Hydroxyurea_for_Children_with_1.pdf

The new england
journal of medicine

n engl j med 380;2 nejm.org January 10, 2019 121

established in 1812 January 10, 2019 vol. 380 no. 2

From Centre Hospitalier Monkole, Kin-
shasa, Democratic Republic of Congo
(L.T.); the Department of Medicine, Uni-
versity Health Network and Mt. Sinai
Hospital, and the University of Toronto,
Toronto (G.T.); the Kenya Medical Re-
search Institute (KEMRI)–Wellcome Trust
Research Program, Kilifi, Kenya (T.N.W.);
the Department of Medicine, Imperial
College London, London (T.N.W.); Hos-
pital Pediátrico David Bernardino, Luanda,
Angola (B.S.); Mbale Clinical Research
Institute and Mbale Regional Referral
and Teaching Hospital–Busitema Univer-
sity, Mbale, Uganda (P.O.-O.); the Divi-
sion of Hematology, Department of Pedi-
atrics, Cincinnati Children’s Hospital (A.L.,
S.E.S., T.S.L., P.T.M., R.E.W.), University
of Cincinnati College of Medicine (A.L.,
P.T.M., R.E.W.), and the Global Health
Center, Cincinnati Children’s Hospital
Medical Center (S.E.S., P.T.M., R.E.W.),
Cincinnati; and Cohen Children’s Medical
Center, New Hyde Park, and the Zucker
School of Medicine at Hofstra/Northwell,
Hempstead — both in New York (B.A.).
Address reprint requests to Dr. Ware at
Cincinnati Children’s Hospital Medical
Center, 3333 Burnet Ave., Cincinnati, OH
45229, or at russell . ware@ cchmc . org.

*A complete list of the REACH Investiga-
tors is provided in the Supplementary
Appendix, available at NEJM.org.

This article was published on December 1,
2018, at NEJM.org.

N Engl J Med 2019;380:121-31.
DOI: 10.1056/NEJMoa1813598
Copyright © 2018 Massachusetts Medical Society.

BACKGROUND
Hydroxyurea is an effective treatment for sickle cell anemia, but few studies have been
conducted in sub-Saharan Africa, where the burden is greatest. Coexisting conditions
such as malnutrition and malaria may affect the feasibility, safety, and benefits of hy-
droxyurea in low-resource settings.
METHODS
We enrolled children 1 to 10 years of age with sickle cell anemia in four sub-Saharan
countries. Children received hydroxyurea at a dose of 15 to 20 mg per kilogram of body
weight per day for 6 months, followed by dose escalation. The end points assessed feasibil-
ity (enrollment, retention, and adherence), safety (dose levels, toxic effects, and malaria),
and benefits (laboratory variables, sickle cell–related events, transfusions, and survival).
RESULTS
A total of 635 children were fully enrolled; 606 children completed screening and began
receiving hydroxyurea at a mean (±SD) dose of 17.5±1.8 mg per kilogram per day. The
retention rate was 94.2% at 3 years of treatment. Hydroxyurea therapy led to significant
increases in both the hemoglobin and fetal hemoglobin levels. Dose-limiting toxic events
regarding laboratory variables occurred in 5.1% of the participants, which was below the
protocol-specified threshold for safety. During the treatment phase, 20.6 dose-limiting
toxic effects per 100 patient-years occurred, as compared with 20.7 events per 100 patient-
years before treatment. As compared with the pretreatment period, the rates of clinical
adverse events decreased with hydroxyurea use, including rates of vaso-occlusive pain
(98.3 vs. 44.6 events per 100 patient-years; incidence rate ratio, 0.45; 95% confidence in-
terval [CI], 0.37 to 0.56), nonmalaria infection (142.5 vs. 90.0 events per 100 patient-years;
incidence rate ratio, 0.62; 95% CI, 0.53 to 0.72), malaria (46.9 vs. 22.9 events per 100
patient-years; incidence rate ratio, 0.49; 95% CI, 0.37 to 0.66), transfusion (43.3 vs. 14.2
events per 100 patient-years; incidence rate ratio, 0.33; 95% CI, 0.23 to 0.47), and death
(3.6 vs. 1.1 deaths per 100 patient-years; incidence rate ratio, 0.30; 95% CI, 0.10 to 0.88).
CONCLUSIONS
Hydroxyurea treatment was feasible and safe in children with sickle cell anemia living in
sub-Saharan Africa. Hydroxyurea use reduced the incidence of vaso-occlusive events, in-
fections, malaria, transfusions, and death, which supports the need for wider access to
treatment. (Funded by the National Heart, Lung, and Blood Institute and others; REACH
ClinicalTrials.gov number, NCT01966731.)

