Controlled trial of a single dose of synthetic surfactant at birth in premature infants weighing 500 to grams

699

David Stevenson, MD, Frans Walther, MD, Walker Long, MD, M a t t h e w Sell, MD, Thomas Pauly, MD, Al ice Gong , MD, David Easa, MD, Arun Pramanik, MD, Michae l LeBlanc, MD, Endla Anday , MD, Ramasubba reddy Dhani reddy, MD, David Burchfield, MD, An thony Corbe t , MB, FRACP, and t he Amer i can Exosurf Neona ta l Study Group I (See A p p e n d i x I)*

From the collaborating centers of the American Exosurf Neonatal Study Group I (see Appen- dix I), Department of Pediatrics, University of North Carolina at Chapel Hill and the Clinical Re- search Division, Burroughs Wellcome Co., Research Triangle Park, North Carolina

In a multicenter, double-bl ind, p lacebo-cont ro l led trial conduc ted at 23 hospi- tals in the United States~ a single prophylact ic 5 ml/kg dose of a synthetic sur- factant (Exosurf Neonatal) or air p lacebo was administered shortly after birth to 215 infants with birth weights of 500 to 699 gm. Despite stratif ication at entry by birth weight and gender, by chance female infants predominated in the air pla- cebo group and male infants predominated in the surfactant group. Among in- fants receiving synthetic surfactant, improvements in oxygen requirements were signif icant at 2 hours after birth (p = 0.014) and persisted for 3 days (p = 0.001); improvements in the alveolar-arterial partial pressure of oxygen gradient were signif icant at 6 hours after birth (p = 0.01) and persisted for 3 days (p = 0.008). Improvements in mean airway pressure were not signif icant at 2 or at 6 hours af- ter birth (p = 0.622 and 0.083, respectively), but became signif icant thereafter and persisted for 3 days ( p = 0.002). Pneumothorax was reduced by slightly more than half (25 vs 11; p = 0.014); death from respiratory distress syndrome (RDS) was also reduced (26 vs 15; p = 0.046). Overall neonatal mortality, how- ever, was not signif icantly reduced (58 vs 46; p = 0.102). Other compl icat ions of RDS and prematurity were not altered, except that pulmonary hemorrhage oc- curred signif icantly more frequently in infants receiving synthetic surfactant (2 vs 12; p = 0.006). These findings indicate that a single prophylact ic dose of syn- thetic surfactant in infants weighing 500 to 699 gm at birth improves lung func- tion, inc idence of air leak, and death from RDS but not overall mortality. The only safety problem identi f ied was an increase in pulmonary hemorrhage. (J PEDIATR 1992;120:$3-$12)

Supported by funds from Burroughs Wellcome Co., Research Tri- angle Park, N.C.

Reprint requests: Anthony Col-bet, MB, FRACP, Department of Pediatrics, Emory University School of Medicine, 2040 Ridgewood Drive, Atlanta, GA 30322.

*In addition, two members of the American Exosurf Neonatal Study Group II participated: David Easa, MD, Kapiolani Medical Center, University of Hawaii, Honolulu, Hawaii, and Arun Pramanik, MD, Louisiana State University Medical Center, Shreveport, La. 9/0/34285

$3

S 4 Stevenson et al. The Journal of Pediatrics February 1992

A-a BPD IVH MAP PDA Po2 RDS

Alveolar-arterial Bronchopulmonary dysplasia Intraventricular hemorrhage Mean airway pressure Patent ductus arteriosus Partial pressure of oxygen Respiratory distress syndrome

Substantial data from randomized trials document that both prophylactic 1-5 and rescue 6-8 treatment with surfac-

tant reduces the morbidity and mortality of premature in- fants. Whether prophylactic or rescue treatment is superior has been controversial, 9' ~0 but recent data from a large, open, but randomized trial indicate that multiple-dose pro- phylactic treatment is superior to multiple-dose rescue treatment in infants less than 30 weeks gestation. 11 A sub- analysis of those data indicated that the greatest reduction in mortality occurred in infants of 24 to 26 weeks gestation.11

In all three of these trials, however, the mean birth weights of enrolled infants exceeded 900 gm, and in two of the trials, 9' ~0 infants of less than 24 weeks were apparently not eligible. As part of a program designed to examine sys- tematically the risks and benefits of the synthetic surfactant Exosurf Neonatal in all premature infants, we have studied both its prophylactic and rescue use in extremely immature infants--those at the edge of viability. Such studies are im- portant, because in the absence of data, clinicians are likely to extrapolate favorable risk to benefit ratios from other premature infant populations 18, 1~ to extremely immature infants, a particularly fragile and vulnerable group.

This report presents results of the first completed trial of Exosurf Neonatal in extremely immature infants. It was a randomized, double-blind comparison of single prophylac- tic doses of air and drug given shortly after birth to 215 in- fants weighing 500 to 699 gm. A preliminary report has ap- peared elsewhere. 13

M E T H O D S

The trial was conducted with a common protocol at 23 hospitals in the United States; participants are listed in the Appendix. Except for birth weights targeted, sample size, and hypothesis to b e tested, the design of this trial was identical to another single-dose prophylactic trial reported recently. 5 The two trials were initiated simultaneously, but the current trial took longer to reach targeted sample size.

