American Journal of Obstetrics and Gynecology (2006) 195, 633–42
www.ajog.org
EDITORS’ CHOICE
Single versus weekly courses of antenatal corticosteroids: Evaluation of safety and efficacy Ronald J. Wapner, MD,a,* Yoram Sorokin, MD,b Elizabeth A. Thom, PhD,c Francee Johnson, RN, BSN,d Donald J. Dudley, MD,e Catherine Y. Spong, MD,f Alan M. Peaceman, MD,g Kenneth J. Leveno, MD,h Margaret Harper, MD, MS,i Steve N. Caritis, MD,j Menachem Miodovnik, MD,k Brian Mercer, MD,l John M. Thorp, MD,m Atef Moawad, MD,n Mary Jo O’Sullivan, MD,o Susan Ramin, MD,p Marshall W. Carpenter, MD,q Dwight J. Rouse, MD,r Baha Sibai, MD,s Steven G. Gabbe, MD,t the National Institute of Child Health and Human Development Maternal Fetal Medicine Units Network Department of Obstetrics and Gynecology,a Drexel University College of Medicine, Philadelphia, PA; Department of Obstetrics and Gynecology,b Wayne State University, Detroit, MI; The Biostatistics Center,c George Washington University Biostatistics Center, Rockville, MD; Department of Obstetrics and Gynecology,d Ohio State University, Columbus, OH; Department of Obstetrics and Gynecology,e University of Utah, Salt Lake City, Utah; The National Institute of Child Health and Human Development,f Bethesda, MD; Department of Obstetrics and Gynecology,g Northwestern University, Chicago, IL; Department of Obstetrics and Gynecology,h University of Texas Southwestern Medical Center, Dallas, TX; Department of Obstetrics and Gynecology,i Wake Forest University School of Medicine, Winston-Salem, NC; Department of Obstetrics and Gynecology,j Magee Womens Hospital, Pittsburgh, PA; Departments of Obstetrics and Gynecology,k Columbia University, New York, NY, and University of Cincinnati, Cincinnati, OH; Department of Obstetrics and Gynecology,l Case Western Reserve University, Cleveland, OH; Department of Obstetrics and Gynecology,m University of North Carolina Chapel Hill, Chapel Hill, NC; Department of Obstetrics and Gynecology,n University of Chicago, Chicago, IL; Department of Obstetrics and Gynecology,o University of Miami, Miami, FL; Department of Obstetrics and Gynecology,p University of Texas Houston, Houston, TX; Department of Obstetrics and Gynecology,q Brown University, Providence, RI; Department of Obstetrics and Gynecology,r University of Alabama, Birmingham, AL; Department of Obstetrics and Gynecology,s University of Tennessee, Memphis, TN; and Department of Obstetrics and Gynecology,t Vanderbilt University, Nashville, TN Received for publication October 10, 2005; revised February 16, 2006; accepted March 21, 2006
Supported by grants from the National Institute of Child Health and Human Development (HD21410, HD21414, HD27869, HD27917, HD27905, HD27860, HD27861, HD27915, HD34122, HD34116, HD34208, HD34136, HD40500, HD40485, HD40544, M01-RR-000080, HD40545, HD40560, HD40512, HD36801). Presented at the annual meeting of the Society for Maternal-Fetal Medicine, 24th Annual Meeting, New Orleans, La, February, 2004. * Reprint requests: Ronald J. Wapner, MD, Department of Obstetrics and Gynecology, Columbia University College of Physicians and Surgeons, 622 W. 168th Street, PH 16-66, New York, NY 10032. E-mail:
[email protected] 0002-9378/$ - see front matter Ó 2006 Mosby, Inc. All rights reserved. doi:10.1016/j.ajog.2006.03.087
634
KEY WORDS Antenatal corticosteroids Prematurity
Wapner et al
Objective: The purpose of this study was to determine if weekly corticosteroids improve neonatal outcome without undue harm. Study design: Women 23 to 32 weeks receiving 1 course of corticosteroids 7 to 10 days prior were randomized to weekly betamethasone or placebo. Results: The study was terminated by the independent data and safety monitoring committee with 495 of the anticipated 2400 patients enrolled. There was no significant reduction in the composite primary morbidity outcome (8.0% vs 9.1%, P = .67). Repeated courses significantly reduced neonatal surfactant administration (P = .02), mechanical ventilation (P = .004), CPAP (P = .05), pneumothoraces (P = .03). There was no significant difference in mean birth weight or head circumference. The repeat group had a reduction in multiples of the birth weight median by gestational age (0.88 vs 0.91) (P = .01) and more neonates weighing less than the 10th percentile (23.7 vs 15.3%, P = .02). Significant weight reductions occurred for the group receiving R4 courses. Conclusion: Repeat antenatal corticosteroids significantly reduce specific neonatal morbidities but do not improve composite neonatal outcome. This is accompanied by reduction in birth weight and increase in small for gestational age infants. Ó 2006 Mosby, Inc. All rights reserved.
