Articles in PresS. Am J Physiol Heart Circ Physiol (April 3, 2009). doi:10.1152/ajpheart.01192.2008
1
Is mean blood saturation (SmbO2) a useful marker of tissue oxygenation?
CLARE E. THORN1, STEPHEN J. MATCHER2, IGOR V. MEGLINSKI2,
ANGELA C. SHORE,1
1
Diabetes and Vascular Medicine, Institute of Biomedical and Clinical
Sciences, Peninsula College of Medicine and Dentistry, University of Exeter
and Plymouth, Exeter, United Kingdom.
2
School of Physics, University of Exeter, Exeter, United Kingdom.
Mean blood saturation, marker of tissue oxygenation?
Clare Thorn
Diabetes and Vascular Medicine
Institute of Biomedical and Clinical Sciences
Peninsula College of Medicine and Dentistry
Barrack Road
Exeter EX2 5AX
United Kingdom
Email: [email protected]
Tel: 01392 403081
Fax: 01392 403027
Copyright © 2009 by the American Physiological Society.
2
ABSTRACT
Increasingly we are monitoring the distribution of oxygen through the microcirculation
using optical techniques such as optical reflectance spectroscopy (ORS) and near infrared
spectroscopy.
Mean blood saturation (SmbO2) and Tissue Oxygenation Index (TOI)
measured by these two techniques respectively evoke a concept of the measurement of
oxygen delivery to tissue. This study aims to establish whether SmbO2 is an appropriate
indicator of tissue oxygenation. Spontaneous fluctuations in SmbO2 observed as changes
in concentration of oxyhaemoglobin [HbO2] and deoxyhaemoglobin [Hb] were measured
by ORS in the skin microcirculation of 30 healthy subjects (15 male, age 21-42 years).
Fourier analysis identified two distinctly different spontaneous falls in SmbO2. The first
type of swing, thought to be induced by fluctuations in arterial blood volume, resulted
from the effects of respiration, endothelial, sympathetic and myogenic activity. There was
no apparent change in [Hb]. In contrast, a second type of swing resulted from a fall in
[HbO2] accompanied by a rise in [Hb] and was only induced by endothelial and
sympathetic activity. Thus the same fall in SmbO2 can be induced by two distinct
responses. A “type I” swing does not suggest an inadequacy in oxygen delivery whilst a
“type II” swing may indicate a change in oxygen delivery from blood to tissue. Mean
blood oxygen saturation alone cannot therefore be accepted as a definitive marker of
tissue oxygenation.
KEYWORDS: optical reflectance spectroscopy, microcirculation,
3
INTRODUCTION
The exquisite structure of the microcirculation is designed to deliver nutrients to every
cell in the human body. The development over the last thirty years of non-invasive optical
techniques such as optical reflectance spectroscopy (ORS) (2, 18, 22, 33) and near
infrared spectroscopy (NIRS) (7, 23) has significantly advanced our understanding of the
haemodynamics of the cutaneous (1, 20), muscle (36), cerebral (11, 13, 29, 48) and
gastrointectinal (15, 38) microcirculation. Furthermore, mean blood saturation (SmbO2)
and Tissue Oxygen Index (StO2) derived by ORS and NIRS respectively evoke a concept
that we can also measure the oxygen delivery to the tissue. These parameters are being
increasing identified as indicators of tissue hypoxia (1, 3, 8, 15, 21, 28, 34, 47) which is a
common end product of circulatory shock and a primary target for resuscitation efforts (5,
37).
The primary goal of this study was to establish whether cutaneous tissue oxygen
saturation (SmbO2) derived by optical reflectance spectroscopy is an appropriate indicator
of tissue oxygenation. A further goal was to explore differences in SmbO2 across
cutaneous sites. The evaluation of SmbO2 in tissue by ORS is made by the spectroscopic
quantitative measurement of the concentration of the two chromophores oxyhaemoglobin
[HbO2] and deoxyhaemoglobin [Hb], using the Beer-Lambert law. The law states that the
attenuation spectrum of a tissue is the summation of the individual spectra of the
constituent absorbers multiplied by their individual concentrations which we wish to
measure. A crucial assumption in this optical technique is that algorithms can be
developed to convert the measured changes in light attenuation at various wavelengths
into corresponding changes in [HbO2] and [Hb]. These algorithms have been evaluated in
the literature (24, 31, 33, 43). With visible wavelengths and small source-detector
spacings, ORS can be implemented to study the haemodynamics of the cutaneous
microcirculation less than one mm below the surface. The derived concentration
parameters [HbO2] and [Hb] can also provide an estimate of changes in blood volume in
the skin obtained from the concentration of total haemoglobin [Hbt], being the sum of
[HbO2] and [Hb]. However, it is important to recognise that ORS calculates the mean
values of [HbO2] and [Hb] across all the vessels of the microcirculation of the skin.
4
Therefore the derived SmbO2 is a mean blood oxygen saturation across arterioles,
capillaries and venules.
