Risk factors for breastfeeding transmission

Polly Clayden, HIV i-Base

Two reports looked at risk factor for postnatal transmission.

Pooled analysis from the Vertical Transmission Study and Ditrame Plus

In an oral abstract Renaud Becquet presented findings from an analysis of breastfeeding pattern and duration and their contribution to postnatal HIV transmission. [1]

Data were pooled from two studies: Ditrame Plus, conducted in urban West Africa, where breastfeeding cessation at 4 months was recommended and where exclusive breastfeeding was rare; and the Vertical Transmission Study which took place in rural South Africa, where promotion of safer breastfeeding means high rates of exclusive breastfeeding, but with longer duration.

This analysis assessed 18-month HIV postnatal transmission. The study included 1195 breastfed infants, HIV-negative at birth. 90% of the infants were breastfed for 3 months. At 6 months 83% of the South African and 38% West African infants children were still breastfed. At 12 months this had declined to 38% and 20% in South Africa and West Africa respectively.

At 3 months of age, 66% of the South African infants had been exclusively breastfed since birth and 55% of West African infants had been predominantly breastfed (defined as breastmilk plus water-based drinks).

The investigators reported the overall postnatal-transmission risk at 18-month in the South African infants was twice that of the West African infants, 9% (95% CI 7-11%) vs 5% (95% CI 3-8%), p=0.03.

In a competing risk analysis allowing for duration of breastfeeding and assuming that all children were breastfed for 18 months, the investigators found the postnatal-transmission risk would be 14% (95% CI 10-18%) in the South African infants and 16% (95% CI 8-28) in the West African infants p=0.32.

The 18-month postnatal-transmission risk was 3.9% (95% CI 2.3-6.5) for children breastfed for less than 6 months, and 8.7% (95% CI 6.8 to 11.0) for those breastfed for more than 6 months; crude HR 2.1 (1.2-3.7), 0.02; adjusted HR 1.8 (0.9 to 3.4), p=0.06 (adjustment included maternal CD4 count).

The investigators found no difference in transmission risk for infants exclusively breastfed to those predominantly breastfed for the same period. But infants receiving solids at least once during the first 2 months of life were 2.9 (1.1-8.0) times more likely to acquire HIV through breastfeeding than infants never exposed to solids in this period (adjusted competing risk analysis, p=0.04).

The investigators wrote: “Breastfeeding duration is a major determinant of postnatal HIV transmission, and safe alternatives to breastmilk beyond 6 months are urgently needed. The postnatal-transmission risk was not different in exclusively and predominantly breastfed children, and we confirm the negative effects of mixed breastfeeding with solids.”


A poster authored by Roger Shapiro and coworkers showed findings from an analysis of early breastfeeding and potential late risk factors for transmission in a randomised infant feeding trial. [2]

In the Mashi study, HIV-positive women were randomised to either breastfeed or formula feed for 6 months. The mothers received AZT from 34 months of gestation and either single-dose NVP or placebo. The infants received single-dose NVP or placebo, and AZT for 1 month (formula-feeding arm) or 6 months (breastfeeding arm). Maternal HAART became available mid-study for women with CD4 cell count <200 cells/mm3 or AIDS.

The investigators reported, of 1116 infants HIV-negative at birth, 6 (1.1%) formula-fed and 7 (1.2%) breastfed were HV-positive by 1 month (p=1.0). They did not find maternal NVP predictive of MTCT either in the first month or later in this study.

Of 549 breastfed infants alive and HIV-negative at 1 month, there were 24 (4.4%) late transmissions: 15 by 4 months, 6 from 4 to 7 months, and in 3 later. Infant feeding patterns were similar from months 1 to 4 regardless of eventual HIV status.

In univariate analysis, associations with late MTCT included higher maternal plasma or breast milk viral load, p=0.0002 and p=0.02, respectively); lower maternal CD4 cell count, p=0.005; infant diarrhea, p=0.03; and infant anemia p=0.001). 3/109 (2.8%) exclusively breastfed infants were infected compared to 20/400 (4.8%) of 400 mixed-fed infants, p= 0.14.

In multivariate analysis, excluding breast milk viral load, maternal viral load (OR 2.7, 95%CI 1.4-55, p=0.005), maternal CD4 cell count (OR 0.08, 95%CI 0.76-1.0, p= 0.06), no electricity in the home (OR 0.1, 95%CI 0.02-1.0, p=0.05), infant diarrhea (OR 3.0, 95% OR 1.2-7.9, p=0.02) and infant anemia (OR 4.1, 95% CI 1.5-11.9, p=0.008) predicted late MTCT.

The investigators reported there were no transmissions in the 34 breastfed infants whose mothers started HAART before delivery. The median baseline maternal CD4 cell count for late transmitters was 225 cells/mm3. There were no late transmissions when baseline maternal viral load was <3500 copies/mL. They noted, “Reverse causality may explain associations between infant illness and late MTCT when HIV testing is infrequent.”


  1. Renaud Becquet R, Bland R,V Leroy V et al. Duration and pattern of breastfeeding and postnatal transmission of HIV: Pooled analysis of individual data from a West and South African cohort study. 15th CROI. February 2008. Boston USA. Oral abstract 46.
  2. Shapiro R, Smeaton L, Lockman S et al. Risk Factors for mother-to-child transmission of HIV-1 from breastfeeding in a randomised clinical trial in Botswana: The Mashi study. 15th CROI. February 2008. Boston USA. Poster abstract 637.

Links to other websites are current at date of posting but not maintained.