Pulse oximetry

The human eye is capable of recognizing hypoxia only when it is severe enough to cause cyanosis. Even under ideal conditions, trained observers cannot detect hypoxaemia until saturation is less than 80%. The difficulty that physicians have, in detecting hypoxaemia, was exemplified in a study of over 14,000 patients being evaluated at the UCLA Emergency Department. Patients were monitored with pulse oximeters, however the result was given to the physicians only after they completed their initial assessment. Necessary subsequent changes in the diagnostic approach and treatment, resulting from an incorrect initial assessment, were found to be very likely with real saturation values around 89%, while they were less common for lower values of saturation, probably because in such cases physicians were indeed able to detect hypoxaemia without the need of a pulse oximeter.

With the proliferation of pulse oximeters in the 1980s, several researchers have shown that hypoxaemia is in fact much more common than previously suspected, with an incidence ranging between 20% and 82%. The significance of an episodic desaturation on the prognosis of the patient is largely unknown. In patients admitted to a general medical centre, Bowton et al. found that a saturation <90%, lasting for at least 5 minutes, occurred in 26% of patients. The mortality rate in patients presenting desaturation in the first 24 hours of hospitalization tended to be three times higher in the subsequent 4-7 months, compared to those who did not desaturate. Although episodic desaturation may simply be a marker of increased risk, rather than the direct cause of the reduced survival rate, the increased mortality was again confirmed in patients with episodic hypoxaemia when the sample of patients was stratified according to disease severity. It remains unknown whether or not early detection and treatment of episodic hypoxia may influence patient outcome.

The pulse oximeter is still one of the most common monitoring devices used in veterinary anaesthesia. Pulse oximeters measures the oxygen saturation of haemoglobin in arterial blood (oxyhaemoglobin), the peripheral pulse rate and often also give a plethysmographic reading. Although all the instruments have been designed for use in humans, some clinical studies have shown that pulse oximeters give a sufficiently accurate reading of oxygen saturation also in small animals.

The saturation of haemoglobin in arterial blood (SpO₂) is of great use in evaluating the amount of oxygen that reaches the tissues, however alone it is not sufficient. One of the main reasons for the popularity of this monitoring instrument, which is required by law for anaesthesia in many parts of the world (unfortunately not in veterinary medicine), is its ability to circumvent the unreliability (and delay) of cyanosis as a sign of desaturation. Any veterinarian who is interested in anaesthesia should understand not only the advantages and limitations of such a monitoring device, but also the physical principles on which it is based.

SOME INFORMATION ON OPTICS

Haemoglobin is a complex protein structure that contains two pairs of polypeptide chains (two alfa chains and two beta chains). Each chain binds with a haem group which contains one atom of iron (in the ferrous state). This protein complex assumes a different form as the degree of oxygenation changes; consequently, the haemoglobin absorption spectrum also changes with the changing of the molecular spatial arrangement. All things considered, a proportionality exists between the amount of oxyhaemoglobin in a sample and the absorption of an electromagnetic wave that passes through the sample itself.

The absorption of an electromagnetic wave through any transparent solution is governed by the Beer-Lambert law. According to this law, the level of absorption of an electromagnetic wave that passes through a transparent homogeneous solution is equal to the product of the thickness of the substance multiplied by the solute concentration and the so-called molar extinction coefficient, which is a constant.

Each spectrophotometric method correlates the electromagnetic wave absorption capacity of the sample with the concentration of the photoactive solute in the solution itself. However, although the functioning of the pulse oximeter is based on the Beer-Lambert law, the pulse oximeter can only be calibrated empirically. This occurs for two reasons: blood is not a homogeneous fluid, and the absorption of electromagnetic waves assumes a non-linear pattern as the concentration of oxyhaemoglobin changes.

HOW IT WORKS

The pulse oximeter is made up of various units: the probe, a unit containing a microprocessor that processes the pulses from the probe, and finally the display. The probe (Fig. 1) emits two monochromatic electromagnetic waves; two waves are necessary because the unknowns in the equation consist of two forms of Hb: oxidized and reduced haemoglobin. The waves must be at a constant frequency in order for the Beer-Lambert law to be applied. Furthermore, the electromagnetic waves used must have a frequency between 600 nm and 1300 nm, and must be chosen so that the two different forms of Hb absorb them in a substantially different manner.

