Anesthesia & Analgesia:
Technology, Computing, and Simulation: Special Article
Respiration Signals from Photoplethysmography
Nilsson, Lena M. MD, PhD
From the Division of Drug Research, Anesthesiology and Intensive Care, Department of Medical and Health Sciences, Linköping University; and Department of Anesthesia and Intensive Care, University Hospital, Linköping, Sweden.
Accepted for publication November 16, 2012.
Published ahead of print February 28, 2013
This Special Article is based on the lecture I gave at the I.A.M.P.O.V. (Innovations and Applications of Monitoring Perfusion, Oxygenation and Ventilation) 2012 Symposium at Yale University, New Haven, CT, June 29 to July 1, 2012.
The author declares no conflicts of interest.
Reprints will not be available from the author.
Address correspondence to Lena Nilsson, MD, PhD, Department of Anesthesia and Intensive Care, University Hospital, S-581 85 Linköping, Sweden. Address e-mail to firstname.lastname@example.org.
Pulse oximetry is based on the technique of photoplethysmography (PPG) wherein light transmitted through tissues is modulated by the pulse. In addition to variations in light modulation by the cardiac cycle, the PPG signal contains a respiratory modulation and variations associated with changing tissue blood volume of other origins. Cardiovascular, respiratory, and neural fluctuations in the PPG signal are of different frequencies and can all be characterized according to their sinusoidal components. PPG was described in 1937 to measure blood volume changes. The technique is today increasingly used, in part because of developments in semiconductor technology during recent decades that have resulted in considerable advances in PPG probe design. Artificial neural networks help to detect complex nonlinear relationships and are extensively used in electronic signal analysis, including PPG. Patient and/or probe-tissue movement artifacts are sources of signal interference. Physiologic variations such as vasoconstriction, a deep gasp, or yawn also affect the signal. Monitoring respiratory rates from PPG are often based on respiratory-induced intensity variations (RIIVs) contained in the baseline of the PPG signal. Qualitative RIIV signals may be used for monitoring purposes regardless of age, gender, anesthesia, and mode of ventilation. Detection of breaths in adult volunteers had a maximal error of 8%, and in infants the rates of overdetected and missed breaths using PPG were 1.5% and 2.7%, respectively. During central apnea, the rhythmic RIIV signals caused by variations in intrathoracic pressure disappear. PPG has been evaluated for detecting airway obstruction with a sensitivity of 75% and a specificity of 85%. The RIIV and the pulse synchronous PPG waveform are sensitive for detecting hypovolemia. The respiratory synchronous variation of the PPG pulse amplitude is an accurate predictor of fluid responsiveness. Pleth variability index is a continuous measure of the respiratory modulation of the pulse oximeter waveform and has been shown to predict fluid responsiveness in mechanically ventilated patients including infants. The pleth variability index value depends on the size of the tidal volume and on positive end-expiratory pressure. In conclusion, the respiration modulation of the PPG signal can be used to monitor respiratory rate. It is probable that improvements in neural network technology will increase sensitivity and specificity for detecting both central and obstructive apnea. The size of the PPG respiration variation can predict fluid responsiveness in mechanically ventilated patients.
Pulse oximetry is one of the most widely used techniques for vital sign monitoring. It is based on the optical technique photoplethysmography (PPG). The PPG signal is complex and composed of different parts, each having the potential to provide clinically useful information. The respiratory synchronous part was first regarded as an artifact disturbing the pulse synchronous variation, which was used in pulse oximetry. However, there are potential clinical applications for the respiratory part as shown by research focusing on these applications. This article gives an overview of factors in the physiologic background and clinical applications of respiratory-derived information using PPG.
The principle of the PPG technique is simple. Light from a light source is scattered and partly absorbed in the tissue. Part of the scattered light emerges through the skin and can be detected by a photosensitive detector. The intensity of the light detected is presented as a plethysmogram.
Blood has a higher absorption coefficient than other tissue components. Changes in tissue blood content can therefore easily be followed by PPG. The stationary part of the plethysmogram, the direct current (DC) part, reflects the optical character of the underlying tissue and the venous blood content. It fluctuates slowly because of varying venous capacity. On top of the DC part, there are small intensity variations, 1% to 5% of the DC level, corresponding to arterial pulsations, called the alternating current (AC) part.
