Antenatal corticosteroids is associated with better postnatal growth outcomes of very preterm infants: A national multicenter cohort study in China

Antenatal corticosteroids (ACS) are standard care to accelerate fetal maturation and reduce neonatal morbidity and mortality among preterm infants, with documented benefits for NRDS, IVH, and NEC. However, their potential effects on postnatal growth and nutrition in very preterm infants (VPIs, gestational age <32 weeks) are not well characterized. In China, approximately 200,000 VPIs are born annually with an overall survival rate of 87.6%. Extrauterine growth restriction (EUGR) remains common (47.3% overall; up to 55.3% at GA <28 weeks), higher than reported in developed countries. Prior retrospective studies suggest ACS may modestly protect against EUGR and accelerate weight gain, but ACS coverage in China (about 75.6%) lags behind developed countries. This study tests the hypothesis that ACS improves postnatal growth and nutrition and reduces EUGR among VPIs in a national multicenter cohort.

Background literature shows ACS reduces mortality and major morbidities (NRDS, IVH, NEC) and may promote gastrointestinal maturation via effects on gut peptides (elevated gastrin and motilin post-feeding). EUGR is linked to adverse growth, neurodevelopmental, and cardiometabolic outcomes. Prior multicenter and cohort studies found ACS associated with lower EUGR risk and faster weight gain in very low birth weight infants, though evidence on nutritional processes (feeding tolerance, timing) is limited. Enteral nutrition adequacy is critical to reduce BPD and EUGR; insufficient early enteral intake is associated with higher BPD even with parenteral compensation. Thus, ACS might facilitate earlier, better-tolerated enteral feeding, improving growth trajectories.

Design: Secondary analysis of a multicenter prospective survey (ChiCTR1900023418) conducted by the Chinese Multicenter EUGR Collaborative Group. Data were prospectively collected at 28 tertiary NICUs across 7 regions in China (September 2019–December 2020) and retrospectively analyzed for this study. Population: Infants with GA <32 weeks, admitted to participating NICUs within 24 hours of birth and hospitalized >14 days. Exclusions: major congenital malformations or genetic/metabolic diseases; death during hospitalization or discharge against medical advice; incomplete data; rescue or repeat ACS courses. Exposure: ACS regimens included betamethasone (2×12 mg IM, 24 h apart) or dexamethasone (4×6 mg IM, 12 h apart). Any ACS defined as ≥1 dose. Complete ACS: full course administered ≥24 h and <7 days before delivery; otherwise partial ACS. Infants were grouped as no ACS, partial ACS, complete ACS. Outcomes: Primary growth outcomes: growth velocity of weight (g/kg/day using exponential model) after regaining birth weight, length and head circumference (HC) gain (cm/week), change in Z-scores (weight, length, HC) from birth to discharge per updated Fenton charts, and EUGR at discharge (weight <10th percentile). Secondary nutritional outcomes: enteral feeding start time, days to full enteral feeds (150 ml/kg/day), cumulative fasting days, breastfeeding, accumulated energy intake in first week, cumulative amino acid and fat emulsion doses in first week, and duration of parenteral nutrition. Major morbidities recorded: NRDS grade III–IV, moderate-to-severe BPD, hsPDA, early- and late-onset sepsis, ROP requiring intervention, NEC (Bell ≥2), IVH grade III–IV. Treatments: IMV use and duration, NIV duration, total oxygen support, postpartum corticosteroid exposure, length of stay. Data collection and quality: Standardized questionnaires, trained personnel, double data entry into EpiData, centralized verification and database lock; ongoing monitoring and query resolution by study leadership. Statistical analysis: Categorical variables compared with chi-square or Fisher's exact tests; continuous variables assessed for normality (Shapiro–Wilk) and compared using Mann–Whitney U or Kruskal–Wallis tests with Bonferroni-adjusted pairwise comparisons. Multivariable linear regressions evaluated associations of ACS categories with nutritional and growth outcomes adjusting for potential confounders: GA at birth, birth weight, gender, mode of delivery, multiple birth, 1- and 5-min Apgar scores, SGA, gestational hypertension, diabetes, breastfeeding, NRDS grade III–IV, moderate-to-severe BPD, and IMV use. Model diagnostics included checks for linearity, independence and distribution of residuals (Durbin–Watson, histograms, scatter plots), homoscedasticity, multicollinearity (tolerance/VIF), and outliers (studentized deleted residuals, leverage, Cook’s distance). Multivariable logistic regressions identified independent factors associated with EUGR with similar covariate adjustment; linearity (Box–Tidwell), goodness of fit (Hosmer–Lemeshow), and collinearity diagnostics performed. ACS modeled with dummy variables (no ACS as reference). Subgroup analyses: comparison of any ACS vs no ACS stratified by GA (24–27, 28–29, 30–31 weeks). Two-tailed P<0.05 considered significant.

