Comparison of first-trimester transvaginal amniocentesis with chorionic villus sampling and mid-trimester amniocentesis.

Between August 1989 and December 1991, 356 patients underwent first-trimester transvaginal amniocentesis (10-12 weeks). The same number of patients referred in the same period for mid-trimester amniocentesis (14-21 weeks) was matched also for maternal age and indication. A third group consisted of the first 356 cases in which chorionic villus sampling (CVS) was attempted. The overall success rate was 99.7 and 100 per cent for early and mid-trimester amniocentesis, respectively, and 97.2 per cent for CVS. The mean harvesting time was 12.8, 11, and 7.9 days, respectively. The percentage of patients rescheduled was 3.4 per cent in first-trimester amniocentesis, 1.7 per cent in mid-trimester amniocentesis, and 6.2 per cent in the CVS group. The early (less than 2 weeks) pregnancy loss was 1.7 and 0.6 per cent in early and mid-trimester amniocentesis, respectively, and 1.7 per cent in CVS. The total pregnancy loss was 3.2, 0.9, and 2.9 per cent, respectively. The rate of preterm birth was 6.0, 5.2 and 6.9 per cent, respectively. The results indicate that CVS has the shortest procedure-result interval, but the highest rescheduling rate. First-trimester amniocentesis has a higher procedure and laboratory success rate but, until otherwise proved, mid-trimester amniocentesis is the most efficient and safest procedure.