Survival of extremely preterm infants has improved through the combined efforts of obstetricians and neonatologists who share a common goal of reducing perinatal mortality.1 The decision whether to begin intensive care for newborns at borderline viability remains controversial, and survival rates for these children vary greatly.2–5 The American Congress of Obstetricians and Gynecologists guidelines state that the decision to begin or withhold support for the newborn at the threshold of viability should be individualized to the newborn's condition at birth, survival and morbidity data, and the parents' preferences.6 Outcome data are specific to each neonatal intensive care unit (NICU) and may be unknown for the individual center, making prenatal counseling difficult.
Women at risk of delivering prematurely at gestational ages more than 25 weeks commonly receive therapies that are known to improve survival of the infant. Although some studies have suggested that a similar approach increases survival in the extremely preterm infant,7,8 interventions are offered less consistently to this population. This inconsistent approach may contribute to the varying survival rates seen among NICUs.
The National Institute of Child Health and Human Development Neonatal Research Network studied a cohort of extremely preterm infants who were admitted to NICUs from 1998 to 2003, and identified variables that were predictive of survival and neurodevelopmental outcome.9 The Neonatal Research Network subsequently established a Web-based calculator that incorporated these variables to predict outcomes. Our study was conducted to estimate whether inborn infants delivered between 22 and 25 weeks of gestation and admitted to the NICU between 2000 and 2009 had observed survival-to-discharge rates that exceeded the predicted rates established during an earlier era and whether survival improved during the decade.
MATERIALS AND METHODS
Study eligibility criteria included all inborn infants born at 22–25 weeks of gestation and admitted to the University of Iowa NICU from January 1, 2000, to December 31, 2009. Similar to the Neonatal Research Network analysis, we included only inborn infants admitted to the NICU who were born without major anomalies, and we excluded those weighing less than 401 g or more than 1,000 g.
Neonatal and maternal characteristics were abstracted from the University of Iowa NICU registry and the medical records. Gestational age was determined by standard clinical guidelines using last menstrual period, prenatal ultrasonography, or physical examination (in decreasing order of consideration). Gestational age was recorded as completed weeks. All included infants were resuscitated and received mechanical ventilation. We excluded from analysis seven newborns who died in the delivery room after failed resuscitation attempts. There were five newborns born at 23 weeks of gestation who were not successfully resuscitated, one each in 2001, 2002, 2005, 2006, and 2008. One newborn born at 24 weeks of gestation was not successfully resuscitated in 2005. One newborn born at 25 weeks of gestation died in the delivery room and had Escherichia coli sepsis.
Observed infant survival-to-discharge rates at each gestational age were compared with predicted survival rates based on gestational age, birth weight, sex, singleton or multiple gestation, and antenatal corticosteroid administration estimation by using the Neonatal Research Network's Web-based outcome calculator (http://www.nichd.nih.gov/about/org/cdbpm/pp/prog_epbo/epbo_case.cfm). Estimated infant survival for each gestational age was described with a box plot with the mean of the sample, median with interquartile range, and whiskers showing ±1.5 times the interquartile range. We calculated the observed percent survival of 22-week-old to 25-week-old infants by birth year and performed an unadjusted linear regression analysis of survival rates by year.
Multivariable logistic regression analysis was used to estimate the association between the observed infant survival to discharge and antenatal variables that have been found previously to be predictive for survival. These results are presented as odds ratios (OR) and 95% confidence intervals (CI). We used available Bayley Scales of Infant Development scores at the ages of 18–22 months (ie, Bayley-II for 2000–2004 and Bayley-III for 2005–2009) in the definition of neurodevelopmental impairment. Neurodevelopment impairment was defined as one or more of the following: Bayley-II mental developmental index less than 70, Bayley-II psychomotor developmental index less than 70, Bayley-III cognitive composite score less than 80, Bayley-III motor score less than 70, cerebral palsy, deafness, or blindness. The Bayley-III composite score of 80 is the mean−2 SD in term controls.10 We used the Mantel-Haenszel χ2 for linear trend test to analyze the relationship between rates of neurodevelopmental impairment and gestational age. Analyses were performed using SAS 9.2. The study was approved by the institutional review board of the University of Iowa.
