One of the most important themes at the International AIDS Conference in Durban was the prevention of mother-to-child transmission of HIV (MTCT). This was very appropriate given that one in four South African women in their peak child-bearing years -- between the ages of 20 and 29 -- are HIV-positive.1 In addition, rates of HIV infection in prenatal clinics in sub-Saharan Africa can run as high as 43%.2 While progress to reduce the rates of MTCT continues to move steadily forward, the conference highlighted the global disparities between developed and developing countries in implementing such interventions.
Rates of transmission in developed countries have dramatically decreased due to widespread efforts to provide pregnant women with access to HIV testing and counseling and information on the benefits of antiretroviral therapy and the risks of breast feeding. In contrast, many developing countries are still struggling to implement small pilot programs and studies that reach only a tiny fraction of HIV-positive pregnant women. Many women in developing countries still have little access to education, testing and counseling and no option but to breast feed, another factor that leads to transmission. While in the US, the emphasis of the Public Health Service Guidelines for the Use of Antiretroviral Agents in Pregnant Women is on the best treatment for the woman's health; in Africa, the focus is solely on preventing MTCT. The only option most African women have is to try and obtain a small amount of antiretroviral medication during pregnancy. While this may interrupt transmission, it does nothing to help the mother combat HIV and there is generally no additional antiretroviral medication available to her after the birth of her baby.
Spreading the Good News
After the results of the pivotal ACTG 076 trial were announced in the mid-1990s, the standard of care for treating pregnant women dramatically changed to incorporate the use of AZT starting between the 14th and 34th week of pregnancy and continuing intravenously during labor and in the newborn for six weeks after delivery. Results of ACTG 076 demonstrated that the AZT regimen could reduce the risk of MTCT by two-thirds. As reported in Durban, widespread implementation of this regimen and other, more potent combination regimens has led to impressive reductions in levels of transmission in industrialized nations.
While the largest proportion of information available is still on the prenatal use of AZT monotherapy, and the longest follow-up has been in AZT-exposed infants, more and more women are using combination therapy during pregnancy. In addition, women are increasingly on treatment prior to conception and are continuing throughout the pregnancy with no interruption during the first trimester. This has raised concerns about possible toxicity of the other antiviral agents. Several studies reported in Durban from the US, Puerto Rico and Europe examined the safety and efficacy of combination therapy.
Carmen Zorilla, MD, from the University of Puerto Rico School of Medicine presented a chart review of 77 women between 1997 and 1999. By 1999, 86% of women were on protease inhibitor-containing regimens. One third of women were on therapy prior to the time of conception and did not discontinue treatment at any time during the pregnancy. The transmission rate was zero in all women and perinatal outcome was good. Most common adverse events in infants were anemia (18%) and candidiasis (13%). The rate of elective cesarean sections increased over time to a peak of 100% in 1998 and 89% in 1999. While C-sections are discussed with each patient, Dr. Zorilla stated, "It is an offer . . . It is becoming more and more accepted that a C-section is not as useful in women with undetectable viral loads on HAART and vaginal delivery is a real option." The American College of Obstetricians and Gynecologists advised in an official statement in May 2000 that data are insufficient to demonstrate a benefit of elective C-sections for pregnant women with viral loads less than 1,000 copies. The rate of preterm deliveries (<37 weeks) was 20%. However, Dr. Zorilla pointed out that 95% of infants were delivered at 35 weeks or more. While this is considered preterm, she stated that at 35 weeks the infants' chances of survival are as good as at full term (40 weeks), as they usually do not need an incubator, have acceptable birth weight and few morbidity risks. Some of the women who delivered prematurely had risk factors for preterm birth, such as intravenous drug use. (Abstract WePpB1303.)
