Noninvasive estimation of human tissue respiration with wavelet ...

ISSN 03621197, Human Physiology, 2012, Vol. 38, No. 4, pp. 396–401. © Pleiades Publishing, Inc., 2012. Original Russian Text © A.I. Krupatkin, 2012, published in Fiziologiya Cheloveka, 2012, Vol. 38, No. 4, pp. 67–73.

Noninvasive Estimation of Human Tissue Respiration with WaveletAnalysis of Oxygen Saturation and Blood Flow Oscillations in Skin Microvessels A. I. Krupatkin Priorov Central Research Institute of Traumatology and Orthopedics, Moscow, 127299 Russia Received December 9, 2011

Abstract—Laser Doppler flowmetry, laser spectrophotometry of oxygen saturation, and the fluorescence determination of the NADH/FAD ratio were carried out in 30 subjects in the upper limb skin zones with and without arteriolovenular anastomoses (AVAs). It was demonstrated that the waveletanalysis of oxygen satu ration and blood flow oscillations in microvessels was an efficient approach to noninvasive estimation of the skin oxygen extraction (OE) and oxygen consumption (OC) rates. OE = (SaO2 – SvO2)/SaO2, where SaO2 (%) and SvO2 (%) are the oxygen saturations of arterial and venular blood, respectively. If the cardiac (Ac, perfusion units, p.u.) to respiratory rhythm amplitude (Ar, p.u.) ratio Ac/Ar ≤ 1, SvO2 = SO2. If Ac/Ar > 1, SvO2 = SO2/(Ac/Ar). OC = Mnutr (SaO2 – SvO2) in p.u. ⋅ %O2, where Mnutr is the nutritive blood flow value in p.u. Mnutr = M/SI, where SI is the shunting index of blood flow in microvessels. The perfusion, OE, and OC values were higher in the skin with AVAs than in the skin without AVAs. The perfusion and oxygen satu ration values were more variable in the skin with AVAs. The oxygen diffusing from the tiniest arterioles and capillaries is the most important for tissue metabolism. The contribution of the total perfusion and the oxygen diffusion from arterioles to tissue metabolism increased under the tissue ischemia conditions. Keywords: waveletanalysis, oxygen extraction, oxygen consumption, oscillations of oxygen saturation DOI: 10.1134/S0362119712040068

The concept of tissue respiration includes gas, pri marily oxygen, exchange in the microcirculatory–tis sue systems, and all the redox processes proceeding in cell mitochondria [1, 2]. The values of oxygen extrac tion (expenditure) and oxygen consumption by tissue according to Fick’s principle are the main physiologi cal parameters characterizing these processes [1]. However, despite the key significance in the physio logical processes, the possibilities of the noninvasive study of human tissue respiration are limited. The spectroscopic approaches, especially the method of laser spectrophotometry, which make it possible to noninvasively estimate the oxygen saturation value in microvessels (SO2, in percent), are promising. SO2 is determined as the percentage of oxyhemoglobin in relation to the sum of oxyhemoglobin and reduced hemoglobin (deoxyhemoglobin). At the same time, the SO2 value characterizes the mixed saturation of the blood, which is brought in contact with the nondiffu sive surface of large and mediumsized arterioles, atre riolovenular anastomoses, and the diffusive surface of the aggregate of terminal arterioles, capillaries, and venules of the measurement zone volume, which is normally 1 mm3. It is clear that the SO2 value cannot always be used for the calculation of venular blood sat uration (SvO2,%) and the oxygen arteriolovenular dif

ference, especially in the skin containing arteriolove nular anastomoses, and serve as a marker of tissue oxy genation [3]. Oscillatory processes play an important role in the functioning of the system of tissue microcirculation [4, 5]. Several frequency ranges of blood flow oscilla tions in microvascular networks, each of a different informationregulatory origin, have been identified [5]. A few attempts to analyze oxygen saturation in microvessels were made using the Fourier analysis of saturation oscillations in the rear forearm skin [3, 6], but no results of the waveletanalysis of these oscilla tions, which is more adapted to the evaluation of ape riodic processes, are available in the literature [7, 8]. Nor are there any attempts to use the oscillatory parameters of oxygen saturation to investigate tissue respiration. The goal of this study was to assess oxygen extrac tion and oxygen consumption in human skin using the waveletanalysis of oxygen saturation and blood flow oscillations in microvessels. METHODS Thirty subjects whose ages varied between 25 and 65 years were examined at an ambient temperature of

