Can a Machine Manage a Premature Baby's Oxygen Better Than a Nurse's Hand?

A 2025 review of sixteen studies looks at automatic oxygen control for babies breathing with gentle support instead of a breathing tube

Premature babies often need extra oxygen, but getting the amount exactly right is surprisingly hard, and both too little and too much can cause harm. A 2025 review by researchers in London gathered sixteen studies testing whether a computerised system that adjusts a baby's oxygen automatically does this better than a nurse turning the dial by hand. The answer, for babies on gentle "non-invasive" breathing support, is that the automatic systems kept oxygen levels in the safe zone more of the time — though no study yet shows whether this leads to healthier babies in the long run.

Why Getting Oxygen Right Is So Difficult

Oxygen is one of the oldest treatments in newborn care, and one of the trickiest. A premature baby's lungs are immature, so many babies need extra oxygen to survive. But oxygen is a double-edged sword. Too little (called hypoxia) can be dangerous and is linked to worse survival. Too much (called hyperoxia) floods the body with harmful molecules and raises the risk of two serious complications of prematurity: a chronic lung condition called bronchopulmonary dysplasia, and an eye condition called retinopathy of prematurity that can damage vision [1]. Because of this, neonatal teams aim to keep each baby's blood-oxygen level — measured painlessly by a sensor on the hand or foot — inside a carefully chosen "target range," not too high and not too low.

The problem is that hitting that target, minute after minute, day after day, is genuinely hard. A landmark study called AVIOx followed very premature babies and found they spent less than half their time inside the intended oxygen range, and that how well a baby stayed on target depended a lot on which nurse was caring for them [2]. This is no criticism of nurses — a premature baby's oxygen needs can swing up and down many times an hour, and one person caring for several babies simply cannot adjust every dial every moment. Yet the target really matters: large trials known as BOOST-II showed that even small differences in the oxygen goal changed how many babies survived and thrived [3], and earlier research had already shown that oxygen affects the developing eye [4]. For decades, families and clinicians simply lived with this gap between the goal and the reality — which is exactly the gap that automatic oxygen control was invented to close. The new review pulls together what we now know about using it in babies on gentle breathing support [5].

How the Automatic System Works

A closed-loop automated oxygen control system — "closed-loop" because it constantly checks its own results and corrects itself — reads the baby's oxygen sensor many times a minute and makes tiny adjustments to the oxygen supply on its own, without waiting for a person [6]. Think of it like the cruise control in a car: instead of the driver constantly tapping the accelerator to hold a steady speed, the system makes hundreds of small corrections automatically, while the driver stays alert and ready to take over. The nurse and doctor remain fully in charge and can step in at any time; the machine simply handles the constant fine-tuning.

The review focused on babies receiving "non-invasive" support — methods such as CPAP (a gentle pressure delivered through soft prongs in the nose) and high-flow nasal cannula — rather than babies with a breathing tube. This matters because neonatal care has been moving toward these gentler methods to avoid the lung injury that a breathing tube can cause. Of the sixteen studies the researchers found, most were small and short, but together they tell a fairly consistent story [5].

What the Studies Found

The main finding is encouraging and easy to state: the automatic systems kept babies in their safe oxygen range for more of the time than hand adjustment did. In one of the first such studies, babies were in target about 90% of the time with the machine, compared with about 82% by hand [7]. Newer systems did even better — in one study babies were on target 81% of the time with automation versus 58% by hand, with fewer dangerous dips and spikes and far fewer manual adjustments needed [8]. The same pattern held for babies on high-flow support, where one study found on-target time of 80% with the machine versus 49% by hand [9]. In concrete terms, that means less time spent too low and less time spent too high — the two situations clinicians most want to avoid. Just as importantly, the automatic systems cut down dramatically on the number of times a nurse had to reach over and adjust the oxygen by hand [7][8]. A premature baby's oxygen can dip and rise many times an hour, and each adjustment is a small task; a system that handles the constant fine-tuning frees the nursing team to spend more time on feeding, comfort, developmental care, and talking with families. This is one of the quiet, practical reasons units are interested in the technology, quite apart from the oxygen numbers themselves.

It also helps to understand why keeping oxygen steady is thought to matter so much. When a baby swings too high, extra oxygen can injure the delicate, still-growing lungs and the blood vessels of the eyes; when a baby dips too low, vital organs are briefly starved of oxygen. Neither extreme is good, and the danger comes not only from where the average sits but from how often a baby bounces between the two. Because the automatic systems reduced time at both the high and the low ends [9], rather than simply trading one problem for the other, researchers see them as a genuinely promising way to smooth out those swings. The different studies used several different machines and computer programs, and one of them worked better when it was set to react more quickly — a hint that the exact design of the system matters, and that not every device will behave the same way [5].

There was an important wrinkle, though. The benefit was not the same at every stage. One study that ran the automatic system for a full 28 days found that in the very first days of life the machine actually let babies dip below target a bit more than hand control did, and only later did it pull ahead [10]. Another study tracking babies over about seven weeks found the clearest advantage in the first two weeks, with little overall difference across the whole stay [11]. The lesson is that the technology is a helper, not a replacement for close human attention — especially in the earliest, most fragile days.

