OBJECTIVE: Determine whether the increased neonatal mortality following repeated courses of antenatal corticosteroids (ANCS), observed in the Thyrotropin-Releasing Hormone (TRH) Trial, was related to confounding maternal risk factors or specific preterm morbidities.
STUDY DESIGN: A post hoc analysis of 595 TRH trial neonates, 26 to 32 weeks' gestation, studied the association between ≥3 courses ANCS and mortality. Potential confounding maternal factors and preterm morbidities were evaluated using logistic regression and log likelihood modeling.
RESULTS: Mortality was 9.2% after ≥3 courses (13/141) vs. 4.8% after 1 or 2 courses (22/454). This association was not explained by maternal factors, or other common preterm morbidities. However, 15/141 infants receiving ≥3 courses (10.6%) had early severe lung disease (ESLD) with 10 deaths, compared to 16/454 of the 1- to 2-course infants (3.5%) with 7 deaths (odds ratio 3.5, p<0.001).
CONCLUSIONS: ESLD, but not maternal risk factors, was associated with increased mortality in preterm infants after ≥3 courses ANCS.
A single course of antenatal corticosteroids (ANCS), administered to women at risk of preterm delivery, has been shown to decrease the incidence of respiratory distress syndrome (RDS), intraventricular hemorrhage (IVH), and neonatal mortality.1,2,3The scientific basis of the use of ANCS has been reviewed.4 Use of ANCS for women threatening to deliver prematurely as well as the need for future research investigating “the short- and long-term benefits and risks of repeated administration of antenatal corticosteroids” was recommended by the National Institutes of Health (NIH) Consensus Development Panel in 1994.1 The beneficial effects of a single course of ANCS begin less than 24 hours after treatment, and persist for at least 7 to 10 days.4,5,6 However, because there are no clinical trial data to establish benefit beyond this interval, many obstetricians elect to treat women with a history of preterm labor with repeated courses every 7 to 10 days.
In 1995, Planer et al.7 conducted a survey of the 1420 members of the Society for Perinatal Obstetricians. Among the 68% of members who responded, 96% reported willingness to give more than one course of antenatal steroids, 50% would retreat even after preterm labor had stopped, and 58% would give a maximum of six or more courses of antenatal steroids. Only 6% of those obstetricians responding to the survey limited ANCS use to two courses.
Ikegami et al.,8 Jobe et al.,9,10 and Marinelli et al.11 have described the effect of repeated ANCS dosing in the fetal sheep model. Cumulative doses were associated with improvement in lung compliance, ventilatory efficiency, and maximal volume but also with symmetric growth retardation as well as alterations in postnatal cortisol, thyroid hormone, and catecholamine levels. In addition, numerous adverse effects of prolonged postnatal steroid use in preterm human infants have been described including growth failure, infection, systemic hypertension and left ventricular hypertrophy,11,12 adrenal suppression,13 intestinal perforation,14 and poor long-term neurodevelopmental outcome.15
In a post hoc analysis of preterm infants in the North American Thyrotropin-Releasing Hormone (TRH) Trial,16 we found no benefit from additional ANCS exposure. Moreover, multiple courses of ANCS were associated with a decrease in birth weight, increased postnatal adrenal suppression, and a significant increase in neonatal mortality. There was also a significant increase in the incidence of the combined outcome of death or chronic lung disease (CLD) at 36 weeks' postmenstrual age. However, all of the difference in outcome was related to higher mortality and not chronic lung disease among survivors. The current analysis examines the potential confounding maternal risk factors and specific morbidities of prematurity associated with mortality in this group of infants.
The effects of repeated courses of ANCS on mortality were evaluated by a post hoc analysis of 595 infants 26 to 32 weeks' gestation born to mothers enrolled in the North American TRH Trial for prevention of lung disease in preterm infants. This randomized controlled trial (1992 to 1996) involved 13 centers and results have been reported.17 This trial was approved by the institutional review boards of all participating study centers, and written informed consent was obtained from all women participating in the trial. Eligible women were in preterm labor between 24 and 30 weeks' gestation and were excluded if there was evidence of bleeding, infection including clinical chorioamnionitis, or hypertension. Women with diabetes, premature rupture of membranes, or carrying multiple gestations were not excluded.
