Since the introduction into clinical practice of midtrimester amniocentesis and first trimester chorionic villus sampling, the obstetrical risks associated with invasive testing have been the subject of multiple studies at the international level. Efficacy of the procedures has been defined in terms of fetal and maternal morbidity and mortality and, over the decades the obstetrical risks quoted to patients has significantly declined. Nevertheless, a quick scan on the internet of different obstetrical practices offering prenatal diagnostic testing reveals markedly wide differences regarding the quoted obstetrical risk of these invasive procedures. In fact, these differences should be viewed positively because they serve as a prototype for issues involving the obstetrical risks of diagnostic testing. It is imperative for the genetic counselor in discussing obstetrical risks of either amniocentesis or CVS to be knowledgeable of the clinical experience of each obstetrician performing such procedures. It is not appropriate to quote a risk from the literature. Counseling must be based on local experiences of specific physicians performing any form of invasive testing. It is interesting to point out that in the United States, exhaustive examination and evaluation by the American College of Pathologists is conducted on cytogenetic laboratories performing chromosome analysis of amniocytes and villi, but there is virtually no professional, external oversight of health professionals performing invasive testing with attendant risks to pregnant women and their fetuses.

In the past a risk of 1% over and above the natural fetal loss rate was quoted. This was based on the only randomized clinical trial of amniocentesis ever done. The procedures were carried out in a specific location by a limited number of physicians using a given needle size. It is well established that risk is associated with experience and needle size so that this single figure is misleading but it is all there was to go on. Meta-analysis of trials comparing amniocentesis with first trimester amniocentesis indicate that there is no material difference between the two types of procedures after allowing for the fetal losses occurring between the late first and early second trimester. Over time there was a sense – not derived from studies as such that the risk was lower than 1% so a figure of 0.5% was often quoted.

In addition to the single randomized trial, risk might be derived by comparing the fetal loss rate in women undergoing the procedure with those who don’t. Unfortunately this type of analysis is subject to strong statistical bias. Hence some groups of women who have the procedures are at higher a priori risk than those who do not. For example, advanced maternal age and abnormal serum biochemistry, factors used to select for the procedure are associated with increased loss. On the other hand among women selected for the procedure and actually go ahead with it are at lower a priori risk because of higher social class. And there is a specific reason for not going ahead: pre-procedure ultrasound shows that the fetus has already expired or there is retroplacental bleeding etc. There is direct evidence of such a bias in the FaSTERtrial where the fetal loss rate is statistically significantly lower in women with positive screening tests who had amniocentesis compared with those who didn't (Eddelman et al, Obstet Gynecol 2006;108(5):1067-72).

Given all these problems of interpretation, probably the best approach is for each physician who carries out large numbers of procedures is to quote the observed fetal loss rate in those having the procedure. The observed rate will include both iatrogenic and natural losses so is conservative (ie biased upward). It could be quoted as "in the last 1000 patients only x had miscarried after y weeks of follow up, so the loss rate due to amnio/CVS is probably less than z% as some would have miscarried anyway".