Procedure-Related Risk of Miscarriage Following Amniocentesis and Chorionic Villus Sampling: A Systematic Review and Meta-Analysis

While it is important for pregnant women to understand the procedure-related risks of amniocentesis and chorionic villus sampling (CVS), which are both commonly performed invasive prenatal diagnosis procedures, current published literature offers inconsistent recommendations and estimated risks. These differences can be attributed to study design and ambiguity in the definition of pregnancy loss. The aim of this study was to accurately determine procedure-related risk of miscarriage following CVS or amniocentesis by estimating both the risk of miscarriage before 24 weeks' gestation in women who undergo one of these 2 procedures and also the background risk of miscarriage in women who do not undergo invasive testing in appropriately controlled studies. A total of 21 studies were selected for final use, after an electronic search of MEDLINE, EMBASE, CINAHL, and The Cochrane Library. Of those, 14 studied amniocentesis, and 7 studied CVS. The search was restricted to studies reported in English, published after the year 2000, and reporting results from at least 1000 invasive procedures. Miscarriage rates were estimated for both the control group and the group undergoing invasive procedures. In the 14 studies reporting on amniocentesis, 124,001 women had 1107 losses, corresponding to a miscarriage rate of 0.70% (95% confidence interval [CI], 0.50%-0.92%). The 771,965 women in the control groups had 6634 losses, with a pooled estimate of 0.70% (95% CI, 0.53%-0.90%). In 7 controlled studies, 324 women miscarried of the 42,716 undergoing amniocentesis (0.81% [95% CI, 0.58%-1.08%]), and 942 women miscarried of the 138,657 who did not undergo invasive procedure (0.67% [95% CI, 0.46%-0.91%]). The difference between the study and control group was not significant (P = 0.1435). In all reported CVS studies, there were 1186 losses in 53,890 women, corresponding to a loss rate of 2.35% (95% CI, 1.68%-3.16%), and a background rate was calculated as 2.26% (95% CI, 0.81%-4.41%). Meta-analysis of 3 controlled studies found no significant difference (P = 0.06385) in rate of miscarriage between women who did not undergo CVS and women who did (0.22% [95% CI, -0.71% to 1.16%]). The risks of attributable loss were found to be 0.1% for amniocentesis and 0.2% for CVS, which are lower than currently stated. These risks could possibly reflect the pregnancy characteristics of the women undergoing these procedures and not related to the invasive procedure itself. The rate of miscarriage could be higher in women undergoing CVS or amniocentesis by nonspecialists or by operators who only occasionally perform such procedures. Miscarriage rates should be reported to monitor unacceptable trends.