Disconnects in presumed heterosexual HIV transmission in Malawi

Abstract

Obtaining large-scale sexual network data is difficult [1] and doing so in a poor country is especially difficult. We congratulate Helleringer and Kohler [2] on their pioneering study of a sexual network on Likoma Island, Malawi. Their approach represents an important step toward appropriate research designs for elucidating HIV transmission dynamics in sub-Saharan Africa. Despite this achievement, their results undermine their assertion that the Likoma sexual network sustains epidemic HIV transmission. In fact, their findings add to the already considerable evidence against penile–vaginal intercourse as the primary driver of HIV transmission in sub-Saharan Africa [3]. In moderate prevalence Likoma, Helleringer and Kohler observed an ‘apparently paradoxical distribution of HIV prevalence’. They found that persons in the core (bicomponents) of the main component of the sexual network were less likely to be HIV infected than those in peripheral positions in the main component or in small network components. This finding is at odds with prior empirical research, which indicates that, when significant HIV transmission occurs, persons located in the core of the corresponding risk network are more likely to be infected. In one high prevalence community, persons in the core of the injection and sexual network had considerably higher HIV prevalence and more frequent risk behaviors than those in peripheral or more isolated positions [4]. In that study, injection and homosexual risk behaviors were positively associated with prevalent HIV infection, but heterosexual risk behaviors were not. Similarly, among heterosexuals in a very low prevalence setting with little ongoing HIV transmission, infected persons were, with one exception, isolates, in small components, or on the periphery of the main component of the risk network [5,6]. In Likoma, it is surprising that self-reported symptoms of sexually transmitted infections (STI) are almost evenly distributed across the sexual network, and are not concentrated in the core as expected. Furthermore, no sexual risk variable (condom use, number of sex partners, STI symptoms, or age difference between sex partners) shows a meaningful association with prevalent HIV infection. Taken together, these observations suggest that the measured sexual network might not underlie HIV transmission in Likoma, and that other unmeasured blood and sexual exposures might be involved. To establish the primary modes of HIV transmission in sub-Saharan Africa with confidence, more appropriate research designs are required. Specifically, it is critical to assess blood and sexual exposures comprehensively (beyond just injections and vaginal intercourse) for incident cases and controls, trace their corresponding contacts and sequence-infected persons' HIV isolates [7–9].