Bates, Elizabeth MD; Rouse, Dwight J. MD, MSPH; Mann, Merry Lynn BS; Chapman, Victoria MPH; Carlo, Waldemar A. MD; Tita, Alan T. N. MD, PhD
Delivery before 39 weeks of gestation is associated with increased neonatal respiratory morbidity compared with delivery at 39 completed weeks (ie, 39 0/7 to 39 6/7).1–3 Because fetal lung maturity reduces the risk of respiratory morbidity, confirmation of fetal lung maturity is a recognized exception to longstanding recommendations against elective delivery before 39 weeks of gestation.4–7 Among a large cohort of U.S. women who underwent elective cesarean delivery at term (ie, at 37 weeks or later), the risks of respiratory morbidity and other adverse neonatal outcomes (including need for cardiopulmonary resuscitation, hypoglycemia, seizures, suspected or proven sepsis, and admission to the neonatal intensive care unit [NICU]) reached their lowest point at 39 to 40 weeks of gestation.3 Despite existing recommendations, in one large study, over a third of elective cesarean deliveries were performed before 39 weeks.3 However, because information on fetal lung maturity testing was lacking in that study, it is unclear whether the increased neonatal morbidity was due to failure to confirm lung maturity or whether delivery before 39 weeks even after confirmed pulmonary maturity is associated with worse neonatal outcomes than delivery at 39 weeks. Therefore, we undertook this study to evaluate the hypothesis that adverse neonatal outcomes are increased when delivery is undertaken before 39 weeks of gestation even after documented fetal lung maturity when compared with delivery at 39 to 40 weeks.
We conducted a retrospective cohort study of women receiving prenatal care and delivering at our center from January 1999 to December 2008. The study was approved by the Institutional Review Board of the University of Alabama at Birmingham. The study period represents the time frame for which results of pulmonary maturity testing were available. Data were derived from our electronic obstetric database, supplemented by direct review of the medical records of women and their newborns as detailed below. Our electronic database has been in operation since 1979 and contains more than 1,000 coded antepartum, intrapartum, and postpartum data fields per patient. Dedicated trained personnel are responsible for entering the antepartum data on an ongoing basis and for conducting audits. Immediately after delivery, physicians or nurse practitioners record the intrapartum and postpartum data on standardized forms.8
Pregnancy outcomes for two groups of women with singleton pregnancies were compared: 1) the early (lung maturity) group included those who delivered between 36 0/7 and 38 6/7 weeks after positive fetal lung maturity testing (based on either the presence of phosphatidyl glycerol or a lethicin-sphingomyelin ratio of 2.0 or higher if nondiabetic or 3.5 or higher if diabetic).6 This group was identified by review of laboratory records for lung maturity testing and verified by direct review of maternal charts; 2) the comparison group included those who delivered at 39 or 40 completed weeks of gestation (without maturity testing). Gestational age was based on the best obstetric estimate used by providers after considering last menstrual period and ultrasound criteria as recommended by the American College of Obstetricians and Gynecologists.9 Mode of delivery in both groups was based on usual obstetric indications. Intended mode of delivery was defined based on the actual mode of delivery and type of labor: intended cesarean delivery included those who had a cesarean delivery without labor; all others delivered either vaginally or by cesarean after induced, augmented, or spontaneous labor were coded as intended vaginal deliveries.
Ascertainment of the characteristics of these women and their pregnancy outcomes was through the obstetric database. Those with either multiple births or factors suggesting fetal compromise, including major congenital anomalies, umbilical cord prolapse, nonreassuring antepartum testing, placental abruption, and oligohydramnios were excluded. All admissions to the NICU were identified for direct chart review by two coauthors (E.B. and M.L.M.) to ascertain neonatal outcomes. Questions or discrepancies regarding outcome abstraction were resolved in consultation with a third author (A.T.N.T.).
