Center Differences and Outcomes of Extremely Low Birth Weight Infants
Objective. Previous multicenter studies have shown significant center differences in neonatal characteristics and morbidities. This study evaluated center differences in outcome at 18 to 22 months among extremely low birth weight (ELBW; 401-1000 g) infants after adjusting for demographics and antenatal interventions, and it identified neonatal interventions associated with outcome differences.
Methods. We assessed the outcome of 2478 liveborn infants who were admitted in 1993 and 1994 to the 12 centers of the Neonatal Research Network of the National Institute of Child Health and Human Development; 1483 (60%) infants survived to 18 to 22 months, and 1151 (78%) had comprehensive evaluations. Logistic regression analyses were performed to identify center differences and the association of 4 neonatal interventions—active resuscitation, postnatal steroids, ventilator treatment for ≤27 days, and full enteral feedings ≤24 days—with adverse outcomes (cerebral palsy, low Bayley scores, and neurodevelopmental impairment [NDI]), after adjusting for demographics and antenatal interventions.
Results. Using bivariate analyses, significant center differences were identified for mortality, antenatal and postnatal interventions, social and environmental variables, neonatal morbidities, and neurodevelopmental outcomes for the 12 centers. After adjustment for maternal and infant demographics and antenatal interventions, the percentage of ELBW infants who had died or had NDI at 18 to 22 months ranged from 52% to 85%. Active resuscitation and postnatal steroids were associated with increases of NDI of 11.8% and 19.3%, whereas shorter ventilation support and shorter time to achieve full enteral feeds were associated with decreases in NDI of 20.7% and 17.3%, respectively.
Conclusion. There are large and disturbing differences among centers in outcomes at 18 to 22 months after adjusting for demographic and antenatal interventions. Center differences in postnatal interventions associated with differences in outcome can provide hypotheses for testing in clinical trials to improve outcome.
- extremely low birth weight
- center differences
- neurologic outcome
- developmental outcome
- cerebral palsy
- follow-up studies
Single-center studies of extremely low birth weight (ELBW) infants have reported diverse neonatal and follow-up data, and multicenter studies1–4 have reported significant differences in neonatal characteristics, morbidities, and care practices. The National Institute of Child Health and Development (NICHD) Neonatal Research Network conducts longitudinal follow-up of ELBW (401 to <1000 g) survivors. This report expands on the body of information available on the heterogeneity of center outcomes and investigates the relationships among center differences, neonatal interventions, and 18- to 22-month outcomes of ELBW infants who were born in 1993 and 1994 at 12 network centers.5 The objectives of this study were 1) to assess the neurodevelopmental outcome differences among centers after adjusting for demographics and antenatal care and 2) to assess the association between specific neonatal interventions and neurodevelopmental outcome. We hypothesized that centers would differ in adjusted neurodevelopmental outcomes and mortality at 18 to 22 months and that center differences in the use of neonatal interventions would be associated with outcome.
Neonatal and survival data were collected on all infants who were liveborn in the center and all infants who were born outside the center and admitted to center NICUs. Families of survivors were invited to participate in their center follow-up programs for a comprehensive assessment. Follow-up data were collected on all long-term survivors who were evaluated in follow-up. The follow-up assessment consisted of a battery of developmental and neurologic assessments, a medical and social history, and parent interviews.5 The Bayley Scales of Infant Development II6 mental and motor tests with derivation of a mental developmental index (MDI) and psychomotor developmental index (PDI) were administered by a certified examiner who was trained to reliability by 1 of 4 psychologists with high interexaminer reliability on the Bayley and previous formal training in test administration. Certification for administration of the Bayley was achieved by the successful completion of 2 videotaped demonstrations of accurate performance and scoring of the Bayley as determined by 1 of the 4 Gold Standard psychologists. Certification is renewed annually for all testers.