Cohort: Of 2,800 screened, 2,514 VPIs included. Any ACS: 77.9% (n=1,959); complete ACS: 48.7% (n=1,224); partial ACS: 29.2% (n=735); no ACS: 22.1% (n=555). Baseline: Complete ACS associated with lower median birth weight and lower male prevalence; higher cesarean, multiple pregnancy, and gestational hypertension. GA, birth length/HC, Apgar scores, SGA, gestational diabetes similar across groups. Morbidities/treatments: Complete ACS had lowest NRDS grade III–IV and moderate-to-severe BPD; lower IMV use and shorter IMV duration. NIV duration longest in partial ACS; total oxygen, postpartum steroid exposure, and length of stay similar. Nutritional outcomes (univariate): Any ACS associated with earlier enteral feeding start, shorter cumulative fasting time, higher breastfeeding rate, and higher first-week energy intake; parenteral amino acid/fat doses, PN duration, and time to full enteral feeds similar. Adjusted nutritional outcomes (multivariable linear regression vs no ACS): - Enteral feeding start time: complete ACS B = −5.808 h (95% CI −10.818, −0.798; P=0.023); partial ACS NS. - Cumulative fasting days: partial ACS B = −0.367 d (−0.731, −0.018; P=0.047); complete ACS B = −0.777 d (−1.115, −0.438; P<0.001). Growth outcomes (univariate): Any ACS group had greater greatest weight loss but similar time to regain birth weight. Weight and length growth velocities faster with ACS; HC velocity similar. Decline in weight Z-score less with ACS; discharge anthropometrics (weight, length, HC) similar. Adjusted growth outcomes (multivariable linear regression vs no ACS): - Weight growth velocity: partial ACS B = +0.500 g/kg/d (0.124, 0.876; P=0.009); complete ACS B = +0.640 g/kg/d (0.293, 0.986; P<0.001). - Length growth velocity: complete ACS B = +0.050 cm/w (0.011, 0.089; P=0.012); partial ACS NS. - Change in weight Z-score (less negative decline): partial ACS B = +0.077 (0.009, 0.144; P=0.026); complete ACS B = +0.170 (0.108, 0.232; P<0.001). EUGR: Any ACS had lower overall EUGR incidence (46.1%) vs no ACS (51.5%; P=0.022). By birth weight, reductions notable in 1,000–1,499 g (53.6% vs 60.6%; P=0.026) and trend in <1,000 g. Adjusted odds for EUGR (no ACS reference): partial ACS aOR 0.636 (95% CI 0.476, 0.851; P=0.002); complete ACS aOR 0.603 (0.460, 0.789; P<0.001). Stratified by GA (any ACS vs no ACS): GA 24–27 weeks aOR 0.480 (0.244, 0.942; P=0.033); GA 28–29 weeks aOR 0.653 (0.427, 0.999; P=0.050); GA 30–31 weeks aOR 0.673 (0.475, 0.954; P=0.026). Overall, ACS—especially complete courses—were associated with earlier enteral feeding, reduced fasting, faster weight gain, improved length velocity (complete only), smaller decline in weight Z-score, and lower EUGR risk. ACS exposure also correlated with reduced severe NRDS and moderate-to-severe BPD and less IMV.

The study demonstrates that ACS exposure in VPIs is associated with improved early enteral nutrition and postnatal growth, addressing the hypothesis that ACS benefits growth and nutrition beyond its known respiratory benefits. Mechanistically, prior evidence suggests ACS may stimulate gastrin and motilin, enhancing gastrointestinal maturation, feeding tolerance, and enabling earlier enteral caloric intake, which can help reduce BPD and EUGR. The observed greater early weight loss with ACS likely reflects enhanced diuresis, yet time to regain birth weight was unchanged and subsequent weight velocity was higher, indicating faster recovery and potential early catch-up. A dose–response was observed: complete ACS consistently yielded earlier feeding, larger reductions in fasting days, higher weight velocity, improved length velocity, and less decline in weight Z-score compared with partial or no ACS. The association of ACS with reduced severe NRDS, decreased IMV use and duration, and lower moderate-to-severe BPD aligns with known ACS respiratory benefits and may further facilitate growth by reducing illness severity. Stratified analyses suggest the protective effect against EUGR is evident particularly in the most immature (24–27 weeks) and also in 30–31 weeks GA infants, with borderline significance at 28–29 weeks, extending prior findings that focused on somewhat more mature gestations. Collectively, these results support optimizing ACS administration and standardized nutritional strategies to improve growth outcomes in VPIs.

ACS exposure in very preterm infants promotes earlier enteral feeding, reduces fasting, accelerates weight gain, lessens decline in weight Z-score, and independently lowers the risk of EUGR, with more pronounced benefits following a complete ACS course. These findings support increasing appropriate ACS utilization and implementing standardized nutritional guidelines to improve postnatal growth in VPIs in China. Future research should refine GA- and birth-weight–specific benefits, assess subgroups such as fetal growth restriction and multiple gestations, and evaluate long-term growth and developmental outcomes.

This was a secondary analysis of prospectively collected data; residual confounding cannot be excluded despite multivariable adjustment. Infants hospitalized <14 days and those who died or were discharged against medical advice were excluded, potentially biasing ACS exposure estimates and associations with morbidities. Detailed ACS regimen data (exact timing relative to delivery, dosing specifics, steroid type) were unavailable, limiting evaluation of regimen-specific effects. Some relevant covariates may be unmeasured. Long-term outcomes were not assessed. Larger studies with detailed exposure characterization and longitudinal follow-up are warranted.