Screening for HIV: a review of the evidence for the U.S. Preventive Services Task Force

Abstract

Background: HIV infection affects 850,000 to 950,000 persons in the United States. The management and outcomes of HIV infection have changed substantially since the U.S. Preventive Services Task Force issued recommendations in 1996.

Purpose: To synthesize the evidence on risks and benefits of screening for HIV infection.

Data sources: MEDLINE, the Cochrane Library, reference lists, and experts.

Study selection: Studies of screening, risk factor assessment, accuracy of testing, follow-up testing, and efficacy of interventions.

Data extraction: Data on settings, patients, interventions, and outcomes were abstracted for included studies; quality was graded according to criteria developed by the Task Force.

Data synthesis: No trials directly link screening for HIV with clinical outcomes. Many HIV-infected persons in the United States currently receive diagnosis at advanced stages of disease, and almost all will progress to AIDS if untreated. Screening based on risk factors could identify persons at substantially higher risk but would miss a substantial proportion of those infected. Screening tests for HIV are extremely (>99%) accurate. Acceptance rates for screening and use of recommended interventions vary widely. Highly active antiretroviral therapy (HAART) substantially reduces the risk for clinical progression or death in patients with immunologically advanced disease. Along with other adverse events, HAART is associated with an increased risk for cardiovascular complications, although absolute rates are low after 3 to 4 years.

Limitations: Data are insufficient to estimate the effects of screening and interventions on transmission rates or in patients with less immunologically advanced disease. Long-term data on adverse events associated with HAART are not yet available.

Conclusions: Benefits of HIV screening appear to outweigh harms. The yield from screening higher-prevalence populations would be substantially higher than that from screening the general population.