Monochorionic (MC) twin pregnancies are at risk for several disorders including twin-to-twin transfusion syndrome (TTTS), twin anemia–polycythemia sequence, and twin reversed arterial perfusion. In addition, MC twin pregnancies are at increased risk of birth weight discordance (BWD). The most common cause of this condition is unequal placental sharing (Chang et al., Reference Chang, Chang, Chao, Hsieh, Wang and Wang2009). The prevalence of BWD in MC twins is increased compared to dichorionic twins, affecting up to 25% of MC twin pregnancies, and is associated with high rates of perinatal mortality and morbidity (Adegbite et al., Reference Adegbite, Castille, Ward and Bajoria2005; Lewi et al., Reference Lewi, Jani, Blickstein, Huber, Gucciardo, Van Mieghem, Done, Boes, Hecher, Gratacos, Lewi and Deprest2008b; Russell et al., Reference Russell, Quintero and Kontopoulos2007).
The obstetric management in MC twins with BWD is not well established and includes expectant management, elective preterm birth, fetoscopic laser coagulation of the vascular anastomoses, and selective feticide (Lewi et al., Reference Lewi, Gucciardo, Huber, Jani, Mieghem, Done, Cannie, Gratacos, Diemert, Hecher, Lewi and Deprest2008a; Lewi et al., Reference Lewi, Jani, Blickstein, Huber, Gucciardo, Van Mieghem, Done, Boes, Hecher, Gratacos, Lewi and Deprest2008b; Lewi et al., Reference Lewi, Gucciardo, Van, De, Beck, Medek, Van, Devlieger, De and Deprest2010). A major factor influencing the obstetrical management, in particular the timing of delivery, is the increased risk (9–15%) of intrauterine fetal demise (IUFD) in the smaller twin (Lewi et al., Reference Lewi, Gucciardo, Huber, Jani, Mieghem, Done, Cannie, Gratacos, Diemert, Hecher, Lewi and Deprest2008a; Lewi et al., Reference Lewi, Jani, Blickstein, Huber, Gucciardo, Van Mieghem, Done, Boes, Hecher, Gratacos, Lewi and Deprest2008b; Lewi et al., Reference Lewi, Gucciardo, Van, De, Beck, Medek, Van, Devlieger, De and Deprest2010). In case of IUFD, which may occur unexpectedly, the larger co-twin may exsanguinate through the vascular anastomoses, resulting in double fetal demise or cerebral injury due to hypoxic injury (Lewi et al., Reference Lewi, Jani, Blickstein, Huber, Gucciardo, Van Mieghem, Done, Boes, Hecher, Gratacos, Lewi and Deprest2008b; Lewi et al., Reference Lewi, Gucciardo, Van, De, Beck, Medek, Van, Devlieger, De and Deprest2010). Even in the absence of co-twin demise, the larger twin has been reported to be at increased risk (up to 37%) of severe cerebral lesions (Gratacos et al., Reference Gratacos, Carreras, Becker, Lewi, Enriquez, Perapoch, Higueras, Cabero and Deprest2004; Ishii et al., Reference Ishii, Murakoshi, Takahashi, Shinno, Matsushita, Naruse, Torii, Sumie and Nakata2009). In addition, several studies in twins in general (not stratified by chorionicity) have shown that the larger twin has an increased risk of respiratory distress syndrome (RDS) compared to the smaller twin due to induced preterm delivery (Canpolat et al., Reference Canpolat, Yurdakok, Korkmaz, Yigit and Tekinalp2006; Webb and Shaw, Reference Webb and Shaw2001).
The aim of our study was to analyze the incidence of neonatal and neurological morbidity in MC twins with BWD. We hypothesized that the larger twin has an increased risk of RDS and severe neonatal morbidity compared to the smaller co-twin.
