The mean head circumference was 289 ± 33 mm. The univariate effect size for antenatal steroid use was −14.5 ± 0.6 mm. Table 3 shows the multivariable model for head circumference as the dependent variable. The multivariable adjusted effect of antenatal steroid was −3.1 ± 0.4 mm (95% CI −3.8, −2.4, P < .001). After birth weight was added to the model, a significant and independent negative association between antenatal corticosteroid use and head circumference remained (−1.2 ± 0.3 mm, 95% CI −1.8, −0.6, P < .001). Multiple gestation was significantly correlated with lower birth weights and smaller head circumference (P < .001).
Figure 1 shows the multivariable effect size of the 12 statistically significant independent variables on head circumference (P < .05). Figure 2 compares the effect sizes of diabetes, preeclampsia, and antenatal steroids in the final multivariable model, with and without birth weight included. Figure 3 shows the effect size of incomplete versus complete courses of steroid therapy on head circumference at birth. Regardless of model, incomplete steroid use had no effect on head circumference at birth. In contrast, all of the models demonstrated a significant negative effect size with at least one completed course of antenatal steroid therapy.
We studied the effects of antenatal corticosteroids on birth weight and head circumference by gestational age using crude and multivariable analysis that included absolute values (birth weight in grams and head circumference in millimeters) and relative values (percent of mean). Figure 4 shows the multivariable‐adjusted effect of antenatal corticosteroids on birth weight by gestational age. The variation in effect sizes by gestational age was significant in both crude and multivariable analyses (P for interaction < .001). This effect increased at later gestational ages (31–34 weeks) in the absolute and relative assessments. The variation in effects on head circumference was similar in the crude analysis but was not significant after adjustment for birth weight.
Use of antenatal corticosteroids varied substantially across the 100 participating sites (mean 57%, interquartile range 49%–67%, range, 3%–88%). Furthermore, the effects of antenatal corticosteroids on birth weight and head circumference varied significantly across sites in crude and multivariable analyses (P for interaction < .001). Multivariable‐adjusted effects on birth weight (mean ± SE) ranged from −791 ± 415 g to 299 ± 219 g (interquartile range −18 ± 36 g to −103 ± 50 g). Most sites (82%) had negative observed effect sizes (ie, lower birth weight was associated with corticosteroid use), and the sites with the most extreme effects (negative or positive) tended to have smaller numbers of cases and thus larger SEs. Inclusion of site in the model did not alter the mean effect of antenatal corticosteroids (adjusted mean effect, −63 ± 6 g). Similar results were found for head circumference. The mean (± SD) number of cases per site was 143 ± 149 (minimum 1, lower quartile 32, median 91, upper quartile 205, maximum 712).
Our findings suggest that antenatal corticosteroid therapy may be associated with a reduction in birth weight and head circumference, independent of other major predictive factors. Our study is unique in terms of its large sample, the wide geographic and racial distribution, and control for multiple confounding variables. Others have observed no effect of antenatal corticosteroid on birth weight3–7 or head circumference1,4,7 because of relatively small samples, great variation in populations, inclusion criteria, corticosteroid use, and study design, and failure to control for many potentially confounding variables. Table 4 reviews the results of previously published studies. Of note, many of the studies in Table 4 compared different corticosteroid exposure groups. Our study controlled for many potentially confounding variables not considered in other studies that may affect newborn head circumference, including birth weight, nulliparity, cesarean section, and presentation.
