Literature concerning early genetic amniocentesis, obtained before 14-15 weeks of gestation, has been examined and compared. Key figures of post-procedural spontaneous fetal losses have been compared as well as cytogenetic data from the published series. Cytogenetically more samples fail to grow in the early weeks, but the diagnostic accuracy seems to be comparable with that of amniocentesis obtained in the later weeks. A sampling method, improving the cell yield by means of filtration, has been developed in order to lower the risk of culture failure and reduce the reporting time in the gestational weeks 11 to 13. The filter technique is under evaluation in a randomized trial. Detection of neural tube defects with early amniotic fluid sampling is possible, but whether the detection rate is comparable to that of standard amniocentesis remains to be seen. Although the safety of early amniocentesis has not yet been evaluated in a randomised study, data from almost 5,000 cases suggest that the procedure related loss rate is approximately 1%, and thus the same as for standard amniocentesis and chorionic villus sampling.