. . . .
ORIGINAL ARTICLE
Screening for haemoglobinopathies on cord blood:
laboratory and clinical experience
Fleur Wolff, Fre¤de¤ric Cotton and Be¤atrice Gulbis
J Med Screen2012;
19
:
116–122DOI: 10.1258/jms.2012.011107
See end of article for authors’ affiliations
. . . .
Correspondence to: Professor Be´atrice Gulbis, Department of Clinical Chemistry, Hoˆpital Erasme, Route de Lennik, 808, Brussels, Belgium; beatrice.gulbis@erasme. ulb.ac.be
Accepted for publication 24 April 2012
. . . .
Objectives Blood from the umbilical cord (cord blood) is screened for haemoglobinopathies in several neonatal screening programmes, as well as before banking as a source of stem cells. We investigated the pre-analytical and analytical aspects of neonatal screening for haemoglobinopathies on liquid cord blood using the Sebia Haemoglobin kit. We give an interpretation of the results as well as a proposed scheme for reporting of the results.
Methods A neonatal screening programme on liquid cord blood has been performed in all labour wards in Brussels since 1994. Using that material, the screening methods of isoelectric focusing and capillary zone electrophoresis were compared using 962 cord blood samples. From December 2008 to December 2010, 47,388 neonatal samples were analysed by capillary electrophoresis as the first-line method for neonatal screening. High-performance liquid chromatography was used as the second-line method.
Results Capillary zone electrophoresis on liquid cord blood enabled the detection of all clinically significant haemoglobin variants, significant levels of Hb Bart’s, andb-thalassaemia major. Among the 47,388 neonatal samples tested, 362 (0.7%) were suspected to be contaminated with maternal blood, but no diagnostic error was reported retrospectively for a major haemoglobinopathy. Recommendations for the interpretation and reporting of results of neonatal screening for haemoglobinopathies using the Sebia Haemoglobin kit are proposed.
Conclusions A routine capillary electrophoresis kit adapted to neonatal screening and liquid cord blood is reliable for screening for haemoglobinopathies. It enables early detection and reporting of all major haemoglobinopathies and most minor ones. It also enables use of a simple scheme to report the results.
INTRODUCTION
C
urrently, blood from the umbilical cord (cord blood) is screened for haemoglobinopathies in two situ-ations: 1) as part of a neonatal screening pro-gramme, and 2) before stem cells from cord blood are added to a cord blood bank. The main purpose of a neonatal screening programme for haemoglobinopathies is to identify sickle cell disorders, because early diagnosis has been demonstrated to reduce mortality during childhood when followed by vaccination, the institution of prophylaxis against infection, and parental education.1,2In 1972, it was demonstrated that haemoglobinopathies could be screened for at birth using blood samples from the umbilical cord.3 Subsequently, several institutions
implemented comprehensive screening programmes for sickle cell disorders using cord blood.4 – 7 However, given
that dried blood spots obtained by heel prick for use with the Guthrie card were already used in neonatal screening programmes for metabolic disorders and thus were readily available, neonatal screening for haemoglobinopathies on cord blood was abandoned in several countries. However,
cord blood is still used in various countries, and especially in Africa.8,9
Cord blood banking is a complex process that requires a carefully designed quality control system that includes several sequential steps.10 All samples of cord blood that are intended for banking have to be screened for haemoglo-binopathies before transplantation, in accordance with the International Standards for Cord Blood Banking.11
To screen effectively for haemoglobinopathies, either for cord blood banking or as part of a neonatal screening pro-gramme, the screening technique used should be able to identify neonates who suffer from a major haemoglobinopa-thy, such as sickle cell disorder, b-thalassaemia major or haemoglobin H disease. This means that the technique chosen must be able to detect homozygotes and compound heterozygotes for all the common haemoglobin (Hb) var-iants, including Hb S, Hb C, Hb D-Punjab, Hb E, Hb O-Arab, and Hb Bart’s. The method must also be sensitive, in order to detect a low level of Hb A. In cases with a major haemoglobinopathy, the cord blood will be excluded from the bank, and such neonates who are identified
through the neonatal screening programme will receive special attention. Those affected by sickle cell syndrome will benefit from an early diagnosis and be given compre-hensive care. In the case of thalassaemia major, neonates will be referred to dedicated health-care staff and their parents will benefit from genetic counselling.
