Effects of antenatal dexamethasone on postnatal serum cortisol levels in late preterm infants

Abstract Objectives To explore whether antenatal dexamethasone impacts postnatal serum cortisol levels in stable late preterm (LPT) infants. Secondary outcomes were to identify short-term hospital outcomes related to antenatal dexamethasone exposure. Methods A prospective cohort study of serial serum cortisol levels in LPT infants within 3 h of birth, and at 1, 3, and 14 postnatal days. Serum cortisol levels were compared between infants exposed to antenatal dexamethasone >3 h and <14 days prior to delivery (aDex) and those who either did not receive dexamethasone or were exposed < 3 h or >14 days prior to delivery (no-aDex). Results Thirty-two LPT infants (aDex) were compared with 29 infants (no-aDEX). Group demographic characteristics were similar. Serum cortisol levels were identical between the groups at all 4-time points. Cumulative antenatal dexamethasone exposure ranged from 0 to 12 doses. Post-hoc analysis of the 24-hour serum cortisol levels indicated a significant difference between 1 to 3 cumulative doses versus 4 or more doses (p = .01). Only 1 infant in the aDex group had a cortisol level <3rd percentile of the reference value. Rates of hypoglycemia (absolute difference [95% CI] − 1.0 [–16.0,15.0]; p = .90) and mechanical ventilation were similar in both groups (absolute difference [95%CI] − 0.3 [–9.3,8.7]; p = .94). No deaths occurred. Conclusion Antenatal dexamethasone administered 14 days prior to delivery did not affect serum cortisol levels and short-term hospital outcomes in stable LPT infants. Exposure to low cumulative doses of dexamethasone resulted in transient low serum cortisol levels compared to 4 or more doses only at 24-hours.


Introduction
Antenatal corticosteroids (ACS) are part of standard therapy in women with anticipated preterm birth to improve outcomes following delivery.Although there is a potential for long-term neurodevelopmental harm [1,2] especially in the extremely preterm infant, the reduction in respiratory distress syndrome from accelerated fetal lung maturation and perinatal and neonatal death, favor treatment [3,4].
Late preterm (LPT) infants (born at 34 to 36 weeks' gestation) are at-risk for postnatal respiratory distress, short and long-term morbidities and mortality when compared to term infants [3].We reported that the rates of mechanical ventilation (MV) and mortality was 4.2 and 3.9-fold higher, respectively, compared to term infants during an observational period of no ACS during LPT gestation [5].Following the ALPS trial [6], ACS use increased substantially in mothers at-risk for LPT delivery [7].Despite the preferential recommendation for antenatal betamethasone, dexamethasone is currently used with the same total dose of 24 mg for any LPT pregnancy where delivery is anticipated within 7 days [8][9][10].
Concerns regarding the routine use of ACS during LPT gestation have been raised [11,12].While extremely preterm infants exposed to ACS may benefit from a reduction in neurodevelopmental sequelae, LPT infants have a higher risk of neurocognitive impairment (adjusted hazard ratio 1.1 [95%CI, 1.05-1.20]),mental and psychiatric disorders [13][14][15][16][17]. Another potential adverse effect of exogenous glucocorticoids is altered regulation of the hypothalamic-pituitary adrenal (HPA) axis [18,19].Such exposure could suppress fetal cortisol production, and extend postnatally to alter normal transition [19].Animal and human observational studies have shown aberrant responses on the HPA axis in infants exposed to ACS, particularly with repeated courses or chronic exposure [20].Therefore, balancing the risks and benefits of routine ACS in mothers who are at-risk for LPT delivery needs careful consideration.Moreover, a substantial proportion of at-risk LPT mothers eventually deliver at term and incur unnecessary ACS exposure [21].Most ACS reports involve the use of betamethasone while the evidence for antenatal dexamethasone (aDex) is limited.The recent WHO ACTION II trial, reported no clinical benefit of aDex for respiratory or any other outcomes in LPT infants over controls [22].
Both betamethasone and dexamethasone are potent glucocorticoids that readily cross the placenta due to ineffective elimination of placental 11ß-hydroxysteroid dehydrogenase type 2 (11ß-HSD2), resulting in supraphysiologic fetal levels [19].Dexamethasone, however, has a relatively short halflife compared to betamethasone [23,24].A recent meta-analysis demonstrated similar clinical outcomes and potential side effects of both agents, but with limited long-term outcomes [25].Data on which drug and dose regimen is best for both mothers and infants, remains inconclusive [25].Since dexamethasone is still used world-wide, it is important to explore whether the recommended regimen impacts cortisol levels of LPT infants and influences respiratory morbidity in resource-limited neonatal settings.
Our primary outcome was to determine postnatal cortisol levels of LPT infants 34 to 36 weeks gestation in a group exposed to aDex within 14 days prior to delivery compared to non-exposed infants.Secondary outcomes were to explore the effect of cumulative doses of aDex on postnatal serum cortisol levels and short-term hospital outcomes.

