Objective: To assess the effect of zidovudine on productive HIV infection of the brain.
Design: To correlate the incidence of HIV-specific neuropathology with zidovudine therapy.
Patients: We examined 192 AIDS cases neuropathologically; 97 had never been treated with zidovudine, 72 had received zidovudine for over 3 months and until death, 23 had their treatment terminated more than 1 month before death.
Results: The incidence of HIV encephalitis/HIV leukoencephalopathy (HIVE/HIVL) and of multinucleated giant cells (MGC) was significantly lower in patients who had received zidovudine than in those who had never received zidovudine. The yearly incidence of HIVE/HIVL increased between 1982 and 1987 probably because of improved survival, and decreased between 1987 and 1990 although the percentage of patients treated with zidovudine increased. Since 1991 the incidence of HIVE/HIVL and of MGC increased slightly. The percentage of patients treated with zidovudine until death decreased and that of patients whose treatment was terminated increased concomitantly. In 1989 and 1990, most patients whose treatment was terminated had MGC and HIVE/HIVL. In 1991 and 1992 this incidence decreased markedly, coinciding with the introduction of dideoxyinosine therapy.
Conclusion: Zidovudine treatment significantly reduces the occurrence of productive HIV infection of the brain in AIDS. Discontinuing zidovudine therapy may favour the occurrence of HIV encephalitis. Substitution therapy with dideoxyinosine also appears to protect against HIV-specific brain pathology.