- Email: [email protected]
Amniocentesis and chorionic villus sampling in twin gestations: which is the best sampling technique?
Research
www. AJOG.org
OBSTETRICS
Amniocentesis and chorionic villus sampling in twin gestations: which is the best sampling technique? Giuliana Simonazzi, MD; Alessandra Curti, MD; Antonio Farina, MD; Gianluigi Pilu, MD; Luciano Bovicelli, MD; Nicola Rizzo, MD OBJECTIVE: To compare the fetal loss rate ⬍24 weeks and the pre-
term premature rupture of the membranes ⬍34 weeks’ gestation according to type of invasive procedure and to sampling techniques in twins.
STUDY DESIGN: Retrospective cohort study of 204 twin pregnancies, who underwent amniocentesis (100) or chorionic villus sampling (104). RESULTS: Fetal loss rate ⬍4 weeks was 3.85% in chorionic villus sampling group and 4.00% in amniocentesis group (P value not significant). According to sampling technique, fetal loss rate was 4.17% (chorionic villus sampling 1 puncture), 2.70% (amniocentesis 1 puncture), 3.75% (chorionic villus sampling 2 punctures), and 4.76% (amniocentesis 2 punctures), (P values not significant). Preterm premature rupture of the
membranes rate ⬍34 weeks was 8.2% chorionic villus sampling group and 10% in amniocentesis group (P value not significant). According to sampling technique, preterm premature rupture of the membranes rate was 12.5% (chorionic villus sampling 1 puncture), 8.1% (amniocentesis 1 puncture), 6.9% (chorionic villus sampling 2 punctures), and 11.1 % (amniocentesis 2 punctures), (P values not significant). CONCLUSION: Double entry technique does not affect significantly the
outcomes evaluated, in both amniocentesis and chorionic villus sampling. Key words: amniocentesis, chorionic villus sampling, fetal loss, Kaplan-Meier algorithm, sampling technique, twin gestation
Cite this article as: Simonazzi G, Curti A, Farina A, et al. Amniocentesis and chorionic villus sampling in twin gestations: which is the best sampling technique? Am J Obstet Gynecol 2010;202:365.e1-5.
T
he incidence of twin pregnancies has been increasing over the past 3 decades1 both due to greater reliance on fertility treatments, such as artificial or nonartificial reproductive technologies, and the rising maternal age secondary to delayed childbearing. Conventional in vitro fertilization (IVF) and intracytoplasmic sperm injection (ICSI) twin rates are 26.9% and 26.02%, respectively, and triplet rates are 2.8% and 2.9%, respectively for an estimated total of approximately 197,000-220,000 infants worldwide.2 Patients with multiple gestations are at higher risk for fetal chromosomal anomalies than those with singletons. The Department of Obstetrics and Gynecology (Drs Simonazzi, Curti, Farina, Pilu, and Rizzo), St. Orsola Malpighi Hospital, University of Bologna, and Tecnobios Prenatale (Dr Bovicelli), Bologna, Italy. Received June 19, 2009; revised Aug. 27, 2009; accepted Nov. 10, 2009. Reprints not available from the authors. 0002-9378/$36.00 © 2010 Mosby, Inc. All rights reserved. doi: 10.1016/j.ajog.2009.11.016
overall probability that a multiple gestation contains an aneuploid fetus is directly related to its zygosity. If monozygotic twins most often have the same karyotype and the risk of an affected fetus approximates the maternal age risk of a singleton, each fetus of a dizygotic twin pair has an independent risk of aneuploidy, so that the pregnancy has approximately twice the singleton risk of an affected fetus.3 Therefore, invasive prenatal diagnosis can be considered an important aspect of the antepartum management of patients with multiple gestation.4 However, the recommendation to undergo this testing needs careful evaluation, in particular in gestations resulting from fertility treatments, because of the possible increased risk of pregnancy loss because of invasive procedures. Previous studies5-15 have reported high success rates in twin gestations for both chorionic villus sampling (CVS) and amniocentesis, whereas the rates of pregnancy loss vary widely. Ghidini et al8 evaluated the risk of amniocentesis in twins, comparing 101 sampled pregnancies with an unsampled control group scanned at a matching gestational age,
and detected no significant difference in total loss rates. The risk of fetal loss in twin pregnancies after CVS varies widely; those studies13-15 that have compared fetal loss rate in singletons and twins found that CVS was not associated with an increased risk of either total pregnancy losses or single fetal losses. Only 2 studies13,15 have compared CVS with amniocentesis in twin pregnancies in contemporaneously sampled groups. They found no difference in the pregnancy loss rate and a potential small benefit for CVS regarding the total fetal loss rate. No studies have compared second-trimester amniocentesis with first-trimester CVS according to sampling techniques: single vs double entry to the uterus. In some cases, 1 entry to the uterus may be made, sampling before the first placenta and then moving the needle to the other placenta in case of CVS or aspirating amniotic fluid from the sac of the first twin and then advanced through the intertwin membrane into the sac of the second one. The hypothesis is that by introducing a single-entry technique, the fetal loss in twins undergoing
APRIL 2010 American Journal of Obstetrics & Gynecology
365.e1
Research
Obstetrics
invasive prenatal diagnosis may be lower and similar to that of singletons.16,17 The aim of this study was to assess the invasive prenatal diagnosis-related risk of pregnancy loss in twin gestations and in particular to compare second-trimester amniocentesis with first-trimester CVS according to sampling techniques.
