Share this article on:

Umbilical Artery Doppler Flow Velocity Waveforms After Transplacental Amniocentesis

Haugen, Guttorm MD, PhD; Helbig, Anne MD; Husby, Henrik MD

Original Research

OBJECTIVE To assess the influence of transplacental versus nontransplacental needle passage during genetic amniocentesis on umbilical artery (UA) pulsatility index (PI) and fetal heart rate (FHR).

METHODS Genetic amniocentesis was performed in 205 women with no major fetal malformations detected by prenatal ultrasound at a median gestational age of 14 weeks and 3 days (range 13 weeks and 1 day to 18 weeks and 6 days). Chromosomal abnormalities were observed in five fetuses. These pregnancies were excluded from further analyses. The study group consisted of 56 of the remaining 200 women in whom amniocentesis had been performed transplacentally. As controls two patients with nontransplacental needle passage were chosen for each woman in the study group, matched for gestational age (± 3 days) and as far as possible for the indication for amniocentesis. The UA PI and the FHR were measured immediately before and after the amniocentesis.

RESULTS Amniocentesis did not cause significant changes in UA PI and FHR within or between the two groups. Division of the study population into three subgroups dependent on gestational age did not alter the results. Pregnancy outcome was similar in the two groups.

CONCLUSION Transplacental needle passage during amniocentesis did not induce any changes in UA PI or FHR relative to a group with nontransplacental amniocentesis.

Transplacental needle passage during genetic amniocentesis does not induce any significant changes in umbilical artery pulsatility index or fetal heart rate.

Department of Obstetrics and Gynecology, Rikshospitalet University Hospital, Oslo, Norway.

Address reprint requests to: Guttorm Haugen, MD, PhD, The National Hospital (Rikshospitalet), Oslo N-0027, Norway; E-mail:

Received May 20, 2002. Received in revised form October 28, 2002. Accepted November 7, 2002.

Earlier reports have shown a significant increase in the incidence of spontaneous abortions in relation to second-trimester amniocentesis after transplacental needle passage relative to after nontransplacental needle passage. 1,2 It is conventional to avoid the placenta if possible during the procedure. The basis of this policy has been challenged by studies showing no difference in the frequency of spontaneous abortions after transplacental needle passage versus nontransplacental needle passage in both second trimester 3,4 and early amniocentesis. 5,6 Though transplacental amniocentesis does not seem to cause any major threat to pregnancy, a possible physiologic or pathophysiologic influence on fetal and umbilicoplacental circulation cannot be excluded.

During the last decade, analyses of first- and early second-trimester fetal Doppler blood flow velocity waveforms have been used in different studies as a predictive parameter in the identification of aneuploid fetuses. 7,8 Only a few studies have used such waveforms to investigate the influence of second-trimester amniocentesis per se on fetal and umbilicoplacental circulation. Weinraub et al 9 showed an increase and Martinez et al 10 no significant alterations in the resistance indices of the umbilical artery (UA) after genetic amniocentesis. These studies were not designed to observe possible alterations in Doppler blood flow dependent on whether the procedure was performed transplacentally or not.

The primary aim of the present study was to assess the influence of transplacental as compared with nontransplacental needle passage during amniocentesis on the Doppler flow velocity waveforms in the UA and fetal heart rate (FHR). Procedure-related complications and pregnancy outcome are also reported.

Back to Top | Article Outline


The study was performed from January 1998 to July 2000 on women with singleton pregnancies between 13 and 20 weeks' gestation scheduled for genetic amniocentesis. The study was performed prospectively to investigate the influence of transplacental needle passage on UA pulsatility index (PI) and FHR during genetic amniocentesis. Inclusion criteria were willingness of the women to participate in the study, ability to perform the procedure given the clinical workload, and no major malformations identified sonographically. During the study period about 1400 amniocenteses were performed (excluding pregnancies with major malformations). Approximately 85% of the women were excluded because of the clinical workload.

The study group comprised women with transplacental amniocentesis. As controls, two patients with non-transplacental needle passage were chosen for each woman in the study group, matched for gestational age (± 3 days). The groups were matched for gestational age because of the decline in UA PI and FHR throughout the early second trimester. 11 To reduce possible confounding variables on pregnancy outcome the groups were also, as far as possible, matched for the indication for amniocentesis. If more than two controls could be matched, the controls were selected at random. The matching was not performed on a one to one basis. The gestational age at amniocentesis was calculated by biometry using biparietal diameter and the current national standards of reference. 12

We assumed a standard deviation (SD) for the PI values during the gestational age of current interest of 0.20. 10 With a two-sided α of .05 and power of 0.85 a total of about 165 women were required to detect a PI difference of 0.10 using the present model of two controls per woman in the study group. For the FHR, for the same number of women and assuming an SD of eight beats per minute, 10 a difference of four beats per minute would be detected at the same level of significance and power.

