The presence and degree of odor identification deficits in 55 HIV-infected (30 asymptomatic, 25 symptomatic) and 29 HIV-negative at-risk control volunteers were examined longitudinally using the University of Pennsylvania Smell Identification Test (UPSIT). Factors other than HIV infection that could account for olfactory loss (i.e., sinusitis or upper respiratory infection) were also considered by obtaining MRI scans of the nasal passages and information from an olfaction questionnaire. No differences were found among groups at the first administration of the UPSIT, with significant differences among groups emerging at the 1-year and 2-year follow-ups. The symptomatic group showed a significant decline in odor identification scores across time, while means for the asymptomatic and control groups remained stable. The presence of sinusitis or an upper respiratory infection appeared to have no effect on odor identification. The implications for these findings in relationship to cognitive decline in neurodegenerative diseases are discussed.