Material and methods
This prospective cohort study was conducted in a tertiary care perinatal center in Thailand.Obstetrical practices follow the American College of Obstetrics and Gynecology (ACOG) recommendations and includes ACS in women who are LPT and expected to deliver within 7 days.The regimen involves 5 mg intramuscular dexamethasone every 12 h up to a maximum of 4 doses.Birth resuscitation follows the 2020 ILCOR guideline.All LPT infants are admitted to neonatal care units supervised by attending neonatologists and have a glucose screen performed during the first day of life.
The study protocol was approved by the institutional research committee and written parental consent was mandated for eligible infants.Inclusion criterion was inborn, LPT infants at 34 to 36 weeks' gestation.We excluded infants with major congenital anomalies, prenatal or postnatal diagnosis of a brain anomaly, maternal ACS apart from dexamethasone, and postnatal corticosteroid exposure.Gestational age was documented at birth based on ultrasonographic assessment during early pregnancy or postnatal clinical assessment.Each infant had 4 blood samples drawn; (1) within 3 h of birth; (2) at 24 h; (3) 72 h; and (4) 14 days of age.Venous blood samples were collected into ethylenediaminetetracetic acid-containing tubes, centrifuged, and analyzed at the central laboratory within 2 h of collection.Serum cortisol was measured by the cobas 8000 modular analyzer series (Roche Diagnostic GmbH, Germany) which employs a solid phase radioimmunoassay.The laboratory system is certified by the College of American Pathologists and standardized in accordance with ISO 15189.Clinical management was dependent on decisions by responsible caregivers and followed institutional policies.

Definitions
A duration of 14 days prior to delivery was chosen based on recommended elapse times for repeat ACS courses in very-preterm gestation and to allow sufficient time for dexamethasone to affect the HPA axis.Based on the pharmacokinetics of intramuscular dexamethasone in healthy women, the peak level of dexamethasone occurs 3 h post administration [23].The aDex LPT infant group was therefore classified as those exposed to ACS within 14 days and >3 h prior to delivery.Mothers who never received dexamethasone or received aDex >14 days or <3 h prior to delivery were classified as the no aDex group.
Diagnosis of fetal growth restriction (FGR) followed the Society for Maternal-Fetal Medicine guidance [26] which includes ultrasonographic estimation of fetal weight or abdominal circumference below the 10 th percentile for gestational age using the Thai population-based fetal growth references.If either parameter was abnormal, doppler blood flow studies of the umbilical artery and amniotic fluid were evaluated.
Small-for-gestational age (SGA) was defined as a birthweight less than the 10 th percentile using the Fenton curves for gestational age and sex [27].
Hypoglycemia was defined as a plasma glucose <40 mg/dL in symptomatic LPT infants.In asymptomatic infants, a plasma glucose <25 or <35 mg/dL during the first 4 h or 4 to 48 h of life, respectively was employed as the hypoglycemic threshold [26].

