The course and outcomes of multiple pregnancy following assisted reproductive technology

Nowadays, multiple pregnancy remains the most important issue in obstetrics due to the development of numerous concomitant complications, primarily miscarriage and preterm birth which largely determine the outcomes of multiple pregnancies [1–5].

The incidence of spontaneous multiple pregnancy is 2.5–3% [6–8]. As a result of the use of assisted reproductive technologies (ART), multiple pregnancies occur 4–16 times more often than in natural fertilization [3, 8, 9] due to the late reproductive age of the mother, the use of ovulation stimulation and in vitro fertilization (IVF) with the transfer of more than one embryo [10]. All of the above confirms the importance of studying the characteristics of the course and outcomes of multiple pregnancies following IVF programs and it determined the purpose of this study.

Materials and Methods

According to the above-mentioned purpose, a total of 669 pregnant women with dichorionic diamniotic twins were examined. The patients were divided into two groups: group I included 266 (39.8%) patients who became pregnant after ART and group II consisted of 403 (60.2%) patients who became pregnant spontaneously. The course and outcomes of pregnancy were evaluated in patients of both groups. The exclusion criterion from the study was the monochorionic type of placentation.

The diagnosis of dichorionic diamniotic twins in all 669 patients was made in the first trimester of pregnancy on the basis of ultrasound imaging of two gestational sacs and two chorions, and the presence of fused chorions confirmed by the twin peak sign (lambda sign) at the base of the inter-twin membrane.

In 53 out of 266 (19.9%) patients of group I (ART) who had 3 or 4 embryos transferred in the IVF program, there was a reduction of 1–2 embryos to twins in the first trimester.

In the future, during pregnancy, the subjects of both groups underwent ultrasound assessment: once every 4 weeks from 16 to 34 weeks, then once every 2 weeks. Birth weight discordance was calculated with the help of the following formula:

(larger twin weight – the smaller twin weight) x 100)/larger twin weight.

The diagnosis of intrauterine growth retardation (IUGR) was made using the ultrasound assessment findings which revealed a decrease in the estimated fetal weight (EFW) of one or both fetuses less than the 10th percentile. Fetometry and ultrasound Doppler assessment of blood flow in case of IUGR were repeated depending on the degree of change in blood flow in the umbilical artery. If the diastolic component of the blood flow spectrum curve in the umbilical artery was positive, fetometry and Doppler assessment were performed twice every 1–2 weeks; if the diastolic component was «zero» or reverse followed by obligatory evaluation of the pulsation index in the venous duct and the middle cerebral artery (MCA), fetometry and ultrasound Doppler were performed weekly up to 26 weeks gestation, and then twice a week.

Fetal deterioration was assessed on the basis of detected changes in blood flow in the UA, venous duct, centralization of fetal blood flow, and oligohydramnios. All patients at 18–20 weeks’ gestation underwent transvaginal echography with measurement of the length of the cervix. Reducing the length of the cervix to less than 25 mm was considered a risk factor for preterm birth. The physical development of children at birth was evaluated according to special standards for the development of fetuses and newborns from multiple pregnancies [11].

Statistical analysis

The IBM SPSS Statistics 23 software package was used for statistical data processing. The D’Agostino– Pearson test was used to determine the normality of the distribution. Data with a normal distribution are presented as an average value (standard deviation), and t-test was used for their comparison. Data with a distribution other than normal are presented as a median (interquartile range) with the Mann–Whitney criterion used for the comparison. Qualitative data are presented as an absolute value (n) and %, Fisher’s exact test was used for their comparison. To assess the relationship between the frequency of gestational complications and the age of patients, the odds ratio (OR) was calculated with a confidence interval (CI) of 95%. The results were considered statistically significant at the level of p<0.05.

Results

The results of our study showed that patients in group I were significantly older (32.43 [4.82] years) than pregnant women in group II (28.74 [4.88] years; p=0.02). In the first trimester, pregnant women of group I (ART), compared with those of group II (spontaneous twins), more often developed complications, such as anemia (29.3% and 19.1%; p<0.001), threatened miscarriage (59.8% and 35.9%; p<0.001), miscarriage (39.8% and 24.6%; p=0.02).

In the second trimester, patients of group I who became pregnant following ART continued being at a high risk of miscarriage compared to those who were pregnant spontaneously. At 14–22 weeks’ gestation miscarriage occurred in 11 out of 669 (1.6%) patients; it was statistically significantly more common in patients of group I (9 out of 266 [3.4%]) than in group II (2 out of 403 [0.5%]; p<0.001). Late miscarriage was caused by the fetal reduction (6 patients from group I), cervical incompetence (3 patients from group I, 2 patients from group II). The rate of late miscarriage in group I after fetal reduction was 11.3% (6 out of 53), and the duration of pregnancy was 32.7 (6.7) weeks, which was less (p=0.03) than in pregnant women of group I without fetal reduction, 34.9 (3.8) weeks.

