ES is the elective treatment for Respiratory Distress Syndrome (RDS) in preterm infants (PT), and of proven usefulness in several studies when administered early after the RDS diagnosis is established. Patients admitted to our Neonatal Intensive Care Unit are transfered from other hospitals, hence, the administration of the first ES dose is mostly made rather late than recommended. Our study's goal is to analyzed late ES adminstration usefulness in terms of neumothorax (NTX) and bronchiopulmonary displasia (BPD) incidence and mortality rate. One hundred and ten PT were prospectively included. Eighty three PT with RDS diagnosis were admitted between April 1988 and March 1991 and conventionally treated without ES (group 0). From April 1991 on all PT with RDS diagnosis (n=27) received ES (group 1), because SE was available in Argentina, we judged that not using it in this patients was unethical. No differences between these groups were found in weight and gestacional age.(p> 0.05). The average age for administration was 19 hours (r= 2 to 57). Statistical analysis was performed with Fisher and t test, depending on each indication. Mortality rate in group 0 was 31.3% vs 17.3% in group 1. This was statistically significant. No PT developped NTX in this latter group, but 18% in group 0 showed this complication (p<0.05). BPD was present in 5% in group 0 vs 28% in group 1 (p<0.05). This data show that the late administration of ES lowered NTX's incidence and mortality rate in our PT with RDS. On the contrary, BPD developped more frequently in this patients, perhaps because of a better outcome in very low weight infants. We conclude that administration of ES is useful even when it is introduced later than recommended.