Newborn screening for α-thalassaemia by a capillary electrophoresis method

Bouva et al. ¹ found that the percentage haemoglobin Bart's (HbBart's), as depicted by the FAST peak, was only a relative indication for the number of α genes affected in α-thalassaemia. No significant difference was demonstrated between HPLC in −α/αα and −α/−α, between −α/−α and ––/αα or between ––/αα and ––/−α genotypes. It seemed that HPLC was technically not a preferable approach in neonatal screening for α-thalassaemia.

Until recently, newborn screening for α-thalassaemia was conducted at our hospital, using a capillary electrophoresis method for quantification of HbBart's in cord blood, which exactly distinguished HbH disease from other milder α-thalassaemia forms. Over 14 months, 6525 post-delivery cord blood samples were collected. Two milliliters of cord blood with ethylenediaminetetraacetic acid as an anticoagulant was obtained. Haemoglobin analysis was performed within 24 hours using an automated capillary electrophoresis system (CapillaryS 2; Sebia, Paris, France; software version 6.2), in which charged molecules are separated at alkaline pH by their electrophoreticmobility, electrolyte pH and electroosmotic flow. Each peak of the Hbs appears in a specific zone. Normal newborn Hb consists of approximately 70–80% Hb F (α₂γ₂) and 20–30% Hb A (α₂β₂). The absence or reduced synthesis of the α-globin chain can lead to the production of abnormal HbBart's in the fetus. In our experience with CapillaryS 2, HbBart's levels as low as 0.1% could be detected. For samples with HbBart's present, genomic DNA was extracted and gap-polymerase chain reaction was used to identify the three common deletional defects (–SEA, −α3.7 and −α4.2). If no deletional mutations were found, reverse dot-blot was used to simultaneously detect two common mutations (Hb Constant Spring and Hb Quong Sze). Otherwise, direct DNA sequencing was used to identify rare point mutations of α-globin gene. The automated capillary electrophoresis detected various amounts of HbBart's in 377 (5.8%) samples. The results of genotyping are summarized in Table 1.

Table 1

HbBart's percentages in cord blood with different genotypes of α-globin gene

Genotype	Number of samples (%)	%HbBart's (range)
Single α-globin gene defect
−α3.7/αα	62 (16.4)	0.39±0.21 (0.1–0.9)
−α4.2/αα	20 (5.3)	0.50±0.16 (0.2–0.9)
αCSα/αα	12 (3.2)	1.59±0.32 (1.0–2.3)
αQSα/αα	10 (2.7)	0.48±0.18 (0.2–0.7)
αTα/αα (ααT /αα)*	5 (1.3)	0.58±0.29 (0.3–0.9)
Two α-globin gene defect
−−SEA/αα	256 (67.9)	3.55±1.04 (1.0–8.6)
−−THAI/αα	2 (0.5)	4.7; 4.6
−α3.7/−α4.2	1 (0.3)	3.7
Hb H disease
−−SEA/−α3.7	6 (1.6)	20.21±2.22 (17.5–22.5)
−−SEA/−α4.2	1 (0.3)	20.6
Normal α-globin gene †
αα/αα	2 (0.5)	0.3; 0.6
Total	377 (100)

*One case of α2CD8 (−C), one case of α2CD31(AGG→AAG), one case of α1CD62 (GTG→GCG) and two cases of α1 CDs117/118(+ATC)

^†Not screened for deletional α⁺-thalassaemia other than the two common Chinese mutations (−α3.7 and −α4.2). This could explain the presence of a small amount of HbBart's

A recent study showed a decreased rate of morbidity if HbH disease is diagnosed at birth, ² making a strong case for newborn screening, especially in areas with high α-thalassaemia prevalence. ³ Newborn screening for HbH disease is based on the detection of HbBart's which is only possible within the newborn period, before γ-globin production switches to β-globin. ⁴

As shown in our study, an elevation of the HbBart's percentage correlated with the number of affected α genes. Significant differences were found in Bart's levels between samples with HbH disease and those with two α-globin gene defect. Using an HbBart's cut-off value of 10%, HbH disease could be differentiated with absolute certainty from other milder α-thalassaemia forms. Our results suggest that analysis of HbBart's using the capillary electrophoresis system could be an efficient screening approach for HbH disease in newborns. This has also been confirmed in other studies. ^5,6