Cervical incompetence was diagnosed at 15 to 28 weeks’ gestation in 65 out of 658 (9.9%) women pregnant with twins. Cervical incompetence was revealed 3.3 times more often in assisted multiple pregnancies (in 44 out of 257 [17.1%]; p<0.001) than in spontaneous twins (in 21 out of 401 [5.2%]). It should be noted that late reproductive age of patients (37 years or more) contributed to the development of cervical incompetence in patients of group I (OR 2.86; 95% CI: 1.37–6.02, p=0.005); the influence of age on the development of cervical incompetence was statistically insignificant in pregnant women of group II (OR 1.33; 95% CI: 0.29–6.00, p=0.12).

The correction of cervical incompetence was performed in 48 out of 65 (73.8%) patients (32 women from group I, 16 women from group II). The cervix was sutured at 15–22 weeks’ gestation (18.3 [3.4] weeks) in 25 out of 48 (52.1%) pregnant women with severe shortening of the cervix (the length of the cervix is less than 15 mm), including 5 women with amniotic sac prolapse. An obstetric pessary was administered in 23 of 48 (47.9%) patients with the cervix shorter less than 25 mm at a later time (22–28 weeks, 25.4 [4.6] weeks). Micronized vaginal progesterone was prescribed to 40 (out of 48) patients with shortened cervix, along with cerclage and administration of an obstetric pessary. The correction of cervical incontinence was not performed in 17 out of 65 (26.2%) pregnant women with amniotic sac prolapse after the 28th week of gestation. Due to the treatment (cerclage, obstetric pessary, vaginal progesterone), pregnancy was prolonged in 48 patients by 14.3 (3.4) weeks, which was more (p=0.02) than in 17 patients who did not receive treatment, 5.5 (1.4) weeks. The obtained results made it possible to draw a conclusion about the effectiveness of the correction of cervical incompetence by means of cerclage in cases of severe shortening of the cervix in the stage up to 22 weeks gestation, including the cases of amniotic sac prolapse and administration of an obstetric pessary at a later stage (after 22 weeks gestation); our findings are consistent with the opinion of modern authors on the need to correct cervical incontinence in multiple pregnancies [7, 12].

According to fetometry data after 31 weeks’ gestation, clinically significant discordant fetal growth (difference in EFW of 20% or more) was detected in 79 of 603 (13.1%) patients from both groups; 23 of them were diagnosed with IUGR (EFW<10th percentile). At the same time, selective IUGR (difference in EFW of more than 25% in case of IUGR of one fetus less than the 10th percentile) at 31–39 weeks’ gestation was detected in 47.8% (11 out of 23); the differences between the groups were statistically not significant (p=0.37).

According to the fetometry data, the incidence of fetal IUGR by the end of the second trimester was 28.3% (186 out of 658) and did not differ significantly between the groups (p=0.23). As pregnancy progressed, the number of patients with IUGR increased from 28.3% to 41.3% (249 out of 603) at 31–39 weeks’ gestation, apparently due to the competitive co-existence of two fetuses. At the same time, fetal IUGR occurred in 185 out of 249 (74.3%) patients of both groups (EFW<5th percentile was revealed in 12 women), and IUGR of both fetuses was observed in 64 (25.7%) patients (EFW <5th percentile was revealed in 8 women). During the first half of pregnancy the incidence of IUGR did not differ between groups; however, at 31–39 weeks gestation, IUGR was more common (p=0.04) in group I (46.2%) than in group II (38.4%).

The incidence of oligohydramnios was statistically significantly higher (p<0.001) in pregnant women with assisted multiple pregnancies (6.6%), in comparison with the group of patients with spontaneous multiple pregnancies (2.7%); this fact can be explained by the higher incidence of IUGR in group I. As a rule, oligohydramnios was observed in patients with severe IUGR (EFW <5th percentile).

Ultrasound Doppler assessment of the women with assisted multiple pregnancies and spontaneous twins in the second and third trimesters showed statistically insignificant differences in pulsation index in the uterine arteries (p=0.23–0.67). There were also no differences in pulsation index in the left and right uterine arteries (p=0.12–0.45). Dicrotic recess in one or both uterine arteries in the middle of pregnancy was observed in 198 women, later preeclampsia developed in 153 (77.3%) of them.