An important lesson in human teratology was learned from use of a steroid‐related drug, diethylstilbestrol; it is therefore surprising that repeated dosing of corticosteroid was generally accepted without more long‐term follow‐up. Corticosteroids have potent effects in the developing fetus; they shift metabolism from cellular growth to precocious differentiation of cellular proteins in at least 15 different tissues.11 Many concerning animal studies suggest that corticosteroid has potential adverse effects on the developing brain. Weichsel demonstrated that corticosteroid exposure during critical periods of brain development caused irreversible effects on brain cell division, differentiation, and myelination, resulting in latent or long‐term physiologic and behavioral effects.12 In the rhesus monkey fetus, exposure to dexamethasone reduced the number of neurons and caused pronounced degenerative changes in the axodendritic synaptic terminals in the hippocampal region of the brain.13 In the rat model, dexamethasone aggravated ischemic neuronal injury14 and increased motoneuronal cell death.15 In their review of the literature, Huang and colleagues conclude that “there is much evidence from small laboratory mammals that shows exogenous corticosteroids retard brain development with a concomitant mosaic of biochemical, structural and behavioral deficits.”16 In a recent review of the effects of antenatal steroids on the developing brain, Whitelaw and Thoresen also cautioned against use of repeated courses of antenatal corticosteroids.17 Placentation and corticosteroid metabolism and pharmacokinetics differ substantially among mice, rats, sheep, and primates. It is therefore important to consider that two nonhuman primate studies involving pregnant rhesus monkeys demonstrated a reduction in somatic and brain growth associated with equivalent doses of corticosteroids used in humans.18,19
Huang and colleagues used equivalent human doses of betamethasone in pregnant sheep and concluded that this drug significantly retarded fetal brain growth, even after one dose.16 They concluded that a single dose of betamethasone significantly reduced brain weight by 10% at term, and a total of four doses reduced brain weight by 21% (P < .05). The reduction in brain weights were much greater at term than during preterm16; this is especially concerning because human studies (including our report) are primarily limited to preterm newborns.1–8 Our study also suggests that the antenatal steroid effect on birth weight and head circumference increases with advancing gestational age (Figure 4). Relatively small reductions in head circumference are associated with more significant reductions in intracranial volume, since the relationship between these two variables is defined by a cubic function. For example, French and colleagues estimated that a change in head circumference of 1 cm (a 4% decrease) was associated with an 11% reduction in intracranial volume.2
The clinical significance of our findings is unknown. It seems especially of concern that antenatal corticosteroids were associated with a reduction in head circumference, even after controlling for birth weight (Figure 2). Thus, it appears that antenatal corticosteroids reduce brain growth more than somatic growth. In contrast, both diabetes and preeclampsia demonstrated a complete and significant effect reversal when birth weight was added to the model (Figure 2). This effect reversal with addition of birth weight to the model lends significant credibility and validity to our database and our statistical models because they are consistent with the well‐known pathophysiologic effects of the two diseases: Diabetes causes macrosomia, and preeclampsia causes asymmetrical fetal growth restriction (head sparing).20 Esplin and colleagues recently reported that multiple courses of antenatal corticosteroids were associated with a delay in long‐term psychomotor development in children with low birth weight.21 Another recent study demonstrated that small head circumference was strongly associated with learning problems in school‐aged children.22 At least two studies suggest that repeated courses of antenatal corticosteroids were not associated with significant adverse effects at 2 years of age.2,9
The Barker hypothesis, which has been confirmed in large, multiracial epidemiologic studies over four continents, states that decreased birth weight is associated with increased risk for adult cardiovascular and metabolic disorders, including hypertension, hyperlipidemia, type 2 diabetes, and death from ischemic heart disease.23,24 Welberg and Seckl reviewed these studies and suggested that the Barker hypothesis could be mediated by fetal brain programming or imprinting from high levels of glucocorticoids in utero.25 The potential long‐term risks of antenatal betamethasone therapy, in terms of fetal brain imprinting or programming or other adverse events, have not been excluded and require further consideration in clinical practice.
We recognize that the effect size of steroid on head circumference in our study is small, less than the intraobserver or interobserver variance of the actual measurement. Antenatal steroid exposure would not be associated with a consistent measurement bias of head circumference in the nursery, and the large sample would therefore overcome random intraobserver or interobserver variances in measurement. Although the clinical significance of this reduction in head circumference associated with antenatal corticosteroid may be legitimately questioned, the error of the measurement is unrelated to and does not negate the statistical significance of this finding.
Selection bias is unlikely to account for the differences in our study. None of the newborns were selected for the study; all nonanomalous newborns 23 to 34 6/7 weeks of gestational age at birth who were admitted to neonatal intensive care units were included. No previous study has controlled for as many of the major obstetric variables as in our study (Tables 2 and 3), nor has any previous study approached even one tenth of the sample size (Table 4). Clinical management, including use of antenatal corticosteroid therapy, was not standardized among the 100 centers. However, multivariable analysis controlled for factors that might influence the clinician's use of antenatal corticosteroids, such as gestational age, preterm premature rupture of membranes, preeclampsia, and diabetes. The fact that clinical management and steroid use was not uniform across the centers more accurately reflects standard clinical practice. This should be considered a strength of the study rather than a limitation because the steroid effect was consistent across most centers despite great variation in practice patterns, geographic location, and racial distribution.
An important limitation of our study is that the actual number of completed corticosteroid courses could not be ascertained. This is especially unfortunate in view of the dramatic dose‐related effects of antenatal steroid observed by Huang and colleagues14 and French and associates.2 However, within the constraints of the Vermont‐Oxford classification (none, incomplete, or complete),8 a dose‐related effect was clearly established. An incomplete course (<48 hours) was not associated with an effect on head circumference, and every model studied in the cohort receiving at least one completed course (≥48 hours) was associated with a significant reduction in head circumference (Figure 3). The use of multiple courses of antenatal steroid therapy was common practice in the 1990s, and patients who received multiple courses in our study may account for the significant effect size in our “complete steroid use” group. Other potential limitations of the study design and database are largely overcome by the very large sample. Further study is needed to determine the clinical significance of our findings.
