Special Feature: Amniocentesis Today — Is It Really A Benign Procedure?
Special Feature
Amniocentesis Today—Is It Really A Benign Procedure?
By John C. Hobbins, MD, Professor and Chief of Obstetrics, University of Colorado Health Sciences Center, Denver, is Associate Editor for OB/GYN Clinical Alert.
Over the last five years, there has been an increased emphasis on screening for aneuploidy, in part due to the ACOG endorsement of the concept of offering nuchal translucency (NT) and biochemical screening to all pregnant patients, and not just those of advanced maternal age (AMA). During the same time period, the numbers of women who were the original recipients of screening attention, those of AMA, increased appreciably. For example, in 1985 only 5% of pregnant women were over 35 years of age, while now it is estimated that that figure approaches 20%.
While noninvasive screening has enabled many AMA patients to avoid invasive sampling, many younger patients, whose risk for trisomy 21 increased above their age-related risk because of the screening, are now weighing the risks and benefits of invasive testing. This is why it is so important to revisit the topic of amniocentesis morbidity.
Older studies
Long ago, the threshold above which a patient was considered to be of "advanced maternal age" was based on vintage data suggesting the risk of amniocentesis to be about 1 in 200, similar to the risk (1 in 280) of a 35- year-old woman. Interestingly, until very recently, most nonrandomized studies yielded results that were quite similar to the often quoted risk related figure of 0.5%. For example, Seeds gathered data from 29 controlled and non-controlled studies and found a raw loss rate of 2.1% and an extrapolated procedure-related loss rate of 0.7%.1
The only randomized trial in the literature came from Denmark.2 The study involved over 4600 patients low risk patients (under 35 years of age), half of whom had amniocenteses under ultrasound guidance, and half not. The study was conducted between 1980 and 1984. The results indicated a procedure-related risk of 1% (1.7% vs 0.7%).
Contemporary studies
The most current meta-analysis of amniocentesis risk was published by Mujezinovic and Alfirevic.3 Twenty-nine articles from 1995 until now were incorporated into their study. The pooled data showed a fetal loss rate within 14 days of amniocentesis to be 0.6% and prior to 24 weeks, 0.9%. Five studies provided control data which showed a relative risk of fetal loss of 1.45. The authors seemed to concentrate on the total loss rate because of the bias associated with the nonrandomized controlled data. However, deep in the discussion section was their estimated procedure-related loss rate of 0.6% (1 in 170), which they pointed out as being lower than the 1% figure in the Danish randomized trial.
Last year, a report emerged which represented a spin off study from the well-known FASTER trial in which outcome data from a cohort of 3096 patients having amniocenteses were compared with those from over 30,000 patients not having this procedure.4 The difference in the loss rates up to 24 weeks between groups was only 1 in 1600, which was out of sync with everything else in the literature. Not surprisingly, this set off some very lively debate regarding how to counsel patients.
Caughey reviewed 16 years worth of experience with amniocentesis at one center and found that the overall fetal loss rate in the 30,893 patients having the procedure was 0.83%.5 Data from a small group of controls (1292) were generated from the same center as part of the FASTER investigation. These patients had a spontaneous loss rate of 0.37%, giving a 0.46% procedure related loss rate. It is of note, however, that the authors indicated that in the last 5 years of the study the extrapolated procedure-related risk dropped to 0.27% (1 in 370).
In the latest issue of Obstetrics & Gynecology, another single center report emerged involving data from 1990 to 2006. Obido at al analyzed information from 11,746 patients who had amniocenteses and 39,811 who did not.6 The authors were interested in fetal death rate and miscarriage rate, which was 0.4%, and the total fetal loss prior to 24, which was 0.97%. The non-amnio group had spontaneous loss rates of 0.26% and 0.87%, respectively. The authors then concluded that there was an apparent procedure-related risk of only 1 in 769.
