We wished to determine whether antiretroviral therapy with zidovudine first received when a person is diagnosed with AIDS was associated with diminished or increased direct health resource utilization. As a measure of health resource utilization, we examined all Medicaid-administered health care charges to adult Maryland residents diagnosed with AIDS from 1987 to 1989 who were part of the Human Immunodeficiency Virus Information System. We specifically compared those persons who first received zidovudine therapy either prior to or within 60 days of diagnosis of AIDS (n = 101) with those who never received zidovudine therapy (n = 279). Median survival time after diagnosis of AIDS in those who received zidovudine was 605 days and in those who did not receive zidovudine 235 days. After diagnosis of AIDS, median per-person lifetime direct health care charges to Medicaid were $66,200 in those who received zidovudine and $31,300 in those who did not receive zidovudine. The median incremental charge per year of life gained in zidovudine users was $34,600 compared with nonusers. Adjusting by proportional hazards regression for age, gender, race/ethnicity, HIV transmission risk group, AIDS-defining diagnosis, and length of follow-up, lifetime Medicaid charges were higher in zidovudine receivers. Compared with patients who did not receive zidovudine, patients who first received zidovudine at the time AIDS was diagnosed incurred higher cumulative lifetime charges, associated principally with longer survival time. The rate of resource utilization was not decreased by zidovudine use.