METHODS
Optical reflectance spectroscopy. The ORS instrumentation used in this study was
developed in the School of Physics, University of Exeter, United Kingdom. Using a
specially designed fibre optic probe, light from a stabilized quartz tungsten halogen white
light source is delivered to the skin via a 200μm diameter quartz fibre. Back scattered
light is then collected in an array of 18 individual 50μm diameter detector fibres
encapsulated in the in-house built probe. Six detector fibres form a concentric circle
250μm from the light source, six fibres have a source-detector spacing of 400μm and a
further six fibres have a source-detector spacing of 800μm. The receiving fibres deliver
the reflected light into a SPEX 270M grating spectrometer and the spectrum is recorded
by a Wright Technologies MK II CCD camera equipped with an EEV CCD 30-11-5-219
charged coupled device camera of 1024 x 256 pixel format. Spectra ranging from 470nm
to 1120nm are acquired in 0.05 seconds, processed using a custom-written Windowsbased software package and analyzed using Matlab (The MathWorks Inc., Natick,
Massachusetts, USA). The protocol was designed to replicate the experimental technique
suggested by Merschbrock and colleagues (33) in which reflectance spectra are obtained
specifically from the blood volume in the skin and not the surrounding interstitium.
Initially a reference spectrum is taken with pressure applied to the probe sufficient to
occlude the superficial blood vessels. This is observed as a disappearance of the
attenuation peaks of blood. The spectrum is attributed to attenuation of the incident light
by absorption and scattering in the interstitium. From the Beer-Lambert law the
attenuation of light in optical densities (OD) is the log10 of the incident light intensity
divided by the transmitted light intensity. Taking the reference spectrum of the
interstitium as the incident light and subsequent ‘interstitium and blood-filled’ spectra as
the transmitted light intensity then attenuation can be derived for the blood alone. To
calculate [HbO2] and [Hb] in the volume of skin sampled by the ORS probe a modified
Beer-Lambert law was implemented to account for attenuation due to scattering. A
simple 4-component multi-linear regression algorithm uses 100 wavelengths between 500
5
and 600nm to produce a least squares fit of each attenuation spectrum for blood alone.
This takes the form of a modified Beer Lambert law:
[
]
A blood (λ ) = ⎛⎜ [Hb ]⋅ α Hb ( λ ) + HbO 2 ⋅ α HbO (λ ) ⎞⎟ ⋅ B ⋅ d + Go + G1 (λ )
2
⎝
⎠
Ablood(λ) is the wavelength dependent attenuation of light intensity by the blood alone;
[Hb], [HbO2] are μmolar concentrations of deoxyhaemoglobin and oxyhaemoglobin in
(μmol per litre of skin tissue); αHb(λ), αHbO2(λ) are wavelength dependent specific
absorption coefficient of HbO2 and Hb (μmolar-1cm-1); d is source-detector spacing (cm);
B is differential pathlength factor; Go is constant scattering loss term for the blood; G(λ)
is linear wavelength dependent scattering loss term for the blood. The combined term B.d
is the true optical path which the scattered light has travelled. Although d is simply the
geometrical distance between the points where the light enters and leaves the skin, the
differential pathlength factor B , being dependent on the amount of scattering in the
medium, is difficult to determine. Techniques for measuring B given large ( >1cm)
values for d, have been described in the literature (9, 10) however a value for B is not
required in this study. The parameter B is a function of wavelength, however in this study
the modified Beer Lambert law is implemented over a narrow spectral range from 500 to
600nm. With a small source-detector of 250µm it is considered that within this spectral
range a derived value of B would not have a significant dependence upon wavelength.
The parameter of interest in this study is the mean blood saturation given by SmbO2 =
[HbO2] x 100/([HbO2]+[Hb]). Hence, taking the ratio of the chromophore concentrations,
B.d appears in the top and bottom of the equation and therefore cancels out.
Experimental protocols. The investigation was carried out with the approval of the
School of Physics Research Ethics Committee, University of Exeter. Thirty healthy
volunteers (14 F, 16 M: age 21-42 years: 27.5 ± 6.5 (mean±SD); blood pressure 110/65 ±
10/8) took part in the study. All volunteers gave written informed consent and SmbO2 was
measured either in dorsal skin of the left forearm or in the skin of the left dorsal index
finger, distal from the knuckle over the proximal phalanx. There were 15 studies on the
6
forearm and 23 studies on the index finger with 8 subjects investigated at both sites. All
subjects meet the inclusion criteria defined as non-smokers, no family history of
cardiovascular disease, normotensive and no medication. Caffeinated drinks were not
permitted for two hours prior to the study. Subjects lay supine and were acclimatized for
1/2 hour to a room temperature of 24 - 260C. After a 30 minute period of acclimatisation
which ensured the subject was adjusted to room temperature and data had stabilised,
baseline data were recorded.
Data analysis.