Surfaetant preparation. The material used was Exosurf Neonatal (Burroughs Wellcome Co., Research Triangle Park, N.C.), which is a synthetic surfactant invented by Clements14; a detailed description is available elsewhere. 8 In brief, Exosurf Neonatal is a mixture of dipalmitoylphos- phatidylcholine, hexadeeanol, and tyloxapol, which has been formulated with sodium chloride as a lyophilized

powder. Exosurf Neonatal is packaged under vacuum in 10 ml glass vials and stored at room temperature. Immedi- ately before use, 8 ml of sterile, preservative-free water is added to the vial. The resulting suspension contains 13.5 mg/ml of dipalmitoylphosphatidylcholine, 1.5 mg/ml of hexadecanol, and 1.0 mg/ml of tyloxapol in 0.1N sodium chloride. Exosurf Neonatal reduces surface tension to low levels in vitro on both the Wilhelmy balance and a modified pulsating bubble surfactometer 15 and improves lung func- tion and survival in premature animal models of respiratory distress syndrome) 5, 16

Patient recruitment and enrollment, inborn infants weigh- ing 500 to 699 gm at birth were eligibl e for the trial at each hospital. Consent for enrollment was sought from all moth- ers, except those with prenatal exclusion criteria, considered likely to deliver a live infant in this weight range. The ex- clusion criteria were as follows: (1) proven fetal lung ma- turity, (2) known fetal malformation or chromosomal abnormality, (3) fetal growth retardation, (4) hydrops re- talis, (5) purulent amnionitis, (6) maternal heroin addic- tion, and (7) obstetric decision not to support the fetus. Liveborn infants for whom consent was obtained were weighed immediately at birth to determine eligibility but were excluded if postnatal exclusion criteria were present. These three criteria were major malformation, three or more minor anomalies, and evidence of hydrops fetalis.

Enrolled infants were stratified by weight and gender into four groups: (1) 500 to 599 gm males, (2) 500 to 599 gm females, (3) 600 to 699 gm males, and (4) 600 to 699 gm females. Within each stratum infants were randomly as- signed to receive either Exosurf Neonatal or air placebo as quickly as possible after birth. Randomization was per- formed by computer-generated random numbers sealed in sequentially numbered, opaque envelopes.

Infants entering the trial had tracheal intubation by the clinical team in the delivery room and were placed on me- chanical ventilation. The settings were standardized for the first 5 minutes of resuscitation: rate 40/rain, peak pressure 25 cm HzO , expiratory pressure 5 cm H20, inspiration time 0.4 second, and 80% oxygen. Peak pressure could be adjusted to achieve satisfactory chest motion, and oxygen adjusted to abolish cyanosis. After 5 minutes the settings were determined by the attending physician.

Surfactant administration. Exosurf Neonatal or air pla- cebo was administered by a separate dosing team, whose members did not reveal the treatment administered and did not participate in subsequent clinical care. A screen was placed around the infant and dose administration occurred in secrecy. After t he addition of water to the Exosurf Neonatal vial and vigorous agitation, the suspension was aspirated from below the froth into a plastic syringe. With the patient supine, 5 ml/kg body weight was administered

Volume 120 Single-dose synthetic surfactant $ 5 .Number 2, Part 2

T a b l e I. Derivation of population enrolled in trial

No. of infants

Eligible by birth weight Enrolled and randomized Infants not entered

Failure to meet inclusion/ 84 exclusion criteria Patient deemed unstable 3 Outborn 63 Prenatal decision to transfer out 2 Parents refused consent 11 Precipitous delivery/missed 133 Total infants not entered

511 215

296

NO. not d o s e d *

2 5

T a b l e I I . Pat ient accountabi l i ty

No. dosed Randomization No. dosed with Exosurf group/analysis with air Neonatal

Air (n = 106) 102 2 Exosurf (n = 109) 0 104 Efficacy analyses

"Intent to treat" 106 109 Primary ("as 103 97

treated")? Safety analyses

"As treated," 109 106 no exclusions

*Patients not given anything grouped with air recipients for the primary ef- ficacy analysis and the safety analyses. ?As treated, after exclusion of patients with unrecognized congenital pneu- monia or major malformations at the time of entry.

by direct instil lation into the t rachea through a special

side-ported endotracheal tube adapter without in terrupt ion

of mechanica l ventilation. Ha l f the dose was given over a 1-

to 2-minute period, and the pat ient was turned 90 degrees

to one side for 30 seconds; the second hal f was then given,

and the pa t ient was turned to the other side. The dosing

team remained in seclusion with the pat ient for 10 minutes

after complet ion of the dose. Care was then re turned to the

clinical team. Suct ioning the endotracheal tube was pro-

scribed by the protocol for 2 hours af ter dose adminis t ra-

tion. The dosing team did not reveal what was instilled for

a m i n i m u m of 2 years af ter completion of enrol lment in the

trial.

