Differential effects of early weaning for HIV-free survival of children born to HIV-infected mothers by severity of maternal disease

Abstract

Background: We previously reported no benefit of early weaning for HIV-free survival of children born to HIV-infected mothers in intent-to-treat analyses. Since early weaning was poorly accepted, we conducted a secondary analysis to investigate whether beneficial effects may have been hidden.

Methods: 958 HIV-infected women in Lusaka, Zambia, were randomized to abrupt weaning at 4 months (intervention) or to continued breastfeeding (control). Children were followed to 24 months with regular HIV PCR tests and examinations to determine HIV infection or death. Detailed behavioral data were collected on when all breastfeeding ended. Most participants were recruited before antiretroviral treatment (ART) became available. We compared outcomes among mother-child pairs who weaned earlier or later than intended by study design adjusting for potential confounders.

Results: Of infants alive, uninfected and still breastfeeding at 4 months in the intervention group, 16.1% who weaned as instructed acquired HIV or died by 24 months compared to 16.0% who did not comply (p = 0.98). Children of women with less severe disease during pregnancy (not eligible for ART) had worse outcomes if their mothers weaned as instructed (RH = 2.60 95% CI: 1.06-6.36) compared to those who continued breastfeeding. Conversely, children of mothers with more severe disease (eligible for ART but did not receive it) who weaned early had better outcomes (p-value interaction = 0.002). In the control group, weaning before 15 months was associated with 3.94-fold (95% CI: 1.65-9.39) increase in HIV infection or death among infants of mothers with less severe disease.

Conclusion: Incomplete adherence did not mask a benefit of early weaning. On the contrary, for women with less severe disease, early weaning was harmful and continued breastfeeding resulted in better outcomes. For women with more advanced disease, ART should be given during pregnancy for maternal health and to reduce transmission, including through breastfeeding.

Trial registration: (ClinicalTrials.gov) NCT00310726.

Conflict of interest statement

Competing Interests: The authors have declared that no competing interests exist.

Figures

Figure 1. Panel A: HIV-free survival among 229 children in the intervention group whose mothers had less severe disease and would not have been eligible for antiretroviral therapy during pregnancy stratified by whether they adhered to the intervention (weaned by the completion of 4 months) (dashed line) or whether they continued breastfeeding after 4 months (solid line) (p = 0.035).

Panel B: HIV-free survival among 97 children in the intervention group whose mothers had advanced disease and would have been eligible for antiretroviral therapy during pregnancy but did not receive it stratified by whether they adhered (weaned by the completion of 4 months) (dashed line) or whether they continued breastfeeding after 4 months (solid line) (p = 0.006).

Figure 2. Relative hazards of HIV infection or death by weaning before 15 completed months stratified by maternal CD4 count during pregnancy.

Hazard ratios greater than 1 indicate that weaning increases the risk of HIV/death. Hazard ratios less than 1 indicate that weaning decreases the risk of HIV/death.