OBJECTIVE: To estimate the natural history of funneling in the second trimester by transvaginal ultrasonograms and whether funneling increases the risk of spontaneous birth. METHODS: Secondary analysis of a blinded, multi-center observational study of women with at least one prior spontaneous preterm birth at 16.0-31.9 weeks who subsequently carried singleton gestations. Cervical length, funneling (membrane prolapse greater than or equal to 5 mm), funnel shape, and dynamic changes were recorded at 16-18 weeks, and then every 2 weeks until 23.9 weeks. Managing obstetricians were blinded to the ultrasonography results. The primary outcome was gestational age at delivery. RESULTS: Five hundred ninety scans were performed in 183 women, of which 60 (33%) had funneling observed on at least one of the serial evaluations. These 60 women delivered at an earlier gestational age at delivery than the 123 women without funneling (31.7±7.9 weeks compared with 36.9±4.4 weeks; P<.001). In the 60 women with funneling on at least one evaluation, the progression over time of internal os cervical anatomy from a "T" to a "V" to a "U" shape was associated with earlier gestational age at delivery, whereas resolution of "V" shape funnels was associated with term delivery. Women with a shortened cervical length less than 25 mm (n=60) had a similar gestational age at birth with or without funneling (30.6±8.0 weeks compared with 31.9±6.6 weeks; P=.59). After controlling for the shortest observed cervical length, largest funnel percent was not a significant independent risk factor. CONCLUSION: The natural history of second-trimester funneling has significant variability and a significant association with earlier gestational age at delivery. As an independent finding, funneling does not add appreciably to the risk of early gestational age at delivery associated with a shortened cervical length. © 2007 The American College of Obstetricians and Gynecologists.