Background: Decreasing infant mortality was a key aim of Millennium Development Goal (MDG) 4. While many regions worldwide made substantial progress, not all attained MDG4. Defining determinants of infant mortality in settings with high rates of infant mortality can inform strategies to further decrease mortality.Methods: Data were analyzed from the Mama Salama Study (MSS), a prospective peripartum cohort study in Western Kenya examining HIV acquisition in pregnancy to 9 months postpartum between 2011 and 2014. Cases of infant death were compared to control infants who survived to 9 months postpartum. Sub-analyses compared neonatal and perinatal mortality cases to controls. Logistic regression was used to identify determinants of infant, neonatal, and perinatal mortality using StataÒ 13 software.Findings: In multivariate case-control comparison of 34 infant deaths and 1053 control infants, independent correlates of infant mortality were preterm delivery (aOR¼3.49, 95% CI 1.68-7.26), twin delivery (aOR¼4.63, 95% CI 1.22-17.55), travel time to clinic greater than 1 hour (aOR¼2.66, 95% CI 1.04-6.84), maternal malaria during pregnancy (aOR¼3.52, 95% CI 1.40-8.86), and maternal chlamydia infection during pregnancy (aOR¼3.76, 95% CI 1.37-10.30). Maternal chlamydia infection was also an independent determinant of neonatal mortality (aOR¼9.56, 95% CI 2.49-36.64).Interpretation: Improved services to detect, treat and prevent maternal and infant chlamydia and malaria, and vigilance in the care of preterm and twin deliveries may decrease infant mortality in high mortality regions.