As mentioned above, techniques that are used for neo-natal screening for haemoglobinopathies should be able to detect heterozygotes for all the common Hb variants. The neonates who are identified by the test will benefit from the knowledge of their own genetic risk, while their parents will benefit from information on their status. Nevertheless, the reporting of carriers remains a subject of debate, because of the potential risk of highlighting false paternity and the psychological complications for the child and family.
The two techniques that are used commonly for neonatal screening for haemoglobinopathies have a high sensitivity and specificity. These are isoelectric focusing (IEF) and high-performance liquid chromatography (HPLC).12 Both will detect most clinically significant Hb variants, as well as the absence or a very low level of Hb A. Capillary electro-phoresis (CE) is another, relatively new, technique that can be used.13The HPLC and CE techniques can be automated fully and provide quantitative results. Other techniques, such as tandem mass spectrometry and immunoassays, are available but the former remains expensive and requires complex sample preparation, whereas the latter is focused on screening for a specific Hb variant only, such as Hb S or Hb C.14,15
A universal standalone neonatal screening programme for haemoglobinopathies on liquid cord blood was set up in Brussels in 1994. The programme includes all neonates born under the care of the Brussels labour wards. Until 2007, cord blood was screened by IEF, with HPLC used as the confirmatory technique when variants were detected. The CE technique has also been demonstrated to be a reliable tool for neonatal and general screening for haemoglobinopathies.13,16
The major objective of the current study was to determine whether liquid umbilical cord blood and a routine capillary zone electrophoresis method are reliable for use in screening for haemoglobinopathies at birth. Our study allowed us to provide an interpretation and to describe limitations in the presumptive identification of a haemoglobinopathy by CE, as well as introducing a scheme for the reporting of results.
MATERIALS AND METHODS
Data on neonates and samples
The samples used in the study were those collected during the universal neonatal screening programme for haemoglo-binopathies that has been in operation in labour wards in Brussels since December 1994. In brief, liquid cord blood samples in ethylenediaminetetraacetate are collected care-fully to avoid contamination with maternal blood. The eth-nicity of the parents and the birth weight (,2.5, 2.5 – 3, .3 – 3.5, .3.5 kg) and gestational age (,32, 32 – 36, .36 – 38,.38 weeks) of the neonates, as well as any trans-fusion given before the test, are recorded. The neonatal
screening programme is approved by each local ethics committee.
Capillary electrophoresis method: normal pattern
and quantification of Hb A
Capillary electrophoresis was performed directly on native cord blood using the Capillarys Hemoglobin Kit (Sebia, Vilvoorde, Belgium), but using a specific CE cord blood pro-gramme on the Capillarys II CE system (Sebia).16Each series of samples included an Hb AF control (Sebia, PN4777).
The effect of gestational age on the percentage of Hb A was evaluated in a subgroup of 402 neonates for whom the exact gestational age was recorded. Normal percentages of Hb A in cord blood samples were determined for neonates with a gestational age greater than 38 weeks who presented a normal pattern of Hb fractions (i.e. the presence of only Hb A and Hb F).
Evaluation of pre-analytical parameters
for screening by CE
First, we evaluated the possibility of contamination of cord blood by maternal blood.
To identify a biological parameter that indicated maternal contamination of cord blood samples, we assessed the effect of such contamination on haematological parameters and Hb pattern. The parameter red cell distribution width (RDW) was quantified on 15 fresh, uncontaminated samples of cord blood to generate a 95% normal range. Subsequently, four samples of cord blood were contami-nated artificially with their respective maternal blood at 5%, 10%, 20% and 40%, and the RDW and percentage of Hb A were determined for each level of contamination.
We then assessed the effect of maternal blood contami-nation on the pattern of Hb fractions. For this test, a cord blood sample with an Hb S level of 16% was diluted at different levels with a normal adult sample (from 5% to 85%).
We also evaluated the effect of a delay between sampling and analysis on the proportion of the different Hb fractions. We determined the percentages of Hb A and Hb S that were obtained by CE and HPLC after 17 cord blood samples from neonates who were heterozygous for Hb S had been stored at ambient temperature for one week.
