ARTICLE

Neonatal Outcomes of Extremely Preterm Infants

From the NICHD Neonatal Research Network
AUTHORS: Barbara J. Stoll, MD,a Nellie I. Hansen, MPH,b WHATS KNOWN ON THIS SUBJECT: The NICHD NRN has
Edward F. Bell, MD,c Seetha Shankaran, MD,d Abbot R. published periodic evaluations of morbidity and mortality rates
Laptook, MD,e Michele C. Walsh, MD, MS,f Ellen C. Hale, RN,
BS, CCRC,a Nancy S. Newman, BA, RN,f Kurt Schibler, MD,g
for VLBW infants. Increased VLBW survival has paralleled
Waldemar A. Carlo, MD,h Kathleen A. Kennedy, MD, MPH,i improvements in prenatal, obsteric and neonatal care, but recent
Brenda B. Poindexter, MD, MS,j Neil N. Finer, MD,k Richard data suggest that a plateau in survival may have been reached.
A. Ehrenkranz, MD,l Shahnaz Duara, MD,m Pablo J.
Snchez, MD,n T. Michael OShea, MD, MPH,o Ronald N. WHAT THIS STUDY ADDS: This study is the rst NRN study to
Goldberg, MD,p Krisa P. Van Meurs, MD,q Roger G. Faix, report outcomes on the basis of GA-specic information, which
MD,r Dale L. Phelps, MD,s Ivan D. Frantz , III, MD,t Kristi L.
Watterberg, MD,u Shampa Saha, PhD,b Abhik Das, PhD,v should be particularly valuable to obstetricians and pediatricians
and Rosemary D. Higgins, MD,w for the Eunice Kennedy as they counsel parents of high-risk infants.
Shriver National Institute of Child Health and Human
Development Neonatal Research Network
aDepartment of Pediatrics, School of Medicine, Emory University,

Childrens Healthcare of Atlanta, Atlanta, Georgia; bStatistics

and Epidemiology Unit, RTI International, Research Triangle
Park, North Carolina; cDepartment of Pediatrics, Carver College
of Medicine, University of Iowa, Iowa City, Iowa; dDepartment of
abstract
Pediatrics, School of Medicine, Wayne State University, Detroit, OBJECTIVE: This report presents data from the Eunice Kennedy
Michigan; eDepartment of Pediatrics, Women and Infants Shriver National Institute of Child Health and Human Development Neo-
Hospital, Brown University, Providence, Rhode Island;
fDepartment of Pediatrics, Rainbow Babies and Childrens natal Research Network on care of and morbidity and mortality rates
Hospital, Case Western Reserve University, Cleveland, Ohio; for very low birth weight infants, according to gestational age (GA).
gDepartment of Pediatrics, School of Medicine, University of

Cincinnati, Cincinnati, Ohio; hDivision of Neonatology, School of METHODS: Perinatal/neonatal data were collected for 9575 infants of
Medicine, University of Alabama at Birmingham, Birmingham, extremely low GA (2228 weeks) and very low birth weight (4011500
Alabama; iDepartment of Pediatrics, University of Texas Medical
School at Houston, Houston, Texas; jDepartment of Pediatrics, g) who were born at network centers between January 1, 2003, and
School of Medicine, Indiana University, Indianapolis, Indiana; December 31, 2007.
kDepartment of Neonatology, University of California, San Diego,

Medical Center, San Diego, California; lDepartment of Pediatrics, RESULTS: Rates of survival to discharge increased with increasing GA
School of Medicine, Yale University, New Haven, Connecticut;
mDepartment of Pediatrics, Miller School of Medicine, University
(6% at 22 weeks and 92% at 28 weeks); 1060 infants died at 12 hours,
of Miami, Miami, Florida; nDepartment of Pediatrics, University with most early deaths occurring at 22 and 23 weeks (85% and 43%,
of Texas Southwestern Medical Center, Dallas, Texas; respectively). Rates of prenatal steroid use (13% and 53%, respec-
oDepartment of Pediatrics, School of Medicine, Wake Forest

University, Winston-Salem, North Carolina; pDepartment of tively), cesarean section (7% and 24%, respectively), and delivery room
Pediatrics, School of Medicine, Duke University, Durham, North intubation (19% and 68%, respectively) increased markedly between
Carolina; qDepartment of Pediatrics, School of Medicine,
Stanford University Palo Alto, California; rDivision of 22 and 23 weeks. Infants at the lowest GAs were at greatest risk for
Neonatology, Department of Pediatrics, School of Medicine, morbidities. Overall, 93% had respiratory distress syndrome, 46%
University of Utah, Salt Lake City, Utah; sDepartment of
Pediatrics, School of Medicine and Dentistry, University of patent ductus arteriosus, 16% severe intraventricular hemorrhage,
Rochester, Rochester, New York; tDivision of Newborn Medicine, 11% necrotizing enterocolitis, and 36% late-onset sepsis. The new
Department of Pediatrics, Floating Hospital for Children, Tufts
Medical Center, Boston, Massachusetts; uDepartment of
severity-based denition of bronchopulmonary dysplasia classied
Pediatrics, University of New Mexico Health Sciences Center, more infants as having bronchopulmonary dysplasia than did the tra-
Albuquerque, New Mexico; vStatistics and Epidemiology Unit, RTI ditional denition of supplemental oxygen use at 36 weeks (68%, com-
International, Rockville, Maryland; and wEunice Kennedy Shriver
National Institute of Child Health and Human Development, pared with 42%). More than one-half of infants with extremely low GAs
National Institutes of Health Bethesda, Maryland had undetermined retinopathy status at the time of discharge. Center
(Continued on last page) differences in management and outcomes were identied.
CONCLUSION: Although the majority of infants with GAs of 24
weeks survive, high rates of morbidity among survivors continue to
be observed. Pediatrics 2010;126:443456

PEDIATRICS Volume 126, Number 3, September 2010 443

Over the previous 2 decades, the Eu- mined as the best obstetric estimate severe ROP in either eye, or undeter-
nice Kennedy Shriver National Institute by using ultrasonography and/or the mined in either eye without severe ROP
of Child Health and Human Develop- date of the last menstrual period. In- in either eye. Complete denitions are
ment Neonatal Research Network trauterine growth restriction, dened included in a footnote to Table 6. The
(NRN) has monitored trends in morbid- as BW of 10th percentile for gender registry was approved by the institu-
ity and mortality rates among very low and GA, was determined by using tional review boards at each center.
birth weight (VLBW) infants born at the growth charts published by Alexander
university centers that constitute the et al.9 Morbidities were dened Statistical Analyses
NRN.1 6 Increased VLBW infant survival in earlier publications,1 6,10,11 includ- All infants were studied for assess-
rates have paralleled improvements ing respiratory distress syndrome, ment of maternal characteristics, neo-
in prenatal, obstetric, and neonatal bronchopulmonary dysplasia (BPD), natal demographic features, interven-
care.7,8 NRN data suggest that a pla- intraventricular hemorrhage (IVH), tions performed soon after birth, and
teau in VLBW infant survival rates periventricular leukomalacia (PVL), survival. Infants who died at 12
might have been reached, despite in- early-onset and late-onset sepsis, ne- hours were excluded from analyses fo-
creased use of prenatal corticosteroid crotizing enterocolitis, patent ductus cused on morbidities diagnosed at
treatment, prenatal antibiotic treat- arteriosus, and retinopathy of prema- 12 hours. For determination of rates
ment, and early neonatal surfactant turity (ROP). Respiratory distress syn- of survival without morbidity, morbid-
treatment.6 Previous NRN reports pre- drome was dened on the basis of ity was dened as severe IVH (grade
sented patient characteristics, inter- clinical features and oxygen or respi- 3), PVL, BPD, necrotizing enterocolitis,
ventions, and outcomes according to ratory support for 6 of the rst 24 stage 3 ROP, or infection (early-
birth weight (BW), with an upper limit hours. onset sepsis, late-onset sepsis, or
of 1500 g. Such BW-specic data may meningitis).
Three denitions of BPD were used,
be skewed by more-mature infants
namely, traditional BPD (supplemental Statistical signicance for unadjusted
with growth restriction. The aim of this
oxygen use at postmenstrual age comparisons was determined by using
study was to evaluate management,
hospital complications, and mortality
[PMA] of 36 weeks); BPD determined by 2 or Wilcoxon tests. Logistic or linear
using the National Institutes of Health regression models were used to as-
rates among infants with gestational
Workshop severity-based diagnostic sess associations with GA, with adjust-
ages (GAs) of 22 to 28 weeks who were
criteria,12 and BPD determined accord- ment for study center and infant BW,
born at NRN centers between 2003 and
ing to physiologic denition.13 Surviv- with statistical signicance deter-
2007.
ing infants who were discharged or mined by using Wald 2 or F tests. Gen-
METHODS transferred before PMA of 36 weeks eralized logit regression models were
were classied on the basis of their used for comparisons involving cate-
Study Population and Clinical status at 36 weeks, if status informa- gorical variables with 2 levels.
Outcomes tion was available, or oxygen use at Risk of death and changes in clinical
Infants born alive at NRN centers in discharge/transfer, if status informa- practice during the study period were
20032007 with GAs of 2207 to 2867 tion was not available. Unless noted assessed by using robust Poisson re-
weeks and BWs of 401 to 1500 g were otherwise, BPD refers to the tradi- gression models14 to produce correct
studied, including those with congeni- tional denition. SEs for the estimated relative risks
tal anomalies. These infants were part Revisions to data collection in 2006 in- (RRs). Additional adjustments for clus-
of the NRN VLBW registry.1 6 cluded questions about maternal cho- tering according to center were not
Research personnel collected mater- rioamnionitis, placental pathologic made because study center was
nal pregnancy/delivery data soon after conditions, nitric oxide use, and ibu- treated as a xed effect in these mod-
birth and infant data from birth to profen use and expanded data collec- els, which also included effects for BW
death, discharge/transfer, or 120 days tion on birth resuscitation and neuro- and GA. To assess linear trends, year
of age (status). For infants with pro- logic, pulmonary, and ophthalmologic was included as a continuous variable,
longed hospitalizations, limited infor- outcomes. In addition to ophthalmo- with adjusted RRs for the change per
mation was collected up to 1 year. logic examination results and inter- year being reported. Initial models in-
Denitions for maternal and infant ventions, the following outcomes, de- cluded terms for interactions between
characteristics were provided in a ned in the manual of operations, each GA and year, to assess whether
manual of operations. GA was deter- were recorded: favorable in both eyes, yearly trends varied according to GA.