It also helps to know where this technology came from. Automatic oxygen control was first tested, and is best proven, in babies who have a breathing tube and are on a ventilator; in that situation it has been linked not just to steadier oxygen but to shorter time on the ventilator and less lung disease [5]. Babies breathing with gentle nasal support are a different situation, though. Air can leak around the soft nasal prongs, babies breathe through their mouths, and very premature babies often have natural pauses in their breathing — all of which make the oxygen harder for any system to predict and control. That is part of why the evidence in gently supported babies, while consistent and encouraging, is not yet as strong as it is for babies on a ventilator, and why doctors are careful not to promise more than the studies actually show.

What This Means for Families

If your baby is cared for in a unit that uses one of these systems, the practical message is reassuring. The evidence consistently shows that automatic oxygen control helps keep a baby's oxygen steadier and within the intended range, which is exactly what the whole team is working toward. It also frees nurses from constantly adjusting the dial, giving them more time for other parts of your baby's care. Importantly, the machine never takes over from the medical team — doctors and nurses set the target, watch closely, and step in whenever needed.

It is just as important to be honest about what we do not yet know. None of the sixteen studies measured whether automatic control actually leads to fewer cases of lung disease or eye disease, or better survival, in babies on non-invasive support [5]. Keeping oxygen steadier is very likely a good thing, and a careful summary of the evidence (a Cochrane review, considered the gold standard for weighing medical research) confirmed that the machines reduce time spent too low and too high on this kind of support — while noting the evidence is only moderately certain [12]. But "very likely helpful" is not the same as "proven to change outcomes," and good doctors are careful about that difference.

What Researchers Are Working On Next

The review's authors are clear about the next step: a large, carefully designed trial that follows babies for the whole time they need breathing support and measures the outcomes that matter most to families — healthy lungs, healthy eyes, and survival — rather than just how well oxygen stays on target over a single day [5]. Such a study would also look at how to handle those tricky first days of life, compare the different types of gentle breathing support, and pair the technology with good staff training. Until then, automatic oxygen control is best understood as a promising, helpful tool that keeps premature babies closer to their oxygen goal — a genuine improvement in day-to-day care whose full long-term benefits researchers are still working to confirm.

References

  1. Perrone S, Bracciali C, Di Virgilio N, Buonocore G. Oxygen Use in Neonatal Care: A Two-edged Sword. Front Pediatr. 2016;4:143. doi:10.3389/fped.2016.00143 ↩
  2. Hagadorn JI, Furey AM, Nghiem TH, et al. Achieved versus intended pulse oximeter saturation in infants born less than 28 weeks' gestation: The AVIOx study. Pediatrics. 2006;118:1574–1582. doi:10.1542/peds.2005-0413 ↩
  3. BOOST-II Australia and United Kingdom Collaborative Groups. Outcomes of Two Trials of Oxygen-Saturation Targets in Preterm Infants. N Engl J Med. 2016;374:749–760. doi:10.1056/NEJMoa1514212 ↩
  4. STOP-ROP Multicenter Study Group. Supplemental Therapeutic Oxygen for Prethreshold Retinopathy of Prematurity (STOP-ROP): Primary outcomes. Pediatrics. 2000;105:295–310. doi:10.1542/peds.105.2.295 ↩
  5. Kaltsogianni O, Dassios T, Greenough A. Closed-Loop Automated Oxygen Control in Preterm Infants Receiving Non-Invasive Respiratory Support. Children (Basel). 2025;12(11):1528. doi:10.3390/children12111528 ↩
  6. Dani C. Automated control of inspired oxygen (FiO2) in preterm infants: Literature review. Pediatr Pulmonol. 2019;54:358–363. doi:10.1002/ppul.24238 ↩
  7. Urschitz MS, Horn W, Seyfang A, et al. Automatic control of the inspired oxygen fraction in preterm infants: A randomized crossover trial. Am J Respir Crit Care Med. 2004;170:1095–1100. doi:10.1164/rccm.200407-929OC ↩
  8. Dargaville PA, Marshall AP, Ladlow OJ, et al. Automated control of oxygen titration in preterm infants on non-invasive respiratory support. Arch Dis Child Fetal Neonatal Ed. 2022;107:39–44. doi:10.1136/archdischild-2020-321538 ↩
  9. Reynolds PR, Miller TL, Volakis LI, et al. Randomised cross-over study of automated oxygen control for preterm infants receiving nasal high flow. Arch Dis Child Fetal Neonatal Ed. 2019;104:F366–F371. doi:10.1136/archdischild-2018-315342 ↩
  10. Schouten TMR, Abu-Hanna A, van Kaam AH, et al. Prolonged use of closed-loop inspired oxygen support in preterm infants: A randomised controlled trial. Arch Dis Child Fetal Neonatal Ed. 2024;109:221–226. doi:10.1136/archdischild-2023-325831 ↩
  11. Dijkman KP, Delbressine JJ, Dieleman JP, et al. Continuous Application of Closed-Loop FiO2-Control in Extremely Preterm Infants: A Matched Cohort Single-Center Study. Pediatr Pulmonol. 2025;60:e71122. doi:10.1002/ppul.71122 ↩
  12. Stafford IG, Lai NM, Tan K. Automated oxygen delivery for preterm infants with respiratory dysfunction. Cochrane Database Syst Rev. 2023;11:CD013294. doi:10.1002/14651858.CD013294.pub2 ↩