Share this article on:

Complications of Third-Trimester Amniocentesis Using Continuous Ultrasound Guidance

Gordon, Michael C. MD; Narula, Komal MD; O'Shaughnessy, Richard MD; Barth, William H. Jr. MD

Original Research

OBJECTIVE The objective of the study was to estimate the risks of third-trimester amniocentesis with continuous ultrasound guidance.

METHODS Cohort study. We reviewed the medical records of women who had an amniocentesis with continuous ultrasound guidance after 30 weeks' gestation at a single institution from January 1991 through December 1994. For procedures performed from January 1991 to February 1994, we obtained information from a chart review. From March 1994 to December 1994, we collected data prospectively. The primary outcome was whether or not there were any complications within 48 hours of the procedure. We also sought to determine any risk factors associated with complications.

RESULTS Complete records and data were available for 562 amniocenteses during the study period. The mean gestational age at the time of amniocentesis was 34.9 weeks. Of the 562 procedures, five (0.8%) were unsuccessful and 50 (9%) required more than one needle stick. The complication rate was 0.7% (95% confidence level (CI) = 0.02%, 1.9%). These included spontaneous labor in a preterm gestation (1), premature rupture of the membranes (1), placental abruption (1), and fetal–maternal hemorrhage (1). No patient required an emergency cesarean delivery and none suffered a perinatal death (95% CI 0, 0.8%). Complications were not associated with the number of needle sticks, the presence of bloody amniotic fluid, or the level of operator experience.

CONCLUSIONS Third-trimester amniocentesis performed with continuous ultrasound guidance has a high success rate and low risk for complications.

Third-trimester amniocentesis with continuous ultrasound guidance has a high success rate and a low complication rate.

Wilford Hall Medical Center, Lackland Air Force Base, Texas; and The Ohio State University College of Medicine, Columbus, Ohio.

Address reprint requests to: Michael C. Gordon, MD, Department of Obstetrics and Gynecology, Wilford Hall Medical Center, 859th MDOS/MMNO, 2200 Bergquist Drive, Suite 1, Lackland AFB, TX 78236–5300; E-mail:

The views expressed in this article are those of the authors and should not be construed as the official policy or position of the United States Air Force, the Department of Defense, or the United States Government.

Received June 12, 2001. Received in revised form September 26, 2001. Accepted October 18, 2001.

Amniocentesis is a common invasive procedure that is used for both diagnostic and therapeutic indications during pregnancy. Information regarding the risks and complications for second-trimester procedures is readily available.1,2 Less information is available regarding the risks of amniocentesis performed in the third trimester. Prior to the use of ultrasound, complications of third-trimester amniocentesis were common and included rupture of the membranes, infection, maternal hemorrhage, uterine vessel injury, fetal or placental hemorrhage, fetal injuries, and fetal distress.3 Most obstetricians now use continuous ultrasound guidance for amniocentesis, but few studies offer information regarding the type and frequency of complications associated with this procedure.4,5 The purpose of this study was to examine the success rate and complications associated with third-trimester amniocentesis when assisted with continuous ultrasound guidance. A secondary goal was to identify any factors that might predispose to complications of the procedure.

Back to Top | Article Outline


We conducted an observational study of all women who underwent a third-trimester amniocentesis from January 1991 through December 1994 at the Ohio State University Hospital. From January 1991 to February 1994, we performed a chart review using a medical record search for procedures performed in our hospital among inpatients and outpatients. From March 1994 to December 1994, data were collected prospectively using a simple checklist at the time of the amniocentesis. We followed these patients until delivery for any late complications. All amniocenteses performed after 30 weeks' gestation were included.

An obstetric resident, a maternal–fetal medicine fellow, or an obstetric staff member performed the amniocenteses. Clinicians performed all of the amniocenteses with continuous ultrasound guidance either by the operator themselves or with the assistance of another obstetrician (Figure 1). All procedures took place in the hospital, either on labor and delivery or in an adjoining antepartum testing suite, and all were performed using 20- or 22-gauge amniocentesis needles. Hospital staff performed a nonstress test after each procedure.



We obtained the following information for each amniocentesis: indication, gestational age at the time of the procedure, whether the needle traversed the placenta, the number of needle passes per procedure, the presence or absence of visible blood in the fluid specimen, the level of training of the operator, and the presence of any complications while monitored after the procedure or within 48 hours. For procedures performed after 37 completed weeks' gestation, we did not consider spontaneous labor or premature rupture of the membranes as complications.

We performed statistical analyses using either χ2 or Fisher exact test where appropriate. A P value <.05 was considered significant. Exact 95th percentile confidence intervals (CI) were calculated for all complication rates.6

Back to Top | Article Outline


We found complete medical records for 562 women who had an amniocentesis after 30 weeks' gestation between January 1991 and December 1994. These included 469 procedures from the retrospective medical record review and 93 from the 10-month prospective observation.

