Résumé : Objectives: To evaluate in twin pregnancy the utility of non-invasive prenatal testing using circulating cell-free fetal DNA (cfDNA) in screening for the three main autosomal fetal trisomies. Methods: cfDNA testing was offered to 492 patients with a twin pregnancy without ultrasound anomaly as a first-line screening test or after routine serum screening. Data were collected prospectively and a retrospective analysis was performed. cfDNA analysis was performed by massively parallel sequencing. The fetal-fraction threshold used for test evaluation was 8%. Regression analysis was performed to investigate the effect on the test failure rate of maternal and pregnancy characteristics, and the performance of the test was also reported. Results: cfDNA analysis was performed as a first-line test (after the first-trimester scan) in 377 patients and following serum screening in 115. Of the 420 pregnancies for which outcome was available and cfDNA screening was assessed, 78.7% were dichorionic–diamniotic. The test failed on the first attempt in 12 (2.9%) pregnancies, and regression analysis demonstrated that only maternal weight was a significant independent predictor of test failure. A result was subsequently achieved in the 10 cases for which a second sample was obtained. cfDNA analysis identified all three cases of trisomy 21 and the only case of trisomy 18. For trisomy 21, the specificity was 99.8% (95% CI, 98.7–100.0%). When considering pregnancies according to whether they were conceived spontaneously or after assisted reproductive technology, there were no significant differences in terms of maternal weight or no-result rate for cfDNA screening between these two groups. Conclusions: In twin pregnancy without fetal ultrasound abnormality, cfDNA screening for trisomies 21, 18 and 13 had a high success rate and good performance. Therefore, in routine practice, cfDNA analysis could be considered as a first- or second-line screening test. Copyright © 2017 ISUOG. Published by John Wiley & Sons Ltd.