Comparison of PPG signals in absolute numbers between subjects or between measurement sites is not possible, because absorption of light is dependent on local factors. PPG signal strength variables, such as amplitude, must therefore be regarded as arbitrary and only relative comparisons can be made.
PPG is easily measured from the skin. Because of low scattering and absorption, there is an “optical window” in skin and most other soft tissues in the region of 600 to 1300 nm. For these wavelengths, the volume and depth of tissue reached by the light will be large. Most clinical optical devices therefore use visible light or near-infrared light. Pulse oximetry is based on the difference in absorption between oxygenated and deoxygenated hemoglobin at 2 wavelengths, red (approximately 660 nm) and infrared (approximately 940 nm). Studies evaluating the PPG respiratory variation have mostly used the same wavelengths. Kamal et al.,1 who conducted an early review, proposed that wavelengths from 600 to 700 nm are the most effective in PPG. DC changes during respiration and breathholding are similar but not identical at wavelengths from 405 to 1064 nm.2 Also, the shape of the AC component differs between recordings at different wavelengths. Longer wavelengths penetrate deeper into the tissue and collect information from different blood vessels than shorter wavelengths.
HISTORY ON PPG
In 1937, Hertzman and Spealman3 reported that PPG could be used to measure blood volume changes in the finger. Some years later, the signal was split into the AC and DC components. Artifacts caused by movement of the probe against the skin were identified in the 1930s and led to the development of positioning devices. Illumination was another issue, because the early light sources gave light with a wide spectrum without constant intensity. Developments in semiconductor technology involving light-emitting diodes, photodiodes, and phototransistors during recent decades have resulted in considerable advances in PPG probe design.
The most widely used application of PPG is pulse oximetry. Squire4 recognized in 1940 that the transmission of red and infrared light through tissue changed with saturation. Thirty years later, a Japanese engineer, Takuo Aoyagi, was able to compute arterial saturation by looking at the ratio between AC and DC at 2 different wavelengths, red and infrared.5 Pulse oximeters were marketed worldwide in the 1980s. By using respiratory modulation of the cardiovascular system, dynamic parameters reflect volume status and predict the value of a fluid bolus. The AC and DC components are used similarly, with an automated function, pleth variability index (PVI), commercialized in 2007.
FREQUENCY COMPONENTS IN THE PPG SIGNAL
Spectral analysis techniques can be used to characterize oscillations by dividing them into their sinusoidal components. Cardiovascular, respiratory, and neural fluctuations in the PPG signal can all be characterized accordingly. One subdivision is shown in Table 1. The ranges are approximate, because, for example, the respiratory rate can lie outside the range of 11 to 18 per minute.
ARTIFACTS IN THE PPG SIGNAL
PPG measurements are sensitive to patient and/or probe-tissue movement artifacts and are frequent sources of signal interference. Physiologic variations such as vasoconstriction, coughing, a deep gasp, or yawn also affect the PPG signal. Low- and intermediate-frequency vasomotor waves might interfere with respiratory waves. There are several ways within the field of computer signal processing to address the difficulties in separating good quality recordings from artifacts. The use of adaptive digital signal processing to separate the venous and arterial absorbance signals has been successful in displaying saturation during patient motion.6 The PPG pulsatile component has the main component focused on efforts to achieve artifact reduction.
ARTIFICIAL NEURAL NETWORK
Artificial neural networks are extensively used in electronic signal analysis, including PPG. In an artificial neural network, a structure of mathematical units (neurons) is constructed. The units form mathematical algorithms. Neural networks gather their knowledge by detecting the patterns and relationships in data and learn through experience, not from programming. They detect complex nonlinear relationships between dependent as well as independent variables.7
INTERACTION BETWEEN RESPIRATION AND CIRCULATION
Respiration and circulation are interrelated. In the late 1800s, Traube8 and Hering9 reported that the arterial blood pressure varied with respiratory frequency. Respiratory-induced variations in intrathoracic pressure are transmitted to the central veins. During inspiration, venous pressure is decreased and venous return is increased. The filling pressure of the right ventricle increases, and right-ventricular stroke volume increases. Blood is pooled in the lungs and the left stroke volume is decreased. Systolic blood pressure decreases, as does, to a smaller extent, diastolic blood pressure, and heart rate increases. With expiration, the blood flow to the left side of the heart is increased, left stroke volume increases, and in the periphery, both the arterial blood pressure and the peripheral venous pressure increase. With positive pressure ventilation, the opposite circulatory effects occur.10
Respiration causes blood volume variations on both the arterial and venous sides. Respiratory variations are more pronounced in venous return than in stroke volume from the right chamber, and even more than in left stroke volume. This is due to the greater buffering capacity of the right heart chamber and of the pulmonary circulation. The transmural right chamber pressure varies by 32%, whereas the mean arterial blood pressure in the carotid artery only varies by 2%.11 The largest blood volume is found on the venous side of the circulatory system, and because this is a low pressure system, it is sensitive to small variations in pressure.
