- Email: [email protected]
Wavelet-analysis of skin temperature oscillations during local heating for revealing endothelial dysfunction
Microvascular Research 97 (2015) 109–114
Contents lists available at ScienceDirect
Microvascular Research journal homepage: www.elsevier.com/locate/ymvre
Wavelet-analysis of skin temperature oscillations during local heating for revealing endothelial dysfunction Sergey Podtaev a, Rodion Stepanov a,c,⁎, Elena Smirnova b, Evgenia Loran b a b c
Institute of Continuous Media Mechanics, Ural Branch of Russian, Academy of Science, Perm, Russian Federation Perm State Medical Academy, Perm, Russian Federation Perm National Research Polytechnic University, Perm, Russian Federation
a r t i c l e
i n f o
Article history: Accepted 17 October 2014 Available online 27 October 2014 Keywords: Endothelial dysfunction Skin blood flow Skin temperature fluctuations Local heating test Wavelet analysis Type 2 diabetes
a b s t r a c t Skin microvessels have proven to be a model to investigate the mechanisms of vascular disease; in particular, endothelial dysfunction. To analyze skin blood flow, high-resolution thermometry can be used because lowamplitude skin temperature oscillations are caused by changes in the tone of skin vessels. The aim of our study was to test the possibilities of wavelet analysis of skin temperature (WAST) for the diagnosis of impaired regulation of microvascular tone in patients with type 2 diabetes. A local heating functional test was used for the assessment of microvascular tone regulation. A control group consisted of healthy male and female volunteers (n = 5 each), aged 39.1 ± 5.3 years. A group of patients with type 2 diabetes comprised thirteen people, seven men and six women, aged 36 to 51 years old (43.2 ± 3.4 years). The diagnosis of diabetes was made according to the criteria of the World Health Organization (WHO). The mean disease duration was 7.36 ± 0.88 years. Skin temperature oscillations, reflecting intrinsic myogenic activity (0.05–0.14 Hz), neurogenic factors (0.02–0.05 Hz) and endothelial activity (0.0095–0.02 Hz) increase greatly during local heating for healthy subjects. In the group of patients with type 2 diabetes, no statistically significant differences in the amplitudes in the endothelial range were observed. Relative changes in the oscillation amplitudes in patients with type 2 diabetes were markedly lower compared to the control group. The latter indicates that the WAST method enables assessment of the state of vascular tone and the effects of mechanisms responsible for regulation of blood flow in the microvasculature. © 2014 Elsevier Inc. All rights reserved.
Introduction Endothelial dysfunction (ED) is one of the universal mechanisms of formation of vascular complications such as atherosclerosis, hypertension, coronary heart disease, and diabetes (Endemann and Schiffrin, 2004). ED is a systemic pathological disorder that manifests primarily in the weakening of the endothelium-dependent vasodilation and microvascular remodeling, which is an early indicator of pathological processes in the cardiovascular system (Lekakis et al., 2011). Microvessels of the skin are convenient and affordable objects of study and, therefore, the skin can be used as a model to investigate the mechanisms of vascular disease; in particular, ED (Holowatz et al., 2008). It has been suggested that skin microvascular dysfunction is related to vascular disorders in other regions of the cardiovascular system; for example, coronary disease correlates with cutaneous endotheliumdependent microvascular reactivity (Ijzerman et al., 2003). There is enough evidence that skin microvascular function is an independent characteristic of cardiovascular disease in patients with type 2 diabetes ⁎ Corresponding author at: Institute of Continuous Media Mechanics, Ural Branch of Russian, Academy of Science, Perm, Russian Federation. E-mail address: [email protected] (R. Stepanov).
http://dx.doi.org/10.1016/j.mvr.2014.10.003 0026-2862/© 2014 Elsevier Inc. All rights reserved.
(Sokolnicki et al., 2006), renal disease (Stewart et al., 2004), peripheral vascular disease (Rossi and Carpi, 2004), and heart failure (Green et al., 2006). Optical methods such as photoplethysmography and laser Doppler flowmetry (LDF) are the most widely used techniques for monitoring skin blood flow (Allen and Howell, 2014; Roustit and Cracowski, 2013). Spectral analysis of the variable component of the LDF signal enables assessment of the state of the vascular tone and the effects of mechanisms responsible for regulation of blood flow in the microvasculature. Within the skin blood flow oscillation range, five sub-bands reflect different factors involved in the regulation of vascular tone: pulse wave (0.5–2 Hz), respiration wave (0.14–0.5 Hz), myogenic oscillations (0.05–0.14 Hz), neurogenic activity (0.02–0.05 Hz), and endothelial mechanisms (0.0095–0.02 Hz) (Shiogai et al., 2010; Rossi et al., 2008). Oscillations in blood flow with a characteristic frequency of around 0.01 Hz are associated with endothelial activity; in particular, nitric oxide (NO) synthesis activity. Studies based on the results of Ach iontophoresis tests (endothelium-independent vasodilator) and sodium nitroprusside (endothelium-independent vasodilator) have confirmed that oscillations of 0.01 Hz are related to endothelial activity (Kvandal et al., 2003). The laser Doppler flowmetry technique has some limitations that limit its application. First of all, there is a high level of artifacts
110
S. Podtaev et al. / Microvascular Research 97 (2015) 109–114
associated with probe displacement, which is a fundamental limitation on a method used for long-term measurements, functional tests and a subsequent spectral analysis. Furthermore, because of the spatial anatomic heterogeneity of the distribution of microvessels in the dermis (Braverman et al., 1990), the reproducibility of LDF measurements may be relatively low (Roustit et al., 2010). Such poor reproducibility at random recording sites measured by a single-point LDF could be improved when an integrated probe, laser Doppler imager or laser speckle contrast imager with an increased recording area is used (Roustit and Cracowski, 2013), but the time-frequency data analysis is particularly challenging in this case. To analyze skin blood flow responses, high-resolution thermometry can be used because the low-amplitude temperature oscillations on the skin surface are caused by changes in the tone of skin vessels (Burton and Taylor, 1940; Mabuchi et al., 1989). A cross-spectral analysis of the variations in blood pressure wave forms and temperature shows a high degree of correlation between the spontaneous fluctuations of skin temperature and the vasomotor activity of small arteries and arterioles in subcutaneous tissues (Shusterman et al., 1997). Application of the wavelet correlation method shows that, in the frequency interval corresponding to myogenic, neurogenic and endothelial mechanisms of regulation of the microvascular tone, temperature fluctuations are correlated with microcirculation fluctuations (Bandrivsky et al., 2004; Podtaev et al., 2008). Weak phase coherence between temperature and blood flow was observed for unperturbed skin, and due to heating it increased in all frequency intervals (Sheppard et al., 2011). Skin temperature measurements are relatively insensitive to mechanical factors. Due to the fact that the function defining the relationship between blood flow and skin temperature has the properties of a lowpass filter, the results obtained for temperature fluctuations are minimally sensitive to noise, and exhibit good reproducibility. In the work of Smirnova et al. (2013), the wavelet analysis of skin temperature (WAST) was used to study the mechanisms of vascular tone violations.