Several neonatal outcomes were examined, including neonatal death, respiratory distress syndrome (RDS), transient tachypnea of the newborn, bronchopulmonary dysplasia, persistent pulmonary hypertension, need for respiratory support (including ventilator support or any other mode of supplemental oxygenation), use of surfactant, metabolic complications including hypoglycemia, hyperbilirubinemia requiring treatment, generalized seizures, necrotizing enterocolitis, hypoxic ischemic encephalopathy periventricular leukomalacia, feeding difficulties, and suspected or proven sepsis. The primary outcome was the composite of all of these outcomes. In addition, we examined a second composite outcome that excluded suspected sepsis, as this was the neonatal outcome with the highest incidence. Two other neonatal outcomes, NICU admission and prolonged hospitalization (defined as neonatal hospital stay more than 4 days), were also examined but were not included in the primary composite outcome. Neonatal outcomes were defined according to standard ICD-9 diagnostic codes assigned by the neonatology attending on discharge, death, or transfer of the neonate. Although we focused on adverse neonatal outcomes severe enough to warrant a NICU admission, we recognized that mild forms of morbidities such as hyperbilirubinemia and suspected sepsis would not always lead to a NICU admission. Therefore, we also performed post hoc direct chart review of a random sample of 200 charts of neonates who were not admitted to the NICU (in a ratio of one early delivery to four at 39–40 weeks). We examined the primary outcome by group to determine whether results were similar to the main findings.
We estimated that there were approximately 450 women delivered before 39 weeks after documented lung maturity during the study period. Comparing them with a minimum of 1,800 women delivered at 39 to 40 weeks (ie, 1:4 ratio) would provide at least 80% power to detect a 4% absolute difference in the primary outcome from a baseline incidence of 7% among births at 39 to 40 weeks with 2-sided α of 0.05. Statistical analyses were accomplished using SAS 9.1. The χ2 and Student t tests were used as appropriate to compare descriptive characteristics. The Mantel-Haenszel χ2 test for trend was used to assess for linear trends in outcomes by gestational age from 36 to 39 or 40 weeks. Logistic regression was used to adjust for potentially confounding covariates including maternal age, ethnicity, parity, neonatal sex, intended mode of delivery, and any medical complication (including diabetes mellitus and chronic hypertension) for the comparison between the early group and the 39- to 40-week group. Birth weight and maternal body mass index (BMI, calculated as weight (kg)/[height (m)]2) were adjusted for in additional models.
There were a total of 490 singleton deliveries at 36 to 38 weeks after positive pulmonary maturity testing and 13,571 singleton deliveries at 39 or 40 weeks during the study period. After applying exclusion criteria (Fig. 1), a total of 13,798 mother-neonate dyads comprised the study population: 459 delivered between 36 and 38 weeks after positive fetal lung maturity testing and 13,339 delivered at 39 to 40 weeks. Reasons for fetal lung maturity testing before 39 weeks were diabetes (with or without hypertension) in 37.1%, previous uterine rupture or surgery not restricted to lower segment (22.9%), chronic hypertension (7.2%), preeclampsia (5.2%), placenta previa (4.1%), elective (4.1%), history of fetal death or abruption (2.8%), cholestasis of pregnancy (2.8%), and other (13.8%). The “other” category included a wide range of indications such as suspected growth restriction; hydramnios; isoimmunization; and maternal cardiac, pulmonary or other medical conditions.
Among women delivered early after fetal lung maturity, 42.5% were at 36 completed weeks, 40.7% at 37 weeks, and only 16.8% at 38 weeks compared with 56.2% at 39 weeks and 43.8% at 40 weeks in the comparison group. Mean gestational age was 37.1± 0.7 weeks for the early group compared with 39.8± 0.6 weeks. Maternal and neonatal characteristics for the two study groups are presented in Table 1. Women in the early delivery group were significantly more likely to be older, white, married, and multiparous. Women in the early group initiated prenatal care earlier—mean gestation at first ultrasonogram was significantly lower and a higher proportion had ultrasonograms before 20 completed weeks of gestation. As expected, these women were also more likely to have obstetric or medical complications such as hypertensive disorders or diabetes. Mean birth weight was significantly higher in the 39- to 40-week group, and a higher proportion of neonates were small for gestational age.