The neurologic examinations were based on the Amiel Tison neurologic assessment.7 Cerebral palsy (CP) was defined as a nonprogressive central nervous system disorder characterized by abnormal muscle tone in at least 1 extremity and abnormal control of movement or posture. Neurologic examinations were performed by experienced certified examiners who had been trained to reliability in the examination procedure in a 2-day “hands on” workshop. Structured interviews were performed to obtain social and economic status information8 and a detailed interim medical history, including data on hearing status and vision status. All definitions, protocols, and procedures are contained in a manual of operations.5
The sample consisted of 2478 ELBW infants who were liveborn between January 1, 1993, and December 29, 1994, and cared for in 1 of 12 participating centers of the NICHD Neonatal Research Network. There were 1530 (61.7%) survivors at discharge, and there were 47 deaths after discharge. Of the 1483 long-term survivors, 1151 (78%) were evaluated at 18 to 22 months’ corrected age, and 332 were lost to follow-up.
Survival characteristics, maternal demographics, neonatal characteristics, morbidities, and 18- to 22-month outcomes were compared across centers in Tables 1 to 4⇓⇓⇓⇓. Center differences in continuous measures were examined using the Kruskal-Wallis test (>2 groups). Differences on categorical measures were analyzed using the χ2 test, or Fisher exact test for cases in which cell size was too small for the χ2 test. P < .05 was considered statistically significant. Because most significant P values were <.0001, differences would still be significant after correction for multiple tests.
Logistic regression models were then run to answer 2 questions. First, what are the center differences in outcome after adjusting for maternal and neonatal demographic characteristics known at the time of birth and antenatal interventions known to have an impact on outcome? The outcomes of interest were MDI <70, PDI <70, CP, neurodevelopmental impairment (NDI), and NDI or death by 18 months. The regression for NDI or death was added because death is a competing outcome variable for major morbidity. NDI was defined as the presence of any of the following: CP, deaf/hard of hearing requiring amplification in both ears, bilaterally blind, MDI <70, or PDI <70. The center with the best outcome for NDI was the reference center for the models (Table 5). The maternal and neonatal demographic characteristics included in the model were birth weight (unit = 100 g), small for gestational age (SGA), multiple birth, congenital anomaly, gender, maternal age (≤19 vs >19 years), education (mother not a high school graduate vs a high school graduate), single versus married, income (≤$20 000 vs >20 000), health insurance, and race (black or Hispanic vs white/other). The antenatal interventions were prenatal care, outborn, antenatal steroids, and cesarean section. In the models for NDI and/or death, congenital malformation, maternal education, income, and insurance were excluded because this information was not available for infants who died. Center differences that persist after adjusting for known demographic and antenatal interventions are more likely to be associated with postnatal interventions.
From the 5 models shown in Table 5, we calculated the individual center expected rates of outcomes adjusted for maternal and neonatal demographics and antenatal interventions and compared them with the observed rates of outcomes (Table 6). Adjusted rates were calculated by substituting the average for the covariates in the logistic regression equation. For example, 7.9606 was used for birth weight (expressed in 100-g increments) because 796.06 g was the average birth weight for the infants who were followed up. For race, 0.5126 was used for black and 0.1234 was used for Hispanic because 51.26% were black and 12.34% were Hispanic. For the center of interest (center 1), 1 was used for the center 1 rate and 0 was used for the other center rates. The logistic regression equation was then evaluated
We then asked a second question. To what degree are center neonatal interventions associated with outcomes? For evaluating the effects of clinical neonatal interventions on the 5 study outcomes, 4 interventions were modeled separately adjusting for maternal and neonatal demographic, antenatal interventions, and center. The 4 practices were active resuscitation defined as chest compressions and medications, postnatal steroids (PNS), ventilator for ≤26 days versus >26 days on the ventilator, and ≤23 days versus >23 days to full enteral feeds. Twenty-six days on a ventilator and 23 days to full feeds were median values. Only active resuscitation could be modeled for NDI or death because PNS, days on ventilator, and days to full feeds were not applicable for infants who died shortly after birth. In addition, maternal education less than high school, income <$20 000, and Medicaid insurance were not included because the information was not collected for the infants who died. Indicator variables were created for each of the 4 interventions. For example, if the infant was on the ventilator ≤26 days, then the indicator variable was set to 1; if the infant was on the ventilator >26 days, then the indicator variable was set to 0.
