University of California, San Francisco Logo

University of California, San Francisco | About UCSF | Search UCSF | UCSF Medical Center

Home > Global Health Literature Digest > Differential Effects
Differential effects of early weaning for HIV-free survival of children born to HIV-infected mothers by severity of maternal disease
Global Health Sciences Literature Digest
Published September 21, 2009
Journal Article

Kuhn L, Aldrovandi GM, Sinkala M, et al. Differential effects of early weaning for HIV-free survival of children born to HIV-infected mothers by severity of maternal disease. Zambia Exclusive Breastfeeding Study (ZEBS). PLoS One. 2009 Jun 26;4(6):e6059.

In Context

Current guidelines from the World Health Organization (WHO) recommend exclusive breastfeeding for HIV-infected women for the first six months of life unless replacement feeding is acceptable, feasible, affordable, sustainable and safe (AFASS) for them and their infants before that time.(1) Estimates from mathematical models and findings from observational studies have suggested early weaning reduces rates of HIV transmission(2,3,4,5,6,7) These studies, however, have not taken into account the risk of mortality in uninfected children secondary to early weaning. The Zambia Exclusive Breastfeeding Study (ZEBS) was a clinical trial of a behavioural intervention involving early weaning at age four months on HIV-free survival of infants born to HIV-infected mothers. Results from this study demonstrated no benefit of early weaning on HIV-free survival.(8) Given the poor acceptance of early weaning in the study population and the incomplete adherence, a true benefit of the intervention may have been masked. The present analysis was designed to evaluate this hypothesis.

Objective

To investigate the consequences of incomplete adherence for inferences about the effects of early weaning, comparing outcomes among mother-child pairs in the intervention and control groups who stopped breastfeeding earlier or later than intended by the study design

Setting

Two antenatal clinics in Lusaka, Zambia

Study Design

Randomized controlled trial, unblinded

Participants

HIV-infected pregnant women who attended either of two study clinics and intended to breastfeed. Women still breastfeeding at one month post-partum (n=958) were randomized.

Primary outcomes

Infant HIV infection or death at 24 months of age. For HIV infection, the midpoint between the last negative and first positive PCR test was imputed as the event time; for deaths, the actual age of death was used.

Methods

The trial evaluated a behavioural intervention to encourage exclusive breastfeeding with abrupt cessation of breastfeeding at four months compared to standard practice. HIV-infected pregnant women who intended to breastfeed were recruited into the study and received single-dose nevirapine prophylaxis. Women who were still breastfeeding at one month post-partum were randomized into two arms: In the control arm, women were encouraged to exclusively breastfeed for six months, gradually introduce complementary foods (not provided) and continue to breastfeed for a duration of their own choosing. In the intervention arm, women were encouraged to breastfeed exclusively for four months and then to stop breastfeeding abruptly. A three-month supply of infant formula, a fortified weaning cereal, and education on safe replacement feeding practices was provided. Mother-infant pairs were followed for 24 months. A standardized questionnaire was administered at study visits to determine feeding behaviors. Breastfeeding duration was determined based on the exact age that breastfeeding was first reported to have completely stopped. Maternal blood samples were obtained at enrollment and were tested for CD4 and CD8 counts. Women were classified retrospectively as having severe disease if their CD4 count was <200 cells/µL or if their CD4 count was 200-349 cells/µL and they had WHO stage III or greater clinical disease. If women did not meet these criteria, they were classified as having less severe disease. Infant blood samples were collected on the day of birth and at regular intervals through 24 months of age and tested by PCR. Information on child deaths was obtained from hospital and clinic records and interviews with caretakers and health care personnel.

The authors conducted secondary analyses to evaluate the effect of incomplete adherence on previously observed associations between early infant weaning and HIV infection or death. Outcomes among mother-child pairs in the intervention and control groups who adhered to the study protocol were compared with those who did not adhere, defined as stopping breastfeeding earlier or later than specified by the study protocol. Cox proportional hazard models were used to investigate potential confounders and effect modifiers of the relation between infant feeding modality and outcome.

Results

Of the 958 mother-child pairs randomized, 328/481 (68.2%) in the intervention arm and 333/477 (69.8%) in the control arm were alive, HIV-uninfected and still breastfeeding at four months. Analysis was restricted to these 661 mother-child pairs. Median duration of breastfeeding was 4.4 months (IQR 4.2-15) in the intervention arm and 16 months (IQR 12-19) in the control arm (P<0.001).

In the intervention group, there was no significant difference in the rate of HIV infection or death by 24 months among those who adhered to the intervention (weaning at 4 months) (16.1%) compared to those who continued breastfeeding (16.0%) (P=0.98). Significant effect modification was observed by severity of maternal disease. Among women with less advanced disease, HIV transmission rates were similar among those who adhered to weaning at 4 months, and those who continued to breastfeed (RH 1.38; 95% CI: 0.34-5.51). However, mortality rates in uninfected children were significantly higher in mothers who adhered compared to those who continued to breastfeed (RH 3.23; 95% CI: 1.10-9.50). Among women with advanced HIV disease, HIV transmission rates were significantly lower among those who adhered to weaning at four months compared to those who continued to breastfeed (RH 0.25; 95% CI: 0.09-0.72). In this group, mortality rates in uninfected children did not differ by weaning practice (RH 0.36; 95% CI: 0.07-1.77). Overall, infants of women with less severe disease during pregnancy had significantly worse outcomes if their mothers were adherent and weaned early compared to infants whose mothers were non-adherent and continued to breastfeed (RH=2.40; 95% CI: 1.04-5.54). However, children born to women with advanced HIV disease had significantly better outcomes if their mothers were adherent (RH=2.60; 95% CI: 1.06-6.36, P=0.036), even after adjusting for confounders.

