A Better Way to Deliver Oxygen to Preemies in the NICU?

William T. Basco, Jr, MD, MS;  Laura Stokowski, RN, MS


February 15, 2019

Oxygen Delivery Methods for Preterm Infants

Hypoxemia (including intermittent hypoxemia) in premature infants has detrimental effects on neurodevelopment and other sequelae. Not only is the total duration of time spent in the hypoxemic range a concern, but the frequent fluctuations between hypoxemia and hyperoxemia are believed to have detrimental vascular effects.[1]

Methods of providing oxygen to spontaneously breathing premature infants are imperfect, and oxygen delivery is often inconsistent. Nasal cannulae are convenient for care, but even with a constant flow rate, low-flow cannulae permit entrainment of ambient air with inhalation, thus altering the hypopharyngeal oxygen concentration and, consequently, the infant's oxygen saturation level.

Oxygen hoods must frequently be shifted aside for care, resulting in wide fluctuations of inspired oxygen concentration. Pumping oxygen into an incubator is a method abandoned decades ago because of difficulties maintaining consistent ambient oxygen levels during caretaking, as opening the door or the portholes allows room air to mix with the oxygen within the incubator compartment. Regardless of method, when fluctuating inspired oxygen concentrations cause the infant's oxygen saturation to drift outside the target range, nurses must manually adjust the delivery of oxygen, essentially chasing the infant's oxygen saturation levels 24 hours a day.

In a recent study,[2] Travers and colleagues aimed to use technology to replace the human element involved in maintaining the oxygenation status of preterm infants. They used two incubator models with integrated servo-controlled ambient oxygen delivery systems that can be digitally set to administer between 22% and 65% oxygen. When the ambient oxygen concentration drops on opening the incubator for caretaking, the system increases oxygen delivery to maintain the present level.

They compared episodes of hypoxemia in spontaneously breathing preterm infants, serving as their own controls, in two oxygen delivery conditions: the incubator with servo-controlled oxygen delivery and a standard low-flow nasal cannula.

The mean gestational age of the study infants was 27 weeks, and the median birth weight was 933 g. The median age at enrollment was 36 days, and all infants had been off mechanical ventilation or continuous positive airway pressure for at least 48 hours before enrollment. An incubator was required for temperature regulation. Initially, 27 infants were enrolled and 25 completed the study. Each infant was studied for a total of 4 days. Half the infants began with oxygen delivery through a nasal cannula, and the other half began with environmental servo-controlled oxygen. The oxygen delivery conditions were alternated every 24 hours to complete four 24-hour treatment periods.

Study Findings

The main outcome of interest was the number of episodes of intermittent hypoxemia in 24 hours, defined as an oxygen saturation level of 85% or less for at least 10 seconds.

Overall, intermittent hypoxemia occurred more frequently when oxygen was delivered by nasal cannula. The total number of intermittent hypoxemia episodes per 24 hours was 117 in the incubator oxygen environment compared with 130 when the infants were receiving nasal cannula oxygen. The proportion of time spent with an oxygen saturation level of 85% or less was slightly longer during nasal cannula treatment, at 6% compared with 5% during servo-controlled oxygen treatment, a statistically significant difference. The number of severe hypoxemia episodes per 24 hours was slightly higher during nasal cannula oxygen delivery, at 48 episodes compared with 47 episodes during servo-controlled oxygen delivery, and bradycardia occurred at the same rate (6 episodes/24 hours), regardless of oxygen delivery approach. The investigators concluded that oxygen delivery using the servo-controlled incubator delivery system resulted in fewer intermittent hypoxemia events compared with oxygen delivery via low-flow nasal cannula. They hypothesize that this benefit resulted from providing a more stable hypopharyngeal oxygen concentration than that achieved with the nasal cannula alone.


These data are slightly out of the realm of everyday pediatric practice, but I thought the study was worth reviewing because of the strength of the design. In a crossover trial such as this, each patient serves as his or her own control, giving us the ability to compare these two treatment approaches by enrolling only 25 infants. A key take-home message is that the daily number of hypoxemic events in these very premature infants is remarkably high, at greater than 110/24 hours, regardless of treatment approach. Bradycardia episodes were less frequent, but even six bradycardic episodes per day would get the attention of parents and treatment providers. It is also worth noting that this study was completed only among infants who required being placed in an incubator for temperature regulation, and one should be careful not to extend these findings to older infants who would not otherwise need incubators.

The idea of automatic control of oxygen delivery has been floated for years, but has been attempted mostly in research settings. The servo-controlled system used in this study measures and controls only the ambient oxygen concentration; it does not take into account the infant's oxygen saturation level. In this respect, it differs from servo-controlled thermoregulation, which increases or decreases the environmental heat output of the incubator to maintain a preset skin temperature. The ambient oxygen system will, therefore, have limited applicability in the neonatal intensive care unit (NICU). A typical infant born at 27 weeks might spend 10-13 weeks in the NICU. Even without chronic lung disease, fluctuation of oxygen saturation levels can be a problem throughout much of this time. Yet, the servo-controlled system tested here could be used only when the infant is spontaneously breathing and requires both an incubator and supplemental oxygen. The rest of the time (which could be many weeks), caretakers would still be the primary means of controlling the concentration of oxygen delivered.

One benefit of servo-controlled oxygen delivery within an incubator is not having to fit small infants with nasal cannulae, as well as avoiding the skin damage often caused by adhesives to keep the cannula in place. Many parents begin providing frequent and prolonged skin-to-skin ("kangaroo") care as soon as their infants no longer require mechanical ventilation or continuous positive airway pressure. If servo-controlled oxygen delivery were in use, skin-to-skin care would still require the use of a nasal cannula or other oxygen source.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: