Timing of Patent Ductus Arteriosus Treatment and Respiratory Outcome

in Premature Infants: A Double-Blind Randomized Controlled Trial

Ilene R. S. Sosenko, MD, M. Florencia Fajardo, MD, Nelson Claure, MSc, PhD, and Eduardo Bancalari, MD
Objective To determine whether early ibuprofen treatment, at the onset of subtle patent ductus arteriosus (PDA)
symptoms, would improve respiratory outcome in premature infants compared with expectant management,
with ibuprofen treatment only when the PDA becomes hemodynamically signicant (HS).
Study design We conducted a randomized double-blind controlled trial of infants with gestational ages 23 to 32
weeks and birth weights 500 to 1250 g who had echocardiography for subtle PDA symptoms (metabolic acidosis,
murmur, bounding pulses). Infants were then randomized to early treatment (blinded ibuprofen; n = 54) or
expectant management (blinded placebo, n = 51). If the PDA became HS (pulmonary hemorrhage, hypotension,
respiratory deterioration), infants received open label ibuprofen. Infants with HS PDA at enrollment were excluded
from the study. Respiratory outcomes and mortality and major morbidities were determined.
Results Early treatment infants received ibuprofen at median age of 3 days; infants in the expectant group
in whom HS symptoms developed (20%) received ibuprofen at median of 11 days. A total of 49% of expectant
infants never required ibuprofen or ligation. No signicant differences were found in the primary outcome (days on
oxygen [O
2
] during the rst 28 days), death, O
2
at 36 weeks, death or O
2
at 36 weeks, intestinal perforation, surgical
necrotizing enterocolitis, grades III and IV intracranial hemorrhage, periventricular leukomalacia, sepsis or retinop-
athy of prematurity.
Conclusion Infants with mild signs of PDA do not benet from early PDA treatment compared with delayed
treatment. (J Pediatr 2012;160:929-35).
A
lthough the ductus arteriosus is critical in utero to allow most blood to bypass the pulmonary circulation while main-
taining fetal systemic blood ow, its persistent patency after birth can cause signicant clinical consequences related to
excessive pulmonary blood ow and compromised systemic blood ow. Whereas in the term infant the ductus con-
stricts within hours of birth, in the premature infant the ductus may remain patent for a prolonged period or fail to close.
1
The current clinical approach to a patent ductus arteriosus (PDA) in the preterm infant is varied and controversial. Although
some clinicians treat a PDA prophylactically, others treat a PDA when mild signs present within a few days after birth. Other
clinicians take an expectant approach and allow for possible spontaneous closure, treating a PDA at a later time, only when
signs indicate hemodynamic signicance. In addition, the denition and management of a hemodynamically signicant (HS)
PDA are both variable and controversial. This wide variation in approach to the PDA is caused by the lack of denitive data on
the optimal time to treat for immediate clinical response and long-term complications, such as bronchopulmonary dysplasia
(BPD) and pulmonary hypertension.
2
There are inherent risks and benets with either approach. The early treatment of PDA may be benecial by eliminating both
early morbidity (including hypotension, respiratory deterioration, intestinal hypoperfusion) and later morbidity (including
BPD). However, because a substantial percentage of PDAs will close without therapy, or may remain open without producing
signicant symptoms, initiating early treatment may unnecessarily expose many infants to cyclo-oxygenase inhibitors or sur-
gical ligation, both of which are associated with adverse effects. This situation might be avoided by delaying PDA treatment and
allowing for the possibility of spontaneous closure without exposure to unnecessary medication or surgery.
Multiple studies have reported an association between PDA and increased risk for BPD.
3-5
Despite studies examining the
different treatment modalities and treatment schedules for PDA, none has denitively shown a signicant effect on reduction
in the incidence of BPD. However, few studies have enrolled a large enough number of infants, particularly extremely prema-
ture infants, or have delayed PDA treatment beyond 7 days. Therefore, groups reported from the literature receiving either
early or late treatment ended up differing by only a few days.
6,7
Thus, the
From the Division of Neonatology, Department of
Pediatrics, University of Miami/Miller School of Medicine,
Miami, FL
Supported by an unrestricted grant from Ovation (now
Lundbeck) Pharmaceuticals and University of Miami
(Project: New Born). The authors declare no conicts of
interest.
Registered at ClinicalTrials.gov (NCT00802685).
0022-3476/$ - see front matter. Copyright 2012 Mosby Inc.