Antenatal corticosteroids improve neonatal morbidity and mortality and have been shown to reduce the risk of respiratory distress syndrome, intraventricular hemorrhage, and mortality by 50% or more.1-4 Despite documented efficacy and clinical use for over 20 years, unanswered questions remain about the duration of the effect, its potential diminution with time, and the usefulness of retreatment. Regardless of the absence of data, many practitioners, in an attempt to ensure that all neonates delivering preterm receive the maximal effect of treatment, readminister therapy when patients remain undelivered more than 7 days after an initial course. Recently, animal and retrospective human cohort studies have suggested harmful effects of repetitive doses of corticosteroids. These include reduced birth weight and head circumference.5-7 Only 1 randomized trial has prospectively evaluated the impact of repetitive doses on birth weight and did not confirm these observations.8 To determine the benefits and risks of repeated corticosteroid administration, the National Institute of Child Health and Human Development (NICHD) Maternal Fetal Medicine Units (MFMU) Network performed a randomized clinical trial to test the hypothesis that when compared with a single course, weekly courses of antenatal corticosteroids would improve neonatal outcome without causing undue harm.
Material and methods This study was a randomized, double-masked, placebocontrolled, multicenter clinical trial performed by 18 centers of the NICHD MFMU Network. The trial was approved by the institutional review boards at all centers and all mothers gave informed consent to participate. Pregnant women with intact membranes between 23 weeks 0 days and 31 weeks 6 days were eligible if they had
received a single full course of betamethasone or dexamethasone between 7 and 10 days earlier, and were at high risk for spontaneous preterm birth, or had the diagnosis of placenta previa or chronic abruption. Exclusions were: preterm premature rupture of the membranes (PPROM) prior to randomization, confirmed fetal lung maturity, chorioamnionitis, a major fetal anomaly, nonreassuring fetal status, systemic corticosteroid use during the current pregnancy, or insulin-dependent diabetes. Gestational age was determined from the last menstrual period provided that ultrasonography confirmed the estimate. When there was discordance, the duration of gestation at randomization was determined from the first sonogram performed. Numbered kits were prepared using randomization sequences created by the independent data coordinating center. The randomization sequences, stratified by clinical center, type of qualifying course, and whether the patient was an inpatient or outpatient, were generated using the urn design. This method minimizes the degree of imbalance in the number of patients assigned to each treatment by increasing the probability of assignment to the treatment that has been selected least often previously.9 At randomization, the patient was assigned to the next sequentially numbered kit. Women were randomized either to weekly courses of betamethasone or an identical-appearing placebo prepared by a centralized research pharmacy. Each course consisted of 2 injections of betamethasone 12 mg (as 6 mg betamethasone sodium phosphate and 6 mg betamethasone acetate) repeated once in 24 hours or matching placebo. Initially, patients received courses until birth or 33 weeks 6 days’ gestation, whichever was sooner. After 67 patients had been enrolled, the number of courses (not including the qualifying course) was limited to 4 because of difficulty in recruitment and published literature suggesting possible harmful effects of multiple courses.5-7
Wapner et al Table I
635
Demographic characteristics of randomized patients by treatment group
Indication for initial course of steroid (No. [%]) Previous spontaneous preterm birth Twin gestation Preterm labor Cervical cerclage Placenta previa or abruption Gestational age at randomization (wks) (mean G SD) Patient age at randomization (yr) (mean G SD) Predominant race (No. [%]) African American Caucasian Other Marital status (No. [%]) Married Divorced Never married Total years of school (mean G SD) Smoked during this pregnancy (No. [%]) Nulliparity (No. [%])
Maternal data were collected at the time of each injection, at delivery, and at discharge. Neonates were followed through discharge or 120 days after birth, whichever occurred first. Using a standardized technique, study personnel measured neonatal length, head and upper arm circumference within 72 hours of life. Cranial ultrasounds were performed on all neonates before 14 days of age. For those born less than 33 weeks’ gestation, ultrasounds were also done at discharge. The primary outcome was a composite endpoint consisting of the presence of any of the following: (1) severe respiratory distress syndrome (RDS), defined as clinical features of RDS with the need for oxygen and respiratory support from 6 to 24 hours or more of age, an abnormal chest x-ray, and either administration of a full course of surfactant or a fraction of inspired oxygen (FiO2) of at least 60%; (2) grade III or IV intraventricular hemorrhage; (3) periventricular leukomalacia; (4) chronic lung disease, defined as the need for supplemental oxygen at 36 weeks’ corrected age in infants born before 34 weeks; or (5) stillbirth or neonatal death. Deidentified charts of all infants who died or were admitted to the neonatal intensive care unit underwent blinded central review by 2 neonatologists and head ultrasounds were read centrally by a panel of blinded radiologists.