Data collection. Data for each enrolled infant were col-

lected by a nurse coordinator and a neonatologist. Informa-

tion was recorded for each mother ' s demographic profile,

medical and obstetric history, and pregnancy, labor, and

delivery. Vent i la tor and oxygen requirements and ar ter ial

blood gas values were recorded at 2, 6, 12, and 24 hours af-

T a b l e I I I . Character is t ics of enrolled infants

Exosurf Air p l a c e b o Neonata l

(n = 109) (n = 106)

Birth weight* (gin) 613 _+ 55 618 _+ 56 Gestational age* (wk) 24.8 + 1.7 24.9 _+ 1.7 Race

White 53 54 Black 39 35 Hispanic 13 12 Other 4 5

Gender Male 53 61 Female 56 45

Gender/weight strata (gin) Male, 500-599 15 19 Female, 500-599 22 19 Male, 600-699 38 42 Female, 600-699 34 26

Delivery Abdominal 46 43 Vaginal 63 63

Multiple gestation 27 24 Membranes ruptured <1 hr 53 58

before delivery Prenatal steroids >24 hr 11 18 Apgar score at 1 mint 3 3 Apgar score at 5 mint 6 6

*Mean _+ SD. "~Median.

T a b l e IV. Baseline venti lator settings and oxygen

requirements before single dose of Exosurf Neonata l or

air placebo

Air Exosurf p lacebo Neonatal (n = 109) (n = 106)

Age at first dose 33 _+ 30 32 _+ 19 (min)

Peak inspiratory pressure 26 _+ 4 27 _+ 6 (cm H20)

Positive end-expiratory 4.9 _+ 2.4 4.7 _+ 0.8 pressure (cm H20)

Ventilator rate 45 + 12 44 + 11 (beats/min)

% inspired 83 _+ 16 85 _+ 14 oxygen

Values are mean _+ SD.

ter bir th and at the bi r th anniversary t ime for days 2

through 28.

For each set of blood gas values, the value for alveolar

oxygen tension was calculated by the formula (713 X

F I O 2 ) - Pco2, except at the single high al t i tude center,

S 6 Stevenson et al, The Journal of Pediatrics February 1992

o~

oy LL

~3 03

E s CD

0

0

-10

-20

-30

Hours Post-Dose 2 6

2 p=0.002O

0.2

0 "F

0.2

n -0.4 <

-0.6

-0.8

ca 1.0 K o -1.2 u_

-1.4

-1.61 O

-1.8~

Hours Post-Dose 2 6

45O

~-~ 350

I ~ 300

p=0 6223 i O~ a_ 250 , <

~[~ 200

0 p=0.0828

p=0.2519

i p=0.0102

[

2 6 Hours Post-Dose

Fig, I. Onset of action. Mean changes (+SE) in oxygen requirements (left panel), MAP (middle panel), and A-a Po2 gradient (right panel) at 2 and 6 hours after administration (first and second time points at which data were collected) for infants randomly assigned to receive a single prophylactic dose of air placebo (open bars) or Exosurf Neonatal (hatched bars). Note that significant improvements in oxygen requirements were present by 2 hours and significant improvements in A-a Po2 gradients were present by 6 hours.

Toble V. Air leak

Air p l a c e b o Exosurf N e o n a t a l (n = 109) (n ~ 106) Re la t ive risk 95% Cl

Pneumothorax 25 1 l 0.456* 0.244, 0.852 Pneumomediastinum 2 1 0.526 0.042, 6.671 Pulmonary interstitial emphysema 46 47 1.045 0.769, 1.420 Pneumopericardium l 4 4.061 0.612, 26.94 Any form of air leak 56 5l 0.932 0.712, 1.218

Relative risk <1 indicates decreased risk with Exosurf Neonatal. CI, Confidence intervals. *p = 0.014.

where local mean barometric pressure minus 47 (for water vapor pressure) was substituted for 713.

Chest radiographs were performed on days 1, 7, and 28. Attending clinicians evaluated those for days 1 and 7; a pe- diatric radiologist (F. Volberg, MD, Bowman Gray School of Medicine, Winston-Salem, N.C.) evaluated those for day 28 and assigned a score based on the Edwards' classifica- tion of the severity of bronchopulmonary dysplasia. 17

Definition of RDS and BPD. The diagnosis of RDS was made at 7 days of age by review of the hospital course. In- fants who died less than 12 hours after birth were consid- ered to have had RDS if death was attributed to RDS by a neonatologist or by an autopsy examination. For those alive at 12 hours, all the following criteria had to be met: (1) after the age of 12 hours 30% or more oxygen was re- quired to maintain arterial Po2 more than 50 mm Hg; (2) after the age of 12 hours a mean airway pressure of 6 cm H20 or more was required; (3) the chest radiograph at age

24 hours was considered to be consistent with RDS by the attending physician; and (4) after complete evaluation, the lung disease was not attributable to another cause. Tran- sient tachypnea was diagnosed in those infants with lung disease requiring less oxygen and mean airway pressure af- ter age 12 hours than was required for infants with RDS. Congenital pneumonia was diagnosed in the presence of lung disease at birth with a positive blood culture.