The Doppler velocity waveforms were obtained using Acuson 128XP (Acuson, Mountain View, CA), Acuson Sequoia, or Vingmed System Five (Vingmed, Horten, Norway). The same machines were used for real-time sonographic monitoring of the amniocentesis. Probes with Doppler carrier frequencies of 3.5–5.0 MHz were used. For Acuson 128XP the spatial peak temporal average intensity was less than 50 mW/cm2. The output intensity from Vingmed System Five and Acuson Sequoia was monitored as mechanical and thermal indices on the display.

The amniocenteses were performed by a freehand technique under ultrasound guidance using a needle with an outer diameter of 0.9 mm. The sample volume of amniotic fluid in milliliters was equal to gestational age in weeks. All patients had only one needle insertion. To avoid fetal bleeding during transplacental needle passage color Doppler was used to visualize chorionic vessels.

The ultrasound Doppler waveforms were obtained from a free-floating loop of the medial part of the cord at an insonation angle as low as possible but always below 30° and in the absence of fetal movements. Fetal heart rate was measured by the M mode of cardiac ventricular motion. The examinations were performed by two of the authors, and the investigator who performed the amniocentesis also performed the ultrasound measurements immediately before and after the procedure. Umbilical artery PI values and FHR were measured as the average of three consecutive heart cycles. The recordings were stored as hard copies. The intra- and interobserver variabilities for the UA PI measured as the coefficients of variation (n = 12) were 4.9% and 6.5%, respectively.

The interval from the end of the sonographic examination before amniocentesis to the end of that performed after the procedure was measured in all women. The interval from the end of the amniocentesis to the post-procedure Doppler measurements was not recorded. To estimate this interval we measured the time used to perform the actual amniocentesis in a separate group of women with transplacental (n = 17) or nontransplacental (n = 34) needle passage. The different time intervals were measured to the nearest half minute.

Follow-up data on the pregnancies and for the neonatal period were obtained by a written questionnaire or telephone interview of the women and from hospital records. The study was approved by the Regional Ethics Committee, and informed consent was obtained from the women in the main study and in the separate study on the interval.

The results were analysed using Statistical Package for Social Sciences 10.0 (SPSS Inc., Chicago, IL). The results of the ultrasound Doppler waveforms and heart rate measurements were analyzed by paired t test for differences within the groups and with unpaired Student t test for differences between the groups. Further analyses of the primary outcome variables were performed by the Levene test for equality of variance, χ2 test, bivariate correlation (Spearman r s), and multiple linear regression. The Mann–Whitney U test and χ2 test were used for analyses of the clinical data. A P value less than .05 was considered statistically significant.

Back to Top | Article Outline


Two hundred five women participated in the study. Indications for amniocentesis were advanced maternal age (38 years or older at estimated date of delivery, according to national policy) (n = 136), earlier child with trisomy (n = 18), maternal anxiety (n = 5), use of antiepileptic drugs (n = 8), or sonographic soft markers of chromosomal abnormalities or minor malformations (n = 38). Chromosomal abnormalities were observed in five of the 205 pregnancies. These fetuses were excluded from further analyses.

The study group consisted of the 56 women in whom amniocentesis was performed transplacentally. The control group comprised 112 of the remaining 144 women. Clinical details for the two groups are presented in Table 1.

Table 1

Table 1

Twenty of the pregnancies with ultrasonographic markers of aneuploidy or minor malformations were included in the final analysis. These observations were made at a routine scan at about 18 weeks except in one fetus with increased nuchal translucency (3.3 mm) at 12 weeks' gestation. Of these 20 pregnancies eight had more than one soft marker or minor malformation. The sonographic observations included choroid plexus cysts (n = 8), renal pelvis dilatation of 7 mm or more (n = 4), increased nuchal translucency (n = 1), nuchal fold greater than 6 mm (n = 1), echogenic intracardiac foci (n = 2), hyperechogenic bowel (n = 2), sandal gap (n = 2), cleft lip or palate (n = 2), talipes (n = 4), micrognathia (n = 1), and localized edema on the front of the thorax (n = 1). After birth no major or other minor malformations were observed in these 20 neonates.

