Background:
HIV infection in San Francisco has occurred primarily among men who have sex with men (MSM). For prevention programs to be effective, localities must monitor trends in HIV incidence. The serologic testing algorithm for recent HIV seroconversion (STARHS) characterizes HIV-infected persons as having recent (< six months) or longstanding infection and can be used to estimate HIV incidence. Unfortunately, after 2004 STARHS was no longer routinely available.

Objective:
Determine the incidence of HIV over time in San Francisco city municipal STD clinic patients

Method:
STARHS was applied to HIV-positive specimens from persons receiving voluntary HIV testing at the San Francisco municipal sexually transmitted disease (STD) clinic from January 2001 through December 2004 to determine trends in HIV incidence. Sexual and other risk behaviors were routinely ascertained from all patients undergoing testing.

Result:
HIV incidence among MSM who did not report injecting drugs was 3.5%/year in 2001, 4.4%/year in 2002, 4.2%/year in 2003, and 3.6%/year in 2004 (p=0.47, Cochran-Armitage chi square). Among MSM who also reported injection drug use, there were no recent infections found in 2001. HIV incidence in MSM who injected drugs was 33.9%/year in 2002, 13.4%/year in 2003, and 8.3%/year in 2004 (p=0.46, Cochran-Armitage chi square).

Conclusion:
Despite years of prevention efforts directed at MSM, HIV incidence remains high and stable in this group, and is especially high among MSM who also inject drugs in San Francisco.

Implications:
Novel HIV prevention efforts are required and a critical assessment of current programs