Acta Pædiatrica ISSN 0803–5253

REGULAR ARTICLE

Intravenous immune globulin reduces the need for exchange transfusions

in Rhesus and AB0 incompatibility
KMN Huizing1,2,3 , J Røislien4 , TWR Hansen (t.w.r.hansen@medisin.uio.no)1
1.Neonatal Intensive Care Unit, Division of Paediatrics, Rikshospitalet University Hospital and Faculty of Medicine, University of Oslo, Norway
2.Faculty of Medicine, University of Maastricht, The Netherlands
3.Department of Paediatrics, Canisius-Wilhelmina Ziekenhuis, Nijmegen, The Netherlands
4.Department of Biostatistics, Institute for Basic Medical Sciences, University of Oslo, Norway

Keywords Abstract
ABO incompatibility, Exchange transfusion,
Aim: To conduct a quality control review of a single institution experience with intravenous immune
Intravenous immune globulin, Neonate, Rh
incompatibility globulin in the treatment of Rhesus and AB0 incompatibility.
Correspondence Methods: Intravenous immune globulin as treatment for Rhesus and AB0 incompatibility was
Thor Willy Ruud Hansen, Division of Paediatrics, introduced in our hospital in 1998. We performed a chart review of 176 infants with Rhesus or AB0
Rikshospitalet University Hospital, N-0027 Oslo,
incompatibility treated in our hospital between 1993 and 2003, divided into a historical control group
Norway.
Tel: +47-23074573 | (1993–1998) and a treatment group (1999–2003). The project was approved through institutional
Fax: +47-23072960 | ethics procedures.
Email: t.w.r.hansen@medisin.uio.no
Results: The use of exchange transfusion as a therapeutic modality was significantly reduced in the
Received cohort treated with intravenous immune globulin (OR 0.11; 95% CI 0.046–0.26, p < 0.001). We
22 February 2008; revised 22 May 2008;
accepted 26 May 2008. found no difference between the intravenous immune globulin group and the infants receiving only
DOI:10.1111/j.1651-2227.2008.00915.x
exchange transfusion as far as the duration of phototherapy. Infants with Rhesus incompatibility had a
higher need for top-up transfusions than those with AB0 incompatibility.

Conclusion: This study supports the evidence from previous studies suggesting that intravenous immune
globulin significantly reduces the need for exchange transfusion in infants with Rhesus and AB0 incompatibility.

The lack of newer, larger randomized controlled studies in

INTRODUCTION
Rhesus (Rh) and/or AB0 blood type incompatibility may
cause haemolysis and hyperbilirubinaemia in newborn in-
fants. Exchange transfusion (1,2) and phototherapy (3–5)
have for decades been the principal therapies to reduce
hyperbilirubinaemia and prevent kernicterus. Exchange
transfusions are not without risks (morbidity 2.8–5.2%, mor-
tality 0.5–3.3%) (6), and efforts have been made to avoid the
procedure.
Several studies have suggested that intravenous im-
munoglobulin (IVIG) might be a useful tool for treating
Rh and/or AB0 blood type immunization (7–11). Yet, two
recent reviews found the evidence to be limited due to a
paucity of large randomized studies (6,12). Nevertheless,
recommendations for IVIG were included in the recently up-
dated American Academy of Pediatrics (AAP) guidelines for
management of hyperbilirubinaemia in the newborn (13).
In 1998, the use of IVIG to treat certain infants with blood
type incompatibility was introduced in the neonatal inten-
sive care unit (NICU) at Rikshospitalet University Hospital,
Oslo, Norway. Because in our view the published data, albeit
limited, were compelling when viewed in light of the biolog-
ical rationale (14–16), we felt that the state of equipoise
required for a randomized controlled trial was not present.
Following this change in treatment policy, the number of
exchange transfusions performed in our NICU appeared to
decrease substantially.
this field and the experience and case volume in our NICU
suggested that a retrospective chart review performed as a
quality assurance study might yield results of more general
interest. Our main hypothesis was that the use of IVIG had
reduced the need for exchange transfusions in infants with
Rhesus and/or AB0 isoimmunization. A secondary hypoth-
esis was that the duration of phototherapy might be reduced
in these children. As IVIG therapy does not, like exchange
transfusions, remove the circulating antibodies in the infant,
we further hypothesized that the development of anaemia
requiring top-up blood transfusions would be more com-
mon in newborns receiving only IVIG. Finally, we wanted
to study the time course of serum bilirubin levels in the dif-
ferent treatment groups.
PATIENTS AND METHODS
The study was performed in the NICU of the Division of
Paediatrics at Rikshospitalet University Hospital in Oslo,
Norway. We cross-referenced the NICU database (‘Neona-
talprogrammet’) with the general hospital database (‘PIMS’)
and identified 214 patients by diagnoses and procedure
codes. The main inclusion criterion was the presence of
Rh and/or AB0 isoimmunization. In the case of Rh isoim-
munization, a positive direct antiglobulin test (DAT) was
required. In AB0 isoimmunization the DAT test may not
be positive because of the weaker antigenicity of the AB0