The results concerning the influence of amniocentesis on UA PI values are presented in Table 2. No significant differences were observed between the two groups before or after amniocentesis, and there were no significant alterations within each group after the procedure. Moreover, no significant differences were observed in the SDs of the PI levels within or between the groups. The percentage of fetuses with a reduction, no change, or an increase in the PI values after the procedure did not significantly differ between the two groups. As shown in Table 3, there was no difference in FHR between the groups, and amniocentesis did not induce changes in FHR within or between the two groups.

Table 2

Table 2

Table 3

Table 3

The UA PI values before amniocentesis were negatively correlated with gestational age (r s = −.41, P < .001). The changes in PI values after amniocentesis (values after minus values before the procedure) were not correlated with gestational age (r s = −.13, P = .09) but negatively correlated to the UA PI values before amniocentesis (r s = −.26, P = .001). In a multiple linear regression analysis with the changes in PI values as the dependent variable, the regression coefficients ± standard error (SE) were 0.031 ± 0.024 for transplacental needle passage (nonsignificant, P = .20), −0.005 ± 0.001 for gestational age in days (P = .001), and −0.270 ± 0.053 for PI values before amniocentesis (P < .001). The adjusted R 2 for the model was 0.14. Table 4 presents the results of the changes in PI values dividing each group into three subgroups based on gestational weeks.

Table 4

Table 4

Fetal heart rate before amniocentesis was negatively correlated with gestational age (r s = −.16, P = .05). The changes in FHR after amniocentesis (values after minus values before the procedure) were positively correlated with gestational age (r s = −.16, P = .05), and negatively correlated to the FHR before amniocentesis (r s = −.32, P < .001). In a multiple linear regression analysis with the changes in FHR as the dependent variable the regression coefficients ± SEs were 0.365 ± 0.799 for the method of needle passage (nonsignificant, P = .65), 0.038 ± 0.040 for gestational age in days (nonsignificant, P = .34), and −0.229 ± 0.057 for the FHR before amniocentesis (P < .001). The adjusted R 2 for the model was 0.09.

The interval from the end of the sonographic examination before amniocentesis to the end of that after amniocentesis statistically differed (P = .02), with a median time of 10.0 minutes (range 4.5–20.0) and 8.5 minutes (5.5–14.5) in the study and control groups, respectively. The changes observed in PI values and FHR after the procedure were not correlated to this interval. In the separate study on the time used on the actual amniocentesis the median time was 4.5 minutes (2.0–7.5) versus 3.5 minutes (3.0–7.0) in the group with transplacental and nontransplacental needle passage, respectively (P = .09).

None of the patients with transplacental needle passage experienced any major complications after amniocentesis. In the control group one woman aborted spontaneously at 20 weeks' gestation. Another patient was hospitalized for 1 week because of watery discharge the day after amniocentesis, after which she had an uncomplicated pregnancy. The frequency of preterm delivery (less than 37 weeks) was six of 56 (10.7%) and five of 111 (4.5%) (P = .18) in the study and control groups, respectively. In the study group three of the women delivered between gestational weeks 30 and 34, whereas all other preterm deliveries occurred after week 34. Median gestational age at delivery was 39 weeks plus 4 days in both groups (range 30 plus 5 to 41 plus 6 in the study group and 35 plus 0 to 42 plus 5 in the control group) (P = .26). Median birth weights were 3620 g (range 1550–5030) and 3680 g (range 2250–5200) (P = .18) in the study and control groups, respectively. Craniosynostosis was diagnosed in two neonates (both nontransplacental group), one infant was treated for epilepsy at the age of 2 months (transplacental group), and three were treated for hip dysplasia (two in the transplacental group and one in the nontransplacental group). In these six pregnancies the amniocenteses were performed at a gestational age of 13 weeks plus 5 days to 15 weeks plus 1 day because of advanced maternal age.

Back to Top | Article Outline


Theoretically, transplacental needle passage might induce the release of vasoactive substances that could influence peripheral vascular impedance and cardiac function in the fetus. Serotonin and thromboxane A2 will be released in the case of clot formation. These substances are known to constrict the umbilicoplacental vessels. 13,14 We did not demonstrate any difference in the measured hemodynamic parameters dependent on whether amniocentesis was performed transplacentally or not. This is in accord with a smaller study on third-trimester amniocentesis. 15 Although the present study cannot exclude the release of and the effects of vasoactive substances after transplacental amniocentesis, they did not have any influence on the UA Doppler velocity waveforms or on FHR.