M ATERIALS AND M ETHODS Between January 2002 and December 2007, 287 consecutive sets of multiple gestations underwent 153 first-trimester CVS (122 dichorionic, 12 monochorionic, 19 triplet) and 134 second-trimester amniocentesis (101 dichorionic, 27 monochorionic, 5 triplets) at the Unit of Prenatal Medicine (University of Bologna, Italy) and at Tecnobios Prenatale (Bologna, Italy). We excluded triplet pregnancies and 11 twin pregnancies (7 in the CVS group and 4 in the amniocentesis group) reduced to single for chromosomal anomalies in 1 fetus. Moreover, based on cytogenetic analysis, 2 pregnancies in both groups were electively terminated for abnormal results. Karyotyping was successfully obtained in all cases and no diagnostic errors were reported. Twenty-one patients in the group undergoing CVS and 23 in the amniocentesis group were lost to follow-up. This is a retrospective cohort study of the remaining 204 women with twin pregnancies of whom 100 underwent amniocentesis and 104 underwent CVS. All the procedures were performed transabdominally; when technically possible, a single entry was performed both in monochorionic and dichorionic pregnancies to reduce the risk of miscarriage. Both twins were sampled in all cases. The main indication for CVS and amniocentesis was fetal chromosomal evaluation for advanced maternal age (ⱖ35 years) and for parental decision. In a few cases, invasive procedures were performed because of the presence of chromosomal anomalies in previous pregnancies. All the individuals had a genetic counseling session provided by certified physicians. For the vast majority of the cases, the choice of procedure depended on the woman’s personal appraisal of the 365.e2
www.AJOG.org risks and benefits of each technique. In some cases, only amniocentesis was available because of a late admission. The decision about the sampling technique (single or double entry to the uterus) was made by the operators, based on placental location for CVS (1 entry in fused placentae; 2 entries in separate placentae) and on the technical possibility to perform a single entry for amniocentesis. CVS was performed at 10-14 weeks’ gestation, using a ultrasound-guided technique and a double coaxial needle system (outer needle 18 gauge, inner needle 20 gauge). The guide needle was first introduced into the placenta to be sampled. Thereafter, an aspiration needle was passed through the guide needle and villi obtained by aspiration through the sampling needle whose tip was moved forward and backward. When a single entry to the uterus was made, the first placenta was sampled first and then the needle was moved to the other placenta, taking care to ensure sampling of both placentae by aspirating the remotest possible areas of the 2 placentae. Amniocentesis was performed at 14-20 weeks’ gestation by a transabdominal free hand ultrasound-guided technique. A 22 gauge 0.7 mm needle was generally used and about 15 mL of amniotic fluid were obtained from each sac. When a single-uterine entry was made, after having aspirated amniotic fluid from the sac of the first twin, the syringe was removed, the stylet replaced in the needle, and the needle then advanced through the intertwin membrane under continuous ultrasound guidance into the sac of the second twin. Ultrasound examination was always performed before the procedure to determine the exact number of fetuses and both placental and fetal location. Fetal heart rate variability was always demonstrated after the procedure. Neither antibiotics nor tocolytics were given after amniocentesis or CVS. Information on pregnancy outcome was obtained from the patients themselves by telephone call several months after the expected date of birth or from maternal units.
American Journal of Obstetrics & Gynecology APRIL 2010
Statistical analysis Descriptive analysis was performed by routine test. The outcomes of interest of the study were the fetal loss rate before 24 weeks and the premature rupture of the membranes (pPROM) ⬍34 weeks’ gestation. The pPROM was diagnosed clinically, in presence of a history of watery vaginal discharge, confirmed on sterile speculum examination. Data were weighted for the number of fetuses observed (n ⫽ 408) because the outcome of the same pregnancy could be different (alive vs lost fetus and/or PPROM of 1 or both the sacs). Power analysis was performed by using the power analysis sample size (PASS) software at a fixed type I error ⫽ 0.05. The fetal loss rate and the pPROM rates were stratified according to type of procedure and to sampling techniques (single vs double entry), and calculated using the Kaplan-Meier algorithm or 2 test. Pregnancies that ended before the end of the follow-up constituted the “censor” group. The log rank test was used to explore differences between the generated subcategories (CVS vs amniocentesis and single vs double punctures). Results were considered statistically significant at P value ⬍ .05. The occurrence of pPROM was evaluated by 2 test.
R ESULTS
Mean gestational age (days ⫾ SD) at the time of invasive procedures was 83 ⫾ 5.98 and 112 ⫾ 9.03 (P ⬍ .001). Table 1 shows the demographic characteristics of the 2 groups. No significant statistical difference was found except for the rate of monochorionic twins. The number of entries according to chorionicity are shown in Table 2. All the surivival rates have been calculated by Kaplan-Meier algorithm. In our cohort of 204 patients with twin pregnancies, who underwent amniocentesis (100) or CVS (104), the total fetal loss rate ⬍24 weeks was 3.92% (Figure 1). The fetal loss rate ⬍24 weeks was 3.85% in the group of women who underwent CVS and 4.00 % in the group who underwent amniocentesis (P ⫽ .95, log rank test) (Figure 2).
Obstetrics
www.AJOG.org
Research
TABLE 1
FIGURE 1
Demographic characteristics of the CVS and amniocentesis groups
Fetal loss <24 weeks’ gestation in the overall series
Variable
CVS (104)
Amniocentesis (100)
P value
Maternal age, y
35.2 ⫾ 3.4
34.3 ⫾ 4.4
ns
Assisted fertilization, n (%)
18 (17.3)
25 (25)
ns
7 (6.73)
23 (23)
⬍ .01
.............................................................................................................................................................................................................................................. ..............................................................................................................................................................................................................................................
Monochorionic, n (%)
..............................................................................................................................................................................................................................................
