Development of customized fetal growth charts in twins

Accepted Manuscript Development of Customized Fetal Growth Charts in Twins Tullio Ghi, Federico Prefumo, Anna Fichera, Mariano Lanna, Enrico Periti, Nicola Persico, Elsa Viora, Giuseppe Rizzo PII: S0002-9378(16)46487-2 DOI: 10.1016/j.ajog.2016.12.176 Reference: YMOB 11476 To appear in: American Journal of Obstetrics and Gynecology Received Date: 9 September 2016 Revised Date: 14 December 2016 Accepted Date: 29 December 2016 Please cite this article as: Ghi T, Prefumo F, Fichera A, Lanna M, Periti E, Persico N, Viora E, Rizzo G, for the Società Italiana di Ecografia Ostetrica e Ginecologica working group on fetal biometric charts, Development of Customized Fetal Growth Charts in Twins, American Journal of Obstetrics and Gynecology (2017), doi: 10.1016/j.ajog.2016.12.176. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. 1 2 3 4 5 6 7 8 9 10 ACCEPTED MANUSCRIPT Manuscript # W16-0651 revised version DEVELOPMENT OF CUSTOMIZED FETAL GROWTH CHARTS IN TWINS Tullio Ghi1, Federico Prefumo2 , Anna Fichera2, Mariano Lanna3, Enrico Periti4, Nicola Persico5, Elsa Viora6, Giuseppe Rizzo7 for the Società Italiana di Ecografia Ostetrica e Ginecologica working group on fetal biometric charts Department of Obstetrics and Gynecology, University of Parma, Italy 11 2. Department of Obstetrics and Gynecology, University of Brescia, Italy 12 3. Department of Obstetrics and Gynecology, University of Milan, Buzzi Children's Hospital Italy 13 4. Department of Obstetrics and Gynecology, Presidio Ospedaliero Centro Piero Palagi, Firenze, Italy 14 5. Department of Obstetrics and Gynecology ‘L. Mangiagalli’, Fondazione IRCCS Ca’ Granda, Ospedale M AN US C 15 RI PT 1. Maggiore Policlinico, Milan, Italy 16 6. Department of Obstetrics and Gynecology, Ospedale Sant’Anna, Turin, Italy 17 7. Department of Obstetrics and Gynecology, University of Rome Tor Vergata, Rome, Italy 18 19 20 21 22 23 24 25 26 27 28 Società Italiana di Ecografia Ostetrica e Ginecologica (SIEOG) working group on fetal biometric charts collaborating authors: Arduini D. Arduino S, Aiello E, Boito S, Celentano C, Chianchiano N, Clerici G., Cosmi E, D’addario V, Di Pietro C, Ettore G, Ferrazzi E, Frusca T, Gabrielli S, Greco P, Lauriola I, Maruotti GM, Mazzocco A, Morano D, Pappalardo E, Piastra A, Rustico M, Todros T, Stampalija T., Visentin S, Volpe N, Volpe P, Zanardini C. Corresponding Author Giuseppe Rizzo, MD TE D The authors report no conflict of interest to declare or financial disclosure Dept Obstetrics and Gynecology 30 Università di Roma Tor Vergata 31 Polo Clinico Assistenziale Santa Famiglia 32 Via dei Gracchi 134 33 00192 Roma Italy 34 Tel +39-06-328331 35 email: giuseppe.rizzo@uniroma2.it 36 word count 5771 37 38 39 AC C EP 29 ACCEPTED MANUSCRIPT 40 Condensation 41 The growth of uncomplicated twin fetuses is infuenced by parental variables and fetal gender and it is 42 reduced in comparison with singletons starting from 26-28 weeks onwards. This reduction is more 43 evident in monochorionic twins. 45 Short versison of the Title 46 Fetal growth in twin pregnancies RI PT 44 AC C EP TE D M AN US C 47 ACCEPTED MANUSCRIPT Background. Twin gestations are at significantly higher risk of fetal growth restriction in comparison 51 with singletons. Using fetal biometric charts customized for obstetrical and parental characteristics may 52 facilitate accurate assessment of fetal growth. 53 Objective(s): To construct reference charts for gestation of fetal biometric parameters stratified by 54 chorionicity and customized for obstetrical and parental characteristics. 55 Study Design: Fetal biometric measurements obtained from serial ultrasound examinations in 56 uncomplicated twin pregnancies delivering after 36 weeks of gestation were collected by 19 Italian fetal 57 medicine units under the auspices of the Società Italiana di Ecografia Ostetrica e Ginecologica. The 58 measurements acquired in each fetus at each examination included biparietal diameter (BPD), head 59 circumference (HC), abdominal circumference (AC) and femur length (FL). Multilevel linear regression 60 models were used to adjust for the serial ultrasonographic measurements obtained and the clustering of 61 each fetus in twin pregnancy. The impact of maternal and paternal characteristics (height, weight, 62 ethnicity), parity, fetal sex and mode of conception were also considered. Models for each parameter 63 were stratified by fetal chorionicity and compared to our previously constructed growth curves for 64 singletons 65 Results: The dataset included 1781 twin pregnancies (dichorionic 66 with 8923 ultrasonographic examination with a median of 5 (range 2-8) observations per pregnancy in 67 dichorionic and 6 in (range 2-11) monochorionic pregnancies. Growth curves of twin pregnancies 68 differed from those of singletons, and differences were more marked in monochorionic twins and during 69 the third trimester. A significant influence of parental characteristics was found. 70 Conclusion(s): Curves of fetal biometric measurements in twins are influenced by parental 71 characteristics. There is a reduction in growth rate during the third trimester. The reference limits for 72 gestation constructed in this study may provide an useful tool for a more accurate assessment of fetal 73 growth in twin pregnancies. monochorionic diamniotic 492) AC C EP TE D M AN US C Abstract RI PT 48 49 50 ACCEPTED MANUSCRIPT 74 Introduction 76 Twin gestations are at significantly higher risk of fetal growth restriction in comparison with singletons, 77 and this may contribute to their increased incidence of the adverse perinatal outcome. Fetal smallness 78 for gestational age may affect one of both fetuses, with an overall incidence estimated at 5%-10% in 79 dichorionic and 15%-25% in monochorionic pairs 1, 2. 80 On this basis an accurate sonographic assessment of fetal biometry is warranted with the aim of 81 detecting cases with substantial growth restriction or discordance, and accordingly guiding the antenatal 82 care. In clinical practice, singleton pregnancy reference charts for ultrasound biometry are often applied 83 to multiple gestations, since specific nomograms for intrauterine growth of twins are few and of 84 uncertain clinical validity. In humans this sounds biologically inappropriate as the growth potential of 85 twins might per se be reduced compared to singletons, being limited by the inability of a woman to cope 86 in late pregnancy with two fetuses growing each at the same rate of a singleton. 87 Most studies have in fact documented a progressive flattening of the fetal growth rate in comparison 88 with singletons starting from 28 to 32 weeks 89 between dichorionic and monochorionic pairs or between uneventful and complicated pregnancies. Very 90 recently some Authors 91 gestations showing a reduced growth rate in monochorionic compared to dichorionic sets. Notably in 92 this study parental factors have not been considered in constructing the nomograms. The use of 93 nomograms customized on the basis of parental factors and fetal sex has been proposed to assess 94 intrauterine fetal growth in singleton gestations. This method compared with standard reference charts 95 has been proven by some to be more efficient in identifying the true small fetuses who are at higher risk 96 of perinatal complications 9-11. 97 The aim of this study was to produce the first longitudinal charts for fetal ultrasound biometry in 98 uncomplicated twins gestations customized for chorionicity and for parental factors. 99 M AN US C RI PT 75 have provided ultrasound biometry charts in a large group of normal twin AC C EP 8 . However, some of these studies failed to differentiate TE D 3-8 ACCEPTED MANUSCRIPT 100 101 Methods 102 Study Population 104 This was a retrospective multicentric study performed in 19 Italian units under the auspices of the 105 Società Italiana di Ecografia Ostetrica e Ginecologica (SIEOG, www.sieog.it). All the units had proven 106 expertise in sonographic assessment of fetal growth and were opted in by the steering committee of the 107 study. Data were obtained from the combined ultrasound and delivery databases of each unit for 108 pregnancies delivered between January 2010 and December 2015. 109 Inclusion criteria were: uncomplicated twin pregnancy of known 110 length in the first 111 two live 112 maternal and paternal height and weight, parity and ethnic group. Gestational age was calculated by 113 CRL of the larger twin using the equation of Robinson and Fleming 114 was based upon the sonographic findings obtained at the first trimester (two placental sites or lambda 115 sign with a single placental site for 116 At that stage accurate labelling of the twins 14 (twin 1 or A vs twin B or 2) was carried out in accordance 117 with placental site (in case of dichorionic pregnancies with two distinct placental masses), fetal position 118 (up and 119 placental mass). Fetal sex was also noted later in pregnancy to facilitate labelling. The maternal weight 120 recorded during the first trimester at the time of the first antenatal visit was considered. 