Comparison between the IEF and CE methods
A set of 962 cord blood samples were tested by both CE and IEF. Isoelectric focusing was carried out using the Perkin Elmer Neonatal Hemoglobin Test Kit, in accordance with the manufacturer’s instructions (PerkinElmer Life Sciences, Zaventem, Belgium). Profiles obtained with IEF were ana-lysed as reported previously.17The presumptive identification of all Hb variants was con-firmed by ion-exchange HPLC on a Variant II Hemoglobin Testing System using the b-Thalassemia Short Program (Bio-Rad, Clinical Diagnostics, Nazareth Eke, Belgium) as described previously.17In brief, the HPLC technique separ-ates Hb fractions, quantifies Hb A2and Hb F, and provides
qualitative determinations of abnormal Hb variants on the basis of their respective retention times.
Evaluation of CE as a first-line screening test
A total of 47,388 cord blood samples were analysed from December 2008 to December 2010 by CE, used as a first-line screening method. For all Hb variants detected by CE, HPLC was used as a second-line method.
Statistical analyses
Statistical analyses were performed with the GraphPad Prism software (version 4, Graph Pad Software, Inc., USA). Values were reported as the median and range. Changes in the per-centage of Hb A in relation to weeks of pregnancy and birth weight were assessed by one-way analysis of variance (ANOVA) on 402 neonates whose gestational age and birth weight had been recorded accurately. The relationship between the percentage of Hb A or Hb Bart’s and gestational age was evaluated by calculating the Spearman correlation coefficient. The effect of storage at room temperature on the proportions of Hb A and Hb S in cord blood samples that were obtained from the 17 neonates who were hetero-zygous for Hb S was assessed with the pairedt-test. APvalue ,0.05 was considered statistically significant. The percen-tages of Hb Bart’s obtained by CE versus HPLC were com-pared by linear regression analysis and Bland – Altman analysis.
RESULTS
Capillary electrophoresis method: normal Hb
pattern and quantification of Hb A
The relative migration times of the normal Hb fractions (i.e. Hb A and Hb F) are shown in Figure 1. The electrophoreto-gram obtained was divided into zones C1 through C12 by standardization relative to the location of Hb F.
The relationship of the level of Hb A with gestational age was determined for cord blood samples from 402 neonates. Birth weight and gestational age were correlated signifi-cantly with the percentage of Hb A, with Spearman coeffi-cients of 0.4124 and 0.5557, respectively (P,0.0001). A significant percentage of Hb A was already present at 26 weeks of gestation (Table 1). For full term neonates, the median (range) percentage of Hb A at birth was 20.5% (9 – 43%).
Evaluation of pre-analytical parameters for
screening by CE
Among the 47,388 cord blood samples tested, the prevalence of maternal blood contamination was estimated. For 331 (0.70%) samples, a peak for Hb A2 (median Hb A2%:
0.7%; range: 0.3 – 2.2%) and a percentage of Hb A greater than 40% (median Hb A%: 43.2%; range: 40 – 89%) were observed; these samples were suspected to be contaminated with maternal blood. Thirty-one samples had a percentage of Hb A .90% (median HbA%: 97.4%; range: 90.8 – 98%) and were considered to be probably maternal blood.
The use of RDW and Hb A% as parameters to detect maternal blood contamination was evaluated. The 2.5th and 97.5th percentiles obtained for RDW in 15 fresh samples of cord blood that were not contaminated were 17% and 21%, respectively, which were significantly differ-ent from the values obtained in a female adult reference population (11% and 13%). The effect of maternal blood contamination at a level of 5% to 40% on the RDW and the percentage of Hb A was evaluated. The RDW and per-centage of Hb A increased significantly only after contami-nation with 40% maternal blood and thus were not useful parameters for use to detect low levels of maternal blood contamination.
The effect of maternal blood contamination on the detec-tion of an abnormal pattern of Hb fracdetec-tions was also evalu-ated. For a cord blood sample that presented an Hb S level of 16%, a level of 1% of Hb S was still detected after maternal blood contamination at 85%.
These results demonstrate that maternal contamination is rare (0.7%), but equivocal results (i.e. a level of Hb A.40% or a very low level of an Hb variant when compared with the
Figure 1 Relative migration times of the different haemoglobin variants separated with a specific cord blood program on the Capillarys 2
Table 1 Percentages of Hb A obtained by capillary electrophoresis at different gestational ages (n¼402). Values are expressed as the median (range)
Gestational age
(weeks) Number ofsamples Median Hb A percentage(range)
26–27 2 5.4 (4.8–6) 31–32 5 9.0 (5.2–15) 33 7 7.0 (6.6–9.2) 34 6 7.8 (7.3–11) 35 7 9.6 (7.5–13) 36 18 12.0 (8.4–26) 37 28 14.0 (10–25) 38 59 17.0 (11–41) 39 85 19.0 (9.1–38) 40 110 21.5 (11–38) 41 69 24.0 (11–43) 42 6 19.5 (16–28)
Hb A level) require repeat testing of the neonate to ensure that no significant haemoglobinopathies have been missed. The effect of a delay between sampling and analysis on the percentages of Hb A and Hb S was also assessed. The time dependence of the change in the percentages of Hb A and Hb S as determined in 17 heterozygous AS cord blood samples is shown in Figure 2. After one week of storage at 20 – 258C, a significant rise in the percentage of Hb A (þ39.4%) was observed with the CE technique (P, 0.0001) and the median percentage of Hb S decreased sig-nificantly ( – 9.2%, P,0.0001). A significant decrease in the median percentage of Hb S was also observed with the HPLC technique ( – 15.6%,P¼0.0038). However, the per-centage of Hb A determined by HPLC remained stable after one week of storage at room temperature (þ5.7%,P .0.05).
Comparison between IEF and CE
We compared the results obtained with the IEF and CE tech-niques for 962 samples of cord blood. The results are shown in Table 2. Among the specimens, 16 Hb variants (1.7%) were identified by both techniques, with no discrepancy in the detection of variants. A low percentage of Hb A could be detected with IEF by visual assessment of the Hb A/Hb F acetylated ratio or the Hb F/Hb A ratio.18On the basis of the 2.5th percentile of the distribution of the percentage of Hb A obtained by CE in normal full-term neonates, IEF and CE identified a low level of Hb A in 43/962 (4.5%) and 23/962 (2.4%) of the samples, respectively. The advan-tage of CE is the quantification of the Hb fractions, whereas interpretation of the Hb A level by IEF is subjective.
Hb Bart’s can be recognized on IEF gels as two or three fast-moving fractions. Hb Bart’s was identified in 63/962 (6.5%) and 23/962 (2.4%) of the samples tested by IEF and CE, respectively. All the Hb Bart’s fractions that were undetected by CE were weak bands when observed by IEF.
Evaluation of CE as a first-line screening test
We report our two years of experience using CE as the primary screening technique for haemoglobinopathies and HPLC as the secondary technique to confirm all abnormal patterns of Hb.
The relative migration times of the clinically significant Hb variants identified by CE are shown in Figure 1. Among the 47,388 samples analysed, 41 neonates (1:1156) had a sickle cell disorder: Hb SS (n¼31), Hb SC (n¼6), Hb SE (n¼1), Hb S-dbthalassaemia (n¼1), Hb S-b8thalassaemia (n¼1), and Hb S-bþthalassaemia (n¼1).
The screening also revealed the presence of heterozygosity forb-globin variants (Hb S, Hb C, Hb D-Punjab, Hb E, and Hb O-Arab) in 975 (1:49) neonates. The most common phe-notype was the Hb S trait (1:58). Some uncommon variants (a- org-globin variants) were also identified in 62 neonates (1:764).
In three cases, discrepancies were observed between the results obtained by CE and HPLC. The first case showed the absence of Hb A when tested by CE, whereas a low per-centage of Hb A (1.4%) was detected by HPLC. A molecular analysis of the sample confirmed the presence of Hb S-bþ
thalassaemia. The second case concerned a neonate who was identified to be heterozygous for Hb S by CE, but homo-zygous for Hb S by HPLC. A sample obtained six weeks later showed a typical profile of homozygosity for Hb S on CE. The third case was tested five times. An Hb AS profile that was suspected from the results of CE was not confirmed with HPLC. However, parental analysis and globin separ-ation analysis demonstrated the presence of the a- or
g-variant in three of the five tests. All these cases were resolved by performing a second-line technique that used another methodology of separation, which is the rec-ommended procedure.
Among the 47,388 samples analysed, 953 neonates (2%) showed a peak for Hb Bart’s in the range of 0.1 to 8.8%, and 229 children (0.5%) had a percentage greater than 2%. On the basis of the Spearman correlation coefficient, neither gestational age nor birth weight was correlated with the per-centage of Hb Bart’s (data not shown). No case of Hb H disease was diagnosed.
