Background:  Patterns of recruitment of sex partners vary greatly by men who have sex with men (MSM) and are at high risk of HIV infection. We investigated the types of venues attended by individuals with recent HIV infection in order to characterize where sex partners were recruited, to compare recruitment in different venue types such as bath-houses and bars or clubs, and by demographics.

Methods:  We enrolled 89 men with recent HIV infection in the San Diego Primary HIV Infection cohort were interviewed using a computer assisted survey instrument, and asked to name bars, clubs, bookstores, circuit parties, and parks attended, and the number of sex partners recruited in each. A pair of venues were considered connected if 1 or more individuals recruited sex partners at both venues; individuals were considered connected if they recruited sex partners in the same venue. The data were analyzed using generalized linear models, and social network analysis methods.

Results:  Of 89 MSM, 52 (58%) recruited sex partners from a total of 74 venues, with more partners recruited from bath-houses than from bars or clubs. Of 74 venues, 62 formed an interconnected network, with bath-houses at the core. Bath-houses exhibited high levels of clustering with other venues, and connections to many other venues. One bath-house (A) was of particular importance in the network; not only were the highest number of partners recruited there, but individuals who recruited sex partners at bath-house A also recruited partners at a total of 23 other venues, including other bath-houses, both in San Diego and Los Angeles, bars, clubs, and bookstores. In contrast, there was no significant correlation of individual network characteristics, such as clustering and degree, and individual factors such as age, ethnicity, employment status, and level of education, suggesting mixing of age, ethnic, and socioeconomic status networks within these venues.

Conclusions:  Many participants reported no recruitment of sex partners over a 3-month period, despite attending many bars and clubs, implying that spread of HIV from these individuals during acute/early infection may be limited. However, many of the participants who reported one or more casual partners either attended bath-houses, or bars or clubs where other individuals who attended bath-houses recruited partners. Work is underway to determine whether clustering of individuals by recruitment venue correlates with clustering of HIV by genetic similarity.

Keywords: Epidemiology; Acute HIV Infection; Sexual networks