PROLONGED CO-TRIMOXAZOLE IN HIV-INFECTED CHILDREN

36
was presumptive;

although an endwith a relatively highpoint review comCD4 T-cell percentagemittee adjudicated

(me-dian, 33% in theall diagnoses, it is
ARROW trial vs. 11%possible that some
in the Children withparticipants had
HIV
Antibiotictuberculosis,
P.
Prophylaxis
[CHAP]jirove-cii
7
or
trial ). Our findingspneumonia,
infection
due
to
contrast with those of a
another
or-ganism,
previous observational
as
previously
13
U.S. study,
which
35
We
showed no increasedreported.
fewer
risk of severe bacterialobserved
of
infection among ART-cases
treated children whotuberculosis
stopped receiv-ing co-among
trimoxazole after goodparticipants who
immune reconsti-tution.stopped receiving
-trimoxazole
However,
studiesco
involving adults in Af-than among those
rica have shown thatwho continued to
there is an ongoing riskreceive it (Table
in
the
of severe bacterialS2
infection
associatedSupplementary
with rela-tively highAppendix);
30-33however,
most
CD4 T-cell counts.
diagnoses
of
Maintaining
tuberculosis
were
prophylaxis against a
wide spectrum of gram-presump-tive, the
were
positive and gram-numbers
small,
and
negative organisms may
tuberculosis
and
there-fore be important
in sub-Saharan Africa,bacterial
where rates of bacterialcoinfection is not
35

unusual.
infection are high.
Like
other One question
4,7,15,24
why
the
studies,
oursis
showed bene-fits fromantibacterial
activity of coco-trimoxazole
prophylaxis in twotrimoxazole
in
countries with highpersists
children
(those
in
rates of resistance to
co-tri-moxazole,
forour study had
been
receiving
reasons that are unclear.
ART for 4 to 5
Co-tri-moxazole may
years by the date
retain
sufficient
of trial closure)
antimicrobial activity to
but wanes in
prevent severe bacterial
15
disease, or its activityadults.
against other pathogens,Children
including
Myco- generally have a
higher risk of
bacterium
34

bacte-rial

tuberculosis
and P. infections
than
jirovecii, may haveadults,
and
been underestimated.immune reconSince
diagnosticstitution is driven
capacity
in
thepredominantly by
ARROW trial wasrepopula-tion
limited, the diagnosis ofwith naive CD4 T
most cases of severecells
in
bacterial
pneumonia

children but not

in
adults;
children
may
therefore
still
have
poor
pathogen-specific
memory
responses
despite long-term
ART.

those
receiving
Stopping
co-placebo.37
The
trimoxazole
high frequency of
prophylaxis had no
infection,
significant effect on
particularly
growth in our study parmalaria,
in
ticipants, in contrast to
children
who
the improved growth
ob-served
amongstopped receiving
children who receivedco-trimoxazole
co-trimoxa-zole,
ascould explain the
compared with thoserates of anemia in
who received placebo,the ARROW trial.
in the CHAP trial,Alter-natively, cowhich involved HIV-trimoxazole may
the
infected children whoreduce
immune
activation
were not receiving
that
impairs
37
ART.
Children have
39
erythropoiesis,
rapid weight and height
9either directly by
increases during ART,
means
of
particularly during theimmunomodulator
38
first
year ;
anyy prop-erties or
additional benefit of co-indirectly
by
trimoxazole after 2reducing
the
years of ART may betranslocation
of
minimal.
intestinal
In
our
study,microbes.15 The
children
and
greater increases
adolescents
who
in white-cell and
stopped receiving coneutrophil counts
trimoxazole had higher
in
participants
rates
of
anemia,
who
stopped
including
potentially
receiving
colife-threatening (grade
trimoxazole,
as
4) events, than those
com-pared
with
who continued cothose
who
trimoxazole. Although
continued
to
this
may
seem
receive it, are
counterin-tuitive, it is
consistent with the
consistent with the
potential
results of the CHAP
myelosuppressive
trial, in which children
effects of coreceiving
co-tritrimoxazole.
moxazole who had not
The
main
previously
received
limitation
of
our
ART
had
greater
trial
is
the
combiincreases
in
the
of the
hemoglobin level thannation

open-label design
and the management-based
primary end point
(hospitalization).
However,
the
open-label design
had the advan-tage
of reducing the
pill burden in the
prophy-laxisstopped
group,
thus
allowing
potential
improvements in
ART adherence to
be
evaluated.
There was no cost
to caregivers, so
there was no bias
from differences
in the ability to
pay for hos-pital
care
between
groups. We cannot
rule
out
differences
in
preventive
or
health-seeking behavior
among
caregivers
or
differences among
clinicians in the
threshold
for
hospitalizing
a
patient.
More
participants in the
prophylaxisstopped
group
might have been
hospitalized
owing to concerns
about
the
cessation of co-tri-

