One of the hallmarks of twin pregnancies is the slower rate of fetal growth when compared with singleton pregnancies during the third trimester. The mechanisms underlying this phenomenon and whether it represents pathology or benign physiological adaptation are currently unclear. One important implication of these questions relates to the type growth charts that should be used by care providers to monitor growth of twin fetuses. If the slower growth represents pathology (ie, intrauterine growth restriction caused uteroplacental insufficiency), it would be preferable to use a singleton growth chart to identify a small twin fetus that is at risk for perinatal mortality and morbidity. If, however, the relative smallness of twins is the result of benign adaptive mechanisms, it is likely preferable to use a twin-based charts to avoid overdiagnosis of intrauterine growth restriction in twin pregnancies. In the current review, we addressed this question by describing the differences in fetal growth between twin and singleton pregnancies, reviewing the current knowledge regarding the mechanisms responsible for slower fetal growth in twins, summarizing available empirical evidence on the diagnostic accuracy of the 2 types of charts for intrauterine growth restriction in twin pregnancies, and addressing the question of whether uncomplicated dichorionic twins are at an increased risk for fetal death when compared with singleton fetuses. We identified a growing body of evidence that shows that the use of twin charts can reduce the proportion of twin fetuses identified with suspected intrauterine growth restriction by up to 8-fold and can lead to a diagnosis of intrauterine growth restriction that is more strongly associated with adverse perinatal outcomes and hypertensive disorders than a diagnosis of intrauterine growth restriction based on a singleton-based chart without compromising the detection of twin fetuses at risk for adverse outcomes caused by uteroplacental insufficiency. We further found that small for gestational age twins are less likely to experience adverse perinatal outcomes or to have evidence of uteroplacental insufficiency than small for gestational age singletons and that recent data question the longstanding view that uncomplicated dichorionic twins are at an increased risk for fetal death caused by placental insufficiency. Overall, it seems that, based on existing evidence, the of use twin charts is reasonable and may be preferred over the use of singleton charts when monitoring the growth of twin fetuses. Still, it is important to note that the available data have considerable limitations and are primarily derived from observational studies. Therefore, adequately-powered trials are likely needed to confirm the benefit of twin charts before their use is adopted by professional societies.
Introduction
One of the hallmarks of twin pregnancies is the slower rate of fetal growth during the third trimester when compared with singleton fetuses. As a consequence, the proportion of twin fetuses identified as small for gestational age (SGA) (defined as fetal weight or birthweight below the 10th percentile for gestational age throughout the current manuscript) based on singleton growth charts is as high as 30% to 53%. ,
The mechanisms underlying fetal growth deceleration in twins and whether this phenomenon represents pathology (ie, intrauterine growth restriction [IUGR]) or physiological adaptation in a predominantly monotocous species when challenged with a multiple pregnancy has been the subject of debate. One important implication of these questions relates to the type of growth charts that should be used to interpret the growth of twin fetuses. If the slower growth of twins represents pathology (ie, IUGR because of the limited capacity of the uteroplacental unit), it would be preferable to use a singleton chart to identify a small twin fetus that is at risk for perinatal mortality and morbidity. If, however, the relative smallness of twins is mainly the result of benign adaptive mechanisms, it is likely preferable to use a twin-based chart, designed to adjust for this physiological adaptation, which must operate in all twins, and thus to avoid overdiagnosis of IUGR in twin gestations. , The direct consequence of reducing the false positive diagnosis of IUGR in twins include a reduction in the utilization of resources (eg, ultrasound examinations), interventions, and patient-partner anxiety. Furthermore, the use of twin charts has the potential to lead to a diagnosis of SGA twin that is more clinically relevant in terms of the association with adverse outcomes. In pursuit of this goal, several groups have developed and published twin-based growth charts ( Figure 1 and Figure 2 ). , , , , ,
Despite these publications, most professional societies do not provide clear recommendations about the standard to assess fetal size and growth in twin pregnancies ( Table 1 ). Some organizations specifically recommend the use of singleton charts, whereas others recommend twin-based charts. In the absence of consistent guidance, most centers default to the use of singleton charts to interpret fetal growth in twins. This current state of affairs and the fear of change are likely driven by the concern that the use of twin-based charts will normalize the slower growth rate of twins when it is in fact pathologic and might therefore compromise the detection of growth restricted twin fetuses at risk for fetal death.
