Monoamniotic twins occur in approximately 1% of monozygotic twin pregnancies. Monoamniotic twins are at increased risk of preterm delivery and fetal/neonatal death. The causes of these deaths include cord accidents, cord entanglement, and structural anomalies of the infants (Lewi, Reference Lewi2010). A perinatal mortality of 30–70% has been reported (Beasley et al., Reference Beasley, Megerian, Gerson and Roberts1999). However, improvement of ultrasonographic resolution has enabled the detection of monoamniotic twinning during an early stage of pregnancy, and intensified monitoring of such twins may have reduced the overall perinatal mortality (Hack et al., Reference Hack, Derks, Schaap, Lopriore, Elias, Arabin, Eggink, Sollie, Mol, Duvekot, Willekes, Go, Koopman-Esseboom, Vandenbussche and Visser2009; Heyborne et al., Reference Heyborne, Porreco, Garite, Phair and Abril2005). However, the prospective risk of intrauterine fetal death (IUFD) according to gestational week has not been extensively studied.
This retrospective study was conducted to review the overall short-term outcome of monoamniotic twins in Japan and to determine the prospective risk of IUFD so as to adequately counsel parents with monoamniotic twins.
Materials and Methods
This study was approved by the Ethics Committee of Hokkaido University Hospital. A total of 101 women with monoamniotic twin pregnancies were studied. Data on these 101 women were abstracted from the Japan Society of Obstetrics and Gynecology Successive Pregnancy Birth Registry System (JSOG-SPBRS), in which approximately 120 secondary and tertiary hospitals participated and provided information on successive deliveries that occurred at ≥22 weeks of gestation in these hospitals. Approximately 60,000 women were registered annually. The available information from this system included maternal age, parity, gestational week at delivery, chorionicity of the placenta, sex of the infants, birthweight, live-born/stillborn, early neonatal death within 7 days after birth, delivery mode, use of assisted reproductive techniques, maternal complications such as preterm premature rupture of membranes (PPROM), pregnancy-induced hypertension (PIH) including gestational hypertension and pre-eclampsia, placental abruption, placenta previa, twin-to-twin transfusion syndrome (TTTS), hyperglycemia, and structural abnormalities of the infants. A possible cause of death was sometimes described as a special reference for infants with perinatal mortality. These 101 women all had monoamniotic twin pregnancies (maternities) that were registered in this system during an 8-year period between January 2002 and December 2009, corresponding to approximately 0.5% of approximately 20,000 twin pregnancies that were registered during the study period.
Confirmation and/or diagnoses of various conditions such as placental amnionicity and TTTS depended on the institutions that participated in this registry system. However, we have guidelines for the management of twin pregnancies in Japan (JSOG and JAOG, Reference Minakami, Hiramatsu, Koresawa, Fujii, Hamada, Iitsuka, Ikeda, Ishikawa, Ishimoto, Itoh, Kanayama, Kasuga, Kawabata, Konishi, Matsubara, Matsuda, Murakoshi, Ohkuchi, Okai, Saito, Sakai, Satoh, Sekizawa, Suzuki, Takahashi, Tokunaga, Tsukahara and Yoshikawa2011). According to the guidelines, it is recommended not to determine chorionicity until the end of 10 weeks of gestation, to count the numbers of chorions and amnions using ultrasonography, to diagnose as diamniotic in cases of monochorionic twins with a thin dividing membrane, to presume monoamniotic twins and repeat the ultrasonography examination in cases with an unrecognizable dividing membrane, to determine amnionicity (mono- or di-) before 14 weeks of gestation, to examine using ultrasonography at least once every 2 weeks in cases with monochorionic diamniotic twin pregnancy, paying attention to discordances in volumes of the amniotic fluid and/or fetal development, to presume TTTS and examine extensively in a case with a tendency toward polyhydramnios in one twin and oligohydramnios in the co-twin, and to diagnose as TTTS when one twin exhibited polyuric polyhydramnios, with the deepest vertical pool measuring at least 8 cm, and the co-twin exhibited oliguric oligohydramnios, with the deepest vertical pool measuring at most 2 cm.
We focused on the short-term outcomes of twins, including IUFD, early neonatal death within 7 days of life (END), and survivors until 7 days after birth (Alive). The prospective risk of at least a single IUFD at gestational week N can be obtained using the following equation: number of all women who experience a single IUFD and a double IUFD at ≥ gestational week N/number of all women who give birth at ≥ gestational week N (Minakami et al., Reference Minakami, Izumi, Tsukahara and Tamada1993).
