Information from two ACTG trials, protocols 076 and 185, has provided some insight into the association between maternal viral load and vertical transmission. In ACTG 076, a three-part AZT protocol reduced the rate of HIV transmission by two-thirds, from about 25% to 8%, in a group of relatively healthy, mostly treatment-naïve pregnant women. (The three-part AZT regimen consisted of 100 mg oral AZT five times daily commenced sometime during the last six months of pregnancy; then IV AZT during labor, and finally, AZT syrup given to the newborn for six weeks.) ACTG 185 studied a cohort of sicker women, with lower CD4 counts, higher viral loads and more time on prior antiviral therapy. The group received the 076 regimen plus either HIVIG (an antibody preparation taken from people who produce high levels of HIV antibodies) or IVIG (a standard antimicrobial preparation from HIV-negative persons). Researchers had expected the rate of transmission to be higher in the 185 population and were surprised to find that it was actually only 4.8% in both treatment arms.
At a symposium at the 5th Conference on Retroviruses and Opportunistic Infections, Lynne Mofenson, M.D., of the National Institute of Child Health and Human Development at NIH, gave an update on both trials (lecture S11). In ACTG 076, transmission occurred across all viral loads, in both the treatment and placebo groups. There was no threshold under which transmission did not occur, including below the limit of detection (1,000 copies/ml). In the placebo group, there was a highly significant correlation between increased vertical transmission and higher viral load, but the association between viral load and transmission was weaker in the treatment group and did not achieve the same statistical significance. The women receiving AZT experienced a two-thirds reduction in vertical transmission at each viral load level. AZT lowered transmission at all HIV RNA levels even at low viral loads, where the women on placebo already exhibited low transmission rates.
Some factors that had been thought to contribute to transmission were found in ACTG 185 to have no independent effect. Among these were mode of delivery (vaginal delivery versus C-section) and duration of membrane rupture (time to delivery after the "bag of waters" breaks). In ACTG 185 analysis, transmission was associated with baseline and delivery viral loads, with each ten-fold (1 log) increment in viral load linked to a three- to four-fold increase in risk of vertical transmission. None of the women with HIV RNA below 500 copies/ml at delivery transmitted, while 5% of those with levels over 500 copies/ml transmitted. According to Dr. Mofenson, the results suggest that in pregnant women receiving AZT treatment, decreasing viral load to low levels or below the limit of quantification might further reduce transmission.
A meta-analysis presented at the Retrovirus Conference (DG Contopoulos-Ioannidis, poster 238) pooled the results of nine studies with 1,115 mother-infant pairs (696 untreated and 419 treated women). The combined rate of transmission for untreated women was 21.3%. Broken down by viral load, the rates of transmission were: 5% for below 1,000 copies/ml, 15% for 1,000 to 10,000 copies/ml and 37% for over 10,000 copies/ml. The predictive value of viral load was greater in cohorts with higher viral loads than in those with lower levels.
Transmission rates for treated women were lower than for untreated women, except at viral load levels below 1,000 copies/ml, where the rate was the same in both groups (5%). In treated women with viral loads of 1,000 to 10,000 copies/ml, transmission was 7%, compared to 18% in treated women with over 10,000 copies/ml. The investigators concluded that maternal viral load is a strong predictor of the average risk in groups of untreated mothers but a modest to poor predictor of transmission in treated mothers.
It is unclear exactly how or when AZT works to interrupt transmission. In ACTG 076, AZT use decreased the risk of transmission regardless of its effect on maternal viral load. Since AZT only modestly reduced viral load in any case, decreases effected by AZT did not account for the reduced transmission. Dr. Mofenson suggested that the effectiveness of AZT might be due in part to the drug's activity in the infant during labor.
In February, the CDC announced that a study in Thailand of short-course AZT treatment recently demonstrated that the rate of transmission is reduced by 51% when the drug is only administered orally after 36 weeks of pregnancy and continued through delivery but not given to the newborn at all. This finding seems to support the theory that interventions around the time of late gestation and delivery (when the majority of transmissions are thought to occur) are most efficacious. With the establishment of the benefit of an abbreviated 076 regimen, the controversial placebo arms of the vertical transmission trials in developing countries will be dropped and replaced with short-course therapy. The recommendation for women in the U.S. to use the more effective and longer 076 protocol will not change.