Society | Recommendation |
---|---|
ACOG-SMFM (United States; 2016, 2020) , | Not specified |
CNGOF (France; 2011, 2015) , | Not specified |
FIGO (2021) | Based on the available evidence it seems reasonable to use twin-specific charts for the assessment of fetal growth in twin gestations, because this has the potential to avoid overdiagnosis of IUGR in this population. |
ISUOG (2016) | The slower growth of twins suggests that specific twin growth charts should be used for documenting and monitoring growth in twin pregnancies. However, the use of specific twin growth charts is controversial because of the concern that the reduced growth in the third trimester observed in most twin pregnancies might be caused by some degree of placental insufficiency, warranting close observation. |
NICE (United Kingdom; 2019) | Not specified |
PSANZ (Australia & New Zealand; 2018) | Not specified |
SOGC (Canada; 2017) | Singleton growth curves currently provide the best predictors of adverse outcome in twins and may be used for evaluating growth abnormalities |
The decision on which chart to use should be based on sound evidence regarding the diagnostic accuracy of the 2 types of charts for adverse perinatal outcomes in twin pregnancies. In the current review, we attempted to answer this question by (1) describing the differences in fetal growth between twin and singleton pregnancies; (2) reviewing the current knowledge about the mechanisms responsible for the slower growth of twins; (3) summarizing available empirical evidence on the diagnostic accuracy of the 2 types of charts for IUGR in twin pregnancies; and (4) addressing the question of whether uncomplicated dichorionic twins are at an increased risk for fetal death when compared with singletons.
Differences in Fetal Growth Between Twin and Singleton Fetuses
Fetal weight and birthweight
Several studies compared the growth of twin and singleton fetuses and provided twin-specific birthweight-based and ultrasound-based charts. , , , , , , , The key characteristics of some of these charts are presented in Table 2 and the 50th percentile curves of these charts are compared in Figures 1 and 2 along with a representative singleton chart for comparison. Despite being derived from different populations, most twin charts are very similar to each other and demonstrate consistent differences in relation to singleton charts. This is true for both birthweight-based charts, in which twin fetuses demonstrate a reduced growth rate starting at approximately 28 to 30 weeks ( Figure 1 ), and ultrasound-based charts, in which the reduced growth of twins is observed at an earlier gestational age of approximately 26 to 28 weeks ( Figure 2 ). Given that birthweight-based charts of singletons are known to underestimate fetal growth in the preterm period because of the higher rate of placental dysfunction and growth restriction among infants born prematurely, , it is likely that the ultrasound-based charts ( Figure 2 ) reflect more accurately the differences in intrauterine growth between twin and singleton fetuses and should be preferred over birthweight-based charts also in the case of twin pregnancies.
Reference | Type | Population | Sample size | Stratified by chorionicity | Exclusion of complicated MC twins a |
---|---|---|---|---|---|
Birthweight-based charts | |||||
Alexander et al, 1998 | BW | United States | Singleton: 3,603,971 Twin: 463,856 Triplet: 18,843 | No | No |
Ananth et al, 1998 | BW | United States | DC:1030 MC: 272 | Yes | No |
Gielen et al, 2008 | BW | Netherlands, Belgium | DC:6310 MC:2144 Singleton: 76471 | Yes | No |
Li et al, 2015 | BW | Australia | 85,436 | No | No |
Bricelj et al, 2017 | BW | Slovenia | 5352 | No | No |
Dai et al, 2017 | BW | China | 54,786 | No | No |
Horst et al, 2020 | BW | Poland | 1317 | No | No |
Ultrasound-based charts | |||||
Min et al, 2000 | U/S, retro | United States | 1831 DC: 1282 MC: 348 | Yes | No |
Ong et al, 2002 | U/S, retro | United Kingdom | 884 | No | No |
Araujo et al, 2014 | U/S, retro | Brazil | DC:176 MC:157 | Yes | Yes |
Shivkumar et al, 2015 | U/S, retro | Canada | DC:540 MC:102 | Yes | Yes |
NICHD, Grantz et al, 2016 | U/S, pros | United States | DC: 171 Singleton: 1731 | N/A (DC only) | No |
STORK, Stirrup et al, 2017 | U/S, retro | United Kingdom | DC:1802 MC:323 | Yes | No |
Gabbay-Benziv et al, 2017 | U/S, retro | United States | 2115 | Yes | No |
Hiersch et al, 2019 | U/S, retro | Canada | DC:434 MC:96 | Yes | Yes |
Sekiguchi et al, 2019 | U/S, retro | Japan | DC:190 MC:174 | Yes | Yes |
Wilkof Segev et al, 2019 | U/S, retro | Israel | 195 DC: 136 MC: 32 | No | No |
Savirón-Cornudella et al, 2020 | U/S, retro | Spain | DC:435 MC:83 | Yes | Yes |
a Refers to pregnancies complicated by twin-to-twin transfusion syndrome or selective intrauterine growth restriction.
