Clinical outcomes of fetal selective reduction in dichorionic triplet pregnancies

Abstract

Background:

It is recommended to reduce triplet pregnancy containing monochorionic (MC) twins to singleton. Given that some couples with infertility are eager to retain twins, better strategy is needed to avoid obstetrical risks and satisfy their strong wish. This retrospective observational study aimed to investigate the outcomes of triplet pregnancy reduction.

Methods:

Subjects with triplet pregnancies who underwent selective reduction between 2016 and 2019 at our hospital were enrolled. A total of 66 subjects with dichorionic triplet (DCT) with MC twins and an MC singleton were divided into two groups: group A (N = 38), reduced to dichorionic diamniotic (DCDA) twins; group B (N = 28), reduced to MC diamniotic (MCDA) twins. Obstetrical and perinatal outcomes were compared between groups.

Results:

Group A had significantly lower rates of early miscarriage (0% vs 14.3%, p = 0.028), cesarean section (81.6% vs 100%, p = 0.041), and late premature delivery (21.1% vs 45.4%, p = 0.047) than group B. Significantly higher rates of full-term delivery (71% vs 36.4%, p = 0.009) and take-home baby (100% vs 78.6%, p = 0.004), and higher gestational age at delivery (median: 38 [36.9, 39.0] vs 35.8 [34.4, 37.0] weeks, p < 0.001), total neonatal weight (2899.7 ± 647.6 vs 2354.4 ± 651.8 g, p < 0.001), weight of twins (2550 vs 2350 g, p = 0.039), and weight of larger neonate in twins (2790 vs 2500 g, p = 0.045) were observed in group A compared to group B.

Conclusion:

DCT reduced to DCDA twins confers better pregnancy outcomes than into MCDA twins. This might benefit for triplet pregnancy subjects who strongly want to retain fraternal twins.

1. INTRODUCTION

With the increasing use of ovulation drugs and assisted reproductive technology (ART), the multiple pregnancy rates have increased significantly in the past three decades.^1–5 However, high percentages of low birth weight and perinatal mortality rate were reported in twins/triples.⁴ The probability of cerebral palsy in one or more of the triplets was 7% to 8%.⁵ In vitro fertilization (IVF) and non-IVF fertility treatments (ie, ovulation induction and ovarian stimulation) are associated with significant risk factors for the genesis of multiple births (twin, triplet, and higher-order births)^6,7 as multiple embryos are transferred in most ART procedures.

Monochorionic (MC) twins share one placenta with vascular anastomoses connecting the circulations of both fetuses. The occurrence of MC twin pregnancies is rare (only 0.4% of the general population), and it occurs in one of 250 pregnancies⁸ and is associated complications due to a shared placental circulation are frequent and severe.⁹ MC twin pregnancies involve a much higher risk of perinatal morbidity and mortality than singletons or dichorionic (DC) twins. The risk of fetal or neonatal death in twin pregnancies is around five times higher than singletons, and it is more generous in identical than nonidentical twins.¹⁰ Development of twin-twin transfusion syndrome (TTTS), selective fetal growth restriction (sFGR), and twin anemia polycythemia sequence (TAPS) leads to increased risk of perinatal morbidity and mortality in MC twins. sFGR results mainly from unequal placental sharing. TTTS and TAPS result from unbalanced shunting of blood between the two fetuses across the placental vascular anastomoses.¹¹ TTTS occurs in 10% to 15% of all MC twins, usually between 16 and 26 weeks,¹² and is a fundamental reason for obstacle and death in MC twin pregnancy.¹²

Previous literature suggested that multifetal pregnancy reduction (MFPR) improves the outcomes of multiple pregnancies, especially high-order multiple pregnancies.^13–17 MFPR is defined as a first trimester or early second trimester procedure for reducing the total number of fetuses in a multifetal pregnancy by one or more.¹⁸ In most cases, the involved gestation will be higher-order multifetal pregnancies, defined by the presence of three or more fetuses. MFPR of MC twins in the second or third trimester is mainly implemented due to TTTS or fetal deformity.^19–24 To improve pregnancy outcomes and decrease neonatal and maternal complications, MFPR is a desirable clinical goal.^{13–16,25–29} Considering the pregnancy complications of MC twins and mother’s safety, it is recommended to reduce triplet pregnancy containing MC twins to singleton.^30,31 Given that some couples with infertility are eager to retain twins, this study aimed to investigate the pregnancy outcomes of selective reduction of DCT in order to find an option to avoid MCDA complications and satisfy the couple’s strong wish.

