Materials and Methods
We identified all deliveries between 23 0/7 and 23 6/7 weeks of gestation at Brigham & Women's Hospital, Boston, Massachusetts, from January 1995 to December 1999 and abstracted information from maternal and infant records. Acceptable dating criteria for inclusion were intrauterine insemination or in vitro fertilization dates, first-trimester ultrasonogram, or firm menstrual dating confirmed by second-trimester ultrasonogram with less than 6% difference in days between the ultrasonographic estimate and that of last menstrual period.14 To avoid violating the statistical assumption of independence inherent in an analysis that includes multiple-gestation pregnancies, we restricted the analysis to singleton pregnancies only. Women were excluded if they presented for elective pregnancy termination or had known fetal death or poor dating criteria.
Clinical characteristics included maternal age, administration of antenatal corticosteroids, gestational age at delivery, and presentation at delivery. Pregnancies were considered treated by steroids if delivery occurred more than 6 hours after the first injection of betamethasone. Maternal race is not universally recorded in the medical records of our institution; we therefore could not reliably add this information to our descriptive variables.
Neonatal records were abstracted for mortality and short-term morbidity, including the respiratory distress syndrome (RDS), intraventricular hemorrhage, cerebral white matter disease, chronic lung disease, necrotizing enterocolitis, and retinopathy of prematurity. Survival was defined as discharge from the neonatal intensive care unit (NICU). Neonatal RDS was defined as the need for supplemental oxygen with radiographic evidence of hyaline membrane disease. Neonates were considered at risk for RDS if they survived long enough to be admitted to the NICU and undergo chest radiography. Chronic lung disease was defined as persistent supplemental oxygen requirement at 36 corrected weeks of age. Neonates were considered at risk of chronic lung disease if they survived to 36 corrected weeks' gestation. Retinopathy of prematurity was evaluated by ophthalmologic examination at or after 4 weeks of life and considered severe if it was grade 3 or higher. Neonates were considered at risk of necrotizing enterocolitis after the second day of life.15 Diagnoses of periventricular leukomalacia and intraventricular hemorrhage were confirmed by ultrasonography after the first 6 hours of life. Grade 3 or 4 hemorrhage was considered severe.
Statistical analyses were done by using the STATA statistical package (Stata Corp., College Station, TX). Fisher exact and Student statistics were used to compare categorical and continuous demographic and clinical characteristics, respectively, of surviving and non-surviving neonates. Exact binomial confidence intervals (CIs) were calculated because of the limited sample. Proportional hazards modeling and Kaplan–Meier survival curves were used to estimate a hazard rate and graphically illustrate the effect of clinically important variables (antenatal corticosteroids and gestational age) on survival. Results were considered statistically significant at P ≤.05.
Thirty-three women with singleton pregnancies met dating criteria for inclusion between 23 0/7 and 23 6/7 weeks' gestation. Six of 33 live born infants died before admission to the NICU. Eleven of 33 infants survived (survival rate 0.33; exact 95% CI 0.18, 0.52). Among 11 survivors, 7 were discharged to their parents, 1 went to a chronic care facility, and 4 were transferred to a level-II NICU closer to their parents' homes. Based on a sample of 33 live births, our analysis was able to detect a 30% difference in survival by gestational age category with a power of 0.95.
Table 1 shows data according to clinical presentation. We found no differences in survival by clinical presentation of parturients. However, survivors were more likely than nonsurvivors to have received at least 6 hours' exposure to antenatal corticosteroids. More advanced gestational age in the 23rd week was associated with increased survival: 0 of 12 at 23 0/7 to 23 2/7 weeks, 4 of 10 at 23 3/7 to 23 4/7 weeks, and 7 of 11 at 23 5/7 to 23 6/7 weeks. Figure 1 shows the Kaplan–Meier survival curves for each gestational-age category.
With respect to short-term neonatal morbidity, we compared survivors with nonsurvivors who lived long enough to become at risk for each morbidity (Table 2). Survivors and at-risk nonsurvivors received a diagnosis of RDS and required exogenous surfactant. All survivors progressed to chronic lung disease. Two of 11 survivors had grade 3 or higher retinopathy of prematurity. Because they died, no nonsurvivor received a diagnosis of chronic lung disease or retinopathy of prematurity.
