Background:
Recent studies observed detectable HIV-1 virus levels in breastmilk (BM) despite undetectable HIV-1 RNA viral load (VL) in plasma. This discordance between HIV VL in plasma and BM could account for residual vertical HIV-1 transmission during lactation. We assessed the association of vertical HIV-1 transmission with HIV VL in plasma and BM, and with tenofovir (TFV) drug levels.
Methods:
This case-control study was nested within the IMPAACT PROMISE 1077BF perinatal HIV trial which evaluated maternal ARV strategies to prevent BM HIV transmission. Cases were mother-infant pairs with infants who had a positive HIV nucleic acid test (NAT) during the breastfeeding period; control pairs had infants who were HIV NAT negative. Two controls were matched for each case by infant sex, study site, maternal age at delivery, and 1077BF component at infection. Maternal plasma and BM collected near an infant’s infection date were assayed for HIV total nucleic acid (TNA; DNA + RNA) VL, DNA VL, RNA VL, and TFV concentration. Conditional logistic regression was used to calculate odds ratios (ORs) and 95% confidence intervals (CIs).
Results:
93 mother-infant pairs (31 cases; 62 controls) were enrolled from Malawi, Uganda, South Africa, Zimbabwe and India. Median maternal age at delivery was 25 years, 39 (42%) were male infants, 57 (61%) were randomized in the postpartum component. Median (Q1, Q3) age of infection was 6 (3, 14) months. Over 70% (22/31) of samples were taken on the same day or within one month of infection. Median (Q1, Q3) maternal plasma VL was 39,228 (4822, 124,886) copies/ml for cases vs 20 (20, 2,104) for controls. BM RNA VL was above lower limit of quantification for 17 (55%) cases and 7 (11%) controls. The odds of infant HIV infection were 2.6 times higher for each log10 increase in maternal plasma RNA VL (95% CI: 1.6-4.5) and 1.8 times higher for each log10 increase in BM RNA VL (95% CI: 1.3-2.6). BM DNA VL was not detected in 26 (84%) cases and 59 (95%) controls. Only 3/14 (21%) case mothers on a TFV-containing regimen had detectable TFV levels in their plasma or BM vs 31/37 (84%) control mothers with detectable TFV in plasma and 29/37 (78%) in BM. In case mothers, plasma TFV levels were 10-fold lower (geometric mean ratio (95% CI): 0.11 (0.04-0.26) compared with controls, and BM TFV levels were 5-fold lower 0.18 (0.08-0.43) in case mothers compared with controls.
Conclusions:
Odds of breastmilk HIV-1 transmission is associated with higher maternal plasma and BM VL and lower TFV levels.