a b s t r a c t

Hydroxyurea for Children with Sickle Cell Anemia
in Sub-Saharan Africa

Léon Tshilolo, M.D., Ph.D., George Tomlinson, Ph.D., Thomas N. Williams, M.D., Ph.D., Brígida Santos, M.D.,
Peter Olupot-Olupot, M.D., Ph.D., Adam Lane, Ph.D., Banu Aygun, M.D., Susan E. Stuber, M.A., Teresa S. Latham, M.A.,

Patrick T. McGann, M.D., and Russell E. Ware, M.D., Ph.D., for the REACH Investigators*

n engl j med 380;2 nejm.org January 10, 2019122

T h e n e w e n g l a n d j o u r n a l o f m e d i c i n e

Sickle hemoglobinopathies are com-mon and life-threatening genetic dis s. Homozygous hemoglobin S (HbSS) is the
most severe genotype, and together with hemo-
globin Sβ0 thalassemia, is called sickle cell ane-
mia.1 On deoxygenation, erythrocytes become
sickle-shaped, rigid, adhesive, and prone to lysis;
blood flow is blocked within small vessels, lead-
ing to ischemic tissue injury.1 In the United States,
approximately 100,000 persons are affected,2 most
of whom have numerous acute and chronic medi-
cal complications that lead to poor quality of life
and early death.1,3 On a global scale, the incidence
of sickle hemoglobinopathies is greatest in sub-
Saharan Africa, with more than 300,000 babies
with sickle cell disease born annually, represent-
ing approximately 1% of births in the region.4

In high-resource settings such as the United
States and Europe, as well as Jamaica and other
Caribbean settings, the early identification of chil-
dren with sickle cell anemia by means of neonatal
screening allows for comprehensive care that in-
cludes simple, effective, and lifesaving interven-
tions with penicillin prophylaxis, pneumococcal
immunizations, and caregiver education.5-7 The
advent of routine transcranial Doppler screening
to identify children at risk for stroke, along with
access to safe erythrocyte transfusions, has further
contributed to marked decreases in morbidity.8 To
date, however, very few settings in sub-Saharan
Africa have established screening programs for
sickle cell anemia, and specialized treatment is
available at only a few large, urban centers.

Hydroxyurea was first shown to induce fetal
hemoglobin production more than 30 years ago9
and is now a Food and Drug Administration–
approved treatment for sickle cell anemia in both
children (Siklos, Addmedica) and adults (Hydrea
and Droxia, Bristol-Myers Squibb). By means of the
induction of fetal hemoglobin and other beneficial
changes, including mild myelosuppression, hy-
droxy urea therapy has been shown to have clinical
efficacy in reducing the incidence of acute vaso-
occlusive events,10,11 ameliorating chronic organ
damage,12 and prolonging survival.13,14 Evidence-
based guidelines from the National Heart, Lung,
and Blood Institute recommend offering hydroxy-
urea treatment to persons with sickle cell anemia
as early as 9 months of age.15

Whether hydroxyurea will be safe and effec-
tive in Africa is unclear. Coexisting conditions
such as malaria, other infectious diseases that
are endemic to the area, and malnutrition may

increase the incidence of toxic effects and limit
treatment responses. We conducted an interna-
tional trial, Realizing Effectiveness across Conti-
nents with Hydroxyurea (REACH), to investigate
the feasibility, safety, and benefits of hydroxyurea
treatment for children with sickle cell anemia liv-
ing in sub-Saharan Africa.