As before, adjusted rates were calculated by substituting the average for the covariates in the logistic regression equation. For the predictor of interest (PNS), 1 was used for the “yes” probability and 0 was used for the “no” probability. All data were analyzed at the Research Triangle Institute (Research Triangle Park, NC).
Table 1 shows the median and range data on outborn status, deaths in the first 12 hours of life, survival to discharge, postdischarge deaths, and infants lost to follow-up. Significant center differences were observed for all variables except postdischarge death. A comparison of demographic characteristics (maternal age, marital status, race, and outborn status) and 10 biological risk factors (including intraventricular hemorrhage [IVH], periventricular leukomalacia [PVL] defined as the presence of cystic lesions, and chronic lung disease [CLD] defined as oxygen at 36 weeks’ gestation) between those who were followed up and those who were lost to follow-up was not significant.
Table 2 shows the maternal characteristics of ELBW infants who were evaluated at 18 to 22 months of age. Significant center variability was identified for maternal age, maternal education, income <$20 000, marital status, Medicaid insurance, race, primary language spoken in home, prenatal care, outborn, antenatal steroids, and cesarean section.
Neonatal characteristics of the infants seen at 18 to 22 months are shown by center in Table 3. The mean and standard deviation are given for the continuous measures with the exception of days in the hospital, for which median and range are reported. For categorical measures, the number is given with the percentage in parentheses. Mean birth weight ranged from 765 to 830 g, and mean gestational age ranged from 26 to 27 weeks. Although SGA9 ranged from 12% to 33% and male gender ranged from 41% to 57%, neither difference was significant. The percentage of infants who were multiples ranged from 8% to 28%. Multiples, congenital malformations, grade 3 to 4 IVH/PVL, threshold retinopathy of prematurity, sepsis, jaundice, surfactant, CLD, number of days to regain birth weight, and the 4 “study” interventions differed among centers. Active resuscitation rates (use of either chest compressions or drugs) ranged from 5% to 28%. Overall at birth, 87.7% of infants were intubated and 75.2% were ventilated by hand (data not shown). PNS administration ranged from 27% to 65%. Mean days on a ventilator ranged from 16 to 40, and mean days to full enteral feeds ranged from 19 to 33 days. Median number of days of hospitalization ranged from 86 to 108.
Table 4 depicts the results for the children who were evaluated at 18 to 22 months. The percentage of children who were evaluated with a normal neurologic examination varied from 63% to 93%. CP ranged from 6% to 30%. There was a 13-point difference for the MDI and an 18-point difference for the PDI between the center with the lowest and highest mean score. Significant differences among centers were also identified in hearing impairment. The percentage of deaths occurring before 18 months is shown as a percentage of infants who died or were evaluated at 18 months and ranged from 34% to 54%. The presence of any NDI ranged from 25% to 69%, and the presence of any NDI or death ranged from 64% to 84%. Participation in early intervention at the time of follow-up, as reported by the caregiver, ranged from 22% to 80% (data not shown).
Table 5 shows the odds ratios and 95% confidence intervals for the centers and demographics and antenatal interventions from the logistic regression models to predict CP, MDI <70, PDI <70, NDI, and NDI or death. Center 2, which had the best adjusted outcome for NDI, is the reference center. In the model for CP, male gender was associated with increased risk and higher birth weight with decreased risk. Antenatal interventions that had protective effects were prenatal care and antenatal steroids. Six centers had significantly higher odds compared with the reference center.
In the model to predict a PDI <70, demographic factors associated with adverse effects were Medicaid insurance, male gender, and multiple birth. Higher birth weight was protective. One intervention, antenatal steroids, was positively associated with higher PDI. All centers had significantly higher odds ratios compared with the reference center.
In the model to predict an MDI <70, low maternal education, Medicaid insurance, male gender, and multiple birth all were associated with adverse outcome. Higher birth weight and SGA were associated with decreased risk. The single antenatal intervention that was positively associated with a higher MDI was prenatal care. Nine centers had significantly higher odds ratios compared with the reference center.