121 Exclusion criteria were: conception by heterologous assisted reproductive 122 chromosomal 123 from a multifetal 124 syndrome (TTTS) or twin anemia-polycytemia sequence (TAPS) pre-existing maternal disease such as 125 hypertension, diabetes, renal and autoimmune M AN US C RI PT 103 known pregnancy dating by crown-rump delivery at or beyond 36 weeks of gestation of birthweight > 5th centile for the national Italian charts information available on . The diagnosis of chorionicity T sign with a single placental site for monochorionic). EP TE D 13 12 AC C right or left) or cord insertion (monochorionic or dichiorionic pregnancies with a single uncertain maternal fetal structural or spontaneous or iatrogenic reduction drug occurrence of twin to twin transfusion the development of obstetric complications ACCEPTED MANUSCRIPT such as pre-eclampsia and gestational diabetes. All the units used the same criteria to define the above 127 mentioned pregnancies complications, according to the guidelines of the Italian National Institute of 128 Health (ISS) for pregnancy care 15. 129 A gestational age interval between 16 and 36 weeks was considered. Longitudinal measurements were 130 required, with a minimum of two sets of measurements for each twin pregnancy. As this was a 131 retrospective analysis of routinely collected anonymized clinical data, no ethical committee approval 132 was necessary according to national regulations. 133 We decided to rely on Italian national standard birthweight charts 134 pregnancies were to exclude due a birthweight below the 5th percentile. In our country we lack 135 customized birthweight charts for twins. RI PT 126 in order to select which twin M AN US C 12 136 137 138 Ultrasound measurements 139 Fetal measurements were all made in accordance with SIEOG guidelines 140 (BPD) and the fetal head circumference (HC) were measured from a cross-sectional view of the fetal 141 head at the level of the thalami, with an angle of insonation of 90° to the midline echoes, a symmetrical 142 appearance of both hemispheres, a continuous midline echo (falx cerebri) broken in middle by the 143 cavum septum pellucidum and no cerebellum visualized. The BPD was measured at the level of the 144 thalami from the outer to the inner edge of the fetal skull. The HC measurements included the outer 145 edge of the proximal calvarial wall and the outer edge of the distal calvarial wall. The abdominal 146 circumference (AC) was measured on a transverse section of the fetal abdomen, showing the stomach 147 bubble, symmetric lower ribs, and the umbilical vein at the level of the portal sinus. The femur length 148 (FL) was measured in its longest axis perpendicular to the transducer direction, with calipers placed at 149 the ends of the ossified diaphysis without including the distal femoral epiphysis. Estimated fetal weight 150 (EFW) was calculated using the Hadlock III formula, that incorporates HC, AC, and FL 17. 151 . The biparietal diameter AC C EP TE D 16 ACCEPTED MANUSCRIPT Statistical analysis 153 Comparison of the characteristics between dichorionic (DC) and monochorionic diamniotic (MCDA) 154 pregnancies was performed using chi square test for categorical variables and t-test or Mann Whitney U 155 test for continuous variables, according to their distribution. For modelling growth curve trajectories of 156 the fetal biometric parameters evaluated we used linear mixed models. The data set considered were 157 hierarchical in nature and a random effect structure that incorporates in the modelling the correlation for 158 both twin-pair and fetus within twin pair was used. The covariates considered in the model as fixed 159 effects were gestational age and other variables potentially influencing the ultrasound measurements as 160 paternal and maternal height (expressed in cm), paternal and maternal weight (expressed in kg), ethnic 161 group (categorized as European, East Asian, Central African and North African) 162 as nulliparous or parous) and gender (categorized as male or female). We performed a logarithmic 163 transformation of gestational age for fitting the models. Using polynomial transformation of different 164 degrees or other method of transformation did not improve the statistical significance. Separate growth 165 curves were built for DC and MCDA twins. These were analyzed in comparison with the growth charts 166 for uncomplicated singleton pregnancies customised for fetal sex, obstetrical and parental characteristics 167 recently developed by SIEOG 168 and singleton pregnancies were evaluated by the Wald test. Statistical analysis was performed using 169 SPSS version 20 (SPSS Inc. Chicago, IL, USA) and R software packages (version 3.1.2, http://www.R- 170 project.org). 172 173 Results , parity (categorized EP TE . Week specific difference in biometric measurements between twins AC C 171 19 18 D M AN US C RI PT 152 174 Complete ultrasound fetal biometric data were obtained from 1781 twin pregnancies including 1289 175 dichorionic (DC) and 439 monochorionic diamniotic (MCDA) gestations who fulfilled the inclusion 176 criteria. Overall 8923 ultrasonographic examinations were available (6640 in DC and 2463 in MCDA). 177 The median number of observations per twin pregnancy was 5 in DC (range 2-8) and 6 in MCDA ACCEPTED MANUSCRIPT (range 2-11). The characteristics of the study population are shown in Table 1. When compared to DC 179 twins, MCDA pregnancies showed a lower incidence of nulliparity (p<0.001), of conception by in vitro 180 fertilization (p<0.001), an earlier gestational age at delivery (p<0.001) and a lower birthweight 181 (p<0.001). No significant differences were found for any other feature considered. 182 Tables 2 to 5 show the fitted regression coefficients and their statistical significance for the biometric 183 variables considered. As expected, gestational age had a significant positive association with all 184 biometric parameters. For BPD, maternal weight (p=0.003) and fetal sex (p<0.0001) were the other 185 associated covariates in DC twins, while in MCDA only the effect of fetal sex (p<0.0001) resulted 186 significant. For HC, maternal weight (p=0.005), maternal height (p=0.004), paternal height (p=0.0015) 187 and fetal sex (p<0.0001) had a significant association in DC twins, while maternal height (p=0.032), 188 paternal height (p=0.05) and fetal sex (p<0.0001) were associated in MCDA twins. Maternal weight 189 (p=0.005), maternal height (p=0.0029) and fetal sex (p<0.0001) resulted significantly related to AC 190 measurements in DC twins, while maternal height (p=0.029) and fetal sex (p=0.0027) showed the same 191 association in MCDA ones. When the FL was analyzed the significant covariates were maternal height 192 (p<0.0001) and paternal height (p<0.0001) in DC and paternal height (p=0.001) in MCDA pregnancies. 193 Since there was a small number of pregnancies in the three non-European groups the data do not allow 194 any comment on the effect of ethnicity on size or growth in twins. The effect size of all the considered 195 covariates in the construction of the mixed regression models are reported in supplemental materials 196 (Supplemental Tables 1-4). 197 Figures 1 and 2 present the growth curves of the biometric parameters considered in DC and MCDA 198 twins, respectively, compared to singletons. Singleton reference limits were constructed using our 199 national growth charts customized for parent characteristics, obstetrical history and fetal sex 200 Similarly in Figure 3 the EFW of DC and MCDA twins were compared to singletons. In order to allow a 201 comparison the same covariates were used for singletons and twins (i.e. European ethnicity for both 202 parents, parity 0, maternal weight 60 kg, maternal height 160 cm, paternal height 180 cm, male fetal 203 sex) and tables (Supplemental Tables 5-9) were generated to allow centile comparison. To allow an easy AC C EP TE D M AN US C RI PT 178 19 . ACCEPTED MANUSCRIPT calculation of the growth curve percentiles in twin pregnancies with different combination of covariates 205 we have created an Excel based file (Additional file 1). 206 The growth curves of DC twin pregnancies appeared to differ significantly from those of singletons, 207 with the reference percentiles of each biometric parameter showing lower values along the whole gesta- 208 tional interval considered. The differences with singleton growth charts were more evident with advanc- 209 ing gestation (Figure 1). When the Wald test was applied to evaluate week-specific differences in the 210 biometric variables between singleton and DC twins, BPD measurements appeared different from 31 211 weeks (p=0.05), HC from 29 weeks (p=0.04), AC from 27 weeks (p=0.05) and FL from 34 weeks 212 (p=0.03) of gestation. 213 Similarly the growth curves of MCDA twin pregnancies appeared to differ significantly from those of 214 singletons, the reference percentiles of each biometric parameter showing lower values along the whole 215 gestational interval considered, Again, the differences with singleton growth charts became more evi- 216 dent with advancing gestation and for some parameter such as AC appeared to increase progressively 217 during the third trimester (Figure 2). Significant differences were evidenced for BDP from 30 weeks 218 (p=0.03), HC from 28 weeks (p=0.05), AC from 26 weeks (p=0.04) and FL from 34 weeks (p=0.05) of 219 gestation. 