EXCHANGE BLOOD TRANSFUSION COMPARED WITH SIMPLE TRANSFUSION
FOR FIRST OVERT STROKE IS ASSOCIATED WITH A LOWER RISK OF
SUBSEQUENT STROKE: A RETROSPECTIVE COHORT STUDY OF 137 CHILDREN
WITH SICKLE CELL ANEMIA
MONICA L. HULBERT, MD, DOUGLAS J. SCOTHORN, MD, PHD, JULIE A. PANEPINTO, MD, MSPH, J. PAUL SCOTT, MD,
GEORGE R. BUCHANAN, MD, SHARADA SARNAIK, MD, ROBERT FALLON, MD, PHD, JEN-YIH CHU, MD, WINFRED WANG, MD,
JAMES F. CASELLA, MD, LINDA RESAR, MD, BRIAN BERMAN, MD, THOMAS ADAMKIEWICZ, MD, LEWIS L. HSU, MD, PHD,
KIMBERLY SMITH-WHITLEY, MD, DONALD MAHONEY, MD, GERALD WOODS, MD, MASAYO WATANABE, MD, AND
MICHAEL R. DEBAUN, MD, MPH
A retrospective cohort study of children with sickle cell anemia (SCA) and strokes was used to test the hypothesis that
exchange transfusion at the time of stroke presentation more effectively prevents second strokes than does simple transfusion.
Children receiving simple transfusion had a 5-fold greater relative risk (95% confidence interval ? 1.3 to 18.6) of second stroke
than those receiving exchange transfusion. (J Pediatr 2006;149:710-2)
In the only study published to date addressing initial treatment for acute stroke, Ohene-
Frempong1described a single-institution series of 34 patients with SCD and strokes. During
the first 48 hours after symptom onset, 57% of these patients received either simple or
exchange transfusion therapy, but treatment-specific outcomes were not examined.1We
is a more effective initial treatment for stroke than simple transfusion.
he optimal strategy for the initial treatment of stroke in children with sickle cell
disease (SCD) has not been identified. The options include simple or exchange
transfusion; however, the potential benefits and risks of these therapies are unknown.
The eligibility criteria, data collection, and study methods have been described
previously.2Patients were identified at 14 medical centers, and their records were reviewed
between 1998 and 2000. Institutional review board approval was obtained according to
guidelines at the participating centers. Patients were included in the analysis if they had
sickle cell anemia (SCA) and had undergone at least 5 years of chronic blood transfusion
therapy after stroke. Exclusion criteria were lack of documentation of stroke, lack of at
least 5 years of transfusion therapy documented in medical records, transfusion therapy
that did not occur primarily at the participating institution, and transfusion therapy
occurring at more than 6-week intervals at any time during follow-up. First and subse-
quent strokes were defined as acute neurologic symptoms and signs lasting longer than 24
hours or, for symptoms lasting less than 24 hours, with an imaging study demonstrating
acute ischemia. Exchange transfusion was defined as either manual exchange transfusion
or automated erythrocytapheresis. Definitions of antecedent or concurrent medical events
were as given previously.2Participating physicians were asked to describe the chronic
transfusion practices (simple transfusion or exchange transfusion) at their institution
Sickle cell anemia
Sickle cell disease
See editorial, p 595, and
related article, p 707
From Washington University School of
Medicine, St. Louis, MO; The University of
Texas Southwestern Medical Center, Dal-
las, TX; Medical College of Wisconsin, Mil-
waukee, WI; Children’s Hospital of Michi-
Children’s Hospital, St. Louis, MO; St. Jude
Children’s Research Hospital, Memphis,
TN; Johns Hopkins University School of
Medicine, Baltimore, MD; Rainbow Babies
& Children’s Hospital, Cleveland, OH;
Emory University School of Medicine, At-
lanta, GA; Children’s Hospital of Philadel-
phia, Philadelphia, PA; Baylor College of
Medicine, Houston, TX; and University of
Missouri–Kansas City, Kansas City, MO.
Supported by the Doris Duke and Robert
Wood Johnson Foundations and the Na-
tional Institutes of Health (training grant
Submitted for publication Aug 18, 2005;
last revision received Apr 20, 2006; ac-
cepted Jun 16, 2006.
Reprint requests: Michael R. DeBaun, MD,
MPH, Washington University School of
Medicine, 4444 Forest Park Blvd, CB
0022-3476/$ - see front matter
Copyright © 2006 Mosby Inc. All rights
during the study period (before 2000). Patients from institu-
tions with a consistent practice of either simple transfusion or
exchange transfusion on a chronic basis were included in the
analysis; those from institutions with a variety of practices
The measure of association was relative risk, because the
study was a retrospective cohort analysis. Fisher’s exact test
and ?2analysis were performed for categorical data. Stroke-
free (event-free) survival and overall survival curves were es-
timated using the Kaplan-Meier method. The log-rank func-
tion test was used to compare stroke-free survival time
distributions for patients receiving initial exchange transfu-
sion and simple transfusion. The level of significance was
specified as P ? .05. Data analyses were performed with SPSS
version 12.0 (SPSS Inc, Chicago, IL).