Single entry, n (%)
24 (23.1)
37 (37)
ns
Loss of both twins ⬍24 wks, n (%)
3 (2.88)
4 (4)
ns
Loss of 1 twin ⬍24 wks, n (%)
1 (0.96)
0 (0)
ns
Delivery ⱕ34 wks, n (%)
36 (34.6)
31 (31)
ns
GA at delivery (wks ⫹ d)
34 ⫹ 0 ⫾ 4 ⫹ 5
34 ⫹ 3 ⫾ 4 ⫹ 4
ns
.............................................................................................................................................................................................................................................. .............................................................................................................................................................................................................................................. .............................................................................................................................................................................................................................................. .............................................................................................................................................................................................................................................. ..............................................................................................................................................................................................................................................
CVS, chorionic villus sampling; GA, gestational age Simonazzi. Amniocentesis and CVS in twin gestations. Am J Obstet Gynecol 2010.
Again, the stratification according with number of entries did not reach a significant value (P ⫽ .61, log rank test) (Figure 3). According to both variables, the fetal loss rate was 4.17% (CVS 1 entry), 2.70% (amniocentesis 1 entry), 3.75% (CVS 2 entries), and 4.76% (amniocentesis 2 entries), without a significant statistical difference (Figures 2-4) (P ⫽ .96, log rank test). However, it must be noted that the power of the survival analysis was quite low (from 0.22-0.26). To reach a power of 0.80, we would have needed a total of 2380 cases. The overall pPROM rate ⬍34 weeks was 9.1%. According to subgroups it was 8.2% in the group of women who underwent CVS and 10% in the group who underwent amniocentesis (P ⫽ .52, 2 test). According also to sampling technique, the pPROM rate was 12.5% (CVS 1 entry), 8.1% (amniocentesis 1 entry), 6.9% (CVS 2 entries), and 11.1% (amniocentesis 2 entries), without a significant statistical difference (P ⫽ .50, 2 test). There were no errors in the samples and in no case did cell culture fail. In 1
case of CVS, it was also necessary to perform an amniocentesis (only in 1 twin) because of a confined chorionic mosaicism, not demonstrated in the fetus (0.9%). On the basis of a cytogenetic analysis, 2 pregnancies in the CVS group (1.9%) and 2 in the amniocentesis group (1.9%) were electively terminated for abnormal results.
C OMMENT Our data suggest that CVS and amniocentesis for prenatal diagnosis of aneuploidy in twin pregnancies do not differ statistically for the fetal loss rate before 24 weeks and the pPROM ⬍34 weeks’ gestation. Moreover, the comparison between single vs double entry to the uterus for both first-trimester CVS and secondtrimester amniocentesis failed to demonstrate statistically significant differences. In particular, amniocentesis with 1 entry scores the lowest risk (2.70%) and a double-entry procedure carries an extra 2% risk (4.70%). Instead, for CVS
Simonazzi. Amniocentesis and CVS in twin gestations. Am J Obstet Gynecol 2010.
the quoted risk is basically the same for 1 or 2 entries being 4.17% and 3.75%, respectively. Lack of statistical difference may be related to the the very low rate of the fetal loss in our study, because a sample size analysis would require about 2500 cases to detect a possible significant difference between CVS and amniocentesis, as well as single vs double entry by log rank test. However, to our knowledge, this study is statistically among the more robust designs reported in the literature. In general, our results compare favorably with the literature. The total fetal FIGURE 2
Fetal loss <24 weeks’ gestation stratified according to technique (CVS vs amniocentesis)
TABLE 2
Number (%) of entries to the uterus according to chorionicity CVS (n ⴝ 104)
Amniocentesis (n ⴝ 100)
Entry to the uterus
Monochorionic
Dichorionic
Single
2 (28.5)
22 (22.7)
7 (31.8)
30 (38.5)
61 (29.9)
Double
5 (71.5)
75 (77.3)
15 (68.2)
48 (61.5)
143 (70.1)
Monochorionic
Dichorionic
Total
..............................................................................................................................................................................................................................................
..............................................................................................................................................................................................................................................
CVS, chorionic villus sampling. Simonazzi. Amniocentesis and CVS in twin gestations. Am J Obstet Gynecol 2010.
Simonazzi. Amniocentesis and CVS in twin gestations. Am J Obstet Gynecol 2010.
APRIL 2010 American Journal of Obstetrics & Gynecology
365.e3
Research
Obstetrics
FIGURE 3
Fetal loss <24 weeks’ gestation stratified according to number of entries
Simonazzi. Amniocentesis and CVS in twin gestations. Am J Obstet Gynecol 2010.
loss rate ⬍24 weeks of 3.85% in the CVS group is in line with the fetal loss rates of 2-4% reported by other authors.4 The 4% fetal loss rate ⬍24 weeks in the amniocentesis group is similar to the fetal loss rate reported by Ghidini et al8 (3.5%) and slightly higher than that of Yukobowich et al9 (2.73%). To date, only 2 studies13,15 have compared CVS with amniocentesis in twin pregnancies in contemporaneously sampled groups, finding no difference in the pregnancy loss rate (3.2% vs 2.9% and 4.5% vs 4.1%, respectively) and a potenFIGURE 4
Fetal loss <24 weeks’ gestation stratified according to technique (CVS vs amniocentesis) and number of entries
Simonazzi. Amniocentesis and CVS in twin gestations. Am J Obstet Gynecol 2010.