All rights reserved. 10.1016/j.jpeds.2011.12.031
BPD Bronchopulmonary dysplasia
FIO
2
Fraction of inspired oxygen
HS Hemodynamically signicant
MV Mechanical ventilation
NEC Necrotizing enterocolitis
O
2
Oxygen
PDA Patent ductus arteriosus
929
optimal timing for PDA therapy might prevent the progres-
sion of lung damage and the development of BPD, while lim-
iting the potential complications of PDA therapy.
The objective of this trial was to determine whether
early ibuprofen treatment, at the onset of subtle PDA
symptoms, would improve respiratory outcome in prema-
ture infants compared with expectant management, with
ibuprofen treatment only if and when the PDA became
HS. We hypothesized that early pharmacologic treatment
of the PDA with ibuprofen would improve respiratory
course (as reected by a reduction in duration of supplemen-
tal oxygen (O
2
) during the rst 28 days) compared with
expectant PDA treatment.
Methods
This prospective double-blind randomized control clinical
trial was conducted from January 2008 to August 2010 at the
tertiary level neonatal intensive care unit at Holtz Chil-
drens/University of Miami/JacksonMemorial Medical Center
(Miami, Florida). The study was approved by the institutional
review board of the University of Miami. Written informed
parental consent was obtained before enrollment. The trial
was registered with ClinicalTrials.gov (NCT00802685).
Eligible infants were those born at Jackson Memorial Hos-
pital or its afliates, with birth weights between 500 and 1250
g and gestational ages between 23 and 32 weeks who were >24
hours old but #14 days old. Infants who were severely small
for gestational age (>3 SD less than the mean birth weight for
gestational age) or had major congenital malformations,
proven sepsis (positive blood culture results), serum creati-
nine level >1.7, oliguria (urine output <1 cc/kg/hr), pulmo-
nary hypertension (with right-to-left PDAshunt), abdominal
pathology (abdominal distension, discoloration, abnormal
abdominal radiograph), bleeding diathesis, terminal condi-
tion (intractable respiratory failure, intractable hypotension,
no expectation of survival beyond 48 hours) were excluded,
as were infants with symptoms of a HS PDA at study enroll-
ment (see denition below).
Sample size was determined by using the primary outcome
of number of days spent on supplemental O
2
during the rst
28 days of life. On the basis of earlier data from our neonatal
intensive care unit, infants in this birth weight category had
a mean duration of 20 8 days of supplemental O
2
during
the rst 28 days. To detect a 20% difference of days of O
2
de-
pendence during the rst 28 days, an estimated 84 infants per
group (168 total) were determined to be needed. On the basis
of the census in the neonatal intensive care unit, we estimated
that approximately 150 babies would be born in this weight
and gestational age category per year and that approximately
100 of them would have PDA. We estimated being able to
conduct this study in a 2- to 3-year period. However, after
two-thirds planned enrollment (105 of 168 infants), NeoPro-
fen (Ovation Pharmaceuticals, Lebanon, New Jersey) was re-
called by the manufacturer and was no longer available in the
United States for clinical or research use. This necessitated
study termination in August of 2010.
Safety and Efcacy Monitoring/Interim Analyses
Safety was overseen by a data safety monitoring board, which
evaluated results when the study was one-third and two-
thirds completed (shortly before study termination) for these
safety issues: death, intestinal perforations, necrotizing
enterocolitis (NEC) requiring surgery, pulmonary hyperten-
sion requiring nitric oxide, severe pulmonary hemorrhage,
and renal failure (dened as creatinine level >2.0).
Interim efcacy analysis of the primary outcome measure
(number of days on O
2
at 28 days) was also conducted at
one-third and two-thirds of enrollment by using the method
of two-sided spending function of OBrien-Fleming.
8
Study Procedure
Consent was obtained as soon as possible after birth for all in-
fants meeting eligibility criteria. After 24 hours of age, but not
after 14 days of age, eligible infants were observed for the pres-
ence of early, mild symptoms of a PDA, specically dened as
metabolic acidosis (base excess <7), meanbloodpressure less
than weeks in gestation or requiring vasopressor support,
bounding pulses/hyperactive precordium, or systolic murmur
(Figure 1; available at www.jpeds.com). When one or more of
the aforementioned symptoms was noted, a color Doppler
echocardiogram was performed. When the initial echo-
cardiogram results were positive for PDA (dened as the
presence of a PDA with either predominantly left-to-right or
bidirectional shunt), infants were stratied according to birth
weight (500-800 g and 801-1250 g) and randomized (via
sealed envelopes prepared by using a random number table)
to one of two study groups. When the initial echocardiogram
results were negative for PDA, the enrolled infants were
observed until 14 days of age, and a repeat echocardiogram
was conducted when PDA symptoms developed; when the
results of this repeat echocardiogram were positive, the baby
was then randomized to one of the two study groups.
