VOLUME 87 MAY

#{149} 1991 NUMBER

#{149} 5

Very Low Birth Weight Outcomes of the

National Institute of Child Health and Human
Development Neonatal Network

Maureen Hack, MB, BCh; Jeffrey D. Horbar, MDII; Michael H. Malloy,

MD*; Jon E. Tyson, MDII ; Elizabeth Wright, PhDf;
and Linda Wright, MD*

From the *Natjona/ Institute of Child Health and Human Development, Bethesda, MD; tthe
George Washington University Biostatistics Center, Washington, DC; §Case Western
Reserve University, Cleveland, Ohio; University of Texas in Dallas; and
llUniversity of Vermont Medical Center, Burlington, Vermont

ABSTRACT. This report describes the neonatal outcomes 1000 g (range 42% to 75%), 87% at 1001 through 1250 g
of 1765 very low birth weight (<1500 g) infants delivered (range 84% to 91%), and 93% at 1251 through 1500 g
(range 89% to 98%). By obstetric measures of gestation,
from November 1987 through October 1988 at the seven
survival was 23% at 23 weeks (range 0% to 33%), 34% at
participating centers of the National Institute of Child
24 weeks (range 10% to 57%), and 54% at 25 weeks
Health and Human Development Neonatal Intensive (range 30% to 72%). Neonatal morbidity included respi-
Care Network. Survival was 34% at <751 g birth weight ratory distress (67%), symptomatic patent ductus arte-
(range between centers 20% to 55%), 66% at 751 through riosus (25%), necrotizing enterocolitis (6%), septicemia
(17%), meningitis (2%), urinary tract infection (4%), and
intraventricular hemorrhage (45%, 18% grade III and
IV). Morbidity increased with decreasing birth weight.
Received for publication Jun 21, 1990; accepted Sep 7, 1990.
Reprint requests to (M.H.) Rainbow Babies and Childrens Hos-
Oxygen was administered for 28 days to 79% of <751-
g birth weight
infants (range between centers 67% to
pital, University Hospitals of Cleveland, 2101 Adelbert Rd,
Cleveland, OH 44106.
100%), 45% of 751- through 1000-g infants (range 20%
to 68%), and 13% of 1001- through 1500-g infants (range
Members of the NICHD Neonatal Research Network: National
5% to 23%). Ventilator support for 28 days was given
Institute of Child Health and Human Development: Sumner J.
to 68% of infants at <751 g, 29% at 751 through 1000 g,
Yaffe, MD; Charlotte Catz, MD; Linda L. Wright, MD; Michael
and 4% at >1000 g. Hospital stay was 59 days for survi-
H. Malloy, MD, MS; George G. Rhoades, MD. George Washing-
vors vs 15 days for infants who died. Sixty-nine percent
ton University Data Coordinating Center: Elizabeth Wright,
of survivors had subnormal (<10th percentile) weight at
PhD; Tavia Gordon; Lynn Onstad, ScM; Elizabeth Phillips,
discharge. The data demonstrate important intercenter
MA, MSc. Women and Infants Hospital of Rhode Island: Wil-
variation of current neonatal outcomes, as well as differ-
ham Oh, MD, Chairman. University ofAlabama at Birmingham:
ences in philosophy of care and definition and prevalence
George Cassady, MD; Joseph Philips III, MD. University of
of morbidity. Pediatrics 1991;87:587-597; very low birth
Vermont Medical Center: Jerold Lucey, MD; Jeffrey Horbar,
weight, neonates, morbidity, mortality.
MD. Case Western Reserve University: Avroy A. Fanaroff, MD;
Maureen Hack, MD. University of Texas Southwestern Medical
Center at Dallas: Jon E. Tyson, MD; Ricardo Uauy, MD. Wayne
State University School of Medicine: Ronald Poland, MD; See-
ABBREVIATIONS. VLBW, very low birth weight; NICHD, Na-
tha Shankaran, MD. Dartmouth Hitchcock Medical Center:
tional Institute of Child Health and Human Development; SGA,
George Little, MD; William Edwards, MD. University of Ten-
small for gestational age.
nessee at Memphis: Sheldon B. Korones, MD; Richard Cooke,
MD. University of Miami, Jackson Memorial Medical Center,
Miami, Florida: Charles R. Bauer, MD; Emmalee S. Bandstra,
MD.
PEDIATRICS (ISSN 0031 4005). Copyright © 1991 by the Neonatal intensive care, introduced in the 1960s
American Academy of Pediatrics. and further refined in the 1970s and 1980s, led to

Downloaded from www.aappublications.org/news by guest on September

PEDIATRICS Vol.24,872018
No. 5 May 1 991 587
improved survival ‘for infants of very low birth VLBW infants in parentheses) included the Uni-
weight (VLBW; <1500 g).’ Many methods of treat- versity ofAlabama at Birmingham (n = 207), Rain-
ment have been introduced following reports of bow Babies and Childrens Hospital at Case West-
their effectiveness on limited numbers of infants em Reserve University (n = 198), the University of
and in many instances without controlled clinical Texas Southwestern Medical Center at Dallas (n =

