Background Persons with sexually transmitted infections (STIs), including HIV, are more likely to have sexual partners with STIs. It is not known whether persons with STIs are more likely to have social contacts with STIs. We compared the prevalence of HIV and STIs in the social networks of STI patients to the social networks of community controls.

Methods The study was based at Kamuzu Central Hospital STI Unit from 2010–2012. Three groups of 45 “seeds” were enrolled: HIV-infected STI patients, HIV-uninfected STI patients, and community controls frequency matched on age, gender, and area. Seeds were asked to recruit < 5 “contacts.” HIV prevalence and STI prevalence among contacts were compared between arms using generalised estimating equations accounting for correlation by seed.

Results Mean number of contacts recruited was 1.3 for HIV-infected clinic seeds, 1.8 for HIV-uninfected clinic seeds, and 2.3 for community seeds (N = 243). The majority of contacts (89%) had never been in a sexual relationship with the recruiting seed. HIV prevalence was higher in HIV-infected clinic seeds’ contacts (31%) compared to community seeds’ contacts (11%) (p = 0.009), but not in HIV-uninfected clinic seeds’ contacts (10%) compared to community seeds’ contacts (11%) (p > 0.9). STI syndrome prevalence was higher in both HIV-infected clinic seeds’ contacts (18%) and HIV-uninfected clinic seeds’ contacts (26%) compared to community seeds’ contacts (8%) (p = 0.04). Differences between arms remained significant for both HIV and STI prevalence when analysis was restricted to non-sexual contacts.

Conclusion Seeds with HIV were more likely to have networks with higher HIV prevalence and seeds with STIs were more likely to have networks with higher STI prevalence, even among non-sexual contacts. Research is needed to understand the social and epidemiologic mechanisms underlying these findings. Social contact recruitment at an STI clinic is one promising way of identifying persons in need of HIV and STI services.