M e t h o d s

Trial Design

We designed this phase 1–2, open-label, interna-
tional trial to assess the feasibility, safety, and
benefits of hydroxyurea treatment in young chil-
dren with sickle cell anemia living in sub-Saharan
Africa. The two-stage trial design has been de-
scribed previously16 and included a built-in pause
in enrollment to ensure the safety of the initial
dose level (Fig. S1 in the Supplementary Appen-
dix, available with the full text of this article at
NEJM.org). Children 1 to 10 years of age were
recruited at four clinical trial sites in sub-Saharan
Africa (Hospital Pediátrico David Bernardino in
Luanda, Angola; Centre Hospitalier Monkole in
Kinshasa, Democratic Republic of Congo; Kilifi
District Hospital in Kilifi, Kenya; and Mbale Re-
gional Referral Hospital in Mbale, Uganda), with
a goal of treating 150 participants per site with
once-daily oral hydroxyurea. At each site, enroll-
ment was paused to allow for the evaluation of
the 3-month incidence of hematologic toxic ef-
fects among the first 53 participants enrolled.16

The protocol (available at NEJM.org) was de-
signed by the authors and approved by all appro-
priate institutional review boards or ethics com-
mittees, as well as by national regulatory groups.
A parent or legal guardian provided written in-
formed consent for each participant to enroll and
begin the screening process. A full list of the in-
vestigative teams, including the physicians, nurs-
es, pharmacists, and laboratory and data person-
nel who received specific protocol training, is
provided in the Supplementary Appendix. Local
trial teams collected the data, which were mon-
itored and analyzed by the data management
center. The first draft of the manuscript was
written by the first and last authors, with contri-
butions by all the authors. All the authors made
the decision to submit the manuscript for publi-
cation. The sponsors had no oversight or involve-
ment in data collection and analysis or in the
writing and submission of the manuscript. The
last author vouches for the accuracy and com-

A Quick Take
is available at

NEJM.org

n engl j med 380;2 nejm.org January 10, 2019 123

H y drox y ur e a for Sick le Cell A nemi a in A fr ic a

pleteness of the data and for the fidelity of the
trial to the protocol.

Drug Treatment and Dose Escalation

Hydroxyurea capsules were donated by Bristol-
Myers Squibb, which had no role in the trial de-
sign or conduct, the data collection or analysis,
or the manuscript preparation or review. After a
2-month screening period that was used to collect
pretreatment clinical and laboratory data, the
starting dose of hydroxyurea was 15 to 20 mg per
kilogram of body weight per day. After 6 months
of treatment, the hydroxyurea dose was esca-
lated by 2.5 to 5.0 mg per kilogram per day every
2 months on the basis of peripheral-blood counts
to determine a maximum tolerated dose, which
was defined as a stable daily dose that caused mild
bone marrow suppression without toxic effects
— typically, an absolute neutrophil count of less
than 4000 per cubic millimeter. To ensure the
safety of the participants, the dose level was
monitored at each visit by trial staff, who entered
the results of the complete blood count and re-
ticulocyte count into an interactive online calcu-
lator tool before the drug was dispensed.

End Points

The primary safety end point was hematologic
dose-limiting toxic effects in the first 3 months
of hydroxyurea treatment in the first 133 children
enrolled at each clinical trial site; the expected
and allowable rates of this end point were 20%
and 30%, respectively, on the basis of published
toxicity data, with type I and II error rates of 10%.16
Protocol-specified thresholds for toxic effects in-
volving laboratory variables included a hemoglo-
bin level of less than 4.0 g per deciliter, an abso-
lute neutrophil count of less than 1000 per cubic
millimeter, an absolute reticulocyte count of less
than 80,000 per cubic millimeter unless the hemo-
globin level was more than 7.0 g per deciliter, and
a platelet count of less than 80,000 per cubic mil-
limeter. Additional trial end points included as-
sessments of feasibility (enrollment, retention, and
adherence), safety (dose levels, toxic effects, and
malaria), and benefits (laboratory variables, sickle
cell–related events, transfusions, and survival).