The diagnosis of BPD was made by analysis of the infant's condition after 28 days of age. A combination of radiographic 17 and clinicaP 8 criteria was used. The follow- ing criteria had to be met: (1) the presence of tachypnea and retractions, (2) the need for supplemental oxygen to main- tain the skin surface Po2 above 60 mm Hg or the pulse oxi- meter reading at more than 85% saturation, and (3) chest radiographic abnormalities rating a score of 4 or more by the Edwards' classification. Among infants who died before 28 days and had autopsies, BPD was diagnosed by the his-

Volume 120 Single-dose synthetic surfactant S 7 Number 2, Part 2

tologic criteria of Bonikos et a1.19 If infants died between 10 and 28 days and did not have an autopsy, BPD was diag- nosed if, at the time of death, they required more than 60% oxygen and 7 cm H20 MAP, if they had persistently abnormal chest radiographs, and if there was no other ex- planation for respiratory failure.

Intraventrleular hemorrhage. A l l infants had cranial ultrasonography performed between days 5 and 7 and repeated as necessary. Copies of all cranial ultrasound studies performed in the first 28 days of life were forwarded to the same pediatric radiologist (F. Volberg, MD) for evaluation. The results were classified as subependymal hemorrhage (grade 1 IVH), IVH without ventricular dila- tion (grade 2 IVH), IVH with ventricular dilation (grade 3 IVH), and periventricular echodensities.

Patent duetus arteriosus. Criteria were established for the diagnosis of PDA, and its presence or absence was recorded each day for 28 days. The diagnosis of PDA was made in the presence of a continuous murmur or a systolic murmur if the infant was on mechanical ventilation. Patients with- out a murmur but mechanically ventilated were assigned this diagnosis if they had any one of the following: hyper- active precordium, bounding pulses, pulse pressure greater than 30 mm Hg, or hepatomegaly. Regardless of ventilation status or the presence or absence of a murmur, PDA was diagnosed if there was cardiomegaly or pulmonary vascu-

lar congestion on the chest radiograph. When available, echocardiography was used to confirm this diagnosis. The criteria were direct visualization or Doppler detection of turbulent flow in the pulmonary artery or retrograde flow in the descending aorta. If the diagnosis of PDA was made, the protocol mandated treatment with at least three doses of

indomethacin. Efficacy outcome measures. The principal efficacy out-

come variable was prospectively designated as death at age 28 days. Secondary efficacy outcomes were death from RDS, occurrence of BPD, and survival without BPD. The protocol specified two analyses of efficacy: a "primary" analysis, in which patients were analyzed as actually treated and excluding infants who had unrecognized congenital pneumonia or major malformation at entry, and an "intent to treat" analysis, which included all patients as random-

ized. Safety outcome measures. The protocol stated that all

other outcome variables would be considered as safety measures, and patients would be analyzed as actually treated without any exclusions. As in all other North American trials of Exosuff Neonatal, the principal safety outcome for this trial was severe IVH, which was defined as grade 3 IVH and/or periventricular echodensities. An independent advisory panel, described in detail elsewhere, s blindly monitored the incidence of severe IVH in the two

1 Air

70 7 ~ 60

5O O

4O

30

80

E 400 ._~ " o

L9 d 300 a . ca

200

I I I I I I I I 20 0 1 2 3 4 5 6 7

Days

500 �9 Air

I I I I I I I I 100 0 1 2 3 4 5 6 7

Days

12

11 ~ �9 Air

4 1 - I I I I I I I I 0 1 2 3 4 5 6 7

Days

F i g . 2. Duration of action. In top, middle, and lower panels, re- spectively, mean oxygen concentrations (in percentages), mean A-a Po2 gradients (in mm Hg), and MAP (in cm H20) (all_+ SE) for the first 7 days of life in infants randomly assigned to receive a sin- gle prophylactic dose of air placebo (dark squares) or Exosurf Neonatal (dark circles) are shown. First-time point plotted is at time zero (immediately before administration) in top and bottom panels. First-time point is 2 hours after birth in middle panel (be- cause blood gases were not required before enrollment). Note that improvements in oxygen concentrations (p=0.001), MAP (p = 0.001), and A-a Po2 gradients (p = 0.001) after a single pro- phylactic dose of Exosurf Neonatal lasted for at least 3 days.

groups. The panel was empowered to break the blind if the incidence of severe IVH was significantly higher in one group and to stop the trial if that group proved to be the one receiving Exosurf Neonatal.

Except for pulmonary hemorrhage, which was identified prospectively as a safety outcome for incidence comparison

S 8 Stevenson et al. The Journal of Pediatrics February 1992

1.0

> . m

>

0.9

0.8

0.7

0.6

0.5

0.4

0.3

0.2

0.1

0.0

T r e a t m e n t G r o u p

- Air

. . . . . = E x o s u r f

1

I I I I

0 50 100 1;0 200 2;0 3;0 3;0 4C)0 450 5()0 Study Day

rival in the Exosurf Neonatal group appeared to be maintained

F i g . 3 . Kaplan-Meier survival curves through l-year adjusted age in infants treated with a single prophylactic dose of air placebo (solid line) or Exosurf Neonatal (broken line). Infants with unrecognized congenital pneumonia or major anomalies at the time of entry are excluded. Note that survival was poor in both groups and that the modest gains in sur-

over time.

but not defined, all other safety outcomes (e.g., apnea, hy- potension, sepsis, and nosocomial pneumonia) were defined prospectively; definitions can be found elsewhere. 8 In addi- tion, the use of a number of medications, such as pancuro- nium, methylxanthines, antibiotics, diuretics, anticonvul- sants, sedatives, and inotropes, was recorded.

