Second-trimester assessment of gestational age in twins: validation of singleton biometry charts




To determine the accuracy of dating twin pregnancies, of between 16 and 26 weeks' gestation, using singleton head circumference (HC) formulae.


This was a retrospective study of 269 singleton and 119 twin non-anomalous pregnancies conceived by in-vitro fertilization (IVF) with a known embryo transfer date. Fetal ultrasound biometry data for HC, obtained using different formulae, were compared with expected fetal HC size for gestation calculated from the date of conception. Similar comparisons were undertaken for femur length (FL) and for transverse cerebellar diameter.


The mean differences in HC between observed ultrasound measurements and those expected from the IVF history were small (1–4 mm) and within the measurement error for both singletons and twins for all formulae. All measurements from the larger and the smaller twins straddled those of singletons, regardless of biometry and formula used. Negative skewing of FL measurements in the smaller twin suggests that fetal growth restriction may occur at this gestation and supports the practice of dating using the HC of the larger twin.


Singleton pregnancy HC charts can be used to date reliably twin pregnancies. The data of the study also suggest that the HC of the larger twin is the most reliable measurement for use in dating. Copyright © 2010 ISUOG. Published by John Wiley & Sons, Ltd.


Gestational age assessment is fundamental to good pregnancy management, irrespective of the number of fetuses. Accurate estimation of gestational age is required for most screening tests, timing of invasive procedures and management of fetal growth restriction (FGR), preterm labor and post-term pregnancy1, 2. Conventionally, pregnancies are dated by the last menstrual period (LMP); however, in about 40% of pregnancies the LMP is either not known or the information is not reliable3, 4. Therefore, recent practice has been to assign gestational age using ultrasound fetal biometry5, 6. The National Institute of Clinical Excellence (NICE) has recently recommended that all pregnancies should be dated by fetal crown–rump length (CRL) between 11 and 14 weeks of gestation and by head circumference (HC) thereafter7. Previous studies have validated the use of standard CRL formulae for use with twin pregnancy8–10. However, there is a paucity of data on the accuracy of published second-trimester biometry formulae in twin pregnancy. Furthermore, there is a lack of consensus about whether the pregnancy should be dated on measurements taken from the larger twin, the smaller twin or on the average measurement of the twins.

The aim of this study was to determine the accuracy of second-trimester HC formulae for singletons to date twin pregnancy from the larger, the smaller or the average twin fetal size. For comparison, femur length (FL) and transverse cerebellar diameter (TCD) measurements were also assessed in these pregnancies.

Patients and Methods

This was a retrospective study of pregnancies conceived by in-vitro fertilization (IVF) or intracytoplasmic sperm injection (ICSI) that were seen in a routine obstetric setting between June 1999 and March 2010. Only non-anomalous dichorionic twin and singleton pregnancies that had undergone second-trimester ultrasound assessment between 16 and 26 weeks' gestation were included in the study. Each measurement from a fetus was considered once between 16 and 26 weeks. IVF/ICSI pregnancies were used because the date of conception was known and the gestational age at the scan could be calculated. In order to correct for any systematic variation in dating and/or early fetal growth that might occur in pregnancies achieved using assisted reproduction techniques, IVF/ICSI singleton pregnancies were used as controls.

Ultrasound examinations were carried out only by trained sonographers, and fetal biometry data for HC, TCD and FL were obtained as previously described11–13. IVF/ICSI gestational (expected) age was calculated using the embryo transfer date as a proxy for the date of conception (day 14).

Fetal biometry ultrasound data, obtained using different HC, FL and TCD formulae, were compared with expected fetal size for gestation calculated from the date of conception14–18. To determine if there was any systematic overestimation or underestimation of fetal size, the 95% CI of the mean difference was calculated for singleton vs. larger, smaller and mean twin fetal biometry measurements. To explore the biometry frequency distribution, Z-scores were calculated as the number of SD by which an observed biometry measurement deviated from the mean for gestation, with a negative value being smaller. Normality of distribution was assessed using the Kolmogorov–Smirnov test. Mean Z-scores were compared between singleton, larger and smaller twins using one-way ANOVA. The Bonferroni post-hoc test was used to explore differences between individual groups.


