Background: The optimal duration of zidovudine administration to prevent perinatal transmission of human immunodeficiency virus type 1 (HIV-1) should be determined to facilitate its use in areas where resources are limited.
Methods: We conducted a randomized, double-blind equivalence trial of zidovudine starting in the mother at 28 weeks' gestation, with 6 weeks of treatment in the infant (the long-long regimen), which is similar to protocol 076; zidovudine starting at 35 weeks' gestation, with 3 days of treatment in the infant (the short-short regimen); a long-short regimen; and a short-long regimen. The mothers received zidovudine orally during labor. The infants were fed formula and were tested for HIV DNA at 1, 45, 120, and 180 days. After the first interim analysis, the short-short regimen was stopped.
Results: A total of 1437 women were enrolled. At the first interim analysis, the rates of HIV transmission were 4.1 percent for the long-long regimen and 10.5 percent for the short-short regimen (P=0.004). For the entire study period, the transmission rates were 6.5 percent (95 percent confidence interval, 4.1 to 8.9 percent) for the long-long regimen, 4.7 percent (95 percent confidence interval, 2.4 to 7.0 percent) for the long-short regimen, and 8.6 percent (95 percent confidence interval, 5.6 to 11.6 percent) for the short-long regimen. The rate of in utero transmission was significantly higher with the two regimens with shorter maternal treatment (5.1 percent) than with the two with longer maternal treatment (1.6 percent).
Conclusions: The short-short zidovudine regimen is inferior to the long-long regimen and leads to a higher rate of perinatal HIV transmission. The long-short, short-long, and long-long regimens had equivalent efficacy. However, the higher rate of in utero transmission with the short-long regimen suggests that longer treatment of the infant cannot substitute for longer treatment of the mother.