396

NONINVASIVE ESTIMATION OF HUMAN TISSUE RESPIRATION

21–22°C in a sitting position after a 30min rest. In order to study skin microcirculation in different infor mationregulatory situations, both healthy individuals without concomitant complaints, including cardio vascular and autonomic system diseases (n = 16), and 14 patients with consequences of upper limb injuries were included in the study group. The hemoglobin content in the blood of all the subjects was within the normal limits. Laser Doppler flowmetry (LDF), spectrophotom etry of oxygen saturation with spectral waveletanaly sis of oxygen saturation and blood flow oscillations, as well as fluorescence determination of the skin’s respi ratory chain components, were carried out using an LAKKM hardware complex (Lazma, Russia). Due to the combined probe, the complex allows the blood flow, oxygen saturation, and fluorescence recordings to be carried out in the same surface area of tissue with a volume of about 1 mm3. The recordings were made for 400 s in the skin of the palmar surface of the digit III distal phalanx (the region of the glabrous skin of the acral zones of the upper limb rich in arteriolovenular anastomoses (AVAs) exceptionally dependent on sym pathetic vasomotor innervation) and the lower arm’s medial surface (the skin without AVAs). LDF was car ried out according to the method described in [9] with a 3mm probe in the nearinfrared channel of laser radiation (wavelength, 0.78 μm; depth of probing, about 1 mm), which enabled us to investigate not only nutritive, but also larger microvessels, including AVAs. The microcirculation index (MI), measured in perfu sion units (p.u.), characterizing the total (arteriolar, capillary, and venular) averaged stationary perfusion of microvessels approximated to their average volume blood flow rate, was assessed. The average MI value or М is proportionate to the erythrocyte count and its aver age linear velocity in the volume probed. The sampling rate of the MI value was 40 Hz. The arterial blood SO2 value (SaO2,%) was determined using the LAKKM pulsoxymetry channel. The averaged oxygen saturation (SO2, %) was recorded in skin microvessels with simul taneous laser spectrophotometry. In order to assess the MI and the SO2 oscillatory component, spectral waveletanalysis of oscillations was used (software 3.0.2.246, Lazma, Russia). The choice of waveletanalysis is explained by the fact that the conventionally employed Fourier transform is adjusted to stationary signals whose static characteris tics do not change with time; in addition, it is difficult to extract a signal with low frequency and amplitude. The microcirculatory blood flow is not continuous; it has a wide frequency range and nonstationary polyharmonic characteristics, and waveletanalysis with high resolu tion for the frequency, time, and amplitude is better suited for analyzing its oscillations [7, 8]. The program used a continuous wavelet transform and the Morle complexvalued wavelet as an analyzing wavelet. The following frequency ranges were analyzed: the endothelial NОdependent band (0.0095–0.02 Hz) HUMAN PHYSIOLOGY

Vol. 38

No. 4

2012

397

[10]; the oscillations determined by the lowfrequency rhythm of impulses of sympathetic adrenergic vaso motor fibers directly innervating skin microvessels (0.02–0.046 Hz) [11, 12]; the total myogenic band (0.05–0.015 Hz) is subdivided into two subbands: the 0.05 to 0.069Hz oscillations determined by the influence on myocytes of the neuropeptides of sensory peptidergic nerve fibers (activated in inflammation, including neurogenic inflammation, moderate hyper thermia) [13] and the 0.07 to 0.15Hz oscillations determined by the myogenic activity of microvessel myocytes [14]; the 0.16 to 0.18Hz oscillations deter mined by parasympathetic or sympathetic cholinergic influences; in the limb tissues, they are variable and are of predominantly passive venular character [15], as well as the oscillations formed outside the microcircu latory system, i.e., the 0.2 to 0.4Hz respiratory waves realized via the venular link and reflecting the respira tory modulation of blood outflow [16, 17]; and the 0.8 to 1.6Hz pulse waves spreading along afferent microvessels [16]. The maximal averaged amplitude of oscillations in the corresponding frequency range (A, p.u.) was determined, e.g., An, the amplitude in the neurogenic sympathetically dependent band (p.u.); Am, the amplitude in the total myogenic band (p.u.); Ac, the cardiac rhythm amplitude (p.u.); and Ar, the respiratory rhythm amplitude (p.u.). In the limb tissues, toneforming arteriolarlevel oscillations up to 0.15 Hz are active. The LAKKM fluorescence diagnostic unit was used to determine the skin fluorescence index (FI) and as a reference method for estimating skin oxidative metabolism. FI = NADH/FAD, where NADH is the fluorescence amplitude (in arbitrary units) of the reduced form of nicotinamideadenine dinucleotide with the maximal fluorescence spectrum at a wave length of about 460 nm and FAD is the fluorescence amplitude (in arbitrary units) of oxidized flavinade nine dinucleotide with the maximal fluorescence spectrum at a wavelength of about 550 nm. The FI value is proportionate to the activity of oxidative metabolism [18–20]. Methodology of calculating skin oxygen extraction using laser spectrophotometry with spectral wavelet analysis of SO2 oscillations. The wellknown equation for calculating oxygen extraction (OE) is OE = (CaO2 – CvO2)/СaO2,, where СаО2 and СvО2 are the arterial and venous blood content of oxygen, respec tively (in mL О2/mL) [1]. Considering the small vari ability in the blood oxygen volume, the СО2 value is proportionate to SO2, and the equation assumes the following form: OE = (SaO2 – SvO2)/SaO2. SO2 shows mixed blood saturation from arterioles to venules inclusive, due to which, in order to single out SvO2, it is necessary to exclude the arterial component (SaO2). In the waveletspectrum of oscillations, the SaO2 value is modulated by the cardiac rhythm; and SvO2, by the respiratory (venular) rhythm. If the Ac/Ar ratio ≤ 1, then SvO2 is taken to be equal to SO2 and

