Practical Pulse Oximetry

Understanding Contemporary Pulse Oximetry

Charlotte Bell, MD Professor and Director of Pediatric Anesthesiology New York University School of Medicine President, Society for Technology in Anesthesia http://www.anestech.org/

The article also available in PDF: 102KB

Published by permission of GASNet Inc. © 2005

Introduction

The modern pulse oximeter has been commercially available for only about two decades. Yet, in that short period of time, this single monitor has become ubiquitous in contemporary anesthesia practice. In fact, its presence is noticed throughout the hospital in acute care settings such as the emergency department and intensive care units. Furthermore, most clinics and offices can produce an oximeter and will routinely record oxygen saturation as the fifth vital sign.

Despite its omnipresence, there has never been publication of evidence-based outcomes to support the use of oximetry. Many would argue that this lack of evidence is also apparent in the use of parachutes – one need not study the usefulness of parachutes for jumping out of airplanes. Certainly health care providers have come to rely on the displayed numeric saturation in a wide variety of medical situations. Understanding how this number is derived, the potential pitfalls and errors, and interpretation of data provided will assist the clinician in maximizing the usefulness of this tool.

History

The first attempts to build oximeters occurred in the 1930s and 40s by Mathes (Germany) and Millikin (England). In the late 1940s, Dr. Earl Wood of the Mayo Clinic first introduced oximeters in the clinical setting. All of these instruments were difficult to set up and calibrate. Hewlett-Packard introduced the first self-calibrating oximeter in the 1970s, using an ear-mounted sensor, fiber optic light source and a heat element to improve local perfusion. The expense and bulkiness of this tool made it unattractive for daily use.

Aoyagi, an engineer with the Nihon Kohden Corporation in Japan was the first to recognize that pulsatile signals measured at two different wavelengths of light could be accurately related to oxygen levels in arterial blood. This oximeter was introduced in the mid 1970s and used a tungsten light source. The pulsatility of arterial blood allows differentiation from absorption of light in other tissue, such as venous blood.

Contemporary pulse oximeters have built on Aoyagi's original theory but use light-emitting diodes (LED), which can generate very large amounts of narrowband light to vastly improve the signal quality. Scott Wilbur at Biox Technology in Colorado (later, Ohmeda and now GE) introduced the first commercially viable pulse oximeter in 1981 using a sensor with LEDs and a photodiode.

How does oximetry work?

Pulse oximeters shine light from two light emitting diodes (LEDs), one at 660 nm and one at 940 nm, through a tissue bed (typically, the finger). Oxygenated hemoglobin and reduced hemoglobin will absorb light differentially from the two LEDs. The photodiode opposite the LED receives the transmitted light. This technique, spectrophotometry, is a tool that allows for the amount of solute (such as hemoglobin) in a liquid to be measured when certain factors, including the path length of light, are known.

However, the path length of light in each patient is variable depending on the size of the finger, toe or ear. Furthermore, light is absorbed by tissue other than arterial blood (venous blood, muscle, other tissue). In order to calculate a useful number, that of hemoglobin saturation in arterial blood, pulse oximeters are uniquely designed.

The oximeter will measure the light from the 660 nm LED and the 940 LED (alternating) from the non-pulsatile tissue bed (direct current or DC component). It will also measure the increased absorbance present during pulsatile flow when the arterial bed expands from the increased systolic volume (alternating current or AC component). The AC component at each wavelength divided by the DC component at each wavelength results in a ratio (R), which is uniquely related to the arterial saturation of hemoglobin.

Using empiric data obtained from human volunteers, pulse oximeter manufacturers have derived graphs that relate the value R to oxygen saturation. Although these algorithms are unique to each manufacturer, an R of 1.0 typically correlates with a saturation display of approximately 85%. Therefore, ambient light or motion, which can add, equally to both the numerator and denominator, may result in an error message in the range of 85% (see limitations below).

