113

ORIGINAL ARTICLE

Neonatal screening for sickle cell disease in Central

Africa: a study of 1825 newborns with a new
enzyme-linked immunosorbent assay test
LeŁon Mutesa, Franc- ois Boemer, Louis Ngendahayo, Stephen Rulisa, Emmanuel K Rusingiza,
Neniling Cwinya-Ay, DeŁogratias Mazina, Pierre C Kariyo,Vincent Bours and Roland Schoos
..................................................................................................

J Med Screen 2007;14:113–116

Objectives To evaluate the feasibility of systematic neonatal screening for sickle cell disease in the
region of Great Lakes in Central Africa using a new approach with limited costs.
Methods Between July 2004 and July 2006, 1825 newborn dried blood samples were collected
See end of article for
authors’ affiliations onto filter papers in four maternity units from Burundi, Rwanda and the East of the Democratic Republic
............... of Congo. We tested for the presence of haemoglobin C and S in the eluted blood by an enzyme-linked
immunosorbent assay (ELISA) test using a monoclonal antibody. All ELISA-positive samples (multiple of
Correspondence to:
Professor Vincent Bours, the median (MoM)X1.5) were confirmed by a simple molecular test. The statistica software version
Department of Human 7.1 was used to create graphics and to fix the MoM cut-off, and the w2 of Pearson was used to compare
Genetics, CHU Sart-Tilman, the genotype incidences between countries.
University of Liège, 4000,
Liège, Belgium;
Results Of the 1825 samples screened, 97 (5.32%) were positive. Of these, 60 (3.28%) samples
[email protected] were heterozygous for Hb S, and four (0.22%) for Hb C; two (0.11%) newborns were Hb SS
homozygotes.
Accepted for publication Conclusions The lower cost and the high specificity of ELISA test are appropriate for developing
8 May 2007
............... countries, and such systematic screening for sickle cell anaemia is therefore feasible.

INTRODUCTION screening programme was initially introduced in maternity

units from Kigali and Butare in Rwanda, and later extended

S
ickle cell anaemia (SS) is the most common and to other major maternities of Bujumbura in Burundi, and
severe haemoglobinopathy in African populations.1 Goma in the East of the Democratic Republic of Congo. The
It is an autosomal recessive genetic disorder char- study protocol was approved by the Institution Review
acterized by a mutation in the gene coding for the b-globin Board of Medical Research.
and production of abnormal haemoglobin (Hb S) in red A standard data form included the identification and
blood cells, which become hard and sticky and are shaped geographic origin of every newborn. The parents were
like sickles. The clinical features of SS include increased informed about the procedures and gave their informed
infection risk, variable degrees of haemolysis leading to pain consent. Dried blood samples were collected by heel prick
and anaemia, and intermittent episodes of vaso-occlusion onto filter paper (S&S, Schleicher & Schuell BioScience,
that cause tissue ischaemia, acute and chronic organs Inc., 10 Optical Avenue.Keene NH, USA) during the first
dysfunction.2 Heterozygous individuals have a sickle cell week of life. Feedback information was given to the
trait (Hb SA), a generally benign, asymptomatic carrier paediatricians and haematologists in charge of the affected
state.3 newborns, and the appropriate medical care was started
In Central Africa, the prevalence of this pathology appears for homozygous patients. The parents were also informed
high (1.65%) and more than half of the affected children die about the risk of recurrence in subsequent pregnancies.
before the age of 5 years. The survivors develop degen- The parents of the heterozygous children were also
erative complications, increased morbidity and reduced life contacted, when possible, for further information. At the
expectancy.4 However, precocious and specialized medical time of the study, it was not possible to organize a
care reduces morbidity and mortality of these patients. systematic genetic screening in these families, but such an
Therefore, we initiated a pilot study to evaluate the approach should be feasible in Rwanda in the near future.
feasibility of a systematic neonatal screening programme
for sickle cell disease using a new ELISA test in the Great
Lakes region of Central Africa.
Immunological analysis by ELISA test
MATERIAL AND METHODS Primary screening was performed using the immunologi-
cal approach by ELISA with a mouse immunoglobulin
Population and samples G (IgG) monoclonal antibody detecting Hb S and Hb C, as
Between July 2004 and July 2006, a neonatal screening previously described.5 The defined multiple of the median
pilot study for sickle cell disease was conducted in three (MoM) value was used to analyse absorbance results
countries of the Great Lakes region in Central Africa. The (MoMX1.5).