Statistical analyses were performed using the statistical software package StatView 5.0 for Macintosh (SAS Institute Inc. Cary, NC, USA). All the data were presented as the mean ± SD.
Results
The mean gestational week at delivery was 31.8 ± 3.7 (Table 1). Ninety (89.2%) women underwent a cesarean delivery. Sixty-four (63.4%) and 32 (31.7%) women gave birth to twins at <34 and <30 weeks of gestation, respectively. Female–female pairs accounted for 55.4% (56/101). Double IUFD and double END occurred in eight and one women, respectively. A total of 28 perinatal deaths (22 IUFDs and 6 ENDs) occurred in 18 women and 174 infants born to 91 women survived 7 days after birth (Tables 1 and 2). Thus, the uncorrected perinatal mortality rate was 13.9% (28/202).
TABLE 1 Demographic Characteristics of 101 Women
Note: ART, assisted reproductive technology; IUFD, intrauterine fetal death; END, early neonatal death within 7 days of life. *For 83 twin pairs in whom both twins survived.
TABLE 2 Specific Causes of 28 Infant Deaths
Note: GW, gestational week; –, absent; #, unknown details; CDH, congenial diaphragmatic hernia; IUFD, intrauterine fetal death; END, early neonatal death within 7 days of life. The cause of death (CE, cord entanglement; TTTS, twin-to-twin transfusion syndrome; SH, subarachnoid hemorrhage; and PH, pulmonary hypertension) were described as special references on the registered forms. Thus, no description of the cause of death does not imply that these events did not occur in other infants with a poor outcome.
A double demise including IUFD/IUFD, IUFD/END, and END/END was more likely to occur during the second trimester, while a single demise (IUFD/Alive or END/Alive) appeared likely to occur at any time during gestation (Figure 1). All 10 double demises occurred exclusively at ≤29 weeks of gestation.
FIGURE 1 Distribution of gestational week at delivery and outcome of infant; ■ (black square), double demise including intrauterine death and early neonatal death within 7 days of life; ▒ (gray square), single demise including intrauterine death or early neonatal death within 7 days of life; □ (white square), both twins alive at 7 days after birth.
A total of 13.9% of women (14/101) who reached gestational week (GW) 22−0/7 experienced at least a single IUFD (Table 2). The prospective risk of at least a single IUFD was 13.9% for women who reached gestational week (GW) 22−0/7, gradually decreasing thereafter but remaining at a relatively high level throughout the third trimester (Figure 2).
FIGURE 2 Prospective risk of at least a single intrauterine death according to gestational week.
As the causes of IUFD and END were not specified for some infants, we focused on maternal complications including PPROM, TTTS, PIH, placental abruption, placenta previa, and structural anomalies of the infants. None of the 101 women developed placental abruption or placenta previa, four (4.0%) developed TTTS, two (2.0%) developed PPROM, and two (2.0%) developed PIH (Table 1). A total of 13 infants (6.4%) had structural anomalies explaining 17.9% (five infants; Cases 8, 10, and 16 in Table 2) of the 28 perinatal deaths. Other structural anomalies such as otocleisis, cleft lip and palate, cleft lip and upper mandible, and one lung agenesis not causing infant death were seen in eight infants. TTTS explained 10.7% (three infants; Cases 9 and 12 in Table 2) of the 28 perinatal deaths. Other causes of death were described in nine infants: seven from cord entanglement (Cases 2, 6, 11, 18 in Table 2) and two from sepsis (Case 3 in Table 2).
Discussion
Although the policies for managing monoamniotic twins may have varied between hospitals participating in the JSOG-SPBRS, the overall perinatal mortality rate was 13.9% (28/202). Malformation of infants explained 5 of the 28 perinatal mortalities in this study. The prospective risk of at least a single IUFD was 13.9% for women who reached 22 weeks of gestation, decreasing during the second trimester but remaining relatively high (around 6.0%) at between 30 and 36 weeks of gestation in this study.