444 STOLL et al
 ARTICLE

Nonsignicant interactions were re- and conrmed more commonly by pla- endotracheal intubation in the delivery
moved, and the models were rerun. cental histologic ndings at lower GAs. room and surfactant therapy varied
Overall, 59% of infants were born according to GA (year-GA interaction:
Participating NRN Study Centers through cesarean section, with the P .01 for each). Rates of intubation
The numbers of infants included from steepest increase in cesarean section and surfactant therapy decreased for
each center were as follows: University delivery rates between GAs of 22 and infants born at 28 weeks. During the
of Alabama, 805 infants; Brown Univer- 24 weeks (7% at 22 weeks and 60% at study period, the proportion of infants
sity, 616 infants; University of Califor- 24 weeks). receiving continuous positive airway
nia, San Diego, 528 infants; Case West- With adjustment for center and BW, pressure (CPAP) therapy at 24 hours
ern Reserve University, 415 infants; there were no differences in racial dis- increased among infants of 24
University of Cincinnati, 974 infants; tribution according to GA (Table 2). weeks, as did the proportion of infants
Duke University, 426 infants; Emory Early neonatal interventions differed who never underwent intubation. Al-
University, 516 infants; Indiana Univer- according to GA (Table 2). At 22 weeks, though the adjusted RR for BPD de-
sity, 720 infants; University of Iowa, 99 only 19% of infants underwent intuba- creased over time among infants who
infants; University of Miami, 515 in- tion and ventilation in the delivery survived to PMA of 36 weeks, the
fants; University of New Mexico, 97 room. Intubation rates increased to change was clinically insignicant.
infants; University of Rochester, 243 in- 68% at 23 weeks and 87% at 24 weeks
fants; Stanford University, 334 infants; and decreased at 24 weeks. Of 856 Neonatal Characteristics and
University of Texas Southwestern Med- infants who received resuscitation Morbidities Among Infants Who
ical Center at Dallas, 488 infants; Uni- drugs and/or chest compressions, Survived >12 Hours
versity of Texas Health Science Center 96% also underwent intubation. Rates Overall, 89% of infants born at GAs of
at Houston, 765 infants; Tufts Univer- of surfactant therapy increased from 22 to 28 weeks survived 12 hours.
sity, 137 infants; University of Utah, 269 17% at 22 weeks to 63% at 23 weeks Substantially more early survivors
infants; Wake Forest University, 465 in- and 90% at 24 weeks. The proportion of born at 22 to 24 weeks received resus-
fants; Wayne State University, 637 in- infants who died at 12 hours de- citation efforts (intubation, drug treat-
fants; Yale University, 526 infants. creased with increasing GA, from 85% ment, and/or chest compression) in
at 22 weeks to 1% to 2% at 27 to 28 the delivery room, compared with in-
RESULTS
weeks (Table 3). Risk of early death fants born at 22 to 24 weeks who died
Study Group was signicantly elevated for infants at 12 hours (22 weeks, 90% vs 7%; 24
A total of 9575 infants with GAs of 22 to born at 22 to 24 weeks, compared with weeks, 91% vs 59%). Signicant differ-
28 weeks and BWs of 401 to 1500 g infants born at 28 weeks (22 weeks, ences in resuscitation efforts between
were born at NRN centers between adjusted RR: 15.76 [95% condence in- those who survived 12 hours and
January 1, 2003, and December 31, terval [CI]: 10.1324.52]; 23 weeks, ad- those who did not were not seen
2007, and are included in this study. justed RR: 9.88 [95% CI: 6.48 15.08]; 24 among infants with GAs of 25 to 27
Overall, 25% of the cohort subjects weeks, adjusted RR: 2.90 [95% CI: 1.90 weeks. Among infants born at 28
were multiple births. 4.43]), but not for infants born at 25 to weeks, a smaller proportion of those
27 weeks. who survived 12 hours received re-
Maternal and Infant suscitation efforts in the delivery
Characteristics, Delivery Room Changes in Clinical Practices room, compared with those who died
Interventions, and Early Deaths Rates of prenatal steroid use in- within 12 hours (48% vs 65%; P .05).
Rates of prenatal steroid use in- creased by 1% per year during the Infants at the lowest GAs were at the
creased with increasing GA, from 13% study period, and rates of cesarean greatest risk for morbidities of prema-
at 22 weeks to 53% at 23 weeks and section delivery increased by 2% per turity (Tables 5 and 6). Overall, 93% in-
85% to 87% at 24 to 28 weeks (Table 1). year (Table 4). Rates of prenatal antibi- fants experienced respiratory dis-
Rates of prenatal antibiotic use were otic use decreased by 3% per year. tress. Rates of mechanical ventilation
lowest for mothers who delivered at 22 These trends did not vary according to at 24 hours decreased from 96% at 22
weeks (51%) and highest for those GA (year-GA interaction: for prenatal weeks to 40% at 28 weeks, and rates of
who delivered at 24 to 25 weeks (73%). steroid therapy, P .47; for cesarean CPAP therapy at 24 hours increased
Chorioamnionitis was documented section delivery, P .37; for prenatal from 0% at 22 weeks to 3% at 23 weeks,
more frequently in maternal records antibiotic treatment, P .66). Rates of 8% at 24 weeks, and 38% at 28 weeks.

PEDIATRICS Volume 126, Number 3, September 2010 445

TABLE 1 Maternal Demographic Features and Perinatal Information According to GA for VLBW Infants Born in NRN Centers Between January 1, 2003,
and December 31, 2007 (Including Infants Who Died Within 12 Hours After Birth)
Characteristic 22 wk 23 wk 24 wk 25 wk 26 wk 27 wk 28 wk Total
(N 421) (N 871) (N 1370) (N 1498) (N 1576) (N 1838) (N 2001) (N 9575)
Mothers age, ya
Mean (range) 27 (2232) 27 (2532) 27 (2531) 27 (2330) 27 (2530) 27 (2531) 27 (2532) 27 (2531)
SD (range) 6.9 (4.29.9) 6.4 (0.07.0) 6.5 (4.78.9) 6.6 (4.97.2) 6.5 (4.67.6) 6.7 (5.17.5) 6.8 (5.27.5) 6.6 (5.77.0)
High school degree, % (range) 68 (0100) 69 (25100) 72 (40100) 70 (3394) 72 (2590) 70 (0100) 72 (4189) 71 (3688)
Medical insurance, % (range)b
Medicaid/public insurance 51 (20100) 42 (20100) 50 (1571) 53 (3369) 50 (2971) 50 (2481) 49 (1978) 49 (2969)
Private insurance 35 (067) 45 (067) 39 (976) 38 (662) 40 (569) 40 (367) 42 (479) 40 (663)
Self-pay/uninsured 13 (040) 13 (051) 11 (042) 8 (038) 9 (030) 9 (022) 8 (031) 9 (131)
Other 1 (013) 1 (011) 1 (010) 1 (011) 1 (010) 1 (015) 1 (015) 1 (013)
1 prenatal visit, % (range)b 88 (33100) 92 (84100) 93 (85100) 94 (86100) 93 (80100) 95 (85100) 95 (89100) 94 (85100)
Diabetes mellitus, % (range)c 3 (020) 3 (017) 3 (010) 4 (08) 5 (215) 5 (011) 6 (115) 5 (38)
Hypertension, % (range)b 8 (026) 10 (028) 14 (040) 20 (938) 25 (1343) 27 (1542) 31 (1147) 22 (1437)
Prepartum hemorrhage, % 21 (067) 27 (076) 21 (940) 22 (556) 19 (054) 17 (535) 16 (623) 20 (932)
(range)b
Prenatal steroid treatment, % 13 (0100) 53 (10100) 85 (49100) 86 (62100) 86 (48100) 87 (57100) 86 (46100) 80 (4597)
(range)b
Prenatal antibiotic treatment, % 51 (2192) 65 (088) 73 (5690) 73 (4894) 68 (45100) 66 (5185) 64 (5089) 67 (5585)
(range)b
ROM 24 h before delivery, % 22 (045) 22 (042) 25 (840) 26 (1340) 28 (1550) 25 (1432) 24 (1634) 25 (1832)
(range)b
Mode of delivery, % (range)b
Vaginal, vertex 60 (20100) 53 (075) 32 (2056) 31 (1944) 33 (1143) 30 (1444) 30 (1248) 34 (1843)
Vaginal, breech 32 (080) 23 (056) 7 (022) 4 (013) 3 (013) 2 (04) 2 (05) 6 (012)
Vaginal, not otherwise specied 1 (033) 1 (06) 0 (00) 1 (05) 1 (02) 1 (01) 1 (01) 1 (01)
Cesarean section 7 (033) 24 (3100) 60 (2480) 65 (5479) 65 (5289) 68 (5586) 68 (4788) 59 (4781)
Infants born in 20062007 N 159 N 321 N 493 N 573 N 583 N 732 N 771 N 3632
Chorioamnionitis documented 28 (0100) 26 (0100) 20 (039) 19 (056) 19 (044) 15 (028) 14 (022) 18 (729)
in mothers medical record,
% (range)b
Placental pathologic evaluation 77 (36100) 86 (50100) 82 (50100) 83 (62100) 80 (46100) 80 (44100) 83 (0100) 82 (58100)
performed, % (range)
Placental pathologic evaluation N 123 N 272 N 401 N 475 N 461 N 585 N 634 N 2951
Histologic chorioamnionitis, 70 (25100) 61 (0100) 59 (0100) 51 (25100) 48 (073) 41 (2361) 34 (857) 48 (2673)
% (range)b
Ranges are across all participating NRN centers. Information was missing as follows: mothers age, 4 infants; mothers education, 2834 infants; mothers medical insurance, 300 infants;
prenatal care, 8 infants; diabetes mellitus, 8 infants; hypertension, 10 infants; prepartum hemorrhage, 8 infants; prenatal steroid treatment, 27 infants; prenatal antibiotic treatment, 30
infants; rupture of membranes date and/or time, 228 infants; mode of delivery, 9 infants; chorioamnionitis, 9 infants; placental pathologic evaluation, 26 infants; histologic chorioamnionitis,
17 infants. P values were determined with the Wald 2 test for differences according to GA, with adjustment for center and BW. ROM indicates rupture of membranes.
a P .05.