All women received a “course” of antenatal steroids at the time of enrollment. This was defined as either two doses of 12 mg intramuscular betamethasone given 24 hours apart, or four doses of 6 mg dexamethasone given every 6 hours. Less than 10% of mothers received treatment with dexamethasone, and these infants were combined with the betamethasone group for all analyses. Treatment of women with ANCS before 24 weeks' gestation was a violation of study protocol. Women were randomized to receive TRH or placebo concurrently with a course of ANCS, usually the first course. Repeated courses of ANCS, as defined above, were given at the discretion of the individual attending obstetricians. The timing of these courses (at weekly or greater than 1-week intervals) was also at the obstetrician's discretion. After birth, infants were treated by protocol with surfactant. Extensive maternal, obstetrical, and neonatal outcome data was collected and entered into a computerized database.
Mortality before 36 weeks' postmenstrual age was compared for live-born neonates 26 to 32 weeks' gestation who had received 1 or 2 courses of ANCS vs. ≥3 courses. Only infants ≥26 weeks' gestation were included in the analysis because they would have had the opportunity to receive ≥3 courses of ANCS by study protocol. Because there was no association between TRH and either neonatal outcome or the number of courses of ANCS received, these two treatment arms were combined for the current analysis. No adjustment for center difference effects was necessary because the frequency of retreatment with ANCS was similar among study centers.
Multiple potential confounding maternal risk factors for mortality were assessed including maternal race, smoking, alcohol or substance abuse, maternal illness, complications of pregnancy, and tocolytic use. In addition, morbidities of prematurity, which might explain the increased neonatal mortality following ≥3 courses of ANCS, were evaluated. RDS was defined as a requirement for supplemental oxygen and either the need for assisted ventilation for >48 hours after birth or radiologic findings consistent with RDS. The severity of respiratory failure was assessed using a ventilatory score of mean airway pressure times inspired oxygen concentration (FiO2). This type of scoring system was previously used by French et al.18 in a study of neonatal effects of multiple courses of antenatal steroids. In the current analysis, early severe lung disease (ESLD) was defined as either death from respiratory failure within 6 hours of life or severe respiratory failure requiring a ventilatory score (mean airway pressure×FiO2) >5 at 6, 12, and 24 hours of life. Initial chest radiograph reports on neonates with ESLD were reviewed when available to rule out other potential sources of severe respiratory failure. Need for pharmacologic blood pressure support and incidence of infection, IVH, patent ductus arteriosus (PDA), and necrotizing enterocolitis (NEC) were also reviewed.
Association between the presence of maternal risk factors and mortality was determined using Fisher's exact test, and between infant characteristics such as gestational age (GA), using t-tests. Associations between ≥3 courses ANCS and maternal and infant characteristics were similarly explored. Best subset logistic regression to predict mortality was used. The effect of adding ANCS was examined using a likelihood ratio test. Log linear modeling was performed to evaluate the associations between ANCS, ESLD and mortality. Finally, Kaplan-Meier curves were used to compare survival times between ≥3 courses ANCS and 1 or 2 courses ANCS.
In the total TRH study population of 1123 neonates, 55% received 1 course of ANCS, 19% received 2 courses, and 26% received ≥3 courses with 24 infants (2% of the total) receiving 8 or more ANCS. Among infants 26 to 32 weeks' gestation in this analysis, 454 (76%) received 1 or 2 courses ANCS and 141 (24%) received ≥3 courses (maximum of 8 courses).
Association Between Mortality and ANCS Exposure
Thirteen of 141 ≥3 ANCS infants died (9.2%) compared to 22 of 454 (4.8%) receiving 1 or 2 courses ANCS. Table 1 shows the association between multiple courses of ANCS and mortality of neonates 26 to 32 weeks after adjusting for gestational age and multiple gestation by logistic regression analysis. Infants who had received ≥3 ANCS had an increased risk of death (odds ratio 3.2, p=0.003) compared with 1 or 2 ANCS. The addition of an interaction term to determine whether the effect of multiple courses of ANCS was different for singleton versus multiple gestations was not statistically significant. Because of the uncertainty regarding persistence of beneficial effects of ANCS beyond 7 days after treatment, a term reflecting the interval from last dose of steroids to delivery was added to the model. There was no significant difference in mortality for infants delivering 7 to 13 days after their last dose of ANCS compared to those delivering within 1 to 6 days.
