Article

Day-by-Day Postnatal Survival in Very Low Birth Weight Infants

The George Washington University Medical Center, Newborn Services, Washington, DC 20037, USA.
PEDIATRICS (Impact Factor: 5.47). 08/2010; 126(2):e360-6. DOI: 10.1542/peds.2009-2810
Source: PubMed

ABSTRACT

Postnatal survival rates of very low birth weight (VLBW) infants are well established for each birth weight or gestational age category. These figures do not differentiate viable infants who survive the first few days of life from extremely immature ones who die shortly after birth. This study aimed to develop standardized curves for day-by-day postnatal survival rates of VLBW infants.
National Inpatient Sample Database and its pediatrics-only subportion were analyzed for the years 1997-2004. Infants with birth weight <1500 g were included in the study. Infants were classified according to their birth weight into 4 groups: <500, 500 to 749, 750 to 999, and 1000 to 1499 g. Postnatal survival rates were calculated for each group at birth and at 1, 2, 3, 4 to 5, 6 to 7, 14, 21, 28, and >28 postnatal days.
Overall survival for infants with birth weight <500 g was 8%. Those who lived through the first 3 days of life had a chance of survival up to 50%. Infants in the 500- to 749-g group had overall survival rate of 50% that increased to 70% if they survived through the third day and 80% by the end of the first week. There was no improvement in the overall survival of any birth weight category over the years of the study.
VLBW infants who survive the first few postnatal days have a considerably better chance for life. We can predict postnatal survival chances for each birth weight category on a day-by-day basis until discharge.

Full-text

Available from: Hany Z Aly, Jan 16, 2016
DOI: 10.1542/peds.2009-2810
2010;126;e360-e366; originally published online Jul 12, 2010; Pediatrics
Mohamed A. Mohamed, Ayman Nada and Hany Aly
Day-by-Day Postnatal Survival in Very Low Birth Weight Infants
http://www.pediatrics.org/cgi/content/full/126/2/e360
located on the World Wide Web at:
The online version of this article, along with updated information and services, is
rights reserved. Print ISSN: 0031-4005. Online ISSN: 1098-4275.
Grove Village, Illinois, 60007. Copyright © 2010 by the American Academy of Pediatrics. All
and trademarked by the American Academy of Pediatrics, 141 Northwest Point Boulevard, Elk
publication, it has been published continuously since 1948. PEDIATRICS is owned, published,
PEDIATRICS is the official journal of the American Academy of Pediatrics. A monthly
. Provided by George Washington Univ on September 20, 2010 www.pediatrics.orgDownloaded from
Page 1
Day-by-Day Postnatal Survival in Very Low Birth
Weight Infants
WHAT’S KNOWN ON THIS SUBJECT: Overall mortality and
postnatal survival rates of VLBW infants are well established for
each birth weight or GA category.
WHAT THIS STUDY ADDS: Day-by-day postnatal survival curves
for various birth weight or GA categories can be predicted until
discharge to home. Trends of mean age at death, mean length of
stay, and first- and third-day mortalities over the years of the
study were determined.
abstract
OBJECTIVES: Postnatal survival rates of very low birth weight (VLBW)
infants are well established for each birth weight or gestational age
category. These figures do not differentiate viable infants who survive
the first few days of life from extremely immature ones who die shortly
after birth. This study aimed to develop standardized curves for day-
by-day postnatal survival rates of VLBW infants.
METHODS: National Inpatient Sample Database and its pediatrics-only
subportion were analyzed for the years 1997–2004. Infants with birth
weight 1500 g were included in the study. Infants were classified
according to their birth weight into 4 groups: 500, 500 to 749, 750 to
999, and 1000 to 1499 g. Postnatal survival rates were calculated for
each group at birth and at 1, 2, 3, 4 to 5, 6 to 7, 14, 21, 28, and 28
postnatal days.
RESULTS: Overall survival for infants with birth weight 500 g was 8%.
Those who lived through the first 3 days of life had a chance of survival
up to 50%. Infants in the 500- to 749-g group had overall survival rate of
50% that increased to 70% if they survived through the third day and
80% by the end of the first week. There was no improvement in the
overall survival of any birth weight category over the years of the study.