Clinicians, investigators and nursing staff were blinded to the
study group to which the baby was assigned and the
medication the baby was receiving. Only the neonatal
pharmacists were aware of the study group of each baby and
were responsible for preparing the blinded ibuprofen or
blinded placebo study drug. The two study groups were:
(1) Early treatment (consisting of study drug, which for
this group was blinded ibuprofen); or (2) Expectant
treatment (consisting of study drug, which for this group
was blinded placebo). When infants in either study group,
after receiving their initial specic blinded study drug,
continued with or developed PDA symptoms that did not
meet criteria of hemodynamic signicance, they were either
observed or received a prolonged course or second course of
their assigned blinded study drug at the discretion of
the clinical team. For both groups, treatment with ibuprofen
(open label) was initiated only when symptoms of a HS PDA
developed and a PDA was conrmed with echocardiography.
Criteria dening a HS PDA included signs of PDA plus the
presence of pulmonary hemorrhage (persistent bloody
secretions from the endotracheal tube) alone, or signs of
PDA plus cardiomegaly and pulmonary edema on chest
THE JOURNAL OF PEDIATRICS www.jpeds.com Vol. 160, No. 6
930 Sosenko et al
radiography plus one of the following: either hypotension not
caused by an identiable cause other than PDA requiring
vasopressor dose >10 mg/kg/min or respiratory failure (not
caused by an identiable cause other than PDA), which was
dened as the presence of at least two of the following
increases in ventilator settings: fraction of inspired oxygen
(FIO
2
) > 0.5, intermittent mandatory ventilation > 40
breaths/min, peak inspiratory pressure > 20 cm H
2
O, or high
frequency ventilation with Paw > 13 and FIO
2
> 0.5. These
settings needed to be maintained for >8 hours and were
required to maintain O
2
saturation between 88% and 95%
and PaCO
2
< 65 mm Hg. The study intervention ended at
day 28. When the PDA remained open after day 28 and the
infant was still requiring mechanical ventilation (MV), the
PDA could be treated with open label ibuprofen or PDA
ligation at the discretion of the clinical team.
Drug Administration
At randomization, infants in the early treatment group
received blinded ibuprofen lysine (NeoProfen, Ovation
[now Lundbeck] Pharmaceuticals, Deereld, Illinois); in-
fants in the expectant treatment group received blinded
placebo (5% dextrose). The dosing schedule for ibuprofen
was an initial dose of 10 mg/kg, followed by two doses of 5
mg/kg each, every 24 hours, by slow intravenous infusion;
dosing of placebo involved equivalent volumes of dextrose
by slow intravenous infusion on the same schedule. When
it was determined that an enrolled infant had persistent
signs of PDA without meeting hemodynamic signicance,
a prolonged course of blinded assigned study drug (two
additional doses of blinded 5 mg/kg ibuprofen or equiva-
lent volume of blinded placebo) or a repeat course of
blinded assigned study drug could be administered at the
discretion of the clinical team. When HS PDA developed
in infants from either study group before 28 days and
PDA was conrmed, infants received unblinded, open-
label ibuprofen and, when contraindicated or unsuccessful,
PDA ligation.
Study Endpoints
Data were collected on maternal and infant demographics
and PDA characteristics of both groups, including age of
PDA presentation, age at randomization, percent of babies
in each group requiring open label ibuprofen or ligation,
time of ibuprofen administration, PDA outcome after 28
days of life, and number in whom contraindications (includ-
ing renal and abdominal pathology) developed. The primary
outcome of this study was the number of days spent on sup-
plemental O
2
by each infant during the rst 28 days (as an in-
dicator of evolving BPD). The need for supplemental O
2
was
dened as O
2
>21%needed to keep saturations $88%. Death
and major respiratory outcomes including need for O
2
at 36
weeks corrected post-menstrual age, death or need for O
2
at
36 weeks, and number of days spent on MV during the rst
28 days, and the need for O
2
or MV at day 28 was assessed.
Secondary outcomes included total duration of O
2
, duration
of MV, need for O
2
>30% at 36 weeks, pneumothorax, pul-
monary interstitial emphysema, postnatal steroids, intestinal
perforation, necrotizing enterocolitis requiring surgery, in-
tracranial hemorrhage, periventricular leukomalacia, reti-
nopathy of prematurity, and sepsis (dened as positive
blood culture results).
Statistical Analysis
Data analyses were conducted on an intention-to-treat basis.