The interval from the end of the sonographic examination before amniocentesis to that after amniocentesis was longer in the transplacental group than in the non-transplacental group. As shown in the separate measurements on time expenditure the difference is mostly due to a longer time used to perform the actual transplacental amniocentesis. Thus, the interval from amniocentesis to the following Doppler measurements is assumed to be quite similar in the two groups. Weinraub et al 9 investigated the change in the UA systolic–diastolic ratio (S/D) at different intervals after amniocentesis. The ratio was slightly increased 10 minutes after the procedure, and a further increase was observed after 2 hours. These increases were abolished after indomethacin treatment, and the authors concluded that the changes in S/D were due to uterine contractions. Because of fetal movements we found it impossible to perform Doppler measurements at an exact interval after amniocentesis. Allowing 2–7.5 minutes for performance of the amniocentesis after the initial sonographic examination, most of our measurements were obtained within the first 5 to 6 minutes after amniocentesis. In vitro perfusion studies on human UAs have shown a maximum vasoconstrictory response 3–5 minutes after serotonin administration. 14 An interval of 5 to 6 minutes will thus coincide with a possible acute effect after the release of vasoactive substances from the placenta after transplacental needle passage, and such an effect would not be disturbed by an effect due to uterine constrictions. We did not observe any differences in PI values or FHR dependent on the measured interval in either group. However, this was not a primary aim of our study, and an effect after a longer time interval cannot be excluded.

After chorionic villus sampling (CVS), Zoppini et al 16 observed a larger variation in FHR and UA PI relative to a control group without any invasive procedure. In the present study there were no significant differences between the two groups either in the SDs of the changes observed or in the percentages of fetuses with an increase or a decrease in the measured values after amniocentesis. However, from the present study we can only make inferences concerning transplacental versus nontransplacental needle passage, and not of amniocentesis as such because we did not include any control group without invasive procedures. Chorionic villus sampling is usually performed at an earlier gestational age than the present study and has to be regarded as a more traumatic procedure to the placenta than amniocentesis. In accordance with that assumption Martinez et al 17 observed a significant increase in UA PI values after CVS at a gestational age of 11 weeks or lower, whereas no changes were observed at a later gestational age. In the present study linear regression revealed a significant influence of gestational age on the changes in UA PI. Amniocentesis might thus represent different physiologic challenges to the fetus dependent on duration of pregnancy. However, the gestational age of the study population was skewed, with fewer pregnancies at higher gestational ages, and further speculations do not seem warranted.

The changes in UA PI and FHR were negatively correlated to UA PI and FHR before amniocentesis, respectively. This implies that high values before amniocentesis were related to a decrease and low values to an increase in the respective variables after the procedure. This phenomenon might be due to the normal variability of UA PI and FHR, which are known to increase during the gestational age represented in the present study. 18 However, to investigate this issue a control group without amniocentesis has to be included.

Besides CVS, cordocentesis has also been observed to induce changes in fetal hemodynamic variables. Third-trimester cordocentesis caused a decrease in FHR and UA PI, whereas in the same study no changes were observed after amniocentesis performed at the same gestational age. 15 Other studies have also shown a decrease in UA PI after cordocentesis that was similar 19 or larger 20 after transplacental versus nontransplacental needle passage. Generally, in the third trimester puncture of the umbilical vein wall seems to induce larger vasoactive changes than transplacental needle passage.

After birth, the included fetuses did not show malformations that might have interfered with the measured values other than those observed at the time of amniocentesis. Especially, no major cardiovascular malformations were observed. Two infants had craniosynostosis. To our knowledge no reports have indicated any relation to amniocentesis. Only one CVS series has reported the occurrence of craniosynostosis after such a procedure. 21 Three neonates (1.8%) were treated for hip dysplasia. Earlier and larger series have reported a lower incidence both in early and in second-trimester amniocentesis, 22 and the incidence has been shown to be unaltered relative to a control group without amniocentesis. 1 The frequency of serious pregnancy complications was within the expected range. One spontaneous abortion occurred in the control group, and the rate of preterm deliveries was nonsignificantly larger in the study group. However, the present study is too small to make any further inference on clinical outcome after transplacental versus nontransplacental needle passage.

In conclusion, transplacental needle passage during amniocentesis did not induce any significant alterations in UA Doppler flow velocity waveforms or in FHR relative to a group with nontransplacental amniocentesis.

