A noninvasive approach by serial ultrasound examination at 12–15, 18, and 30 weeks of gestation can be used to exclude homozygous α 0 -thalassemia-induced fetal anemia. At 12–15 weeks of gestation, the predictive values for the fetal cardio-thoracic ratio were better than that for the placental thickness. At 16–20 weeks of gestation, measuring middle cerebral artery peak systolic velocity is associated with a low false-positive rate. However, the false-positive rate of this noninvasive approach can be about 3%, requiring an invasive test to confirm the diagnosis. A false-negative may result in a delay in diagnosis. The success of this noninvasive approach depends on an accurate measurement of the fetal cardiothoracic ratio which can be improved by adequate training and subsequent quality control. Currently, there is a lack of data reporting the performance of a noninvasive approach before 12 weeks of gestation.
Highlights
- •
A noninvasive approach by ultrasonography can be used to exclude homozygous α 0 -thalassemia-induced fetal anemia.
- •
In the first trimester, the predictive values for the fetal cardio-thoracic ratio are better than that for the placental thickness.
- •
In the second trimester, the use of middle cerebral artery peak systolic velocity is associated with a low false positive rate.
- •
This non-invasive approach is associated with a small false positive rate, requiring an invasive test to confirm the diagnosis.
Introduction
α 0 -thalassemia couples are at risk of having a fetus with homozygous α 0 -thalassemia or Hb-Bart’s disease. A prenatal diagnosis is necessary because it is associated with poor fetal outcomes and major maternal complications, including preeclampsia, postpartum hemorrhage or even death. Conventionally, a prenatal diagnosis is achieved by a DNA study following an invasive procedure which is either chorionic villus sampling (CVS) or amniocentesis, but both procedures are associated with a small miscarriage rate (<0.5%). Alternatively, a noninvasive approach using serial ultrasound (US) examinations for features of fetal anemia can effectively differentiate anemic and unaffected fetuses. While the majority of Hb-Bart fetuses will develop some signs of anemia during the first trimester which will then lead to an early confirmation of the diagnosis with CVS, fetuses without any sign of anemia before 24 weeks of gestation are unlikely to be affected, and hence an invasive test can be avoided (Lam et al., 1997a) (Leung et al., 2004 and 2006). In this chapter, the principle of US prediction, the accuracy of individual US markers and the strategy of the monitoring will be reviewed.
Principle of ultrasound prediction
In normal fetuses, the expression of ζ-globin is limited to the yolk-sac of primitive erythroblasts before eight weeks of gestation, and then switched to α-globin expression (Liebhaber et al., 1996). In fetuses affected by homozygous α 0 -thalassemia in which all the four α-globin genes are absent, hemoglobin (Hb) F (α2γ2) cannot be produced, resulting in anemia from this gestation period onwards. The majority of the Hb in these affected fetuses is Hb-Bart’s (γ 4 ) that does not release oxygen as effectively as the Hb F (Bernini LF et al., 1998). Approximately 10–20% of the Hb is embryonic Hb Portland 1 ( ζ 2 γ 2 ) as a result of the continuous expression of the embryonic ζ-globin genes (Kutlar et al., 1989).
Because of the presence of anemia and hypoxia in an affected fetus, cardiac output increases resulting in cardiomegaly. Placental hyperplasia/extramedullary hematopoiesis occurs resulting in placentomegaly. When cardiac failure subsequently develops, cardiomegaly further increases, placenta edema and later hydrops appears. It is well known that middle cerebral artery peak systolic velocity (MCA-PSV) is increased in fetal anemia because of an increased cardiac output and a decline in blood viscosity (Moise et al., 1990) and fetal hypoxia. These ultrasonographic features of fetal anemia, including placentomegaly, fetal cardiomegaly and increased MCA-PSV form the basis for prenatal exclusion of a fetus affected by homozygous α 0 -thalassemia.
At-risk groups
α 0 -thalassemia couples, both being carriers, have a 25% risk of having a fetus affected by homozygous α 0 -thalassemia because it is an autosomal recessive disorder. α 0 -thalassemia is common in Southeast Asia and recorded in the Mediterranean region as well as other parts of the world because of population migration. Universal prenatal screening for thalassemia using maternal mean corpuscular volume (MCV) or mean corpuscular hemoglobin (MCH) has been advocated for at-risk populations because of their ethnic origin. The details are described in another chapter.
The noninvasive approach using ultrasonography is applicable in single fetus as well as twin pregnancies (Leung et al., 2005). In pregnancies conceived after the preimplantation genetic diagnosis for α-thalassemia, this approach is particularly useful to confirm normality without an invasive procedure (Chan et al., 2006).
Gestation
High predictability of this noninvasive approach has been reported from 12 weeks of gestation onward (Leung et al., 2006). First trimester US exclusion is preferred to that of second trimester because of an earlier relief of maternal anxiety in an unaffected pregnancy and avoiding the disadvantages of termination of an affected pregnancy at a later gestation.