Early pulse oximeters used frequencies of 660 nm (absorbed to a lesser extent by oxyhaemoglobin than by reduced haemoglobin) and 940 nm (with a reverse absorption). Among the reasons which prompted the selection of these frequencies was the fact that the arrival of LEDs made available a (almost) monochromatic source of electromagnetic waves at low cost and safe for the patient. LED emissions are in reality not exactly monochromatic, however since the absorption curve of oxyhaemoglobin is rather flat at a wavelength of around 940 nm (infrared radiation), such variations do not have a substantial effect on the reading. The selection of 660 nm (red radiation) depends instead on the fact that the difference in absorption between the two forms of Hb, at that wavelength, is maximum.

In addition to the two transmitting diodes, the probe also contains a photodiode that picks up the signal that filters through, or is reflected by the tissue on which the probe is applied. The photodiode is positioned so that it receives the signal from the diodes perpendicularly to its surface, once the signal has crossed any extremity of the patient.

The probe must be constructed so that it exerts a slight but steady pressure on the tissues; in no way should the signal produced by the diode reach the photodiode bypassing the patient. Since the thickness and colour of the extremity on which the probe is applied may vary, it is important that the energy emitted by the diodes is variable in such a way that the photodiode receives a transmissible signal in all conditions.  The probe is connected to a processing unit by means of cables that must be shielded from electrical interferences.

Although the above probe is the one most commonly used, there are other probes that work by picking up waves reflected from the tissues rather than those transmitted directly by the diodes. In these instruments, the diodes and photodiodes are in line, not requiring the "clamp" configuration  typical of probes which pick up the diode transmitted waves; this solution allows to position the probes on more parts of the body (in humans), or in body cavities in animals.

Once amplified and converted from analog to digital, the signal is processed by a microprocessor.

The microprocessor isolates the pulsatile component of the signal (Fig. 2) and separates the plethysmographic components detected by the two diodes. Once the pulsatile component has been isolated, the ratio between the amplitude of the red radiation and the infrared radiation is calculated (red:infrared ratio) (Fig. 3).

For each value of this ratio the system can correlate a saturation value derived from experimental studies on human volunteers and stored in an EPROM.  In fact, for ethical reasons, experimental data have not been collected beyond an 80% saturation; beyond such threshold saturation values have been extrapolated mathematically.

The microprocessor also regulates the functioning of the diodes according to the input signal received, it activates the alarms, stores value trends and controls the display.

IMPLICATIONS OF HAEMOGLOBIN SATURATION AND ADVANTAGES IN CLINICAL USE

Haemoglobin desaturation is a warning sign, signalling the existence of a problem which at some level is preventing the normal transport of oxygen from the mouth of the patient to the tissue where the probe of the pulse oximeter is placed.

The amount of oxygen supply to the tissues in one minute is the result of the following equation:

Flow of O2 = 1.39 x CO x SpO2 x Hb concentration + 0.003 PaO2.

As the equation shows, the saturation of haemoglobin is just one of the three factors involved; this suggests that the presence of a normal saturation is not sufficient to be able to conclude that the oxygen supply to the body is normal. If cardiac output is rarely measured in a veterinary patient under anaesthesia, haemoglobin is instead easily measured, and it is therefore easy to assess how a state of anaemia may consistently reduce the transport of oxygen in spite of a 100% saturation.

The possibility of identifying hypoxia with a continuous, non-invasive, quick and inexpensive monitoring device is of extraordinary importance in anaesthesia, emergency and intensive care, and explains the great success that pulse oximetry has enjoyed over the years. It should also be noted that the saturation level correlates to the oxygen content in a directly proportional manner per volume unit, while the use of partial pressure of oxygen in arterial blood (PaO₂), used for the same purpose, is much less precise in determining the severity of the hypoxia. This is caused by the trend of the haemoglobin dissociation curve. On the other hand, when using high fractions of inhaled oxygen, the readings of the pulse oximeter do not absolutely provide an early identification of the onset of problems, while the PaO₂ resulting from a blood gas analysis proves to be more sensitive (once again due to the dissociation curve). Although the advantages of using a pulse oximeter during anaesthesia on a patient being supplied with high fractions of oxygen are limited, its use in patients with severe respiratory disease is definitely recommended. Furthermore, the heart rate detected with the pulse oxymeter in a patient under anaesthesia and with good peripheral perfusion is more reliable than the rate detected by an ECG, as the latter is more prone to interferences and errors in counting.

