Demographic characteristics of sources of HIV-1 transmission in Zambia

BackgroundIn the last decade, universally available antiretroviral therapy (ART) has led to greatly improved health and survival of people living with HIV in sub-Saharan Africa, and appears to have contributed to reduced rates of new infections. Individuals acting as sources of infection need to be characterised to design effective prevention strategies. MethodsWe used viral genomes to investigate the demographic characteristics of sources of HIV-1 infection. Between 2014 and 2018, the HPTN 071 PopART study was conducted to quantify the public health benefits of ART. Viral samples from 7,124 study participants in Zambia were deep-sequenced as part of HPTN 071-02 PopART Phylogenetics, an ancillary study. We identified 300 likely HIV-1 transmission pairs and investigated the source individuals in those pairs to better understand transmission in the general population. FindingsAfter demographic weighting, 59.4% of transmissions were male to female, with 43{middle dot}2% (95% CI: 36{middle dot}8%-49{middle dot}7%) of transmissions being from males aged 25-40. Overall, men transmitted 2.09-fold (2{middle dot}06-2{middle dot}29) more infections per capita than women, a ratio peaking, when stratified by source age, at 5.88 (2{middle dot}78-15{middle dot}8) in the 35-39 age group. 17{middle dot}4% of sources (12{middle dot}5%-22{middle dot}4%) carried viruses resistant to first-line ART. 12{middle dot}9% (8{middle dot}5%-17{middle dot}3%) of transmissions linked individuals from different communities in the trial. InterpretationHIV-1 transmission in the HPTN 071 study communities comes from a wide range of age and sex groups, and that there is no outsized contribution of importation or drug resistance mutations to new infections. Men aged 25-40, under-served by current treatment and prevention service, should be prioritised for HIV testing and ART. FundingNational Institute of Allergy and Infectious Diseases, US Presidents Emergency Plan for AIDS Relief, International Initiative for Impact Evaluation, Bill & Melinda Gates Foundation, National Institute on Drug Abuse, and National Institute of Mental Health Research in context panelO_ST_ABSEvidence before this studyC_ST_ABSWe searched PubMed, with no date or language filters, to identify previous quantitative studies investigating the role of age, sex, mobility and drug resistance mutations (separately or in combination) as drivers of new HIV transmissions in heterosexual transmission in sub-Saharan Africa. Observational studies were considered along with those using phylogenetic or mathematical modelling methodologies. In observational studies, having an older partner or a migrant partner is frequently identified as a risk factor for HIV acquisition, particularly in women, but this is a slightly separate question to the quantification of the overall frequency of these demographics amongst the sources of new infections. The most recent studies of drug resistance have shown an increasing prevalence, particularly in NNRTI resistance. The ability to use phylogenetics to investigate HIV transmission is reasonably recent, and the ability to use it to reconstruct who infected whom in transmission pairs is more recent still. One previous, influential study in South Africa posited a key role of men in their 30s in infecting very young women, and being themselves infected by women in their 30s, but that work did not use a methodology that was able to conclusively reconstruct direction of transmission. A more recent study in Botswana found similar age distributions for male and female sources of transmission, with the average individual being in their late 30s or early 40s. However, these ages were recorded at the time of study enrolment and do not take into account the time from infection to sampling. The paper also showed that, in that setting, the majority of transmissions were between members of the same community. A number of other phylogenetic studies have been concerned with the very specific dynamics of HIV-1 in the fishing communities of Lake Victoria. No previous phylogenetics work to our knowledge has considered the distribution of drug resistance mutations in sources, and none considered the interaction between source characteristics. We were unable to find any previous mathematical modelling studies for which the characterisation of sources according to these variables was a major focus. Added value of this studyOur methodology uses a phylogenetic approach to identify likely transmission pairs and the direction of transmission between their members. We find a total of 300 pairs, larger than any previous study. We demonstrate a novel and simple new approach to accounting for potential sampling bias. We also employ a methodology that allows us to estimate the ages of the individuals involved at the time of transmission, rather than that of sampling, countering a key bias in previous approaches. Partly for this reason, our age profiles for sources peak at earlier ages than in previous work, with an average in the early 30s for male sources and mid-20s for females. We fail to confirm the existence of a "renewal cycle" of transmission involving a major contribution from women in older age groups. We also examine the contribution of outside-community transmission and drug resistant mutations, and, for the first time, show that these three characteristics (age/sex, migration, and drug resistance) operate on separate axes and do not cluster together. We use our results to calculate the relative contribution of male to female sources to transmission in age bands, finding that this grows to a peak in the 35-39 age group in which men are responsible for almost six times as many new infections per capita as women. Implications of all the available evidenceThe heterosexual HIV epidemic in sub-Saharan Africa appears to be maintained by transmissions from young women and slightly, but not dramatically, older men. The contributions of sources transmitting drug-resistant virus, or sources who reside outside a focal community, is not particularly large, and there is no disproportionate contribution from individuals who share any combination of a high-risk age group, a residence outside the community, and drug-resistant virus. In generalised HIV epidemics, it is tempting to attempt to identify particular demographic groups, of relatively small sizes, for whom intensive targeting of prevention measures will have a major effect on transmission in the general population. The current state of the evidence suggests that this may not be possible, as the demographic profile of sources of transmission is not dissimilar to that of the general population. While it may be more difficult and resource-intensive to design universal interventions for the whole population, there may be no shortcuts.