Introduction Late diagnosis of the human immunodeficiency virus (HIV) is a major concern epidemiologically, socially and for national healthcare systems. Although the association of certain demographics with late HIV diagnosis has been reported in several studies, the association of other factors, including clinical and phylogenetic factors, remains unclear. In the present study, we conducted a nationwide analysis to explore the association of demographics, clinical factors, HIV‐1 subtypes/circulating recombinant form (CRFs) and genetic clustering with late HIV diagnosis in Japan, where new infections mainly occur among young men who have sex with men (MSM) in urban areas. Methods Anonymized data on demographics, clinical factors and HIV genetic sequences from 39.8% of people newly diagnosed with HIV in Japan were collected by the Japanese Drug Resistance HIV‐1 Surveillance Network from 2003 to 2019. Factors associated with late HIV diagnosis (defined as HIV diagnosis with a CD4 count <350 cells/μl) were identified using logistic regression. Clusters were identified by HIV‐TRACE with a genetic distance threshold of 1.5%. Results Of the 9422 people newly diagnosed with HIV enrolled in the surveillance network between 2003 and 2019, 7752 individuals with available CD4 count at diagnosis were included. Late HIV diagnosis was observed in 5522 (71.2%) participants. The overall median CD4 count at diagnosis was 221 (IQR: 62–373) cells/μl. Variables independently associated with late HIV diagnosis included age (adjusted odds ratio [aOR] 2.21, 95% CI 1.88–2.59, ≥45 vs. ≤29 years), heterosexual transmission (aOR 1.34, 95% CI 1.11–1.62, vs. MSM), living outside of Tokyo (aOR 1.18, 95% CI 1.05–1.32), hepatitis C virus (HCV) co‐infection (aOR 1.42, 95% CI 1.01–1.98) and not belonging to a cluster (aOR 1.30, 95% CI 1.12–1.51). CRF07_BC (aOR 0.34, 95% CI 0.18–0.65, vs. subtype B) was negatively associated with late HIV diagnosis. Conclusions In addition to demographic factors, HCV co‐infection, HIV‐1 subtypes/CRFs and not belonging to a cluster were independently associated with late HIV diagnosis in Japan. These results imply the need for public health programmes aimed at the general population, including but not limited to key populations, to encourage HIV testing.

Background: HIV-1 circulating recombinant form (CRF) 01_AE is the second major subtype in Japan. Our previous study indicated that CRF01_AE was predominantly circulating in heterosexuals/injecting drug users (IDUs). With implications of increased CRF01_AE infections among men who have sex with men (MSM), this study sought to investigate whether the transmission dynamics of CRF01_AE infections in Japan have changed. Methods: Sequences from 8032 newly diagnosed HIV-1-infected individuals were analysed. For 614 (7.6%) of CRF01_AE cases, clusters were identified and categorised by transmission risks. Median times to the most recent common ancestors (tMRCA) were estimated. Results: The individuals were predominantly Japanese (64%) and male (72%). MSM became the predominant transmission risk from 2014. Thirty transmission clusters (TCs) and 48 pairs, including 40% of individuals, were identified. MSM were approximately five times more likely to be in a TC compared to heterosexuals, and were the major contributors to TCs. tMRCA data suggest that MSM TCs emerged from 1996 and became predominant around 2000. Conclusions: CRF01_AE has spread among MSM, with frequent and continuous cluster formations, and MSM has become the predominant transmission risk. Our study suggested that CRF01_AE transmission has shifted from heterosexuals/IDUs to MSM. Prevention measures targeting key populations should be considered for controlling CRF01_AE spread.