Continuous variables in groups were compared with the stu-
dent t test or Mann-Whitney rank sum test as appropriate,
and categorical variables were compared with c
2
test and
Fisher exact test for the pooled data from all infants and
within each strata. Mantel-Haenszel statistics were used to
test for independence between the treatments and dichoto-
mous outcome variables controlling for the co-variate of
birth weight strata. The Mantel-Haenszel common OR and
95% CI are reported. Statistical analysis was conducted
with IBM SPSS Statistics software version 19 (IBM Corp, Ar-
monk, New York and SPSS Inc, Chicago, Illinois), SAS (SAS
Institute, Cary, North Carolina), and nQuery software (Sta-
tistical Solutions; Saugus, Massachusetts). A P value <.05
was considered to be signicant. Statistical analysis was per-
formed with the Biostatistics Collaboration and Consulting
Core of the Division of Biostatistics, Department of Epidemi-
ology and Public Health of the University of Miami Miller
School of Medicine.
Results
During the study period, there were 399 infants between
500 and 1250 g born at Jackson Health System Hospitals
(Figure 2; available at www.jpeds.com). Of those infants,
consent was given for 179 infants, but 6 of the consents were
withdrawn. One hundred ve infants were randomized, and
68 infants were not randomized because of presentation
with the exclusion criterion of HS PDA at randomization
(n = 10), absence of PDA on echocardiography (n = 55), or
other reasons (suspected NEC, oliguria, early death, n = 3).
The characteristics of the infants in the early and the
expectant treatment groups, stratied in two weight
categories (500-800 g and 801-1250 g) and analyzed as one
composite weight group, are seen in Table I. Of the 105
infants enrolled in the study, 54 were randomized to the
early group and 51 to the expectant group (Figure 2).
When stratied by birth weight, 57 infants were enrolled
in the 801 to 1250 g group, compared with 48 infants
in the #800 g group. There were no signicant differences
in infant demographics between early and expectant
infants in gestational age, sex, race, antenatal steroid
exposure, maternal chorioamnionitis, Apgar score at 5
minutes, and early respiratory course. Although more
early infants in the composite weight group were on O
2
at
24 hours compared with expectant infants, this difference
disappeared by 72 hours, and there were no differences in
proportion of infants needing MV at 24 hours or needing
surfactant therapy between groups.
June 2012 ORIGINAL ARTICLES
Timing of Patent Ductus Arteriosus Treatment and Respiratory Outcome in Premature Infants: A Double-Blind
Randomized Controlled Trial
931
The response of the PDA to early compared with expec-
tant treatment is summarizedinTable II. PDAwas diagnosed
in infants in both groups with echocardiography at a median
of day 3 of life, and infants were randomized to either early
treatment (blinded ibuprofen) or expectant treatment
(blinded placebo) at day 3 to 4, indicating that all early
treatment infants received their rst dose of ibuprofen at
a relatively early age. A signicantly greater number of
infants in the expectant group (blinded placebo) were
given a second course of blinded study drug compared with
infants in the early treatment group (blinded ibuprofen).
Of infants in the expectant group, only 20% needed open
label ibuprofen during the rst 28 days for HS PDA and
received it at a median age of 11 days. Thirteen percent of
infants who had received early ibuprofen (early group),
also required open label ibuprofen in the rst 28 days for HS
PDA. There were no signicant differences in groups for the
need for open label ibuprofen (for HS PDA) or need for
surgical ligation in the rst 28 days of age.
There were no signicant differences between the early
treatment group and the expectant treatment group in
any of the respiratory outcomes listed in Table III, except
a higher proportion of infants in the early group required
>30% inspired O
2
at 36 weeks. Specically, the primary
outcome, number of days on O
2
during the rst 28 days,
did not differ in the treatment groups. There were no
differences in total days on O
2
or total days on MV, BPD
(dened as O
2
requirement at 36 weeks post-menstrual
age), and the composite outcome of death or O
2
at 36
weeks between expectant and early treatment groups.
Death until discharge did not differ in the groups. Pulmo-
nary complications, including need for postnatal steroids and
pneumothorax, were not different, although there was more
pulmonary interstitial emphysema in the early treatment
group. There were no differences in other complications of
prematurity, including grades III and IV intracranial hemor-
rhage, preventricular leukomalacia, necrotizing enterocolitis
(requiring surgery), spontaneous intestinal perforation, sep-
sis or retinopathy of prematurity $ stage 3 (Table IV).
