Background:  The circumstances surrounding the emergence of HIV-1 group M, subtype B (the predominant strain of AIDS virus in Europe, Haiti, the United States, and the rest of the Americas) remain unclear. Here we report a set of “fossil” HIV-1 sequences that provide definitive evidence of where, when, and how subtype B emerged.

Methods:  We recovered complete HIV-1 env gene sequences from specimens collected in 1982-1983 from 5 Haitian AIDS patients, all recent immigrants to the United States who were among the first recognized AIDS victims. We tested the hypothesis of a Haitian origin for subtype B by conducting a detailed Bayesian Markov chain Monte Carlo phylogenetic analysis, including a “relaxed” molecular clock analysis, using an alignment of the fossil sequences plus 117 previously published subtype B env sequences from 19 different countries.

Results:  The hypothesis of a US or other non-Haitian origin of subtype B is strongly rejected (p <0.001) in favor of a Haitian origin (p = 0.999). HIV-1 moved from Africa to Haiti in a single patient in or around 1966 (1962-1970). It then spread there for some years before first successfully dispersing elsewhere. Almost all non-Haitian subtype B infections around the world can be traced to a single migration of the virus out of Haiti in or around 1969 (1966-1972), a key turning point in the history of the AIDS pandemic. One exception is the subtype B epidemic in Trinidad and Tobago, which emanated from a separate, single-patient introduction from Haiti.

Conclusions:  Our findings establish Haiti as the country with the oldest HIV/AIDS epidemic outside Sub-Saharan Africa. Because of its 40-year history, the HIV-1 epidemic in Haiti exhibits a greater range of viral genetic diversity than the rest of the world’s subtype B strains combined, a fact relevant to vaccine design. The timing of the Haitian origin of the epidemic supports the idea that the genesis of subtype B occurred with the return of 1 of the many Haitian professionals who worked in the Congo in the 1960s. The timing of the subsequent single-patient initiation of the United States and worldwide epidemic shows conclusively that HIV-1 was circulating in the United States for over a decade before the recognition of AIDS in 1981. Our results suggest that the global spread of HIV-1 involves more inertia than previously supposed, with major outbreaks hinging on rare, single transmission events. They also provide compelling independent corroboration of an early 20^th century M-group ancestor.