Inadequate Timing Limits the Benefit of Antenatal Corticosteroids on Neonatal Outcome: Retrospective Analysis of a High-Risk Cohort of Preterm Infants in a Tertiary Center in Germany.

Introduction A common problem in the treatment of threatened preterm birth is the timing and the unrestricted use of antenatal corticosteroids (ACS). This study was performed to evaluate the independent effects of the distinct timing of antenatal corticosteroids on neonatal outcome parameters in a cohort of very low (VLBW; 1000 - 1500 g) and extreme low birth weight infants (ELBW; < 1000 g). We hypothesize that a prolonged ACS-to-delivery interval leads to an increase in respiratory complications. Materials and Methods Main data source was the prospectively collected single center data for the German nosocomial infection surveillance system (KISS) between 2015 and 2018. Multivariate regression analysis was performed to determine independent effects of the ACS-to-delivery interval on the need for ventilation, surfactant or the occurrence of bronchopulmonary dysplasia, neonatal sepsis or necrotizing enterocolitis. Subgroup analysis was performed for ELBW and VLBW neonates. Results A total of 239 neonates were included. We demonstrate a significantly increased risk of respiratory distress characterized by the need for ventilation (OR 1.045; CI 1.011 - 1.080) and surfactant administration (OR 1.050, CI 1.018 - 1.083) depending on the ACS-to-delivery interval irrespective of other confounders. Every additional day between ACS and delivery increased the risk for ventilation by 4.5% and for surfactant administration by 5%. Subgroup analysis revealed significant differences of respiratory complications in VLBW infants. Conclusions Our data strongly support the deliberate use and timing of antenatal corticosteroids in pregnancies with threatened preterm birth versus a liberal strategy. When given more than 7 days before birth, each day between application and delivery increases is relevant concerning major effects on the infant. Especially VLBW preterm neonates benefit from optimal timing. The Author(s). This is an open access article published by Thieme under the terms of the Creative Commons Attribution-NonDerivative-NonCommercial License, permitting copying and reproduction so long as the original work is given appropriate credit. Contents may not be used for commecial purposes, or adapted, remixed, transformed or built upon. ( https://creativecommons.org/licenses/by-nc-nd/4.0/ ).

Introduction

Administration of antenatal corticosteroids (ACS) in threatened premature birth (PTB) before 34 weeks of gestation (wks) is stated to be the most effective treatment known for preventing serious preterm complications in the newborn including respiratory distress syndrome (RDS), intraventricular hemorrhage (IVH) and necrotizing enterocolitis (NEC), as well as reducing overall mortality 1 ,  2 . Betamethasone 12 mg twice or dexamethasone 4 × 6 mg decrease mortality and morbidity markedly for preterm infants especially in very low birthweight infants (VLBW < 1500 g) 3 ,  4 . Data of extremely low birthweight infants (ELBW < 1000 g) are still rare 3 ,  5 ,  6 . ACS enhances maturation of pulmonary structure and function 1 ,  7 ,  8 . Nevertheless, it is still unknown how long these beneficial effects last. There is evidence that the benefits of corticosteroids diminish over time. On the other hand the absence of ACS or incomplete ACS are risk factors for poor neonatal outcomes 9 . Current literature suggests a window of optimal efficacy between 48 hours and seven days after administration of the first dose of ACS 5 ,  10 ,  11 ,  12 . Administration of a full course of ACS in PTB is recommended by worldwide guidelines 13 ,  14 ,  15 ,  16 , being used as a quality marker for delivery room practice despite the unpredictable nature of PTB, which leads to a liberal application of ACS to all PTB 17 ,  18 . Currently, only 23 to 40% of ACS are administered in the window of optimal efficacy, which puts a large number of infants at risk for detrimental outcomes 10 ,  19 ,  20 .

The objective of this study was to evaluate the effects of ACS-to-delivery interval on neonatal outcome parameters in a high-risk cohort of VLBW infants below 34 wks.

Materials and Methods

Cohort composition

From January 1, 2015 to December 31, 2018 487 women were treated in the tertiary perinatal center of Jena university hospital because of threatened PTB before 34 wks. 248 women delivered after 34 wks or with a birthweight > 1500 g and were therefore excluded ( Fig. 1 ).