51

N ENGL J MED 370;1 NEJM.ORG JANUARY 2, 2014

ew
T
Engl
h
andnejm.org
Downloaded
from
e
Journ
on
N
al of

Medi
January 23, 2014.
cine
For personal use
only. No other uses
without permission.
Copyright 2014

Massachusetts
Medical Society.
All rights
reserved.

T h e NE W E NGL A ND JOU R NA L o f M E DICINE

after more than

6 months of
moxazole
ART need
prophylaxis,
to
be
realthough it is
considered.
equally
plauDespite
the
sible that more
inclusion
of
participants in
participants
the prophylaxiswith more than
continued group
2 years of ART
were
and high CD4
hospitalized
T-cell counts
owing
to
(median count
concerns about
in participants
breakthrough
>5 years of
infection during
age, 720 per
prophylaxis.
cubic
The consistency
millimeter),
across multiple
continued pro(in-cluding
phylaxis with
severe)
end
co-trimoxazole
points
and
led
to
subgroups sugsustained
gests
an
reductions in
underlying
hospitalization
effect. Even if
for a range of
the real effect is
smaller
thanin-fections, in
where
observed,
co-areas
malaria
is
not
trimoxazole is
endemic
as
inex-pensive,
well
as
in
costing only a
areas
where
it
few cents per
is
endemic,
day.
no
Our findingswith
evidence
of
highlight
the
importance ofvariation
co-trimoxazole across
ages
in
treatingand with conHIV-infected sistent benefits
children
and
adolescents inwith respect to
sub-Saharan secondary end
Africa, even inpoints.
the ART era.Continued coCurrent WHOtrimoxazole
recommendatio
prophylaxis
11
ns
had
an
that co-acceptable
trimoxazole
safety profile
prophylaxis
and was not
can be stoppedas-sociated
in
childrenwith
poor
older than 5adherence to
years of ageART.
who
have
adequate CD4
T-cell
reconstitution
(>350 CD4 T
cells per cubic
millimeter)

These
findings argue
strongly
for
long-term cotrimoxazole
prophylaxis in
conjunction
with ART in
children
and
adolescents in
sub-Saharan
Africa,
until
further
evidence
is
available
to
guide decisions
about cessation
of prophylaxis.
Wheth-er the
benefits of cotrimoxazole
will wane over
the long term is
unknown, but
given
the
persis-tent risk
we observed,
co-trimoxazole
will prob-ably
continue
to
provide
a
benefit after 2
years. Although
an increase in
the use of cotrimoxazole
prophylaxis
carries the risk
of
increased
bacterial
resistance
in
locations where
co-trimoxazole
is still used for
first-line
treatment,
resistance rates
are
already
1,4,7

high,
and
infections
occurring in our
study
participants
who
had
stopped
receiving
prophylaxis
were
often
treated
with
broad-spectrum antibiotic

GlaxoSmithKline
donated
first-line
drugs
for
the
Antiretroviral
Research
for
Watoto (ARROW)
trial and provided
funding
for
virologic assays.
Disclosure
forms provided
by the authors are
available with the
full text of this
article at
NEJM.org.
We thank the
participants,
caregivers, and
staff from all the
centers
participating in the
ARROW trial.
8.]
Walker
AS, et
C
al.
Wi hintu
Effect
ktor
C,
of
SZ,
Bhat
cotrim
Sassan GJ,
oxazol
Walk e on
Morok er
causes
ro M, AS,
of
Grant
et al. death,
AD, et Cohospit
al.
trimo al
Effica
xazol
cy of e as admiss
ions
trimet
proph and
hopri
ylaxi
antibio
m-sul- s
phame again tic use
in
thoxaz st
HIVole
oppor infecte
prophy tunist
d
laxis
ic
childre
to
infect n.
decrea ions
AIDS
se
in
2007;
morbi
HIV21:77dity
infect 84.
and
ed
mortal Zamb
Su
ity in ian
tcliffe
HIVchildr CG,
1van
en
infecte (CH
Dijk
d
JH,
AP):
patient a
Bolton
s with doubl C,
tuberc
Persau
eulosis
d D,
blind
in
rando Moss
Abidja mise
WJ.
n,
Effecti
d
Cte
place veness
dIvoir boof ane:
a contr
tiretro
rando
viral
olled
mised
therap
trial.
control Lany
led
among
cet
trial.
2004; HIVLancet 364:1 infecte
1999;3 865d
53:146 71.
childre
9-75.
n
in
M
[Erratu
subuleng
m,
Sahara
Lancet a V, n
1999;3 Ford
Africa.
53:207 D,
Lancet

