Phylogenetic analysis has shown that some HIV positive men who self-report as heterosexual have viruses that cluster only with men who have sex with men (MSM). We characterised this group.
HIV pol sequences were obtained from the UK HIV Drug Resistance Database. To these were added the ten publicly available sequences closest to each UK sequence. Clusters were selected in maximum likelihood phylogenies for analysis in BEAST. Networks were then created by linking together nodes if sequences shared a common ancestor within the previous 5 years in time-resolved phylogenies (Figure 1). Potential nondisclosed MSM (pnMSM) were identified as self-reported heterosexual men who clustered only with men. We compared the centrality (a measure of connectedness and importance) of pnMSM and MSM and calculated assortativity (the propensity for nodes sharing attributes to link) by self-reported risk group. Finally, we evaluated whether pnMSM linked MSM and heterosexuals.
In total, 49772 subtype A1, B and C pol sequences were analysed. Of these, 14405 linked within 5 years and were represented in the network, including 38452 MSM, 1743 female HET and 1341 male HET. We identified 223 network clusters comprising 955 MSM and 249 pnMSM. pnMSM represented 18.6% of linked self-reported heterosexual men, more than twice the proportion of women clustering with MSM (131/1743; 7.5%). pnMSM were more likely to be infected with subtype B than heterosexual men (p<0.0001) and more likely to be Black-African than both MSM and heterosexual men (p<0.0001). pnMSM were less likely to be diagnosed with a recent infection than MSM (12.5% vs 74.9%, p <0.0001) and slightly older (p<0.05). Betweenness centrality was lower for pnMSM than for MSM (2.37 vs 4.11, p<0.005), indicating that they were in peripheral positions in MSM clusters. Assortativity by risk group was higher than expected (-0.124 vs -0.196, p<0.05) indicating that pnMSM linked to each other. We found that self-reported male heterosexuals were much more likely than female heterosexuals to link MSM and heterosexuals (Fisher’s exact test; p<0.0005; OR 2.24).
We have shown that pnMSM do not behave like MSM or male heterosexuals. This has implications both in understanding HIV epidemiology in the UK and for prevention. pnMSM appear to have fewer partners and to preferentially partner with other pnMSM. They are at higher risk for HIV than male heterosexuals and may put female partners at risk by linking the MSM and heterosexual epidemics.