Repeat (n = 252)
Placebo (n = 243)
119 (47.2) 48 (19.0) 159 (63.1) 33 (13.1) 31 (12.3) 28.0 G 2.4 25.8 G 6.0
115 (47.3) 53 (21.8) 171 (70.4) 38 (15.6) 21 (8.6) 28.1 G 2.3 25.9 G 5.9
107 (42.5) 81 (32.1) 64 (25.4)
106 (43.6) 75 (30.9) 62 (25.5)
148 (58.7) 23 (9.1) 81 (32.1) 11.9 G 2.4 57 (22.6) 30 (11.9)
131 (53.9) 18 (7.4) 94 (38.7) 12.0 G 2.6 54 (22.2) 30 (12.3)
Those delivering after this gestational age were expected to have an incidence of the primary outcome of 2% to 4%. Hence, for all randomized patients a primary outcome rate of 11.5% was anticipated for patients assigned to placebo. With 80% power and a type I error rate of 5% (2-sided), detection of a 30% reduction for patients assigned to repeat steroids required a sample size of 1200 patients in each group. The primary safety outcomes to be evaluated were birth weight and head circumference. To account for the dependence of birth weight on gestational age, it was prespecified that birth weights would be converted to multiples of the gestational age specific median, using the standards of Alexander.10 It was estimated that a total of 500 patients would provide 90% power with a 5% type I error rate to detect a mean difference in birth weight of 150 g and at least 80% power to detect a 5% difference in birth weight multiples of the median (MOMs), assuming a standard deviation of 20%. A previous study performed by the participating centers of the MFMU Network indicated that the standard deviation for head circumference in this population would be 2.5 cm. It was estimated that a sample size of 500 patients would yield at least 90% power to detect a 0.75 cm difference in head circumference.
Interim analysis Sample size Previous network studies and an internal feasibility study demonstrated that approximately 25% of patients meeting entry criteria would deliver by 32 weeks’ gestation and have a primary outcome rate of 30% to 40%.
The trial was monitored by an external data and safety monitoring committee (DSMC) appointed by the NICHD. For interim analysis of the primary endpoint, the group sequential method of Lan and DeMets with a generalization of the O’Brien-Fleming boundary as the
636
Wapner et al
Figure
Outline of Patient Enrollment and Randomization.
spending function for the type I error was chosen to determine whether there was a significant difference.11
Statistical analysis Study outcomes were evaluated on an intent-to-treat basis. Categorical variables were compared using the chi-square or Fisher exact tests. For categorical infant
outcomes, including the primary outcome, the statistical unit was pregnancy rather than infant, because of the correlation between twins. A pregnancy was credited with an outcome if either or both twins experienced that outcome. Continuous variables were compared using the Wilcoxon rank sum test for maternal data and the Wei-Lachin procedure,12 an extension of the Wilcoxon rank sum test adjusting for the correlation between
Wapner et al Table II
637
Maternal/obstetric outcome
Outcome
Repeat (n = 250)
Placebo (n = 242)
Any side effects* (No. [%]) Bruising (No. [%]) Pain at injection site (No. [%]) Lump at injection site (No. [%]) GI upset (No. [%]) Insomnia (No. [%]) Contractions (No. [%]) Cushingoid appearance (No. [%]) Gestational age at delivery (wk) (mean G SD) Latency from randomization to delivery (d) (mean G SD) Clinical chorioamnionitis (No. [%]) Postpartum endometritis (No. [%]) Abnormal 1 hour OGTTy (No. [%]) Preeclampsia/gestational hypertension (No. [%]) Preterm premature rupture of membranes (No. [%]) Cesarean section (No. [%]) Preterm delivery !37 weeks (No. [%]) Preterm delivery !32 weeks (No. [%])
68 (27.0) 13 (5.2) 28 (11.1) 0 (0.0) 6 (2.4) 11 (4.4) 1 (0.4) 3 (1.2) 34.8 G 3.8 47.4 G 28.9
135 (55.6) 33 (13.6) 96 (39.5) 14 (5.8) 17 (7.0) 5 (2.1) 5 (2.1) 0 (0.0) 34.8 G 3.9 47.0 G 27.1
8 6 50 15 37 93 157 60
6 10 37 8 35 91 157 52
(3.2) (2.4) (26.9) (6.0) (14.8) (37.2) (62.8) (24.0)
(2.5) (4.1) (21.5) (3.3) (14.5) (37.6) (64.9) (21.5)
Relative risk (95% CI)
P value ! .001
0.38 0.28 0.00 0.34 2.12 0.19
(0.21-0.70) (0.19-0.41) (0.00-0.26) (0.14-0.85) (0.75-6.02) (0.02-1.64) .88 .85
1.29 0.58 1.25 1.82 1.02 0.99 0.97 1.12
(0.45-3.66) (0.21-1.57) (0.86-1.81) (0.78-4.20) (0.67-1.57) (0.79-1.24) (0.85-1.11) (0.81-1.55)
.63 .28 .24 .16 .92 .93 .63 .51
* Side effects were assessed in the full randomized cohort (252 in the repeat group; 243 in the placeo group). y Three hundred fifty-eight patients had oral 1 hour GTT administration post randomization.