1. Banks BA, Merrill JD, Cnaan A, et al., and the North American Thyrotropin-Releasing Hormone Study Group. Multiple courses of antenatal corticosteroids and outcome of premature neonates. Am J Obstet Gynecol 1999;181:709–17.
2. 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.
3. Pratt L, Waschbusch L, Ladd W, Gangnon R, Hendricks SK. Multiple vs. single betamethasone therapy. Neonatal and maternal effects. J Reprod Med 1999;44:257–64.
4. Elimian A, Verma U, Visintainer P, Tejani N. Effectiveness of multidose antenatal steroids. Obstet Gynecol 2000; 95:34–6.
5. Abbasi 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.
6. Thorp JA, Jones AMH, Hunt C, Clark R. Multiple dose antenatal betamethasone and mother-baby outcome. Am J Obstet Gynecol 2001;184:196–202.
7. Shelton SD, Boggess KA, Nurtha AP, Groff AO, Herbert WNP. Repeated fetal betamethasone treatment and birth weight and head circumference. Obstet Gynecol 2001;97;301–4.
8. Bloom SL, Sheffield JS, Mcintire DD, Leveno KJ. Antenatal dexamethasone and decreased birth weight. Obstet Gynecol 2001;97:485–90.
9. Thorp JA, O'Connor M, Jones AMH, Hoffman EL, Belden B. Does perinatal phenobarbital exposure affect developmental outcome at age 2 years? Am J Perinatol 1999; 162:51–60.
10. Crowley PA. Antenatal corticosteroid therapy: A meta-analysis of the randomized trials, 1972 to 1994. Am J Obstet Gynecol 1995;173:322–34.
11. Ballard PL, Ballard RA. Scientific basis and therapeutic regimens for use of antenatal glucocorticoids. Am J Obstet Gynecol 1995;173:254–62.
12. Weichsel ME. The therapeutic use of glucocorticoid hormones in the perinatal period: Potential neurological hazards. Ann Neurol 1977;25:364–6.
13. Uno H, Lohmiller L, Thieme C, et al. Brain damage induced by prenatal exposure to dexamethasone in fetal rhesus macaques. I. Hippocampus. Dev Brain Res 1990; 53:157–67.
14. Tsubota S, Adachi N, Chen J, Yorozuya T, Nagaro T, Arai T. Dexamethasone changes brain monoamine metabolism and aggravates ischemic neuronal damage in rats. Anesthesiology 1999;90:515–23.
15. Prodanov D, Mantchev G, Iliev A, et al. Effects of dexamethasone in rat neonatal model of axotomy-induced motoneuronal cell death. Arch Physiol Biochem 1998;106:355–61.
16. Huang WL, Beazley LD, Quinlivan JA, Evans SF, Newnham JP, Dunlop SA. Effect of corticosteroids on brain growth in fetal sheep. Obstet Gynecol 1999;94:213–8.
17. Whitelaw A, Thoresen M. Antenatal steroids and the developing brain. Arch Dis Child Fetal Neonatal Ed 2000; 83:F154–7.
18. Johnson JWC, Mitzner W, London WT, Palmer AE, Scott R. Betamethasone and the rhesus fetus: Multisystemic effects. Am J Obstet Gynecol 1979;133:677–84.
19. Novy MJ, Walsh SW. Dexamethasone and estradiol treatment in pregnant rhesus macaques: Effects on gestational length, maternal plasma hormones and fetal growth. Am J Obstet Gynecol 1983;145:920–31.
20. Doubilet PM, Benson CB, Callen PW. Ultrasound evaluation of fetal growth. In: Fleischer AC, Manning FA, Jeanty P, Romero R, eds. Sonography in obstetrics and gynecology. 6th edition. New York: McGraw-Hill, 2001: 206–20.
21. Esplin MS, Fausett MB, Smith S, et al. Multiple courses of antenatal steroids are associated with a delay in long-term psychomotor development in children with birth weights ≤ 1500 grams [abstract]. Am J Obstet Gynecol 2000;182:S24.
22. Stathis SL, O'Callaghan M, Harvey J, Rogers Y. Head circumference in ELBW babies is associated with learning difficulties and cognition but not ADHD in the school-aged child. Dev Med Child Neurol 1999;416:375–80.
23. Barker, DJP, Osmond C, Goldings J, Kuh K, Wadsworth MEJ. Growth in utero, blood pressure in childhood and adult life, and mortality from cardiovascular disease. Br. Med J 1989;298;564–7.
24. Barker DJP, Winter PD, Osmond C, Margetts B, Simmonds SJ. Weight in infancy and death from ischaemic heart disease. Lancet 1989;2;577–80.
© 2002 by The American College of Obstetricians and Gynecologists.
25. Welberg LA, Seckl JR. Review Article: Prenatal stress, glucocorticoids and the programming of the brain. J Neuroendocrinol 2001;13:113–28.