CONCLUSIONS: VLBW infants who survive the first few postnatal days
have a considerably better chance for life. We can predict postnatal
survival chances for each birth weight category on a day-by-day basis
until discharge. Pediatrics 2010;126:e360–e366
AUTHORS: Mohamed A. Mohamed, MD, MS, MPH,
a
Ayman
Nada, MD,
b
and Hany Aly, MD
a
a
Newborn Services, The George Washington University Medical
Center and the Children’s National Medical Center, Washington,
DC; and
b
Institute of Childhood Studies, Ain Shams University,
Cairo, Egypt
KEY WORDS
neonatal mortality, postnatal survival, Healthcare Cost and
Utilization Project, National Inpatient Sample Database
ABBREVIATIONS
VLBW—very low birth weight
HCUP—Healthcare Cost and Utilization Project
NIS—National Inpatient Sample Database
ICD-9 International Classification of Disease, Ninth Revision
GA— gestational age
LOS—length of stay
www.pediatrics.org/cgi/doi/10.1542/peds.2009-2810
doi:10.1542/peds.2009-2810
Accepted for publication Apr 15, 2010
Address correspondence to Mohamed A. Mohamed, MD, The
George Washington University Medical Center, Newborn
Services, 900 23rd St NW, Suite G-2092, Washington, DC 20037.
E-mail: mmohamed@mfa.gwu.edu
PEDIATRICS (ISSN Numbers: Print, 0031-4005; Online, 1098-4275).
Copyright © 2010 by the American Academy of Pediatrics
FINANCIAL DISCLOSURE: The authors have indicated they have
no financial relationships relevant to this article to disclose.
e360 MOHAMED et al
. Provided by George Washington Univ on September 20, 2010 www.pediatrics.orgDownloaded from
Page 2
Changes in the perinatal treatment of
preterm infants, including prenatal re-
ferral to tertiary care centers, prena-
tal steroid treatment, postnatal as-
sisted ventilation, and surfactant
therapy, have been associated with a
substantial increase in the survival of
very low birth weight (VLBW) infants.
1–4
The majority of infants who are born at
500 g die in the delivery room or
shortly after birth
3–6
; however, mortal-
ity rates of the mature ones who sur-
vive through the crucial early postna-
tal period is not well established.
5–9
Practitioners need graphs for day-by-
day postnatal survival of preterm in-
fants when counseling parents before
or during their stay in NICUs.
During the past 25 years, the incidence
of preterm birth in the United States
has increased from 9.5% to 12.5%.
1
Despite the significant improvement in
the survival of extreme preterm in-
fants from the 1980s to the 1990s,
there was no additional improvement
from the early 1990s until 1999.
10–12
Whether this is attributable to the ag-
gressive resuscitation of nonviable in-
fants, resulting in a delay in their time
of death, is not yet clear. Trends for the
change in time of death for preterm
infants have not been explored in re-
cent years.
In this study, a large cohort of VLBW
infants who were born between 1997
and 2004 were used to answer the fol-
lowing questions: (1) What are the
postnatal survival odds of VLBW in-
fants measured at various postnatal
ages for each birth weight category?
(2) Is there an improvement in the sur-
vival of VLBW infants in recent years?
(3) Is there a change in the time of
death concomitant with advances in
management over the years of the
study? (4) Is there a significant de-
crease in the length of hospital stay of
survivors until discharge to home dur-
ing the study period? (5) Is there a sig-
nificant change in the proportion of
various birth weight categories among
surviving VLBW infants?
METHODS
We used the de-identified data sets
produced by Healthcare Cost and Utili-
zation Project (HCUP) of the Agency for
Healthcare Research and Quality. Data
sets are from an all-payer database
that annually collects information
from millions of inpatient hospitaliza-
tion records at time of discharge.
These data sets reflect 1000 hospi-
tals from across the United States with
various care levels (primary, tertiary),
type of insurance (public, private), and
academic (university, general) set-
tings. The HCUP produces the National
Inpatient Sample Dataset (NIS) and its
pediatrics-only version, KID. Both data
sets include 100 data elements for
each hospital stay, such as primary and
secondary diagnoses (precoded by us-
ing the International Classification of
Disease, Ninth Revision [ICD-9]) and pri-
mary and secondary procedures (pre-
coded by using Current Procedural Ter-
minology). The KID data set has identical
data elements to those included in the
NIS data set.
13,14
The KID data set is avail-
able only for the years 1997, 2000, and
2003; we used the data for these 3 years.
