This valuable tool has optimized care for newborns with chronic lung disease.

In the management of critically ill neonates, measurement of arterial oxygenation is frequently required to prevent hypoxia or hyperoxia.1 Hypoxia may lead to pulmonary vasoconstriction and pulmonary hypertension. In addition, the resulting alterations in systemic blood flow may lead to neurologic and other organ damage.2 Hyperoxia is associated with oxygen free radical production, which may cause cellular and tissue damage.3 In neonatology, the most common example of this process is seen in preterm infants with immature retinal vascularization. Hyperoxia has been associated with damage to the retina, resulting in the retinopathy of prematurity.4

Oxygen supplementation is critical to the survival of many infants with respiratory disease. In the neonatal intensive care unit (NICU), when oxygen therapy is used alone or in addition to other supportive therapies (such as mechanical ventilation, surfactant replacement therapy, or inhaled nitric oxide), there is a risk of rapid change in the patient’s oxygen saturation, requiring an immediate response from the clinician. An effective method for monitoring arterial oxygenation levels continuously in these patients is a high priority. Direct blood gas sampling from indwelling umbilical or peripheral arterial lines to measure Po2 and oxygen saturation is considered to be the gold standard for accuracy. This method, however, only provides intermittent oxygen monitoring, is invasive, and (in the neonatal population) can lead to significant blood loss and erroneous results if an improper sampling technique is used.5

The ideal monitor would offer hypoxia, hypoxemia, and hyperoxia detection, minimal false alarms, and information storage. It would also be noninvasive, continuous, self-calibrating, and easy to use.

The first widely accepted noninvasive oxygen monitoring system used in infants was the transcutaneous oxygen (tcPo2) monitor. Introduced in the 1970s, the tcPo2 monitor uses electrodes that measure Po2 through the skin. This method has some disadvantages, however, including frequent calibrations, a long stabilization period, slow response time, inaccuracy in older infants with bronchopulmonary dysplasia, and the risk of skin burns from the heated electrode. Hypotension, hypovolemia, hypothermia, and acid-base abnormalities also affect the accuracy of the tcPo2 electrode.6 Despite these limitations, the importance of tcPo2 monitoring in detecting hyperoxia must not be overlooked.7


Pulse oximetry (Spo2) combines spectrophotometry, plethysmography, and microprocessor technology to determine arterial oxygen saturation. Oxygenated hemoglobin and deoxygenated hemoglobin have different light-absorption characteristics. Oxygenated hemoglobin absorbs less light in the red band (600 to 750 nm) and more in the infrared band (850 to 1,000 nm) than deoxygenated hemoglobin.8 Pulse oximeter probes use a light source consisting of two light-emitting diodes (LEDs), one emitting red light (at 660 nm), the other, infrared light (at 940 nm). This light is transmitted across a tissue bed (finger or toe). A photodetector placed opposite the LEDs (Figure 1, page 50) measures the intensity of the transmitted light across the vascular bed. During transillumination of the tissue bed, there are periodic changes in both the length of the light path and the tissue absorbance. This is because of the volume input of pulsatile arterial blood. This pulsatile surge represents the inflow of oxygenated hemoglobin into the tissue bed.8 A plethysmographic waveform is generated (Figure 2). The peaks and troughs of this waveform are detected by measuring the transmitted light many times per second. By dividing the absorbency values at the peaks by those at the troughs, a pulse-added absorbency is obtained. This value is independent of the absorbency characteristics of the nonpulsatile parts of tissue (tissue, bone, and venous and capillary blood). The red-to-infrared ratio of these pulse-added values is translated into a digital signal that is displayed as the saturation percentage, along with the pulse rate.


Studies9,10 have established the efficacy of pulse oximetry in accurately monitoring oxygen saturation in term and preterm infants with respiratory disease. Furthermore, studies11,12 have shown that there is a good correlation between pulse oximetry measurements and laboratory co-oximeter values of 75 percent to 95 percent. This range of oxygenation is encountered in most clinical situations.

