Objective: To assess the complications of third-trimester amniocentesis for fetal karyotyping in women unwilling to accept the fetal loss risks of second-trimester amniocentesis. Methods: Retrospective study of singleton pregnancies that underwent a third-trimester amniocentesis for karyotyping. 150 complete charts between 1998 and 2005 were reviewed. Results: The indications were: isolated abnormal second-trimester biochemical markers (n = 57), isolated maternal age >38 years (n = 46), integrated risk (maternal age, first-trimester nuchal translucency, second-trimester maternal serum markers) >1/250 (n = 22), history of chromosomal abnormality (n = 17) or maternal choice (n = 8). The median maternal age and gestational age at sampling were: 40 years (23–48), 32.4 weeks (29.7–37.1). Median interval between amniocentesis, definitive result of amniocentesis, and delivery were 14 days (7–42), and 49 days (10–67) respectively. There were no abnormal karyotypes and no termination of pregnancy. Six women out of 150 (4%) had spontaneous labor before 36 weeks (2% after 36 weeks). Conclusion: The risk of spontaneous labor before 37 weeks after late amniocentesis is 4% (2% before 36 weeks). This technique provides a late but safe reassurance to women who are unwilling to accept the risks of earlier fetal karyotyping. This is of interest to countries such as France where legislation permits late termination of pregnancy.