HIV-1 RNA levels and the development of clinical disease. North American Lamivudine HIV Working Group.

OBJECTIVE: To assess the prognostic value of HIV RNA levels for predicting clinical disease independently of the CD4 lymphocyte count in patients on antiretroviral therapy.

DESIGN: Cohort of HIV-infected patients from two trials of lamivudine therapy.

PATIENTS: For 620 patients randomized in the North American NUCA3001 and NUCA3002 trials of lamivudine, HIV RNA levels were measured (median, seven measures per patient) and CD4 counts were assessed at a central laboratory (median, 13 counts per patient). Patients were in the 1993 Centers for Disease Control and Prevention (CDC) stages A (n = 439), B (n = 135) or C (n = 46) at baseline.

OUTCOME MEASURES: For patients who were in CDC stage A at baseline we considered the ability of HIV RNA levels and CD4 counts to predict the development of CDC stage B or C disease. A Cox proportional hazards model was used. In a second analysis, patients who were AIDS-free at baseline were considered, and the endpoint was AIDS (CDC stage C).

RESULTS: Patients' initial CD4 counts ranged (5-95% centiles) from 104 to 529 x 10(6)/l (median, 274 x 10(6)/l) and HIV RNA levels from 1900 to 339680 copies/ml (median, 44240 copies/ml). For the first analysis, with CDC stage B or C disease as endpoint, both the most recent HIV RNA level and CD4 count predicted the development of clinical disease [relative hazard (RH) for HIV RNA, 1.96 per 10-fold difference in HIV RNA; 95% confidence interval (CI), 1.41-2.73; P = 0.0001; and RH for CD4 count, 1.82 per twofold difference in CD4 count; 95% CI, 1.27-2.56; P = 0.0009]. When both HIV RNA and CD4 count were included in a multiple regression model, both markers provided information additional to that given by the other (RH for HIV RNA, 1.75; 95% CI, 1.23-2.50; P = 0.002; and RH for CD4 count, 1.40; 95% CI, 0.95-2.07; P = 0.09). In the second analysis, with AIDS as endpoint, both HIV RNA level (P = 0.02) and CD4 count (P = 0.004) were independently associated with clinical progression. These results were essentially unchanged after adjustment for treatment arm (zidovudine/lamivudine versus control arms).

CONCLUSION: The HIV RNA level shows ability to predict the development of clinical disease and may thus be of importance in addition to the CD4 count in patient monitoring.