Cause of death. If death occurred, the primary cause was determined by a blinded investigator taking into account the clinical history. For patients who died and underwent post- mortem examination, the autopsy cause of death was sub- stituted for the clinically determined cause of death in the

analyses to follow. Follow-up phase. All infants were followed throughout

the hospital course. After discharge, contact was main- tained by correspondence, telephone, and outpatient clinic visits until the ages of 1 and 2 years. At these times perti- nent information with regard to hospital admissions, surgi- cal operations, oxygen and ventilator use, and retinal examinations was obtained, and a complete physical exam- ination and neurodevelopmental assessment were per- formed. These results will be reported separately, except the survival data, which are included herein.

Statistical methods. When the trial was planned, it was calculated that 198 infants, half in each group, would be required to provide a 90% chance of detecting a 40% improvement in the mortality rate of surfactant-treated in-

fants. 2~ Survival was displayed graphically with the Kaplan-Meier procedure. Efficacy outcomes were analyzed with the Cochran-Mantel-Haenszel test, which controlled for the four randomization strata, el Oxygen and MAP re-

quirements and alveolar-arterial Po2 gradients at 2 and 6 hours after administration (the first two time points at

which such data were collected) were compared by non- paired t tests of change from baseline. Oxygen and ventila- tor requirements and A-a Po2 gradients for the first 72 hours after entry were compared by calculation of area under the curve for the first 72 hours after log transformation to achieve a normal distribution and then by analysis of cova- riance with the baseline value as covariate. 22 Differences were considered significant at p < 0.05. Differences in safety outcomes were analyzed by 95% confidence intervals; for those outcomes for which the 95% confidence interval did not overlap 1 (i.e., safety outcomes that were statisti- cally different), Cochran-Mantel-Haenszel tests were used to determine p values for the observed difference.

R E S U L T S

The trial was conducted at 23 American hospitals for 26 months (Oct. 10, 1987, to Dec. 15, 1989). This report is limited to the first 28 days after birth, except that the mortality rate through 1 year of adjusted age is also in-

cluded.

Volume 120 Single-dose synthetic surfactant S 9 Number 2, Part 2

T a b l e Vl. Causes of death for all patients as treated

Cause

"As treated, no exclusions"

Air Exosurf p lacebo Neonatal (n = 109) (n = 106)

Days l to 28 RDS 26 15" Pulmonary air leaks 7 6 IVH 2 7 BPD 3 l Necrotizing enterocolitis 0 2 Sepsis 5 4 Other 17 16

Total 60 51 Days 29 to 1 yr

Pulmonary air leaks l 0 IVH 1 1 BPD 2 7 Infection 2 4 Renal failure 3 0 Other 0 1 Unknown 0 1

Total 9 14

*p = 0.046.

Study accountability. During the enrollment period, 511

infants born at the participating hospitals were eligible by

weight. Of these, 296 infants were not enrolled for reasons

given in Table I; thus, 215 infants entered the trial; 106 in-

fants were randomly assigned to receive air placebo and 109

to receive Exosurf Neonatal (Table II). Two infants

randomly assigned to air placebo mistakenly received Ex-

osurf Neonatal, and 2 other infants randomized to air did

not receive anything. Five infants randomly assigned to Ex-

osurf Neonatal were not given either air placebo or Exosurf

Neonatal (Table II). A total of 15 infants had unrecognized

congenital pneumonia or an unrecognized major malfor-

mation at entry: 6 among those who received air placebo or

nothing and 9 among those who received Exosurf Neonatal.

For the as-treated analyses of efficacy and for the safety

analyses, the 7 infants not treated with anything (2 of whom

were randomly assigned to air placebo and 5 of whom were

randomly assigned to Exosurf Neonatal) were considered

air placebo recipients (Table II).

The independent advisory panel examined interim blinded

data from this trial on seven occasions during the study pe-

riod, and each time concluded that the study should

continue. Baseline comparabi l i ty /Despi te prospective stratifica-

tions by gender, by chance the air placebo group had a pre-

ponderance of female infants, whereas the Exosurf Neo-

natal group had a preponderance of male infants (Table

III). Also, by chance, the use of prenatal corticosteroids for

A

100 F A ~ A ~ A

80

r (9 ~ 6O s

"6 o~

40

20

0

Fig. 4. Incidence of death (black), survival with BPD (cross- hatched), and survival without BPD (dotted) by weight/gender stratum in infants treated with a single prophylactic dose of air placebo (A) or Exosurf Neonatal (E). Infants with unrecognized congenital pneumonia or major anomalies at the time of entry are excluded; infants alive but whose BPD status is unknown are indi- cated by white (white). Note that for males weighing 500 to 599 gin, the incidence of death is greater than 80% (A), irrespective of treatment. In the other three gender/weight strata, infants who re- ceived a single prophylactic dose of Exosurf Neonatal appeared to have increased survival. In the 500 to 599 gm female (B) and 600 to 699 gm male infants (C), the extra survivors in the Exosurf Neo- natal group appeared to have BPD, but in the 600 to 699 gm fe- male infants (D), the increase in survival was not accompanied by an increase in BPD. Thus, for 600 to 699 gm female infants (D), a single prophylactic dose of Exosurf Neonatal appeared to increase survival without BPD. (Strata: A, males, 501 to 599 gm, B, females, 501 to 599 gm, C, males, 600 to 699 gm, D, females, 600 to 699 gm.)