Demographics significantly associated with adverse effects for NDI were low maternal education, Medicaid insurance, male gender, and multiple birth. Higher birth weight, SGA, and prenatal care were protective. Eight centers differed from the reference center.
The final model predicted death or impairment. Maternal education, Medicaid insurance, income, and congenital malformation were not available for infants who died, and these variables were not included in this model. As in the previous models, male gender and multiple birth remained risk factors, and higher birth weight and SGA were protective. Hispanic ethnicity was also associated with increased risk. Prenatal care and antenatal steroids both were associated with decreased risk. Nine centers differed from the reference center. In all 5 models, there were significant center effects compared with the reference center after controlling for demographics and center antenatal interventions.
On the basis of the models shown in Table 5, observed rates of adverse outcomes compared with the expected rates of adverse outcomes after adjustment for demographics and antenatal interventions for each center were calculated and are shown in Table 6. Adjusted rates of CP ranged from 4.5% to 23.1%, MDI <70 ranged from 14.5% to 62.6%, and PDI <70 ranged from 7.2% to 44.4%. The adjusted rate of NDI varied considerably and ranged from 23.6% to 71%. The adjusted rate of NDI or death ranged from 52.0% to 84.7%. Although center 2 has the lowest observed and expected rates of CP, MDI <70, PDI <70, and NDI, center 11 had a virtually identical adjusted rate of NDI or death.
Table 7 shows the associations of 4 neonatal interventions on outcome after adjusting for center, maternal and neonatal demographics, and antenatal interventions. Separate logistic regression models were run for each outcome/intervention combination. The model coefficients, the P values, and the odds ratios with confidence intervals are shown for each model. The adjusted rates of an adverse outcome are shown relative to the presence or absence of the intervention. These are calculated from the logistic regression model using average values for maternal and neonatal demographics and antenatal interventions as described earlier. Active neonatal resuscitation was associated with an increase in the rate of CP by 10%, MDI <70 by 10%, PDI <70 by 12%, NDI by 12%, and NDI or death by 11%. For example, PNS were associated with an increase in the rates of CP by 7%, MDI <70 by 14%, PDI <70 by 20%, and NDI by 19%. Fewer than 26 days of ventilation was associated with a decrease in the rate of CP by 9%, MDI <70 by 17%, PDI <70 by 19%, and NDI by 21%. Fewer days to achieve full enteral feed was associated with a decrease in the rate of an MDI <70 by 13%, PDI <70 by 9%, and NDI by 17%.
Striking differences remained in center outcomes after adjusting for demographics and antenatal interventions. These findings suggest that differences in center neonatal practices strongly influence outcome. Although other multicenter studies have reported significant differences in neonatal outcomes,1–4,9–12 this is the first study to report on the relative impact of center neonatal interventions on 18- to 22-month outcomes of ELBW infants after adjusting for maternal and neonatal demographics and antenatal interventions.
Using bivariate analyses, significant differences were identified among the 12 centers for almost all social and demographic variables, antenatal interventions, postnatal interventions, and 18- to 22-month outcomes, consistent with other multicenter neonatal studies.1–4,9–12 The 2 socioeconomic markers in our models that are associated with poverty (less than a high school education and Medicaid health insurance) both were predictors of low Bayley scores and major impairments in our cohort. Although black race ranged from a low of 3% to a high of 82%, white race 6% to 76%, and Hispanic ethnicity 0% to 61%, the only finding of race/ethnicity was an association between Hispanic ethnicity and increased NDI or death. The limited findings attributed to race are consistent with those of others, suggesting that poverty rather than race is an important predictor of adverse outcome.13,14 An important demographic factor in our models, multiple births, was associated with increased rates of low PDI, low MDI, NDI, and death or NDI. This variable deserves additional investigation because of the steadily increasing rates of multiples in NICUs associated with assisted reproductive technology.15–17 The rates of ELBW multiples at the centers ranged from 8% to 28%. The network database during this time period did not collect information on assisted reproductive technology. Neonatal morbidities that are known to contribute to outcome, which varied among the centers, were IVH, PVL, sepsis, and jaundice. These variables were not entered into the models because our primary goal was to identify the association of neonatal interventions with outcome. In addition, we identified differences among centers in antenatal interventions, which we controlled for in our multivariate analyses. It is impossible to measure or adjust for all potential confounders, and our findings are undoubtedly affected to some degree by residual confounding.