We identified a total of 326 extremely preterm infants who were admitted to the University of Iowa Hospitals and Clinics NICU between 2000 and 2009, including both inborn and outborn neonates. We excluded infants who were transferred into our institution (n=67), those with major congenital anomalies (n=10), those weighing less than 401 g or more than 1,000 g (n=11), and those with missing data regarding the timing of antenatal corticosteroid administration needed to predict outcome using the Neonatal Research Network calculator (n=6). Five of the excluded newborns met more than one of the exclusion criteria. Our final study cohort included 237 infants.
The characteristics of women and their infants are shown in Table 1. The overall observed infant survival to discharge rate was 79.9%.
We compared the observed percent survival to the predicted survival estimates in Figure 1. Each infant's survival likelihood was calculated using the Neonatal Research Network outcomes calculator and then grouped into five subgroups by increasing survival estimates (shown in gray). These estimates then were compared with the observed percent survival for each of the predicted ranges (shown in black). For example, the 41–60% group represents the distribution of values for infants with predicted survival rates between 41% and 60% as compared with our observed survival rate of 80%. The observed survival rates were higher than those predicted by the Neonatal Research Network outcomes calculator in all subgroups except for 81–100%, the highest survival group. In a separate analysis, unadjusted survival rates for the decade by gestational age (compared with predicted rates) were found to be: 22 weeks, 33% (compared with 19%); 23 weeks, 58% (compared with 38%); 24 weeks, 87% (compared with 58%); and 25 weeks, 85% (compared with 70%).
We next examined our data using univariable analysis for antenatal factors that may have contributed to the observed survival. We first examined survival rates by antenatal corticosteroid exposure and gestational age. Among all infants delivered at 22–25 weeks of gestational age, 81.5% of those who received any antenatal corticosteroids survived; for those who received no antenatal corticosteroids, 40% survived. The number needed to treat with any antenatal corticosteroids to prevent one death was 2.4.
The difference in survival rate between those who received any antenatal corticosteroids and those who did not was not statistically significant for the 22-week, 23-week, and 24-week infants, but was statistically significant for the 25-week infants (P=.003). This result may be attributable to inadequate power at younger gestational age, because 96% of the infants were exposed to steroids and because the sample size in the 22-week to 23-week groups was small (n=47). We next considered route of delivery as an antenatal factor that may affect survival rates. Route of delivery made no significant difference in survival in our population.
We next calculated survival for infants born at 22–23 weeks of gestation and 24–25 weeks of gestation for each year of the study. The unadjusted survival rates and a linear regression line for all of the infants (22–25 weeks of gestational age) over the decade are shown in Figure 2. To adjust survival by year for other predictors of survival, a stepwise logistic regression analysis was performed, as shown in Table 2. Exposure to antenatal corticosteroids (OR 5.27, 95% CI 1.26–22.08), female sex (OR 3.21, 95% CI 1.42–7.26), gestational age by week (OR 1.89, 95% CI 1.27–2.81), 1-minute Apgar score (OR 1.39, 95% CI 1.15–1.69), and year of birth (OR 1.17, 95% CI 1.02–1.34) were all significant predictors, and mode of delivery and birth cohort number were not when adjusted for other variables. Grade IV intraventricular hemorrhage was a significant predictor of reduced likelihood of survival (OR 0.14) but was removed from the model because parents were offered withdrawal of support for these newborns.
Neurodevelopmental outcome at 18–22 months was evaluated next in 122 available surviving infants, as shown in Table 3. The performance of a neurodevelopmental evaluation was consistent among the gestational age groups, ranging from 63.6% to 66.7%. As expected, neurodevelopment impairment decreased with increasing gestational age (P<.043). In addition, cognitive impairment (Bayley-II mental developmental impairment less than 70 or Bayley-III cognitive composite score less than 80) ranged from 14.3% in neonates at 23 weeks of gestation to 50% in neonates at 22 weeks of gestation. There were only two neonates born at 22 weeks of gestation who survived and underwent developmental evaluation. Both were appropriate for gestational age male neonates who received antenatal steroids less than 12 hours before vaginal delivery. Neither had blindness, hearing loss, or cerebral palsy. One had a grade II intraventricular hemorrhage and, on neurodevelopmental testing, had a mental developmental index score of 50 and a psychomotor developmental index score of 69. The second newborn had a patent ductus arteriosus and stage 3 retinopathy of prematurity, and on neurodevelopmental testing had a mental developmental index score of 98 and a psychomotor developmental index score of 57.