1362 
C 2008 The Author(s)/Journal Compilation 
C 2008 Foundation Acta Pædiatrica/Acta Pædiatrica 2008 97, pp. 1362–1365
Huizing et al.
antigens. Therefore, a positive DAT was not required for
these patients. Nineteen patient charts could not be found,
and another 19 infants did not meet the inclusion cri-
teria and were excluded. A total of 176 newborn in-
fants with Rh and/or AB0 isoimmunization born between
1993 and 2003 and treated in our NICU were thus in-
cluded in this study, and data were collected from their
charts.
In addition to a positive DAT in Rh-immunized infants,
our treatment criteria included a total serum bilirubin (TSB)
exceeding, or shortly destined to exceed, the in-house limits
for performing an exchange transfusion. These limits were
described by a graph that started at 70 μmol/L at birth and
increased in a curve to 350 μmol/L at 72 h of life, remaining
at that level thereafter. An IVIG dose of 500 mg/kg, which
could be repeated, was chosen based on the study by Rübo
et al. (8).
Patients were divided into two major cohorts: a historical
control group born between 1993 and 1998 (cohort 1) and
a treatment group born between 1999 and 2003 (cohort 2).
The infants received either phototherapy alone, a combi-
nation of phototherapy and IVIG, phototherapy plus one
or more exchange transfusions, or phototherapy combined
with IVIG plus one or more exchange transfusions.
The primary end point of our study was the number of
patients receiving exchange transfusions in the NICU. The
duration of phototherapy, the need for subsequent top-up
blood transfusions, and the evolution of haemoglobin and
bilirubin values over time formed the secondary end points.
The project was approved through the ethics and data
security procedures mandated by the hospital. The parents
of the patients were informed of the study and gave their
consent on an opt-out basis. Consent was denied for one
infant only.
STATISTICS
Data are presented as means (±SD) or proportions unless
stated otherwise. The baseline characteristics of both co-
horts were compared with the chi-square test. The analysis
of whether IVIG reduces the need for exchange transfu-
sions in infants with Rh or AB0 isoimmunization was done
by a chi-square test as well as a multiple logistic regres-
sion in order to adjust for possible confounders. After log-
transformation, the phototherapy data were analysed with
linear regressions, both univariate and multiple, as well as
in a separate model based on Akaike’s Information Crite-
rion (17). To assess the need for top-up blood transfusion,
several logistic regressions were performed, both univariate
and multiple, adjusted for the duration of phototherapy to
avoid surrogate effects.
The time course of haemoglobin and serum bilirubin lev-
els was analysed using generalized linear mixed models
(GLMM) (18). GLMM can be viewed as an extension and
combination of regression analysis and ANOVA, which in
addition allows for treating single patients as the unit of
observation. The analyses were performed with SPSS 14.0
(Chicago, IL, USA) and R (http://www.r-project.org).