220 Comparing DC with MCDA pregnancies, the measurements of each biometric index appeared slightly 221 smaller in the latter group, with differences being statistically significant only for AC after 33 weeks of 222 gestation (p=0.03) EP AC C 223 TE D M AN US C RI PT 204 224 Comment 225 Principal findings 226 In a large population of uncomplicated dichorionic and monochorionic twin pregnancies we documented 227 a different growth pattern in comparison with singleton fetuses, with a flattening of the biometric curve 228 starting at 26-28 weeks of gestation for all biometric parameters. Differences with singleton charts were 229 larger in monochorionic twins, progressively increasing during the third trimester for some parameters ACCEPTED MANUSCRIPT such as AC. Moreover, as previously shown in singletons 19, 20, a relationship between fetal biometric 231 data and parental characteristic and fetal gender was documented in both dichorionic and monochorionic 232 twins. 233 Clinical and research implications 234 The use of twin-specific customized growth charts for ultrasound biometry may allow a more accurate 235 assessment of the intrauterine biometry of twins for clinical purposes. In particular this approach may 236 help the provider in distinguishing cases of true fetal smallness among a subgroup of pregnancies whose 237 intrauterine growth potential compared with singleton is per se reduced. On this basis a precise 238 sonographic diagnosis of fetal growth restriction among twin gestations is considered as a cornerstone to 239 optimize their clinical management and to reduce the risk of adverse outcomes. Surprisingly, in common 240 practice the reference charts for the intrauterine growth of twins are very often those in use for the 241 evaluation of singletons. A recent theory has recently suggested that constraints to maternal metabolism 242 increase in pregnancy may limit fetal growth this may further explain why the intrauterine growth rate 243 in twins might be reduced in comparison with singletons 244 twin size charts has been claimed by some as a more reliable tool to assess the intrauterine fetal growth 245 in multiple gestation 246 range of fetal growth downward, has the potential to mask truly growth restricted twins and increase 247 perinatal morbidity from failure to recognize growth restriction. However, having selected as a reference 248 standard a large group of uncomplicated twin gestations delivered close to term with a fetal birthweight 249 of both twins above the 5th percentile of population standards, this should reduce if not abolish the risk 250 of overlooking or masking a fetal growth restriction of one of both fetuses using these charts. This is 251 simply because the biometric data used to produce these charts come from super healthy twin 252 altough the fetal measurements may appear smaller than those of a singleton a good placental function is 253 in fact required to a normal twin to fit in our curves. The choice to exclude those twin pregnancies 254 whose birthweight of one of both fetuses was below the 5th percentile of population standards was made 255 with the aim of maintaining a low threshold to define fetal smallness in twins. Using the 5th percentile M AN US C RI PT 230 . On this basis the construction of specific . In principle, adjusting for multiple pregnancy, thereby shifting the normal AC C EP 5, 22-24 TE D 21 ACCEPTED MANUSCRIPT 256 at 36 weeks or beyond (rather than the 10th percentile as in singletons) as the lower limit to define 257 smallness at birth should account for the reduced intrauterine size of normally growing twins compared 258 to singletons. At the same time it should limit the risk of overlooking fetal growth restriction of twins 259 and considering as biologically normal for two fetuses what is a pathologically reduced growth pattern 260 25 261 We are aware that monochorionic and dichorionic pregnancies have different rates of IUGR and also 262 that the threshold of physiological intertwin discordance of biometric data is varies according to the 263 chorionicity On this point we feel that the use of growth reference charts which are customized for 264 chorionicity may help the clinician also in the interpretation of the intertwin discordance as it should 265 more accurately reflect the specific intrauterine growth pattern of dichorionic and monochorionic twin 266 gestations. In other words using a reference charts which have been specifically designed for dichorionic 267 and monochorionic twin pregnancies the clinician will be able to assess more accurately the degree of 268 intertwin discordance and to determine, in a clinical context, if this difference should be considered 269 physiological or pathological. We decided to include the measurements obtained from both twins at each 270 visit rather than select the measurement of the largest twin. We are aware of the potential risk of 271 downgrading the reference interval for fetal growth, and that this may eventually decrease the sensitivity 272 in detecting antenatally pathological fetal smallness. However, the strict inclusion criteria of our study 273 population should reduce the risk of overlooking fetal growth restriction. RI PT M AN US C D TE EP AC C 274 . 275 Previous Studies 276 Some older studies have shown smaller values for all biometric parameters obtained sonographically in 277 twin gestations compared to singletons. Ong et al. 278 used a single random measurement of the dataset to construct the intrauterine nomograms. In their series 279 the AC values of twins were smaller in comparison with singleton gestations only after 32 weeks of 280 gestation, whereas BPD appeared reduced along the whole pregnancy. However, in the aforementioned 281 study the fetal growth charts were not adjusted for the chorionicity, as the differentiation between 6 assessed 884 twins between 1986 and 1999, and ACCEPTED MANUSCRIPT 282 dichorionic and monochorionic placentae has become accurate only more recently. 283 In 2012 Liao et al. 284 longitudinal prospective study, without differentiating for chorionicity. Using a multilevel regression 285 approach they constructed specific charts for all biometric parameters, whose values appeared smaller 286 compared with those obtained in singleton after 28 weeks. 287 Stirrup et al. 8, using a large database of twin pregnancies, retrospectively built reference charts for all 288 fetal biometric parameters from 14 weeks to term adjusting for chorionicity. Similarly to our findings, 289 they found that ultrasound measurements of fetal growth showed a significant reduction in twin 290 pregnancies, particularly in the third trimester, compared with singletons. Also in their cohort, this 291 reduction was more marked in MCDA gestations. However, they also included complicated twin 292 gestations such as those with twin to twin transfusion or fetal growth restriction, which contributed to 293 the construction of the reference growth charts. This should be acknowledged as a methodological 294 limitation in building the twin specific nomograms for the intrauterine fetal growth. 295 Recently ultrasound based estimated fetal weight reference charts have been retrospectively built in 642 296 uncomplicated dichiorionic and monochorionic twin pregnancies 7. In this study the reference centiles of 297 fetal weight were significantly lower among monochorionic compared to dichiorionic twins along the 298 whole pregnancy. Furthermore, similarly to previous studies, a significant flattening of the intrauterine 299 fetal weight curve in twins compared to singleton in the third trimester was reported, starting earlier in 300 the monochorionic than in the dichiorionic group (28 vs 32 weeks). Finally, a recent study from the 301 National Institute of Child Health and Human Development has shown that compared with singleton 302 fetuses, dichorionic twin fetuses have a progressively asymmetrical slower growth, beginning around 32 303 weeks of gestation 3. 304 In this study, as previously proposed by others 305 parameters obtained at ultrasound and not only the estimated fetal weight. We believe indeed that this is 306 a more appropriate approach when developing fetal growth charts as some parameters may vary 307 according to the ethnicity or the constitutional characteristics of the parents, these differences not being assessed a smaller group of 125 uncomplicated diamniotic twin gestations in a AC C EP TE D M AN US C RI PT 4 26, 27 , we opted to customize all the fetal biometric ACCEPTED MANUSCRIPT specifically reflected by the changes in estimated fetal weight 7. Some of our findings related to the 309 association between fetal biometric parameters and parental characteristics are not easy to interpret: HC 310 has many more significant associations than the BPD. This is not biologically plausible, and might 311 easily be explained by the fact that the variance in BPD measurements is smaller than for HC and thus 312 may not have a sufficient power to reveal associations of HC. Moreover, the fact that maternal weight 313 does not seem to affect significantly the fetal growth charts of twins, as opposed to singletons 19, 20, may 314 be explained by the fact that the mean maternal weight in twins is larger than in singletons. 