• Aucun résultat trouvé

regimen of oral zidovudine to reduce vertical transmission of HIV in breastfed children in Côte d’Ivoire and Burkina Faso:

N/A
N/A
Protected

Academic year: 2021

Partager "regimen of oral zidovudine to reduce vertical transmission of HIV in breastfed children in Côte d’Ivoire and Burkina Faso:"

Copied!
7
0
0

Texte intégral

(1)

in the infant at 6 months was 18·0% in the zidovudine group (n=192) and 27·5% in the placebo group (n=197;

relative efficacy 0·38 [95% CI 0·05–0·60]; p=0·027).

Adjustment for centre, period of recruitment, mode of delivery, maternal CD4-cell count, duration of labour, prolonged rupture of membranes, and duration of breastfeeding did not change the treatment effect. The proportions of women taking more than 80% of the planned maximum dose were 75% before delivery, 81%

during labour, and 83% post partum, without statistical difference between the groups. No major adverse biological or clinical event was reported in excess among women and children of the zidovudine group.

Interpretation A short course of oral zidovudine given during the peripartum period is well accepted and well tolerated, and provides a 38% reduction in early vertical transmission of HIV-1 infection despite breastfeeding.

L a n c e t1999; 353: 7 8 6 – 9 2 See Commentary page ???

I n t r o d u c t i o n

Mother-to-child or vertical transmission of HIV leads to 1500 new paediatric HIV infections each day in Africa.1 Most observational studies estimate the risk of HIV vertical transmission to be 20–35% in developing countries in the absence of intervention.2 P o s t n a t a l transmission through breastfeeding is thought to account for most of the excess risk observed in Africa compared with more developed countries.3

The efficacy of zidovudine in reducing HIV vertical transmission by two-thirds was shown by the multicentre AIDS Clinical Trials Group (ACTG) 076/Agence Nationale de Recherches sur le SIDA (ANRS) 024 clinical trial in the USA and France.4T h i s result, obtained in the absence of breastfeeding, has led to public-health recommendations and to increasing use of zidovudine monotherapy in pregnant women. As a consequence, vertical transmission rates now do not exceed 5% in this context.5 – 7

The results of that trial are not directly applicable in developing-country settings, for the following reasons:8 the unknown efficacy of zidovudine monotherapy in pregnant women if breastfeeding cannot be discontinued; logistic difficulties, including the need for early prenatal visits; intravenous infusion during labour;

and cost. Simpler and more affordable regimens of S u m m a r y

Background Zidovudine reduces the rate of vertical transmission of HIV in non-breastfed populations. We assessed the acceptability, tolerance, and 6-month efficacy of a short regimen of oral zidovudine in African populations practising breastfeeding.

Methods A randomised double-blind placebo-controlled trial was carried out in public clinics of Abidjan, Côte d’Ivoire, and Bobo-Dioulasso, Burkina Faso. Eligible participants were women aged 18 years or older, who had confirmed HIV-1 infection and pregnancy of 36–38 weeks duration, and who gave written informed consent.

Exclusion criteria were severe anaemia, neutropenia, abnormal liver function, and sickle-cell disease. Women were randomly assigned zidovudine (n=214; 300 mg twice daily until labour, 600 mg at beginning of labour, and 300 mg twice daily for 7 days post partum) or matching placebo (n=217). The primary outcome was the diagnosis of HIV-1 infection in the infant on the basis of sequential DNA PCR tests at days 1–8, 45, 90, and 180.

We compared the probability of infection at a given age in the two groups. Analyses were by intention to treat.

Findings Women were enrolled between September, 1995, and February, 1998, when enrolment to the placebo group was stopped. Analysis was based on 421 women and 400 lifeborn infants. Baseline demographic, clinical, and laboratory characteristics were similar in the two groups. The Kaplan-Meier probability of HIV i n f e c t i o n

6-month eff i c a c y, tolerance, and acceptability of a short

regimen of oral zidovudine to reduce vertical transmission of HIV in breastfed children in Côte d’Ivoire and Burkina Faso:

a double-blind placebo-controlled multicentre trial

François Dabis, Philippe Msellati, Nicolas Meda, Christiane Welffens-Ekra, Bruno You, Olivier Manigart, Valériane Leroy, Arlette Simonon, Michel Cartoux, Patrice Combe, Amadou Ouangré, Rosa Ramon,

O d e t t e K y - Z e r b o , Crépin Montcho, Roger Salamon, Christine Rouzioux, Philippe Van de Perre, Laurent Mandelbrot, for the DITRAME Study Group*

*Other members listed at end of paper

Unité INSERM 330, Université Victor Segalen Bordeaux 2, Bordeaux, France (Prof F Dabis MD, V Leroy MD, R Salamon MD);

ORSTOM Petit Bassam, Abidjan (P Msellati MD); Centre Muraz, OCCGE, Bobo-Dioulasso, Burkina Faso (N Meda MD,

O Manigart MSc, A Simonon MSc, M Cartoux MD, A Ouangré PharmD, O Ky-Zerbo MPH, P Van de Perre MD); Centre Hospitalier

Universitaire de Yopougon, Abidjan, Côte d’Ivoire

(C Welffens-EkraMD); CeDReS, Programme PAC-CI, Abidjan (B YouPhD, P Combe PharmD, R Ramon MD, C Montcho MSc);

Laboratoire de Virologie, Centre Hospitalier, Universitaire Necker–Enfants Malades, Paris (C Rouzioux PharmD); and Maternité Port-Royal, Centre Hospitalier Universitaire Cochin Port-Royal, Paris, France (L Mandelbrot MD)

Correspondence to:Prof François Dabis, Unité INSERM 330, Université Victor Segalen Bordeaux 2, 33076 Bordeaux Cedex, France

(e-mail: Francois.dabis@dim.u-bordeaux2.fr)

(2)

zidovudine started to be evaluated in 1995 in several developing countries.9The release in February, 1998, of the initial results of the zidovudine trial in Bangkok, Thailand, by the US Centers for Disease Control and Prevention was a first step in this respect.1 0 T h e reduction by half of the vertical transmission rate in the absence of breastfeeding had two immediate consequences: a worldwide policy of zidovudine prophylaxis was adopted1 1and the placebo groups were discontinued in all of the trials under way in Africa under various sponsorships.1 2 We report the initial results of the DITRAME ANRS 049a trial carried out in Abidjan, Côte d’Ivoire, and Bobo-Dioulasso, Burkina Faso, to evaluate a short course of oral zidovudine among breastfeeding West African women.