Data Collection and Storage

All the trial data were collected and entered into
a REDCap Internet-based data-capture system.
Separate REDCap environments were developed
in English, Portuguese, and French, which were

the official languages at each clinical site.16 Data
on clinical adverse events of grade 2 or higher were
collected during both the screening (pretreatment)
phase and the treatment phase. Trial monitoring
used a remote system and on-site evaluations. Data
on adverse events were collected and curated, with
review of source documentation whenever avail-
able. Deaths were further evaluated with the use
of the World Health Organization (WHO) verbal
autopsy form that has previously been validated
for children with sickle cell anemia.17

Statistical Analysis

The Simon two-stage design for the primary safety
end point (hematologic dose-limiting toxic effects
in the first 3 months of hydroxyurea treatment)
has been described previously.16 Secondary end
points regarding feasibility, safety, and benefits
were summarized with means and standard de-
viations or percentages, as appropriate. We used
a competing-risk approach to estimate the cumu-
lative incidence of death or withdrawal. Labora-
tory values at baseline and at 12 months were
compared by Student’s t-test. Rates of clinical ad-
verse events during the pretreatment phase and
during the treatment phase were presented as the
number of events per 100 patient-years and com-
pared by the incidence rate ratio with 95% con-
fidence intervals, all of which were calculated
from Poisson regression with the use of general-
ized estimating equations to account for cluster-
ing and overdispersion. A similar approach was
used to show the effect of increasing treatment
duration, with events and time at risk during treat-
ment divided into consecutive 6-month intervals.
There were no adjustments for multiple compari-
sons. All the statistical analyses were performed
with the use of R software, version 3.4.4.

R e s u l t s

Enrollment and Retention

All four trial sites proceeded to the second stage
of the trial and met their enrollment goals. A total
of 635 children had consent provided by a parent
or guardian and entered screening, 606 children
completed screening and began receiving hydroxy-
urea treatment, and 600 children (99.0%) com-
pleted 3 months of the trial treatment (Fig. 1). The
overall retention rate in the trial was 94.2% at
3 years of treatment (Fig. 2). A total of 33 children
(5.4%) withdrew from the trial after treatment
initiation. More than 98% of the trial visits were

n engl j med 380;2 nejm.org January 10, 2019124

T h e n e w e n g l a n d j o u r n a l o f m e d i c i n e

completed, including 91% within the scheduled
visit window.

Hydroxyurea Dose

The initial mean (±SD) dose of hydroxyurea that
was administered was 17.5±1.8 mg per kilogram
per day, which was within the protocol-directed
starting dose range of 15 to 20 mg per kilogram
per day. The initial dose was fixed for 6 months in
all the participants to allow for the assessment of
laboratory and clinical adverse events. Dose esca-
lation began at month 6, and to date, 515 children
(85.0%) have reached a mean maximum tolerated
dose of 22.5±4.9 mg per kilogram per day, with
the dose ranging from 18.9±4.2 mg per kilogram
per day in Angola to 25.3±4.8 mg per kilogram
per day in Uganda. The mean time to reaching the
maximum tolerated dose was 11±4 months, and
the mean overall treatment duration was 29±9
months. The rate of adherence to medication was
assessed at each scheduled visit, and more than
90% of assessments documented no missed dos-
es. A brief drug shortage between February and
April 2016 affected the administration of the
medication at three sites but had no influence on
the primary end point (Table S1 in the Supplemen-
tary Appendix).