trials. Furthermore, they have not been uniformly 290), Wayne State University School of Medicine
applied. It is thus not surprising that differences in (n = 327), University of Tennessee at Memphis (n
survival and morbidity have been reported among = 315), the University of Miami, Jackson Memorial
neonatal intensive care units.2’3 These differences Medical Center, Miami, (n = 315), and the Univer-
have been compounded further by differences in sity of Vermont Medical Center and Dartmouth
sociodemographic factors, prenatal care, infant Hitchcock Medical Center combined as one collab-
characteristics, and data collection. orating center (n = 113).
In 1987 the National Institute of Child Health Data were abstracted by research nurses from the
and Human Development (NICHD) initiated a neo- mothers’ and infants’ charts. Sociodemographic,
natal intensive care network to conduct multicenter pregnancy, and delivery data were obtained soon
studies of therapeutic interventions. A data base of after birth and neonatal and outcome data were
all the VLBW infants was developed to include gathered until discharge or death. In instances
selected sociodemographic factors, perinatal con- where infants were transferred to secondary level
ditions, infant birth data, delivery room and neo- units, survival data were collected until the infant
natal practice, and outcome. The objective of this was discharged home or to a chronic care facility.
“generic” data base was to survey neonatal practice, A data coordinating center (George Washington
morbidity, and mortality and to provide baseline University) was responsible for all aspects of the
information for planning randomized clinical trials biostatistical design, analysis, and data manage-
for VLBW infants. ment of the studies in concert with a committee
In this report we describe the care and outcome that included the principal and coinvestigators at
of 1765 VLBW infants delivered in the seven par- each center, as well as a chairperson who was
ticipating NICHD Neonatal Intensive Care Net- independent of the participating institutions, and
work centers. The variability in care provided in representatives from the Center for Research for
different centers reflects neonatal practice and out- Mothers and Children, Prevention Research
comes of the late 1980s and raises important ques- Program, and the Pregnancy and Perinatology
tions about current regimens of neonatal care. In- Branches of the NICHD. The generic data forms
depth analyses of specific types of care and out- and a compendium of definitions were developed
comes have appeared as abstracts and will be the by a generic data subcommittee consisting of mem-
subject of future reports.49 bers from the above groups. The data coordinating
center conducted all interim and final statistical
POPULATION AND METHOD OF DATA analyses.
COLLECTION For the purposes of this report, birth weight-
specific survival results are presented according to
During the 12-month period from November
World Health Organization recommendations in
1987 through October 1988, 2004 VLBW newborns
four 250-g birth weight subgroups. All livebirths of
were admitted to the participating perinatal cen-
501 g birth weight and above, including those who
ters; 1765 were inborn. Because of the variability
remained in the delivery room, were included in the
in age of admission and outcome of transferred
analyses. Because of space constraints, neonatal
infants, the present report pertains only to the 1765
morbidity and care are examined in three birth
infants with birth weight between 501 and 1500 g
weight subgroups, 501 through 750 g, 751 through
delivered at the perinatal centers. 1000 g, and 1001 through 1500 g.
The seven participating neonatal centers were
selected for having a predominantly inborn popu-
SOCIODEMOGRAPHIC DESCRIPTORS
lation and previous experience in the planning and
conduct of clinical trials. No attempt was made to Risk factors associated with VLBW infants, such
select the centers for geographic or sociodemo- as lack of prenatal care, being black, and/or having
graphic representation in the United States or cx- an unmarried mother, were prevalent among the
emplary survival rates. All were tertiary academic network units (Table 1). Twenty-two percent of the
perinatal centers with pediatric residency and neo- mothers were younger than 20 years of age. In
natal fellowship training programs. The participat- contrast, in the United States in 1987 only 12.4%
ing perinatal centers (with the number of inborn of all mothers were aged 20 years and younger.1#{176}

588 VERY LOW BIRTH WEIGHT

Downloaded OUTCOMES
from www.aappublications.org/news by guest on September 24, 2018
Sixty-two percent of the mothers were black; how- TABLE 1. Sociodemographic and Birth Data (n =

ever, two centers had 16% and 27% Hispanic births,

1765): Percent Prevalence (Range Among Centers)
respectively. Fifteen percent of nonbiack mothers Maternal factors
and 25% of black mothers had no prenatal care. Age <20 y 22 (10-29)
Race (black) 62 (6-86)
Overall, 11% of black mothers in the United States
Married 33 (12-66)
received no prenatal care in 1986.10 No prenatal care 21 (6-30)
Infant birth data
INFANT BIRTH WEIGHT AND GESTATIONAL Mean birth wt, g 1059 (1039-1078)
AGE Distribution
501-750 g 20 (17-25)
Twenty percent of the population weighed be- 751-1000 g 22 (16-26)
tween 501 and 750 g, 22% between 751 and 1000 g, 1001-1250 g 27 (23-31)
1251-1500 g 32 (27-35)
and 58% between 1000 and 1500 g (Table 1).
Mean gestational age 28.1 (27.6-28.5)
Gestational age was determined both by obstet- by obstetric meas-
rical history (the date of the last menstrual period, ures, wk*
obstetric and, when available, ultrasonographic pa- Distribution
rameters) and the physical component of the Bal- s23 wk 6 (2-7)
24-25 wk 15 (13-20)
lard assessment.” The physical measures of the
26-27 wk 22 (18-25)
Ballard were selected because they are reported to 28-31 wk 44 (39-50)
provide a more accurate assessment and are less 32+ wk 13 (7-19)
likely to overestimate gestational age than are neu- Mean gestational age 29.3 (28.7-30.4)
rologic measures.12 Nevertheless the mean gesta- by physical com-
ponent of Ballard,
tional age by obstetric measures was 28.1 weeks,
wkt
compared with 29.3 weeks by Ballard physical as- Distribution
sessment. Twenty-one percent of the population s23 wk 0 (0-3)
were considered to be of 25 weeks’ gestation or less 24-25 wk 7 (2-13)
by obstetric measures, compared with 7% by Ba!- 26-27 wk 18 (13-29)
28-31 wk 54 (49-62)
lard assessment. This difference between obstetric
32+ wk 20 (7-36)
measures of gestational age and the Ballard physi- 1n 1733.
cal assessment, with apparent overestimation of
tn= 1601.
gestational age by Ballard assessment, was evident
at each week of gestation up to 32 weeks.
Infants weighing less than 10th percentile at each Obst.tc& Measures (n.17M)

gestational age were defined as growth retarded or Ballard Physical Eu,. (n-1602)

small for gestational age (SGA) at birth.’3 Twenty-

two percent of the total VLBW population were
considered SGA by obstetric measures compared
with 42% when gestational age was calculated from
% of
the physical Ballard. No infants were considered Population

