Association of Adverse Neighborhood Exposures With HIV Viral Load in Pregnant Women at Delivery.

Importance: Racial disparities in maternal morbidity and mortality are in large part driven by poor control of chronic diseases. The association between adverse neighborhood exposures and HIV virologic control has not been well described for women with HIV during pregnancy. Objective: To evaluate the association between adverse neighborhood exposures and HIV viral load at delivery. Design, Setting, and Participants: This population-based cohort study assessed HIV surveillance data for pregnant women with HIV who had live deliveries in Philadelphia from January 1, 2005, through December 31, 2015. Data analyses were completed in August 2020. Exposures: Neighborhood exposures included extreme poverty, educational attainment, crime rates (using separate and composite measures), and social capital categorized above or below the median. Each neighborhood exposure was modeled separately to estimate its association with elevated HIV viral load. Main Outcomes and Measures: The main outcome was elevated HIV viral load of ≥200 copies/mL at delivery. We hypothesized that adverse neighborhood exposures would be associated with higher odds of having an elevated viral load at delivery. Confounders included birth year, age, race/ethnicity, previous birth while living with HIV, and prenatal HIV diagnosis. Prenatal care and substance use were considered potential mediators. We used logistic mixed effects models to estimate the association between neighborhood exposures and elevated viral load, adjusting for confounders in Model 1 and confounders and mediators in Model 2. Results: There were 905 births among 684 women with HIV, most of whom were aged 25 to 34 years (n = 463 [51.2%]) and were Black non-Hispanic (n = 743 [82.1%]). The proportion of women with elevated viral load decreased from 58.2% between 2005 and 2009 to 23.1% between 2010 and 2015. After adjusting for confounders in Model 1, higher neighborhood education was associated with lower odds of having an elevated viral load (adjusted odds ratio [AOR], 0.70; 95% CI, 0.50-0.96). More violent crime (AOR, 1.51; 95% CI, 1.10-2.07), prostitution crime (AOR, 1.46; 95% CI, 1.06-2.00), and a composite measure of crime (AOR, 1.44; 95% CI, 1.05-1.98) were positively associated with having a higher HIV viral load. These associations remained after adjusting for mediators in Model 2. In addition, the AOR for intermediate prenatal care varied between 1.93 (95% CI, 1.28-2.91) and 1.97 (95% CI, 1.31-2.96), whereas the AOR for inadequate prenatal care varied between 3.01 (95% CI, 2.05-4.43) and 3.06 (95% CI, 2.08-4.49) across regression models. Conclusions and Relevance: In this cohort study, adverse neighborhood exposures during pregnancy and poor engagement in prenatal care were associated with poor virologic control at delivery. These findings suggest that interventions targeted at improving maternal health need to take the social environment into consideration.