Optimization of Saturation Targets And Resuscitation Trial (OptiSTART)

Purpose: Oxygen is the most commonly used drug during preterm resuscitation, but its use has been linked to mortality and several morbidities such as bronchopulmonary dysplasia (BPD), intraventricular hemorrhage (IVH), and retinopathy of prematurity (ROP).Preterm infants at birth are more vulnerable to oxygen toxicity caused by exposure to high concentrations of oxygen during resuscitation, but are also equally vulnerable to the adverse effects of suboptimal oxygen use in the delivery room. To balance the need to give sufficient O2 to correct hypoxia and to avoid excess O2, the NRP recommends initiating preterm resuscitation with low FiO2 (≤ 0.3) and subsequent titration to achieve specified target O2 saturations (SpO2). These SpO2 targets (Sat50) are based on approximated 50th percentile SpO2 values of minute per minute pre-ductal SpO2 observed in healthy term newborns who were born vaginally at sea level, breathing spontaneously and did not need any oxygen at birth. Current SpO2 targets are based on expert opinion. In the absence of studies comparing different SpO2 targets, Sat50 was selected by experts at the International Liaison Committee On Resuscitation (ILCOR) and NRP to balance the need for O2 supplementation and avoid excess O2 exposure. Ranges were adjusted by NRP so that they could be memorized easily. However, a recently published study suggested that the currently recommended Sat50 targets are lower than 50th percentile SpO2 values for term infants who undergo delayed cord clamping. Recent studies suggest that the currently recommended Sat50 SpO2 targets may be too low to promote an adequate drop in pulmonary vascular resistance in preterm infants and may be harmful. Recent studies raise concern that the Sat50 strategy may be associated with poorer breathing efforts, higher IVH and higher mortality. The majority of research in O2 use during preterm resuscitation has focused on the starting FiO2, while SpO2 targets have received very little attention. The investigators propose a paradigm shift by changing the research focus to determine the optimal target SpO2 range instead of the initial FiO2. The investigators propose a novel O2 strategy where resuscitation is initiated with 0.3 FiO2 and is titrated every 30 seconds to meet target SpO2 roughly based on the 75th percentile of Dawson curves (Sat75) compared to the 50th percentile of Dawson curves (Sat50). Preliminary data from principal investigator's NICHD K23 funded pilot RCT of 75 preterm infants <31 weeks' gestational age (GA) demonstrated that infants randomized to 75th percentile SpO2 goals (Sat75) had a lower incidence of SpO2 <80% at five minutes and lower duration of bradycardia and PPV in the DR compared to infants randomized to 50th percentile SpO2 goals (Sat50). Although not powered for this analysis, Sat75 infants had less pulmonary hypertension and higher survival without BPD. The investigators hypothesize that the Sat75 strategy results in better pulmonary transition, less time with low SpO2 and lower oxidative stress resulting in a lower incidence of neonatal morbidities such as BPD, IVH, ROP and higher survival without BPD.

Hypothesis: The primary hypothesis is that delivery room resuscitation of preterm infants < 31 weeks' GA with Sat75 targets compared to Sat50 targets will result in increased survival without BPD at 36 weeks' postmenstrual age (PMA).

Study design: A prospective multicenter randomized controlled trial of Sat75 versus Sat50 will be conducted. As per the current NRP guidelines, resuscitation will be initiated with 0.3 FiO2 in both intervention and control groups. The study intervention is the Sat75 where FiO2 will be titrated every 30 seconds by 0.1-0.2 to achieve target SpO2 that approximates the 75th percentile SpO2 observed in healthy term newborns. The control group is the Sat50 where FiO2 will be titrated by 0.2-0.3 every 30 seconds to achieve the NRP recommended target SpO2 which approximates the 50th percentile SpO2 observed in healthy term newborns. In the rare event, where chest compressions are needed during delivery room resuscitation, the FiO2 will be increased to 1.0 as per current NRP guidelines. Once the heart rate (HR) is stabilized, the FiO2 will be reduced to meet the target SpO2. Apart from the randomly assigned target SpO2, resuscitation will follow current NRP guidelines. Assignment to the Sat75 or Sat50 group will use a 1:1 allocation ratio and stratification by GA: 23 0/7 to 27 6/7 weeks and 28 0/7 to 30 6/7 weeks. Demographics, perinatal variables, resuscitation characteristics and neonatal morbidity-mortality data and data of Neurodevelopmental follow-up at around 2 years of age will be electronically recorded using the web-based data entry system developed by the Data Coordinating Center.

Analysis plan: An intention to treat analysis of all randomized infants who received study intervention and have a post-randomization study measurement will be conducted. Baseline characteristics of the newborns will be compared between treatment groups. Descriptive statistics such as mean, SD, median, and range will be used to summarize continuous variables in each study group; frequencies and percentages will be presented for categorical variables. The groups will be formally compared using Student's t-tests and Wilcoxon rank-sum tests, as appropriate, for continuous variables. The chi-square test of association and Fisher's exact test will be used for categorical variables as appropriate.

The primary analysis will be based on the chi-square test of association. Rates of survival without BPD will be presented with 95% confidence intervals. Multivariable logistic regression models (for odds ratios), or modified Poisson regression models with robust standard errors (for risk ratios), will be used to adjust for any imbalanced baseline characteristics. Any characteristics identified as being imbalanced at baseline will be considered as covariates. The primary analysis plan assumes that the intra-class correlation among infants of multiple gestations will be negligible. As sensitivity analyses, generalized estimating equations with robust standard errors will be used to account for clustering effects within siblings. Statistical analysis will be performed with SAS software version 9.4 (SAS Institute, Cary NC). A two-sided 0.05 level of significance will be used and no interim analysis is planned. Data collection protocols will be implemented to minimize missing data. The investigators will conduct 'Per Protocol' analysis and 'As Treated' analysis strictly as secondary analyses to evaluate the impact of time spent outside the group target SpO2 on clinical outcomes. Further, the investigators will conduct pre-planned analyses to examine possible site by testing site by treatment interactions.