Two reports similar to ours with respect to the maternity number of monoamniotic twins are available (Hack et al., Reference Hack, Derks, Schaap, Lopriore, Elias, Arabin, Eggink, Sollie, Mol, Duvekot, Willekes, Go, Koopman-Esseboom, Vandenbussche and Visser2009; Heyborne et al., Reference Heyborne, Porreco, Garite, Phair and Abril2005): one included 96 cases in which both twins were alive at 12 weeks of gestation and were delivered between January 1993 and December 2003 (Heyborne et al., Reference Heyborne, Porreco, Garite, Phair and Abril2005); and another included 98 women at ≥20 weeks of gestation between January 2000 and December 2007 (Hack et al., Reference Hack, Derks, Schaap, Lopriore, Elias, Arabin, Eggink, Sollie, Mol, Duvekot, Willekes, Go, Koopman-Esseboom, Vandenbussche and Visser2009). For comparison purposes, only data for women with ≥22 weeks of gestation were abstracted from these two studies (Hack et al., Reference Hack, Derks, Schaap, Lopriore, Elias, Arabin, Eggink, Sollie, Mol, Duvekot, Willekes, Go, Koopman-Esseboom, Vandenbussche and Visser2009; Heyborne et al., Reference Heyborne, Porreco, Garite, Phair and Abril2005) (Table 3). Table 3 underscores the common features of the three studies. The perinatal mortality rates were very comparable among the three studies, as shown in Table 3. The crude perinatal mortality is around 15% for women with monoamniotic twins at ≥22 weeks of gestation, although several small studies have reported better outcomes (Baxi & Walsh, Reference Baxi and Walsh2010; Cordero et al. Reference Cordero, Franco and Joy2006; DeFalco et al., Reference DeFalco, Sciscione, Magerian, Tolossa, Macones, O'Shea and Pollock2006; Dias et al., Reference Dias, Mahsud-Dornan, Bhide, Papageorghiou and Thilaganathan2010; Pasquini et al., Reference Pasquini, Wimalasundera, Fichera, Barigye, Chappell and Fisk2006; Rodis et al., Reference Rodis, McIIveen, Egan, Borgida, Turner and Campbell1997).
TABLE 3 Comparison of Infant Outcome in Three Reports Regarding Monoamniotic Twins
Note: §these two studies included neonatal deaths within 28 days of life, while the present study included early neonatal deaths within 7 days of life; *includes 5 infants with major anomalies and 1 therapeutic feticide; **attending perinatologist felt that these deaths were due to cord entanglement not TTTS; ¶includes double intrauterine deaths, single intrauterine deaths followed by neonatal co-twin death, and double neonatal deaths; †includes intrauterine death and neonatal death; ‡includes true knot.
As shown in the present study, the prospective risk of at least a single IUFD remained high (around 6.0%) at between 30 and 36 weeks of gestation. Although cord entanglement has been implicated as a cause of fetal/neonatal death, cord entanglement is observed frequently irrespective of the presence or absence of fetal/neonatal demise (Cordero et al. Reference Cordero, Franco and Joy2006; DeFalco et al., Reference DeFalco, Sciscione, Magerian, Tolossa, Macones, O'Shea and Pollock2006; Dias et al., Reference Dias, Mahsud-Dornan, Bhide, Papageorghiou and Thilaganathan2010; Pasquini et al., Reference Pasquini, Wimalasundera, Fichera, Barigye, Chappell and Fisk2006; Rodis et al., Reference Rodis, McIIveen, Egan, Borgida, Turner and Campbell1997; Suzuki et al., Reference Suzuki, Kaneko and Araki2001): among 11 monoamniotic twin pregnancies, cord entanglement was seen in two (67%) of three pregnancies with IUFD versus seven (88%) of eight pregnancies without IUFD (Suzuki et al., Reference Suzuki, Kaneko and Araki2001); among 36 pregnancies, cord entanglement was seen in 15 (42%) (Cordero et al., Reference Cordero, Franco and Joy2006); among 20 pregnancies with two live structurally normal twins at 20 weeks of gestation, 19 (95%) already had cord entanglement at <20 weeks gestation and all 40 twins survived perinatally (Pasquini et al., Reference Pasquini, Wimalasundera, Fichera, Barigye, Chappell and Fisk2006); among 23 pregnancies, cord entanglement was seen in 10 (47.6%) of 21 pregnancies with both-twin survival versus one (50.0%) of