www.jmedscreen.com Journal of Medical Screening 2007 Volume 14 Number 3

114 Mutesa et al.

Restriction fragment length polymorphism All ELISA-positive samples (MoM above 1.5) were
checked by a PCR-restriction molecular test, and the final
All positive samples were analysed by restriction fragment
result was considered as positive if confirmed by the
length polymorphism (RFLP). Genomic DNA was isolated
molecular analysis. Among the 48 positive samples from
from dried blood samples with the Instagene Dry Blood Kit
Rwanda, 28 were Hb S heterozygous carriers (Hb SA,
(BioRad, California, USA). The primers were designed as
2.54%); two were Hb C carriers (Hb CA, 0.18%) and one
follows: 50 -TAGGGTTGGCCAATCTACTC-30 for forward, and
had SS anaemia (Hb SS, 0.09%) (Table 2). From Burundi,
50 -TTAGGGTTGCCCATAACAGC-30 for reverse. The poly-
out of the 38 ELISA-positive patients, 28 were Hb S
merase chain reaction (PCR) conditions, as previously
heterozygous carriers (Hb SA, 4.40%), while one had sickle
reported, were used to amplify a 444 bp fragment of the
cell disease (Hb SS, 0.16%).
b-globin gene covering the two mutation sites, Hb S and Hb
We received only 84 samples from the Democratic
C, at codons 6 and 26, respectively. The 444 bp fragment of
Republic of Congo, a small number that does not allow
the b-globin gene was digested with restriction endonuclease
any appropriate allele frequency evaluation. However, the
enzymes, Bsu36I and BseRI (New England Biolabs, Massa-
Democratic Republic of Congo Hb S incidence was higher
chusetts, USA), for discrimination of the two haemoglobin
(8.33% of positive ELISA tests and 7.14% of confirmed
mutations. The restricted samples were then checked on a
positive samples), and the difference was statistically
3% agarose gel electrophoresis (BioRad, California, USA).
significant between countries (P ¼ 0.002 based on the ELISA
tests, Table 1, and P ¼ 0.036 on the basis of molecular
Statistical analysis confirmations, Table 2). Indeed, among the 84 samples from
the Democratic Republic of Congo, 11 were ELISA positive
The statistica software version 7.1 was used to create and the PCR test revealed that four of these were Hb S
graphics and to fix the MoM cut-off level. The data were heterozygotes (Hb SA, 4.76%), while two were Hb C
then analysed by the Epi Info software 3.3.2 version. The w2 heterozygotes (Hb CA, 2.38%).
of Pearson was used to compare the genotype incidences For the whole cohort, our results indicated that the
between countries (threshold of error ¼ 0.05). overall incidence of SS in the area was 0.11%. The
prevalence of sickle cell trait (Hb SA) was 3.23%, whereas
the haemoglobin C carriers were less frequent (0.22%).
RESULTS
As described in a previous study on a large cohort of
From July 2004 to July 2006, 1825 (914 females and 911 Caucasians,5 samples containing Hb S or C had MoM values
males) newborn dried blood samples were collected from in excess of 2.0, and the reference cut-off was fixed at that
major maternity units in Rwanda, Burundi and the East of level for screening in that population. In this African
the Democratic Republic of Congo. Among these samples, population, we tested different hypotheses at various levels
97 (5.32%) newborns screened positive for haemoglobin S of MoM ranging between 1.5 and 3.4 (Figure 1). Ninety-
or C using an ELISA test with MoM values between 1.5 and seven patients had MoM values over a cut-off of 1.5; of
3.4 (Table 1). these 97 patients, 66 (68%) were confirmed by PCR to carry

Table1 Sample characteristics and incidence of ELISA-positive

18
tests (MoM>1.5)
16
Numbers of
Number of observations

samples Incidence of Hb S (ELISA) 14

12
n % n % P value AS-AC SS AA
10
Sex 8
Female 914 50.1 38 4.16
Male 911 49.9 59 6.48 6
4
Total 1825 100 97 5.32
2
Ethnic origin 0
Burundi 637 34.9 38 5.97
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4

DRCw
1.
1.
1.
1.
1.
2.
2.
2.
2.
2.
2.
2.
2.
2.
2.
3.
3.
3.
3.
3.

