The aim of the study was to assess—while controlling for individual risk characteristics—how certain social network structural characteristics (degree, eigenvector, and betweenness centrality) are related to HIV infections. Injecting drug users (N = 299) in Vilnius, Lithuania were recruited using incentivized chain referral sampling for a cross-sectional study. Sociometric social links were established between participants, and UCINET was used to calculate network measures. HIV prevalence was 10 %, and all except two knew they were infected. Of the five variables that remained significant in the final multivariate model, one showed temporal cumulative infection risk (more years since first drug injecting), three reflected informed altruism (always using condoms, less distributive syringe sharing and having not more than one sex partner), and one pointed to the importance of social network structure (betweenness centrality, indicating bridge populations). Loess regression indicates that betweenness may have the highest impact on HIV prevalence (about 60 vs. 20 % estimated HIV prevalence for the highest betweenness centrality values vs. highest age values). This analysis contributes to existing evidence showing both potential informed altruism (or maybe social desirability bias) in connection with HIV infection, and a link between HIV infection risk and the role of bridges within the social network of injecting drug user populations. These findings suggest the importance of harm reduction activities, including confidential testing and counseling, and of social network interventions.