agents, such asand improved

ceftriaxone.
methods
of
Long-term
drug-use foreoperational
casting
and
research wouldprocurement of
usefully addressco-trimoxazole
both
theseat
health
issues.
facilities
to
Achieving
prevent
widespread
depletion
of
benefits
fromstocks.40
co-trimoxazole Supported by the
for children inU.K.
Medical
ART pro-gramsResearch Council
will
requireand the U.K.
Department
for
additional
International
training
ofDevelopment. ViiV
and
health workersHealthcare
in adults
with
REFEREN tuberculosi
CES
s in rural
Anglaret South
Africa.
X, Chne G,
AIDS
Attia A, et al.
2005;19:16
Early
3-8.

1.

chemoproph
Mermi
ylaxis with
trimethoprim n J, Lule J,
Ekwaru JP,
sulphametho et al. Efxazole
forfect of coHIV-1trimoxazol
infected
e
adults
inprophylaxi
Abidjan,
s
on
Cte
morbidity,
dIvoire: amortality,
ran-domised CD4-cell
trial. Lancetcount, and
1999;353:14 viral load
63-8.
in
HIV

4.

in
2. Badri M,infection
rural

Ehrlich
R,
Ugan-da.
Wood
R,
Lancet
Maartens G.
2004;364:1
Initiating co428-34.
trimoxazole
Nunn
prophylaxis
in
HIV-AJ, Mwaba
infected
P, Chintu
patients
inC, Mwinga
Africa:
anA,
eval-uation Darbyshire
of
theJH, Zumla
provisional A. Role of
WHO/UNAI coDS
trimoxazol
recommenda e
tions. AIDSprophylaxi
2001;15:114 s
in
3-8.
reducing

5.

3. Grimwa mor-tality
in
HIV

de K, Sturm
infected
AW, Nunn
adults
AJ, Mbatha
D, Zungubeing
D, Gilks CF.treated for
Effective- tuberculosi
ness
ofs:
cotrimoxazo randomise
d clinical
le
prophylaxis trial. BMJ
on mortality2008;337:a
257.

6.

7.

9.

8.

Infect Dis
2008;8:477tion
in
89.
[Erratum, infants and
children:
Lan-cet
Infect Distowards
2009;9:736. uni-versal
access
]
recommen
Adations for
ntiretro a
public
viral
health
therapy approach.
for HIVGeneva:
infec- World
Health
Organizati
on, 2010
(http://ww
w
.
who.int/hi
v/pub/pae
diatric/inf
ants2010/
en/index.h
tml).

11.

e prophy-Mellad
laxis
foro M,
HIVet al.
related
Withdr
infections awal
among
of
children, Pneum
adolescentsocystis
and adultsjirovec
in
re-ii prosourcephylax
limited
is in
settings: HIVrecommen infecte
dations ford
a
publicchildre
health
n
approach. under
Geneva: highly
World
active
Health
antiret
Organizati roviral
on, 2006therap
(http://
y.
www.who.i AIDS
nt/hiv/pub/ 2005;1
plhiv/ctx/e 9:2103
n/).
-8.

Guidelines
on
cotrimoxazol

Urschel S,Nach
Ramos J,man S,

10.

12.

13.

Gona
P,
Dank
ner
W, et
al.
The
rate
of
serio
us
bacte
rial
infect
ions
amon
g
HIVinfect
ed
childr
en
with
immune
recon
stituti
on
who
have
disco

Downloaded from

52
N ENGL
J MED

370;1
NEJM.O
RG
JANUAR
Y 2,

2014

The New England Journal of Medicine

ntinued
opport
unistic
infecti
on
prophy
laxis.
Pediatr
ics
2005;1
15(4):
e488e494.

14.
Suthar
AB,
Granic
h R,
Mermi
n
J,
Van
Rie A.
Effect
of
cotrim
oxazol
e on
mortal
-

14
Massac
husetts
Medica
l
Society
. All
rights
reserve
d.