I. GENERAL PRINCIPLES. Both invasive and noninvasive techniques are used to monitor respiratory health in the clinical setting. Although both methods have limitations, monitoring of oxygenation and ventilation is critical to assess respiratory function. Invasive techniques, including blood gas monitoring, allow (i) assessment of pulmonary gas exchange; (ii) determination of hemoglobin oxygen saturation and arterial oxygen content; and (iii) evaluation, although limited, of adequacy of tissue oxygen delivery. Noninvasive techniques may be less specific but allow easier determination of serial measurements and identification of trends.

II. OXYGEN USE AND MONITORING. Causes of hypoxemia include hypoventilation, mismatch of ventilation and perfusion, diffusion impairment, and shunt. Supplemental oxygen is most effective for diffusion abnormalities but ineffective to treat hypoventilation. In emergency situations, sufficient oxygen to abolish cyanosis should be administered. Oxygen monitoring with pulse oximetry should be initiated as soon as possible, and the concentration of oxygen should be adjusted to maintain saturation values within a targeted range. An oxygen blender and pulse oximeter should be used whenever supplemental oxygen is administered. Monitoring of oxygen use is necessary to reduce both hypoxic injury to tissues and to minimize oxidative injury to the lungs or the immature retina of the preterm infant.

A. Arterial blood gas (ABG) measurements. Arterial PO₂ (PaO₂) and PCO₂ (PaCO₂) are direct indicators of efficiency of pulmonary gas exchange in
infants with acute lung disease. PaO₂ measured under steady state conditions from an indwelling catheter is the “gold standard” for oxygen monitoring.

1. Usual values. Most sources consider 50 to 80 mm Hg to be an acceptable target range for newborn PaO₂. Preterm infants who require respiratory support may exhibit wide swings in PaO₂ values. In such circumstances, a single blood gas value may not accurately reflect the overall trend of oxygenation.

2. Sampling. To minimize sampling and dilutional artifacts, ABG samples should be collected in dry heparin syringes that are commercially available for this purpose. Most blood gas analyzers allow determination of blood gas values as well as other whole blood parameters on 0.2- to 0.3-mL samples. Samples should be analyzed within 15 minutes or preserved on ice if sent to a remote laboratory site. Blood gas sampling by percutaneous puncture is used when the need for measurement is infrequent or an indwelling catheter is not available. However, the discomfort of the puncture may result in agitation and a fall in PaO₂ so that the value obtained underestimates the true steady state value.

B. Noninvasive oxygen monitoring provides real-time trend data that is particularly useful in infants exhibiting frequent swings in PaO₂ and oxygen saturation. Noninvasive devices also may reduce the frequency of blood gas sampling in more stable patients.

1. Pulse oximetry is the primary tool for noninvasive oxygen monitoring in newborns. Pulse oximeters provide continuous measurement of hemoglobin oxygen saturation (SpO₂) with a high level of accuracy (±3%) when compared to control values measured by co-oximetry, at least down to the range of 70%.

a. General characteristics. Oximeters depend on different absorption characteristics of oxygenated versus reduced hemoglobin for various wavelengths of light. Differences in transmission of two (usually red and near infrared [IR]) or more wavelengths through tissues with pulsatile blood flow are measured. Using the measured values, the proportion of oxygenated and reduced hemoglobin is calculated and displayed as percent saturation. Modern pulse oximeters can efficiently discriminate artifactual values from valid measurements. Sensitivity of detection of hypoxemia by pulse oximeters is dependent on the averaging time of the oximeters; shorter averaging times detect hypoxemia more sensitively compared to longer averaging times.

b. Disadvantages. Pulse oximetry does not measure the PaO₂ and thus is insensitive in detecting hyperoxemia. Due to the shape of the oxyhemoglobin dissociation curve, if SpO₂ is >95%, PaO₂ is unpredictable. Under such conditions, PaO₂ may be >100 mm Hg. Patient movement and the low amplitude pulse wave of small preterm infants may introduce artifacts that result in false episodes of desaturation, although software modifications have reduced this problem. Other potential sources of artifact include inappropriate sensor placement, presence of high-intensity light (some phototherapy devices), fetal hemoglobin values >50%, and presence of carboxyhemoglobin or methemoglobin.

c. Targeted saturation values. The optimal range of oxygen saturation, especially for preterm infants, remains unknown. One systematic review and meta-analysis of several well-designed multicenter randomized controlled trials concluded that targeting oxygen saturations to 85% to 89% is associated with an increased risk of mortality and necrotizing enterocolitis but a lower risk of retinopathy of prematurity. However, more recent meta-analyses suggest that too much uncertainty exists to make a clear recommendation. In older studies that targeted oxygen saturation values in preterm infants after the immediate newborn period (supplemental therapeutic oxygen for prethreshold retinopathy of prematurity [STOP-ROP,] benefits of oxygen saturation targeting [BOOST]), SpO₂ values >95% in preterm infants receiving supplemental oxygen was associated with increased need for prolonged supplemental oxygen. Determination of an optimal target saturation in preterm infants may be elusive because some babies may be more vulnerable to oxidative injury and others to hypoxia. These vulnerabilities may affect end organs (e.g., the eye, the brain, and the gut) differently and may vary over time with organ maturation.