PLASMA VIRAL LOAD AND CD4(-1 INFECTION() LYMPHOCYTES AS PROGNOSTIC MARKERS OF HIV)

Background: The rate of disease progression among persons infected wit h human immunodeficiency virus type 1 (HIV-1) varies widely, and the r elative prognostic value of markers of disease activity has not been d efined. Objective: To compare clinical, serologic, cellular, and virol ogic markers for their ability to predict progression to the acquired immunodeficiency syndrome (AIDS) and death during a 10-year period. De sign: Prospective, multicenter cohort study. Setting: Four university- based clinical centers participating in the Multicenter AIDS Cohort St udy. Patients: 1604 men infected with HIV-1. Measurements: The markers compared were oral candidiasis (thrush) or fever; serum neopterin lev els; serum beta(2)-microglobulin levels; number and percentage of CD3( +), CD4(+), and CD8(+) lymphocytes; and plasma viral load, which was m easured as the concentration of HIV-1 RNA found using a sensitive bran ched-DNA signal-amplification assay. Results: Plasma viral load was th e single best predictor of progression to AIDS and death, followed (in order of predictive strength) by CD4(+) lymphocyte count and serum ne opterin levels, serum beta(2)-microglobulin levels, and thrush or feve r. Plasma viral load discriminated risk at all levels of CD4(+) lympho cyte counts and predicted their subsequent rate of decline. Five risk categories were defined by plasma HIV-l RNA concentrations: 500 copies /mL or less, 501 to 3000 copies/mL, 3001 to 10 000 copies/mL, 10 001 t o 30 000 copies/mL, and more than 30 000 copies/mL. Highly significant (P < 0.001) differences in the percentages of participants who progre ssed to AIDS within 6 years were seen in the five risk categories: 5.4 %, 16.6%, 31.7%, 55.2%, and 80.0%, respectively. Highly significant (P < 0.001) differences in the percentages of participants who died of A IDS within 6 years were also seen in the five risk categories: 0.9%, 6 .3%, 18.1%, 34.9%, and 69.5%, respectively. A regression tree incorpor ating both HIV-1 RNA measurements and CD4(+) lymphocyte counts provide d better discrimination of outcome than did either marker alone; use o f both variables defined categories of risk for AIDS within 6 yea rs t hat ranged from less than 2% to 98%. Conclusions: Plasma viral load st rongly predicts the rate of decrease in CD4(+) lymphocyte count and pr ogression to AIDS and death, but the prognosis of HIV-infected persons is more accurately defined by combined measurement of plasma HIV-1 RN A and CD4(+) lymphocytes.