SGA at 24 weeks, whereas at 25 weeks, 9% were

considered SGA by obstetric measures compared
with 35% by Ballard (Fig 1). Inasmuch as only
babies up to 1500 g birth weight were included, all
23242526272829303132333435
those with a gestational age of more than 33 weeks
Gestational Age (wks)
were considered SGA.
Fig 1. Percent of population classified as small for ges-
tational age (birth weight <10th percentile for gestational
SURVIVAL age),13 according to obstetric measures and the physical
component of the Ballard.
Of the 1765 infants, 74% survived until dis-
charged home or to a chronic care facility (n = 6).
Although the <751-g birth weight group constituted to 98%) birth weight, respectively. Large intercen-
only 20% of births, they accounted for 50% of the ter differences in survival were evident mostly in
mortality. Survival was 34% for infants <751 g the lower birth weight and gestational age
birth weight (center range 20% to 55%), 66% for subgroups.
those 751 through 1000 g (range 42% to 75%), 87% Because small increments in birth weight may be
for those 1001 through 1250 g (range 84% to 91%), critical at the lowest birth weights, outcome was
and 93% for those 1251 through 1500 g (range 89% further examined by 100-g birth weight subgroups

Downloaded from www.aappublications.org/news by guest on September 24, 2018 ARTICLES 589

(Fig 2). Survival was 18% at 500 through 600 g after admission but may also indicate that many
(range 0% to 42%), 30% at 601 through 700 g (range perinatologists have not accepted the evidence of
14% to 45%), and 56% at 701 through 800 g (range steroid effectiveness.14
33% to 74%). For infants who were actively sup-
ported at birth with assisted ventilation via endo- Mode of Delivery
tracheal intubation, survival was 21% at 501
The cesarean section rate was high at all centers
through 600 g (range 0% to 44%), 33% at 601
and comparable to national trends.’5 Forty-seven
through 700 g (range 9% to 50%), and 53% at 700
percent of the infants were delivered by cesarean
through 800 g (range 31% to 73%), respectively.
section, 47% were vaginal vertex, and only 6%
Ten percent of all deaths occurred after 28 days
vaginal breech deliveries (Table 2). The cesarean
of life. Eight-one percent of the total population
section rate decreased as birth weight decreased to
survived to 1 week of age, 77% to 28 days, and 74%
32% for <751-g infants. A similar pattern emerged
to discharge. The infants in the lowest birth weight
when only singleton births were examined.
categories tended to die earlier (Fig 3); however, 8%
of <751-g birth weight deaths occurred after 28
days of life. Delivery Room Resuscitation
Gestational age-specific survival, presented in Fifty-six percent of infants received delivery
Fig 4, demonstrates the differences in survival rates room respiratory support via endotracheal intuba-
depending on whether gestational age was esti- tion and assisted ventilation: 77% of those with
mated by obstetric measures or the Ballard physical birth weight <751 g, 77% with birth weight 751
criteria. Using obstetric measures, survival at 23 through 1000 g, and 42% with birth weight 1001
weeks was 23% (range 0% to 33%); at 24 weeks, through 1500 g (Table 2). The aggressiveness of
34% (range 10% to 57%); and at 25 weeks, 54% delivery room care for infants of<751 g birth weight
(range 30% to 72%). By the Ballard physical esti- varied; two centers provided assisted ventilation to
mation, survival at 24 weeks was 26% (range 0% to practically every infant whereas the remaining cen-
100%, n = 1!) and at 25 weeks, 41% (range 0% to tars gave such care to approximately two thirds of
65%). Because the Ballard overestimates gesta- infants. Drugs were used for delivery room resus-
tional age, infants considered by Ballard to be of a citation in 10% of deliveries (range 2% to 22%),
certain gestational age are in fact younger; thus the more frequently in the lowest birth weight group.
higher mortality by Ballard compared with obstet-
nc measures of gestation. At 28 weeks and above
RESPIRATORY MORBIDITY
there tends to be greater concordance in survival
rates between the two methods. Classification of Respiratory Distress

ANTENATAL AND DELIVERY ROOM CARE The classic clinical presentation of hyaline mem-
brane disease or respiratory distress syndrome
Antenatal Steroids (grunting respirations, retractions, increased oxy-
Antenatal steroids were given to 16% (range 1%
gen requirement, and diagnostic radiographic find-
ings prior to 6 hours of age) is rarely seen in <1000-
to 33%) of mothers, of whom 62% received a full
g birth weight infants, especially those receiving
course (Table 2). This low rate of steroid use may
assisted ventilation in the delivery room and posi-
relate to the fact that many mothers deliver soon
tive end-expiratory pressure in the neonatal inten-
sive care unit. It is thus difficult to clinically diag-
I tuvgr Aovvg
nose respiratory distress syndrome in extremely low
birth weight infants without biochemical evidence
of surfactant deficiency. At the request of center
participants, a condition, referred to as respiratory
insufficiency of prematurity, was defined as a re-
S
quirement for oxygen and usually ventilator or con-
Survival
stant positive pressure support in the absence of
classic respiratory distress syndrome, pneumonia,
or other pulmonary pathology. Overall, 67% of the
population was defined as having respiratory dis-
tress syndrome and/or respiratory insufficiency of
Birthweight (grams) prematurity. The use of these two definitions varied
Fig 2. Percent survival by 100-g birth weight subgroups. widely between centers. Two centers frequently

590 VERY LOW Downloaded

BIRTH from www.aappublications.org/news
WEIGHT OUTCOMES by guest on September 24, 2018
 1.0

0.8

Birthweight (grams)
0.6
tO

-C 501-750

. 751-1000

a 0.4
U 1001-1250

. 1251-1500

0.2

I 50 100 150 200 250 300

Age of Death (days)

Fig 3. Life table for age at death by birth weight.