It was shown that the response to the cold pressor test in patients with type 2 diabetes and with impaired glucose tolerance differs essentially from that of healthy subjects in the endothelial frequency range. Local heating of the skin is one of the commonly used functional tests for the assessment of the skin blood flow. A pronounced vasodilation in human skin is achieved with local heating to 42 °C for 20–30 min (Johnson and Kellog, 2010). Violation of the mechanisms of vasodilation during the local heating test is observed in patients with types 1 and 2 diabetes (Khan et al., 2000; Sandeman et al., 1991). Thermal vasodilation heating decreases in patients with a high risk of type 2 diabetes; for example, in subjects with fasting hyperglycemia (Jaap et al., 1994), and in women with previous gestational diabetes (Hannemann et al., 2001). Local heating is a fairly simple and effective test to evaluate the mechanisms involved in regulating the vascular tone, in particular, the endothelial dysfunction. There are several reasons why this test has not found wide application in practice; however, when combined with an easy-to-implement technique WAST, it can be successfully used in routine clinical studies, especially for screening and early detection of ED. The purpose of this work is to study the possibilities of the method WAST for the diagnosis of impaired regulation of microvascular tone in patients with type 2 diabetes during the local heating test. Subjects and methods Subjects A control group consisted of healthy male and female volunteers (five each), aged 39.1 ± 5.3 years. The exclusion criteria included any significant neurological, cardiovascular, dermatological, and psychological history, diabetes, cigarette smoking and pregnancy. They were all normotensive subjects with normal levels of blood glucose, and their average BMI was 24.7 ± 1.2 kg/m2. A group of patients with type 2 diabetes comprised thirteen people, seven men and six women, aged 36 to 51 years old (43.2 ± 3.4 years). The diagnosis of diabetes was made according to the criteria of the World Health Organization (WHO). The mean disease duration was 7.36 ± 0.88 years. The study excluded patients with known microvascular complications (myocardial infarction, stenocardia, violations of cerebral and peripheral circulation) and severe microvascular disorders (pre- and proliferative retinopathy, diabetic foot). Distal diabetic polyneuropathy with the presence of positive symptoms was detected in all patients. Oral hypoglycemic therapy was given to seven people, three patients were treated only with insulin, and another three patients received the combination (insulin plus tablet formulations). The average level of glycated hemoglobin at the time of the survey was 9.7 ± 0.7%. On average, systolic blood pressure (SBP) was 130.0 ± 1.9 mm Hg, diastolic blood pressure (DBP) – 82.3 ± 1.2 mm Hg, and BMI – 28.7 ± 1.0 kg/m2. The study of biochemical parameters of patients with type 2 diabetes revealed elevated cholesterol (6.6 ± 0.34 mmol/l), triglycerides (3.7 ± 0.69 mmol/l), low-density lipoprotein (3.9 ± 0.2 mmol/l), and the reduced level of high density lipoproteins (1.2 ± 0.1 mmol/l). Elevated urine microalbumin level (20–50) was detected in one female patient. The study protocol was approved by the local ethics committee of the Perm State Medical Academy. All subjects gave written informed consent. Measurement procedure
Fig. 1. (a) Temperature variations during the local heating test (solid line — initial measurements, and dotted line — measurements with the subtracted exponential trend), (b) and (c) wavelet coefficients for the data shown by solid and dotted lines, respectively.
The tests were carried out at a room temperature of 22.5 ± 0.5 °С. Measurements were made after a fast and four-hour abstinence from smoking. The patients did not take any medication affecting vascular tone (nitrates or calcium antagonists). During the local heating test, the participants lay in the supine position. Skin temperature was measured on the palm surface of the distal phalanx of the index finger of the right hand. The output signals of the temperature sensor (HRTS5760, Honeywell International, Inc., USA) were transmitted after
S. Podtaev et al. / Microvascular Research 97 (2015) 109–114
111
Fig. 2. The signal recovered by the inverse wavelet-transform according to the frequency band (dotted line) and the amplitude characteristic obtained using expression (3) (solid line).
amplification to the 18-bit bipolar analog-to-digital converter (ADS 7793, Analog Devices) scaled to ± 5 V with sampling frequency of 200 Hz. For the temperature range 25–45 °C, with consideration for signal-to-noise amplifier ratio, the actual resolution of temperature was 0.001 °C. During the measurements, all necessary precautions were taken to reduce the effect of external heat flows on the thermistor
recording the skin temperature. The thermistor was placed in a specially designed plastic case (20 × 30 × 10 mm3) filled with a material with low thermal conductivity (λ b 0.02Wt / (m × K)) for its protection against ambient temperature variations. The case also allowed the sensors to be fixed on the skin surface with a medical plaster, which prevents sensor displacements during measurements.
Fig. 3. Initial signal and its wavelet transform for the healthy participants (at left) and for the patient with diabetes (at right).
112
S. Podtaev et al. / Microvascular Research 97 (2015) 109–114
A miniature resistance heater (2 × 2 × 1 mm 3 ) was mounted 7 mm off the temperature sensor in the same plastic case. Once the heating element was turned on and reached a steady state, thermal equilibrium between the heater, the sensor and the top layer of the skin was established. In order to maintain the desired temperature, the constant power mode was used. The time needed to establish thermal equilibrium after turning on the heating element was about 200 s. Acquiring temperature data from the sensor began after mounting the case with a sensor on the skin. Skin temperature measurements were carried out continuously for 10 min before the local heating, for 2 min after turning on the heater and on reaching the temperature of about 40 ± 0.7 °C, and then for 10 min while keeping this thermal regime.
Software and statistical analysis The application of advanced signal processing techniques provides more detailed answers to the questions related to the analysis and interpretation of signals. In our study, we discuss the use of a continuous wavelet transform. Despite its higher computational complexity compared to the Fourier transformation, the continuous wavelet transform makes the assessment of signal amplitude characteristics in allocated frequency bands more convenient. The continuous wavelet transform of the function T(t) is given by the equation:
wða; t Þ ¼
1 jaj
Z∞
0 t 0 −t 0 T t ψ dt ; a
ð1Þ
−∞
where the mother wavelet ψ(t) is scaled by the parameter a and transferred at time t '. The choice of the form depends upon the subsequent problems. The applied Morlet wavelet ψ(t) = exp(2iπt − t2/2) offers good resolution in both frequency and time. We are interested here to know how temperature oscillation amplitudes change before and after heating at time tp = 600 in the frequency bands corresponding to myogenic oscillations (0.05–0.14 Hz), neurogenic activity (0.02–0.05 Hz), and endothelial mechanisms (0.0095– 0.02 Hz). The onset of heating is characterized by a sharp monotonic change in temperature, which contributes much to the wavelet coefficients of all the frequencies near time tp. We subtract this trend using the exponential model of the form a + be− t/c, where a, b and с are found by fitting the data within the 100 s interval after tp. The results of such transformation are given in Fig. 1, where one can see that the subtraction of the trend makes the signal clearer. The signal amplitude assessment in the frequency band presents more difficulties than in the case of a single frequency. For a given frequency, it is sufficient to take w(afix, t); for the frequency band, we should perform an inverse wavelet transform to recover the signal. 1 e F ðt Þ ¼ Cψ
0 t 0 −t dt 0 da w a; t ψ ; a a
aZ max
a min
ð2Þ
where amin and amax are defined by the frequency band limits. The standard deviation of the recovered signal e F ðt Þ during the examined interval of time may serve as a characteristic of the signal amplitude (Smirnova et al., 2013). In this paper, we present an alternative approach that allows us to avoid the inverse transform. It reads as
Aðt Þ ¼
1 Cψ
aZ max
a min
0 da w a; t : a
ð3Þ
Table 1 The oscillation amplitudes Ae, An and Am (*102, °С) before and during heating for healthy subjects and for patients with type 2 diabetes.
Fig. 4. The ratio of amplitudes before (A1) and after (A2) local heating for three frequency bands. Dashed line shows A2 = 2A1 dependence. (For interpretation of the references to color in this figure, the reader is referred to the web version of this article.)
