Background: Previous studies of teenage primiparas have found little or no association between young maternal age and preterm birth. However, the risk of preterm birth in teenage multiparas should not be overlooked because of the high rate of repeat teenage pregnancies. Objective: To compare the risk of preterm birth in teenage and adult multiparas. Design: Cross-sectional analysis of US Natality Files, 1990 to 1996. Methods: We calculated the risk of very preterm birth (Ͻ33 weeks' gestation) for multiparas aged 10 to 20 years compared with 25-year-olds, stratified by age and race/ethnicity. Adjusted odds ratios (AORs) were estimated controlling for maternal education, marital status, prenatal care, and previous preterm births. Effects of smoking and interpregnancy interval were analyzed separately. Results: Throughout adolescence, multiparas face higher AORs for very preterm births. For white non-Hispanic multiparas compared with 25-year-old multiparas, 10to 14-year-olds had an AOR of 4.22 (95% confidence interval [CI], 2.26-7.88), 15-to 17-year-olds had an AOR of 2.19 (95% CI, 1.99-2.42), 18-and 19-year-olds had an AOR of 1.69 (95% CI, 1.58-1.80), and 20-year-olds had an AOR of 1.33 (95% CI, 1.24-1.41). A similar pattern of decreasing AOR with increasing maternal age was observed for black non-Hispanic and Hispanic mothers, although wide race/ethnicity disparities exist. Adjusting for maternal smoking and short interpregnancy interval did not change these results. Conclusions: Risk of very preterm birth in teenage multiparas is associated with young age after controlling for other risk factors. Interventions to prevent repeat pregnancies and the associated risk of premature birth deserve high priority.