Association of ≥3 ANCS With Maternal Factors
Exposure to ≥3 courses ANCS was associated with a number of potentially confounding maternal and obstetrical factors. Neonates who had received ≥3 courses ANCS had an increased mean gestational age at delivery (29.2 vs. 28.5 weeks, p<0.001) and were more likely to be from multiple gestation pregnancies (31% for multiples vs. 22% for singletons, p=0.02), diabetic mothers (53% vs. 22%, p=0.001), or mothers treated with tocolytic agents including betamimetics (66% vs. 19%, p=0.001), magnesium sulfate (58% vs. 20%, p=0.001), or prostaglandin inhibitors (76% vs. 21%, p=0.001). Exposure to ≥3 courses ANCS was less common among mothers who smoked (17% vs. 26%, p=0.03), and among African American mothers (16% vs. 24%, p=0.01).
The association between potential confounding maternal risk factors and neonatal mortality was evaluated in univariate analysis (Table 2). Only tocolytic exposure, specifically betamimetics, magnesium sulfate, or prostaglandin inhibitors, and receiving ≥3 courses ANCS were associated with an increased mortality. However, tocolytic exposure was not significantly associated with mortality when added to the logistic regression model including ≥3 courses ANCS, gestational age, and multiple gestations shown in Table 1. Three or more courses of ANCS continued to be significantly associated with mortality in models adjusting for tocolytic exposure. Randomization to TRH versus placebo treatment was not associated with either neonatal mortality or the number of courses of ANCS received (data not shown). In addition, as expected, infants who died had a lower mean gestational age than survivors (27.7 vs. 28.7 weeks, p<0.001).
ANCS and Preterm Neonatal Morbidities
Common morbidities of preterm infants were evaluated to determine their possible contribution to the increased mortality seen with ≥3 courses of ANCS. Figure 1a shows the frequency of preterm morbidities for the total study population, with the frequency of neonatal morbidities among only the infants who died shown in Figure 1b. The incidences of nonrespiratory neonatal morbidities of prematurity was similar for ≥3 ANCS when compared to 1 or 2 ANCS. Although the overall incidence of RDS was not associated with number of courses of ANCS, ESLD was more common among neonates who had received more than 3 courses of ANCS (11% vs. 3.5% in the 1- or 2-course group; odds ratio 3.3, p<0.001) (Table 3). Ten of 13 (77%) of the neonates who died after receiving ≥3 courses of ANCS had severe ESLD, compared to only 7 of 22 (32%) of neonates who died after 1 or 2 courses. All infants with ESLD received early treatment with surfactant. Chest radiographs, when available, were consistent with severe RDS. A small number of these infants were also given the clinical diagnosis of pulmonary hypoplasia. Unfortunately, without lung pathology, it was impossible to confirm this diagnosis.
Figure 2 shows survival curves grouped by number of courses of ANCS. Survival time after birth was significantly shorter for the ≥3-course group (p=0.03 by Kaplan-Meier analysis). There were five infants in the ≥3 ANCS and three in the 1 or 2 ANCS who died of pulmonary causes in the first 6 hours of life. Four ≥3 ANCS and one of the 1 or 2 ANCS died in the delivery room of ventilatory failure and received the presumptive diagnosis of pulmonary hypoplasia. Unfortunately, postmortem exams were not available on these infants and the diagnosis cannot be confirmed by lung pathology.
Log Linear Modeling
A log linear model was fit to ANCS, ESLD, and mortality. The most parsimonious well-fitting model included an association between ESLD and mortality (p<0.001) and an association between ANCS and ESLD (p<0.001), but did not include a direct association between ANCS and mortality in the presence of these two associations. Table 3 shows the frequency of ESLD and mortality, stratified by the number of courses of ANCS. ESLD was associated with an increased risk of mortality, 67% in the ≥3-course group vs. 44% in the 1- or 2-course group. By log linear modeling, the increased incidence of ESLD explained the association between ≥3 courses of ANCS and mortality.
