RECOMMENDATIONS FROM THE MEETING
ON MOTHER-TO-INFANT TRANSMISSION OF HIV
BY USE OF ANTIRETROVIRALS,
The number of children infected with HIV is
increasing with the
ever-expanding AIDS pandemic. To date, it is estimated that more
than
1 million children have been infected with
HIV, most of them through
mother-to-infant (MTI) transmission. During
the period 1990-2000, the
World Health Organization (WHO) projects that
as many as 5-10 million
children will be HIV-infected at birth or through
breast-feeding, the
majority of them in sub-Saharan
documentation of a clear reduction of the risk for MTI transmission of
HIV by use of zidovudine
(ZDV) in pregnancy constitutes a major
breakthrough, opening new areas for research and intervention in
this
field.
An interim analysis of a phase III
randomized, placebo-controlled
trial (ACTG 076), conducted in
the
evaluate the efficacy, safety and tolerance of zidovudine (ZDV) for
the prevention of MTI
transmission of HIV has demonstrated a clear
reduction of the risk for MTI
transmission for the group who received
ZDV. The study was conducted by the Pediatric AIDS
Clinical Trials
Group (ACTG) of the
National Institute of Allergy and Infectious
Diseases (NIAID) in
collaboration with the National Institute of Child
Health and Human Development (NICHD) in the
Institut National de la Santé et de
la Recherche Médicale (INSERM) and
the Agence Nationale
de la Recherche sur le SIDA (ANRS) in
Eligible participants were HIV-infected
pregnant women who had
received no antiretroviral treatment during the current
pregnancy, had
no clinical indications for maternal antepartum
ZDV therapy, and had
baseline CD4+ lymphocyte counts greater than 200 cells/mm3.
The ZDV
regimen consisted of antepartum ZDV (100 mg orally 5 times daily)
initiated between 14 and 34 weeks gestation and continued
throughout
the remainder of pregnancy, followed by intrapartum intravenous ZDV
(loading dose 2
mg/kg starting in labour, followed by continuous
infusion of 1 mg/kg/hour until delivery), followed by oral
administration of ZDV (syrup 2 mg/kg every
6 hours for 6 weeks
beginning 8 to 12 hours after birth) to the infant. The primary
study
endpoint, HIV infection of the infant, was defined by one
positive HIV
culture obtained from peripheral blood. Specimens for viral
culture
were obtained from the infants at birth, 12, 24 and 78
weeks
postpartum.
At the time of the interim analysis, 477
women had been enrolled. The
median age was 25 years (range 15 to 43), the median CD4+
lymphocyte
count was 550 cells/mm3 (range 200 to 1,818), and the
median
gestational age at entry was 26 weeks. Maternal demographics
revealed
a predominantly minority population: only 19% were
white non-Hispanic.
The baseline characteristics of the women
were balanced between the
two randomized groups. There were 364 infants with
sufficient data to
be included in the interim efficacy analysis, 180 in the
ZDV group,
184 in the placebo group. As of this analysis, 13 infants in the ZDV
group and 40 in the placebo group were HIV-infected (i.e.
had at least
one positive HIV culture). Based on a Kaplan-Meier
estimate at 18
months, the transmission rate in the placebo group was 25.5%
whereas
the rate in the ZDV group was 8.3%.
This corresponded to a 67.5%
relative reduction in transmission risk. This risk reduction
was
highly statistically significant (two-sided p=0.000056).
Reported maternal and infant side effects
were balanced between the
two randomized groups with the one exception that
hemoglobin levels
were lower for infants in the ZDV
group. The mean decrease in
hemoglobin was less than 1 gram/dl, did not require transfusion,
and
resolved within 12 weeks after completion of ZDV therapy. The study
currently provides no information regarding any late effects of
ZDV in
infants, including those who do not become infected with HIV.