more than 24 hours before delivery was slightly more com-

mon among the Exosurf Neonatal group (11 vs 18). Oth-

erwise, the randomization procedure was successful. The air

placebo and active treatment groups were comparable in

birth weight, gestational age, route of delivery, and Apgar

scores (Table IID. Baseline oxygen requirements and ven-

tilator settings at the time of administration (30 minutes

after birth) were also equivalent in the two groups (Table

IV).

Respiratory distress syndrome. The incidence of RDS was

not reduced among infants receiving synthetic surfactant

(76 vs 78). A significant reduction in the concentration of

oxygen needed to maintain normal arterial Po2 was present

by 2 hours after birth, the first t ime point at which data were

recorded (Fig. 1, left panel, p = 0.014). Improvements in

the A-a Po2 gradient were significant by 6 hours after birth,

S 1 0 Stevenson et al. The Journal of Pediatrics February 1992

Table VII. Efficacy analyses

"Intent to t reat" "As t r e a t e d "

af ter exclusions*

Air p l a c e b o Exosurf N e o n a t a l Air p l a c e b o Exosurf N e o n a t a l (n = 106) (n = 109) p (n = 103) (n = 97) p

Mortality <10 days 50 48 0.320 52 39 0.079 Mortality --<28 days 56 55 0.322 58 46 0.102 Mortality -<l yr 65 69 0.839 66 59 0.491 BPD by 28 days (whether alive or dead) 23 27 0.765 22 24 0.799 BPD in 28-day survivors 17 26 0.210 16 23 0.433 Dead by 28 days or alive 73 81 0.794 74 69 0.567

with BPD through 28 days Day-28 survival without 30 27 0.794 26 27 0.567

BPD

*For this analysis, as specified in the protocol, 15 infants who had unrecognized congenital pneumonia or major malformations at entry were excluded.

Table VIII. Safety outcomes: Other complications associated with prematurity and RDS for all patients as treated

Air p l a c e b o Exosurf N e o n a t a l (n = 109) (n = 106) Relat ive risk 95% Cl

IVH 39 45 1.065 0.796, 1.425 PVED 5 12 2.511 0.865, 7.287 Grade 3 IVH/PVED 20 19 0.868 0.490, 1.540 Seizures 12 9 0.772 0.331, 1.797 PDA 53 57 1.085 0.839, 1.403 Necrotizing enterocolitis 2 4 2.215 0.419, 11.72 Sepsis 33 37 1.147 0.783, 1.679 Nosocomial pneumonia 11 11 0.965 0.417, 2.231 Meningitis 4 6 1.400 0.392, 5.006 Arterial hypotension 76 81 1.086 0.925, 1.276 Apnea of prematurity 37 35 1.008 0.699, 1.452 Pulmonary hemorrhage 2 12" 6.078 1.699, 21.74

Relative risk > l indicates increased risk with Exosurf Neonatal C1, Confidence intervals; CI not overlapping I indicates a statistically significant difference in the two groups; PVED, periventricular echodensities. *p = 0.006.

the second t ime point at which data were recorded (Fig. 1,

right panel, p = 0.010). Improvements in M A P were not

significant by 6 hours after birth (Fig. 1, middle panel, p = 0.082) but became significant thereafter. Improve-

ments in oxygen concentration, MAP, and A-a Po2 gradi-

ent were maintained during the first 3 days of life (Fig. 2,

p < 0.001). Reductions in peak inspiratory pressure and in-

termittent mandatory ventilation requirements were also

maintained through the first 3 days of life (data not shown,

p < 0.05). Pneumothorax was reduced by slightly more

than half (Table V: 25 vs 11, p = 0.014); death from RDS

was also reduced (Table VI: 26 vs 15, p = 0.046).

Overall mortality rate, BPD, and survival without BPD.

By the end of day 28, no significant differences in overall

mortality, incidence of BPD, or survival without BPD were

observed in the two groups (Table VII). Nevertheless, the

Kaplan-Meier survival curve appeared to indicate improved

survival in the surfactant group (Fig. 3). To study this pos-

sibility further, overall mortality, BPD, and survival with-

out BPD were compared in each of the four strata separately

(Fig. 4). Among males weighing 500 to 599 gm, death was

virtually universal regardless of treatment. In females

weighing 500 to 599 gm and males weighing 600 to 699 gin,

a single dose of synthetic surfactant appeared to be associ-

ated with a modest reduction in overall mortality but an in-

crease in BPD. In females weighing 600 to 699 gin, a single

dose of synthetic surfactant appeared to be associated with

a modest reduction in overall mortality without a change in

the incidence of BPD, which resulted in an improvement in

survival without BPD.

