OBJECTIVE: To compare the use of midtrimester vaginal polymorphonuclear cell (PMN) counts on Gram-stain preparations with PMN to epithelial cell (EPI) ratios for the prediction of subsequent spontaneous preterm birth. METHODS: From a prospective cohort study, 83 women with a spontaneous preterm birth at less than 35 weeks of gestation (cases) were compared with a control group of 108 women who delivered at term. Vaginal Gram-stain preparations were obtained at the initial study visit (20-25 weeks). Vaginal PMN and epithelial EPI cell counts were assessed on 5 nonadjacent oil immersion fields under x 1,000 magnification. Vaginal PMN/EPI ratios for each evaluated slide field were calculated to control for intraslide variation in cellular density. RESULTS: Mean delivery gestational age was 31.5 ± 3.5 weeks for the cases and 39.8 ± 1.2 weeks for the controls. Mean PMN counts were similar in the cases (13 ± 20 cells per oil immersion field) and the controls (10 ± 14 cells per oil immersion field) (P = .17). The mean PMN/EPI ratio, however, was significantly higher among the cases (3.4 ± 6.0) than among the controls (1.8 ± 2.4, P = .01). When PMN counts and PMN/EPI ratios were dichotomized by the 95th percentile cutoff (based on control values), the PMN/EPI ratio was significantly associated with subsequent development of spontaneous preterm birth at less than 35 weeks of gestation (odds ratio 3.8, 95% confidence interval 1.3-11.2), whereas PMN counts were not (odds ratio 2.1, 95% confidence interval 0.7-6.1). CONCLUSION: Midtrimester vaginal PMN counts on Gramstain preparations were not significantly associated with subsequent spontaneous preterm birth in this population. However, the ratio of PMN to EPI counts, which provides internal standardization of variation in slide cellular density, was significantly higher in women with subsequent spontaneous preterm birth at less than 35 weeks. © 2005 by The American College of Obstetricians and Gynecologists.