The ORS data from the source-detector spacing of 250μm, considered to sample from a
skin depth to ~200μm (32), were analysed. Periods of 500 seconds artefact-free baseline
data were analysed in Matlab. Examples of the signals of [HbO2], [Hb], [Hbt] and the
derived SmbO2 are given in Figures 1 and 2. Trends in the data below 0.005 Hz identified
from the moving average of a 200 second window were removed and the signals were
multiplied by a Hamming window to reduce spectral leakage. Amplitude spectra for
[HbO2], [Hb] and [Hbt] were derived by fast Fourier analysis (Matlab) with a range of
0.005 Hz to 3.3 Hz. Each subject’s raw data were manually examined to identify patterns
in fluctuation in [HbO2], [Hb], [Hbt] and hence SmbO2. Quantifiable changes in each
individual SmbO2 swing for each subject were identified by the parameters illustrated
schematically in Figure 3. These were used to identify differential markers between type
I and type II swings.
Statistics
Data presented in the text are mean ± standard deviation (SD). Statistical analysis was
performed by SPSS version 13.0 (SPSS Inc., Chicago, USA) with data sets tested for
normality using the Kolmogororv-Smirnov test. For normally distributed data, group
comparisons were made by parametric unpaired t-test. For non-normally distributed data
and sample sizes less than 12, non-parametric statistics were applied.
7
RESULTS
The mean blood saturation SmbO2 derived by ORS was analysed from the 23 studies in
the finger and 15 studies in the forearm. The mean SmbO2 was higher in the finger (63 ±
11%) compared to the forearm (45 ± 10%) p<0.01, unpaired t-test. The SmbO2 was not
found to be constant over time but oscillated with an amplitude (mean ± SD) of 16 ± 9.%
in the finger and 11 ± 6 % in the forearm averaged across all subjects. Fourier analysis
and manual inspection of the data revealed low frequency oscillations in SmbO2 at less
than 0.05 Hz. These oscillations in saturation were associated with two different patterns
of changes in tissue blood volume [Hbt]. We have defined these as “type I” and “type II”
swings in SmbO2. A type I swing in SmbO2 is characterized by changes in total blood
volume [Hbt] predominantly arising from changes in [HbO2] with little change in [Hb]
(Figure 1). A type II swing in SmbO2 is characterized by an anti-phase swing in [HbO2]
and [Hb] with some small changes in total blood volume [Hbt] synchronised
predominantly with changes in [HbO2] (Figure 2). The mean and standard deviation for
the six parameters defined to describe a SmbO2, swing (t1, t2, g1, g2, min, max) are
presented in Table 1. These results are averaged separately for type I and type II swings
across all subjects. None of the parameters t1, t2, g1, g2 and min and max were
significantly different when measured in a type I SmbO2 swing compared to a type II
SmbO2 swing (unpaired t-test, p>0.05). There were 5 subjects demonstrating both type I
and type II swings and again there was no significant difference between any of these
parameters (paired Wilcoxon sign-ranked p>0.05). There were no differences in type I
swing parameters (t1, t2, g1, g2 and min and max) from the forearm skin compared to
finger skin parameters however, type II swing differ according to cutaneous site. The
time for SmbO2 to rise to a maximum (t2) was significantly shorter in the finger than in
the forearm (p=0.019); SmbO2 rose at a faster rate (g2) in the finger than in the forearm
(p=0.013) and the SmbO2 falls in the finger was faster (g1) than in the forearm (p<0.047).
A schematic diagram to illustrate these differences is shown in Figure 4.
The oscillation frequencies of type I and type II swings in finger and forearm determined
by Fourier analysis, and the number of subjects with type I and type II swings of that
8
frequency are shown in Figure 5. These data from all 30 subjects demonstrate that all
type II swings occur at < 0.04 Hz where as type I swings can occur up to 0.32 Hz.
DISCUSSION
This study has evaluated the mean blood saturation SmbO2 across arterioles, capillaries
and venules in the cutaneous microcirculation of dorsal finger and forearm as measured
by ORS. The SmbO2 was lower in the forearm (45 ± 10%) compared to the finger (63 ±
11%) and was not constant over time but fluctuated. Although the range of skin
temperatures across subjects varied from 28.0 to 33.9 0C the maximum change of skin
temperature within a study was only 0.9 0C. These SmbO2 values are similar to other ORS
studies of normal healthy human skin which identified that the saturation of haemoglobin
in dermal vessels were greater in acral skin of the extremities compared with proximal
sites. For example, Caspary and colleagues reported mean(SD) oxygen saturation of
65(11.9)% for volar forearm and 90(3.9)% for tip of index finger (4) whilst others report
mean oxygen saturations of 72.9(12.2)% for lateral forearm (33). All studies were
performed with a skin temperature of 26-310C, though unfortunately the researchers did
not specify the source-detector spacings. In contrast, different depths at one site, as
opposed to different sites, reveal similar saturations. Mean(SD) oxygen saturations above
the medial malleolus of 39.6(19.5)% to a depth of 200μm and 31.0(12.4)% to a depth of
1mm have been reported at a room temperature of 200C but with no skin temperature
reported (18). The skin above the medial malleolus of the leg is taken to be non-acral and
the saturations derived are lower than those reported for fingers and forearms.