Although it is well established that a single course of ANCS decreases mortality and morbidity in the preterm infant, there is little information on the benefits and risks associated with multiple courses of ANCS. In this post hoc analysis of preterm infants enrolled in a multicentered clinical trial, we found that ≥3 courses of ANCS were associated with increased neonatal mortality after adjusting for gestational age and multiple gestations, with an adjusted odds ratio of 3.2. This increased mortality was not explained by maternal risk factors including diabetes, hypertension, substance abuse, prolonged rupture of membranes, chorioamnionitis, tocolytic use, or placenta abruptio or previa.
In the preterm neonate, three or more courses of ANCS were associated with an increased risk of ESLD, but not total RDS, PDA, IVH, NEC, or need for pharmacologic blood pressure support. Three or more courses of ANCS was no longer associated with increased mortality after adjusting for the effects of ESLD. In addition we found that neonatal outcome was similar for infants delivered 1 to 6 days vs. 7 to 13 days after their last dose of ANCS. Because it is difficult to distinguish severe RDS from pneumonia, we cannot exclude the possibility that some of this severe lung disease was due to pulmonary infection. However, we did not observe an increased rate of clinical sepsis with repeated courses of ANCS.
Because of a lack of autopsy data in this study, it is impossible to determine the exact lung pathology leading to increased ESLD and mortality following multiple courses of ANCS. It is possible that in some infants there is a detrimental effect of repeated early exposure to corticosteroid on lung development during gestation. Treatment of newborn rats with postnatal dexamethasone during the period of alveolarization resulted in lung hypoplasia and alveolar simplification, as well as an increased risk of hypoxia-induced pulmonary hypertension.19,20,21,22,23,24 Pratt et al.25 have described deleterious effects on growth in the rabbit.
In the preterm lamb model, however, multiple courses of ANCS conveyed additional pulmonary beneficial effects, with improved pulmonary function testing, but also were associated with growth retardation and increased fetal loss.8,9,10 It is clear that steroids have a maturational effect on the lung, leading to acute improvement in pulmonary function. With this maturation, though, comes a decrease in lung growth and alveolarization. The apparent discrepancy between our findings and the effects of repeated courses of antenatal steroids seen in the lamb model may be due to a species difference in this balance between the beneficial effects on maturation and the adverse effects on growth. It is also important to note that the majority of infants in our multiple-course group were exposed to steroids early in gestation, and for a prolonged period. This may make them more vulnerable to deleterious growth effects. Lastly, all infants exposed to multiple courses of steroids did not have severe lung disease. This may reflect a genetic or constitutional variability in susceptibility to the adverse effect of steroids on growth.
In a retrospective analysis of 477 singleton infants less than 33 weeks' gestation born in Western Australia between 1990 and 1992, multiple courses of ANCS were associated with a reduction in birth weight and head circumference.18 When comparing single to repeated courses of ANCS, no additional benefit was seen with respect to mortality, respiratory outcomes, and growth or severe disability at 3 years. In addition, although the power of that study was limited with only 40 infants receiving multiple courses of ANCS, there was a trend toward increased incidence of severe hyaline membrane disease defined as mean airway pressure×FiO2 >8 (15% for multiple courses vs. 9.9% for one course) and an increased incidence of severe chronic lung disease (19% for multiple courses vs. 11% for one course). This was true despite increased mean gestational age among neonates receiving multiple courses of ANCS. Vermillion et al.26 have also reported increased neonatal sepsis in infants with multiple courses who had prolonged rupture of membranes.
In contrast, Abassi et al.27 conducted a retrospective review of 369 singleton infants 24 to 34 weeks gestation. They reported decreased RDS among infants receiving greater than two courses of ANCS. However, the exact distribution of gestational ages in each group was not presented, and it would be unlikely, with current obstetrical practices, for a large number of 24 and 25 weeks gestation infants to have received more than two courses of ANCS. Inclusion of these extremely preterm infants, with nearly 100% incidence of RDS, could have increased the observed RDS rate in the 1- or 2-course group.
Although the associations described in this post hoc analysis are of concern, inferences regarding a causal relationship are purely speculative. It is possible that other unidentified prenatal factors that might make an obstetrician particularly concerned about a fetus would increase the use of multiple courses of ANCS. Women who received ≥3 courses were more likely to have tocolytic exposure, an independent predictor of mortality. However, in adjusted analysis, ≥3 courses of ANCS was a stronger predictor, and continued to be significant even after adjusting for tocolytic exposure. One would also expect that mothers with excellent access to medical care would be more likely to receive multiple courses of ANCS. However, this potential bias would be expected to lead to improved outcome with multiple courses of ANCS rather than increased mortality.