Because of these differences, the use of singleton charts in twin pregnancies may lead to a relatively large proportion of suspected growth restricted twin fetuses because of what seems to be growth deceleration (or “falling off the curve”) ( Figure 3 ) or because of twin fetuses being classified as SGA ( Figure 4 ). ,
Individual biometric indices
The growth pattern of individual biometric indices between twin and singleton fetuses have been compared previously. , , , The Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD) Fetal Growth Studies likely reflect these differences best because these charts were derived prospectively from the same US population of 171 women with twin pregnancies and 2334 women with low-risk singleton pregnancies ( Figure 5 ), whereas most other studies used previously published singleton charts for the purpose of comparison. When compared with singletons, twin fetuses demonstrate reduced growth of the abdominal circumference starting at approximately 28 weeks, whereas the growth of head circumference and femur length is similar between twins and singletons across gestation ( Figure 5 ). As a consequence, the mean head circumference to abdominal circumference ratio is progressively higher for twins than for singletons starting around 33 weeks of gestation ( Figure 5 ). This observation suggests that twins not only grow slower during the third trimester, but also experience a relatively asymmetric growth pattern when compared with singletons.
Dichorionic vs monochorionic twins
Several studies compared fetal growth between dichorionic and monochorionic twin pregnancies. Although some studies found that monochorionic twins experience reduced growth rates when compared with dichorionic twins, possibly because of the smaller placental mass of monochorionic twins, , the differences were very small ( Figure 6 ). , , , , Furthermore, it should be noted that any differences observed between dichorionic and monochorionic twins could be confounded by the higher rate of complications in monochorionic twins (eg, twin-to-twin transfusion syndrome and selective IUGR) because most studies did not exclude monochorionic twin gestations complicated by these conditions. Finally, the effect of chorionicity on fetal growth seems to be too small to be of clinical significance ( Figure 6 ). For these reasons, and for clarity and simplicity, it seems reasonable that the same dichorionic twin-based chart could be applied to both dichorionic and monochorionic twin pregnancies.
Presenting vs nonpresenting twin
The second (nonpresenting) twin has been reported to be smaller on average than the first (presenting) twin by a mean of 25 to 91 g, and these differences have been reported to persist up to the age of 5 years and 12 years for dizygotic and monozygotic twins, respectively. This finding has been attributed to a higher rate of placental maternal vascular malperfusion pathology in the nonpresenting twin, possibly because of the fetal position and the greater distance of its placenta from the uterine arteries.
Mechanisms Responsible for Fetal Growth Deceleration of Twins—Pathology or Physiology?
A longstanding debate exists about the mechanisms underlying the slower growth of twin fetuses and on whether this phenomenon is pathologic (and thus merits closer monitoring) or a benign physiological adaptation that optimizes the survival of twin fetuses by decreasing their nutritional demands (and thereby decreasing their risk for fetal death) and uterine overdistention (and thereby reducing the risk for prematurity and the associated neonatal mortality). , , The main explanations that have been proposed for this phenomenon are reviewed below.