Before heating
During heating
p1–2
Healthy subjects Endothelial Neurogenic Myogenic
0.92 ± 0.44 0.59 ± 0.41 0.17 ± 0.10
4.55 ± 2.69 2.60 ± 1.38 0.83 ± 0.46
0.0003 0.0004 0.0006
Type 2 diabetes Endothelial Neurogenic Myogenic
1.19 ± 0.70 0.56 ± 0.40 0.17 ± 0.14
1.77 ± 0.76 1.00 ± 0.52 0.40 ± 0.40
NS 0.024 0.006
S. Podtaev et al. / Microvascular Research 97 (2015) 109–114
where A(t) serves as the envelope of e F ðt Þ (Fig. 2). For the characteristic of oscillations, we obtain the parameters Ae, An and Am in each frequency band. Apart from the average value of A(t) within the examined interval, we can compute the standard deviation essential for assessing an error. The original algorithms of the wavelet analysis were realized in C++. The data are represented as M ± SD, where M is the average value and SD is the standard deviation. The groups were compared using the non-parametric statistical tests (the Mann–Whitney U-test). The Wilcoxon test made it possible to compare the paired data, and the p values b 0.05 were considered statistically significant. The statistical analysis was performed using Mathematica 7.0 and statistical software Statistica 8.0.
Results Fig. 3 shows the typical skin temperature as a function of time for a healthy person who had undergone the local heating test. In the regime without external heating (0 … 600 s), temperature oscillations caused by blood flow oscillations in skin microvessels are registered. During the test (600 … 1300 s), the temperature increased concurrently with the increase in the oscillation amplitude in all three frequency bands (wavelet plane in Fig. 3b). The response to the local heating test studied in patients with type 2 diabetes was essentially different from that of healthy subjects. During heating, the oscillation amplitude of the skin temperature was significantly lower. An example of temperature variation and the corresponding wavelet-plane are shown in Figs. 3c and d. We note that the exponential heating rate varies a lot among participants. For the case shown in Fig. 3a the coefficient, c, is about 25 s, and it is 40 s for the case in Fig. 3b. However, the difference between the groups of healthy participants and patients with type 2 diabetes is not significant enough for an interpretation, because the averaged coefficients, c, are 35 ± 11 s and 40 ± 8 s, respectively. Figs. 4a, b, c present the graphs of amplitudes of the wavelet coefficients Ae, An and Am before and during the test in the myogenic, neurogenic and endothelial frequency bands for the control group (red dots) and the group of patients with diabetes (black dots). The initial oscillation amplitudes Ae, An and Am in all three frequency bands did not exhibit significant differences between the groups. In healthy people, the amplitudes during the local heating test differed from the initial amplitudes in all three frequency ranges (Table 1). In the group of patients with diabetes, no significant differences in the endothelial frequency range were observed (Table 1). A comparison of relative changes in amplitudes κ x ¼
Ax2 −Ax1 , Ax1
which defined for each
frequency range, shows a significant difference between the groups (Figs. 5a, b, c).
113
Discussion The study carried out using the WAST method has indicated that in the group of healthy subjects, skin temperature oscillation amplitude increases greatly in all three frequency ranges during the local heating test. By contrast, in the group of patients with diabetes, no significant increase in the oscillation amplitude in the endothelial range was observed. Furthermore, relative changes in the oscillation amplitudes were markedly lower compared to the control group within all frequency ranges. We propose that these results can be considered a violation of the mechanisms of vasodilation induced by the endothelial dysfunction. Based on the measurements of responses to local heating, the phases of initial and prolonged vasodilation have been defined. Initial vasodilation is related to the axon-reflex mechanisms. It is blocked by local anesthetics directly in the hot zone of the skin and manifests itself most vividly on heating to 34–35 °C (Stephens et al., 2001). The second phase begins when NO is released from the vascular endothelium. In general, the vasodilatory response to local heating depends primarily on the second phase. It is shown that the second-phase dilation decreases on blocking the skin NO-synthase with L-NAME prior to heating (Minson et al., 2001). The characteristic time intervals of axon-reflex responses (2–3 min) are compatible with the period when the temperature increases because of heating. In our investigation, because of the temperature rise inertia, we did not get a chance to register a manifestation of the axon reflex. The duration of the heating test is shorter than the time interval taken to observe the “plateau” discussed in the previous investigations. We intentionally minimized the duration of the test for the sake of convenience in the practical usage. We found that 7–8 endothelial oscillations with 70–90 s periodicity are sufficient for statistically robust results. We checked that the errors slightly decrease for the twice-longer tests, but confirm that it insignificantly improves the scope of the method. Nevertheless, this is the limitation of the study. Violation of the mechanisms of vasodilation during the local heating test has been observed in patients with types 1 and 2 diabetes (Khan et al., 2000; Sandeman et al., 1991). It is now recognized that chronic hyperglycemia plays a major role in the initiation of diabetic vascular complications. More and more data confirm the damaging effect of “glucose toxicity” on the secretory capabilities of an insular apparatus (Aronson, 2008). The location of the endothelium on the border with the blood flow makes it vulnerable to a variety of risk factors for vascular complications, including hyperglycemia (McCall, 1992). It is known that insulin resistance and the violation of NO production are closely linked (Marchesi et al., 2009; Youn et al., 2014). The manifestation of endothelial dysfunction is a violation of its vasoregulating function (Avogaro et al., 2011). Endothelial dysfunction is one of the earliest signs of vascular lesions in diabetic patients and can be detected in the
Fig. 5. Relative changes in amplitudes Ke (a), Kn (b), Km (c) for healthy subjects and for patients with type 2 diabetes.
114
S. Podtaev et al. / Microvascular Research 97 (2015) 109–114
early stages of the disease, just before the appearance of atherosclerotic plaques (Storey et al., 2001), i.e. at the stage of impaired glucose tolerance (Smirnova et al., 2013). In the case of carbohydrate metabolism violation, protein kinase C (PKC) is activated in endothelial cells. PKC increases the permeability of endotheliocytes to proteins, attenuates the endothelium-dependent relaxation of blood vessels, and enhances lipid peroxidation, whose products also inhibit the vasodilatory endothelial function (Grattagliano et al., 2008). A reduction in the production of vasodilating factors and an increase in the production of vasoconstrictors in the endothelium lead to an increased vasoconstrictor response (Widlansky et al., 2003). Nowadays, the concept of endothelial dysfunction as a key element of vascular lesions in patients with diabetes has been formulated. However, the lack of clear diagnostic criteria (biochemical, instrumental) for endothelial dysfunction allows us to continue a search for the optimal methodology for the study of various aspects of violations (Celermajer, 2008). The WAST method may be proposed as a tool to assess endothelial dysfunction before clinical manifestations appear. However, further validation is needed to be performed while blocking endothelium-dependent pathways, e.g. by using microdialysis. Conflict of interest statement The authors have no conflicts of interest to declare. Acknowledgments This work was supported by the Russian Science Foundation No. 1415-00809. References Allen, J., Howell, K., 2014. Microvascular imaging: techniques and opportunities for clinical physiological measurements. Physiol. Meas. 35, 91–141. Aronson, D., 2008. Hyperglycemia and Pathobiology of Diabetic Complications. Cardiovascular Diabetology: Clinical, Metabolic and Inflammatory Facets. In: Fisman, E.Z., Holon, A. Tenenbaum (Eds.), Tel-Hashomer, pp. 1–114. Avogaro, A., Albiero, M., Menegazzo, L., Kreutzenberg, S., Fadini, G., 2011. Endothelial dysfunction in diabetes. The role of reparatory mechanisms. Diabetes Care 34 (2), 285–290. Bandrivsky, A., Bernjak, A., McClintock, P., Stefanovska, A., 2004. Wavelet phase coherence analysis: application to skin temperature and blood flow. Cardiovasc. Eng. 4, 89–93. Braverman, I.M., Keh, A., Goldminz, D., 1990. Correlation of laser Doppler wave patterns with underlying microvascular anatomy. J. Investig. Dermatol. 95, 283. Burton, A.C., Taylor, R.M., 1940. A study of the adjustment of peripheral vascular tone to the requirements of the regulation of body temperature. Am. J. Physiol. 129, 566–577. Celermajer, D.S., 2008. Reliable endothelial function testing: at our fingertips? Circulation 117 (19), 2428–2430. Endemann, D.H., Schiffrin, E.L., 2004. Endothelial dysfunction. J. Am. Soc. Nephrol. 15, 1983–1992. Grattagliano, I., Palmieri, V.O., Portincasa, P., Moschetta, A., Palasciano, G., 2008. Oxidative stress-induced risk factors associated with the metabolic syndrome: a unifying hypothesis. J. Nutr. Biochem. 19, 491–504. Green, D.J., Maiorana, A.J., Siong, J.H., Burke, V., Erickson, M., Minson, C.T., Bilsborough, W., O'Driscoll, G., 2006. Impaired skin blood flow response to environmental heating in chronic heart failure. Eur. Heart J. 27, 338–343. Hannemann, M.M., Liddell, W.G., Shore, A.C., Clark, P.M., Tooke, J.E., 2001. Vascular function in women with previous gestational diabetes mellitus. J. Vasc. Res. 39, 311–331. Holowatz, L.A., Thompson-Torgerson, C.S., Kenney, W.L., 2008. The human cutaneous circulation as a model of generalized microvascular function. J. Appl. Physiol. 105, 370–372.