For the years 1998, 1999, 2001, 2002, and
2004, for which the KID data set is un-
available, we used the NIS data set. In this
study, we extracted data on birth weight
categories, gestational age (GA) catego-
ries (whenever available), gender, race,
source of admission, maternal diag-
noses related to this admission, neona-
tal diagnoses accumulated during the
hospitalization, disposition at discharge,
length of hospital stay (LOS), and age at
death. This study was approved by the
institutional review board at The George
Washington University Medical Center.
Patient Selection
All infants who were admitted to the
NICU with birth weight of 1500 g
were included in the study. In both data
sets, birth weight was reported ac-
cording to ICD-9 diagnostic codes as
follows: 765.01, 765.02, 765.03, 765.04,
and 765.05 for the birth weight catego-
ries 500, 500 to 749, 750 to 999, 1000
to 1249, and 1250 to 1499 g, respec-
tively. GA was reported according to
the ICD-9 diagnostic codes as follows:
765.21, 765.22, 765.23, 765.24, 765.25,
and 765.26 for GA 24, 24, 25 to 26, 27
to 28, 29 to 30, and 31 to 32 weeks,
respectively. ICD-9 diagnostic codes
for GA at birth were introduced into
these data only as of late 2002. Survival
rates were calculated for infants in
each birth weight category at day of
birth, representing overall survival
rate, and at 1, 2, 3, 4 to 5, 6 to 7, 8 to 14,
15 to 21, 22 to 28, and 28 postnatal
days. Analysis included infants who
were in-born or transferred from
other hospitals. We excluded the fol-
lowing infants: (1) infants who were
out-transported from the delivery hos-
pital because of their unknown out-
comes and to avoid duplicate inclusion
and (2) infants who had any of the fol-
lowing diagnoses that may affect their
clinical course or outcome: small for
GA for infants 1000 to 1499 g as per ICD-
9 codes found in the data; congenital
anomalies of central nervous system;
congenital heart diseases except
patent ductus arteriosus; congenital
lung diseases and diaphragmatic her-
nia; abdominal wall defects; renal dys-
plasia; multiple congenital anomalies;
and infants with trisomy 13, 18, or 21.
HCUP data did not include infants who
never had admission records, such as
sill births or those who died in the de-
livery room and were not admitted to
the NICU.
Data Management
Microsoft Excel 2003 (Redmond, WA)
was used to calculate postnatal sur-
vival rates for each birth weight or GA
category. Percentages of day-by-day
postnatal survival among VLBW infants
were calculated by using the following
ARTICLES
PEDIATRICS Volume 126, Number 2, August 2010 e361
. Provided by George Washington Univ on September 20, 2010 www.pediatrics.orgDownloaded from
Page 3
formula: [(percentage of overall sur-
vivals at time of discharge)/(percent-
age of overall mortality among infants
who died that is declining for each
postnatal day), for the same birth
weight or GA category].
SAS 9.1.3 (SAS Institute, Cary, NC) was
used to measure significance of differ-
ences in overall and first- and third-
day mortality, change in time of death,
LOS among survivors, and proportion
of each birth weight among overall
survivors for each birth weight or GA
category across the years of study.
Mantel-Haenszel
2
test and regres
-
sion analysis models were used to ex-
amine significant trends over the
years.
RESULTS
Data sets had 115 350 infants who
weighed 1500 g. We excluded
12 857 infants: 2599 infants were
small for GA; 1993 infants had central
nervous system anomalies; 6966 had
congenital heart disease; 1255 had
lung anomalies; 163 had congenital
lung diseases and diaphragmatic
hernia; 182 had abdominal wall de-
fects; 49 had renal dysplasia; 82 had
multiple congenital anomalies; and
640 had trisomy 13, 18, or 21. The
final number used for the analysis
was 102 493 VLBW infants. In this co-
hort, 10.6% weighed 500 g, 18.4%
weighed between 500 and 749 g,
16.9% weighed between 750 and
999 g, and 54.1% weighed between
1000 and 1499 g. Demographic char-
acteristics of the birth weight cate-
gories are shown in Table 1.