84 4.6 11 13.10
Rwanda 1104 60.5 48 4.35 0.002 MoM cut-off values
Figure 1 Results of ELISA test with MoM values X1.5. The
Total 1825 100 97 5.32 distribution of MoM values for AS-AC and SS samples, as confirmed
The P value refers to a comparison of the incidence between the three countries by the molecular test, are shown, as well as the numbers of false-
w
DRC ¼ Democratic Republic of Congo positive samples when a cut-off of 1.5 is chosen

Table 2 Incidence of Hb S and Hb C heterozygotes and homozygotes (n ¼1825)

Burundi DRC Rwanda Total
Genotype
n % n % n % n %
AA 608 95.45 78 92.86 1073 97.19 1759 96.38
AS 28 4.40 4 4.76 28 2.54 60 3.28
AC 0 0.00 2 2.38 2 0.18 4 0.22
SS 1 0.16 0 0.00 1 0.09 2 0.11

Total 637 34.90 84 4.60 1104 60.50 1825 100

DRC ¼ Democratic Republic of Congo

Journal of Medical Screening 2007 Volume 14 Number 3 www.jmedscreen.com

Neonatal screening for sickle cell disease 115

100

90

80

70

Percentage
60 Cut-off
50

40

30

20

10

0
1.5 1.6 1.7 1.8 1.9 2.0 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9 3.0 3.1 3.2 3.3 3.4
MoM cut-off values
% of total false % of loss of true Logarithmic graph of
positive samples positives cumulative frequencies

Figure 2 Acceptable level for an MoM cut-off value. The first curve (square symbols) indicates the
percentage of false-positive samples for the indicated cut-off values, relative to the total number of
false-positive samples (n ¼ 31) detected at a cut-off of 1.5. The second curve (triangle symbols)
represents the percentage of lost positive samples for the indicated cut-off values. The total number
of positive samples for a cut-off value of 1.5 was 66

100

90
Cumulative frequencies of positive samples

80

70

60

50

40

30

20

10

0
1.5 1.6 1.7 1.8 1.9 2.0 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9 3.0 3.1 3.2 3.3 3.4
MoM cut-off values
Logarithmic graph trend
Figure 3 Cumulative percentages of detected positive samples for the indicated MoM cut-off
values, relative to the number of positive samples detected with a cut-off of 1.5

Hb S or C mutations and we expect to have detected over DISCUSSION

90% of positive patients, heterozygotes and homozygotes
(Table 2, Figures 1 and 2). At this cut-off, 31/97 (32%) false- The main objective of neonatal screening is the early
positive results were generated. However, when the cut-off detection of newborns with sickle cell disease in order to
was fixed at 1.8, the rate of false-positives decreased start appropriate medical care, including penicillin prophy-
markedly (9.4%), but we lost 13.64% (9 heterozygous laxis around 2–3 months of age,6 preferably before the onset
patients) of positive samples. Beyond an MoM cut-off of 2, of symptoms. In the African population, although the
the number of false-positive results was slightly lower symptoms of sickle cell disease occur during the first year
(4 instead of 6 at a cut-off of 1.8), but the detection of Hb of life in 80% of the children, diagnosis is rarely achieved
S- and C-positive patients decreased significantly as we before the age of 2–3 years. As a consequence, it is likely that
identified only 43 out of 66 positive samples (a loss of 34.8% many children die undiagnosed in early infancy from
of positive samples) (Figures 2 and 3). meningitis, pneumonia or acute anaemia, as did one of