Variations were identified in center practices of all 4 postnatal interventions. Variations in active resuscitation (5%–28%) and in survival rates (51%–72%) reflect in part differences in management style. The death rate at <12 hours (that may be an indicator for initial aggressiveness of care) ranged from 11% to 27%. Center 2, with the highest death rate at <12 hours, had the lowest percentage of survivors with NDI at 18 to 22 months. The relationships among center death rates, active resuscitation, and outcome, however, were variable. Center 11 had the lowest death rate at <12 hours, even among infants <750 g, and the second lowest NDI rate (38%). In contrast, center 3 had the lowest rate of active resuscitation (5%), the highest death rate ≤120 days, and the highest NDI rate. The associations of active resuscitation in our regression models with increased rates of CP, low PDI, NDI, and death or NDI may reflect that the most distressed infants require more aggressive resuscitation. Additional evaluation of candidacy of very low birth weight infants for active resuscitation is indicated.
The study of Lee et al11 evaluated outcomes of infants of 22 to 25 weeks’ gestation in 12 hospitals in the Canadian NICU Network. They found that delivery room deaths ranged from 8.5% to 57.4% and neonatal survival ranged from 31.8% to 78.6%. They did not find that variation in delivery room deaths had an impact on neonatal morbidities. Lorenz et al,18 however, specifically evaluated the effects of aggressive perinatal management of very low birth weight infants at 23 to 26 weeks’ gestation by comparing a cohort in the United States with a cohort in the Netherlands. They found that infants in the US cohort were more commonly delivered by cesarean section (30.1% vs 5.6%) and treated with mechanical ventilation (95.4% vs 64.7%). Although survival at 2 years of age was significantly greater in the US cohort compared with the Dutch cohort (47.3% vs 22.9%), the prevalence of disabling CP among survivors was 5 times higher in the US cohort. These data demonstrate the impact of antepartum and neonatal care on survival and long-term morbidity.
PNS administration for CLD or for weaning from ventilators was associated with poor outcomes. PNS use increased the rate of CP by 7%, MDI <70 by 14%, PDI <70 by 20%, and NDI by 19%. These findings are concerning and support the growing evidence that aggressive use of PNS is deleterious to ELBW infants. Although we do not have data on duration of PNS administration, our findings are consistent with other recent human and animal studies reporting negative effects of PNS on neurodevelopmental outcome.19–26
Days of assisted ventilation ranged from a mean of 16 to 40 days (median: 26 days). Fewer than 26 days of ventilation was associated with lower rates of CP, low MDI, low PDI, NDI, and death or NDI. Infants with bronchopulmonary dysplasia, a known risk factor for poor outcome, require prolonged ventilation. Our retrospective data cannot differentiate between severity of pulmonary disease requiring ventilation and variations in ventilation management. Prospective evaluation of ventilation protocols is needed.
Although the management goal in the NICU is to provide sufficient nutrients to support rates of growth and nutrient acceleration equal to intrauterine rates, ELBW infants often do not tolerate enteral feeds for 2 to 3 weeks after birth. Prolonged administration of parenteral feeds has been shown to be associated with delays in weight gain and the onset of postnatal growth restriction (NICHD).27 Our data suggest that achieving full enteral nutrition earlier (≤23 days) is associated with better neurodevelopmental outcome.
Significant variability was demonstrated in each of the center-adjusted neurodevelopmental outcomes. Of particular interest was the variation in CP rate (range: 4.5%–23.1%), MDI scores <70 (range: 14.5%–62.6%), and PDI scores <70 (range: 7.2%–44.4%). It could be speculated that some of these differences are secondary to variable assessment techniques. However, the NICHD Neonatal Research Network follow-up study has standardized definitions and protocols to establish inter-rater reliability for the neurologic and Bayley assessments. Therefore, we are confident that differences observed are valid.