IJzerman, R.G., de Jongh, R.T., Beijk, M.A.M., van Weissenbruch, M.M., Delemarre-van de Waal, H.A., Serné, E.H., Stehouwer, C.D.A., 2003. Individuals at increased coronary heart disease risk are characterized by an impaired microvascular function in skin. Eur. J. Clin. Investig. 33, 536–542. Jaap, A.J., Hammersley, M.S., Shore, A.C., Tooke, J.E., 1994. Reduced microvascular hyperaemia in subjects at risk of developing type 2 (non-insulin-dependent) diabetes mellitus. Diabetologia 37, 214–216. Johnson, J.M., Kellog, D.L.J., 2010. Local thermal control of the human cutaneous circulation. J. Appl. Physiol. 109, 1229–1238. Khan, F., Elhadd, T.A., Greene, S.A., Belch, J.J., 2000. Impaired skin microvascular function in children, adolescents, and young adults with type 1 diabetes. Diabetes Care 23, 215–220. Kvandal, P., Stefanovska, A., Veber, M., Kvermmo, H., Kirkeboen, K., 2003. Regulation of human cutaneous circulation evaluated by laser Doppler flowmetry, iontophoresis, and spectral analysis: importance of nitric oxide and prostaglandines. Microvasc. Res. 65, 160–171. Lekakis, J., Abraham, P., Balbarini, A., Blann, A., Boulanger, C.M., et al., 2011. Methods for evaluating endothelial function: a position statement from the European Society of Cardiology Working Group on Peripheral Circulation. Eur. J. Cardiovasc. Prev. Rehabil. 18 (6), 775–789. Mabuchi, K., Chinzei, T., Nasu, Y., Yonezawa, T., Fujimasa, I., Atsumi, K., 1989. Frequency analysis of skin temperature and its application for clinical diagnosis. Biomed. Thermol. 9, 30–33. Marchesi, C., Ebrahimian, T., Angulo, O., Paradis, P., Schiffrin, E.L., 2009. Endothelial nitric oxide synthase uncoupling and perivascular adipose oxidative stress and inflammation contribute to vascular dysfunction in a rodent model of metabolic syndrome. Hypertension 54, 1384–1392. McCall, A.L., 1992. The impact of diabetes on the CNS. Diabetes 41 (5), 557–570. Minson, C.T., Berry, L.T., Joyner, M.J., 2001. Nitric oxide and neurally mediated regulation of skin blood flow during local heating. J. Appl. Physiol. 91 (4), 1619–1626. Podtaev, S., Morozov, M., Frick, P., 2008. Wavelet-based correlations of skin temperature and blood flow oscillations. Cardiovasc. Eng. 8, 185–189. Rossi, M., Carpi, A., 2004. Skin microcirculation in peripheral arterial obliterative disease. Biomed. Pharmacother. 58, 427–431. Rossi, M., Carpib, A., Galettaa, F., Franzonia, F., Santoroa, G., 2008. Skin vasomotion investigation: a useful tool for clinical evaluation of microvascular endothelial function? Biomed. Pharmacother. 62, 541–545. Roustit, M., Blaise, S., Millet, C., Cracowski, J.L., 2010. Reproducibility and methodological issues of skin post-occlusive and thermal hyperemia assessed by single-point laser Doppler flowmetry. Microvasc. Res. 79, 102–108. Roustit, M., Cracowski, J., 2013. Assessment of endothelial and neurovascular function in human skin microcirculation. Trends Pharmacol. Sci. 7, 34. Sandeman, D.D., Pym, C.A., Green, E.M., Seamark, C., Shore, A.C., Tooke, J.E., 1991. Microvascular vasodilatation in the feet of newly diagnosed non-insulin dependent diabetic patients. BMJ 302, 1122–1123. Sheppard, L.W., Vuksanović, V., McClintock, P.V.E., Stefanovska, A., 2011. Oscillatory dynamics of vasoconstriction and vasodilation identified by time-localized phase coherence. Phys. Med. Biol. 56, 3583–3601. Shiogai, Y., Stefanovska, A., McClintock, P.V.E., 2010. Nonlinear dynamics of cardiovascular ageing. Phys. Rep. 488, 51–110. Smirnova, E., Podtaev, S., Mizeva, I., Loran, E., 2013. Assessment of endothelial dysfunction in patients with impaired glucose tolerance during a cold pressor test. Diab. Vasc. Dis. Res. 10 (6), 489–497. Shusterman, V., Anderson, K.P., Barnea, O., 1997. Spontaneous skin temperature oscillations in normal human subjects. Am. J. Physiol. 273, 1173–1181. Sokolnicki, L.A., Roberts, S.K., Wilkins, B.W., Basu, A., Charkoudian, N., 2006. Contribution of nitric oxide to cutaneous microvascular dilation in individuals with type 2 diabetes mellitus. Am. J. Physiol. Endocrinol. Metab. 292, 314–318. Stephens, D.P., Charkoudian, N., Benevento, J.M., Johnson, J.M., Saumet, J.L., 2001. The influence of topical capsaicin on the local thermal control of skin blood flow in humans. Am. J. Physiol. Regul. Integr. Comp. Physiol. 281 (3), R894–R901 (Sep). Stewart, J., Kohen, A., Brouder, D., Rahim, F., Adler, S., Garrick, R., Goligorsky, M.S., 2004. Noninvasive interrogation of microvasculature for signs of endothelial dysfunction in patients with chronic renal failure. Am. J. Physiol. Heart Circ. Physiol. 287, 2687–2696. Storey, A.M., Perry, C.J., Petrie, J.R., 2001. Endothelial dysfunction in type 2 diabetes. Br. J. Diabetes Vasc. Dis. 1, 22–27. Widlansky, M.E., Gokce, N., Keaney, J.F., Joseph, A., 2003. The clinical implications of endothelial dysfunction. Vita J. Am. Coll. Cardiol. 7, 42. Youn, J.Y., Siu, K.L., Lob, H.E., Itani, H., Harrison, D.G., Cai, H., 2014. Role of vascular oxidative stress in obesity and metabolic syndrome. Diabetes 63 (7), 2344–2355.