The percentage of Hb Bart’s was quantified in samples from 32 neonates with both the CE and HPLC techniques. The linear regression of the percentages for Hb Bart’s that were obtained with CE and HPLC showed a considerable magnitude of error, with a slope of 0.6726 (95% confidence interval: 0.1305 – 1.215), a y-intercept of 8.397 (95% confi-dence interval: 6.358 – 10.44), and a coefficient of corre-lation of 0.1762. The Bland – Altman analysis showed consistently that HPLC gave a higher percentage of Hb Bart’s than CE, with a bias of 7.34 (standard deviation: 2.95). To suspect Hb H disease, the cut-off of Hb Bart’s that is used for CE should be much lower than that used for HPLC: it should be above 10%.
A proposal for the reporting of results is given in Table 3.
DISCUSSION
In the present study, we investigated whether the use of liquid cord blood and CE is reliable in screening for haemo-globinopathies. Our results showed that liquid cord blood was generally not contaminated with maternal blood and was suitable for testing. The Capillarys Hemoglobin Kit from Sebia is a fully automated system that allows easy interpretation of the Hb fractions, as well as the detection of all major and most minor haemoglobinopathies. However, as recommended, a second-line test that used a different methodology was necessary to confirm the results obtained by CE.19
Screening of cord blood for sickle cell anaemia was reported first in 1972. The major disadvantage of the use of this type of sample is probably the risk of contamination by maternal blood,3 which leads to a significant risk of a false-negative in relation to the assignment of disease phe-notype. A comparison between liquid and dried blood for neonatal haemoglobinopathy screening was reported in 1994. It was concluded that both types of sample were equally reliable for use in neonatal haemoglobinopathy screening.20Our experience showed that maternal contami-nation would need to be very high to alter the neonatal Hb profile and this was rarely observed (331/47,388; 0.7%). Unfortunately, the 95% reference intervals for the biological
parameters RDW and percentage of HbA were too wide to allow them to be used to exclude all contaminated samples. The use of cord blood samples has some advantages that are important to consider: (1) the large volume of the sample, which allows confirmatory techniques (e.g. DNA testing) to be applied; (2) the high quality and stability of the sample; (3) regardless of the technique used, degraded Hb components are often present when dried blood spots are analysed, and considerable experience is required for the correct interpretation of separation profiles;21 and (4) the result is available very rapidly after birth.
The Brussels neonatal screening programme, which uses cord blood, enables maternity staff to be informed of the results and confirmatory samples to be obtained from chil-dren suspected to be affected by haemoglobinopathies before the infants are discharged home.17 In Africa, where women are usually discharged one day after delivery, cord blood is the easiest material to collect. Moreover, African physicians have reported that the collection of cord blood is more acceptable to families than heel pricks. The cost of sampling in a tube or on filter paper is comparable, but
the processing of a tube of blood is much easier and can follow the analytical process applied in a routine laboratory. Owing to extensive immigration, sickle cell disorders are now the most prevalent serious inherited diseases in Belgium. The main goals of screening for haemoglobinopa-thies are to detect neonates with a sickle cell disorder so that preventive medical treatments can be implemented, and to detect minor haemoglobinopathies so that genetic counselling can be provided. In addition to neonatal screen-ing for haemoglobinopathies, samples of cord blood that are intended for the cord blood blank must be screened for sickle cell disorders and thalassaemia. The quality control procedures for standard cord blood banking recommend the use of HPLC, IEF or CE. The method used to screen for a haemoglobinopathy must be able to characterize Hb A, Hb A2, Hb S, and Hb C in the heterozygous or
homozy-gous state. Units of cord blood that are derived from neo-nates who are found to have a sickle cell disorder or a
b-thalassaemia major are discarded. Units of cord blood from neonates who are found to have a minor haemoglobi-nopathy are accepted but listed as such.11
To assess the capacity of the Capillarys Hemoglobin Kit to screen for haemoglobinopathies on umbilical cord blood, we compared the results from 962 neonates obtained in parallel by the IEF and CE methods and tested the use of CE as a first-line screening test during a two-year period. Using the CE method, the median percentage of Hb A in normal full-term neonates was 20.5%, and the percentage increased sig-nificantly with gestational age (17, 19 and 21.5% at 38, 39 and 40 weeks of gestational age, respectively). These data were in agreement with those published by Mantikou et al., who found median percentages of Hb A of 18.2, 19.5 and 21.9% in cord blood samples after 38, 39 and 40 weeks of gestation, respectively.22
Both the IEF and CE methods performed equally well in detecting structural Hb variants. Compared with CE, the Figure 2 Seventeen cord blood samples were kept at room temperature for one week. Changes in Hb A% measured (a) by CE and (b) by HPLC, or in Hb S% measured (c) by CE and (d) by HPLC
Table 2 Comparison of the results obtained by isoelectric focusing (IEF) and capillary electrophoresis (CE) on 962 samples IEF CE Normal Hb (s) 840 900 Abnormal Hb 16 16 Hb Bart’s 63 23 Low level of Hb A 43 23
Total samples analysed 962 962
Hb Bart’s was recognized on gels obtained by IEF as two or three rapidly migrating bands.