Fig. 1

 Cohort composition. Abbreviations: ACS – antenatal corticosteroids; NICU – neonatal intensive care unit; wks – weeks of gestation.

We included 239 inborn neonates with a birthweight < 1500 g and a GA < 34 wks treated in our Neonatal Intensive Care Unit (NICU). Seven of those received no ACS, 47 received incomplete ACS and 185 received a complete course of ACS. A full course of ACS was defined as two doses of betamethasone with 24 hours in between doses and delivery at more than 48 hours after first dose. Only one dose or delivery before 48 hours after first dose was defined as incomplete course. Window of efficacy was defined as a delivery between 48 h and seven days after the first 12 mg betamethasone dosage (group ACS ≤ 7 days). The other group, out of this window, was defined as ACS > 7 days. No rescue courses or multiple dose regimens were used in our institution.

In the 185 cases of completed ACS neonatal outcomes were compared between those born ≤ 7 days (n = 90) and those born > 7 days after beginning of ACS (n = 95). To further differentiate concerning the effects of maturity and birthweight the cohort was stratified for ELBW (< 1000 g) and VLBW (1000 – 1500 g) subgroups.

We do confirm that any research activities during this study were performed according to ethical standards of the Declaration of Helsinki and the protocol for this research project has been approved by the Ethics Committee of the Friedrich-Schiller-University.

Ascertainment of clinical parameters and perinatal outcome

The main source for neonatal parameters was the prospectively collected data for the German national nosocomial infection surveillance system, known as KISS (Krankenhaus-Infektions-Surveillance-System) and combined with clinical patient records retrospectively. KISS data were only entered by one of this studyʼs investigators. Annual training and skill-tests for the investigators were ensured. Regular quality control meetings and reports from the KISS center secured data integrity. We assessed all infants < 1500 g birth weight up until either death or reaching a weight of 1800 g. Parameters included were need for CPAP or need for surfactant via catheter or endotracheal tube and need for ventilation, as well as outcome parameters like mortality, neonatal sepsis, NEC and bronchopulmonary dysplasia (BPD) for neonatal complications and age at birth, fetal birthweight, singleton or multiple pregnancies. We decided to use objective parameters like ventilation and surfactant instead of the somewhat subjective radiographic interpretation of RDS 16 .

Neonatal hypotrophy was defined as a birth weight below 10th percentile for gestational age or the Z-Score below − 1.282 according to the revised Fenton Preterm Growth Charts and is represented as small for gestational age (SGA) in the tables 21 .

BPD was diagnosed based on Walsh criteria (need for supplemental oxygen or ventilation at 36/0 wks) 22 and NEC on Neo-KISS criteria (German neonatal surveillance system) 23 . The infant morbidity was calculated using the CRIB-Score (Clinical Risk Index for Babies), which is a robust predictor of neonatal morbidity. It is a semiquantitative score that consists of six items (birthweight, gestational age [GA], severe malformation, neonatal acidosis and maximum and minimum oxygen requirements in the first 12 hours of life) with a minimum of zero and a maximum of 23 achievable points 24 .

Pregnancy-related data including indication for delivery, mode of delivery and GA at birth were retrieved from our perinatal database. Diagnosis of delivery was grouped: 1st PTB caused by preterm premature rupture of membranes (PPROM) and/or chorioamnionitis; 2nd fetal growth restriction (FGR), 3rd maternal indications (e.g., pre-eclampsia, HELLP, placenta praevia) and 4th others (e.g., placental abruption). Mode of delivery included C-section and any type of vaginal delivery.