Methods
All consecutive MC diamniotic twins with BWD delivered at our center between April 2002 and April 2011 were included in this retrospective study. Inter-twin BWD was defined as a difference in birth weight ≥25% and was calculated using the following formula: (birth weight larger twin – birth weight smaller twin)/birth weight larger twin × 100%. Perinatal outcome data on all MC twins delivered at our center are prospectively entered in our dedicated database. We used the database to retrieve all required medical information for this study. We included only MC twin pregnancies who delivered >24 weeks’ gestation and resulted in two live-born twins. Exclusion criteria were: MC twins with TTTS, monoamnionicity, congenital anomaly, and MC triplets (or higher order). Diagnosis of TTTS was based on internationally accepted standardized antenatal ultrasound criteria (Senat et al., Reference Senat, Deprest, Boulvain, Paupe, Winer and Ville2004).
We recorded the presence of early or late estimated fetal weight (EFW) difference (defined as the detection of EFW discordance before or after 20 weeks’ gestation, respectively) and the presence of selective intrauterine growth restriction (sIUGR; defined as an EFW below the 10th percentile in one twin) (Hadlock et al., Reference Hadlock, Harrist and Martinez-Poyer1991).
All included cases with sIUGR were retrospectively classified according to the classification system recently proposed by Gratacos et al. based on the umbilical artery Doppler flow characteristics: type I (persistently positive end-diastolic flow), type II (persistently absent or reversed end-diastolic flow (AREDF), or type III (intermittent AREDF) (Gratacos et al., Reference Gratacos, Lewi, Munoz, Acosta-Rojas, Hernandez-Andrade, Martinez, Carreras and Deprest2007).
Primary outcome was the incidence of neonatal mortality and severe neonatal morbidity, particularly RDS. Severe neonatal morbidity was defined as a composite outcome including any of the following: RDS, patent ductus arteriosus, necrotizing enterocolitis (NEC), and/or sepsis.
RDS was defined as respiratory failure requiring mechanical ventilation (including continuous positive airway pressure) and oxygen supplementation in the first 24 hours, and typical radiological findings on chest X-rays. Diagnosis of NEC was made according to Bell's criteria (Bell et al., Reference Bell, Ternberg, Feigin, Keating, Marshall, Barton and Brotherton1978). Neonatal sepsis was diagnosed in clinically ill neonates with positive blood cultures.
Secondary outcome was the incidence of cerebral injury detected on neonatal cranial ultrasound examination. After birth, cranial ultrasound scans were performed routinely in all MC twins according to our previously described clinical protocol (Lopriore et al., Reference Lopriore, Wezel-Meijler, Middeldorp, Sueters, Vandenbussche and Walther2006). Intraventricular hemorrhages (IVH) were classified according to Volpe (Volpe, Reference Volpe and Volpe2001) and periventricular leucomalacia (PVL) was graded according to de Vries et al. (de Vries et al., Reference de Vries, Eken and Dubowitz1992). Mild cerebral lesions on cranial ultrasound scans were defined as the presence of at least one of the following: IVH grade I or II, PVL grade I, subependymal pseudocysts, and lenticulostriate vasculopathy. Severe cerebral lesions were defined as the presence of at least one of the following: IVH grade III or IV, PVL ≥ grade II, porencephalic cysts, and ventricular dilatation. Ventricular dilatation was present when the width of one or both lateral ventricles exceeded the 97th percentile (Levene, Reference Levene1981). The study was exempt from Institutional Review Board Approval as ethical approval and informed consent in the Netherlands is not needed for anonymized studies with medical charts.
Statistics
Paired t-tests were used to compare normally distributed values between two groups. Differences between dichotomous variables were assessed with the McNemar test. Results of categorical variables were compared using chi-square test. Logistic regression analysis was performed with weight order (larger vs. smaller twin), birth order (first-born vs. second-born twin), and gender as independent variables and RDS or severe neonatal morbidity as dependent variables.
The results of the logistic models were expressed as odds ratio (OR) and 95% confidence intervals (CI). A p-value <.05 was considered to indicate statistical significance. All statistical data were analyzed using SPSS statistics version 17.0 (SPSS Inc., Chicago, IL, USA).