The unadjusted rate of the primary composite adverse neonatal outcome was significantly more common in the early delivery group as was the second composite outcome, that is, excluding suspected sepsis (Fig. 2). Both composite adverse outcomes were more than twofold higher in neonates delivered at less than 39 weeks compared with those delivered at 39 or 40 weeks; unadjusted relative risk (RR) was 2.4 (95% confidence interval [CI] 1.7–3.5) and 2.8 (CI 1.9–4.2), respectively. The incidence of the primary composite outcome decreased with increasing gestational age (P for trend <.001): 9.2% (CI 5.9%–14.1%) at 36, 3.2% (CI 1.5%–6.8%) at 37, 5.2% (CI 2.0%–12.6%) at 38, and 2.5% (CI 2.2%–2.8%) at 39–40 weeks.
There were no perinatal deaths in the early lung maturity group, whereas there were 24 (0.18%) in the 39- to 40-week group, of which 22 (91.7%) were stillbirths; this was not significantly different. There was one case of intraventricular hemorrhage and seven cases of hypoxic ischemic encephalopathy in the 39- to 40-week group; neither of these outcomes was present in the early group. There was no case of necrotizing enterocolitis or periventricular leukomalacia in either group, and no differences in the incidence of persistent pulmonary hypertension, generalized seizures, or need for cardiopulmonary resuscitation between groups. The unadjusted incidence proportions of other individual adverse outcomes included in the composite, as well as admission to the NICU and prolonged hospitalization, were more frequent in the early group compared with the 39- to 40-week group (Table 2).
The primary composite adverse outcome was further stratified by each of three factors: intended vaginal or cesarean mode of delivery, presence or absence of maternal diabetes, and dating ultrasonogram obtained at or before 20 weeks compared with after 20 weeks. As shown in Figure 3, the primary outcome was more frequent in the early group for each subgroup of these characteristics. In addition, the primary outcome was associated with early delivery regardless of the presence (RR 1.7, 95% CI 1.1–2.8) or absence (RR 2.4, CI 1.4–4.2) of any medical complication including diabetes and hypertension.
Odds ratios (ORs) for the association between adverse neonatal outcomes (composite or individual) and delivery group (39- to 40-week group as referent) from logistic regression models adjusting for maternal age, race, parity, medical complications (hypertensive or diabetes), intended mode of delivery, and neonatal sex are presented in Table 3. Earlier delivery after confirmed lung maturity testing was associated with a 1.6-fold increase in the primary outcome and significant increases in other adverse neonatal outcomes by as much as 7.6-fold for RDS and more than 11-fold for treated hyperbilirubinemia. Also significantly associated with the primary outcome were maternal age as a continuous variable (OR 1.03; CI 1.01–1.05), male neonate (OR 1.25; CI 1.01–1.55), nulliparity (OR 2.15; CI 1.69–2.74), presence of medical complications (OR 1.55; CI 1.21–2.00), and intended vaginal delivery compared with intended cesarean delivery (OR 0.49; CI 0.36–0.68).
When we adjusted for actual (instead of intended) mode of delivery, the results for the outcomes presented in Table 3 were materially unchanged. For example, early delivery after lung maturity was associated with a 1.7-fold (CI 1.1–2.7) and 1.9 fold (CI 1.2–3.1) increase in the primary composite and alternative composite outcomes, respectively; the adjusted ORs for cesarean delivery were 3.6 (CI 2.8–4.6) and 4.1 (CI 3.1–5.5), respectively. Further adjustment of this model by type of labor (induced, spontaneous, none) gave similar results; compared with no labor, induced (adjusted OR 1.43; CI 0.95–2.15) and spontaneous labor (adjusted OR 1.3; CI 0.89–1.91) were not significantly associated with the primary outcome.
