Background: The influence of class I and class II HLA specificities and of the concordance between maternal and infant HLA on vertical HIV-1 transmission is a controversial issue. 

Methods:  HLA typing was performed with serological and molecular methods in samples of mothers and infants enrolled in the Ariel study, a cohort including 203 mother-infant pairs.  HLA effects were evaluated alone and in multivariate modeling considering also other known predictors of perinatal transmission (maternal viral load, antiretroviral therapy, duration of ruptured membranes, chorioamnionitis).  

Results:  Modest associations were seen with specific alleles (increased risk with infant B67 and B58 and maternal DR1; decreased risk with maternal B12), but these were not statistically significant after adjusting for multiple comparisons. Mother-infant concordance at any class I locus was a strong predictor of transmission (odds ratio [OR] 4.16; p= 0.028). Transmission was not associated with class II concordance. Class I HLA concordance retained its importance after adjusting for maternal viral load, antiretroviral therapy, duration of rupture of membranes or histological chorioamnionitis.  In multivariate modeling, only class I concordance (OR 3.59; p= 0.069) and chorioamnionitis (OR 3.79; p= 0.030) were retained as independent predictors of transmission. 

Conclusions:  HLA alleles, and in particular the class I concordance between maternal and neonatal HLA, may regulate the risk of perinatal HIV-1 transmission.