Inferring transmission dynamics from HIV-1 genealogies

With a national prevalence of 4.9% in the adult population, the HIV-1 epidemic in Kenya is the fifth largest in the world. HIV-1 prevalence is more than three-fold higher among HIV key populations – including men who have sex with men (MSM), people who inject drugs (PWID), and female sex workers (FSW) than in the general heterosexual (HET) population. However, the contribution of different risk groups in the propagation of the epidemic has not been investigated. Also, the
epidemic is geographically heterogeneous (65% of all new infections occur in nine out of the 47 counties in Kenya). Yet, the rates of HIV-1 transmission between geographic regions have not been described. Also, data are lacking on how levels and trends of HIV drug resistance (HIVDR) in Kenya compare among individuals of different risk groups, with or without antiretroviral therapy (ART) exposure. The primary objective was to phylogenetically describe virus transmission within and between risk groups (MSM, PWID, FSW, and HET) and geographic locations as well as to determine levels of HIV-1 drug resistance over time within and between risk groups in Kenya. A secondary objective was to phylogenetically characterise transmission patterns in a paediatric HIV-1 outbreak in Pakistan.
In the first objective, clustering patterns in Kenya indicated that HIV-1 transmission between risk groups was rare – where most HIV-1 transmission occurs within-risk groups. In addition, when HIV-1 (infrequently) jumped between risk populations, virus jumps from HET to key populations were more common than vice-versa. There was significant West-to-East transmission (i.e. from high-to-low HIV-1 prevalence regions) in the mixed epidemic. Interestingly, Coast and Nairobi provinces were suggested to be important geographic hubs of HIV-1 dissemination in the MSM-specific HIV-1 sub-epidemic. HIVDR analysis revealed that overall pre-treatment HIVDR increased from 6.9% in 1986-2005 to 24.2% in 2016-2020. This was associated with increased non-nucleoside reverse transcriptase inhibitors (NNRTI) resistance in all risk groups. DRMs of any kind were found in treatment naïve HET (13.9%, 95% CI: 12.7-15.2), FSW (19.9%, 95% CI: 15.8-24.6), MSM (15.1%, 95% CI: 9.7-21.9), PWID (31.0%, 95% CI: 19.5-44.5), and children (41.3%, 95% CI: 30.1-53.3). PWID and children were more likely than HET to have DRMs (aOR, 3.5, 95% CI: 1.7-5.4, p<0.001, and aOR, 3.0, 95% CI: 1.8-4.8, p<0.001), respectively. No integrase strand transfer inhibitors (INSTI) drug resistance was detected. Hence, current INSTI-based ART regimens may remain effective in controlling HIV-1 in Kenya. In the secondary objective, clustering patterns in the Pakistani paediatric HIV-1 outbreak revealed multiple introductions of HIV-1 and no phylogenetic HIV-1 mixing between children and
key populations. Findings may be relevant for HIV-1 control in Kenya and Pakistan.