To determine possible differences in hepatic artery flow between trisomy 21 and euploid fetuses at 11–13 weeks' gestation.
To determine possible differences in hepatic artery flow between trisomy 21 and euploid fetuses at 11–13 weeks' gestation.
Hepatic artery pulsatility index (PI) and peak systolic velocity (PSV) were measured in fetuses at low risk of aneuploidies (n = 350) and another group at high risk, including 283 euploid and 47 with trisomy 21. The association of hepatic artery PI and PSV with trisomy 21, fetal nuchal translucency (NT) thickness, tricuspid regurgitation, and reversed a-wave in the ductus venosus was investigated.
In the low-risk group, the median hepatic artery PSV was 10.0 cm/s and the 95th centile was 14.3 cm/s. The distribution of hepatic artery PI was skewed, but for PI of 2 or more the distribution was Gaussian. In 325 (92.9%) cases, the PI was 2 or more (high PI) and in 25 (7.1%) it was below 2 (low PI). In 33 (70.2%) of the trisomy 21 pregnancies, the PSV was above the 95th centile and the PI was below 2. Multiple regression analysis showed that in the prediction of hepatic artery PSV there were significant contributions from fetal karyotype, tricuspid regurgitation, and reversed a-wave in the ductus venosus, but not delta NT, pregnancy-associated plasma protein-A, or free β-human chorionic gonadotrophin.
Trisomy 21 at 11–13 weeks is associated with increased flow in the hepatic artery. Copyright © 2011 John Wiley & Sons, Ltd.
In fetal life, the liver is a vital organ with both metabolic and hematopoietic activities. Normally, > 90% of the blood supply to the liver is from the umbilical and portal veins and < 10% comes directly from the hepatic artery which is a branch of the celiac trunk from the descending aorta. In fetal hypoxemia, there is an increase in the fraction of umbilical venous blood shunted to the heart through the ductus venosus, with consequent decrease in the blood supply to the liver (Kilavuz and Vetter, 1999; Dubiel et al., 2003). Reduced perfusion of the liver leads to local accumulation of adenosine which in turn acts directly on the hepatic artery to cause vasodilatation and consequent compensatory increase in the blood supply to the liver (Ebbing et al., 2008). This hepatic arterial buffer response, which aims to maintain liver perfusion under adverse consequences, shows the importance of the liver in fetal survival.
Recent evidence suggests that the hepatic arterial buffer response may be apparent from the first trimester of pregnancy (Bilardo et al., 2010). The hepatic artery pulsatility index (PI) was measured at 11–13 weeks in 59 fetuses at high risk for aneuploidies and found to be significantly lower in those with aneuploidies and adverse outcome than in the group with normal outcome.
The aims of this study are to first establish a normal range of hepatic artery PI and peak systolic velocity (PSV) at 11–13 weeks' gestation, second, determine possible significant differences in PI and PSV between trisomy 21 and euploid fetuses, and third, examine the relation of PI and PSV with fetal nuchal translucency (NT) thickness and Doppler flow pattern in the ductus venosus and across the tricuspid valve.
This was a prospective study in two groups of singleton pregnancies undergoing screening for aneuploidies between week 11 and 13 weeks and 6 days by a combination of maternal age, maternal serum free β-human chorionic gonadotrophin (β-hCG) and pregnancy-associated plasma protein-A (PAPP-A), ultrasonographic measurement of fetal NT thickness, and Doppler assessment of the flow pattern in the ductus venosus and across the tricuspid valve (Snijders et al., 1998; Kagan et al., 2008; Kagan et al., 2009; Maiz et al., 2009).
The entry criteria for the first group of 350 fetuses were NT thickness < 3 mm, no tricuspid regurgitation or reversed a-wave in the ductus venosus, estimated risk for trisomy 21 < 1 in 300 and no obvious fetal defects at the 11–13 weeks and 20–22 weeks scan (low-risk group). The entry criteria for the second group of 330 fetuses were: increased risk for aneuploidies and fetal karyotyping which was either normal (n = 283) or trisomy 21 (n = 47). In these cases, the ultrasound and Doppler investigations were carried out within 1 h before chorionic villus sampling (CVS) for karyotyping (pre-CVS group).
