OBJECTIVE: Our purpose was to create a model for predicting amnionitis and rapid delivery in preterm labor patients by use of amniotic fluid interleukin-6 and clinical parameters. STUDY DESIGN: Amniotic fluid was cultured and analyzed, and a clinical score (incorporating gestational age, amniotic fluid Gram stain, glucose, leukocyte esterase, and maternal serum C-reactive protein) was determined in 111 patients diagnosed with preterm labor. Statistical analysis involved t tests, χ 2, logarithmic regression, and multivariate regression analysis ( P ≤ .05). RESULTS: The incidence of positive amniotic fluid cultures was 8.7% (9 of 103 patients). Patients with positive cultures of the amniotic fluid had a shorter delivery interval (4.8 ± 7.5 vs 28.9 ± 25.4 days, P < .001). Patients with elevated amniotic fluid interleukin-6 (≥7586 pg/ml) were more likely to have a positive amniotic fluid culture (relative risk = 8.8, 95% confidence interval = 1.6 to 47.4, P < .001) and to be delivered within 2 days (relative risk = 16.8, 95% confidence interval = 4.5 to 62.7, P < .001). Stepwise multivariate regression analysis yielded a model using interleukin-6, cervical dilatation, and gestational age ( r 2 = 0.63, P < .001) with a specificity of 100% for predicting delivery within 2 days of amniocentesis. CONCLUSIONS: A mathematical model using maternal amniotic fluid interleukin-6 seems to be a useful clinical tool for quantifying the interval to preterm birth for patients in preterm labor. (Am J Obstet Gynecol 1998;179:172-8.)