Statistical analysis

For the main group comparison all inborn infants with GA < 34 wks and birthweight < 1500 g were considered (n = 239). For the subgroup comparisons and multivariate logistic regression analysis we included only preterm infants after a full course of ACS (n = 185). In order to get unbiased estimates in the multivariate logistic regression model, a prior sample size estimation was performed for the primary outcome variables ventilation and surfactant use 25 . χ 2 test or Fisher exact test were used to compare categorical data. Since most of the continuous data were not normally distributed, we used the median and interquartile range for data presentation and description. Mann-Whitney U test or Kruskal-Wallis test were performed to compare continuous data between groups. Adjusted odds ratios (ORs) for estimating the association of the ACS-to-delivery interval (days) and gestational age at birth (wks) with neonatal complications (ventilation, surfactant, CPAP, BPD, neonatal sepsis and NEC) were determined using logistic regression. ORs with 95% confidence interval (CI) are presented. The potential confounders like fetal sex, birthweight and sepsis were included as covariates in the statistical analysis as presented in footnotes. Regression models were tested for an overall predictive value using Omnibus-Tests and only significant overall predictive models are presented 26 . A p-value < 0.05 was considered to indicate statistical significance (2-tailed). Statistical analysis was performed with SPSS 24.0.

Results

Within the observation period, 346 of the 487 (71%) women with threatened PTB and receiving ACS gave birth between 23/0 and 34/0 wks. Although this rate is quite better than reported by others 27 ,  28 it means that in hindsight 135 women unnecessarily received ACS.

A total of 239 neonates met our inclusion criteria (delivery in our hospital; NICU admission; born between 23/0 and 34/0 wks and a birth weight < 1500 g), of which 185 received a complete course of ACS. Seven neonates did not receive ACS prior birth and 47 neonates received an incomplete ACS only. There was no significant difference between these groups regarding fetal sex, numbers of multiple pregnancies, gestational age at birth (weeks) and birth weight, but in the mode of delivery and delivery indications.

As expected, newborns with missing or incomplete ACS demonstrated more adverse outcomes with higher mortality, NEC rates and increasing need for invasive ventilation. The entire cohort characteristics and groups comparisons are presented in Table 1 .

Table 1  Cohort characteristics and univariate analysis of the subgroups: no ACS, incomplete ACS and complete ACS (≤ 7 days and > 7days).

	Entire cohort	No ACS	Incomplete ACS	ACS ≤ 7 days	ACS > 7 days	p-value †	p-value ‡
Data are n (%) or median and interquartile range (IQR) unless otherwise specified. Significant findings (p < 0.05) are highlighted in bold.† Comparison between the four groups. ‡ Comparison ACS ≤ 7 days and ACS > 7days only. ACS – antenatal corticosteroids; PPROM – preterm premature rupture of membranes; CPAP – continuous positive airway pressure; SGA – small for gestational age; NEC – necrotizing enterocolitis; BPD – bronchopulmonary dysplasia; CRIB- clinical risk index for babies
Total, n	239	7	47	90 (48.6)	95 (51.4)
Sex
Male	128 (53.6)	5 (71.4)	27 (57.4)	47 (52.2)	49 (51.6)	0.706	1
Female	111 (46.4)	2 (28.6)	20 (42.6)	43 (47.8)	46 (48.4)
Multiples						0.032	0.022
Twins	72 (30.1)	2 (28.6)	20 (42.6)	19 (21.1)	33 (32.6)
Triplets	4 (1.7)	–	–	–	4 (4.2)
Delivery indication						< 0.001	0.163
PPROM or chorioamnionitis	143 (59.8)	5 (71.4)	28 (59.6)	46 (51.1)	64 (67.4)		–
Fetal growth restriction	49 (20.5)	–	2 (4.3)	28 (31.1)	19 (20)		–
Maternal indications	26 (10.9)	1 (14.3)	6 (12.8)	11 (12.2)	8 (8.4)		–
Others	21 (8.8)	1 (14.3)	11 (23.4)	5 (5.6)	4 (4.2)		–
Mode of delivery						0.021	0.125
Vaginal	20 (8.4)	2 (28.6)	7 (14.9)	8 (8.9)	3 (3.2)
C-Section	219 (91.6)	5 (71.4)	40 (85.1)	82 (91.1)	98 (96.8)
Gestational age (weeks)	28 (26 – 30)	27 (26 – 32)	26 (25 – 29)	28 (26 – 30)	29 (27 – 30)	0.002	0.042
ACS-to-delivery interval (days)	5 (2 – 12)	–	1 (1 – 2)	4 (3 – 5)	17 (10 – 26)	< 0.001	< 0.001
Birth weight (g)	1020 (790 – 1305)	1030 (855 – 1290)	945 (750 – 1310)	950 (683.75 – 1245)	1180 (935 – 1385)	0.003	< 0.001
SGA	56 (23.4)	3 (42.9)	7 (14.9)	28 (31.1)	18 (18.9)	0.060	0.063
CPAP	226 (95.4)	6 (85.7)	42 (91.3)	84 (94.4)	94 (98.9)	0.105	0.109
Surfactant	150 (63.3)	5 (71.4)	32 (69.6)	53 (59.6)	60 (63.2)	0.679	0.651
Ventilation	76 (32.2)	3 (42.9)	23 (50)	21 (23.6)	29 (30.9)	0.017	0.320
Sepsis	21 (8.8)	–	3 (6.4)	14 (15.6)	4 (4.3)	0.036	0.012
NEC	12 (5)	–	6 (12.8)	5 (5.6)	1 (1.1)	0.025	0.112
BPD	27 (11.3)	1 (14.3)	4 (8.5)	15 (16.7)	7 (7.4)	0.214	0.068
Neonatal death	18 (7.5)	2 (28.6)	7 (14.9)	7 (7.8)	2 (2.1)	0.007	0.093
CRIB-Score	2 (1 – 7)	6 (1 – 15)	3 (1 – 7.25)	2 (1 – 7)	1 (1 – 5)	0.003	0.013