This variation across cutaneous sites could arise from the different mechanisms involved
in blood flow regulation in acral and non-acral skin. In the finger there are arteriovenous
anastomoses (AVAs) that directly connect arterioles to venules and are under high basal
sympathetic vasoconstrictor drive (17, 30). Vasodilation produces shunting of arterial
blood through the AVAs into the venous compartment. As SmbO2 measures a mean
saturation across arterioles, capillaries and venules then this increase in [HbO2] in the
venules of the finger will produce a higher SmbO2. Conversely, the time-averaged mean
SmbO2 in the forearm would be lower due to the lack of AVAs. It is also possible that a
9
lower cutaneous blood flow in the forearm reduces the washout of [Hb] or results in
increased extraction consequently lowering the time-averaged SmbO2. A reduced
cutaneous flow in the forearm may also explain alterations we have seen in the type II
swings in the finger and forearm where the SmbO2 appears to change at a slower rate and
over a longer time period in the forearm than in the finger.
Preliminary visual inspection of [HbO2], [Hb] and [Hbt] identified two distinct responses
inducing swings in SmbO2. A type I swing reflecting a change in total blood volume [Hbt]
predominantly arising from a change in [HbO2] with little change in [Hb]. A type II
swing reflects anti-phase oscillations in [HbO2] and [Hb] with some small changes in
total blood volume synchronised predominantly with changes in [HbO2]. Research often
focuses on the study of saturation of blood in different tissues as an index of tissue health.
However, it is easy to incorrectly extrapolate that a tissue containing a high fraction of
HbO2 i.e. a high saturation, is well supplied with oxygen. For oxygen exchange from
blood to tissue to have taken place then [HbO2] must fall and [Hb] must rise i.e a
reduction in saturation. Oxygen saturation by definition relies on constant blood volume
if it is to be interpreted as a measure of oxygen delivery. However, optical reflectance
spectroscopy and near infrared spectroscopy have successfully demonstrated the
spontaneous fluctuations in blood volume observed as vasomotion (12, 39, 45). The
authors believe that the present observational study, without the involvement of
interventions, demonstrates the impact that these spontaneous fluctuations in blood
volume may have upon the derivation of mean blood saturation. It has been shown that a
measured change in SmbO2 can occur either as a type I or type II swing and that there
appears to be no significant difference in the appearance of these changes in SmbO2 in
either the finger or forearm. However, a difference occurs in the change in [Hb]
suggesting that the type I and type II swings may indicate contrasting levels of oxygen
exchange. For type I swings the [Hb] remains relatively constant indicating a relatively
constant metabolic demand and oxygen usage, despite the apparent swing in SmbO2. This
is, therefore, presumably due to a change in arterial blood volume arising from either
arterial vasodilation and constriction or capillary recruitment. This observation contrasts
with the literature that identifies 70% of total blood volume to be contained in the venous
10
compartment. For type I swings to be linked with changes in the venous rather than the
arterial network we would expect to observe coherent changes in [HbO2] and [Hb]. This
we do observe in blood volume changes related to the respiratory frequency and can be
seen as the high frequency component in [HbO2] and [Hb] in Figure 1A. Indeed, the
respiratory-induced oscillations of near infrared absorption in tissues has been presented
as a method of measuring venous saturation (14). Francheschini and colleagues
emphasise the importance of verifying that [HbO2] and [Hb] oscillate in phase at the
respiratory frequency. However, our study has identified oscillations in [HbO2] without
changes in [Hb] and this could only occur in the venous compartment if the mean blood
saturation was 100%. We therefore suggest that that type I swings arise from changes in
arterial blood volume alone. Although using a different technique and tissue type,
research using positron emission tomography has demonstrated that cerebral blood
volume changes during hypercapnia and hypocapnia are caused by changes in arterial
blood volume without changes in venous and capillary blood volume (19).
In a type II swing the [HbO2] and [Hb] appear to change in anti-phase. This indicates a
change in oxygen extraction arising from either altering oxygen demand or a change in
cutaneous blood flow. This complex relationship between parameters affecting the mean
blood saturation has been studied in cerebral oxygenation (TOI) by NIRS (44). Tachtsidis
derived an equation demonstrating the direct relationship between TOI with the
arterial/venous volume ratio, oxygen consumption and the indirect relationship with
cerebral blood flow. It is based upon a simple two compartment, arterial and venous
system and the Fick principle. Although in this study the tissue volume is small and
cutaneous blood flow heterogeneous, this model can still provide insight into the complex
nature of the parameters that can affect mean blood saturation. The equation can be
rewritten to describe the parameters affecting SmbO2 as determined by ORS in the
cutaneous microcirculation:
⎞
⎛ Vv
⎟⎟
S mb O 2 = SaO2 − ⎜⎜
⎝ Va + Vv ⎠
CMRO2
⎛
⋅ ⎜⎜
−2
⎝ k ⋅ SBF ⋅ Hb ⋅ 10
[
]
⎞
⎟⎟ × 100%
⎠
11
Where SaO2 is arterial saturation (%); Va is arterial blood volume per unit mass of tissue
(l.g-1); Vv is venous blood volume per unit mass of tissue (l.g-1); CMRO2 is oxygen
consumption (ml of oxygen min-1); k (Hüfner’s number) is oxygen binding capacity per
1g haemoglobin (ml); SBF is skin blood flow (ml.min-1) and Hb is the haemoglobin
concentration (g of Hb.dl-1of blood).
