The rate at which South Carolina women infected with HIV transmitted the disease to their newborns declined dramatically in the three years after the Public Health Service recommended a regimen of zidovudine therapy to reduce perinatal transmission of the virus.1 In 1993, the year before results of a large-scale clinical trial led the federal government to recommend zidovudine use for infected women during pregnancy and delivery, and for their infants during the perinatal period, 13% of babies born to women with HIV acquired the virus; for the years 1995-1997, the proportion was 6%. Furthermore, the transmission rate was significantly lower if zidovudine had been used at all three recommended times than if it had not.
To study the relationship between zidovudine use and perinatal transmission of HIV, analysts examined data from a statewide registry that includes all reported cases of HIV and AIDS; they supplemented this information by reviewing the medical records of infected women and their infants. The comparison of transmission rates before and after the government recommendations were issued included all mother-child pairs in whom the women were known to be HIV-positive before giving birth, received prenatal and delivery care in South Carolina and had a live birth: a total of 88 pairs in 1993 and 217 in 1995-1997. Subsequent analyses of the 1995-1997 cohort excluded women whose records lacked data on CD4 counts and illicit drug use during pregnancy (factors that could affect transmission), leaving a sample of 149 women.
In 1993, the rate of perinatal HIV transmission was 13%; for the three-year period 1995-1997, the rate was 6% (3-8% annually), and the difference was statistically significant. During the later period, the rate was significantly lower if the complete zidovudine regimen had been followed for the mother and child (3%) than if it had not (15%).
Between 1995 and 1997, the proportion of mother-child pairs who received all three parts of the recommended regimen increased steadily and significantly from 68% to 93%. Using individual logistic regression models, which did not control for potentially confounding variables, the analysts examined the effects of a wide range of factors on the likelihood of receipt of the complete regimen. According to the results, the likelihood that mother-child pairs had received all three parts of the treatment were significantly reduced if the mother had used illicit drugs or had had a sexually transmitted disease diagnosed during pregnancy (odds ratios, 0.4 and 0.1, respectively).
Other findings, while not achieving statistical significance, suggest additional relationships: The likelihood that women and their infants received the complete recommended treatment appears to be elevated among women with the lowest CD4 counts, residents of rural areas, women who experienced premature rupture of membranes more than four hours before delivery, those who had had adequate prenatal care, those lacking private insurance and those who delivered by cesarean.
In addition, the analysts found that the infant's infection status was significantly associated only with receipt of the complete zidovudine regimen: The odds of infection were reduced by 80% (odds ratio, 0.2) in mother-child pairs who had received zidovudine at all three recommended times. While the remaining findings were not statistically significant, they suggest that the risk of transmission is elevated among women with low CD4 counts, those who engaged in risky behaviors or had a sexually transmitted disease diagnosed during pregnancy, those who had had inadequate prenatal care and those who had received a prescription for zidovudine before becoming pregnant.
Summing up their findings, the researchers point out that their population-based study not only documents a decline in perinatal HIV transmission but also "demonstrates how quickly a small, southern, rural state responded to the recommendations for the use of zidovudine in HIV-1-infected pregnant women and their children." If transmission rates are to be reduced further, they conclude, interventions will be needed to address such problems as inadequate prenatal care and pregnant women's use of illicit drugs.--D. Hollander
REFERENCE
1. Harris NS et al., Zidovudine and perinatal human immunodeficiency virus type 1 transmission: a population-based approach, Pediatrics, 2002, 109(4), <http:// www.pediatrics.org/cgi/content/full/109/4/e60>.