2. METHODS

2.1. Subjects

This is a retrospective study. Subjects diagnosed with triplet pregnancy and underwent selective mechanical reductions in our hospital between 2016 and 2019 were enrolled. This study was approved by the ethics committee of Shanghai Jiai Genetics and IVF Center (No. 2015-36). A written informed consent was obtained from all the participants. All methods involved were carried out by relevant guidelines and regulations of Shanghai Jiai Genetics and IVF Center.

2.2. Diagnosis and counseling

A triplet pregnancy was diagnosed by transvaginal ultrasound during the sixth week of pregnancy. The attending physician informed the couple of the potential risks of multiple pregnancies to the fetus and maternity and introduced the surgical procedure and reduction risks. One week later, the high-order multiple pregnancies were confirmed again by transvaginal ultrasound. All patients underwent counseling on the risks and benefits of reduction again and were advised to reduce triplet pregnancies to singletons or twins. However, if the couples with DCT pregnancies containing MC twins were strongly eager to retain the twins, the potential risks of MC twins to the pregnancy and its uncertainty of selective reduction were explained to the couples. Decisions regarding reduction and the final number of fetuses were made by the patients after counseling. All reductions were carried out by one highly skilled physician.

2.3. Perioperative medication

Anticoagulant drugs, such as aspirin and heparin, were discontinued 1 week before the operation. On the day of surgery, patients stopped taking oral or vaginal progesterone drugs and changed to an intramuscular injection of 40 mg progesterone twice a day. During the perioperative period, an intravenous infusion of cefotaxime sodium (1.5 g in 500 mL saline) was given. Two days after surgery, they were given oral cefradine capsules (500 mg) four times a day for five consecutive days.

2.4. Strategies for choosing the fetus to undergo reduction

All the 66 pregnant women underwent reduction surgery at 7 weeks and 6 days of gestation, and the fetuses to be reduced were selected before surgery. Abnormal or suspected abnormal fetuses were identified by ultrasonography (Fig. 1A), with the following characteristics: (1) relatively small gestational sac or embryo length, (2) gestational sac close to the internal cervical orifice, (3) abnormal echo in the fetal sac, and (4) fetal sac surrounded by fluid.

Fig. 1.

Representative ultrasound images of group A. A, One day before reduction. B, During reduction. The target fetus F3 was reduced by puncture needle (red arrow). C, Two weeks after reduction. F1 = the monochorionic singleton; F2 = the kept fetus of the monochorionic twins; F3 = the reduced fetus of the monochorionic twins; GS1 = gestational sac-1; GS2 = gestational sac-2.

2.5. Reduction procedure

Transvaginal puncture was used instead of transabdominal puncture for the following reasons: (1) smaller gestation age; (2) transvaginal ultrasound is clearer than abdominal ultrasound; (3) shorter route; (4) easier to fix the transvaginal transducer during reduction. During the procedure, the patients were placed in the lithotomy position, with an empty urinary bladder. The vagina was cleaned with povidone-iodine, and the embryos were visualized using a 5 to 7 MHz transvaginal ultrasound transducer (Aloka Ultrasound Machine SSD-3500; Aloka Corporation Inc., Tokyo, Japan) to verify their number, position, size, and heart activity. The entire procedure was conducted using an echo tipped needle (PTC needle, 16 gauge; HAKKO Co Ltd., Chikuma-shi, Nagano, Japan; Meal hub). A puncture guideline was used to determine the puncture path and depth, and the needle was inserted through the vagina, bladder (in some patients), and myometrium to the fetus’s cardiac pulsation area be reduced. By rotating the needle core, the needle tube and needle core were made scissor-shaped. Next, the needle tube and needle core were rotated synchronously until the fetus’s heartbeat to be reduced disappeared. The needle core was then pulled out and reinserted, the reduced fetal fragments were sucked out, and the needle core was withdrawn (Fig. 1B). Transvaginal ultrasonography was performed 3, 10, and 28 days after surgery to evaluate the surgical effect (Fig. 1C), and obstetric follow-up was performed until delivery.

2.6. Statistical analysis

Baseline categorical variables were compared between the two groups by using Chi-squared/Fisher exact tests to detect any differences. For continuous variables, the data distributions were determined by using the Kolmogorov-Smirnov tests. The normally distributed variables were compared by using independent t test. Mann-Whitney U test were used for nonnormally distributed variables. Categorical data are presented as number (percentage), continuous data with normal distribution are presented as the mean ± SD, and the data without normal distribution are presented as median (Q1, Q3). A two-sided p value of <0.05 was regarded as statistically significant. Data management and statistical analyses were conducted by using SAS version 9.4 software (SAS Institute, Inc., Cary, NC).