The average values of pulsation index in the umbilical arteries were higher in induced twins (group I) than in spontaneous twins (group II) at the following periods: 20–21 weeks (1.32 [0.17] and 1.08 [0.10], respectively; p=0.01), 24–25 weeks (1.18 [0.23] and 1.03 [0.12]; p=0.03), 35–36 weeks (0.97 [0.29] and 0.83 [0.15]; p=0.02), 37–38 weeks (0.95 [0.25] and 0.82 [0.16]; p=0.03). However, it should be emphasized that in 127 out of 257 (49.4%) patients of group I, pulsation index in the uterine and umbilical arteries did not exceed the 5th and 95th percentiles of the standard indicators developed by us [13]. There were isolated blood flow disorders in the mother–placenta–fetus (MPF) system in 98 (38.1%) pregnant women, and combined disorders in several parts of the MPF system in 32 (12.5%) women. It should be noted that blood flow indicators in the MPF system were better in pregnant women of group II: 14% of patients had isolated disorders, and only 5.2% of women had combined disorders (versus 12.5% in group I; p<0.001). It should be noted that combined impairments of the MPF system were detected in 17 out of 20 (85%) pregnant women of groups I and II with severe fetal IUGR (EFW <5th percentile).

Preeclampsia developed in 253 of 658 (38.4%) patients with twins. Although there were no statistically significant differences in the incidence of preeclampsia between the groups of pregnant women (p=0.32), there was an earlier (p=0.02) onset of preeclampsia and a higher incidence of severe forms (p=0.03) in group I (23.4 [5.6] weeks; 22.8%) than in group II (26.3 [6.5] weeks;14.3%). The severity of preeclampsia was an indication for early delivery at 22–31 weeks’ gestation in 6 patients. Late reproductive age (37 years or more) contributed to the development of severe forms of preeclampsia in both group I (OR 2.15; 95% CI:1.04–4.45, p=0.005) and group II (OR 3.37; 95% CI:1.33–8.54, p=0.001).

The incidence of anemia in pregnant women with twins increased during pregnancy from 23.2% (in the first trimester) to 42.4% (in the third trimester), but there were no statistically significant differences in the incidence of this complication in the third trimester between groups I and II (p=0.12).

Gestational diabetes mellitus (GDM) was detected more often in pregnant women of group I than in patients of group II: 13.6% versus 8.5% (p=0.01).

Premature detachment of the normally located placenta or placenta previa accompanied by bleeding of varying severity at 24–37 weeks’ gestation was observed in 27 of 658 (4.1%) patients with twins. Although the differences between the groups are not statistically significant (p=0.42), bleeding due to placenta previa occurred in patients of group I at an earlier (p=0.02) gestation period (32.5 [1.9] weeks) than in women of group II (34.6 [1.5] weeks).

One of the main problems of multiple pregnancies which largely influences perinatal outcomes is preterm birth. In our study, preterm birth occurred in 317 of 658 (48.2%) women at 22 to 37 weeks’ gestation. Thus, preterm birth occurred in almost half of pregnant women with twins, and these results are consistent with the data of the world literature, namely 50–60% [3, 4, 14, 15].

It should be noted that extremely and early preterm birth (at 22–31 weeks’ gestation) occurred in 8.4% of pregnant women with twins; preterm birth, including late preterm birth (at 31–37 weeks), occurred in 39.8% of cases. The latter can be explained by the high incidence of induced labor associated with various complications of multiple pregnancies.

It is noteworthy that the frequency of extremely and early preterm birth at 22–31 weeks was more often observed in pregnant women in group I (ART): 13.2% versus 5.2% of pregnant women in group II with spontaneous twins (p=0.01). The difference in the frequency of preterm birth at 31–37 weeks between groups of pregnant women was less remarkable (46.3% and 35.7%), although it was statistically significant (p=0.04). Half of the pregnant women with twins (341 out of 658; 51.8%) had a timely delivery, while the patients who became pregnant after using ART (40.5%) were less likely to have a timely delivery than those with spontaneous twins (59.1%) (p=0.03).

Antenatal losses in pregnant women with twins were noted in 1.3% of cases (17 out of 1316 fetuses). During 22–31 weeks’ gestation, 9 fetuses died, and during 32–38 weeks gestation, there were 17 fatal outcomes. In 6 pregnant women with twins, both children died antenatally, and one of the twins died in 5 women. IUGR less than the 5th percentile and asphyxia were detected in 12 of the 17 (70.6%) dead fetuses, dissociation in EFW with the second fetus more than 25% (selective IUGR) was detected in 5 cases (29.4%). Pregnant women after using ART had antenatal losses 2.1 times more often than women with spontaneous twins: 1.9% (10 out of 514) and 0.9% (7 out of 802), respectively (p<0.001).