C 2008 The Author(s)/Journal Compilation 
C 2008 Foundation Acta Pædiatrica/Acta Pædiatrica 2008 97, pp. 1362–1365
Intravenous immune globulin reduces the need for exchange transfusions
RESULTS
Baseline characteristics
The cohorts did not differ in baseline characteristics such
as DAT positivity, maternal and infant AB0 blood type
and Rhesus factor (Table 1). There were no differences be-
tween the cohorts regarding mode of delivery, presentation
at delivery, gestational age, growth, birth weight, gender,
smoking, alcohol or drug use during pregnancy (results not
shown).
Need for exchange transfusion
A large, significant difference was found between cohorts
1 and 2 as regards the need for exchange transfusions (chi-
square test; p < 0.001). An exchange transfusion was needed
for 49% (39/80) of the infants in cohort 1 whereas this was
only the case for 12% (12/96) of the newborns in cohort 2.
The difference was still significant after adjusting for AB0
and Rh incompatibility in a multiple logistic regression (p <
0.001, OR 0.11, 95% CI 0.046–0.26).
Duration of phototherapy
Children receiving either exchange transfusion or IVIG
needed significantly more phototherapy than children who
received neither of them (p = 0.0090 and p = 0.027, respec-
tively). However, compared to the 39 h median duration
of phototherapy in our entire study population, the added
1.3 h is of no real importance in everyday medicine.
We found no difference in the duration of phototherapy
between infants who received IVIG and those who were
treated with exchange transfusion (p = 0.74). In the year
2000, a new phototherapy unit was introduced in the hospi-
tal, but an independent samples t-test showed that this did
not affect the results (p = 0.77).
Top-up blood transfusion
Infants with Rh incompatibility had a higher risk of needing
a top-up blood transfusion than those with AB0 incompat-
ibility. (OR 7.3, 95% CI 1.8–29.3, p = 0.0050 vs. OR 0.14,
95% CI 0.038–0.49, p = 0.0020). The apparent lack of dif-
ference between those treated with IVIG versus exchange
transfusion (p = 0.14) may be due to insufficient power, as
the performed analysis is not a proper equality calculation
(19,20).
The major predictive factor for needing a top-up blood
transfusion was the duration of phototherapy (p < 0.001).
Infants who needed phototherapy for > 80 h (i.e. >2 times
the median duration of phototherapy in our study) had
significantly greater need for such transfusions than those
treated for a shorter period (OR 30.4, 95% CI 7.3–146.7,
p < 0.001).
Evolution of haemoglobin values over time
Infants treated with IVIG developed lower Hb values than
those without (p = 0.0042). Children with AB0 incompat-
ibility had higher Hb levels over time than infants with Rh
incompatibility (p = 0.0080). Cohort and exchange transfu-
sion did not significantly impact the change of haemoglobin
over time.
1363
Intravenous immune globulin reduces the need for exchange transfusions Huizing et al.

Table 1 Baseline characteristics. Chi-square or t-tests for difference between cohorts

Cohort 1∗ Cohort 2∗ p-value
N = 80 N = 96

Maternal AB0 blood type A 14 14 0.94

B 4 4
AB 5 7
0 56 70
Missing 1 1
Maternal rhesus factor (D) − 39 34 0.07
+ 41 62
Jaundice in previous children No 43 55 0.86
Yes 26 30
Not known 11 11
DAT − 15 19 0.97
+ 59 76
Missing 6 1
Infant AB0 blood type A 53 60 0.73
B 12 18
AB 3 3
0 7 13
Missing 5 2
Infant Rhesus factor (D) − 5 6 0.99
+ 75 89
Missing 0 1
AB0 compatibility Comp.† 19 30 0.28
Incomp‡ 55 63
Missing 6 3
Rh compatibility Comp. 44 65 0.12
Incomp. 36 30
Missing 0 1
∗
Cohort 1: born 1993–1998, cohort 2: born 1999–2003. SGA = small for gestational age; AGA = appropriate for gestational age; LGA = large for gestational
age. † Comp. = compatible; ‡ Incomp. = incompatible.