315 The clinical usefulness of customization has been the object of debate in the last years 10, 28, 29. However 316 a number of publications have shown that in singleton pregnancies the use of customized growth charts 317 is more accurate in identifying the true small fetuses whose risk of perinatal complications is actually 318 increased. 10, 11. Also in multiple pregnancies, the use of customized birth weight charts for twins rather 319 than those for singletons seems more accurate in predicting adverse fetal and neonatal outcomes 320 Following the publication of the large prospective INTERGROWTH-21st study 321 demonstrate a significant impact of the ethnicity on the variability of fetal biometric data, the use of 322 normative universal growth charts has been claimed as more appropriate. However the concept of an 323 optimal fetal growth pattern that should ideally be followed by each fetus has been challenged from a 324 theoretical point of view 325 INTERGROWTH-21st standards may be less effective than population 326 indentifying those small fetuses at risk of perinatal mortality or morbidity. Similar evidence is currently 327 not available for twin pregnancies and also the current study does not allow to conclude that customized 328 biometric models, in comparison to population-derived reference ranges, perform better in terms of their 329 ability to identify individual fetuses at risk of adverse perinatal outcomes. The clinical usefulness of 330 these models should be evaluated in a clinical trial and only if they are shown to be superior should they 331 be considered for use in a clinical context. Altough such a validation can be carried also in retrospect, 332 only a prospective study would be able to provide a convincing demonstration that the use of 333 customized charts produces a measurable benefit in terms of reduction of perinatal morbidity or 30 5, 24 . , which failed to . Moreover, recent evidence from singleton pregnancies suggests that the 32 or customized 9 charts in AC C EP 31 TE D M AN US C RI PT 308 ACCEPTED MANUSCRIPT mortality compared to the use of the standard curves, as recently shown for singletons 9. 335 Strenghts and limitations 336 The main strength of our study is that complicated twin pregnancies were excluded in order to construct 337 unbiased reference charts. In particular as the objective of this study was to build the normal intrauterine 338 biometric charts of healthy uncomplicated dichorionic and monochorionic twin gestations, we decided a 339 priori to exclude from the retrospective data collection the twins whose birthweight was below the 5th 340 centile for population standards, and those who were delivered before 36 weeks. This may affect the 341 construction of the reference interval and determine a selection bias between monochorionic and 342 dichorionic pairs. However the rationale for this choice was to avoid the data contamination with 343 measurements obtained from complicated twin pregnancies. 344 Thanks to a multilevel regression model that takes into account fixed and random effects, we adjusted 345 our curves for chorionicity, parental variables and fetal gender. In particular, a main difference from 346 previous studies is that these two latter factors, both constitutional variables of both parents and fetal 347 gender, were considered in the model and were shown to have a significant impact on the different fetal 348 biometric data, as previously documented in singleton 20. The decision to include also paternal variables 349 seems biologically plausible due to the presumably relevant contribution of the father to the fetal growth 350 however in this study the genetic paternity was based upon maternal report and remained EP TE D M AN US C RI PT 334 19 unproven. Differently from our previous study on singleton gestations , no association was found 352 between the twins biometry and the parental ethnicity, although the lack of significance may depend on 353 the low number of non-European women enrolled in the current study. 354 The participation to this multicentric trial was arbitrarily restricted to units with long standing 355 experience in obstetric ultrasound whose operators are certified by the Italian Society of Ultrasound in 356 Obstetrics and Gynecology, and this is certainly a further strength point of this study. Finally this is to 357 date the study with the largest number of biometric data longitudinally collected in twins used to 358 construct the nomograms of intrauterine fetal growth. 359 Among the main weaknesses of this study it is the retrospective design which prevented us from AC C 351 ACCEPTED MANUSCRIPT validating our growth curves in the clinical practice and assessing if this tool allows a more accurate 361 assessment of intrauterine twins biometry, thus reducing the risk of adverse perinatal outcome. However 362 the design of the study which by definition has retrospectively selected the largest group of normal 363 uneventful twin gestation to construct the reference charts did not allow to test the clinical usefulness of 364 these customized curves in the management of twin gestations. Moreover the rather homogenous racial 365 mix, with approximately 90% of women of European origin probably led to an underestimation of the 366 effect of parental racial origin on twin growth, which was not statistically significant for any biometric 367 parameter. Some population studies on the customization of twin birth weight charts have actually 368 proven an effect of maternal ethnicity on birthweight 369 certainty on paternal data is an additional limitation of our model which supports customization of fetal 370 biometry in accordance to the characteristics of both parents. We customized our growth curves 371 according to the maternal weight prepregnancy weight at the time of the first ultrasound scan. Altough 372 the pre-pregnancy weight is a more reliable index of maternal characteristics independently from the 373 effect of the pregnancy, its exact value may be uncertain or unknown to the 374 pragmatically decided to use the weight which was actually measured by the midiwife and reported in 375 the antenatal notes. Furthermore, availability of the full set of study variables was a criterion of 376 inclusion, and all participating centers shared only datasets containing this information: we were 377 therefore unable to analyze details on the exclusions and the number of each type of exclusion in order 378 to assess the representativity of the population. 379 The two fetuses within the twin pair were in fact treated as two independent fetuses and provided two 380 distinct set of measurement for each visit. However as previously suggested by others 381 measurements within a twin pair are not completely independent from each other and they are correlated 382 to each other. This is biologically consistent with the fact that we may sonographically diagnose 383 chorionicity but not zygosity and this latter factor may significantly affect the interdependency of the 384 biometric data within a dichorionic twin pair. On this base, the use of a regression mixed model which 385 accounts for the correlation of twin measurements has been suggested by some when assessing fetal RI PT 360 M AN US C 23 . The unavailability of the father or the lack of AC C EP TE D therefore we 33 the ACCEPTED MANUSCRIPT biometry and growth of dichorionic twin gestation. This has been done also in our study as specified in 387 the methods section although we acknowledge that this method, despite being widely used, cannot 388 completely adjust for those biological and environmental factors which determine the interdependency 389 of the biometric data within a twin pair. 390 We decided to include the dataset obtained from both twins at each visit rather than select the 391 measurement of the best twin. We are aware of the potential risk of downgrading the reference interval 392 for fetal growth and that this may eventually decrease our sensitivity in detecting antenatally a 393 pathological fetal smallness. However having built our curves with the biometric data of superhealthy 394 uncomplicated dichorionic and monochorionic twin gestations should keep high enough our reference 395 interval reducing the risk of overlooking fetal growth restriction 25. Moreover, a recent analysis suggests 396 that increasing intertwin birthweight discordance is not associated with long-term neuropscychological 397 disadvantages. However it carries an increased risk of neonatal complications and infant mortality which 398 might be, at least in part, iatrogenic 399 discordant twins who may benefit from increased intervention better than currently used standards. 400 A major issue remains whether fetal growth in twins should be measured against a singleton reference: 401 given the higher morbidity associated with twins, correcting for the presence of twins might not be 402 appropriate. However, there is evidence that optimal birthweights are different in twins and in singletons 403 5, 23, 24 404 study is needed to prove that our curves are superior to singleton or non-customized twin curves in 405 clinical practice. . To this effect, our standards may help to better identify those EP TE D 34 M AN US C RI PT 386 AC C , and this justifies adopting specific size and growth references for twins. A prospective validation 406 407 Conclusion 408 In conclusion, this large retrospective study has confirmed that the intrauterine growth of uncomplicated 409 twin pregnancies is reduced in comparison with singletons starting from 26-28 weeks. This reduction is 410 more evident in monochorionic twins. The growth pattern of the fetal biometric parameters is ACCEPTED MANUSCRIPT significantly influenced by parental variables and fetal gender. The reference ranges for gestation 412 constructed in this study may provide an useful tool for a more accurate assessment of fetal growth in 413 twin pregnancies. AC C EP TE D M AN US C RI PT 411 ACCEPTED MANUSCRIPT 414 References: 415 1. 