637 
Risk Factors for Mother-to-Child Transmission of HIV-1 from Breastfeeding in a Randomized Clinical Trial in Botswana: The Mashi Study
Roger Shapiro*1, L Smeaton2, S Lockman3, I Thior4, R Rossenkhan4, C Wester4, L Stevens4, J Leidner2, J Makhema4, and M Essex5
1Beth Israel Deaconess Med Ctr, Boston, MA, US; 2Harvard Sch of Publ Hlth, Ctr for Biostatistics in AIDS Res, Boston, MA, US; 3Brigham and Women`s Hosp, Boston, MA, US; 4Botswana Harvard Sch of Publ Hlth AIDS Initiative Partnership for HIV Res and Ed, Gaborone; and 5Harvard Sch of Publ Hlth, Boston, MA, US
Background: Early breastfeeding and potential late
risk factors for mother-to-child HIV transmission (MTCT) were evaluated in a
randomized infant feeding trial.
Methods: HIV-infected women were randomized to
breastfeed or formula feed for 6 months. Mothers received antenatal and
intrapartum zidovudine (ZDV), and single-dose nevirapine (NVP) or placebo.
Infants received single-dose NVP or placebo, and ZDV prophylaxis for 1 month (formula-fed
arm) or 6 months (breastfed arm). Infant HIV DNA polymerase chain reaction (PCR)
was performed at birth, 1, 4, 7, and 12 months. Monthly visits occurred through
7 months. Exclusive breastfeeding (no liquids or solids, but 1 allowed liquid lapse)
was defined at 5 months. Maternal HAART became available mid-study (CD4 cell
count <200 cells/mm3 or AIDS). Estimates were by Kaplan-Meier
methods and Cox proportional hazards modeling.
Results: Of 1116 infants at risk, 6 (1.1%) formula-fed
and 7 (1.2%) breastfed became infected between birth and 1 month (p = 1.0).
Maternal single-dose NVP receipt did not predict MTCT either in the first month
or thereafter. Of 549 breastfed infants alive and HIV-uninfected at 1 month, in
24 (4.4%) late HIV transmissions occurred; in 15 by 4 months, in 6 from 4 to 7
months, and in 3 thereafter. Infant feeding patterns were similar from months 1
to 4 regardless of eventual HIV status. Univariate associations with late MTCT
included higher maternal plasma or breast milk HIV-1 RNA (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
(2.8%) of 109 exclusively breastfed infants became infected compared with 20
(4.8%) of 400 mixed-fed infants (p = 0.14). In multivariable analysis
(excluding breast milk HIV-1 RNA), maternal HIV-1 RNA (p = 0.01),
maternal CD4 cell count (p = 0.06), no electricity in the home (p =
0.05), infant diarrhea (p = 0.04), and infant anemia (p = 0.004)
predicted late MTCT. No MTCT occurred in 34 breastfed infants whose mothers
started HAART by delivery. Median baseline maternal CD4 cell count for late
transmitters was 225 cells/mm3. No late MTCT occurred when baseline
maternal plasma HIV-1 RNA was <3500 copies/mL.
Conclusions: With maternal and infant antiretroviral
prophylaxis and partial HAART availability, breastfeeding was not a risk for
MTCT in the first month of life. Late MTCT was rare at higher maternal CD4 cell
count or lower HIV-1 RNA. Reverse causality may explain associations between
infant illness and late MTCT when HIV testing is infrequent.
|