Discussion
Currently, the clinical approach to a PDA is variable between
institutions, on the basis of the lack of denitive evidence-
based data. It remains unclear as to what represents the opti-
mal time to treat a PDA: prophylactic treatment in infants
younger than a certain gestational age, early treatment at
the onset of mild symptoms, or later when the PDA becomes
HS, in relation to immediate clinical consequences, and long-
term outcomes, particularly BPD. There are potential com-
plications of medical PDA therapy, such as renal dysfunction
and intestinal perforation. There are also complications of
surgical PDA closure, such as cardiopulmonary dysfunction,
Table II. Randomization and PDA evolution
All infants Birth weight 500-800 g Birth weight 801-1250 g
Early (n = 54) Expectant (n = 51) Early (n = 24) Expectant (n = 24) Early (n = 30) Expectant (n = 27)
PDA 1st positive echocardiogram, day 3 (2-5) 3 (2-5) 3 (2-4) 3 (2-4) 3 (2-5) 4 (2-5)
Randomization, day 3 (2-5) 4 (2-5) 3 (2-4) 3 (2-4) 4 (2-5) 4 (2-6)
Repeat course of blinded study drug, rst 28 days* 9 (17) 24 (47) 5 (21) 12 (50) 4 (13) 12 (44)
Open-label ibuprofen, rst 28 days 7 (13) 10 (20) 5 (21) 7 (29) 2 (7) 3 (11)
Postnatal day 15 6 11 6 13 6 12 5 19 3 10 8
Surgical ligation, rst 28 days 4 (7) 4 (8) 4 (17) 3 (13) 0 (0) 1 (4)
Open-label ibuprofen, after 28 days 7 (14) 13 (26) 4 (18) 8 (33) 3 (10) 5 (19)
Surgical ligation, after 28 days 4 (8) 6 (12) 3 (14) 5 (21) 1 (3) 1 (4)
Results expressed as mean SD, median (10th-90th percentile), or number (%).
*4.44 (1.79-10.98) expectant group compared with early group, Mantel-Haenszel common OR estimate.
Table I. Population characteristics
All infants Birth weight 500-800 g Birth weight 801-1250 g
Early (n = 54) Expectant (n = 51) Early (n = 24) Expectant (n = 24) Early (n = 30) Expectant (n = 27)
Birth weight, g 854 204 842 203 663 85 666 92 1007 125 997 135
Gestational age, weeks 26 (23-28) 25 (24-29) 24 (23-26) 24 (23-26) 27 (25-29) 27 (25-30)
Female 29 (54) 21 (41) 14 (58) 11 (46) 15 (50) 10 (37)
Black 33 (61) 30 (59) 17 (71) 19 (79) 16 (53) 11 (41)
Antenatal steroids 46 (85) 45 (88) 22 (92) 23 (96) 24 (80) 22 (82)
Maternal chorioamnionitis 7 (13) 9 (18) 5 (21) 5 (21) 2 (7) 4 (15)
5 minute Apgar score 7 (5-9) 7 (4-9) 6 (5-8) 7 (3-9) 8 (5-9) 7 (4-9)
On O
2
at 24 hours 27 (52) 18 (35) 13 (54) 12 (50) 15 (50) 6 (22)
On O
2
at 72 h 24 (44) 23 (45) 13 (54) 12 (50) 11 (37) 11 (41)
On MV at 24 hours 36 (67) 39 (77) 20 (83) 19 (79) 16 (53) 20 (74)
Surfactant administration 39 (72) 38 (75) 16 (67) 18 (75) 23 (77) 20 (74)
Results expressed as mean SD, median (10th-90th percentile), or number (%).
THE JOURNAL OF PEDIATRICS www.jpeds.com Vol. 160, No. 6
932 Sosenko et al
chylothorax, and vocal cord and diaphragmatic paralysis. An
important consideration is that a considerable number of
PDAs will close spontaneously (approximately 15% sponta-
neous closure in infants <25 weeks and approximately 75%
at >29 weeks).
9
We undertook this study to clarify this clin-
ical dilemma.
Excessive pulmonary blood ow in an immature lung with
ongoing maturation of alveolar and vascular structures can
result in abnormal development of the pulmonary vessels.
10
Furthermore, studies in premature baboons indicate that in-
fants with persistence of the PDA have delayed alveolar devel-
opment and decreased alveolar surface area compared with
infants with early closure of the PDA by using ibuprofen.
11
Both of these aspects of developmental derangement, vascu-
lar and alveolar, are features of BPD in human infants.
12,13
However, despite strong evidence that increased ow and
pressure in the developing pulmonary vasculature caused
by a PDA can produce morphologic and functional alter-
ations in the lung, clinical evidence linking PDA with the de-
velopment of BPD remains largely epidemiological.
5
Because
infants with PDA are exposed to multiple risk factors for lung
injury (ie, MV, O
2
exposure, infection, inammation), these
factors are important confounders in the association of PDA
and increased risk for BPD.