416 twins, and triplets in the United States? Clin Obstet Gynecol 417 2. 418 monochorionic versus dichorionic twin pregnancies: clinical implications of a large Dutch cohort study. 419 BJOG 420 3. 421 Studies. Am J Obstet Gynecol 2 422 4. 423 fetal ultrasound biometry in twin pregnancies. Clinics (Sao Paulo) 424 5. 425 charts for twin gestations to optimize identification of small-for-gestational age fetuses at risk of 426 intrauterine fetal death. Ultrasound Obstet Gynecol 427 6. 428 curves for size. BJOG 429 7. 430 twins. Am J Obstet Gynecol 431 8. 432 Fetal growth reference ranges in twin pregnancy: analysis of the Southwest Thames Obstetric Research 433 Collaborative (STORK) multiple pregnancy cohort. Ultrasound Obstet Gynecol 434 9. 435 birthweight standards for identification of perinatal mortality and morbidity. Am J Obstet Gynecol 436 437 Alexander GR, Kogan M, Martin J, Papiernik E. What are the fetal growth patterns of singletons, -25. RI PT Hack KE, Derks JB, Elias SG, et al. Increased perinatal mortality and morbidity in -67. Grantz KL, Grewal J, Albert PS, et al. Dichorionic twin trajectories: the NICHD Fetal Growth M AN US C -221.e16. Liao AW, Brizot Mde L, Kang HJ, Assuncao RA, Zugaib M. Longitudinal reference ranges for -5. Odibo AO, Cahill AG, Goetzinger KR, Harper LM, Tuuli MG, Macones GA. Customized growth ! -42. D ! TE Ong S, Lim MN, Fitzmaurice A, Campbell D, Smith AP, Smith N. The creation of twin centile !-8. AC C EP Shivkumar S, Himes KP, Hutcheon JA, Platt RW. An ultrasound-based fetal weight reference for ! e1-9. Stirrup OT, Khalil A, D'Antonio F, Thilaganathan B, Southwest Thames Obstetric Research C. ! -7. Anderson NH, Sadler LC, McKinlay CJ, McCowan LM. INTERGROWTH-21st vs customized e1-7. 10. Gardosi J. Customized charts and their role in identifying pregnancies at risk because of fetal ACCEPTED MANUSCRIPT ! 438 growth restriction. J Obstet Gynaecol Can 439 11. 440 rates in England and regional uptake of accreditation training in customised fetal growth assessment. 441 BMJ Open 442 12. 443 compared with other European studies. J Pediatr Gastroenterol Nutr 444 13. 445 J Obstet Gynaecol 1975 446 14. 447 labeling of twin pregnancies on ultrasound. Ultrasound Obstet Gynecol 448 15. 449 Guida 20). Roma: Istituto Superiore di " 450 16. 451 $ 452 17. 453 use of head, body, and femur measurements--a prospective study. Am J Obstet Gynecol 454 18. Steer PJ. Race and ethnicity in biomedical publications. BJOG 455 19. Ghi T, Cariello L, Rizzo L, et al. Customized fetal growth gharts for parents' characteristics, race, 456 and parity by quantile regression analysis: a cross-sectional multicenter Italian study. J Ultrasound Med Gardosi J, Giddings S, Clifford S, Wood L, Francis A. Association between reduced stillbirth !! ! RI PT Bertino E, Spada E, Occhi L, et al. Neonatal anthropometric charts: the Italian neonatal study ! !-61. -10. M AN US C Robinson HP, Fleming JE. A critical evaluation of sonar "crown-rump length" measurements. Br Dias T, Ladd S, Mahsud-Dornan S, Bhide A, Papageorghiou AT, Thilaganathan B. Systematic ! ! -3. Istituto Superiore di Sanità. Sistema Nazionale per le Linee Guida. Gravidanza fisiologica (Linea # 2011. D Società Italiana di Ecografia Ostetrica e Ginecologica. Linee Guida SIEOG 2010. Cento (FE): TE 2010. !!!-7. -7. AC C EP Hadlock FP, Harrist RB, Sharman RS, Deter RL, Park SK. Estimation of fetal weight with the ! 457 -15. !-92. 458 20. Gardosi J, Mongelli M, Wilcox M, Chang A. An adjustable fetal weight standard. Ultrasound 459 Obstet Gynecol 460 21. 461 human altriciality. Proc Natl Acad Sci U S A -74. Dunsworth HM, Warrener AG, Deacon T, Ellison PT, Pontzer H. Metabolic hypothesis for -6. ACCEPTED MANUSCRIPT 462 22. Zhang J, Mikolajczyk R, Lei X, Sun L, Yu H, Cheng W. An adjustable fetal weight standard for 463 twins: a statistical modeling study. BMC Med 464 23. 465 sectional birthweight data. Twin Res Hum Genet 466 24. 467 standards: do singletons and twins need separate standards? Am J Epidemiol 468 25. 469 Gynaecol 470 26. 471 growth of the fetal head, abdomen and femur. Eur J Obstet Gynecol Reprod Biol 472 27. 473 Ultrasound Obstet Gynecol 474 28. 475 birthweight standards. Paediatr Perinat Epidemiol 476 29. 477 fetal growth? J Obstet Gynaecol Can 478 30. 479 on serial ultrasound measurements: the Fetal Growth Longitudinal Study of the INTERGROWTH-21st 480 Project. Lancet 481 31. 482 misconception? Am J Obstet Gynecol 483 32. 484 stillbirths. Ultrasound Obstet Gynecol 2016. 485 33. ! Gielen M, Lindsey PJ, Derom C, et al. Twin-specific intrauterine 'growth' charts based on cross-35. RI PT Joseph KS, Fahey J, Platt RW, et al. An outcome-based approach for the creation of fetal growth -24. !-23. M AN US C Blickstein I. Is it normal for multiples to be smaller than singletons? Best Pract Res Clin Obstet Johnsen SL, Wilsgaard T, Rasmussen S, Sollien R, Kiserud T. Longitudinal reference charts for -85. Pang MW, Leung TN, Sahota DS, Lau TK, Chang AM. Customizing fetal biometric charts. ! -6. D Hutcheon JA, Zhang X, Platt RW, Cnattingius S, Kramer MS. The case against customised TE -6. EP Hutcheon J. Do customized birth weight charts add anything but complexity to the assessment of ! -13. AC C Papageorghiou AT, Ohuma EO, Altman DG, et al. International standards for fetal growth based ! -79. Hanson M, Kiserud T, Visser GH, Brocklehurst P, Schneider EB. Optimal fetal growth: a ! !! e1-4. Poon LC, Tan MY, Yerlikaya G, Syngelaki A, Nicolaides KH. Birthweight in live births and Wright D, Syngelaki A, Staboulidou I, Cruz Jde J, Nicolaides KH. Screening for trisomies in ACCEPTED MANUSCRIPT 486 dichorionic twins by measurement of fetal nuchal translucency thickness according to the mixture 487 model. Prenat Diagn 488 34. 489 pregnancies affected by weight discordance. Acta Obstet Gynecol Scand 2016 Nov 18. doi: 490 10.1111/aogs.13062. ! -21. RI PT Vedel C, Oldenburg A, Worda K, et al. Short and long term perinatal outcome in twin 491 AC C EP TE D M AN US C 492 ACCEPTED MANUSCRIPT 493 Table 1. Characteristics of the twin pregnancies according to chorionicity. Data are expressed as 494 mean±SD or No (%) Monochorionic diamniotic twin N=1289 N=492 Mother maternal age (years) 34.23±5.48 963 (74.68%) height (cm) 165.53±6.15 weight (kg) 62.72±10.73 M AN US C nulliparous RI PT Dichorionic twin Ethnic gruop P value 32.63±5.25 0.679 313 (63.16%) 0.001 165.01±6.01 0.236 61.07±10.99 0.258 European 1192 (92.40%) 442 (89.7%) East Asian 9 (0.70%) 23 (4.8%) 41 (3.20%) 9 (1.8) 47 (3.7%) 18 (3.7) 0.222 422 (32.37%) 54 (10.98%) 0.0001 177.48±8.28 177.08±6.70 0.355 1204 (93.4%) 443 (90.2) 8 (0.6%) 24 (4.8%) 27 (2.1%) 6 (1.1) 50 (3.9%) 19 (3.9) 0.16 37.36±0.710 36.7±0.65 0.001 2648.37±308.34 2516.40±328.41 0.001 637 (49.4%) 227 (46.2) 652 (50.6) 265 (53.8%) Central African North African Conception by in vitro fertilization (IVF) D Father TE height Ethnic group East asia Central African EP European AC C North African Fetus/newborn gestational age at delivery (weeks) Birthweight( g) sex male female 495 496 0.216 ACCEPTED MANUSCRIPT 497 498 499 500 Table 2. Mixed regression models for biparietal diameter (BPD) in dichorionic and monochorionic diamniotic twins. Parameter Estimate Std. Error t p intercept -158.23 0.76 209.25 0.0001 log gestational age 68.57 0.19 366.23 0.0001 mother weight 0.02 0.01 2.99 0.003 sex (female) -0.71 0.13 5.39 0.0001 Intercept -157.64 log gestational age 68.68 sex (female) -0.78 AC C EP TE D 501 502 503 M AN US C Monochorionic diamniotic RI PT Dichorionic 0.73 217.42 0.0001 0.22 313.98 0.0001 0.20 3.94 0.0001 504 505 506 507 508 ACCEPTED MANUSCRIPT Table 3. Mixed regression models for head circumference (HC) in dichorionic and monochorionic diamniotic twins. Parameter Estimate Std. Error t p intercept -591.74 7.72 673.18 0.0001 log gestational age 244.33 0.48 mother weight 0.07 0.03 mother height 0.08 0.04 father height 0.07 0.03 sex (female) -2.69 Intercept RI PT Dichorionic 0.0001 2.80 0.005 2.00 0.04 2.44 0.015 0.34 7.86 0.001 -622.31 12.41 50.42 0.0001 log gestational age 245.53 0.62 403.76 0.0001 mother height 0.14 0.07 2.16 0.032 father height 0.17 0.06 2.84 0.005 -3.72 0.76 4.93 0.0001 M AN US C 510.55 D Monochorionic diamniotic TE sex (female) AC C EP 509 510 511 512 513 514 ACCEPTED MANUSCRIPT Table 4. Mixed regression models for abdominal circumference (AC) in dichorionic and monochorionic diamniotic twins. Parameter Estimate Std. Error t p intercept -646.97 8.51 76.06 0.0001 log gestational age 257.15 0.62 413.39 0.0001 mother weight 0.06 0.03 2.99 0.05 mother height 0.16 0.05 1.91 0.0029 sex (female) -1.68 0.43 3.88 0.0001 Intercept -643.36 log gestational age 255.82 mother height 0.17 sex (female) -2.10 AC C EP TE D 515 516 517 M AN US C