Dr. Thorne from the Institute of Child Health in London presented data on 2,633 women enrolled and prospectively followed in the European Collaborative Study between 1985 and 1999. Whereas in 1994 all women received AZT monotherapy, by 1998/9 24% received double therapy and 41% triple or quadruple therapy. A growing number of women became pregnant while on therapy (12% in 1998/9). These women had been on antiretrovirals for an average of 25 months before pregnancy. Antiretroviral exposure in utero was not associated with prevalence or pattern of congenital abnormalities but was associated with moderate to severe anemia in the infant. The overall rate of MTCT was 5.4% in 1998/9 compared with 16% before 1994. Dr. Thorne is currently working on an updated analysis including infants delivered up to June 2000, which demonstrates that the rate of MTCT since 1998 has dropped even further to 2.1%. (Abstract MoOrC240.) Finally, in a late breaker presentation from the Institute of Human Virology, results were reported from WITS (Women and Infants Transmission Study) on 1,482 women prospectively followed from 1990 to 1999. MTCT rates markedly declined based on potency of the antiretroviral combination: 20.7% for no treatment; 7.7% for AZT monotherapy following the 076 protocol; 3.9% for combination therapy without protease inhibitors and 1.1% for HAART. Transmission rates increased with increasing viral loads at delivery: 0.9% for under 400 copies; 6.4% for 400-3,000 copies; 11.3% for 3,000-40,000 copies; 21.1% for 40,000-100,000 copies and 30.1% for over 100,000 copies. There were two women whose viral loads were less than 400 at delivery who transmitted: one received HAART for one month before delivery but was taking illicit drugs during pregnancy; the other received combination therapy but the duration of membrane rupture (time since breaking of the bag of waters) was greater than 24 hours. (According to previously reported WITS data, the risk of MTCT nearly doubles when the membranes rupture more than four hours prior to delivery.3) (Abstract LbOr4.)
All in all, good progress was reported toward maximal reduction of MTCT. Some conference presenters even referred to the potential "elimination" of pediatric HIV in the developed world. This was largely based on the widespread acceptance of HIV testing in pregnant women and the use of HAART. One CDC study in close to 6,500 perinatally HIV-exposed children born in 1994 to 1998, documented that the percentage of mothers tested for HIV before birth had increased from 80% to 96% (Abstract MoOrC239). Once a woman knows her HIV status, she can make an informed decision about treatment options for herself and her baby. This was reflected in the increasing use of HAART during pregnancy. The use of more potent regimens does not appear to be associated with increased adverse events in mother or child and does appear to be associated with reduced transmission (independent of viral load at delivery). The theme at the conference was how these interventions could be modified for the developing world for women who do not have access to the wide range of drugs available in industrialized nations.
Encouraging Results from MTCT Studies in the Developing World
Results from numerous international studies showed promise in reducing MTCT with simpler and less-expensive regimens than those used in developed countries. Thailand has been especially progressive in designing and implementing prevention programs to address MTCT. At the conference, data were reported on the Perinatal HIV Prevention Trial, a large-scale placebo-controlled trial comparing four treatment regimens using different durations of AZT therapy: (1) long-long arm: AZT starting at 28 weeks for mother, oral AZT during labor and for six weeks in the infant; (2) long-short arm: AZT starting at 28 weeks for mother, oral AZT during labor and for 3 days in the infant; (3) short-long arm: AZT starting at 35 weeks for mother, oral AZT during labor and for 6 weeks in the infant; (4) short-short arm: AZT starting at 35 weeks for mother, oral AZT during labor and for 3 days in the infant.
After an interim analysis, the short-short regimen was found to be significantly less effective than the long-long and was dropped. When further analysis was undertaken to evaluate the other three treatment arms in the 1,437 women enrolled, transmission rates across the regimens were found to be comparable. MTCT was as follows: 6.5% in the long-long arm, 4.7% in the long-short arm and 8.6% in the short-long arm. However, while equivalence between the long-long and long-short arms was established, equivalence with the short-long arm was only border-line. Shortening the maternal treatment period was associated with reduced overall efficacy and higher infection rate at birth (5% for short maternal treatment, 1.8% for long treatment). The authors conclude that both the long-short and long-long regimen appear safe and effective and are simpler and cheaper than the original 076 protocol. While treatment of the infant for six weeks may not add benefit when the mother receives longer prenatal treatment, it may prevent some infections when mothers receive shorter treatment. (Abstract LbOr3.)