398

KRUPATKIN

Table 1. Results of LDF and laser spectrophotometry of the upper extremity skin Parameters

Skin without AVAs

Skin with AVAs

SO2, %

62.3 ± 2.5

71.2 ± 4.6*

SaO2, %

96.4 ± 1.1

96.4 ± 1.1

SvO2, %

52 ± 3.6

38 ± 6.7*

0.46 ± 0.04

0.6 ± 0.08*

OE Mtotal, p.u.

3.6 ± 1

Mnutr, p.u.

1.6 ± 0.5

9 ± 2.4*

OC, p.u. % O2

71 ± 9.8

533.7 ± 86*

FI

1.08 ± 0.07

24.3 ± 5.9*

1.33 ± 0.1*

Note: See “Methods” for explanation of the abbreviations here and in Table 2. * Significant difference (p > 0.05) from skin without AVAs; in the remaining cases, p > 0.05.

OE = (SaO2 – SO2)/SaO2. This variant predominates in most cases of recordings from the skin without AVAs. If the Ac/Ar ratio > 1, SvO2 = SO2/(Ac/Ar) and OE = (SaO2 – SvO2)/SaO2. This variant predomi nates in most cases of recordings from the skin with AVAs. In the cases of resonance oscillations in the active frequency bands, the cardiac and/or respiratory rhythm amplitudes may not be expressed in the spec trum, and the SvО2 calculation has some specific fea tures. If resonance is detected in the zone of afferent arterioles (the ranges of oscillations of endothelial and sympathetic origin), the SO2 value reflects the arteri olar SО2. Since the oscillations in the nutritive link (the total myogenic range [21]) are not expressed, we consider that the maximal oxygen extraction occurs in the process of recording and the SvO2 value approxi mates 0; OE = SаО2/SаO2 = 1. In the cases of resonance of oscillations in the total myogenic and respiratory bands, SvО2 = SO2 and OE = (SаО2 – SO2)/SаO2. If SvO2 = SO2 in the skin zones with AVAs, an additional confirmation by the shunting index (SI) is necessary: SvO2 = SO2/SI, where SI = 1 + (An/Am). In the absence of oscillations in the sympathetically dependent range, An = 0 and SvO2 = SО2. In the presence of these oscillations, the SvO2 value is calculated with the SI. Methodology of calculation of skin oxygen con sumption using LDF and laser spectrophotometry with spectral waveletanalysis of SO2 and blood flow oscilla tions. According to Fick’s principle and with regard to the use of the SO2 value, oxygen consumption (OC) = Q × (SaO2 – SvO2), where Q is the volume blood flow rate (in mL/min × 100 g) and (SaO2 – SvO2) is the oxygenbased arteriovenous difference (as a percent

