RESUMO
OBJECTIVES: To investigate whether combining clinical data with the serological testing algorithm for recent HIV seroconversion (STARHS) reliably identifies otherwise unrecognized recent infections and observe their trends. DESIGN: Incorporation of STARHS into routine HIV diagnosis. METHODS: STARHS was applied to serum collected between 1996 and 2005 at HIV diagnosis and routine clinical/laboratory markers of recent infections were determined. The recent infections were identified by conventional means, by STARHS, and by both combined. RESULTS: Of 1526 infections diagnosed, 812 were new. Of these, 604 were in men who have sex with men (MSM); 208 in heterosexuals; 88% had serum available for STARHS, which identified 88 incident infections that would otherwise have been unrecognized (12% of all new infections, 34% of all recent infections). Of these, 88% reported recent high-risk sex; 47% reported seroconversion symptoms. STARHS confirmed recent infections in 71 of 74 (96%) known to be infected within 6 months by conventional methods. Combining both approaches, recent infections increased over time from 26% (1996) to 45% (2005) [P < 0.001]. STARHS results from 3% new diagnoses and 8% previous diagnoses were deemed false incident (associated with antiretroviral therapy, advanced disease or undetectable viral load). False incident results were only inexplicable in two individuals. CONCLUSION: Adjunctive use of STARHS with clinical data identified a high and increasing proportion of new HIV diagnoses as recent infections, confirming significant ongoing transmission. Since 2002, 50% of new diagnoses among MSM were recent infections. Identification of additional recent infections by STARHS enables effective intervention that may benefit the individual and reduce onward transmission.