The possible origins of the type I and type II swings can be identified in this equation.
With the type I swing the arterial blood volume Va is seen to fluctuate such that arterial
vasodilation i.e. an increase in Va, induces an increase in SmbO2, conversely arterial
vasoconstriction induces a fall in SmbO2. This can occur without any alteration to the
oxygen consumption CMRO2. The type II swings are more complicated as a change in
oxygen extraction can be induced by either a change in cutaneous blood flow or a change
in oxygen demand. As the type I and type II swings have been shown to be
indistinguishable by their change in SmbO2 this leads to the hypothesis that the
measurement of SmbO2 may not by itself be a definitive marker of the level of oxygen
supply to a tissue. It may in fact be more useful to take a measure of [Hb] as an indication
of tissue oxygen exchange.
The data have also been analysed to investigate the frequencies of the oscillations in
SmbO2. The histograms in Figure 5 illustrate how type II swings only occur at frequencies
< 0.04 Hz whilst the type I swings occur at a range of frequencies up to 0.32 Hz. Our
understanding of the nature and origin of rhythmic fluctuations in the cardiovascular
system has been advanced with the development of optical tools such as laser Doppler
fluximetry (LDF) (27, 40) and intra-vital microscopy (6, 35, 41). These techniques have
studied the spontaneous rhythmic vasodilation and vasoconstriction of blood vessels,
known as vasomotion, through the measurement of oscillations in the concentration of
moving red blood cells (flux) or vessel diameter respectively. The research has revealed
five characteristic frequencies of oscillation in the cardiovascular system at 1 Hz, 0.3 Hz,
0.1 Hz, 0.04 Hz and 0.01 Hz (42) with a possible sixth frequency at less than 0.01 Hz
(26). The origins of these oscillations are thought to arise from endothelial activity at <
0.02 Hz; sympathetic activity between 0.02 and 0.05 Hz and myogenic activity between
12
0.05 and 0.15 Hz (42). This study is, to our knowledge, the first to analyze oscillations in
cutaneous mean blood saturation. It suggests that the type II swings are likely to arise
from endothelial and sympathetic activity (< 0.04 Hz oscillations); whilst the type I
swings appear to be derived not only from endothelial and sympathetic activity but also
from myogenic and respiratory influences. The effect of respiration is thought to act
predominantly upon the venous compartment (14) and would therefore alter the
arterial/venous volume ratio to induce type I swings, though this would also induces
some change in [Hb]. The passive contraction of the venous compartment induced by
inspiration could increase the arterial/venous volume ratio and hence increase SmbO2.
Similarly with the myogenic response arising from transluminal pressure changes, this
might dominate the arterial compartment and thus alter arterial/venous volume ratio. It is
interesting that type II swings in SmbO2 possibly linked with oxygen extraction appear
only to be induced by endothelial and sympathetic activity. Theoretical modeling of
vasomotion has suggested that these oscillations might ensure adequate oxygen delivery
to all tissues (25, 46) with the largest effect seen for low frequencies < 0.05 Hz related
specifically to endothelial and sympathetic activity (16).
In summary, optical reflectance spectroscopy has clearly demonstrated oscillations in the
mean oxygen saturation SmbO2 that occur spontaneously in the dorsal skin of the forearm
and index finger. Although these changes in SmbO2 appear to be similar in all cases, there
are in fact two identifiably separate responses inducing these fluctuations. Type I
oscillations in SmbO2 result from changes in total blood volume predominantly arising
from changes in [HbO2] with little change in [Hb]. This may be purely due to arteriolar
vasodilation and not reflect changes in tissue oxygen delivery and usage. These swings
are influenced by endothelial, sympathetic, myogenic and respiratory effects. Type II
oscillations in SmbO2 result from an anti-phase swing in [HbO2] and [Hb] with some small
changes in total blood volume synchronised predominantly with changes in [HbO2]. Type
II swings may reflect changes in oxygen supply and/or usage and appear to result only
from endothelial and sympathetic activity. The time-averaged mean oxygen saturation
SmbO2 of finger skin (63%) is significantly greater than that of forearm skin (45%). This
study in human skin concludes that changes in SmbO2 do not necessarily reflect changes
13
in the surrounding tissue oxygen consumption. Fluctuations in SmbO2 can be induced by
two separate responses that cannot be distinguished unless changes in [HbO2] and [Hb]
are also monitored. Although this preliminary study is based upon observational
evidence, it highlights the imperative need for further theoretical and experimental
modeling of this parameter before it can be routinely used as a definitive clinical marker
of oxygen delivery to tissue. The simultaneous measurement of tissue perfusion using
laser Doppler fluximetry will provide further information on the haemodynamics of type
I and type II swings providing it is interrogating the same tissue volume. It is suggested
that caution should be taken when attempting to extrapolate clinically significant
information from measurements of mean blood saturation alone.