Variability in compliance for follow-up has been reported to affect reports of disability rates.28 In the current study, identified postdischarge deaths at centers were low, ranging from none to 6%, and did not differ significantly by center. In addition, although follow-up rates in this study ranged from 56% to 94%, a comparison of neonatal biological and demographic characteristics identified no differences between infants who were assessed and those who were lost to follow-up. We cannot rule out a bias, however, within the population of infants who were evaluated because additional demographic factors (maternal education and insurance status) were not available for the infants who were lost to follow-up, and the actual morbidity rate of the “lost subjects” at 18 months remains unknown.
In summary, significant differences in center demographics, antenatal interventions, neonatal interventions, and outcomes were identified. The rate of NDI after adjusting for demographics and antenatal interventions ranged from 24% to 71%, and the rate of NDI or death ranged from 52% to 85%. Our evaluation of the independent effects of 4 neonatal interventions demonstrated that aggressive resuscitation and PNS were associated with increases of NDI of 11.8% and 19.3%, and shorter ventilation support and shorter time to achieve full enteral feeds were associated with decreases in NDI of 20.7% and 17.3%, respectively. Both center demographics and illness severity, and center differences in clinical practice influence outcome. Identification of significant adjusted differences in center NDI rates and center use of neonatal interventions that affect NDI support the need for hypothesis-driven clinical trials to assess the efficacy of current neonatal interventions.
This study was supported by the National Institute of Child Health and Human Development through Cooperative Agreements: U10 HD27904, U10 HD27856, U01 HD36790, U10 HD21385, U10 HD21415, U10 HD21373, U10 HD27853, U10 HD21397, U10 HD27880, U10 HD27881, U10 HD21364, U10 HD27851, U10 HD27871; and CRC grants M01 RR 00750, M01 RR 08084, M01 RR 00070, M01 RR 00997, M01 RR 06022
Presented in part at the annual meeting of the Society for Pediatric Research; May 1–5, 1999; San Francisco, CA.
NICHD Neonatal Research Network: Betty R. Vohr, MD, Follow-Up Study Chairman; Alan Jobe, MD, Neonatal Network Chair; Case Western Reserve University (U10 HD21364), Maureen Hack, MD1, Avroy A. Fanaroff, MB, BCh, Harriet Friedman, MA2; University of Cincinnati (U10 HD27853), Jean J. Steichen, MD1, Edward F. Donovan, MD, Tari Gratton, RN2; Emory University (U10 HD27851), Neal P. Simon, MD1, Barbara J. Stoll, MD, Ellen Hale, RN; Indiana University (U10 HD27856), Anna M. Dusick, MD1, James A. Lemons, MD, Julie Ahlrichs, RN; University of Miami (U10 HD21397), Charles R. Bauer, MD1, Wendy Griffin, RN; National Institute of Child Health and Human Development, Linda L. Wright, MD1, Sumner J. Yaffe, MD, Elizabeth M. McClure, MEd; University of New Mexico (U10 HD27881), Lu-Ann Papile, MD1, Jean Lowe, PhD2, Barbara Woodward, MPH, OTR/L; Stanford University (U10 HD27880), Barry E. Fleisher, MD1, David K. Stevenson, MD, M. Bethany Ball; University of Tennessee at Memphis (U10 HD21415), Sheldon B. Korones, MD, Henrietta Bada, MD; University of Texas Southwestern Medical Center at Dallas (U10 HD21373), Sue Broyles, MD1, Jon E. Tyson, MD, Jackie Hickman, RN; Wayne State University (U10 HD21385), Virginia Delaney-Black, MD1, Seetha Shankaran, MD, Debra Driscoll, RN; Women and Infants’ Hospital (U10 HD27904), Betty R. Vohr, MD1, William Oh, MD, Terri Leach, MEd2; Yale University (U10 HD27871), Richard A. Ehrenkranz, MD, Linda C. Mayes, MD1, Elaine Sherwonit, MSN, CPNP.1Follow-up Study Principal Investigator;2Gold Standard Certified Bayley Examiner.
- Received February 22, 2002.
- Accepted October 22, 2003.