In both groups of fetuses, the hepatic artery was assessed by transabdominal ultrasound using the following criteria: (1) the examinations were undertaken during fetal quiescence, (2) the magnification of the image was such that the fetal thorax and abdomen occupied the whole screen, (3) a right ventral mid-sagittal view of the fetal trunk was obtained and color flow mapping was used to demonstrate the umbilical vein, ductus venosus, descending aorta, and hepatic artery (Figure 1), (4) the pulsed Doppler sample was first set at 2.0 mm and placed so that it included both the ductus venosus and adjacent upper part of the hepatic artery (to ensure that this vessel rather than the celiac trunk is sampled) and it was then reduced to 1.0 mm to include only the hepatic artery, (5) the insonation angle to the hepatic artery was < 30°, (6) the filter was set at a high frequency (120 Hz) to avoid contamination from adjacent veins, (7) the sweep speed was high (2–3 cm/s) so that the waveforms were widely spread, and (8) the pulsed wave pulse repetition frequency was set low (2.2–3.3 Hz) allowing better assessment of the PSV. When three similar consecutive waveforms were obtained, the PSV and PI were measured by the software of the machine after manual tracing.
The agreement and bias for the measurements of hepatic artery PSV and PI by a single examiner and between two different examiners were investigated from the study of 50 cases which were selected at random from the study population. One operator (A) who made the original measurements repeated each measurement and a second operator (B) made the measurements once. The operators were not aware of the measurements of each other and operator A when making the measurements on the second occasion was not aware of measurements on the first occasion.
Continuous and categorical variables were compared using the Mann–Whitney U-test with post hoc Bonferroni correction and χ2-test or Fisher's exact test, respectively. Proportions were compared by the χ2-test. Normality of distributions was tested by the probability plots and Kolmogorov–Smirnov test.
Hepatic artery PSV required square root (sqrt) transformation to approximate a Gaussian distribution. In the low-risk group, multiple regression analysis was used to determine the factors amongst maternal characteristics and gestation that provided significant contribution in predicting sqrt hepatic artery PSV. The only factor providing a significant contribution was fetal crown-rump length (CRL) (see Section on Results) and therefore the hepatic artery PSV in each patient in the pre-CVS and the trisomy 21 group was expressed as a difference from the expected normal mean for the fetal CRL (delta value) in the low-risk group. Multiple regression analysis was used to determine the significant contributors to delta hepatic artery PSV among the continuous variables delta NT, PAPP-A multiple of the median (MoM) and free β-hCG MoM, and categorical variables fetal karyotype (euploid or trisomy 21), tricuspid flow (regurgitation or normal), and ductus venosus flow (a-wave reversed or normal).
In the low-risk group, the distribution of hepatic artery PI was skewed, but for PI of 2 or more the distribution was Gaussian with a median of 2.61 (Figure 2). In 325 (92.9%) cases, the PI was 2 or more (high PI) and in 25 (7.1%) it was below 2 (low PI). In the pre-CVS group, logistic regression analysis was used to determine whether fetal karyotype, delta NT, log10 PAPP-A MoM, log10 free β-hCG MoM, tricuspid regurgitation, and reversed a-wave in the ductus venosus had a significant contribution in predicting whether the patient belonged to the low-PI group. Non-parametric analysis was used to examine the significance of the relationship between hepatic artery PSV and PI in the outcome groups.
The Bland–Altman analysis was used to compare the degree of agreement and bias between measurements by a single examiner and two different examiners for hepatic artery PSV and PI (Bland and Altman, 1986). Paired t-test was used to compare the significance of difference between these paired measurements.
The statistical software package SPSS 15.0 (SPSS, Chicago, IL, USA) and XLSTAT-Pro 2010 (Addinsoft, New York, NY, USA) were used for data analyses.
The maternal and fetal characteristics of the study groups are compared in Table 1.
|Low-risk group||Pre-chorionic villus sampling groups|
|Characteristics||(n = 350)||Euploid (n = 283)||Trisomy 21 (n = 47)|
|Maternal age in years, median (IQR)||33.0 (28.8–36.0)||36.0 (32.0–39.0)*||37.8 (36.0–40.0)*|
|Body mass index in kg/m2, median (IQR)||23.3 (21.3–25.8)||23.8 (21.9–26.4)||23.7 (21.1–27.3)|
|Crown-rump length in mm, median (IQR)||62.5 (57.1–68.0)||66.5 (60.0–74.0)*||67.8 (59.3–75.7)*|
|Caucasian, n (%)||298 (85.1)||225 (79.5)||40 (85.1)|
|African, n (%)||26 (7.4)||23 (8.1)||3 (6.4)|
|South Asian, n (%)||16 (4.6)||16 (5.7)||2 (4.3)|
|East Asian, n (%)||6 (1.7)||11 (3.9)||2 (4.3)|
|Mixed, n (%)||4 (1.1)||8 (2.8)||0|
|Cigarette smoker, n (%)||19 (5.4)||24 (8.5)||3 (6.4)|
|Spontaneous, n (%)||327 (93.4)||264 (93.3)||43 (91.5)|
|Ovulation drugs, n (%)||23 (6.6)||19 (6.7)||4 (8.5)|
In the low-risk group, regression analysis showed that in the prediction of sqrt hepatic artery PSV there was a significant contribution from fetal CRL but not from maternal body mass index (BMI) (p = 0.363), racial origin (p = 0.285), mode of conception (p = 0.475), or smoking status (p = 0.873).
