CROI: Large Zambian study finds early abrupt weaning of HIV-exposed infants does not improve HIV-free survival

This article is more than 17 years old. Click here for more recent articles on this topic

Abruptly weaning HIV-exposed infants after four months of exclusive breastfeeding does not improve HIV-free survival over normal, gradual weaning according to results from the Zambia Exclusive Breastfeeding Study (ZEBS) presented at the Fourteenth Conference on Retroviruses and Opportunistic Infections last week in Los Angeles.

Although mothers who weaned earlier were less likely to transmit HIV to their child in this randomised controlled study, early abrupt weaning was also associated with a higher mortality rate compared to traditional weaning practices, particularly in the subset of infants who were HIV-infected. Furthermore, according to Dr Moses Sinkala of the Lusaka District Health Management Team in Zambia, abrupt weaning was not associated with as great a reduction in post-natal HIV transmission as originally expected.

“Our results do not support current recommendations that encourage early cessation of breastfeeding for HIV-infected mothers,” said Dr Sinkala. In particular, he believes that guidelines need to unambiguously recommend that mothers of infants known to be HIV-infected should continue breastfeeding until they are two years of age.


Breastfeeding is the best way to feed an infant, and protects against diarrhoea, pneumonia and malnutrition, which can occur with replacement feeding. In addition, studies have shown clear survival advantages when mothers breastfeed their infants in resource limited settings.


exclusive breastfeeding

Feeding an infant only breast milk, with no other liquids or solids, for the first six months of life.

mother-to-child transmission (MTCT)

Transmission of HIV from a mother to her unborn child in the womb or during birth, or to infants via breast milk. Also known as vertical transmission.


The result of a statistical test which tells us whether the results of a study are likely to be due to chance and would not be confirmed if the study was repeated. All p-values are between 0 and 1; the most reliable studies have p-values very close to 0. A p-value of 0.001 means that there is a 1 in 1000 probability that the results are due to chance and do not reflect a real difference. A p-value of 0.05 means there is a 1 in 20 probability that the results are due to chance. When a p-value is 0.05 or below, the result is considered to be ‘statistically significant’. Confidence intervals give similar information to p-values but are easier to interpret. 

intent to treat analysis

All participants in a clinical trial are included in the final analysis, in the groups they were originally assigned to, whether or not they actually completed their course of treatment. This method provides a better estimate of the real-world effect of a treatment than an ‘on treatment’ analysis.

inter-quartile range

The spread of values, from the smallest to the largest. The inter-quartile range (IQR) only includes the middle 50% of values and measures the degree of spread of the most common values.

However, given the risk of mother to child HIV transmission through breastmilk, guidelines were developed by WHO/UNICEF to encourage HIV-infected mothers to avoid breastfeeding whenever possible. If this was not possible from birth, then HIV-infected women were encouraged to stop breastfeeding within the first months of life — as soon as replacement-feeding should become acceptable, feasible, affordable, sustainable and safe (AFASS).

“But all these guidelines were being made without any experimental data, to either support or refute this recommendation,” said Dr Sinkala. “So how do you give a woman with HIV enough information for her to be able to make an informed choice? Some programmes recommend the early cessation of breastfeeding, but in the end, which one is really safer?”

ZEBS was therefore proposed to see whether four months of exclusive breastfeeding, followed by rapid weaning (within 24 hours or as rapidly as possible), was feasible in Zambia and to evaluate whether it would provide the same breastfeeding advantages, including reducing child mortality), as weaning normally over the course of continued breastfeeding — while also effectively reducing the risk of mother to child transmission of HIV (MTCT) at 24 months.

Methods and baseline characteristics

At two clinics in Lusaka, pregnant women were screened for HIV infection, and those testing positive were given single-dose nevirapine 200mg at labour onset. Neonates were given 2mg/kg within 24 hours of birth. All women were given extensive counselling to exclusively breastfeed.

One month after birth, women infant pairs were randomly assigned to either Group A, abrupt weaning at month four or Group B, which would wean the infant when and how they normally would. The anticipated sample size was 1,000 HIV-positive women/infant pairs. HIV infection in the infants in both groups was diagnosed by PCR and the pairs were followed for 24 months. The primary study outcome was length of infant HIV-free survival.

At the time of randomisation, there were 958 mother-infant pairs were included, with 481 in Group A, and 477 in Group B. Follow-up was highly successful, with only about 85% lost to follow-up. In Group A, 404 infants, and in Group B, 411 infants are remaining in the study or have reached the primary endpoints (HIV, death or 24 months).