While the shorter course Thai regimens cost significantly less than 076, they are still too costly for many parts of the developing world. In response to the need for even less expensive interventions, HIVNET 012 was undertaken. This trial examined the safety and efficacy of very short course AZT versus even shorter course nevirapine in a breastfeeding study population in Uganda. The nevirapine regimen consisted of single dose to the mother at the onset of labor and a single dose to the infant within 72 hours of birth. The AZT regimen consisted of an oral dose administered at the onset of labor and continued through delivery and one week of AZT twice daily to the infant.
Nevirapine has several important characteristics that make it a good candidate for use in preventing MTCT. It is considerably more potent than AZT and able to suppress HIV replication much more effectively and quickly. It is rapidly absorbed when taken orally, entering the bloodstream and beginning to work almost immediately. It has been well established that nevirapine can cross the placental barrier. It also has a long half-life (61 to 66 hours in pregnant women and 45 to 54 hours in babies), meaning that therapeutic levels of the drug persist and continue to work days after the drug is taken.
Preliminary data from HIVNET 012 were released last year (see Treatment Issues, Winter 1999/2000). Final transmission rates and safety data were presented in Durban for 619 women through infant age 12 months. Results were not significantly different to the preliminary analysis. Transmission rates at birth were essentially equivalent in both arms: 8.1% for nevirapine versus 10.3% for AZT. However, after 12 months in this breast feeding population, MTCT rates had increased to 15.7% in the nevirapine arm versus 24.1% in the AZT arm. Nevirapine showed a 39% reduction in MTCT compared to AZT. In addition, the benefits of nevirapine occurred despite breast feeding by the women in the study. The reason for this is unclear. Both regimens appear to be well tolerated. According to the researchers no adverse event was definitely or probably related to the study drugs. (Abstract LBOr1.)
An eagerly anticipated report at the conference was from the South African Intrapartum Nevirapine Trial (SAINT). This large study examining short-course treatments in over 1,300 women was particularly relevant because it was conducted in South Africa. The successful outcome of the trial indicates that the positive results demonstrated in HIVNET 012 and other international trials are reproducible in South African women. SAINT randomized study participants to receive either nevirapine or AZT/3TC. The SAINT nevirapine regimen was slightly different to that used in HIVNET 012 in that the mother received two doses instead of one. Not all women breast fed, 40% opted to bottle feed.
The nevirapine arm consisted of one dose during labor and a second dose 24 to 48 hours after delivery to the mother and a single dose to the infant. The AZT/3TC arm consisted of AZT/3TC at onset of labor and through delivery and then twice daily to the mother and infant for one week. Rates of MTCT between the arms were comparable at birth: 8.2% for nevirapine and 6.8% in the AZT/3TC arm. By week eight, transmission had increased to 14% in the nevirapine arm and 10.8% in the AZT/3TC arm. This difference was not statistically significant. Both regimens appeared relatively safe. There were no serious treatment-related adverse events in either regimen in infants or mothers. Neither liver problems nor significant skin rash were documented in the mothers. Two children experienced skin rashes, but these were not clinically significant. (Abstract TuOrB356.)
