IntroductionRespiratory care of extremely preterm infants remains a challenge. The majority require invasive mechanical ventilation (MV), which is a contributing factor in developing bronchopulmonary dysplasia (BPD). It is important to keep MV to a minimum but there have been concerns that attempting extubation too early increases the risk for atelectasis, re-intubation, and further lung trauma. The aim of this study was to compare two different approaches to extubation. MethodsSingle-center, retrospective cohort study including infants born at 22 + 0–25 + 6 weeks during 2005–2009 and 2011–2015, before and after implementing guidelines recommending delayed extubation. Primary outcomes were BPD, duration of MV and length of hospital stay. ResultsEighty-eight infants in the early era and 102 infants in the late era were included. Infants in the first period were younger at first extubation attempt, and a higher number of infants were extubated within 24 h, 72 h, and one week after birth. The number of infants re-intubated and postnatal age at re-intubation did not differ between the groups. The incidence of severe BPD was 28% in the early period compared to 48% in the later (p < 0.01). Infants in the late period had longer duration of MV (17 vs 27 days, p < 0.01) but similar length of hospital stay (118 vs 123, p = 0.21). ConclusionAfter implementing guidelines recommending delayed extubation, the incidence of severe BPD was higher and the duration of MV was longer. This supports the strategy to attempt extubation early even in extremely preterm infants.