365.e4
www.AJOG.org tial small benefit for CVS regarding the total fetal loss rate in one study13 (4.9% vs 9.3%), but not in the other15 (10.2% vs 8.8%). As our study also stratified for the number of entries, the fetal loss rate may be more accurate. A single-entry CVS carries a numerically almost double risk of pPROM rate than a double entry (12.5% vs 6.9%), but this difference does not reach a statistically significant value. In our series, there were no errors in the samples. This is probably because before performing any invasive procedure in twins in general, and CVS in particular, a thorough ultrasound examination is always performed to evaluate and document the chorionicity, location of the fetuses, their related placenta, and the location of cord insertions. Moreover, when a CVS is performed by a singleentry technique, the remotest possible areas of the placentae are aspirated to ensure both are sampled. Another potential problem after CVS is the risk of placental-confined mosaicism. We encountered this complication in 1 case (0.9%): an amniocentesis (only in 1 twin) was then performed and the mosaicism was not confirmed in the fetus. As shown in Table 2, there is a disproportion between the subcategories generated according to corionicity (29 monochorionic vs 175 dichorionic pregnancies) but the choice of sampling technique, in terms of number of entries, was never based on the number of placentae. In addition, both twins were always sampled. We acknowledge that the lack of a control group of unsampled twins is a potential limitation of our study. As reported by Caughey et al,18 in the nonrandomized setting, there are 2 major issues that may cause bias in the findings. The first is the different gestational ages at the time of the procedures (particularly between 9 and 15 weeks), because gestational age is strongly associated with pregnancy loss rate, and the second is the demographic characteristics of the patients. In our study, even though CVS was performed 4-6 weeks earlier than amniocentesis, the overall pregnancy loss rate was similar to that of amniocentesis.
American Journal of Obstetrics & Gynecology APRIL 2010
However, because of the low number of “events” in respect with the total number of cases, it was not possible to properly adjust for the possible effect on the survival rate because of the gestational age at the time of enrollment. Again, we were not able to evaluate any other source of bias for the limited clinical information available. However, the population hereby enrolled is quite homogeneous for ethnicity and social status. Despite these limitations, we believe these findings may be of value in counseling parents of twins because of the increasing number of gestations resulting from fertility programs and the high risk of chromosomal abnormalities in twin pregnancies. The choice of which invasive procedure to perform in twin pregnancies is based on several factors: indications, gestational age, technical difficulties, and operator experience. The fetal loss (⬍24 weeks’ gestation) and the pPROM rates (⬍34 weeks’ gestation) of CVS appear comparable to those of amniocentesis and the double-entry technique does not increase significantly procedure-related complications. Given the risk of sampling error and cross-contamination, mainly in CVS, the double-entry technique is preferable to avoid diagnostic f mistakes. REFERENCES 1. Russell RB, Petrini JR, Damus K, Mattison DR, Schwarz RH. The changing epidemiology of multiple births in the United States. Obstet Gynaecol 2003;101:129-35. 2. Adamson GD, de Mouzon J, Lancaster P, Nygren KG, Sullivan E, Zegers-Hochschild F. World collaborative report on in vitro fertilization, 2000. Fertil Steril 2006;85:1586-622. 3. Jenkins TM, Wapner RJ. The challenge of prenatal diagnosis in twin pregnancies. Curr Opin Obstet Gynecol 2000;12:87-92. 4. Weisz B, Rodeck CH. Invasive diagnostic procedures in twin pregnancies. Prenat Diagn 2005;25:751-8. 5. Palle C, Andersen JW, Tabor A, Lauritsen JG, Bang J, Philip J. Increased risk of abortion after genetic amniocentesis in twin pregnanies. Prenat Diagn 1983;3:83-9. 6. Pijpers L, Jahoda MG, Vosters RP, Niermeijer MF, Sachs ES. Genetic amniocentesis in twin pregnancies. BJOG 1988;95:323-6. 7. Anderson RL, Goldberg JD, Golbus MS. Prenatal diagnosis in multiple gestation: 20 years’ experience with amniocentesis. Prenat Diagn 1991;11:263-70.
Obstetrics
www.AJOG.org 8. Ghidini A, Lynch L, Hicks C, Alvarez M, Lockwood CJ. The risk of second- trimester amniocentesis in twin gestations: a case-control study. Am J Obstet Gynecol 1993;169: 1013-4. 9. Yukobowich E, Anteby EY, Cohen SM, Lavy Y, Granat M, Yagel S. Risk of fetal loss in twin pregnancies undergoing second trimester amniocentesis. Obstet Gynaecol 2001;98:231-4. 10. Tóth-Pál E, Papp C, Beke A, Bán Z, Papp Z.Genetic amniocentesis in multiple pregnancies. Fetal Diagn Ther 2004;19:138-44. 11. Millaire M, Bujold E, Morency AM, Gauthier RJ. Mid-trimester genetic amniocentesis in twin pregnancy and the risk of fetal loss. Obstet Gynaecol 2006;28:512-8.
12. Rochon M, Stone J. Invasive procedures in multiple gestations. Curr Opin Obstet Gynecol 2003;15:167-75. 13. Wapner RJ, Johnson A, Davis G, Urban A, Morgan P, Jackson L. Prenatal diagnosis in twin gestations: a comparison between second trimester amniocentesis and first trimester chorionic villus sampling. Obstet Gynaecol 1993;82: 49-56. 14. Brambati B, Tului L, Guercilena S, Alberti E. Outcome of first trimester chorionic villus sampling for genetic investigation in multiple pregnancy. Ultrasound Obstet Gynaecol 2001;17: 209-16. 15. Antsaklis A, Souka AP, Daskalakis G, Kavalakis Y, Michalas S. Second-trimester amnio-
Research
centesis vs. chorionic villus sampling for prenatal diagnosis in multiple gestations. Ultrasound Obstet Gynaecol 2002;20:476-81. 16. Van Vugt JM, Nieuwint A, van Geijn HP. Single-needle insertion: an alternative technique for early second-trimester genetic amniocentesis. Fetal Diagn Ther 1995;10:178-81. 17. Sebire NJ, Noble PL, Odibo A, Malligiannis P, Nicolaides KH. Single uterine entry for genetic amniocentesis in twin pregnancies. Ultrasound Obstet Gynaecol 1996;7:26-31. 18. Caughey A, Hopkins LM, Norton ME. Chorionic villus sampling compared wit amniocentesis and the difference in the rate of pregnancy loss. Obstet Gynecol 2006;108:612-6.
