Demographics of people who transmit HIV-1 in Zambia: a molecular epidemiology analysis in the HPTN-071 PopART study

Abstract

BACKGROUND In the last decade, universally available antiretroviral therapy has led to reduced HIV incidence in sub-Saharan Africa. Sources of remaining transmission need to be characterised to design effective prevention strategies. METHODS We used phylogenetics to understand the population characteristics of people who are sources of infection. HIV samples from 6,864 individuals from Zambia were deep-sequenced as part of HPTN 071-02 (PopART) Phylogenetics between 2014 and 2018. We identified 300 likely directed transmission pairs and analysed their sources to better understand transmission in the general population. RESULTS After demographic weighting of the recipient population to match the estimated total population infected during the trial period, 59.4% (95% CI: 53.1%-65.8%) of transmissions were male-to-female, with 43.1% (36.6%-49.5%) of transmissions from males aged 25-40. Since the adult HIV prevalence was 2.0 times higher in women than men, the per-capita transmission rate was 2.93 times higher per infected male than per infected female. 25.6% (19.9%-31.3%) of sources were estimated to have themselves been infected less than a year before the transmission event. 16.8% (12%-21.7%) of sources transmitted viruses resistant to first-line ART. 13% (8.6%-17.4%) of transmissions occurred between individuals from different study communities. CONCLUSIONS Our findings suggest that HIV transmission in the study communities took place as part of common sexual mixing, and that there was no outsized contribution of importation, from drug resistance, or recent infection. Men aged 25-40 should be the focus of treatment for prevention interventions, even if linking them to care requires more effort.

Growing gender disparity in HIV infection in Africa: sources and policy implications

“…Empiric incidence data published over the last decade documents widespread declining incidence across the African continent 24 , and UNAIDS mathematic models from eastern and southern Africa have estimated a 43% reduction in incidence regionally since 2010 1 . Our findings are also compatible with genderstratified HIV incidence data from prospective African population-level studies that indicate greater differences in rates of new infections between men and women in the same cohort over calendar time 24 , data from population surveillance studies and HIV treatment and prevention trials showing lower levels of viral suppression among men compared to women with HIV 64,65 , and phylogenetic studies from Botswana 28 and Zambia 29 . Together, these observations suggest that the principal characteristics of the evolving HIV epidemic summarized in Table 1 likely apply more broadly in similar rural and semi-urban populations across Eastern and Southern Africa.…”

“…To characterize the population transmission flows by age and gender that underly observed shifts in incidence, we deep-sequenced virus from 2174 participants with HIV and sufficient viral load for sequencing (Supplementary Table S5 25 ). By embedding genomic surveillance into a population-based cohort study, deep-sequence sampling coverage was high relative to typical pathogen sequencing studies, which is essential for reconstructing transmission events (Supplementary Table S6 29,34,35,36,37,38 ).…”

Longitudinal population-level HIV epidemiologic and genomic surveillance highlights growing gender disparity of HIV transmission in Uganda

“…Changing drivers of HIV infection in Africa men compared to women with HIV 64,65 , and phylogenetic studies from Botswana 28 and Zambia 29 . Together, these observations suggest that the principal characteristics of the evolving HIV epidemic summarized in Table 1 likely apply more broadly in similar rural and semi-urban populations across Eastern and Southern Africa.…”

Subnational HIV incidence trends in Malawi: large, heterogeneous declines across space

“…This assumption is consistent with a recent analysis of viral genetic data suggesting that HIV transmission in SSA is highly local. 50 Second, we omitted age structure from the compartmental model. Age is a critical determinant of HIV infection risk and mortality, but explicitly representing age resulted in computationally intractable number of compartments for our inference framework.…”