Our objective was to determine if placental histologic acute inflammation is related to maternal and fetal serum cytokine levels in preterm labor, using a data set previously constructed blinded to histopathologic information. To this goal in 1992, 32 consecutive patients at 20-36 weeks with progressive labor and tocolytic failure were recruited. Maternal serum sampled during the active phase of labor, and fetal (umbilical vein) serum were assayed by ELISA for levels of soluble interleukin-1 beta (IL-1 beta), soluble interleukin-2 receptor (IL-2 R), and interleukin 6 (IL-6) (T-Cell Diagnostics). Acute placental inflammation was scored by two groups blinded to clinical data, and the average scores analyzed for relationships to serum cytokine levels. Weighted kappa values, reflecting interobserver agreement in scoring of acute inflammation, were: amnion 0.84; choriodecidua 0.84; umbilical cord 0.85; and chorionic plate 0.73. Fetal levels of IL-1 beta and IL-2 R were higher with grade 3-4 acute amnionitis than with grades 0-2 (p = 0.022 and p = 0.023). Fetal levels of all three cytokines were higher in grade 3-4 umbilical vasculitis (IL-1 beta p = 0.008, IL-2 R p = 0.01, and IL-6 p = 0.03). In contrast, maternal serum cytokine levels were not associated with presence or severity of histologic evidence of acute placental inflammation. Histologic acute inflammation was not related to duration of labor, interval from membrane rupture to delivery, and presence or duration of antibiotic therapy. We conclude that fetal serum, but not maternal serum cytokine levels, are correlated with histologic evidence of acute placental inflammation, and may reflect a predominant placental origin of the cytokines.