M e t h o d s Study population

The trial was carried out in two large cities with good access to and coverage of antenatal care services. In Abidjan, the economic capital of Côte d’Ivoire (3 million inhabitants), the Yopougon district is the most densely populated, with 500 0 0 0 inhabitants, a university hospital, and a network of public clinics offering prenatal, obstetric, and paediatric care. Bobo- Dioulasso has 400 000 inhabitants and is the second largest city in Burkina Faso. It has a public hospital and neighbourhood prenatal clinics. 1995 HIV seroprevalence estimates in pregnant women are available through unlinked serosurveys and range from 7·5% in Bobo-Dioulasso1 3 to 14·2% in A b i d j a n .1 4Vertical transmission rates were estimated at 24·7%

for HIV-1 and 1·2% for HIV-2 in a cohort study in Abidjan in 1 9 9 2 .1 5

The choice of the study protocol was partly based on the results of a feasibility survey carried out in this population in 1 9 9 5 .1 6 The study was a randomised double-blind placebo- controlled clinical trial. It started as a phase II tolerance trial and was extended as a phase III efficacy study, including the phase II observations with extended follow-up. The study protocol was approved by the Ethical Committee of the National AIDS Control Programme in Côte d’Ivoire, the National AIDS Control Committee in Burkina Faso, and in France by the Ethical Committee of the Bordeaux University Hospital, and the Institutional Review Board of the ANRS.

All pregnant women attending the selected prenatal clinics were systematically offered pretest counselling and HIV testing if they were at least 18 years old; presented before 32 weeks of gestation according to the best possible method of estimation, ultrasound examination, history of last menstrual period, or uterine height at clinical examination; lived within the city limits; and planned to give birth in one of the study clinics.

A blood sample was obtained for the detection of HIV infection from each woman who signed an informed consent form. All serum samples were screened for HIV-1 and HIV-2 antibodies by a commercial ELISA (Genelavia Mixt, Diagnostics Pasteur, Paris, France or Murex ICE 1-O-2, Murex Biotech Ltd, Dartford, UK). Confirmation and discrimination were obtained by use of a commercial synthetic peptide ELISA (Peptilav 1-2, Diagnostics Pasteur). Results were given to women who returned for post-test counselling, generally within 2 weeks. A second sample was then obtained from each woman who reacted to the first HIV test to verify her HIV status with a rapid discriminant test (Multispot, Diagnostics Pasteur).

Women who tested positive for HIV-1 or whose serological reaction was dually reactive for HIV-1 and HIV-2 were systematically asked to enter the trial. After informed consent had been given, blood was collected for haemoglobin electrophoresis, complete blood count, and measurement of liver function. Reasons for exclusion were sickle-cell markers SS, CC, or S C haemoglobin, haemoglobinaemia below 7 g/dL,

absolute neutrophil count below 0·7531 09/L, and alanine and aspartate aminotransferases above 2·5 times the standard value for the laboratory. Women who met all the inclusion criteria and accepted the study principles were scheduled for enrolment between 36 and 38 weeks of gestation.

Study methods

The oral regimen of zidovudine or matching placebo was prepartum treatment twice daily, a single oral loading dose at the beginning of labour, and a 7-day course of postpartum treatment twice daily. Zidovudine formulation was in 250 mg tablets for phase II, replaced by 300 mg tablets for phase III because the 250 mg formulation was no longer available from the manufacturer. Thus, the prepartum and postpartum treatment regimens were either 500 mg or 600 mg per day, and the labour dose was 500 mg or 600 mg once. The intrapartum treatment package was given to each woman at inclusion.

Intrapartum treatment was self-initiated by the woman as soon as labour started, either at home or in the delivery room. No treatment was given to the newborn infant.

Sequentially numbered sealed packages were prepared by an independent central pharmacy according to the randomisation list drawn by the independent statistician with SAS version 6.11. Block randomisation (in blocks of ten) was used with stratification by centre, to assign eligible women to one of the study regimens. The study drugs were stored in prepackaged envelopes containing 1 week of treatment with written i n s t r u c t i o n s .

Demographic and clinical information, and CD4- lymphocyte count were obtained at entry to the study. Women were given a weekly prepartum appointment for interview, clinical follow-up, and drug distribution, to allow documentation of compliance and tolerance. Women were asked to come to the study clinic to give birth and be admitted to hospital for at least 2 4 h. Delivery characteristics were recorded. Postpartum treatment was initiated, with an appointment 1 week later at the end of the treatment period, then at days 45, 90, and 180. If delivery occurred anywhere other than the study clinic, postpartum treatment and follow- up were initiated as soon as possible. Clinical follow-up of the child followed the same schedule as for the mother. Feeding practices were reported at each visit.

Biological follow-up of the women and children included complete blood count at days 8 and 45 post partum, and assessment of liver function at the same dates during phase II.