Time course of serum bilirubin levels
AB0 incompatibility (p = 0.034) and Rh incompatibility
(p = 0.0010) were statistically significant explanatory vari-
ables, whereas cohort, IVIG and exchange transfusion were
not. A subgroup analysis comparing the group receiving only
IVIG with the group receiving only exchange transfusion did
not show any statistically significant difference (p = 0.17).
DISCUSSION
A significant reduction in the number of exchange transfu-
sions occurred after the introduction of IVIG as a treatment
for infants with AB0 or Rh immunization in the NICU of
Rikshospitalet University Hospital in Oslo, Norway. Pre-
vious studies had indicated that administration of IVIG in
neonates with AB0 or Rh isoimmunization reduces the need
for exchange transfusions, but the total number of infants
included in those studies was limited (306 infants in five
studies) (7–11). Our results provide further support for this
concept and confirm that IVIG is a good treatment for blood
type incompatibility in neonates.
As would be expected in a retrospective study, some chal-
lenges were encountered in the analyses. Thus, patients born
during the first months after IVIG was introduced were
not always treated according to the new treatment proto-
col. If these infants had been treated according to protocol,
we probably would have had more cases in our IVIG-only
group. Thus, our results most likely underestimate the real
effect of IVIG.
It has been hypothesized that infants with blood type in-
compatibility treated with IVIG have a greater risk of devel-
oping late anaemia and needing top-up blood transfusions
(8,10). While Alcock and Liley (6) found no significant dif-
ference, Gottstein and Cooke’s analysis (12) suggested an
increased need for top-up blood transfusions in infants who
had received IVIG relative to those treated with exchange
transfusion (RR 8.0, 95% CI 1.0–62.2). We found no sta-
tistically significant difference in the need for top-up blood
transfusions between infants treated with these two modal-
ities. However, this may be due to a lack of power. As the
infants included in this study were often transferred to lo-
cal hospitals after they had been stabilized, our in-house
data on late top-up blood transfusions might be incomplete.
Therefore, parents were asked by mail questionnaire about
blood transfusions received elsewhere. Some parents may
not have remembered a transfusion or returned the ques-
tionnaire. Lack of recollection may be more likely in the his-
torical cohort, but would have skewed the data in disfavour

1364 
C 2008 The Author(s)/Journal Compilation 
C 2008 Foundation Acta Pædiatrica/Acta Pædiatrica 2008 97, pp. 1362–1365
Huizing et al.
of IVIG. Thus, the evidence on this point is still equivocal,
and further research is needed before conclusions can be
drawn.
The greater need for top-up blood transfusions in infants
with Rh versus AB0 incompatibility shown in our study is
compatible with the biological differences between these
two conditions. Neonatal erythrocytes compared to adult
contain fewer A or B antigenic sites, and relatively few an-
tibodies bind to the cells. Also, because A and B substance
are present in other tissues, anti-A and anti-B antibodies
disappear from the blood of newborns within the first days
of life (21). Conversely, the Rh antigen is only found on
erythrocytes and the Rh antibody will remain attached to
these cells until they are destroyed by the reticuloendothe-
lial system. Thus, a greater destruction of red cells would be
expected in Rh-immunized versus AB0-immunized infants.
A point of interest is the strong positive relationship we
found between the duration of phototherapy and the need
for top-up transfusions (p < 0.001). This probably reflects a
greater severity of disease with more pronounced haemol-
ysis. In our population, infants who needed phototherapy
for more than twice the median duration in the entire
study group were the most likely candidates to need such
transfusions.
In conclusion, our data provide further evidence in favour
of IVIG treatment in lieu of exchange transfusion as the
primary treatment modality in newborns with AB0 or Rh
isoimmunization. IVIG decreases the need for exchange
transfusion without increasing the duration of phototherapy.
Infants who were Rh immunized and/or needed extended
phototherapy should be watched closely for development of
transfusion-requiring anaemia.
ACKNOWLEDGEMENTS
We are especially grateful to Ishtiaq Khushi for develop-
ing our data recording program. For retrieving the patient
files we would like to thank Turid Haugen and the staff
of the archives of the Rikshospitalet University Hospital,
Oslo, Norway. Finally, Professor Carlos E. Blanco from the
‘Academisch Ziekenhuis Maastricht’, the University hospi-
tal of Maastricht, the Netherlands is thanked for valuable
discussions.
References
1. Wallerstein H. Treatment of severe erythroblastosis by
simultaneous removal and replacement of blood of the
newborn. Science 1946; 103: 583–4.
2. Diamond LK, Allen FH Jr., Thomas WO Jr. Erythroblastosis
fetalis: VII. Treatment with exchange transfusion. N Engl J
Med 1951; 244: 39–49.

