Birth weight threshold for postponing preterm birth Ralph T. DePalma, MD, Kenneth J. Leveno, MD, Mary Ann Kelly, RN, M. Lynne Sherman, RN, and Thomas J. Carmody, PhD Dallas, Texas OBJECTIVE: The study was designed to determine the birth weight threshold at which obstetric efforts intended to delay delivery might potentially improve rates of neonatal morbidity and mortality among pregnancies delivered after spontaneous preterm labor or rupture of the membranes. STUDY DESIGN: We studied 1147 singleton infants with birth weights between 1000 and 2499 gm and whose only complication was spontaneous preterm labor or preterm rupture of the membranes. The Mantel-Haenszel X2 statistic was used to evaluate trends for neonatal mortality and several indexes of morbidity. RESULTS: The birth weight threshold for neonatal mortality was 1600 gm (p < 0.001). For neonatal morbidity the threshold was between 1600 and 1900 gm (p < 0.008). CONCLUSION: Aggressive obstetric attempts to prevent preterm birth for infants whose weights exceed 1900 gm offers few apparent potential benefits. (AM J OasTET GVNECOL 1992;167:1145-9).

Key words: Birth weight, preterm birth Obstetric approaches to prematurity are governed by expectations for neonatal outcome. During the past 20 years a great variety of innovations in obstetric care have been focused on preventing prematurity. Although prematurity is often defined as delivery at ::;37 weeks, two American collaborative investigations on obstetric approaches for prematurity established that only those infants ::;34 weeks could potentially benefit from delayed delivery." 2 Throughout these two decades expectations for neonatal outcome have progressively improved, with increased availability and continued innovation in neonatal care. Not only has the frontier for neonatal survival been pushed earlier into pregnancy, but survival of larger premature infants has become as good as that for term infants. For example, in 1963 Jackie Kennedy was delivered of a 34-week infant who died of respiratory distress. Patrick Kennedy weighed 4 pounds 1OV2 ounces; death in this weight category in a contemporary intensive care unit is extremely rare. This success story for neonatal care has prompted us to evaluate the birth weight threshold at which obstetric efforts intended to delay delivery might potentially improve neonatal mortality and morbidity.

Material and methods Between Jan. I, 1990, and May 31, 1991, 20,813 women delivered at Parkland Memorial Hospital.

From the Department of Obstetrics and Gynecology, The University of Texas Southwestern Medical Center. Presented at the Twelfth Annual Meeting of the Society of Perinatal Obstetricians, Orlando, Florida, February 3-8, 1992. Reprint requests: Ralph T. DePalma, MD, University of Texas Southwestern Medical Center, Department of Obstetrics and Gynecology, 5323 Harry Hines Blvd., Dallas, TX 75235-9032. 616139672

Nurses attending each delivery completed a perinatal data sheet, and research nurses assessed the data for consistency and completeness before electronic storage. Similarly, research nurses completed a neonatal data sheet for each infant discharged. These data were then linked to intrapartum results with a computerized network. There were a total of 1475 pregnancies (7.1 %) resulting in singleton infants weighing between 1000 and 2499 gm. From this group 1147 women (5.5%) were selected for study, these included only those pregnancies complicated by spontaneous pre term labor and preterm ru pture of the membranes and with a live fetus on admission. We chose this range of birth weights to correspond to approximately the 50th percentiles for 27 and 35 weeks, respectively.3 Not included were 149 multiple gestations and 328 singleton gestations. The singleton gestations were excluded because of medical or obstetric complications necessitating immediate delivery. The most common indications for exclusion were hypertension (n = 283), diabetes (n = 27), clinical chorioamnionitis on admission (n = 27), placenta previa (n = 18), and abruptio placentae (n = 6). Some of these women had multiple indications for exclusion. Statistics. The Mantel-Haenszel X2 statistic was used to evaluate for trends in neonatal mortality and morbidity. Specifically, linearly increasing or decreasing percentages of these neonatal outcomes were analyzed in relation to birth weight groups. The two highest birth weight categories were compared first, then the three highest categories, and so on, until all categories were compared. This procedure was used to establish a threshold birth weight below which the trend of increasing mortality or morbidity became apparent. Total days in hospital and in intensive care were com-

1146 DePalma et al.

October 1992 Am J Obstet Gynecol

Table I. Perinatal mortality in 1147 singleton pregnancies, with birth weights between 1000 and 2499 gm and with live fetuses on admission Birth weight

1000-1299

gm

No. of infants No. of perinatal deaths, total No. of perinatal deaths, anomalies No. of neonatal deaths, corrected* No. of fetal deaths, corrected*

1300-1599

gm

1600-1899

gm

1900-2199

gm

2200-2499

gm

Total

600 (22%) 1 (0.2%)

1147 (100%) 25 (2)

58 (2%) 8 (14%)