www.jmedscreen.com Journal of Medical Screening 2007 Volume 14 Number 3

116 Mutesa et al.

the two homozygous patients detected during the present Our data provide information about the incidence of
study. In addition, neonatal screening can alert the parents sickle cell disease in the region of Great Lakes in Central
of an affected baby about the risk for the next pregnancy Africa, and they suggest that a neonatal screening pro-
and the need for antenatal diagnosis. gramme based on an ELISA test is suitable and should be
However, in Central Africa, the prevalence of inherited extended at the national level in all countries of Central
sickle cell disorders has been underestimated.7 Neonatal Africa, especially in Rwanda and Burundi, where the sickle
screening is necessary to obtain correct numbers. In this cell disease has been previously considered to be rare.
pilot study, a new immunological technique, using an ELISA
test recognizing the Hb S and Hb C, was performed as the ................
primary screening method. As shown in Figure 2, an Authors’ affiliations
acceptable detection level of Hb S and Hb C was obtained Léon Mutesa, Assistant, Human Genetics, CHU, University of Liège,
at the MoM cut-off of 1.8 in this population. These values Belgium
Franc- ois Boemer, Assistant, Human Genetics, CHU, University of
are slightly different from our reported data in a mostly Liège, Belgium
Caucasian population, which showed that an MoM cut-off Louis Ngendahayo, Medical Doctor, Department of Pathology,
reduced at 2.0 markedly increased the number of false- CHU, National University of Rwanda, Rwanda
positives (12.9–44.2%) and did not allow any additional Stephen Rulisa, Medical Doctor, Department of Pediatrics, CHU,
National University of Rwanda, Rwanda
detection of Hb S or C.5 Moreover, in the present study, two Emmanuel K Rusingiza, Medical Doctor, Department of Pediatrics,
children were found to be homozygous for Hb S, with MoM CHU, National University of Rwanda, Rwanda
values of 2.3 and 2.4, respectively, which were different Neniling Cwinya-Ay, Medical Doctor, Department of Pediatrics,
from those reported in a Caucasian population.5 These CHU, National University of Rwanda, Rwanda
differences could be explained by population disparities. Déogratias Mazina, Research associate, Department of Public
Health, CHU, University of Liège, Belgium
Indeed, it has been shown that there are differences Pierre C Kariyo, Medical Doctor, Department of Pediatrics, CHU,
between normal values obtained from subjects with University of Burundi, Burundi
European or African ancestry.8 Therefore, different refer- Vincent Bours, Professor, Human Genetics, CHU, University of
ence ranges need to be considered for these two groups, and Liège, Belgium
Roland Schoos, Research associate, Human Genetics, CHU,
they may have implications for how different ethnic groups
University of Liège, Belgium
could respond to this genetic testing.
The ELISA test was validated as a simple and rapid
technique for neonatal screening with limited costs com-
pared with commonly used high-performance liquid chro-
REFERENCES
matography (HPLC) and isoelectric focusing (IEF) 1 Fleming AF. The presentation, management and prevention of crisis in sickle
cell disease in Africa. Blood Rev 1989;3:18–28
techniques, and it would be applicable by laboratories with 2 Serjeant GR. The clinical features of sickle cell disease. Baill Clin Haematol
limited facilities. The incidence of sickle cell disease in the 1993;6:93–15
studied population is high, and thus a neonatal screening 3 Lane PA. Sickle cell disease. Pediatr Clin North Am 1996;43:639–64
4 Koko J, Dufillot D, M’Ba-Meyo J, et al. Mortality of children with sickle cell
programme appears fully justifiable, particularly as prophy- disease in a pediatric department in Central Africa]. Arch Pediatr
lactic measures lead to a reduction of mortality and 1998;5:965–9
morbidity in childhood. It is important to mention that a 5 Boemer F, Vanbellinghen JF, Bours V, et al. Screening for sickle cell disease
on dried blood: a new approach evaluated on 27,000 Belgian newborns.
correct sampling is most important as insufficient amounts J Med Screen 2006;13:132–6
of blood on the filter paper can generate false-negative 6 Vichinsky E, Hurst D, Earles A, et al. Newborn screening for sickle cell
disease: effect on mortality. Pediatrics 1988;81:749–55
results. 7 Njamnshi AK, Wonkam A, Djientcheu Vde P, et al. Stroke may appear to be
Recently, in a series of 987 Rwandan neonates screened rare in Saudi-Arabian and Nigerian children with sickle cell disease, but
for red blood cell disorders using IEF method, the incidence not in Cameroonian sickle cell patients. Br J Haematol 2006;133:210;
author reply 211.
of sickle cell trait was 2.7%, but no other Hb variant was 8 Beutler E, West C. Hematologic differences between African-Americans
detected.9 In West Africa, 4.7% children were found to be and whites: the roles of iron deficiency and alpha-thalassemia on
heterozygous for Hb S, whereas 0.3% were heterozygous hemoglobin levels and mean corpuscular volume. Blood 2005;106:740–5
9 Munyanganizi R, Cotton F, Vertongen F, et al. Red blood cell disorders in
for Hb C.10 In our population, sickle cell trait was found Rwandese neonates: screening for sickle cell disease and glucose-6-
in 3.28% and Hb C trait in 0.22% of screened newborns, phosphate dehydrogenase deficiency. J Med Screen 2006;13:129–31
10 Masmas TN, Garly ML, Lisse IM, et al. Inherited hemoglobin disorders in
and these incidences were thus close to the previously Guinea-Bissau, West Africa: a population study. Hemoglobin 2006;
reported ones. 30:355–64

Journal of Medical Screening 2007 Volume 14 Number 3 www.jmedscreen.com