We found that observed survival rates increased over the study decade and that the Neonatal Research Network outcomes calculator underestimated the survival of 22-week to 24-week infants. These findings highlight the need for each NICU to regularly evaluate preterm infant survival and neurodevelopmental outcomes and to base counseling and resuscitation policies on those outcomes. Recently, the Nuffield Council on Bioethics established recommendations that the parents' wishes take precedence at 23 weeks and established resuscitation as the standard at 24 weeks.11 The American Congress of Obstetricians and Gynecologists does not define limitations of medical capability but rather recommends that counseling for parents of extremely preterm newborns should include informing them that newborns delivered before 24 weeks of gestation are less likely to survive and that those who do survive “are not likely to survive intact.”6
Most established resuscitation guidelines are based on gestational age rather than on factors unique to each newborn and institution. The Neonatal Research Network study challenged this by identifying antenatal factors other than gestational age that were predictive of survival and neurodevelopment outcomes.9 Bader et al12 also raised concerns regarding parental counseling based only on gestational age. Both of these studies suggest that parents need to receive targeted counseling that is specific to the individual pregnancy. Our results also question the validity of resuscitation guidelines that are based solely on gestational age.
We found that the Neonatal Research Network outcomes calculator underestimated survival of infants born at 22–24 weeks of gestation cared for at our institution. This finding may be attributable partly to the high rate of antenatal corticosteroids administered to mothers at risk of delivering extremely preterm newborns. In 1972, it first was reported that antenatal corticosteroids reduced the incidence of respiratory distress syndrome in preterm infants.13 Two decades later, the National Institutes of Health issued a statement that all mothers at risk for delivering between 24 and 34 weeks of gestation should receive antenatal corticosteroids.14 These gestational age guidelines remain unchanged despite a recently updated American Congress of Obstetricians and Gynecologists committee opinion addressing the use of multiple course of antenatal corticosteroid therapy for fetal lung maturation.15 Although there has been a trend toward administration of antenatal corticosteroids for fetuses at risk of being delivered before 24 weeks of gestation, this practice remains controversial. Recent data suggest a beneficial effect of antenatal steroids on infant outcome when administered to mothers carrying fetuses of less than 24 weeks of gestation.7,16 Mori et al7 report decreased mortality in a Japanese cohort of 22-week to 23-week infants (adjusted hazard ratio 0.72 with 95% CI 0.53–0.97) as well as 24-week to 25-week infants (adjusted hazard ratio 0.65 with 95% CI 0.59–0.86) after antenatal corticosteroid administration. Zayek et al16 report improved survival in 22-week and 25-week infants but not 23-week and 24-week infants between two epochs (January 1, 1998, to June 30, 2003, compared with July 1, 2003, to December 31, 2008) that showed an increase in both cesarean delivery and antenatal corticosteroid use.
Our analysis showed a beneficial effect of higher 1-minute Apgar scores on survival rates, unlike previous studies that found the 5-minute Apgar score (but not the 1-minute score) to be predictive of survival.17,18 Our findings support the idea that survival is influenced by condition at birth, which may be affected by aggressive obstetric care. Simply stated, the extremely preterm newborn delivered in a physiologically stable condition has a better chance of survival.
There are a number of limitations to our study. First, our study was retrospective, using data from our neonatal registry. Second, our study was not designed to capture data pertaining to newborns whose parents decided to forego intensive care or who were not admitted to the NICU (ie, the seven newborns who died in the delivery room after attempted resuscitation). Third, the structure of the registry data precluded us from tracking infants who were transferred to other hospitals and evaluated at other centers after discharge. Last, our study period covered two different versions of Bayley Scales (BSID-II Bayley 1993 and Bayley-III Bayley 2006). Researchers have found that the mean Bayley-III cognitive score is approximately 10 points higher than the mean BSID-II mental developmental index in a similar population of term infants.10
In summary, our analysis found that the Neonatal Research Network outcomes calculator underestimated observed survival of 22-week to 24-week infants and that the adjusted likelihood of survival increased during 2000–2009. In addition, we found increased survival if antenatal corticosteroids were administered, the infant was female, additional gestational weeks attained, and as the decade progressed. These findings reflect improvement in survival of extremely preterm inborn infants who were admitted to the NICU and whose parents elected aggressive treatment. Because the survival and neurodevelopmental outcomes of infants born at 22 weeks of gestation were not as favorable as those for infants born at 23 week to 25 weeks of gestation, aggressive treatment of these children continues to be controversial. Consistent with other recent studies that report improved survival of infants born at less than 24 weeks of gestation, this study suggests that the use of antenatal corticosteroids and aggressive treatment should be considered for these children.
