The standard zidovudine regimen given to expectant mothers and their newborn children is no more effective in preventing perinatal transmission of HIV-1 than regimens in which the course of treatment for either the mothers or their infants is abbreviated, according to a randomized trial conducted in northern Thailand.1 When both mothers and infants receive abbreviated zudovudine treatment, however, the transmission rate is more than double the rate for mother-infant pairs receiving the standard regimen (11% vs. 4%).

To determine whether the standard regimen of zidovudine could be shortened without increasing the risk of mother-to-child HIV transmission, researchers recruited 1,437 HIV-infected women at 27 rural and provincial hospitals in northern Thailand between June 1997 and December 1999. The women, who were enrolled at 28 weeks of gestation, agreed to feed their infants formula rather than breast milk, which could be a vehicle for HIV transmission. Participants were randomly assigned to four treatment groups. Women assigned to the standard regimen received oral zidovudine starting at 28 weeks' gestation and continuing through delivery; the drug was administered to their infants for the first six weeks after their birth. The second group tested abbreviated regimens for both mother and child: Women were given zidovudine from 35 weeks of gestation through delivery, and their infants received zidovudine for only three days following birth. The third regimen called for the standard treatment for the mother and the shortened treatment for the child; in the fourth, only the mother's treatment was shortened. Analysis was based on the participants' assigned regimen rather than on their actual compliance.

The characteristics of the participating women were similar across treatment groups: Eight in 10 were aged 20-30 years, six in 10 had had only one child, more than nine in 10 had no symptoms of HIV infection or AIDS, and none had previously received antiretroviral drugs. Compliance was excellent among women in all groups: More than nine in 10 took at least 80% of the prescribed zidovudine doses. CD4 counts were lower for women who enrolled later in the study than for those who enrolled earlier.

An interim analysis of mother-infant pairs enrolled before December 4, 1998, showed that HIV transmission was significantly higher when both mother and child received shortened treatment (11% vs. 4%); the investigators therefore discontinued assignments to that treatment group. The infants of women who had been assigned to that group on or after the cutoff date but had not yet delivered were given the standard regimen and excluded from the efficacy analysis.

The HIV transmission rates for the group in which only the mother received abbreviated treatment and the group in which only the infant received abbreviated treatment did not differ significantly from the rate for the group receiving the standard treatment (9%, 5% and 7%, respectively). An analysis adjusting for time of enrollment (i.e., before or after randomization was limited to three groups) produced similar, nonsignificant results, as did analyses using HIV transmission or infant death within six months of birth as their endpoint.

The investigators then collapsed the four groups into two, comparing the transmission rate for mothers who received the standard regimen to the rate among mothers who received the abbreviated regimen. They found that mothers who received abbreviated treatment with zidovudine were more than twice as likely as those who received the standard treatment to transmit HIV to their infant (5% vs. 2%). The results did not change after adjustment for time of enrollment.

No women dropped out of the study because of zidovudine intolerance or toxicity. The risk of maternal death, stillbirth or serious complications for women receiving the short course of zidovudine did not differ significantly from the risk for women receiving the standard regimen. The four groups of infants had similar rates of health problems and death.

According to the investigators, their results provide clear evidence that pregnant women should be given the long regimen whenever possible. They suggest that a three-day course of treatment is appropriate for infants whose mothers are treated with zidovudine starting at 28 weeks' gestation, but recommend the long regimen for infants whose mothers start treatment late in pregnancy.

The authors of an accompanying editorial, however, observe that the findings have little significance for countries where "zidovudine prophylaxis has been largely superseded by prophylaxis with a combination of antiretroviral agents."2 They also question the feasibility and effectiveness of a long maternal zidovudine regimen in countries with limited resources, noting that "initiation of treatment at 28 weeks of gestation not only is more costly in terms of drugs but also requires earlier and more frequent antenatal care and thus necessitates a strong health care infrastructure."--A. Hirozawa


1. Lallemant M et al., A trial of shortened zidovudine regimens to prevent mother-to-child transmission of human immunodeficiency virus type 1, New England Journal of Medicine, 2000, 343(14):982-991.

2. Peckham C and Newell ML, Preventing vertical transmission of HIV infection, New England Journal of Medicine, 2000, 343(14):1036-1037.