b P .001.

c P .01.

The risk of BPD was inversely related IVH, 6% grade 2 IVH, 7% grade 3 IVH, 9% 28 weeks); 11% of infants developed
to GA at birth. Because of the inclusion grade 4 IVH, 2% ventriculomegaly with- necrotizing enterocolitis (Table 6).
of infants with mild BPD (oxygen ther- out IVH, and 2% other abnormalities. Patent ductus arteriosus was diag-
apy for 28 days but use of room air at PVL was observed for 3% of infants nosed for 46% of infants, of whom 71%
36 weeks), more infants were classi- with sonograms performed in the rst were treated with indomethacin, 13%
ed as having BPD with the new, 28 days and 4% with sonograms per- ibuprofen (2006 2007), and 27% sur-
severity-based, denition of BPD (new formed after 28 days. Rates of abnor- gical closure. Among 7313 infants who
denition, 68%; traditional denition, mal ultrasound ndings decreased were still in the hospital at 28 days,
42%; physiologic denition, 40%). with increasing GA. 94% underwent an ophthalmologic ex-
Most infants who survived 12 hours Sepsis was diagnosed more frequently amination before hospital discharge,
underwent 1 cranial ultrasound at the lowest GA (rates of early-onset death, or transfer. Of the 6866 with ex-
evaluation within 28 days; 64% of re- sepsis were 6% at 22 weeks and 1% at amination ndings, 59% were diag-
sults were normal (Table 6). Overall, 28 weeks, and rates of late-onset sep- nosed as having ROP (96% at 22 weeks
10% of sonograms indicated grade 1 sis were 58% at 22 weeks and 20% at and 32% at 28 weeks), and 12% under-

446 STOLL et al
 ARTICLE

TABLE 2 Infant Demographic Features and Delivery Information According to GA for VLBW Infants Born in NRN Centers Between January 1, 2003, and
December 31, 2007 (Including Infants Who Died Within 12 Hours After Birth)
Characteristic 22 wk 23 wk 24 wk 25 wk 26 wk 27 wk 28 wk Total
(N 421) (N 871) (N 1370) (N 1498) (N 1576) (N 1838) (N 2001) (N 9575)
BW, ga
Mean (range) 511 (473621) 581 (549639) 651 (609677) 744 (709791) 854 (737891) 960 (9191009) 1082 (10221207) 836 (789903)
SD (range) 66.9 (30.4122) 92.0 (55.4139) 105 (90.6125) 135 (107162) 163 (133183) 189 (164218) 206 (160229) 241 (218259)
Male, % (range)a 58 (093) 55 (43100) 53 (4070) 53 (4681) 53 (4563) 55 (3766) 51 (3658) 53 (4758)
Race/ethnicity, %
(range)
Black, non-Hispanic 45 (0100) 38 (081) 41 (085) 41 (081) 39 (486) 38 (289) 36 (087) 39 (384)
Black, Hispanic 0 (00) 1 (010) 1 (010) 1 (06) 1 (05) 1 (03) 1 (05) 1 (03)
White, non-Hispanic 30 (080) 37 (063) 34 (470) 34 (071) 36 (462) 40 (379) 41 (588) 37 (571)
White, Hispanic 19 (067) 20 (0100) 18 (076) 19 (088) 19 (073) 18 (174) 17 (067) 18 (170)
American Indian/ 1 (020) 0 (00) 1 (040) 1 (013) 1 (028) 1 (010) 1 (030) 1 (020)
Alaska native
Asian/Pacic 4 (043) 3 (054) 3 (037) 3 (023) 3 (021) 3 (019) 3 (023) 3 (027)
islander
1 race/other 1 (019) 1 (014) 2 (026) 1 (021) 2 (022) 1 (09) 1 (011) 1 (017)
Intrauterine growth 0 (00) 4 (016) 6 (030) 8 (014) 8 (120) 10 (515) 9 (015) 8 (510)
restriction, %
(range)a
Multiple birth, % 28 (048) 30 (11100) 25 (732) 21 (640) 22 (840) 25 (040) 28 (1637) 25 (1834)
(range)a
Delivery room
resuscitation, %
(range)
Endotracheal 19 (0100) 68 (10100) 87 (53100) 82 (5398) 75 (3292) 65 (3190) 47 (1082) 67 (4185)
intubationa
Resuscitation druga 3 (020) 8 (032) 9 (032) 6 (028) 5 (022) 4 (019) 2 (07) 5 (116)
Chest compressiona 3 (040) 10 (024) 13 (040) 10 (137) 7 (022) 6 (015) 4 (014) 8 (219)
Apgar score of 3, %
(range)a
At 1 mina 89 (0100) 73 (50100) 53 (3071) 44 (2563) 36 (2253) 32 (1748) 23 (1230) 42 (2953)
At 5 mina 86 (0100) 49 (089) 20 (040) 12 (325) 8 (022) 7 (114) 4 (09) 16 (325)
Admission
temperature,
Ca,b
Mean (range) 34.7 (31.337.0) 35.0 (33.236.6) 35.4 (34.237.0) 35.8 (34.836.9) 36.1 (35.137.0) 36.2 (35.137.1) 36.2 (35.137.2) 35.9 (34.837.0)
SD (range) 1.7 (0.13.2) 1.7 (0.11.9) 1.4 (0.71.5) 1.1 (0.61.3) 1.0 (0.51.2) 0.9 (0.51.1) 0.9 (0.41.2) 1.2 (0.71.3)
Surfactant therapy, % 17 (0100) 63 (10100) 90 (58100) 88 (72100) 85 (56100) 78 (4394) 65 (4186) 76 (5888)
(range)a
Ranges are across all participating NRN centers. Information was missing as follows: gender, 2 infants; race/ethnicity, 24 infants; intrauterine growth restriction, 2 infants; endotracheal
intubation, 9 infants; resuscitation drug, 13 infants; chest compression, 13 infants; Apgar score at 1 minute, 78 infants; Apgar score at 5 minutes, 76 infants; temperature, 1097 infants.
a P .001 from the Wald 2 test for differences according to GA, with adjustment for center and BW. Differences in BW were adjusted for center effects only. Race/ethnicity was tested as

black, white, or other.

b Infant temperature at initial admission to the nursery for infants born in 20032005 and rst temperature reading obtained within 60 minutes after birth for infants born in 2006 2007.