Statistical analysis
A sample size was calculated from a previous study of serum cortisol levels among infants who were and were not exposed to ACS [27].Based on a type I error of 0.05, a power of 80%, and estimated drop-out rate of 50%, 30 infants in each group were required.Demographic characteristics between the groups were analyzed using chi-square test, Fisher's exact test, or Mann-Whitney U test according to the type and distribution of each variable.Serum cortisol levels are presented in the median [25 th percentile (P25), 75 th percentile (P75)].Cortisol level comparisons at each time point were evaluated by a Mann-Whitney U test.Low serum cortisol levels were determined using the median and 3 rd percentile levels of reference values for age of healthy-term infants [28].The magnitude of LPT infants with low cortisol levels and hospital outcomes are presented as absolute difference [95% confidence interval; 95%CI] and relative risk [95%CI].
Since the number of infants in each individual cumulative antenatal dexamethasone dose was relatively small, we grouped cumulative doses of dexamethasone throughout pregnancy into 0, 1 to 3, and 4 or more doses to represent no exposure, an incomplete course, and complete course, respectively.Comparisons of cortisol levels within cumulative dose groups were assessed using Kruskal-Wallis test.Dunn's test was used for pairwise comparisons of serum cortisol across the 3 groups of cumulative doses.Analyses were performed using IBM SPSS Statistics 26.0.A pvalue <.05 was considered statistically significant.

Results
From 1 March 2022 to 30 June 2022, 61 LPT infants were enrolled who were delivered to 53 mothers.Figure 1 outlines patient enrollment in the study.Forty-three infants (70.5%) had aDex.Thirty-two infants (51%) received aDex within 14 days and > 3 h prior to delivery (aDex group) and 29 infants (48%) were not exposed within 14 days (no-aDex group).Twenty-six of 29 infants (89.7%) in the no-aDex group received ACS but did not qualify within the study definition.Eighteen and 8 infants received aDex <3 h and >14 days prior to delivery, respectively.Table 1 shows group demographic characteristics.There were no statistical differences except a higher rate of SGA infants in the no-aDex group (37.9% vs 12.5%, p ¼ .02).Diagnosis of FGR was noted in only 1 infant in each group which was statistically insignificant (p ¼ 1.00).Despite similar gestational age distribution, there was an insignificant trend toward lower birthweight in the no-aDex group (2383.4± 478.5 vs 2184.1 ± 339.8 aDex vs no-aDex group; respectively, p ¼ .06).However, the incidence of low-birthweight infants (BW <2500 g) was significantly higher in the no-aDex group [25 (86.2%) versus 20 (62.5%) in aDex, respectively, p ¼ .04].
Table 2 presents postnatal cortisol levels at each postnatal age.There was no difference in the median cortisol level between the aDex and no-aDex groups at every time point.Additionally, the proportion of infants who had cortisol levels less than the median reference values for age was similar between the groups.Only 1 infant in each group (aDex;3.2% vs. no-aDex; 3.4%) had cortisol levels less than the 3 rd percentile for age at 24 h of life (p ¼ .96).
Total cumulative dexamethasone doses throughout pregnancy ranged from 0 to 12 doses.Figure 2 shows the effect of cumulative aDex doses on cortisol levels at each postnatal age.There was no statistical difference across the range of cumulative dexamethasone  Table 3 shows short-term hospital outcomes.The group incidences of morbidities and mortality were similar.The incidence of respiratory distress requiring positive-pressure support was also similar between the groups for both noninvasive (absolute difference À0.6 [-24.5, 14.3]; p ¼ .64)and mechanical ventilation (absolute difference À0.3 [-9.3, 8.7]; p ¼ .94).No infant required high-frequency ventilation.Diagnosis of hypoglycemia was similar (absolute difference À1.0 [-16.0,15.0;p ¼ .90].The rate of hypoglycemia requiring intravenous glucose was 1 (3.1%; aDex) and 1 (3.4%; no-aDex, p ¼ .96).One infant in the aDex group received hydrocortisone for hemodynamic instability due to E.coli sepsis on the 1 st day of life and was excluded from subsequent cortisol monitoring.