We conclude that randomized controlled trials are warranted to assess the safety and benefit of repeated courses of ANCS. This study adds support to the recent NIH consensus statement that repeated courses of ANCS should not be given outside of the structure of a clinical trial.
NIH Consensus Development Panel. Effect of corticosteroids for fetal maturation on perinatal outcomes JAMA 1995 273: 413–7
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
Crowley P . Antenatal corticosteroid therapy: a meta-analysis of the randomized trials, 1972 to 1994 Am J Obstet Gynecol 1995 173: 322–35
Ballard PL, Ballard RA . Scientific basis and therapeutic regimens for use of antenatal glucocorticoids Am J Obstet Gynecol 1995 173: 1 254–62
Polk DH, Ikegami M, Jobe AH, Sly P, Kohan R, Newnham J . Preterm lung function after retreatment with antenatal betamethasone in preterm lambs Am J Obstet Gynecol 1997 176: 308–15
Ikegami M, Polk DH, Jobe AH . Minimum interval from fetal betamethasone treatment to postnatal lung responses in preterm lambs Am J Obstet Gynecol 1996 174: 1408–13
Planer BC, Ballard RA, Ballard PL et al. Antenatal corticosteroid (ANCS) use in preterm labor in the USA Pediatr Res 1996 39: 110A
Ikegami M, Jobe AH, Newnham J, Polk DH, Willet KE, Sly P . Repetitive prenatal glucocorticoids improve lung function and decrease growth in preterm lambs Am J Respir Crit Care Med 1997 156: 178–84
Jobe AH, Wada N, Berry LM, Ikegami M, Ervin MG . Single and repetitive maternal glucocorticoid exposures reduces fetal growth in sheep Am J Obstet Gynecol 1998 178: 880–5
Jobe AH, Newnham J, Willet K, Sly P, Ikegami M . Fetal versus maternal and gestational age effects of repetitive antenatal glucocorticoids Pediatrics 1998 102: 1116–25
Marinelli KA, Burke GS, Herson VC . Effects of dexamethasone on blood pressure in premature infants with bronchopulmonary dysplasia J Pediatr 1997 130: 594–602
Evans N . Cardiovascular effects of dexamethasone in the preterm infant Arch Dis Child 1994 70: F25–30
Ford LR, Willi SM, Hollis BW, Wright NM . Suppression and recovery of the neonatal hypothalmic–pituitary–adrenal axis after prolonged dexamethasone therapy J Pediatr 1997 131: 722–6
Garland JS, Alex CP, Pauly TH et al. A three-day course of dexamethasone therapy to prevent chronic lung disease in ventilated neonates: a randomized trial Pediatrics 1999 104: (1 Pt 1) 91–9
Yeh T, Lin Y, Huang C et al. Early dexamethasone therapy in preterm infants: a follow-up study Pediatrics 1998 101: 5 E7 May
Banks BA, Cnaan A, Morgan MA et al and the North American TRH Study Group. Multiple courses of antenatal corticosteroids and outcome of premature neonates Am J Obstet Gynecol 1999 181: 709–17
Ballard RA, Ballard PL, Cnaan A et al for the North American TRH Study Group. Antenatal thyrotropin-releasing hormone for prevention of lung disease in preterm infants N Engl J Med 1998 338: 493–8
French NP, Hagan R, Evans SF, Godfrey M, Newnham JP . Repeated antenatal corticosteroids: size at birth and subsequent development Am J Obstet Gynecol 1999 180: 114–21
Le Cras TD, Markham NE, Abman S . Dexamethasone treatment of infant rats causes lung hypoplasia and increases the severity of hypoxia-induced pulmonary hypertension in adults Pediatr Res 1999 45: 4 309A
Massaro DJ, Massaro GD . The regulation of the formation of the pulmonary alveoli In: Bland RD, Coalson JJ, editors. Chronic Lung Diseases of Early Infancy New York: Marcel Dekker 1998 pp 479–92
Massaro D, Massaro GD . Dexamethasone accelerates postnatal alveolar wall thinning and alters wall composition Am J Physiol 1986 251: R218–24