We conducted a series of validity and other analyses. When adjusted further for birth weight, the OR for the association between the primary outcome and early delivery increased slightly to 1.83 (CI 1.16–2.87). In analyses adjusting additionally for BMI in the primary outcome model, BMI was not a significant predictor and the results were limited by the relatively large number of those missing BMI information. Also the primary results were materially unchanged by excluding all neonates with low birth weight (less than 2,500 g). An examination of different cutoffs of the lecithin-sphingomyelin ratio was severely limited by the size of the early group but suggested that a ratio of at least 4:5 is required for adverse neonatal outcome to be highly unlikely.
The review of 200 randomly selected charts of neonates who were not admitted to the NICU supported our main results. Neonates in the early group were more likely to experience one or more adverse outcomes compared with the 39- to 40-week group (RR 1.5, CI 1.0–2.4). The majority of these were mild forms of hyperbilirubinemia or suspected sepsis.
A large body of available evidence indicates that delivery before 39 weeks of gestation is associated with increased risks of adverse neonatal outcomes compared with delivery at 39 or 40 weeks,1–3,10,11 and documentation of fetal lung maturity before early delivery lowers this risk.4–7 Even so, early delivery even after positive fetal lung maturity was associated with a nearly twofold increase in the primary composite adverse neonatal outcome in our large cohort. Furthermore, several individual adverse neonatal outcomes including RDS, treated hyperbilirubinemia, hypoglycemia and suspected or proven sepsis were significantly increased (by up to 7.6-fold for RDS and over 11-fold for treated hyperbilirubinemia) with early delivery. A composite outcome excluding suspected sepsis (the most frequent adverse neonatal outcome) was also increased, as were admission to the NICU and prolonged hospitalization. The increase in the primary composite outcome persisted regardless of the presence or absence of diabetes, chronic hypertension, or other maternal medical illnesses. As expected the frequency of adverse outcomes was higher among those with diabetes compared with nondiabetics. Adverse outcomes were also increased in the early group regardless of intended vaginal compared with cesarean mode of delivery, and the presence or absence of a dating ultrasonogram before 20 weeks of gestation. Taken together these findings are consistent with relative immaturity at 36 to 38 weeks (regardless of lung maturity) compared with 39 to 40 weeks and lower threshold for admission to the NICU and for invasive sepsis work-ups (suspected sepsis).
These findings should be considered in the light of our study's limitations. First, the retrospective study design limited our ability to identify and correct potential misclassification of neonatal outcomes. However, the extent of such bias would have to differ markedly by study group to negate our findings. Second, direct chart review of adverse neonatal outcomes focused primarily on outcomes associated with a NICU admission. Therefore, our results emphasize the more severe spectrum of each adverse outcome. Nonetheless, review of outcomes among the random sample of non-NICU admissions found a consistent increase in adverse neonatal outcomes among those delivered early after positive lung maturity testing. Third, confounding remains a possibility because patients were not randomized into policies of early delivery after fetal lung maturity compared with delaying delivery until 39 weeks. Finally, our study does not fully address the risk of stillbirth associated with either delivery strategy. Although there was no stillbirth in the early delivery group compared with 22 (0.16%) in the 39- to 40-week group (not statistically significant), this reflects the fact that fetal lung maturity testing is undertaken only when there is a viable fetus. Therefore, the true risk of stillbirth is underestimated in the early group—assuming that one stillbirth occurred; the risk would be 0.22%. Multiple comparisons are not a major concern because we specified a primary composite outcome a priori. The results for individual outcomes are internally consistent with our primary finding of increased adverse neonatal outcome with early delivery. We note that severe adverse outcomes such as necrotizing enterocolitis or intraventricular hemorrhage that are more likely to be associated with increased risk of death or long-term impairment were not prevalent.