APRIL 2010 American Journal of Obstetrics & Gynecology
365.e5
www. AJOG.org
OBSTETRICS
Amniocentesis and chorionic villus sampling in twin gestations: which is the best sampling technique? Giuliana Simonazzi, MD; Alessandra Curti, MD; Antonio Farina, MD; Gianluigi Pilu, MD; Luciano Bovicelli, MD; Nicola Rizzo, MD OBJECTIVE: To compare the fetal loss rate ⬍24 weeks and the pre-
term premature rupture of the membranes ⬍34 weeks’ gestation according to type of invasive procedure and to sampling techniques in twins.
STUDY DESIGN: Retrospective cohort study of 204 twin pregnancies, who underwent amniocentesis (100) or chorionic villus sampling (104). RESULTS: Fetal loss rate ⬍4 weeks was 3.85% in chorionic villus sampling group and 4.00% in amniocentesis group (P value not significant). According to sampling technique, fetal loss rate was 4.17% (chorionic villus sampling 1 puncture), 2.70% (amniocentesis 1 puncture), 3.75% (chorionic villus sampling 2 punctures), and 4.76% (amniocentesis 2 punctures), (P values not significant). Preterm premature rupture of the
membranes rate ⬍34 weeks was 8.2% chorionic villus sampling group and 10% in amniocentesis group (P value not significant). According to sampling technique, preterm premature rupture of the membranes rate was 12.5% (chorionic villus sampling 1 puncture), 8.1% (amniocentesis 1 puncture), 6.9% (chorionic villus sampling 2 punctures), and 11.1 % (amniocentesis 2 punctures), (P values not significant). CONCLUSION: Double entry technique does not affect significantly the
outcomes evaluated, in both amniocentesis and chorionic villus sampling. Key words: amniocentesis, chorionic villus sampling, fetal loss, Kaplan-Meier algorithm, sampling technique, twin gestation
Cite this article as: Simonazzi G, Curti A, Farina A, et al. Amniocentesis and chorionic villus sampling in twin gestations: which is the best sampling technique? Am J Obstet Gynecol 2010;202:365.e1-5.
T
he incidence of twin pregnancies has been increasing over the past 3 decades1 both due to greater reliance on fertility treatments, such as artificial or nonartificial reproductive technologies, and the rising maternal age secondary to delayed childbearing. Conventional in vitro fertilization (IVF) and intracytoplasmic sperm injection (ICSI) twin rates are 26.9% and 26.02%, respectively, and triplet rates are 2.8% and 2.9%, respectively for an estimated total of approximately 197,000-220,000 infants worldwide.2 Patients with multiple gestations are at higher risk for fetal chromosomal anomalies than those with singletons. The Department of Obstetrics and Gynecology (Drs Simonazzi, Curti, Farina, Pilu, and Rizzo), St. Orsola Malpighi Hospital, University of Bologna, and Tecnobios Prenatale (Dr Bovicelli), Bologna, Italy. Received June 19, 2009; revised Aug. 27, 2009; accepted Nov. 10, 2009. Reprints not available from the authors. 0002-9378/$36.00 © 2010 Mosby, Inc. All rights reserved. doi: 10.1016/j.ajog.2009.11.016
overall probability that a multiple gestation contains an aneuploid fetus is directly related to its zygosity. If monozygotic twins most often have the same karyotype and the risk of an affected fetus approximates the maternal age risk of a singleton, each fetus of a dizygotic twin pair has an independent risk of aneuploidy, so that the pregnancy has approximately twice the singleton risk of an affected fetus.3 Therefore, invasive prenatal diagnosis can be considered an important aspect of the antepartum management of patients with multiple gestation.4 However, the recommendation to undergo this testing needs careful evaluation, in particular in gestations resulting from fertility treatments, because of the possible increased risk of pregnancy loss because of invasive procedures. Previous studies5-15 have reported high success rates in twin gestations for both chorionic villus sampling (CVS) and amniocentesis, whereas the rates of pregnancy loss vary widely. Ghidini et al8 evaluated the risk of amniocentesis in twins, comparing 101 sampled pregnancies with an unsampled control group scanned at a matching gestational age,
and detected no significant difference in total loss rates. The risk of fetal loss in twin pregnancies after CVS varies widely; those studies13-15 that have compared fetal loss rate in singletons and twins found that CVS was not associated with an increased risk of either total pregnancy losses or single fetal losses. Only 2 studies13,15 have compared CVS with amniocentesis in twin pregnancies in contemporaneously sampled groups. They found no difference in the pregnancy loss rate and a potential small benefit for CVS regarding the total fetal loss rate. No studies have compared second-trimester amniocentesis with first-trimester CVS according to sampling techniques: single vs double entry to the uterus. In some cases, 1 entry to the uterus may be made, sampling before the first placenta and then moving the needle to the other placenta in case of CVS or aspirating amniotic fluid from the sac of the first twin and then advanced through the intertwin membrane into the sac of the second one. The hypothesis is that by introducing a single-entry technique, the fetal loss in twins undergoing
APRIL 2010 American Journal of Obstetrics & Gynecology
365.e1
Research
Obstetrics
invasive prenatal diagnosis may be lower and similar to that of singletons.16,17 The aim of this study was to assess the invasive prenatal diagnosis-related risk of pregnancy loss in twin gestations and in particular to compare second-trimester amniocentesis with first-trimester CVS according to sampling techniques.