Children’s blood samples were also obtained for diagnosis of HIV infection. Maternal plasma and breastmilk samples were routinely collected and tested for HIV RNA in the first consecutive 164 mothers. Results of this subsample will be published elsewhere.

Medical and psychosocial support was also available from study teams between scheduled visits. All prescribed medications, hospital stays, and transportation were offered free of charge during the study period. Long-term follow-up, scheduled for 2 years, is still under way.

Capillary blood was collected in microtainer tubes containing edetic acid from neonates between days 1 and 8.

Peripheral blood (1–2 mL) was taken at days 45, 90, and 180.

Peripheral-blood mononuclear cells were separated by gradient centrifugation. A nested PCR assay was done with three pairs of HIV-1-specific oligonucleotide primers (Eurogentec, Herstal, Belgium) according to previously described t e c h n i q u e s .1 7 The amplified sequences were located in g a g (GAG881–882/SK38–39) and p o l (H1POL42–45/43–44 and POL001–004/002–003) regions of the HIV-1 genome. All tests were done in duplicate, with positive and negative controls, and distilled water as an internal control. Validation of nested PCR amplification used calibrated DNA from 8E5 cells as positive controls with a detection level of five copies of proviral DNA per 106cells. In samples negative by the two p o l p r i m e r pairs, DNA quality was verified by PCR with HLA-DQa primers. PCR was classified as positive when a positive signal was obtained for at least two of the three primer pairs. All

(3)

samples collected at day 180, or earlier when not available, were systematically processed for PCR, which was then applied to all the preceding samples for a given child if the first sample tested gave a positive result. In Bobo-Dioulasso, samples were also subsequently analysed by commercial quantitative HIV plasma RNA PCR test (Amplicor HIV Monitor, version 1.5, Roche Diagnostics Systems, Branchburg, NJ, USA), which gave concordant results.

According to international standards,1 8the diagnosis of HIV- 1 infection in the child was defined on the basis of one positive PCR. The first positive test in the series allowed estimation of the timing of infection.1 8Absence of infection was defined by a negative PCR on the last available sample or by a negative serological result with the techniques reported for maternal H I V screening and applied to samples collected every 3 months between 9 and 18 months of age.1 9 Children who had no

sample available for PCR and could not be followed up beyond 6 months of age were classified as having unknown HIV s t a t u s .

Maternal severe anaemia was defined by haemoglobin concentration below 7 g/dL at days 8 or 45 post partum and neutropenia by a neutrophil count below 0·531 09/L. Infant anaemia was defined by haemoglobinaemia below 10 g/dL at days 1–8 or below 8 g/dL at day 45, and neutropenia by a neutrophil count below 0·531 09/L at one of these two measures. All maternal (¶42 days post partum) and infant deaths were taken into account.

Estimation of compliance with treatment was based on weekly interviews and pill count. Good compliance was defined by a treatment including above 80% of the number of pills that should have been taken based on its duration. Because zidovudine increases the mean corpuscular volume even after short periods of treatment, we used this variable measured at each maternal haematological investigation to assess compliance indirectly.

Statistical analysis

The sample size (n=780) was calculated to detect a 40%

reduction in the vertical transmission rate at 6 months of age from a baseline rate of 25%, with a type I error of 5% (two- sided test), power of 90%, and 20% loss to follow-up.2 0 W e chose the intention-to-treat approach for analysis, keeping all available records. One interim analysis was planned with 500 observations. Because enrolment was stopped in February, 1998, the study was terminated after review of the randomised observations by the independent data and safety monitoring board.

Group comparisons used Student’s t test or the non- parametric Mann-Whitney test for quantitative variables and x2 test or Fisher’s exact test for qualitative variables when appropriate. The probabilities of infection at a given age, continued breastfeeding at a given age, and infant death were estimated by the Kaplan-Meier survival technique, and comparisons used the log-rank test. Cox multivariate proportional hazards model was used to study zidovudine relative efficacy. Point estimates of summary statistics are reported with their 95% CI. To guarantee statistical independence of observations, we randomly selected only one infant from each multiple pregnancy.

R e s u l t s Study population

Details of the HIV screening programme and the difficulties encountered with acceptability of HIV s c r e e n i n g2 1and uptake in the trial2 2have been reported elsewhere and are summarised in figure 1. Between September, 1995, and February, 1998, 431 women were enrolled in the trial, 164 during the phase II inclusion period (September, 1995, to October, 1996) and 267 during phase III (February, 1997, to February, 1998). Ten women were excluded from analysis b e c a u s e their serological diagnosis of HIV i n f e c t i o n could not be confirmed in time. This report is based on 421 maternal records, 248 from the Abidjan centre and 173 from the Bobo-Dioulasso centre (figure 1, table 1).

Among the 421 pregnant women, 406 livebirths and eight stillbirths were recorded. There were seven p a i r s of twins, one with a stillbirth (zidovudine g r o u p ) , and six with two livebirths each (figure 1, t a b l e 1). In each case one liveborn twin was included in the analysis.

Five women were dually reactive to HIV-1 and HIV-2.

The zidovudine and placebo groups showed similar distributions and values for all baseline maternal variables (table 1). 76% of pregnant women were aged below 30 years, and 24·5% were primiparous.

T h e median number of pregnancies was three, and Figure 1: Trial profile

(4)

1 8 % of women were nulliparous. Two women were at clinical AIDS stage at enrolment. The median CD4-cell count was 545/mL; the CD4-cell count was b e l o w 2 0 0 /mL in 8·3% of women and above 500/mL in 56·9%. Haemoglobin was of the AA type in 82% of women; the remainder had either AC or AS or sickle-cell trait.