Contents lists available at ScienceDirect
Microvascular Research journal homepage: www.elsevier.com/locate/ymvre
Wavelet-analysis of skin temperature oscillations during local heating for revealing endothelial dysfunction Sergey Podtaev a, Rodion Stepanov a,c,⁎, Elena Smirnova b, Evgenia Loran b a b c
Institute of Continuous Media Mechanics, Ural Branch of Russian, Academy of Science, Perm, Russian Federation Perm State Medical Academy, Perm, Russian Federation Perm National Research Polytechnic University, Perm, Russian Federation
a r t i c l e
i n f o
Article history: Accepted 17 October 2014 Available online 27 October 2014 Keywords: Endothelial dysfunction Skin blood flow Skin temperature fluctuations Local heating test Wavelet analysis Type 2 diabetes
a b s t r a c t Skin microvessels have proven to be a model to investigate the mechanisms of vascular disease; in particular, endothelial dysfunction. To analyze skin blood flow, high-resolution thermometry can be used because lowamplitude skin temperature oscillations are caused by changes in the tone of skin vessels. The aim of our study was to test the possibilities of wavelet analysis of skin temperature (WAST) for the diagnosis of impaired regulation of microvascular tone in patients with type 2 diabetes. A local heating functional test was used for the assessment of microvascular tone regulation. A control group consisted of healthy male and female volunteers (n = 5 each), aged 39.1 ± 5.3 years. A group of patients with type 2 diabetes comprised thirteen people, seven men and six women, aged 36 to 51 years old (43.2 ± 3.4 years). The diagnosis of diabetes was made according to the criteria of the World Health Organization (WHO). The mean disease duration was 7.36 ± 0.88 years. Skin temperature oscillations, reflecting intrinsic myogenic activity (0.05–0.14 Hz), neurogenic factors (0.02–0.05 Hz) and endothelial activity (0.0095–0.02 Hz) increase greatly during local heating for healthy subjects. In the group of patients with type 2 diabetes, no statistically significant differences in the amplitudes in the endothelial range were observed. Relative changes in the oscillation amplitudes in patients with type 2 diabetes were markedly lower compared to the control group. The latter indicates that the WAST method enables assessment of the state of vascular tone and the effects of mechanisms responsible for regulation of blood flow in the microvasculature. © 2014 Elsevier Inc. All rights reserved.
Introduction Endothelial dysfunction (ED) is one of the universal mechanisms of formation of vascular complications such as atherosclerosis, hypertension, coronary heart disease, and diabetes (Endemann and Schiffrin, 2004). ED is a systemic pathological disorder that manifests primarily in the weakening of the endothelium-dependent vasodilation and microvascular remodeling, which is an early indicator of pathological processes in the cardiovascular system (Lekakis et al., 2011). Microvessels of the skin are convenient and affordable objects of study and, therefore, the skin can be used as a model to investigate the mechanisms of vascular disease; in particular, ED (Holowatz et al., 2008). It has been suggested that skin microvascular dysfunction is related to vascular disorders in other regions of the cardiovascular system; for example, coronary disease correlates with cutaneous endotheliumdependent microvascular reactivity (Ijzerman et al., 2003). There is enough evidence that skin microvascular function is an independent characteristic of cardiovascular disease in patients with type 2 diabetes ⁎ Corresponding author at: Institute of Continuous Media Mechanics, Ural Branch of Russian, Academy of Science, Perm, Russian Federation. E-mail address: [email protected] (R. Stepanov).
http://dx.doi.org/10.1016/j.mvr.2014.10.003 0026-2862/© 2014 Elsevier Inc. All rights reserved.
(Sokolnicki et al., 2006), renal disease (Stewart et al., 2004), peripheral vascular disease (Rossi and Carpi, 2004), and heart failure (Green et al., 2006). Optical methods such as photoplethysmography and laser Doppler flowmetry (LDF) are the most widely used techniques for monitoring skin blood flow (Allen and Howell, 2014; Roustit and Cracowski, 2013). Spectral analysis of the variable component of the LDF signal enables assessment of the state of the vascular tone and the effects of mechanisms responsible for regulation of blood flow in the microvasculature. Within the skin blood flow oscillation range, five sub-bands reflect different factors involved in the regulation of vascular tone: pulse wave (0.5–2 Hz), respiration wave (0.14–0.5 Hz), myogenic oscillations (0.05–0.14 Hz), neurogenic activity (0.02–0.05 Hz), and endothelial mechanisms (0.0095–0.02 Hz) (Shiogai et al., 2010; Rossi et al., 2008). Oscillations in blood flow with a characteristic frequency of around 0.01 Hz are associated with endothelial activity; in particular, nitric oxide (NO) synthesis activity. Studies based on the results of Ach iontophoresis tests (endothelium-independent vasodilator) and sodium nitroprusside (endothelium-independent vasodilator) have confirmed that oscillations of 0.01 Hz are related to endothelial activity (Kvandal et al., 2003). The laser Doppler flowmetry technique has some limitations that limit its application. First of all, there is a high level of artifacts
110
S. Podtaev et al. / Microvascular Research 97 (2015) 109–114
associated with probe displacement, which is a fundamental limitation on a method used for long-term measurements, functional tests and a subsequent spectral analysis. Furthermore, because of the spatial anatomic heterogeneity of the distribution of microvessels in the dermis (Braverman et al., 1990), the reproducibility of LDF measurements may be relatively low (Roustit et al., 2010). Such poor reproducibility at random recording sites measured by a single-point LDF could be improved when an integrated probe, laser Doppler imager or laser speckle contrast imager with an increased recording area is used (Roustit and Cracowski, 2013), but the time-frequency data analysis is particularly challenging in this case. To analyze skin blood flow responses, high-resolution thermometry can be used because the low-amplitude temperature oscillations on the skin surface are caused by changes in the tone of skin vessels (Burton and Taylor, 1940; Mabuchi et al., 1989). A cross-spectral analysis of the variations in blood pressure wave forms and temperature shows a high degree of correlation between the spontaneous fluctuations of skin temperature and the vasomotor activity of small arteries and arterioles in subcutaneous tissues (Shusterman et al., 1997). Application of the wavelet correlation method shows that, in the frequency interval corresponding to myogenic, neurogenic and endothelial mechanisms of regulation of the microvascular tone, temperature fluctuations are correlated with microcirculation fluctuations (Bandrivsky et al., 2004; Podtaev et al., 2008). Weak phase coherence between temperature and blood flow was observed for unperturbed skin, and due to heating it increased in all frequency intervals (Sheppard et al., 2011). Skin temperature measurements are relatively insensitive to mechanical factors. Due to the fact that the function defining the relationship between blood flow and skin temperature has the properties of a lowpass filter, the results obtained for temperature fluctuations are minimally sensitive to noise, and exhibit good reproducibility. In the work of Smirnova et al. (2013), the wavelet analysis of skin temperature (WAST) was used to study the mechanisms of vascular tone violations.