4,5,14
The approach to PDA treatment, specically treatment
modalities and treatment schedules, in relation to long-
term pulmonary outcome remains in question. In the large
prospective study of Schmidt et al, prophylactic treatment
with indomethacin had no effect on BPD, assessed as O
2
de-
pendence at 36 weeks post-conception.
14
Focusing on PDA
treatment rather than prophylaxis, Clyman conducted
a meta-analysis of several small trials of early compared
with late pharmacologic PDA therapy and reported that
infants who received early treatment for PDA were at lower
risk for BPD and reduced duration of MV when compared
with infants receiving delayed treatment.
6
None of these stud-
ies was specically designed to examine BPD as the primary
outcome. More recently, Van Overmeire et al administered
indomethacin either early (day 3 0.5) or late (day
7 1.7) in relatively large premature infants (approximately
29 weeks, approximately 1200 g) and found no signicant
differences in respiratory outcome between early and late
groups.
7
However, in this study, infants in the late treatment
group had lower FiO
2
and mean airway pressure require-
ments from days 3 through 8 in infants born before 28 weeks
of gestation. Most recently, Benitz reviewed contemporary
literature and strongly concluded that there was no denitive
evidence of benet from medical or surgical PDA closure in
Table III. Respiratory outcomes assessed from birth until day 28, 36 weeks PMA, and discharge or death
All infants Birth weight 500-800 g Birth weight 801-1250 g
Early (n = 54) Expectant (n = 51) Early (n = 24) Expectant (n = 24) Early (n = 30) Expectant (n = 27)
Days on O
2
rst 28 days 21 (7-27) 19 (526) 25 (14-27) 23 (14-27) 18 (6-26) 14 (2-24)
Total days on O
2
39 (9-133) 37 (10-97) 85 (17-146) 57 (26-119) 31 (7-62) 23 (4-59)
Days on MV rst 28 days 10 (1-28) 8 (1-28) 24 (7-28) 23 (7-28) 3 (0-12) 4 (0-15)
Today days on MV 12 (1-68) 13 (1-59) 45 (5-76) 35 (7-68) 5 (1-18) 6 (0-21)
On O
2
at 36 weeks PMA 17 (33) 16 (33) 13 (59) 10 (46) 4 (13) 6 (22)
Death or O
2
at 36 weeks PMA 19 (35) 18 (35) 15 (63) 12 (50) 4 (13) 6 (22)
On O
2
>30% at 36 weeks PMA* 9 (17) 2 (4) 7 (32) 2 (9) 2 (7) 0 (0)
Death or O
2
>30% at 36 weeks PMA

11 (20) 4 (8) 9 (38) 4 (17) 2 (7) 0 (0)

PMA, postmenstrual age.
Results expressed as median (10th-90th percentile) or number (%).
Variables O
2
at 36 weeks PMA and O
2
>30% at 36 weeks PMA as percent of all infants alive at assessment age.
*0.17 (0.03-0.88) expectant compared with early group, Mantel-Haenszel common OR estimate.
0.27 (0.07-0.99) expectant compared with early group, Mantel-Haenszel common OR estimate.
Table IV. Mortality and morbidity
All infants Birth weight 500-800 g Birth weight 801-1250 g
Early (n = 54) Expectant (n = 51) Early (n = 24) Expectant (n = 24) Early (n = 30) Expectant (n = 27)
Death to discharge 4 (7) 6 (12) 4 (17) 5 (21) 0 (0) 1 (4)
Postnatal steroids 6 (11) 5 (10) 5 (21) 3 (13) 1 (3) 2 (7)
Pneumothorax 4 (8) 3 (6) 2 (8) 2 (8) 2 (7) 1 (4)
Pulmonary interstitial emphysema* 9 (17) 2 (4) 8 (33) 2 (8) 1 (3) 0 (0)
IVH grades III and IV 6 (11) 7 (14) 5 (21) 6 (25) 1 (3) 1 (4)
PVL 4 (8) 3 (6) 3 (13) 2 (8) 1 (3) 1 (4)
NEC (requiring surgery) 5 (9) 2 (4) 4 (17) 1 (4) 1 (3) 1 (4)
Spontaneous intestinal perforation 2 (4) 4 (8) 1 (4) 3 (13) 1 (3) 1 (4)
Sepsis 22 (42) 23 (45) 13 (54) 13 (54) 9 (31) 10 (37)
ROP stage $3

7 (14) 4 (8) 7 (32) 4 (18) 0 (0) 0 (0)

IVH, intracranial hemorrhage; PVL, periventricular leukomalacia; ROP, retinopathy of prematurity.
Results expressed as number (%).