C 2008 The Author(s)/Journal Compilation 
C 2008 Foundation Acta Pædiatrica/Acta Pædiatrica 2008 97, pp. 1362–1365
Intravenous immune globulin reduces the need for exchange transfusions
3. Cremer RJ. Influence of light on the hyperbilirubinaemia of
infants. Lancet 1958: 1: 1094–7.
4. Obes-Polleri J, Hill WS. La fototerapia en las icterı́cias del
recien nacido. Rev Chil Ped 1964; 25: 638.
5. Lucey JF, Ferreiro M, Hewitt J. Prevention of
hyperbilirubinemia of prematurity by phototherapy. Pediatrics
1968; 40: 1047–54.
6. Alcock GS, Liley H. Immunoglobulin infusion for isoimmune
haemolytic jaundice in neonates. Cochrane Database Syst Rev
2002; 3: CD003313.
7. Sato K, Hara T, Kondo T, Iwao H, Honda S, Ueda K.
High-dose intravenous gammaglobulin therapy for neonatal
immune haemolytic jaundice due to blood type
incompatibility. Acta Paediatr Scand 1991; 80: 163–6.
8. Rubo J, Albrecht K, Lasch P, Laufkotter E, Leititis J, Marsan
D, et al. High-dose intravenous immune globulin therapy for
hyperbilirubinemia caused by Rh haemolytic disease. J Pediatr
1992; 121: 93–9.
9. Daolu T, Ovali F, Samanci N, Bengisu E. High-dose
intravenous immunoglobulin therapy for rhesus haemolytic
disease. J Int Med Res 1995; 23: 264–71.
10. Alpay F, Sarici SU, Okutan V, Erdem G, Ozcan O, Gokcay E.
High-dose intravenous immunoglobulin therapy in neonatal
immune haemolytic jaundice. Acta Paediatr 1999; 88: 216–9.
11. Miqdad AM, Abdelbasit OB, Shaheed MM, Seidahmed MZ,
Abomelha AM, Arcala OP. Intravenous immunoglobulin G
(IVIG) therapy for significant hyperbilirubinaemia in AB0
hemolytic disease of the newborn. J Matern Fetal Neonatal
Med 2004; 16: 163–6.
12. Gottstein R, Cooke RW. Systematic review of intravenous
immunoglobulin in haemolytic disease of the newborn. Arch
Dis Fetal Neonatal Ed 2003; 88: F6–10.
13. American Academy of Pediatrics Subcommittee on
Hyperbilirubinemia. Management of hyperbilirubinemia in the
newborn infant 35 or more weeks of gestation. Pediatrics
2004; 114: 297–316.
14. Imbach P, Barandun S, d’Apuzzo V, Baumgartner C, Hirt A,
Morell A, et al. High-dose intravenous gammaglobulin for
idiopathic thrombocytopenic purpura in childhood. Lancet
1981; 1:1228–31.
15. Berlin G, Selbing A, Ryden G. Rhesus haemolytic disease
treated with high-dose intravenous immunoglobulin. Lancet
1988; 325: 1153.
16. De la Camara C, Arrieta R, Gonzalez A, Iglesias E, Omenaca
F. Intravenous immunoglobulin as the sole prenatal treatment
for severe Rh immunization. N Engl J Med 1988; 318: 519.
17. Akaike H. A new look at the statistical model identification.
IEEE Trans Aut Ctr 1974; 19: 716–23.
18. McCulloch CE, Searle RS. Generalized, linear and mixed
models. New York: John Wiley & Sons, 2001.
19. Laster LL, Johnson MF. Non-inferiority trials: the ‘at least as
good as’ criterion. Statistics Med 2003; 22: 187–200.
20. Snappin SM. Commentary: noninferiority trials. Curr Control
Trials Cardiovasc Med 2000; 1: 19–21.
21. Zipursky A, Bowman JM. Isoimmune hemolytic diseases. In:
Nathan DG, Oski FA, editors. Hematology of infancy and
childhood. 4th ed. Philadelphia: W.B. Saunders Company,
1993: 44–73.
1365