The venous plexus contains the major fraction of the cutaneous blood volume. Dermal blood flow is controlled via vasoconstriction mediated by the sympathetic nervous system. Arteriovenous anastomoses that are located at acral sites have dense sympathetic innervations.12 Sympathetic efferent activity in skin is coherent with respiration.13 These bursts appear more often during the inspiratory phase and are usually not interrupted during periods of apnea. The magnitude of sympathetic contribution to the respiratory PPG signal is small compared with the dominating changes in venous return and cardiac output.
Another phenomenon observed in the relationship between circulation and respiration is “entrainment” or “phase locking.” This is a process by which an oscillating system adapts to the frequency of another oscillator, leading to synchronization of the 2 systems.14
For continuous or long-term registration of vital signs, noninvasiveness is attractive. Because the PPG signal includes a respiratory component (Fig. 1), there are several possible clinical applications for using the PPG signal.
Respiratory-induced intensity variations (RIIVs) contained in the baseline of the PPG signal have been well documented (Fig. 2),15–19 and several methods for detecting respiratory rates from PPG are based on the RIIV component.20–22 Qualitative RIIV signals may be used for monitoring purposes regardless of age, gender, anesthesia, and mode of ventilation.23,24 Different algorithms can be used to extract the respiratory rate: min-max, peak-to-peak, or pulse shape. An automatic algorithm using wavelet analysis techniques in volunteers with respiratory rates varying from 6 to 19 breaths per minute had a maximal error of 8%.20 Respiratory rate extraction using particle filter can obtain near-continuous breathing rates.25
The amplitude of the RIIV signal varies with ventilatory pressure26 and subsequently with tidal volume (Fig. 3).18,19 Under experimental conditions, tidal volume variations can be followed by RIIVs.17 The amplitude is also affected by respiratory rate; amplitude decreases with increased respiratory rate.17 Automatic algorithms are influenced strongly by the amplitude of the respiratory signal and accordingly a higher respiratory rate leads to a lower signal-to-interference ratio for the RIIVs extracted from the PPG signal. The detecting capability decreases with increasing respiratory rate.27
RIIV is more prominent during positive pressure ventilation than during spontaneous breathing.28 This difference might be an effect of increased pressure variation in the thorax, but also by reduced sympathetic activity from the anesthesia.
The small size of the probe, the possibility of continuous monitoring, and few side effects are some explanations to perform studies that focus on PPG breath detection in infants.29–31 In the study by Johansson et al.,29 motion disturbances were common, but during periods of nondisturbed recordings, the rates of overdetected and missed breaths using PPG were 1.5% and 2.7%, respectively. Those authors used a noncommercial reflection mode probe placed on the thigh. Wertheim et al.31 reliably monitored respiratory rate from pleth data from a commercially available pulse oximeter during natural quiet sleep.
Another respiratory synchronous component is the frequency modulation of the heart rate, known as respiratory sinus arrhythmia. Respiratory rate can be estimated by extracting the PPG-derived heart rate variability. This method is only reliable in cases of spontaneous ventilation and in individuals with sinus cardiac rhythm.
Amplitude modulations of the pulse caused by cardiac stroke volume variations affected by respiration can also be identified by PPG. The amplitudes of both the systole and diastole have been evaluated.27
Pulse transit time is often measured as the time between the R-peak of the electrocardiogram and the onset of the peripheral pulse detected by a pulse oximeter and can be derived from signals present in standard monitoring setups. By using the beat-to-beat respiratory fluctuations in pulse transit time, almost 90% of the breaths can be detected during spontaneous breathing.32
To decrease the error rate of detecting respiratory rhythms, a combination of detection principles can be used. Johansson33 combined RIIV, systolic waveform, diastolic waveform, respiratory sinus arrhythmia, and pulse amplitude in a neural network. For clinical respiration monitoring applications based on PPG signals, the integration of respiratory information from several main variations seems to be a realistic approach.20,27
When breathing movements disappear, the intrathoracic pressure variations that drive the circulatory variation synchronously with respiration are gone. Although the rhythmic RIIV signal disappears, there are fluctuations that can interfere with registration of respiration. Irregular fluctuations of low amplitude can be observed in the peripheral venous pressure and RIIV signals during apnea.18 Also, slower arterial pressure waves persist during apnea.34 During apnea without respiratory movements, the respiratory rhythm of the sympathetic nerve activity seems to persist, influencing the vascular system at a periodicity of 0.1 Hz.34 In spectral analysis during apnea, low frequency content is present in arterial blood pressure and in sympathetic nerve activity.35 Low frequency autonomic rhythms are also unaffected by differences in respiratory rate.35 The detection of central apnea is scarcely evaluated and more studies in the clinical setting are warranted.