twins, for infants. All P values are 2-sided without adjustment for multiple testing. Before starting the study, it was decided that women delivering very prematurely and those receiving multiple repetitive courses of steroids would be of particular interest. Thus, analyses of outcomes stratified by gestational age and number of courses were planned.
Results Recruitment began in March, 2000. At the second interim analysis, the DSMC recommended that enrollment be halted because of a tendency towards decreased birth weight in the repeat steroid group without any evident reduction in the primary morbidity outcome and also because of difficulties in recruitment. For this interim analysis, the primary outcome was evaluated in a cohort of 282 women randomized before a given date such that final infant outcomes were available. Emerging literature on the safety of repeat courses of antenatal steroids from cohort and animal studies was also considered. All women in the trial at the time of the decision to stop in April 2003 were allowed to complete their assigned courses.
Patient population Table I describes the demographic characteristics of patients participating in this study. Figure outlines patient recruitment; 495 patients were randomized, with 252 in the repeat steroid group and 243 in the placebo group. There were no significant differences in demographic
parameters. The number of study courses were equally distributed between the treatment groups. A total of 63.4% of patients received 4 or more study courses. The total number of courses received was related to the latency from the first study dose to delivery. The group receiving 1 study course remained in utero a median of 5 days, 2 courses 13 days, 3 courses 25 days, and 4 or more courses had a median latency period of 59 days. One hundred twelve patients (22.6%) delivered at less than 32 weeks’ gestation, with 60 randomized to repeated corticosteroids and 52 to placebo. In this cohort, the repeated steroid group was similar to the controls in mean gestational age at delivery (29.4 vs 28.8 weeks), gestational age at randomization (27.1 vs 26.7 weeks), latency period to delivery (15.4 vs 14.6 days), percent male sex, percent twin gestations, and total number of study courses (median = 2 for each group).
Maternal outcomes There were no significant differences between the 2 groups in gestational age at delivery (34.8 weeks for each group), time from randomization until delivery (47.4 vs 47.0 days), or the occurrence of preterm premature rupture of the membranes, preterm delivery, glucose intolerance, chorioamnionitis, postpartum endometritis, or preeclampsia (Table II). There was no difference in post-injection contractions or maternal perceived fetal movement. Maternal side effects were less common overall with repeated steroids (27.0 vs 55.6%, P ! .0001). Significant reductions in bruising, pain at the injection site, lumps at the injection site, and gastrointestinal upset
638
Wapner et al
Table III
Neonatal outcomes for all patients
Neonatal outcome
Repeat (n = 250)
Placebo (n = 242)
Relative risk (95% CI)
P value
CompositedSevere RDS, IVH III-IV, PVL, CLD, perinatal death (No. [%]) Perinatal death (No. [%]) IVH III-IV/PVL* (No. [%]) Severe RDS (No. [%]) CLD (No. [%]) Pulmonary outcomes RDS (No. [%]) On ventilator support (No. [%]) Use of surfactant (No. [%]) Any CPAP (No. [%]) BPD (No. [%]) Pneumothorax (No. [%]) Neurologic outcomes IVH* (No. [%]) Seizures (No. [%]) Any ROP (No. [%]) Cardiovascular outcomes PDA (No. [%]) Pressor or volume support (No. [%]) Infections Sepsis (No. [%]) Pneumonia (No. [%]) Gastrointestinal outcomes NEC (No. [%])
20 (8.0)
22 (9.1)
0.88 (0.49-1.57)
.67
3 0 6 14
(1.2) (0) (2.4) (5.6)
6 2 10 15
(2.5) (0.87) (4.1) (6.2)
0.48 0.00 0.58 0.90
(0.12-1.91) (0.00-1.93) (0.21-1.57) (0.45-1.83)
.33 .50 .28 .78
24 36 29 44 16 1
(9.6) (14.4) (11.6) (17.6) (6.4) (0.40)
32 60 46 60 26 7
(13.2) (24.8) (19.0) (24.8) (10.7) (2.9)
0.73 0.58 0.61 0.71 0.60 0.14
(0.44-1.20) (0.40-0.84) (0.40-0.94) (0.50-1.00) (0.33-1.08) (0.01-0.85)
.21 .004 .02 .05 .08 .03
15 (6.5) 1 (0.4) 15 (6.0)
18 (7.8) 3 (1.2) 20 (8.3)
0.83 (0.43-1.61) 0.32 (0.03-3.08) 0.73 (0.38-1.38)
.59 .37 .33
7 (2.8) 17 (6.8)
14 (5.8) 31 (12.8)
0.48 (0.20-1.18) 0.53 (0.30-0.93)
.10 .02
11 (4.4) 10 (4.0)
18 (7.4) 9 (3.7)
0.59 (0.29-1.23) 1.08 (0.44-2.60)
.15 .87
10 (4.0)
11 (4.5)
0.88 (0.38-2.03)
.76
* Ultrasounds were available for 230 patients in the repeat group and 230 patients in the placebo group.
