Glucocorticoids in perinatal medicine: Misguided rockets?☆

See related article, p. 9 .

Once the rockets are up, who cares where they come down? ‘That’s not my department,’ says Wernher von Braun.

Tom Lehrer as quoted by W. A. Silverman (page 5 )1

Glucocorticoids are one of the most potent classes of drugs used in perinatal medicine. Neonatologists like results, and glucocorticoids have immediate physiologic effects on lung function of preterm infants, which are clinically apparent (even to parents) and demonstrable in clinical trials.2 They get the rockets up. Finer et al3 summarize recent information indicating that there are significant short-term (days to weeks) and long-term (1 to 2 years) adverse effects that should make us seriously question where and how the rockets come down. The problem with integrating the information from the multiple clinical trials into a consensus recommendation for clinical practice results from 2 overlapping factors. Glucocorticoid treatment effects (beneficial or adverse) are perhaps as sensitive to dose and duration of treatment as any class of pharmacologic agents, and the very preterm infant is at a developmental stage that may be particularly susceptible to glucocorticoid effects.4 Those effects likely are secondarily modified by common superimposed events in the preterm infant such as antenatal or postnatal proinflammation or infection and catabolic state. Most clinical trials have used high-dose dexamethasone (acute asthma doses) that were weaned over periods as long as 42 days. Infants have been treated early (<72 hours) in attempts to prevent chronic lung disease, and infants have been treated later (7 to 28 days) in attempts to minimize progression of chronic lung disease. The trials in aggregate provide little guidance on dose, when to initiate treatment, or duration of treatment. Almost any treatment schedule imaginable has been tried, but few have heeded the general guidelines to initially try the lowest possible doses of glucocorticoids to achieve a therapeutic effect.5

If one chooses to use glucocorticoids, it seems to me that there should be a clear short-term clinical goal (perhaps extubation), and if that goal is not achieved within 3 to 4 days, then therapy should be stopped. Furthermore, the “standard” initial dose of 0.5 mg/kg/d (equivalent to 12 mg of cortisol) is very high relative to the 24-hour basal production rate of about 0.5 mg cortisol for a preterm infant weighing 1 kg and a stress production rate of perhaps 1.5 mg/d. The recently terminated National Institute Child Health and Human Development trial used an initial dose of 0.15 mg/kg/d dexamethasone and found an increased incidence of gastrointestinal perforations, demonstrating complications at lower doses.6 Initial doses of 0.1 to 0.2 mg/kg/d dexamethasone should be sufficient and may be too high. Watterberg et al7 recently reported in a small trial that 1 mg of hydrocortisone per day, begun at <48 hours of age and continued for 9 days followed by 0.5 mg of hydrocortisone per day for 3 days, increased survival without supplemental oxygen at 36 weeks’ postconceptional age. This strategy is based on the frequent occurrence of low plasma cortisol values in very low birth weight infants (adrenal insufficiency?). The issue of minimal effective dose is ripe for more clinical research.

What should be the criteria for good targeting of where the rockets come down? Short-term benefits can be important, and a decreased mortality rate for infants when treatment was initiated between 7 and 21 days has been demonstrated by meta-analysis.2 However, the overriding concerns of Finer et al3 are the discouraging outcomes at 1 to 2 years. There is no follow-up information to school age or older ages. A point that needs emphasis is that there are only a few reports in which reasonable numbers of infants were followed up—and those numbers are small. The glucocorticoid doses used for the trials that reported adverse outcomes are worth noting. Yeh et al8 gave 0.5 mg/kg/d dexamethasone, beginning at 8 hours of age, with the dose tapered over 3 weeks. O’Shea et al9 randomized infants to 0.5 mg/kg/d dexamethasone between 15 and 25 days of age and tapered the dose over 42 days. Shinwell et al10 began 0.25 mg/kg dexamethasone within 12 hours of age and gave 6 doses over 3 days. These trials used high doses very early after birth, long courses of dexamethasone, or both. I agree with Finer et al3 that all future trials should include a plan for collecting and reporting follow-up results.

There also are some concerns about where the rockets are landing after the use of antenatal glucocorticoids. A single course of antenatal glucocorticoids for women at risk for preterm delivery before 32 weeks should be the standard of care.11 However, repetitive courses of glucocorticoids at weekly intervals may have no benefit and may increase the risk of adverse outcomes.12, 13

All the available data concerning the clinical use of repetitive courses of glucocorticoids are from retrospective analyses, which do little more than generate hypotheses that need to be tested in prospective randomized trials. Those trials are now underway and should include assessments at follow-up. A recent report about choice of antenatal glucocorticoid is of particular concern. Baud et al14 reported that antenatal dexamethasone (4 doses of 6 mg every 12 hours) increased periventricular leukomalacia relative to antenatal betamethasone (2 doses of 12 mg of betamethasone given 24 hours apart). These glucocorticoids and dosage schedules are the 2 standard treatment courses used for antenatal glucocorticoids, although an “accelerated” betamethasone course is also used with no apparent added benefit. Fetal heart rate variability and breathing patterns change more after administration of betamethasone than after administration of dexamethasone to mothers.15 In sheep, fetal infusions of betamethasone cause preterm labor, whereas fetal infusions of dexamethasone do not consistently cause preterm labor.16 The clinical data suggest that antenatal betamethasone is more effective than dexamethasone in preventing death.17 Although these glucocorticoids are stereoisomers thought to have equivalent pharmacologic effects,5 there are a number of unexplained differences. Baud et al14 suggested that the 0.15 mg sulfite preservative per milligram of dexamethasone might be responsible for the increased incidence of periventricular leukomalacia. I think this not a likely explanation. Sulfites are common food additives and are present in cardiovascular drugs, and 3 g of protein for hyperalimentation contains about 20 mg of sulfites. The maternal exposure will be small relative to the postnatal exposure of very preterm infants. The weight of evidence is that betamethasone is the preferred glucocorticoid for antenatal use. Perhaps we should question our choice of dexamethasone for postnatal use.

Glucocorticoids will continue to be used in perinatal medicine. However, Finer et al3 are correct to point out that we need to be very concerned about long-term outcomes. Future studies need to focus on the perhaps unrealistic goal of finding the minimum dose and duration of treatments that will be effective. Adverse neurodevelopmental outcomes no doubt result from the effects of these potent agents on developing systems and should come as no surprise. We should not accept short-term benefits as demonstrating therapeutic gains without understanding the implications for neurodevelopmental outcomes. We should not view glucocorticoids from Galen’s perspective.

All who drink of this treatment recover in a short time, Except those whom it does not help, who all die. It is obvious, therefore, that it fails only in incurable cases

Galen as quoted by W. A. Silverman (page 3 )1