As more and more women take antiretroviral drugs to protect their own health, the 076 protocol is becoming increasingly irrelevant as pregnant women come to care providers with extensive prior AZT exposure. The impact that prior AZT use will have on the efficacy of the 076 protocol remains to be seen. In the analysis of risk factors performed for ACTG 185, it was determined that duration of prior AZT use was not associated with vertical transmission. In contrast, a group in France that conducted a study of 665 women, found that AZT treatment of the mother before pregnancy considerably reduced the impact of preventive treatment. The transmission rate was significantly higher among pretreated mothers, 20% versus 5%, possibly due to the emergence of AZT-resistant HIV in pretreated mothers (MJ Mayaux et al. The Journal of Pediatrics. Dec. 1997; 131(6):857-62).
A study presented in January 1997 by Dr. Scott Eastman at the 4th Conference on Retroviruses and Opportunistic Infections (abstract 516) evaluated participants from ACTG 076 to determine the prevalence of AZT resistance at entry and delivery. Neither the baseline prevalence nor the development of resistance were found to be overwhelming. In addition, a mutation at codon 70 of the reverse transcriptase gene (an AZT-induced mutation) in the mother, was not strongly associated with transmission.
Different results were seen in a cohort of 142 women with moderately advanced disease (median CD4 count of 315) from the Women and Infants Transmission Study (WITS). It is important to note that this substudy was conducted from 1992 to 1994, prior to the recommendation that pregnant HIV-positive women follow the 076 protocol. So, while all the participants were taking AZT for their own health, they were not following the three-part 076 protocol and might be expected to have a higher rate of transmission.
Robert Colgrove, M.D., from Beth Israel Deaconess Hospital in Boston, presented study results this year at the 5th Retrovirus Conference (abstract 265). About 25% of women in this group had HIV with at least one AZT-associated resistance mutation. Thirty-four percent had taken AZT prior to pregnancy. Lower maternal CD4 count and higher maternal viral load were associated with having AZT resistance. There was a 20% transmission rate in the group and there were cases of AZT-resistant virus being transmitted to the infants.
A multivariate analysis was done to correct for confounding factors and found a strong correlation between the incidence of AZT-resistance mutations and transmission. When the higher maternal viral loads arising from AZT's lack of effect on resistant HIV was separated out, the resistance to AZT in itself imparted a five-fold increase in the risk of transmission. Other factors independently associated with risk of transmission were total lymphocyte count and duration of ruptured membranes (see also S. Welles. Annual Meeting of the Infectious Disease Society of America, Sept. 1997, abstract 18).
Note that the WITS study found that HIV mutations conferring AZT resistance had a significant impact, whereas ACTG 185 did not find duration of prior AZT use to be independently associated with higher risk of transmission. Seth Welles, M.D., of the University of Minnesota, one of the WITS study's investigators, pointed out that previous AZT use by the mother was not as predictive as the presence of AZT-resistance mutations because not everyone who takes AZT will develop resistance. He stressed that the 076 protocol is highly effective and works well in women who are relatively healthy and largely treatment naïve, but that in women with more advanced disease, resistance will be more of a problem.
The question that needs to be addressed is whether the risk of transmission can safely be reduced by other antiretrovirals, especially in AZT-resistant women. Two posters at the 5th Retrovirus Conference reported on the results of a small phase I ACTG test of ddI use in pregnancy (posters 225 and 226). ACTG 249 examined the effect of ddI in 12 HIV-positive pregnant women receiving the drug from weeks 26 to 36 gestation. ddI was found to cross the placenta, but not as efficiently as AZT. The extent of placental transfer was greater than 80% with AZT, compared to 30% with ddI. ddI was well tolerated during pregnancy. Most importantly, no drug-related toxicities were observed in the mothers, fetuses or infants. This study was not designed to evaluate the effect of ddI on the rate of transmission.
If the rate of vertical transmission is to remain low as increasing numbers of women are exposed to AZT, further studies are required. But because of the potential for damage to the developing fetus (see next), these studies must proceed with caution.