3. RESULTS

3.1. Patient characteristics

A total of 66 subjects were included in this study, with 38 cases of DCT reduced to DCDA twins (group A) and 28 cases of DCT reduced to MCDA twins (group B). Seven subjects in group B were pregnant with natural conception after ovarian stimulation, and the rest received embryo transfer. The baseline characteristics are listed in Table 1. The mean age at receiving embryo transfer was 33.4 ± 3.5 and 33.8 ± 5.0 years, respectively. Group B had significantly higher percentages of spontaneous abortion after MFPR (21.4% vs 0%, p = 0.004) compared to group A, especially during the first trimester (14.29% vs 0%, p = 0.028). No significant difference was displayed in age, BMI, blood pressure, or obstetric complications between groups.

Table 1.

Baseline characteristics

Characteristic	Group Aa (N = 38)	Group Bb (N = 28)	p
Age at egg retrieval, yr	33.2 ± 3.2	33.1 ± 4.7	0.911
Age at embryo transfer, yr	33.4 ± 3.5	33.8 ± 5.0	0.720
No. of embryos transferred	2.0 (2.0-2.0)	2.0 (2.0-2.0)	0.234
BMI, kg/m2	21.0 (18.7-22.9)	20.6 (19.5-22.3)	0.734
SBP, mmHg	110.0 (105.0-119.0)	119.0 (108.0-121.0)	0.077
DBP, mmHg	72.4 ± 4.2	74.3 ± 4.4	0.095
Have hypertension	1/38 (2.6)	2/26 (7.1)	0.160
Spontaneous abortion after MFPR (%)	0 (0.0)	6 (21.4)	0.004
During the first trimester	0 (0.0)	4 (14.3)	0.028
During the second trimester	0 (0.0)	2 (7.1)	0.176
Obstetric complications
Total obstetric complications, %	10 (26.3)	5 (17.9)	0.555
Gestational hypertension	2 (5.3)	1 (3.6)	1.000
Gestational diabetes	6 (15.8)	1 (3.6)	0.224
Hyperthyroidism	0 (0.0)	0 (0.0)	NA
Hypothyroidism	1 (2.6)	1 (3.6)	1.000
Postpartum hemorrhage	1 (2.6)	1 (3.6)	1.000
Thrombocytopenia	0 (0.0)	1 (3.6)	0.424
Placenta previa	1 (2.6)	0 (0.0)	1.000
Placenta residue	0 (0.0)	0 (0.0)	NA

Continuous data with normal distribution are presented as the mean ± SD, the data without normal distribution are presented as median (Q1, Q3). Categorical data are presented as number (%). Significant values are showing in bold.

BMI = body mass index; DBP = diastolic blood pressure; MFPR = multifetal pregnancy reduction; NA = without Chi-square test due to <2 level; SBP = systolic blood pressure.

^aTriplets with multiple pregnancy twins and multiple pregnancy singletons reduced to dichorionic diamniotic twins.

^bTriplets with multiple pregnancy twins and multiple pregnancy singletons reduced to monochorionic diamniotic twins.

3.2. Obstetrical and perinatal outcomes

The obstetrical and perinatal outcomes are presented in Table 2. A total of 22 MC singletons and 16 from the remaining MC twins in group A, and four MC singletons and 18 MCDA twins in group B were born. Group A had significantly lower rates of late preterm delivery (21.1% vs 45.4%, p = 0.047) and cesarean section (81.6% vs 100%, p = 0.041), and significantly higher rates of full-term delivery (71.0% vs 36.4%, p = 0.009) and take-home baby (100% vs 78.6%, p = 0.004). As expected, the median of gestational age at delivery (38.00 [36.9, 39.0] vs 35.8 [34.4, 37.0] weeks, p = 0.001) and mean of total neonatal weight (2899.72 ± 647.55 vs 2354.38 ± 651.81 g, p < 0.001) in group A were significantly higher than those in group B. Group B had significantly higher percentage of neonates with very low birth weight (12.5% vs 0%, p = 0.012), lower median weight of twins (2350 [2025, 2575] vs 2550 [2250, 2800], p = 0.039), and median weight of larger neonate in twins (2500 [2300, 2600] vs 2790 [2400, 2975], p = 0.045) compared to group A. No significant differences in neonatal morbidities were found between groups.

Table 2.