M ATERIALS AND M ETHODS Between January 2002 and December 2007, 287 consecutive sets of multiple gestations underwent 153 first-trimester CVS (122 dichorionic, 12 monochorionic, 19 triplet) and 134 second-trimester amniocentesis (101 dichorionic, 27 monochorionic, 5 triplets) at the Unit of Prenatal Medicine (University of Bologna, Italy) and at Tecnobios Prenatale (Bologna, Italy). We excluded triplet pregnancies and 11 twin pregnancies (7 in the CVS group and 4 in the amniocentesis group) reduced to single for chromosomal anomalies in 1 fetus. Moreover, based on cytogenetic analysis, 2 pregnancies in both groups were electively terminated for abnormal results. Karyotyping was successfully obtained in all cases and no diagnostic errors were reported. Twenty-one patients in the group undergoing CVS and 23 in the amniocentesis group were lost to follow-up. This is a retrospective cohort study of the remaining 204 women with twin pregnancies of whom 100 underwent amniocentesis and 104 underwent CVS. All the procedures were performed transabdominally; when technically possible, a single entry was performed both in monochorionic and dichorionic pregnancies to reduce the risk of miscarriage. Both twins were sampled in all cases. The main indication for CVS and amniocentesis was fetal chromosomal evaluation for advanced maternal age (ⱖ35 years) and for parental decision. In a few cases, invasive procedures were performed because of the presence of chromosomal anomalies in previous pregnancies. All the individuals had a genetic counseling session provided by certified physicians. For the vast majority of the cases, the choice of procedure depended on the woman’s personal appraisal of the 365.e2
www.AJOG.org risks and benefits of each technique. In some cases, only amniocentesis was available because of a late admission. The decision about the sampling technique (single or double entry to the uterus) was made by the operators, based on placental location for CVS (1 entry in fused placentae; 2 entries in separate placentae) and on the technical possibility to perform a single entry for amniocentesis. CVS was performed at 10-14 weeks’ gestation, using a ultrasound-guided technique and a double coaxial needle system (outer needle 18 gauge, inner needle 20 gauge). The guide needle was first introduced into the placenta to be sampled. Thereafter, an aspiration needle was passed through the guide needle and villi obtained by aspiration through the sampling needle whose tip was moved forward and backward. When a single entry to the uterus was made, the first placenta was sampled first and then the needle was moved to the other placenta, taking care to ensure sampling of both placentae by aspirating the remotest possible areas of the 2 placentae. Amniocentesis was performed at 14-20 weeks’ gestation by a transabdominal free hand ultrasound-guided technique. A 22 gauge 0.7 mm needle was generally used and about 15 mL of amniotic fluid were obtained from each sac. When a single-uterine entry was made, after having aspirated amniotic fluid from the sac of the first twin, the syringe was removed, the stylet replaced in the needle, and the needle then advanced through the intertwin membrane under continuous ultrasound guidance into the sac of the second twin. Ultrasound examination was always performed before the procedure to determine the exact number of fetuses and both placental and fetal location. Fetal heart rate variability was always demonstrated after the procedure. Neither antibiotics nor tocolytics were given after amniocentesis or CVS. Information on pregnancy outcome was obtained from the patients themselves by telephone call several months after the expected date of birth or from maternal units.
American Journal of Obstetrics & Gynecology APRIL 2010
Statistical analysis Descriptive analysis was performed by routine test. The outcomes of interest of the study were the fetal loss rate before 24 weeks and the premature rupture of the membranes (pPROM) ⬍34 weeks’ gestation. The pPROM was diagnosed clinically, in presence of a history of watery vaginal discharge, confirmed on sterile speculum examination. Data were weighted for the number of fetuses observed (n ⫽ 408) because the outcome of the same pregnancy could be different (alive vs lost fetus and/or PPROM of 1 or both the sacs). Power analysis was performed by using the power analysis sample size (PASS) software at a fixed type I error ⫽ 0.05. The fetal loss rate and the pPROM rates were stratified according to type of procedure and to sampling techniques (single vs double entry), and calculated using the Kaplan-Meier algorithm or 2 test. Pregnancies that ended before the end of the follow-up constituted the “censor” group. The log rank test was used to explore differences between the generated subcategories (CVS vs amniocentesis and single vs double punctures). Results were considered statistically significant at P value ⬍ .05. The occurrence of pPROM was evaluated by 2 test.
R ESULTS
Mean gestational age (days ⫾ SD) at the time of invasive procedures was 83 ⫾ 5.98 and 112 ⫾ 9.03 (P ⬍ .001). Table 1 shows the demographic characteristics of the 2 groups. No significant statistical difference was found except for the rate of monochorionic twins. The number of entries according to chorionicity are shown in Table 2. All the surivival rates have been calculated by Kaplan-Meier algorithm. In our cohort of 204 patients with twin pregnancies, who underwent amniocentesis (100) or CVS (104), the total fetal loss rate ⬍24 weeks was 3.92% (Figure 1). The fetal loss rate ⬍24 weeks was 3.85% in the group of women who underwent CVS and 4.00 % in the group who underwent amniocentesis (P ⫽ .95, log rank test) (Figure 2).
Obstetrics
www.AJOG.org
Research
TABLE 1
FIGURE 1
Demographic characteristics of the CVS and amniocentesis groups
Fetal loss <24 weeks’ gestation in the overall series
Variable
CVS (104)
Amniocentesis (100)
P value
Maternal age, y
35.2 ⫾ 3.4
34.3 ⫾ 4.4
ns
Assisted fertilization, n (%)
18 (17.3)
25 (25)
ns
7 (6.73)
23 (23)
⬍ .01
.............................................................................................................................................................................................................................................. ..............................................................................................................................................................................................................................................
Monochorionic, n (%)
..............................................................................................................................................................................................................................................
Single entry, n (%)
24 (23.1)
37 (37)
ns
Loss of both twins ⬍24 wks, n (%)
3 (2.88)
4 (4)
ns
Loss of 1 twin ⬍24 wks, n (%)
1 (0.96)
0 (0)
ns
Delivery ⱕ34 wks, n (%)
36 (34.6)
31 (31)
ns
GA at delivery (wks ⫹ d)
34 ⫹ 0 ⫾ 4 ⫹ 5
34 ⫹ 3 ⫾ 4 ⫹ 4
ns
.............................................................................................................................................................................................................................................. .............................................................................................................................................................................................................................................. .............................................................................................................................................................................................................................................. .............................................................................................................................................................................................................................................. ..............................................................................................................................................................................................................................................
