Randomised comparison of amniocentesis and transabdominal and transcervical chorionic villus sampling.

We have compared three methods of prenatal diagnosis in two large obstetric centres in Denmark. Women were randomly assigned transabdominal (TA) chorionic villus sampling (CVS), transcervical (TC) CVS, or second-trimester amniocentesis (AC); women at high genetic risk were randomised between the two CVS groups only. Analysis of 45 epidemiological variables showed the three procedure groups to be similar at enrollment. All women were followed up until completion of pregnancy. Among 3079 women at low genetic risk total fetal loss rates were 10.9% for TC CVS, 6.3% for TA CVS, and 6.4% for AC (p < 0.001). More women had bleeding after the procedure in the CVS groups (p < 0.001), whereas more amniotic fluid leakage (p < 0.001) was reported after AC. No uterine infections occurred in any group. No case of oromandibular-limb abnormality was seen in the CVS groups, but 1 child in the AC group had aplasia of the right hand. The two CVS approaches were compared among 2882 women at low and high genetic risk who were found to have cytogenetically normal fetuses. Rates of unintentional loss after the procedure were 7.7% for TC CVS and 3.7% for TA CVS (p < 0.001; 95% Cl of difference 2.3-5.8%). At baseline ultrasound scanning after establishment of optimum sampling conditions, more TC than TA procedures (p < 0.001) were judged not to be feasible. We found that TA CVS allows better access to the placental site than TC sampling, is an easier skill to acquire, and has the potential that more villi can be aspirated when needed. The risk of fetal loss is similar after TA CVS and AC. However, losses after AC are at a later stage and are therefore more distressing. TA procedures remain the first choice for prenatal diagnosis. Since, in our hands, TC sampling carries a greater risk to the fetus, we have abandoned TC CVS in our two study centres.