14 (3·3%) women were lost to follow-up before delivery (figure 1, table 1). The median duration of the prepartum treatment was 21 days in both groups (p=0·72; table 2). 51 women (21 zidovudine group, 30 placebo group) had no prepartum treatment. Treatment duration was 7 days or less in 23·1% of women. Overall, 59·7% of the 407 women followed up until labour gave birth in one of the study clinics, 28·3% in another c l i n i c , and 10·1% gave birth at home, without difference between the groups (table 1). Median duration of labour was 8 h in both groups (p=0·70). Prolonged rupture of membranes (>4 h) was reported in 27·4%

o f the 354 women on whom records were available.

Caesarean section was done in ten cases (for emergency indications in six). Intrapartum treatment was reported in 81·3% of women who gave birth, without difference between the groups (p=0·91; table 2).

The median duration of postpartum maternal treatment was 7 days in both groups (table 2).

1 1 women in the zidovudine group and 16 in the placebo group did not receive any postpartum treatment. 11 children in the zidovudine group and four in the placebo group were bottlefed from birth (p=0·07). The probability of continued breastfeeding at 3 months was 85·9% in the zidovudine group and 89·8% in the placebo group; the probability had fallen to 74·0% and 78·4%, respectively, at 6 months ( p = 0 · 2 7 ) .

Adherence to treatment

Cumulative treatment was calculated and compared with the maximum number of pills to be taken according to duration (table 2). 78·1% of the zidovudine group and 75·4% of the placebo group (p=0·52) took more than 80% of the planned prepartum treatment, and 86% of women took at least 50% of cumulative treatment overall. Intrapartum treatment was either not given or given properly, with an overall use in 81·3% of women. 85·0% of the zidovudine group and 81·5% of the placebo group (p=0·35) took more than 80% of planned postpartum treatment. Maternal mean corpuscular volume (table 2) increased by 4 · 0 mL over the treatment period in the zidovudine group and decreased by 0·15 mL in the placebo group (p<0·0001 for the comparison of changes between g r o u p s ) .

T o l e r a n c e

There was no interruption of treatment other than in women lost to follow-up. Two women in the placebo group and one in the zidovudine group died in the first

Zidovudine Placebo p

Pregnant women randomised

Total 209 212

Median (range) age in years 25 (18–42) 24 (18–43) 0·61*

Median (range) parity 1 (0–9) 1 (0–8) 0·85*

36–38 weeks of gestation at enrolment 202 (96·7%) 197 (92·9%) 0·09†

Duration of gestation at enrolment (weeks) 32–38 34–38

Median (range) haemoglobin (g/dL)§ 9·6 (7·0–14·2) 9·5 (6·7–12·3) 0·23*

Median (range) CD4-cell count (/mL)§ 568 (11–1480) 535 (14–1870) 0·70*

<200 CD4 cells/mL 17 (8·3%) 17 (8·2%) 0·96†

AIDS clinical stage§ 1 (0·5%) 1 (0·5%) 1·0‡

Women who gave birth

Total 201 206 0·57†

At home 21/201 (10·4%) 20/206 (9·7%) 0·80†

Prolonged rupture of membranes (>4 h) 44/171 (25·7%) 53/183 (29·0%) 0·50†

Caesarean section 6/201 (3·0%) 4/206 (1·9%) 0·50‡

Pregnancy outcomes

Total 203 211

Stillbirths 1 7 0·07‡

Livebirths 202 204

Number of pairs of twins 2 4

Median (range) postpartum maternal 185 (0–185) 185 (0–185) 0·14*

follow-up (days)

*Mann-Whitney test. †x2test. ‡Fisher’s exact test. §At baseline.

Table 1: Baseline and follow-up characteristics of participants

Zidovudine Placebo p

Pregnant women assessed*

Total 209 212

Median (range) duration of prepartum 21 (2–73) 21 (1–62) 0·72†

treatment (days)

>80% of prepartum treatment pills taken 163 (78·1%) 160 (75·4%) 0·52‡

Pregnant women followed up until labour

Total 201 206 0·57‡

Two intrapartum treatment pills taken 163 (81·1%) 168 (81·6%) 0·91‡

Women who gave birth to a live baby

Total 200 200 0·12‡

Median (range) duration of postpartum 7 (0–7) 7 (0–7) 0·57†

maternal treatment (days)

>80% of prepartum treatment pills taken 170 (85·0%) 163 (81·5%) 0·35‡

Mean (SD) maternal mean corpuscular volume (mL)

At inclusion 85·9 (7·5) 85·6 (8·2) 0·71§

At day 8 post partum 90·1 (7·7) 85·3 (7·7) <0·0001§

*Five women were excluded from analysis after randomisation in each group. Eight women in zidovudine group and six in placebo group were lost to follow-up immediately after enrolment and are excluded from this analysis. †Mann-Whitney test. ‡x2test. §Student’s ttest.

Table 2: Adherence to treatment

Zidovudine Placebo p Women who gave birth

Total 201 206

Women with severe anaemia (day 8 or 45) 11 8 0·47*

Women with severe neutropenia (day 8 or 45) 1 1 0·99†

Maternal deaths (by 42 days) 1 2 1·0†

Children

Total followed up >1 day 200 200

Children with severe anaemia 10 12 0·67*

Children with severe neutropenia 1 4 0·37†

Infant deaths

Total 17 26 0·16‡

Neonatal deaths 0–28 days 8 4 0·04†§

Deaths 29–180 days 9 22

*x2test. †Fisher’s exact test. ‡Log-rank test of probabilities of death at 6 months.

§Comparison of proportion of neonatal deaths by treatment group among all infant deaths. Severe anaemia was defined for women as haemoglobin <7 g/dL and for infants <10 g/dL on days 1–8 or <8 g/dL on day 45. Severe neutropenia was defined as counts of <0·53109/L for all participants.