We did not detect a difference in proportions of survivors and at-risk nonsurvivors diagnosed with necrotizing enterocolitis. One of 11 survivors and 8 of 12 at-risk nonsurvivors developed severe intraventricular hemorrhages (grade 3 or 4) (P = .03). A significant difference in periventricular leukomalacia was also found between survivors and at-risk nonsurvivors: 1 of 11 and 7 of 12, respectively (P = .03) (Table 2). The survivor with severe intraventricular hemorrhage did not develop periventricular leukomalacia. All nonsurvivors admitted to the NICU died after decisions to withdraw care after consultation with the family because of worsening clinical prognoses, rather than after calamitous events and unsuccessful resuscitation.
Cox proportional hazards were calculated to examine the effect on survival of advancing gestation in the 23rd week and antenatal corticosteroids while controlling for gestational age. In a model for gestational age in which the 23 0/7 to 23 2/7 category was the reference group, the survival hazard ratio was statistically significant and improved with each successive gestational age category: 0.26 (95% CI 0.09, 0.76) at 23 3/7 to 23 4/7 weeks (P = .014) and 0.10 (95% CI 0.03, 0.34) at 23 5/7 to 23 6/7 weeks (P < .0001), suggesting successive improvements in survival with advanced gestation in the 23rd week. In the second model, the hazard ratio for corticosteroid administration was initially significant (0.37 [95% CI 0.15, 0.90]) as a bivariate but became nonsignificant with introduction of control variables for gestational age in the 23rd week (0.66 [95% CI 0.25, 1.76]). This change suggests that the apparent beneficial effect of antenatal corticosteroids are a proxy variable for more advanced gestational age.
Our dating criteria were among the most strict in the current literature,2–12 yet are reasonably attainable in routine clinical practice. Approximately one third of pregnancies delivered at 23 weeks survived to discharge from the NICU. Most survivors were discharged to their parents. Our observed survival rate was consistent with that found by Hack and Fanaroff2 in their review of neonatal outcomes at similar gestational ages. Survival was limited to neonates born in the latter half of the 23rd week.
All survivors in our series developed retinopathy of prematurity and chronic lung disease. It has been suggested that children with chronic lung disease might be at increased risk of death because of sequelae of pulmonary insufficiency.2 However, pulmonary abnormalities also have been reported to resolve in children who survived extremely preterm birth.16 In our study, periventricular leukomalacia and severe intraventricular hemorrhage were present in surviving neonates at rates consistent with those in other studies of 23-week outcomes.3,8,17 The presence of those neurologic findings increases the risk of long-term neurologic sequelae and poor cognitive development.18,19 Most survivors were free of severe intraventricular hemorrhage, whereas most nonsurvivors had it, suggesting that severe intraventricular hemorrhage was a poor prognostic indicator of survival at 23 weeks' gestation. All of the NICU deaths resulted from withdrawal of life support, so the prospect of a poor neurologic outcome in the setting of severe intraventricular hemorrhage might strongly bias parents and practitioners against continued care compared with other morbid conditions in the 23rd week.
The rate of necrotizing enterocolitis was consistent with that published for the same birth weight category.2 Severe retinopathy of prematurity occurred in few neonates and at a rate below that published for the same gestational age.2 The reasons for the lower rate of severe retinopathy of prematurity are unclear and might have been a function of the small sample. Antenatal corticosteroids had no discernable independent effect on survival at 23 weeks. That finding was noted in other studies of extremely preterm neonates2 and may indicate a limitation.
Our analysis was limited by sample size. Negative findings should therefore be interpreted with caution. Our data pertain only to singleton pregnancies and might not apply to 23-week fetuses from multifetal pregnancies. Our study represents the experience at a single institution with an aggressive approach toward perinatal and pediatric intervention. Not all institutions, practitioners, or parents would be comfortable with this level of intervention, especially when long-term outcomes of infants born at 23 weeks' gestation remain unknown.
Our data provide additional information for counseling parturients who present with impending 23-week deliveries and their families. Further research into long-term morbidity and quality of life among these infants is needed before equitable and ethical approaches to viability in the 23rd week of gestation can be agreed on.