*0.16 (0.03-0.85) Expectant vs. Early, Mantel-Haenszel common OR estimate.
Analyzed as percent of all infants alive at assessment age.
June 2012 ORIGINAL ARTICLES
Timing of Patent Ductus Arteriosus Treatment and Respiratory Outcome in Premature Infants: A Double-Blind
Randomized Controlled Trial
933
premature infants.
2
Few, if any, of the previously reported tri-
als of early compared with late medical treatment for PDA
have actually delayed PDA treatment after 7 days. Thus, of
major importance is determining whether an early treatment
approach to PDA (as soon as mild PDA symptoms appear)
would improve respiratory outcome compared with an
expectant approach (either awaiting spontaneous PDAclo-
sure or treatment at the presentation of hemodynamic signif-
icance), which was the purpose of this study.
When this trial was designed to answer the aforemen-
tioned question, it was hypothesized that early treatment
would improve respiratory outcome (primary endpoint:
days of O
2
dependence during the rst 28 days) compared
with expectant treatment. The study was powered to de-
tect a 20% difference of days of O
2
dependence during the
rst 28 days (indicative of early evolution to BPD) in
groups. By using earlier data from our population, we esti-
mated that 84 infants per group (168 total) would be needed
and that the study could be completed within a 3-year pe-
riod. Ibuprofen (NeoProfen) was selected as the medical
therapy because of similar efcacy prole, but reduced renal
adverse effects compared with indomethacin.
15
We analyzed
the enrolled infants at study termination and found no dif-
ferences in days of O
2
dependence during the rst 28 days,
BPD (dened as O
2
requirement at 36 weeks), or death or
BPD in the study groups, both analyzed as a whole or as sep-
arate stratied weight groups, although there was a higher
number of infants on O
2
>30% at 36 weeks in the early
treatment group. This nding was contrary to our hypothe-
sis that early treatment would be benecial for respiratory
outcome. There were no major differences in other compli-
cations of prematurity. Additional statistical analysis and
extrapolation with the effect size demonstrated by the en-
rolled infants (not reported here) indicated that no statisti-
cal difference would be present in respiratory outcomes
(days of O
2
dependence during the rst 28 days, BPD at
36 weeks) if the study enrollment had continued to its initial
goal of 168 infants. The study specically excluded infants
whose initial PDA presentation was that of hemodynamic
signicance. Randomizing these infants to early placebo
was not feasible because of the possibility that delayed treat-
ment could have been detrimental in this sicker group of in-
fants. Thus, the ndings of this study cannot be directly
extrapolated to those infants who have early HS PDA as de-
ned here.
The strengths of this study include having a high number
of enrolled infants (despite early termination) compared
with earlier studies and the lowest mean gestational age
and birth weight (approximately 25-26 weeks, approxi-
mately 850 g) of the study population. We blinded the in-
tervention so that clinicians were kept unaware of the study
group of each infant. In addition, the study resulted in
a spread of treatment of 8 days (rather than #4 in earlier
studies) between early ibuprofen (which in this study
was at day 3-4) compared with expectant management
(only 20% needed open label ibuprofen for HS PDA, at
a median of day 11). Also, this was a single site study,
designed with input from a group of neonatologists whose
clinical approach to the premature infant was largely consis-
tent. Finally, this study allowed assessment of spontaneous
PDA closure rate, found to be 49% in this very low birth
weight population of infants. A major weakness, as afore-
mentioned, was the premature study termination before
reaching the enrollment goal because of ibuprofen unavail-
ability. The choice of ibuprofen, rather than indomethacin,
might also be considered a weakness of this study. Despite
data indicating similar rates of PDA closure with either
drug, a recent meta-analysis demonstrated an increased
risk of BPD with ibuprofen compared with indomethacin
treatment.
16
Taking the strengths and weaknesses in account, the
study results suggest a lack of benet from early PDA treat-
ment at the onset of mild clinical signs compared with delay
of PDA treatment until the onset of clear hemodynamic
signs. On the basis of these results, plus the nding of
a high spontaneous PDA closure rate, delaying therapy until
a HS PDA is present could prevent unnecessary and poten-
tially toxic intervention for PDA in critically ill premature
infants. n
The authors acknowledge the excellent assistance of Dr Deepthi Ala-
pati and research associates Carmen DUgard and Sylvia Vanbus-
kirk, the biostatistical support of Dr Hua Li and Dr Shari
Messinger, the pharmacy support of Kamal Behbahani and Jorge
Chivite, the cardiology support of Dr Sethuraman Swaminathan
and the neonates, families, faculty, and staff of the Project: NewBorn
NICU/Holtz Childrens Hospital of the University of Miami/Jackson
Memorial Hospital.