Back to Top | Article Outline


1. Tabor A, Madsen M, Obel EB, Philip J, Bang J, Nørgaard-Pedersen B. Randomised controlled trial of genetic amniocentesis in 4606 low-risk women. Lancet 1986;i:1287–93.
2. Porreco RP, Young PE, Resnik R, Cousins L, Jones OW, Richard T, et al. Reproductive outcome following amniocentesis for genetic indications. Am J Obstet Gynecol 1982;143:653–60.
3. Crane JP, Kopta MM. Genetic amniocentesis: Impact of placental position upon the risk of pregnancy loss. Am J Obstet Gynecol 1984;150:813–6.
4. Bombard AT, Powers JF, Carter S, Schwartz A, Nitowsky HM. Procedure-related fetal losses in transplacental versus nontransplacental genetic amniocentesis. Am J Obstet Gynecol 1995;172:868–72.
5. Bravo RR, Shulman LP, Phillips OP, Grevengood C, Martens PR. Transplacental needle passage in early amniocentesis and pregnancy loss. Obstet Gynecol 1995; 86:437–40.
6. Tharmaratnam S, Sadek S, Steele EK, Harper MA, Nevin NC, Dornan JC. Transplacental early amniocentesis and pregnancy outcome. Br J Obstet Gynaecol 1998;105:228–30.
7. Martinez Crespo JM, Comas C, Ojuel J, Puerto B, Borrell A, Fortuny A. Umbilical artery pulsatility index in early pregnancies with chromosome anomalies. Br J Obstet Gynaecol 1996;103:330–4.
8. Matias A, Gomes C, Flack N, Montenegro N, Nicolaides KH. Screening for chromosomal abnormalities at 10–14 weeks: The role of ductus venosus blood flow. Ultrasound Obstet Gynecol 1998;12:380–4.
9. Weinraub Z, Avrech OM, Golan A, Zabow P, Ron-El R, Bukovsky I, et al. Indomethacin and amniocentesis-induced changes in fetal flow velocity waveforms. Ultrasound Obstet Gynecol 1992;2:104–6.
10. Martinez JP, Comas C, Ojuel J, Puerto B, Borrell A, Fortuny A. Doppler assessment of umbilical flow after genetic amniocentesis. Early Hum Dev 1996;44:105–11.
11. Guzman ER, Schulman H, Karmel B, Higgins P. Umbilical artery Doppler velocimetry in pregnancies of less than 21 weeks' duration. J Ultrasound Med 1990;9:655–9.
12. Eik-Nes SH, Grøttum P, Jørgensen NP, Løkvik B. Normal range curves for BPD and MAD. Drammen, Norway: Scan-Med a/s, 1983.
13. Bjøro K, Stray-Pedersen S. In vitro perfusion studies on human umbilical arteries. 2. Effects of prostacyclin and a thromboxane A2 mimetic. Acta Obstet Gynecol Scand 1986;65:467–71.
14. Haugen G, Rognerud H. Doppler flow velocity waveforms and vasoactive effects of serotonin in human umbilical arteries. Gynecol Obstet Invest 2001;51:22–7.
15. Capponi A, Rizzo G, Rinaldo D, Arduini D, Romanini C. The effects of fetal blood sampling and placental puncture on umbilical artery and fetal vessels blood flow velocity waveforms. Am J Perinatol 1996;13:185–90.
16. Zoppini C, Ludomirsky A, Godmilow L, Weiner S, Maislin G, Donnenfeld AE. Acute hemodynamic effects induced by chorionic villus sampling: A preliminary investigation. Am J Obstet Gynecol 1993;169:902–7.
17. Martinez JP, Comas C, Ojuel J, Borrell A, Puerto B, Fortuny A. Immediate changes in umbilical blood flow after transcervical chorionic villus sampling performed by biopsy forceps. Prenat Diagn 1996;16:223–9.
18. Ursem NTC, Struijk PC, Hop WCJ, Clark EB, Keller BB, Wladimiroff JW. Heart rate and flow velocity variability as determined from umbilical Doppler velocimetry at 10–20 weeks of gestation. Clin Sci 1998;95:539–45.
19. Capponi A, Rizzo G, Pasquini L, Turri E, Arduini D, Romanini C. Indomethacin modifies the fetal hemodynamic response induced by percutaneous umbilical blood sampling. Am J Obstet Gynecol 1997;177:758–64.
20. Hecher K, Stettner H, Spernol R, Harrington K, Szalay S. The effect of cordocentesis on umbilical and middle cerebral blood flow velocity waveforms. Br J Obstet Gynaecol 1993;100:828–31.
21. Silver KK, MacGregor SN, Muhlbach LH, Knutel TA, Kambich MP. Congenital malformations subsequent to chorionic villus sampling—outcome analysis of 1048 consecutive procedures. Prenat Diagn 1994;14:421–7.
22. Wilson RD, Johnson JM, Dansereau J, Singer J, Drinnan SL, Winsor EJT, et al. Randomised trial to assess safety and fetal outcome of early and midtrimester amniocentesis. Lancet 1998;351:242–7.
© 2003 by The American College of Obstetricians and Gynecologists. Published by Wolters Kluwer Health, Inc. All rights reserved.