Multivariate logistic regression analysis for complete ACS with confounders

Multivariate logistic regression analysis (including confounders: sex and birth weight) in the subgroup of complete ACS showed a significant effect of the ACS-to-delivery interval (time between first dose and delivery) on the need for ventilation (OR 1.045; CI 1.011 – 1.080) and surfactant administration (OR 1.050, CI 1.018 – 1.083) as shown in Table 2 . When looking at the subgroup of complete ACS within the window of efficacy no significant independent effect of ACS-to-delivery interval could be demonstrated for the need for ventilation (OR 1.065; CI 0.724 – 1.566) nor for surfactant administration (OR 1.290; CI 0.874 – 1.904). Each additional day between ACS and delivery increased the risk for ventilation by 4.5% and surfactant by 5% despite all the other confounders.

Table 2  Multivariate regression analysis of the association of ACS-to-delivery, gestational age and neonatal complications.

Outcome variable	Influencing variable	Adjusted ORs*	CI (95%)	Adjusted ORs †	CI (95%)
* Adjusted for sex and birth weight; † adjusted for sex, birth weight and sepsis; significant findings (p < 0.05) are highlighted in bold. ACS – antenatal corticosteroids; CPAP – continuous positive airway pressure; BPD – bronchopulmonary dysplasia; NEC – necrotizing enterocolitis
Ventilation (n = 183)	ACS-to-delivery interval (days)	1.045	1.011 – 1.080	1.048	1.013 – 1.084
	Gestational age (weeks)	0.604	0.466 – 0.783	0.592	0.454 – 0.772
Surfactant (n = 184)	ACS-to-delivery interval (days)	1.050	1.018 – 1.083	1.050	1.018 – 1.083
	Gestational age (weeks)	0.544	0.426 – 0.695	0.545	0.427 – 0.696
BPD (n = 185)	ACS-to-delivery interval (days)	1.032	0.970 – 1.099	1.026	0.958 – 1.097
	Gestational age (weeks)	0.684	0.479 – 0.977	0.657	0.440 – 0.979
Sepsis (n = 184)	ACS-to-delivery interval (days)	0.948	0.860 – 1.045
	Gestational age (weeks)	0.961	0.689 – 1.341
NEC (n = 184)	ACS-to-delivery interval (days)	0.834	0.613 – 1.135	0.824	0.596 – 1.141
	Gestational age (weeks)	1.240	0.734 – 2.097	1.369	0.775 – 2.418

Gestational age at birth also revealed an independent impact on the need for ventilation (OR 0.604, CI 0.466 – 0.783) and for surfactant use (OR 0.544 CI 0.426 – 0.695) as well as for BPD (OR 0.684; CI 0.479 – 0.977). Each additional week of intrauterine maturation decreased the risk for ventilation by 38%, for surfactant application by 47% and for the occurrence of BPD by 31% as expected. Adding sepsis as a potential confounder as presented in the right column of Table 2 did not affect prior results.