Submitted for publication Sep 16, 2011; last revision received Nov 23, 2011;
accepted Dec 16, 2011.
Reprint requests: IleneR. S. Sosenko, MD, Department of Pediatrics/Neonatology
(R131), University of Miami Miller School of Medicine, POBox 016960, Miami, FL
33101. E-mail: isosenko@miami.edu
References
1. Clyman RI. Mechanisms regulating the ductus arteriosus. Biol Neonate
2006;89:330-5.
2. Benitz WE. Treatment of persistent patent ductus arteriosus in preterm
infants: time to accept the null hypothesis? J Perinatol 2010;30:241-52.
3. Brown ER, Stark A, Sosenko IR, Lawson EE, Avery ME. Bronchopulmo-
nary dysplasia: possible relationship to pulmonary edema. J Pediatr
1978;92:982-4.
4. Rojas MA, Gonzalez A, Bancalari E, Claure N, Poole C, Silva-Neto G.
Changing trends in the epidemiology and pathogenesis of neonatal
chronic lung disease. J Pediatr 1995;126:605-10.
5. Gonzalez A, Sosenko IRS, Chandar J, Hummler H, Claure N,
Bancalari E. Inuence of infection on patent ductus arteriosus and
chronic lung disease in premature infants weighing 1000 g or less. J Pe-
diatr 1996;128:470-8.
6. Clyman RI. Recommendations for the postnatal use of indomethacin: an
analysis of four separate treatment strategies. J Pediatr 1996;128:601-7.
7. Van Overmeire B, Van de Brock H, Van Laer P, Weyler J,
Vanbaesebrouck P. Early versus late indomethacin treatment for patent
ductus arteriosus in premature infants with respiratory distress syn-
drome. J Pediatr 2001;138:205-11.
8. OBrien PC, Fleming TR. A multiple testing procedure for clinical trials.
Biometrics 1979;35:549-56.
THE JOURNAL OF PEDIATRICS www.jpeds.com Vol. 160, No. 6
934 Sosenko et al
9. Koch J, Hensley G, Roy L, Brown S, Ramaciotti C, Rosenfeld CR. Prev-
alence of spontaneous closure of the ductus arteriosus in neonates at
a birth weight of 1000 g or less. Pediatrics 2006;117:1113-21.
10. Mata-Greenwood E, Meyrick B, Soifer SJ, Fineman JR, Black SM.
Expression of VEGF and its receptors Flt-1 and Flk-1/KDR is altered
in lambs with increased pulmonary blood ow and pulmonary hyper-
tension. Am J Physiol Lung Cell Mol Physiol 2003;285:L222-31.
11. McCurnin D, Seidner S, Chang LY, Waleh N, Ikegami M, Petershack J,
et al. Ibuprofen-induced patent ductus arteriosus closure: physiologic,
histologic, and biochemical effects on the premature lung. Pediatrics
2008;121:945-56.
12. Husain AN, Siddiqui NH, Stocker JT. Pathology of arrested acinar devel-
opment in postsurfactant bronchopulmonary dysplasia. Human Pathol
1998;29:710-7.
13. Thebaud B, Abman SH. Bronchopulmonary dysplasia: where have all the
vessels gone? Role of angiogenic growth factors in chronic lung disease.
Am J Respir Crit Care Med 2007;175:978-85.
14. Schmidt B, Davis P, Moddemann D, Ohlsson A, Roberts RS, Siagal S,
et al. the TIPP investigators. Long term effects of indomethacin prophy-
laxis in extremely low birth weight infants. New Engl J Med 2001;344:
1966-72.
15. Katakam LI, Cotton CM, Goldberg RN, Dang CN, Smith PB. Safety
and effectiveness of indomethacin versus ibuprofen for treatment of
the patent ductus arteriosus. Am J Perinatol 2010;27:425-9.
16. Jones LJ, Craven PD, Attia J, Thakkinstian A, Wright I. Network
meta-analysis of indomethacin versus ibuprofen versus placebo for
PDA in preterm infants. Arch Dis Child Fetal Neonatal Ed 2011;
96:F45-54.
June 2012 ORIGINAL ARTICLES
Timing of Patent Ductus Arteriosus Treatment and Respiratory Outcome in Premature Infants: A Double-Blind
Randomized Controlled Trial
935
Figure 1. Study ow diagram.
Figure 2. Consolidated Standards of Reporting Trials
(CONSORT) diagram.
THE JOURNAL OF PEDIATRICS www.jpeds.com Vol. 160, No. 6
935.e1 Sosenko et al