went treatment for ROP (50% at 22 Survival and Morbidity Rates (All signicant for infants born at 24 to 27
weeks and 2% at 28 weeks). A total of 9575 Infants) weeks, compared with 28 weeks (24
2630 infants evaluated in 2006 2007 weeks, adjusted RR: 2.52 [95% CI: 2.10
had ROP outcomes recorded at the Rates of survival to discharge in-
3.04]; 27 weeks, adjusted RR: 1.23 [95%
time of discharge or 120 days of age. creased with increasing GA, from 6%
CI: 1.011.49]). Rates of survival to dis-
Among those infants, 39% had favor- at 22 weeks to 92% at 28 weeks (72%
charge according to GA did not change
able outcomes, 7% had unfavorable overall) (Fig 1 and Table 3). Infants
during the study period (Table 4).
outcomes with severe ROP requiring born at 22 to 23 weeks had 3 times
treatment, and 53% had undetermined the risk of death, compared with in- Neonatal morbidities occurred fre-
ROP outcomes (ie, had not reached the fants born at 28 weeks (22 weeks, ad- quently among survivors. Rates of sur-
threshold for surgery or were still im- justed RR: 3.88 [95% CI: 3.18 4.73]; 23 vival with morbidity decreased from
mature and required further examina- weeks, adjusted RR: 3.56 [95% CI: 2.95 100% at 22 weeks to 92% at 23 weeks,
tion) (Table 6). 4.30]). RRs decreased but remained 91% at 24 weeks, 80% at 25 weeks, 66%

PEDIATRICS Volume 126, Number 3, September 2010 447

TABLE 3 Mortality Rates According to GA for VLBW Infants Born in NRN Centers Between January 1, 2003, and December 31, 2007
% (Range)
22 wk 23 wk 24 wk 25 wk 26 wk 27 wk 28 wk Total
(N 421) (N 871) (N 1370) (N 1498) (N 1576) (N 1838) (N 2001) (N 9575)
Survived 6 (050) 26 (253) 55 (20100) 72 (5090) 84 (61100) 88 (76100) 92 (88100) 72 (5595)
Died 94 (50100) 74 (4798) 45 (080) 28 (1050) 16 (039) 12 (024) 8 (012) 28 (545)
Time of deatha
12 h 85 (0100) 43 (090) 11 (044) 5 (019) 3 (011) 1 (05) 2 (07) 11 (125)
1224 h 2 (06) 3 (07) 2 (06) 1 (03) 1 (02) 1 (02) 1 (01) 1 (02)
13 d 1 (08) 9 (030) 6 (011) 3 (025) 2 (08) 1 (06) 1 (04) 3 (07)
47 d 2 (023) 4 (020) 4 (011) 3 (07) 1 (08) 1 (06) 1 (02) 2 (05)
814 d 2 (050) 5 (050) 5 (020) 3 (09) 2 (06) 2 (019) 1 (05) 3 (18)
1528 d 1 (015) 4 (016) 7 (015) 4 (08) 3 (011) 2 (05) 2 (07) 3 (06)
29 d 1 (08) 6 (017) 10 (030) 8 (015) 5 (010) 4 (08) 2 (05) 5 (19)
Survived N 25 N 226 N 748 N 1078 N 1319 N 1616 N 1847 N 6859
Survived without morbidityb 0 (00) 8 (014) 9 (018) 20 (043) 34 (049) 44 (1965) 57 (674) 37 (750)
Died N 396 N 645 N 622 N 420 N 257 N 222 N 154 N 2716
Respiratory support 82 (40100) 77 (0100) 66 (2196) 68 (0100) 73 (42100) 66 (20100) 60 (0100) 72 (2995)
withheld/withdrawn
before deathc
Died at 12 h N 359 N 375 N 147 N 72 N 46 N 27 N 34 N 1060
Respiratory support 85 (40100) 85 (43100) 79 (0100) 86 (0100) 78 (0100) 85 (0100) 79 (25100) 84 (53100)
withheld/withdrawn
before deathd
Ranges are across all participating NRN centers.
a Proportions among all infants including survivors.

b Proportions among infants who survived. Morbidities included severe IVH, PVL, BPD, necrotizing enterocolitis, infections, and ROP stage 3.

c Proportions among infants who died. Data on respiratory support withheld/withdrawn were missing for 52 infants.

d Proportions among infants who died within 12 hours. Data on respiratory support withheld/withdrawn were missing for 2 infants.

at 26 weeks, 56% at 27 weeks, and 43% hort of all preterm infants born at GAs formation on decision-making pro-
at 28 weeks. Infection and BPD were of 27 weeks in Sweden in 2004 2007 cesses to help explain differences. In
the most-frequent morbidities. Al- demonstrated survival rates higher our cohort, rates of active obstetric in-
though unadjusted rates of survival than rates reported for other coun- tervention, as indicated by prenatal
without major morbidity seemed un- tries or reported previously for Swe- steroid administration and cesarean
changed, the adjusted RR for survival den.20 Our study reviewed neonatal section delivery, increased markedly
without morbidity increased over time morbidity and mortality rates for a after 23 weeks of gestation. Prenatal
(Table 4). The median length of hospi- large cohort of extremely preterm in- steroid use was almost twice as fre-
tal stay among survivors was 84 days, fants, to evaluate changes in clinical quent for infants born at GAs of 24 to 28
and lengths of stay decreased with in- practice and contemporary outcomes weeks, compared with infants born
creasing GA, from 141 days at 22 weeks at US academic centers. Although pre- earlier. Similarly, rates of neonatal in-
to 63 days at 28 weeks (P .001). PMA vious reports from the NRN used BW as terventions and intensive care, mea-
at discharge decreased from 42 weeks the reference for morbidity and sur- sured as active resuscitation with ven-
for surviving infants born at GAs of 22 vival rates, the current study assessed tilation in the delivery room, increased
weeks to 37 weeks for those born at 28 outcomes according to GA. Apprecia- substantially between 22 and 23 weeks
weeks (Fig 2).
tion of GA-based outcomes is particu- (19% vs 68%). Rates of death at 12
DISCUSSION larly valuable for prenatal counseling hours, which in part reect willing-
and physician/family decision-making. ness to provide intensive care to the
Although VLBW infant mortality rates
The decisions to provide active obstet- most-immature infants, decreased
in the United States decreased sub-
stantially in the 1980s and early ric care and to initiate neonatal inten- with increasing GA, from 85% of in-
1990s,3 6,1518 most reports, including sive care for the most-immature in- fants at 22 weeks to 2% of infants at 28
ndings for this cohort, failed to dem- fants remain controversial. Center weeks.
onstrate further progress in reducing differences in obstetric/early neonatal In-hospital morbidity rates remain
neonatal morbidity and mortality interventions were identied, but we high among extremely preterm in-
rates.6,1619 In contrast, a population co- did not collect sufciently detailed in- fants, and morbidities contribute

448 STOLL et al
 ARTICLE

TABLE 4 Clinical Practice Indicators and Survival Rates According to Birth Year for 9575 VLBW Infants Born in NRN Centers Between January 1, 2003,
and December 31, 2007
Characteristic Percent Adjusted RR P
(95% CI)
2003 2004 2005 2006 2007
(N 1919) (N 1992) (N 2032) (N 1900) (N 1732)
Prenatal steroid treatment, all infants 81 76 80 79 83 1.01 (1.001.02) .01
Prenatal antibiotic treatment, all infants 72 68 68 63 66 0.97 (0.960.98) .001
Cesarean section, all infants 57 58 62 60 60 1.02 (1.001.03) .01
Delivery room endotracheal intubation
22 wk 17 22 21 16 19 0.98 (0.851.13) .8
23 wk 69 67 67 68 67 0.99 (0.961.02) .6
24 wk 89 89 88 83 86 0.98 (0.971.00) .04
25 wk 86 86 79 81 77 0.97 (0.950.99) .001
26 wk 83 78 68 76 69 0.95 (0.940.97) .001
27 wk 69 70 64 58 61 0.96 (0.940.98) .001
28 wk 54 58 45 41 38 0.90 (0.870.93) .001
All infants 71 72 66 63 62
Surfactant therapy
22 wk 18 17 21 13 18 0.97 (0.831.13) .7
23 wk 65 66 63 63 59 0.98 (0.941.01) .2
24 wk 91 88 89 88 93 1.00 (0.991.01) .9
25 wk 89 89 85 91 87 0.99 (0.981.00) .2
26 wk 89 83 84 87 82 0.99 (0.971.00) .045
27 wk 78 80 75 77 78 1.00 (0.981.01) .6
28 wk 73 70 61 62 59 0.94 (0.920.96) .001
All infants 78 77 74 75 74
Survived to discharge
22 wk 6 7 5 3 8
23 wk 27 21 33 27 21
24 wk 56 53 55 55 54
25 wk 71 72 70 75 71
26 wk 82 87 82 83 84
27 wk 88 86 88 88 91
28 wk 94 89 93 93 92
All infants 72 70 71 73 72 1.00 (0.991.01) .3
Survived 12 h N 1709 N 1762 N 1818 N 1689 N 1537
Never intubated, all infants who survived 12 ha 8 8 10 10 11 1.12 (1.071.18) .001
Necrotizing enterocolitis, all infants who survived 12 h 9 11 11 11 12 1.04 (0.991.09) .1
Survived 24 h N 1685 N 1738 N 1785 N 1678 N 1532
CPAP therapy at 24 h
22 wk 0 0 0 0 0
23 wk 5 4 4 1 1 0.73 (0.511.05) .09
24 wk 5 7 11 6 8 1.09 (0.951.24) .2
25 wk 12 14 22 18 23 1.16 (1.071.25) .001
26 wk 27 32 32 25 34 1.01 (0.961.07) .6
27 wk 32 35 35 37 40 1.05 (1.01.09) .04
28 wk 36 36 36 41 40 1.04 (1.01.08) .06
All infants who survived 24 h 23 25 26 27 29
Survived 72 h N 1626 N 1672 N 1721 N 1631 N 1478
Late-onset sepsis, all infants who survived 72 h 36 38 37 36 33 0.98 (0.961.0) .1
Cranial sonography performed within 28 d after birth N 1660 N 1708 N 1749 N 1646 N 1497
Severe IVH, all infants with sonograms 16 16 14 17 16 0.96 (0.931.0) .05
Infants who underwent cranial imaging before and/or after 28 d N 1665 N 1714 N 1752 N 1651 N 1500
PVL, all infants with imaging ndings 4 5 5 4 5 0.93 (0.871.0) .06
Survived to PMA of 36 wk N 1426 N 1455 N 1483 N 1421 N 1280
BPD, infants who survived to PMA of 36 wk 43 42 40 43 43 0.94 (0.920.95) .001
Survived to discharge N 1385 N 1403 N 1445 N 1383 N 1243
Survived without morbidityb
22 wk 0 0 0 0 0
23 wk 14 10 3 5 9
24 wk 5 10 10 8 11
25 wk 22 20 21 17 20
26 wk 32 38 34 31 34
27 wk 44 44 46 42 44
28 wk 58 55 62 55 54
All infants who survived to discharge 37 37 38 35 36 1.04 (1.021.06) .001
Information was missing as follows: prenatal steroid treatment, 27 infants; prenatal antibiotic treatment, 30 infants; cesarean section delivery, 9 infants; delivery room endotracheal
intubation, 9 infants; surfactant therapy, 10 infants; never intubated, 3 infants; necrotizing enterocolitis, 1 infant; CPAP therapy, 14 infants; late-onset sepsis, 2 infants; severe IVH, 9 infants;
PVL, 1 infant; BPD, 42 infants; survived without morbidity, 32 infants. RRs and P values were determined for the change per year from a modied Poisson model that included effects for study
center, infant BW and GA, and year and, where signicant, effects for the year-GA interaction (delivery room intubation, surfactant therapy and CPAP therapy at 24 h). RRs are shown for all
infants overall in cases in which the year-GA interaction was not signicant and separately for infants born at each GA in cases in which the interaction was signicant. The year-GA interaction
could not be assessed for the category of never intubated because of small sample sizes.
a Never used conventional or high-frequency ventilator or underwent nasal synchronized intermittent mandatory ventilation.