Causes of hypoxia are generally related to problems involving the respiratory system (hypoventilation, ventilation-perfusion mismatch, low diffusion, etc.), while diseases of the cardiovascular system, such as right-left shunt and embolism, or even the reduction of the fraction of inhaled oxygen, can cause desaturation.

PLETHYSMOGRAPHY

Together with the percentage of saturation and heart rate, each pulse oximeter should include a plethysmographic tracing. The most important function of this tracing is to show that the instrument is working properly. Should the plethysmographic curve not be similar to the arterial pressure wave, the saturation value cannot be considered reliable. However, the presence of such a wave is in itself no guarantee of a correct SpO₂ value.

A very important factor for the interpretation of plethysmography is to know whether the signal is shown “exactly as it is being collected" or if it is being "normalized". In this latter case, no amplitude changes of the wave are present and, until the device is able to detect the signal, the wave remains equal over time. In some instruments, in which the signal is normalized, the degree of amplification is shown on the display.

The plethysmographic wave is the result of the variation of the transmitted or reflected light energy that reaches the photodiode. Many factors may affect this level of brightness:

• Changes in the amount of blood in the tissue
• Orientation of the erythrocytes
• Concentration of the erythrocytes
• Formation of rouleaux
• Local blood velocity
• Distance between the light source and the photodiode
• Arterial flow and venous flow

It is easy to understand that factors related to local perfusion are of considerable importance in influencing this wave, with the result that plethysmography might be scarcely correlated with systemic perfusion. For example, if the probe is applied to the skin, plethysmography will be greatly influenced by factors affecting the blood flow in the cutaneous vascular bed. As is known, the factors that influence the cutaneous blood flow are numerous and often peculiar to this district of the body, but not to others. In veterinary anaesthesia, the tongue is the most frequently used site for positioning the probe. Compared to the skin, the flow in this district is less prone to local variations and can thus better represent the influence that systemic factors, such as cardiac function and vascular compliance, can exert on peripheral blood flow.

LIMITATIONS OF PULSE OXIMETRY

Pulse oximetry is a rapid, non-invasive, continuous and economical method for the identification of hypoxia. The method, however, has considerable limitations, especially in veterinary practice. It should also be emphasized that this monitoring technique does not provide any information on the capability of the tissues to receive and use the oxygen transported by the arterial circulation.

One of the biggest problems in the use of a pulse oximeter is its sensitivity to patient movements. This, along with the fact that in most of our animals the only area where the probe can be positioned is the tongue, limits its use to patients who are at least deeply sedated, thus excluding most conscious patients who would benefit from such monitoring. Various methods have been devised to prevent motion artefacts, but with little success. An innovative technological approach, called Masimo signal extraction technology (SET™; Masimo Corporation, Mission Viejo, California, USA), has been recently introduced in order to differentiate the true signal from the artefact caused by noise and low perfusion. This technique incorporates new algorithms for the processing of red and infrared light signals, making it possible to more clearly separate signal from noise. When the test was conducted on healthy volunteers undergoing standardized movements, Masimo SET ™ presented much lower error rates (as a percentage of time in which the device showed errors exceeding 5%, 7% and 10%) and dropout rates (defined as the percentage of time in which the saturimeter did not provide any data), compared to Nelcor N-200 and N-3000 oximeters (Nellcor Puritan Bennett, Pleasanton, California, USA), in all test conditions. In 50 operated patients, Dumas et al. observed that the frequency of activation of the alarm of the pulse oximeter was halved with the Masimo SET™ system versus a conventional oximeter (Nellcor N-200). The performance improvement was particularly evident under conditions of sudden, non-rhythmic movements and tremors, with a 22 fold reduction in the signal loss observed over time. Manufacturers use sophisticated softwares to try to identify and remove artefacts and, by adopting "averaging" methods, they try to provide the saturation value as a mean of a series of reliable measurements. This process takes 10-20 seconds, during which abrupt changes in saturation can go undetected. 

The algorithms included in these software programmes heavily condition the reliability and the capability of the device to read in "difficult" clinical conditions. Readings may also be affected by a malpositioning of the probe (a problem called the "penumbra effect"), which is much more difficult to identify by the device itself. The consequence of this problem is typically an underestimation of saturation, caused by a different distance between the two diodes and the tissue under examination. Problems of malpositioning are quite common when using probes that are not designed for use in animals.