Slower growth of twins as pathology
The first and probably most intuitive explanation is that the slower growth of twins during the third trimester is caused by a failure of the uteroplacental unit to meet the nutritional requirements of 2 fetuses late in gestation. , One argument used to support this hypothesis is the reported higher risk for fetal death in late gestation in twin pregnancies when compared with singleton pregnancies, although more recent, robust studies have challenged this observation (as described below). The relatively asymmetrical growth pattern of twins compared with singletons (as described previously) may also be supportive of this argument. In contrast, some have disagreed with this explanation and argued that twins already demonstrate reduced growth early in the third trimester at a time when nutrient supply across the placenta is unlikely to be a limiting factor. , , In addition, several studies found that twins diagnosed as SGA based on singleton charts are less likely to have placental histopathological abnormalities than SGA singletons.
A second explanation is that the slower growth of twins is caused by the physical constraint imposed by the uterine size, , although others argue that this mechanism is unlikely given the high compliance of the uterine wall. , The third explanation, known as the placental crowding hypothesis proposed by Bleker et al to describe this phenomenon, is that early concurrence during placental development (ie, the simultaneous development of more than 1 placenta within the uterine cavity) in twin pregnancies limits primary placental growth and, subsequently, fetal growth.
Slower growth of twins as a physiological adaptation
The mean birthweight of singleton fetuses at 38 weeks’ gestation according to North American charts is approximately 3250 g. , , If fetuses in multifetal pregnancies were to follow the same growth trajectory as singletons, the combined fetal weight at 38 weeks in twin pregnancies would have been approximately 6500 g. Such a singleton-like growth trajectory might put twin fetuses at an increased risk for fetal death because the uteroplacental circulation may fail to sustain such a large fetal mass , , and it might increase the risk for preterm birth because of uterine overdistention. Therefore, teleologically, a slower fetal growth pattern in multifetal pregnancies may represent an evolutionary mechanism to optimize perinatal survival by decreasing the risk for undernutrition and uterine overdistention. ,
In line with this rationale, a fourth explanation suggests that the slower growth of twins is a physiological phenomenon that is the result of fetal programming early in pregnancy. , This emerging paradigm suggests that events occurring early in gestation have a critical role in determining the fetal growth trajectory in twin gestations. This hypothesis is supported by evidence that twins may down-regulate their growth rate early in gestation (as early as 16 weeks) , and by studies on the effect of a vanishing twin and fetal reduction in twin gestations on the growth of the surviving twin. , For example, the birthweight following an early reduction of a high-order multiple pregnancy to twins is reported to be lower than nonreduced twin gestations. indicating that fetal reduction, even when occurring in the first trimester, does not completely reverse the slower fetal growth that characterizes multifetal gestations. These findings suggest that early programming, in contrast with a late uterine environment, plays an important role in determining fetal growth throughout gestation. The nature of these early processes is not fully understood. Hormonal mechanisms such as the glucose-insulin and the hypothalamic-pituitary-adrenal axes, which are linked to IUGR in singletons, , , have been suggested to play a role. , Other suggested mechanisms include epigenetic modifications, such as changes in the fetal hypothalamic proopiomelanocortin and glucocorticoid receptor genes and insulin growth factor (IGF) and IGF receptor genes. , This interpretation therefore provides support for the use of twin-based rather than singleton-based charts to avoid overdiagnosis of IUGR in twin gestations.
In summary, the mechanisms responsible for the slower growth of twins and whether this phenomenon represents a pathologic growth restriction or a benign physiological adaptation remains unclear. , This is evident in the literature by the use of terms such as “adaptive growth restriction” or “physiologic growth restriction,” further contributing to the confusion with regards to the nature of the mechanisms responsible for this phenomenon. Therefore, the answer to the question of which chart should be used to interpret the growth of twin fetuses should be based on empirical evidence on whether SGA twins face the same risks as SGA singletons and on the diagnostic accuracy of twin-based charts in comparison with singletons-based charts for adverse perinatal outcomes (eg, fetal death and neonatal mortality and morbidity) or other proxy measures for IUGR in twin pregnancies.
Diagnostic Accuracy of Twin-Based Charts Compared with Singleton-Based Charts for Intrauterine Growth Restriction in Twin Pregnancies
Several studies have attempted to address the question of which type of chart should be preferentially used for twin pregnancies by comparing the outcomes of twin fetuses diagnosed as SGA using singleton-based charts with the outcomes of twin fetuses diagnosed using twin-based charts. Other studies addressed the question of whether the relative smallness of twins represents pathology or a physiological adaptation by comparing the outcomes of SGA twin fetuses (diagnosed based on singleton charts) with the outcomes of SGA singleton fetuses. We summarize hereafter and in Table 3 some of the available evidence to answer each of these 2 questions.