CVS, chorionic villus sampling; GA, gestational age Simonazzi. Amniocentesis and CVS in twin gestations. Am J Obstet Gynecol 2010.
Again, the stratification according with number of entries did not reach a significant value (P ⫽ .61, log rank test) (Figure 3). According to both variables, the fetal loss rate was 4.17% (CVS 1 entry), 2.70% (amniocentesis 1 entry), 3.75% (CVS 2 entries), and 4.76% (amniocentesis 2 entries), without a significant statistical difference (Figures 2-4) (P ⫽ .96, log rank test). However, it must be noted that the power of the survival analysis was quite low (from 0.22-0.26). To reach a power of 0.80, we would have needed a total of 2380 cases. The overall pPROM rate ⬍34 weeks was 9.1%. According to subgroups it was 8.2% in the group of women who underwent CVS and 10% in the group who underwent amniocentesis (P ⫽ .52, 2 test). According also to sampling technique, the pPROM rate was 12.5% (CVS 1 entry), 8.1% (amniocentesis 1 entry), 6.9% (CVS 2 entries), and 11.1% (amniocentesis 2 entries), without a significant statistical difference (P ⫽ .50, 2 test). There were no errors in the samples and in no case did cell culture fail. In 1
case of CVS, it was also necessary to perform an amniocentesis (only in 1 twin) because of a confined chorionic mosaicism, not demonstrated in the fetus (0.9%). On the basis of a cytogenetic analysis, 2 pregnancies in the CVS group (1.9%) and 2 in the amniocentesis group (1.9%) were electively terminated for abnormal results.
C OMMENT Our data suggest that CVS and amniocentesis for prenatal diagnosis of aneuploidy in twin pregnancies do not differ statistically for the fetal loss rate before 24 weeks and the pPROM ⬍34 weeks’ gestation. Moreover, the comparison between single vs double entry to the uterus for both first-trimester CVS and secondtrimester amniocentesis failed to demonstrate statistically significant differences. In particular, amniocentesis with 1 entry scores the lowest risk (2.70%) and a double-entry procedure carries an extra 2% risk (4.70%). Instead, for CVS
Simonazzi. Amniocentesis and CVS in twin gestations. Am J Obstet Gynecol 2010.
the quoted risk is basically the same for 1 or 2 entries being 4.17% and 3.75%, respectively. Lack of statistical difference may be related to the the very low rate of the fetal loss in our study, because a sample size analysis would require about 2500 cases to detect a possible significant difference between CVS and amniocentesis, as well as single vs double entry by log rank test. However, to our knowledge, this study is statistically among the more robust designs reported in the literature. In general, our results compare favorably with the literature. The total fetal FIGURE 2
Fetal loss <24 weeks’ gestation stratified according to technique (CVS vs amniocentesis)
TABLE 2
Number (%) of entries to the uterus according to chorionicity CVS (n ⴝ 104)
Amniocentesis (n ⴝ 100)
Entry to the uterus
Monochorionic
Dichorionic
Single
2 (28.5)
22 (22.7)
7 (31.8)
30 (38.5)
61 (29.9)
Double
5 (71.5)
75 (77.3)
15 (68.2)
48 (61.5)
143 (70.1)
Monochorionic
Dichorionic
Total
..............................................................................................................................................................................................................................................
..............................................................................................................................................................................................................................................
CVS, chorionic villus sampling. Simonazzi. Amniocentesis and CVS in twin gestations. Am J Obstet Gynecol 2010.
Simonazzi. Amniocentesis and CVS in twin gestations. Am J Obstet Gynecol 2010.
APRIL 2010 American Journal of Obstetrics & Gynecology
365.e3
Research
Obstetrics
FIGURE 3
Fetal loss <24 weeks’ gestation stratified according to number of entries
Simonazzi. Amniocentesis and CVS in twin gestations. Am J Obstet Gynecol 2010.
loss rate ⬍24 weeks of 3.85% in the CVS group is in line with the fetal loss rates of 2-4% reported by other authors.4 The 4% fetal loss rate ⬍24 weeks in the amniocentesis group is similar to the fetal loss rate reported by Ghidini et al8 (3.5%) and slightly higher than that of Yukobowich et al9 (2.73%). To date, only 2 studies13,15 have compared CVS with amniocentesis in twin pregnancies in contemporaneously sampled groups, finding no difference in the pregnancy loss rate (3.2% vs 2.9% and 4.5% vs 4.1%, respectively) and a potenFIGURE 4
Fetal loss <24 weeks’ gestation stratified according to technique (CVS vs amniocentesis) and number of entries
Simonazzi. Amniocentesis and CVS in twin gestations. Am J Obstet Gynecol 2010.