L:l Pneumopericardium was extremely rare and di-

I agnosed in only 1% of the population. No birth
t:3 Ballard hysical Esam
00

T
weight-specific differences were noted. Of the 15
80 infants so diagnosed, 5 survived.

60
(7,
Survival
RESPIRATORY CARE PRACTICE
40

20
Oxygen Administration

Use of oxygen and assisted ventilation was most

<23 23 24 25 26 27 2831 32+
common in the 751- through 1000-g birth weight
Gest,slinl Age )wks(
group, because some extremely immature <751-g
Fig 4. Survival by gestational age according to the oh- infants were not treated (Table 2). For children
stetric measures and the physical component of the Bal- who were treated, the mean duration of oxygen
lard (Ballard not applicable at <24 weeks’ gestation).
administration in the <751-g birth weight group
was 41 days vs 15 days in the >1000-g group (Table
classified infants, even those weighing more than
1000 g, as having respiratory insufficiency of pre- 3). The overall prevalence ofoxygen administration

maturity, whereas two centers rarely used this di- for 28 days was 26% (range 16% to 38%). This
agnosis. increased with decreasing birth weight and gesta-
tion. Furthermore, large intercenter differences
Pneumothorax and Pneumopericardium were evident (Table 2). Figure 5 illustrates the

A pneumothorax occurred in 10% of infants, duration of oxygen administration by birth weight.

more commonly in the lower birth weight groups. Seventy-nine percent of the <751-g birth weight
Only 7 of the 182 pneumothoraces occurred in infants received oxygen for 28 days or more (center
infants who were not receiving assisted ventilation, range 67% to 100%) compared with 45% (range
all of whom weighed >1000 g (one was receiving 20% to 68%) of 751- through 1000-g and 13% (range
intratracheal continuous positive airway pressure, 5% to 23%) of >1000-g birth weight infants, re-
three were receiving nasal continuous positive air- spectively. Thirty-six percent of the <751-g survi-
way pressure, and three were receiving oxygen by vors received oxygen for 84 days (3 months) or
hood). more, compared with only 2% ofthe >1000-g group.

Downloaded from www.aappublications.org/news by guest on September 24, 2018 ARTICLES 591

TABLE 2. Frequency of Conditions and Care: Percent Prevalence (Range Among Centers by Birth Weight)
501-1500g 501-750g 751-l000g 1001-1500g
(n = 1765) (n = 349) (n = 382) (n = 1034)
Delivery room care
Antenatal steroids 16 (1-33) 8 (0-17) 19 (3-42) 18 (1-40)
Mode of delivery
Vaginal vertex 47 (41-53) 51 (32-63) 41 (25-53) 47 (40-52)
Vaginal breech 6 (1-10) 16 (5-24) 7 (0-12) 3 (1-6)
Cesarean section 47 (37-55) 32 (13-48) 52 (38-73) 50 (43-54)
Delivery room resuscitation
Endotracheal intubation 56 (30-67) 77 (58-97) 77 (38-91) 42 (15-50)
Drugs given 10 (2-22) 18 (6-53) 12 (0-28) 7 (0-11)
Infant birth data
Apgar score <6 at 1 mm 60 (55-63) 89 (82-97) 74 (61-83) 45 (42-48)
Apgar score <6 at 5 mm 25 (18-32) 56 (45-65) 31 (16-45) 13 (8-19)
Multiple birth 18 (15-30) 17 (12-30) 22 (14-29) 17 (11-34)
Intrauterine growth failure* 22 (11-31) 27 (15-42) 21 (10-24) 20 (9-29)
Respiratory morbidity
Respiratory distress syndrome (RDS) 43 (12-79) 39 (5-95) 48 (6-88) 42 (18-72)
Respiratory insufficiency 27 (4-54) 51 (16-92) 42 (8-87) 13 (0-29)
RDS and/or respiratory insufficiency 67 (52-80) 85 (73-100) 87 (63-96) 53 (34-72)
Pneumonia 7 (3-13) 9 (5-20) 10 (4-21) 6 (1-9)
Pneumothorax 10 (6-15) 14 (8-22) 14 (9-33) 8 (4-12)
Pneumopericardium 1 (0-2) 1 (0-5) 2 (0-4) 1 (0-2)
Apnea treated with xanthenest 57 (44-69) 87 (75-100) 77 (60-90) 47 (35-63)
Oxygen administration 82 (71-93) 77 (54-100) 93 (85-100) 80 (59-89)
Ventilator support 70 (55-78) 82 (54-100) 89 (69-100) 60 (37-72)
Tracheostomy 1 (0-2) 1 (0-5) 2 (0-5) 0 (0-2)
Oxygen at 28 d* 26 (16-38) 79 (67-100) 42 (21-68) 13 (5-23)
Ventilator support at 28 d 16 (4-25) 68 (25-87) 29 (8-50) 4 (0-7)
Jaundice
Peakbilirubin(meanmg/lOOmL)t 10(9-11) 9(6-10) 9(8-11) 10 (9-12)
Phototherapyt 77 (27-99) 85 (50-100) 81 (18-100) 75 (28-98)
Exchange transfusionst 4 (0-10) 8 (0-19) 5 (0-11) 3 (0-9)
Patent ductus arteriosus
Symptomatic 25 (14-46) 32 (15-55) 41 (22-69) 17 (8-36)
Fluid restrictions 83 (55-100) 85 (47-100) 85 (62-100) 80 (50-100)
Indomethacin 77 (45-93) 82 (43-95) 74 (38-94) 76 (48-92)
Surgery 27 (10-49) 41 (0-86) 25 (14-50) 19 (3-45)
Necrotizing enterocolitis 6 (1-13) 3 (0-7) 9 (0-21) 7 (0-11)
Infection
Septicemia 17 (7-26) 22 (12-40) 25 (9-33) 13 (4-21)
Meningitis 2 (0-3) 2 (0-4) 3 (0-7) 1 (0-3)
Urinary tract infection 4 (0-7) 4 (0-17) 6 (2-14) 3 (0-5)
Central nervous system
Seizures 6 (2-9) 10 (2-22) 8 (0-17) 3 (1-5)
Ultrasonogram performed 85 (77-97) 68 (44-96) 86 (77-96) 91 (75-97)
Intraventricular hemorrhage, II grade
I 17 (8-24) 16 (4-36) 17 (8-23) 18 (9-25)
II 10 (3-40) 15 (4-35) 15 (2-56) 7 (2-34)
III 11 (4-15) 16 (4-33) 14 (2-22) 8 (3-13)
IV 7 (3-14) 16 (0-36) 12 (5-29) 4 (1-7)
Periventricular leucomalaciall 8 (4-18) 11 (0-23) 7 (0-19) 7 (0-16)
Indwelling catheters
Umbilical artery catheter 60 (43-74) 72 (48-89) 75 (50-88) 50 (29-67)
Umbilical venous catheter 14 (2-44) 26 (3-70) 17 (0-58) 9 (0-29)
Central venous catheter 8 (2-28) 13 (3-64) 9 (2-33) 5 (1-15)
Nutrition
Parenteral nutritiont 81 (60-99) 97 (90-100) 94 (80-100) 75 (51-98)
Weight <10th percentile at discharge# 69 (53-78) 96(88-100) 83 (56-100) 63 (44-73)
* Calculated by obstetric measures.
t Survivors.
:j: Of infants alive at 28 days.
§ Percent of symptomatic patent ductus arteriosus treated.
II Infants who had an ultrasonogram (n = 1440), most severe bleed.
#{182}
Of children who had an ultrasonogram at 3-4 weeks.
# Discharge home from tertiary centers only.