It was shown that the response to the cold pressor test in patients with type 2 diabetes and with impaired glucose tolerance differs essentially from that of healthy subjects in the endothelial frequency range. Local heating of the skin is one of the commonly used functional tests for the assessment of the skin blood flow. A pronounced vasodilation in human skin is achieved with local heating to 42 °C for 20–30 min (Johnson and Kellog, 2010). Violation of the mechanisms of vasodilation during the local heating test is observed in patients with types 1 and 2 diabetes (Khan et al., 2000; Sandeman et al., 1991). Thermal vasodilation heating decreases in patients with a high risk of type 2 diabetes; for example, in subjects with fasting hyperglycemia (Jaap et al., 1994), and in women with previous gestational diabetes (Hannemann et al., 2001). Local heating is a fairly simple and effective test to evaluate the mechanisms involved in regulating the vascular tone, in particular, the endothelial dysfunction. There are several reasons why this test has not found wide application in practice; however, when combined with an easy-to-implement technique WAST, it can be successfully used in routine clinical studies, especially for screening and early detection of ED. The purpose of this work is to study the possibilities of the method WAST for the diagnosis of impaired regulation of microvascular tone in patients with type 2 diabetes during the local heating test. Subjects and methods Subjects A control group consisted of healthy male and female volunteers (five each), aged 39.1 ± 5.3 years. The exclusion criteria included any significant neurological, cardiovascular, dermatological, and psychological history, diabetes, cigarette smoking and pregnancy. They were all normotensive subjects with normal levels of blood glucose, and their average BMI was 24.7 ± 1.2 kg/m2. A group of patients with type 2 diabetes comprised thirteen people, seven men and six women, aged 36 to 51 years old (43.2 ± 3.4 years). The diagnosis of diabetes was made according to the criteria of the World Health Organization (WHO). The mean disease duration was 7.36 ± 0.88 years. The study excluded patients with known microvascular complications (myocardial infarction, stenocardia, violations of cerebral and peripheral circulation) and severe microvascular disorders (pre- and proliferative retinopathy, diabetic foot). Distal diabetic polyneuropathy with the presence of positive symptoms was detected in all patients. Oral hypoglycemic therapy was given to seven people, three patients were treated only with insulin, and another three patients received the combination (insulin plus tablet formulations). The average level of glycated hemoglobin at the time of the survey was 9.7 ± 0.7%. On average, systolic blood pressure (SBP) was 130.0 ± 1.9 mm Hg, diastolic blood pressure (DBP) – 82.3 ± 1.2 mm Hg, and BMI – 28.7 ± 1.0 kg/m2. The study of biochemical parameters of patients with type 2 diabetes revealed elevated cholesterol (6.6 ± 0.34 mmol/l), triglycerides (3.7 ± 0.69 mmol/l), low-density lipoprotein (3.9 ± 0.2 mmol/l), and the reduced level of high density lipoproteins (1.2 ± 0.1 mmol/l). Elevated urine microalbumin level (20–50) was detected in one female patient. The study protocol was approved by the local ethics committee of the Perm State Medical Academy. All subjects gave written informed consent. Measurement procedure
Fig. 1. (a) Temperature variations during the local heating test (solid line — initial measurements, and dotted line — measurements with the subtracted exponential trend), (b) and (c) wavelet coefficients for the data shown by solid and dotted lines, respectively.
The tests were carried out at a room temperature of 22.5 ± 0.5 °С. Measurements were made after a fast and four-hour abstinence from smoking. The patients did not take any medication affecting vascular tone (nitrates or calcium antagonists). During the local heating test, the participants lay in the supine position. Skin temperature was measured on the palm surface of the distal phalanx of the index finger of the right hand. The output signals of the temperature sensor (HRTS5760, Honeywell International, Inc., USA) were transmitted after
S. Podtaev et al. / Microvascular Research 97 (2015) 109–114
111
Fig. 2. The signal recovered by the inverse wavelet-transform according to the frequency band (dotted line) and the amplitude characteristic obtained using expression (3) (solid line).
amplification to the 18-bit bipolar analog-to-digital converter (ADS 7793, Analog Devices) scaled to ± 5 V with sampling frequency of 200 Hz. For the temperature range 25–45 °C, with consideration for signal-to-noise amplifier ratio, the actual resolution of temperature was 0.001 °C. During the measurements, all necessary precautions were taken to reduce the effect of external heat flows on the thermistor
recording the skin temperature. The thermistor was placed in a specially designed plastic case (20 × 30 × 10 mm3) filled with a material with low thermal conductivity (λ b 0.02Wt / (m × K)) for its protection against ambient temperature variations. The case also allowed the sensors to be fixed on the skin surface with a medical plaster, which prevents sensor displacements during measurements.
Fig. 3. Initial signal and its wavelet transform for the healthy participants (at left) and for the patient with diabetes (at right).
112
S. Podtaev et al. / Microvascular Research 97 (2015) 109–114
A miniature resistance heater (2 × 2 × 1 mm 3 ) was mounted 7 mm off the temperature sensor in the same plastic case. Once the heating element was turned on and reached a steady state, thermal equilibrium between the heater, the sensor and the top layer of the skin was established. In order to maintain the desired temperature, the constant power mode was used. The time needed to establish thermal equilibrium after turning on the heating element was about 200 s. Acquiring temperature data from the sensor began after mounting the case with a sensor on the skin. Skin temperature measurements were carried out continuously for 10 min before the local heating, for 2 min after turning on the heater and on reaching the temperature of about 40 ± 0.7 °C, and then for 10 min while keeping this thermal regime.
Software and statistical analysis The application of advanced signal processing techniques provides more detailed answers to the questions related to the analysis and interpretation of signals. In our study, we discuss the use of a continuous wavelet transform. Despite its higher computational complexity compared to the Fourier transformation, the continuous wavelet transform makes the assessment of signal amplitude characteristics in allocated frequency bands more convenient. The continuous wavelet transform of the function T(t) is given by the equation:
wða; t Þ ¼
1 jaj
Z∞
0 t 0 −t 0 T t ψ dt ; a
ð1Þ
−∞
where the mother wavelet ψ(t) is scaled by the parameter a and transferred at time t '. The choice of the form depends upon the subsequent problems. The applied Morlet wavelet ψ(t) = exp(2iπt − t2/2) offers good resolution in both frequency and time. We are interested here to know how temperature oscillation amplitudes change before and after heating at time tp = 600 in the frequency bands corresponding to myogenic oscillations (0.05–0.14 Hz), neurogenic activity (0.02–0.05 Hz), and endothelial mechanisms (0.0095– 0.02 Hz). The onset of heating is characterized by a sharp monotonic change in temperature, which contributes much to the wavelet coefficients of all the frequencies near time tp. We subtract this trend using the exponential model of the form a + be− t/c, where a, b and с are found by fitting the data within the 100 s interval after tp. The results of such transformation are given in Fig. 1, where one can see that the subtraction of the trend makes the signal clearer. The signal amplitude assessment in the frequency band presents more difficulties than in the case of a single frequency. For a given frequency, it is sufficient to take w(afix, t); for the frequency band, we should perform an inverse wavelet transform to recover the signal. 1 e F ðt Þ ¼ Cψ
0 t 0 −t dt 0 da w a; t ψ ; a a
aZ max
a min
ð2Þ
where amin and amax are defined by the frequency band limits. The standard deviation of the recovered signal e F ðt Þ during the examined interval of time may serve as a characteristic of the signal amplitude (Smirnova et al., 2013). In this paper, we present an alternative approach that allows us to avoid the inverse transform. It reads as
Aðt Þ ¼
1 Cψ
aZ max
a min
0 da w a; t : a
ð3Þ
Table 1 The oscillation amplitudes Ae, An and Am (*102, °С) before and during heating for healthy subjects and for patients with type 2 diabetes.
Fig. 4. The ratio of amplitudes before (A1) and after (A2) local heating for three frequency bands. Dashed line shows A2 = 2A1 dependence. (For interpretation of the references to color in this figure, the reader is referred to the web version of this article.)