External light sources can alter the reading of pulse oximeters. To overcome this drawback, during the read cycle a phase is programmed by the manufacturer during which both diodes are turned off. This makes it possible to measure the level of ambient lighting, so that it can be taken into account during the subsequent reading.

Hypoperfusion and peripheral vasoconstriction can substantially reduce the volume of pulsatile blood, to the extent of creating serious problems in less sophisticated equipment. It has also has been proven that a differential pressure (difference between systolic and diastolic pressure) lower than 20 mmHg seriously affects the accuracy of the measurement and even the ability of the device to give a reading.

The accuracy of a pulse oximeter is based on the assumption that all the haemoglobin in the blood is haemoglobin A. In clinical reality, the blood has variable amounts of other types of haemoglobin that not only do not transport oxygen, but may also make the reading less accurate. Carboxyhaemoglobin and methaemoglobin are normally present in small amounts, but under certain conditions their concentration can be detected. Their concentration can only be measured with a CO-oximeter. The presence of substances such as methylene blue and indocyanine green may affect the accuracy of the reading. It is worth remembering that even under ideal conditions, in the reading range between 70% - 100% the deviation from real saturation is still in the order of 2-3%.

In animals, pigmented areas of the body are not usable for the monitoring of saturation. This appears to be due to the size of the pigment granules, which in animals appears to be larger than in the human species.

RECENT ADVANCES

Morphological analysis of the plethysmographic waveform

In spite of its simple appearance, the waveform of a pulse oximeter is a very complex signal which contains much more information than what we have discussed up to now.As previously mentioned, the signal that reaches the receiver of the probe consists of two components: a continuous absorption (CA) and an alternating absorption (AA) component. The continuous component is the non-pulsatile component of the plethysmographic signal and is inversely proportional to the diffusion and absorption of light by the tissues on which the probe is placed, including the non-pulsatile blood (arterial and venous). The alternating component is the pulsatile component of the plethysmographic waveform. One hypothesis is that the AA component represents the blood in the arterial side of circulation. This wave is also conditioned by the volume of tissue being analyzed, which varies with the variation of blood flow passing through it. The greater the volume of blood (vasodilation), the more light is absorbed. Thus, the less light that passes through the tissue, the lower the current generated by the photodetector. Thus, during systole, the amount of light transmitted through the finger is lower than during diastole, and hence the original plethysmographic signal looks like a mirror image of a waveform of the arterial blood pressure. To make it easier for doctors to interpret the plethysmographic waveform, most devices invert the image on the display. Furthermore, the plethysmographic waveform has an amplification that often self-adjusts to the needs. Consequently, the potential physiological information which may derive from the continuous and the alternating component are lost (not used). It is clear that the plethysmographic waveform provides important physiological information. The variations of the AC and DC components of the waveform of the plethysmograph are correlated to the vasomotor tone [12-14]. The continuous component is also affected by breathing and may contain information about the volemic status of patients [15-16]. Recent advances in pulse oximetry have concentrated on the morphological analysis of the plethysmographic waveform.

A variable which is derived from the plethysmographic waveform is the perfusion index (PI). PI is defined as AC / DC × 100% of the plethysmographic waveform, and is now present in some pulse oximeters. In general terms, PI reflects the peripheral vasomotor tone. Low PI suggests peripheral vasoconstriction (or severe hypovolaemia) and high PI suggests vasodilation. The PI is sensitive to different factors, such as the temperature of the finger, the use of vasoactive drugs, the sympathetic tone of the nervous system (pain, anxiety, and so on) and stroke volume.

Another variable derived from the waveform of the plethysmograph is the plethysmographic variability index (PVI). The PVI is a relatively new parameter currently provided by only one manufacturer of pulse oximeters (Masimo). The PVI quantifies the variability of the plethysmographic waveform caused by breathing, and it has been developed to be a surrogate measure of intravascular volume. It is defined as (PI max - PI min) / PI max × 100%. Several similar parameters (variation of the impulse pressure, delta up/down, systolic variability in arterial pressure) have previously been derived from the waveform of arterial pressure.

Predicting fluid responsiveness using a non-invasive device has offered a new approach to haemodynamic management, previously available only with the use of invasive monitoring [13]. The PVI, as described above, is an easy clinical measurement that is continuously available. Cannesson et al. have reported that a PVI value> 14% is predictive of the response to fluids. Future studies should define the utility of PVI in guiding the management of intravascular volume.

