Pregnancy Loss Rates After Midtrimester Amniocentesis
Abstract & Commentary
By John C. Hobbins, MD, Professor and Chief of Obstetrics and Gynecology, University of Colorado Health Sciences Center, Denver, is Associate Editor for OB/GYN Clinical Alert.
Professor and Chief of Obstetrics, University of Colorado Health Sciences Center, Denver
Dr. Hobbins reports no financial relationship to this field of study.
Synopsis: There was no significant difference in loss rates between those undergoing amniocentesis and those not undergoing amniocentesis.
Source: Eddleman KA, et al. For the First and Second Trimester Evaluation of Risk (FASTER) Trial Research Consortium. Pregnancy loss rates after midtrimester amniocentesis. Obstet Gynecol. 2006;108:1067-1072.
The lead paper in the November issue of Obstetrics and Gynecology should stimulate lively discussion regarding the counseling of patients of advanced maternal age (AMA), and, if taken at face value, could impact the field of prenatal diagnosis, in general.
The study involved patients enrolled in the First and Second Trimester Evaluation of Risk for aneuploidy (FASTER) trial. Of the 35,003 patients being screened with first trimester nuchal translucency exams, first trimester biochemistry, and second trimester quad screening, a cohort of 3096 patients chose to have invasive sampling through amniocentesis. Because of the comprehensive study design, it was possible to have complete follow-up in these pregnancies through 23 weeks of gestation, as well as in the remaining 31,907, who were used as controls.
The loss rate in the amniocentesis group was 1%, compared with 0.94 % in the control group, giving an extremely low procedure-related risk of amniocentesis of 6 per 10,000. Based on the results, the authors felt that "these contemporary data will likely have a significant effect on how patients are counseled in current clinical practice," and they even challenged the "arbitrary" concept of offering amniocentesis to only those whose age-related risk for Down syndrome was similar to, or greater than, the commonly quoted amniocentesis risk figure of 1 in 200.
Commentary
If the risk of amniocentesis is really 6 per 10,000, then this would represent a risk that is less than the risk of Down syndrome in virtually every patient who is pregnant today. However, before scrapping the old paradigm, I would like to inject a note of caution. The study results, although being the latest published on amniocentesis risk, are completely out of sync with everything else in the literature. For example, Seeds scoured the literature for data on the risk of "contemporary" second trimester amniocentesis and came up with 29 studies that involved more than 1000 amniocenteses in each study. Five of the studies had control data for comparison, and pooled data from these studies showed a loss rate in the amnio group that exceeded controls by 0.6%. In the entire combined sample of 33,790 patients having amniocenteses there was a derived procedure related rate of 0.7%, when compared with the available control data. Both these figures were similar to the frequently quoted risk of 1 in 200.
It is granted that these figures were not based on randomized trials, but neither was the above FASTER trial randomized. In fact, the only randomized trial in the literature surfaced 20 years ago and involved over 4600 patients under 35 years of age, half of whom were randomly allocated to have ultrasound directed amniocentesis. The other half, after having an ultrasound scan demonstrating a normal fetus, were simply observed until 26 weeks of gestation. The loss rate in the amniocentesis group was 1.7%, compared with 0.7% in the control group, giving a procedure related loss rate of 1% — a result that had the same attention-getting effect as the FASTER trial results are sure to attract, but with opposite effect.
Last, another contemporary trial with very similar methodology as the FASTER trial did not have the same procedure related results. The Serum, Urine, and Ultrasound Screening Study (SURUSS) incorporated data from 47,000 patients, some of whom had amniocentesis. The very lengthy report was predominantly focused on the efficacy of screening through noninvasive methods. However, by digging into the report one will find that the loss rate to 24 weeks in the amniocentesis group was 1.8% vs. 0.79% in those who did not have procedures (with confidence levels that did not attain statistical significance). The resulting 1.0% procedure-related risk was far different than the FASTER trial results in a contemporary setting with similar patient numbers, experienced operators, and very similar methodology.
So, what's going on here? First, in order to be assured that a group having amniocentesis is no different than those who did not, the groups ideally would have to be randomized (a task that, probably, no one would like to take on today). Second, when dealing with a prevalence of fetal loss that is very low to start with, you need to have very large numbers of patients in each group to detect whether or not there is a true difference between groups. The addition of, let's say, only 7 more losses among the 3096 amniocentesis patients in the FASTER trial would change the procedure related loss rate to about 1 in 300.
To embellish further, an initial rendition of the FASTER trial data was presented at the Society of Maternal Fetal Medicine (SMFM) meeting in 2004. The abstract was based on a data set from the first 1605 amniocenteses analyzed. There were only 10 losses in this group, representing a rate of 0.62%. However, when the study was expanded to include 1491 more patients having amniocenteses, there were 21 losses, giving a raw loss rate in the new group of 1.4%, and a total loss rate in the final published amniocentesis group of 3096 patients of 1.0%. The reason the derived procedure-related risk remained low was that the addition of 5800 patients not having amniocentesis (representing only an increase in 20%) raised the spontaneous loss rate in the control group from 0.47% to 0.94%.
The point here is that a few changes in either column can confound the interpretation of this type of study.
Even those who have extensive experience with amniocentesis occasionally find their needle tips in areas where they do not exactly wish them to be, and virtually everyone has had instances in which a fetal loss or ruptured membranes was related in time to an amniocentesis. Yes — this could well have happened as a spontaneous loss, unrelated to the procedure, but now, to say that one can do up to 1600 amniocenteses, on average, before expecting one procedure-related loss, seems completely unrealistic, and goes against everything out there in the literature. A figure of 1 in 200 may be too high to quote for some centers, but it is hoped that this study will be put in proper perspective before a case is made for amniocentesis for all.
References
- Eddleman KA, et al. For the First and Second Trimester Evaluation of Risk (FASTER) Trial Research Consortium. Pregnancy loss rates after midtrimester amniocentesis. Obstet Gynecol. 2006;108:1067-1072.
- Seeds JW. Diagnostic mid trimester amniocentesis: how safe? Am J Obstet Gynecol. 2004;191:607-615.
- Tabor A, et al. Randomised controlled trial of genetic amniocentesis in 4606 low-risk women. Lancet. 1986;1:1287-1293.
- Wald N, et al. First and second trimester antenatal screening in Down's syndrome: the results in the Serum, Urine and Ultrasound Screening Study (SURUSS). Health Technology Assessment. 2003;7:1-77.
You have reached your article limit for the month. Subscribe now to access this article plus other member-only content.
- Award-winning Medical Content
- Latest Advances & Development in Medicine
- Unbiased Content