First trimester ultrasound exclusion
Placental thickness
Historically, placental thickness (PT) was the first ultrasonographic parameter studied for the prediction of an affected pregnancy (Ghosh et al., 1994). At that time, an US examination of the fetal heart in the first trimester was difficult using the old US technology.
When the thickest part of the placenta is found subjectively after scanning up and down, the transducer is placed perpendicular to the placenta so that an echogenic line can be visualized on its surface. Measurements of the PT are then taken in the longitudinal and transverse sections ( Figure 1 a and b) (Ghosh et al., 1994), and the maximum PT is used.
In a study on 231 at-risk pregnancies (Ghosh et al., 1994) of which 60 (26.0%) were affected, the sensitivity in detecting affected pregnancies at cutoff of mean PT plus 2 SD was higher after than that before 12 weeks of gestation (0.95 vs 0.72) while the false-positive rate remained low around 3%.
In another study on 832 at-risk pregnancies of which 168 (20.2%) were affected (Leung et al., 2006), the sensitivity for PT in the prediction of affected pregnancies at a cutoff of >18mm at 12–15 weeks of gestation was 77.1% only while the false-positive rate was as high as 19.0% . The results of this study appeared to be more realistic than the previous one in view of a larger sample size, more US operators and the limitations of prediction using PT, as detailed below. The mean (SD) of PT of affected and unaffected pregnancies in various gestation periods was given in Table 1 (unpublished data generated from the same study).
| Ultrasonographic measurements | Unaffected Mean (SD) | Affected Mean (SD) | Difference Mean (95% CI) |
|---|---|---|---|
| CTR | |||
| 12–13 weeks | 0.45 (0.04) n = 215 | 0.54 (0.03) n = 57 | 0.08–0.10 |
| 14–15 weeks | 0.44 (0.05) n = 91 | 0.56 (0.04) n = 24 | 0.10–0.14 |
| 16–17 weeks | 0.46 (0.03) n = 178 | 0.56 (0.04) n = 23 | 0.09–0.12 |
| 18–19 weeks | 0.47 (0.04) n = 127 | 0.60 (0.04) n = 22 | 0.12–0.15 |
| 20–21 weeks | 0.47 (0.03) n = 97 | 0.57 (0.05) n = 19 | 0.08–0.13 |
| 22–23 weeks | 0.49 (0.05) n = 72 | 0.62 (0.09) n = 19 | 0.09–0.18 |
| 24–25 weeks | 0.49 (0.04) n = 58 | 0.64 (0.10) n = 15 | 0.10–0.21 |
| PT (mm) | |||
| 12–13 weeks | 15.7 (3.0) n = 213 | 20.5 (4.7) n = 57 | 3.5–6.1 |
| 14–15 weeks | 17.7 (3.3) n = 88 | 20.3 (4.5) n = 23 | 0.6–4.7 |
| 16–17 weeks | 20.9 (3.8) n = 177 | 23.6 (5.3) n = 21 | 0.2–5.2 |
| 18–19 weeks | 22.4 (5.1) n = 124 | 26.5 (5.1) n = 22 | 1.7–6.4 |
| 20–21 weeks | 23.6 (4.1) n = 94 | 27.2 (3.7) n = 18 | 1.5–5.7 |
| 22–23 weeks | 26.3 (3.1) n = 72 | 35.7 (10.1) n = 19 | 4.5–14.3 |
| 24–25 weeks | 32.0 (5.9) n = 57 | 42.5 (12.4) n = 15 | 3.5–17.5 |
When the placenta is adjacent to a focal myometrial contraction or located in the fundus or lateral uterine wall, difficulty in defining the placental base or getting a true perpendicular plane to the placenta may be encountered, resulting in an over- or underestimation of the PT (Ghosh et al 1994). Besides, a false negative result can occur in an affected pregnancy when a placenta is large but not thick (Ghosh et al 1994).
Fetal cardiothoracic ratio
In view of the limitations of an accurate measurement of PT and its prediction, the fetal cardiothoracic ratio (CTR) was the second ultrasonographic parameter studied (Lam et al., 1997a). When a subcostal four-chamber heart view is obtained, the fetal transverse cardiac diameter (TCD) at the level of the atrio-ventricular valves between the epicardial surfaces at diastole and the transverse fetal thoracic diameter (TD) are measured ( Figure 2 a and b) (Lam et al., 1997a). The fetal CTR equals TCD/TD. At 12-14 weeks of gestation, an optimal view of the fetal CTR is usually obtained through an abdominal scan using a 5 or 7 mHz curvilinear transducer (Lam et al., 1997a), or through vaginal ultrasonography using a 5 or 7 mHz vector transducer (Lam et al., 1997a).