When the airway is partly obstructed, an increase in the force of respiratory movements takes place and RIIV is more prominent.28 Pulse transit time is inversely proportional to blood pressure, and the decreases in blood pressure that occur with inspiration correspond to a lengthening in pulse transit time. In patients with obstructive sleep apnea, the size of these inspiratory swings in pulse transit time correlates well with the degree of inspiratory effort.36 Obstructive respiratory events also provoke a relative bradycardia and a PPG pulse amplitude increase followed by a rapid increase in heart rate and an intense vasoconstriction after the end of obstruction.37 In a pilot study, a neural network signal analysis reached a sensitivity of 75% and specificity of 85% in detecting airway obstruction by PPG.38 This lack of sensitivity is a major limitation in a clinical setting. Collection of a larger and more variable set of empirical data as well as improvements in neural network technology might improve future performance.
The RIIV in the baseline of the PPG signal is sensitive to hypovolemia.26 The RIIV amplitude depends on the ventilatory changes in central venous pressure that are transmitted to the peripheral venous pressure. The venous distensibility is increased when the pressure in the veins is decreased, and the amplitude of RIIV is accordingly enlarged.39
The arterial pressure waveform changes with hypovolemia. This has been explained by the sensitivity of left-chamber stroke volume to increased respiratory-induced fluctuations in preload. There is a high correlation, r = 0.85, between systolic pressure variation and pulse oximeter (AC) amplitude variation in patients receiving positive pressure ventilation during changes in blood volume.40 During hypovolemia, the PPG respiratory variation is noticed earlier than that in systolic blood pressure variation, and also lasts longer during resuscitation, indicating that PPG is a more sensitive method for detection of hypovolemia during anesthesia.41
Using frequency domain analysis, Alian et al.42 showed that during progressive hypovolemia in spontaneously breathing subjects, there was a shift in the amplitude density of ear PPG from the cardiac to the respiratory frequencies. When hypovolemia resulted in decreasing blood pressure, there was a shift in ear PPG modulation from autonomic to respiratory.42
Respiratory variations in left-ventricular stroke volume are predictive of a response to volume expansion in patients with regular sinus rhythm under positive pressure ventilation with a tidal volume of 7 to 8 mL/kg. A beat-to-beat measurement of the PPG pulse synchronous waveform amplitude (POP) allows determination of maximal (POPmax) and minimal (POPmin) amplitude over a single respiratory cycle (Fig. 4). The respiratory synchronous variation ΔPOP (%) = 100 × ([POPmax − POPmin]/[(POPmax + POPmin)/2]) can be recorded from a commercial pulse oximeter attached to a finger, if the automatic gain is disabled or held constant to allow RIIV at the baseline to emerge.43 ΔPOP is strongly related to the respiratory variation in pulse pressure, ΔPP.44 ΔPP is calculated over a respiratory cycle in the same way as ΔPOP. Although both are accurate predictors of fluid responsiveness,43 the threshold values for discrimination between fluid responders and nonresponders are not numerically identical between ΔPP and ΔPOP.
Predicting fluid responsiveness is also of importance in nonanesthetized patients. One study conducted in spontaneously breathing volunteers showed that ΔPOP can reflect changes in ventricular preload.45
RIIV is associated with the movement of venous blood and could be used for measurements assessing preload and fluid responsiveness. A theoretical advantage is that venous modulation should be unaffected by cardiac arrhythmia. Future studies could explore this field.46
PLETH VARIABILITY INDEX
A continuous measure of the respiratory AC component, the PVI, has been available for approximately 5 years. Perfusion index (PI) is calculated from the infrared signal, using the formula (AC/DC) × 100 and reflects the amplitude of the pulse oximeter waveform. PVI (%) is a measure of the dynamic change in PI that occurs during 1 or more complete respiratory cycles, calculated as ([PImax − PImin]/PImax) × 100.