The obstetrical and perinatal outcomes after excluding spontaneous abortions after MFPR (n = 60)

Characteristic	Group Aa (N = 38)	Group Bb (N = 22)	p
Total preterm delivery, %	11 (29.0)	14 (63.6)	0.009
Preterm delivery (28–34 wk)	3 (7.9)	4 (18.2)	0.405
Preterm delivery (34-37 wk)	8 (21.1)	10 (45.4)	0.047
Full-term delivery, %	27 (71.0)	8 (36.4)	0.009
Take baby home rate, %	38/38 (100.0)	22/28 (78.6)	0.004
Survival rate, %	38/38 (100.0)	22/28 (78.6)	0.004
One survivor	22 (57.9)	4 (18.2)	0.003
Two survivors	16 (42.11)	18 (81.8)	0.003
Cesarean section rate, %	31 (81.6)	22 (100.0)	0.041
Gestational age at delivery, wk	38.0 (36.9-39.0)	35.8 (34.4-37.0)	0.001
Neonatal weight
Total neonatal weight, g	2899.7 ± 647.6	2354.4 ± 651.8	<0.001
Very low birth weight (<1500 g)	0 (0.0)c	5 (12.5)c	0.012
Weight of twins, g	2550.0 (2250.0-2800.0)	2350.0 (2025.0-2575.0)	0.039
Weight of larger neonate in twins, g	2790.0 (2400.0-2975.0)	2500.0 (2300.0-2600.0)	0.045
Weight of smaller neonate in twins, g	2425.0 (2057.5-2550.0)	2200.0 (1800.0-2400.0)	0.189
Weight difference of twins, g	400.0 (237.5-465.0)	250.0 (200.0-350.0)	0.128
Neonatal morbidities
Total neonatal morbidities	1 (2.6)	1 (4.5)	1.000
Ventricular septal defect	1 (2.6)	0 (0.0)	1.000
Renal dysplasia	0 (0.0)	0 (0.0)	NA
Neonatal pneumonia	0 (0.0)	1 (4.5)	0.424
Neonatal jaundice	0 (0.0)	0 (0.0)	NA
Neonatal hypoglycemia	0 (0.0)	0 (0.0)	NA

Continuous data with normal distribution are presented as the mean ± SD. Categorical data are presented as number (%). Significant values are showing in bold.

MFPR = multifetal pregnancy reduction; NA = without Chi-square test due to <2 level.

^aTriplets with multiple pregnancy twins and multiple pregnancy singletons reduced to dichorionic diamniotic twins.

^bTriplets with multiple pregnancy twins and multiple pregnancy singletons reduced to monochorionic diamniotic twins.

^cProportion among total neonatal.

The number of fetal demise events in the two groups is summarized in Table 3. In group A, 22 fetal demises after MFPR occurred, and all were from the remaining MC twins, whereas 21 out of them happened during the first trimester and the other one during the second trimester. In group B, 16 fetal demises after MFPR occurred: 11 (from seven twins) happened during the first trimester, four (from two twins) during the second trimester, and one during the preterm delivery.

Table 3.

Fetal demises after MFPR

Fetal demise	Group Aa (N = 22)	Group Bb (N = 16)
From the MC twin	22	16
During the first trimester	21 (95.45)	11 (68.75)
During the second trimester	1 (4.5)	4 (25)
During labor	0	1 in preterm delivery (6.25)
From MC singleton	0	-

Data are presented as n (%). - indicates MC singleton has been selectively reduced by MFPR, so no MC singleton after MFPR for fetal demise estimation.

MC = monochorionic; MFPR = multifetal pregnancy reduction.

^aTriplets with multiple pregnancy twins and multiple pregnancy singletons reduced to dichorionic diamniotic twins.

^bTriplets with multiple pregnancy twins and multiple pregnancy singletons reduced to monochorionic diamniotic twins.