In an early study on 62 pregnancies at 13–14 weeks (Lam et al., 1997a), using a CTR cutoff level of > or = 0.5, only 75% of affected pregnancies were detected at 13–14 weeks at a false-positive rate of 8%. This suboptimal sensitivity was attributed to the suboptimal imaging of the CTR using transabdominal sonography in some of the cases (Lam et al., 1997a). In another study by the same group on 135 at-risk pregnancies of which 43 (31.9%) were affected, transvaginal sonography was added when an accurate imaging of CTR was not possible using transabdominal sonography (Lam et al., 1997a). The same CTR cutoff level of > or = 0.5 was 100% sensitive with 0 false-positive rate (Lam et al., 1999a).
In a study on 832 at-risk pregnancies of which 168 (20.2%) were affected (Leung et al., 2006), the sensitivity for the fetal CTR at 12–15 weeks of gestation was higher than that for the placental thickness (97.5% vs 77.1%) at a lower false-positive rate (9.1% vs 19.0%) (Leung et al., 2006). The sensitivity of 97.5% in this study appeared to be more realistic than 100% in the above study (Lam et al., 1999a) because of a larger sample size and more US operators. The mean (SD) of CTR of affected and unaffected fetuses in various gestation periods is given in Table 1 (unpublished data generated from the same study).
In another study on 154 at-risk pregnancies, unsatisfactory imaging was encountered in five cases, four of whom were at 11 weeks’ gestation (Zhen et al., 2015). Of the remaining 149 cases with optimal imaging, the sensitivity and specificity for the fetal CTR were 97.1% and 100%, respectively (Zhen et al., 2015). These results highlight the importance of optimal imaging and an accurate measurement of CTR.
Nuchal translucency, MCA-PSV and other ultrasound parameters
In a study on 94 at-risk fetuses of which 32 (34.0%) were affected (Lam et al., 1999c), the median nuchal translucency (NT) multiple of median (MoM) of these affected fetuses ( Figure 3 ) was significantly higher than that of the normal controls by 19% or 0.3–0.4mm (p < 0.001), but with an extensive overlap of NT between cases and controls. It seems that NT is not an accurate US marker for the prediction of homozygous α 0 -thalassemia. When an increased NT is found in an at-risk fetus, the possibility of chromosomal abnormality as well as homozygous α 0 -thalassemia should be raised (Lam et al., 1999c).
In a study on 80 at-risk pregnancies of which 19 (23.8%) were affected (Lam and Tang, 2002), the median (interquartile range) MCA-PSV at 12–13 weeks of gestation was significantly higher in the affected than the unaffected pregnancies [19 (7) vs 14 (6) cm/s; p=0.001], but there was a high overlap in the measurements between the affected and unaffected pregnancies. Thus, it seems that MCA-PSV is not an accurate US marker for homozygous α 0 -thalassemia. Nevertheless, Leung et al (2005) reported that MCA-PSV was increased in an affected fetus of a dichorionic twin while normal in an unaffected one at 14 weeks of gestation. According to the experience of the authors, it is more difficult to accurately measure MCA-PSV in the first than the second trimester because of more fetal movement and smaller size of MCA resulting in an inaccurate placement of the Doppler gate along the direction of MCA.
Other abnormal US features, including limb reduction defects related to early ischemic damage (Lam and Tang, 2000), echogenic bowel (Lam et al., 1999b), increased cardiac flow (Lam et al., 1999d), or abnormal ductus venosus Doppler (Lam et al., 2001) are not sensitive enough for the prediction of affected fetuses.
First trimester ultrasound exclusion
Placental thickness
Historically, placental thickness (PT) was the first ultrasonographic parameter studied for the prediction of an affected pregnancy (Ghosh et al., 1994). At that time, an US examination of the fetal heart in the first trimester was difficult using the old US technology.
When the thickest part of the placenta is found subjectively after scanning up and down, the transducer is placed perpendicular to the placenta so that an echogenic line can be visualized on its surface. Measurements of the PT are then taken in the longitudinal and transverse sections ( Figure 1 a and b) (Ghosh et al., 1994), and the maximum PT is used.
In a study on 231 at-risk pregnancies (Ghosh et al., 1994) of which 60 (26.0%) were affected, the sensitivity in detecting affected pregnancies at cutoff of mean PT plus 2 SD was higher after than that before 12 weeks of gestation (0.95 vs 0.72) while the false-positive rate remained low around 3%.
In another study on 832 at-risk pregnancies of which 168 (20.2%) were affected (Leung et al., 2006), the sensitivity for PT in the prediction of affected pregnancies at a cutoff of >18mm at 12–15 weeks of gestation was 77.1% only while the false-positive rate was as high as 19.0% . The results of this study appeared to be more realistic than the previous one in view of a larger sample size, more US operators and the limitations of prediction using PT, as detailed below. The mean (SD) of PT of affected and unaffected pregnancies in various gestation periods was given in Table 1 (unpublished data generated from the same study).