Before heating
During heating
p1–2
Healthy subjects Endothelial Neurogenic Myogenic
0.92 ± 0.44 0.59 ± 0.41 0.17 ± 0.10
4.55 ± 2.69 2.60 ± 1.38 0.83 ± 0.46
0.0003 0.0004 0.0006
Type 2 diabetes Endothelial Neurogenic Myogenic
1.19 ± 0.70 0.56 ± 0.40 0.17 ± 0.14
1.77 ± 0.76 1.00 ± 0.52 0.40 ± 0.40
NS 0.024 0.006
S. Podtaev et al. / Microvascular Research 97 (2015) 109–114
where A(t) serves as the envelope of e F ðt Þ (Fig. 2). For the characteristic of oscillations, we obtain the parameters Ae, An and Am in each frequency band. Apart from the average value of A(t) within the examined interval, we can compute the standard deviation essential for assessing an error. The original algorithms of the wavelet analysis were realized in C++. The data are represented as M ± SD, where M is the average value and SD is the standard deviation. The groups were compared using the non-parametric statistical tests (the Mann–Whitney U-test). The Wilcoxon test made it possible to compare the paired data, and the p values b 0.05 were considered statistically significant. The statistical analysis was performed using Mathematica 7.0 and statistical software Statistica 8.0.
Results Fig. 3 shows the typical skin temperature as a function of time for a healthy person who had undergone the local heating test. In the regime without external heating (0 … 600 s), temperature oscillations caused by blood flow oscillations in skin microvessels are registered. During the test (600 … 1300 s), the temperature increased concurrently with the increase in the oscillation amplitude in all three frequency bands (wavelet plane in Fig. 3b). The response to the local heating test studied in patients with type 2 diabetes was essentially different from that of healthy subjects. During heating, the oscillation amplitude of the skin temperature was significantly lower. An example of temperature variation and the corresponding wavelet-plane are shown in Figs. 3c and d. We note that the exponential heating rate varies a lot among participants. For the case shown in Fig. 3a the coefficient, c, is about 25 s, and it is 40 s for the case in Fig. 3b. However, the difference between the groups of healthy participants and patients with type 2 diabetes is not significant enough for an interpretation, because the averaged coefficients, c, are 35 ± 11 s and 40 ± 8 s, respectively. Figs. 4a, b, c present the graphs of amplitudes of the wavelet coefficients Ae, An and Am before and during the test in the myogenic, neurogenic and endothelial frequency bands for the control group (red dots) and the group of patients with diabetes (black dots). The initial oscillation amplitudes Ae, An and Am in all three frequency bands did not exhibit significant differences between the groups. In healthy people, the amplitudes during the local heating test differed from the initial amplitudes in all three frequency ranges (Table 1). In the group of patients with diabetes, no significant differences in the endothelial frequency range were observed (Table 1). A comparison of relative changes in amplitudes κ x ¼
Ax2 −Ax1 , Ax1
which defined for each
frequency range, shows a significant difference between the groups (Figs. 5a, b, c).
113
Discussion The study carried out using the WAST method has indicated that in the group of healthy subjects, skin temperature oscillation amplitude increases greatly in all three frequency ranges during the local heating test. By contrast, in the group of patients with diabetes, no significant increase in the oscillation amplitude in the endothelial range was observed. Furthermore, relative changes in the oscillation amplitudes were markedly lower compared to the control group within all frequency ranges. We propose that these results can be considered a violation of the mechanisms of vasodilation induced by the endothelial dysfunction. Based on the measurements of responses to local heating, the phases of initial and prolonged vasodilation have been defined. Initial vasodilation is related to the axon-reflex mechanisms. It is blocked by local anesthetics directly in the hot zone of the skin and manifests itself most vividly on heating to 34–35 °C (Stephens et al., 2001). The second phase begins when NO is released from the vascular endothelium. In general, the vasodilatory response to local heating depends primarily on the second phase. It is shown that the second-phase dilation decreases on blocking the skin NO-synthase with L-NAME prior to heating (Minson et al., 2001). The characteristic time intervals of axon-reflex responses (2–3 min) are compatible with the period when the temperature increases because of heating. In our investigation, because of the temperature rise inertia, we did not get a chance to register a manifestation of the axon reflex. The duration of the heating test is shorter than the time interval taken to observe the “plateau” discussed in the previous investigations. We intentionally minimized the duration of the test for the sake of convenience in the practical usage. We found that 7–8 endothelial oscillations with 70–90 s periodicity are sufficient for statistically robust results. We checked that the errors slightly decrease for the twice-longer tests, but confirm that it insignificantly improves the scope of the method. Nevertheless, this is the limitation of the study. Violation of the mechanisms of vasodilation during the local heating test has been observed in patients with types 1 and 2 diabetes (Khan et al., 2000; Sandeman et al., 1991). It is now recognized that chronic hyperglycemia plays a major role in the initiation of diabetic vascular complications. More and more data confirm the damaging effect of “glucose toxicity” on the secretory capabilities of an insular apparatus (Aronson, 2008). The location of the endothelium on the border with the blood flow makes it vulnerable to a variety of risk factors for vascular complications, including hyperglycemia (McCall, 1992). It is known that insulin resistance and the violation of NO production are closely linked (Marchesi et al., 2009; Youn et al., 2014). The manifestation of endothelial dysfunction is a violation of its vasoregulating function (Avogaro et al., 2011). Endothelial dysfunction is one of the earliest signs of vascular lesions in diabetic patients and can be detected in the
Fig. 5. Relative changes in amplitudes Ke (a), Kn (b), Km (c) for healthy subjects and for patients with type 2 diabetes.
114
S. Podtaev et al. / Microvascular Research 97 (2015) 109–114
early stages of the disease, just before the appearance of atherosclerotic plaques (Storey et al., 2001), i.e. at the stage of impaired glucose tolerance (Smirnova et al., 2013). In the case of carbohydrate metabolism violation, protein kinase C (PKC) is activated in endothelial cells. PKC increases the permeability of endotheliocytes to proteins, attenuates the endothelium-dependent relaxation of blood vessels, and enhances lipid peroxidation, whose products also inhibit the vasodilatory endothelial function (Grattagliano et al., 2008). A reduction in the production of vasodilating factors and an increase in the production of vasoconstrictors in the endothelium lead to an increased vasoconstrictor response (Widlansky et al., 2003). Nowadays, the concept of endothelial dysfunction as a key element of vascular lesions in patients with diabetes has been formulated. However, the lack of clear diagnostic criteria (biochemical, instrumental) for endothelial dysfunction allows us to continue a search for the optimal methodology for the study of various aspects of violations (Celermajer, 2008). The WAST method may be proposed as a tool to assess endothelial dysfunction before clinical manifestations appear. However, further validation is needed to be performed while blocking endothelium-dependent pathways, e.g. by using microdialysis. Conflict of interest statement The authors have no conflicts of interest to declare. Acknowledgments This work was supported by the Russian Science Foundation No. 1415-00809. References Allen, J., Howell, K., 2014. Microvascular imaging: techniques and opportunities for clinical physiological measurements. Physiol. Meas. 35, 91–141. Aronson, D., 2008. Hyperglycemia and Pathobiology of Diabetic Complications. Cardiovascular Diabetology: Clinical, Metabolic and Inflammatory Facets. In: Fisman, E.Z., Holon, A. Tenenbaum (Eds.), Tel-Hashomer, pp. 1–114. Avogaro, A., Albiero, M., Menegazzo, L., Kreutzenberg, S., Fadini, G., 2011. Endothelial dysfunction in diabetes. The role of reparatory mechanisms. Diabetes Care 34 (2), 285–290. Bandrivsky, A., Bernjak, A., McClintock, P., Stefanovska, A., 2004. Wavelet phase coherence analysis: application to skin temperature and blood flow. Cardiovasc. Eng. 4, 89–93. Braverman, I.M., Keh, A., Goldminz, D., 1990. Correlation of laser Doppler wave patterns with underlying microvascular anatomy. J. Investig. Dermatol. 95, 283. Burton, A.C., Taylor, R.M., 1940. A study of the adjustment of peripheral vascular tone to the requirements of the regulation of body temperature. Am. J. Physiol. 129, 566–577. Celermajer, D.S., 2008. Reliable endothelial function testing: at our fingertips? Circulation 117 (19), 2428–2430. Endemann, D.H., Schiffrin, E.L., 2004. Endothelial dysfunction. J. Am. Soc. Nephrol. 15, 1983–1992. Grattagliano, I., Palmieri, V.O., Portincasa, P., Moschetta, A., Palasciano, G., 2008. Oxidative stress-induced risk factors associated with the metabolic syndrome: a unifying hypothesis. J. Nutr. Biochem. 19, 491–504. Green, D.J., Maiorana, A.J., Siong, J.H., Burke, V., Erickson, M., Minson, C.T., Bilsborough, W., O'Driscoll, G., 2006. Impaired skin blood flow response to environmental heating in chronic heart failure. Eur. Heart J. 27, 338–343. Hannemann, M.M., Liddell, W.G., Shore, A.C., Clark, P.M., Tooke, J.E., 2001. Vascular function in women with previous gestational diabetes mellitus. J. Vasc. Res. 39, 311–331. Holowatz, L.A., Thompson-Torgerson, C.S., Kenney, W.L., 2008. The human cutaneous circulation as a model of generalized microvascular function. J. Appl. Physiol. 105, 370–372.