Spontaneous vaginal delivery occurred in 262 out of 658 (39.8%) patients. It should be noted that pregnant women of group I gave birth spontaneously less often: 19.1% versus 53.1% of patients from group II (p<0.001). During spontaneous vaginal delivery, the patient experienced the following complications: poor progress of labor in 50 out of 262 (19.1%) patients, discoordinated labor in 24 (9.2%) patients, placental abruption in 8 (3.1%) patients; the differences between the groups were statistically insignificant (p=0.12, p=0.09, p=0.36, respectively).

Operative delivery was performed in the period of 22–39 weeks in 387 out of 658 (58.8%) patients of both groups. More often operative delivery was done at 31–36 weeks (64.9%) than at 37–39 weeks (56.6%) gestation (p=0.04). In pregnant women with assisted multiple pregnancies, cesarean section was performed more frequently (80.2%) than in patients with spontaneous twins (45.1%) (p<0.001). The frequency of urgent cesarean delivery did not differ statistically significantly between groups I and II: 62.1% and 70.7%, respectively (p=0.07). At the same time, the frequency of planned cesarean section in pregnant women of group I was 37.9% compared to 29.3% in group II (p=0.02).

The indications for planned cesarean delivery in patients with twins were the following: non-vertex presentation of one or both fetuses, IUGR of one or both fetuses, preterm birth, severe preeclampsia and lack of the effect from the therapy, poor progress of labor, and scarring on the uterus after cesarean section. In pregnant women of group I (ART), there were additional indications for cesarean delivery: the patient’s late reproductive age (over 37 years), a long history of infertility, and ineffective previous attempts of IVF.

Cesarean section for the delivery of the second fetus was performed in 9 out of 658 (1.4%) patients (2 cases in group I, 7 cases in group II) due to fetal collision (1), umbilical cord prolapse (2), and the transverse position of the second fetus (6).

Hypotonic bleeding in the postpartum period occurred in 45 of 658 (6.8%) women in labor. It should be noted that bleeding was 3.5 times less frequent (p<0.001) in group I (in 2.7%) than in group II (in 9.4%), which can apparently be associated with a higher frequency of spontaneous vaginal delivery in patients with spontaneous multiple pregnancies, as well as a higher frequency of planned cesarean delivery in group I. Bleeding occurred 2.4 times more often (p=0.01) during spontaneous delivery (10.7%) than during cesarean section (4.5%).

A total of 658 patients had 1,285 live births out of 1,316 (97.6%) children: 95.5% in group I (ART), 99% in group II (spontaneous twins). Intranatally, 14 (1.1%) fetuses died. Almost all the fetuses that died during childbirth were from group I, 13 (2.5%) versus 1 (0.1%) (p<0.001). Both fetuses died in 5 patients during childbirth, and one fetus died in 4 patients. Severe IUGR was revealed in 11 fetuses that died intranatally (EFW <5 percentile), asphyxia in breech presentation was detected in 3 fetuses that died intranatally.

Almost half of the infants (604 out of 1285) were born prematurely: 89 babies were born at 22–31 weeks (52 in group I, 37 in group II), and 515 babies were born at 31–37 weeks (232 in group I, 283 in group II). There was a timely delivery of 681 (53%) children: 207 (40.3%) in group I, 474 (59.1%) in group II (p=0.02). The analysis of the physical development indicators of 1285 newborns showed that in premature infants (604) these indicators did not differ statistically significantly in the observed groups (p=0.18–0.45). Full-term neonates of group I (207) compared with full-term infants of group II (474) had less body weight (2691.6 [318.4] and 2747.5 [342.3] g, respectively; p=0.02), weight and height parameters (55.9 [5.1] and 56.6 [5.4]; p=0.03) and chest circumference (31.2 [1.7] and 31.6 [1.7] cm; p=0.03).

The total weight of children of 6000 g or more was 2.6 times less frequently detected in group I: 2.7% versus 7.1% in group II (p<0.001), which also seems to be associated with a lower incidence of hypotonic bleeding in group I. It is worth noting that a third of patients (15 out of 45) with hypotonic bleeding had a total weight of neonates exceeding 6000 g.

The Apgar score in 604 preterm infants of groups I and II at 1 minute was 7.66 (1; 8) points, at 5 minutes – 8.55 (4; 10) points; in 681 full-term infants – 7.90 (5; 9) and 8.85 (6; 10) points, respectively; the differences between the groups were statistically insignificant (p=0.25). Asphyxia was diagnosed more often (p<0.001) in children born at 22–31 weeks gestation (in 38 of 66 infants; 57.6%) and 31–36 weeks gestation (in 20 infants; 30.3%) than in full-term infants (in 8; 12.1%).