Results
A total of 47 consecutive MC twin pregnancies with BWD ≥ 25% and two live-born twins were delivered at our center during the study period. In almost all MC pregnancies with BWD (94%, 44/47), the smaller twin was growth restricted (EFW < p10). The MC pregnancies with BWD and sIUGR were classified as type I (32%, 14/44), type II (27%, 12/44), and type III (41%, 18/44), according to the staging system from Gratacos et al. (Gratacos et al., Reference Gratacos, Lewi, Munoz, Acosta-Rojas, Hernandez-Andrade, Martinez, Carreras and Deprest2007). Mean gestational age at birth was 33.2 weeks (range 28–37 weeks). Baseline characteristics are shown in Table 1.
Baseline Characteristics in the 47 MC Twin Pregnancies With Birth Weight Discordance
Note: aValue given as mean ± SD; bReported for pregnancies with gestational at delivery ≤34 weeks; sIUGR = selective intrauterine growth restriction; EFW = estimated fetal weight; UA = umbilical artery; AREDF = absent or reversed end diastolic flow.
The incidence of severe neonatal morbidity in the larger and smaller twin was 38% (18/47) and 19% (9/47), respectively (OR 2.66, 95% CI 0.94–7.44, p = .04) and was due primarily to the higher incidence of RDS, 32% (15/47) and 6% (3/47), respectively (OR 6.88, 95% CI 1.66–32.83, p < .01). After logistic regression analysis, being the larger twin was the only significant risk factor for RDS, whereas birth order or gender were not associated with increased risk of RDS. Detailed information on neonatal morbidity and mortality is presented in Table 2.
Neonatal Morbidity and Mortality in the Smaller Versus Larger Twin
Note: Severe neonatal morbidity is defined as respiratory distress syndrome, patent ductus arteriosus, necrotising enterocolitis, or sepsis.
Cranial ultrasound scans were performed in 98% (92/94) of neonates. We were not able to perform an ultrasound scan in one twin pair delivered at 36 weeks’ gestation, because both infants were quickly discharged from the hospital due to good clinical condition.
The overall incidence of mild and severe cerebral lesions was 18% (17/92) and 1% (1/92), respectively, and was similar between larger and smaller twins. Detailed information on cerebral lesions is shown in Table 3.
Cerebral Lesions Detected on Neonatal Cranial Ultrasound in the Smaller Versus Larger Twin
Note: IVH = intraventricular hemorrhage; PVL = periventricular leukomalacia; n.a. = not applicable.
Severe cerebral lesions were detected in only one neonate, the larger twin in a MC twin pregnancy with Gratacos type III, delivered at 28 + 6 weeks’ gestation due to fetal distress of the smaller twin. Antenatal steroids had not been administered because of acute deterioration and lack of time. The larger twin suffered from severe respiratory failure due to RDS grade III requiring high-frequency oscillation ventilation and surfactant administration. She gradually developed pulmonary interstitial emphysema and required intensive mechanical ventilation and high levels of oxygen. On day 15, extensive bilateral cystic PVL grade III was detected. These cystic lesions were not detected on previous cranial ultrasounds performed on day 1, 3, 7, and 12. Due to increasing respiratory insufficiency and extensive cerebral lesions, intensive care support was withdrawn and the neonate died on day 16. Postmortem cerebral magnetic resonance imaging confirmed the ultrasound findings. The estimated time of origin of the cerebral lesions was 1 week after birth.