IJzerman, R.G., de Jongh, R.T., Beijk, M.A.M., van Weissenbruch, M.M., Delemarre-van de Waal, H.A., Serné, E.H., Stehouwer, C.D.A., 2003. Individuals at increased coronary heart disease risk are characterized by an impaired microvascular function in skin. Eur. J. Clin. Investig. 33, 536–542. Jaap, A.J., Hammersley, M.S., Shore, A.C., Tooke, J.E., 1994. Reduced microvascular hyperaemia in subjects at risk of developing type 2 (non-insulin-dependent) diabetes mellitus. Diabetologia 37, 214–216. Johnson, J.M., Kellog, D.L.J., 2010. Local thermal control of the human cutaneous circulation. J. Appl. Physiol. 109, 1229–1238. Khan, F., Elhadd, T.A., Greene, S.A., Belch, J.J., 2000. Impaired skin microvascular function in children, adolescents, and young adults with type 1 diabetes. Diabetes Care 23, 215–220. Kvandal, P., Stefanovska, A., Veber, M., Kvermmo, H., Kirkeboen, K., 2003. Regulation of human cutaneous circulation evaluated by laser Doppler flowmetry, iontophoresis, and spectral analysis: importance of nitric oxide and prostaglandines. Microvasc. Res. 65, 160–171. Lekakis, J., Abraham, P., Balbarini, A., Blann, A., Boulanger, C.M., et al., 2011. Methods for evaluating endothelial function: a position statement from the European Society of Cardiology Working Group on Peripheral Circulation. Eur. J. Cardiovasc. Prev. Rehabil. 18 (6), 775–789. Mabuchi, K., Chinzei, T., Nasu, Y., Yonezawa, T., Fujimasa, I., Atsumi, K., 1989. Frequency analysis of skin temperature and its application for clinical diagnosis. Biomed. Thermol. 9, 30–33. Marchesi, C., Ebrahimian, T., Angulo, O., Paradis, P., Schiffrin, E.L., 2009. Endothelial nitric oxide synthase uncoupling and perivascular adipose oxidative stress and inflammation contribute to vascular dysfunction in a rodent model of metabolic syndrome. Hypertension 54, 1384–1392. McCall, A.L., 1992. The impact of diabetes on the CNS. Diabetes 41 (5), 557–570. Minson, C.T., Berry, L.T., Joyner, M.J., 2001. Nitric oxide and neurally mediated regulation of skin blood flow during local heating. J. Appl. Physiol. 91 (4), 1619–1626. Podtaev, S., Morozov, M., Frick, P., 2008. Wavelet-based correlations of skin temperature and blood flow oscillations. Cardiovasc. Eng. 8, 185–189. Rossi, M., Carpi, A., 2004. Skin microcirculation in peripheral arterial obliterative disease. Biomed. Pharmacother. 58, 427–431. Rossi, M., Carpib, A., Galettaa, F., Franzonia, F., Santoroa, G., 2008. Skin vasomotion investigation: a useful tool for clinical evaluation of microvascular endothelial function? Biomed. Pharmacother. 62, 541–545. Roustit, M., Blaise, S., Millet, C., Cracowski, J.L., 2010. Reproducibility and methodological issues of skin post-occlusive and thermal hyperemia assessed by single-point laser Doppler flowmetry. Microvasc. Res. 79, 102–108. Roustit, M., Cracowski, J., 2013. Assessment of endothelial and neurovascular function in human skin microcirculation. Trends Pharmacol. Sci. 7, 34. Sandeman, D.D., Pym, C.A., Green, E.M., Seamark, C., Shore, A.C., Tooke, J.E., 1991. Microvascular vasodilatation in the feet of newly diagnosed non-insulin dependent diabetic patients. BMJ 302, 1122–1123. Sheppard, L.W., Vuksanović, V., McClintock, P.V.E., Stefanovska, A., 2011. Oscillatory dynamics of vasoconstriction and vasodilation identified by time-localized phase coherence. Phys. Med. Biol. 56, 3583–3601. Shiogai, Y., Stefanovska, A., McClintock, P.V.E., 2010. Nonlinear dynamics of cardiovascular ageing. Phys. Rep. 488, 51–110. Smirnova, E., Podtaev, S., Mizeva, I., Loran, E., 2013. Assessment of endothelial dysfunction in patients with impaired glucose tolerance during a cold pressor test. Diab. Vasc. Dis. Res. 10 (6), 489–497. Shusterman, V., Anderson, K.P., Barnea, O., 1997. Spontaneous skin temperature oscillations in normal human subjects. Am. J. Physiol. 273, 1173–1181. Sokolnicki, L.A., Roberts, S.K., Wilkins, B.W., Basu, A., Charkoudian, N., 2006. Contribution of nitric oxide to cutaneous microvascular dilation in individuals with type 2 diabetes mellitus. Am. J. Physiol. Endocrinol. Metab. 292, 314–318. Stephens, D.P., Charkoudian, N., Benevento, J.M., Johnson, J.M., Saumet, J.L., 2001. The influence of topical capsaicin on the local thermal control of skin blood flow in humans. Am. J. Physiol. Regul. Integr. Comp. Physiol. 281 (3), R894–R901 (Sep). Stewart, J., Kohen, A., Brouder, D., Rahim, F., Adler, S., Garrick, R., Goligorsky, M.S., 2004. Noninvasive interrogation of microvasculature for signs of endothelial dysfunction in patients with chronic renal failure. Am. J. Physiol. Heart Circ. Physiol. 287, 2687–2696. Storey, A.M., Perry, C.J., Petrie, J.R., 2001. Endothelial dysfunction in type 2 diabetes. Br. J. Diabetes Vasc. Dis. 1, 22–27. Widlansky, M.E., Gokce, N., Keaney, J.F., Joseph, A., 2003. The clinical implications of endothelial dysfunction. Vita J. Am. Coll. Cardiol. 7, 42. Youn, J.Y., Siu, K.L., Lob, H.E., Itani, H., Harrison, D.G., Cai, H., 2014. Role of vascular oxidative stress in obesity and metabolic syndrome. Diabetes 63 (7), 2344–2355.