Adverse Events

Laboratory monitoring at scheduled visits and at
unscheduled visits for illness identified dose-limit-
ing toxic effects during the screening phase (a total
of 2012 complete blood counts were performed
over a period of 111 patient-years) and during the
treatment phase (a total of 13,589 complete blood
counts were performed over a period of 1469 pa-
tient-years). Hematologic dose-limiting toxic ef-
fects during the first 3 months (the primary safety
end point) occurred in 5.1% of the participants
overall, with the rate at each site being well below
the protocol-specified thresholds for toxic events
regarding laboratory variables. The rates of these
safety events differed across sites, ranging from
0.8% to 8.3% (P = 0.01 by Fisher’s exact test) (Ta-

606 Initiated hydroxyurea treatment

635 Patients had consent provided by a parent
or guardian and underwent screening

29 Were withdrawn during screening
11 Were ineligible
3 Had consent withdrawn
3 Relocated
6 Did not adhere to study regimen
2 Were lost to follow-up
4 Died

600 Completed 3 mo of study treatment

6 Were withdrawn during mo 0–3
1 Was ineligible
2 Did not adhere to study regimen
1 Was lost to follow-up
2 Died

588 Completed 12 mo of study treatment

12 Were withdrawn during mo 3–12
1 Was ineligible
1 Relocated
1 Did not adhere to study regimen
3 Were lost to follow-up
6 Died

415 Completed 24 mo of study treatment
162 Had not completed 24 mo of study

treatment before database lock

11 Were withdrawn during mo 12–24
4 Relocated
1 Did not adhere to study regimen
1 Was lost to follow-up
5 Died

235 Completed 36 mo of study treatment
177 Had not completed 36 mo of study

treatment before database lock

3 Were withdrawn during mo 24–36
1 Did not adhere to study regimen
2 Died

5 Completed 48 mo of study treatment
229 Had not completed 48 mo of study

treatment before database lock

1 Died after 36 mo

Figure 1. Screening, Enrollment, and Follow-up
of the Participants.

The numbers of patients who withdrew from the trial
or died are from the group of patients who completed
treatment up to the respective time point.

n engl j med 380;2 nejm.org January 10, 2019 125

H y drox y ur e a for Sick le Cell A nemi a in A fr ic a

ble S2 in the Supplementary Appendix), and the
random-effects pooled estimate was 4.5% (95%
confidence interval [CI], 2.3 to 8.8). No signifi-
cant differences between the screening and treat-
ment phases were found with regard to the indi-
vidual laboratory toxic-effect variables (Table 1),
but slight differences were noted between sites
(data not shown).

Laboratory Benefits

Laboratory variables at baseline revealed anemia
with expected leukocytosis and reticulocytosis
(Table 2). After 1 year of hydroxyurea treatment,
the participants had significant increases in the
hemoglobin level (increase of 1.0 g per deciliter;
95% CI, 0.8 to 1.0), the mean corpuscular volume
(increase of 13 fl; 95% CI, 12 to 13), and the fetal
hemoglobin level (increase of 12.5%; 95% CI, 11.8
to 13.1). During hydroxyurea treatment, the white-
cell count, absolute neutrophil count, and absolute
reticulocyte count significantly decreased, reflect-
ing the intended mild bone marrow suppression,
and these effects were sustained over time (Ta-
ble 2). Similar results were recorded at each indi-
vidual clinical trial site (data not shown).

Clinical Benefits

Comparison of event rates between the screening
and treatment phases revealed significant reduc-
tions in the incidence of all clinical adverse events
(308.4 vs. 170.7 events per 100 patient-years; inci-
dence rate ratio, 0.54; 95% CI, 0.48 to 0.62) and
serious adverse events (10.8 vs. 4.4 events per
100 patient-years; incidence rate ratio, 0.47; 95% CI,
0.25 to 0.90) during hydroxyurea treatment. The
investigators did not consider any of the adverse
events or serious adverse events to be related to
hydroxyurea treatment.

The overall rate of sickle cell–related events was
significantly reduced (114.5 vs. 53.0 events per 100
patient-years; incidence rate ratio, 0.47; 95% CI,
0.38 to 0.57), and the rates of vaso-occlusive pain
and the acute chest syndrome were both reduced
(Table 1). The rates of infection also declined, in-
cluding rates of nonmalaria infection (142.5 vs.
90.0 events per 100 patient-years; incidence rate
ratio, 0.62; 95% CI, 0.53 to 0.72) and severe
infection of grade 3 or higher (28.9 vs. 8.0 events
per 100 patient-years; incidence rate ratio, 0.28;
95% CI, 0.19 to 0.42).

