Current guidelines recommend consideration of CD4 cell percentage as well as CD4 cell count in therapeutic decisions. The relative value of CD4 cell count compared with CD4 cell percentage in predicting risk of AIDS-defining illnesses (ADIs) in the post-HAART (highly active antiretroviral therapy) era is unknown. Data from an observational clinical cohort of adult HIV-infected patients were used to assess the risk of developing an ADI associated with specific absolute CD4 counts (CD4) and CD4%'s (CD4%) using all CD4-CD4% pairs obtained after January 1996. The incidence of developing an ADI was assessed over a maximum of 6 months after the CD4-CD4% pair was obtained. Using multivariable negative binomial regression, the incidence rate ratio (IRR) for developing an ADI by CD4 and CD4% categories was computed. A total of 15,736 CD4-CD4% pairs from 2185 patients who developed 608 ADIs was analyzed. The IRR for developing an ADI by absolute CD4 was 17.9 (95% CI: 13.2, 24.4) events/100 person-years for <50 cells/mm, 6.2 (95% CI: 4.4, 7.9) for 50-100 cells/mm, and 2.7 (95% CI: 1.9, 4.0) for 100-200 cells/mm, compared with the referent stratum of 200-350 cells/mm. Without adjustment for absolute CD4, the IRR was 14.4 (95% CI: 9.3,22.6) for CD4% <7%, 3.7 (95% CI: 2.4,5.9) for 7-14%, 1.9 (95% CI: 1.1, 3.1) for 15-21%, compared with the referent stratum of >21%. However, in a multivariable analysis adjusting for absolute CD4, CD4%, and other clinical and demographic variables, the absolute CD4 but not the CD4% was associated strongly with developing an ADI. The results suggest that CD4% adds little further predictive information after accounting for the absolute CD4 count for the short-term risk of developing an ADI. The absolute CD4 count is the more important measure of immune status and is preferred over the CD4% for making treatment decisions in HIV-infected adults.