There is a paucity of studies that specifically compare risks of adverse neonatal outcomes between neonates delivered early after lung maturity testing to the baseline risks at 39 or 40 weeks. Several reports have examined the risk of adverse outcomes among neonates delivered early with and without maturity testing.12–15 Some reports suggest that, among well-dated pregnancies at or near term (36–38 weeks), the use of fetal lung maturity testing does not reduce the incidence of respiratory distress syndrome12,13 and is not cost-effective.14 A recent survey of obstetric providers in the United States revealed that many do not think amniocentesis for fetal lung maturity before near-term (or early-term) delivery is beneficial, and this belief has contributed to a decline in amniocentesis for fetal lung maturity.16 However, one study based on the FLM TDX II test showed that delivery gestational age and fetal lung maturity result were independent determinants of the risk of RDS,15 suggesting that there is no threshold value at which the risks converged. For example, for each reported value of the FLM TDX-II, delivery at 37 weeks was consistently associated with a higher risk of RDS compared with delivery at 39 weeks.15 Our results are consistent with those reported in this study. Quite similar to findings from the recent report from the Eunice Kennedy Shriver National Institute of Child Health and Human Development Maternal-Fetal Medicine Units Network,3 early delivery was more common among older, white, married women and those who initiated prenatal care early (with earlier dating ultrasound examination). Ironically, these characteristics may be assumed to place women at lower risk for adverse outcomes.
We believe the differences between our study groups are attributable in large measure to the difference in gestational age rather than to the indications for fetal lung maturity testing. At our institution, antepartum testing is routine for pregnancies with known risk factors for fetal death. Consistent with American College of Obstetricians and Gynecologists recommendations, the fetal lung maturity testing group did not include pregnancies with fetal (or maternal) indications mandating delivery such as nonreassuring antenatal fetal testing.6 We also excluded pregnancies with indications for immediate delivery from the comparison group. There were fewer small-for-gestational age neonates in the early delivery group, likely owing to close surveillance and delivery without lung maturity testing. Therefore, we studied pregnancies that were not at immediate fetal risk; fetal lung maturity testing was done to prevent future fetal or maternal compromise. Indications for lung maturity testing were present in both study groups. For example, among patients with mild preeclampsia diagnosed before 37 weeks (when we typically deliver), a proportion accept fetal lung testing and delivery at 36 weeks whereas others choose to wait until 37 weeks. Overall, our findings suggest that in the absence of ongoing concern about fetal death or maternal well-being if the pregnancy continued, delivery should be delayed until 39 weeks. Although fetal lung maturity testing in our study was based on the lethicin-sphingomyelin ratio, we would expect similar results with the other tests because they have negative predictive values (95% or greater) comparable to the lecithin-sphingomyelin ratio at the cutoffs of 2 to 3.5 used in this study.6
It is well established that elective delivery before 39 weeks is associated with a preventable increase in neonatal morbidity and admissions to the NICU (and increased health care costs).1–3,10,11,17,18 Moreover, because purely elective deliveries by definition occur in pregnancies at lowest risk of fetal death, the risk of unexplained stillbirth (less than 0.05%) is considerably lower than observed in this study.19,20 Therefore, although our study population did not include a sizable proportion of purely elective deliveries (a result of center-specific practices), a natural extension of our findings is that purely elective fetal lung maturity testing and early delivery should be avoided. Replication of these findings in a cohort of women undergoing purely elective delivery is recommended.
1. Morrison JJ, Rennie JM, Milton PJ. Neonatal respiratory morbidity and mode of delivery at term: influence of timing of elective caesarean section. Br J Obstet Gynaecol 1995;102:101–6.
2. Zanardo V, Simbi AK, Franzoi M, Soldà G, Salvadori A, Trevisanuto D. Neonatal respiratory morbidity risk and mode of delivery at term: influence of timing of elective caesarean delivery. Acta Paediatr 2004;93:643–7.