Univariate analysis: Comparison of complete ACS ≤ 7 days vs. ACS > 7 days

Timing of ACS was within the window of efficacy in 48.6% (n = 90) of the cases (ACS ≤ 7 days). Both subgroups ACS ≤ 7 days and ACS > 7 days did not differ in fetal sex, delivery indications and mode of delivery but were significantly different in GA (28 wks [IQR 26 – 39] vs. 29 wks [IQR 27 – 30]; p < 0.05), birth weight (950 g [IQR 838.75 – 1245] vs. 1180 g [IQR 935 – 1385]; p < 0.01) and CRIB-Score (2 [IQR 1 – 7] vs. 1 [IQR 1 – 5]; p < 0.05). Less multiples were seen in the group with ACS ≤ 7 days (21.1% vs. 36.6%; p < 0.05).

The timing of ACS showed a strong trend towards the need for less CPAP, surfactant, and ventilation in the subgroup ACS ≤ 7days. No statistically significant effect was seen on the incidence of NEC, BPD or neonatal death. Remarkably, there were more cases of neonatal sepsis (15.6% vs. 4.3%; p < 0.05) in the group with ACS ≤ 7 days compared to the group with ACS > 7 days ( Table 1 ).

Univariate analysis of complete ACS: VLBW and ELBW subgroups

To countervail the birthweight discrepancies in the main group, a subgroup analysis was performed comparing ELBW and VLBW infants ( Table 3 ). Within these groups the efficacy of ACS timing was evaluated. The mean intervals between ACS and delivery were significantly different when comparing ACS ≤ 7 days and ACS > 7 days in all subgroups.

Table 3  Comparison of the subgroups ACS ≤ 7 days and ACS > 7 days depending on birthweight categories: ELBW and VLBW.

	ELBW < 1000 g			VLBW ≥ 1000 – 1500 g
Variable	ACS ≤ 7 days	ACS > 7 days	p-value †	ACS ≤ 7 days	ACS > 7 days	p-value †
Data are n (%) or median and interquartile range (IQR) unless otherwise specified. Significant findings (p < 0.05) are highlighted in bold.† Comparison between the subgroups ACS ≤ 7 days and ACS > 7 days using nonparametric tests for each subgroup. ACS – antenatal corticosteroids; ELBW – extreme low birthweight; VLBW – very low birthweight; PPROM – preterm premature rupture of membranes; SGA – small for gestational age; CPAP – continuous positive airway pressure; NEC – necrotizing enterocolitis; BPD – bronchopulmonary dysplasia, CRIB – clinical risk index for babies
Total, n	51	31		39	64
Sex			0.820			0.1
Male	27 (52.9)	15 (48.4)		20 (51.3)	34 (53.1)
Female	24 (47.1)	16 (51.6)		19 (48.7)	30 (46.9)
Multiples
Twins	10 (19.6)	5 (16.1)	0.775	9 (23.1)	26 (40.6)	0.033
Triplets				–	4 (6.3)
Delivery indication			0.966			0.033
PPROM or chorioamnionitis	30 (58.8)	19 (61.3)		16 (41.0)	45 (70.3)
Fetal growth restriction	17 (33.3)	9 (29.0)		11 (28.2)	10 (15.6)
Maternal indications	3 (5.9)	2 (6.5)		8 (20.5)	6 (9.4)
Others	1 (2)	1 (3.2)		4 (10.3)	3 (4.7)
Mode of delivery			0.245			0.632
Vaginal	6 (11.8)	1 (3.2)		2 (5.1)	2 (3.1)
C-Section	45 (88.2)	30 (96.8)		37 (94.9)	62 (96.9)
Gestational age (weeks)	26 (24 – 27)	27 (25 – 27)	0.215	30 (29 – 32)	29 (29 – 30)	0.112
ACS-to-delivery interval (days)	4 (3 – 5)	13 (10 – 22)	< 0.001	4 (3 – 5)	20.5 (10 – 32)	< 0.001
Birth weight (g)	720 (550 – 920)	840 (675 – 935)	0.074	1290 (1150 – 1384)	1295 (1172.5 – 1427.5)	0.509
SGA	17 (33.3)	7 (22.6)	0.330	11 (28.2)	11 (17.2)	0.219
CPAP	47 (92.2)	31 (100)	0.292	37 (97.4)	63 (98.4)	1
Surfactant	43 (84.3)	25 (80.6)	0.765	10 (26.3)	35 (54.7)	0.007
Ventilation	19 (37.3)	13 (43.3)	0.642	2 (5.3)	16 (25)	0.014
Sepsis	11 (21.6)	4 (13.3)	0.555	3 (7.7)	–	0.052
NEC	4 (7.8)	1 (3)	0.646	1 (2.6)	–	0.379
BPD	13 (25.5)	6 (19.4)	0.598	2 (5.1)	1 (1.6)	0.555
Neonatal death	6 (11.8)	0	0.078	1 (2.6)	2 (3.1)	1
CRIB-Score	6 (2 – 9)	7 (2 – 8)	0.535	1 (0 – 2)	1 (0 – 2)	0.884