A low percentage of Hb A was reported with IEF by assessing the ratio Hb A/Hb F acetylated or the ratio Hb F/Hb A as described previously.18
IEF method identified more neonatal samples with a low percentage of Hb A (4.5% versus 2.3%). A previous report revealed that visual examination of IEF gels gave some false-positive results for samples determined to have low proportions of Hb A and suggested the use of a quantitative technique to confirm the initial result.20
The presence of Hb Bart’s was also detected more often by IEF than by CE (6.5% versus 2.3%). However, all the Hb Bart’s fractions that were undetected by CE were weak bands on IEF. Of the 47,388 neonates screened with CE, Hb Bart’s was detected in 2%. The level ranged from 0.1% to 8.8% and was not correlated with gestational age or birth weight, as demonstrated previously.17Following con-sideration of the ethnic origin of the child and the parental phenotype, if available, the presence of Hb Bart’s at a level higher than 2% was reported. The level of Hb Bart’s is related to the number of defective or missing a-genes. By using the cut-off value of 2% Hb Bart’s for CE, all the neo-nates who had at least two defects ina-globin genes would normally be detected. In fact, Munkongdeeet al. found that the mean+standard deviation percentage of Hb Bart’s in cord blood samples obtained from neonates with two
a-globin gene defects were 3.6+1.3% and 4.6+0.5% for a-thalassaemia 2 homozygotes (-a/-a) and for
a-thalassaemia 1 heterozygotes ( –/aa), respectively.23 Heterozygousa-thalassaemia-2 (-a/aa) occurs when there is only one gene missing. These individuals are known as silent carriers, and in these cases a small percentage (1% to 4%) of Hb Bart’s can be detected by IEF. Such cases might be missed by CE, whereas levels of Hb Bart’s that are higher than 4%, which are observed in neonates sus-pected of having twoa-globin gene defects, can be detected by CE.23The linear regression analysis of the results for Hb
Bart’s obtained by CE and HPLC demonstrated a poor corre-lation, which is consistent with previously reported findings. The cut-off levels proposed for Hb Bart’s to detect neonates with Hb H disease are 25%24,25 and 29%23 for HPLC and CE, respectively.
As recommended, any abnormality that is detected by a first-line screening method must be confirmed in the same sample by a second method.19 This second method must be based on a different principle of separation. The main combinations that are used are IEF followed by HPLC or vice versa. This approach enables the reliable identification of most haemoglobinopathies. Indeed, problems can occur during neonatal screening for haemoglobinopathies if only one technique is performed without a confirmatory tech-nique. Some of the problems are of minor importance; for example, identifying a neonate initially as having an Hb S trait and demonstrating subsequently with HPLC that the infant actually carries an a- org-globin variant. However, other problems can have dramatic consequences that cause a delay in the optimal management of a child and their family. This is illustrated by the example of a neonate for whom a heterozygous Hb S profile was suspected initially with CE. The HPLC analysis did not confirm the initial result and instead showed a homozygous Hb S profile. The anomaly might be explained by the significant rise in the percentage of Hb A (þ39.4%) that was observed only with CE after samples had been stored for one week at ambient temperature, with a concomitant and significant decrease in the percentage of Hb S ( – 9.2%). This apparent increase in the percentage of Hb A during storage could be attributed to the presence of products of Hb degradation that migrated simultaneously with Hb A on CE. In contrast, upon analysis with HPLC, the percentages of Hb A in the same samples