age of saturation). The surface of the microcirculatory bed of minute arterioles and capillaries makes the main contribution to oxygen diffusion. That is the rea son why the nutritive blood flow value (Mnutr, p.u.) rather than the total perfusion value (M, p.u.) was used as the Q value equivalent. Thus, the oxygen consump tion was calculated using the equation OC = Mnutr × (SaO2 – SvO2), in p.u. ⋅ % О2. Mnutr was calculated according to the equation Mnutr= M/SI, where SI is the shunting index. SI = SI1 + SI2. For skin zones without AVA, SI1 = Amax/Am, where Amax is the maximal amplitude of the dominant oscillations in the active range of frequen cies up to 0.15 Hz (p.u.). For zones with AVAs, SI1 = 1 + (An/Am). SI2 = Apass/Am, where Apass is the maxi mal amplitude of oscillations in the ranges of passive cardiac and respiratory oscillations (p.u.). The SI2 value was included for the SI calculation if only the condition SI2 ≥ 1 held. It should be noted that the OC equation includes the perfusion rate value, due to which the OC value calculated according to Fick’s principle reflects the oxygen consumption rate. The results were statistically processed using the Biostat 4.03 software; the Mann–Whitney test was used to compare two samples; and Spearman’s rank order correlation coefficient (r) and Pearson’s linear correlation coefficient (k) were used to analyze corre lations. RESULTS AND DISCUSSION The examples of the waveletspectrum recordings of oxygen saturation are shown in Fig. 1, and the quantitative parameter values are shown in Table 1. As follows from the data presented, the perfusion, OE, and OC values in zones with AVAs exceeded the same AVAfree skin values. The greatest differences were found in the perfusion values and, as a conse quence, in the OC values calculated with Mnutr. It was in part determined by the choice of the thin skin of the lower arm’s medial surface with lowlevel perfusion as an AVAfree zone. In other skin areas without AVAs, these values may be higher. In addition, the perfusion and saturation values were more variable in the skin with AVAs. This may be related to the involvement of AVAs, absolutely dependent on sympathetic regula tion in the realization of the thermoregulatory func tion and, as a consequence, on the high variability in distribution of the blood flow between the shunting and nutritive pathways. The SvO2 values obtained as a result of the study are consistent with the data in the literature data on oxy gen saturation in the mammalian cutaneous postcap illary venules, in particular, in the hamster cheek pouch devoid of AVAs [22]. Applying the phosphores cence extinction method, the authors of [23] revealed the partial oxygen tension values to be 27.3 ± 2.1 and 25.5 ± 2.2 mmHg for small (about 20 μm in diameter) and larger venules (about 50 μm in diameter), respec HUMAN PHYSIOLOGY

Vol. 38

No. 4

2012

NONINVASIVE ESTIMATION OF HUMAN TISSUE RESPIRATION (a)

399

SO2

2.81 2.46 2.11 1.76 1.41 1.06 0.70 0.35

0.06

0.11

0.30

0.50 0.70 0.901.10 1.60 SO2

0.30

0.50 0.70 0.901.10 1.60

(b) 0.68 0.60 0.51 0.43 0.34 0.26 0.17 0.09

0.06

0.11

Fig. 1. The waveletspectra of oxygen saturation oscillations in skin microvessels with (a) arteriolovenular anastomoses and (b) without them. The ordinate shows the saturation oscillation amplitudes (%); the abscissa shows the oscillation frequencies (Hz). (a) The maximal cardiac rhythm amplitude frequency is 1.19 Hz; the maximal respiratory rhythm amplitude frequency is 0.21 Hz; the maximal cardiac rhythm amplitude (Ac) is 2.54%; the maximal respiratory rhythm amplitude (Ar) is 1.07%; the Ac/Ar ratio = 2.37. (b) The maximal cardiac rhythm amplitude frequency is 1.1 Hz; the maximal respiratory rhythm amplitude frequency is 0.25 Hz; the maximal cardiac rhythm amplitude (Ac) is 0.16%; the maximal respiratory rhythm amplitude (Ar) is 0.5%; the Ac/Ar ratio = 0.32.

tively [23]. The authors of [24] reported 22 ± 3.5 and 30.8 ± 5.5 mmHg for the same vessels [24]. The agree ment between these data and the SvO2 values is about 50%. For objectivity of the significance of the OC calcu lation, the correlations between OC and FI were inves tigated (Table 2). Table 2 shows that Mnutr should be used instead of the total blood flow value for the OC calculation. Only HUMAN PHYSIOLOGY

Vol. 38

No. 4

2012

in these cases, the correlation between OC and FI was significant. This gives evidence that the oxygen diffus ing from small arterioles and capillaries plays an important role in metabolism. The oxygen diffusing from larger arterioles may be involved not only in the tissue oxidative processes but diffuse back into the capillaries or be utilized for the metabolic needs of the vessel wall itself, although the latter pathway is less probable [25].

