Background Although antenatal corticosteroids (ACS) represent the paradigm of “evidence-based medicine”, their efficacy in case of twin pregnancy is not established, and available results are conflicting.

This study aimed at evaluating the association between ACS and neonatal outcomes in twins, comparing the results with those in singletons.

Methods A cohort of neonates 24–29 weeks gestational age (GA) without congenital anomalies, born in 2005–2012 and assisted in 90 hospitals adhering to the Italian Neonatal Network, was analysed.

Outcomes were: death, grade 3–4 intraventricular haemorrhage (IVH), cystic periventricular leukomalacia (PVL). Logistic regression models, adjusting for GA, sex and birthweight and clustering for hospitals, were used. Results were also checked adjusting for a propensity score of receiving ACS.

Results We studied 13029 infants (mean GA 27.1 weeks; mean birthweight 964 g); 81.4% were treated with ACS (any dose); 29% were twins. Twins were treated more often than singletons (84.4 vs. 80.2%, p < 0.001).

Among twins, ACS were associated with a reduction of in-hospital death (adjusted Odds ratio, aOR = 0.75; 95% CI: 0.58–0.98) and IVH (aOR = 0.52; 95% CI: 0.40–0.68) but not PVL (aOR = 0.92; CI: 0.62–1.36).

The effect of ACS was always smaller in twins than in singletons (72% lower for death, 23% for IVH, and 15% for PVL).

Analyses using the propensity score approach yielded similar results.

Conclusions This large cohort allowed us to clarify that ACS prophylaxis is efficacious also in twins but the magnitude of the effect is always lower than in singletons. Given the high prevalence of twins among preterm infants, this matter should be further investigated.