Historical Overview

While evaluating the role of fetal exposures to corticosteroids on labor in sheep, Liggins reported in 1969 that fetal dexamethasone infusions caused preterm delivery of lambs that had unanticipatedly good lung aeration. That result was consistent with other early reports that corticosteroids could mature developing organ systems in other animal models. His observation in sheep was quickly evaluated in a clinical trial reported in 1972 that randomly assigned 282 women to a maternal intramuscular treatment with the drug betamethasone (Celestone) that was used to suppress inflammation. Celestone Soluspan is a mixture of two prodrug forms of the fluorinated corticosteroid betamethasone–betamethasone phosphate and a micronized suspension of betamethasone acetate. When injected intramuscularly, the betamethasone phosphate is rapidly dephosphorylated while the betamethasone acetate is slowly deacylated over many hours. Both compounds yield free betamethasone, which crosses the placenta to achieve a prolonged fetal exposure when given as a two-dose treatment of 12 mg at a 24-hour interval. This level of detail is important to understand today’s approach for ANS dosing (see the following text).

The Liggins and Howie trial demonstrated decreased RDS and mortality in infants of an average gestational age of 35 weeks at delivery. More than 20 trials were completed by 1993 that in aggregate demonstrated comparable effects without clear risks. However, ANS were not widely used in the United States until recommended by a National Institutes of Health Consensus Conference in 1994. The majority of trials tested Celestone against placebo, but dexamethasone phosphate, a similar drug given as 4 doses of 6 mg every 12 hours, was used in some trials as summarized by the definitive meta-analysis of Roberts et al. in 2006 that was updated in 2017. The primary outcomes of most of the trials were decreased RDS and mortality, although ANS also had benefits for decreased intraventricular hemorrhage and necrotizing enterocolitis. ANS became the standard of care for all women at risk of preterm delivery and ANS were recognized as a major advancement in perinatal medicine. There was minimal information developed for drug selection or dose in animal models or humans, and ANS remain unapproved by the U.S. Food and Drug Administration.

Mechanism of Action of ANS on Fetal Lungs

Multiple experimental studies with animals, human fetal lung explants, and isolated lung cells followed the initial Liggins observation in fetal sheep. Exposure to steroids induced the enzymes that contribute to surfactant synthesis and increased surfactant lipid and protein components in tissue and air spaces that became the explanation for the effects of ANS on fetal lungs. However, in large animal models such as sheep and primates, improved lung function as assessed by increased gas volumes and improved lung mechanics occur within 15 to 24 hours, whereas surfactant amounts do not increase for 3 to 5 days. ANS decrease lung mesenchyme tissue volume, and the barrier function of the air space epithelium improves rapidly after ANS in animal models, which contributed to improved lung function soon after ANS treatment. At the molecular level, ANS upregulate and downregulate the expression of multiple genes, with the net effects of improved lung function after preterm delivery. These multiple effects certainly cause improved lung function, but the ANS response is not simply an acceleration of normal lung development. Animal fetuses exposed to ANS have a transient decrease in saccular/alveolar septation and microvascular arborization that is similar to the phenotype of delayed lung development associated with BPD. Alveolarization and microvascular development “catches up” in fetal sheep if they remain undelivered. There is no information about how ANS-induced arrests in lung development associated with improved lung function might alter subsequent lung development after preterm delivery. Empirically very preterm infants exposed to ANS who do not have BPD seem to fare well, although their lung function is not equivalent to the lung function of term infants. Further work using newer research techniques is needed to learn the effect of ANS on the fetal lung.

Dated Randomized Controlled Trials and Current Patient Populations

When ANS were initially used, RDS was a lethal disease for most infants and very preterm infants with or without RDS had very high mortality. The single ANS treatment trials versus placebo reported more than 25 years ago have RDS outcomes for just 102 randomized patients that delivered before 28 weeks’ gestational age ( Table 20.1 ). In contrast, today RDS at more than 28 weeks’ gestation is generally easily managed with surfactant and current techniques for respiratory support. The recommendation at a 1994 Consensus Conference was to administer ANS for women at risk of preterm delivery between 24 and 34 weeks’ gestational age, even though there were minimal data to support that recommendation. We lack the randomized controlled trial (RCT) evidence that use of ANS benefits the pulmonary outcomes of the population of infants most targeted to receive ANS today.

