Incidence and timing of pregnancy losses: Relevance to evaluating safety of early prenatal diagnosis

Abstract

Knowing the frequency and timing of pregnancy loss during normal gestation is integral to evaluating the safety of prenatal diagnostic techniques. That preclinical loss rates are high in humans has long been suspected, but in the past decade new data concerning these losses have become available. Cohort studies indicate that many women who show positive beta-HCG assays never show clinical evidence of pregnancy. Cytogenetic abnormalities have also recently been documented in 20% of ostensibly normal in vitro fertilization embryos. All the above are consistent with the sentinel studies of Hertig and Rock, who showed high frequencies of morphological abnormalities in preimplantation embryos. The frequency of fetal losses after clinical recognition of pregnancy is 12-15%; however, more sensitive (ultrasonographic) methods of detecting fetal demise now indicate that most clinically recognized pregnancies occur prior to 8-9 weeks, being retained in utero 2-3 weeks prior to expulsion. Loss rates are influenced by maternal age, smoking, alcohol, and other confounding variables that if not taken into account could yield spurious results concerning safety of prenatal diagnostic techniques. After 8 weeks gestation the likelihood of losing a viable pregnancy is only 3% and after 16 weeks only 1%.