The most frequent use for the pulse oximeter is the detection of hypoxemia.13 The relationship between Po2 and oxygen saturation is depicted by the oxygen- hemoglobin dissociation curve (Figure 3). Along the steep portion of the curve, relatively small changes in Po2 lead to large changes in oxygen saturation. Pulse oximetry provides instantaneous information that is sensitive to any change in the infant’s oxygen status in that important range in which there is a risk of hypoxia. As a result, clinical procedures such as weaning from supplemental oxygen, adjusting positive end-expiratory pressure during mechanical ventilation, endotracheal intubation, and endotracheal suction can be carried out using the pulse oximeter to detect early episodes of desaturation that are not yet apparent to the clinician. Continuous pulse oximetry can also detect an acute decompensation in clinical status. For example, a sudden drop in Spo2 may indicate the development of a pneumothorax, pneumopericardium, or blocked endotracheal tube.

Infants being treated with exogenous surfactant, inhaled nitric oxide, and high- frequency ventilation may experience dramatic improvements in oxygenation. Pulse oximetry allows the clinician to titrate the fraction of inspired oxygen (Fio2) rapidly to a predetermined Spo2 value in order to prevent prolonged hyperoxia.

Infants with persistent pulmonary hypertension of the newborn (PPHN) have increased pulmonary vascular resistance that prevents normal pulmonary blood flow. This causes a right-to-left shunting of blood across the patent foramen ovale and patent ductus arteriosus.14 The diagnosis of PPHN is usually confirmed by echocardiography, along with analysis of blood gas samples from preductal and postductal sites. A simpler way to detect this right-to-left shunting is to use two pulse oximeters and measure preductal and postductal Spo2. In one study15 it was found that arterial saturation in the right arm (preductal) of at least 3%above the lower limb (postductal) is evidence of right-to-left ductal shunting. During this critical phase of the infant’s disease, the rapid response time of the pulse oximeter at detecting a right-to-left ductal shunt may lead to earlier treatment of PPHN.

Clinical situations outside the NICU in which pulse oximetry is used in infants include surgery, cardiac catheterization, neonatal transport, outpatient management of chronic lung disease, and sleep studies for the detection of neonatal apnea.

Pulse oximetry depends on adequate peripheral perfusion. In low-cardiac-output states or shock, the oximeter may not detect a pulse waveform. Most pulse oximeters require a pulse pressure of more than 20 mm Hg and a systolic blood pressure greater than 30 mm Hg to operate reliably.7

Pulse oximeters use calibration curves derived from healthy volunteers. Low saturations (from less than 70 percent to 80 percent) cannot be obtained from these calibrations; therefore, they are extrapolated from measurements at higher saturations. This approach tends to overestimate actual oxygen saturation values that are less than 70 percent.7 The accuracy of pulse oximetry is questionable at these low saturations.16

Because the pulse oximeter operates on two wavelengths, it can detect only oxygenated and deoxygenated hemoglobin. It does not take into account other types of hemoglobin, such as carboxyhemoglobin (HbCO), methemoglobin (MetHb), and fetal hemoglobin (HbF). It has been shown there is a slight inaccuracy in Spo2 values in the presence of high HbF levels. This degree of error is acceptable for clinical practice.10 MetHb is produced during inhaled nitric oxide therapy; high levels of MetHb can cause Spo2 readings to stay around 855 independent of actual oxygen saturation. Increasing levels of HbCO cause slight overestimation of the actual oxygen saturation as measured by pulse oximetry.7

Because of the shape of the oxygen-hemoglobin dissociation curve, pulse oximetry is not ideal in preventing hyperoxia. This is of great importance in the care of premature infants who are at risk for development of the retinopathy of prematurity. As oxygen saturation increases to more than 90 percent, the curve flattens (Figure 3, page 50). Along this flat part of the curve, Po2 can increase dramatically with only a small change in oxygen saturation. The measurement of Pao2 is therefore important in these infants. A current practice in NICUs, aimed at preventing hyperoxia in the premature infant, is to set the high alarm limit on the pulse oximeter at 90 percent to 92 percent (the point at which the O2Hb curve starts to flatten). The infant’s Fio2 can then be titrated to keep the Spo2 at around 90 percent, with periodic verification of Pao2.