Monochorionic diamniotic RI PT Dichorionic 12.85 50.07 0.0001 0.79 324.81 0.0001 0.08 2.20 0.029 0.94 2.22 0.027 518 519 ACCEPTED MANUSCRIPT Tab 5 Mixed regression models for femur length (FL) in dichorionic and monochorionic diamniotic twins. Parameter Estimate Std. Error t p Dichorionic -163,20 2,12 76,99 0,0001 log gestational age 60,17 0,13 469,35 0,0001 mother height 0,05 0,01 4,34 0,0001 father height 0,04 0,01 4,51 0,0001 Intercept -157,43 2,64 -59,585 0,0001 log gestational age 60,37 0,17 365,67 0,0001 father height 0,05 0,01 3,214 0,001 520 521 AC C EP TE D 522 M AN US C Monochorionic diamniotic RI PT intercept 523 ACCEPTED MANUSCRIPT LEGENDS 524 Figure 1: Estimated 5th, 50th and 95th percentiles for BPD (a), HC (b), AC (c) and FL (d) in DC twins 526 (red lines) as obtained from linear mixed models. Data are compared with corresponding reference 527 percentiles in singleton pregnancies (black lines). In both groups values were customized for the same 528 paternal and obstetrical covariates and for fetal sex. 529 RI PT 525 Figure 2: Estimated 5th, 50th and 95th percentiles for BPD (a), HC (b), AC (c) and FL (d) MCDA twins 531 (blue lines) as obtained from linear mixed models. Data are compared with corresponding reference 532 percentiles for in singleton pregnancies (black lines). In both groups values were customized for the 533 same paternal and obstetrical covariates and for fetal sex. M AN US C 530 534 Figure 3: Estimated 5th, 50th and 95th percentiles for estimated fetal weight in DC twins (panel a red 536 lines) and MCDA twins (panel b blue lines). Data are compared with corresponding reference 537 percentiles for in singleton pregnancies (black lines). In both groups values were customized for the 538 same paternal and obstetrical covariates and for fetal sex. AC C EP TE D 535 AC C EP TE D M AN US C RI PT ACCEPTED MANUSCRIPT AC C EP TE D M AN US C RI PT ACCEPTED MANUSCRIPT AC C EP TE D M AN US C RI PT ACCEPTED MANUSCRIPT AC C EP TE D M AN US C RI PT ACCEPTED MANUSCRIPT AC C EP TE D M AN US C RI PT ACCEPTED MANUSCRIPT AC C EP TE D M AN US C RI PT ACCEPTED MANUSCRIPT AC C EP TE D M AN US C RI PT ACCEPTED MANUSCRIPT AC C EP TE D M AN US C RI PT ACCEPTED MANUSCRIPT AC C EP TE D M AN US C RI PT ACCEPTED MANUSCRIPT AC C EP TE D M AN US C RI PT ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT What is known about the subject The intrauterine growth potential of twins is expected to be reduced compared to singletons, being limited by the inability of a woman to cope in late pregnancy with two fetuses growing each at the same rate of a singleton. Some studies in the past have in fact documented a progressive flattening RI PT of the fetal growth rate in comparison with singletons starting from 28 to 32 weeks. However, some of these studies failed to differentiate between dichorionic and monochorionic pairs or between uneventful and complicated pregnancies. M AN US C What is new In this study the first intrauterine biometric charts of healthy uncomplicated twin gestations adjusted for chorionicity, parental variables and fetal gender have been constructed. On a large number of longitudinal observation the intrauterine growth of uncomplicated twin pregnancies has been shown to be reduced in comparison with singletons starting from 26-28 weeks. This reduction is more evident in monochorionic twins. The growth pattern of the fetal biometric parameters is D significantly influenced by parental variables and fetal gender TE Why is this important EP The use of twin-specific customized growth charts for ultrasound biometry should be carefully considered when assessing the intrauterine biometry of twins for clinical purposes, in order to refine AC C the diagnosis of growth abnormalities and take more appropriate clinical decisions. RI PT ACCEPTED MANUSCRIPT AC C EP TE D M AN U SC DEVELOPMENT OF CUSTOMIZED FETAL GROWTH CHARTS IN TWINS Tullio Ghi, Federico Prefumo , Anna Fichera, Mariano Lanna, Enrico Periti, Nicola Persico, Elsa Viora, Giuseppe Rizzo on behalf of the Società Italiana di Ecografia Ostetrica e Ginecologica (SIEOG) working group on fetal biometric charts AJOG 2016 ACCEPTED MANUSCRIPT RI PT Background I AC C EP TE D M AN U SC • Twin gestations are at significantly higher risk of fetal growth restriction in comparison with singletons • The growth potential in twin pregnancies may be limited by anatomical, physiological and metabolic factors. As a consequence the use of specific twin biometric chart rather than those contructed for singleton may be more effective in identifying growth discrepancies Tullio Ghi, Federico Prefumo , Anna Fichera, Mariano Lanna, Enrico Periti, Nicola Persico, Elsa Viora, Giuseppe Rizzo on behalf of the Società Italiana di Ecografia Ostetrica e Ginecologica (SIEOG) working group on fetal biometric charts AJOG 2016 ACCEPTED MANUSCRIPT RI PT Background II AC C EP TE D M AN U SC • In singletons the use of nomograms customized on the basis of parental factors and fetal sex resulted more efficient in identifying abnormal fetal growth than using standard reference charts • Specific normograms have already been constructed for twins pregnancies Tullio Ghi, Federico Prefumo , Anna Fichera, Mariano Lanna, Enrico Periti, Nicola Persico, Elsa Viora, Giuseppe Rizzo on behalf of the Società Italiana di Ecografia Ostetrica e Ginecologica (SIEOG) working group on fetal biometric charts AJOG 2016 ACCEPTED MANUSCRIPT RI PT Background III SC • Limitations of previous studies are AC C EP TE D M AN U – inclusion of complicated pregnancies (e.g. twin to twin transfusion or fetal growth restriction) – no customization for parental and obstetrical characteristics Tullio Ghi, Federico Prefumo , Anna Fichera, Mariano Lanna, Enrico Periti, Nicola Persico, Elsa Viora, Giuseppe Rizzo on behalf of the Società Italiana di Ecografia Ostetrica e Ginecologica (SIEOG) working group on fetal biometric charts AJOG 2016 ACCEPTED MANUSCRIPT RI PT Objective AC C EP TE D M AN U SC • To construct reference charts of fetal biometric parameters based on a population of uncomplicated twin pregnancies stratified by chorionicity and customized for obstetrical and parental characteristics • To evalute differences with singleton pregnancies Tullio Ghi, Federico Prefumo , Anna Fichera, Mariano Lanna, Enrico Periti, Nicola Persico, Elsa Viora, Giuseppe Rizzo on behalf of the Società Italiana di Ecografia Ostetrica e Ginecologica (SIEOG) working group on fetal biometric charts AJOG 2016 ACCEPTED MANUSCRIPT RI PT Study Design AC C EP TE D M AN U SC • Retrospective multicentric study • Inclusion criteria: uncomplicated twin pregnancy of known chorionicity; dating by crown-rump length in the first trimester; known pregnancy outcome; delivery at or beyond 36 weeks of gestation of two live fetuses; birthweight > 5th centile for the national Italian • Serial recordings (at least two ultrasonographic examinations for each pregnancy) of biparietal diameter, head circumference, abdominal circumference and femur length. • Linear mixed models were used for modelling growth curve trajectories of the fetal biometric parameters • Comparison with singleton biometric customized nomograms (Ghi T, et al J Ultrasound Med 2016;35:83-92) Tullio Ghi, Federico Prefumo , Anna Fichera, Mariano Lanna, Enrico Periti, Nicola Persico, Elsa Viora, Giuseppe Rizzo on behalf of the Società Italiana di Ecografia Ostetrica e Ginecologica (SIEOG) working group on fetal biometric charts AJOG 2016 ACCEPTED MANUSCRIPT RI PT Results AC C EP TE D M AN U SC • 1781 twin pregnancies (1289 dichorionic, 492 monochorionic diamniotic) fullfilled the inclusion criteria • Overall 8923 ultrasonographic examinations were available (6640 in dichorionic and 2463 in monochorionic diamniotic) in a gestational age range between 16 and 36 weeks . • The median number of observations per twin pregnancy was 5 in dichorionic (range 2-8) and 6 in monochorionic diamniotic (range 2-11) Tullio Ghi, Federico Prefumo , Anna Fichera, Mariano Lanna, Enrico Periti, Nicola Persico, Elsa Viora, Giuseppe Rizzo on behalf of the Società Italiana di Ecografia Ostetrica e Ginecologica (SIEOG) working group on fetal biometric charts AJOG 2016 ACCEPTED MANUSCRIPT RI PT Significant differences in the characteristics of the pregnancies studied according to chorionicity M AN U N=1289 Monochorionic diamniotic twin N=492 p= SC Dichorionic twin 963 (74.68%) 313 (63.16%) 0.001 Conception by In Vitro Fertilization 422 (32.37%) 54 (10.98%) 0.0001 gestational age at delivery (weeks) 37.36±0.710 36.7±0.65 0.001 Birthweight( (g) 2648.37±308.34 2516.40±328.41 0.001 AC C EP TE D nulliparous Tullio Ghi, Federico Prefumo , Anna Fichera, Mariano Lanna, Enrico Periti, Nicola Persico, Elsa Viora, Giuseppe Rizzo on behalf of the Società Italiana di Ecografia Ostetrica e Ginecologica (SIEOG) working group on fetal biometric charts AJOG 2016 ACCEPTED MANUSCRIPT Results RI PT Significant covariates other than gestational age resulting from mixed regression models Mother weight Fetal sex Head Circumference Mother height Mother weight Father height Fetal sex EP TE D M AN U Biparietal Diameter SC dichorionic Monochorionic diamniotic Fetal sex Mother weight Father height Fetal sex Mother height Fetal sex Femur length Father height AC C Abdominal Cirumference Mother height Mother weight Fetal sex