b Morbidities included BPD, severe IVH, PVL, necrotizing enterocolitis, ROP stage 3, and infections (early-onset sepsis, late-onset sepsis, or meningitis). Proportions were determined among

survivors. Of the 25 surviving infants born at GA of 22 weeks, none survived without major morbidity.

PEDIATRICS Volume 126, Number 3, September 2010 449

TABLE 5 Pulmonary Morbidities According to GA for VLBW Infants Who Were Born in NRN Centers Between January 1, 2003, and December 31, 2007, and
Survived 12 Hours After Birth
Characteristic % (Range)
22 wk 23 wk 24 wk 25 wk 26 wk 27 wk 28 wk Total
(N 62) (N 496) (N 1223) (N 1426) (N 1530) (N 1811) (N 1967) (N 8515)
Respiratory distress syndromea 95 (75100) 98 (75100) 98 (64100) 97 (77100) 94 (61100) 90 (50100) 86 (55100) 93 (6099)
Surfactant therapya 97 (50100) 97 (90100) 95 (83100) 90 (72100) 86 (58100) 78 (4395) 65 (4186) 82 (6493)
Pneumothorax 15 (040) 11 (033) 11 (023) 9 (420) 7 (015) 5 (014) 4 (09) 7 (313)
Pulmonary hemorrhageb 16 (050) 15 (050) 13 (640) 10 (328) 7 (220) 4 (014) 3 (07) 7 (318)
Postnatal steroid treatmenta 15 (050) 18 (050) 20 (060) 14 (044) 9 (030) 6 (014) 2 (06) 10 (024)
Never intubateda,c 0 (00) 1 (06) 1 (02) 2 (08) 5 (014) 12 (040) 23 (644) 9 (222)
Respiratory support at 24 h for infants who N 55 N 471 N 1192 N 1414 N 1520 N 1804 N 1962 N 8418
survived 24 h
Conventional or high-frequency 96 (0100) 94 (83100) 89 (71100) 76 (5795) 61 (4392) 49 (2174) 40 (2061) 62 (4783)
ventilationa,d
Nasal SIMVb,d 0 (00) 1 (06) 2 (016) 3 (020) 3 (018) 2 (012) 3 (016) 3 (014)
CPAP therapya,d 0 (00) 3 (010) 8 (029) 18 (530) 30 (549) 36 (1279) 38 (1766) 26 (846)
Use of oxygen alonea,d 2 (0100) 1 (06) 1 (07) 2 (07) 2 (013) 3 (010) 5 (015) 3 (19)
Infants who survived to PMA of 36 wk N 27 N 241 N 790 N 1121 N 1344 N 1648 N 1852 N 7023
BPD (oxygen use at 36 wk)a,e 85 (0100) 73 (35100) 69 (31100) 55 (20100) 44 (19100) 34 (1376) 23 (988) 42 (2089)
Infants in hospital at PMA of 36 wk or N 27 N 231 N 774 N 1088 N 1284 N 1565 N 1739 N 6708
discharged/transferred at 3336 wk
Severity-based BPDa,f
Mild BPD 15 (0100) 26 (050) 26 (067) 37 (062) 35 (058) 28 (052) 16 (035) 27 (538)
Moderate BPD 30 (0100) 35 (0100) 34 (068) 29 (970) 26 (571) 20 (455) 15 (057) 23 (860)
Severe BPD 56 (0100) 39 (0100) 37 (0100) 26 (386) 17 (444) 13 (030) 8 (029) 18 (340)
Infants born in 20062007 N 19 N 174 N 438 N 547 N 566 N 728 N 754 N 3226
Inhaled nitric oxide treatmentb,g 11 (050) 8 (050) 10 (054) 8 (027) 7 (025) 3 (012) 3 (014) 6 (019)
Infants who survived to PMA of 36 wk N9 N 83 N 274 N 422 N 482 N 650 N 691 N 2611
BPD by physiologic denitiona,h 89 (50100) 70 (0100) 68 (0100) 55 (19100) 44 (6100) 31 (0100) 22 (0100) 40 (1582)
Ranges are across all participating NRN centers. Proportions are among all infants who survived 12 hours, except as noted. Information was missing as follows: respiratory distress
syndrome, 5 infants; surfactant treatment, 7 infants; pulmonary hemorrhage, 2 infants; postnatal steroid treatment, 41 infants; never intubated, 3 infants; ventilator use at 24 hours, 13
infants; nasal synchronized intermittent mandatory ventilation at 24 hours, 14 infants; CPAP at 24 hours, 14 infants; oxygen alone at 24 hours, 14 infants; nitric oxide use, 1 infant. P values
were determined with the Wald 2 test for differences according to GA, with adjustment for center and BW or, for nasal synchronized intermittent mandatory ventilation and inhaled nitric
oxide use, with adjustment for BW only. SIMV indicates synchronized intermittent mandatory ventilation.
a P .001.

b P .05.

c Never used conventional or high-frequency ventilator or underwent nasal synchronized intermittent mandatory ventilation.

d Proportions among infants who survived 24 hours after birth. Use of oxygen alone at 24 hours was dened as receiving supplemental oxygen without conventional or high-frequency

ventilation, nasal synchronized intermittent mandatory ventilation, or CPAP therapy.

e Proportions among infants who survived to PMA of 36 weeks and had nonmissing outcome data (N values shown). BPD could not be determined for 42 infants.

f Proportions among infants who were still in the hospital at PMA of 36 weeks or, if discharged or transferred before 36 weeks, were in the hospital for 28 days and until PMA of 33 weeks.

N values are shown for infants with nonmissing outcome data. Severity-based BPD could not be determined for 68 infants. More information about severity-based BPD is presented in the
text.
g Proportions among infants born in 2006 2007.

h Proportions among infants born in 2006 2007 who survived to PMA of 36 weeks and had nonmissing outcome data (N values shown). BPD according to the physiologic denition could not

be determined for 90 infants. Information on how BPD was determined according to the physiologic denition is presented in the text.

to adverse neurodevelopmental out- individual infants and has limited data less prophylactic use of surfactant,
comes. The majority of infants studied on the severity of illness at birth, infor- and alternative modes of respiratory
experienced a major complication dur- mation that might permit more- support. Rates of endotracheal intuba-
ing the initial hospitalization, with the detailed evaluation and understanding tion in the delivery room decreased in
risk of morbidity being inversely re- of center differences. Reducing the recent years among infants of 24
lated to GA at birth. Center differences high rates of in-hospital morbidity weeks, with a corresponding increase
in the proportions of infants with spe- among extremely low GA infants who in CPAP therapy use at 24 hours of life.
cic morbidities were noted. At the are provided ongoing intensive care At GA of 28 weeks, use of surfactant
lowest GAs (2224 weeks), small num- remains a challenge for clinicians and decreased in the most-recent years.
bers of infants at some centers con- investigators. Furthermore, the proportion of infants
tributed to the variability. The registry who survived 12 hours without ever
does not collect data on the reasons To reduce rates of BPD, attention is be- undergoing intubation and ventilation
behind the choice of interventions for ing paid to avoidance of intubation, increased with increasing GA and