Three-hundred and ninety six IVF pregnancies that had undergone second-trimester ultrasound assessment between 16 and 26 weeks' gestation were seen during the study period. This group contained 269 singleton and 119 dichorionic twin pregnancies. Eight monochorionic pregnancies were excluded from the analysis as this limited number of cases precludes confidence in the statistical analysis. The mean decimal gestational age (SD) at inclusion was 21.57 (1.09) weeks for singleton pregnancies and 21.35 (1.22) weeks for twin pregnancies (P = 0.564). The proportion of ICSI pregnancies was not significantly different between the two groups (singleton = 11.76%; twins = 10.41%; P = 0.217).

The mean differences in fetal size between ultrasound measurements (observed) vs. IVF history (expected) for HC, FL and TCD are shown in Table 1 and in Figure 1. For the majority of measurements and formulae, the larger and smaller twin measurements straddled those of singletons, with the average twin size being closest to singleton measurements. The correlation between observed and expected HC measurements in the larger twin is shown in Figure 1.

Figure 1.

Scatter-plot of observed (from ultrasound) and expected (from in-vitro fertilization history) head circumference (HC) measurements of the larger twin using the Verburg formula17.

Table 1. Differences between observed (determined using ultrasound) and expected (determined from the in-vitro fertilization (IVF) history) biometry measurements that were calculated, using various formulae, for singleton, larger twin and smaller twin IVF fetuses
 Difference in biometry measurements (mm, mean (95% CI))
Biometry formulaSingleton (n = 269)Larger twin (n = 119)Smaller twin (n = 119)Average twin size (n = 119)
  1. FL, femur length; HC, head circumference; TCD, transverse cerebellar diameter.

HC formulae 
 Chitty et al.14− 1.510 (−2.516 to − 0.496)1.050 (−0.270 to 2.363)− 4.280 (−5.524 to − 3.035)− 1.620 (−2.820 to − 0.415)
 Verburg et al.170.695 (−0.322 to 1.713)3.273 (1.939 to 4.606)− 2.053 (−3.312 to − 0.794)0.609 (−0.608 to 1.828)
 Salomon et al.162.027 (1.118 to 3.047)4.301 (2.962 to 5.639)− 1.025 (−2.264 to 0.214)1.638 (0.428 to 2.848)
FL formulae 
 Chitty et al.15− 0.140 (−0.390 to 0.109)0.399 (0.094 to 0.703)− 1.266 (−1.779 to − 0.753)− 0.521 (−0.989 to 0.051)
 Verburg et al.170.213 (−0.037 to 0.465)0.771 (0.4465 to 1.077)− 0.719 (−1.113 to − 0.326)0.026 (−0.298 to 0.349)
 Salomon et al.16− 0.522 (−0.794 to − 0.294)− 0.005 (−0.311 to 0.302)− 1.501 (−1.899 to − 1.103)− 0.756 (−1.083 to − 0.429)
TCD formulae 
 Serhatlioglu et al.18− 1.636 (−1.784 to − 1.487)− 1.691 (−2.310 to − 1.071)− 2.492 (−3.096 to − 1.890)− 2.092 (−2.698 to − 1.486)
 Verburg et al.17− 0.027 (−0.158 to 0.104)0.409 (0.226 to 0.593)− 0.416 (−0.609 to − 0.222)− 0.003 (−0.174 to 0.169)

The Z-scores of HC and TCD measurements in singletons and twins were normally distributed. For FL measurements, the Z-score distribution was negatively skewed only for singletons (P = 0.003) and for the smaller twin (Figure 2; P = 0.020). ANOVA showed significant differences among HC, TCD and FL Z-score groups. The Bonferroni post-hoc test results for individual comparisons of the mean Z-scores are shown in Table 2. There were significant differences for most comparisons except singleton vs. average twin size for HC, FL and TCD.

Figure 2.

Frequency distribution of femur length Z-scores of the smaller twin, using the Verburg formula17. The distribution violates assumption of normality (Kolmogorov–Smirnov test, P = 0.020). n = 119; mean (SD): − 0.39 (1.188).