Non-RCT information Regarding ANS and Outcomes

Other information is available to support ANS use at early gestational ages. Very immature animal models and in vitro explants of human fetal lungs have maturational responses to steroids, indicating that receptors and response pathways are present from early gestational ages. Very large databases from neonatal networks are being extensively mined for outcomes of very low-birth-weight (VLBW) infants who were and were not exposed to ANS. These analyses focus on mortality and outcomes such as intraventricular hemorrhage but not respiratory outcomes soon after birth, presumably because most infants at less than 28 weeks’ gestation age receive some respiratory support and are imprecisely coded as having RDS. A weakness of the studies is a lack of information about the severity of the early respiratory distress and any accurate categorization as to whether early deaths are caused by RDS.

Travers et al. reported that ANS decreased mortality for all gestational ages from 24 to 34 weeks. The data are for the period from 2009 to 2013 for 61,571 infants, with about 84% of infants exposed to ANS, except at 34 weeks when the percent of exposure was 45% ( Fig. 20.1 ). ANS exposure resulted in a remarkably consistent decreased mortality even at 34 weeks’ gestation. A similar report from Carlo et al. documented decreased mortality and intraventricular hemorrhage for more than 10,000 infants for the gestational age range from 22 to 25 weeks with 74% of the population exposed to ANS. No respiratory outcomes are reported other than BPD, which is increased by use of ANS ( Fig. 20.2 ). The association of increased BPD with ANS is quite uniform across the RCTs and epidemiology studies. The generally accepted explanation is that ANS improves survival in the most marginal infants who are at increased risk of developing BPD. However, the arrest of septation and microvascular development with ANS may contribute to this adverse lung outcome. The severity of BPD relative to ANS has not been analyzed and is worthy of evaluation.

Epidemiology of a large clinical experience for the effect of ANS on infant mortality (2009–2013). Across gestations from 24 to 34 weeks, ANS was associated with a similar decrease in infant death, but the number needed to treat increased greatly as gestational age increased.

Data from Travers CP, Clark RH, Spitzer AR, et al. Exposure to any antenatal corticosteroids and outcomes in preterm infants by gestational age: prospective cohort study. BMJ . 2017;356:j1039.

Association of antenatal steroid use and the bronchopulmonary dysplasia (BPD) outcome. Across very early gestational ages, there was no decrease in BPD, and the overall effect was an increase in BPD. Data for 10,541 infants cared for in the Eunice Kennedy Shriver National Institute of Child Health and Human Development Neonatal Research Network from 1993 to 2009.

Data from Carlo WA, McDonald SA, Fanaroff AA, et al. Association of antenatal corticosteroids with mortality and neurodevelopmental outcomes among infants born at 22 to 25 weeks’ gestation. JAMA . 2011;306:2348–2358.

The optimal interval from ANS treatment to delivery for maximal benefit based on the RCTs is 1 to 7 days. A new analysis of a large European data set not randomized to ANS indicates that maximal benefit for decreasing death may be an exposure interval as short as 18 to 36 hours, with benefit after 3 hours. Such rapid effects would be more consistent with acute steroid effects on cardiovascular function rather than for lung maturation.

A concern with the associations between ANS and the clinical outcomes is that the great majority of women are treated with ANS. The women not given ANS differ in ways that cannot be reliably adjusted for in the analyses. The women not exposed to ANS have different causes and severities of problems related to prematurity. Many likely deliver before ANS can be given. The epidemiologic analyses may overestimate the benefits of ANS at early gestational ages. Another concern is treatment creep. The RCT data to support early gestational treatment with ANS are minimal, and RCTs for the use of ANS at “previable” gestational ages are nonexistent. Again, the epidemiology supports improved outcomes at gestations of 22 and 23 weeks, but that information likely reflects highly selected use of ANS in pregnancies being actively managed to achieve infant survival.