14
ACKNOWLEDGMENTS
We would like to thank all volunteers in the School of Physics, University of Exeter. C.
Thorn was supported by a grant from the Darlington Trust. Present addresses of authors
are: S.Matcher, The Kroto Institute, University of Sheffield, UK; I. Meglinski, Cranfield
Health, Cranfield University, UK.
15
1.
Beckert S, Witte MB, Konigsrainer A, and Coerper S. The impact of the
Micro-Lightguide O2C for the quantification of tissue ischemia in diabetic foot ulcers.
Diabetes Care 27: 2863-2867, 2004.
2.
Benaron DA, Parachikov IH, Cheong WF, Friedland S, Rubinsky BE, Otten
DM, Liu FW, Levinson CJ, Murphy AL, Price JW, Talmi Y, Weersing JP,
Duckworth JL, Horchner UB, and Kermit EL. Design of a visible-light spectroscopy
clinical tissue oximeter. J Biomed Opt 10: 44005, 2005.
3.
Bykowski J, Kollias N, and LaMuraglia GM. Evaluation of peripheral arterial
occlusive disease and postsurgical viability using reflectance spectroscopy of skin.
Microvascular Research 67: 152-158, 2004.
4.
Caspary L, Thum J, Creutzig A, Lübbers DW, and Alexander K. Quantitative
reflection spectrophotometry: spatial and temporal variation of Hb oxygenation in human
skin. Int J Microcirc Clin Exp 15: 131-136, 1995.
5.
Cohn SM, Nathens AB, Moore FA, Rhee P, Puyana JC, Moore EE, and
Beilman GJ. Tissue oxygen saturation predicts the development of organ dysfunction
during traumatic shock resuscitation. J Trauma 62: 44-54; discussion 54-45, 2007.
6.
Colantuoni A, Bertuglia S, and Intaglietta M. Observation of vasomotion
through an implanted window, enabling comparison between anesthetized and wake
states. Am J Physiol 246: H508-H517, 1984.
7.
Cope M. The development of a near infrared spectroscopy system and its
application for non invasive monitoring of cerebral blood and tissue oxygenation in the
newborn infant.: University of London, 1991.
8.
Creteur J. Muscle StO2 in critically ill patients. Curr Opion Crit Care 14: 261266, 2008.
9.
Delpy DT, Cope M, Vanderzee P, Arridge S, Wray S, and Wyatt J. Estimation
of Optical Pathlength through Tissue from Direct Time of Flight Measurement. Physics
in Medicine and Biology 33: 1433-1442, 1988.
10.
Duncan A, Meek JH, Clemence M, Elwell CE, Tyszczuk L, Cope M, and
Delpy DT. Optical Pathlength Measurements on Adult Head, Calf and Forearm and the
Head of the Newborn-Infant Using Phase-Resolved Optical Spectroscopy. Physics in
Medicine and Biology 40: 295-304, 1995.
11.
Edwards AD, Wyatt JS, Richardson CE, Delpy DT, Cope M, and Reynolds
EOR. Cotside measurement of cerebral blood flow in ill newborn infants by near infrared
spectroscopy Lancet 2: 770-771, 1988.
12.
Elwell CE, Springett R, Hillman E, and Delpy DT. Oscillations in cerebral
haemodynamics. Implications for functional activation studies. Adv Exp Med Biol 471:
57-65, 1999.
13.
Ferrari M, De Marchis C, Giannini I, Nicola A, Agostino R, Nodari S, and
Bucci G. Cerebral blood volume and haemoglobin oxygen saturation monitoring in
neonatal brain by near infrared spectroscopy. Adv Exp Med Biol 200, 1986.
14.
Franceschini MA, Boas DA, Zourabian A, Diamond SG, Nadgir S, Lin DW,
Moore JB, and Fantini S. Near-infrared spiroximetry: noninvasive measurements of
venous saturation in piglets and human subjects. J Appl Physiol 92: 372-384, 2002.
16
15.
Friedland S, Soetikno R, and Benaron D. Reflectance spectrophotometry for
the assessment of mucosal perfusion in the gastrointestinal tract. Gastrointest Endosc
Clin N Am 14: 539-553, ix-x, 2004.
16.
Goldman D and Popel AS. A computational study of the effect of vasomotion on
oxygen transport from capillary networks. J Theor Biol 209: 189-199, 2001.
17.
Hale AR and Burch GE. The arteriovenous anastomoses and blood vessels of
the human finger. Medicine 39: 191-240, 1960.
18.
Hanna GB, Newton DJ, Harrison DK, Belch JJ, and McCollum PT. Use of
lightguide spectrophotometry to quantify skin oxygenation in a variable model of venous
hypertension. Br J Surg 82: 1352-1356, 1995.
19.
Ito H, Ibaraki M, Kanno I, Fukuda H, and Miura S. Changes in the arterial
fraction of human cerebral blood volume during hypercapnia and hypocapnia measured
by positron emission tomography. J Cereb Blood Flow Metab 25: 852-857, 2005.
20.
Ives CL, Harrison DK, and Stansby GS. Prediction of surgical site infections
after major surgery using visible and near-infrared spectroscopy. Adv Exp Med Biol 599:
37-44, 2007.