Mother weight Father height Tullio Ghi, Federico Prefumo , Anna Fichera, Mariano Lanna, Enrico Periti, Nicola Persico, Elsa Viora, Giuseppe Rizzo on behalf of the So Italiana di Ecografia Ostetrica e Ginecologica (SIEOG) working group on fetal biometric charts AJOG 2016 AC C EP 80 70 60 50 FL (mm) TE D M AN U SC RI PT ACCEPTED MANUSCRIPT 40 30 20 10 14 18 22 26 30 34 38 gestational age (weeks) Figure 1: Estimated 5th, 50th and 95th percentiles for biparietal diameter (BPD) (a), head circumference (HC) (b), abdominal circumference (AC) (c) and femur length (FL) (d) in dichorionic twins (red lines) as obtained from linear mixed models. Data are compared with corresponding reference percentiles in singleton pregnancies (black lines). In both groups values were customized for the same covariates(maternal weight 60 kg, maternal height 160 cm, paternal height 180 cm, male fetal sex) Tullio Ghi, Federico Prefumo , Anna Fichera, Mariano Lanna, Enrico Periti, Nicola Persico, Elsa Viora, Giuseppe Rizzo on behalf of the Società Italiana di Ecografia Ostetrica e Ginecologica (SIEOG) working group on fetal biometric charts AJOG 2016 AC C EP TE D M AN U SC RI PT ACCEPTED MANUSCRIPT Figure 2: Estimated 5th, 50th and 95th percentiles for biparietal diameter (BPD) (a), head circumference (HC) (b), abdominal circumference (AC) (c) and femur length (FL) (d)i inmonochorionic diamniotic twins (blue lines) as obtained from linear mixed models. Data are compared with corresponding reference percentiles in singleton pregnancies (black lines). In both groups values were customized for the same covariates(maternal weight 60 kg, maternal height 160 cm, paternal height 180 cm, male fetal sex) Tullio Ghi, Federico Prefumo , Anna Fichera, Mariano Lanna, Enrico Periti, Nicola Persico, Elsa Viora, Giuseppe Rizzo on behalf of the Società Italiana di Ecografia Ostetrica e Ginecologica (SIEOG) working group on fetal biometric charts AJOG 2016 3200 3200 2800 2800 2000 2400 SC 2400 RI PT 3600 Estimated fetal weight (grams) 3600 2000 M AN U Estimated fetal weight (grams) ACCEPTED MANUSCRIPT 1600 1200 TE D 800 0 18 22 26 30 34 gestational age (weeks) 38 AC C 14 EP 400 1600 1200 800 400 0 14 18 22 26 30 34 gestational age (weeks) 38 Figure 3: Estimated 5th, 50th and 95th percentiles for estimated fetal weight in DC twins (red lines) and MCDA twins (blue lines). Data are compared with corresponding reference percentiles for in singleton pregnancies (black lines). In both groups values were customized for the same covariates(maternal weight 60 kg, maternal height 160 cm, paternal height 180 cm, male fetal sex) Tullio Ghi, Federico Prefumo , Anna Fichera, Mariano Lanna, Enrico Periti, Nicola Persico, Elsa Viora, Giuseppe Rizzo on behalf of the Società Italiana di Ecografia Ostetrica e Ginecologica (SIEOG) working group on fetal biometric charts AJOG 2016 ACCEPTED MANUSCRIPT Results RI PT Comparison with singletons charts* M AN U SC Statistically significant differences were evidenced by the Wald test from the following gestational ages onwards dichorionic > 31 weeks Head Circumference > 29 weeks EP TE D Biparietal Diameter Femur length AC C Abdominal Cirumference > 27 weeks > 34 weeks monochorionic diamniotic > 30 weeks > 28 weeks > 26 weeks > 34 weeks * Ghi T, et al. Customized fetal growth charts for parents' characteristics, race, and parity by quantile regression analysis: a cross-sectional multicenter Italian study. J Ultrasound Med 2016;35:83-92. Tullio Ghi, Federico Prefumo , Anna Fichera, Mariano Lanna, Enrico Periti, Nicola Persico, Elsa Viora, Giuseppe Rizzo on behalf of the Società Italiana di Ecografia Ostetrica e Ginecologica (SIEOG) working group on fetal biometric charts AJOG 2016 ACCEPTED MANUSCRIPT Conclusions AC C EP TE D M AN U SC RI PT • Curves of fetal biometric measurements in twins are influenced by parental characteristics and fetal sex. • There is a reduction in growth rate during the third trimester. • The reference limits for gestation constructed in this study from uncomplicated twin pregnancies may provide an useful tool for the identification of fetal growth abnormalities in twin pregnancies. . Tullio Ghi, Federico Prefumo , Anna Fichera, Mariano Lanna, Enrico Periti, Nicola Persico, Elsa Viora, Giuseppe Rizzo on behalf of the Società Italiana di Ecografia Ostetrica e Ginecologica (SIEOG) working group on fetal biometric charts AJOG 2016 ACCEPTED MANUSCRIPT Dichorionic twins RI PT BPD (mm) HC (mm) AC (mm) FL (mm) EFW (grams) 10 30.3 108.0 85.7 16.5 108 EP TE D M AN U SC 16 0 160 60 180 0 AC C gestational age (weeks) gestational age (days) mother height (cm) mother weight (kg) father height (cm) sex (M=0;F=1) 5 28.9 103.8 81.3 15.4 99 Centiles 50 33.1 116.1 94.4 18.6 126 90 35.9 124.2 103.0 20.7 148 95 37.3 128.4 107.5 21.8 160 Monochorionic biamniotic twins gestational age (weeks) gestational age (days) mother height (cm) mother weight (kg) father height (cm) sex (M=0;F=1) ACCEPTED MANUSCRIPT BPD (mm) HC (mm) AC (mm) FL (MM) EFW (grams) 10 30.0 103.7 83.5 16.3 104 Centiles 50 32.8 112.0 93.1 18.4 123 AC C EP TE D M AN US C RI PT 16 0 160 60 180 0 5 28.5 99.4 78.5 15.2 95 90 35.6 120.2 102.8 20.5 145 95 37.0 124.5 107.7 21.6 158 ACCEPTED MANUSCRIPT Supplememtal Table 1: Effect size expressed as the sum of squares of the covariates considered for the construction of the mixed regrussion model for biparietal diameter in DC and MCDA twin pregnancies. p 0.083 0.003 0.310 0.001 0.675 0.197 0.443 MCDA Sum of Squares 30.611 53.346 13.653 556.721 4.706 2.747 3.114 D TE EP AC C p 0.128 0.093 0.412 0.0001 0.551 0.648 0.236 RI PT DC Sum of Squares 32.234 305.164 11.896 692.645 2.039 19.32 6.780 M AN US C Covariate Mother height Mother weight Father height sex Ethnic origin mother Ethnic origin father parity ACCEPTED MANUSCRIPT Supplemental Table 2: Effect size expressed as the sum of squares of the covariates considered for the construction of the mixed regression model for head circumference in DC and MCDA twin pregnancies. p 0.04 0.003 0.015 0.001 0.613 0.768 0.175 MCDA Sum of Squares 660.963 130.701 1169.978 3337,949 23.277 19.412 126.302 D TE EP AC C p 0.032 0.354 0.005 0.0001 0.695 0.823 0.362 RI PT DC Sum of Squares 1203.528 305.164 987.367 3069.603 31.51 11,17 236.08 M AN US C Covariate Mother height Mother weight Father height sex Ethnic origin mother Ethnic origin father parity ACCEPTED MANUSCRIPT Supplemetal Table 3 : Effect size expressed as the sum of squares of the covariates considered for the construction of the mixed regression model for abdominal circumference in DC and MCDA twin pregnancies. TE EP AC C MCDA Sum of Squares 974.303 423.453 166.348 780.032 485.954 187.926 405.374 p 0.029 0.085 0.283 0.027 0.067 0.316 0.094 RI PT p 0.0029 0.05 0.175 0.001 0.168 0.270 0.08 M AN US C DC Sum of Squares 1203.528 739.472 236.084 3069.603 336.030 215.612 541.957 D Covariate Mother height Mother weight Father height sex Ethnic origin mother Ethnic origin father parity ACCEPTED MANUSCRIPT Supplemental Table 4: Effect size expressed as the sum of squares of the covariates considered for the construction of the mixed regression model for femur length in DC and MCDA twin pregnancies. TE EP AC C MCDA Sum of Squares 21.345 0.629 237.948 16.577 5.589 7.346 0.323 M AN US C p 0.0001 0.810 0.0001 0.235 0.597 0.664 0.554 p 0.123 0.784 0.0001 0,213 0.423 0.396 0.847 RI PT DC Sum of Squares 49.145 0.554 140.885 13.602 3.123 2.024 3.374 D Covariate Mother height Mother weight Father height sex Ethnic origin mother Ethnic origin father parity ACCEPTED MANUSCRIPT 95 37.3 41.6 45.7 49.6 53.4 56.9 60.3 63.6 66.8 69.8 72.7 75.6 78.3 81.0 83.6 86.1 88.5 90.9 93.3 95.5 97.8 5 28.5 32.5 36.3 39.9 43.2 46.4 49.4 52.2 54.9 57.5 59.9 62.3 64.5 66.7 68.7 70.7 72.6 74.4 76.1 77.8 79.4 10 30.0 34.0 37.9 41.5 44.9 48.1 51.2 54.1 56.9 59.5 62.1 64.5 66.8 69.0 71.2 73.2 75.2 77.1 79.0 80.8 82.5 50 32.8 36.9 40.9 44.6 48.1 51.5 54.7 57.7 60.6 63.4 66.1 68.7 71.2 73.6 76.0 78.2 80.4 82.5 84.6 86.5 88.5 90 35.6 39.8 43.9 47.7 51.3 54.8 58.1 61.3 64.4 67.3 70.2 73.0 75.6 78.2 80.7 83.2 85.5 87.9 90.1 92.3 94.5 SC 90 35.9 40.1 44.2 48.0 51.7 55.2 58.5 61.7 64.8 67.7 70.6 73.3 76.0 78.5 81.0 83.5 85.8 88.1 90.4 92.5 94.7 M AN U 50 33.1 37.2 41.2 44.9 48.4 51.7 54.9 58.0 60.9 63.7 66.4 69.0 71.5 73.9 76.2 78.4 80.6 82.7 84.8 86.8 88.7 TE 10 30.3 34.4 38.1 41.7 45.1 48.3 51.4 54.3 57.0 59.7 62.2 64.6 66.9 69.2 71.3 73.4 75.4 77.3 79.2 81.0 82.7 EP 5 28.9 32.9 36.6 40.1 43.4 46.5 49.5 52.3 55.0 57.6 60.0 62.4 64.6 66.8 68.8 70.8 72.7 74.5 76.3 78.0 79.6 AC C weeks 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 Monochorionic diamniotic D Dichorionic RI PT Supplemental Table 5: Estimated 5th. 10th. 50th. 90th and 95th percentiles for biparietal diameter in dichorionic. monochorionic biamnioticand singleton pregnancies. In all the groups values were customized for the same covariates(maternal weight 60 kg. maternal height 160 cm. paternal height 180 cm. male fetal sex. and for singleton also parity 0. race european) 95 37.0 41.3 45.4 49.3 53.0 56.5 59.9 63.2 66.3 69.4 72.3 75.1 77.9 80.6 83.2 85.7 88.2 90.6 93.0 95.3 97.6 Singleton 5 29.7 33.7 37.5 41.1 44.4 47.7 50.7 53.7 56.5 59.2 61.8 64.3 66.7 69.0 71.2 73.4 75.5 77.5 79.5 81.4 83.3 10 30.1 34.2 38.0 41.7 45.1 48.4 51.5 54.5 57.4 60.1 62.8 65.3 67.8 70.1 72.4 74.6 76.8 78.8 