV (PLWH) in two districts