乌鸦传媒

Challenging existing recommendations, delivery room resuscitation with high initial levels of oxygen for very preterm infants may reduce their odds of death, a systematic review and individual participant data network meta-analysis suggested.

Using data from 12 randomized trials and more than 1,000 infants younger than 32 weeks' gestational age showed that a high initial fractional inspired oxygen (FiO₂) of 90% or greater was associated with a 55% reduction in mortality (OR 0.45, 95% credible interval [CrI] 0.23-0.86) when compared with a low FiO₂ of 30% or less, though with low certainty of evidence.

Furthermore, that high initial FiO₂ was also associated with a lower odds of mortality versus an intermediate FiO₂ of 50% to 65% (OR 0.34 95% CrI 0.11-0.99), this time with very low certainty of evidence, reported James Sotiropoulos, MD, of the University of Sydney in Australia, and colleagues.

The findings diverge from current recommendations and suggest that "further evaluation of high initial FiO₂ is required," the researchers wrote in .

"Whilst promising and potentially practice-changing, these findings will need to be confirmed in future larger studies," Sotiropoulos said in a .

The number needed to treat to prevent one additional death was 16 for the high-versus-low oxygen comparison and 11 for the high-versus-intermediate comparison.

"Ensuring very premature infants get the right treatment from the beginning sets them up to lead healthy lives," he said. "The goal is to find the right balance -- how do we give enough oxygen to prevent death and disability, but not damage vital organs."

Historically, an FiO₂ of 100% had been used for all infants requiring resuscitation at birth, the study authors explained in their introduction, but more recent data showing room air to be associated with reduced mortality in term and near-term infants shifted practice.

"This spurred changes to in 2010 to recommend the judicious use of blended oxygen for preterm infants despite there being no conclusive evidence for this vulnerable population," wrote Sotiropoulos and colleagues.

For preterm infants, a group that often requires oxygen supplementation to prevent hypoxia, those guidelines favor starting at low oxygen concentrations (FiO₂ 21% to 30%) and then titrated to reach oxygen saturation (SpO₂) targets of 80% to 85% at 5 minutes and 85% to 95% at 10 minutes of life, they noted.

Still, the researchers encouraged caution when interpreting their results given the large body of evidence detailing potential harms at higher oxygen concentrations.

"Excessive oxygen exposure may increase the formation of harmful reactive oxygen species, causing irreversible organ and tissue damage," wrote Sotiropoulos and co-authors.

"For this reason, benefits of higher oxygen must be balanced with potential risks of hyperoxia," they continued. "Without careful monitoring of the evolution of SpO₂ and titration of FiO₂, the excessive use of high concentrations of oxygen may be harmful. Our findings do not support the historic practice of administering a static concentration of 100% oxygen; rather, they suggest that high initial FiO₂ that is then titrated down to the lowest effective dose of oxygen is probably superior to starting with low FiO₂ and then titrating up."

For their study, Sotiropoulos and colleagues included 12 randomized trials conducted from 2005 to 2019 that compared two or more initial oxygen concentrations for delivery room resuscitation, ultimately including 1,055 infants under 32 weeks' gestational age with individual patient data.

Ten of the trials were conducted in high-income countries and two in middle-income countries, and trials featuring low or intermediate initial FiO₂ were permitted if the delivered oxygen concentration was titrated based on clinical status.

Of the included infants, 48% were girls, with a median gestational age of 28 weeks. Average birth weight was 1.1 kg, nearly two-thirds were delivered via c-section, 17% were from multiple-birth pregnancies, and 10% were small for gestational age. Most received low oxygen (n=507), followed by high oxygen (n=378) or intermediate oxygen (n=170).

The primary outcome was mortality, and across the 11 studies (n=1,003) that contributed to this endpoint, deaths occurred in 12%. The investigators reported that "there was no evidence of differential effects of treatment across gestational ages or according to infant sex (post hoc, primary outcome only) when examining treatment-covariate interactions," but noted limited power to detect such interactions.

For the secondary outcomes of chronic lung disease, severe intraventricular hemorrhage, and retinopathy of prematurity, they found no differences between the different initial oxygen concentrations (based on 11 trials).

Nine studies provided data on oxygen saturation outcomes and showed that infants receiving higher initial oxygen concentrations had a more than threefold greater odds of achieving an SpO₂ of at least 80% at 5 minutes compared with the lower initial oxygen (OR 3.67, 95% CrI 1.15-12.21). Mean difference in SpO₂ at 5 minutes was 9.02 (95% CrI 2.08-16.31) favoring high versus low initial oxygen.

Limitations included a high risk of bias in the studies (in three studies for the primary outcome), mostly due to a lack of blinding, and low certainty evidence on account of the few mortality events. Sotiropoulos and co-authors also noted that few infants were born younger than 24 weeks' gestation, limiting the generalizability for this group.

Comment