Background: Demand for early prenatal diagnosis of genetic abnormalities is anticipated to increase. For many families, early diagnosis is critical because elective termination would only be considered during the first trimester. Fetal karyotyping or other genetic testing usually requires amniocentesis or chorionic villus sampling (CVS). Based on studies performed about 20 years ago (soon after these tests were first introduced), CVS was reported to be associated with an increased rate of fetal loss compared with amniocentesis. The difference was reported to be lower, but still significant, in procedures performed before 20 weeks' gestation. For both procedures, results are greatly influenced by physician experience. Caughey and colleagues conducted a study to determine current rates of fetal loss in patients undergoing CVS and amniocentesis compared with those who did not, as well as if the difference in loss rate between CVS and amniocentesis has changed over the past two decades.

The Study: The authors performed a retrospective cohort study of women who presented to a major U.S. medical center from 1983 to 2003 for prenatal diagnosis because of advanced maternal age, elevated risk on triple screening, or request for prenatal diagnosis. Cases of multiple gestation, fetal aneuploidy, or major fetal anomaly were excluded from the study. Experienced practitioners performed all procedures. The primary outcome was pregnancy loss before 24 weeks' gestation. Follow-up rates were 98.5 percent for CVS patients, 98.1 percent for amniocentesis, and 91.8 percent for patients who received second trimester counseling and ultrasonography but no invasive diagnostic procedure.

Results: The 9,886 women who underwent CVS were older and more likely to be white and multiparous compared with the 30,893 women who had amniocentesis. As anticipated, the overall pregnancy loss after CVS (3.12 percent) was significantly higher than the loss after amniocentesis (0.83 percent). The rates of fetal loss declined over time for both procedures, but the rate of decline was significantly greater for CVS. Between the periods 1983 to 1987 and 1998 to 2003, fetal loss rates for CVS fell from 4.36 to 1.93 percent. For amniocentesis, the corresponding fall was from 0.91 to 0.64 percent. When multivariate logistical regression was used to compare the loss rate after the two procedures over time, the odds ratio decreased and became statistically and clinically nonsignificant for the period 1998 to 2003.

Conclusions: The authors conclude that the danger of fetal loss following amniocentesis or CVS has declined in recent years, and that the loss rate for CVS is no longer significantly higher than that for amniocentesis. The authors believe these results will assist families and health care professionals in making difficult decisions about antenatal diagnoses, especially during the first trimester.