Comprehensive Assessment of Blood and Sexual Exposures Needed for Rigorous Investigation of HIV Transmission

E ditor: Ssemwanga and colleagues examined the relationship between the phylogenetic similarity of HIV-infected women's viruses and their geographic proximity in Kampala, Uganda, in 2008–2009. ¹ From their article, we cannot determine exactly how many sets of HIV-infected women were classified as working or living in the same area of Kampala, nor can we ascertain how many sets had phylogenetically similar viruses. Nonetheless, the four dyads (1, 4, 6, and 7) described in some detail seem to be only a fraction of both the geographically and phylogenetically defined sets of women. The authors assumed that women who lived or worked in the same area were linked sexually, although they reported no data that women in the same sets had sex partners in common or were reachable within a small number of steps in a time-ordered sexual network ² prior to their diagnoses.

It seems to us that Ssemwanga and colleagues have a golden opportunity to extend their study into a rigorous investigation of how these women acquired HIV. In 2004, ^3,4 and in dozens of publications since, we outlined the elements of a research design for determining the modes of HIV transmission with confidence. These elements include assessing blood and sexual exposures comprehensively in incident HIV cases and controls, tracing contacts corresponding to these exposures, and sequencing infected persons' HIV isolates. We successfully implemented a simple version of this design in an investigation of hepatitis C virus transmission among injection drug users. ⁵ The need for such designs is underlined by the accumulating evidence that diverse blood exposures are associated with prevalent and incident HIV infection in sub-Saharan Africa, independent of sexual behaviors. ^{6 –14}

We therefore urge Ssemwanga and colleagues to reinterview all sets of women with phylogenetically similar viruses, regardless of where they lived or worked, about possible blood and sexual exposures to HIV during a fixed period prior to their diagnoses (such as the 5 preceding years). Blood exposures include punctures and cuts received in formal and informal health care, cosmetic care, rituals, and injection (and other) drug use. Interviews should include detailed questions about where, when, and why they received such punctures and cuts and who administered them. ¹³ Questions about sexual exposures should focus on eliciting women's anal and vaginal sex partners and partners' identifying information, dates of contact, and condom use by type of sexual practice.

With such data, the authors could evaluate whether pairs of women who had phylogenetically similar viruses were more likely to be linked than pairs with phylogenetically dissimilar viruses in terms of blood exposures (e.g., exposures received at the same place at a similar date) and sexual exposures (within a small number of steps in a time-ordered sexual network) consistent with the dates of a pair's HIV diagnoses. Data on both types of exposures are likely to be incomplete due to the length of time since exposure and the volume of exposures, but the same was true for the first HIV contact investigation in the United States. ¹⁵ We hope that Ssemwanga and colleagues will persevere as the U.S. investigators did and thereby significantly advance the understanding of HIV transmission in Kampala's high-risk women and, by inference, other high-risk women in Africa.