Causes of death. The primary causes of death, as deter-

mined by a blinded neonatologist at each hospital or by

Volume 120 Single-dose synthetic surfactant S 1 1 Number 2, Part 2

postmortem examination when available, are listed in Ta- ble VI. As mentioned, a significant reduction in the number of deaths attributed to RDS was observed among those treated with Exosurf Neonatal (p = 0.046), but otherwise the causes of death did not differ in the two groups.

Complications of prematurity and respiratory distress syndrome. The complications of prematurity and RDS, such as IVH and PDA (Table VIII), and the use of medications (data not shown) were unchanged by treatment with Exo- surf Neonatal. The sole exception was a significant increase in the number of infants with puhnonary hemorrhage (p = 0.006).

D I S C U S S I O N

The primary hypothesis tested in this trial, that a single prophylactic dose of Exosurf Neonatal administered shortly after birth to infants weighing 500 to 699 gm would reduce the neonatal mortality rate by 40%, was not proved true. This result stands in contrast to the findings of an identical trial conducted simultaneously in the same centers in infants weighing 700 to 1100 gin, in whom a 39% reduction in neonatal mortality (p = 0.022) and a 43% reduction in 1-year mortality (p = 0.002) were observed, s

This apparent failure could be related to several factors, including (1) the gender imbalance at baseline favoring the air placebo group; (2) a longer duration of surfaetant defi-

ciency in younger infants, necessitating more than one dose to see a significant effect on overall mortality rates; (3) in- adequate power of the study, particularly if immaturity in nonpulmonary organ systems affects overall mortality more strongly in this weight group than in larger infants (as is in fact likely the case); and (4) the type of surfactant used.

Although the 21% reduction in overall neonatal mortal- ity observed in the Exosurf Neonatal group was not statis- tically significant, it would be incorrect to conclude that a single prophylactic dose in this weight group is without clinical benefit. Physiologic benefit was evident within 2 hours of birth and was maintained through 3 days of age. The incidence of RDS (which occurred in more than 70% of both treated and control groups) was not reduced, but pneumothorax and death from RDS were reduced by approximately one half. in contrast, even in larger infants, no reductions in death from RDS (much less overall mor- tality) were observed after single prophylactic doses of var- ious mammalian surfactants. 2325

The only safety problem identified in this study was a significant increase in pulmonary hemorrhage in the Exo- surf Neonatal-treated group. In design of the trial, pulmo- nary hemorrhage was not prospectively defined; thus, infants with blood-tinged tracheal secretions and no clini- cal deterioration were categorized similarly to infants with

massive pulmonary hemorrhage, dramatic radiographic changes, and marked clinical deterioration. Two other ar- ticles in this supplement address the pathogenesis and pre- vention of this problem. In essence, the excess pulmonary

hemorrhage observed appears to be caused by enhanced left-to-right ductal shunting after surfactant replacement, which in some infants results in hemorrhagic pulmonary edema. This complication is likely preventable by early treatment with indomethacin. No infant in this study died of pulmonary hemorrhage.

No increases in sepsis, nosocomial pneumonia, or brain hemorrhage were observed in these extremely fragile

infants. The particularly poor outcome of male infants weighing

500 to 599 gm with or without surfactant might be expected. The next step will be to determine whether additional doses of surfactant and early attention to PDA closure can alter the currently dismal outcome of male infants in this birth weight group.

Ongoing studies of developmental outcome after surfac- rant replacement in these very tiny babies appear to indicate that at 1 year mild to moderate handicap is reduced in in- fants weighing less than 750 gm and Bayley scores are in- creased. 26 These preliminary observations, if confirmed,

would provide a strong rationale for the aggressive use of surfactant in this weight class even if overall mortality is not reduced. The latter possibility seems unlikely if more than one dose were administered, given that three prophylactic doses provide further reductions in overall mortality in comparison with one dose in 700 to I100 gm infants127 A randomized trial designed to explore the risk and benefits of up to six doses of Exosurf Neonatal in this weight group is

under way.

We thank the clinical coordinating staff at each participating hospital and the clinical monitoring staff at Burroughs Wellcome for careful conduct of this trial. We also thank Kate Esch and the Clinical Data Processing Department and Diane Wold, PhD, and the Department of Clinical Statistics at Burroughs Wellcome for excellent data processing and statistical efforts.

REFERENCES

1. Merritt TA, Hallman M, Bloom BT, et al. Prophylactic treat- ment of very premature infants with human Surfactant. N Engl J Meal 1986;315:785-90.

2. Ten Centre Study Group. Ten centre trial of artificial surfac- tant (artificial lung expanding compound) in very premature babies. Br Med J 1987;294:991-6.

3. Phibbs RH, Ballard RA, Clements JA, et al. Initial clinical trial of Exosurf, a protein-free synthetic surfactant for pro- phylaxis and early treatment of hyaline membrane disease. Pediatrics, 1991;88:1-9.