Reference | Design & setting | Study group(s) | Control group | Study outcomes | Rate of SGA in study vs control group | Outcomes in study vs control group | Comments or limitations |
---|---|---|---|---|---|---|---|
Outcome of SGA twins based on twin chart vs singleton chart | |||||||
Mendez-Figueroa, 2018 | Retrospective multicenter study, United States (Maternal-Fetal Medicine Units Network) (N=7673) | SGA twins (BW ≤10th percentile) based on twin birthweight chart of Ananth et al, 1998 | SGA twins (BW ≤10th percentile) based on singleton birthweight chart of Alexander et al, 1996 | Primary outcome: Composite neonatal outcome (5 min Apgar <4, RDS, mechanical ventilation, IVH grade III–IV, NEC grade 2–3, neonatal sepsis, PVL, seizure, stillbirth, or neonatal death) | 4% vs 33% | The odds of the primary outcome among SGA twins (using AGA twins as reference) was higher when SGA was diagnosed using twin chart vs singleton chart (aOR, 1.68; 95% CI, 1.23–2.29 for twin chart vs aOR, 1.19; 95% CI, 1.08–1.32 for singleton chart). | Chorionicity was unknown Most of data was derived from registries for cesarean deliveries |
Shea et al, 2021 | Retrospective single-center (United States) (N=1460) | Group 1: SGA DCDA twins (EFW<10th percentile) based on both the NICHD ultrasound twin chart and the Hadlock ultrasound singleton chart Group 2: SGA DCDA twins (EFW<10th percentile) based on the Hadlock ultrasound singleton chart only | Non-SGA based on either chart | Mild neonatal morbidity: O 2 support or CPAP for <72 h, hypoglycemia, hypocalcemia, hyperbilirubinemia, or IVH grade I–II. Severe neonatal morbidity: NEC grade 2A+, IVH grade III–IV, BPD, mechanical ventilation, or neonatal death | Group 1: 8.1% Group 2: 8.8% | Twins diagnosed as SGA based on both charts (group 1) had greater odds of the study outcomes than those diagnosed as SGA based on only a singleton chart (group 2) or non-SGA twins (mild neonatal morbidity: 57.6% vs 40.3% vs 37.8%, respectively; P <.001; aOR, 2.38; 95% CI, 1.38–4.13 for group 1 vs non-SGA; severe composite: 10.2% vs 2.3% vs 3.4%, respectively; P =.002; aOR, 2.82; 95% CI, 1.16–6.88 for group 1 vs AGA). Twins diagnosed as SGA based on only singleton chart (group 2) were not at increased risk for neonatal mortality compared with non-SGA twins (mild composite: aOR, 1.17; 95% CI, 0.70–1.97; severe composite: aOR, 0.76; 95% CI, 0.17–3.37). | No data on Doppler studies |
Kalafat et al, (STORK) 2019 | Retrospective multicenter (United Kingdom): Cohort 1 (STORK) n=2150 Cohort 2 (SGH) n=759 | SGA (EFW<10th percentile) twins based on ultrasound chorionicity-specific twin chart | Cohort 1 control: SGA (EFW <10th percentile) based on noncustomized singleton birthweight chart Cohort 2: Control 1—SGA (EFW <10th percentile) based on noncustomized singleton birthweight chart ; Control 2—SGA (EFW <10th percentile) based on customized singleton chart | Stillbirth | Cohort 1: 11.6% vs 14.2% Cohort 2: 7.1% (study) vs 8.5% (control 1) vs 12.8% (control 2) | Cohort 1: Detection rate of EFW<10th% for stillbirth – 44% using twin chart vs 48% using singleton chart ( P =.48) Detection rate of EFW<third% for stillbirth – 35% using twin chart vs 38% using singleton chart ( P =.99) Cohort 2: Detection rate of EFW<10th percentile for stillbirth—47% using twin chart, 47% using noncustomized singleton chart, and 53% using customized singleton chart ( P =.99). Detection rate of EFW<third percentile for stillbirth—32% using twin chart, 26% using noncustomized singleton chart, and 37% using customized singleton chart ( P =.99). | Small number of stillbirth cases. Twin charts were originally derived from the STORK cohort |