Expected sqrt hepatic artery PSV = 2.878 + 0.005 × fetal CRL in mm; R2 = 0.011, p = 0.029
In the pre-CVS group, in both the trisomy 21 and euploid fetuses, compared to the low-risk group, the delta hepatic artery PSV, delta NT, and serum free β-hCG were higher and maternal serum PAPP-A was lower (Table 2, Figure 3). In 37 (78.7%) of the 47 fetuses with trisomy 21, the delta hepatic artery PSV was above the 95th centile of the low-risk group.
|Characteristics||Low-risk group (n = 350)||Pre-CVS euploid (n = 283)||Pre-CVS trisomy 21 (n = 47)|
|Delta hepatic artery PSV, median (IQR)||− 0.15 (−1.50 to 1.59)||0.26 (−1.17 to 2.27)*||11.7 (8.0 to 16.0)*|
|Low hepatic arteryPI, n (%)||25 (7.1)||49 (17.3)*||35 (74.5)*|
|Fetal delta NT, median (IQR)||0.23 (0.03 to 0.48)||0.44 (0.09 to 0.96)*||2.23 (1.31 to 4.67)*|
|PAPP-A MoM, median (IQR)||1.03 (0.73 to 1.48)||0.78 (0.46 to 1.44)*||0.70 (0.35 to 1.15)*|
|Free β-hCG, median (IQR)||1.00 (0.73 to 1.41)||1.24 (0.71 to 1.96)*||1.96 (1.30 to 3.35)*|
|Tricuspid regurgitation, n (%)||2 (0.6)||33 (11.7)*||32 (68.1)*|
|Reversed a-wave in ductus venosus, n (%)||1 (0.3)||36 (12.7)*||22 (46.8)*|
Multiple regression analysis in the total population showed that in the prediction of delta hepatic artery PSV there were significant contributions from fetal karyotype (p < 0.0001), tricuspid regurgitation (p = 0.001), and reversed a-wave in the ductus venosus (p = 0.002) but not delta NT (p = 0.229), log10 free β-hCg MoM (p = 0.188), or log10 PAPP-A MoM (p = 0.779).
In the low-risk group, the distribution of hepatic artery PI was skewed, but for PI of 2 or more the distribution was Gaussian (Figure 2). The median PI in the high-PI and low-PI groups did not change significantly with fetal CRL (p = 0.099, p = 0.921). The median, 5th and 95th centiles in the high-PI group were 2.61, 2.11, and 3.09, respectively.
In trisomy 21 fetuses, the proportion of cases with low PI was higher than in the low-risk group (74.5 vs 7.1%, p < 0.0001). Similarly, the proportion of euploid fetuses with low PI was increased (17.3 vs 7.1%, p < 0.0001).
In the pre-CVS group, logistic regression analysis showed significant contribution for belonging in the low-PI group provided by fetal karyotype (p < 0.0001), but not delta NT (p = 0.917), tricuspid regurgitation (p = 0.236), reversed a-wave in ductus venosus (p = 0.521), log10 PAPP-A MoM (p = 0.194), or log10 free β-hCG MoM (p = 0.240).
In all groups, there was a significant inverse association between hepatic artery PSV and PI (low-risk: Spearman's rho ρ = − 0.292, p < 0.0001; euploid: ρ = − 0.490, p < 0.0001; trisomy 21: r = − 0.586, p < 0.0001; Figure 4). In 33 (70.2%) of the 47 fetuses with trisomy 21 there was low PI and high PSV.