4. Bose C, Corbet A, Bose J, et al. Improved outcome at 28 days

S 1 2 Stevenson et al. The Journal of Pediatrics February 1992

of very Low birth weight infants treated with a single dose of synthetic surfactant. J PEDIATR 1990;117:947-53.

5. Corbet A J, Bucciarelli R, Goldman SA, Mammel MA, Wold D, Long WA. Decreased mortality in small premature infants treated at birth with a single dose of synthetic surfactant: a multicenter trial. J PZDIATR 1991;118:277-84.

6. Hallman M, Merritt TA, Javenpaa A-L, et al. Exogenous hu- man surfactant for treatment of severe respiratory distress syndrome: a randomized clinical trial. J PED1ATR 1985; 106:963-9.

7. Collaborative European Multieenter Study Group. Surfaetant replacement therapy for severe neonatal respiratory distress syndrome: an international randomized clinical trial. Pediat- rics 1988;82:683-91.

8. Long W, Thompson T, Sundell H, et al. Effects of two rescue doses of a synthetic surfactant on mortality rate and survival without bronchopulmonary dysplasia in 700 to 1350 gram in- fants with respiratory distress syndrome. J PEDIATR 1991; 118:595-605.

9. Dunn MS, Shennan AT, Zayack D, Possmayer F. Bovine sur- factant replacement therapy in neonates of less than 30 weeks' gestation: a randomized controlled trial of prophylaxis versus treatment. Pediatrics 1991;87:377-86.

10. Merritt TA, Hallman M, Berry C, et al. Randomized, placebo- controlled surfactant given at birth versus rescue administra- tion in very low birth weight infants with lung immaturity. J PEDIATR 1991 ; t l 8:581-94.

11. Kendig JW, Notter RH, Cox C, et al. A comparison of surfac- tant as immediate prophylaxis and as rescue therapy in new- born infants of less than 30 weeks gestation. N Engl J Med 1991;324:865-71.

12. Long W, Corbet A, Cotton R, et al., the American Exosurf Neonatal Study Group I, and the Canadian Exosurf Neonatal Study Group. A controlled clinical trial of a synthetic surfac- tant in infants weighing 1250 g or more with respiratory dis- tress syndrome. N Engl J Med 1991;325:1696-703.

13. Long W, Stevenson D, Pauly T, LeBlanc M, and the Ameri- can Exosurf Neonatal Study Group I. Effects of a single pro- phylactic dose of Exosurf neonatal in 215 500-700 gram infants [Abstract]. Pediatr Res 1991;29:223A.

14, Long WA, Sanders RL. New treatment methods in idiopathic respiratory distress syndrome: replacement of surface active material. Clin Crit Care Med 1988;13:21-56.

15. Tooley WH, Clements JA, Muramatsu K, Brown CL, Schlueter MA. Lung function in prematurely delivered rabbits treated with a synthetic surfactant. Am Rev Respir Dis 1987; 136:651-6.

16. Durand D J, CLyman RI, Heymann MA, et al. Effects of a pro- tein-free synthetic surfactant on the survival and pulmonary function of preterm lambs. J PEDIATR 1985;107:775-80.

17. Edwards DK. Radiology of hyaline membrane disease, tran- sient tachypnea of the newborn, and bronchopulmonary dys- plasia. In: Farrel PM, ed. Lung development: biological and clinical perspectives. New York: Academic Press, 1982;2:47- 89.

18. O'Brodovich HM, Mellins RB. Bronchopulmonary dysplasia. Am Rev Respir Dis 1985;132:694-709.

19. Bonik0s DS, Bonsch DS, Northway WH Jr, Edwards DK. Bronchopulmonary dysplasia: the pulmonary pathologic se- quel of necrotizing bronchiolitis and pulmonary fibrosis. Hum Pathol 1976;7:643-66.

20. Friedman IM, Furberg CD, DeMetz DL. Fundamentals of clinical trials. Littleton, Mass.: PSG Publishing, 1985:83-107.

2t. Fleiss JL. Statistical methods for rates and proportions. New York: John Wiley & Sons, 1981:160-87.

22. Dixon W J, Massey FJ. Introduction to statistical analysis. New York: McGraw-Hill, !969:222-36.

23. Soll RF, Hoekstra RE, Fangman J J, et al. Multicenter trial of single dose modified bovine surfactant extract (Survanta) for prevention of respiratory distress syndrome. Pediatrics 1990; 85:1092-102.

24. Kwong MS, Egan EA, Notter RH, et al. Double blind clinical trial of calf lung surfactant extract for the prevention of hya- line membrane disease in extremely immature infants. Pediat- rics 1985;76:585-92.

25. Kendig JW, Notter RH, Cox C, et al. Surfactant replacement at birth: final analysis of a clinical trial and comparisons with similar trials. Pediatrics 1988;82:756-62.

26. Walter D, McGuinness G, Bose C, et al. Double blind one year follow up in 1450 infants randomized to Exosurf Neonatal or air [Abstract]. Pediatr Res 1991;29:270A.

27. Gerdes J, CookL, Beaumont E, et al. Effects of three vs one prophylactic doses of Exosurf Neonatal in 700-1100 gram in- fants [Abstract]. Pediatr Res 1991;29:214A.