- Reprint requests to (B.R.V.) Women and Infants Hospital, 101 Dudley St, Providence, RI 02905. E-mail:
Dr Tyson’s current affiliation is the University of Texas Medical School at Houston, Houston, Texas.
- ↵Hack M, Horbar JD, Malloy MH, Tyson JE, Wright E, Wright L. Very low birth weight outcomes of the National Institute of Child Health and Human Development Neonatal Network. Pediatrics.1991;87 :587– 597
- Avery ME, Tooley WH, Keller JB, et al. Is chronic lung disease in low birth weight infants preventable? A survey of eight centers. Pediatrics.1987;79 :26– 30
- Horbar JD, Rogowski J, Plsek PE, et al, for the NIC/Q Project Investigators of the Vermont Oxford Network. Collaborative quality improvement for neonatal intensive care. Pediatrics.2001;107 :14– 22
- ↵Vohr BR, Wright LL, Dusick AD, et al. Neurodevelopmental and functional outcome of extremely low birth weight (ELBW) infants in the NICHD Neonatal Network 1993–1994. Pediatrics.2000;105 :1216– 1226
- ↵The Psychological Corporation. Bayley Scales of Infant Development. 2nd ed. San Antonio, TX: Harcourt Brace & Company; 1993
- ↵Amiel-Tison C. Neuromotor status. In: Taeusch HW, Yogman MW, eds. Follow-up Management of the High-Risk Infant. Boston, MA: Little, Brown & Company; 1987:115–126
- ↵Hollingshead AB. Four Factor Index of Social Status. New Haven, CT: Yale University Press; 1975
- ↵Richardson D, Tarnow-Mordi WO, Lee SK. Risk adjustment for quality improvement. Pediatrics.1999;103(suppl E) :255– 265
- ↵Lee SK, Chan K, Ohlson A, Synnes AR, Peliowski A, Koravangater S. Canadian NICU Network. Variations in treatment decisions and outcomes of infants ≤ 25 weeks gestation: evidence from Canada. Pediatr Res.1999;45(suppl) :247A
- ↵Lorenz JM, Paneth N, Jetton JR, den Duden L, Tyson JE. Aggressive vs expectant care of the extremely premature newborn: outcome and resource utilization. Pediatr Res.1999;45(suppl) :249A
- ↵Thebaud B, Lacaze-Masmonteil T, Watterberg K. Postnatal glucocorticoids in very preterm infants: “the good, the bad, and the ugly”? Pediatrics.2001;107 :413– 415
- Yeh TF, Lin YJ, Hsieh WS, et al. Early postnatal dexamethasone therapy for the prevention of chronic lung disease in preterm infants and respiratory distress syndrome: a multicenter clinical trial. Pediatrics.1997;100 :3
- Yeh TF, Lin YJ, Huang CC, et al. Early dexamethasone therapy in preterm infants: a follow-up study. Pediatrics.1998;101(5) . Available at: pediatrics.org/cgi/content/full/101/5/e7
- Halliday HL, Ehrenkranz RA. Early postnatal (<96 hours) corticosteroids for preventing chronic lung disease in preterm infants. Cochrane Database Syst Rev.2001;(1) : CD001146
- Halliday HL, Ehrenkranz RA. Moderately early (7–14 days) postnatal corticosteroids for preventing chronic lung disease in preterm infants. Cochrane Database Syst Rev.2001;(1) : CD001144
- ↵Halliday HL, Ehrenkranz RA. Delayed (>3 weeks) postnatal corticosteroids for chronic lung disease in preterm infants. Cochrane Database Syst Rev.2001;(1) : CD001145
- ↵Ehrenkranz RA, Vohr BR, Dusick AM, Wright LL, Wrage L, Poole K, and the NICHD Neonatal Research Network. Longitudinal growth of hospitalized ELBW infants and outcome in early childhood. Pediatr Res.2000;47 :307A
- ↵Tin W, Fritz S, Wariyar U, Hey E. Outcome of very preterm birth: children reviewed with ease at 2 years differ from those followed up with difficulty. Arch Dis Child Fetal Neonatal Ed.1998;79 :F83– F87
- Copyright © 2004 by the American Academy of Pediatrics