were seen with repeated steroids. This group had 3 cases of maternal Cushingoid appearance (1.2%).
Efficacy There was no significant reduction in the occurrence of the primary outcome with repeated steroids (Table III). However, all infant morbidities were decreased reaching significance for surfactant use, the need for mechanical ventilation, the need for pressure or volume support, and the occurrence of a pneumothorax. There was no elevation in neonatal systolic or diastolic blood pressure in the repeated steroids group. In the cohort of neonates delivering less than 32 weeks’ gestation, the repeat steroid group had a reduced incidence of the primary outcome (Table IV). This difference was not statistically different. Significant improvements in specific pulmonary and cardiovascular outcomes were seen with repeated steroids in this subpopulation.
Neonatal anthropometric outcomes Table V presents neonatal measurements, including birth weight and head circumference, overall and for singletons. Fetal growth was also evaluated based on number of courses of steroids received (0 to 3 vs 4 or more). There was a 95 g reduction in birth weight (P = .09) in
the repeat steroid group, which was significant when birth weights were analyzed as gestational age specific multiples of the median (P = .01). When the birth weight was evaluated for those receiving 0 to 3 courses and those receiving 4 or more courses (not including the qualifying course), there was a significant reduction only in the 4 or more course group. Similar results were obtained for length at birth, both when expressed as an absolute measurement and as a multiple of the gestational age specific mean from published standards.13 Ponderal index was not significantly different between the 2 treatment groups, overall or by subgroup. There was no difference in the absolute head circumferences or gestational age specific multiples of the median either overall or in the 4 or more course group.14 Birth weight below the 10th and 5th percentiles for gestational age was significantly more common with repeated steroids. This again was predominantly caused by the group of infants exposed to 4 or more courses. The results were similar when twins and singletons were analyzed separately.
Comment Our trial evaluated both the efficacy and safety of repetitive courses of antenatal corticosteroids. We found a reduction in birth weight and an increase in small for
Wapner et al Table IV
639
Neonatal outcomes for deliveries less than 32 weeks
Neonatal outcome
Repeat (n = 60)
Placebo (n = 52)
Relative risk (95% CI)
P value
CompositedSevere RDS, IVH III-IV, PVL, CLD, perinatal death (No. [%]) Perinatal death (No. [%]) IVH III-IV, PVL (No. [%]) Severe RDS (No. [%]) CLD (No. [%]) Pulmonary outcomes RDS (No. [%]) On ventilator support (No. [%]) Use of surfactant (No. [%]) Any CPAP (No. [%]) BPD (No. [%]) Pneumothorax (No. [%]) Neurologic outcomes IVH (No. [%]) Seizures (No. [%]) Any ROP (No. [%]) Cardiovascular outcomes PDA (No. [%]) Pressor or volume support (No. [%]) Infections Sepsis (No. [%]) Pneumonia (No. [%]) Gastrointestinal outcomes NEC (No. [%])
14 (23.3)
20 (38.5)
0.61 (0.34-1.08)
.08
1 0 5 11
(1.7) (0) (8.3) (18.3)
5 2 10 14
(9.6) (3.8) (19.2) (26.9)
0.17 0.00 0.43 0.68
(0.02-1.44) (0.00-1.66) (0.16-1.19) (0.34-1.37)
.09 .21 .09 .28
21 28 23 32 13 0
(35.0) (46.7) (38.3) (53.3) (21.7) (0)
23 39 33 41 25 4
(44.2) (75.0) (63.5) (78.8) (48.1) (7.7)
0.79 0.62 0.60 0.68 0.45 0.00
(0.50-1.25) (0.46-0.85) (0.41-0.88) (0.51-0.89) (0.26-0.79) (0.00-0.84)
.32 .002 .008 .005 .003 .04
10 (16.7) 0 (0) 15 (25.0)
15 (28.8) 3 (5.8) 20 (38.5)
0.58 (0.28-1.17) 0.00 (0.00-1.13) 0.65 (0.37-1.13)
.12 .10 .13
5 (8.3) 10 (16.7)
13 (25.0) 20 (38.5)
0.33 (0.13-0.87) 0.43 (0.22-0.84)
.02 .009
10 (16.7) 6 (10.0)
18 (34.6) 7 (13.5)
0.48 (0.24-0.95) 0.74 (0.27-2.07)
.03 .57
9 (15.0)
6 (11.5)
1.3 (0.50-3.41)
.59