365.e4
www.AJOG.org tial small benefit for CVS regarding the total fetal loss rate in one study13 (4.9% vs 9.3%), but not in the other15 (10.2% vs 8.8%). As our study also stratified for the number of entries, the fetal loss rate may be more accurate. A single-entry CVS carries a numerically almost double risk of pPROM rate than a double entry (12.5% vs 6.9%), but this difference does not reach a statistically significant value. In our series, there were no errors in the samples. This is probably because before performing any invasive procedure in twins in general, and CVS in particular, a thorough ultrasound examination is always performed to evaluate and document the chorionicity, location of the fetuses, their related placenta, and the location of cord insertions. Moreover, when a CVS is performed by a singleentry technique, the remotest possible areas of the placentae are aspirated to ensure both are sampled. Another potential problem after CVS is the risk of placental-confined mosaicism. We encountered this complication in 1 case (0.9%): an amniocentesis (only in 1 twin) was then performed and the mosaicism was not confirmed in the fetus. As shown in Table 2, there is a disproportion between the subcategories generated according to corionicity (29 monochorionic vs 175 dichorionic pregnancies) but the choice of sampling technique, in terms of number of entries, was never based on the number of placentae. In addition, both twins were always sampled. We acknowledge that the lack of a control group of unsampled twins is a potential limitation of our study. As reported by Caughey et al,18 in the nonrandomized setting, there are 2 major issues that may cause bias in the findings. The first is the different gestational ages at the time of the procedures (particularly between 9 and 15 weeks), because gestational age is strongly associated with pregnancy loss rate, and the second is the demographic characteristics of the patients. In our study, even though CVS was performed 4-6 weeks earlier than amniocentesis, the overall pregnancy loss rate was similar to that of amniocentesis.
American Journal of Obstetrics & Gynecology APRIL 2010
However, because of the low number of “events” in respect with the total number of cases, it was not possible to properly adjust for the possible effect on the survival rate because of the gestational age at the time of enrollment. Again, we were not able to evaluate any other source of bias for the limited clinical information available. However, the population hereby enrolled is quite homogeneous for ethnicity and social status. Despite these limitations, we believe these findings may be of value in counseling parents of twins because of the increasing number of gestations resulting from fertility programs and the high risk of chromosomal abnormalities in twin pregnancies. The choice of which invasive procedure to perform in twin pregnancies is based on several factors: indications, gestational age, technical difficulties, and operator experience. The fetal loss (⬍24 weeks’ gestation) and the pPROM rates (⬍34 weeks’ gestation) of CVS appear comparable to those of amniocentesis and the double-entry technique does not increase significantly procedure-related complications. Given the risk of sampling error and cross-contamination, mainly in CVS, the double-entry technique is preferable to avoid diagnostic f mistakes. REFERENCES 1. Russell RB, Petrini JR, Damus K, Mattison DR, Schwarz RH. The changing epidemiology of multiple births in the United States. Obstet Gynaecol 2003;101:129-35. 2. Adamson GD, de Mouzon J, Lancaster P, Nygren KG, Sullivan E, Zegers-Hochschild F. World collaborative report on in vitro fertilization, 2000. Fertil Steril 2006;85:1586-622. 3. Jenkins TM, Wapner RJ. The challenge of prenatal diagnosis in twin pregnancies. Curr Opin Obstet Gynecol 2000;12:87-92. 4. Weisz B, Rodeck CH. Invasive diagnostic procedures in twin pregnancies. Prenat Diagn 2005;25:751-8. 5. Palle C, Andersen JW, Tabor A, Lauritsen JG, Bang J, Philip J. Increased risk of abortion after genetic amniocentesis in twin pregnanies. Prenat Diagn 1983;3:83-9. 6. Pijpers L, Jahoda MG, Vosters RP, Niermeijer MF, Sachs ES. Genetic amniocentesis in twin pregnancies. BJOG 1988;95:323-6. 7. Anderson RL, Goldberg JD, Golbus MS. Prenatal diagnosis in multiple gestation: 20 years’ experience with amniocentesis. Prenat Diagn 1991;11:263-70.
Obstetrics
www.AJOG.org 8. Ghidini A, Lynch L, Hicks C, Alvarez M, Lockwood CJ. The risk of second- trimester amniocentesis in twin gestations: a case-control study. Am J Obstet Gynecol 1993;169: 1013-4. 9. Yukobowich E, Anteby EY, Cohen SM, Lavy Y, Granat M, Yagel S. Risk of fetal loss in twin pregnancies undergoing second trimester amniocentesis. Obstet Gynaecol 2001;98:231-4. 10. Tóth-Pál E, Papp C, Beke A, Bán Z, Papp Z.Genetic amniocentesis in multiple pregnancies. Fetal Diagn Ther 2004;19:138-44. 11. Millaire M, Bujold E, Morency AM, Gauthier RJ. Mid-trimester genetic amniocentesis in twin pregnancy and the risk of fetal loss. Obstet Gynaecol 2006;28:512-8.
12. Rochon M, Stone J. Invasive procedures in multiple gestations. Curr Opin Obstet Gynecol 2003;15:167-75. 13. Wapner RJ, Johnson A, Davis G, Urban A, Morgan P, Jackson L. Prenatal diagnosis in twin gestations: a comparison between second trimester amniocentesis and first trimester chorionic villus sampling. Obstet Gynaecol 1993;82: 49-56. 14. Brambati B, Tului L, Guercilena S, Alberti E. Outcome of first trimester chorionic villus sampling for genetic investigation in multiple pregnancy. Ultrasound Obstet Gynaecol 2001;17: 209-16. 15. Antsaklis A, Souka AP, Daskalakis G, Kavalakis Y, Michalas S. Second-trimester amnio-
Research
centesis vs. chorionic villus sampling for prenatal diagnosis in multiple gestations. Ultrasound Obstet Gynaecol 2002;20:476-81. 16. Van Vugt JM, Nieuwint A, van Geijn HP. Single-needle insertion: an alternative technique for early second-trimester genetic amniocentesis. Fetal Diagn Ther 1995;10:178-81. 17. Sebire NJ, Noble PL, Odibo A, Malligiannis P, Nicolaides KH. Single uterine entry for genetic amniocentesis in twin pregnancies. Ultrasound Obstet Gynaecol 1996;7:26-31. 18. Caughey A, Hopkins LM, Norton ME. Chorionic villus sampling compared wit amniocentesis and the difference in the rate of pregnancy loss. Obstet Gynecol 2006;108:612-6.
APRIL 2010 American Journal of Obstetrics & Gynecology
365.e5