BackgroundMany infants born at less than 34 weeks of gestational age will require resuscitation in the delivery suite. Yet, different resuscitation techniques are specified in different national guidelines, likely reflecting a limited evidence base. One difference is the length of mechanical inflation initially delivered to infants either via a facemask or endotracheal tube. Some guidelines specify short inflations delivered at rates of 40–60/min, others recommend initial inflations lasting 2–3 s or sustained inflations lasting for ≥ 5 s for initial resuscitation. Research has shown that tidal volumes > 2.2 mL/kg (the anatomical dead space) are seldom generated unless the infant’s respiratory effort coincides with an inflation (active inflation). When inflations lasting 1–3 s were used, the time to the first active inflation was inversely proportional to the inflation time. This trial investigates whether a sustained inflation or repeated shorter inflations is more effective in stimulating the first active inflation.MethodsThis non-blinded, randomised controlled trial performed at a single tertiary neonatal unit is recruiting 40 infants born at < 34 weeks of gestational age. A 15-s sustained inflation is being compared to five repeated inflations of 2–3 s during the resuscitation at delivery. A respiratory function monitor is used to record airway pressure, flow, expiratory tidal volume and end tidal carbon dioxide (ETCO2) levels. The study is performed as emergency research without prior consent and was approved by the NHS London-Riverside Research Ethics Committee. The primary outcome is the minute volume in the first minute of resuscitation with secondary outcomes of the time to the first active inflation and ETCO2 level during the first minute of recorded resuscitation.DiscussionThis is the first study to compare a sustained inflation to the current UK practice of five initial inflations of 2–3 s.Trial registrationClinicalTrials.gov, NCT02967562. Registered on 15 November 2016.Electronic supplementary materialThe online version of this article (doi:10.1186/s13063-017-2311-y) contains supplementary material, which is available to authorized users.