gestational age neonates among fetuses exposed to repeat courses of antenatal steroids. We also demonstrated significant improvements in specific pulmonary and cardiovascular outcomes; especially among infants delivering very preterm. The mechanisms underlying the altered growth and its clinical significance are unknown. While the weight and length reduction are of concern, no diminution in other growth parameters including head circumference, arm circumference, or ponderal index were seen. Since this reduction is only seen among infants exposed to 4 or more courses, a threshold exposure leading to a metabolic change may be required before weight reduction occurs. Alternatively, a lower exposure may have an impact on growth but would not be detectable because pregnancies exposed to 3 or fewer courses delivered relatively soon after treatment. Animal studies of repetitive doses of maternal steroids suggest a dose-related weight reduction with growth temporarily arrested or slowed with each course.15 These studies also demonstrate catch-up growth if the fetus remains in utero. There is little information in humans on the long-term consequences of growth delay induced by repetitive exposure to corticosteroids. In a retrospective cohort study, French et al demonstrated catch-up growth at 3 years of age and no increased risk of neurologic sequelae in babies born less than the 10th percentile following in
utero exposure to 3 or more courses.6 Other retrospective studies have not been as reassuring,5 but may be biased by confounders since pregnancies remaining in utero for 3 or 4 weeks after an episode of preterm labor may be different than those of a similar gestational age delivering rapidly. Hence, the long-term significance of our observations will have to await the planned 24-month postdelivery physical and neurodevelopmental analysis. We were unable to demonstrate that weekly repetition of corticosteroids was effective in reducing overall infant morbidity in exposed pregnancies. However, our findings are of low power because of the termination of this trial before the final sample size was accrued. This may have led to the incorrect rejection of certain observed effects as nonsignificant. This decision was difficult and relied on the independent DSMCs evaluation of the risk/benefit ratio after evaluating preliminary data as well as information accumulating in the literature. The committee also considered that recruitment was scheduled to take another 6 years since the primary outcome rate was lower than expected. A detailed description of this decision is discussed elsewhere.16 Despite our reduced sample size, reductions in pulmonary and cardiovascular morbidities were seen, especially among the subgroup of infants delivering less than 32 weeks’ gestation. Neonates of women exposed to repeat steroids had a significant reduction in the need for
640 Table V
Wapner et al Anthropometric measurements by study group and number of courses All infants
Singletons
Outcome
Repeat (N = 296)
Placebo (N = 294)
P value
Repeat (N = 202)
Placebo (N = 190)
P value
Birth weight (gm) Birth weight MOMs* Head circumference (cm) Length (cm) Length MOMsy Ponderal Index (kg/m3) Arm circumference (cm) Birth weight ! 10th percentile – no. (%) Birth weight ! 5th percentile – no. (%) Head circumference ! 10th percentile – no. (%) Birth weight (MOMs) by number of courses 1 study course, N = 73 2 study courses, N = 51 3 study courses, N = 98 R4 courses, N = 368
2194.3 G 0.88 G 30.6 G 44.2 G 0.98 G 24.4 G 8.8 G 59 (23.7)
2289.6 G 0.91 G 30.8 G 44.7 G 0.99 G 24.5 G 8.9 G 37 (15.3)
.09 .01 .25 .08 .10 .88 .62 .02
2365.1 G 0.91 G 31.1 G 45.1 G 0.99 G 24.8 G 9.1 G 39 (19.3)
2468.6 G 0.95 G 31.3 G 45.7 G 1.0 G 24.9 G 9.3 G 16 (8.4)
.19 .003 .31 .12 .01 .87 .46 .002
762.3 0.16 3.1 4.6 0.06 3.6 1.9
791.8 0.15 3.3 5.1 0.05 3.5 1.9
816.8 0.15 3.3 5.0 0.05 3.5 1.9
32 (12.9)
21 (8.7)
.14
21 (10.4)
9 (4.7)
.04
25 (10.4)
20 (8.7)
.53
16 (8.1)
12 (6.6)
.58
0.91 0.90 0.95 0.86
G G G G
0.12 0.18 0.20 0.15
0.94 0.96 0.88 0.90
G G G G
0.16 0.14 0.16 0.14