Finally, another interesting study from South Africa looked at short courses of antiretroviral agents that have not been as widely investigated as AZT, 3TC and nevirapine. Glenda Gray, MD, of the Chris Hani Baragwanath Hospital in Soweto, presented results of study A1455-094, which was sponsored by Bristol-Myers Squibb to look at the safety and efficacy of its two drugs, ddI and d4T, in comparison to an AZT monotherapy arm. The regimens consisted of ddI versus d4T versus ddI/d4T versus AZT initiated in the mother at 34 to 36 weeks gestation and continued through delivery and for six weeks exclusively in the infants. This study is being conducted in a non-breast feeding cohort, which, in conjunction with the fact that the infants are receiving treatment for a full six weeks, may account for the very favorable outcomes. In the preliminary data analysis of about 200 women at six weeks, there was a significant reduction in transmission to 3.6% in the overall study group. Rates of MTCT were equivalent across all study arms. Maternal and infant safety evaluations demonstrated no significant treatment-associated toxicities. Dr. Gray stated that there was good transfer of the drugs across the placenta. This is encouraging early data and shows that other antiretroviral agents appear safe and effective and give women additional options. However, estimated cost of the regimens are $60 to $100 per mother/infant pair, which is higher than the HIVNET 012 regimen. (Abstract TuOrB355.)
Access to Nevirapine for the Developing World
The cost of the HIVNET 012 nevirapine regimen is only about $4 US per mother/infant pair. This makes it "one of the few deliverable and sustainable strategies for prevention of HIV transmission resource-poor settings," according to the investigators of the study in a 1999 Lancet article.4 In addition, safety data from about 700 mother/infant pairs has shown no important adverse reaction with single dose nevirapine, according to a more recent Lancet article.5 The National Institute of Allergy and Infectious Diseases (NIAID), the sponsor of the HIVNET 012 study, issued a press release last year stating that if the single-dose nevirapine regimen is "implemented widely in developing countries, it potentially could prevent some 300,000 to 400,000 newborns per year from beginning life infected with HIV."6
Immediately before the International AIDS Conference in July, Boehringer Ingelheim, the manufacturer of nevirapine, announced that it would make the drug available free of charge to developing countries for a period of five years for the prevention of MTCT. While this lowers the cost barrier for offering this intervention, other obstacles still remain. Access to health care services is required: such basics as voluntary counseling and testing are not available to all pregnant women in developing countries. In addition, many women who are tested refuse to return for their results. Presumably, the availability of affordable and effective interventions to reduce the risk of MTCT would be an incentive for more women to agree to counseling and testing. A combination of counseling and testing, with the immediate provision of nevirapine, could increase the number of women treated.
However, the South African government seems reluctant to commit to making nevirapine more widely available. The South African Minister of Health has stated, "Nevirapine currently should not be used outside approved research environments," and instead offered only to "expand the research sites on nevirapine to all the provinces so that we can improve our understanding of the operational challenges that would attend any introduction of antiretrovirals to prevent MTCT in the public health system." The South African group TAC is calling for stronger measures from the government to make the drug more widely available.
The Downside of Nevirapine Use
While the efficacy and low cost of short-course nevirapine has generated much excitement throughout the global community, recent data presented at the Durban conference unmasks a drawback associated with its use as described in HIVNET 012: drug resistance. Two studies demonstrated development of nevirapine-resistant mutations in the virus of the mother and the infant after exposure to the single-dose regimen.
J. Brooks Jackson, MD, from the Johns Hopkins Medical Institutes in Baltimore, presented resistance data from HIVNET 012. Genotyping was conducted on 30 treatment-naive study subjects who received a single dose of nevirapine at the onset of labor. This revealed that at six weeks postpartum, seven (23%) of the 30 women had developed nevirapine mutations. The common K103N resistance mutation, which confers cross-resistance to all non-nucleoside analogs (NNRTIs), was found in all seven of the women. This raises the question of whether nevirapine would be effective in preventing MTCT in a second pregnancy. Data were presented demonstrating that the presence of the nevirapine resistance mutations tended to decline in the mothers over time (two-six months). Whether resistant virus would reemerge quickly enough after nevirapine dosing in a second pregnancy to impede efficacy is unclear. The investigators expressed the opinion that the HIVNET 012 regimen could probably be used effectively in subsequent pregnancies. The nevirapine-resistance mutations were also detected in three of the seven infected infants. The clinical significance of the mutations in the infants is as unclear, with no effect on one-year mortality rates. (Abstract LbOr13).
