will a randomised controlled trial detect them?
Controlled trials have been instigated, comparing the results and complications of chorionic villus sampling (CVS) and amniocentesis. At present most attention has focused on the miscarriage rate, which itself requires large numbers to provide studies of reasonable power. This paper describes a case of stillbirth due to massive placental abruption following CVS. The size of trial needed to provide sufficient power to detect a range of realistic postulated changes in the incidence of this or other important possible complications is calculated. It would appear that the size of sample required is very large, and that present trials will not succeed in detecting possible important differences in late obstetric, neonatal, and other long term complications of the procedure.