In the VLBW subgroup (birthweight 1000 – 1500 g) ACS ≤ 7 days reduced the risk for a RDS with a reduced need for surfactant application (26.3% vs. 54.7%; p < 0.01) and for ventilation (5.3% vs. 25.0%; p < 0.05) significantly. In the VLBW group there were significant differences concerning the indication for delivery, because more children in the ACS ≤ 7days group were treated for FGR (28.2% vs. 15.6%) whereas more cases of PPROM (41.0% vs. 70.3%; p > 0.05) and multiples (23.1% vs. 46.9%) could be seen in the ACS > 7 days group. We did not see differences in neonatal complications: death, BPD, NEC and sepsis in both subgroups. In the ELWB subgroup < 1000 g (n = 82) no differences of patient characteristics, indications, nor perinatal complications could be found. There was again a trend towards the need for less mechanical ventilation in the ACS ≤ 7 days subgroup.

A visualization of the cohort and the main respiratory complications (surfactant use and mechanical ventilation) is summarized in Fig. 2 , according to birthweight and ACS-to-delivery interval. The scatter plots show in a case by case manner the higher percentage of interventions (e.g. more green dots) in the ACS > 7 days group (right upper quadrant) in VLBW infants (1000 – 1499 g – right quadrant) than in the ACS < 7 days group (right lower quadrant).

Fig. 2

 Scatterplots of respiratory complications ( a  mechanical ventilation, b  surfactant use) matched to birthweight and ACS-to-delivery interval, horizontal line at 7 days of ACS-to-delivery interval, vertical line at birthweight of 1000 grams. Empty plots mean no complication, full plots mean complication occurred. Abbreviations: ACS – antenatal corticosteroids.

Discussion

This analysis of a large well characterized cohort of high-risk preterm neonates below 1500 g birthweight could demonstrate a significantly increased risk of severe RDS characterized by the need for ventilation (OR 1.045; CI 1.011 – 1.080) and surfactant administration (OR 1.050; CI 1.018 – 1.083) depending on the ACS-to-delivery interval irrespective of other confounders. Every additional day between ACS and delivery increased the risk for ventilation by 4.5% and for surfactant administration by 5% in our cohort ( Fig. 2 ). However, this could not be shown when the delivery occurred within 7 days in the window of optimal efficacy. Additionally, every week of maturation decreased the risk of pulmonary complications. This finding emphasizes the importance of a careful timing as close to delivery as possible versus a liberal use of ACS in all PTB. We focused on the patient group of ELBW and VLBW infants as those are at the highest risk of complications and long-term sequelae due to their immaturity 23 ,  29 ,  30 .

Our results also confirm previous data 3 ,  4 and demonstrate that a completed course of ACS reduces neonatal mortality by 83% compared to no ACS and by 67% compared to an incomplete ACS course. We also report a significant reduction of NEC and the need for ventilation as shown in Table 1 in agreement to recent studies and meta-analysis 1 ,  3 ,  31 . This underlines the need and benefits of a full course of ACS before the delivery of a preterm infant whenever possible.