Table 2. Comparison of differences in mean biometry Z-scores using one-way ANOVA and the post-hoc Bonferroni test for multiple comparisons using the Verberg formulae17
 Larger twinSmaller twinAverage twin size
Head circumference
 Singleton0.011< 0.0051.000
 Larger twin< 0.0050.041
 Smaller twin0.041
Femur length
 Singleton0.070< 0.0051.000
 Larger twin< 0.0050.022
 Smaller twin0.022
Transverse cerebellar diameter
 Larger twin< 0.0050.025
 Smaller twin0.025


This is the largest study to evaluate the validity of second-trimester singleton biometry charts to date twin pregnancy. The findings of the study demonstrate that the variation in fetal HC size between singleton and twin pregnancies at 16 to 26 weeks' gestation is unlikely to be of clinical significance. The maximum difference in HC between the singleton and the larger twin was about 5 mm, which is equivalent to a difference of only 2 days at this gestational time-point. Additionally, the HC measurements of the smaller twin and the larger twin ‘straddled’ that of singletons, consistent with the expectation that twins are similar in size to singletons and that the difference in their size represents normal physiological variation. Only one previous study has attempted to assess second-trimester twin pregnancy dating using HC19. In that study, 67 twin and 152 singleton IVF pregnancies were examined and it was concluded that the average HC of twins was the best predictor of IVF age, without evaluating the distribution of the larger and the smaller twin sizes individually19. This group then suggested the use of a composite formula, incorporating HC, FL and abdominal circumference (AC), for dating using the average twin size. In a follow-up study, the same authors evaluated average twin CRL and their proposed formula in a subsequent cohort containing 81 twin and 104 singleton IVF pregnancies9. They subsequently reported that CRL was superior to their specific second-trimester formula, but that this difference was marginal, amounting to a gestational age difference of 1 day. Two other studies have attempted to validate second-trimester singleton biometry charts to date twin pregnancy, but both assessed only biparietal diameter and FL20, 21.

The use of IVF/ICSI, rather than spontaneously conceived pregnancies, to test the validity of second-trimester singleton biometry to date twin pregnancies could be questioned. However, only IVF and ICSI pregnancies have a known date of conception, permitting comparisons of observed and expected fetal sizes. Furthermore, singleton IVF/ICSI pregnancies were used as controls to correct for any systematic differences or errors that could result from the use of assisted reproduction. This study dated both singleton and twin pregnancies on the same basis (embryo transfer date). The gestational age differences that occur when using IVF/ICSI dates are only a few days and are equivalent to both the measurement error of ultrasound and the variation in size between the larger and the smaller twin22. It is conventional to date IVF pregnancies using the oocyte retrieval date, but this practice is not based on published evidence. As a result of the emerging complexities of IVF/ICSI with frozen embryos, early/late embryos and blastocyst transfer, previous studies have even suggested that these pregnancies should be routinely dated using the first-trimester CRL8, 23, 24. Given that the differences in fetal size are only slightly larger in the second trimester, dating could also be based on ultrasound biometry measurements in the absence of first-trimester scan findings.

The issue of whether the CRL from the larger or the smaller twin should be used to date spontaneously conceived twin pregnancies is unresolved. Multiple pregnancies are at increased risk of FGR compared with singletons, and the purpose of pregnancy dating is to allow accurate assessment of growth in subsequent scans25. Estimating the due date is a secondary goal in this instance, as most multiple pregnancies deliver before 40 weeks of gestation. The finding that the HC and TCD of the smaller twin were only 1 day behind a singleton pregnancy would initially suggest that twin pregnancies do not exhibit significant FGR at 18–26 weeks' gestation. This assertion is also supported by the finding that the Z-scores for HC and TCD measurements were normally distributed without evidence of a negative skew. Furthermore, this insignificant difference indicates that the HC and TCD size differences in the smaller twin were caused by physiological variation, rather than by pathology of placental insufficiency. However, the FL measurements of the smaller twin were negatively skewed, introducing the possibility of FGR as a cause. Indeed, the observation of second-trimester short FL as an early feature of FGR from placental insufficiency has been made previously in non-IVF pregnancies26, 27. Given the possibility of growth restriction in smaller twins at this gestation, it would appear to be unwise to date twin pregnancy measurements on the smaller or the average twin size. It is relatively infrequent for a twin to be pathologically large, and a policy of dating by the larger twin would seem to be the most effective.