Usefulness of pp65 antigenemia for the early diagnosis of cytomegalovirus disease in patients with AIDS.

An observational cohort study was performed to assess the effectiveness of a cytomegalovirus antigenemia (CMV-Ag) assay designed to predict clinical CMV disease in patients with AIDS. Eighty-six HIV-infected patients with CD4+ cell counts of < 100/mm3, positive CMV IgG, and no previous CMV disease were enrolled. Thirty-eight (44%) patients had at least one positive CMV antigenemia test, ten of whom eventually developed CMV focal disease. CMV disease was diagnosed in 13 (15%) patients. The CMV antigenemia assay was positive in ten of these 13 patients. Using a cut-off value of five positive cells in every 150,000 leukocytes sampled, the CMV antigenemia assay had a positive predictive value of 89% and a negative predictive value of 94%. The median time from the first positive CMV antigenemia test to the onset of CMV disease was 102 days. CMV disease probability at 6 months in patients with a CMV antigenemia value > or = 5 was 77.8% versus 6% in patients with CMV antigenemia value < 5 (log-rank test = 48.345; P < 0.001). Several independent factors were associated with the development of CMV disease: CMV antigenemia > or = 5 cells (hazard ratio: 20.44), CD4+ count < or = 25/mm3 (HR: 3.12), and sexual transmission of HIV infection (hazard ratio, 3.15). CMV antigenemia seems to be a good predictor of CMV disease in patients with AIDS.