USE OF OXIMETERS FOR REGIONAL ANAESTHESIA

Local anaesthetics cause sympathectomy which, if induced during regional anaesthesia, causes peripheral vasodilation. This vasodilation can be quantified using the waveform of the plethysmograph (PI). The PI has therefore been considered  as a predictive factor for the success of regional anaesthesia. To monitor intravascular PI two recent studies used the vascular tone alterations induced by epinephrine in patients undergoing epidural anaesthesia during general anaesthesia. These studies have shown that the PI was a good indicator for intravascular adrenaline injection under general anaesthesia (intravascular adrenaline injection induced a significant reduction of the PI).

A manufacturer has explored the use of the plethysmographic waveform amplitude to estimate the nociception-antinociception balance in the anaesthetized patient. This plethysmographic index (not commercially available) reflects the activation of the sympathetic nervous system and, potentially, is an indicator of inadequate analgesia. The use of this plethysmographic index as a guide in the use of intraoperative analgesics will be studied in the future.

MULTIWAVELENGTH ANALYSIS

Carboxyhaemoglobin and measurements of methaemoglobin

 As previously explained, commonly used pulse oximeters use two working frequencies in order to determine the arterial saturation, measured as the ratio of oxyhaemoglobin and total haemoglobin.This stems from the theory that there are only two types of haemoglobin that absorb light in the blood: oxyhaemoglobin and reduced haemoglobin. However, other species of haemoglobin, such as  methaemoglobin, carboxyhaemoglobin and deoxyhaemoglobin can absorb  visible electromagnetic waves. A recent development in pulse oximetry platforms uses multiple wavelengths of light to analyze different haemoglobins. The use of eight or more wavelengths makes it possible to measure the concentrations of carboxyhaemoglobin and methaemoglobin in humans in a non-invasive and continuous way [19].

Measuring total haemoglobin                                                                                                                

In 2009, another innovation in multi-pulse oximetry was announced.The use of 12 or more wavelengths of light allows the continuous and non-invasive monitoring of total haemoglobin. Future studies will evaluate the clinical utility of this device.

Improvement in data analysis and  alternative technologies                                                                           

Motion, low perfusion, and optical interference can cause significant errors in pulse oximeters, resulting in loss of data, inaccurate readings, and false alarms.Most doctors agree that motion artefacts and the resulting false alarms were the most significant drawback of pulse oximetry in an intensive care environment. Motion causes a low signal-to-noise ratio, with a decreased SpO2 due to the interference of venous saturation. This venous component is exacerbated by the low perfusion. In addition, all pulse oximeters have been empirically calibrated using desaturation tests in healthy volunteers, in a range of oxyhaemoglobin ranging from 100% to 70%. Therefore the calibration of the device is very rigid and does not take into account other variables peculiar to the system, for example the type of probe being used. Some sensors include a small digital memory chip that contains the specific calibration curves of the sensor.

For pulse oxymeters to be a useful clinical tool, they must provide, in real-time and continuously, accurate measurements in a wide range of values of arterial oxygen saturation during all types of patient movements: continuous, intermittent, aperiodic, rhythmic, and during low perfusion. Significant differences seem to exist in the capacity of currently available technologies to tolerate movement and to achieve these objectives; this said, all recently developed technologies seem to tolerate movement better than conventional technologies. However, in spite of the progress made in these areas, it is likely that future improvements will be relatively small because of the excellence of the measurements currently available.

IMPLICATIONS FOR CLINICAL PRACTICE

Recent advances in pulse oximetry have brought these devices into new and important clinical situations. They have evolved from the relatively simple monitoring of arterial oxygen saturation to technologically advanced instruments capable of measuring different types of haemoglobin and physiological parameters. Currently, pulse oximeters can be useful in haemodynamic management, in regional anaesthesia, in  monitoring haemoglobin and carboxyhaemoglobin and to measure methaemoglobin A. The new generation of pulse oximeters are able to provide accurate measurements in difficult situations, such as low perfusion, the presence of motion artefacts and low arterial oxygen saturation. The plethysmographic waveform is used to analyze some new parameters that could have a significant impact on future clinical practice and, no doubt, new derivated plethysmographic parameters will be developed and studied. Clinical research will be needed to define the usefulness of these technologies and to identify new monitoring opportunities.