592 VERY LOW BIRTH WEIGHT

Downloaded OUTCOMES
from www.aappublications.org/news by guest on September 24, 2018
 TABLE 3. Duration of Vario us Therapies: Days (Range) b y Center by Bi rth Weight
501-1500g 501-750g 751-l000g 1001-1500g
Ventilator support 15 (9-23) 24 (9-47) 20 (15-27) 9 (4-17)
Oxygen administration 24 (17-38) 41 (28-78) 33 (25-41) 15 (9-31)
Umbilical artery catheter 5 (4-7) 6 (4-8) 6 (3-8) 5 (4-6)
Umbilical venous catheter 4 (1-6) 5 (1-11) 3 (1-3) 3 (1-7)
Central venous lines 34 (15-65) 35 (19-53) 48 (21-129) 26 (1-38)
Phototherapy* 5 (3-7) 6 (2-9) 6 (3-8) 5 (2-7)
Parenteral nutrition* 19 (12-26) 33 (22-49) 25 (15-35) 15 (8-19)
Days to regain birth wt* 15 (7-21) 19 (8-26) 18 (9-27) 14 (7-19)
Length of hospital stay* 59 (4074) 114 (95-140) 80 (39-130) 46 (32-60)
Length of hospital stayt 15 (9-23) 10 (1-33) 19 (7-41) 20 (5-47)
* Survivors.
t Deaths.

Birthweight (grams)

.
Birlhweight (grams)

I 00
x Ol7F4) x 501-750
0 ;l 0 751-1000
1(01/131, U -
80 . 1001-1500

92/136

Children

supplemental
02
60

40
- T12+ % of
Children
Ventilator
on

38/128

20
14/125 5/125
: 7/259

28 56 84 112
56 84
Days on Oxygen
Days on Ventilator

Fig 5. Percent of children receiving supplemental oxy-

Fig 6. Percent of children ventilator dependent by du-
gen by duration of time receiving oxygen and birth ration oftime on ventilator and birth weight. The denom-
weight. The denominator indicates the number of surviv- inator indicates the number of surviving children at each
ing children at each time period.
time period.

Ventilator Support
bation, this low rate of tracheostomy reveals an
Many infants who did not receive initial respi- extremely conservative approach to prophylactic
ratory support in the delivery room later required tracheostomy, as well as a very low rate of intuba-
assisted ventilation. Ventilator support was most tion-associated subglottic stenosis.
frequently used in the 751- through 1000-g group,
of which 90% received such care. Overall 70% of
JAUNDICE
the VLBW infants required ventilator support for
periods ranging from 1 to 379 days (mean 15 days). Bilirubin levels and treatment for jaundice are
The duration of ventilator support increased with presented for survivors only because the majority
decreasing birth weight (Table 3). Sixty-eight per- of infants who die, do so before jaundice develops.
cent of <751-g infants received ventilator support The mean peak total bilirubin level ranged from 9-
for 28 days or more, compared with 29% of those 11 mg/100 mL, with little difference in the birth
born at 751 through 1000 g and 4% of those >1000 weight subgroups (Table 2). Phototherapy was used
g. Eleven percent of the <751-g group vs 5% of the in 77% of infants, with a slight increase among the
751- through 1000-g group and only 0.2% of the <750-g group. Four percent of infants received ex-
1001 through 1500-g birth weight group received change transfusions, with a wide variation among
ventilator support for 84 days or more (Fig 6). centers; center 5 had the highest rate (10%) whereas
centers 2 and 4 had the lowest rates (1% each)
Tracheostomy (Table 3).
Tracheostomy was performed on 15 infants, 10 There were large differences among centers in
of whom survived. Considering that 16% of infants the mean maximal bilirubin levels attained (Fig 7).
alive at 28 days and 5% of those alive at 56 days These differences in treating and/or preventing
were still being ventilated via endotracheal intu- jaundice probably reflect the uncertainty about the