John Sullivan, MD, from the University of Massachusetts Medical School, presented a second study about nevirapine resistance. Dr. Sullivan analyzed genotypes of NNRTI-naive women participating in the US trial PACTG 316, which evaluated the efficacy of a single dose of nevirapine versus placebo to the mother and infant to reduce MTCT in women receiving open-label antiretroviral treatment. Genotyping was conducted on 37 women for resistance mutations. The investigators found that a nevirapine-resistance mutation was found in one woman prior to administration of nevirapine, and three of the 37 women developed nevirapine resistance post-nevirapine administration. It is important to note that this study is still blinded so investigators do not know if the women with the nevirapine mutations were being dosed with the drug or placebo. (Abstract LbOr14.)
These findings are disappointing because they suggest that a single-dose administration of nevirapine may set women and HIV-positive infants at a future disadvantage to not only nevirapine but to the entire class of NNRTIs prior to ever starting antiretroviral therapy. However, in resource-poor settings where NNRTIs are not expected to be available in the near future, the costs and operational advantages of the nevirapine regimen may make it an attractive option. Both Drs. Jackson and Sullivan stressed that the occurrence of the resistance mutations should not be used as an excuse by governments to delay implementation of short-course regimens to prevent MTCT in the developing world.
While single-dose nevirapine may still be advantageous in preventing MTCT in developing countries, it is probably not a good option for women in developed countries who are already on effective treatment. The only case where the benefit might outweigh the risk of resistance to US women would be for an HIV-positive woman whose status was identified late in her pregnancy. In this scenario, according to the Public Health Service (PHS) Guidelines, a women would be offered four options, one of which is a single dose of nevirapine at the onset of labor followed by a single dose for the newborn at age 48 hours. (The other three options consist of short-course AZT/3TC, AZT monotherapy or AZT/nevirapine.) Women would be evaluated after delivery to determine if antiretroviral treatment would be indicated for their own health.
Lynne Mofenson, MD, of the National Institutes of Health, suggested some theoretical ways that one might reduce the likelihood of the development of resistance. "The hypothesis is that if one kept viral replication very low in the mother during the period that nevirapine drug levels decrease and become negligible (probably about one week or so) than resistance shouldn't develop. Adding Combivir (AZT/3TC) for one week to the mother (and then stopped) is one possible choice because of ease of administration. Another choice would be the initiation of HAART very soon after delivery. This should inhibit viral replication and hence development of resistance as well, if not better, than my hypothetical Combivir suggestion. HAART would continue in the mother for therapeutic purposes. My suggestion of one week of Combivir (in conjunction with single dose nevirapine) is more relevant to developing countries where HAART is not available. In actuality in the US, for most women, it would be recommended that the mother initiate HAART postpartum and continue it for her own treatment."
About the "M" in MTCT and Breast Feeding
The whole issue of short-course regimens that provide antiretroviral medication to the mother only during pregnancy generated a great deal of controversy at the Durban conference. The keynote speaker of one session paraphrased this concern with his question "Where is the 'M' in MTCT (mother-to-child transmission)?" A short-course treatment is basically an intervention that uses the woman's body to deliver treatment to the fetus. It is of no benefit to the woman and as stated above, may increase viral resistance. This could actually make things worse for the woman by limiting future treatment options. In addition, it was noted that only providing MTCT interventions without continued treatment for the mother will increase the number of AIDS orphans who frequently become street children. With the number of AIDS orphans worldwide at over 10 million, it seems in the best interests of the infants and the communities to address the mother's health needs. (Session Sy03.)
Dr. Gray, an investigator in many of the South African studies, posed several questions. Is it ethical to implement MTCT interventions without treating the mother? Is it ethical not to offer voluntary testing and counseling if there are no treatment options? Is it ethical not to inform pregnant women they are at risk for HIV? She advocated that women have a right to information and education, no matter what their circumstances are. And while there are ethical dilemmas to overcome, there are also viable interventions to prevent MTCT right now. She concluded by stating, "let us not find another reason not to prevent MTCT." (Session Sy03.)