Effects on Survival

Analyses of additional key clinical events revealed
significant reductions during hydroxyurea treat-
ment in the rate of malaria infections (46.9 vs. 22.9
events per 100 patient-years; incidence rate ratio,
0.49; 95% CI, 0.37 to 0.66), blood transfusion (43.3
vs. 14.2 events per 100 patient-years; incidence
rate ratio, 0.33; 95% CI, 0.23 to 0.47), and death
(3.6 vs. 1.1 events per 100 patient-years; incidence
rate ratio, 0.30; 95% CI, 0.10 to 0.88). When
grouped into 6-month time intervals, the rates of
multiple life-threatening clinical events declined
rapidly after the initiation of hydroxyurea therapy
and dose escalation, with a sustained or improved
benefit. Reductions in event rates over time were
noted for all sickle cell–related clinical events,
malaria, vaso-occlusive pain, transfusion, the
acute chest syndrome, and death from any cause
(Fig. 3).

D i s c u s s i o n

In this trial involving children with sickle cell ane-
mia living in sub-Saharan Africa, we found that
hydroxyurea treatment was feasible, reasonably
safe, and had both laboratory and clinical benefits.
Specifically, as compared with pretreatment rates,
the rates of clinical events, including vaso-occlusive
pain, infection, malaria, transfusion, and death,
declined after 1 year of hydroxyurea treatment.

Enrollment was robust at all the clinical trial
sites, with enthusiasm and support from local
clinical research teams and populations of pa-
tients, which showed the feasibility of conducting
a large-scale clinical trial in sub-Saharan Africa.
Hematologic dose-limiting toxic effects during the
first 3 months of treatment (the primary safety end
point) occurred in only a small number of par-
ticipants, and the hydroxyurea dose was then
safely escalated toward a maximum tolerated dose,
similar to treatment protocols in the United
States.15 Expected hematologic benefits occurred,
with significant increases in the hemoglobin and
fetal hemoglobin levels. Significant reductions
were observed in the incidence rates of sickle cell–
related clinical events, including vaso-occlusive
pain, and major clinical events including infec-
tion, malaria, transfusion, and death. No serious
adverse events or deaths were considered by the
investigators to be related to hydroxyurea treat-

n engl j med 380;2 nejm.org January 10, 2019126

T h e n e w e n g l a n d j o u r n a l o f m e d i c i n e

ment (Tables S3 and S4 in the Supplementary
Appendix).

In the United States and Europe, hydroxyurea
has emerged as a potent, disease-modifying thera-
py with regulatory approvals for use in both chil-
dren and adults. Although the primary mode of
action of hydroxyurea is through the induction
of fetal hemoglobin, this drug has multiple salu-
tary effects on erythrocytes, leukocytes, and even
endothelium that make it a beneficial oral treat-
ment for this life-threatening disease, especially at
the maximum tolerated dose.18 The WHO in-
cludes hydroxyurea on its Model Lists of Essential
Medicines for children and adults for the treatment
of sickle hemoglobinopathies,19,20 which provides
an impetus for widespread use on a global scale.
To date, however, few studies have been completed
in sub-Saharan Africa or other low-income set-
tings where the burden of sickle cell disease is
greatest.21 In addition to several small studies of
hydroxyurea conducted in Nigeria,22-24 the Novel
Use of Hydroxyurea in an African Region with
Malaria (NOHARM) trial was a randomized, dou-

ble-blinded, placebo-controlled trial that was con-
ducted at a single, large, urban site in Uganda,
which showed that the rates of malaria were not
higher with hydroxyurea than with placebo.25
However, the overall incidence of malaria in the
NOHARM trial was very low, whereas the par-
ticipants in the REACH trial lived in urban and
rural areas of Africa that have a much higher in-
cidence of malaria than Kampala, Uganda, where
the NOHARM trial was conducted.