400

KRUPATKIN

Table 2. Correlation between FI and OC Skin without AVAs

Skin with AVAs

Correlation co efficient 1*

2

1*

2

3

k

0.93; p = 0.003

0.38; p = 0.46

0.85; p = 0.002

0.44; p = 0.2

0.57; p = 0.086

r

0.9; p = 0.005

0.31; p = 0.3

0.88; p = 0.001

0.42; p = 0.2

0.58; p = 0.073

Note: In variant 1, the OC value was calculated according to the methodology described. In variant 2, the total perfusion value Mtotal was used instead of the Mnutr value. In variant 3, the SO2 value was used instead of the SvO2 value. * Significant correlation (p > 0.05).

At the same time, oxygen diffusion from arterioles may be of importance under tissue ischemia condi tions. Note that in the subjects whose Mnutr value in the skin without AVAs was extremely low (less than 1 p.u.), the correlation between FI and OC was also decreased. Only when М was substituted for Mnutr did the correla tion attain the values shown in Table 2. Thus, in the skin without AVAs, the contribution of oxygen diffus ing from arterioles to tissue metabolism increased against the background of the concomitant ischemic component; under these conditions, the M value played the role of Mnutr. Thus, the data in Table 2 confirm the validity of the methodology of SvO2 and OC calculation according to the abovementioned equation. In cases of tissue ischemia and a sharp Mnutr decrease to less than 1 p.u., it is expedient to use the M value for the OC calculation. No correlation between the OE and FI values was found (р > 0.05). However, the correlation between the Mnutr and FI values (k = 0.85; p = 0.02) was shown in the skin without AVAs. CONCLUSIONS Thus, the dynamic approach using the wavelet analysis of oxygen saturation and microvascular blood flow oscillations is efficient for noninvasive estimation of tissue respiration. The equations for the calculation of the extraction of skin oxygen and the oxygen con sumption rate were first proposed and verified. In skin with AVAs, the perfusion, the OE, and OC values sig nificantly exceeded the same AVAfree skin values. The greatest differences were noted in the perfusion values and, hence, in the OC values calculated with the nutritive blood flow value. In skin with AVAs, the perfusion and saturation values were more variable compared to skin without AVAs. The oxygen diffusing from small arterioles and capillaries plays the most important role in tissue metabolism. With the associ ated ischemic component, the contribution of total perfusion and the oxygen diffusing from arterioles to tissue metabolism increased.

REFERENCES 1. Fundamental’naya i klinicheskaya fiziologiya (Basic and Clinical Physiology), Kamkin, A.G. and Kamenskii, A.A., Eds., Moscow: Akademiya, 2004. 2. Son’kin, V.D. and Tambovtseva, R.V., Razvitie myshechnoi energetiki i rabotosposobnosti v ontogeneze (Development of Muscular Energetics and Working Capacity in Ontogenesis), Moscow: Knizhnyi Dom Librokom, 2011. 3. Thorn, C.E., Matcher, S.J., Meglinski, I.V., and Shore, A.C., Is Mean Blood Saturation a Useful Marker of Tissue Oxygenation?, Am. J. Physiol. Heart Circ. Physiol., 2009, vol. 296, no. 5, p. H1289. 4. Bernjak, A., Clarkson, P.B., McClintock, P.V., and Stefanovska, A., LowFrequency Blood Flow Oscilla tions in Congestive Heat Failure and after β1Blockade Treatment, Microvasc. Res., 2008, vol. 76, no. 3, p. 224. 5. Krupatkin, A.I., The Problem of Information Value in Microvascular Networks, Hum. Physiol., 2011, vol. 37, no. 3, p. 312. 6. Thorn, C.E., Kyte, H., Slaff, D.W., and Shore, A.C., An Association between Vasomotion and Oxygen Extraction, Am. J. Physiol. Heart Circ. Physiol., 2011, vol. 301, no. 2, p. H442. 7. Stefanovska, A., Dynamics of Blood Oxygenation Gives Better Insight into Tissue Hypoxia than Aver aged Values, Am. J. Physiol. Heart Circ. Physiol., 2009, vol. 296, no. 5, p. H1224. 8. Tankanag, A.V. and Chemeris, N.K., Adaptive Wave letAnalysis of Human Skin Peripheral Blood Flow Oscillations, Biofizika, 2009, vol. 54, no. 3, p. 537. 9. Lazernaya dopplerovskaya floumetriya mikrotsirkulyatsii krovi. Rukovodstvo dlya vrachei (Laser Doppler Flow metry of Blood Microcirculation. Doctors’ Manual), Krupatkin, A.I. and Sidorov, V.V., Eds., Moscow: Meditsina, 2005. 10. Kvandal, P., Stefanovska, A., Veber, M., et al., Regula tion of Human Cutaneous Circulation Evaluated by Laser Doppler Flowmetry, Iontophoresis, and Spectral Analysis: Importance of Nitric Oxide and Prostaglan dins, Microvasc. Res., 2003, vol. 65, p. 160. 11. Kastrup, J., Buhlow, J., Lassen, N.A., Vasomotion in Human Skin before and after Local Heating Recorded with Laser Doppler Flowmetry. A Method for Induc HUMAN PHYSIOLOGY