The baseline characteristics were well matched between groups. The median age of the women in the study was 26 years old. Only 7% were employed (as other than a home-maker). 85% were married, 50% had a husband that was employed. 40% had electricity at home, 14% had a refrigerator. Only 6% had a water tap inside their house, 10% had a water tap outside but on their property. Many were quite poor — 23% had been without food for more than one day in the prior month.

The median CD4 cell count at baseline was 329 cells/mm3 (interquartile range (IQR) 207-471). The median viral load prior to exclusive breastfeeding was 39,334 copies per ml (IQR 9,922-131,317). 76% had exclusively breastfeed their previous child for at least four months.

Discussion and implications

Audience members asked what provision was made for replacement foods for the infants, and Dr Sinkala said the study investigators had made certain that the mothers had adequate replacement food supplies and even directly provided food whenever there was a problem (such as when infants were failing to thrive). Its not clear that this was a proactive enough strategy to prevent nutritional problems in these infants, though clearly the infants in this study had better access to replacement foods than most HIV-exposed infants without access to dedicated programmatic nutritional interventions.

The findings provide data which could help mothers choose which feeding options are best, considering her individual immune status and whether her infant is known to be HIV-infected or not.

For example, “the outcomes were better if breastfeeding continued,” said Dr Sinkala. “We believe that counselling for mothers of known HIV-infected children should unambiguously encourage to breastfeed until the second year of age.”

In addition, the results strongly indicate that encouraging mothers to wean early may not produce better infant health outcomes in the absence of alternative feeding options that have been demonstrated to be AFASS. Thus the study could also have programmatic implications in countries where replacement-feeding interventions are not believed to feasible.

“Early cessation of breastfeeding has got a price. It has got a significant programmatic cost including access to breastmilk substitutes, whether mothers can afford them or not. And it also carries many social risks including disclosure of HIV status, as well as stigmatisation. These costs and risks may be justifiable, if net benefit to infant health can be achieved but we didn’t see that in this study," concluded Dr Sinkala.

More on the implications of this and other infant feeding studies presented at the Conference will be discussed in a separate article.


“If you can give mothers proper education and support, they can exclusively breastfeed,” said Dr Sinkala. Uptake of exclusive breastfeeding in the study was high with 80% still exclusively breastfeeding at month four (non-exclusive breastfeeding was defined as reporting giving any liquid, including water, or solids, other than medicine, to the infant yesterday or at least once per week).

Failing to exclusively breastfeed within the first four months of life was a major risk factor for HIV transmission, with 10.1% of those infants who non-exclusively breastfed becoming postnatally infected within the first four months of life compared to 4% of those who were exclusively breastfed, with a time dependent hazard ratio of 3.4 (95% confidence interval (CI) 1.7-7.2).

Of those mothers who were still exclusively breastfeeding at 4 months of age, 59.3% of mothers in Group A successfully stopped breastfeeding as directed. Of these, 70% weaned within 0-2 days, 25% within two days to a week, 4% within 7-14 days and 1% slightly over 15 days. However, some mothers assigned to group A appeared to disregard their randomisation with 29.7% still breastfeeding at one year, and 24.1% at 15 months. In Group B, however, women weaned much more gradually. At one year, only 32.9% had weaned, and only 45.7 had weaned at 15 months, with a median duration of 16 months of breastfeeding. Both groups have comparable CD4 cell counts and viral loads.

At 24 months, no significant survival difference was found between infants in the two study groups. If restricted to infants who were HIV-negative at four months, in the intent to treat analysis at 24 months the early weaning study group experienced 17% HIV infection or mortality (n=329) vs. 19% (n=331) in the extended breastfeeding group (p = 0.21). Dr Sinkala said there was also no difference in an “as practiced” analysis (comparing the groups by reported weaning practice).

Looking at maternal stage of disease and at compliance in Group A, there was a benefit to continuing to breastfeed in the subset of asymptomatic mothers with higher CD4 cell counts (above 350), HIV-free survival among HIV-negative children who continued to be breastfed was significantly higher than those who were abruptly weaned at four months (p=0.03).

Continued breastfeeding was also better for the infants with HIV-positive test results before four months of age, with dramatically better survival in Group B (p=0.01).


Sinkala M et al. No benefit of early cessation of breastfeeding at 4 months on HIV-free survival of infants born to HIV-infected mothers in Zambia: the Zambia Exclusive Breastfeeding Study. Fourteenth Conference on Retroviruses and Opportunistic Infections, Los Angeles, abstract 74, 2007.