With full enrollment, high rates of adherence
to trial visits and medication use, and a retention
rate of nearly 95% in the trial over a period of
3 years, our trial showed that hydroxyurea treat-
ment was both feasible and safe in sub-Saharan
Africa. Despite frequent laboratory monitoring,
dose-limiting toxic effects were uncommon in the
first 3 months of treatment, and then the incidence
did not increase further (Table 1). After 6 months
of fixed-dose hydroxyurea therapy, dose escala-
tion was not associated with unacceptable side
effects or with toxic effects regarding laboratory
variables and reached daily doses that were similar

Figure 2. Retention of Participants in the Trial.

The shaded area represents the 95% confidence interval for death or withdrawal from the trial. The inset shows the
same data on an enlarged y axis.

C
u

m
u

la
ti

ve
I

n
ci

d
en

ce
(

%
)

100

80

60

20

40

0
0.0 0.5 1.0 1.5 2.0 2.5 3.53.0

Years since Hydroxyurea Initiation

No. of Participants
Cumulative No. of Trial

Withdrawals
Cumulative No. of Deaths

606
0

0

596
6

4

587
11

8

560
13

11

374
16

13

71
17

16

130
17

15

329
17

14

10

8

9

7

6

4

3

1

5

2

0
0.0 0.5 1.0 1.5 2.0 2.5 3.53.0

Death or withdrawal

Death
Withdrawal

n engl j med 380;2 nejm.org January 10, 2019 127

H y drox y ur e a for Sick le Cell A nemi a in A fr ic a

T
ab

le
1

. L
ab

o
ra

to
ry

a
n

d
C

lin
ic

al
A

d
ve

rs
e

E
ve

n
ts

.*

E
ve

n
t

S
cr

ee
n

in
g

P
h

as
e

T
re

at
m

en
t

P
h

as
e

In
ci

d
en

ce
R

at
e

R
at

io

(9
5%

C
I)

N
o

. o
f

E
ve

n
ts

N
o

. o
f

P
ar

ti
ci

p
an

ts
R

at
e

N
o

. o
f

E
ve

n
ts

N
o

. o
f

P
ar

ti
ci

p
an

ts
R

at
e

no
. o

f e
ve

nt
s

pe
r

10
0

pa
ti

en
t-

yr
no

. o
f e

ve
nt

s
pe

r
10

0
pa

ti
en

t-
yr

D
o

se
-l

im
it

in
g

to
xi

c
ef

fe
ct

23
20

20
.7

30
2

15
9

20
.6

1.
00

(
0.

62
–1

.6
1)

A
n

em
ia

8
8

7.
2

8
6

6
1

5.
9

0.
82

(
0.

40
–1

.7
1)

R
et

ic
u

lo
cy

to
p

en
ia

3
3

2.
7

7
2

6
1

4.
9

1.
82

(
0.

58
–5

.6
8)

N
eu

tr
o

p
en

ia
1

1
0.

9
4

8
4

0
3.

3
3.

59
(

0.
50

–2
5.

71
)

T
h

ro
m

b
o

cy
to

p
en

ia
11

11
9.

9
9

6
6

1
6.

5
0.

66
(

0.
34

–1
.2

5)

C
lin

ic
al

a
d

ve
rs

e
ev

en
t

34
2

23
4

30
8.

4
25

07
52

7
17

0.
7

0.
54

(
0.

48
–0

.6
2)

G
ra

d
e

2
29

3
21

2
26

4.
2

23
54

51
3

16
0.

2
0.

60
(

0.
52

–0
.6

8)

G
ra

d
e

3
48

43
43

.3
15

0
11

2
10

.2
0.

24
(

0.
17

–0
.3

3)

G
ra

d
e

4
1

1
0.

9

3

3
0.

2
0.

23
(

0.
02

–2
.1

8)

S
er

io
u

s
ad

ve
rs

e
ev

en
t

12
12

10
.8

6
5

5
8

4.
4

0.
47

(
0.

25
–0

.9
0)

G
ra

d
e

3
8

8
7.

