Short course AZT for breast-feeding mothers: warning of viral rebound when AZT stopped

The use of AZT (zidovudine, Retrovir) by mothers to prevent the transmission of HIV to their babies lowers HIV viral load in breast milk in the first week after birth, according to a study published in the October 15^th edition of The Journal of Infectious Diseases. However, the study also established that when women stopped taking AZT prophylaxis, viral load in breast milk rose within weeks to levels comparable to those seen in untreated women, increasing infant exposure to the virus and the risks of infection with HIV. Because of this finding the investigators suggest that longer-term antiretroviral treatment using combinations of drugs may be warranted during the early breastfeeding period.

Breast milk and HIV transmission

It is estimated that the infants of untreated HIV-positive mothers ingest over 300,000 cell-free HIV particles a day and 25,000 HIV-infected cells a day if they are breast-fed. The ingestion of a litre of HIV-infected breast milk by an infant is thought to carry the same risk of HIV transmission as an instance of unprotected sexual intercourse between adults.

A study conducted in Thailand and published in 1998 demonstrated that a short course of AZT monotherapy was effective at preventing mother-to-child transmission of HIV. In this study, however, mothers did not breast-feed.

In much of Africa there is no safe alternative to breastfeeding and in 1999 a study in Cote d’Ivoire and Burkina Faso showed that, in the short-term at least, a short course of AZT monotherapy was effective at preventing mother-to-baby transmission of HIV even if the mother breast-fed.

However, it was unknown if AZT reduced HIV viral load in breast milk as well as in plasma and how durable this reduction in HIV replication would be.

DITRAME 049a – AZT prophylaxis, breast milk and mother-to-baby HIV transmission

The DITRAME 049a trial randomised HIV-positive pregnant women to receive short-course AZT prophylaxis or a placebo from week 36-38 of pregnancy to day seven post-delivery. The investigators compared levels of HIV RNA in the breastmilk of women who both did and did not transmit HIV to their infants at day 8 post-delivery, day 45, day 90, day 180 and then every three months if the child was still being breastfed until 18 months of age.

Blood samples were also collected to see if there was any relationship between plasma viral load and viral load in breastmilk and the risk of mother-to-baby HIV transmission. An ultrasensitive assay was used to quantify HIV viral load as low as 10 copies/ml in breastmilk.

Data were available for 384 women, who were recruited to the study between 1995 and 1998. In total, 98 mothers (25%) transmitted HIV to their babies.

Results

The investigators restricted their analysis to the 28 mothers who infected their infants with HIV who had provided a breastmilk sample at day 8 post-delivery. A random selection of 130 mothers who did not transmit HIV to their infants and who also provided a breast milk sample at day 8 was also included in the investigators' analysis.

At day 8, HIV viral load was detectable in the breastmilk of 54% of women who received the placebo and 33% of women who were treated with AZT prophylaxis (p = 0.011). Levels of HIV were lower in the breast milk of the women who received AZT (median 25 copies/ml) than in the breast milk of the women who took the placebo (median 53 copies/ml, p = 0.012).

By day 45, however, the median levels of HIV viral load in breast milk were comparable for both the AZT and placebo arms of the study (27 vs. 25 copies/ml).

At days 8, 45, and 90, HIV was more frequently detected in the breastmilk of the mothers who transmitted HIV to their infants (78%, 82%, 100%) than those who did not (39%, 46%, 41%). At all three time points the difference between the frequency of HIV detection between transmitters and non-transmitters was statistically significant (day eight, p

Furthermore, when HIV was detected, the mothers transmitting HIV has significantly higher levels of HIV in their breast milk than the mothers who did not infect their infants with HIV (day eight, 197 vs. 32 copies/ml, p

The investigators then restricted their analysis to a sample of 80 women (20 of who transmitted HIV, and 60 who did not) with complete data on CD4 cell count at entry, plasma HIV viral load eight days after delivery, HIV viral load in breastmilk eight days after delivery, and HIV viral load in breastmilk at day 45. This established that HIV viral load in breastmilk at day 8 (p = 0.009) and mean difference in viral load in breast milk between day 8 and 45 (p = 0.005) were significantly associated with postnatal HIV transmission. A strong trend was also found between maternal HIV plasma viral load at day eight and transmission (p = 0.056).

Although HIV was detected more frequently in the breast milk of women who received the placebo at day 8 than in that of the women who took AZT, the investigators stress that this was no longer the case by day 45. At day 8 the median viral load of women who took AZT and transmitted HIV was 56 copies/ml compared to a median viral load of 1608 copies/ml to transmitters who took the placebo (p = 0.002). By day 45 the median viral load in breast milk for both the transmitters who took AZT and the placebo was comparable (471 vs. 346 copies/ml, p = 0.72).

However, at day 45, the median viral load in breast milk for mothers, across the study, who did not transmit HIV to their infants was 20 copies/ml.

Conclusions – viral “burst” after AZT stopped

The investigators write, “we generated data that reveal three distinct and novel observations: (1) maternal [AZT] prophylaxis has a significant effect on levels of HIV-1 RNA in breastmilk; (2) [AZT] prophylaxis and withdrawal have a significant effect on the temporal evolution of expression of HIV-1 RNA in breastmilk, with a viral burst on interruption of treatment; and (3) both baseline viral levels of HIV-1 RNA in breast milk and the increment of viral RNA in breastmilk at significantly associated with postnatal mother-to-child HIV transmission.”

Because HIV replication in breastmilk increased to levels comparable to the placebo group by day 45 after AZT prophylaxis was discontinued at day 7, the investigators conclude by suggesting that "longer-term maternal antiretroviral prophylaxis may be an option to prevent transmission by breast-feeding during the early postnatal period, when levels of HIV-1 RNA in breast milk are the highest."

However, mothers exposed to AZT alone for a long period are at risk of developing resistance, so current operational guidelines do not recommend the use of AZT alone throughout the breast-feeding period. Instead studies are looking at the use of combinations of drugs.