We prospectively examined a cohort of HIV-positive persons with an accurately estimated date of HIV seroconversion who were infected through injecting drug use or sexual contact to estimate the proportion of long-term nonprogressors (LTNP), considering four definitions of LTNPs. We also evaluated whether factors such as gender, age, and HIV-exposure category were associated with being LTNP; we determined the overlap among the definitions and compared the CD4 and CD8 counts and the CD4/CD8 decline among LTNPs and "moderate" and "fast" progressors. Of the 528 persons selected for analysis, 2 to 4% were considered LTNPs, depending on the definition. The proportion of LTNPs varied by definition, and there was little overlap among definitions. The LTNPs did not appear to differ from "moderate" and "fast" progressors with regard to main demographic characteristics, and they showed a better trend of immunological parameters, appearing to have a slower progression rather than a permanently arrested infection.