PVI During Controlled Ventilation
PVI has been shown to predict fluid responsiveness in mechanically ventilated patients including infants in the operating theater47–50 or in intensive care.51 Patients have mostly been studied under very stable conditions because stimulation such as nociceptive input can induce changes in vasomotor tone that changes PI. PVI is not able to distinguish between changes in PI induced by respiration from other changes. In a recent study, Hood and Wilson49 found that PVI could also accurately predict fluid responsiveness during intraoperative dynamic conditions.
The use of PVI-guided fluid management throughout surgery is associated with lower lactate levels during major abdominal surgery, but this approach does not reduce the length of hospital stay or complication rate.52 PVI is increased when pneumoperitoneum is established and cannot predict fluid responsiveness.53
Signal quality, body temperature, vasoactive drug infusion, depth of anesthesia, presence of nociceptive input or surgical stress, and spontaneous movements may have an impact on PI. PVI is dependent on PI.54,55 Decreases in PI after skin incision are accompanied by increases in the PVI.55 In patients showing low PI measurements, PVI must be interpreted with caution. Higher perfusion states, reaching a PI >4%, improve the discrimination between fluid volume responders and nonresponders.54 The use of norepinephrine negatively impairs the ability to guide intravascular fluid administration by PVI.56
The PVI value depends on the size of the tidal volume and on the application of positive end-expiratory pressure.57 A problem with the use of PVI to guide fluid administration arises from difficulty in finding an optimal cutoff value that can be recommended. Values from 9.5%48 to 14%47,52 have been reported from patients under anesthesia and a threshold value of 17%51 was found in deeply sedated patients in intensive care.
Hood and Wilson49 compared PVI measured from the finger and earlobe, and they found that, although the PI was very low at the earlobe compared with the finger, the PVI could predict fluid responsiveness when measured in the supine position in anesthetized patients after a postinduction period of 5 minutes and before surgery was started but not during surgery. Cutoff values at the 2 measurement positions were similar, 10 and 9.5, respectively.
PVI During Spontaneous Breathing
The irregularity of spontaneous breathing with varying frequency and tidal volumes from breath to breath together with lower intrathoracic pressure variations has led researchers to question the use of PVI as a dynamic indicator for fluid responsiveness during spontaneous breathing. Furthermore, it is probable that factors influencing PI are more frequent in spontaneously breathing individuals. In contrast to the number of studies of PVI during controlled ventilation, data on PVI during spontaneous breathing are very limited. There are only a few studies showing a potential for use of PVI as a dynamic indicator. In one study, the preanesthesia PVI value correlated with the decrease in mean arterial blood pressure after induction of anesthesia.58 In that study, great caution was applied to reduce mental stress to ensure that the patients were relaxed. Passive leg raising gives a rapid and reversible preload challenge similar to an infusion of 300 to 500 mL of colloids. Keller et al.59 showed that a threshold PVI value of >19% was a weak predictor of response to passive leg raising in spontaneously breathing volunteers. Although PVI values increase from 18 to 22 after blood donation or a hemodialysis session, the sensitivity to detect hypovolemia is only 45%.60 The usefulness of PVI during spontaneous breathing must be regarded as questionable.
Noninvasiveness, the possibility of continuous monitoring, and few negative side effects are attractive attributes of PPG. The technique has already spread into different areas of health care. Methods for detecting respiratory rate are often based on the baseline component of the signal, and several publications have shown a high detecting capability. Detection of breaths in adult volunteers had a maximal error of 8%. A combination of respiratory information from other main variations in the signal can increase the robustness and help its clinical use. As breathing movements disappear during central apnea, the phasic PPG respiratory signal vanishes. There is, however, a risk that autonomic nerve activity can influence the PPG signal and interfere with the reading. More studies exploring central apnea detection are required. With an obstructed airway, the amplitude of the PPG signal increases, but at the same time the hemodynamic pattern changes. Obstructive apnea can be traced by PPG with a sensitivity of 75% and a specificity of 85%. Improvements in neural network technology together with combining the respiratory rate and the arterial saturation monitoring might increase sensitivity and specificity for detecting both central and obstructive apnea. An interesting perspective for PPG is the sensitivity for detecting hypovolemia. PVI is a continuous measure of the respiratory modulation of the pulse oximeter waveform and has been shown to predict fluid responsiveness in mechanically ventilated patients. The PVI value depends on the size of the tidal volume and on positive end-expiratory pressure. Cutoff values from 9.5% to 14% have been reported.