two with IUFD (DeFalco et al., Reference DeFalco, Sciscione, Magerian, Tolossa, Macones, O'Shea and Pollock2006); and cord entanglement was noted in all 13 and 18 monoamniotic twin pregnancies in two reports (Dias et al., Reference Dias, Mahsud-Dornan, Bhide, Papageorghiou and Thilaganathan2010; Rodis et al., Reference Rodis, McIIveen, Egan, Borgida, Turner and Campbell1997). Thus, the presence of cord entanglement itself appeared not to affect the outcome of infants with monoamniotic twins. However, monoamniotic twins were likely to exhibit fetal distress or non-reassuring fetal surveillance: 36 (45.6%) of 79 women and 10 (43.5%) of 23 women underwent a prompt delivery because of fetal distress or non-reassuring fetal surveillance in two reports (DeFalco et al., Reference DeFalco, Sciscione, Magerian, Tolossa, Macones, O'Shea and Pollock2006; Heyborne et al., Reference Heyborne, Porreco, Garite, Phair and Abril2005). Therefore, although not yet verified, many investigators consider that inpatient management and the intensive monitoring of monoamniotic twins and early delivery at 32 to 34 weeks of gestation may improve outcome (Allen et al., Reference Allen, Windrim, Barrett and Ohlsson2001; DeFalco et al., Reference DeFalco, Sciscione, Magerian, Tolossa, Macones, O'Shea and Pollock2006; Dias et al., Reference Dias, Mahsud-Dornan, Bhide, Papageorghiou and Thilaganathan2010; Heyborne et al., Reference Heyborne, Porreco, Garite, Phair and Abril2005; Lewi, Reference Lewi2010; Pasquini et al., Reference Pasquini, Wimalasundera, Fichera, Barigye, Chappell and Fisk2006; Rodis et al., Reference Rodis, McIIveen, Egan, Borgida, Turner and Campbell1997).
In the present study, structural anomalies — including one set of conjoined twins — explained a large fraction of the 17.9% perinatal mortality, consistent with previous studies (Table 3) (Allen et al., Reference Allen, Windrim, Barrett and Ohlsson2001; Cordero et al., Reference Cordero, Franco and Joy2006; Dias et al., Reference Dias, Mahsud-Dornan, Bhide, Papageorghiou and Thilaganathan2010; Hack et al., Reference Hack, Derks, Schaap, Lopriore, Elias, Arabin, Eggink, Sollie, Mol, Duvekot, Willekes, Go, Koopman-Esseboom, Vandenbussche and Visser2009; Heyborne et al., Reference Heyborne, Porreco, Garite, Phair and Abril2005). Apparently, structural anomalies in monoamniotic twins explain a larger fraction of infant deaths than TTTS. TTTS occurs in 10% of cases and is the main cause of death in monochorionic diamniotic twins (Lewi et al., Reference Lewi, Jani, Blickstein, Huber, Gucciardo, Van Mieghem, Doné, Boes, Hecher, Gratacós, Lewi and Deprest2008; Morikawa et al., Reference Morikawa, Yamada, Yamada, Sato and Minakami2011). However, TTTS is less common in monoamniotic twins, with incidences of 3.1% (3/96) (Heyborne et al., Reference Heyborne, Porreco, Garite, Phair and Abril2005), 4.0% (4/101) in the present study, 6.1% (6/98) (Hack et al., Reference Hack, Derks, Schaap, Lopriore, Elias, Arabin, Eggink, Sollie, Mol, Duvekot, Willekes, Go, Koopman-Esseboom, Vandenbussche and Visser2009), 8.3% (3/36) (Cordero et al., Reference Cordero, Franco and Joy2006), and 9.1% (1/11) (Suzuki et al., Reference Suzuki, Kaneko and Araki2001) reported to date.
In conclusion, the present study demonstrated that the prospective risk of IUFD was 13.9% for women who reached gestational week 22−0/7, gradually decreasing thereafter but remaining at between 4.5% and 8.0% between gestational week 30−0/7 and 36−0/7, implying that a considerable number of women experience IUFD at ≥30 weeks of gestation. Structural anomalies of infants explained 17.9% (5/28) of perinatal deaths, but no structural anomalies were seen in four IUFD infants born to four mothers at ≥30 weeks in the present study. Literature review suggested that the presence of cord entanglement is not a reliable risk factor for perinatal deaths. Inpatient management and the intensive monitoring of monoamniotic twins irrespective of the presence or absence of cord entanglement and early delivery at 32 to 34 weeks of gestation may improve outcome.