Vol. 38

No. 4

2012

NONINVASIVE ESTIMATION OF HUMAN TISSUE RESPIRATION

12.

13.

14.

15.

16. 17.

18.

tion of Vasomotion, Int. J. Microcirc. Clin. Exp., 1989, vol. 8, p. 205. Krupatkin, A.I., The Influence of the Sympathetic Innervation on the Skin Microvascular Tone and Blood Flow Oscillations, Hum. Physiol., 2006, vol. 32, no. 5, p. 584. Krupatkin, A.I., Influence of Sensory Peptidergic Innervation on Human Skin Blood Flow Oscillations in the Range 0.047–0.069Hz, Hum. Physiol., 2007, vol. 33, no. 3, p. 296. Krupatkin, A.I., Blood Flow Oscillations at a Fre quency of about 0.1 Hz in Skin Microvessels Do Not Reflect Sympathetic Regulation of Their Tone, Hum. Physiol., 2009, vol. 35, no. 2, p. 183. Silverman, D.G. and Stout, R.G., Distinction between AtropinSensitive Control of Microvascular and Car diac Oscillatory Activity, Microvasc. Res., 2002, vol. 63, p. 196. Krupatkin, A.I., Cardiac and Respiratory Oscillations of the Blood Flow in Microvessels of the Human Skin, Hum. Physiol., 2008, vol. 34, no. 3, p. 323. Kirilina, T.V., Krasnikov, G.V., Tankanag, A.V., et al., RespirationDependent Blood Flow Oscillations in Human Skin Microcirculation System, Region. Blood Circ. Microcirc., 2009, no. 2, p. 58. Chance, B., Schoener, B., Oshino, R., et al., Oxida tion–Reduction Ratio Studies of Mitochondria in FreezeTrapped Samples. NADH and Flavoprotein

HUMAN PHYSIOLOGY

Vol. 38

No. 4

2012

19.

20.

21. 22.

23.

24.

25.

401

Fluorescence Signals, J. Biol. Chem., 1979, vol. 254, no. 11, p. 4764. Skala, M.C., Riching, K.M., GendronFitzpatrick, A., et al., In vivo Multiphoton Microscopy of NADH and FAD Redox States, Fluorescence Lifetimes and Cellu lar Morphology in Precancerous Epithelia, Proc. Natl. Acad. Sci. U.S.A., 2007, vol. 104, no. 49, p. 19494. Zhuo, S., Chen, J., Jiang, X., et al., Visualizing Extra cellular Matrix and Sensing Fibroblasts in Human Der mis by Nonlinear Spectral Imaging, Skin Res. Technol., 2007, vol. 13, no. 4, p. 406. Krupatkin, A.I., Dynamic Oscillatory Circuit of Regu lation of Capillary Hemodynamics, Hum. Physiol., 2007, vol. 33, no. 5, p. 595. Svensjo, E., The Hamster Cheek Pouch as a Research Model of Inflammation, in Microvascular Research, Shepro, D., Ed., Elsevier Academic, 2006, vol. 1, p. 195. Carvalho, H. and Pittman, R.N., Longitudinal and Radial Gradients of PO2 in the Hamster Cheek Pouch Microcirculation, Microcirculation, 2008, vol. 15, no. 3, p. 215. Bertugliya, S. and Dzhusti, A., Microcirculation and Oxygen Measurement Using the Phosphorescence Extinction Technique, Region. Blood Circ. Microcirc., 2002, no. 1, p.73. Pittman, R.N., Regulation of Tissue Oxygenation, in Morgan and Claypool Life Sciences, 2011, p. 89.