Blanco LN, Massaro GD, Massaro D . Alveolar dimensions and number: developmental and hormonal regulation Am J Physiol 1989 257: L240–7
Blanco LN, Frank L . The formation of alveoli in rat lung during the third and fourth postnatal weeks: effect of hyperoxia, dexamethasone, and deferoxamine Pediatr Res 1993 334–40
Rush MG, Riaz-Ul-Haq, Chytil F . Opposing effects of retinoic acid and dexamethasone on cellular retinol-binding protein ribonucleic acid levels in the rat Endocrinology 1991 129: 705–9
Pratt L, Magness RR, Phernetton T, Hendricks SK, Abbott DH, Bird IM . Repeated use of betamethasone in rabbits: effects of treatment variation on adrenal suppression, pulmonary maturation, and pregnancy outcome Am J Obstet Gynecol 1999 180: 995–1005
Vermillion ST, Soper DE, Chasedunn-Roark J . Neonatal sepsis after betamethasone administration to patients with premature rupture of membranes Am J Obstet Gynecol 1999 181: 320–7
Abassi S, Hirsch D, Davis J et al. Effect of single versus multiple courses of antenatal corticosteroids on maternal and neonatal outcome Am J Obstet Gynecol 2000 182: 1243–9
We thank the investigators (see Appendix listing), physicians and nurses at all of the TRH study centers for their participation in the original TRH study. We thank Christine Coburn for her role as study coordinator and Peggy McDonald for her assistance in preparation of the manuscript.
This study was supported by grants RO1-HD29201, P50-HL56401, MO1-RR00240, MO1-RR00040, MO1-RR00042, and MO1-RR01271 and by Perinatal Associates, Inc., and the Children's Hospital of Eastern Ontario Research Institution.
In addition to the authors, the following members of the North American Thyrotropin-Releasing Hormone Study Group participated in this study: J. Pinto-Martin, B. Planer, M. McCarthy and E. Escobar, The Children's Hospital of Philadelphia, Philadelphia; M. Morgan, E. Anday, K. Mooney, and M. Johnson, Hospital of the University of Pennsylvania, Philadelphia; W. Parer, R. Phibbs, N. Newton and J. Milar, University of California, San Francisco and San Francisco Medical Center, San Francisco; J. Padbury, M. Ross, D. Polk and S. Harrington, Harbor-University of California, Los Angeles Medical Center, Torrence; D. Davis, K. Ash and J. Frank, Ottawa General Hospital, Ottawa; E. Tyrala and L. Chan, Temple University Hospital, Philadelphia; J. Lioy and R. Librizzi, West Jersey Hospital, Voorhees; M. Hart, J. Garbaciak and E. Ramthun, St. Joseph's Hospital, Phoenix; S. Sawai and C. Carballo, Good Samaritan Hospital, Phoenix; F. Mannino, T. Moore and E. Milan, University of California, San Diego Medical Center, San Diego; H. Schneider and D. Block, Kaiser-Permanente Medical Center, San Diego; J. Keith and M. Rivera-Alsina, Naval Medical Center, San Diego; N. Ragavan and N. Dunn, Abington Memorial Hospital, Philadelphia; V. Bhutani, S. Weiner and M. Grous, Pennsylvania Hospital, Philadelphia.
About this article
Cite this article
Banks, B., Macones, G., Cnaan, A. et al. Multiple Courses of Antenatal Corticosteroids are Associated With Early Severe Lung Disease in Preterm Neonates. J Perinatol 22, 101–107 (2002) doi:10.1038/sj.jp.7210604
Neonatal steroids induce a down-regulation of tenascin-C and elastin and cause a deceleration of the first phase and an acceleration of the second phase of lung alveolarization
Histochemistry and Cell Biology (2014)
The Journal of Steroid Biochemistry and Molecular Biology (2013)
Quantitative Ultrasound Measurements in Premature Infants at 1 Year of Age: The Effects of Antenatal Administered Corticosteroids
Calcified Tissue International (2011)
BMC Pregnancy and Childbirth (2009)
The Indian Journal of Pediatrics (2007)