Table 3 Presumptive diagnosis and confirmation of major haemoglobinopathies screening results obtained by capillary electrophoresis (CE) and confirmation by high-performance liquid chromatography (HPLC) on umbilical cord blood
Major haemoglobinopathies and potentially clinically significant disorders Capillary
electrophoresis
Hb(s) pattern Presumptive diagnosis Confirmation by HPLCon the same sample If both CE
/HPLC results are consistent Ask for a new sample
Hb FþHb S Hb FþHb SþHb C Hb FþHb SþHb D-Punjab Hb FþHb SþHb O-Arab Hb FþHb SþHb E Hb FþHb SþHb X Hb FþHb SþHb A with Hb S%.Hb A%
Sickle cell disorder Yes Immediately and parents’ samples if phenotype/genotype unknown
Hb F only Hb FþHb A with
Hb A,6 %
Possibleb-thalassaemia major or intermedia or prematurity or hereditary persistence of Hb F
Yes If full-term neonate, retest immediately with parents’ samples if phenotype/genotype unknown
If premature, within three months or immediately for parents’ samples if phenotype/genotype unknown Hb FþHb A with Hb
Bart’s.10% a-thalassaemia trait or Hb Hdisease Yes Immediately, and parents’ samples ifphenotype/genotype unknown Hb FþHb E
Hb FþHb XþHb Y Hb FþHb C Hb FþHb X
Consider possible mild to severe
haemoglobinopathy Yes Within three months, and parents’ samples ifphenotype/genotype unknown
were not altered after one week of storage at room tempera-ture (þ5.7%). The combination of CE as the primary method for screening with HPLC as the secondary technique maintained the same specificity and sensitivity as IEF/HPLC but offered high throughput and more efficient automation of the first step, thus reducing laboratory workload.
From December 2008 to December 2010, the estimated incidence of sickle cell disorders among neonates born in Brussels was 1:1156. On the basis of comparison with results published previously, the incidence of this mono-genic disorder appears to be increasing in Belgium. From December 1994 to June 1998 and from June 1998 to December 2007, the estimated incidence of sickle cell dis-orders was 1:2103 and 1:1559, respectively.7,26A diagnosis of sickle cell disorder in a neonate is reported immediately to the local co-ordinator of haemoglobinopathies screening and the medical staff of the maternity service concerned. The phenotype is confirmed by retesting the affected child immediately.
In the Brussels screening programme, heterozygosity for an Hb variant is reported to the local co-ordinator by regular mail within two weeks. To confirm the initial result and to exclude the possibility of maternal contami-nation of cord blood, all heterozygous profiles for clinically relevant abnormal Hb are re-tested three months after the infant’s birth. If possible, an analysis of the Hb pattern of the parents is also performed. The detection and reporting of minor haemoglobinopathies with clinically relevant var-iants in a neonate provides information that is likely to be relevant to future pregnancies of the mother.
In conclusion, the use of samples of liquid umbilical cord blood and capillary zone electrophoresis, i.e. the Capillarys Hemoglobin Kit from Sebia, to screen for haemoglobinopa-thies is reliable. Maternal contamination of cord blood is rare. Any abnormality that is detected by CE must be con-firmed in the same sample by an additional technique. At the very least, all neonates with a clinically significant Hb variant, with no Hb A, or with a high level of Hb Bart’s should be retested to confirm the initial result. In the neo-natal screening programme used in Brussels, no infant has been reported to have a normal Hb pattern at birth but found subsequently to be positive for a major haemoglobi-nopathy or a haemoglobihaemoglobi-nopathy trait. It enables confirma-tory samples to be obtained from children suspected of having a haemoglobinopathy before the infants are dis-charged home.
. . . .