Discussion
Fetal growth restriction in singletons has been shown to enhance fetal lung maturation and reduce the risk of RDS due to increased endogenous steroid production caused by chronic intrauterine stress (Gluck and Kulovich, Reference Gluck and Kulovich1973; Torrance et al., Reference Torrance, Derks, Scherjon, Wijnberger and Visser2009; van et al., 2009). In twin pregnancies, the risk of RDS is also reported to be associated with gender and birth order, male infants being more susceptible for RDS and second-born twins being more at risk for perinatal asphyxia and respiratory failure (Canpolat et al., Reference Canpolat, Yurdakok, Korkmaz, Yigit and Tekinalp2006; Webb and Shaw, Reference Webb and Shaw2001). In this study, we found that the risk of RDS in the larger twin in MC twins with BWD is increased compared to the smaller twin. Our findings in MC twins confirm several studies performed in twins in general, irrespective of chorionicity. This is the first study to show this correlation in MC twins. Chorionicity is highly relevant to this finding as MC twins are identical twins and potential bias caused by inter-twin genetic differences is thus excluded.
In a study of preterm discordant twin pairs, Yinon et al. found that the incidence of RDS was significantly higher in the presence of sIUGR compared to discordant twin pairs without sIUGR, 35% (17/46) versus 8% (4/50) (Yinon et al., Reference Yinon, Mazkereth, Rosentzweig, Jarus-Hakak, Schiff and Simchen2005). However, Yinon et al. did not report the difference in RDS between the larger and smaller twins. In a retrospective study of 124 twin pairs (chorionicity not described), Webb and Shaw found that the need for oxygen at 4 hours in twins above 28 weeks gestation was strongly associated with being the larger twin (OR 1.9; 95% CI 1.03–3.46) (Webb and Shaw, Reference Webb and Shaw2001). In a large study of 266 twin pairs (chorionicity not reported), Canpolat et al. reported that being the larger twin was a more important risk factor for RDS (OR 4.5; 95% CI 2.2–9.2) than being male or second-born twin (Canpolat et al., Reference Canpolat, Yurdakok, Korkmaz, Yigit and Tekinalp2006).
In our study, the MC thus identical twins acted as each other's controls for gestational age at birth and maternal factors (including antenatal corticosteroid treatment) which may affect risk of RDS. After logistic regression analysis, birth order and gender were not associated with increased risk of RDS and only weight order (larger vs. smaller twin) was associated with RDS.
This study also shows that the risk of severe cerebral injury in MC twins with BWD is low and does not differ between larger and smaller twins. This is in accordance with several other studies (Chang et al., Reference Chang, Chang, Chao, Lien, Cheng and Chueh2010; Lopriore et al., Reference Lopriore, Slaghekke, Vandenbussche, Middeldorp, Walther and Oepkes2008). However, some studies have shown the opposite, reporting a high incidence (up to 20%) of parenchymal damage in MC twins with sIUGR (Gratacos et al., Reference Gratacos, Carreras, Becker, Lewi, Enriquez, Perapoch, Higueras, Cabero and Deprest2004; Ishii et al., Reference Ishii, Murakoshi, Takahashi, Shinno, Matsushita, Naruse, Torii, Sumie and Nakata2009; Quintero et al., Reference Quintero, Bornick, Morales and Allen2001). Cerebral damage appeared to be particularly prominent in the larger twin in MC pregnancies with intermittent AREDF (Gratacos et al., Reference Gratacos, Carreras, Becker, Lewi, Enriquez, Perapoch, Higueras, Cabero and Deprest2004; Ishii et al., Reference Ishii, Murakoshi, Takahashi, Shinno, Matsushita, Naruse, Torii, Sumie and Nakata2009). Several explanations for the discrepancy in results can be envisaged, such as methodological differences between the studies (e.g., different cranial ultrasound regimens) and the small number of patients included in all these studies, preventing accurate determination of the incidence of cerebral injury. An important factor, which may strongly influence the rate of cerebral injury, is gestational age at delivery. In the cohort from Gratacos et al., mean gestational age at delivery was lower (30.7 weeks’ gestation) (Gratacos et al., Reference Gratacos, Carreras, Becker, Lewi, Enriquez, Perapoch, Higueras, Cabero and Deprest2004) than in our cohort and in the study from Chang et al. (33.2 and 33.4 weeks’ gestation, respectively) (Chang et al., Reference Chang, Chang, Chao, Lien, Cheng and Chueh2010). Lower gestational age rather than sIUGR or Doppler flow patterns may account for the discrepancy in results between the studies. Detailed analysis of the timing of detection of cystic PVL is required to determine whether cerebral injury in MC twins with sIUGR and/or BWD is due to antenatal factors (such as abnormal Doppler measurements of the umbilical artery) or postnatal factors (such as prematurity). Detection of cystic PVL early after birth (within 1–2 weeks) suggests an antenatal etiology of cerebral injury, whereas late detection (from 2 weeks after delivery onward) suggests a postnatal etiology. In the large twin with cystic PVL reported in this study, the cysts were detected at day 15, suggesting a postnatal etiology associated with prematurity. We hypothesize that in analogy with increased risk of RDS due to lung immaturity, larger twins may also be more susceptible to PVL due to brain immaturity.