Day-by-Day Survival Rates
Overall survival for VLBW infants in this
cohort was 77.5%. When the infants
who weighed 500 g were excluded,
the overall survival for infants who
weighed 500 to 1499 g was 85.6%. Dur-
ing the first day of life, 34.8% of the
mortality took place; by the end of the
first 3 days of life, 58.2% of mortality
had occurred. Almost 90% of all mor-
talities occurred by the 28th day. For
infants who weighed 500 g, only 8%
survived until discharge to home. Most
(71.9%) of the mortality took place dur-
ing the first day of life; by the end of the
third day, 85.6% of the mortality had
occurred. For infants who weighed 500
to 749 g, the overall survival was
50.8%, with 19.6% mortality occurring
during the first day of life and 31.4%
occurring by third day of life. Infants
who weighed 750 to 1000 g had a better
survival rate of 85.1%. Mortality within
the first day of life was 3.1%, and mor-
tality by third day of life was 6.4%.
Graphical representation of the day-
by-day postnatal survival for each
birth weight category is shown in Fig 1.
Table 2 depicts day-by-day postnatal
mortality for various birth weight
categories.
Trends in Mortality Over the Years
Overall mortality declined from 23.4%
in 1997 to 22% in 2004 (P .001). When
we excluded infants who weighed
FIGURE 1
Percentage of survivals according to postnatal age in each birth weight category (N 102 493). f
indicates 500 g; , 500 to 749 g; Œ, 750 to 999 g; F, 1000 to 1499 g.
TABLE 1 Characteristics of Study Population, %
Characteristic Weight, g
500
(n 10 915)
500–749
(n 18 863)
750–999
(n 17 433)
1000–1499
(n 55 282)
Female gender 49.51 49.71 46.86 49.43
Race
White 29.85 30.35 33.92 35.41
Black 26.15 23.67 20.91 19.48
Hispanic 12.95 13.86 13.51 13.41
Other 32.05 32.12 31.66 31.70
Cesarean delivery 13.45 35.54 45.81 49.56
Twins 14.74 12.58 13.30 16.87
Twins 2.08 2.13 3.19 4.45
Transport 3.04 18.18 20.90 14.98
Severe IVH 53.47 53.70 42.25 18.22
Pulmonary hemorrhage 1.68 5.81 3.46 0.74
NEC 1.21 7.29 8.37 3.94
Maternal hypertension 0.89 1.02 1.52 2.44
Chorioamnionitis 2.53 1.34 0.80 0.58
Antipartum hemorrhage 1.74 1.29 1.06 1.13
IVH indicates intraventricular hemorrhage; NEC, necrotizing enterocolitis.
e362 MOHAMED et al
. Provided by George Washington Univ on September 20, 2010 www.pediatrics.orgDownloaded from
Page 4
500 g, there was no significant
change in mortality (P .25). There
was no significant change for better
survival in any of the birth weight cat-
egories across the years of the study
(Table 3).
For the first day of life, mortality in-
creased over the years for infants
who weighed 500g(P .04) but
did not change in other birth weight
categories. By the third day of life,
mortality had decreased for infants
who weighed 500 and 500 to 749 g
over the years (P .001 and P .01,
respectively; Table 3).
LOS Over the Years
There was an increase of 2 days in
time until death in the overall popula-
tion from 1997 to 2004 (P .001). Time
until death increased only for infants
with birth weight 500 and 500 to
749g(P .016 and P .001, respec-
tively). The LOS until discharge to home
for survivors increased in all birth
weight categories (Table 4).
Demographic Composition of
Survivors
Only 1% of surviving VLBW infants
weighed 500 g, and 12% weighed
500 to 749 g. There was no apparent
change in these proportions among
survivors across the years of study
(Table 5).
Postnatal Survival According to GA
Curves to show day-by-day postnatal
survival for VLBW infants by using GA
24, 24, 25 to 26, and 27 to 28 weeks
are shown in Fig 2. This graph was de-
veloped by using data from 2002–2004
whenever available. The trends for
postnatal survival according to GA
groups were generally similar to those
that were developed by using birth
weight categories.
DISCUSSION
This study provides the day-by-day
postnatal survival curves for VLBW in-
fants classified according birth weight
or GA categories. Because clinical con-
ditions of these infants are unstable
early in postnatal life, their chance to
survive substantially changes over
time; therefore, these data provide cli-
nicians and parents with a better un-
derstanding of infants’ survival in the
TABLE 2 Mortality in VLBW Infants in First 28 Days of Life, %
Parameter Day of Life At
Discharge,
%
P
1 3 7 14 28
All 1500 g (n 102 493) 12.01 16.46 18.16 19.57 20.93 22.53 .001
500–1499 g (n 91 578) 5.01 8.37 10.01 11.41 12.80 14.40 .001
500g(n 10 915) 71.94 85.55 87.84 89.34 90.40 92.10 .001
500–749 g (n 18 863) 19.55 31.40 36.54 40.70 44.61 49.18 .001
750–999 g (n 17 433) 3.13 6.40 8.31 10.19 12.34 14.91 .001
1000–1499 g (n 55 282) 0.57 1.08 1.31 1.75 2.05 2.33 .001
Accumulated postnatal mortality for VLBW infants included in the study in the first 28 days of their NICU stay. Represented
as proportion from total at every specified day.