1. Hack M, Fanaroff AA. Outcomes of extremely-low-birth-weight infants between 1982 and 1988. N Engl J Med 1989;321:1642–7.
2. Hack M, Fanaroff AA. Outcomes of children of extremely low birthweight and gestational age in the 1990s. Early Human Dev 1999;53:193–218.
3. Allen MC, Donohue PK, Dusman AE. The limit of viability—neonatal outcome of infants born at 22 to 25 weeks' gestation. N Engl J Med 1993;329:1597–601.
4. Perinatal care at the threshold of viability. American Academy of Pediatrics Committee on Fetus and Newborn. American College of Obstetricians and Gynecologists Committee on Obstetric Practice. Pediatrics 1995;96:974–6.
5. Draper ES, Manktelow B, Field DJ, James D. Prediction of survival for preterm births by weight and gestational age: retrospective population based study. BMJ 1999;319:1093–7.
6. Bottoms SF, Paul RH, Iams JD, Mercer MD, Thom EA, Roberts JM, et al. Obstetric determinants of neonatal survival: Influence of willingness to perform cesarean delivery on survival of extremely low-birth-weight infants. Am J Obstet Gynecol 1997;176:960–6.
7. Emsley HCA, Wardle SP, Sims DG, Chiswick ML, D'Souza SW. Increased survival and deteriorating developmental outcomes in 23- to 25-week-old gestation infants, 1990–4 compared with 1984–9. Arch Dis Child 1998;78:99–104.
8. Hussain N, Galal M, Ehrenkranz RA, Herson VC, Rowe JC. Pre-discharge outcomes of 22-27 weeks gestational age infants born at tertiary care centers in Connecticut: Implications for perinatal management. Conn Med 1998;62:131–7.
9. Kramer WB, Saade GR, Goodrum L, Belfort M, Moise KJ Jr. Neonatal outcome after active perinatal management of the very premature infant between 23 and 27 weeks' gestation. J Perinatol 1997;17:439–43.
10. O'Shea TM, Kleinpeter KL, Goldstein DJ, Jackson BW, Dillard RG. Survival and developmental disability in infants with birth weights of 501 to 800 grams, born between 1979 and 1994. Pediatrics 1997;100:982–6.
11. Holtrop PC, Ertzbischoff LM, Roberts CL, Batton DG, Lorenz RP. Survival and short-term outcome in newborns of 23 to 25 weeks' gestation. Am J Obstet Gynecol 1994;170:1266–70.
12. Stevenson DK, Wright LI, Lemons JA, Oh W, Korones SB, Papile LA, et al. Very low birth weight outcomes of the National Institute of Child Health and Human Development Neonatal Research Network, January 1993 through December 1994. Am J Obstet Gynecol 1998;179:1632–9.
13. Morse SB, Haywood JL, Goldenberg RL, Bronstein J, Nelson KG, Carlo WG. Estimation of neonatal outcome and perinatal therapy use. Pediatrics 2000;105:1046–50.
14. Hadlock FP. Ultrasound determination of menstrual age. In: Callen PW, ed. Ultrasonography in obstetrics and gynecology. 3rd ed. Philadelphia: W.B. Saunders, 1994:86–101.
15. Berseth CL, Abrams SA. Special gastrointestinal concerns. In: Taeusch HW, Ballard RA, eds. Avery's diseases of the newborn. 7th ed. Philadelphia: WB Saunders, 1998:965–78.
16. Hansen T, Corbet, A. Chronic lung disease. In: Taeusch HW, Ballard RA, eds. Avery's diseases of the newborn. 7th ed. Philadelphia: WB Saunders, 1998:634–47.
17. Cooke RWI. Improved outcome for infants at the limits of viability. Eur J Pediatr 1996;155:665–7.
18. Roth SC, Bauldwin J, McCormick DC, Edwards AD, Townsend J, Stewart AL, et al. Relation between ultrasound appearance of the brain of very preterm infants and neurodevelopmental impairment at eight years. Dev Med Child Neurol 1993;35:755–68.
19. Whitaker AH, Feldman JF, Van Rossem R, Schonfeld IS, Pinto-Martin JA, Torre C, et al. Neonatal cranial ultrasound abnormalities in low birth weight infants: Relation to cognitive outcomes at six years of age. Pediatrics 1996;98:719–29.