The mean gestational age (± standard deviation) at the time of amniocentesis was 34.9 ± 2.2 weeks, with a range from 30 to 42 weeks. Indications for performing the amniocenteses are listed in Table 1. Obstetric residents performed most of the procedures with supervision by obstetric staff or fellows. Third- or fourth-year residents performed 244, first- or second-year residents 184, and staff or fellows 134. Most commonly, residents performed the procedure assisted by other residents with indirect (hands-off) supervision by the fellow or staff.

Table 1

Table 1

The number of needle passes per amniocentesis ranged from one to four with a mean number of attempts of 1.1. Most (512, 91%) required only a single pass, whereas 50 (9%) required a second attempt. Eight women (1.4%) required three or four passes. We were unable to demonstrate a difference in the need for multiple attempts based on the level of training of the operator (resident versus staff/fellow, 10.5% versus 5.2%, P = .087). In addition, the specific postgraduate year (PGY) level of the resident or whether a fellow or staff performed the amniocentesis was not related (P > 05) to the mean number of attempts, the need for more than one attempt, or the incidence of blood-stained amniotic fluid (see Table 2). Only five of the 562 amniocenteses (0.8%, 95% CI 0.3%, 2.1%) were unsuccessful. Of these five unsuccessful amniocenteses, the procedure was abandoned after three attempts in one instance, two attempts in two instances, and after a single attempt in two instances. The reasons for the unsuccessful attempts were not well recorded in the charts. Only one complication occurred in women having unsuccessful amniocentesis (Table 3, case 3). All five unsuccessful attempts were performed by obstetric residents.

Table 2

Table 2

Table 3

Table 3

We obtained clear amniotic fluid in 490 (87%) of the procedures, bloody fluid in 55 (10%), and meconium-stained fluid in eight (1.4%). The color of the fluid was not recorded for four (0.7%) of the amniocenteses, and no fluid was obtained in the five unsuccessful attempts. The incidence of bloody amniotic fluid was higher if more than one needle stick was necessary (30% versus 8%, P = .001) or if the needle traversed the placenta (40% versus 3.6%, P = .001).

There were a total of four complications for a rate of 0.71% (95% CI 0.2%, 1.9%). Table 3 contains the details of these cases. All of the mothers and infants had a good outcome, and none had an emergency cesarean delivery. All four amniocenteses with complications were performed by senior level residents (PGY 3 and 4) with a single needle stick. We were unable to demonstrate a significant difference in the complication rates between resident and staff/fellow physicians (0.93% versus 0, P = .58) or in the presence of clear versus bloody amniotic fluid (0.41% versus 1.8%, P = .27). We could not determine whether the complications were associated with oligohydramnios or traversing the placenta because this information was not recorded for three of the four cases with complications. In cases 1 and 2, the amniotic fluid leakage and the preterm labor both occurred within 24 hours of the amniocentesis. The complications in cases 3 and 4 were first suspected due to the difficulty in obtaining the amniotic fluid and later confirmed to be due to a finding of abnormal fetal heart rate tracings immediately after the performance of the procedure.

No perinatal deaths or emergent cesarean deliveries for nonreassuring fetal heart rates occurred within 48 hours of amniocentesis (0, 95% CI 0, 0.8%). Among 468 cases with information regarding pregnancy outcomes beyond 48 hours, no patients experienced complications of the amniocentesis.

Back to Top | Article Outline


Before the use of ultrasonography, the risks of third-trimester amniocentesis were unacceptably high (0.4–10.8%) and included fetal injury, fetal death, and even, though rare, maternal death.3

Piiroinen and colleagues reported a 1.8% incidence of premature rupture of membranes, but no serious complications among 501 third-trimester amniocenteses performed immediately after ultrasound examination (but without the use of continuous real time guidance).7

Most obstetricians now perform amniocenteses with continuous ultrasound guidance. We preformed a medical literature search using OVID from the years 1970 to 2001 using the search term “amniocentesis” and could locate only two series documenting the risks of a third-trimester amniocentesis with continuous ultrasound guidance.4,5 Haeusler and colleagues reported no complications within 48 hours of an ultrasound-guided amniocentesis among 194 women with gestational diabetes between 24 and 35 weeks' gestation.5 Among 962 third-trimester procedures reported by Stark and coauthors, six (0.7%) required urgent delivery.4 In the latter study, staff obstetricians performed all of the procedures with continuous ultrasound guidance by an experienced sonographer. Although residents performed the majority of amniocenteses in our series, the complication rate was similar to that noted by Stark and colleagues. More importantly, no authors have reported a fetal death due to third-trimester amniocentesis when the procedure has been performed with continuous ultrasound guidance.

No patient in the current series required an emergency cesarean delivery (0, 95% CI 0, 0.8%) immediately after the amniocentesis. Significantly, of the four procedures with complications, all became apparent during the period of monitoring and observation immediately after the amniocentesis. In the series described by Stark and colleagues, six women (0.7%, 95% CI 0.16, 1.24) required an urgent delivery before completion of fetal maturity studies.4 Five of these required emergency cesarean delivery either for fetal heart rate abnormalities (n = 3), placental bleeding (n = 1), abruptio placentae (n = 1), or uterine rupture (n = 1). Although our data suggest that performance of a third-trimester amniocentesis may be safe in a setting where immediate delivery is not available, we still urge caution. Both studies describe complications with the potential to require an immediate delivery. Unfortunately, neither study identifies factors that could be used to determine which pregnancies are at higher risk for such complications.