Table 3: Biological tolerance to treatment and vital status of women and children post partum

Zidovudine Placebo p

(n=200) (n=200)

Mean (SD) birthweight (g) 2915 (432) 2860 (417) 0·21*

Low birthweight (<2500 g) 28/194 (14·4%) 31/190 (16·3%) 0·61†

Congenital malformations 5 (2·5%) 1 (0·5%) 0·22‡

Median (range) gestational age 39·5 (34–42) 39 (33–42) 0·11§

(weeks)

Prematurity score of Finnström 7/168 (4·2%) 15/174 (8·6%) 0·09†

<37 weeks

Intrauterine growth retardation|| 12/168 (13·1%) 22/174 (12·6%) 0·90†

*Student’s ttest. †x2test. ‡Fisher’s exact test. §Mann-Whitney test. ||Low birthweight in an infant of >37 weeks gestational age.

Table 4: Neonatal characteristics

(5)

6 weeks post partum. Two women, one in each group, died with HIV-related symptoms after 6 weeks post partum. Two women in the zidovudine group and seven in the placebo group reported severe signs and symptoms during the treatment period, including fever, diarrhoea, nausea, or muscular pain. Severe but transient anaemia was diagnosed post partum in 5·6%

of women in the zidovudine group and 4·0% in the placebo group (p=0·47; table 3). One woman in each group developed severe but transient neutropenia.

The two groups showed little, if any, difference in anthropometry of neonates or adverse perinatal outcomes (table 4). 17 infant deaths were reported in the zidovudine group and 26 in the placebo group, with a higher proportion of neonatal deaths in the zidovudine group (table 3). The probability of death was 4·0% at 4 weeks and 8·8% at 6 months in the zidovudine group, versus 2·0% and 13·6%, respectively, in the placebo group (p=0·16). Ten (5·2%) neonates in the zidovudine group and 12 (6·2%) in the placebo group developed severe anaemia, but this abnormality was corrected by day 45. Severe neutropenia was less common (table 3), and again was corrected by day 45.

E f f i c a c y

Efficacy analysis is based on follow-up of 400 liveborn children. Eight children in the zidovudine group and three in the placebo group were of unknown HIV infection status because they died soon after birth and no blood sample was available (overall loss to follow-up 2·7%). HIV infection was found in 33 children born to 180 women in the zidovudine group and in 52 children born to 175 women in the placebo group.

The probability of being infected during the first 6 months of life was significantly higher in the children of placebo-group mothers than in the children of zidovudine-group mothers (log-rank test p=0·027; table

5, figure 2). The efficacy of zidovudine was thus estimated at 38% (95% CI 5–60) in reducing vertical transmission of HIV-1 .

In Cox multivariate analysis that took account of the variables listed in table 6, zidovudine reduced vertical transmission by 42% (5–65; p=0·028). Prolonged rupture of membranes multiplied the risk of transmission by almost 2, independently of the treatment effect, and low maternal CD4-cell count also increased the transmission rate independently of treatment effect. Duration of breastfeeding did not influence the treatment effect.

D i s c u s s i o n

We report the efficacy of a short regimen of oral zidovudine in reducing vertical transmission of HIV-1 in a breastfed population, with about 40% reduction in the rate of transmission estimated at 6 months of age. The demonstrated efficacy of this zidovudine regimen was obtained with only a short course of treatment in a population with generally mild HIV-1-related immunodeficiency. Treatment was well accepted and well tolerated.

Several factors strengthen these conclusions. The t w o randomised groups were similar in terms of a l l relevant characteristics at baseline. Losses to follow- up did not exceed 6% (women and children combined) and data were generally available for more than 95% of the records. The estimate of the vertical transmission rate in the placebo group is compatible with those reported elsewhere,2 including for Côte d’Ivoire.1 5 Finally, our findings are consistent with other trial results showing higher efficacy in non-breastfed p o p u l a t i o n s4 , 2 3 and a similar study in another breastfed population in Côte d’Ivoire.2 4 The reduction in transmission rate is close to the assumption made for the sample-size calculation.

Probabilities of infection over time in the placebo group were consistent with previous estimates in R w a n d a2 5 and the Democratic Republic of Congo (former Zaïre).2 6Maximum efficacy of zidovudine was

Hazard ratio (95% CI) p

Zidovudine vsplacebo 0·58 (0·35–0·95)* 0·028

Prolonged rupture of membranes (>4 h) 1·93 (1·15–3·22) 0·015 Maternal CD4-cell count at inclusion† 1·20 (1·09–1·32) 0·0004

Duration of labour‡ 0·99 (0·96–1·02) 0·492

Duration of breastfeeding§ 1·07 (0·58–1·96) 0·818

Abidjan vsBobo-Dioulasso 1·13 (0·69–1·84) 0·634

Phase II vsphase III 0·96 (0·57–1·61) 0·878

Caesarean section 1·43 (0·19–10·73) 0·741

*Corresponding to a relative efficacy of 0·42 (95% CI 0·05–0·65). †For a decrease of 100/mL. ‡For an increase of 1 h. §For an increase of 1 month of the duration of breastfeeding. Zidovudine group: 25 infected/149. Placebo group: 43 infected/162.

Table 6: Determinants of overall HIV-1 vertical transmission according to treatment group

Figure 2: Probability of HIV-1 infection in infants according to treatment group during the first 6 months of life

Kaplan-Meier analysis of zidovudine (n=192) versus placebo (n=197).

Timepoint Children at risk Cumulative number infected Probability (%) of infection (95% CI) Relative efficacy*

(95% CI)

Zidovudine Placebo Zidovudine Placebo Zidovudine Placebo

Day 0 192 197 0 0 . . . . . .