21.
Ives CL, Harrison DK, and Stansby GS. Tissue oxygen saturation, measured by
near-infrared spectroscopy, and its relationship to surgical-site infections. Br J Surg 94:
87-91, 2007.
22.
Ji S, Chance B, Nishiki K, Smith T, and Rich T. Micro-light guides: a new
method for measuring tissue fluorescence and reflectance. Am J Physiol 236: C144-156,
1979.
23.
Jobsis FF. Noninvasive infrared monitoring of cerebral and myocardial oxygen
suffiency and circulatory parameters Science 198: 1264-1267, 1977.
24.
Johns M, Giller CA, and Liu HL. Determination of hemoglobin oxygen
saturation from turbid media using reflectance spectroscopy with small source-detector
separations. Appl Spectrosc 55: 1686-1694, 2001.
25.
Kislukhin VV. Regulation of oxygen consumption by vasomotion. Math Biosci
191: 101-108, 2004.
26.
Kvandal P, Landsverk SA, Bernjak A, Stefanovska A, Kvernmo HD, and
Kirkeboen KA. Low-frequency oscillations of the laser Doppler perfusion signal in
human skin. Microvasc Res 72: 120-127, 2006.
27.
Kvernmo HD, Stefanovska A, Bracic M, Kirkeboen KA, and Kvernebo K.
Spectral analysis of the laser Doppler perfusion signal in human skin before and after
exercise. Microvasc Res 56: 173-182, 1998.
28.
Leung FW. Endoscopic reflectance spectrophotometry and visible light
spectroscopy in clinical gastrointestinal studies. Dig Dis Sci 53: 1669-1677, 2008.
29.
Leung TS, Aladangady N, Elwell CE, Delpy DT, and Costeloe K. A new
method for the measurement of cerebral blood volume and total circulating blood volume
using near infrared spatially resolved spectroscopy and indocyanine green: application
and validation in neonates. Pediatr Res 55: 134-141, 2004.
30.
Lossius K, Eriksen M, and Walloe L. Fluctuations in blood flow to acral skin in
humans: connection with heart rate and blood pressure variability. J Physiol 460: 641655, 1993.
17
31.
Matcher SJ, Elwell CE, Cooper CE, Cope M, and Delpy DT. Performance
Comparison of Several Published Tissue near-Infrared Spectroscopy Algorithms. Anal
Biochem 227: 54-68, 1995.
32.
Meglinsky IV and Matcher SJ. Modelling the sampling volume for skin blood
oxygenation measurements. Med Biol Eng Comput 39: 44-50, 2001.
33.
Merschbrock U, Hoffmann J, Caspary L, Huber J, Schmickaly U, and
Lübbers DW. Fast wavelength scanning reflectance spectrophotometer for noninvasive
determination of hemoglobin oxygenation in human skin. Int J Microcirc Clin Exp 14:
274-281, 1994.
34.
Olsson C and Thelin S. Regional cerebral saturation monitoring with nearinfrared spectroscopy during selective antegrade cerebral perfusion: diagnostic
performance and relaionship to postoperative stroke. J Thorac Cardiovasc Surg 131: 371379, 2006.
35.
Oude Vrielink HH, Slaaf DW, Jan Tangelder G, Weijmer-Van S, and
Reneman RS. Analysis of vasomotion waveform changes during pressure reduction and
adenosine application. Am J Physiol 258: H29-H37, 1990.
36.
Pereira MI, Gomes PS, and Bhambhani YN. A brief review of the use of near
infrared spectroscopy with particular interest in resistance exercise. Sports Med 37: 615624, 2007.
37.
Puyana JC and Pinsky MR. Searching for non-invasive markers of tissue
hypoxia. Crit Care 11: 116, 2007.
38.
Ramirez FC, Padda S, Medlin S, Tarbell H, and Leung FW. Reflectance
spectrophotometry in the gastrointestinal tract: limitations and new applications. Am J
Gastroenterol 97: 2780-2784, 2002.
39.
Reinhard M, Wehrle-Wieland E, Grabiak D, Roth M, Guschlbauer B,
Timmer J, Weiller C, and Hetzel A. Oscillatory cerebral hemodynamics--the macro- vs.
microvascular level. Journal of the neurological sciences 250: 103-109, 2006.
40.
Rossi M, Carpi A, Galetta F, Franzoni F, and Santoro G. The investigation of
skin blood flowmotion: a new approach to study the microcirculatory impairment in
vascular diseases? Biomed Pharmacother 60: 437-442, 2006.
41.
Rucker M, Strobel O, Vollmar B, Spitzer WJ, and Menger MD. Protective
skeletal muscle arteriolar vasomotion during critical perfusion conditions of
osteomyocutaneous flaps is not mediated by nitric oxide and endothelins. Langenbecks
Arch Surg 388: 339-343, 2003.
42.
Stefanovska A and Bracic M. Reconstructing cardiovascular dynamics. Control
Eng Practice 7: 161-172, 1999.
43.