86 (3%) 7 (8%)

140 (5%) 8 (6%)

263 (10%) 1 (0.4%)

2 (3%)

3 (3%)

8 (6%)

1 (0.4%)

0

14 (0.1%)

6 (10%)

3 (3%)

0

0

1 (0.2%)

10 (0.9%)

0

1 (l %)

0

0

0

1 (0.1%)

*Malformations excluded.

pared across all birth weights with analysis of variance, and then individual categories were compared with Tukey's standardized range test. Because of the highly skewed distribution of total days in hospital and intensive care, the analysis of variance results were confirmed by the Mantel-Haenszel X2 statistic. The X2 statistic was used to test for an association between measures of mortality and morbidity and maternal race or infant sex.

Results The racial demographics for the 1147 pregnancies now reported included 602 (52%) black women, 298 (26%) Hispanic women, and 235 (20%) white women. The remaining 2% were Southeast Asian. All infants in this investigation were delivered at Parkland Hospital. A total of 498 (43%) women were nulliparous with a mean age and parity of 23 years and 1, respectively. Pre term ruptured membranes complicated 493 (43%) of the pregnancies. A total of 631 (55%) of the infants were female. Cesarean section was the route of delivery for 241 (21 %). Shown in Table I are perinatal deaths according to 300 gm birth weight increments in the 1147 pregnancies reported. There were a total of 25 deaths for an overall mortality rate of 22 per 1000. However, 14 (56%) of these deaths were the result of major congenital anomalies. Also shown in Table I are neonatal and fetal deaths according to birth weight groups, corrected to exclude malformations. The corrected mortality rate was nine per 1000, and there was only one neonatal death among infants weighing > 1600 gm. This 2300 gm infant was born after preterm labor with intact membranes and died of septic shock caused by group B streptococci. There was one fetal death; this fetus weighed 1475 gm and died of unknown cause during expectant management for pre term ruptured membranes. The corrected mortality rate was significantly as so-

ciated with weight (p < 0.001). Although no trend of deaths varying from birth weight was seen for birth weights from 1600 to 2499 gm, a significant trend was detected when the 1300 to 1599 gm category was added to the analysis and again when the 1000 to 1299 gm category was added. This pattern suggests that the death rate for infants with birth weights> 1600 gm is constant and lower than for those with birth weights 1600 gm. Because of the small number of infants with birth weights between 1000 and 1600 gm, this subset was divided into two groups to make comparisons more meaningful. For consistency the larger infants were also placed into 300 gm groups. In this investigation, which included 1147 infants whose birth weights were between 1000 and 2499 gm, there was only one death among 1003 infants whose birth weights exceeded 1600 gms, and this death was not directly attributable to complications of immaturity. The birth weight threshold for neonatal morbidity was 1600 to 1900 gm. For example, serious respiratory insufficiency as evidenced by the need for mechanical ventilation decreased significantly at birth weights of ~ 1600 gm with the statistical threshold at 1900 gm, and almost half of these higher birth weight infants

(nine of20 whose birth weights exceeded 1600 gm) had verified sepsis as opposed to pulmonary immaturity as a likely explanation of respiratory distress. Similarly, other serious neonatal morbidity (i.e., bronchopulmonary dysplasia and intraventricular hemorrhage) were also significantly correlated with the 1600 gm threshold. Both seizures and necrotizing enterocolitis were associated with a statistical threshold of 1900 gm. However, there were no deaths attributable to either of these among infants whose birth weights exceeded 1600 gm. The study population now reported consisted predominately of black infants (52%), although Hispanic and whites constituted 26% and 20%, respectively. Some investigators have concluded that maternal race and fetal sex have significant impact on the survival of premature infants. In the American collaborative study of dexamethasone for fetal lung maturation,7 only white singleton female infants whose gestational ages between 30 and 34 weeks benefited from corticosteroid therapy. Similarly, Sappenfield et al." reported that in 1980 the neonatal mortality rate was 1.6 times higher for black infants compared with white ones in spite of the fact that the survival rate at any given gestational age is greater for low-birth-weight black infants. This was explained to a large extent by the greater frequency of preterm births among blacks compared with whites. Wilcox and Russell,9 however, reported that the apparent superior survival for low-birth-weight black infants compared with white infants is reversed and becomes lower when population-specific standards for smallness are applied. We analyzed neonatal mortality and morbidity in relation to maternal race and infant sex for the pregnancies reported and found no significant correlations. Similar results for neonatal mortality were reported by Goldenberg et al. IO Specifically, no significant differences in survival associated with infant race or sex were