450 STOLL et al
 ARTICLE

TABLE 6 Rates of Infections and Other Morbidities According to GA for VLBW Infants Who Were Born in NRN Centers Between January 1, 2003, and
December 31, 2007, and Survived 12 Hours After Birth
Characteristic % (Range)
22 wk 23 wk 24 wk 25 wk 26 wk 27 wk 28 wk Total
(N 62) (N 496) (N 1223) (N 1426) (N 1530) (N 1811) (N 1967) (N 8515)
Early-onset sepsisa 6 (067) 4 (020) 4 (09) 2 (07) 2 (06) 2 (06) 1 (04) 2 (14)
Meningitisb 0 (00) 5 (025) 3 (012) 4 (015) 3 (09) 1 (05) 1 (05) 3 (08)
Late-onset sepsisa,c 58 (0100) 62 (086) 55 (2974) 46 (2467) 35 (1453) 27 (1552) 20 (436) 36 (1851)
NECb 5 (033) 12 (050) 15 (022) 13 (524) 9 (025) 10 (021) 8 (320) 11 (419)
NEC managed medicallyd 67 (50100) 31 (0100) 39 (0100) 52 (10100) 48 (17100) 47 (2375) 58 (0100) 48 (21100)
NEC treated surgicallyd 33 (050) 69 (0100) 61 (0100) 48 (090) 52 (083) 53 (2577) 42 (0100) 52 (079)
PDAa 55 (13100) 54 (21100) 60 (3180) 55 (2592) 48 (2188) 42 (1480) 32 (1360) 46 (2678)
Indomethacin therapy for PDAe,f 82 (0100) 73 (0100) 76 (2596) 72 (2994) 69 (794) 68 (3794) 67 (3195) 71 (3391)
Surgical treatment of PDAa,e 50 (0100) 43 (077) 40 (1362) 33 (1365) 24 (644) 16 (033) 12 (033) 27 (1041)
Infants in hospital at 28 d N 30 N 277 N 874 N 1197 N 1386 N 1683 N 1866 N 7313
ROP examination performeda,g 93 (50100) 91 (71100) 93 (50100) 94 (78100) 96 (67100) 95 (87100) 92 (6899) 94 (8299)
ROP diagnoseda,h 96 (50100) 88 (0100) 89 (50100) 79 (2994) 65 (2081) 49 (1875) 32 (556) 59 (3575)
ROP stage 3a,h 57 (0100) 48 (0100) 42 (2577) 25 (1154) 14 (029) 7 (014) 3 (011) 16 (628)
Intervention/surgical treatment for 50 (0100) 40 (0100) 35 (1758) 17 (040) 8 (021) 4 (09) 2 (07) 12 (422)
ROPa,h
Infants in hospital with weight measured N 24 N 215 N 736 N 976 N 1106 N 1231 N 1204 N 5492
at PMA of 36 wk
Growth failure at 36 wka,i 92 (50100) 91 (0100) 85 (67100) 83 (63100) 79 (3398) 76 (4298) 73 (4496) 79 (5997)
Cranial ultrasonography performed 85 (50100) 92 (67100) 95 (87100) 97 (85100) 98 (92100) 98 (94100) 98 (90100) 97 (93100)
within 28 d after birth
Sonogram ndings within 28 da,j N 53 N 454 N 1163 N 1385 N 1499 N 1781 N 1925 N 8260
Normal 32 (0100) 41 (1374) 49 (1470) 57 (3084) 65 (3690) 70 (5083) 77 (5091) 64 (4379)
IVH grade 1 13 (040) 9 (050) 11 (043) 9 (017) 11 (023) 10 (524) 10 (032) 10 (523)
IVH grade 2 13 (050) 9 (025) 9 (029) 8 (219) 5 (014) 5 (014) 4 (025) 6 (212)
IVH grade 3 8 (033) 15 (047) 12 (520) 8 (015) 7 (014) 6 (015) 4 (010) 7 (313)
IVH grade 4 30 (067) 21 (050) 14 (033) 13 (336) 7 (031) 5 (117) 3 (015) 9 (423)
Ventriculomegaly, no IVH 4 (033) 3 (013) 3 (06) 3 (06) 2 (09) 2 (06) 1 (05) 2 (04)
PVL within 28 db,k 6 (033) 4 (025) 3 (011) 4 (018) 3 (08) 2 (08) 2 (05) 3 (16)
Infants born in 20062007 N 19 N 174 N 438 N 547 N 566 N 728 N 754 N 3226
PDAa,l 53 (0100) 52 (13100) 56 (0100) 55 (20100) 51 (12100) 43 (080) 34 (063) 47 (2378)
Ibuprofen therapy for PDAl 0 (00) 13 (064) 16 (050) 16 (060) 13 (064) 12 (044) 11 (060) 13 (052)
Infants in hospital at 28 d N 11 N 92 N 320 N 471 N 508 N 678 N 718 N 2798
ROP examination performedf,m 91 (50100) 91 (71100) 92 (50100) 94 (75100) 95 (67100) 96 (82100) 93 (68100) 94 (82100)
ROP outcome at statusa,n
Determined, favorable in both eyes 10 (0100) 27 (0100) 28 (062) 31 (086) 38 (0100) 46 (0100) 46 (5100) 39 (883)
Determined, severe ROP in either eye 30 (0100) 30 (0100) 21 (067) 11 (038) 5 (025) 3 (09) 1 (09) 7 (020)
Undetermined ROP status in either 60 (0100) 43 (0100) 51 (082) 58 (14100) 57 (0100) 51 (0100) 53 (095) 53 (888)
eye (neither had severe ROP)
Ranges are across all participating NRN centers. Proportions are among all infants who survived 12 hours, except as noted. Information was missing as follows: patent ductus arteriosus,
4 infants; indomethacin therapy for patent ductus arteriosus, 38 infants; surgical treatment for patent ductus arteriosus, 2 infants, necrotizing enterocolitis, 1 infant; early-onset sepsis, 1
infant; meningitis, 1 infant; late-onset sepsis, 2 infants; ROP examination performed, 1 infant; ROP, 1 infant; ROP stage 3, 7 infants; intervention/surgical treatment for ROP, 32 infants. P
values were determined with the Wald 2 test for differences according to GA, with adjustment for center and BW or, for early-onset sepsis, meningitis, and ibuprofen therapy for patent
ductus arteriosus, with adjustment for BW only. PDA indicates patent ductus arteriosus; NEC, necrotizing enterocolitis.
a P .001.

b P .01.

c Proportions among infants who survived 3 days after birth.

d Proportions among infants with necrotizing enterocolitis.

e Proportions among infants with patent ductus arteriosus.

f P .05.

g Proportions among infants who were still in the hospital at 28 days of life.

h Proportions among infants who were still in the hospital at 28 days and underwent a ROP examination.

i Growth failure was dened as 10th percentile for gender at PMA of 36 weeks. Proportions were determined among infants who were alive and in the hospital at PMA of 36 weeks and who

had weight measurements at PMA of 35 to 37 weeks. Among those in the hospital at 36 weeks, weight data were missing for 101 infants with other measurements taken at PMA of 36 weeks,
weight was measured but not in the 35- to 37-week period for 21 infants, and growth failure could not be determined for 2 infants with missing gender information.
j From the sonogram with the most-severe ndings. Proportions were determined among those who had sonograms performed within 28 days after birth. Categories shown are mutually

exclusive. Some infants with IVH also had PVL, that is, 18 (2%) of 832 infants with grade 1 IVH, 21 (4%) of 506 infants with grade 2 IVH, 31 (5%) of 598 infants with grade 3 IVH, and 117 (16%)
of 710 infants with grade 4 IVH. Ten (6%) of 173 infants with ventriculomegaly and no IVH also had PVL.
k Proportions indicate the proportions of infants with PVL overall, with and without IVH, among infants who had sonograms performed within 28 days after birth.

l Proportions were determined among infants born in 2006 2007. Ibuprofen treatment proportions are among infants who had patent ductus arteriosus. Information was missing as follows:

patent ductus arteriosus, 1 infant; ibuprofen treatment, 1 infant.

m Proportions among infants born in 2006 2007 who were still in the hospital at 28 days of life. Information was missing for ROP examination for 1 infant.

n Proportions among infants born in 2006 2007 who were still in the hospital at 28 days and had ROP examinations performed. An assessment of ophthalmologic outcome at the time of

status was made as follows:(1) favorable, dened as one of the following in each eye: vessels mature, vessels in zone III in 2 consecutive examinations, acute ROP of stage 1 or 2 in zone III
in 2 consecutive examinations, or ROP in zone II or zone III but determined to be clearly regressing; (2) unfavorable, dened as severe ROP on the basis of one of the following in either eye:
received surgical treatment for ROP, met criteria for undergoing surgery, or retinal detachment resulting from ROP; or (3) undetermined outcome, dened as one of the following in either
eye: immature vessels in zone I or zone II, immature vessels reaching zone III in any single examination, stage 1 or 2 ROP in zone III at any single examination, stage 2 or 3 ROP in zone II or
III not regressing, or active ROP in zone I or zone II. Data on outcome at status was missing for 13 infants.