80.8 82.8 84.7 50 34.2 38.4 42.4 46.2 49.8 53.2 56.5 59.6 62.5 65.4 68.1 70.8 73.3 75.8 78.1 80.4 82.7 84.8 86.9 88.9 90.9 90 36.2 40.5 44.5 48.3 51.9 55.4 58.6 61.8 64.8 67.7 70.4 73.1 75.7 78.1 80.5 82.8 85.1 87.2 89.3 91.4 93.4 95 40.8 45.2 49.3 53.2 56.9 60.4 63.8 67.0 70.0 73.0 75.8 78.5 81.1 83.6 86.1 88.4 90.7 92.9 95.1 97.2 99.2 ACCEPTED MANUSCRIPT RI PT Supplemental Table 6 Estimated 5th. 10th. 50th. 90th and 95th percentiles for head circumference in dichorionic. monochorionic biamniotic and singleton pregnancies. In all the groups values were customized for the same covariates (maternal weight 60 kg. maternal height 160 cm. paternal height 180 cm. male fetal sex. and for singleton also as parity 0. race european) 95 128.4 143.7 158.2 171.9 185.0 197.5 209.5 220.9 231.9 242.5 252.8 262.6 272.2 281.4 290.4 299.2 307.6 315.9 324.0 331.8 339.5 5 99.4 114.0 127.7 140.5 152.7 164.2 175.1 185.5 195.3 204.7 213.7 222.3 230.5 238.4 245.9 253.1 260.1 266.7 273.1 279.3 285.2 10 103.7 118.4 132.2 145.2 157.5 169.2 180.2 190.8 200.9 210.5 219.7 228.5 236.9 245.1 252.9 260.4 267.6 274.6 281.3 287.7 294.0 50 112.0 126.9 140.9 154.2 166.8 178.8 190.2 201.1 211.5 221.6 231.2 240.5 249.4 258.0 266.3 274.4 282.2 289.7 297.1 304.2 311.1 SC 90 124.2 139.4 153.7 167.2 180.1 192.4 204.2 215.4 226.2 236.6 246.6 256.3 265.6 274.6 283.4 291.8 300.1 308.1 315.9 323.5 330.9 90 120.2 135.4 149.6 163.2 176.1 188.4 200.1 211.4 222.2 232.6 242.7 252.4 261.8 270.9 279.8 288.4 296.8 304.9 312.8 320.6 328.2 M AN U 50 116.1 130.9 144.9 158.1 170.6 182.6 193.9 204.8 215.2 225.2 234.7 244.0 252.8 261.4 269.7 277.7 285.5 293.0 300.3 307.4 314.2 TE 10 108.0 122.5 136.1 149.0 161.1 172.7 183.7 194.1 204.1 213.7 222.9 231.7 240.1 248.2 256.0 263.6 270.9 277.9 284.7 291.2 297.6 EP 5 103.8 118.1 131.6 144.3 156.2 167.6 178.4 188.6 198.4 207.8 216.7 225.3 233.5 241.4 249.0 256.3 263.3 270.1 276.6 282.9 289.0 AC C weeks 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 Monochorionic diamniotic D Dichorionic 95 124.5 139.7 154.1 167.8 180.9 193.3 205.2 216.7 227.7 238.4 248.7 258.6 268.3 277.6 286.8 295.6 304.3 312.7 321.0 329.1 337.0 Singleton 5 110.3 124.2 137.4 149.9 161.7 173.0 183.7 194.0 203.8 213.2 222.3 231.0 239.3 247.4 255.3 262.8 270.1 277.2 284.1 290.8 297.3 10 114.3 128.2 141.4 153.9 165.7 177.0 187.7 198.0 207.8 217.2 226.3 235.0 243.3 251.4 259.3 266.8 274.1 281.2 288.1 294.8 301.3 50 122.9 137.8 151.9 165.2 177.8 189.8 201.2 212.1 222.6 232.6 242.3 251.6 260.5 269.1 277.5 285.5 293.3 300.9 308.2 315.4 322.3 90 131.9 147.3 161.7 175.4 188.4 200.8 212.6 223.8 234.6 244.9 254.9 264.4 273.6 282.5 291.1 299.4 307.4 315.2 322.8 330.1 337.3 95 139.2 154.8 169.5 183.4 196.5 209.1 221.0 232.5 243.4 253.9 263.9 273.6 283.0 292.0 300.7 309.1 317.3 325.2 332.9 340.3 347.5 ACCEPTED MANUSCRIPT 95 107.5 123.6 138.9 153.4 167.3 180.6 193.4 205.7 217.6 229.1 240.2 251.1 261.6 271.9 281.9 291.7 301.3 310.7 319.9 328.9 337.9 5 78.5 93.6 107.8 121.1 133.6 145.3 156.4 166.9 176.9 186.3 195.2 203.6 211.6 219.2 226.5 233.3 239.8 246.0 251.9 257.4 262.7 10 83.5 98.7 113.1 126.5 139.2 151.2 162.6 173.4 183.7 193.4 202.7 211.6 220.0 228.1 235.8 243.2 250.3 257.0 263.5 269.7 275.6 50 93.1 108.6 123.3 137.1 150.2 162.7 174.6 186.0 196.8 207.3 217.3 227.0 236.3 245.3 253.9 262.3 270.4 278.3 286.0 293.4 300.6 90 102.8 118.5 133.5 147.6 161.2 174.1 186.6 198.5 210.0 221.2 231.9 242.4 252.5 262.4 272.0 281.4 290.6 299.6 308.4 317.1 325.6 SC 90 103.0 119.0 134.1 148.4 162.0 175.1 187.6 199.6 211.1 222.3 233.1 243.5 253.7 263.5 273.1 282.4 291.5 300.4 309.1 317.6 326.0 M AN U 50 94.4 110.0 124.7 138.6 151.8 164.3 176.3 187.7 198.7 209.2 219.2 228.9 238.3 247.3 256.0 264.5 272.6 280.5 288.2 295.7 302.9 TE 10 85.7 101.0 115.3 128.8 141.5 153.6 165.0 175.8 186.2 196.0 205.4 214.4 222.9 231.1 239.0 246.5 253.8 260.7 267.4 273.7 279.9 EP 5 81.3 96.3 110.5 123.7 136.2 148.0 159.2 169.7 179.7 189.2 198.2 206.8 215.0 222.8 230.2 237.3 244.0 250.4 256.6 262.4 268.0 AC C weeks 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 Monochorionic diamniotic D Dichorionic RI PT Supplemental Table 7 Estimated 5th. 10th. 50th. 90th and 95th percentiles for abdominal circumference in dichorionic. monochorionic biamniotic and singleton pregnancies. In all the groups values were customized for the same covariates (maternal weight 60 kg. maternal height 160 cm. paternal height 180 cm. male fetal sex. and for singleton also as parity 0. race european) 95 107.7 123.6 138.7 153.1 166.9 180.0 192.7 205.0 216.8 228.3 239.5 250.3 260.9 271.3 281.4 291.3 301.1 310.6 320.1 329.3 338.5 Singleton 5 84.6 99.6 113.8 127.3 140.0 152.1 163.7 174.7 185.3 195.4 205.2 214.5 223.6 232.3 240.7 248.9 256.7 264.4 271.8 279.0 286.0 10 86.0 101.3 115.8 129.5 142.5 154.9 166.7 178.0 188.7 199.1 209.0 218.6 227.8 236.7 245.3 253.6 261.7 269.5 277.0 284.4 291.5 50 97.4 113.6 129.0 143.4 157.2 170.3 182.7 194.6 206.1 217.0 227.5 237.6 247.4 256.8 265.9 274.6 283.2 291.4 299.4 307.2 314.7 90 104.8 121.4 137.2 152.0 166.1 179.5 192.3 204.5 216.2 227.5 238.2 248.6 258.6 268.3 277.6 286.6 295.3 303.8 312.0 320.0 327.7 95 118.7 135.7 151.7 166.8 181.1 194.7 207.8 220.2 232.1 243.5 254.5 265.0 275.2 285.0 294.5 303.6 312.5 321.1 329.4 337.5 345.4 ACCEPTED MANUSCRIPT 95 21.9 25.6 29.0 32.3 35.5 38.5 41.5 44.3 46.9 49.6 52.1 54.6 56.8 59.1 61.3 63.3 65.4 67.4 69.2 71.2 73.1 5 10 50 15.2 18.8 22.1 25.3 28.3 31.1 33.8 36.4 38.8 41.2 43.4 45.6 47.6 49.6 51.5 53.3 55.1 56.8 58.4 60.0 61.5 16.3 19.9 23.3 26.5 29.5 32.4 35.1 37.7 40.2 42.6 44.9 47.1 49.2 51.2 53.1 55.0 56.8 58.6 60.3 61.9 63.5 18.4 22.1 25.5 28.8 31.9 34.8 37.6 40.3 42.9 45.3 47.7 50.0 52.2 54.3 56.4 58.3 60.2 62.1 63.9 65.7 67.4 M AN U 90 20.8 24.4 27.9 31.1 34.3 37.3 40.2 43.0 45.6 48.2 50.6 53.0 55.2 57.5 59.6 61.6 63.7 65.6 67.4 69.3 71.1 D 50 18.6 22.2 25.7 28.9 32.0 34.9 37.7 40.4 43.0 45.4 47.8 50.1 52.3 54.4 56.4 58.4 60.3 62.1 63.9 65.7 67.4 TE 10 16.4 20.0 23.5 26.7 29.7 32.6 35.3 37.8 40.4 42.7 45.0 47.1 49.3 51.3 53.2 55.2 56.9 58.7 60.5 62.1 63.6 EP 5 15.3 18.9 22.3 25.6 28.5 31.4 34.0 36.5 39.1 41.3 43.5 45.6 47.8 49.7 51.5 53.5 55.2 56.9 58.7 60.2 61.7 AC C weeks 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 Monochorionic diamniotic 90 SC Dichorionic RI PT Supplemental Table 8 Estimated 5th. 10th. 50th. 90th and 95th percentiles for femur length in dichorionic. monochorionic biamniotic and singleton pregnancies. In all the groups values were customized for the same covariates(maternal weight 60 kg. maternal height 160 cm. paternal height 180 cm. male fetal sex. and for singleton also as parity 0. race european) 20.5 24.2 27.7 31.1 34.2 37.3 40.1 42.9 45.6 48.1 50.6 52.9 55.2 57.4 59.6 61.6 63.6 65.6 67.5 69.4 71.2 95 21.6 25.4 28.9 32.3 35.5 38.5 41.4 44.2 46.9 49.5 52.0 54.4 56.8 59.0 61.2 63.3 65.4 67.4 69.4 71.3 73.2 Singleton 5 16.3 19.9 23.2 26.4 29.4 32.3 35.1 37.7 40.2 42.6 44.9 47.1 49.3 51.3 53.3 55.3 57.2 59.0 60.7 62.4 64.1 10 16.7 20.4 23.8 27.0 30.0 32.9 35.7 38.4 40.9 43.3 45.7 47.9 50.1 52.2 54.2 56.1 58.0 59.9 61.6 63.4 65.1 50 19.5 23.2 26.8 30.1 33.3 36.3 39.2 41.9 44.6 47.1 49.5 51.9 54.1 56.3 58.4 60.4 62.4 64.3 66.1 67.9 69.7 90 21.2 25.0 28.6 32.0 35.2 38.2 41.1 43.9 46.6 49.1 51.6 53.9 56.2 58.4 60.5 62.6 64.6 66.5 68.4 70.2 71.9 95 23.2 27.0 30.6 34.0 37.2 40.3 43.2 46.0 48.7 51.3 53.7 56.1 58.4 60.6 62.7 64.8 66.8 68.7 70.6 72.4 74.2 ACCEPTED MANUSCRIPT 90 148 194 250 316 392 479 576 685 805 936 1079 1234 1399 1576 1763 1961 2168 2386 2612 2847 3091 95 160 211 271 343 425 520 626 745 877 1021 1179 1349 1532 1728 1935 2155 2387 2630 2884 3147 3421 5 95 125 161 203 251 306 366 433 505 583 667 756 850 948 1051 1156 1266 1377 1491 1607 1725 10 104 137 176 222 275 334 401 475 555 642 736 835 942 1053 1170 1291 1417 1547 1680 1816 1955 50 123 162 208 263 326 397 477 566 664 770 886 1010 1142 1283 1431 1587 1750 1919 2095 2277 2465 90 145 191 246 310 384 469 565 671 789 919 1059 1211 1375 1550 1735 1932 2139 2356 2582 2818 3063 SC 50 126 166 214 270 334 407 489 580 680 789 906 1033 1168 1311 1463 1621 1787 1960 2139 2324 2515 M AN U 10 108 142 182 229 283 345 413 488 571 661 757 860 969 1083 1204 1329 1460 1594 1733 1875 2020 TE 5 99 130 167 211 260 316 378 446 521 601 688 780 877 979 1085 1195 1309 1427 1547 1670 1794 EP 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 AC C weeks Monochorionic diamniotic D Dichorionic RI PT Supplemental Table 9 Estimated 5th. 10th. 50th. 90th and 95th percentiles for fetal weigth (grams) in dichorionic. monochorionic biamniotic and singleton pregnancies. In all the groups values were customized for the same covariates(maternal weight 60 kg. maternal height 160 cm. paternal height 180 cm. male fetal sex. and for singleton also as parity 0. race european) 95 158 208 267 337 418 511 615 732 862 1004 1159 1328 1509 1704 1911 2131 2363 2608 2864 3131 3409 Singleton 5 107 140 179 226 279 339 407 482 565 656 754 859 971 1091 1217 1350 1490 1635 1786 1942 2103 10 110 145 187 236 292 356 427 507 595 690 794 905 1024 1151 1284 1425 1572 1725 1885 2050 2220 50 135 179 232 294 366 447 539 641 754 876 1009 1151 1303 1464 1633 1811 1997 2190 2390 2596 2808 90 155 206 267 338 421 515 620 738 866 1007 1159 1321 1495 1678 1871 2073 2283 2501 2726 2957 3195 95 188 249 321 406 504 614 738 875 1025 1188 1363 1550 1748 1957 2175 2404 2640 2884 3136 3393 3656