The extreme sensitivity of these monitors, along with the location of the probes, sometimes allows motion artifact to produce false Spo2 readings and false alarms. This may be a particular problem in an active neonate, rendering the pulse oximeter inaccurate. A solution to this problem is to increase the sampling interval of the monitor. To decrease this type of artifact, synchronization of the pulse waveform with the QRS complex has been developed for newer models.13


Pulse oximetry is a simple and reliable technique for the continuous, noninvasive monitoring of oxygenation in newborn infants and babies with chronic lung disease. This valuable tool has optimized care for these infants. Although there is still a need to know Po2 in some infants, and there are certain limitations that need to be understood clearly for the proper application of this technique, pulse oximetry adds an important degree of control to oxygen management.

Krishna Mullahoo, RRT, is the ECMO coordinator, division of newborn medicine, Montreal Children’s Hospital, in Canada.


  1. Hay W. Pulse oximetry in neonatal medicine. Clin Perinatol. 1991;18:441-472.
  2. Martin RJ, Klaus MH, Fanaroff AA. Respiratory problems. In: Klaus MH, Fanaroff AA, eds. Care of the High Risk Neonate. Philadelphia, Pa: WB Saunders; 1986:171-201.
  3. Korones SB. Complications. In: Goldsmith JP, Karotkin EH, eds. Assisted Ventilation of the Neonate. Philadelphia, Pa: WB Saunders; 1981:245-271.
  4. Silverman W. Retrolental fibroplasia: a modern parable. In: Oliver TK, ed. Monographs in Neonatology. New York, NY: Grune & Stratton;1980.
  5. Thorkelsson T, Hoath S. Accurate micromethod of neonatal blood sampling from peripheral artery catheters. J Perinatol. 1995;15:43-46.
  6. Deckardt R, Steward DJ. Noninvasive arterial hemoglobin oxygen saturation versus transcutaneous oxygen tension monitoring in the preterm infant. Crit Care Med. 1984;12:935-939.
  7. Poets CF, Southall DP. Noninvasive monitoring of oxygenation in infants and children: practical considerations and areas of concern. Pediatrics. 1994;93:737-746.
  8. Alexander CM, Teller LE, Gross JB. Principles of pulse oximetry: theoretical and practical considerations. Anesth Analg. 1989;68:368-376.
  9. Durand M, Ramanathan R. Pulse oximetry for continuous oxygen monitoring in sick newborns. J Pediatr. 1986;109:1052-1056.
  10. Jennis MS, Peabody JL. Pulse oximetry: an alternative method for the assessment of oxygenation in newborn infants. Pediatrics. 1987;79:524-528.
  11. Praud J, Carolfilis A, Bridley F, et al. Accuracy of two wavelength pulse oximetry in neonates and infants. Ped Pulm. 1989;6:180-182.
  12. Fanconi S, Doherty P, Edmonds JF, et al. Pulse oximetry in pediatric intensive care: comparison with measured saturations and transcutaneous oxygen tension. J Pediatr. 1985;107:362-366.
  13. Wahr JA, Tremper KK, Diab M. Pulse oximetry. Respiratory Care Clinics of North America. 1995;1:77-105.
  14. Graves ED III, Redmond CR, Arensman RM. Persistent pulmonary hypertension in the neonate. Chest. 1988;93:638-641.
  15. Macdonald PD, Yu VY. Simultaneous measurement of preductal and postductal oxygen saturation by pulse oximetry in hyaline membrane disease. Arch Dis Child. 1992;67:1166-1168.
  16. Welch JP, DeCesare R, Hess D. Pulse oximetry: instrumentation and clinical applications. Respiratory Care. 1990;35:584-601.