Beyond CD4 counts: Multi-biomarker assessment of estimated time since HIV-1 infection and risk factors for late diagnosis.

Abstract

Late diagnosis (LD) of HIV, defined as CD4 count <350 cells/μL or AIDS at diagnosis, remains a substantial public health challenge, with half of individuals in Europe diagnosed late. How long these individuals have lived with HIV remains unclear. We applied a multi-biomarker model (MBM) to estimate the time since infection (ETI) in newly diagnosed individuals classified as LD. We conducted a prospective cohort study of newly diagnosed people with HIV in Sweden between November 2017 and December 2022. Individuals were classified as LD or non-LD according to an updated consensus definition. We applied an MBM combining CD4 count, pol polymorphism counts, pol next-generation sequencing diversity, HIV-1 limiting-antigen avidity assay and last negative HIV test to probabilistically derive ETI distributions. More than half (52.4%) were classified as LD. Risk factors for LD included age >50 years (adjusted odds ratio (aOR) 2.94, 95% CI 1.42-6.09), origin in Asia and Pacific (aOR 4.53, 95% CI 1.81-11.32) and HIV acquisition in Asia and Pacific (aOR 2.28, 95% CI 1.08-4.83). Men who have sex with men had reduced risk of LD versus persons with heterosexual acquisition (aOR 0.38, 95% CI 0.22-0.64). ETI in 344 individuals revealed diagnostic delays; 41% (95% CrI 36.7-46.0) were diagnosed within 1 year: 71% of non-LDs (CrI 95% 65.0-77.7) versus 12% of LDs (95% CrI 8.4-16.2). Among LDs, over 60% remained undiagnosed beyond 3 years. Multi-biomarker ETI reveals that most HIV-1 infections remain undiagnosed for over 1 year. The updated LD definition correlates better with time since infection and ETI distributions show substantial diagnostic delays. Expanded testing strategies targeting heterosexual and older populations are urgently needed.