We had hoped to have the answer to the following question by now: where should we target our oxygen saturation goals for premature neonates born at <28 weeks of gestational age? The hope was that data gathered from 5 trials – Surfactant, Positive Pressure, and Oxygenation Randomized Trial (SUPPORT), Benefits of Oxygen Saturation Targeting (BOOST II) (consisting of three separate trials: UK, Australia, and New Zealand), and Canadian Oxygen Trial (COT) would definitely guide practice. But that data has not. One potential complicating factor has been the question of how to deal with an “anomaly” in the pulse oximeter calibration software, discovered while BOOST II and COT were still ongoing, but SUPPORT was completed. In brief, is has been argued that the oximeter software—once corrected—led to better separation in achieved saturations between babies randomized to 85%-89% and 91%-95%. It has also been arguedthat with this better separation in treatment groups, the increased mortality in babies randomized to 85%-89% becomes clearer. However, this interpretation of the study results has been questioned. In this volume of The Journal, Whyte et al asked whether the revised oximeter calibration software did in fact allow for a “cleaner” separation in the BOOST II treatment groups. Why is this study important? BOOST II consisted of three separate trials: UK, Australia, and New Zealand. When BOOST II reported that the revised software allowed for “greater separation,” they analyzed data from subjects randomized in all three trials both before and after the oximeter software change. Looking at the babies enrolled after the software change, a cleaner separation of achieved pulse oximeter readings was achieved. However, the New Zealand trial had completed their enrollment prior to the software change, and thus could not contribute any data to the postsoftware change group. Thus, Whyte et al ask whether data from subjects enrolled pre- and postsoftware change from just the UK and Australia trials support the hypothesis that the revised software results in a cleaner separation in treatment groups, thus potentially explaining the clearer signal for increased mortality in babies randomized to the 85%-89% in BOOST II after the software change. Their conclusion is riveting. The data presented here challenge our interpretation of these trials, and forces thoughtful discussion about how to move forward with oxygen saturation target goals in the neonatal intensive care unit. Article page 382 ▶ Benefits of Oxygen Saturation Targeting Trials: Oximeter Calibration Software Revision and Infant SaturationsThe Journal of PediatricsVol. 182PreviewIt has been reported in the 3 Benefits of Oxygen Saturation Targeting (BOOST-II) trials that changes in oximeter calibration software resulted in clearer separation between the oxygen saturations in the two trial target groups. A revised analysis of the published BOOST-II data does not support this conclusion. Full-Text PDF