Repeated ANS Treatments

The pregnancies of many women considered likely to deliver 1 to 7 days after ANS may continue for weeks or until term. A recent report noted that more than 50% of women given ANS did not deliver within 7 days. The fact that the majority of women do not deliver in an optimal window for efficacy and many deliver at term is of substantial concern if this “off- target” treatment has adverse effects on the mother or infant. Multiple trials have evaluated Celestone for repeated ANS given at weekly intervals, at 2-week intervals, or when preterm delivery appeared to be imminent more than 7 days after the initial treatment. The benefits of steroids seem to be optimal at 1 to 7 days after ANS because some of the maturational effects seem to fade after 7 days, and repeated weekly treatments in animal models augment lung benefits. The concerns with repeated treatments are cumulative injury to the fetal lung and neurodevelopment. A meta-analysis of 10 trials that randomized 4733 pregnancies demonstrated modest improvements in respiratory outcomes with no concerns for neurodevelopment. However, several trials reported effects of repeated courses of ANS on birth weight and length, and one trial was stopped early for perceived adverse fetal growth effects. There are fewer concerns about a single repeated dose of ANS, but the use of multiple weekly treatments likely is excessive. Further, a single dose of Celestone may be adequate for retreatment rather than the two-dose treatment.

Concerns Regarding Drug and Dose Choices for ANS

Liggins chose the 50/50 mixture of betamethasone phosphate plus betamethasone acetate for maternal intramuscular treatments to match the prolonged exposures with fetal infusions in sheep. This dose and treatment interval were not validated by standard pharmacology. The other ANS treatment recommended by the World Health Organization (WHO) is 4 doses of 6-mg dexamethasone phosphate. There have been fewer clinical evaluations of this less expensive and widely available drug for ANS. Notably, there are no trials for repeated courses of treatment using dexamethasone. Worldwide, different drug and dosing schedules are used without validation—for example, the United Kingdom uses betamethasone phosphate, an untested drug. Based on measurements of others indicating that drug doses are too high and our experience with betamethasone acetate, the slow-release component of Celestone, it is likely that the perinatal community is using the wrong drugs at higher doses than necessary to achieve the pulmonary benefits of ANS. In fetal sheep models, 25% of the total betamethasone exposure given as one dose of betamethasone acetate is equivalent to the current clinical dose, and lower doses may be equally effective as a lung maturational agent. Further, dexamethasone phosphate or betamethasone phosphate given as two doses is less effective for lung maturation in sheep models. Concerns about risks of ANS should be minimized if lower dosing strategies can be developed.

Long-Term Outcomes After ANS

The 30-year outcomes of the original Liggins and Howie trial were reassuring, but those infants were more mature than current populations. One concern is that the follow-up from the early trials focused on the infants who were delivered before term, with no follow-up of those who received off-target treatments and were delivered close to or at term. Braun et al. observed that newborns delivered after 34 weeks after receiving ANS treatment weighed less than unexposed newborns. There are many other cohort or convenience comparisons of variables (e.g., responses to stress and aortic stiffness) that suggest adverse effects of ANS in later life. Although these outcomes are consistent with results from animal models, a substantial concern is that the human populations have not been randomly assigned to ANS treatment. Because the great majority of women eligible for ANS are treated, a comparison population of unexposed children may not be representative. All observations are further confounded by prematurity. Global neurodevelopmental outcomes are reassuring even for infants exposed to repeated courses of ANS.

Antenal Steroid Use in Low-Resource Environments

Most of the premature infants born worldwide are born in low-resource environments and have a very high mortality rate. The use of ANS is very spotty despite their use as a priority of the WHO and other groups attempting to improve maternal and infant outcomes. The recently completed Antenatal Corticosteroid Trial illustrated the difficulties and risks of using ANS in low-resource environments. This trial randomly assigned almost 100,000 women in Pakistan, Africa, India, and Guatemala to the WHO-recommended ANS treatment of four doses of betamethasone after identification of risk of preterm delivery. Major problems were a lack of gestational age determination and identification of women at risk. Use of ANS did not improve outcomes for the smallest infants who had a perinatal mortality risk of more than 350 per 1000 births. More problematic was the increased mortality of off-target ANS-exposed infants who were large and mature, perhaps because of increased infection. This experience is a caution that all exposed infants need to be followed for outcomes, not just those delivered prematurely. A safe treatment in one environment may be neither safe nor effective in a different environment.