Background:  Reports of HIV-1 super-infection have mostly involved recent seroconverters, in whom the incidence of apparent superinfection is comparable to the incidence of primary HIV infection. In contrast, super-infection among chronically infected persons is rare. The prevalence of primary and co-infection in North America is not known.

Methods:  The Positive Partners study in San Francisco was a prospective cohort of 435 HIV⁺ individuals who enrolled with a seroconcordant sexual partner (partner Group A) or reported unprotected intercourse with multiple unenrolled seropositive partners (partner Group B). Eligibility was based on reported unprotected intercourse (UI) with HIV⁺ partners in the past year. Pol, gag, and tat population sequences were analyzed phylogenetically; among individuals suppressed on ART sequences were derived from proviral DNA.

Results:  In Group A, 242 persons involving 178 partnerships were followed, 86% of the cohort reported more than 2 years since HIV diagnosis (median = 9.7 years). A total of 229.5 person-years of observation occurred and 16,987 episodes of UI were reported. An additional 48 persons contribute to 836 Group B partnerships and 3294 episodes of UI. Despite this high level of exposure no super-infection in Group A linked to a virologically characterized enrollment partner was detected. Apparent dual infections, indicated by phylogenetically divergent virus populations at separate timepoints, were observed in 10 individuals primarily in DNA (9 of 175) of virologically suppressed participants and 1 in RNA (1 of 115).

Conclusions:  The lack of systemic super-infection in partnerships with known exposure to divergent viruses (Group A, 0 of 230 person-years) and only 1 systemic SEDI not acquired from an enrolled partner in a monogamous individual (Cohort B, 0 of 836 UAI partnerships) was less than expected in the absence of mechanisms blocking HIV super-infection. Given high rates of primary co-infection reported in subtype B epidemics (24%, US) evidence of divergent viral strains, especially in DNA during suppressive therapy may reflect frequent primary co-infection. Intensive virological assessment of recently infected individuals for evidence of primary co-infection and follow-up to document sequential expression of co-infecting viral variants would clarify inquiry into whether appearance of divergent viral variants reflects HIV-1 super-infection or SEDI of initially co-infected persons.