Birth weight threshold

Volume 167 Number 4, Part 1

found in 2061 infants whose birth weights were between 500 and 2600 gm (70% were black) and who were delivered between 1979 and 1981 at the Hospitals of the University of Alabama in Birmingham. Resnick et al. 11 analyzed 16,183 infants admitted to 10 neonatal intensive care centers and concluded that mortality was not related to maternal race or infant sex for birth weights > 1500 gm. However, among 4590 infants whose birth weights were between 1000 and 1500 gm, including approximately one third who were neonatal transports, they reported higher survival rates for female and black infants. Our results do not confirm those of Resnick et al. for infants weighing 1500 gm. Indeed, they reported the birth weight threshold for mortality to be 1500 gm. Both of these latter results by Resnick et al. are similar to those now reported. For the purpose of this study birth weight rather than gestational age was reported because of the inability in many instances to accurately estimate gestational age. Goldenberg et al. 12 reported on the survival of infants born between 1978 and 1984 at various gestational ages and found that the gestational age beyond which little could be gained was 34 weeks. Between 27 and 34 weeks' gestation there was an incremental improvement in survival that averaged 4% per week and ranged between 1% and 6%. Neonatal survival for infants with birth weights between 1550 and 2200 gm (31 to 34 weeks' gestation) increased from 93% to 98%. In the study now reported there were no neonatal deaths in 403 infants with birth weights between 1600 and 2200 gm, which corresponds to gestational ages between 31 and 34 weeks as described by Goldenberg et al. Although our results are based on birth weight rather than gestational age and because it is possible that disorders of fetal growth may influence birth weight and consequently have an impact on neonatal survival, we deliberately excluded those pregnancies at greater risk for growth disorders, including infants with malformations. It must also be considered that neonatal survival during these critical weeks of gestation has improved since 1984.

1149

In conclusion, we are somewhat surprised by the remarkably good outcomes for infants whose birth weights exceeded 1000 gm. At our institution the birth weight threshold at which postponement of delivery might be beneficial in terms of preventing neonatal death appears to be 1600 gm, independent of maternal race or infant sex. The threshold for neonatal morbidity, which included respiratory distress and necrotizing enterocolitis, was 1600 to 1900 gm. We believe that aggressive attempts at postponing delivery in infants whose birth weights exceed 1900 gm offers few benefits because of the efficacy of current neonatal management. As we have discussed, neonatal survival and morbidity can be influenced by a multitude of factors and it is most certainly prudent that health care providers exercise caution in the application of the results now reported. REFERENCES 1. Merkatz IR, Peter ]B, Barden TP. Ritodrine hydrochloride: a beta-mimetic agent for use in preteI'm labor. II. Evidence of efficacy. Obstet Gynecol 1980;56:7. 2. Collaborative Group on Antenatal Steroid Therapy: Effect of antenatal dexamethasone administration on the prevention of respiratory distress syndrome. AM] OBSTET GYNECOL 1981;141:276. 3. Brenner WE, Edelman DA, Hendricks CH. A standard of fetal growth for the United States of America. AM] OBSTET GYNECOL 1976;126:555. 4. Swischuk LE. Imaging of the newborn, infant and young child. ed 3. Baltimore: Williams & Wilkins, 1989:953. 5. Guzzetta F, Shackelford GD, Volpe S, Perlman]M, Volpe ]]. Periventricular intra parenchymal echodensities in the premature newborn: critical determinant of neurologic outcome. Pediatrics 1986;78:995. 6. Cox SM, Williams ML, Leveno K]. The natural history of preteI'm ruptured membranes: what to expect of expectant management. Obstet Gynecol 1988;71 :558-62. 7. Avery ME, Aylward G, Creasy R, Little AB, Stripp B. Update on prenatal steroid for prevention of respiratory distress. AM] OBSTET GYNECOL 1986;155:2. 8. Sappenfield WM, Buehler JW, Binkin NJ, et ai. Differences in neonatal and postneonatal mortality by race, birthweight, and gestational age. Public Health Rep 1987; 102: 182-92. 9. Wilcox A, Russell I. Why small black infants have a lower mortality rate than small white infants: the case for population-specific standards for birth weight. ] Pediatr 1990; 116:7-10. 10. Goldenberg RL, Nelson KG, Hale CD, et al. Survival of infants with low birth weight and early gestational age, 1979 to 1981. AM] OBSTET GYNECOL 1984;149:508-11. 11. Resnick MB, Carter RL, Ariet M, et al. Effect of birth weight, race, and sex on survival of low-birth-weight infants in neonatal intensive care. AM] OBSTET GYNECOL 1989;161:184-7. 12. Goldenberg RL, Nelson KG, Davis RO, Koski]. Delay in delivery: influence of gestational age and the duration of delay on perinatal outcome. Obstet Gynecol 1984;64: 480-4.

Birth weight threshold for postponing preterm birth.

The study was designed to determine the birth weight threshold at which obstetric efforts intended to delay delivery might potentially improve rates o...
535KB Sizes 0 Downloads 0 Views