Day 3 182 186 5 10 2·6 (0·3–4·9) 5·1 (2·0–8·2) 0·49 (0–0·83)

Day 8 173 173 14 21 7·5 (3·7–11·2) 10·8 (6·4–15·1) 0·32 (0–0·66)

Day 45 155 145 28 42 15·1 (9·9–20·2) 21·8 (15·9–27·6) 0·34 (0–0·59)

Day 90 145 131 31 48 16·8 (11·4–22·1) 25·1 (19·0–31·3) 0·37 (0·01–0·60)

Day 180 133 113 33 52 18·0 (12·4–23·5) 27·5 (21·1–33·9) 0·38 (0·05–0·60)

*12hazard ratio (univariate analysis). Eight children in the zidovudine group and three in the placebo group were of unknown HIV infection status and were excluded from the analysis.

Table 5: Probability of HIV-1 infection in children during the first 6 months of life

(6)

obtained before day 8. Curves of acquisition of infection in the two groups were very similar after 1 week.

Increased postnatal transmission was not observed in the zidovudine group by 6 months of age. This rebound effect was theoretically possible with an increase in viral load when maternal zidovudine treatment stopped.2 7 Reduction in transmission obtained with this peripartum intervention was thus compatible with reports in non-breastfed populations.4 , 2 3Identification of the time of infection is difficult when the first PCR- positive result is obtained at around day 45, as in more than 40% of our cases. Our data do not allow us to make conclusions about the respective contribution of each component of the treatment. The similarity of our findings with those of the other Côte d’Ivoire trial2 4d o e s not allow us to draw conclusions on the efficacy of a single dose versus repeated doses of oral zidovudine during labour and 1 week of postpartum maternal treatment. Efficacy was influenced by maternal CD4- cell count independently of treatment effect, as in other t r i a l s ,4 , 2 3 , 2 4but the frequency of severe immunodeficiency was low, as is generally the case in developing c o u n t r i e s .2 3 , 2 4 We cannot yet report on the long-term efficacy of the intervention because breastfeeding continues in this cohort.2 8 Owing to the threat of late postnatal transmission,2 9long-term efficacy may best be assessed from combination of several trial reports, as has been done for observational studies.3 0

No serious adverse pregnancy outcome was reported in excess at birth or during follow-up in the zidovudine group. Haematological toxic effects were uncommon, and abnormal findings were always transient. This profile is compatible with a report of the ACTG 076/ANRS 024 trial, which looked at long-term safety of zidovudine.3 1The fact that severe anaemia was rare, despite the low haemoglobin concentrations at enrolment, shows that zidovudine monotherapy can be offered to most African pregnant women, many of whom present with mild anaemia,3 2as long as iron and folate supplementation is systematically given, as it was in our trial.

Adherence to the overall intervention was good.

Compliance to prepartum treatment was not optimum, but haematological indices showed that at least some zidovudine had actually been taken. The main limiting factor for intrapartum treatment was that 40% of the women did not reach the study clinic to give birth.

Information to patients should therefore emphasise the importance of taking the oral loading dose of zidovudine as soon as labour starts. Treatment was not repeated during labour, but this procedure may be complex and will not be suitable for women who will give birth in uncontrolled conditions.2 4 Efforts to identify simpler antiretroviral regimens are needed. Such investigations will be difficult to design because the results of this trial and others set a new standard of care for developing c o u n t r i e s ,3 3 against which new treatments should be c o m p a r e d .

The public-health implications of our findings can be discussed in the light of the latest WHO guidelines.3 4A 3-week regimen of zidovudine monotherapy at the end of pregnancy and during labour can be recommended on the basis of its proven efficacy, but infant-feeding options should be discussed with each woman to guarantee maximum efficacy. Programmes to lower the rate of vertical transmission of HIV require not simply

drug delivery but also comprehensive and upgraded antenatal, obstetric, and postnatal care, including maternal HIV screening and alternatives to b r e a s t f e e d i n g .1 7 , 3 5 The cost of such a programme is therefore not only the cost of the drug.3 6This trial and o t h e r s2 3 , 2 4 are a first step towards a global strategy for maternal and child health programmes in countries with high HIV s e r o p r e v a l e n c e .

C o n t r i b u t o r s

F Dabis, P Msellati, and N Meda cowrote the protocol and coordinated the study. C Welffens-Ekra was an obstetrician primary coinvestigator.

B You, C Montcho, O Manigart, and A Simonon were responsible for molecular biology. V Leroy contributed to design and was in charge of statistical analysis. M Cartoux contributed to design and organisation of antenatal HIV screening procedures. P Combe and A Ouangré supervised laboratory activities. R Ramon and O Ky-Zerbo were trial monitors. R Salamon provided methodological expertise in trial design and project direction. C Rouzioux assisted in the implementation of molecular biology and organised quality control. P Van de Perre contributed to the design and conduct of the research, and provided direction to all virological aspects of the project. L Mandelbrot cowrote the protocol and was an obstetrician primary coinvestigator. F Dabis wrote the paper.

DITRAME (DIminution de la Transmission Mere-Enfant) Study G r o u p

Other investigators—L Dequae-Merchadou, R Lassalle (Bordeaux Coordination Unit); A Bazie, A M Cassel Beraud, B Dao,

L Gautier-Charpentier, F D Ky, B Nacro, O Sanou, I Sombié, F T a l l , S Tiendrebeogo, Y Traore, D Valea, S Yaro (Bobo-Dioulasso Centre);

and D Bonnard, R Camara, M Dosso, N Elenga, G Gourvellec, J B Kottan, R Likikouet, V Noba, M Timité, I Viho.