4. DISCUSSION

This study indicates that selective reduction of DCTs to retain DCDA twins in confers a more appropriate pregnancy and delivery outcome than retaining MCDA twins. It does not increase the risk of neonatal morbidity or obstetric complications. Based on the results and the existing literature, it was proposed that there are two strategies for MFPR with MCDA twins. First, according to the Chinese Society for Reproductive Medicine (CSRM) guideline for multiple pregnancy reduction (2016)³² by the Committee of Chinese Society of Reproductive Medicine, physicians should advise couples to reduce the number of births and retain singletons as much as possible. Several studies have shown a perinatal mortality rate three to seven times higher in MC twins than in DC twins.^30,31 The perinatal mortality rate of DCT with MC twins is twice the DC twins, and the incidence of neonatal morbidity in MC twins is seven times that in DC twins.^30,31 Besides, the incidence of congenital malformations in MC twins is 2.5 times more than the DC twins or singletons.¹⁵ Second, many couples with DCT pregnancies containing MC twins and MC singleton strongly want to retain twins. In such cases, physicians may recommend reducing one from the MC twins to obtain the DCDA twins, according to this study’s results. In this regard, the method of reducing MC singleton and retaining MCDA twins (group B) are not recommended for DCT pregnancy due to its significantly higher abortion rate (21.43%) and shorter gestational age than reducing MCDA twins to retain DCDA twin with two singletons (group A). Selective reduction of one monozygotic twin fetus was thought to be infeasible in MC twins due to its unique characteristics, as the risks of early TTTS happened due to vascular anastomoses of 96% vessels in the single placental bed.^16,33 However, previous studies showed the miscarriage rates ranged from 9.65% to 47.3% in pregnancies retaining MC twins, 5% to 33.3% in pregnancies with reduction of the MC twins, and 5.8% to 15% in conservative management pregnancies.^34–36 Besides, single intrauterine fetal demise in multiple pregnancies has profound consequences for the surviving twin, including a 30% to 50% risk of death.^37,38 The spontaneous abortion rate and premature delivery rate (<32 weeks) of pregnancies with reduction of the MC twins (group A) in this study were observed lower than those with retaining the MC twins (group B) or reported by Antsaklis et al³⁹ (8% and 11%). There were also no cases of very low birth weight infants in group A, better than in group B (16%) and Antsaklis et al³⁹ (11%). Therefore, we believe that MFPR with reduction of the MC twins is a feasible strategy in DCT pregnancies.

Severe neurological injury between 18% and 24% has been reported in single intrauterine fetal demise survivors.^37,40 In the delivery outcome, none of the 16 neonates’ twins in group A had any apparent neurological damage or other complications under the selective reduction of MC twins from DCT, which might be due to the low gestational age at the time of the reduction, whereas intertwin transfusion was not completely established.

We used mechanical reduction techniques rather than the transabdominal injection of 10% potassium chloride solution because of the following reasons: (1) in early gestation age, it is more convenient to perform mechanical reduction; (2) shorter manipulation time; (3) diffusion of 10% potassium chloride solution toward vessels or tissues will influence the rest fetus. After a gestation of 7 to 8 weeks, ultrasound can show chorionic thickening and enhanced echo in the decidua basalis, which allows physicians to identify the following: (1) the number of amniotic sac, (2) Siamese, (3) abnormal conditions of the fetuses, such as umbilical cord cysts, (4) the location of the original placenta, and (5) which fetus is nearest to the cervical inner mouth, thus decreasing the risk of placenta previa in the retained fetus. Some studies have suggested that reducing the fetus close to the cervix may increase the risk of spontaneous abortion after surgery,³ but retaining the fetus close to the cervix may increase the chance of placenta previa. In the present study, it was prioritized to reduce the fetus nearest to the cervical inner mouth. The results showed that the spontaneous abortion rate after the operation was not high and that placenta previa was low in the third trimester. Among the 66 cases, only two underwent spontaneous abortion within 10 days of the operation. One case of placenta previa was found in the third trimester. The low rate of short-term spontaneous abortion after the operation has been related to the adequate luteal support given before and after the operation, the younger gestational age during the operation, the experience of the operating physician, the clean removal of reduced fetal debris during the operation, and the short operation time. Additionally, before the operation, fetal abnormalities, hydrohystera, and the position relationship between each gestational sac and uterine hemorrhage were examined using transvaginal ultrasound to help physicians and patients formulate an operation plan. It may also help lessen the rate of spontaneous abortion after the operation, decrease obstetric complications, and improve pregnancy outcomes.

Selective reduction is mainly performed at gestational 11 to 14 weeks because chorionicity can be correctly diagnosed by ultrasound.⁴¹ Bora et al⁴² reported very high agreement in the diagnosis of chorionicity and amnionicity in twin pregnancies by transvaginal ultrasound at gestational 7 to 9 weeks and 11 to 14 weeks, suggesting similar level of accuracy.

There are some limitations in this study. First, the lack of triple pregnancy cases without MFPR as a control group might lead to uncomprehensive analysis. Second, the small sample size of each group might cause bias of the result and decrease the power of statistical analysis.

In conclusion, our study suggested that the pregnancy outcomes of DCTs reduced to DCDA twins in the first trimester are relatively safer than DCTs reduced to MCDA twins, showing acceptable pregnancy outcomes. This strategy might be a suitable decision for patients with triplet pregnancies who strongly want to retain fraternal twins.