The publication of these results prompted WHO
to convene an
international meeting on the prevention of MTI
transmission of HIV by
use of antiretrovirals,
attended by over 50 scientists and
representatives from research funding agencies, drug regulatory
agencies, and pharmaceutical companies, which took place in
from
the preliminary data on the efficacy of ZDV in preventing MTI
transmission of HIV, to define the current public health
implications
of the ACTG 076 results, to
review potential alternative
antiretroviral drug regimens, more adapted to circumstances in
developing countries, and to propose coordination of the
research
efforts of the various agencies and institutions who plan
intervention
studies in this area. The recommendations made by the group
that
convened in
1. ACTG 076 has
demonstrated that MTI transmission of HIV can be
reduced by use of ZDV. Therefore,
the concept of reducing MTI
transmission of HIV by use of antiretrovirals
has been shown to be
valid.
2. It should be emphasized that the results
of ACTG 076 are only
directly applicable to a specific population. Moreover, the ZDV
regimen employed in the ACTG 076
study has a number of features (cost,
logistical issues, among others) which limit its general
applicability. Therefore, no global recommendations regarding use
of
ZDV to prevent MTI transmission
of HIV can be made.
The study population in ACTG
076 consisted of asymptomatic
HIV-infected pregnant women with CD4+
lymphocyte counts higher than
200 cells/mm3, living in
the
from minority populations (only 19 percent were white
non-Hispanic).
They did not breast feed their children, with
one exception. Although
it is anticipated that the ZDV
regimen used in ACTG 076 would also
reduce MTI transmission when given
to HIV-infected pregnant women with
AIDS-related symptoms and/or low CD4+
lymphocyte counts (< 200
cells/mm3), the magnitude of this effect in
this population with a
presumably higher viral load is unknown. It should also be noted
that
the MTI transmission rate in the
ACTG 076 placebo group (25%) is
higher than the transmission rates usually found in European
populations (15-20%). This may be the result of chance, or may
indicate differences in the study populations which may impair
the
generalization of the results.
In ACTG 076,
pregnant women were available to start ZDV treatment
between 14 and 34 weeks of gestational age. Such a ZDV regimen,
starting early in pregnancy, is not suitable for women who
would not
present at health facilities before delivery.
This regimen is also costly (US$ 1,000 to
1,500 per treatment), and
requires intravenous administration during delivery. Both
features
make the ACTG 076 ZDV regimen difficult to apply in many developing
country situations.
Nevertheless, it is clear that where
availability, cost and logistic
factors are not limiting factors, ZDV
should be offered to
HIV-infected pregnant women for the purpose
of preventing MTI
transmission of HIV.
3. At present, there is no information
regarding potential long-term
ZDV toxicity to infected and uninfected
infants. Neither do we know
the implications of (repeated) ZDV
treatment during pregnancy for the
efficacy of later ZDV treatments of
the mothers. Attempts should be
made to undertake long-term follow-up of mothers who
received
prolonged antiretroviral treatment during pregnancy and of
their
infants, to ascertain whether there are long-term adverse
effects.
In vitro and animal experiments have revealed
potential genotoxic and
carcinogenic effects of ZDV and other
available nucleoside analogue
reverse transcriptase inhibitors. ZDV
induced chromosomal
abnormalities in cultured lymphocytes at 2 µg/ml, and was mutagenic
in
mouse lymphoma cells at 1,000 µg/ml. Moreover, vaginal squamous cell
neoplasms were described in mice and rats who had received ZDV doses
equivalent to 3 and 24 times the ordinary human exposure.
Long-term
follow-up of infants who were exposed in utero
is thus required.
Follow-up of mothers is advised to determine
the implications of
(repeated) ZDV treatment during pregnancy on the likelihood of
emergence of ZDV-resistant viral
strains.
4. Since there is
currently no safety or efficacy information
available for other antiretrovirals
used in pregnancy, their use in
attempting to prevent MTI transmission
of HIV should be limited to
clinical trials.
5. To increase the applicability of antiretrovirals in the reduction
of MTI transmission of HIV, it
is essential to explore simpler and
less costly drug regimens in the full spectrum of
HIV-infected
pregnant women. Such regimens, including interventions
restricted to
the intrapartum period, should
be urgently studied in randomized
controlled trials. Separate studies on safety and
pharmacokinetics may
have to be done in populations from which no such data are
available.