In the control group, significant interaction also was observed between severity of maternal disease and weaning practice. Among infants of mothers with less severe disease, stopping breastfeeding before 24 months was associated with a significant increased risk of HIV infection or death (RH 3.41; 95% CI: 1.52-7.65). However, among infants of mothers with more advanced disease, weaning was not associated with the outcome after adjusting for maternal viral load and child birth weight.

Among infants born to HIV-infected mothers with less advanced disease in either arm, breastfeeding remained protective until 15 months of age. Using breastfeeding to 15 months as the reference, weaning earlier was associated with reduced risk of HIV infection or death when maternal CD4 counts were low and increased risk of HIV infection or death when maternal CD4 counts were high. A threshold CD4 count of 306 cells/µL (95% CI: 220-394) was determined as the point at which weaning at 15 months switched from protective to harmful. Early weaning was determined to be protective if women had CD4 counts below threshold and harmful if women had higher CD4 counts.

Conclusions

Authors state that even after taking into account the dilution introduced by poor adherence, the original inference of no benefit of early weaning is correct if no distinction is made among mothers at different stages of HIV progression. Gains achieved in transmission reduction due to early weaning were offset by increases in mortality in the group overall. However, differential effects of early weaning by the severity of maternal disease were observed. For women with more advanced disease, HIV transmission rates were significantly lower among those who weaned early, resulting in a net benefit of early weaning. For women with less severe disease, however, early weaning was harmful and continued breastfeeding resulted in better outcomes. The authors conclude that infants born to HIV-infected women who have high CD4 counts are at increased risk if breastfeeding is stopped at any time before 16 months and that when maternal CD4 counts exceed 300 cells/µL there is a net benefit of continued breastfeeding for HIV-free survival of infants.

Quality Rating

The ZEBS trial was of moderate quality. The trial was randomized and authors note that randomization assignment was prepared by the study statistician and was accessed through a computer program. Given the nature of the behavioural intervention, it was impossible to blind participants and study team to the treatment arm. However, the primary outcome of interest was not deemed to be subject to bias based on lack of blinding. In the original analysis, intention to treat analyses were performed. One study limitation was the lack of acceptance of early weaning by study participants, leading to poor adherence to the intervention. Lack of adherence has the potential to bias results toward the null, and the present analysis was performed to evaluate whether beneficial effects were masked by adherence issues. This secondary analysis examined a large number of confounders.

Programmatic Implications

The authors state that for women with less severe disease, breastfeeding should be sustained through 16 months. For women with more severe HIV disease, authors argue that there is little justification for shortening the duration of breastfeeding if access to ART is in place (it was not available in the public sector in Zambia when this study was done). They suggest that ART should be initiated in these women, both for their own health and to reduce transmission to the infant. However, the authors note that further research should be conducted on the administration of antiretrovirals for the duration of breastfeeding to evaluate late postnatal breastfeeding transmission. Results from this study do not fully support the current WHO infant feeding guidelines for HIV-infected women. Future infant feeding recommendations should consider these study results.

References

  1. WHO. HIV and Infant Feeding Technical Consultation Held on behalf of the Inter-agency Task Team (IATT) on Prevention of Infections in Pregnant Women, Mothers and their Infants: Consensus Statement. Geneva, October 25-27, 2006. Abstract not available.
  2. Kuhn L, Stein Z. Infant survival, HIV infection and feeding alternatives in less developed countries. Am J Public Health 1997;87:926-31.
  3. Nagelkerke NJ, Moses S, Embree JE, Jenniskens F, Plummer FA. The duration of breastfeeding by HIV-1-infected mothers in developing countries: balancing benefits and risks. J Acquir Immun Defic Syndr Hum Retrovirol 1995;8:176-81.
  4. Ekpini ER, Wiktor SZ, Satten GA, et al. Late postnatal mother-to-child transmission of HIV-1 in Abidjan, Cote d'Ivoire. Lancet 1997;349:1054-9.
  5. Miotti PG, Taha TE, Kumwenda NI, et al. HIV transmission through breast feeding: a study in Malawi. JAMA 1999;282:744-9.
  6. Fawzi W, Msamanga G, Spiegelman D, et al. Transmission of HIV-1 through breastfeeding among women in Dar es Salaam, Tanzania. JAIDS 2002;31:331-8.
  7. Taha T, Hoover DR, Kumwenda NI, et al. Late postnatal transmission of HIV-1 and associated factors. J Infect Dis 2007;196:10-14.
  8. Kuhn L, Aldrovandi GM, Sinkala M, et al. Effects of early, abrupt cessation of breastfeeding on HIV-free survival of children in Zambia. N Engl J Med 2008;359:130-41.