Optimal obstetrical management in MC twins with BWD and/or sIUGR is not yet well established. Different management options have been proposed, including expectant management with close surveillance, elective preterm birth, fetoscopic laser coagulation of the vascular anastomoses, and selective feticide (in complicated cases) (Lewi et al., Reference Lewi, Gucciardo, Van, De, Beck, Medek, Van, Devlieger, De and Deprest2010). Management decisions are often based on the risk of fetal demise (which is often difficult to predict and may occur unexpectedly) in the smaller twin and on the (controversial) risk of cerebral morbidity in the larger twin. Each management option has its pros and cons that may all affect the incidence of perinatal mortality and cerebral injury in survivors. Elective preterm delivery due to fear of deterioration of the fetal condition of the smaller twin and IUFD prevents fetal demise. However, this may increase the risk of neonatal morbidity (including cerebral injury) due to prematurity, particularly in the larger twin. Expectant management and delayed delivery may reduce prematurity-related morbidity, at a possible cost of the risk of sudden IUFD. Fetoscopic surgical interventions protect the larger twin against exsanguination in case of fetal demise of the smaller twin, but may also lead to complications such as premature rupture of the membranes and preterm delivery. In addition, performance of laser coagulation of vascular anastomoses is technically more challenging in the absence of a polyhydramnios–oligohydramnios situation. In addition, fetoscopic laser may also be a complicated way to achieve a selective reduction as the unequal placental sharing is unlikely to improve the situation for the smaller twin.
Our study was not designed to address obstetrical management. Our data may support the hypothesis that delaying delivery of a high-risk growth restricted twin reduces the risk of adverse effects such as RDS in the larger twin, without increasing the risk of cerebral injury. However, the lack of cranial injury findings may be explained by the study design alone and cannot be extrapolated to antenatal care. Recently, Weisz et al. reported severe cranial injury in 11% of infants but 22% of pregnancies in a small prospective study (Weisz et al., Reference Weisz, Hogen, Yinon, Gindes, Shrim, Simchen, Schiff and Lipitz2011).
Further studies, preferably randomized trials, are warranted to guide fetal medicine specialists and help determining optimal obstetrical management in MC twin pregnancies with BWD and/or sIUGR, including optimization of diagnostic tools (Doppler ultrasound, EFW measurements, cardiotocogram).
Our data should be interpreted with care due to the retrospective nature of the study and the relatively small sample size.
In conclusion, we report a higher risk of RDS and severe neonatal morbidity in the larger twin in MC twins with BWD. As this may seem to be counterintuitive, it is important to be aware of this risk in antenatal counseling, as the small twin may be perceived by parents (and physicians) to be at greater risk of morbidity. In addition, we report no increased incidence of severe cerebral injury in the larger twin, but suggest that the possible association reported in the literature may in part have an ‘iatrogenic’ etiology due to early induced delivery and subsequent injury caused by prematurity. Further studies are urgently needed to determine the optimal management and timing of delivery in these high-risk pregnancies.
Disclosure of Interests
The authors report no conflict of interest.