PEDIATRICS Volume 126, Number 3, September 2010 451

 100

90

80

70

60
Percentage, %

50

40

30

20

10

0
22 23 24 25 26 27 28 Overall
Gestational Age, wk

FIGURE 1
Survival to discharge according to GA among 9575 VLBW infants born in NICHD NRN centers between January 1, 2003, and December 31, 2007. The thin lines
indicate ranges across centers.

50

45
Length of Stay or PMA at Discharge, wk

40

35

30

25

20

15

10

0
22 23 24 25 26 27 28
Length of stay 20.1 18.3 16.7 14.7 12.6 10.7 9.0
PMA at discharge 42.1 41.3 40.7 39.7 38.6 37.7 37.0
Gestational Age at Birth

FIGURE 2
Median length of hospitalization (in weeks) and median PMA at discharge (in weeks) according to GA at birth among 6859 VLBW infants who were born in
NICHD NRN centers between January 1, 2003, and December 31, 2007, and survived to discharge.

452 STOLL et al
 ARTICLE

more-recent year of birth. With sub- hort, GA was estimated on the basis of derscores the need for hypothesis-
stantially increased use of CPAP ther- ultrasound ndings. The authors of the driven clinical trials to assess the
apy, it was surprising that overall Swedish study noted that a limitation efcacy of current neonatal interven-
rates of BPD were unchanged, al- of the use of ultrasonography to deter- tions.2124 Clinicians and investigators
though the adjusted RR for BPD de- mine GA is that erroneously low GAs are challenged to identify and to test
creased over the study period. might be estimated for infants with currently available interventions and
This is the rst study to report ophthal- growth restriction. Given the decrease resources that yield consistently lower
mologic status as favorable, unfavor- in mortality rates with increasing GA, morbidity and mortality rates at some
able, or undetermined at the time of underestimation of GA by as little as 1 centers, so that we can improve rates
the last in-hospital examination. Al- week might explain in part the differ- of survival without major morbidities
though 7% of all infants had severe ence in mortality rates between the 2 and reduce long-term neurodevelop-
ROP, the rate was 30% for infants with cohorts. Greater use of prenatal ste- mental impairments for all infants.
GAs of 22/23 weeks. Of concern, 53% of roid treatment at all GAs and of surfac-
ACKNOWLEDGMENTS
infants had undetermined ophthalmo- tant therapy at 22 to 23 weeks also
The National Institutes of Health pro-
logic status at the last examination be- might have contributed to differences
vided grant support for the NRN Ge-
fore discharge. This nding has impli- between the 2 cohorts.
neric Database Study (Recruitment
cations for discharge planning and During the 5-year study period, there 20032007). This study was supported
underscores the importance of a med- was no substantial improvement in in part by PHS grant UL1 RR025008
ical home, to ensure careful ophthal- rates of survival to discharge for ex- from the Clinical and Translational Sci-
mologic follow-up monitoring of these tremely low gestation infants born at ence Award program, National Insti-
vulnerable infants after discharge NRN centers. However, each additional tutes of Health, National Center for Re-
home or transport to a community week of GA at birth had substantial search Resources.
hospital. survival advantage; the most marked
The following investigators, in addition
Although ours is not a population- changes were between GAs of 22 and
to those listed as authors, participated
based study, we included all extremely 25 weeks, with survival rates increas- in this study: NRN Steering Committee
low gestation births at 20 academic ing from 6% to 72%. Furthermore, chairs: Alan H. Jobe, MD, PhD, Univer-
centers across the United States that rates of survival to discharge without sity of Cincinnati (20032006); Michael
together represent 110 000 live major morbidities increased dramati- S. Caplan, MD, Pritzker School of Medi-
births per year, an annual birth cohort cally between 22 and 25 weeks, with cine, University of Chicago (2006
equal in size to the Swedish national continued steady improvement for 2007); Case Western Reserve Universi-
cohort described recently.20 The rate each additional week of gestation. PMA ty: Rainbow Babies and Childrens
of extremely low gestation birth was at discharge for VLBW infants, a proxy Hospital (National Institutes of Health
vefold higher in our NRN cohort (10 measure of length of stay and a reec- grants GCRC M01 RR80 and U10
births at 27 weeks per 1000 infants) tion of the cost of care, was inversely HD21364): Avroy A. Fanaroff, MD; Cin-
than in the Swedish cohort (2.3 births related to GA at birth. Each additional cinnati Childrens Hospital Medical
at 27 weeks per 1000 infants). This week of GA at birth reduced PMA at Center: University of Cincinnati Hospi-
remarkable difference may be ex- discharge by almost 1 week and total tal and Good Samaritan Hospital (Na-
plained in part by Swedens universal length of hospital stay by 2 weeks, a tional Institutes of Health grants GCRC
health insurance, with free prenatal reection of both severity of illness M01 RR8084 and U10 HD27853): Ed-
care and associated social services, as and complications of prematurity ward F. Donovan, MD; Kate Bridges, MD;
well as an ethnically more homoge- among these very immature infants. Barbara Alexander, RN; Cathy Grisby,
neous and somewhat older pregnant Although adjusted RRs for survival BSN, CCRC; Marcia Worley Mersmann,
population. The high rates of prematu- without morbidity increased over time, RN, CCRC; Holly L. Mincey, RN, BSN; Jody
rity in our cohort underscore the im- the burden of in-hospital complica- Hessling, RN; Duke University: Univer-
portance of the current health care de- tions remained high. Retrospective sity Hospital, Alamance Regional Medi-
bate in the United States. Survival analyses of center differences and cal Center, and Durham Regional Hos-
rates for extremely low gestation in- benchmarking studies to identify best pital (National Institutes of Health
fants born at NRN centers are lower performance have been unable to grants GCRC M01 RR30 and U10
than those reported from Sweden. For identify modiable practices that con- HD40492): C. Michael Cotten, MD, MHS;
nearly all infants in the Swedish co- sistently improve outcomes, which un- Kathy J. Auten, MSHS; Melody B. Lohm-