Discussion
We reported a 25% incidence of mechanical ventilation in LPT infants who developed respiratory distress prior to the adoption of the ACS protocol in our institution [5].Despite recognition of the possible side effects of ACS in both mothers and infants, 70.5% of LPT infants received antenatal dexamethasone at   some point during pregnancy.Although the sample size in the present study was smaller compared to our previous report [5], the 3% mechanical ventilation rate aligns closely with the 1% rate among LPT infants who received antenatal dexamethasone in the WHO ACTION II trial [29].Therefore, antenatal dexamethasone ameliorated but did not completely eliminate short-term respiratory morbidity, However, in the present cohort, no infant required high-frequency ventilation compared to 0.86% in our previous observational study conducted prior to the use of antenatal dexamethasone.Therefore, antenatal dexamethasone may decrease the rate of very severe respiratory failure.However, the use of antenatal dexamethasone to reduce short-term respiratory morbidity must be carefully weighed against the potential, as yet unknown, long-term neurodevelopmental and metabolic risks.Certain maternal demographic characteristics may affect postnatal neonatal cortisol levels.Twin pregnancies are prone to preterm birth and the benefit of antenatal corticosteroids in LPT twin pregnancies remains unclear [30,31].However, antenatal corticosteroids are widely used in this population including use in the major WHO ACTION II trial.We, therefore, chose to include twins to reflect the generalizability of the use of antenatal dexamethasone.The cesarean section rate in both groups (73.3% and 65.2% in aDex and no-aDex groups, respectively) was relatively high.However, given the spectrum of complicated pregnancies, both maternal and fetal conditions led to obstetrical decisions to expedite delivery before term gestation.In a previous prospective study in our center, before the era of antenatal dexamethasone use in LPT pregnancy, we reported a cesarean section rate of 75% in late preterms [5].A recent study by Ben-David A, et al. [30] reported a cesarean rate of 76% in late preterm twin pregnancies who received antenatal corticosteroids.Hence, our increased cesarean section rate was likely due to the referral of only high-risk pregnancies to our tertiary care center during the COVID-19 pandemic combined with obstetricians' preference for cesarean section to both decrease the risk of COVID-19 transmission and maternal perinatal complications.
Previous studies demonstrate transient cortisol suppression in very preterm infants exposed to antenatal betamethasone [32,33].This effect is more pronounced with multiple courses and those born within 3 days after the last dose of maternal betamethasone [34].Similar transient cortisol suppression12 h after birth with antenatal dexamethasone has also been documented but the authors failed to report on the dose of dexamethasone utilized [35].Kiran et al. investigated the effects of single and multiple courses of antenatal dexamethasone (dose not stated) on the HPA axis of newborns between 29-35 weeks of gestational age.Basal, post-adrenocorticotropic hormone and incremental responses in cortisol were similar, although one infant who received 3 antenatal dexamethasone courses died [36].Since none of the studies systematically evaluated the correlation between antenatal dexamethasone dosage and postnatal cortisol levels specifically in LPT infants, we explored the short-term effects of our study.
Due to our definition of exposure within 14 days prior to delivery, 18 infants (29.5%) received dexamethasone before they reached 34 weeks of gestation.Since we attempted to explore the effect of the recent exposure to antenatal dexamethasone in late-preterm infants, we chose gestational age at birth rather than gestational age at the time of dexamethasone administration.This aligns with the ACOG recommendation of giving antenatal corticosteroids to pregnant women less than 34 weeks gestation who are at-risk of preterm delivery and to those who are 34-36 weeks and have not received a previous course of antenatal corticosteroids [8].
The effect of antenatal dexamethasone is potentially governed by multiple factors such as dose regimen, duration of first exposure, time interval, or the number of cumulative doses.We chose the duration of the first dose as the proxy for exposure in this study and also to assess the effect of cumulative doses.The dose and timing of ACS that affects fetal brain development and the HPA axis remain uncertain.Jobe, et al. [23] correspondingly demonstrated that the median time to reach peak plasma concentration in healthy women was 3 h after intramuscular dexamethasone while 2 smaller studies reported <1 h in pregnant women [37,38].The mean half-life (T1/2) of intramuscular dexamethasone in healthy women is 5.2 h [23].Therefore, we chose a cutoff of 3 h to ensure transplacental infant transfer.Our chosen maximum time of 14 days represents dexamethasone effect relative to LPT gestational age and the waned effect of early ACS during early preterm gestation.The other potential confounder was the effect of perinatal stress on postnatal cortisol levels which is not uncommon in mothers' delivering at LPT gestation.However, infants in our cohort had good Apgar scores and were hemodynamically stable except for one excluded infant who received inotropic agents and hydrocortisone.The serum cortisol levels in our study therefore represent heathy LPT infants which in general was similar between the groups and within normal limits compared to reference values of healthy term infants (Table 2) [28].
