A woman's partner and the characteristics of their partnership can play an important role in the health of her pregnancy. Yet, with the notable exception of intimate partner violence, there has been little previous research addressing the associations between partner- or partnership-related factors and birth outcomes. This analysis tested the hypothesis that risk factors related specifically to partner or partnership characteristics increased the risk for preterm birth. Between 2003 and 2005, a total of 580 preterm cases (20-36 weeks gestational age at delivery) and 633 term controls (≥37 weeks) were selected from women delivering at an obstetric hospital in Lima, Peru. Each woman completed a confidential, structured interview and provided biological specimens within 48 h after delivery. Multivariable logistic regression was used to assess associations between partner and partnership characteristics and preterm birth. After adjustment for behavioral, demographic, and obstetric risk factors, ever having had a partner with a history of drug use (aOR = 1.91, 95% CI 1.22-2.99), ever having had anal sex (aOR = 1.40, 95% CI 1.07-1.84), having a current partner with a history of visiting prostitutes (aOR = 1.69, 95% CI 1.22-2.33), and perceiving one's current partner as a "womanizer" (aOR = 1.34, 95% CI 1.02-1.77) were significantly associated with an elevated risk of preterm birth when tested in separate models. These four factors were then used to create a composite partnership risk score, which showed an increasing dose-response relationship with preterm birth risk (per additional partner risk factor: aOR = 1.31, 95% CI 1.16-1.49). These results highlight the importance of considering a broader set of risk factors for preterm birth, specifically those related to a woman's partner and partnership characteristics. Further research could clarify the specific mechanisms through which these partner and partnership characteristics may increase the risk of preterm birth. © 2009 Elsevier Ltd. All rights reserved.