Both our success rate (99.2%) and that of Stark and colleagues (98.4%), are higher than that reported by Piiroinen and coauthors (87%), where the amniocentesis was done without continuous ultrasound guidance.5,7 This suggests that continuous ultrasound guidance may improve the success rate of third-trimester amniocentesis. Continuous ultrasound guidance may also reduce the number of needle sticks needed for a successful amniocentesis. The average number of needle sticks in our study was 1.1 per patient, which compares favorably to the rate of 1.4–1.6 reported before the use of continuous ultrasound.7–9

Resident obstetricians performed the majority of procedures in the current series, but always with the supervision of staff or fellows. Our complication rate and rate of fetal loss was unaffected by the level of experience of the person performing the amniocentesis, but the power of this study to detect a significant difference is low. The level of experience also did not affect the number of attempts necessary for a successful amniocentesis or the incidence of a bloody tap. Our study gave us a 52% power to detect a significant difference in the need to perform more than one attempt when comparing staff/ fellows versus residents. We recognize the potential for confounding factors in this finding because more complicated amniocenteses were possibly performed by more senior level operators, and the amount of direct supervision of residents was difficult to discern. This low complication rate and high success rate may not apply to an operator without special expertise who performs the procedure without supervision from more highly trained physicians. However, unlike genetic amniocenteses for which complication rates are associated with the level of experience of the operator,10,11 our study suggests that special expertise in third-trimester amniocenteses may not be necessary, but additional studies are needed to confirm this finding. Although this is the second largest study to examine the risk of third-trimester amniocentesis, a much larger study would be needed to better define the risk factors due to the very low complication rate. For example, to achieve a power of 80% to detect a 50% decrease in the complication rate from those procedures performed by residents versus those performed by staff/ fellows or between the risks associated with clear or bloody amniotic fluid, we would need 4,000 subjects in each group. It is unlikely that such a study could be performed in a single institution.

As expected, the incidence of a bloody tap was significantly higher if the procedure required a second pass or if the needle traversed the placenta. The amniocenteses that resulted in a bloody tap, however, did not have a higher complication rate although, again, the power to detect a difference is low. We were unable to determine whether traversing the placenta resulted in a higher complication rate because this information was not recorded for three of the four patients with complications. Stark and colleagues noted that half of their patients with complications had required a transplacental amniocentesis.4 Consequently, we recommend avoiding the placenta when possible. If this is not possible one should take care to minimize lateral movements of the needle and avoid obvious overlying fetal vessels by passing the needle through the placenta at a site distant to the umbilical cord insertion.

In summary, our study demonstrates that third-trimester amniocentesis with continuous ultrasound guidance is safe and almost always successful. Obstetricians may use this information to counsel women regarding the risks and benefits of assessing fetal pulmonary maturity status with amniocentesis in the third trimester. Finally, the small but recognized risk of immediate complications supports a period of observation and fetal monitoring after the procedure.

Back to Top | Article Outline


1. The Canadian Early and Mid-Trimester Trial Group. Randomized trial to assess safety and fetal outcome of early and midtrimester amniocentesis. Lancet 1998;351:242–7.
2. Reece EA. Early and midtrimester genetic amniocentesis, safety and outcomes. Obstet Gynecol Clin North Am 1997;24:71–81.
3. Galle PC, Meis PJ. Complications of amniocentesis, a review. J Reprod Med 1982;27:149–55.
4. Stark CM, Smith RS, Lagrandeur RM, Batton DG, Lorenz RP. Need for urgent delivery after third-trimester amniocentesis. Obstet Gynecol 2000;95:48–50.
5. Haeusler MC, Konstantiniuk P, Dorfer M, Weiss P. Amniotic fluid insulin testing in gestational diabetes: Safety and acceptance of amniocentesis. Am J Obstet Gynecol 1998;179:917–20.
6. Fleiss J. Statistical methods for rates and proportions. 2nd ed. New York: John Wiley and Sons, 1981.
7. Piiroinen O, Erkkola R, Gronroos M. Low-risk amniocentesis in the third trimester under ultrasound control. Eur J Radiol 1984;4:309–11.
8. Teramo K, Sipinen S. Spontaneous rupture of fetal membranes after amniocentesis. Obstet Gynecol 1978;52:272–5.
9. Gordon HR, Deukmedjian AG. Suprapubic vs. periumbilical amniocentesis. Am J Obstet Gynecol 1975;12:287–90.
10. Anandakumar C, Wong YC, Annapoorna V, Arulkumaran S, Chia D, Bongso A, et al. Amniocentesis and its complications. Aust N Z J Obstet Gynaecol 1992;32:97–9.
11. Blessed WB, Lacoste H, Welch RA. Obstetrician-gynecologists performing genetic amniocentesis may be misleading themselves and their patients. Am J Obstet Gynecol 2001;184:1340–4.
© 2002 by The American College of Obstetricians and Gynecologists. Published by Wolters Kluwer Health, Inc. All rights reserved.