Name: Lena M. Nilsson, MD, PhD.
Contribution: This author designed the Special Article, searched the literature, and wrote the manuscript.
Attestation: Lena Nilsson approved the final manuscript.
This manuscript was handled by: Dwayne R. Westenskow, PhD.
1. Kamal AA, Harness JB, Irving G, Mearns AJ. Skin photoplethysmography: a review. Comput Methods Programs Biomed. 1989;28:257–69
2. Spigulis J, Gailite L, Lihachev A, Erts R. Simultaneous recording of skin blood pulsations at different vascular depths by multiwavelength photoplethysmography. Appl Opt. 2007;46:1754–9
3. Hertzman AB, Spealman CR. Observations on the finger volume pulse recorded photo-electrically. Am J Physiol. 1937;119:334–5
4. Squire JR. Instrument for measuring quantity of blood and its degree of oxygenation in the web of the hand. Clin Sci. 1940;4:331–9
5. Aoyagi T, Kishi M, Yamaguchi K, Nakajima S, Hirai H, Takase H, Kuse A. New pulsed-type earpiece oximeter (author’s transl) [in Japanese]. Kokyu To Junkan. 1975;23:709–13
6. Barker SJ, Shah NK. The effects of motion on the performance of pulse oximeters in volunteers (revised publication). Anesthesiology. 1997;86:101–8
7. Patel JL, Goyal RK. Applications of artificial neural networks in medical science. Curr Clin Pharmacol. 2007;2:217–26
8. Traube L. Uber periodische Tatigkeitsanderungen des Vasomotorischen und Hemmungs-Nervenzentrums. Cbl Med Wiss. 1865;56:881–5
9. Hering E. Uber Athembewegungen des Gefasssytems. Sitzungb d k Akad d W math naturw. 1869;60:829–56
10. Macefield VG, Wallin BG. Modulation of muscle sympathetic activity during spontaneous and artificial ventilation and apnoea in humans. J Auton Nerv Syst. 1995;53:137–47
11. Santamore WP, Amoore JN. Buffering of respiratory variations in venous return by right ventricle: a theoretical analysis. Am J Physiol. 1994;267:H2163–70
12. Lossius K, Eriksen M, Walløe L. Fluctuations in blood flow to acral skin in humans: connection with heart rate and blood pressure variability. J Physiol (Lond). 1993;460:641–55
13. Cogliati C, Magatelli R, Montano N, Narkiewicz K, Somers VK. Detection of low- and high-frequency rhythms in the variability of skin sympathetic nerve activity. Am J Physiol Heart Circ Physiol. 2000;278:H1256–60
14. Censi F, Calcagnini G, Lino S, Seydnejad SR, Kitney RI, Cerutti S. Transient phase locking patterns among respiration, heart rate and blood pressure during cardiorespiratory synchronisation in humans. Med Biol Eng Comput. 2000;38:416–26
15. Vegfors M, Ugnell H, Hök B, Oberg PA, Lennmarken C. Experimental evaluation of two new sensors for respiratory rate monitoring. Physiol Meas. 1993;14:171–81
16. Johansson A, Oberg PA. Estimation of respiratory volumes from the photoplethysmographic signal. Part I. Experimental results. Med Biol Eng Comput. 1999;37:42–7
17. Johansson A, Oberg PA. Estimation of respiratory volumes from the photoplethysmographic signal. Part 2. A model study. Med Biol Eng Comput. 1999;37:48–53
18. Nilsson L, Johansson A, Kalman S. Respiratory variations in the reflection mode photoplethysmographic signal: relationships to peripheral venous pressure. Med Biol Eng Comput. 2003;41:249–54
19. Nilsson L, Johansson A, Kalman S. Macrocirculation is not the sole determinant of respiratory induced variations in the reflection mode photoplethysmographic signal. Physiol Meas. 2003;24:925–37
20. Leonard PA, Douglas JG, Grubb NR, Clifton D, Addison PS, Watson JN. A fully automated algorithm for the determination of respiratory rate from the photoplethysmogram. J Clin Monit Comput. 2006;20:33–6
21. Lindberg LG, Ugnell H, Oberg PA. Monitoring of respiratory and heart rates using a fibre-optic sensor. Med Biol Eng Comput. 1992;30:533–7