Comment
Even though one can find fault, for a variety of reasons, with every one of the above studies, the condensed information above is all that we have. The older studies are just that—old, and the amniocenteses were performed when ultrasound equipment and operator experience were, in some cases, not up to today's standards. The newest studies contain control groups that may not be ideal for an "apples to apples" comparison. Also, seemingly large studies still do not have enough numbers to provide the statistical clout to address the real risk of amniocentesis. For example, the FASTER trial results were first reported at the annual meeting of the Society of Maternal-Fetal Medicine in 2004 based on data from 1600 amniocenteses and 26,000 controls. The amniocentesis group had 10 losses yielding a rate of 0.62%. When 1500 more amniocenteses were added in their final publication, there were 21 additional losses, yielding a loss rate in the later group of 1.4%. However, the additional 20% more patients added to the control group for the final publication also had a surprisingly high rate of losses, raising the total loss rate from 0.47% to 0.97%. Therefore, with a more than doubling effect on both study and control patients, the final difference between groups was insignificant. In other words, only a few extra losses in either column can skew the statistics appreciably when the prevalence of fetal loss is so low to start with.
Last, and perhaps the most important drawback of the most recent study and the FASTER trial data is the fact that there is an inherent loss rate of about 40% in aneuploid fetuses prior to 24 weeks. It is assumed that many pregnancies in the amniocentesis group ended in termination of pregnancy before some aneuploid fetuses, who might have been destined to abort spontaneously, did so. This would alter the relationship between the numerator (amniocentesis) and the denominator (controls) to benefit the amnio group, thus artificially lowering the procedure-related loss rate.
Unfortunately, it is unlikely that a proper randomized trial will be possible to launch today to contain enough numbers to have any meaning, and we will have to deal with existing, yet "iffy," data that give us an unmanageable fetal loss rate range of between 1 in 100 and 1 in 1600. If the outliers were removed, one would arrive upon a rough middle ground—a procedure related rate of about 1 in 400, which is close to the number estimated in the Caughey report. Recently, we have been quoting our patients a risk range for amniocentesis of between 1 in 200 and 1 in 400.
Based on the FASTER trial data some providers have even recommended that the word "invasive" be removed from counseling sessions dealing with amniocentesis. However, I would argue that most women about to have any needle placed into their uterus would take issue with this procedure being labeled as "non-invasive." Many women would gladly have this procedure performed in order to obtain information that is virtually 100% accurate regarding the karyotype of their fetuses, while other women would not want to take any risk to find this out —especially if their management would not be altered appreciably by the amniocentesis results. It is the latter group, and those in the large middle, "undecided," category, who have been seeking reasonably accurate reassurance through noninvasive screening methods, with some reported sensitivities well over 90%. However, once these patients have a single numbered risk for trisomy 21 and 18, they often expect, for comparison, a similarly clean numerical risk for amniocentesis. Coming in with a "range" can certainly make counseling more complicated, but life is not perfect, and this is the best we can offer in the absence of concrete data.
References
- Seeds, JW. Diagnostic midtrimester amniocentesis: how safe? Am J Obstet Gynecol. 2004;191:607-615.
- Tabor A, et al. Lancet. 1986;1:1287-1293.
- Mujezinovic F, Alfirevic Z. Obstet Gynecol. 2007;110:687-694.
- Edlelman KA, et al; for the First and Second Trimester Evaluation of Risk (FASTER) Trial Research Consortium. Pregnancy loss rates after mid-trimester amniocentesis. Obstet Gynecol. 2006;108:1067-1072.
- Caughey AB, et al. Obstet Gynecol. 2006;108;612-616.
- Odibo AO, et al. Obstet Gynecol. 2008; 111:589-595.
Subscribe Now for Access
You have reached your article limit for the month. We hope you found our articles both enjoyable and insightful. For information on new subscriptions, product trials, alternative billing arrangements or group and site discounts please call 800-688-2421. We look forward to having you as a long-term member of the Relias Media community.