Background: Although recent infant ventilators, including the Servo 300 (Siemens-Elema), have several features that may improve neonatal respiratory outcomes, rigorous clinical evaluations of this new technology have yet to be performed. Objective: To explore the hypothesis that ventilation and oxygenation may be more efficient with the Servo 300. Methods: We retrospectively identified all infants who were switched from a Sechrist IV 100B or Infant Star to a Servo 300 during their stay in our NICU between August, 1993 and December, 1995. Exclusions were: Lack of at least one arterial blood gas (ABG) during each 12 hour period before and after the switch, and peri-operative switches. The ventilator efficiency index (VEI) and the arterial/alveolar oxygen tension (a/A) ratio were computed for each ABG in the 24 hour period surrounding the switch. Multiple observations for each infant were aggregated within each 12 hour period before and after the switch to produce time weighted pairs of VEI and a/A ratio. These were compared by Wilcoxon Signed Rank test. The following cointerventions were recorded during the 24 hour observation period: New prescriptions of narcotic analgesics, inotropes, dexamethasone, indomethacin and neuromuscular blocking agents, surfactant, volume expansion, reintubation, drainage of a pneumothorax and handbagging. Results: There were 69 infants with mean BW (SD) of 1873 (1204) g, and mean (SD) GA of 31.2 (5.7) weeks. Median age at time of switch was 21 hours. The most commonly used ventilation modes on the Servo 300 were Pressure Control (n=43) and SIMV (Pressure Control) + Pressure Support(n=19). Mean (SD) VEI improved from 0.13 (0.07) pre-switch to 0.16 (0.11) post-switch; p=0.001. Mean (SD) a/A ratio improved from 0.19 (0.1) to 0.26(0.13); p=0.0001. Pre-switch prescriptions of narcotics (p=0.009), inotropes(p=0.015) and handbagging (p=0.041) were associated with reduced benefits from the ventilator switch. None of the cointerventions further increased the observed benefit associated with the Servo 300. Conclusion: These uncontrolled data justify the design of controlled trials to test the efficacy of the Servo Ventilator 300.