TABLE 3 Postnatal Mortality of VLBW Infants Over the Years of the Study, %
Parameter Year P for
Trend
1997
(n 16 751)
1998
(n 7243)
1999
(n 7628)
2000
(n 20 228)
2001
(n 7054)
2002
(n 8089)
2003
(n 25 880)
2004
(n 9620)
Overall mortality
Overall 23.36 23.24 23.27 21.76 22.70 24.04 21.89 21.98 .001
500–1499 g 14.65 14.66 14.99 13.88 14.23 15.64 14.23 13.97 .240
500 g 92.33 91.58 90.12 92.81 93.13 93.00 91.32 92.74 .990
500–749 g 50.66 48.98 49.83 47.80 46.28 53.97 48.19 49.97 .450
750–999 g 14.40 16.24 13.71 14.32 15.63 15.25 15.69 14.41 .300
1000–1499 g 2.30 2.08 2.56 2.19 2.18 2.89 2.36 2.25 .510
First-day mortality
Overall 12.52 11.83 11.18 12.1 12.28 12.99 11.64 11.55 .070
500–1499 g 5.18 4.88 4.52 5.02 5.03 5.76 4.94 4.62 .540
500 g 70.70 67.20 65.00 75.97 72.52 70.62 71.76 72.80 .040
500–749 g 20.82 18.79 17.27 19.23 19.13 21.95 19.20 19.34 .650
750–999 g 3.06 2.91 2.47 3.60 3.48 3.85 2.93 2.58 .750
1000–1499 g 0.45 0.57 0.53 0.65 0.43 0.86 0.57 0.49 .370
Third-day mortality
Overall 17.73 16.97 16.52 16.14 16.64 17.34 15.59 15.67 .001
500–1499 g 8.94 8.44 8.24 8.28 8.32 9.14 8.10 7.73 .005
500 g 87.37 84.90 83.45 87.07 85.87 82.87 82.82 85.79 .001
500–749 g 34.00 31.39 29.95 30.87 30.39 33.67 30.19 30.58 .011
750–999 g 6.65 5.98 5.15 6.71 6.79 7.55 6.49 5.10 .680
1000–1499 g 1.02 1.08 1.29 1.09 0.82 1.34 1.01 1.16 .850
ARTICLES
PEDIATRICS Volume 126, Number 2, August 2010 e363
. Provided by George Washington Univ on September 20, 2010 www.pediatrics.orgDownloaded from
Page 5
NICU. A few studies, usually with small
sample size, described postnatal sur-
vival rates of these VLBW infants. Coo-
per et al
15
used a cohort of 1925 infants
who were born at 1 medical center be-
tween 1986 and 1994 to draw curves
for postnatal survival rates in weekly
intervals. Similarly, Jones et al
16
con-
structed actuarial survival rates clas-
sified by gender, by using a sample of
Canadian VLBW infants from 17 nurser-
ies during the period 1996 –1997. Our
study is novel in several aspects. First,
it included a significantly larger sam-
ple size (n 102 351) by using a na-
tionally collected database. Second,
the study drew curves for these in-
fants’ survival rates during the most
critical and unstable first week of life
on a day-by-day basis. Third, the study
included infants with a birth weight of
500 g and those with a GA 24
weeks. Fourth, it provided a trend for
mortality that expanded to a more re-
cent period, 1997–2004, whereas pre-
vious studies were before 2000. Fifth,
these data included infants from
1000 hospitals sampled from 36
states with different types of hospitals,
payment options, care levels, and man-
agement strategies. Such diversity
better reflects the entire nation’s med-
ical, geographic, ethnic, social, and
economic conditions.
The Neonatal Network of the Eunice
Kennedy Shriver National Institute of
Child Health and Human Development re-
ported a decline in VLBW infant mortality
from 26% in 1988 to 20% in 1991 and 16%
in 1995 and 1996.
17–19
Similarly, Horbar et
al
20
demonstrated an