Amniocentesis before 15 weeks' gestation: outcome, risks, and technical problems.

Between January 1, 1979, and May 30, 1986, 4750 amniocenteses were performed by a single physician at the University of California, Davis, Medical Center. Of these procedures 541 (11.4%) were considered early amniocenteses (performed before the fifteenth week since last menstrual period). The medical records were reviewed for maternal age, amniocentesis indication, ultrasound findings, location of the placenta, location of the needle insertion (transplacental versus nontransplacental), color of the amniotic fluid, results of prenatal testing, complications of procedure, and pregnancy outcome. Complete follow-up data were possible for 308 subjects and obtained for 298 (96.8%). There were 11 miscarriages (3.6%), two stillbirths (0.7%), and one neonatal death (0.3%), resulting in a total postprocedural loss rate of 4.7% (14/298). If those patients with a preamniocentesis history of bleeding are eliminated, the total postprocedural rate of loss is 3.3% (7/211). Miscarriage within 2 weeks of amniocentesis occurred in 1.7% of procedures (5/298). No significant difference in pregnancy outcome was noted between transplacental and nontransplacental amniocentesis.