2
2

4
2

3
1.

6
0.

23
(

0.
10

–0
.5

1)

G
ra

d
e

4
0

0
0.

0
2

5
2

1
1.

7
N

A

G
ra

d
e

5
(d

ea
th

)
4

4
3.

6
1

6
1

6
1.

1
0.

30
(

0.
10

–0
.8

8)

C
lin

ic
al

s
ic

kl
e

ce
ll–

re
la

te
d

e
ve

n
t

12
7

10
7

11
4.

5
77

9
35

9
53

.0
0.

47
(

0.
38

–0
.5

7)

V
as

o
-o

cc
lu

si
ve

p
ai

n
o

r
d

ac
ty

lit
is

10
9

92
98

.3
65

5
32

3
44

.6
0.

45
(

0.
37

–0
.5

6)

A
cu

te
c

h
es

t
sy

n
d

ro
m

e
o

r
p

n
eu

m
o

n
ia

10
10

9.
0

7
3

5
6

5.
0

0.
55

(
0.

28
–1

.0
5)

A
cu

te
s

p
le

n
ic

s
eq

u
es

tr
at

io
n

2
2

1.
8

1
6

1
2

1.
1

0.
60

(
0.

13
–2

.7
5)

S
tr

o
ke

2
2

1.
8

1
1

1
0

0.
7

0.
42

(
0.

10
–1

.6
7)

O
th

er
s

ic
kl

e
ce

ll–
re

la
te

d
e

ve
n

t
3

3
2.

7
1

5
1

3
1.

0
0.

38
(

0.
11

–1
.3

4)

A
n

y
ev

en
t

o
f

gr
ad

e
≥

3
26

23
23

.4
10

4
8

5
7.

1
0.

30
(

0.
19

–0
.4

8)

A
d

d
it

io
n

al
c

lin
ic

al
e

ve
n

ts

M
al

ar
ia

52
50

46
.9

33
6

20
0

22
.9

0.
49

(
0.

37
–0

.6
6)

N
o

n
m

al
ar

ia
in

fe
ct

io
n

15
8

13
8

14
2.

5
13

22
42

9
90

.0
0.

62
(

0.
53

–0
.7

2)

A
n

y
in

fe
ct

io
n

o
f

gr
ad

e
≥

3
32

31
28

.9
11

8
9

6
8.

0
0.

28
(

0.
19

–0
.4

2)

S
ep

ti
ce

m
ia

16
15

14
.4

3
7

3
4

2.
5

0.
18

(
0.

09
–0

.3
2)

T
ra

n
sf

u
si

o
n

48
43

43
.3

20
9

14
9

14
.2

0.
33

(
0.

23
–0

.4
7)

*
T

h
e

sc
re

en
in

g
p

h
as

e
in

cl
u

d
ed

1
11

p
at

ie
n

t-
ye

ar
s,

a
n

d
t

h
e

tr
ea

tm
en

t
p

h
as

e
14

69
p

at
ie

n
t-

ye
ar

s.
T

h
e

co
n

fi
d

en
ce

i
n

te
rv

al
s

w
er

e
n

o
t

ad
ju

st
ed

f
o

r
m

u
lt

ip
le

c
o

m
p

ar
is

o
n

s.
N

A
d

en
o

te
s

n
o

t
ap

p
lic

ab
le

.

n engl j med 380;2 nejm.org January 10, 2019128

T h e n e w e n g l a n d j o u r n a l o f m e d i c i n e

T
ab

le
2

. L
ab

o
ra

to
ry

E
ff

ec
ts

o
f

H
yd

ro
xy

u
re

a
T

re
at

m
en

t
d

u
ri

n
g

th
e

T
ri

al
.*

V
ar

ia
b

le
M

o
n

th
0

M
o

n
th

3
M

o
n

th
6

M
o

n
th

1
2

M
o

n
th

2
4

M
o

n
th

3
6

C
h

an
ge

f
ro

m

M
o

n
th

s
0

to
1

2

(9
5%

C
I)

†

H
em

o
gl

o
b

…