Respiratory disorders of various severity at birth were detected in 492 of 1285 (38.3%) children. Due to respiratory failure, 142 out of 492 (28.9%) newborns needed artificial lung ventilation, and 156 (31.7%) infants needed additional ventilation with constant positive pressure. The differences between the groups are statistically insignificant (p=0.46).

Such diagnoses as hydrops fetalis, II-IV degree intraventricular hemorrhage, vegetative and visceral dysfunction and other pathologies were found in 202 out of 1285 (15.7%) children. Differences between the groups were statistically insignificant (p=0.12). In 43 (3.3%) children, congenital malformations were detected (interventricular septal defect, testicular dropsy, polydactyly, etc.); differences between groups I and II were also statistically insignificant (p=0.32), in 18 out of 491(3.7%) infants and 25 out of 794 infants (3.1%).

In the postnatal period, 22 out of 604 (3.6%) premature infants died (mainly at 24–29 weeks): 3.9% of infants in group I; 3.4% of infants in group II (p=0.24). The cause of death was prematurity, respiratory failure, IV degree intraventricular hemorrhage and intrauterine infection.

Thus, the total perinatal loss in pregnant women with twins was 4% (53 out of 1316) in both groups. The indicators of postnatal death of children were almost identical, but the total losses were statistically significantly different between the observation groups due to antenatal and intranatal losses and were 2.8 times higher in the group of patients who became pregnant after assisted technologies in comparison with those pregnant with twins spontaneously, 6.6% versus 2.4% (p=0.01).

Discussion

Asourresearchhasshown, pregnantwomenwithtwins resulting from the use of ART compared with women pregnant spontaneously with twins had a statistically significantly higher rate of gestational complications: anemia in the first trimester, late miscarriage, cervical incompetence, preterm birth, IUGR of a fetus or fetuses at 31–39 weeks, oligohydramnios, and GDM; our results are consistent with the literature data [5, 7, 9, 12, 16]. The late reproductive age of the patients (37 years or more) in both groups was also found to have a statistically significant effect on the development of severe forms of preeclampsia; in group I (ART), late reproductive age was an additional risk factor for the development of cervical incompetence. It is the higher incidence of multiple gestational complications that can be associated with the fact that the indicators of physical development of full-term infants in group I (ART) were lower than ones in newborns from spontaneous twins.

Obstetricians and neonatologists pay particular attention to the condition of infants born to mothers after the use of ART [13, 17–19]. A number of authors [18, 20] note that the condition of newborns in multiple pregnancies after IVF was assessed as satisfactory on the Apgar scale 1.6–2 times less often than in spontaneous multiple pregnancies. In our study, the number of children born with asphyxia did not differ between groups. This confirms our point of view that the condition of infants is more affected by the gestational age at birth than by the type of conception.

In some studies, the relationship between the use of ART and the subsequent development of malformations in infants has been revealed [21, 22]. After a long-term observations, A. Pinborg et al. [23] showed that in recent years there has been a tendency to a decrease in the frequency of malformations in children born after using ART, which can be associated with the improvement of reproductive technologies, as well as with the reduction in the duration of infertility and the younger age of parents. We support the opinion of A. Pinborg et al. [23], considering that in our study the differences in the incidence of malformations in infants in the groups were statistically insignificant.

The study of R. Klemetti et al. [24] showed that perinatal mortality was higher in patients with assisted multiple pregnancies than in spontaneous ones. At the same time, S.R. Murray et al. [25] indicated that ART do not directly affect this indicator, but perinatal mortality in multiple pregnancies is associated primarily with prematurity. We tend to agree with both authors. In our study, perinatal losses were 2.8 times greater in pregnancy resulting from the use of ART than in spontaneous multiple pregnancies (6.6% and 2.4%, respectively), which is probably due to the higher incidence of gestational complications in patients with assisted multiple pregnancies; according to our data, it is prematurity that largely determines the unfavorable perinatal outcomes in pregnant women in the ART group.

Conclusion

Multiple pregnancy following ART in comparison with spontaneously conceived pregnancy is in the highest risk group due to the incidence of multiple gestational complications.

The use of ART does not affect the health of twin newborns, the incidence of various diseases and malformations. However, the total perinatal losses in patients with assisted multiple pregnancies are 2.8 times greater than in spontaneous twins, mainly due to antenatal and intranatal losses

The higher frequency of adverse perinatal outcomes in multiple pregnancies following ART in comparison with spontaneously conceived twins once again highlights the necessity to limit the number of transferred embryos to one.