As it is currently estimated that over 50% of
MTI transmission occurs
around delivery, the most cost-effective intervention using
antiretroviral drugs would be an intrapartum
treatment. This treatment
would also have the advantage of being suitable for women
who do not
present at health facilities before delivery, and would
therefore be
better adapted to many "real life" situations,
both in the
industrialized world and in developing countries. Ideally, this
treatment should be cheap, easy to administer, given in one or
a few
doses, safe, and effective. Nevirapine
(NVP), a non-nucleoside reverse
transcriptase inhibitor, and related compounds, have a rapid and
dramatic impact on viral load, and may thus be suitable
candidates for
this type of interventions. Studies are still needed on
safety and
tolerance in pregnant women before efficacy trials can be
initiated.
Randomized controlled trials offer the best
evaluation of new
treatment regimens. The use of historical controls is strongly
discouraged, due to the wide changes with time in the study
population
(differences in the
distribution of disease stages), the circulating
viral strains, the diagnostic tools used for the
ascertainment of the
HIV infection status of infants (antibodies
in earlier cohorts
compared to direct viral markers in more recent cohorts), and
the
treatment practices (increased use of zidovudine
in HIV-infected
pregnant women with AIDS-related symptoms and/or low CD4+
lymphocyte
counts).
In populations where no data on the safety
and the pharmacokinetic
properties of antiretroviral drugs exist, such studies may need
to be
done. There are known variations in drug metabolism and
tolerance
across ethnic groups. In addition, concurrent other diseases
may
decrease tolerance.
6. Since the ZDV
regimen studied in ACTG 076 is not applicable in
those parts of the world where most MTI
transmission of HIV occurs,
placebo-controlled trials offer the best option for obtaining rapid
and scientifically valid results.
Most of MTI
transmission of HIV occurs in the developing world, where
the ZDV regimen used in ACTG 076 is not applicable because of its cost
and operational requirements. In those parts of the
world, the choice
of a placebo for the control group of a randomized trial
would be
appropriate as there is currently no effective alternative for
HIV-infected pregnant
women. In addition, the difference
between the
MTI transmission rates of HIV between two treatment
groups is likely
to be maximized when a placebo is used for the control
group, thereby
decreasing the sample size required to document statistically
significant differences between the two groups.
Placebo-controlled
trials would therefore guarantee the most expeditious
identification
of appropriate interventions.
7. Detailed information regarding the timing
and mechanisms of
transmission should continue to be sought so that optimal
interventions can be developed. Other approaches to prevent MTI
transmission of HIV such as vaginal disinfection, caesarean
section,
and active and passive immunization, should also be
studied.
Although there is increasing evidence for
late transmission of HIV
during pregnancy, it would be important for the further
development of
interventions, and particularly intrapartum
interventions, to know the
exact timing of transmission around delivery.
Vaginal disinfection, if shown to be
effective, would have several
advantages with regard to its applicability. It is a cheap and
probably safe intervention, which may be done in all women at
delivery
regardless of their HIV infection status (no need for HIV
testing),
and which may have beneficial effects in the prevention
of other
infectious diseases transmitted to infants during delivery.
However,
because the anticipated absolute decrease in MTI transmission rate of
HIV due to vaginal disinfection would
presumably be low (less than
5%), the sample size required to document
statistically significant
differences with a non-intervention group would be larger than
2,000
evaluable mother-infant pairs.
The analysis of data from the European
Collaborative Study indicates a
possible halving of MTI transmission
rate of HIV by use of caesarean
section. However, these results should be interpreted with
caution:
there was a large heterogeneity of MTI
transmission rates of HIV
across centers with similar caesarean section rates, and
data from
other large prospective studies do not always confirm these
results.
Randomized trials using caesarean section as
an intervention would be
necessary before any recommendations are made. These trials
would
require large sample sizes (more than 2,000 evaluable mother-infant
pairs if ZDV regimens are
provided to participants), and would require
careful procedures to randomize pregnant women to a mode of
delivery.