.16 .11 .26 .006
1-3 study courses Birth weight (gm) Head circumference (cm) Length (cm) Length MOMs Ponderal Index (kg/m3) Arm circumference (cm) Birth weight ! 10th percentile – no. (%) Birth weight ! 5th percentile – no. (%) Head circumference ! 10th percentile – no. (%)
773.9 0.16 3.1 4.6 0.05 3.7 1.8
4C study courses
Repeat (N = 105)
Placebo (N = 117)
P value
Repeat (N = 191)
Placebo (N = 177)
P value
1820.7 G 29.0 G 42.2 G 1.00 G 22.8 G 8.1 G 9 (10.3)
1879.5 G 29.0 G 42.1 G 0.99 G 23.6 G 8.0 G 10 (10.8)
.57 .74 .97 .13 .06 .77 .93
2399.6 G 31.6 G 45.4 G 0.97 G 25.3 G 9.2 G 50 (30.9)
2560.6 G 32.1 G 46.5 G 0.99 G 25.1 G 9.5 G 27 (18.1)
.01 .11 .006 .004 .51 .14 .009
810.4 3.4 5.4 0.06 3.6 1.9
851.1 3.8 5.9 0.05 2.7 1.9
650.6 2.4 3.6 0.05 3.4 1.8
617.0 2.1 3.4 0.06 3.8 1.6
4 (4.6)
8 (8.6)
.28
28 (17.3)
13 (8.7)
.03
7 (8.4)
10 (11.6)
.49
18 (11.4)
10 (6.9)
.18
Data expressed as mean G standard deviation unless indicated otherwise. N = number of neonates. * Multiples of the gestational age specific median birthweight. y Multiples of the gestational age specific mean length.
ventilatory support (RR 0.58; CI 0.40-0.84; P = .004), surfactant use (RR 0.61; CI 0.40-0.94; P = .02), occurrence of a pneumothrorax (RR 0.14; CI 0.01-0.85; P = .03), and need for pressor or volume support (RR 0.53; CI 0.30-0.93; P = .02). A number of other morbidities, including components of the composite, were also reduced but a larger sample size would be required to confirm this. A benefit of repeat steroids to very preterm neonates has previously been suggested by Guinn et al.8 In their randomized trial of single versus repetitive courses, a reduction in the composite outcome of severe RDS, BPD, sepsis, severe IVH/PVL, NEC, and mortality was found in neonates delivering less than 28 weeks. In addition, a
statistically significant reduction in the frequency of severe RDS was seen in the entire population. These findings support animal studies that demonstrate significant improvements in pulmonary function in preterm sheep exposed to multiple weekly courses of corticosteroids compared with only a single course.7,15 It is not possible from our data to determine whether the improved outcomes are related to a cumulative effect of the multiple courses or to a reduction in the length of time from the last course to delivery. It is probable that the improvement results from both. In vitro studies demonstrate that the production of surfactant diminishes with time after the removal of steroids from lung cell culture,17 suggesting that a component of the
Wapner et al improved pulmonary performance occurs because of continuous steroid-induced surfactant production. Alternatively, antenatal corticosteroids also induce structural maturational changes in the lung parenchyma,15 which are unlikely to be reversible and may advance with each course. Our findings suggest that routine weekly repetition of steroids to all women at high risk for preterm birth, in an attempt to assure maximum exposure to those that deliver early cannot be justified, may be harmful, and requires treatment of many infants who receive little or no benefit. For example, in our cohort, 77% of the pregnancies eligible for repetitive steroids delivered at 32 weeks and beyond. Among those in the placebo arm delivering beyond 32 weeks, morbidity was exceedingly infrequent. Death and chronic lung disease each occurred in 0.5% of neonates and no cases of IVH grades III/IV or PVL occurred. A benefit of retreatment with antenatal corticosteroids appears to exist for infants delivering less than 32 weeks’ gestation. A trial of repeating steroids every 14 days is presently underway and may identify an efficacious alternative with reduced fetal exposure.18 Alternatively, ‘‘rescue’’ therapy in which retreatment is administered only when preterm birth seems inevitable is another approach. However, this needs further evaluation of both its efficacy and feasibility before it can be recommended.
