Objective: To evaluate the effectiveness of a bundle to reduce unplanned extubations (UE) in ventilated newborn infants (NB) and to verify the factors associated to UE.Method: Intervention study with a historical control group in a university hospital neonatal intensive care unit (NICU) between June 2014-May 2015 (Period I) and September 2015-August 2016 (Period II). All ventilated NB were included except those with facial malformations. The bundle (new tracheal tube fixation model, team training, identification of NB at risk of UE, and debriefing after UE episodes) was implemented between Periods I and II. Rates of UE/100 NB ventilated-day were compared between periods for the entire sample and according to the cause: accidental or by medical indication. Factors associated to the first UE episode of each NB were studied by logistic regression.Results: A total of 231 intubations were performed in 120 infants in Period I (gestational age 33.6 ± 4.7 W; birth weight: 2020 ± 929 g) and 212 intubations in 131 infants in Period II (34.2 ± 4.7 W; 2080 ± 997 g). UE occurred in 19.9% and 14.6% of the NB, in Periods I and II, respectively. Accidental extubation and change of the tube by medical indication were observed in 58.7% and 41.3% of UE in Period I and in 51.6% and 48.4% in Period II. Higher birth weight, lower SNAPPE-II score, and daytime period were associated with a lower chance of UE in all newborns.Conclusion: The bundle did not reduce the UE in NB ventilated in NICU but continued control of UE rates is crucial for improved care, especially for immature and critically neonates.What is new about the paper? The study presented a strategy for assessing the causes of unplanned extubations in a Neonatal Intensive Care Unit, considering not only the accidental extubations, but aldo the medical ordered extubations, which contributes to the definition of actions for the reduction of unplanned extubations in the NICU setting.