Limitations

Pulse oximetry only monitors a single step in the physiologic process of oxygenation. The first step, oxygen delivery to the patient, is monitored by the devices and alarms built into anesthesia machines. The pulse oximeter monitors the amount of oxygen that the lungs deliver to the arterial tree. At the present time, there are no instruments used in daily practice for measuring delivery to tissue or mitochondrial utilization. Nevertheless, oximetry continues to monitor an important step in oxygen uptake, delivery and utilization.

Many users believe that pulse oximetry is a beat-to-beat monitor because of the presence of the plethysmograph which correlates with heartbeat. However, intrinsic to the monitor's ability to calculate arterial saturation is the need to average over time in order to capture pulsatility and differentiate absorption of light in arterial blood versus background absorption in venous blood and tissue. Most oximeters offer the ability to change averaging times. In anesthetized, non-moving patients, a short averaging time will allow for more rapid recognition of changes in saturation. Longer averaging times when the potential for error is present (such as excess ambient light or movement) will minimize error, but also mean that the monitor will not respond as rapidly to changes in saturation.

Time to desaturation and resaturation are also dependent on the location of sensor placement. Central placement (such as the ear) responds quickest, noting desaturation within approximately 10 seconds. Changes in saturation for finger sensors may take 30-60 seconds and up to 90 seconds may be required for sensors placed on the toe.

Oxygenated and reduced hemoglobin are not the only substances that can absorb light at the 660 and 940 wavelengths. Certain biologically useful dyes (methylene blue, indocyanine green, and indigo carmine) are also absorbed at similar wavelengths and can produce false numbers on the oximeter. Nail polish, especially in brown/red tones similar to hemoglobin, may affect the accuracy of the monitors.

Carboxyhemoglobin and methemoglobin are also absorbed at similar wavelengths and the oximeter is incapable of separating these dyshemoglobins. Patients who are at risk for high concentrations of either (such as occurs in patients with smoke inhalation or carbon monoxide poisoning) should be monitored using hemoximetry. Hemoglobin F (fetal) appears to not affect the normal operation of the monitor, probably because the heme moiety is unchanged from hemoglobin A, although the globin part of the molecule is different from adult hemoglobin. For the same reasons, the monitor should function normally with sickle hemoglobin (Hemoglobin S). However, the aggregation of cells during sickle crisis may give false readings.

A particular source of error is noted when the signal-to-noise ratio changes. This can occur if the AC (arterial or pulsatile) component decreases, if the DC (venous or other tissue) component increases, or if artifact affects the AC component. The most common cause of decrease in the AC component occurs in the presence of low perfusion. Although the monitor can amplify the signal received up to 109, the background noise (DC) will also be amplified leading to a false ratio (R) and therefore a false saturation. If very low perfusion occurs, the monitor may be unable to calculate any value. This monitor failure, also called data drop out, may be indicative of extremely low perfusion. Therefore, any situation with data drop out should be an indication to immediately ascertain patient stability with other hemodynamic monitors rather than assuming that the oximeter "isn't working."

An increase in the DC component may be seen with bright ambient light. The oximeter photodiode cannot differentiate the source of the light. For this reason, each LED shines alternately (and then off) to allow the monitor to recognize the light source. When bright light shines on the sensor, the photodiode recognizes light throughout the cycle, again affecting the calculation of AC/DC and the resulting ratio. This source of error can be easily corrected by covering the extremity where the sensor is attached.

Finally, artifact that affects the recognition of the pulsatile component includes venous pulsation or motion of the extremity. Venous pulsations can occur with tricuspid regurgitation, in the presence of an intra-aortic balloon or when the sensor is merely wrapped too tightly around an extremity. Different types of motion (for example, active movement or shivering) have resulted in either data drop out (monitor failure) or error in calculation of the AC/DC ratio.

Developers of newer generation oximeters have attempted to correct the signal-to-noise ratio problems in an effort to minimize data drop out and to improve accuracy. These improvements primarily lie in manufacturing changes made in the machine's algorithm, or the way in which the oximeter calculates the saturation. As a result, these newer oximeters may be able to display an accurate saturation in the face of movement or very low perfusion. The trade-off for this improved performance may be longer averaging times and the need to track an accurate pulse rate to perform the calculations. Some models are also displaying "perfusion index", actually a measure of path length change and a display of pulsatile signal amplitude. This signal may reflect blood flow locally at the sensor site, helping the clinician to place the sensor in an ideal location for optimal function. Current research is looking at the potential for using the signal as a monitor of vascular perfusion or even noninvasive cardiac output.