Authors’ affiliations
Fleur Wolff, Laboratory of Clinical Chemistry, Hoˆpital Erasme, Universite´ Libre de Bruxelles, Brussels, Belgium
Fre´de´ric Cotton,Laboratory of Clinical Chemistry, Hoˆpital Erasme, Universite´ Libre de Bruxelles, Brussels, Belgium
Be´atrice Gulbis,Laboratory of Clinical Chemistry, Hoˆpital Erasme, Universite´ Libre de Bruxelles, Brussels, Belgium
REFERENCES
1 Gaston MH, Verter JI, Woods G,et al.Prophylaxis with oral penicillin in children with sickle cell anemia. A randomized trial.N Engl J Med
1986;314:1593–9
2 Leˆ Phu Quoc, Ferster A, Cotton F,et al.Sickle cell disease from Africa to Belgium, from neonatal screening to clinical management.Medecine Tropicale2010;70:467–70
3 Pearson HA, O’Brien RT. Sickle cell testing programs.J Pediatr
1972;81:1201–4
4 Pearson HA, O’Brien RT, McIntosh S, Aspnes GT, Yang MM. Routine screening of umbilical cord blood for sickle cell diseases.JAMA
1974;227:420–1
5 Sergeant BE, Forbes M, Williams LL, Sergeant GR. Screening cord bloods for detection of sickle cell disease in Jamaica.Clin Chem
1974;20:666–9
6 Evans DI, Blair VM. Neonatal screening for hemoglobinopathy: results in 7691 Manchester newborns.Arch Dis Child1975;51:127–30 7 Garrick MD, Dembure P, Guthrie R. Sickle-cell anemia and other
hemoglobinopathies. Procedures and strategy for screening employing spots of blood on filter paper as specimens.N Engl J Med
1973;288:1265–8
8 Kafando E, Nacoulma E, Ouattara Y,et al.Neonatal haemoglobinopathy screening in Burkina Faso.J Clin Pathol2009;62:39–41
9 Tshilolo L, Aissi LM, Lukusa D,et al.Neonatal screening for sickle cell anaemia in the Democratic Republic of the Congo: experience from a pionner project on 31204 newborns.J Clin Pathol2009;62:35–8 10 Rebulla P, Lecchi L. Cord blood banking and accreditation.ISBT Science
Series2007;2:91–5
11 NetCord-FACT International Standards for Cord Blood Collection. Processing and release for administration. Fourth Edition, 2010 12 Bain BJ. Neonatal/newborn haemoglobinopathy screening in Europe and
Africa.J Clin Pathol2009;62:53–6
13 Renom G, Mereau C, Maboudou P, Pe´rini JM. Potential of the Sebia capillarys neofast automated system for neonatal screening of sickle cell disease.Clin Chem Lab Med2009;47:1423–32
14 Boemer F, Ketelslegers O, Minon JM, Bours V, Schoos R. Newborn screening for sickle cell disease using tandem mass spectrometry.Clin Chem2008;54:2036–41
15 Mutesa L, Boemer F, Ngendahayo L,et al.Neonatal screening for sickle cell disease in Central Africa : a study in 1825 newborns with an enzyme-linked immunosorbent assay test.J Med Screen2007;14:113–6 16 Cotton F, Malaviolle X, Vertongen F, Gulbis B. Evaluation of an automated capillary electrophoresis system in the screening for hemoglobinopathies.
Clin Lab2009;55:217–21
17 Gulbis B, Tshilolo L, Cotton F, Lin C, Vertongen F. Newborn screening for haemoglobinopathies: the Brussels experience.J Med Screen1999;6:11–5 18 Black J. An isoelectrofocusing method to detect haemoglobin variants in
newborn blood samples including the beta-thalassemias.Hemoglobin
1988;12:681 –9
19 NHS antenatal and neonatal newborn screening programmes. http:// www.kcl-phs.org.uk/haemscreening/(last accessed 28 July 2008) 20 Papadea C, Eckman JR, Kuehnert RS, Platt AF. Comparison of liquid and
dried blood for neonatal hemoglobinopathy screening: laboratory and programmatic issues.Pediatrics1994;93:427–32
21 Mantikou E, Harteveld CL, Giordano PC. Newborn screening for hemoglobinopathies using capillary electrophoresis technology: Testing the Capillarys Neonat Fast Hb device.Clin Biochem2010;43:1345–50 22 Mantikou E, Arkesteijn SG, Beckhoven van JM, Kerkhoffs JL, Harteveld CL,
Giordano PC. A brief review on newborn screening methods for hemoglobinopathies and preliminary results selecting beta-thalassemia carriers at birth by quantitative estimation of the HbA fraction.Clin Biochem2009;42:1780–5
23 Munkongdee T, Pichanun D, Butthep P,et al.Quantitative analysis of Hb Bart’s in cord blood by capillary electrophoresis system.Ann Hematol
2011;90:741 –6
24 Michlitsch J, Azimi M, Hoppe C,et al.Newborn screening for hemoglobinopathies in California.Pediatr Blood Cancer
2009;52:486 –90
25 Fucharoen S, Winichagoon P, Wisedpanichkij R,et al.Prenatal and postnatal diagnoses of thalassemias and hemoglobinopathies by HPLC.
Clin Chem1998;44:740–8
26 Gulbis B, Cotton F, Ferster A,et al.Neonatal haemoglobinopathy screening in Belgium.J Clin Pathol2009;62:49–52