Abidjan health facilities—Anonkoua-Koute, Ouassakara, Yopougon, and Yopougon-Attie Health Centres, and Centre Hospitalier Universitaire de Yopougon.

Bobo-Dioulasso health facilities—Accart-Ville, Farakan, and Social Security Health Centres, and Centre Hospitalier National Sourô Sanou.

Data and safety monitoring board—J-F Delfraissy, D C o s t a g l i o l a , C Chouquet, B Bazin, P Lepage, B Masquelier, K Toure Coulibaly.

A c k n o w l e d g m e n t s

We thank J-P Levy, J Dormont, F Agid, C Canon, M-J Commoy, H Pollard, and J-L Chassaing (ANRS); C Marchal, B Floury, J P Louis, and F Fumoux (French Ministry of Cooperation); I M Coulibaly (National AIDS Authority, Côte d’Ivoire); J G Ouango and M Ouedraogo (National AIDS Authority, Burkina-Faso); D Buriot (WHO Abidjan) for support and encouragement; F Rousseau and G Walckenaer (Glaxo Wellcome, France), M Gartland and J-M Steens (Glaxo Wellcome International) for trial drugs; and M-C Saux and B Fallourd (Bordeaux University Hospital) for drug packaging; I Bely and M-P Martin (INSERM U330, Bordeaux) for secretarial and administrative assistance; and, especially the women and children who by their active participation made this research possible.

This trial was sponsored by ANRS, Paris, France and the French Ministry of Cooperation, within the Coordinated Action AC12.

R e f e r e n c e s

1 UNAIDS and WHO. Report on the global HIV/AIDS epidemic—

June 1998. Geneva: UNAIDS/WHO, 1998: 48–50.

2 The Working Group on Mother-To-Child Transmission of HIV.

Rates of mother-to-child transmission of HIV-1 in Africa, America and Europe: results from 13 perinatal studies. J Acquir Immune Defic Syndr Hum Retrovirol 1995; 8: 5 0 6 – 1 0 .

3 Van de Perre P. Postnatal transmission of human immunodeficiency virus type 1: the breast feeding dilemma. Am J Obstet Gynecol 1 9 9 5 ; 173: 4 8 3 – 8 7 .

4 Connor EM, Sperling RS, Gelber R, et al. Reduction of

maternal–infant transmission of human immunodeficiency virus type 1 with zidovudine treatment. N Engl J Med 1994; 331: 1 1 7 3 – 8 0 . 5 Mayaux MJ, Teglas JP, Mandelbrot L, et al. Acceptability and

impact of zidovudine for prevention of mother-to-child human immunodeficiency virus-1 transmission in France. J Pediatr 1 9 9 7 ; 131: 8 5 7 – 6 2 .

6 The European Collaborative Study. Therapeutic and other interventions to reduce the risk of mother-to-child transmission of HIV-1 in Europe. Br J Obstet Gynaecol 1998; 105: 7 0 4 – 0 9 . 7 Simonds RJ, Steketee R, Nesheim S, et al. Impact of zidovudine use

on risk factors for perinatal transmission of HIV. Perinatal AIDS Collaborative Transmission Studies. AIDS 1998; 12: 3 0 1 – 0 8 .

(7)

8 Dabis F, Mandelbrot L, Msellati P, Van de Perre P. Zidovudine to decrease mother-to-child transmission of HIV-1: is it good for developing countries? AIDS 1995; 9: 2 0 4 – 0 6 .

9 Newell ML, Gray G, Bryson Y. Prevention of mother-to-child transmission of HIV-1 infection. AIDS 1997; 11 (suppl A): S165–72.

1 0 CDC. Administration of zidovudine during late pregnancy and delivery to prevent perinatal HIV t r a n s m i s s i o n — T h a i l a n d , 1996–1998. MMWR Morb Mortal Wkly Rep 1998; 47: 1 5 1 – 5 4 . 1 1 UNAIDS. Prevention of HIV transmission from mother to child.

Planning for programme implementation. Report from a meeting, Geneva, March 23–24, 1998. Geneva: UNAIDS; 1998 ( h t t p : / / w w w . u n a i d s . o r g / h i g h b a n d / d o c u m e n t / m o t h e r - t o - c h i l d / i n d e x . h t m l ) .

1 2 Marshall E. Controversial trial offers hopeful result. Science 1 9 9 8 ; 279: 1 2 9 9 .

1 3 Sangaré L, Meda N, Lankoandé S, et al. HIV infection among pregnant women in Burkina Faso: a nationwide survey. Int J S T D AIDS 1997; 8: 6 4 6 – 5 1 .

1 4 Sylla-Koko F, Anglaret X, Traore-Anaky MF, et al. Séroprévalence de l’infection par le VIH dans les consultations prénatales d’Abidjan, Côte d’Ivoire, 1995. Med Mal Infect 1997; 27: 1 2 7 – 2 8 .

1 5 Adjorlolo-Johnson G, De Cock KM, Ekpini E, et al. Prospective comparison of mother-to-child transmission of HIV-1 and HIV-2 in Abidjan, Ivory Coast. J A M A1994; 272: 4 6 2 – 6 6 .

1 6 Cartoux M, Msellati P, Rouamba O, et al. Acceptability of interventions to reduce mother-to-child transmission of HIV-1 in West Africa. J Acquir Immune Defic Syndr Hum Retrovirol 1996; 1 2 : 2 9 0 – 9 2 .

1 7 Van de Perre P, Simonon A, Msellati P, et al. Postnatal transmission of human immunodeficiency virus type 1 from mother to infant: a prospective cohort study in Kigali, Rwanda. N Engl J Med 1 9 9 1 ; 325: 5 9 3 – 9 8 .