In addition, interventions based on the use
of caesarean section to
decrease MTI transmission rate of
HIV may raise several difficulties,
such as possible post-operative complications in
HIV-infected women,
the exposure of the medical personnel to HIV, and the
cost of the
intervention.
8. A particular issue in many affected
populations is the need to
breast feed. Where breast-feeding is recommended
irrespective of HIV
status, studies must be designed to accommodate this policy.
In places where breast-feeding is widespread
and recommended, clinical
trials should enrol breast-feeding
women. Antiretroviral drug
regimens, when given during pregnancy and at delivery, may
prevent HIV
infections which would be later transmitted to the infant
through
breast-feeding. The overall efficacy of the antiretroviral treatment
on MTI transmission of HIV can
therefore be ascertained only after
termination of breast-feeding.
Also, there is a need for trials evaluating
the efficacy of
antiretroviral therapy given in the early postpartum period to the
mother and/or to the child, as there are indications that
this period
is associated with the highest rate of transmission of
HIV through
breast milk.
9. Any study conducted should be part of a
research strategy which may
reasonably be expected to lead to interventions which will be
affordable, feasible and sustainable in the same setting.
10. The study of interventions and, if shown
to be effective, their
implementation, will require HIV testing to be offered to pregnant
women attending for antenatal care. This testing must be
voluntary,
accompanied by pre- and post-test counselling,
and confidentiality
must be assured.
11. Since the
primary endpoint of studies to prevent MTI
transmission
of HIV is the infection status of the infant(s), studies
should be
designed with follow-up sufficiently long for accurate
ascertainment
of this endpoint. In particular, if breast-feeding is practised, the
follow-up may need to extend past discontinuation of
breast-feeding
because of the possibility of late transmission.
12. Every large intervention study should be
monitored by an
independent Data and Safety Monitoring Board.
13. To ensure that the most relevant
scientific questions are
addressed, and to achieve complementarity,
there is a need for world
wide coordination of research activities. WHO should
assume this
coordinating role, as it is in a unique position to do so.
The rapid spread of the HIV pandemic urges
scientists to develop
effective preventive tools. Several research questions still
need to
be answered to identify the most appropriate
antiretroviral drug
regimens for the prevention of MTI
transmission of HIV. Most of these
research questions will be addressed in the form of efficacy
trials
evaluating various antiretroviral drug regimens adapted to
specific
study populations. World wide coordination is necessary to
guarantee
that all pivotal research questions are addressed, and to
avoid
unnecessary duplication of activities. It is recommended that WHO
assumes this coordinating role.
----------------------------------------------------------------------
During the meeting, several protocols for
placebo-controlled trials
evaluating the efficacy of antiretrovirals
for the prevention of MTI
transmission of HIV were reviewed (see Table 1 in annex). As
discussed
earlier, the most promising interventions are the two
protocols using
oral nevirapine started at the
beginning of labour.
Table 1
Experimental treatments suggested for
placebo-controlled clinical
trials depending on the timing of the intervention and the
breast-feeding practices of the study population, with an estimation
of the sample size required.
Artificial
feeding Breast-feeding
Interventions ZDV
PO 4 weeks ZDV PO 4 weeks
starting during the prenatally prenatally
antenatal period
Increased ZDV
during delivery during delivery
ZDV PO 1 week to ZDV PO 1 week
the infant postnatally to the
(optional) mother
and to the
infant
Sample size: 219
* 2 Sample size§: 288 to
603 * 2
Interventions NVP
PO during NVP PO during delivery
starting at delivery
delivery NVP PO 1 week
NVP PO 1 week to postnatally to
the
the infant mother and to the
(optional) infant
Sample size: 468 Sample size§: 458 to
* 2 1,200 * 2
ZDV= zidovudine
NVP= nevirapine
§The smaller sample size is based on the
assumption that
antiretroviral treatment decreases the transmission of HIV through
breast-feeding by 30%. The larger sample size is based on the
assumption that antiretroviral treatment has no effect on the
transmission of HIV through breast-feeding.