Background:
In Japan, approximately 30% of newly diagnosed HIV-infected individuals were identified after the onset of AIDS symptoms, suggesting that many people living with HIV (PLHIV) remain undiagnosed. Understanding factors that are associated with late HIV diagnosis (LHD), which is a major concern worldwide, is essential in achieving the 95-95-95 goal in Japan. Recently, we found that LHD in Japan was associated with a transmission clustering pattern. We present the possibility of identifying LHD-involving clusters in transmission-network analysis using our large-scale monitoring of domestic transmission clusters (dTCs) in Japan.
Methods:
We monitored the dynamics of HIV-1 dTCs in Japan using cases collected by the Japanese HIV Drug Resistance Surveillance Network. dTCs dynamics and the network structure of newly diagnosed cases in Japan were monitored by our search program for HIV nationwide clusters by sequence (SPHNCS), which identifies the transmission link with the genetic distance estimated by the protease-reverse transcriptase sequences. We recruited 9,722 newly diagnosed cases between 2003 and 2021 from our surveillance network, and identified their dTC affiliation, network structure, and chronological tree. The relationship of the transmission network patterns to LHD and outbreak involvements, which were recognized by the CD4 count and chronological tree, were investigated.
Results:
At the end of 2021, 566 subtype B and 105 CRF01_AE dTCs were registered in SPHNCS. Of these, seven and 30 clusters were recognized as LHD and outbreak involved clusters, respectively. At least 5,594 (57.5%) of newly diagnosed PLHIV in Japan were cases of LHD. Individuals in dTCs were more likely to be LHD cases than those not in dTCs (aOR, 0.80; p< 0.01). While mature transmission networks usually follow the power law with the lower degree being more frequent, outbreak clusters were characterized by the opposite degree distribution of the graph. In the degree distribution of the graph containing LHD, we observed a spike indicating a node at the center of the star-like structure in the network.
Conclusions:
Our results indicated that the graph structure of HIV transmission networks may imply involvements in late HIV diagnosis and outbreaks. This suggests the possibility that monitoring dTCs dynamics using network analysis can quickly identify a local population where an outbreak occurred or testing was delayed, to promote preventive measures.