One such dilemma relates to breast feeding. While the short-course interventions may prevent MTCT at birth, in most cases the protective benefit seems to fade with time. This effect is most marked in the PETRA study. PETRA is a very large multi-site African study evaluating several different short-course AZT/3TC regimens (see Treatment Issues, February 1999). The majority of mothers in the study breast fed (70%). PETRA basically demonstrated that the two-drug combination initiated at 36 weeks, during delivery and for one week postpartum for mother and infant reduced transmission by 50% at six weeks to 9.2% when compared to placebo. However, more recent data from Durban indicate that this benefit is essentially lost during breast feeding. At month 18, there was no statistically significant difference between any of the treatment arms when compared to placebo arm: rates of MTCT had risen to 20% to 25% in the treatment arms compared to 26% in the placebo arm. (Abstract LbOr5.)
The high cost of formula, social stigma associated with not breast feeding and lack of sanitary conditions make formula feeding an infeasible option for many women in the developing world. A recent study done by Dr. Anna Coutsoudis has for this reason looked at exclusive breast feeding as a means to reduce post-delivery MTCT. In exclusive breast feeding, mother's milk is the only intake the child has for the first six months of life. In mixed feeding, mother's milk is supplemented by juice, water, and solids. In her study, Dr. Coutsoudis compared rates of MTCT among 551 HIV-positive women who were divided into exclusive breast feeding and non-breast feeding arms. Her study found that exclusive breast feeding for three months compared to mixed breast feeding lowers risk of MTCT by 44%. Dr. Coutsoudis said that "mixed" breast feeding had the potential to introduce contaminants, which could irritate the babies' stomach lining and allow easier access of the virus through the babies' gut. (Abstract LbOr6 and Session DB05.)
While Coutsoudis' results are culturally appealing for many women, there is also contradictory evidence that suggests that exclusive breast feeding may not have a protective effect. Exclusive breast feeding needs further evaluation before it can be identified and promoted as a means to reduce MTCT in the developing world. Additional technologies are being developed to pasteurize expressed milk (LbPp122) and to add alkyl sulfides to inactivate HIV in expressed milk (LbPp123), which may offer other options. Breast feeding of any kind is not recommended for women in developed countries who have access to formula.
In Africa the rate of MTCT is estimated at 21-43%. It is irrefutable that many of the short-term treatments studied can reduce these rates significantly. Widespread implementation of these regimens would have a huge impact on the hundreds of thousands of HIV-positive infants born in the developing world each year. The South Africa government and others have been presented with compelling evidence to act. While the breast feeding debate will continue and the emergence of drug resistance requires further research, there was a resounding consensus at the conference that these uncertainties should not interfere with the implementation of comprehensive MTCT-prevention programs. It should also be clear that preventing MTCT is only the first step and progress must be made towards providing care for the mother herself, so that the M is no longer missing from MTCT. As demonstrated by studies from industrialized nations, which show rates MTCT dropping lower and lower towards zero, what is best for the mother's health is best for the baby.
1. UNAIDS Fact Sheet HIV/AIDS in Africa. www.unaids.org/fact_sheets/files/Africa_Eng.html
2. UNAIDS Report on the Global HIV/AIDS Epidemic, June 2000, p. 126
3. Landesman S et al. The New England Journal of Medicine. June 20, 1996; 334(25):1617-23
4. Guay LA et al. The Lancet. September 4, 1999; 354(9181):795-802
5. Wood E et al. The Lancet. June 17, 2000; 355(9221): 2095-100
6. Department of Health and Human Services, July 14, 1999. http://www.niaid.nih.gov/cgi-shl/simple/release.cfm
Back to the GMHC Treatment Issues September/October 2000 contents page.