Proctor et al, 2019 | Retrospective single-center (Canada) 1520 DC twin pregnancies, 48,943 singleton pregnancies | Pregnancies with SGA (BW<10th percentile) DC twins based on twin birthweight charts | Pregnancies with SGA (BW<10th percentile) DC twins based on singleton birthweight chart | Maternal hypertensive disorders of pregnancy | 5.5% vs 21.3% | SGA in twins based on singletons chart was not associated with hypertensive disorders (aRR, 1.3; 95% CI, 0.9–1.7). SGA in twins based on twins chart was associated with hypertensive disorders (aRR, 2.2; 95% CI, 1.6–3.1), and the magnitude of this association was similar to that observed in singleton pregnancies (aRR, 2.2; 95% CI, 2.0–2.4). | A singleton cohort was included as a positive control |
Outcome of SGA twins (using singleton-chart) vs SGA singletons | |||||||
Walker et al, 2020 | Retrospective population-based study (United States) | SGA late-preterm twins (BW<10th percentile) using birthweight singletons chart (n=3363) | SGA late-preterm singletons (BW<10th%) (n=7283) | Fetal death, neonatal death, neonatal morbidity | N/A | SGA twins had a lower risk of fetal death and neonatal death compared with SGA singletons (aOR, 0.16; 95% CI, 0.08–0.31 and aOR, 0.14; 95% CI, 0.03–0.63, respectively). | SGA was defined based on ICD-9 codes. No information regarding which growth charts was used to diagnose SGA. |
Kilpatrick et al, 1996 | Retrospective single center- (United States) | SGA (BW<10th percentile) twins based on singletons birthweight chart (n=193) | SGA (BW<10th%) singletons based on singletons birthweight chart (n=76) | Perinatal mortality | N/A | SGA twins had a significantly lower rate of perinatal mortality compared with SGA singletons (36 per 1000 vs 105 per 1000, P <.02). | |
Odibo et al, 2005 | Retrospective single center- (United States) | SGA (BW<10th percentile) twins based on a singletons birthweight chart (n=99) | SGA (BW<10th percentile) singletons based on a singletons birthweight chart matched 1:4 by gestational age at birth (n=396) | Primary: perinatal mortality Secondary: NICU admission, RDS, IVH grade 3–4, NEC, PVL, and length of NICU stay | N/A | SGA twins had a higher rate of perinatal mortality than SGA singletons (130 per 1000 vs 60 per 1000, aOR, 2.2; 95% CI, 1.1–5.7). | |
Kibel et al, 2017 | Retrospective single-center (Canada) | SGA (BW<10th percentile) twins based on singleton birthweight chart (n=532) | SGA (BW<10th percentile) singletons based on singleton chart (n=954) | Placental pathology | N/A | Compared with SGA singletons, SGA twins were less likely to have any placental pathology (aOR, 0.37; 95% CI, 0.29–0.46), placental weight<10th percentile (aOR, 0.13; 95% CI, 0.08–0.20), maternal vascular malperfusion pathology (aOR, 0.24; 95% CI, 0.18–0.30) or fetal vascular malperfusion pathology (aOR, 0.62; 95% CI, 0.48–0.82). | |
Outcome of twins and singletons in relation to absolute birthweight | |||||||
Joseph et al, 2009 | Retrospective population-based study (United States) | Twin infants born at 36–42 wk (n=256,988) | Singletons infants born at 36–42 wk (n= 17,554,934) | Range of absolute birthweight at each gestational week that is associated with the lowest risk of neonatal mortality or serious morbidity (5-min Apgar score of <4, assisted ventilation for 30 min or more, or neonatal seizures). | N/A | The lower threshold of the optimal range of birthweight for twins was lower by 152 g (95% CI, 121–183 g) than the corresponding threshold for singletons across all gestational weeks and was substantially lower than the 10th percentile according to a singletons birthweight chart. |