Future directions

The science of photoplethysmography, on which pulse oximetry is based, holds promise for being able to reveal other solutes in blood with noninvasive monitors. The potential for recognizing dyshemoglobins (carboxyhemoglobin, methemoglobin, sulfhemoglobin, etc) immediately comes to mind. Continuous trending in hemoglobin concentration may also be possible. It is likely that signal strength or perfusion index will be useful for more than just measurement of local pulsatile vascularity for optimal sensor positioning. The plethysmogram signal may be able to offer insight into more than just regional perfusion, but also well-being of operative sites (perfusion adequacy) or of the whole organism (cardiac output).

The author acknowledges contributions to this article from the following review on GASNet: Goldman JM, Pologe JA, Keller JP Jr. Pulse Oximetry. www.gasnet.org/pomtp/index.php

References and additional reading

1. Barker SJ and Tremper KK. Pulse oximetry. In Anesthesia Equipment, Principles and Applications. Eds: Ehrenwerth J and Eisenkraft JB. Mosby-Year Book, Inc. St Louis. 1993.
2. Barker SJ and Tremper KK. Pulse oximetry: applications and limitations. Int Anesthesiol Clin. 1987;25:155-75.
3. Freund PR, Overand PT, Cooper J, Jacobson L, Rosse S, Walker B, Posner KL, Cheney FW. A prospective study of intraoperative pulse oximetry failure. J Clin Monit. 1991;7:253-8.
4. Hamber EA, Bailey PI, James SW, Wells DT, Lu JK, Pace NL. Delays in the detection of hypoxemia due to site of pulse oximetry probe placement. J Clin Anesth. 1999;11:113-8.
5. Kawagishi T, Kanaya N, Nakayama M, Kurosawa S, Namiki A. A comparison of the failure times of pulse oximeters during blood pressure cuff-induced hypoperfusion in volunteers. Anesth Analg. 2004;99:793-6.
6. Lee LA. Domino KB. The closed claims project. Has it influenced anesthetic practice and outcome. Anesthesiol Clin North America. 2002;20:485-501.
7. Malviya S, Reynolds PI, Voepel-Lewis T, Siewert M, Watson D, Tait AR, Tremper K. False alarms and sensitivity of conventional pulse oximetry versus the Masimo SET technology in the pediatric postanesthesia care unit. Anesth Analg.2000;90:1336-40.
8. Pologe JA. Pulse oximery: technical aspects of machine design. Int Anesthesiol Clin. 1987;25:137-53.
9. Pologe JA, Raley DM. Effects of fetal hemoglobin on pulse oximetry. J Perinatol. 1987;7:324-6.
10. Pinero A, Illana J, Garcia-Palenciano C. Canizarese F, Canteras M, Canadillas V, Duran E, Parilla P. Effect on oximetry of dyes used for sentinel lymph node biopsy. Arch Surg. 2004;139:1204-7.
11. Raemer DB, Elliott WR, Topulos GP, Philip JH. The theoretical effect of Carboxyhemoglobin on the pulse oximeter. J Clin Monit. 1989;5:246-9.
12. Tobin RM, Pologe JA, Batchelder PB. A characterization of motion affecting pulse oximetry in 350 patients. Anesth Analg. 2002;94:S54-61.
13. Trivedi NS, Ghouri AF, Lai E, Shah NK, Barker SJ. Pulse oximeter performance during desaturation and resaturation: a comparison of seven models. J Clin Anesth. 1997;9:184-8.
14. Webb RK, Ralston AC Runciman WB. Potential errors in pulse oximetry. II. Effects of changes in saturation and signal quality. Anaesthesia. 1991;46:207-12.

Last updated: 8 June 2005Created
Legal notice	© GE Healthcare 2008 ISSN 1795-6269 (Web) ISSN 1795-6277 (CD)	Webmaster