In the univariate analysis, reduced respiratory complications could be seen as a strong trend but failed to reach statistical significance. Remarkably, the group ACS ≤ 7 days comprised infants were born in mean 9 days earlier (28 wks vs. 29 wks; p < 0.05), with lower BW (950 g vs. 1180 g; p < 0.01) and seen to be sicker (CRIB Score 2 vs. 1; p < 0.05) than in the group ACS > 7 days, which are all known risk factors for respiratory complications. This imbalance could be a limitation due to the retrospective character of the analysis, which is not stratified to GA of ACS prophylaxis. There were also more cases of FGR in the group ACS ≤ 7 days (31.1% vs. 20%). Neonatal sepsis occurred more often when ACS was given within 7 days to birth. Lower GA, lower BW and FGR are individual risk factors for neonatal sepsis, which seems to be a possible explanation for this effect in our cohort 23 ,  32 . FGR fetuses were excluded from most studies 3 ,  32 since it is hypothesized that higher catecholamine and steroid levels due to the fetal stress in FGR flatten the ACS effect 33 .

A subgroup analysis according to birthweight < 1000 g (ELBW) and 1000 – 1500 g (VLBW) showed that VLBW infants did need significantly less ventilation (5.3% vs. 25%; p < 0.05) and surfactant (26.3% vs. 54.7%; p < 0.01) if they were born within 7 day after ACS. We were able to show this positive effect in the group ACS ≤ 7 days, although FGR is known to be a risk factor for pulmonary complications 3 ,  32 ,  34 and the distribution of neonates with hypotrophy in the two groups was somewhat unbalanced (28.2% vs. 17.2%, but n. s.). Adding hypotrophy as an independent risk factor to the multivariate regression did not change the results of our analysis concerning the effect of the ACS-to-delivery interval. Nevertheless, our results must be interpreted cautiously because of a higher rate of PPROM cases (41 vs. 70.3%; p < 0.05) in the VLBW group, albeit not neonatal sepsis (VLBW 7.7% vs. 0%; n. s.). Our data consistently showed a strong trend towards less severe RDS as seen in less need for ventilation in the ELBW group with timely ACS but failed to reach statistical significance. Nearly all infants needed surfactant therapy initially independent of ACS timing in this subgroup. These infants are in the highest risk group for neonatal complications (median age both 26 vs. 27 wks and median weight 720 vs. 840 g) but constitute a major part of our study population (82/185 infants). It is still unclear to what extent these extremely premature infants benefit from ACS. An explanation for the uncertain effect of ACS on outcome could be the extreme immaturity of the pulmonary system 3 ,  35 . Additionally, the multitude of treatment options in NICU treatment nowadays may reduce the prominent effectiveness of ACS 16 . Our findings suggest that these infants also seem to benefit from a timely ACS administration, adding valuable data for extremely premature infants 3 ,  5 ,  6 .

One strength of our study is the large number of VLBW and ELBW infants for which data is rare and treatment algorithms are still unclear. As a single center study, we do have high-quality data records, clear treatment algorithms by SOP and therefore little intra-operator variability. Nevertheless, the knowledge about ACS administration might have biased the treatment of infants. To account for these problems objective parameters like need of ventilation and surfactant use were considered. The non-randomized character of the study caused imbalances in the group composition, which we could partially countervail by stratified subgroup analysis. This could limit the generalizability of our results to other delivery populations.

These limitations can only be overcome by a randomized controlled trial to verify these results, but this seems ethically questionable from our point of view.

Conclusion

In conclusion, our data strongly support the deliberate use and timing of ACS in women at risk of preterm birth in contrast to unrestricted ACS administration as scattershot. Quality improvement measures should therefore focus on administration of ACS within the window of efficacy instead of administration of ACS to every woman with threatened PTB. Each additional day in between ACS and delivery is relevant concerning major effects on the infant. VLBW infants especially seem to profit from ACS in the window of efficacy concerning respiratory complications. Since prospective studies on ACS seem not ethically feasible, we believe that our results based on prospective collected register data (KISS) are of high value.