Downloaded from www.aappublications.org/news by guest on September 24, 2018 ARTICLES 593

 Bilirubin

t:: <lomgm%
100
.
90- 14-I6mgm%

80-
U >l6mgm%

70

60-

% of
Population 50

40-

30-

20-

10-

0-
TOTAL
COHORT

CENTER

Fig 7. Mean maximum bilirubin level reached among survivors by birth weight and center.

indications for phototherapy and the long-term ef- 750 g, 48% of those of 751 through 1000 g, and 49%
fects of hyperbilirubinemia. of those of 1001 through 1500 g birth weight, re-
spectively).
PATENT DUCTUS ARTERIOSUS
SEPTICEMIA AND LOCALIZED INFECTIONS
The occurrence of asymptomatic patent ductus
arteriosus was not determined because the infants Septicemia was defined as one positive blood
were not examined at specific ages nor were echo- culture obtained in the presence of clinical signs of
cardiograms routinely performed on every child. infection. According to this definition, 17% of in-
Symptomatic patent ductus arteriosus was, how- fants had at least one episode of septicemia during
ever, noted for 25% of the population, with a wide the hospital stay. The sepsis rate for infants born
variation among centers in both the prevalence of at <1001 g was nearly twice as high as that for
symptomatic patent ductus arteriosus and its treat- infants born at >1000 g (22% vs 13%, respectively)
ment (Table 2). Treatment included fluid restric- (Table 2).
tion in 83%, indomethacin therapy in 77%, and Meningitis was identified in only 2% ofthe entire
surgery in 27% of infants. There were no birth group of infants (range 0% to 3%). The rate of
weight-specific differences in the use of fluid re- meningitis among infants with culture-proven sep-
striction or indomethacin; however, surgery was sis was 5% (14/306). Of the 31 infants with men-
performed more often for the lowest birth weight ingitis, 14 had both positive blood cultures and
infants. spinal fluid cultures whereas 17 had only positive
spinal fluid cultures. Urinary tract infections were
NECROTIZING ENTEROCOLITIS diagnosed in 4% of the population (range 0.5% to
7.1%). Only 4 (0.2%) of the 1765 infants had diag-
Six percent (range 1% to 13%) of infants devel- nosed infections of the bones or joints.
oped necrotizing enterocolitis defined as the pres-
ence of intramural air on roentgenogram, clinical
INTRAVENTRICULAR HEMORRHAGE AND
perforation, or stricture following an episode of
OTHER CENTRAL NERVOUS SYSTEM
suspected necrotizing enterocolitis (Bell’s criteria
PATHOLOGY
stage III or greater16) (Table 2). Two centers had a
very high prevalence (11% and 13%, respectively). During 1987 and 1988 we had no standard pro-
Of the 111 infants with necrotizing enterocolitis, 51 tocol concerning screening for intraventricular
(46%) had surgery (18% of those of 501 through hemorrhage; however, 85% of the population had

Downloaded from www.aappublications.org/news by guest on September 24, 2018

594 VERY LOW BIRTH WEIGHT OUTCOMES
at least one head ultrasonogram performed during ferred to community hospitals, was 59 days, com-
the hospital stay. Forty-five percent of infants with pared with 15 days for infants who died (Table 3).
cerebral ultrasonograms had a periventricular or Survivors <751 g birth weight had the longest stay
parenchymal hemorrhage documented, of which (mean 114 days). However, because the very long
18% were grade III or IV bleeds.’7 The reported hospital stay of the smallest survivors is offset by
higher rate of intraventricular hemorrhage with those in this birth weight category who die early,
decreasing birth weight is evident in Table 2; 32% only 19% of total hospital days were used by infants
of <751-g infants had grade III or IV intraventric- in the 501- through 750-g birth weight category,
ular or parenchymal hemorrhage compared with compared with 27% by infants in the 751- through
12% of those with >1000 g birth weight. The rela- 1000-g category and 54% in the 1001- through 1500-
tively high rate of intraventricular hemorrhage in g birth weight category.
this series might be related to the fact that infants
at lowest risk for an intraventricular hemorrhage DISCUSSION
may not have had an ultrasonogram performed.
These data, collected with a uniform protocol,
NUTRITION AND GROWTH illustrate neonatal outcomes and care practice for
VLBW infants born in the late 1980s. They do not
Measures used to assess the adequacy of nutrition represent a geographically defined cohort; however,
and growth included the time to regain birth weight they are very similar in birth weight and gestational
and the proportion of infants who failed to achieve age distribution to VLBW infants born in the
a discharge body weight greater than the 10th per- United States in 1986.’ Our population did have
centile for postmenstrual age.13’18 Results are pre-
more black infants than national VLBW data (62%
sented only for infants discharged home from the vs 41%, respectively). This might bias comparison
tertiary centers. The mean age to regain birth
of our mortality and morbidity rates to national
weight was 19 days for the <751-g group compared
data. Furthermore, the selection of a VLBW pop-
with 14 days for the >1000-g group (Table 3). Sixty-
ulation with the exclusion of infants of more than
eight percent of infants weighed less than the 10th
1.5 kg birth weight has a marked effect on the
percentile for postmenstrual age at the time of
generalizability of the data to infants of 32 weeks’
discharge (58% of children born appropriate for
gestation or greater, and to a lesser degree to those
gestational age, compared with 97% of those born
between 29 and 31 weeks’ gestation.
SGA).
The data demonstrate important intercenter var-
iation as well as differences in the philosophy of
DURATION OF HOSPITAL STAY
care and definition and prevalence of morbidity. It
Centers 1 and 2 transferred 42% and 29% of their appears that the practice of neonatal medicine re-
infants, respectively, to secondary-level units prior mains in part an art rather than an exact science.
to discharge home whereas the remaining centers Important examples of varying approaches to care
transferred only 1 % to 8%. Five infants were trans- include the delivery room care of the extremely
ferred to a chronic care facility. The mean hospital immature infant, ventilator utilization, use of in-
stay for surviving infants, including those trans- dwelling lines, and the treatment of symptomatic

