Objective: The goal of this study was to examine associations between demographic, behavioral, and clinical variables and mother-to-child HIV transmission in 15 US jurisdictions for birth years 2005 through 2008.
Methods: The study used Enhanced Perinatal Surveillance system data for HIV-infected women who gave birth to live infants. Multivariable logistic regression was used to assess variables associated with mother-to-child transmission.
Results: Among 8054 births, 179 infants (2.2%) were diagnosed with HIV infection. Half of the births had at least 1 missed prevention opportunity: 74.3% of infected infants, 52.1% of uninfected infants. Among 7757 mother-infant pairs with sufficient data for analysis, the odds of having an HIV-infected infant were higher for women who received late testing or no prenatal antiretroviral medications (odds ratio: 2.5 [95% confidence interval (CI): 1.5-4.0] and 3.5 [95% CI: 2.0-6.4], respectively). The odds for mothers who breastfed were 4.6 times (95% CI: 2.2-9.8) the odds for those who did not breastfeed. The adjusted odds for women with CD4 counts <200 cells per microliter were 2.4 times (95% CI: 1.4-4.2) those for women with CD4 counts ≥500 cells per microliter. The odds for women who abused substances were twice (95% CI: 1.4-2.9) those for women who did not.
Conclusions: The odds of having an HIV-infected infant were higher among HIV-infected women who were tested late, had no antiretroviral medications, abused substances, breastfed, or had lower CD4 cell counts. Increases in earlier HIV diagnosis, substance abuse treatment, avoidance of breastfeeding, and use of prenatal antiretroviral medications are critical in eliminating perinatal HIV infections in the United States.