We studied the relationship between maternal second-trimester serum ferritin concentrations and preterm delivery. The 312 preterm delivery cases, studied in aggregate and in subgroups [spontaneous preterm labour, preterm premature rupture of membranes, medically induced preterm delivery, moderate preterm delivery (gestational age at delivery 34-36 weeks) and very preterm delivery (gestational age at delivery <34 weeks)] were compared with 424 randomly selected women who delivered at term. Maternal ferritin concentrations, measured in serum collected at 17 weeks gestation on average, was determined using a two-site chemiluminometric immunoassay. Using multiple logistic regression, we derived maximum likelihood estimates of adjusted odds ratios (OR) and 95% confidence intervals [CI]. Elevation in maternal second-trimester ferritin was weakly associated with the risk of preterm delivery overall. After adjusting for possible confounding by maternal age, race/ethnicity, parity, Medicaid payment status and smoking during the index pregnancy, the OR for extreme quartiles (>64.5 vs. <26.0 ng/mL) of ferritin was 1.3 [95% CI 0.8, 2.1]. Stratified analyses indicated that elevated maternal serum ferritin was associated with an increased risk of preterm premature rupture of membranes (OR = 2.1; 95% CI 1.1, 4.1), but not with spontaneous preterm labour (OR = 0.9; 95% CI 0.4, 1.7) or medically induced preterm delivery (OR = 1.1; 95% CI 0.6, 2.0). The relationship between elevated maternal second-trimester serum ferritin concentrations and preterm delivery was strongest for spontaneous very preterm deliveries (<34 weeks gestation). Women with ferritin concentrations in the highest decile (>96 ng/mL) experienced a 2.7-fold increased risk of delivering before 34 completed weeks, compared with women with concentrations <26.0 ng/mL. These results are consistent with some previous reports, and further underline the potential for heterogeneity in the aetiology of preterm delivery.