The participating institutions identified 164 children
with SCA, a history of stroke, and at least 5 years of blood
transfusion therapy. Of these, 137 met all inclusion criteria;
46% were male. The mean age at initial stroke was 6.3 years
(range, 1.4 to 14 years), and the mean follow-up was 10.1
years (range, 5 to 24 years). Initial strokes occurred between
May 1972 and March 1995. In this cohort, 31 of the 137
patients (23%) had a recurrent stroke while receiving chronic
blood transfusion therapy (2.2 events per 100 patient years).2
Signs and Symptoms of Stroke
Motor deficits were the most common presenting sign
of initial stroke in 125 children for whom symptom data were
available. Of these 125 patients, 107 (86%) experienced weak-
ness/paresis, 31 (25%) experienced seizures, 28 (22%) had
dysarthria/aphasia, 24 (19%) reported headache, 14 (11%)
had sensory deficits, and 7 (5%) exhibited altered conscious-
ness (unresponsive or responsive only to painful stimuli).
Time to Presentation and Initial Treatment
Most patients came to medical attention within 1 day
after onset of stroke symptoms. Of the 124 patients in whom
duration of symptoms was recorded, 65% presented within 1
day, 20% within 1 to 3 days, and 15% after more than 3 days
after symptom onset (range, 4 days to 3 months).
Exchange blood transfusion was the most common
initial treatment for first stroke, regardless of the duration of
symptoms before presentation (Table). The proportion of
patients receiving exchange transfusion at first stroke presen-
tation increased during the study period: 0 of the 4 presenting
before 1980, 7 of the 10 presenting between 1980 and 1985,
23 of the 33 presenting between 1985 and 1990, and 37 of the
43 presenting after 1990 (n ? 90; Pearson’s ?2test; P ? .012).
Exchange Transfusion and Risk of Second Stroke
The method of initial blood transfusion at presentation
with stroke symptoms was associated with the risk of recur-
rent stroke. Initial treatment information was available for 52
patients who presented within 24 hours of stroke symptom
onset. Recurrent strokes occurred in 57% (8/14) of patients
treated with simple transfusion, compared with 21% (8/38) of
those treated with exchange transfusion. Simple transfusion
was associated with a 5-fold greater incidence of recurrent
stroke compared with exchange transfusion (relative risk [RR]
? 5.0; 95% confidence interval [CI] ? 1.3 to 18.6) (Figure).
The proportion of patients without a medical antecedent
event (ie, fever, acute anemia, acute chest syndrome, hyper-
tension, or exchange transfusion)2was similar in the simple
and exchange transfusion groups (79% and 74%, respectively;
2-sided Fisher’s exact test; P ? 1).
Among patients without a medical antecedent event
who presented within 24 hours of symptom onset, those who
received simple blood transfusion were 8 times more likely to
experience subsequent stroke (RR ? 8.0; 95% CI ? 1.7 to
38.8) compared with patients treated with exchange transfu-
sion (8 of 11 patients vs 7 of 28 patients). There were not
enough patients presenting with 1 to 3 days or greater than 3
days of symptoms to permit similar subset analyses in these
categories. The incidence of recurrent strokes for the entire
cohort did not decrease significantly by time period of first
stroke (before 1980, 1980 to 1984, 1985 to 1989, 1990 and
after; Pearson’s ?2test; P ? .33).
We further evaluated the type of chronic blood transfusion
therapy practiced at each institution during the study period.
Seven sites, with a total of 65 patients, routinely used simple
transfusions, and 2 sites with 24 patients routinely used manual
exchange transfusions. Of the patients who presented within 24
hours and initially received exchange transfusion (n ? 38), 18
received chronic simple transfusion, 11 received chronic ex-
change transfusion, and 9 received an unknown type of chronic
transfusion. Among the patients who received exchange trans-
fusion acutely, those who also received exchange transfusion on
a chronic basis experienced fewer second strokes (0/11 vs 7/18
patients who received chronic simple transfusion; Fisher’s exact
test; P ? .026). The RR of second stroke was decreased for
patients who received both acute and chronic exchange transfu-
sions (RR ? 0.61; 95% CI ? 0.42 to 0.88).
Table. Initial treatment of first overt stroke in
children with SCA did not vary with duration of
symptoms before presentation to medical attention
(n ? 90; Pearson’s ?2test; P ? .31)
Duration of symptoms
< 1 day
(n ? 52)(n ? 22)
1 to 3 days
> 3 days
(n ? 16)
00 1 (6%)
One patient was not treated when the stroke was diagnosed after 3 months of symptoms;
this patient was excluded from subsequent analyses.
Exchange Blood Transfusion Compared With Simple Transfusion For First Overt Stroke Is Associated With
A Lower Risk Of Subsequent Stroke: A Retrospective Cohort Study Of 137 Children With Sickle Cell Anemia711
DISCUSSION Download full-text
We have demonstrated that the proportion of children
treated with exchange transfusion for acute strokes has increased
since the procedure was first described by Lusher et al in 19763
and that this treatment is now the most common method of
initial therapy for acute stroke. In addition, we have provided
preliminary evidence indicating that initial exchange transfusion
is associated with a decreased risk of second strokes.