1. Stoll BJ, Hansen NI, Bell EF, Shankaran S, Laptook AR, Walsh MC, et al.. Neonatal outcomes of extremely preterm infants from the NICHD Neonatal Research Network. Pediatrics 2010;126:443–56.
2. Lee SK, McMillan DD, Ohlsson A, Pendray M, Synnes A, Whyte R, et al.. Variations in practice and outcomes in the Canadian NICU network: 1996–1997. Pediatrics 2000;106:1070–9.
3. Vohr BR, Wright LL, Dusick AM, Perritt R, Poole WK, Tyson JE, et al.. Center differences and outcomes of extremely low birth weight infants. Pediatrics 2004;113:781–9.
4. Cotten CM, Oh W, McDonald S, Carlo W, Fanaroff AA, Duara S, et al.. Prolonged hospital stay for extremely premature infants: risk factors, center differences, and the impact of mortality on selecting a best-performing center. J Perinatol 2005;25:650–5.
5. Kusuda S, Fujimura M, Sakuma I, Aotani H, Kabe K, Itani Y, et al.. Morbidity and mortality of infants with very low birth weight in Japan: center variation. Pediatrics 2006;118:e1130–8.
6. Perinatal care at the threshold of viability. ACOG Practice Bulletin No. 38. American College of Obstetricians and Gynecologists. Obstet Gynecol 2002;100:617–24.
7. Mori R, Kusuda S, Fujimura M, Neonatal Research Network Japan. Antenatal corticosteroids promote survival of extremely preterm infants born at 22 to 23 weeks of gestation. J Pediatr 2011;159:110–4.e1.
8. Carlo WA, McDonald SA, Fanaroff AA, Vohr BR, Stoll BJ, Ehrenkranz RA, et al.. Association of antenatal corticosteroids with mortality and neurodevelopmental outcomes among infants born at 22 to 25 weeks' gestation. JAMA 2011;306:2348–58.
9. Tyson JE, Parikh NA, Langer J, Green C, Higgins RD, National Institute of Child Health and Human Development Neonatal Research Networkl. Intensive care for extreme prematurity–moving beyond gestational age. N Engl J Med 2008;358:1672–81.
10. Anderson PJ, De Luca CR, Hutchinson E, Roberts G, Doyle LW. Underestimation of developmental delay by the new Bayley-III Scale. Arch Pediatr Adolescent Med 2010;164:352–6.
11. The ethics of premature delivery. Lancet 2006;368:1844.
12. Bader D, Kugelman A, Boyko V, Levitzki O, Lerner-Geva L, Riskin A, et al.. Risk factors and estimation tool for death among extremely premature infants: a national study. Pediatrics 2010;125:696–703.
13. Liggins GL, Howie RN. A controlled trial of antepartum glucocorticoid treatment for prevention of respiratory distress syndrome in preterm infants. Pediatrics 1972;50:515–25.
14. Effect of corticosteroids for fetal maturation on perinatal outcomes. NIH Consensus Statement 1994;12:1–24.
15. Antenatal corticosteroid therapy for fetal maturation. ACOG Committee Opinion No. 475. American College of Obstetricians and Gynecologists. Obstet Gynecol 2011;117(2Pt1):422–4.
16. Zayek MM, Trimm RF, Hamm CR, Peevy KJ, Benjamin JT, Eyal FG. The limit of viability: a single regional unit's experience. Arch Pediatr Adolescent Med 2011;165:126–33.
17. Drage JS, Kennedy C, Schwarz BK. The Apgar score as an index of neonatal mortality. A report from the collaborative study of cerebral palsy. Obstet Gynecol 1964;24:222–30.
© 2012 The American College of Obstetricians and Gynecologists
18. Mori R, Shiraishi J, Negishi H, Fujimura M. Predictive value of Apgar score in infants with very low birth weight. Acta Paediatr 2008;97:720–3.