Automated oxygen delivery for preterm infants with respiratory dysfunction
Isabella G Stafford,Nai Ming Lai,Kenneth Tan
DOI: https://doi.org/10.1002/14651858.CD013294.pub2
2023-11-30
Abstract:Background: Many preterm infants require respiratory support to maintain an optimal level of oxygenation, as oxygen levels both below and above the optimal range are associated with adverse outcomes. Optimal titration of oxygen therapy for these infants presents a major challenge, especially in neonatal intensive care units (NICUs) with suboptimal staffing. Devices that offer automated oxygen delivery during respiratory support of neonates have been developed since the 1970s, and individual trials have evaluated their effectiveness. Objectives: To assess the benefits and harms of automated oxygen delivery systems, embedded within a ventilator or oxygen delivery device, for preterm infants with respiratory dysfunction who require respiratory support or supplemental oxygen therapy. Search methods: We searched CENTRAL, MEDLINE, CINAHL, and clinical trials databases without language or publication date restrictions on 23 January 2023. We also checked the reference lists of retrieved articles for other potentially eligible trials. Selection criteria: We included randomised controlled trials and randomised cross-over trials that compared automated oxygen delivery versus manual oxygen delivery, or that compared different automated oxygen delivery systems head-to-head, in preterm infants (born before 37 weeks' gestation). Data collection and analysis: We used standard Cochrane methods. Our main outcomes were time (%) in desired oxygen saturation (SpO2) range, all-cause in-hospital mortality by 36 weeks' postmenstrual age, severe retinopathy of prematurity (ROP), and neurodevelopmental outcomes at approximately two years' corrected age. We expressed our results using mean difference (MD), standardised mean difference (SMD), and risk ratio (RR) with 95% confidence intervals (CIs). We used GRADE to assess the certainty of evidence. Main results: We included 18 studies (27 reports, 457 infants), of which 13 (339 infants) contributed data to meta-analyses. We identified 13 ongoing studies. We evaluated three comparisons: automated oxygen delivery versus routine manual oxygen delivery (16 studies), automated oxygen delivery versus enhanced manual oxygen delivery with increased staffing (three studies), and one automated system versus another (two studies). Most studies were at low risk of bias for blinding of personnel and outcome assessment, incomplete outcome data, and selective outcome reporting; and half of studies were at low risk of bias for random sequence generation and allocation concealment. However, most were at high risk of bias in an important domain specific to cross-over trials, as only two of 16 cross-over trials provided separate outcome data for each period of the intervention (before and after cross-over). Automated oxygen delivery versus routine manual oxygen delivery Automated delivery compared with routine manual oxygen delivery probably increases time (%) in the desired SpO2 range (MD 13.54%, 95% CI 11.69 to 15.39; I2 = 80%; 11 studies, 284 infants; moderate-certainty evidence). No studies assessed in-hospital mortality. Automated oxygen delivery compared to routine manual oxygen delivery may have little or no effect on risk of severe ROP (RR 0.24, 95% CI 0.03 to 1.94; 1 study, 39 infants; low-certainty evidence). No studies assessed neurodevelopmental outcomes. Automated oxygen delivery versus enhanced manual oxygen delivery There may be no clear difference in time (%) in the desired SpO2 range between infants who receive automated oxygen delivery and infants who receive manual oxygen delivery (MD 7.28%, 95% CI -1.63 to 16.19; I2 = 0%; 2 studies, 19 infants; low-certainty evidence). No studies assessed in-hospital mortality, severe ROP, or neurodevelopmental outcomes. Revised closed-loop automatic control algorithm (CLACfast) versus original closed-loop automatic control algorithm (CLACslow) CLACfast allowed up to 120 automated adjustments per hour, whereas CLACslow allowed up to 20 automated adjustments per hour. CLACfast may result in little or no difference in time (%) in the desired SpO2 range compared to CLACslow (MD 3.00%, 95% CI -3.99 to 9.99; 1 study, 19 infants; low-certainty evidence). No studies assessed in-hospital mortality, severe ROP, or neurodevelopmental outcomes. OxyGenie compared to CLiO2 Data from a single small study were presented as medians and interquartile ranges and were not suitable for meta-analysis. Authors' conclusions: Automated oxygen delivery compared to routine manual oxygen delivery probably increases time in desired SpO2 ranges in preterm infants on respiratory support. However, it is unclear whether this translates into important clinical benefits. The evidence on clinical outcomes such as severe retinopathy of prematurity are of low certainty, with little or no differences between groups. There is insufficient evidence to reach any firm conclusions on the effectiveness of automated oxygen delivery compared to enhanced manual oxygen delivery or CLACfast compared to CLACslow. Future studies should include important short- and long-term clinical outcomes such as mortality, severe ROP, bronchopulmonary dysplasia/chronic lung disease, intraventricular haemorrhage, periventricular leukomalacia, patent ductus arteriosus, necrotising enterocolitis, and long-term neurodevelopmental outcomes. The ideal study design for this evaluation is a parallel-group randomised controlled trial. Studies should clearly describe staffing levels, especially in the manual arm, to enable an assessment of reproducibility according to resources in various settings. The data of the 13 ongoing studies, when made available, may change our conclusions, including the implications for practice and research.