TABLE 4. Classification of Infants According to Birth Weight-Specific Survival and

Major Neonatal Morbidity*
501-1500g 501-750g 751-l000g 1001-1250g 1251-1500g
(n = 1765) (n = 349) (n = 382) (n = 480) (n = 554)
Died 462 (26) 231 (66) 130 (34) 61 (13) 40 (7)
Survived
Morbidity
CLD 106 (6) 31 (9) 33 (9) 23 (5) 19 (3)
CLD and NEC 6 (0.3) 1 (0.3) 1 (0.3) 3 (0.6) 1 (0.2)
CLD and IVH 47 (3) 13 (4) 13 (3) 13 (3) 8 (1.5)
NEC 61 (4) 3 (0.9) 16 (4) 26 (5) 16 (3)
NEC and IVH 10 (0.6) 1 (0.3) 4 (1) 2 (0.4) 3 (0.5)
IVH 115 (7) 16 (5) 32 (8) 40 (8) 27 (5)
No morbidity 958 (54) 53 (15) 153 (40) 312 (65) 440 (79)
* Values are given as number (percent). CLD, chronic lung disease defined as oxygen
requirement at 36 weeks postmenstrual age; NEC, necrotizing enterocolitis; IVH, grade
III-IV intraventricular hemorrhage.

Downloaded from www.aappublications.org/news by guest on September 24, 2018

ARTICLES 595
patent ductus arteriosus and jaundice. These dif- improve growth during the critical perinatal period
ferences in patient care together with sociodemo- of development.
graphic differences in the patient populations result
in wide differences in the rates of morbidity. This
is most noticeable for chronic lung disease, necro-
ACKNOWLEDGMENTS
tizing enterocolitis, intraventricular hemorrhage,
and jaundice. Future reports will attempt to eluci- We express our appreciation to the nurse coordinators
and data managers for their invaluable contribution to
date some of the factors associated with the differ-
this work.
ent survival rates and morbidity between centers
as well as interrelationships between the outcomes
we have measured.49
Additional important demographic and clinical REFERENCES

aspects of care highlighted by these data include 1. Stahiman MT. Newborn intensive care: success or failure?
the lack of prenatal care, especially among black J Pediatr. 1984;105:162-167
2. Horbar JD, McAuliffe TL, Adler SM, et al. Variability in
mothers; the inaccuracy of the postnatal assess-
28-day outcomes for very low birth weight infants: an analy-
ment of gestational age among extremely low birth
sis of 11 neonatal intensive care units. Pediatrics.
weight infants which makes any outcomes based on 1988;82:554-559
these data difficult to interpret; the problems of 3. Avery ME, Tooley WH, Keller JB, et al. Is chronic lung
care of <751-g birth weight infants, including the disease in low birth weight infants preventible? A survey of
eight centers. Pediatrics. 1987;79:26-30
decision to resuscitate and their high mortality and
4. Shankaran 5, Bandstra E, Edwards W, et al. Incidence and
morbidity rates; and the semantics related to the severity of intraventricular hemorrhage (IVH) in a large
definition of respiratory distress and oxygen de- population of very low birthweight neonates. Pediatr Res.