Stratonnikov AA and Loschenov VB. Evaluation of blood oxygen saturation in
vivo from diffuse reflectance spectra. Journal of Biomedical Optics 6: 457-467, 2001.
44.
Tachtsidis I. Experimental measurements of cerebral haemodynamics and
oxygenation and comparisons with a computational model: a near-infrared spectroscopy
investigation: PhD Thesis, University College London, 2005.
45.
Thorn CE, Matcher SJ, and Shore AC. Combined optical and near infrared
reflectance measurements of vasomotion in both skin and underlying muscle. Proc SPIE
6434: 643421-643428, 2007.
46.
Tsai AG and Intaglietta M. Evidence of flowmotion induced changes in local
tissue oxygenation. Int J Microcirc Clin Exp 12: 75-88, 1993.
18
47.
Wang TL and Hung CR. Role of tissue oxygen saturation monitoring in
diagnosing necrotizing fascilitis of the lowwer limbs. Ann Emerg Med 44: 222-228, 2004.
48.
Wolf M, Ferrari M, and Quaresima V. Progress of near-infrared spectroscopy
and topography for brain and muscle clinical applications. J Biomed Opt 12: 062104,
2007.
19
Figure 1. An example of spontaneous swings in total blood volume [Hbt] in dorsal finger
resulting from changes in [HbO2] predominantly and not [Hb] (A). The subsequent
swings in SmbO2 identified as type I swings (B).
Figure 2. An example of out of phase swings in [HbO2] and [Hb] in dorsal finger (A).
The subsequent swings in SmbO2 identified as type II swings (B). There are also some
small changes in total blood volume [Hbt] synchronised predominantly with changes in
[HbO2].
Figure 3. Schematic diagram of parameters determined in SmbO2 swings where t1 is time
for SmbO2 to fall to minimum (s); t2 is time for SmbO2 to rise to maximum (s); min is
minimum SmbO2 in swing (%); max is maximum SmbO2 in swing (%); g1 is gradient of
fall in SmbO2 (%s-1); g2 is gradient of rise in SmbO2 (%s-1).
Figure 4 Scaled schematic graph of averaged type I and type II swings in the finger and
forearm. Significant differences between these traces are given in Table 1.
Figure 5. Histograms of number of subjects with type I and type II swings in SmbO2 at
different frequencies of oscillation, each bin width 0.02 Hz.
20
Figure 1. An example of spontaneous swings in total blood volume [Hbt] in dorsal finger
resulting from changes in [HbO2] predominantly and not [Hb] (A). The subsequent
swings in SmbO2 identified as type I swings (B).
21
Figure 2. An example of out of phase swings in [HbO2] and [Hb] in dorsal finger (A).
The subsequent swings in SmbO2 identified as type II swings (B). There are also some
small changes in total blood volume [Hbt] synchronised predominantly with changes in
[HbO2].
22
max
g2
SmbO2 (%)
g1
min
t1
t2
time (s)
Figure 3. Schematic diagram of parameters determined in SmbO2 swings where t1 is time
for SmbO2 to fall to minimum (s); t2 is time for SmbO2 to rise to maximum (s); min is
minimum SmbO2 in swing (%); max is maximum SmbO2 in swing (%); g1 is gradient of
fall in SmbO2 (%s-1); g2 is gradient of rise in SmbO2 (%s-1).
23
Figure 4 Scaled schematic graph of averaged type I and type II swings in the finger and
forearm. Significant differences between these traces are given in Table 1.
24
number
10
type I swing finger
5
0
0.02
0.12
0.22
frequency (Hz)
0.32
number
10
type I swing forearm
5
0
0.02
0.12
0.22
frequency (Hz)
0.32
number
10
type II swing finger
5
0
0.02
0.12
0.22
frequency (Hz)
0.32
number
10
type II swing forearm
5
0
0.02
0.12
0.22
frequency (Hz)
0.32
Figure 5. Histograms of number of subjects with type I and type II swings in SmbO2 at
different frequencies of oscillation, each bin width 0.02 Hz.
25
t1
s
t2
s
g1
%s-1
g2
%s-1
min
%
max
%
mean
SD
Forearm mean
SD
17.72
9.39
21.18
8.48
11.35
6.37
12.67
7.56
1.18
0.58
0.77
0.37
1.98
1.53
1.76
0.82
54
16
38
12
70
9
49
10
Type II swings
t1
s
t2
s
g1
%s-1
g2
%s-1
min
%
max
%
20.78
8.62
29.45
16.08
9.67a
2.64
18.21a
9.78
1.19b
0.61
0.50b
0.41
2.40c
1.41
0.84c
0.54
50
19
36
14
72
12
49
10
Type I swings
Finger
Finger
mean
SD
Forearm mean
SD
p=0.019 for t2 in finger vs. t2 in forearm in type II swings, bp=0.047 for g1 in finger vs. g1 in
a
forearm in type II swings, cp=0.013 for g2 in finger vs. g2 in forearm in type II swings
Table 1. Comparison between finger and forearm of the averaged parameters (t1, t2, g1,
g2, min, max) illustrated in Figure 3 for type I and type II swings in SmbO2