22. Nakajima K, Tamura T, Miike H. Monitoring of heart and respiratory rates by photoplethysmography using a digital filtering technique. Med Eng Phys. 1996;18:365–72
23. Nilsson L, Johansson A, Kalman S. Respiration can be monitored by photoplethysmography with high sensitivity and specificity regardless of anaesthesia and ventilatory mode. Acta Anaesthesiol Scand. 2005;49:1157–62
24. Nilsson L, Goscinski T, Johansson A, Lindberg LG, Kalman S. Age and gender do not influence the ability to detect respiration by photoplethysmography. J Clin Monit Comput. 2006;20:431–6
25. Lee J, Florian JP, Chon KH. Respiratory rate extraction from pulse oximeter and electrocardiographic recordings. Physiol Meas. 2011;32:1763–73
26. Murray WB, Foster PA. The peripheral pulse wave: information overlooked. J Clin Monit. 1996;12:365–77
27. Li J, Jin J, Chen X, Sun W, Guo P. Comparison of respiratory-induced variations in photoplethysmographic signals. Physiol Meas. 2010;31:415–25
28. Sluiter ER, Rowaan CJ, Dorlas JC, Nijboer JA, Jorritsma FF, Blom JA, Beneken JE. Peak-to-peak detector for the arterial pulsations in the plethysmogram. II. Results of its use during anaesthesia. Med Biol Eng Comput. 1981;19:267–70
29. Johansson A, Oberg PA, Sedin G. Monitoring of heart and respiratory rates in newborn infants using a new photoplethysmographic technique. J Clin Monit Comput. 1999;15:461–7
30. Olsson E, Ugnell H, Oberg PA, Sedin G. Photoplethysmography for simultaneous recording of heart and respiratory rates in newborn infants. Acta Paediatr. 2000;89:853–61
31. Wertheim D, Olden C, Savage E, Seddon P. Extracting respiratory data from pulse oximeter plethysmogram traces in newborn infants. Arch Dis Child Fetal Neonatal Ed. 2009;94:F301–3
32. Johansson A, Ahlstrom C, Lanne T, Ask P. Pulse wave transit time for monitoring respiration rate. Med Biol Eng Comput. 2006;44:471–8
33. Johansson A. Neural network for photoplethysmographic respiratory rate monitoring. Med Biol Eng Comput. 2003;41:242–8
34. Hagbarth KE, Vallbo AB. Pulse and respiratory grouping of sympathetic impulses in human muscle-nerves. Acta Physiol Scand. 1968;74:96–108
35. Badra LJ, Cooke WH, Hoag JB, Crossman AA, Kuusela TA, Tahvanainen KU, Eckberg DL. Respiratory modulation of human autonomic rhythms. Am J Physiol Heart Circ Physiol. 2001;280:H2674–88
36. Pitson DJ, Sandell A, van den Hout R, Stradling JR. Use of pulse transit time as a measure of inspiratory effort in patients with obstructive sleep apnoea. Eur Respir J. 1995;8:1669–74
37. Haba-Rubio J, Darbellay G, Herrmann FR, Frey JG, Fernandes A, Vesin JM, Thiran JP, Tschopp JM. Obstructive sleep apnea syndrome: effect of respiratory events and arousal on pulse wave amplitude measured by photoplethysmography in NREM sleep. Sleep Breath. 2005;9:73–81
38. Knorr-Chung BR, McGrath SP, Blike GT. Identifying airway obstructions using photoplethysmography (PPG). J Clin Monit Comput. 2008;22:95–101
39. Dorlas JC, Nijboer JA. Photo-electric plethysmography as a monitoring device in anaesthesia: application and interpretation. Br J Anaesth. 1985;57:524–30
40. Shamir M, Eidelman LA, Floman Y, Kaplan L, Pizov R. Pulse oximetry plethysmographic waveform during changes in blood volume. Br J Anaesth. 1999;82:178–81
41. Partridge BL. Use of pulse oximetry as a noninvasive indicator of intravascular volume status. J Clin Monit. 1987;3:263–8
42. Alian AA, Galante NJ, Stachenfeld NS, Silverman DG, Shelley KH. Impact of central hypovolemia on photoplethysmographic waveform parameters in healthy volunteers part 2: frequency domain analysis. J Clin Monit Comput. 2011;25:387–96
43. 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