pendence and its implications, together with the 1990;27:225. Abstract

5. Uauy R, Fanaroff A, Korones 5, Phillips L, Philips J.
pervasive problem of chronic lung disease requiring
Necrotizing enterocolitis (NEC) in VLBW infants: biode-
prolonged oxygen utilization. mographic and clinical correlates. Pediatr Res. 1990;27:228.
Despite gains in the technical aspects of neonatal Abstract
care, very few advances have been made in main- 6. Malloy MH, Onstad L, Wright E. Cesarean sections (CSEC)
and birth outcome in very low birth weight infants (VLBW).
taming adequate growth and nutrition. The poor
Pediatr Res. 1990;27:248. Abstract
growth attainment at discharge, similar to that 7. Bandstra ES, Bauer CR, Onstad L, Lucey JF. Retinopathy
reported,2#{176} is indicative of the combined effects of of prematurity: prevalence and clinical correlates in very-
inadequate nutrition, severity of acute illness, and low-birth-weight (VLBW) infants. Pediatr Res. 1990;27:239.
Abstract
persistence of chronic disease.
8. Tyson JE, Wright E. Resource requirements and survival
Table 4 presents a summary of mortality and the with intensive care of inborn infants 501-800 g birth weight
major morbidity encountered by surviving VLBW (BW). Pediatr Res. 1990;27: 259. Abstract
infants including chronic lung disease defined as 9. Horbar JD, Onstad L. Survival for inborn infants 501-1500
grams at birth: variation among centers. Pediatr Res.
an oxygen requirement at 36 weeks corrected age,21
1990;27:245. Abstract
necrotizing enterocolitis and grade III or IV intra- 10. Wegman ME. Annual summary of vital statistics-1988.
ventricular hemorrhage. Only 15% of 501- through Pediatrics. 1989;84:943-956
750-g birth weight infants, compared with 79% of 11. Ballard JL, Novak KK, Driver M. A simplified assessment
of fetal maturation of newly born infants. J Pediatr.
1251- through 1500-g birth weight infants, survived
1979;95:769-774
and were free of these major complications. Surfac- 12. Constantine NA, Kraemer HC, Kendall-Tackett KA, Ben-
tant was not in use at the time these data were nett FC, Tyson JE, Gross RT. Use ofphysical and neurologic
collected but is currently being widely used to pre- observations in assessment of gestational age in low-birth-
weight infants. J Pediatr. 1987;110:921-928
vent or treat respiratory distress syndrome in all
13. Arbuckle TE, Sherman GJ. An analysis of birth weight by
seven centers. These data thus constitute important gestational age in Canada. Can Med Assoc J. 1989;140:157-
baseline information to evaluate the effects of sur- 165
factant on mortality, respiratory morbidity, and 14. Crowley P, Chalmers I, Keirse MJNC. The effects of corti-
costeroid administration before preterm delivery: an over-
other outcomes.
view of the evidence from controlled trials. Br J Obstet
Challenges for the future include improvement Gynaecol. 1990;97:11-25
in antenatal care and accurate documentation of 15. Malloy MH, Rhoads GG, Schramm W, Land G. Increasing
antenatal gestational age; treatment modes to pre- cesarean section rates in very low-birth weight infants.
JAMA. 1989;262:1475-1478
vent the high rates of septicemia, intraventricular
16. Bell MJ, Ternberg JL, Feigin RD, et al. Neonatal necro-
hemorrhage, necrotizing enterocolitis, and chronic tiziing enterocolitis: therapeutic decisions based upon clini-
lung disease; and improved methods of nutrition to cal staging. Ann Surg. 1978;187:1-7

596 VERY LOW BIRTH WEIGHT

Downloaded OUTCOMES
from www.aappublications.org/news by guest on September 24, 2018
17. Papile L, Burstein J, Burstein R, Koffler H. Incidence and Printing Office; 1988;1:88-1123. US Dept of Health and
evolution of subependymal and intraventricular hemor- Human Services publication (PHS) 88-1123
rhage: a study of infants with birth weights less than 1500 20. Hack M, Merkatz IR, McGrath 5K, Jones PK, Fanaroff AA.
gm. J Pediatr. 1978;92:529-534
Catch-up growth in very low birthweight infants. AJDC.
18. Wingerd J, Schoen EJ, Solomon IL. Growth standards in
1984;138:370-375
the first two years of life based on measurements of white
and black children in a prepaid health care program. Pedi-
21. Shennan AT, Dunn MS, Ohlsson A, Lennox K, Hoskins
atrics. 1971;47:808-825 EM. Abnormal pulmonary outcomes in premature infants:
19. National Center for Health Statistics. Vital Statistics of the prediction from oxygen requirement in the neonatal period.
United States, 1986: Natality. Washington, DC: Government Pediatrics. 1988;82:527-532

FACULTY SCHOLARS’ AWARDS

Each year the William T. Grant Foundation makes awards to up to five

investigators whose research contributes to understanding the development and
well-being of children, adolescents, and youth. Awards are for five (5) years,
totaling $175,000 including indirect costs.
The goal of the Faculty Scholars’ Program is to promote children’s develop-
ment to healthy and productive adulthood by supporting investigators in a
variety of fields on topics such as problem behaviors in school-age children.
Applicants should be junior or pre-tenure, but established investigators (with a
record of publication), not yet in tenure positions. Award recipients will be
called William T. Grant Faculty Scholars. Applicant institutions and individuals
should obtain the brochure outlining the application procedure from:

Robert J. Haggerty, MD
President
William T. Grant Foundation
515 Madison Avenue
New York, New York 10022-5403

Deadline for applications for 1992 awards is July 1, 1991.

ARTICLES 597
Downloaded from www.aappublications.org/news by guest on September 24, 2018
 Very Low Birth Weight Outcomes of the National Institute of Child Health and
Human Development Neonatal Network
Maureen Hack, Jeffrey D. Horbar, Michael H. Malloy, Linda Wright, Jon E. Tyson and
Elizabeth Wright
Pediatrics 1991;87;587

Updated Information & including high resolution figures, can be found at:
Services http://pediatrics.aappublications.org/content/87/5/587
Permissions & Licensing Information about reproducing this article in parts (figures, tables) or in its
entirety can be found online at:
http://www.aappublications.org/site/misc/Permissions.xhtml
Reprints Information about ordering reprints can be found online:
http://www.aappublications.org/site/misc/reprints.xhtml

Downloaded from www.aappublications.org/news by guest on September 24, 2018

 Very Low Birth Weight Outcomes of the National Institute of Child Health and
Human Development Neonatal Network
Maureen Hack, Jeffrey D. Horbar, Michael H. Malloy, Linda Wright, Jon E. Tyson and
Elizabeth Wright
Pediatrics 1991;87;587

The online version of this article, along with updated information and services, is located on
the World Wide Web at:
http://pediatrics.aappublications.org/content/87/5/587

Pediatrics is the official journal of the American Academy of Pediatrics. A monthly publication, it has
been published continuously since 1948. Pediatrics is owned, published, and trademarked by the
American Academy of Pediatrics, 141 Northwest Point Boulevard, Elk Grove Village, Illinois, 60007.
Copyright © 1991 by the American Academy of Pediatrics. All rights reserved. Print ISSN: 1073-0397.

Downloaded from www.aappublications.org/news by guest on September 24, 2018