The recent clinical trials indicate that EA is feasible. The use of ancillary tools, such as ultrasound, requires only a slight modification of the previously established techniques for MA. The amniotic fluid can be obtained and cultured. Combining the reported populations sampled, the procedural and cytogenetic failure rate were 2.0% and 0.3% respectively. The necessary information for prenatal diagnosis in the situations listed previously can be obtained. The one exception to this would be the patient at increased risk for NTD. As noted above, diagnostic levels for AFAFP have been established beginning at 13 weeks gestation. Therefore, amniocentesis before 13 weeks gestation should not be considered an option for these patients. The majority of the findings in these "pilot" studies have been promising, but the issue regarding the safety of EA has not been answered. Harrison et al. theorized that the function of the physiology hydramnios was to "provide a distention growth stimulus to the uterus ... and conversely to share in the maintenance of uterine inhibition." The amniotic fluid is constantly being replaced, having a complete turnover approximately every 3 hours. Does the reduction in volume, for even a short period of time, change the "distensive forces" enough to increase the pregnancy loss rate? Later publications have associated MA with increased rates of congenital orthopedic anomalies and neonatal pulmonary compromise. The true existence of these complications continues to be debated in the literature. Two of the authors have indicated that none of the reported neonates exposed to EA have had such anomalies.(ABSTRACT TRUNCATED AT 250 WORDS)