3. Tita AT, Landon MB, Spong CY, Lai Y, Leveno KJ, Varner MW, et al; Eunice Kennedy Shriver NICHD Maternal-Fetal Medicine Units Network. Timing of elective repeat cesarean delivery at term and neonatal outcomes. N Engl J Med 2009;360:111–20.
4. Fetal maturity assessment prior to elective repeat cesarean delivery. ACOG Committee Opinion No. 98. American College of Obstetricians and Gynecologists. Washington, DC: ACOG; 1991.
5. Cesarean delivery on maternal request. ACOG Committee Opinion No. 394. American College of Obstetricians and Gynecologists. Obstet Gynecol 2007;110:1501.
6. Fetal lung maturity. ACOG Practice Bulletin No. 97. American College of Obstetricians and Gynecologists. Obstet Gynecol 2008;112:717–26.
7. Induction of labor. ACOG Practice Bulletin No. 107. American College of Obstetricians and Gynecologists. Obstet Gynecol 2009;114(2 pt 1):386–97.
8. Tita AT, Biggio JR, Chapman V, Neely C, Rouse DJ. Perinatal and maternal outcomes in women with sickle or hemoglobin C trait. Obstet Gynecol 2007;110:1113–9.
9. American Academy of Pediatrics and the American College of Obstetricians and Gynecologists. Guidelines for perinatal care. 6th ed. Washington, DC: ACOG; 2007.
10. Bailit JL, Gregory KD, Reddy UM, Gonzalez-Quintero VH, Hibbard JU, Ramirez MM, et al. Maternal and neonatal outcomes by labor onset type and gestational age. Am J Obstet Gynecol 2010;202:245.e1–12.
11. Wilmink FA, Hukkelhoven CW, Lunshof S, Mol BW, van der Post JA, Papatsonis DN. Neonatal outcome following elective cesarean section beyond 37 weeks of gestation: a 7-year retrospective analysis of a national registry. Am J Obstet Gynecol 2010;202:250.e1–8.
12. Kjos SL, Berkowitz KM, Kung B. Prospective delivery of reliably dated term infants of diabetic mothers without determination of fetal lung maturity: comparison to historical control. J Matern Fetal Neonatal Med 2002;12:433–7.
13. Luo G, Norwitz ER. Revisiting amniocentesis for fetal lung maturity after 36 weeks' gestation. Rev Obstet Gynecol 2008;1:61–8.
14. Myers ER, Alvarez JG, Richardson DK, Ludmir J. Cost-effectiveness of fetal lung maturity testing in preterm labor. Obstet Gynecol 1997;90:824–9.
15. Parvin CA, Kaplan LA, Chapman JF, McManamon TG, Gronowski AM. Predicting respiratory distress syndrome using gestational age and fetal lung maturity by fluorescent polarization. Am J Obstet Gynecol 2005;192:199–207.
16. McGinnis KT, Brown JA, Morrison JC. Changing patterns of fetal lung maturity testing. J Perinatol 2008;28:20–3.
17. van den Berg A, van Elburg RM, van Geijn HP, Fetter WP. Neonatal respiratory morbidity following elective caesarean section in term infants: a 5-year retrospective study and a review of the literature. Eur J Obstet Gynecol Reprod Biol 2001;98:9–13.
18. Hansen AK, Wisborg K, Uldbjerg N, Henriksen TB. Risk of respiratory morbidity in term infants delivered by elective caesarean section: cohort study. BMJ 2008;336:85–7.
19. Smith GC, Pell JP, Dobbie R. Caesarean section and risk of unexplained stillbirth in subsequent pregnancy. Lancet 2003;362:1779–84.
20. Wood SL, Chen S, Ross S, Sauve R. The risk of unexplained antepartum stillbirth in second pregnancies following caesarean section in the first pregnancy. BJOG. 2008;115:726–31.