The bias (mean difference) and 95% limits of agreement between paired measurements of hepatic artery PSV and PI by the same examiner and by two different examiners are shown in Table 3. Paired t-test to assess the repeatability of measurements showed that there was no significant difference in the measurement of either hepatic artery PSV or PI by the same examiner (p = 0.133 and p = 0.731) or by two different examiners (p = 0.828 and p = 0.269). The Pearson correlation coefficient between the difference and the mean of measurements by the same examiner for hepatic artery PSV was − 0.144 [95% confidence interval (CI): − 0.406 to 0.140] and for hepatic artery PI was 0.083 (95% CI: − 0.200 to 0.353). The respective values for two different examiners were − 0.503 (95% CI: − 0.685 to − 0.261) and 0.213 (95% CI: − 0.070 to 0.464).
|Measurement||Mean difference (95% LOA) (95% CI)|
|Hepatic artery PSV|
|Intra-observer variability||− 0.11[−1.09 (−1.23 to − 0.95), 0.87 (0.73 to 1.01)]|
|Inter-observer variability||0.02 [−0.99 (−1.14 to − 0.84), 1.02 (0.87 to 1.17)]|
|Hepatic artery PI|
|Intra-observer variability||− 0.01[−0.28 (−0.32 to − 0.24), 0.27 (0.23 to 0.31)]|
|Inter-observer variability||0.02 [−0.26 (−0.30 to − 0.22), 0.31 (0.27 to 0.35)]|
The findings of this study demonstrate that first, in trisomy 21 fetuses at 11–13 weeks the fetal hepatic artery PSV is increased and PI is decreased and second, in both euploid and trisomic fetuses high hepatic artery PSV is associated with tricuspid regurgitation and reversed a-wave in the ductus venosus.
In our low-risk group, the median hepatic artery PSV at 11–13 weeks was 10.0 cm/s and the PI was 2 or more in about 90% of cases. One previous study investigating normal fetuses reported that the hepatic artery PSV increased with gestation from a mean of about 18 cm/s at 21 weeks to 33 cm/s at 38 weeks (Ebbing et al., 2008). Such an increase in velocity between the first and second trimesters is compatible with the results of a study investigating vascular development during liver organogenesis which reported that between 10 and 25 weeks of gestation, there is progressive development of arteries and intra-portal capillaries (Gouysse et al., 2002). As for hepatic artery PI, one previous study reported that this increased with gestation to reach a maximum at around 33 weeks decreasing thereafter (Ebbing et al., 2008), whereas another study of normal fetuses at 27–41 weeks reported that the PI decreased with gestation to reach a minimum at 33 weeks and increased thereafter (Dubiel et al., 2003).
In about 80% of fetuses with trisomy 21, PSV was high and PI was low suggesting the presence of increased number of arterioles or vasodilatation in the branches of this vessel with concomitant increase in flow. The observed increased flow in the hepatic artery may either be the consequence or the cause of increased hepatic hematopoietic activity. Animal studies have showed that the fetal hepatic hematopoiesis is modulated by arterial blood flow to the liver (Kunisaki et al., 2006). Trisomy 21 is associated with a uniquely high frequency of acute megakaryoblastic leukemia in early childhood and this may be the consequence of disturbed liver hematopoiesis manifested in striking expansion in megakaryocyte–erythroid progenitors during fetal life (Tunstall-Pedoe et al., 2008).
An alternative mechanism for increased hepatic artery flow in trisomy 21 is the hepatic arterial buffer response to hypoxemia. This is supported by the finding of an association between increased hepatic artery PSV and reduced PI with reversed a-wave in the ductus venosus in both euploid and trisomic fetuses. A previous first-trimester study has also documented an inverse association in impedance to flow in the hepatic artery and ductus venosus (Bilardo et al., 2010). It is possible to speculate that a common underlying mechanism for the association between increased hepatic artery flow and tricuspid regurgitation is cardiac dysfunction resulting in abnormal flow in the ductus venosus and secondary dilatation in the vascular tree supplied by the hepatic artery.
Irrespective of the underlying mechanism of the Doppler findings in the hepatic artery of fetuses with trisomy 21, assessment of flow in this vessel may be useful in first-trimester screening for aneuploidies. Assessment of the hepatic artery is easy to perform for sonographers with extensive experience in the 11–13 weeks' scan and Doppler examination of the ductus venosus. However, successful assessment necessitates adherence to a standard technique. As in the case of Doppler assessment of flow in the ductus venosus and across the tricuspid valve examination of the hepatic artery could potentially be undertaken in all cases or this may be reserved for those with an intermediate risk after first-line combined screening by maternal age, fetal NT, and serum free β-hCG and PAPP-A (Nicolaides et al., 2005; Nicolaides, 2011). The performance of hepatic artery Doppler in screening for trisomy 21 will ultimately be determined by large prospective studies.
This study was supported by a grant from the Fetal Medicine Foundation (Charity No: 1037116).