44. Cannesson M, Besnard C, Durand PG, Bohé J, Jacques D. Relation between respiratory variations in pulse oximetry plethysmographic waveform amplitude and arterial pulse pressure in ventilated patients. Crit Care. 2005;9:R562–8
45. Delerme S, Renault R, Le Manach Y, Lvovschi V, Bendahou M, Riou B, Ray P. Variations in pulse oximetry plethysmographic waveform amplitude induced by passive leg raising in spontaneously breathing volunteers. Am J Emerg Med. 2007;25:637–42
46. Alian AA, Shelley KH. Respiratory physiology and the impact of different modes of ventilation on the photoplethysmographic waveform. Sensors (Basel). 2012;12:2236–54
47. Cannesson M, Desebbe O, Rosamel P, Delannoy B, Robin J, Bastien O, Lehot JJ. Pleth variability index to monitor the respiratory variations in the pulse oximeter plethysmographic waveform amplitude and predict fluid responsiveness in the operating theatre. Br J Anaesth. 2008;101:200–6
48. Zimmermann M, Feibicke T, Keyl C, Prasser C, Moritz S, Graf BM, Wiesenack C. Accuracy of stroke volume variation compared with pleth variability index to predict fluid responsiveness in mechanically ventilated patients undergoing major surgery. Eur J Anaesthesiol. 2010;27:555–61
49. Hood JA, Wilson RJ. Pleth variability index to predict fluid responsiveness in colorectal surgery. Anesth Analg. 2011;113:1058–63
50. Renner J, Broch O, Gruenewald M, Scheewe J, Francksen H, Jung O, Steinfath M, Bein B. Non-invasive prediction of fluid responsiveness in infants using pleth variability index. Anaesthesia. 2011;66:582–9
51. Loupec T, Nanadoumgar H, Frasca D, Petitpas F, Laksiri L, Baudouin D, Debaene B, Dahyot-Fizelier C, Mimoz O. Pleth variability index predicts fluid responsiveness in critically ill patients. Crit Care Med. 2011;39:294–9
52. Forget P, Lois F, de Kock M. Goal-directed fluid management based on the pulse oximeter-derived pleth variability index reduces lactate levels and improves fluid management. Anesth Analg. 2010;111:910–4
53. Høiseth LØ, Hoff IE, Myre K, Landsverk SA, Kirkebøen KA. Dynamic variables of fluid responsiveness during pneumoperitoneum and laparoscopic surgery. Acta Anaesthesiol Scand. 2012;56:777–86
54. Broch O, Bein B, Gruenewald M, Höcker J, Schöttler J, Meybohm P, Steinfath M, Renner J. Accuracy of the pleth variability index to predict fluid responsiveness depends on the perfusion index. Acta Anaesthesiol Scand. 2011;55:686–93
55. Takeyama M, Matsunaga A, Kakihana Y, Masuda M, Kuniyoshi T, Kanmura Y. Impact of skin incision on the pleth variability index. J Clin Monit Comput. 2011;25:215–21
56. Biais M, Cottenceau V, Petit L, Masson F, Cochard JF, Sztark F. Impact of norepinephrine on the relationship between pleth variability index and pulse pressure variations in ICU adult patients. Crit Care. 2011;15:R168
57. Desebbe O, Boucau C, Farhat F, Bastien O, Lehot JJ, Cannesson M. The ability of pleth variability index to predict the hemodynamic effects of positive end-expiratory pressure in mechanically ventilated patients under general anesthesia. Anesth Analg. 2010;110:792–8
58. Tsuchiya M, Yamada T, Asada A. Pleth variability index predicts hypotension during anesthesia induction. Acta Anaesthesiol Scand. 2010;54:596–602
59. Keller G, Cassar E, Desebbe O, Lehot JJ, Cannesson M. Ability of pleth variability index to detect hemodynamic changes induced by passive leg raising in spontaneously breathing volunteers. Crit Care. 2008;12:R37
60. Schoonjans A, Forget P, Labriola L, Deneys V, Jadoul M, Pingaut I, De Kock M. Pleth variability index combined with passive leg raising-induced pulse pressure variation to detect hypovolemia in spontaneously breathing patients. Acta Anaesthesiol Belg. 2010;61:147–50
© 2013 International Anesthesia Research Society