weeks gestation provided a blood sample and completed a food frequency and sociodemographic/health behavior questionnaire. Both serum (n = 1825) and red blood cell (n = 887) folate were analyzed using immunoassay techniques. Dietary folate was poorly correlated with serum (r = -0.08) and red blood cell (-0.05) folate, even after accounting for supplement use. Serum folate was moderately associated with red blood cell folate r = 0.43. The lowest tertile of serum folate was consistently associated with preterm birth even after adjustment for supplement use, smoking, and pregravid BMI (AOR = 2.0, 95% CI = 1.4, 2.8), with a stronger association among non-supplement users (AOR = 2.9, 95% CI = 1.2, 7.3). There was no association with dietary intake. The unadjusted odds ratio for the lowest tertile of red cell folate was near the null for the entire sample as well as for supplement users, but was 1.56 (95% CI = 0.8, 2.8) among non-supplement users. The adjusted odds ratio was 1.24 (95% CI = 0.84, 1.8) for the entire sample, which did not change much after stratification by supplement use (AOR = 1.24 supplement vs. 1.39 non-supplement users). Serum folate was associated with a preterm birth when measured in the second trimester of pregnancy, but the discordant results between the short (serum) and long (red blood cell) term markers of folate status and preterm birth warrant further investigation.Young maternal age and poor pregnancy outcomes: revisiting the association