PEDIATRICS Volume 126, Number 3, September 2010 453

eyer, RN, MSN; Emory University: Chil- Shirley S. Cosby, RN, BSN; University of HD21373): Jon E. Tyson, MD, MPH; Es-
drens Healthcare of Atlanta, Grady Me- California, San Diego: Medical Center ther G. Akpa, RN, BSN; Patty A. Cluff, RN;
morial Hospital, and Emory Crawford and Sharp Mary Birch Hospital for Anna E. Lis, RN, BSN; Georgia E. Mc-
Long Hospital (National Institutes of Women and Newborns (National Insti- David, RN; Claudia I. Franco, RNC, MSN;
Health grants UL1RR025008 and U10 tutes of Health grant U10 HD40461): Beverly Foley Harris, RN, BSN; Sarah
HD27851): David P. Carlton, MD; Ann M. Maynard R. Rasmussen, MD; Paul R. Martin, RN, BSN; Maegan C. Simmons,
Blackwelder, RNC, BS, MS; Michelle Tid- Wozniak, MD; Kathy Arnell, RNC; Clar- RN; Patti Pierce Tate, RCP; University of
well, BSN; Eunice Kennedy Shriver Na- ence Demetrio, RN; Chris Henderson, Utah: University Hospital, Latter Day
tional Institute of Child Health and Hu- RCP, CRTT; Wade Rich, BSHS, RRT; Uni- Saints Hospital, and Primary Chil-
man Development: Stephanie Wilson versity of Iowa: Childrens Hospital (Na- drens Medical Center (National Insti-
Archer, MA; Indiana University: Indiana tional Institutes of Health grants GCRC tutes of Health grants GCRC M01 RR64
University Hospital, Methodist Hospi- M01 RR59 and U10 HD53109): John A. and U10 HD53124): Bradley A. Yoder,
tal, Riley Hospital for Children, and Widness, MD; Karen J. Johnson, RN, MD; Karen A. Osborne, RN, BSN, CCRC;
Wishard Health Services (National In- BSN; University of Miami: Holtz Chil- Jennifer J. Jensen, RN, BSN; Cynthia
stitutes of Health grants GCRC M01 drens Hospital (National Institutes of Spencer, RNC; Kimberlee Weaver-
RR750 and U10 HD27856): James A. Health grants GCRC M01 RR16587 and Lewis, RN, BSN; Wake Forest University:
Lemons, MD; Diana D. Appel, RN, BSN; U10 HD21397): Ruth Everett-Thomas, Baptist Medical Center, Forsyth Medi-
Dianne E. Herron, RN; Lucy C. Miller, RN, RN, MSN; University of New Mexico: cal Center, and Brenner Childrens
BSN, CCRC; Leslie Richard, RN; Leslie D. Health Sciences Center (National Insti- Hospital (National Institutes of Health
Wilson, BSN, CCRC; Faithe Hamer, BS; tutes of Health grants GCRC M01 RR997 grants GCRC M01 RR7122 and U10
RTI International (National Institutes of and U10 HD53089): Conra Backstrom HD40498): Robert G. Dillard, MD; Nancy
Health grant U01 HD36790): W. Kenneth Lacy, RN; University of Rochester: Gol- J. Peters, RN, CCRP; Wayne State Uni-
Poole, PhD; Betty K. Hastings; Elizabeth isano Childrens Hospital at Strong versity: Hutzel Womens Hospital and
M. McClure, MEd; Jeanette ODonnell (National Institutes of Health grants
Childrens Hospital of Michigan (Na-
Auman, BS; Carolyn Petrie Huitema, GCRC M01 RR44 and U10 HD40521):
tional Institutes of Health grant U10
MS; Kristin M. Zaterka-Baxter, RN, BSN; Linda J. Reubens, RN, CCRC; Erica Bur-
HD21385): Rebecca Bara, RN, BSN; Ger-
Stanford University: Dominican Hospi- nell, RN; University of Tennessee (Na-
aldine Muran, RN, BSN; Women and In-
tal, El Camino Hospital, and Lucile tional Institutes of Health grant U10
fants Hospital of Rhode Island (Na-
Packard Childrens Hospital (National HD21415): Sheldon B. Korones, MD; Uni-
tional Institutes of Health grant U10
Institutes of Health grants GCRC M01 versity of Texas Southwestern Medical
HD27904): William Oh, MD; Angelita M.
RR70 and U10 HD27880): David K. Center at Dallas: Parkland Health and
Hensman, RN, BSN; Yale University:
Stevenson, MD; Marian M. Adams, MD; Hospital System and Childrens Medi-
Yale-New Haven Childrens Hospital
M. Bethany Ball, BS, CCRC; Melinda S. cal Center Dallas (National Institutes of
(National Institutes of Health grants
Proud, RCP; Andrew W. Palmquist, RN; Health grants GCRC M01 RR633 and
Tufts Medical Center: Floating Hospital U10 HD40689): Abbot R. Laptook, MD; CTSA UL1 RR24139, GCRC M01 RR6022,
for Children (National Institutes of Charles R. Rosenfeld, MD; Walid A. Sal- and U10 HD27871): Patricia Gettner,
Health grants GCRC M01 RR54 and U10 hab, MD; Gaynelle Hensley, RN; Melissa RN; Monica Konstantino, RN, BSN;
HD53119): Brenda L. MacKinnon, RNC; H. Leps, RN; Nancy A. Miller, RN; Univer- JoAnn Poulsen, RN; Janet Taft, RN, BSN.
Ellen Nylen, RN, BSN; University of Ala- sity of Texas Health Science Center at We are indebted to our medical and nurs-
bama at Birmingham: Health System Houston: Medical School, Childrens ing colleagues and the infants and their
and Childrens Hospital of Alabama Memorial Hermann Hospital, and Lyn- parents who agreed to take part in this
(National Institutes of Health grants don Baines Johnson General Hospital/ study. We thank Avroy Fanaroff for thought-
GCRC M01 RR32 and U10 HD34216): Harris County Hospital District (Na- ful review of the manuscript and Mazie Tin-
Monica V. Collins, RN, BSN, MaEd; tional Institutes of Health grant U10 sley for manuscript preparation.
REFERENCES
1. Hack M, Horbar JD, Malloy MH, Tyson JE, 2. Hack M, Wright LL, Shankaran S, et al. Very- 3. Fanaroff AA, Wright LL, Stevenson DK, et al.
Wright E, Wright L. Very low birth weight low-birth-weight outcomes of the National Very-low-birth-weight outcomes of the Na-
outcomes of the National Institute of Child Institute of Child Health and Human Devel- tional Institute of Child Health and Human
Health and Human Development Neonatal opment Neonatal Network, November 1989 Development Neonatal Research Network,
Network. Pediatrics. 1991;87(5): to October 1990. Am J Obstet Gynecol. 1995; May 1991December 1992. Am J Obstet Gy-
587597 172(2):457 464 necol. 1995;173(5):14231431

454 STOLL et al
 ARTICLE

4. Stevenson DK, Wright LL, Lemons JA, et al. 11. Stoll BJ, Hansen NI, Higgins RD, et al. Very low search Network. Intensive care for extreme
Very low birth weight outcomes of the Na- birth weight preterm infants with early onset prematurity: moving beyond gestational
tional Institute of Child Health and Human neonatal sepsis: the predominance of Gram- age. N Engl J Med. 2008;358(16):16721681
Development Neonatal Research Network, negative infections continues in the National 19. Meadow W, Lee G, Lin K, Lantos J. Changes in
January 1993 through December 1994. Am J Institute of Child Health and Human Develop- mortality for extremely low birth weight in-
Obstet Gynecol. 1998;179(6):16321639 ment Neonatal Research Network, 20022003. fants 1990s: implications for treatment de-
Pediatr Infect Dis J. 2005;24(7):635 639
5. Lemons JA, Bauer CR, Oh W, et al. Very low cisions and resource use. Pediatrics. 2004;
12. Jobe AH, Bancalari E. Bronchopulmonary 113(5):12231229
birth weight outcomes of the National Insti-
dysplasia. Am J Respir Crit Care Med. 2001;
tute of Child Health and Human Develop- 20. EXPRESS Group, Fellman V, Hellstrm-Westas
163(7):17231729
ment Neonatal Research Network. January L, Norman M, et al. One-year survival of ex-
1995 through December 1996. Pediatrics. 13. Walsh MC, Yao Q, Gettner P, et al. Impact of a tremely preterm infants after active perina-
2001;107(1): Available at: http://www. physiologic denition on bronchopulmo-
tal care in Sweden. JAMA. 2009;301(21):
pediatrics.org/cgi/content/full/107/1/e1 nary dysplasia rates. Pediatrics. 2004;
22252233
114(5):13051311
6. Fanaroff AA, Stoll BJ, Wright LL, et al. Trends 21. Vohr BR, Wright LL, Dusick AM, et al. Center
14. Zou G. A modied Poisson regression ap-
in neonatal morbidity and mortality for very differences and outcomes of extremely low
proach to prospective studies with binary
low birthweight infants. Am J Obstet Gy- birth weight infants. Pediatrics. 2004;
data. Am J Epidemiol. 2004;159(7):702706
necol. 2007;196(2):147.e1147.e8 113(4):781789
15. Costeloe K, Hennessy E, Gibson AT, Marlow
7. Muglia LJ, Katz M. The enigma of spontane- 22. Lee SK, Chan K, Ohlsson A, et al. Variations in
N, Wilkinson AR; EPICure Study Group. The
ous preterm birth. N Engl J Med. 2010; treatment decisions and outcomes of in-
EPICure Study: outcomes to discharge
362(6):529 535 from hospital for infants born at the fants 25 weeks gestation: evidence from
8. Iams JD, Romero R, Culhane JF, Goldenberg threshold of viability. Pediatrics. 2000; Canada. Pediatr Res. 1999;45(4):247A
RL. Preterm birth: primary, secondary, and 106(4):659 671 23. Synnes AR, Chien L-Y, Peliowski A, Baboolal
tertiary interventions to reduce the morbid- 16. Horbar JD, Badger GJ, Carpenter JH, et al. R, Lee SK; Canadian NICU Network. Varia-
ity and mortality of preterm birth. Lancet. Trends in mortality and morbidity for very tions in intraventricular hemorrhage inci-
2008;371(9607):164 175 low birth weight infants, 19911999. Pediat- dence rates among Canadian neonatal ICUs.
9. Alexander GR, Himes JH, Kaufman RB, Mor J, rics. 2002;110(1):143151 J Pediatr. 2001;138(4):525531
Kogan M. A United States national reference 17. Chan K, Ohlsson A, Synnes A, et al. Survival, 24. Walsh M, Laptook A, Kazzi SN, et al. A cluster-
for fetal growth. Obstet Gynecol. 1996;87(2): morbidity, and resource use of infants of 25 randomized trial of benchmarking and mul-
163168 weeks gestational age or less. Am J Obstet timodal quality improvement to improve
10. Stoll BJ, Hansen N, Fanaroff AA, et al. Late- Gynecol. 2001;185(1):220 226 rates of survival free of bronchopulmonary
onset sepsis in VLBW neonates: the expe- 18. Tyson JE, Parikh NA, Langer J, Green C, Hig- dysplasia for infants with birth weights of
rience of the NICHD Neonatal Research gins RD; National Institute of Child Health less than 1250 grams. Pediatrics. 2007;
Network. Pediatrics. 2002;110(2):285291 and Human Development Neonatal Re- 119(5):876 890

PEDIATRICS Volume 126, Number 3, September 2010 455

(Continued from rst page)
KEY WORDS
extremely low gestation, very low birth weight, morbidity, death
ABBREVIATIONS
VLBWvery low birth weight
BPDbronchopulmonary dysplasia
BWbirth weight
CIcondence interval
GAgestational age
IVHintraventricular hemorrhage
ROPretinopathy of prematurity
RRrelative risk
NICHDNational Institute of Child Health and Human Development
NRNNeonatal Research Network
CPAPcontinuous positive airway pressure
PVLperiventricular leukomalacia
PMApostmenstrual age
www.pediatrics.org/cgi/doi/10.1542/peds.2009-2959
doi:10.1542/peds.2009-2959
Accepted for publication May 13, 2010
Address correspondence to Barbara J. Stoll, MD, Department of Pediatrics, Emory University School of Medicine and Childrens Healthcare of Atlanta, 2015
Uppergate Dr, Atlanta, GA 30322. E-mail: barbara_stoll@oz.ped.emory.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 nancial relationships relevant to this article to disclose.
Funded by the National Institutes of Health (NIH).

456 STOLL et al