Acknowledgments Other members of the National Institute of Child Health and Human Development Maternal-Fetal Medicine Units Network are as follows: Drexel University College of Medicine: M. DiVito, A. Sciscione, V. Berghella, P. Trauffer, M. Pollock, M. Talucci; Wayne State University: M. Dombrowski, G. Norman, A. Millinder, C. Sudz, D. Driscoll; George Washington University, Biostatistics Center: L. Mele, F. Galbis-Reig, A. Das, L. Leuchtenburg, D. Johnson; Ohio State University: J. Iams, M. Landon, S. Meadows, P. Shubert; University of Utah: M. Varner, K. Anderson, A. Guzman, A. Crowley, M. Fuller; The National Institute of Child Health and Human Development: D. McNellis, K. Howell, S. Pagliaro; Northwestern University: G. Mallett; University of Texas, Southwestern Medical Center: D. Weightman, L. Fay-Randall, P. Mesa; Wake Forest University: P. Meis, M. Swain, C. Moorefield; Magee Womens Hospital: T. Kamon, K. Lain, M. Cotroneo; Columbia University: V. Pemberton, S. Bousleiman, F. Malone, M. D’Alton; Case Western Reserve University: P. Catalano, C. Milluzzi, C. Santori; University of North Carolina, Chapel Hill: K. Moise, K. Dorman; University of Chicago: P. Jones, G. Mallett; University of Miami; D. Martin, F. Doyle; University of Texas
641 Health Science Center at Houston: L. Gilstrap, M. C. Day; Brown University: D. Allard, J. Tillinghast; University of Alabama: A. Northen, K. Bailey, W. Andrews; University of Cincinnati: H. How, N. Elder, B. Alexander, W. Girdler; University of Tennessee: B. Mabie, R. Ramsey. In addition to the authors, the following subcommittee members participated in protocol development/ coordination between clinical research centers (Michelle DiVito, MSN) and protocol/data management and statistical analysis (Lisa Mele, ScM). Dorothy Bulas, MD, Children’s National Medical Center, JoAnn Seibert, MD, University of Arkansas, and Michael DiPietro, MD, University of Michigan, centrally reviewed all cranial ultrasounds. Susan McCune, MD, Children’s National Medical Center, and Rosemary Higgins, MD, NICHD, reviewed neonatal hospital records.
References 1. Crowley P. Prophylactic corticosteroids for preterm birth (Cochrane Review). The Cochrane Library, Issue 3. Oxford: Update Software; 2002. 2. Liggins GC, Howie RN. A controlled trial of antepartum glucocorticoid treatment for prevention of the respiratory distress syndrome in premature infants. Pediatrics 1972;50:515-25. 3. NIH Consensus Development Panel. Effect of corticosteroids for fetal maturation on perinatal outcomes. JAMA 1995;273:413-8. 4. NIH Consensus Development Conference Statement. Antenatal corticosteroids revisited: repeat courses. August 2000. 5. Murphy K, Aghajafari F. Single versus repetitive courses of corticosteroids: what do we know? Clin Obstet Gynecol 2003;46: 161-73. 6. French NP, Hagan R, Evans SF, Mullan A, Newnham JP. Repeated antenatal corticosteroids: size at birth and subsequent development. Am J Obstet Gynecol 1999;180:114-21. 7. Aghajafari F, Murphy KE, Matthews S, Ohlsson A, Amankwah K, Hannan M, et al. Repeated doses of antenatal corticosteroids in animals: a systematic review. Am J Obstet Gynecol 2002;186: 843-9. 8. Guinn DA, Atkinson MW, Sullivan L, Lee M, MacGregor S, Parilla BV, et al. Single vs weekly courses of antenatal corticosteroids for women at risk of preterm delivery: a randomized controlled trial. JAMA 2001;286:1581-7. 9. Wei LF, Lachin JM. Properties of urn randomization in clinical trials. Controlled Clin Trials 1988;9:345-64. 10. Alexander GR, Kogan MD, Himes JH. 1994-1996 U.S. singleton birth weight percentiles for gestational age by race, Hispanic origin, and gender. Matern Child Health J 1999;3:225-31. 11. Lan KKG, DeMets DL. Discrete sequential boundaries for clinical trials. Biometrika 1983;70:659-63. 12. Wei LJ, Lachin JM. Two-sample asymptotically distribution-free tests for incomplete multivariate observations. JASA 1984;79:653-61. 13. Usher R, McLean F. Intrauterine growth of live-born Caucasian infants at sea level: standards obtained from measurements in 7 dimensions of infants born between 25 and 44 weeks of gestation. J Pediatr 1969;74:901-10. 14. Lubchenco LO, Hansman C, Boyd E. Intrauterine growth in length and head circumference as estimated from live births at gestational ages from 26 to 42 weeks. Pediatrics 1966;37:403-8.
642 15. Jobe AH, Ikegami M. Fetal responses to glucocorticosteroids. In: Mendelson CR, editor. Endocrinology of the lung. Totawa: Humana Press; 2000. p. 45-57. 16. Thom EA, Klebanoff MA. Issues in clinical trial design: stopping a trial early and the large and simple trial. Am J Obstet Gynecol 2005;193:619-25.
Wapner et al 17. Tan RC, Ikegami M, Jobe AH, Yao LY, Possmayer F, Ballard PL, et al. Developmental and glucocorticoid regulation of surfactant protein mRNAs in preterm lambs. Am J Physiol 1999;277: L1142-8. 18. MACS Trial sponsored by Maternal Infant and Reproductive Health Units, University of Toronto, Toronto Canada.