Regarding cumulative doses, there was no effect on cortisol levels throughout pregnancy.Cortisol levels across 0 to 4 doses were within the reference range.The differences in serum cortisol at 1 day of life, can be attributed to relatively low cumulative levels with 1 to 3 doses compared to 4 or more doses.Although there was the same trend between no doses versus 4 or more doses, the statistical insignificance of this post-hoc comparison (p ¼ .13)was likely due to the small number of infants who were not exposed to dexamethasone.Hence, neither recent nor widely spread dexamethasone administration affects serum cortisol in LPT infants." Davis et al. reported that moderately-late preterm infants exposed to antenatal betamethasone showed blunted responses of the HPA axis to stressors with lower cortisol levels [39].Both betamethasone and dexamethasone have similar potency with differences in pharmacological properties such as protein-binding capacity and formulation [24].Intramuscular dexamethasone has a shorter half-life compared to betamethasone phosphate (5.2 vs 10.2 h) respectively.It could explain the lower risk of cortical suppression with antenatal dexamethasone and perhaps relative as yet undetermined long-term safety.
The ALPS trial found the risk of neonatal hypoglycemia was 1.6-fold higher in the betamethasone group compared to controls (95% confidence interval [CI]: 1.46-1.96)[40].The postulated mechanism for hypoglycemia was high levels of C-peptide and insulin [41].In the WHO ACTION II trial, the incidence of neonatal hypoglycemia was 8.2% in the dexamethasone group with no difference compared to the control arm (RR [95%CI]: 1.09 [0.65 À 1.81]).[29] We found approximately the same rate of hypoglycemia in the aDex group (9.4%) with no difference compared to the no-aDex group.Therefore, aDex can be considered relatively safe regarding the occurrence of neonatal hypoglycemia.
This is the first study to demonstrate no effect of cumulative doses of antenatal dexamethasone on cortisol levels in LPT infants, and it may be safer than betamethasone.Although some infants in the aDex group received dexamethasone prior to 34 weeks, the majority (27/32; 84.3%) received treatment within LPT gestation with no differences in hospital outcomes between the groups.
Certain limitations may limit generalizability of this study.First, the minimum and maximum timing of antenatal dexamethasone to classify the drug-exposed group may have resulted in contamination of the no-aDex group who received antenatal dexamethasone, but not within our classified time.Therefore, an alternative interpretation of our results may be that cortisol changes occur very quickly (<3 h) following antenatal dexamethasone exposure.Yet, we demonstrated no effect on cortisol levels in both recent exposure and total cumulative doses.Second, our regimen of 5 mg, instead of the standard dose of 6 mg, may have diminished the effect on HPA axis suppression.However, this obstetric dosing regimen is both convenient and reduces wastage (5 mg dexamethasone/ampule).Fourth, we observed cortisol levels in healthy LPT infants which may be altered by stressors experienced during neonatal intensive care.Fifth, our sample size was designed to explore differences in cortisol levels, and we compensated for a drop-out rate post-hospital discharge to ensure the validity of our results.The study was therefore underpowered to identify differences in other clinical outcomes and for each cumulative dose.Finally, we did not evaluate long-term neurodevelopmental outcomes which is important to ensure the true safety of antenatal dexamethasone.

Conclusion
LPT infants exposed to aDex within 14 days of delivery had similar postnatal serum cortisol levels compared to the no-aDex group.Exposure to low cumulative doses of dexamethasone resulted in transient low serum cortisol levels compared to 4 or more doses only at 24-h.There were no differences in short-term clinical outcomes between the groups including neonatal hypoglycemia and the need for respiratory support.

Figure 1 .
Figure 1.Flow diagram of patient recruitment.

Figure 2 .
Figure 2. Effect of cumulative antenatal dexamethasone doses on serum cortisol levels in late preterm infants (a) within 3 h of life; (b) at 24 h; (c) at 72 h; (d) at 14 days of life.Ã P-value <.05 is statistically significant.

Table 1 .
Maternal and infant demographic characteristics of late preterm infants exposed (aDex) and not exposed (no-aDex) to antenatal dexamethasone.

Table 2 .
Comparison of serum cortisol levels in late preterm infants exposed (aDex) and not exposed (no-aDex) to antenatal dexamethasone (N ¼ 61).

Table 3 .
Comparison of hospital outcomes in late preterm infants exposed (aDex) and not exposed (no-aDex) to antenatal dexamethasone (N ¼ 61).