Given the retrospective nature of this study, we can only
postulate why exchange transfusion may decrease the risk of
subsequent stroke. Perhaps initial treatment with exchange
transfusion limits the extent of acute cerebral ischemic injury,
thereby putting the patient at lower risk of recurrent stroke. This
hypothesis is consistent with the observation by Pegelow et al4
indirectly demonstrating that the size of a cerebral infarct may
influence the risk of subsequent stroke. Patients with overt, as
opposed to silent, strokes have larger infarcts and greater risk of
recurrent infarcts.4In patients with overt strokes, the rate of
second overt stroke in patients not receiving chronic blood trans-
fusion therapy was 7.4 events per 100 patient years,5whereas the
rate of overt stroke for those with preexisting silent strokes was
1.03 events per 100 patient years.4
These results also indicate a possible benefit of ex-
change transfusion as the preferred method of chronic blood
transfusion therapy. These results must be interpreted cau-
tiously, because the data were not collected contemporane-
ously and are available for only a subset of the cohort. Fur-
thermore, there is no firm biological basis for this association.
Only a prospective study evaluating the type of chronic blood
transfusion therapy among patients initially treated with ex-
change transfusion at the time of acute stroke would be able
to address the contribution of chronic exchange transfusions
to secondary prevention of stroke.
Our results also indicate that the time to presentation to
medical attention after onset of neurologic symptoms is delayed
by more than 1 day in 1/3 of patients. Recently, Katz et al6
identified the need for improved education for children with
SCD and their families regarding the risk of stroke. Lack of
awareness that SCD is associated with increased risk of stroke
was reported by 64% of caregivers, and 46% could not name any
symptoms of stroke.7These findings and our data suggest that
current methods used by pediatricians and pediatric hematolo-
gists to educate families about stroke risk, symptoms, and the
need for treatment may be ineffective or inadequate.
This study’s limitations are primarily related to its retro-
spective design. A cause-and-effect relationship between the
type of treatment and subsequent stroke cannot be determined.
We cannot exclude the possibility that children receiving ex-
change transfusion had a smaller volume of ischemic tissue and
possibly a lower risk of second stroke due to more limited
cerebral injury. Multivariate analysis of second stroke risk was
not possible, because of the small number of second strokes. We
were not able to evaluate whether resolution of stroke symptoms
was affected by the method of initial treatment. The reduced risk
part to a cohort effect, because care of children who have sus-
tained stroke has improved over time. However, only children
who received blood transfusion therapy at least every 6 weeks
were included, to ensure that the patients in this cohort received
uniform care. We detected no significant decrease in the occur-
rence of subsequent stroke when evaluated by time period of first
stroke, further supporting the role of the initial transfusion pro-
cedure, as opposed to improvements in ongoing care, as being
associated with a lower risk of second stroke.
Results from this study indicate that exchange blood
transfusion is the most common form of initial therapy at the
time of presentation with stroke symptoms, and that it may
prevent recurrent stroke in patients with SCA.
We thank Dr. Wanda Shurney and Dr. Kwaku Ohene-Frem-
pong for assisting with the data collection.
and therapeutic considerations. Semin Hematol 1991;28:213-9.
therapy for at least five years after initial stroke. J Pediatr 2002;140:348-54.
Lusher JM, Haghighat H, Khalifa AS. A prophylactic transfusion pro-
gram for children with sickle cell anemia complicated by CNS infarction.
Am J Hematol 1976;1:265-73.
Pegelow CH, Macklin EA, Moser FG, Wang WC, Bello JA, Miller ST,
et al. Longitudinal changes in brain magnetic resonance imaging findings in
children with sickle cell disease. Blood 2002;99:3014-8.
Powars D, Wilson B, Imbus C, Pegelow C, Allen J. The natural history
of stroke in sickle cell disease. Am J Med 1978;65:461-71.
Katz ML, Smith-Whitley K, Ruzek SB, Ohene-Frempong K. Knowledge
of stroke risk, signs of stroke, and the need for stroke education among children
with sickle cell disease and their caregivers. Ethnicity and Health 2002;7:115-23.
Ohene-Frempong K. Stroke in sickle cell disease: demographic, clinical,
Figure. Initial simple transfusion for first overt stroke in children with
SCA who presented within 24 hours of symptom onset is associated with
increased risk of recurrent stroke compared with exchange transfusion.
Initial exchange transfusion, n ? 38; initial simple transfusion, n ? 14;
RR ? 5.0; 95% CI ? 1.3 to 18.6; log-rank test; P ? .02. All children
received scheduled chronic blood transfusion therapy for at least 5 years
after the first stroke. Solid line, simple transfusion; broken line, exchange
transfusion; dashes, censored events.
712 Hulbert et al The Journal of Pediatrics • November 2006