Suggested readings

1. Comroe JH, Bothello S. The unreliability of cyanosis in the recognition of arterial anoxemia. Am J Med Sci. 1947;214:1–9.
2. Mower WR, Sachs C, Nicklin EL, et al. Effect of routine emergency department triage pulse oximetry screening on medical management. . Chest. 1995;108:1297–1302.
3. Roe PG, Jones JG. Causes of oxyhaemoglobin saturation instability in the postoperative period. Br J Anaesth. 1993;71:481–487.
4. Moller JT, Svennild I, Johannessen NW, et al. Perioperative monitoring with pulse oximetry and late postoperative cognitive dysfunction. Br J Anaesth. 1993;71:340–347.
5. Bowton DL, Scuderi PE, Haponik EF. The incidence and effect on outcome of hypoxemia in hospitalized medical patients. Am J Med. 1994;97:38–46.
6. Matthews NS, Hartke S, Allen JC Jr. An evaluation of pulse oximeters in dogs, cats and horses. Vet Anaesth Analg. 2003 Jan;30(1):3-14.
7. Moyle TJB Pulse Oximetry. 2002, BMJ Books London UK
8. Jubran A. Pulse oximetry. Tobin MJ (ed). Principles and Practice of Intensive Care Monitoring. New York: McGraw Hill, Inc.; 1998:261–287.
9. Pollard V, Prough DS. Signal extraction technology: a better mousetrap? [Editorial]. Anesth Analg. 1996;83:213–214.
10. Barker SJ, Shah NK. Effects of motion on the performance of pulse oximeters in volunteers. Anesthesiology. 1997;86:101–108.
11. Dumas C, Wahr JA, Tremper KK. Clinical evaluation of a prototype motion artifact resistant pulse oximeter in the recovery room. . Anesth Analg. 1996;83:269–272.
12. Shelley KH, Murray WB, Chang D: Arterial-pulse oximetry loops: a new method of monitoring vascular tone. J Clin Monit. 1997, 13:223–8.
13. Landsverk SA, Hoiseth LO, Kvandal P, Hisdal J, Skare O, Kirkeboen KA: Poor agreement between respiratory variations in pulse oximetry photoplethysmographic waveform amplitude and pulse pressure in intensive care unit patients. Anesthesiology. 2008, 109:849–55.
14. Hummler HD, Engelmann A, Pohlandt F, Högel J, Franz AR: Decreased accuracy of pulse oximetry measurements during low perfusion caused by sepsis: Is the perfusion index of any value?Intensive Care Med. 2006, 32:1428–31.
15. Cannesson M, Attof Y, Rosamel P, Desebbe O, Joseph P, Metton O, Bastien O, Lehot JJ: Respiratory variations in pulse oximetry plethysmographic waveform amplitude to predict fluid responsiveness in the operating room. Anesthesiology. 2007, 106:1105–11.
16. Shelley KH, Jablonka DH, Awad AA, Stout RG, Rezkanna H, Silverman DG: What is the best site for measuring the effect of ventilation on the pulse oximeter waveform?Anesth Analg. 2006, 103:372–7.
17. Mowafi HA, Ismail SA, Shafi MA, Al-Ghandi AA: The efficacy of perfusion index as an indicator for intravascular injection of epinephrine-containing epidural test dose in propofol-anesthetized adults. Anesth Analg. 2009, 108:549–53
18. Mowafi HA, Arab SA, Ismail SA, Al-Ghamdi AA, Al-Metwalli RR: Plethysmographic pulse wave amplitude is an effective indicator for intravascular injection of epinephrine-containing epdirual test in sevoflurane-anesthetized pediatric patients. Anesth Analg. 2008, 2008:1536–41.
19. Barker SJ, Curry J, Redford D, Morgan S: Measurements of carboxyhemoglobin and methemoglobin by pulse oximetry: a human volunteer study. Anesthesiology. 2007, 105:892–7.
20. PettersonMT, Begnoche VL, Graybeal JM: The effect of motion on pulse oximetry and its clinical significance. Anesth Analg. 2007, 105 (Suppl 6):S78–84.
21. Aoyagi T, Fuse M, Kobayashi N, Machida K, Miyasaka K: Multiwavelength pulse oximetry: theory for the future. Anesth Analg. 2007, 105 (Suppl 6):S53–8.