Introduction Perinatal healthcare has improved over the past decades, whereas care to extremely low gestational age neonates (ELGAN) has become increasingly “active”. However, whether such a trend translates into a benefit in terms of neonatal neurological morbidity is still discussed.

Patients/Design Historical cohorts study of ELGAN admitted in a tertiary neonatal unit between January 1997 and December 1999 (period I, n  =  147) and between January 2004 and December 2006 (period II, n  =  266). Neonatal morbidity and cerebral ultrasound scan data have been prospectively recorded. Statistical analysis included Student’s t test and χ² analyses.

Results (Mean ± SD) birth weight (946 ± 219 g vs 897 ± 238 g, p<0.01) and gestational age (27 ± 1 weeks vs 26.7 ± 1.3 weeks, p<0.01) was decreased in period II. Severe cerebral haemorrhage (grades III/IV) decreased from 13.6% to 6.7% (p<0.01). Cystic periventricular leucomalacia was not different between periods I and II (3.4% vs 4.9%), as were persistent periventricular echogenicities and/or isolated ventricular enlargement (4.7% vs 5.6%). Survival without cerebral ultrasound abnormalities: 24–26 weeks: period I, II (15/37, 61/111; p = 0.18); 27–28 weeks: period I, II (76/110, 128/155; p = 0.02).

Conclusion In this study, neonatal neurological morbidity in ELGAN decreased during the current decade. Further studies may address medium and long-term outcomes of temporal trends.