1 8 Dabis F, Msellati P, Newell M-L, et al. Methodology of intervention trials to reduce mother to child transmission of HIV with special reference to developing countries. A I D S1995; 9 (suppl A): S67–74.

1 9 Dabis F, Msellati P, Dunn D, et al. Estimating the rate of mother- to-child transmission of HIV: report of a workshop on

methodological issues Ghent (Belgium), 17–20 February 1992.

AIDS 1993; 7: 1 1 3 9 – 4 8 .

2 0 Pocock S. The size of a clinical trial. In: Clinical trials: a practical approach. New York: John Wiley & Sons, 1983: 123–41.

2 1 Cartoux M, Msellati P, Meda N, et al. Attitude of pregnant women toward HIV testing in West Africa: Abidjan, Côte d’Ivoire and Bobo Dioulasso, Burkina Faso, 1995–1996. AIDS 1998; 12: 2 3 3 7 – 4 4 .

2 2 Msellati P, Ramon R. Viho I, et al. Prevention to mother-to-child transmission of HIV in Africa: uptake of pregnant women in a clinical trial in Abidjan, Côte d’Ivoire. AIDS 1998; 12: 1 2 5 7 – 5 8 . 2 3 Shaffer N, Chuachoowong R, Mock PA, et al. Short-course

zidovudine for perinatal HIV transmission in Bangkok, Thailand:

a randomised placebo-controlled trial. Lancet 1999; 353: 7 7 3 – 8 0 . 2 4 Wiktor SZ, Ekpini E, Karon JM, et al. Short-course oral zidovudine

for prevention of mother-to-child transmission of HIV-1 in Abidjan, Côte d’Ivoire: a randomised trial. Lancet 1999; 353: 7 8 1 – 8 5 . 2 5 Simonon A, Lepage P, Karita E, et al. An assessment of the timing

of mother-to-child transmission of human immunodeficiency virus type 1 by means of polymerase chain reaction. J Acquir Immune Defic Syndr 1994; 7: 9 5 2 – 5 7 .

2 6 Bertolli J, St Louis EM, Simonds RJ, et al. Estimating the timing of mother-to-child transmission of human immunodeficiency virus in a breast feeding population in Kinshasa, Zaire. J Infect Dis 1996; 1 7 4 : 7 2 2 – 2 6 .

2 7 Van de Perre P, Meda N, Cartoux M, et al. Zidovudine and breast- feeding. AIDS Patient Care STD 1997; 11: 4 – 5 .

2 8 Msellati P, Meda N, Welffens-Ekra C, et al. Zidovudine and reduction of HIV vertical transmission in Africa. Am J Public Health (in press).

2 9 Ekpini ER, Wiktor SZ, Satten GA, et al. Late postnatal mother-to- child transmission of HIV-1 in Abidjan, Côte d’Ivoire. Lancet 1 9 9 7 ; 349: 1 0 5 4 – 5 9 .

3 0 Leroy V, Newell ML, Dabis F, et al. International multicentre pooled analysis of late postnatal mother-to-child transmission of HIV-1 infection. Lancet 1998; 352: 5 9 7 – 6 0 0 .

3 1 Sperling RS, Shapiro DE, McSherry GD, et al. Safety of the maternal–infant zidovudine regimen utilized in the Pediatric AIDS Clinical Trial Group 076 study. AIDS 1998; 12: 1 8 0 5 – 1 3 . 3 2 Meda N, Dao B, Ouangre A, for the DITRAME Study Group

(ANRS 049 Clinical Trial). HIV, maternal anemia and perinatal intervention using zidovudine. Int J Gynecol Obstet 1998; 61: 6 5 – 6 6 . 3 3 Susser M. The prevention of perinatal HIV transmission in the less-

developed world. Am J Public Health 1998; 88: 5 4 7 – 4 8 . 3 4 WHO. Recommendations on the safe and effective use of short-

course ZDV for prevention of mother-to-child transmission of HIV.

Wkly Epidemiol Rec 1998; 73: 3 1 3 – 2 0 .

3 5 WHO. HIV and infant feeding. Guidelines for decision makers.

[WHO/FRH/NUT/CHD 98.1]. Geneva: WHO, 1998.

3 6 Newell ML, Dabis F, Tolley K, Whyne D. Cost-effectiveness and cost benefit in the prevention of mother-to-child transmission of HIV in developing countries. AIDS 1998; 12: 1 5 7 1 – 8 0 .

Références

Documents relatifs

In univariate logistic regression analysis, maternal CD4 cell count at inclusion, low birth weight, mean viral load at inclusion and at day 8 post-partum were signi®cantly

Figure1a: Conceptual model for socio-economic, clinical and biological risk factors for mother – to – child transmission of HIV-1---18 Figure1b: Trends in selected

This proportion was greater than the 7.8% previously reported in Cameroon among children born to uninfected women [22], but was comparable to the figures reported in children born

Participants reviewed available evidence on the efficacy and safety of antiretroviral prophylactic regimens used for MTCT-prevention in resource-limited settings.They concluded that

A two-dose regimen of nevirapine (single dose to the mother at onset of labour and single dose to the infant within 72 hours of birth) is increasingly being used because of its low

Il a également été montré par HPLC-MS que la FAAH n’ouvre pas le cycle du β-lactame, ce qui suggère l’absence de liaison covalente entre l’enzyme et son inhibiteur..

This exten- sion induces a canonical weak distributive law DP ⇒ PD , which in turn corresponds to a canonical weak lifting of the powerset monad to the category of Eilenberg-Moore

In order to address these questions we designed a model of the Base Excision Repair pathway in Kappa, a rule based formalism for modeling protein-protein and protein-DNA